Chem. Rev.

2010, 110, 1213–1232 1213

Spore Photoproduct: A Key to Bacterial Eternal Life

Céline Desnous,† Dominique Guillaume,‡ and Pascale Clivio*,‡

ICSN, UPR CNRS 2301, 1 Avenue de la Terrasse, 91198 Gif sur Yvette Cedex, France and UMR CNRS 6229,
51 Rue Cognacq Jay, 51096 Reims Cedex, France

Received June 1, 2007

Contents 7. Chemical Synthesis of SP and Derivatives 1225

7.1. Synthesis of SP 1226
1. Introduction 1213 7.2. Synthesis of SPSIDE and SPTIDE 1226
2. Sporulation and Its Implications 1215 8. DNA Photoproduct Repair 1227
3. Major UV-Induced DNA Photoproducts 1215 8.1. Conventional DNA Photoproduct Repair 1227
3.1. Direct UV Radiation Absorption by DNA 1215 8.1.1. Nucleotide Excision Repair (NER) 1228
Nucleobases
8.1.2. Base Excision Repair (BER) 1228
3.1.1. Cycloaddition Reactions 1215
8.1.3. Photolyase-Induced Repair 1228
3.1.2. Photohydration Reactions 1216
8.1.4. Repair of DNA Strand Breaks 1228
3.2. UV-A-Induced Photosensitization 1217
8.1.5. Others 1228
3.2.1. Energy Transfer from a Photosensitizer 1217
8.2. SPDNA Repair 1229
3.2.2. Type I Photosensitization 1217
9. Conclusions 1230
3.2.3. Type II Photosensitization 1217
10. Abbreviations 1230
3.3. DNA Damage Caused by As Yet Unidentified 1218
Mechanisms 11. Acknowledgments 1230
3.3.1. Guanine Oxidation 1218 12. References 1230
3.3.2. Strand Break Formation 1218
4. UV-C-Induced DNA Photoproduct Formation in 1218
1. Introduction
Bacteria The cells of all living organisms are continuously
4.1. In Bacterial Vegetative Cells 1218 exposed to a large variety of harmful agents that can
4.2. In Bacterial Spores 1218 induce either temporary or permanent deleterious effects.
4.3. Stereochemical Consideration of SPDNA 1219 In response to these genotoxic stresses, only those cells
Formation endowed with efficient defense mechanisms have been
4.4. Formation Mechanism of SPDNA 1219 able to overcome life’s challenges and evolutionary
5. Factors Influencing SPDNA Formation in Spores 1220 demands and survive.
5.1. Small Acid-Soluble Spore Proteins 1221 By directly targeting DNA nucleobases, solar UV
5.2. Dipicolinic Acid 1222 radiation is a ubiquitous and particularly potent physical
5.3. Pressure/Hydration Level 1222 agent capable of altering the genome integrity. The most
5.4. Temperature 1222 deleterious UV wavelengths are those in the 200-280 nm
5.5. Conclusions 1222 range (UV-C). Fortunately for humanity, although UV-C
permeates space, it does not in fact reach the surface of
6. Artificial Production of SPDNA, SPTIDE, and SPSIDE 1223
the earth thanks to the presence of ozone and other
6.1. UV Irradiation of Bacterial Vegetative Cells 1223 protective layers high in the atmosphere. Nevertheless,
6.2. UV Irradiation of Isolated DNA 1223 there are important industrial uses for artificially produced
6.2.1. Isolated DNA 1223 UV-C, mostly exploiting its germicidal properties.
6.2.2. Isolated DNA Complexed with R/β-Type 1224 Though UV radiation-induced DNA damage (UV dam-
SASP age or photodamage) has had an indisputable evolutionary
6.2.3. Isolated DNA in the Presence of Ca-DPA 1224 function, from a conservative standpoint the only cells
6.2.4. Isolated DNA in the Presence of R/β-Type 1225 able to transmit their genetic material unaltered are those
SASP and Ca-DPA capable of minimizing the frequency of the photodamage
6.3. UV-C Irradiation of Oligodeoxynucleotides 1225 and/or those possessing accurate DNA repair pathways.
6.4. Irradiation of Thymidine 1225 The frequency of photodamage occurrence can be attenu-
6.4.1. By Direct 254 nm Irradiation 1225 ated by extranuclear constituents. In addition, inaccurate
6.4.2. By Photosensitization 1225 damage repair can result from two distinct phenomena:
6.4.3. By γ-Irradiation, Heavy Ion, or Electron 1225 either overwhelming photodamage or a deficient DNA
Bombardment damage repair system. Both these phenomena can be
induced by the long-term UV exposure of DNA. Once
* To whom correspondence should be addressed. Phone: (33) 326 918 027. the cell chromosomal material has been UV damaged, the
Fax: (33) 326 918 029. E-mail: pascale.clivio@univ-reims.fr.
†
ICSN, UPR CNRS 2301. cell cycle often undergoes dramatic modification. In
‡
UMR CNRS 6229. humans, unrepaired photodamage can lead to mutation and
10.1021/cr0781972  2010 American Chemical Society
Published on Web 11/05/2009
1214 Chemical Reviews, 2010, Vol. 110, No. 3
Céline Desnous, born in Toulouse, France (1979), graduated in 2002
from the Ecole Nationale Supérieure de Chimie de Paris. In 2005, she
received her Ph.D. degree in Organic Chemistry from the University of
Orsay (Paris XI), working under the supervision of Dr. P. Clivio on spore
photoproduct chemistry. She is now working as a research engineer for
the world’s leading manufacturer of technical and creation papers
ArjoWiggins, in one of its research centers based in Apprieu, France.
Dominique Guillaume was born in France in 1959. He obtained his PharmD
degree in 1983 and Ph.D. degree in 1986 from Université de Reims
Champagne-Ardenne (URCA). He spent 1986-1988 as a postdoctoral
fellow in the laboratory of Daniel H. Rich at UW-Madison and became an
assistant professor in 1988 at the school of pharmacy, Université Paris
V, France. He was a visiting faculty member in 1997 at Washington
University in Saint Louis. In 1999, he became Full Professor of Medicinal
Chemistry and moved to URCA in 2005, where his research field includes
peptide and nucleic acid chemistry. He has authored, or coauthored, more
than 80 publications.
cancer; premature cell apoptosis is another frequent
outcome. In the vegetal kingdom, another frequently
observed consequence is growth reduction. The equilib-
rium of fragile terrestrial ecosystems can be strongly
affected in the long term. Similarly, UV damage affecting
fungi, bacteria, or phytoplankton can lead to changes in
soil microbial communities, the biomass, and in aquatic
ecosystems.
Bacterial spores represent a curious and fascinating
lifeform. Their resistance to UV-C radiation is much
higher than that of their corresponding vegetative forms.1
In addition, bacterial spores conserve their capacity for
germination for an extremely long time.2,3 Hence, concerns
regarding the efficiency of UV-C-assisted sterilization4,5 and,
consequently, the actual usefulness of UV-C in controlling
the biological safety6 of food and the subsequent minimiza-
tion of the bioterrorism threat7 have recently been addressed.
Desnous et al.
Pascale Clivio studied Pharmaceutical Sciences at the University of
Bourgogne in Dijon between 1980 and 1985. In 1989, she obtained her
Ph.D. degree in Natural Product Chemistry at the University of Reims
Champagne Ardenne (URCA) under the supervision of Professor Monique
Zeches. In 1989, she got a permanent research position at CNRS (ICSN,
Gif sur Yvette) to work with Dr. Jean-Louis Fourrey in the field of nucleic
acids chemistry and photochemistry. In 1997, she received a one-year
fellowship from the International Agency for Research on Cancer to work
with Professor John-Stephen Taylor at the Washington University in Saint
Louis, Missouri. She recently moved to the FRE CNRS 2715, URCA.
Her current research interests include the study of conformational impact
on DNA photochemistry, the chemistry of DNA repair, and the synthesis
of DNA photoproducts for biological studies. Another area of research
interest is the synthesis of fat nucleosides.
The potentially elevated risk of bacterial planetary contami-
nation through space exploration is also currently of great
interest.8,9
Spores also have to face UV-A and -B radiation in order
to survive. The deleterious effect of UV on spores is a
consequence of either direct effects on their DNA
components or the results of the formation of reactive
oxygen species (ROS). Bacterial spore resistance to UV-A
and -B is attributed mainly to the presence of melanin-
like absorbing pigments in the spore outer layers. Bacterial
spore resistance to UV-C radiation has been associated
with the unique photochemical behavior of its DNA. In
vegetative bacterial cells, UV-C radiation induces DNA
damage principally through pyrimidine dimerization,
which produces cyclobutane pyrimidine dimers (CPDs)
and (6-4) photoproducts ((6-4) PPs). In contrast, in
bacterial spores a single and totally different photoproduct,
5-(R-thyminyl)-5,6-dihydrothymine, is produced by UV-C
irradiation. During the 5 years between the initial isolation
of this photoproduct and its structural elucidation, it was
simply called the “spore photoproduct” or “SP”, and this
name is still commonly used. Damage can be repaired
during germination through the action of a specific enzyme
named SP lyase. Formation of the spore photoproduct and
its subsequent ability to undergo such efficient repair
explains the low vulnerability of bacterial spores to UV-C
exposure.
Particularly during the last 15 years,10-17 numerous reviews
have covered various aspects of the UV resistance of spores.
Only the spore photoproduct itself has not been the subject
of its own review. This is particularly regrettable in light of
the pivotal involvement of this photoproduct in the strong
UV-C resistance of bacterial spores and the importance of
its chemistry in helping to decipher DNA-specific behaviors.
Hence, this review will focus on the chemistry of the spore
photoproduct from its formation to its repair and including
its chemical synthesis.
Spore Photoproduct
2. Sporulation and Its Implications
In response to environmental conditions such as nutrient deple-
tion, some microorganisms initiate a unique survival process called
sporulation.18-20 This process begins with the formation of a
predivisional cell whose gene expression is governed by the Spo0A
protein. This is followed by an asymmetric division. As a result, a
large mother cell and smaller cell called a forespore are formed.
Each has its own individual but interconnected gene expression
program. The forespore, which will become the core of the mature
spore, acquires a second membrane through engulfment by the
mother cell. The space between the two membranes is later filled
in by cell wall material to give the spore cortex. Finally, an outer
shell of protein (the coat) is laid on the external surface of the
developing spore to provide an additional protective layer. Mean-
while, during its maturation, the forespore undergoes two additional
major transformations. Both of these are of the utmost importance
for the protection of its DNA: (1) dehydration, due in part to the
uptake of pyridine-2,6-dicarboxylic acid (dipicolinic acid, DPA)
from the mother cell, and (2) compacting of its chromosomal
material by a family of small, acid-soluble spore proteins (SASP).
In the final step, the mature spore is released into the environment
through lysis of the mother cell.
Spores represent a unique form of life since most of them possess
only a single chromosome and display no apparent metabolism, a
state defined as cryptobiotic. Spore dormancies of longer than 25
million years3 and possibly even 250 million years21,22 have
been reported. Despite their apparent lack of life some spores
play an essential ecological and environmental role.23
Not all bacteria genera possess the ability to sporulate. The
spore-forming Bacillus (and to a lesser extent Clostridium)
species are the most studied. Observations of this species have
often been generalized and extended to other spore-forming
genera. However, species with extreme UV resistance have been
recently isolated.24 This clearly supports the existence in the
spores of some bacterial species with specific biochemical
processes much more complex than previously anticipated.
Under appropriate circumstances, spore dormancy ends
and the spore germinates25 into a vegetative cell whose
genome, despite the possible accumulation of DNA UV
damage during dormancy, is preserved to an amazing extent
relative to the genome of the mother cell.
3. Major UV-Induced DNA Photoproducts
UV radiation is arbitrarily divided in three wavelength
domains: UV-C encompasses the 220-280 nm, UV-B the
280-320 nm, and UV-A the 320-380 nm range. DNA
absorbs mainly in the UV-C and to a smaller extent in the
UV-B range. It barely absorbs in the UV-A domain.
Consequently, only UV-B and -C have any direct effect on
DNA. However, UV-A radiation is capable of indirectly
damaging DNA through the participation of a photosensitizer.
In this review, we will focus only on DNA photoproducts
formed by low-intensity continuous UV irradiation that
mostly involves the first excited states of the nucleobases.
This is in contrast to the high-intensity radiation delivered
by lasers which produces nucleobase radical cations. Numer-
ous reviews have already covered the photochemistry of
DNA.26-34 In this section, we will only discuss the major
classes of DNA photoproducts. Further references can be
found in the reviews mentioned above.
Chemical Reviews, 2010, Vol. 110, No. 3 1215
3.1. Direct UV Radiation Absorption by DNA
Nucleobases
Both UV-C and -B photon absorption give rise essentially
to cycloaddition and photohydration reactions. Pyrimidine
nucleobases (thymine and cytosine) are the most sensitive
to UV-B and -C radiation.
3.1.1. Cycloaddition Reactions
At the dipyrimidine sites (1-4) in DNA, cyclobutane
pyrimidine dimers (CPDs) and (6-4) pyrimidine pyrimidone
photoproducts ((6-4) PPs) are principally formed by direct
photon absorption (Figure 1).
CPDs are the most abundant dipyrimidine photoproducts.
They result from a [2 + 2] cycloaddition reaction between
the C5-C6 double bonds of two adjacent pyrimidines
(Figure 1). Depending on the nature of the C4 (O or NH)
and C5 (CH3 or H) pyrimidine substituent, four different PPs
(5-8) are formed. These PPs consequently can have a cis-
syn stereochemistry since pyrimidines in DNA adopt an anti
glycosidic bond conformation. Cytosine-derived CPDs (6-8)
are unstable and spontaneously deaminate at their C4
position, thus leading to secondary photoproducts belonging
to the uracil-derived CPD family. Under UV-C radiation,
CPDs can revert to their parent nucleobases.
Figure 1. UV-induced c,s CPDs in DNA.
(6-4) Pyrimidine pyrimidone photoproducts are the second most
abundant PPs formed in DNA. They derive from a sequence-
specific Paternó-Büchi reaction between the C5-C6 double bond
of the pyrimidine on the 5′ side and the exocyclic double bond at
the C4 atom of the pyrimidine on the 3′ side, i.e., the carbonyl of
a thymine residue or the imine tautomeric form of a cytosine
residue (Figure 2). The resulting four-membered ring intermediate
(oxetane or azetidine) is unstable and undergoes a ring-opening
reaction leading to (6-4) PPs (9-12) (Figure 2). Formation of
the four-membered ring intermediate was ascertained using the
photochemistry of thio-nucleobases that lead to a more stable
thietane intermediate.35 (6-4) PPs 11 and 12 involving a
cytosine at their 5′ end can spontaneously deaminate, leading
to secondary photoproducts.
After UV-A or -B absorption, (6-4) PPs can be converted
into their corresponding Dewar (Dwr) valence isomer (13-16)
(Figure 3). This latter adduct can revert to its (6-4) isomer
under UV-C radiation.
Under prolonged UV-C exposure, the (6-4) pyrimidine
pyrimidone motif can undergo a ring contraction reaction,
leading to a 2-imidazolone (5-4) pyrimidone derivative.36
1216 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 2. UV-induced (6-4) PPs (9-12) in DNA.
Figure 3. UV-induced Dewar PPs in DNA.
To a minor extent, purine nucleobases are also photoreactive.
However, only adenine has been shown to give rise to cycloaddition
reactions with adjacent nucleobases in DNA. Under UV-C irradia-
tion the unstable azetidine 18 is generated at diadenine sites (17)
by a cycloaddition reaction between the N7-C8 double bond of
the 5′-end residue and the C5-C6 double bond of the 3′-end
adenine residue (Figure 4). Azetidine 18 can then either spontane-
ously rearrange to afford 19 or undergo a hydration reaction,
ultimately leading to 20 (Figure 4).
Figure 4. Formation of UV-induced AA PPs in DNA.
Desnous et al.
Sequence-specific photoaddition of adenine residues with
vicinyl thymine residues can also occur. Thus far only the
ring-expanded derivative 23 produced from the TA site 21
has been isolated. Derivative 23 results from ring opening
of the cyclobutane adduct 22 generated by UV-C-induced
cycloaddition between the C5-C6 double bond of the 5′-
end thymine residue and the C5-C6 double bond of the 3′-
end adenine residue (Figure 5). The X-ray structure of the
TA adduct at the dinucleotide level is now available.37
3.1.2. Photohydration Reactions
Photohydration reactions usually involve pyrimidine bases.
Hydration results in the saturation of the C5-C6 double bond
and in the nonstereoselective substitution of the C6 atom by a
hydroxyl group. Cytosine residues are the most prone to this
kind of photoreaction. The resulting unstable cytosine hydrate
derivative 24 can either revert through dehydration to the starting
compound or deaminate to afford the uracil hydrate residue 25.
The latter dehydrates much less readily than 24 (Figure 6).38,39
The UV-C-induced hydration reaction of thymine and purine
residues in DNA leading, respectively, to 5-hydroxy-5,6-
dihydrothymine and 4,6-diamino-5-formamidopyrimidine or
2,6-diamino-4-hydroxy-5-formamidopyrimidine derivatives has
been reported.40 However, these results have been contested.32
Spore Photoproduct
Figure 5. Formation of UV-induced TA photoproduct 23 in DNA.
Figure 6. Hydrolytic deamination of cytosine hydrate derivative
24 in DNA.
In summary, pyrimidine as well as purine nucleobases are
susceptible to UV radiation, and consequently, several types
of photoproducts can be formed through radiation exposure
of isolated DNA. However, in cellular DNA, CPD and (6-4)
PPs are by far the most frequently produced adducts.
Photohydrates, on the other hand, are generated in much
smaller quantities. The formation of purine photoproducts
in cellular DNA is yet to be confirmed.
3.2. UV-A-Induced Photosensitization
UV-A damage of DNA is an indirect effect arising from the
action of so-far unidentified photosensitizers in cells. Three general
pathways can lead to the formation of DNA photoproducts.
3.2.1. Energy Transfer from a Photosensitizer
Once excited by the UV-A radiation, the photosensitizer can
directly transfer its energy to DNA, most likely via a triplet energy
transfer mechanism. This type of reaction affects mainly dithymine
sites producing as the major photoproduct c,s T[CPD]T (5). The
cytosine-derived photoproducts c,s T[CPD]C (6) and c,s C[CPD]T
(7) are generated in lower yields. It should be noted that this
excitation process does not afford (6-4) PPs since they result from
singlet-state excitation.
Figure 7. UV-A-induced guanine-oxidized PPs in DNA by type I photosensitization.
Chemical Reviews, 2010, Vol. 110, No. 3 1217
3.2.2. Type I Photosensitization
One-electron transfer from DNA nucleobases to the excited
photosensitizer (Type-I photosensitization) can lead to nucleo-
base radical cations that can subsequently undergo either
hydration or deprotonation. In DNA, guanine is the nucleobase
most prone to one-electron oxidation. This can be directly
attributed to the fact that guanine has the lowest ionization
potential or indirectly because transfer through the DNA of a
positive charge originating from another nucleobase radical
cation can occur. Consequently, adenine and pyrimidine nucleo-
bases are less susceptible to this photochemical process.
Nevertheless, adenine and pyrimidine nucleosides can be
targeted by a one-electron oxidation process.
3.2.2.1. Oxidized Purines. Hydration of the guanine
radical cation 26 leads to the formation of the 8-hydroxy-
7-yl-guanine oxidizing radical 27. Upon reduction this in turn
yields 2,6-diamino-4-hydroxy-5-formamidopyrimidine (Fapy-
Gua, 28) in DNA (Figure 7). Oxidation of 27 gives the 8-oxo-
7,8-dihydroguanine derivative (8-oxoGua, 29). Deprotonation
of 26 leads to an oxazolone derivative 30 through a complex
cascade of reactions.
3.2.2.2. Oxidized Pyrimidines. In DNA, the one-electron
oxidation of pyrimidines is a minor pathway that ultimately
leads to the formation of 5-(hydroxymethyl)uracil, 5-formyl-
uracil, barbiturate, as well as hydantoin residues.
3.2.3. Type II Photosensitization
Finally, the excited photosensitizer can also generate singlet
oxygen (type II photosensitization) that further reacts with
nucleobases in the DNA to give oxidized products. Only
guanine residues have been shown to be susceptible to this
mechanism, producing the 8-oxoGua derivative (29) (Figure
8).
1218 Chemical Reviews, 2010, Vol. 110, No. 3 Desnous et al.

Figure 8. UV-A-induced 8-oxoGua formation in DNA by type II photosensitization.

3.3. DNA Damage Caused by As Yet Unidentified
Mechanisms
3.3.1. Guanine Oxidation
Exposure of DNA to UV-B or -C radiation also results in
the formation of some of the 8-oxoGua derivative 29. The
mechanism of its formation could involve either the generation
of ROS or a one-electron oxidation. This kind of photoinduced
oxidation reaction is, however, a minor process compared to
the formation of CPDs in these wavelength domains.
3.3.2. Strand Break Formation
DNA strand breakage (single-strand breaks (SSBs) and
double-strand breaks (DSBs)) is also observed after exposure
to UV-A, -B, and -C radiation,28,41-43 although it is a minor
process compared to the formation of CPDs. Possibly, UV-A
and -B irradiation triggers a complex oxidative process involving
the production of the highly reactive hydroxyl radical by the
mitochondria and type I photosensitization which ultimately
leads to DNA strand breakage.34,44 Strand breaks following
UV-C irradiation would originate from a different mechanism
involving the excited sugar phosphate backbone.45
4. UV-C-Induced DNA Photoproduct Formation in
Bacteria
4.1. In Bacterial Vegetative Cells
In bacterial vegetative cell DNA, CPDs and (6-4) PPs
are the major lesions formed as a result of UV-C irradiation
(Figure 1).26,27,46
Not all of these PPs are generated with the same efficiency:
the major adducts formed are the c,s T[CPD]T 5, T[6-4]C
10, and c,s T[CPD]C 6 and/or the c,s C[CPD]T 7.27,46 In B.
subtilis vegetative cells, for a UV dose of 0.1 J/cm2, c,s
T[CPD]T 5 formation involves 5% of the total amount of
thymine residues.11
residues from different DNA strands represents less than 1%
of the intrastrand SP for a UV dose of 1 J/cm2.48
In the DNA of B. subtilis spores, for a UV-C dose of 0.1
J/cm2, spore photoproduct formation involves 7.5% of the
total amount of thymine residues while c,s T[CPD]T (5)
represents less than 0.2%.11 Interestingly, when spores of B.
subtilis are exposed to both UV-B and -A radiation, spore
photoproduct formation still remains the major event even
though production is much lower (a 103- and 106-fold
decrease, respectively) compared to 254 nm irradiation.49,51
Strictly speaking, the term SP refers exclusively to the
dipyrimidine nucleobase structure 31. However, the term is
frequently used rather loosely to refer to diastereomeric
mixtures in which the N1 atom of each nucleobase is
substituted. In this review, the acronym SP will be used
exclusively for the nucleobase dimer structure 31. Acronyms
SPSIDE and SPTIDE will be used for the dinucleoside analogue
(32) and dinucleotide analogues (33, 34) of SP, respectively.
The term ias-SPDNA will be used for damaged oligode-
oxynucleotides (ODNs) or DNA 35 and 36 containing at least
one intrastrand SP-type lesion, and the term irs-SPDNA will
describe damaged ODNs or DNA 37 and 38 containing at
least one interstrand SP-type lesion. If the intra- or internature
of the spore photoproduct in DNA is not known, the acronym
SPDNA will simply be used.

4.2. In Bacterial Spores

After UV-C irradiation only very small amounts of CPDs,
(6-4) PPs, and single- and double-strand breaks are produced
in spore DNA. Some indirect evidence derived from the
slight UV sensitivity of nonhomologous end joining repair
mutants has recently confirmed the minor formation of strand
breaks induced in spore DNA.47 For a UV dose of both 148
and 1.6 J/cm2,49 the cumulative effect of CPDs, (6-4) PPs,
and single- and double-strand breaks represents less than 1%
of all the DNA damage in the spores. Since this UV dose
range far exceeds the minimum dose required to kill 90%
of spores, the physiological effects of CPDs, (6-4) PPs, and
single- and double-strand breaks are considered negligible.49
The PP formed in spore DNA is almost exclusively the 5-(R-
thyminyl)-5,6-dihydrothymine (or spore photoproduct (SP)).50
It arises almost entirely from two adjacent thymidine residues
on the same DNA strand. SP formation between thymidine
Spore Photoproduct
4.3. Stereochemical Consideration of SPDNA
Formation
The chiral compound SP (31) was isolated in 1965
following the acid hydrolysis of SPDNA.50 Since the hydrolytic
step resulted in the loss of the optically pure sugar moiety,
no mention of the enantiomeric purity of the SP-5,6-
dihydrothymine moiety C5 position was given.52
Four isomers of ias-SPDNA (35 and 36, Figure 9) can be
expected within the DNA strand depending on (1) the
glycosidic bond conformation of the parent thymines at the
time of the C5-CH2 bond formation and (2) the 5′-end or
3′-end location of the thymine residue to be saturated.
Theoretically, each ias-SPDNA can result from two sets of
adjacent thymidine conformers (Figure 9). The configuration
of the 5,6-dihydrothymine residue C-5 atom is dictated by
the syn/anti conformation of the glycosidic bond of its
thymidine precursor, while the 5-R-thyminyl moiety can
result from a syn or anti glycosidic bond conformation of
the corresponding thymidine.
The formation of ias-SPDNA within the double helix is
likely to be regio- and stereospecific since SPTIDE is observed
as a single peak in the HPLC-MS/MS chromatogram of
enzymatically digested SPDNA.53 In the A- or B-DNA double
helix the glycosidic bond conformation of the nucleobases
is in the anti domain. In the dry state, DNA adopts an A
conformation. Thus, since SPTIDE isolated from spore DNA
and from UV-C-irradiated dry DNA display indistinguishable
HPLC and mass fragmentation patterns,48 the nucleobases
Figure 9. Possible glycosidic bond conformers at dithymine sites in DNA and induced structures of ias-SPDNA isomers.
Chemical Reviews, 2010, Vol. 110, No. 3 1219
in the DNA of spores are most likely in the anti glycosidic
bond conformation domain. Therefore, only 35b and 36a
can be formed. Recent experimental results using 32 as a
model substrate for repair studies have identified the C5 S
isomer of 32 as being diastereoselectively repaired by the
SP lyase.54,55 This result has led to the suggestion of a C5 S
configuration in ias-SPDNA, and consequently, the dihy-
drothymine moiety location should be at the 3′ end of the
dinucleotide motif as in 36a (Figure 9). However, extensive
NMR studies carried out on SPTIDE have recently demon-
strated that the dihydrothymine moiety is located at the 5′
end of the dinucleotide motif and that its C5 configuration
is R.56 Therefore, 35b is the correct structure for ias-SPDNA.
Repair experiments consistent with this result have also been
reported.57 Thus, ias-SPDNA results from bond formation
between the methyl group of the 3′-end thymine and the C5
atom of the 5′-end thymine residues.
4.4. Formation Mechanism of SPDNA
Two mechanisms are currently envisioned to explain
SPDNA formation after UV-C irradiation. The methylene link
formation could follow a reaction between two ‘close in
space’ but independently UV-generated thymine-derived
radicals: the 5-R-thyminyl radical (37) and the 5,6-dihy-
drothymin-5-yl radical (38) (Scheme 1, path A).52 Indeed,
evidence has been given for the possible formation by UV
of these two different types of radicals from thymine.58-60
Nevertheless, the simultaneous formation of these two
1220 Chemical Reviews, 2010, Vol. 110, No. 3 Desnous et al.

Scheme 1 of PPs per 104 bases as a function of the UV dose. It is

important to note that the calculated values have no
experimental significance when the dose unit lies outside the
range of the linear part of the curve. Clearly, a comparison
between the two methods is only possible if the sequence of
the irradiated DNA is known and if the quantitative results
expressed as percent of thymine have been calculated using
the linear part of the curve reporting the number of PPs per
104 bases as a function of the UV dose. There are several
limitations that need to be considered when comparing
quantitative data since they depend on both the DNA
concentration and the specific experimental setup used for
irradiation. In addition, some early studies reported the
amount of PPs formed for a single UV dose. However, since
damage formation is not necessarily a linear function of
the UV dose, this isolated information cannot be used to
calculate the normalized amount of PPs formed per UV dose
unit. Consequently, comparisons between quantitative studies
different types of radicals at dithymine sites appears to be are generally less than meaningful.
an event of low probability. In addition, neither the oxidative Irradiation of spores with the full spectrum of sunlight at
products derived from 37 and 38 or the dimerized adducts the Earth’s surface leads to the formation of SPDNA.51
derived from two identical radicals have ever been isolated. However, quantitatively SPDNA formation is much higher on
The second hypothesis was proposed after analyzing UV-C radiation: radiation at 254 nm is 103 times more
photolysis experiments using methyl-deuterated thymidine. efficient for SPDNA induction than that at 313 nm (UV-B)
These experiments suggested a concerted mechanism involv- and 106 times more efficient than that at 365 nm (UV-A).51
ing the methyl group of one thymidine and the double bond Maximum SPDNA formation is reached near 260 nm.65
of the second thymidine (Scheme 1, path B).26 Because 254 nm monochromatic UV radiation accurately
However, neither of these two hypotheses explains the reproduces the effects of UV-C on DNA, the 254 nm
respective role or difference in photoreactivities of the wavelength emitted by low-pressure mercury lamps, also
thymine residues (at either 5′ or 3′) in ias-SPDNA formation. called germicidal lamps, is the one most commonly used in
Nor is the influence of the pyrimidine C4 substituent fully laboratory studies.
explained, since no cytosine-derived SP-like PPs have ever Although in an early study no qualitative difference was
been isolated either in vivo or in vitro even though in the observed in PP formation in the irradiated DNA spores, either
latter case a large range of conditions has been studied. dry or in an aqueous suspension,65 large variations in the
Excited thymine residues can also be generated by quantitative yield of PPs have been recently reported (Table
photosensitization. Benzophenone and pyridopsoralens61,62 1).63 When wild-type spores are irradiated in the dry state,
as well as the physiologically relevant calcium dipicolinate55,63,64 PP formation appears to be dramatically lowered (15-fold)
have all been identified as photosensitizers capable of compared to irradiation in an aqueous suspension. Even
performing this function. though, as previously mentioned, experimental changes
inherent in the different irradiation procedures may explain
5. Factors Influencing SPDNA Formation in Spores Table 1. Quantitative (Quant) and Qualitative (Qual) Yields of
Any comprehensive description of SPDNA formation must Intrastrand Bipyrimidine Photoproducts in Spores and Mutants
Exposed to UV-C Radiation in Aqueous or Dry State
take into account its independent but complementary qualita- Conditionsa
tive and quantitative aspects.
From a qualitative standpoint, only the ratio (expressed ias-SPDNA 35/36 c,s T[CPD]T 5 T[6-4]C 10
in percent) of the total number of spore photoproducts vs Quant Qual Quant Qual Quant Qual total refs
the total number of PPs is considered. Therefore, the quali- sporeb
tative aspect represents only the ‘formation ability’ of SPDNA wild type
compared to the other PPs. wet 365 99.7 0.3 0.08 ND 366.2 48
Conversely, the quantitative aspect represents the ability 252 99.6 0.69 0.27 0.10 0.04 253.1 63
of the thymine residues to give rise to SPDNA. To quantify dry 15.9 98.4 0.14 0.86 0.07 16.16 63
SPDNA formation, two methods are used. The first method R/β-type SASP Q
requires the use of DNA which has either 3H- or 14C-labeled wet 176 64.1 67 24.4 6 2.2 274.4 48
108 37.3 78.4 27 71.2 24.6 289.8 63
thymine residues. After acid hydrolysis of SPDNA the results dry 4.3 35.8 3.2 26.7 3.0 25 12.0 63
are expressed as the ratio (%) of the released SP versus the
Ca-DPA Q
total number of thymine residues initially present in the DNA wet 23.3 71.2 5.3 16.2 0.82 2.5 32.7 63
strand. The second method uses HPLC-mass spectrometry dry 1.6 42.1 1.4 36.8 0.13 3.4 3.8 63
and standard solutions. It measures the amount of SPTIDE (for R/β-type SASP Q
ias-SPDNA) and/or SPSIDE (for irs-SPDNA) released after
enzymatic digestion of the SPDNA obtained for a particular Ca-DPA Q
wet 4.0 1.9 78.0 37.6 82.1 39.6 207.4 63
UV dose range. This amount is then normalized to 104 bases, dry 0.31 8.9 1.2 34.3 1.3 37.1 3.5 63
and the final quantification, expressed as the number of PP a
per 104 bases per UV dose unit (J/cm2), can be calculated Expressed in lesions per 104 bases per J/cm2. b Terms indexed Q
refer to the corresponding deficient spore.
using the initial linear part of the curve, giving the number
Spore Photoproduct
these observed quantitative yield variations, it is nonetheless
clear that the irradiation state has little or no influence on
the respective distribution of PPs since both in the solid state
and in suspension SPDNA remains the main PP generated in
spore DNA.
Since the aqueous environment has only a weak influence
on SPDNA formation, the impact of several other factors has
been studied. Thus far, the most important factor is believed
to be the association of spore DNA with specific spore
proteins,namely,‘smallacid-solublesporeproteins’(SASP).19,66
5.1. Small Acid-Soluble Spore Proteins
SASP are a group of spore proteins of low molecular
weight (5-11 kDa, 60-75 residues). This group constitutes
up to 10% of total spore proteins of dormant spores although
it may comprise 20% of the spore core soluble proteins, a
quantity sufficient to saturate the spore DNA. SASP, which
are synthesized during sporulation, are rapidly degraded
during germination and thereby provide the amino acids
necessary for the bacterial protein synthesis.
The SASP group is composed of different types of
proteins: for B. subtilis the two principal ones are termed
R/β and γ. For other species, the corresponding SASP are
named R/β- and γ-type SASP. Multiple R/β-type SASP exist;
they all have a very similar amino acid sequence and a
molecular weight of 5-7 kDa. R/β-Type SASP are mono-
mers in solution. They are weak, nonspecific DNA-binding
proteins and when bound to DNA form a dimer.67 In the
spore this binding induces a strong modification of the UV-C
DNA photoreactivity. In addition, R/β-type SASP have a
significant effect on gene expression during sporulation and
germination.68
A single γ-type SASP (molecular weight around 8-11
kDa) exists, and it does not appear to influence the DNA
photoreactivity. Its sole function appears to be supplying the
necessary amino acids during the germination step.
In R/β-type SASP-less spore mutants, UV-C-induced
SPDNA formation is quantitatively reduced by 50% compared
to wild-type spores.69 Concomitantly, the formation of c,s
T[CPD]T 5 undergoes a 40-fold quantitative increase.69
A subsequent and more accurate determination of the
nature of the PPs formed by UV-C radiation together with
their relative distribution has provided a better view of the
impact of R/β-type SASP on spore DNA photoreactivity.48,63
Even though R/β-type SASP have no significant influence
on the global amount of PP formed (i.e., their quantitative
impact is negligible), they still dramatically drive the
photochemistry of spore DNA toward ias-SPDNA formation
(i.e., their qualitative impact is large). Indeed, in the presence
of R/β-type SASP, ias-SPDNA represents more than 99% of
all the dipyrimidine PPs formed whereas in R/β-type SASP-
less spore mutants ias-SPDNA represents only 37-64% of the
dipyrimidine PPs.48,63 Interestingly, these studies also show
that ias-SPDNA formation is not fully prevented in R/β-type
SASP-less spore mutants. Although R/β-type SASP-less
spores still contain up to 15% of R/β-type SASP, this
observation clearly indicates that, in vivo, factors other than
SASP contribute to SPDNA formation. The amount of PPs
formed in experiments using R/β-type SASP-less spores is
significantly reduced when the irradiation is performed in
the dry state.63
Linking R/β-type SASP-modified DNA photoreactivity
either to a DNA conformational change or to the consequence
of an induced dehydrated state or to both has been proposed.
Chemical Reviews, 2010, Vol. 110, No. 3 1221
However, this question is still a matter of debate, and the
precise reason for the observed modified DNA photoreac-
tivity is still far from being fully understood. In living cells,
DNA adopts a B conformation. The hypothesis of a different
DNA conformation in spores was made as early as 1965.50
The early suggestion that spore DNA could adopt an A
conformation was formulated 3 years later70 and then
supported by in vitro experiments (IR and CD).71 Conse-
quently, for several years spore DNA has been described as
being in an A conformation, and this conformational switch
has been held responsible for the specific photochemical
behavior of spore DNA. However, electron microscopy
studies have indicated that the base pair per helical turn in
R/β-type SASP-bound DNA is not significantly different
from that of vegetative cell DNA.72 This has led to the
proposal of an A-like conformation for spore DNA,72,73
although more recent cryoelectron microscopy studies sup-
port a conformation close to the B form for spore DNA.74
A detailed picture of the structural interactions occurring
between R/β-type SASP and DNA has become recently
available, nicely completing the few pieces of the puzzle
that had been previously identified. R/β-Type SASP bind
only to double-stranded (ds) DNA (or ds ODNs).75 Although
R/β-type SASP exhibit a random coil conformation in the
absence of DNA,76 upon binding to ds DNA, they become
structured and hence R helical.76 Crystallographic resolution
of a R/β-type SASP bound to a 10 bp DNA duplex has
revealed that the DNA helix adopts an A-like conformation
with the base pair planes essentially parallel to each other
and normal to the helix axis, an average value for the twist
of 31.5°, and all sugar puckering in the C3′-endo conforma-
tion. However, because binding of R/β-type SASP to DNA
widens the minor groove, the rise per base pair of SASP-
bound DNA is now identical to that for B-DNA. Therefore,
spore DNA adopts a “A-B-DNA” conformation.67 Binding
to DNA facilitates SASP dimerization, is cooperative, and
follows both local conformational changes of ds DNA around
the bound R/β-type SASP and also SASP-SASP interac-
tions76 mediated by the N-terminal amino acid residues.77
Crystallographic observations have shown that when bound
to DNA, each R/β-type SASP monomer comprises two long
helical segments, one lying on the edge of the DNA minor
groove and the other located in the minor groove and
connected by a turn region.67 The C-terminal residues are
also implicated in DNA binding but so far only through
unidentified interactions.78 Both N and C termini are devoid
of secondary structure.67 R/β-Type SASP binding to ds DNA
is nonsequence specific even though it is modulated by the
ds DNA sequence. Binding of R/β-type SASP to DNA
encompasses four75,76 to six base pairs79,67 and forms a helical
coating around the DNA that greatly increases the DNA
stiffness.72 Cryoelectron microscopy experiments have re-
vealed that these helical filaments (nucleoprotein helices) are
tightly packed in a toroidal conformation by interdigitation
of R/β-type SASP domains from adjacent helices.74 Such
assembly is stabilized by hydrophobic interactions and
induces a substantial dehydration in the immediate vicinity
of the DNA.74,67 This is believed to modify its photochem-
istry. Indeed, lack of water is known to change the DNA
reactivity, and for example, the dehydrated state of the spore
core is most likely responsible for the R/β-type SASP
protection of cytosine against hydrolytic deamination.80
The highly dehydrated state of R/β-type SASP-bound
DNA is currently believed to be the major factor responsible
1222 Chemical Reviews, 2010, Vol. 110, No. 3
for specific spore DNA photochemistry. This hypothesis is
in good agreement with the observations made for isolated
DNA81,82 (section 6.2.1). Nevertheless, the fundamental
molecular and structural effects of the absence of water on
the photoreactivity of DNA in spores remain unknown. An
unidentified conformational change induced by dehydration
or an unusual SASP-bound DNA rigidity cannot as yet be
ruled out.
5.2. Dipicolinic Acid
Because SPDNA is formed in R/β-type SASP-less mutants,
R/β-type SASP is clearly not the only factor governing SPDNA
formation in spores. Among other factors influencing SPDNA
formation is dipicolinic acid, which is present in the spore
as a Ca2+ chelate (39). Calcium dipicolinate constitutes 15%
of the dry spore weight.83 It strongly absorbs in the UV-C
at 271 and 278 nm and to a lesser extent at 263 nm.83,84
During sporulation, DPA is synthesized in the mother cell
by DPA synthetase. This enzyme is composed of the two
subunits spoVFA and spoVFB, also called DpaA and
DpaB.85 DPA is then internalized into the spore core probably
through proteins encoded by the spoVA operon and excreted
in an early stage of spore germination.86
The role of Ca-DPA as a photosensitizer in spore SPDNA
formation has been demonstrated using DPA-less spore
mutants.87 In these mutants, depending on the experimental
conditions, a 4-7-fold quantitative decrease of SPDNA
formation is observed (Table 1).63 Additionally, although ias-
SPDNA remains the major PP produced (42-71%), its
formation is less selective in these mutants than in wild-
type spores. The concomitant formation of 5 is also observed
(16-36%) (Table 1).63
The involvement of 39 as a photosensitizer in spores is in
agreement with in vitro experiments performed with
Ca-DPA.55,63,87 The participation of Ca-DPA in SPDNA
formation may involve a selective triplet-state energy transfer
from the UV-C-excited Ca-DPA to the thymine bases.63 Such
selective energy transfer could explain the exclusive involve-
ment of dithymine in SPDNA formation and consequently the
absence of spore photoproducts involving cytosine residues.48,63
A full understanding of the mechanism of the involvement
of Ca-DPA remains elusive. Full clarification would first
require a detailed examination of the excited-state properties
of Ca-DPA.
Interestingly, Ca-DPA also induces a decrease in the core
hydration state.88 This has led to the suggestion that 39 may
also act indirectly and participate with R/β-type SASP in
the highly dehydrated state of the spore core. In addition,
Ca-DPA may bind, probably through intercalation to DNA.
This could provide an additional hydration reduction mech-
anism in the immediate vicinity of the spore DNA.89 The
involvement of 39 in the conformational modification of
spore DNA has also been raised. However, for the present
at least this hypothesis is not strongly supported and the role
of Ca-DPA is considered to be chiefly as a photosensitizer.63
In summary, among the chemical factors governing the
spore DNA photoreactivity, R/β-type SASP and Ca-DPA
Desnous et al.
appear to be the most critical. Their cumulative absence in
spore mutants results in a dramatic decrease of SPDNA
formation and leads to spore DNA photochemical behavior
qualitatively and quantitatively closely resembling that of
vegetative cell DNA. When mutant spores lacking both R/β-
type SASP and Ca-DPA are irradiated in the wet state, SPDNA
is formed in only a 2% yield of the total dimeric PPs, whereas
c,s T[CPD]T (5) and T[6-4]C (10) are formed in yields of
38% and 40%, respectively (Table 1).63 The decrease in the
amount of PPs in the absence of both R/β-type SASP and
DPA is even more pronounced when spore mutants are
irradiated in the dry state (Table 1).63
5.3. Pressure/Hydration Level
Among the physical factors evaluated that are known to
influence SPDNA formation is that of pressure. This is because
of the particular concerns regarding possible interplanetary
contamination by spores.
Under the Earth’s atmospheric pressure (101.3 kPa), the
UV-C irradiation of spores leads almost exclusively to the
formation of SPDNA. Under an ultrahigh vacuum of 2 × 10-6
Pa with a 254 nm UV dose of 0.5 J/cm2, SPDNA remains the
main photoadduct (69%) but c,s T[CPD]T (5) and t,s
T[CPD]T (40) are also formed in qualitative yields of 21%
and 10%, respectively.65,90,91 Under ultrahigh vacuum, the
quantitative formation of SPDNA is reduced by ca. 30%
compared to that under atmospheric conditions.90 As ob-
served in heat denaturation experiments, the partial DNA
denaturation in the spore following the extreme dehydrated
state induced by low pressure can explain the formation of
t,s T[CPD]T 40.90,92 The photoreactivity of DNA in spores
under a medium vacuum (1-2 Pa) is similar to that observed
under ultrahigh vacuum.91
5.4. Temperature
Temperature is the other physical parameter identified as
influencing SPDNA formation at dithymine sites in spores.
Because high temperature causes DNA denaturation, only
low-temperature effects can be studied. Quantitatively, the
optimum temperature for SPDNA production is -80 °C.93 At
this temperature, SPDNA formation is twice that observed at
22 °C and four times that observed at -196 °C for a UV-C
dose of 0.02 J/cm2. The combined effects of the humidity
and temperature of the spore core on SPDNA formation have
not yet been explored.
5.5. Conclusions
Two physical (temperature and pressure) and three chemi-
cal (R/β-type SASP, water, and Ca-DPA) factors have been
clearly identified as being particularly critical for SPDNA
Spore Photoproduct Chemical Reviews, 2010, Vol. 110, No. 3 1223

Figure 10. Representation of the influence of chemical (green) and physical (pink) factors on SPDNA formation in bacterial spores.

formation. Although the exact elucidation at the molecular Table 2. Quantitative (Quant) and Qualitative (Qual) Yields of
Intrastrand Bipyrimidine Photoproducts in Isolated DNA
level of each step of the complex cascade of events linking Exposed to UV-C Radiation under Aqueous Conditions or in the
these factors to SPDNA formation remains, dehydration, Dry State in Both the Presence or the Absence of r,β-Type
whether intrinsic or induced, of the spore DNA environment SASP and Ca-DPAa
appears to be the most critical factor governing SPDNA ias-SPDNA 35/36 c,s T[CPD]T 5 T[6-4]C 10
formation. The different parameters identified so far are Quant Qual Quant Qual Quant Qual total refs
presented Figure 10.
solution
ND 238 42.3 187 33.3 562 48
6. Artificial Production of SPDNA, SPTIDE, and 29.7 40.4 15 20.4 73.5 96
SPSIDE dry
18 7.3 111 45.1 27 11 246.1 48
The intriguing photochemical properties of the DNA 3.19 10.6 9.72 32.3 10.1 33.5 30.13 53
present in spores have encouraged (bio)chemists to explore 0.7 3.3 7.8 37.1 7.4 35.2 21 63
the specificity of SPDNA production and to define the R/β-type SASP
conditions permitting its formation outside the spores. solution 25 23.6 49 46.2 2 1.9 106 48
Oligonucleotide models containing a definite number of spore dry 29 66.5 5 11.5 0.2 0.5 43.6 48
photoproducts at known locations are valuable tools in DPA
understanding the SPDNA mechanism of formation for ac- dry 14.9 22.7 30.4 46.3 10.4 15.8 65.7 63
curately studying its biological properties as well as for a 4
Expressed in lesions per 10 bases per J/cm . 2
unraveling its repair processes.
Table 3. Yields of SPDNA and c,s T[CPD]T 5 in Isolated DNA
6.1. UV Irradiation of Bacterial Vegetative Cells Produced by UV-C Irradiationa
Whereas it has been well established that the UV-C dose,
SPDNA, c,s T[CPD]T 5, J/cm2 refs
irradiation of bacterial vegetative cells at room temperature
does not lead to SPDNA, this latter lesion does form in the -100 °C 0.25 0.5 0.1 97
-99 °C 1.4 2 (+ U[CPD]T) 0.05 98
DNA of E. coli cells engineered to synthesize R/β-type SASP -100 °C W/EG 0.5 0.8 0.1 97
(3% of total thymine for a UV dose of 2.5 J/cm2).94 When solution 0.7 4.3 1 100
E. coli cells are exposed to UV-C irradiation at -79 °C, dry 3.1 2.2 1 100
SPDNA is also produced (1% of the total thymine for a UV 0.25 ND 0.1 87
dose of 0.2 J/cm2).95 This result confirms that neither R/β- solution + R/β-type SASP 3.5 0.7 1 100
dry + R/β-type SASP 5 <0.5 1 100
type SASP nor Ca-DNA is absolutely necessary for SPDNA 0.5 ND 0.1 87
formation. dry + R/β-type SASP + 3.4 ND 0.1 87
Ca-DPA
6.2. UV Irradiation of Isolated DNA a
Expressed as percent of the total thymine.
The formation of SPDNA from isolated DNA has also been
reported. The physical and chemical parameters that influence detected after UV-C irradiation. Qualitatively, the distribution
its formation in spores are also critical for its formation in of the PPs formed is c,s T[CPD]T 5 (40-42%), T[6-4]C
isolated DNA. 10 (20-33%), c,s T[CPD]C 6 (15-25%), and less than 7%
each of T[6-4]T 9, c,s C[CPD]T 7, and c,s C[CPD]C 8
6.2.1. Isolated DNA (Table 2).48,96
In vitro, SPDNA formation can be achieved by the UV-C At low temperatures, the UV-C irradiation of a frozen
irradiation of isolated DNA whose photochemical behavior aqueous solution of isolated E. coli DNA or Haemophilus
depends on both the temperature and the level of hy- influenzae-transforming DNA induces the formation of c,s
dration. T[CPD]T 5 and SPDNA in an approximately 2:1 ratio (Table
The photochemical behavior of plasmid and calf thymus 3).97,98 The optimum temperature for SPDNA formation lies
DNA has been studied both in an aqueous environment and between -100 and -120 °C. Hence, E. coli DNA
at room temperature. Under these conditions no SPDNA is photochemical behavior is qualitatively similar both in
1224 Chemical Reviews, 2010, Vol. 110, No. 3
vivo and in vitro, and SPDNA formation does not appear
to be influenced by factors intrinsic to E. coli cells.
Concurrently with the discovery of SPDNA, it was observed
that dehydration promoted SPDNA formation within isolated
DNA.50 In pUC19 plasmid DNA irradiated in the dry state,
SPDNA formation represents quantitatively 3.1% of the total
thymine (Table 3).94 Recent observations have allowed a
more precise description of the effect of dehydration on
SPDNA formation. UV-C irradiation of isolated plasmid or
calf thymus DNA in the dry state and at room temperature
leads to ias-SPDNA (35 or 36) formation (3-10% qualitative
yield) together with c,s T[CPD]T 5 (32-45%) (Table
2).48,53,63 pUC19 plasmid films obtained from air-dried 10
mM sodium phosphate buffer solution also afforded irs-
SPDNA (37 or 38), which represents 1% of the ias-SPDNA.48
In a dry film of calf thymus DNA prepared from a solution
in deionized water, irs-SPDNA and ias-SPDNA have been
obtained in equimolar amounts.99 This difference in qualita-
tive distribution has been attributed to a high salt to DNA
ratio, a condition that promotes the A-DNA conformation,
which in turn favors irs-SPDNA formation.48
Promotion of SPDNA formation from isolated DNA after
dehydration is consistent with the pioneering observations made
by Rahn and Hosszu,81 who used DNA films to study the
influence of the degree of humidity at 25 °C. Above 65%
relative humidity, the photochemical behavior of DNA films
is similar to that observed in solution and does not lead to
the formation of SPDNA. Below 65% relative humidity, SPDNA
formation occurs and is accompanied by a 2-fold quantitative
reduction in the yield of c,s T[CPD]T (5). The maximum
efficiency for SPDNA formation is observed at a relative
humidity of 40%.81
The influence of the isolated DNA water environment was
also studied by adding varying amounts of a series of alcohols.
At ethanol (EtOH) concentrations exceeding 60-70%, at room
temperature, SPDNA formation was detected. Optimum SPDNA
formation was observed for an EtOH concentration of 80%82,99
in a similar quantitative yield to that for UV-C irradiation of
heat-denatured DNA.82 Such results are of the utmost impor-
tance in eliminating the DNA conformation as a determining
factor for SPDNA production.82 Analysis of the SPDNA structure
revealed that UV-C irradiation of an 80% ethanolic solution
of DNA furnished ias- and irs-SPDNA (35 or 36 and 37 or
38, respectively) in a 9:1 ratio.99
The impact of the presence of alcohol and low temperature
on SPDNA formation has been studied using a 1:1 water/
ethylene glycol solution. Compared to the yield obtained with
a water solution irradiated at the same temperature, a 2-fold
enhancement of the quantitative yield of SPDNA was observed
around -100 °C under UV-C irradiation.97
6.2.2. Isolated DNA Complexed with R/β-Type SASP
Because R/β-type SASP and Ca-DPA are known to
influence SPDNA formation within the spore, their influence
in vitro has also been analyzed.
The formation of R/β-type SASP-DNA complexes (5:1
wt/wt) leads to a moderate to strong quantitative reduction
in the DNA UV-C reactivity.48,100 However, examination of
the respective distribution of the PPs has clearly established
that the in vitro binding of DNA to R/β-type SASP
dramatically modifies the photochemical specificity observed
for spore DNA. Binding of R/β-type SASP to DNA induces
a 3.5% quantitative formation of SPDNA (versus 0.7% in the
absence of R/β-type SASP) with respect to the total thymine
Desnous et al.
(Table 3). A reverse trend is observed with c,s T[CPD]T (5)
(4.3-0.7% of the total thymine) (Table 3).100 If all the
dipyrimidine PP formed is now considered, the SPDNA yield
of formation reaches 24% (SPDNA is not formed by DNA
UV-C irradiation in the absence of SASP) (Table 2).48
Interestingly, R/β-type SASP binding to DNA also induces
a decrease in the qualitative yield of formation of T[6-4]C
(10) (from 33% to 6%). The respective qualitative distribu-
tion of the other photoproducts remains substantially un-
changed. This indicates that the formation efficiency of all
other PPs has been similarly altered.48
From a qualitative standpoint, however, the in vitro
induction of SPDNA formation following binding of R/β-
type SASP to DNA is less efficient than in vivo. Here,
SPDNA represents 99% of the dipyrimidine PPs formed.
R/β-Type SASP-DNA binding is consequently important
for SPDNA formation, but other factors are also clearly
involved in vivo.
The UV-C irradiation of dry (i.e., films of) R/β-type SASP-
DNA complexes yields principally SPDNA. Under these
conditions, SPDNA represents 5% of the total thymine while
c,s T[CPD]T 5 represents ca. 0.1% (Table 3).100 SPDNA
represents 66% of the dimeric pyrimidine PPs versus 24%
in the solution state (Table 2). Concomitantly, c,s T[CPD]T
(5) formation strongly decreases (from 46% to 11% of the
PP pool).48 Interestingly, dehydration of the R/β-type SASP-
DNA complex induces a quantitative 2.4-fold reduction of
the PPs formed compared to the solution irradiation results.
A reduction of the same order is observed between the
amount of SPDNA obtained through irradiation of isolated
DNA in solution versus the amount of SPDNA obtained by
irradiation of isolated DNA in the dry state. The combined
action of R/β-type SASP binding and dehydration leads to a
quantitative 12.9-fold decrease in PP formation. Such a
decrease suggests a cumulative effect of the dry state (2.3-
fold) and R/β-type SASP binding (5.6-fold). Consequently,
R/β-type SASP binding to DNA may quantitatively reduce
∼5-fold the formation of PPs in solution as well as in the
dry state, and dehydration may reduce by 2-fold the
formation of PPs in either free or R/β-type SASP-bound
DNA. Again, it is important to keep in mind that the observed
quantitative fluctuations could, in part, be the result of
fluctuations induced by the experimental conditions.
6.2.3. Isolated DNA in the Presence of Ca-DPA
The influence of Ca-DPA (39) on the formation of PPs
in isolated DNA both in solution and in the dry state has
been studied recently using calf thymus DNA.63 In aqueous
solution, the presence of 39 induces a quantitative decrease
in PP formation. This reduced DNA photoreactivity, at-
tributed to UV-C radiation absorption by Ca-DPA, was not
observed on the UV-C irradiation of dry films of DNA
prepared in the presence of 1.5 mM Ca-DPA. Irradiation of
such dry films induces a quantitative 3-fold increase of PPs,
among which SPDNA, c,s T[CPD]T (5), and T[6-4]C (10)
qualitatively represent 23%, 46%, and 16%, respectively.
These correspond to about an 8-fold increase, a 1.2-fold
increase, and a 2-fold decrease compared to dry DNA films
irradiated without Ca-DPA (Table 2). Such a PP distribution
is fully in line with the proposed formation of SPDNA and
c,s T[CPD]T (5) via a DPA-mediated triplet energy trans-
fer process and of a (6-4) PPs via a singlet-state mecha-
nism.63
Spore Photoproduct
6.2.4. Isolated DNA in the Presence of R/β-Type SASP
and Ca-DPA
Irradiation of the pUC19 plasmid in the dry state with a UV
dose of 0.1 J/cm2 gives in the presence of R/β-type SASP and
Ca-DPA an approximately 10-fold enhancement in SPDNA forma-
tion from a R/β-type SASP/DNA dry complex and a 20-fold
enhancement from dry DNA (Table 3).87 Obviously, partial
UV-C absorption by SASP and/or Ca-DPA can also occur,
making an accurate quantitative comparison difficult.
6.3. UV-C Irradiation of Oligodeoxynucleotides
SPDNA has also been generated by UV-C irradiation of
ODNs in the single-stranded (ss) or double-stranded (ds)
state. Irradiation at 280 nm of the dA:dT duplex in an
equimolecular water/ethylene glycol solution at -196 °C
leads to the formation of SPDNA in a quantitative yield of
3% of the total thymine for a UV dose of 1.5 × 10-6 J/cm2.
However, under these conditions c,s T[CPD]T (5) is still the
main PP formed with a yield 6.5% of the total thymine.97
An SPDNA altered specifically at one site has been prepared
using the 35-bp ds of 5′-CCCGGGATCCTCTAGAGT-
TGACCTGCAGGCATGC-3′, a ds ODN that contains only
one TT site. Whereas no SPDNA was formed under irradiation
of an aqueous solution of this ds ODN at 254 nm, irradiation
at the same wavelength of a film prepared under 10% relative
humidity resulted exclusively in SPDNA formation, affecting
5% of the total thymine at a UV dose of 1.6 J/cm2.101 ias-
SPDNA has also recently been produced by UV-C irradiation
in the presence of Ca-DPA of a film of the ss ODN
5′-GGTTGG-3′ obtained by lyophilization.55
It has been reported that dry films of the dinucleoside
monophosphate of thymine (TpT, 41) afford SPTIDE (33
or 34) after UV irradiation in the presence of Ca-DPA or
Na2-DPA as photosensitizers.64 However, no yield was
reported. Without Ca-DPA, no (or only traces of) SPTIDE is
formed in dry films.61,64 Analytical and enzymatic repair
studies carried out on SPTIDE (33 or 34) prepared by
irradiation of TpT/Ca-DPA have shown that it is formed
diastereoselectively. The configuration at the C5-5,6-dihy-
dropyrimidine position is R as observed in ias-SPDNA (35 or
36). The 5,6-dihydropyrimidine moiety is located at the 5′
end, and the configuration of its C5 position is R (5-S 33).56
6.4. Irradiation of Thymidine
6.4.1. By Direct 254 nm Irradiation
Two different experimental conditions have been reported
for the preparation of SPSIDE 32 by the simple exposure of
thymidine to 254 nm irradiation (42). The irradiation can be
performed using either a frozen aqueous solution102 or a thin
solid film of 42.103
Chemical Reviews, 2010, Vol. 110, No. 3 1225
6.4.1.1. In Water at -78 °C. The UV-C irradiation of
frozen aqueous solutions of thymidine 42 yields SPSIDE 32
in a qualitative yield from ca. 13%63,102,103 to 38%99 depend-
ing on the irradiation dose. CPD PPs are always the major
photoproducts formed under these conditions, and the two
C5-diastereomers of SPSIDE 32 are produced in equal
amounts.99 Modification of either the aqueous solution ionic
strength or pH qualitatively promotes SPSIDE formation.102
In a 1 M NaCl or in a 0.1 N NaOH aqueous solution, 32 is
formed in 22% yield, compared to 14% in an aqueous
solution.102
6.4.1.2. In the Dry State at Room Temperature. SPSIDE
32 was isolated in a qualitative yield of 28% when a solid
film of 42 (obtained by evaporation of a methanolic solution)
was irradiated at 254 nm.103 Using thymidine films prepared
by lyophilization, Douki et al. reported a similar distribution
yield and the equimolar formation of the two C5-diastere-
omers.99 On the other hand, only one diastereomer, albeit of
unknown configuration, was obtained in a 78% qualitative
yield after the UV-C irradiation of thymidine films prepared
from a thymidine ethanolic solution.99
6.4.2. By Photosensitization
The formation of SPSIDE from a molecule as simple as
thymidine (42) has stimulated the search for new preparative
conditions. The UV-A (365 nm) irradiation of thymidine
films prepared by the evaporation of a thymidine/pyridop-
soralen methanolic solution leads to a diastereomeric mixture
of SPSIDE 32 in 3% quantitative yield (18% qualitative
yield).62 When benzophenone is used as the photosensitizer,
the 350 nm irradiation of films of 42 (obtained by evaporation
of an ethanolic solution) leads to only one SPSIDE diastere-
omer of unknown C5-configuration.61
Ca-DPA (39) has also been used as a photosensitizer.63
Addition of 1-9 mM of 39 to a frozen aqueous solution of
thymidine 42 induces an increase in the qualitative SPSIDE
yields from 23% to 39%. However, the quantitative reaction
yield decreases when the concentration of 39 increases.63
6.4.3. By γ-Irradiation, Heavy Ion, or Electron
Bombardment
The formation of SPSIDE (32) has also been observed when
different radiation sources were used. Indeed, the γ-irradia-
tion of frozen aqueous solutions of thymidine (42) at -78
°C leads to 32 formation in low yield.104 Electronic or O7+
heavy ion bombardment of compressed thymidine pellets also
leads to 32 in unreported yields.105 However, these irradiation
conditions also generate many radical reaction products.
7. Chemical Synthesis of SP and Derivatives
Work toward the chemical synthesis of SPSIDE (32) is
motivated by the need for precisely identified SPDNA for use
in mechanistic studies. The chemical synthesis of SP and its
derivatives has naturally been seen as an alternative to the
1226 Chemical Reviews, 2010, Vol. 110, No. 3
irradiation method. Adequately functionalized SP derivatives
can then be incorporated into ODNs at selected positions
and in defined sequence contexts.
7.1. Synthesis of SP
Two strategies have been used for the chemical synthesis
of SP (31). The first of these involves the bridging of two
pyrimidine base derivatives; the second is building the 5,6-
dihydropyrimidine skeleton onto a pre-existing pyrimidine
base. The dihydrothymine C5-exocyclic methylene bond
required in the first strategy has been created both under
electrophilic conditions and by the reaction of an anion
intermediate with an electrophilic center. Using the latter
alternative, Bergstrom et al. successfully condensed 6-ami-
nothymine 43 and 5-(trifluoroacetoxymethyl)uracil 44 to
obtain the unstable 5,6-dihydro-6-imino-5-(R-thyminyl)-
thymine 45 (Scheme 2). Reduction of 45 with NaBH3CN
under acidic conditions allowed its conversion into 31, which
was isolated in an overall yield of 2.7%.106
Scheme 2
Taking advantage of the acidic character of the H5 atom
of the 5,6-dihydrothymine moiety, the C5-CH2 bond has
also been formed using anion chemistry (Scheme 3). Ac-
cordingly, 31 was prepared in seven steps with an overall
Scheme 3
Scheme 4
Desnous et al.
yield of 3%. In the presence of LDA, the N1,N3-diprotected
derivative of 5,6-dihydrothymine 46 was transformed into
an anion intermediate that subsequently reacted with the
N1,N3-diprotected 5-formyluracil derivative 47. Formation of
the alcohol 48 resulted in a diastereomeric mixture formed
in 37% yield. A two-step reduction of 48 via a bromide
intermediate afforded the tetraprotected SP derivative 49 in
58% yield. Compound 49 was finally fully deprotected by
treatment with CAN followed by TFA to afford 31 in 6%
yield calculated from 46.107
For the synthesis of 50, the N1,N1’,N3,N3′-tetramethyl
analogue of SP, the 5,6-dihydrothymine moiety was con-
structed (Scheme 4). Halogenation of 5-hydroxymethyluracil
(51) afforded 52 in quantitative yield. It was then reacted
with the diethyl methylmalonate anion to yield 53. This was
subsequently N,N’-dimethylated and then converted into the
achiral barbiturate derivative 54 by treatment with urea under
basic conditions. Dimethylation using MeI yielded 55, and
controlled reduction of one of the two amide carbonyl groups
afforded the tetramethyl SP analogue 50.108
Although 50 is not a suitable starting compound for
producing the necessary intermediates for the synthesis of
SPDNA, its successful synthesis validates this synthetic
strategy.
7.2. Synthesis of SPSIDE and SPTIDE
For the preparation of ODNs containing the spore pho-
toproduct, direct access to SPSIDE or SPTIDE is more valuable.
SPTIDE 33 has been prepared using the nucleophilic
synthetic approach (Scheme 5). The N3-protected 5,6-
dihydrothymidine derivative 56 and the thymidine derivative
57 were independently prepared from thymidine 42 in
70-93% and 68-84% yield, respectively.109,54,110 Using
photobromination compound 57 was converted to its bromide
analogue 58 in 53-60% yield.109,54,110 This was then reacted
with the anion (LDA) of 56 to afford the diastereomeric
Spore Photoproduct Chemical Reviews, 2010, Vol. 110, No. 3 1227

Scheme 5

mixture 59. A two-step deprotection of 59 using TBAF SPSIDE derivative (67). Indeed, deprotonation of 63 using sec-
followed by treatment with SnCl4 and an intermediate HPLC BuLi leads to a carbanion that can react with nucleoside
separation afforded the two SPSIDE 32 diastereomers C5 S aldehyde 64 to afford 65 as a diastereomeric mixture in 44%
and C5 R separately in 8% each.54,110 Their respective yield. The deoxygenation of 65 which is necessary to obtain
configuration was indirectly determined using an oct-4- the SPSIDE analogue 67 was achieved with a 14% yield via
enedioate rigid cyclic derivative.110 From a slightly modified the xanthate intermediate 66 using H3PO2 and AIBN as
procedure, Chandra et al. were able to prepare 32 in ∼50% reducing agents (Scheme 6).111
yield from pure diastereomers 59.57 NMR analysis of each Interestingly, SPSIDE phosphoramidites 68 have been
C5 diastereomer of 32, their precursors, and a cyclic successfully incorporated into several oligonucleotides to
phosphotriester derivative allowed these authors to determine investigate the properties of SPDNA analogues.112 Amazingly,
the configuration at C5 of 32.57 Their assignment is inverse the resulting SPDNA analogues incorporate the 5,6-dihy-
to that previously reported by Friedel et al.110 drothymine moiety at the 3′ end of the dinucleoside motif.
Selective deprotection by HF in acetonitrile of the TES
group of the “target 5′-end sugar residue” of 59 followed
by primary alcohol protection by DMTC1 afforded 60 in
43% yield. Phosphorylation of the 3′ position gave 61 in
93% yield. This was followed by desilylation and cyclization
to give the dinucleotide analogue 62 in 99% yield. The
diastereomeric mixture was successfully separated using
chromatography. Deprotection of the phosphate group of 62
led to the disappearance of the phosphorus chiral center, and
finally, deprotection with SnCl4 followed by NH4OH resulted
8. DNA Photoproduct Repair
in the isolation of the C5 R and C5 S diastereomers 33 in
74% yield (Scheme 5).109
8.1. Conventional DNA Photoproduct Repair
Protection of the N3 atom of the 5,6-dihydrothymine Persistent DNA photoproducts can preclude replication and
moiety is not absolutely essential in order to prepare the transcription of DNA, thus leading essentially to cytotoxicity
1228 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 6
and ultimately cell death. Alternatively, low-fidelity replica-
tion and transcription of DNA, leading to heritable coding
changes, can occur as a result of incorrect nucleotide
incorporation opposite to the damaged nucleobases.113-115
This phenomenon is nevertheless sometimes considered as
a repair pathway and is referred to as mutagenic repair116 or
trans lesion synthesis.117,118
Stricto sensu repair of DNA photoproducts is achieved
by at least five DNA repair pathways generally dependent
on the type of PP. Repair of UV-induced DNA damage has
been specifically reviewed,116 while other broader reviews
cover DNA repair in mammals,119 humans,117 or plants.118,120
Because of the abundance of relevant literature, these repair
modes will only be listed and only the most recent reviews
cited.
8.1.1. Nucleotide Excision Repair (NER)121,122
Operating in both prokaryotes (including Bacillus species)
and eukaryotes, NER displays a wide substrate specificity
since it repairs dipyrimidine photoproducts and bulky DNA
adducts induced by environmental mutagens and chemo-
therapeutic drugs. Although the number of proteins interven-
ing in NER is different in prokaryotes and eukaryotes (4
versus at least 30, respectively), the general repair processes
is identical for these two kingdoms: after recognition of the
lesion, an ODN patch containing the photoproduct, which
has about 10 nucleotides in prokaryotes and about 24-32
nucleotides in eukaryotes, is excised from the damaged DNA
strand. The resulting gap is then filled by the action of a
Desnous et al.
polymerase that uses the complementary strand as a template,
and the resulting nick is sealed by a DNA ligase. Depending
on the location of the photoproduct within the genome, i.e.,
whether it lies in a transcribed or nontranscribed gene, two
distinct repair subpathways have been recognized. These
subpathways, which differ in the lesion recognition step, are
called transcription-coupled NER (TC-NER) and global
genome NER (GG-NER).
8.1.2. Base Excision Repair (BER)123-126
As observed in NER, the BER pathway is also shared by
prokaryotes and eukaryotes and involves the participation
of several enzymes. Distinctively, BER processes nonbulky
monomeric base lesions. Oxidized pyrimidine and purine
photoproducts are two examples of damage commonly
repaired by BER.
In brief, the first step in BER is the recognition and
excision of the damaged base from the corresponding
nucleoside residue through hydrolysis of its N-glycosidic
bond by DNA glycosylases. A nick or gap is then generated
by pathways using either monofunctional or bifunctional
glycosylases. In the case of a nick, a nucleotide insertion/
abasic site elimination process follows, whereas for a gap, a
nucleotide insertion follows. Complete DNA repair is finally
achieved by DNA ligases.
8.1.3. Photolyase-Induced Repair127,128
Photolyases are a group of enzymes that use visible light
for repair and photoreactivate dipyrimidine photoproducts
through an electron transfer mechanism. They have been
identified in some prokaryotes and eukaryotes including
fishes, plants, and aplacental mammals. This family of
enzymes has not been identified in B. subtilis. DNA repair
by photolyases is a single-enzyme process that requires two
cofactors: a light-harvesting photoreceptor and FADH-. It
should be noted that unlike NER and BER, photolyase-
induced repair results from interbase covalent bond breakage
and not from an excision/synthesis process. Consequently,
this type of repair should really be considered as a retro-
damage pathway. CPD- and (6-4) PP-containing DNA is
repaired by specific photolyases.
8.1.4. Repair of DNA Strand Breaks129-136
Repair of SSBs is controlled by mechanisms which
although as yet unclear are likely to involve a subset of the
repair enzymes also implicated in BER. Two distinct major
pathways, homologous recombination (HR) and nonhomolo-
gous end joining (NHEJ), have been identified for the repair
of DSBs. HR retrieves the coding information from a DNA
molecule that shares extensive sequence homology with the
damaged DNA and is therefore an accurate repair pathway.
In contrast, NHEJ, whose occurrence in prokaryotes has been
identified only recently, is more prone to errors since it
rejoins the DNA ends independently of any sequence
homology. In germinating spores, DSBs are principally
repaired by NHEJ.20,47
8.1.5. Others137,138
N-Glycosylases specific to CPD and (6-4) PPs have been
discovered in phages, eukaryotes, and prokaryotes including
Bacillus species. These enzymes may be part of an alternative
DNA excision repair pathway specific to UV-induced dipy-
Spore Photoproduct
rimidine PPs. The nick resulting from the glycosylase activity
could possibly be further processed by the general BER
pathway.
8.2. SPDNA Repair
The nucleotide excision repair pathway (NER) and repair
by the SP lyase enzyme are the two major distinct routes
currently thought to be the main processes for the repair of
SPDNA.139,140 Homologous recombination is a third albeit
minor repair process.141,142 Because NER is nonspecific and
has already been the subject of numerous reviews (in the
very similar E. coli system121 and in prokaryotes143), we will
focus only on the repair of SPDNA by SP lyase.
SP lyase is specific to the repair of SPDNA.101 SP lyase
from only three origins has been studied so far. That from
B. subtilis is currently the most characterized, while those
from Geobacillus stearothermophilus and Clostridium
acetobutylicum have been characterized more recently.
SP lyase was first cloned in 1993144 and is encoded by
the splB gene in B. subtilis144 and C. acetobutylicum145 and
the splG gene in G. stearothermophilus.55 SP lyase requires
strictly anaerobic conditions and can function in B. subtilis
as a monomer of ca. 43 kDa146 or as a homodimer.147 The
homodimer appears to be the active form for G. stearother-
mophilus SP lyase.55 The SP lyase is synthesized during
sporulation148 and is present in the developing spore. During
spore germination, SP lyase specifically binds to SPDNA and
regiospecifically cleaves the methylene bridge of SPDNA,
returning it to the dithymine-containing DNA without
excision of the SP lesion.101,149 The binding of SP lyase to
SPDNA is sequence context independent but structure specific.
It encompasses a 9 bp region surrounding the SP lesion and
causes significant distortion of the DNA. This is presumably
because the dinucleotide part of SPDNA flips out from the
interior of the helix.101 The SP motif can also be repaired at
the SPSIDE and SPTIDE level.54,55,64,57
Scheme 7
Chemical Reviews, 2010, Vol. 110, No. 3 1229
SP lyase belongs to the radical S-adenosylmethionine
(SAM) enzyme superfamily (for recent reviews, see refs
150-153). It uses an iron-sulfur cluster as a cofactor to
provide a one-electron reduction of a second cofactor,
S-adenosylmethionine (AdoMet, 69), to yield a 5′-deoxyad-
enosyl radical, which in turn initiates an hydrogen-atom
abstraction reaction.154 The iron-sulfur cluster consists of
four iron ions of variable oxidation states and four inorganic
sulfide ions ([4Fe-4S]) arranged in a cubane-type structure.
Three of the iron atoms are coordinated to three conserved
cysteine residues of the C91X3C95X2C98 motif of the SP lyase
amino acid sequence. The fourth iron is crucial for the
interaction between the cluster and 69 and is coordinated to
the nitrogen and a carboxylate oxygen of the methionyl
group of AdoMet, 69.145 There is one cluster per
monomeric form.55,146 The redox scheme involved is
e-
[Fe4S4]+ S [Fe4S4]2+ : the reduced form is the active one,
so that in vitro SP lyase needs anaerobic conditions for its
activity. However, since Bacillus species are aerobes, this
raises the question of the influence of oxygen on SP lyase
activity in vivo.
The current knowledge concerning the repair mechanism
of SP lyase is presented in Scheme 7. After binding to the
reduced cluster, AdoMet 69 receives an electron and
undergoes a reductive cleavage, leading to the 5′-deoxyad-
enosyl radical 70 and methionine (71). This radical then
abstracts an hydrogen atom from the C6 position of the 5,6-
dihydropyrimidine moiety of SPDNA to yield 5′-deoxyad-
enosine 72. The resulting SPDNA C6 radical 73 undergoes a
β-scission to regenerate the two parent thymines and 5′-
deoxyadenosyl radical 70 after hydrogen abstraction from
5′-deoxyadenosine 72 (Scheme 7, path a). As a consequence
of the formation of 73, the energy barrier required to break
the inter-pyrimidine C5-CH2 bond is dramatically lowered
to 6.2 kcal/mol.155 For the final step, density functional theory
calculations favor a two-step mechanism involving an
1230 Chemical Reviews, 2010, Vol. 110, No. 3 Desnous et al.

interthymine hydrogen-atom transfer (Scheme 6, path b).155 DMT dimethoxytrityl

Then, the 5′-deoxyadenosyl radical 70 and methionine 71 DPA dipicolinic acid
recombine to regenerate SAM 69 with loss of an electron DSB double-strand break
back to the iron-sulfur cluster: this completes the catalytic GG-NER global genome nucleotide excision repair
HR homologous recombination
repair cycle. Indeed, the role of 69 as a catalytic cofactor LDA lithium diisopropylamide
rather than as a cosubstrate has recently been definitively MSMT 1-(mesitylene-2-sulfonyl)-3-nitro-1,2,4-triazole
addressed.146 Support for this repair mechanism was initially NER nucleotide excision repair
provided by chemical model experiments108 and confirmed NHEJ nonhomologous end joining
by tritium label transfer experiments using either C6-tritiated PMB p-methoxybenzyl
SPDNA156 or 5′-tritiated S-AdoMet.146 The catalytic role of ROS reactive oxygen species
69 was inferred from three pieces of evidence: (1) the SASP small, acid-soluble spore proteins
absence of 5′-deoxyadenosine 72 from the repair reaction, SEM 2-(trimethylsilyl)ethoxymethyl
(2) the repair of a larger amount of SP than the amount of SSB single-strand break
69 present, and finally (3) the full incorporation of the tritium TBAF tetra-n-butylammonium fluoride
TBDMS tert-butyldimethylsilyl
label of 5′-tritiated 69 into the repaired thymine residues. TBDPS tert-butyldiphenylsilyl
In addition to the critical C91X3C95X2C98 motif, SP lyase TC-NER transcription-coupled nucleotide excision repair
contains a fourth conserved cysteine residue at position 141. TES triethylsilyl
Even if this residue does not participate in the cluster TFA trifluoroacetic acid
formation, cys 141 appears to play an essential role because
in mutants lacking this cysteine residue SP lyase activity is
lost.157 Cysteine 141 would be crucial for the conversion of
11. Acknowledgments
the allylic radical intermediate 74 to the parent thymine by We thank the CNRS for a doctoral fellowship for C.D.
preventing its reaction with exogenous free radicals.158
12. References
9. Conclusions (1) Setlow, P. Comments Mol. Cell. Biophys. 1988, 5, 253.
The unique photoreactivity of spore DNA has been (2) Kennedy, M. J.; Reader, S. L.; Swierczynski, L. M. Microbiology
1994, 140, 2513.
described in this review. The exact nature and influence of (3) Cano, R. J.; Borucki, M. K. Science 1995, 268, 1060.
all the factors inducing this remarkable behavior are as yet (4) Atrih, A.; Foster, S. J. Int. Dairy J. 2002, 12, 217.
unknown. Their discovery and role remains a true challenge (5) Marshall, M. M.; Hayes, S.; Moffett, J.; Sterling, C. R.; Nicholson,
for the scientific community. W. L. Appl. EnViron. Microbiol. 2003, 69, 683.
(6) Guerrero-Beltrán, J. A.; Barbosa-Cánovas, G. V. Food Sci. Technol.
Since spore photoproduct formation under the influence Int. 2004, 10, 137.
of UV-A and -B can be in part prevented by spore outer (7) Nicholson, W. L.; Galeano, B. Appl. EnViron. Microbiol. 2003, 69,
layer components, spore photoproduct formation under UV-C 1327.
radiation and its subsequent efficient repair also raises the (8) Horneck, G.; Mileikowsky, C.; Melosh, H. J.; Wilson, J. W.;
Cucinotta, F. A.; Gladman, B. Viable transfer of microorganisms in
enigmatic question of the rationale for the conservation of the solar system and beyond. In Astrobiobiology The Quest for the
this process in some Bacillus species and also of the Conditions of Life; Horneck, G., Baumstark-Khan, C., Eds.; Springer:
preservation of some of its associated molecules (SASP, Ca- Berlin, Heidelberg, 2002; p 57.
DPA). Whereas it can be easily understood that a UV-C (9) Nicholson, W. L.; Schuerger, A. C.; Setlow, P. Mutat. Res. 2005,
571, 249.
protective system was necessary for prokaryote survival when (10) Setlow, P. J. Bacteriol. 1992, 174, 2737.
UV-C was reaching the Earth’s surface, UV-C has been (11) Setlow, P. DNA structure, spore formation, and spore properties. In
filtered by high atmospheric layers for millions of years and Regulation of Bacterial Differentiation; Piggot, P. J., Moran, C. P.,
Youngman, P., Eds.; American Society of Microbiology: Washington,
no longer naturally reach the Earth’s surface. Even if under D.C., 1994; p 181.
sunlight (i.e., UV-B + UV-A) SPDNA is formed in small (12) Setlow, P. Annu. ReV. Microbiol. 1995, 49, 29.
amounts in bacterial spore DNA, such wavelengths lead to (13) Nicholson, W. L. Recent Res. DeVel. Microbiol. 1997, 1, 125.
other photoproducts whose biological importance appears (14) Nicholson, W. L.; Munakata, N.; Horneck, G.; Melosh, H. J.; Setlow,
P. Microbiol. Mol. Biol. ReV. 2000, 64, 548.
also critical and consequently whose repair is highly (15) Setlow, P. EnViron. Mol. Mutagen. 2001, 38, 97.
important for spore survival. Hence, SPDNA formation and (16) Nicholson, W. L.; Fajardo-Cavazos, P.; Rebeil, R.; Slieman, T. A.;
repair could be viewed as a complex but seldom used Riesenman, P. J.; Law, J. F.; Xue, Y. Antonie Van Leeuwenhoek
protective pathway whose genetic information has, however, 2002, 81, 27.
(17) Setlow, P. J. Appl. Microbiol. 2006, 101, 514.
been neatly preserved in the spore single chromosome. It (18) Driks, A. Cell. Mol. Life Sci. 2002, 59, 389.
can be safely assumed that because factors essential for (19) Setlow, P. Trends Microbiol. 2007, 15, 172.
SPDNA formation such as R/β-type SASP and Ca-DPA are (20) Wang, S. T.; Setlow, B.; Conlon, E. M.; Lyon, J. L.; Imamura, D.;
also used in sporulation and spore DNA heat protection, they Sato, T.; Setlow, P.; Losick, R.; Eichenberger, P. J. Mol. Biol. 2006,
358, 16.
have necessarily been preserved through evolution. Conser- (21) Vreeland, R. H.; Rosenzweig, W. D.; Powers, D. W. Nature 2000,
vation of the complex and highly specific SP lyase activity 407, 897.
necessary for the repair of SPDNA is by far the more surprising (22) Hazen, R. M.; Roedder, E. Nature 2001, 411, 155.
and could even be considered an ancient process. (23) Nicholson, W. L. Cell. Mol. Life Sci. 2002, 59, 410.
(24) Link, L.; Sawyer, J.; Venkateswaran, K.; Nicholson, W. Microb. Ecol.
2004, 47, 159.
(25) Setlow, P. Curr. Opin. Microbiol. 2003, 6, 550.
10. Abbreviations (26) Cadet, J.; Vigny, P. The photochemistry of nucleic acids. In
Bioorganic Photochemistry: Photochemistry and the Nucleic Acids;
AIBN 2,2′-azobisisobutyronitrile Harry, M., Ed.; J. Wiley: New York, 1990; Vol. 1, p 1.
BER base excision repair (27) Cadet, J.; Anselmino, C.; Douki, T.; Voituriez, L. J. Photochem.
BuLi butyllithium Photobiol. B 1992, 15, 277.
CAN cerium ammonium nitrate (28) Gorner, H. J. Photochem. Photobiol. B 1994, 26, 117.
Spore Photoproduct
(29) Ruzsicska, B. P.; Lemaire, D. E. DNA photochemistry. In CRC
Handbook of Organic Photochemistry and Photobiology; Horspool
W. M., Song, P. S., Ed.; CRC Press, Inc.: Boca Raton, FL, 1995; p
1289.
(30) Davies, R. J. Biochem. Soc. Trans. 1995, 23, 407.
(31) Cadet, J.; Berger, M.; Douki, T.; Morin, B.; Raoul, S.; Ravanat, J. L.;
Spinelli, S. Biol. Chem. 1997, 378, 1275.
(32) Ravanat, J. L.; Douki, T.; Cadet, J. J. Photochem. Photobiol. B 2001,
63, 88.
(33) Cadet, J.; Courdavault, S.; Ravanat, J. L.; Douki, T. Pure Appl. Chem.
2005, 77, 947.
(34) Cadet, J.; Sage, E.; Douki, T. Mutat. Res. 2005, 571, 3.
(35) Clivio, P.; Fourrey, J.-L.; Gasche, J.; Favre, A. J. Am. Chem. Soc.
1991, 113, 5481.
(36) Thomas, M.; Guillaume, D.; Fourrey, J. L.; Clivio, P. J. Am. Chem.
Soc. 2002, 124, 2400.
(37) Davies, R. J.; Malone, J. F.; Gan, Y.; Cardin, C. J.; Lee, M. P.; Neidle,
S. Nucleic Acids Res. 2007, 35, 1048.
(38) Boorstein, R. J.; Hilbert, T. P.; Cunningham, R. P.; Teebor, G. W.
Biochemistry 1990, 29, 10455.
(39) Douki, T.; Vadesne-Bauer, G.; Cadet, J. Photochem. Photobiol. Sci.
2002, 1, 565.
(40) Doetsch, P. W.; Zasatawny, T. H.; Martin, A. M.; Dizdaroglu, M.
Biochemistry 1995, 34, 737.
(41) Rapp, A.; Greulich, K. O. J. Cell Sci. 2004, 117, 4935.
(42) Ghosh, R.; Amstad, P.; Cerutti, P. Mol. Cell. Biol. 1993, 13, 6992.
(43) Wehner, J.; Horneck, G. J. Photochem. Photobiol. B 1995, 28, 77.
(44) Gniadecki, R.; Thorn, T.; Vicanova, J.; Petersen, A.; Wulf, H. C.
J. Cell. Biochem. 2000, 80, 216.
(45) Gut, I. G.; Farmer, R.; Huang, R. C.; Kochevar, I. E. Photochem.
Photobiol. 1993, 58, 313.
(46) Jagger, J. Ultraviolet inactivation of biological systems. In Photo-
chemistry and Photobiology of Nucleic Acids; Wang S. Y., Ed.;
Academic Press: New York, San Francisco, London, 1976; Vol. II
(Biology), p 147.
(47) Moeller, R.; Stackebrandt, E.; Reitz, G.; Berger, T.; Rettberg, P.;
Doherty, A. J.; Horneck, G.; Nicholson, W. L. J. Bacteriol. 2007,
189, 3306.
(48) Douki, T.; Setlow, B.; Setlow, P. Photochem. Photobiol. 2005, 81,
163.
(49) Slieman, T. A.; Nicholson, W. L. Appl. EnViron. Microbiol. 2000,
66, 199.
(50) Donnellan, J. E.; Setlow, J. R. B. Science 1965, 148, 308.
(51) Tyrrell, R. M. Photochem. Photobiol. 1978, 27, 571.
(52) Varghese, A. J. Biochem. Biophys. Res. Commun. 1970, 38, 484.
(53) Douki, T.; Cadet, J. Photochem. Photobiol. Sci. 2003, 2, 433.
(54) Friedel, M. G.; Berteau, O.; Pieck, J. C.; Atta, M.; Ollagnier-de-
Choudens, S.; Fontecave, M.; Carell, T. Chem. Commun. 2006, 445.
(55) Pieck, J. C.; Hennecke, U.; Pierik, A. J.; Friedel, M. G.; Carell, T.
J. Biol. Chem. 2006, 281, 36317.
(56) Mantel, C.; Chandor, A.; Gasparutto, D.; Douki, T.; Atta, M.;
Fontecave, M.; Bayle, P. A.; Mouesca, J. M.; Bardet, M. J. Am. Chem.
Soc. 2008, 130, 16978.
(57) Chandra, T.; Silver, S. C.; Zilinskas, E.; Shepard, E. M.; Broderick,
W. E.; Broderick, J. B. J. Am. Chem. Soc. 2009, 131, 2420.
(58) Eisinger, J.; Shulman, R. G. Proc. Natl. Acad. Sci. U.S.A. 1963, 50,
694.
(59) Pershan, P. S.; Shulman, R. G.; Wyluda, B. J.; Eisinger, J. Science
1965, 148, 378.
(60) Alcantara, R.; Wang, S. Y. Photochem. Photobiol. 1965, 4, 473.
(61) Douki, T.; Court, M.; Cadet, J. J. Photochem. Photobiol. B 2000,
54, 145.
(62) Moysan, A.; Viari, A.; Vigny, P.; Voituriez, L.; Cadet, J.; Moustacchi,
E.; Sage, E. Biochemistry 1991, 30, 7080.
(63) Douki, T.; Setlow, B.; Setlow, P. Photochem. Photobiol. Sci. 2005,
4, 591.
(64) Chandor, A.; Berteau, O.; Douki, T.; Gasparutto, D.; Sanakis, Y.;
Ollagnier-de-Choudens, S.; Atta, M.; Fontecave, M. J. Biol. Chem.
2006, 281, 26922.
(65) Lindberg, C.; Horneck, G. AdV. Space Res. 1992, 12, 275.
(66) Setlow, P. Annu. ReV. Microbiol. 1988, 42, 319.
(67) Lee, K. S.; Bumbaca, D.; Kosman, J.; Setlow, P.; Jedrzejas, M. J.
Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 2806.
(68) Setlow, B.; McGinnis, K. A.; Ragkousi, K.; Setlow, P. J. Bacteriol.
2000, 182, 6906.
(69) Setlow, B.; Setlow, P. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 421.
(70) Stafford, R. S.; Donnellan, J. E., Jr. Proc. Natl. Acad. Sci. U.S.A.
1968, 59, 822.
(71) Mohr, S. C.; Sokolov, N. V.; He, C. M.; Setlow, P. Proc. Natl. Acad.
Sci. U.S.A. 1991, 88, 77.
(72) Griffith, J.; Makhov, L.; Santiago-Lara, M. L.; Setlow, P. Proc. Natl.
Acad. Sci. U.S.A. 1994, 91, 8224.
(73) Setlow, P. Mol. Microbiol. 1992, 6, 563.
Chemical Reviews, 2010, Vol. 110, No. 3 1231
(74) Frenkiel-Krispin, D.; Sack, R.; Englander, J.; Shimoni, E.; Eisenstein,
M.; Bullitt, E.; Horowitz-Scherer, R.; Hayes, C. S.; Setlow, P.;
Minsky, A.; Wolf, S. G. J. Bacteriol. 2004, 186, 3525.
(75) Setlow, B.; Sun, D.; Setlow, P. J. Bacteriol. 1992, 174, 2312.
(76) Hayes, C. S.; Peng, Z. Y.; Setlow, P. J. Biol. Chem. 2000, 275, 35040.
(77) Hayes, C. S.; Alarcon-Hernandez, E.; Setlow, P. J. Biol. Chem. 2001,
276, 2267.
(78) Kosman, J.; Setlow, P. J. Bacteriol. 2003, 185, 6095.
(79) Nicholson, W. L.; Setlow, B.; Setlow, P. J. Bacteriol. 1990, 172,
6900.
(80) Sohail, A.; Hayes, C. S.; Divvela, P.; Setlow, P.; Bhagwat, A. S.
Biochemistry 2002, 41, 11325.
(81) Rahn, R. O.; Hosszu, J. L. Biochim. Biophys. Acta 1969, 190, 126.
(82) Patrick, M. H.; Gray, D. M. Photochem. Photobiol. 1976, 24, 507.
(83) Germaine, G. R.; Murrell, W. G. Photochem. Photobiol. 1973, 17,
145.
(84) Nudelman, R.; Bronk, B. V.; Efrima, S. Appl. Spectrosc. 2000, 54,
445.
(85) Daniel, R. A.; Errington, J. J. Mol. Biol. 1993, 232, 468.
(86) Tovar-Rojo, F.; Chander, M.; Setlow, B.; Setlow, P. J. Bacteriol.
2002, 184, 584.
(87) Setlow, B.; Setlow, P. Appl. EnViron. Microbiol. 1993, 59, 640.
(88) Paidhungat, M.; Setlow, B.; Driks, A.; Setlow, P. J. Bacteriol. 2000,
182, 5505.
(89) Lindsay, J. A.; Murrell, W. G. Curr. Microbiol. 1986, 13, 255.
(90) Lindberg, C.; Horneck, G. J. Photochem. Photobiol. B 1991, 11, 69.
(91) Nicholson, W. L.; Setlow, B.; Setlow, P. Astrobiology 2002, 2, 417.
(92) Douki, T. J. Photochem. Photobiol. B 2006, 82, 45.
(93) Donnellan, J. E., Jr.; Hosszu, J. L.; Rahn, R. O.; Stafford, R. S. Nature
1968, 219, 964.
(94) Setlow, B.; Hand, A. R.; Setlow, P. J. Bacteriol. 1991, 173, 1642.
(95) Smith, K. C. Biologically important damage to DNA by photoprod-
ucts other than cyclobutane-type thymine dimers. In Radiation
Research; Silini, G., Ed.; North-Holland Publishing Co.: Amsterdam,
1967; p 756.
(96) Douki, T.; Cadet, J. Biochemistry 2001, 40, 2495.
(97) Rahn, R. O.; Hosszu, J. L. Photochem. Photobiol. 1968, 8, 53.
(98) Rahn, R. O.; Setlow, J. K.; Hosszu, J. L. Biophys. J. 1969, 9, 510.
(99) Douki, T.; Laporte, G.; Cadet, J. Nucleic Acids Res. 2003, 31, 3134.
(100) Nicholson, W. L.; Setlow, B.; Setlow, P. Proc. Natl. Acad. Sci. U.S.A.
1991, 88, 8288.
(101) Slieman, T. A.; Rebeil, R.; Nicholson, W. L. J. Bacteriol. 2000, 182,
6412.
(102) Varghese, A. J. Biochemistry 1970, 9, 4781.
(103) Varghese, A. J. Photochem. Photobiol. 1971, 13, 357.
(104) Shaw, A. A.; Cadet, J. J. Chem. Soc., Perkin Trans. 2 1990, 2063.
(105) Gromova, M.; Balanzat, E.; Gervais, B.; Nardin, R.; Cadet, J. Int. J.
Radiat. Biol. 1998, 74, 81.
(106) Bergstrom, D. E.; Rash, K. F. J. Org. Chem. 1979, 44, 1414.
(107) Nicewonger, R.; Begley, T. P. Tetrahedron Lett. 1997, 38, 935.
(108) Mehl, R. A.; Begley, T. P. Org. Lett. 1999, 1, 1065.
(109) Kim, S. J.; Lester, C.; Begley, T. P. J. Org. Chem. 1995, 60, 6256.
(110) Friedel, M. G.; Pieck, J. C.; Klages, J.; Dauth, C.; Kessler, H.; Carell,
T. Chem.sEur. J. 2006, 12, 6081.
(111) Desnous, C. Ph.D. Thesis, Université de Paris-Sud, Faculté des
Sciences d’Orsay, Orsay, France, 2005.
(112) Burckstummer, E.; Carell, T. Chem. Commun. 2008, 4037.
(113) Lehmann, A. R. FEBS Lett. 2005, 579, 873.
(114) Hogg, M.; Wallace, S. S.; Doublie, S. Curr. Opin. Struct. Biol. 2005,
15, 86.
(115) Saxowsky, T. T.; Doetsch, P. W. Chem. ReV. 2006, 106, 474.
(116) Sinha, R. P.; Hader, D. P. Photochem. Photobiol. Sci. 2002, 1, 225.
(117) Christmann, M.; Tomicic, M. T.; Roos, W. P.; Kaina, B. Toxicology
2003, 193, 3.
(118) Kimura, S.; Sakaguchi, K. Chem. ReV. 2006, 106, 753.
(119) Schärer, O. D. Angew. Chem., Int. Ed. Engl. 2003, 42, 2946.
(120) Bray, C. M.; West, C. E. New Phytol. 2005, 168, 511.
(121) Reardon, J. T.; Sancar, A. Prog. Nucleic Acid Res. Mol. Biol. 2005,
79, 183.
(122) Gillet, L. C.; Schärer, O. D. Chem. ReV. 2006, 106, 253.
(123) Nilsen, H.; Krokan, H. E. Carcinogenesis 2001, 22, 987.
(124) Berti, P. J.; McCann, J. A. Chem. ReV. 2006, 106, 506.
(125) Hitomi, K.; Iwai, S.; Tainer, J. A. DNA Repair 2007, 6, 410.
(126) Hazra, T. K.; Das, A.; Das, S.; Choudhury, S.; Kow, Y. W.; Roy, R.
DNA Repair 2007, 6, 470.
(127) Sancar, A. Chem. ReV. 2003, 103, 2203.
(128) Kanai, S.; Kikuno, R.; Toh, H.; Ryo, H.; Todo, T. J. Mol. EVol.
1997, 45, 535.
(129) Dianov, G. L.; Parsons, J. L. DNA Repair 2007, 6, 454.
(130) Caldecott, K. W. DNA Repair 2007, 6, 443.
(131) Fortini, P.; Dogliotti, E. DNA Repair 2007, 6, 398.
(132) Karran, P. Curr. Opin. Genet. DeV. 2000, 10, 144.
(133) Jackson, S. P. Carcinogenesis 2002, 23, 687.
1232 Chemical Reviews, 2010, Vol. 110, No. 3 Desnous et al.

(134) Ishino, Y.; Nishino, T.; Morikawa, K. Chem. ReV. 2006, 106, 324. (149) Van Wang, T. C.; Rupert, C. S. Photochem. Photobiol. 1977, 25,
(135) Hefferin, M. L.; Tomkinson, A. E. DNA Repair 2005, 4, 639. 123.
(136) Pitcher, R. S.; Wilson, T. E.; Doherty, A. J. Cell Cycle 2005, 4, 675. (150) Sofia, H. J.; Chen, G.; Hetzler, B. G.; Reyes-Spindola, J. F.; Miller,
(137) Yasui, A.; McCready, S. J. BioEssays 1998, 20, 291. N. E. Nucleic Acids Res. 2001, 29, 1097.
(138) Lloyd, R. S. Prog. Nucleic Acid Res. Mol. Biol. 1999, 62, 155. (151) Fontecave, M.; Atta, M.; Mulliez, E. Trends Biochem. Sci. 2004,
(139) Munakata, N.; Rupert, C. S. J. Bacteriol. 1972, 111, 192. 29, 243.
(140) Xue, Y.; Nicholson, W. L. Appl. EnViron. Microbiol. 1996, 62, 2221. (152) Layer, G.; Heinz, D. W.; Jahn, D.; Schubert, W. D. Curr. Opin. Chem.
(141) Moeller, R.; Douki, T.; Cadet, J.; Stackebrandt, E.; Nicholson, W. L.; Biol. 2004, 8, 468.
Rettberg, P.; Reitz, G.; Horneck, G. Int. Microbiol. 2007, 10, 39. (153) Marsh, E. N.; Patwardhan, A.; Huhta, M. S. Bioorg. Chem. 2004,
(142) Munakata, N.; Rupert, C. S. Mutat. Res. 1975, 27, 157. 32, 326.
(143) Truglio, J. J.; Croteau, D. L.; Van Houten, B.; Kisker, C. Chem. (154) Rebeil, R.; Sun, Y.; Chooback, L.; Pedraza-Reyes, M.; Kinsland,
ReV. 2006, 106, 233. C.; Begley, T. P.; Nicholson, W. L. J. Bacteriol. 1998, 180, 4879.
(144) Fajardo-Cavazos, P.; Salazar, C.; Nicholson, W. L. J. Bacteriol. 1993, (155) Guo, J.-D.; Luo, Y.; Himo, F. J. Phys. Chem. B 2003, 107, 11188.
175, 1735. (156) Cheek, J.; Broderick, J. B. J. Am. Chem. Soc. 2002, 124, 2860.
(145) Chandor, A.; Douki, T.; Gasparutto, D.; Gambarelli, S.; Sanakis, Y.; (157) Fajardo-Cavazos, P.; Rebeil, R.; Nicholson, W. L. Curr. Microbiol.
Nicolet, Y.; Ollagnier-de-Choudens, S.; Atta, M.; Fontecave, M. C. R. 2005, 51, 331.
Chimie 2007, 10, 756. (158) Chandor-Proust, A.; Berteau, O.; Douki, T.; Gasparutto, D.; Ollagnier-
(146) Buis, J. M.; Cheek, J.; Kalliri, E.; Broderick, J. B. J. Biol. Chem. de-Choudens, S.; Fontecave, M.; Atta, M. J. Biol. Chem. 2008, 283,
2006, 281, 25994. 36361.
(147) Rebeil, R.; Nicholson, W. L. Proc. Natl. Acad. Sci. U.S.A. 2001, 98,
9038.
(148) Pedraza-Reyes, M.; Gutierrez-Corona, F.; Nicholson, W. L. J.
Bacteriol. 1994, 176, 3983. CR0781972
 Chem. Rev. 2010, 110, 1233–1277 1233

Well-Architectured Poly(dimethylsiloxane)-Containing Copolymers Obtained by

Radical Chemistry
Emmanuel Pouget, Jeff Tonnar, Patrice Lucas, Patrick Lacroix-Desmazes, François Ganachaud, and
Bernard Boutevin*
Institut Charles Gerhardt, UMR5253 CNRS/UM2/ENSCM/UM1, Ingénierie et Architectures Macromoléculaires, Ecole Nationale Supérieure de Chimie
de Montpellier, 8 rue de l’Ecole Normale, 34296 Montpellier Cedex 5, France

Received February 29, 2008

Contents 5.2. Surface Properties 1266

5.3. Mechanical Properties 1268
1. Introduction 1233 5.4. Thermal Stability 1268
2. Brief Overview of Synthesis Routes to 1236 5.5. Solution Properties 1269
Functionalized Polysiloxanes
5.6. Other Applications 1269
2.1. Synthesis Routes to Polysiloxane 1236
5.6.1. Supercritical CO2 1269
2.2. Functionalization of Polysiloxane 1237
5.6.2. Solid Electrolyte 1270
3. Copolymers Obtained by Conventional Radical 1238
Polymerization 5.6.3. Polymer Blends 1270
3.1. Polysiloxane Macroinitiators 1239 5.6.4. Nanocomposites 1270
3.1.1. Redox-Initiated Polymerization 1239 6. Conclusion 1270
3.1.2. Photoinitiated Polymerization 1240 7. Abbreviations 1272
3.1.3. Ozonide Macroinitiators 1241 8. References 1273
3.1.4. Bis-(silyl pinacolate) Macroinitiators 1241
3.1.5. Peroxycarbamate Macroinitiators 1241 1. Introduction
3.1.6. Peroxyester Macroinitiators 1242 Polyorganosiloxanes have been produced industrially since
3.1.7. Azo Macroinitiators 1242 the early 1940s, with the development of methylchlorosilane
3.2. Polysiloxane Macromonomers 1245 synthesis (Rochow-Muller direct synthesis) at the industrial
3.2.1. Synthesis of Polysiloxane Macromonomers 1245 scale. This was the basis for a rapid increase in sales
and Copolymerization worldwide, beginning in the United States. In 1992, the total
3.2.2. Multifunctional Macromonomers 1247 production of silicone was estimated to be up to 600 000
3.2.3. Macromonomers Reactivity in 1249 tons per year for a business volume estimated to 5.5 billion
Copolymerization dollars.1 In 2002,2 the overall production was estimated to
3.3. Polysiloxane Macrotransfer Agents 1250 be 2 million tons (Western Europe 33%, North America 34%,
3.3.1. Transfer to PDMS 1250 Asia 28%) for a total value of 8 billion euros. In Western
3.3.2. Silicone Containing Core-Shell Particles 1250 Europe, the production of silicone has been estimated to be
via Radical Polymerization 139 000 tons of fluids, 210 000 tons of sealants, 20 000 tons
3.3.3. Chemical-Initiated Telomerization 1252 of resins, and 139 000 tons of elastomers. One reason for
the high development of these silicon-based polymers and
4. Copolymers Obtained by Controlled Radical 1253
Polymerization materials is the diverse molecular structures of polyorga-
nosiloxanes and the intimate relationship between structure
4.1. Iniferter 1253
and properties not easily achievable by other classes of
4.2. NMP: Nitroxide-Mediated Polymerization 1256 polymers. Thousands of industrial products have been
4.3. ATRP: Atom-Transfer Radical Polymerization 1258 commercialized ranging from linear chains (fluids) to slightly
4.3.1. Main Features 1258 (rubbers) and highly (resins) cross-linked networks, leading
4.3.2. Block Copolymers 1258 to a wide range of applications. Silicone fluids are applied
4.3.3. Graft Copolymers 1260 in areas like proceeding aids (26%), personal care (24%),
4.4. ITP: Iodine-Transfer Polymerization 1261 paints and coatings (10%), paper coatings (15%), mechanical
4.5. RAFT: Reversible Addition-Fragmentation 1263 fluids (7%), and textile (5%). Due to their exceptional
Chain Transfer properties, silicone elastomers and resins have found ap-
4.5.1. Block Copolymers 1263 plications in automotive (20%), electrical fitting (15%),
4.5.2. Graft Copolymers 1263 medical/health (14%), appliances (9%), consumer goods
4.6. Conclusion on CRP Techniques 1264 (9%), textile coating (7%), business machines (5%), coatings
5. Some Properties and Some Applications 1265 (7%), and moldmarking (7%) areas.2
5.1. Phase Segregation Properties 1265 Polyorganosiloxanes exhibit exceptional properties as a
result of the constitutive unit Si-O-Si of the polymer
* To whom correspondence should be addressed. E-mail: backbone. The value of the Si-O bond length is equal to
bernard.boutevin@enscm.fr. 1.64 ( 0.03 Å,3 substantially smaller than that of the Si-O
10.1021/cr8001998  2010 American Chemical Society
Published on Web 12/14/2009
1234 Chemical Reviews, 2010, Vol. 110, No. 3
Emmanuel Pouget was born in Montpellier, France, in 1981. He graduated
from the National School of Chemistry of Montpellier in 2004. He obtained
his Ph.D. degree in 2007 at the University of Montpellier II, working under
the direction of Professor B. Boutevin and the cosupervision of Dr. F.
Ganachaud and Dr. P. Lacroix-Desmazes in the laboratory of Macromo-
lecular Engineering and Chemistry (IAM) led by Professor B. Boutevin.
His research covered silicone modifications (by radical polymerization or
chemical grafting in mass, solvent, or dispersed aqueous media) to obtain
nanostructured materials. In 2008, he joined BLUESTAR Silicones, where
he is currently working as a research manager in the chemistry of silicones.
Jeff Tonnar was born in Luxembourg, in 1979. He graduated from the
National School of Chemistry of Montpellier in France in 2004. He obtained
his Ph.D. degree in 2007 at the University Montpellier II, working under
the direction of Dr. P. Lacroix-Desmazes in the laboratory of Professor
Bernard Boutevin. His research covered iodine-mediated polymerization
[(reverse) ITP], especially in dispersed aqueous media (suspension,
miniemulsion, and ab initio emulsion polymerization). In 2007, he obtained
a postdoctoral research position in the laboratory of Professor S. Armes
(Sheffield University, U.K.). There he worked on the synthesis of
polymer-silica nanocomposites in emulsion polymerization. In 2008, he
joined CIBA, where he is currently working as a project manager on
emulsion polymerization for paper coatings. In 2008 he received the Thesis
Award from Groupe Français d’Etudes et d’Applications des Polymères.
bond length calculated from the additivity of the atomic radii
(1.83 Å). This is, on one hand, indebted to the partial double-
bond character of the Si-O bond3 and, on the other hand,
to the substantial ionic character (40-50%) of the Si-O
bond (electronegativity of Si ) 1.8 and O ) 3.5). This is
the reason why polyorganosiloxanes are considered to be
“organic-inorganic” elastomers compared to pure organic
elastomers (polybutadiene, polyisoprene). The main draw-
back that arises from the polarity of the Si-O bond is the
sensitivity of polysiloxanes to hydrolysis in acidic or basic
conditions. The high value of the Si-O-Si angle (140°) and
the value of the Si-O bond length partly explain the
exceptionally low glass transition temperature observed for
polyorganosiloxanes. For instance, PDMS exhibits a glass
Pouget et al.
Patrice Lucas was born in Versailles, France, in 1978. He graduated from
the National School of Chemistry of Clermont-Ferrand in France in 2002.
He obtained his Ph.D. degree in 2007 at the University Montpellier II,
working under the direction of Professor J.-J. Robin in the laboratory of
Professor Bernard Boutevin. His research dealt with silicone and polyamide
thermoplastic elastomer and the compatibilization of these two polymers
during reactive extrusion using a radical pathway. In 2006, he joined
NANOLEDGE (Montreal, Canada), where he is currently working as a
project manager on nanoparticles functionalization and integration in
thermosets polymers for high-performance structural materials. He is a
member of TRFA (Thermoset Resin Formulators Association).
Patrick Lacroix-Desmazes was born in Caen, France, in 1968. He
graduated from the National School of Chemistry of Montpellier, France,
in 1992. He obtained his Ph.D. degree in 1996 at the University Claude-
Bernard Lyon I, working in the laboratory of Professor Alain Guyot on the
use of macromonomers as stabilizers in dispersion polymerization in polar
media. After postdoctoral research on suspension polymerization, done
with BP Chemicals in Wingles in 1997, he joined the laboratory of Professor
Bernard Boutevin as a CNRS research scientist. In 1999, he developed
RITP, a promising method for controlled/living radical polymerization. He
received his Habilitation Degree in 2004. He was awarded the 2004
Innovative Research ADER Award in collaboration with Solvin Co. His
research interests cover living radical polymerizations (photoiniferters, NMP,
ATRP, RAFT, ITP, RITP), including in dispersed media (emulsion,
dispersion, suspension polymerization), as well as the synthesis and use
of polymers in liquid or supercritical carbon dioxide.
transition temperature of -123 °C due to torsional motion
along the backbone. It has been demonstrated that this very
low intermolecular force results in a large molar volume
(75.5 cm3 mol-1)4 and a low cohesion energy density.5 The
low surface tension, surface energy, solubility parameter, and
dielectric constant observed for PDMS can also be ascribed
to the low intermolecular forces between PDMS chains. The
presence of apolar methyl groups around the Si-O-Si
polymer backbone explains their high lipophilic and hydro-
phobic characters. The value of the Si-O bond dissociation
energy (BDE) (110 kcal mol-1)3 is high, compared to the
Poly(dimethylsiloxane)-Containing Copolymers
François Ganachaud, 40 this year, studied chemistry at CPE Lyon, where
he graduated as an engineering chemist. During his Ph.D work, granted
in 1997 from the University Claude-Bernard Lyon I, he proposed new
polymer latex/oligonucleotide complexes, under the supervision of Christian
Pichot. After a one-year postdoctoral position in Sydney at the KCPC
headed by Professor Robert G. Gilbert, he returned in France to take a
CNRS research position in Paris with Professor Hémery from 1999 to
2003. He then moved to Montpellier in the laboratory of Professor Bernard
Boutevin, where he completed his habilitation in 2004. His current research
interests are (i) polymer synthesis in aqueous dispersed media via ionic
polymerization, (ii) synthesis of functional silicones and related materials,
and (iii) emulsification of various solutes by Ouzo effect.
Bernard Boutevin earned his Chemical Engineering degree from the School
of Chemistry in Montpellier (France) and then received his Ph.D. degree
in 1975 from Montpellier University, where he became a CNRS researcher.
The topics of his research first concerned telomerization: kinetics, chemical
modifications, and their use in architectured polymers (block and grafted
copolymers). The concept was first applied to fluorinated monomers, and
it represents almost one-half of his scientific articles. Nowadays, his
research is directed toward telechelic oligomers, and all of the methods
of living radical polymerization are investigated (NMP, ATRP, ITP, RITP,
and RAFT) in order to prepare original architectured polymers such as
block, grafted, gradient, or alternated. In addition, phosphonated and
silicone-based monomers and (co)polymers are of increasing interest. He
has supervised more than 100 Ph.D. students and published 2 books,
more than 450 peer review articles, 25 reviews or chapters of books, has
been invited to more than 35 international conferences, and is coinventor
of more than 100 patents. In addition, more than 10 products coming
from his research group have been or are marketed by companies. He is
also a consultant for many companies, including Arkema, and he was a
consultant for Dow Corning (United States) and Daikin (Japan). He is a
member of the American Chemical Society and Groupe Français d’Etudes
et d’Applications des Polymères.
BDE of the C-O bond (85.5 kcal mol-1), C-C bond (82.6
kcal mol-1), and Si-C bond (76 kcal mol-1),6 which explains
the excellent thermal stability of polyorganosiloxanes. In
addition, PDMS has a good gas permeability7 and is
transparent to visible and UV light (the methyl groups do
Chemical Reviews, 2010, Vol. 110, No. 3 1235
not absorb radiation above 300 nm). Polyorganosiloxanes
are also resistant to ozone and corona discharge,3,8 exhibit
an excellent film-forming ability, and, among other proper-
ties, show some release action, surface activity, and chemical
and physiological inertness.8,9
This review focuses on PDMS copolymers because they
are the most common polyorganosiloxane copolymers.
However, the well-developed chemistry of silicone enables
one to replace the methyl groups by a wide variety of
functional groups (e.g., phenyl or 1,1,1-trifluoropropyl). This
can be done either randomly or all along the polysiloxane
backbone. Tailored properties can be achieved by judicious
choice of the substituents, depending on the targeted proper-
ties and applications.
Despite their unique properties, PDMS rubbers require
extremely high molecular weights to develop useful
mechanical properties. Also, chemical cross-linking is not
sufficient, and it is necessary to add finely divided high
surface area silica to obtain interesting properties.8,10–18
Recently, the use of physical cross-linking (hydrogen-bond
formation between two complementary groups attached to
the silicone backbone) allowed the synthesis of PDMS
rubbers owning exceptional properties without addition of
silica.19–21 These assemblies are thermoreversible and some-
times dissolve by adding a polar solvent. Another way of
preparing silica-free silicone rubbers is to prepare copolymers
that associate elastomeric segments (low Tg polymers, such
as PDMS) with thermoplastic segments (high Tg) to generate
thermoplastic elastomers (TPE) with improved mechanical
properties.22,23
Due to the exceptional properties of PDMS, interest indeed
has grown in associating PDMS with other polymeric entities
to obtain new materials. This strategy implies the synthesis
of copolymers with various well-controlled architectures:
block, multiblock, or graft copolymers. In the literature, the
elastomeric property is undoubtedly the most desired
characteristic to be conferred to copolymers. In addition,
amphiphilic copolymers can be obtained by associating
PDMS with a hydrophilic polymer. The incorporation of
a polysiloxane block in another polymer is also utilized
to modify the properties of the other polymer, making its
surface more hydrophobic, tuning the gas permeability or
flexibility of the material.
Many different synthesis routes exist to prepare polyor-
ganosiloxane block or graft copolymers. Anionic polymer-
izations and coupling reactions between a polysiloxane block
and another polymer have largely been treated in several
reviews24,25 and will not be examined here. On the other hand,
radical polymerization involving a polysiloxane backbone
has not been extensively described in the literature. Radical
polymerization has the advantage that it does not require the
same purity of reagents as anionic polymerization, which
makes it widely used at an industrial level. Moreover, radical
polymerization is applicable to a large range of monomers:
styrenics, acrylates and methacrylates, vinyl esters (including
vinyl acetate (VAc)), acrylamides, halogenated vinyl mono-
mers, and so on. In addition, radical polymerization can be
used in dispersed aqueous media, limiting the use of organic
solvents. Lastly, radical polymerization has also already been
used to prepare a wide range of different copolymer
architectures: diblock, triblock, multiblock, graft, or star
copolymers.
The aim of this review is thus to give a complete overview
of polysiloxane-containing block or graft copolymers pre-
1236 Chemical Reviews, 2010, Vol. 110, No. 3 Pouget et al.

Scheme 1. General Scheme of the Synthesis of Mono-, Di-, and Multifunctional Polysiloxanesa

a
D3 and D4 holds for hexamethylcyclotrisiloxane and octamethylcyclotetrasiloxane, respectively, whereas G holds for a functional group.

pared by radical polymerization of vinylic monomers. 2.1. Synthesis Routes to Polysiloxane

Following this introduction, we give a brief overview of the
chemistry of polysiloxanes needed for a good understanding
of the following sections. The third part deals with conven-
tional radical polymerization (PDMS macroinitiators, PDMS
macromonomers, and PDMS macrotransfer agents) followed
in the fourth part by controlled radical polymerization (IniFer-
Ter, NMP, ATRP, ITP, RAFT). The last section summarizes
some properties and applications of polysiloxane-containing
copolymers from studies quoted in this review.
2. Brief Overview of Synthesis Routes to
Functionalized Polysiloxanes
The rich chemistry of polysiloxanes offers many possibilities
to tailor the synthesis of well-controlled functionalized polysi-
loxanes. The amount, nature, and position of the functional
groups in the polysiloxane backbone determine the architecture
of the final copolymers. Since the synthesis of polysiloxanes
with controlled end groups and molar masses has been reviewed
by several groups,24–30 we will only give a brief summary of
the most popular synthesis routes.
Scheme 2. Synthesis of Monofunctional PDMS by Butyllithium-Initiated Ring-Opening Polymerization of D3
Polysiloxanes are synthesized using three principal routes:
ring-opening polymerization (ROP), polycondensation, and
redistribution (Scheme 1). The ring-opening polymerization of
cyclosiloxanes enables one to synthesize high molecular weight
siloxanes with better precision than the polycondensation and
redistribution methods. The most common cyclic siloxane
monomers are octamethylcyclotetrasiloxane (Me2SiO)4 (D4) and
hexamethylcyclotrisiloxane (Me2SiO)3 (D3).
The living anionic polymerization of D3 (Scheme 1, route
1),26 leading to ω-monofunctional polysiloxanes, is initiated
by strong inorganic, organic, or organometallic bases.
Typically, butyllithium (BuLi) initiates the polymerization
by first forming a silanolate anion which further propagates
by addition of D3 (Scheme 2). The counterion is usually an
alkali metal (here Li+), but it can also be a tertiary
ammonium or phosphonium cation. The functional group is
introduced during the deactivation of the silanolate ion using
a functional chlorosilane (Scheme 2).25 This technique
enables a good control of the molecular weight and the chain-
end functionality. Here, a promoting solvent such as tet-
Poly(dimethylsiloxane)-Containing Copolymers
Scheme 3. Synthesis of an r,ω-Diiodo-Functionalized Poly(dimethylsiloxane) with an Increased Degree of Polymerization by
Redistribution of D4 in the Presence of TONSIL EX0096 (sulfonic-acid-modified diatom clay) as Catalyst
rahydrofuran is required to initiate the polymerization of D3.
As will be depicted in the present review, it is possible to
initiate the polymerization of D3 with chemical entities of
larger interest than the butyl anion.
In comparison with D3, the controlled ring-opening po-
lymerization of D4 has not been widely studied, mainly
because of the lower reactivity of D4 and a less efficient
control over molecular weights and functionality. The main
technique of ring-opening polymerization of D4 involves
cheap mineral acid or base catalysts to prepare difunctional
polysiloxanes, which can be further used to prepare triblock
or multiblock copolymers. Nevertheless, the introduction of
pendant reactive side groups into a polysiloxane backbone
is quite easy by copolymerizing D4 with a functionalized
D4 monomer, for which one or both methyl groups on the D
unit are substituted by reactive groups. Number-average
molecular weights can be easily controlled by adjusting the
molar ratio of D4 over disiloxane transfer agent.24
Another way to prepare R,ω-difunctional polysiloxanes
is to hydrolyze dichloro- or dialkoxysilanes and to tune the
water content to control the molecular weight of the product
(Scheme 1, route 2).25 Silanol end groups are obtained by
the controlled hydrolysis of the chloro end groups. Hydro-
genopolysiloxanes are prepared by condensation reactions
of dichlorodimethylsilane in the presence of chlorodimeth-
ylsilane. This process is only applied in industry as a basis
for the synthesis of intermediaries in the silicone industry.
Generally, a mixture of cyclic and linear silicone is obtained
which is then polymerized to obtain longer silicone oils.31
Polycondensation of bis-silanol is also a powerful tech-
nique to obtain telechelic functional PDMS, again using acid
or base catalysts. The use of a functional end blocker allows
the preparation of functional telechelic polymers. For
instance, Graiver et al.32 emulsified silanol-terminated oli-
gomers in water with dodecylbenzene sulfonic acid (DBSA)
acting as both surfactant and acid catalyst. A minimum of
25 h of reaction at 22 °C yielded very high molecular weight
macromolecules.
The last technique used to obtain functional polysiloxanes
is to polymerize D4 and/or D4G in the presence of a
functionalized chain stopper or a R,ω-functionalized polysi-
loxane and a small amount of acidic or basic catalyst. Both
ROP of the cyclosiloxanes and subsequent redistribution of
the chains are necessary to exert a relatively good control
of molecular weights and functionality.33,34 As an example,
Chemical Reviews, 2010, Vol. 110, No. 3 1237
Scheme 4. General Scheme of the Hydrosilylation Reaction
our group recently carried out the redistribution of D4 with
an R,ω-diiodo-functionalized poly(dimethylsiloxane) to ob-
tain a functional poly(dimethylsiloxane) with a longer chain
length (Scheme 3).35 Redistribution has also been used to
prepare R,ω-dicarboxypropyl-,36 diaminopropyl-,37 and di-
hydroxybutyl-38functionalized poly(dimethylsiloxane). It is
performed by reacting D4 and the corresponding M2X (X-
SiMe2-O-SiMe2-X) under acidic (R,ω-dicarboxypropyl-func-
tionalized PDMS) or basic conditions (diaminopropyl- and
diaminobutyl-functionalized PDMS).
2.2. Functionalization of Polysiloxane
For a more complete overview of the transformation of
the polysiloxane chain ends, the reader is redirected to the
very complete review of Yilgor et al.24 The main reactions
used to transform the polysiloxane chain ends are hydrosi-
lylation, esterification, amidification, and nucleophilic sub-
stitution reactions, all of which are rapidly presented here.
Hydrosilylation is the most popular reaction for function-
alizing polysiloxanes owing to the ease of synthesis of
starting materials: mono-, di-, or multifunctional PDMS
containing Si-H groups. Hydrosilylation is the reaction
between a hydridosilyl function and either a CdC double
bond or another π bond. It is catalyzed by various metals,
such as chloroplatinic acid, by radical precursors, such as
di-tert-butyl peroxide, in some cases by amine complexes
or aluminum chloride, or other catalysts used for more
specific reactants.39 The general scheme of hydrosilylation
is given in Scheme 4.
For instance, to obtain R,ω-dihydroxypropyl PDMS, one
can hydrosilylate allyl alcohol with R,ω-dihydridosilyl PDMS
(Scheme 5).40 One might note here that some functional
groups require protection (typically acid, amine, or alcohol).
Besides hydrosilylation, conventional reactions of organic
chemistry like esterification, amidification, or nucleophilic
substitution can be used to transform the polysiloxane chain
ends (Scheme 6, routes 1-3). The reaction between an
isocyanate-functionalized molecule and an amino- or a
1238 Chemical Reviews, 2010, Vol. 110, No. 3 Pouget et al.

Scheme 5. Hydrosilylation of Allyl Alcohol onto a SiH Scheme 7. General Mechanism of the Synthesis of the
Functional PDMS40 r,ω-Diiodo PDMS Macrotransfer Agent Useful for ITP
Polymerization of Styrene and VAc (MSA is methane
sulfonic acid)

hydroxy-functionalized polysiloxane is also widely used

(Scheme 6, route 4).
For instance, our group41 carried out the esterification of
2-bromopropionic acid on R,ω-dihydroxypropyl PDMS.
Bromine was afterward substituted by iodine to yield an
active macrotransfer agent involved in iodine-transfer po-
lymerization of VAc and styrene41 (Scheme 7).
Amidification is a common route to prepare azo macroini-
tiators.42 Typically, R,ω-diaminopropyl PDMS is reacted with
4,4′-azobis-4-cyanopentanoyl chloride (Scheme 8).
selected depending on the desired reactivity of the PDMS
Nucleophilic substitution of chlorine by sodium dieth-
macroreagent. Different types of precursors used in conven-
yldithiocarbamate in a poly(chloromethylheptamethyltetra-
tional radical polymerization or controlled radical polymer-
siloxane) was carried out by Inoue et al.43 to prepare a macro-
ization are presented in the following sections.
iniferter (initiation transfer termination) (Scheme 9).
The reaction of an excess of diisocyanate with an R,ω-
diaminopropyl PDMS gives an R,ω-diisocyanato PDMS 3. Copolymers Obtained by Conventional Radical
which can be further reacted with tert-butyl hydroperoxide Polymerization
to obtain a peroxycarbamate macroinitiator (Scheme 10; see Conventional radical polymerization involving silicone
section 3.1.5 for details and references). polymers can follow various strategies. The first section is
The authors are not willing to make an exhaustive listing dedicated to the numerous examples of studies making use
of polysiloxane functionalization. More specific reactions of of PDMS macroinitiators, the decomposition of which leads
functionalization exist and will be described when required to the formation of block and graft copolymers. The second
all along this review. section is an overview of the synthesis of silicone mac-
In conclusion, synthesis of the polysiloxane backbone and romonomers and their reactivity and copolymerization. In
of course the technique of functionalization have to be well the last section, grafting through some “transfer to PDMS”

Scheme 6. General Scheme of the Esterification, Amidification, Nucleophilic Substitution, and Functionalization with an
Isocyanate Functionalized Molecule
Poly(dimethylsiloxane)-Containing Copolymers
Scheme 8. Synthesis of an Azo Macroinitiator by Reaction
of an r,ω-Diaminopropyl PDMS with 4,4′-Azobis-4-
cyanopentanoyl Chloride
Scheme 9. Synthesis of a PDMS Macroiniferter by
Nucleophilic Substitution of Chlorine by Sodium
Diethyldithiocarbamate
(in other words the random generation of radicals onto the
silicone backbone to initiate a polymerization) and telom-
erization will be discussed, including processes implemented
in aqueous media to prepare core-shell particles.
Scheme 10. General Mechanism of the Synthesis of PDMS Peroxycarbamate Macroinitiator
Scheme 11. Generation of r-Acyl Carbon-Centered Radical
Scheme 12. General Scheme for the Synthesis of Aldehyde Functional Polysiloxanes
Chemical Reviews, 2010, Vol. 110, No. 3 1239
3.1. Polysiloxane Macroinitiators
3.1.1. Redox-Initiated Polymerization
The first step of the redox-initiated polymerization is the
formation of an active radical on the PDMS chains followed
by the polymerization of the monomer from the so-called
PDMS macroinitiator. This technique allows the synthesis
of graft and block copolymers containing PDMS.
Most of the studies were carried out by Graiver et al.44–51
The method is based on a redox initiator system that
generates free radicals from enolates of aldehydes and
ketones using copper(II) salts (typically copper(II) 2-ethyl
hexanoate or copper(II) octanoate). This particular redox
system is known to yield R-acyl carbon-centered radicals
(Scheme 11).52,53 Graiver et al.44–51 found that in the presence
of vinyl monomers, these radicals can initiate polymerization,
leading to homopolymers with useful chain-end functionality
as well as block and graft copolymers.54–56 The first step of
the synthesis is the preparation of carbonyl functional
polysiloxanes achieved by ozonolysis of a hexenyl-containing
PDMS copolymer.45 The ozonide intermediate was reacted
with zinc and acetic acid to obtain the aldehyde functionality
(Scheme 12).
The polymerization of the vinyl monomer starting from
the carbonyl functionality of the polysiloxane is initiated by
a redox system: a soluble organic salt of CuII, a tertiary amine
such as triethylamine and dimethylphenylamine, a stabilizer
for CuII ions such as pyridine, and a stabilizer for CuI ions
such as triphenylphosphine. This strategy enabled the
authors to synthesize a PEA-b-PDMS-b-PEA (PEA )
poly(ethyl acrylate)) in 1 h at 70 °C44 and other different
block and graft PDMS copolymers: PMMA-b-PDMS-b-
PMMA,46,47,49 PS-b-PDMS-b-PS,46,47,49 PVPy-b-PDMS-b-
PVPy,50,51 PDMS-g-PS, PDMS-g-PtBuMA,48 PDMS-g-Pt-
BuA,48 PDMS-g-PMAA,46 PDMS-g-PAA,46 PDMS-g-PM-
MA,48 and PDMS-g-PVPy.50,51
The polymerization of N-isopropylacrylamide (NIPAM)
initiated by the (NH4)2CeIV(NO3)6 -R,ω-dihydroxyalkyl(poly-
dimethylsiloxane) redox pair was also studied to obtain a
thermosensitive triblock copolymer PNIPAM-b-PDMS-b-
PNIPAM.57 This polymerization was performed at 30 °C in
hexane containing a small amount of 1,2-dichloroethane to
increase the solubility of NIPAM (heterogeneous solution
1240 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 13. Mechanism of Formation of Alkyl Radicals by Type II Photoinitiators
Scheme 14. Macro-Photoinitiators Used by Pouliquen et al.62,63
polymerization). Low conversions (<30%) were obtained
after 48 h of reaction. The formation of PNIPAM homopoly-
mer chains initiated and/or terminated by the free radicals
generated from monomer-CeIV complexes and the presence
of unreacted PDMS oligomers were observed.
3.1.2. Photoinitiated Polymerization
Photoinitiated polymerization is accomplished by incor-
porating light-sensitive initiators in the polymer structure:
photoreactive chromophores such as azo initiator58 (see azo
macroinitiator, section 3.1.7) or diethyldithiocarbamate43 (see
iniferter, section 4.1). Such initiators can be classified in two
groups: type I photoinitiators lead to active radicals by
monomolecular cleavage (e.g., benzoin ethers),43,58–61 whereas
type II photoinitiators62,63 generate active radicals by hydro-
Pouget et al.
gen photoabstraction, preferably induced by exciplex forma-
tion in the presence of an amine co-initiator (e.g., benzophe-
none/tertiary amine system).64–67 The mechanism of primary
alkyl radicals formation for type II photoinitiators is depicted
in Scheme 13.
To the best of our knowledge, only two studies dealt with
radical photopolymerization using type II photoinitiators.
They were carried out by Pouliquen et al.,62,63 who synthe-
sized the photoinitiators given in Scheme 14. The photo-
cross-linking of PDMS was performed by copolymerizing a
monoacrylate (2-ethylhexyl acrylate) and a diacrylate (1,6-
hexanediol diacrylate). By comparing the single molecules
to the polymeric system, an increase in initiation efficiency
was noticed for the polymer. This “macromolecular effect”
was even greater when using a polysiloxane containing only
Poly(dimethylsiloxane)-Containing Copolymers
Scheme 15. Synthesis of Ozonide-Functionalized
Polysiloxane
2-benzoyl-benzoic ester groups associated with free 4-dim-
ethylaminobenzoic ester in a photosensitive composition (first
photoinitiator in Scheme 14).
Photoinitiated polymerization has been only scarcely
studied to obtain PDMS-containing copolymers. For instance,
photopolymerization of acceptor-donor systems containing
a PDMS backbone has never been investigated. Likely, recent
developments in photoinitiator-free photopolymerization68,69
could be easily adapted to the synthesis of novel photocurable
compositions containing polysiloxane.
3.1.3. Ozonide Macroinitiators
To our knowledge, the use of ozonide-functionalized
polysiloxanes to prepare silicone/organic copolymers by
radical polymerization is described in only one patent
assigned to Dow Corning Corp.70 The first step of the
synthesis is the preparation of the ozonide-containing pol-
ysiloxane, accomplished by ozone bubbling through a
solution of polysiloxane containing a reactive double bond.
As an example, a 5-hexenyl-terminated polysiloxane was
dissolved in dichloromethane at -15 °C and ozone was
introduced at the bottom of the solution. The structure of
the ozonide (Scheme 15) was confirmed by 13C NMR. The
second step consists of the radical polymerization of a vinyl
monomer such as ethyl acrylate or acrylamide. The former
monomer was polymerized at 70 °C for 1 h in the presence
of the ozonide-terminated polysiloxane to obtain a poly(ethyl
acrylate)-b-PDMS-b-poly(ethyl acrylate) triblock copolymer.
Furthermore, the siloxane chain appeared to moderate the
decomposition rate of the ozonide group: the ozonide
obtained from 1-hexene with the same method as before was
found to be quite unstable and rapidly decomposed through
an uncontrolled exothermic reaction.
3.1.4. Bis-(silyl pinacolate) Macroinitiators
Bis(silyl benzopinacolate)s have been introduced as mac-
roinitiators by Crivello et al.71–77 through the hydrosilylation
of the bis(dibenzylvinylsilyl)pinacolate with R,ω-hydride-
terminated PDMS (Scheme 16).
The authors studied the kinetics of thermolysis of bis(silyl
pinacolate), depending on the nature of the groups (phenyl
or methyl) of the thermolyzable moiety. The cleavage
between pinacolate groups creates two active radicals which
initiate the radical polymerization of styrene,73 methyl
methacrylate (MMA),73 and tert-butyl methacrylate.77 More-
over, the absence of homopolymer was evidenced by
chromatography.72 The major drawback of this method is
the formation, in the final copolymer, of Si-O-C bonds
known to be prone to hydrolysis.78,79
Chemical Reviews, 2010, Vol. 110, No. 3 1241
Scheme 16. Synthesis of Benzopinacolate Macroinitiator by
Hydrosilylation
3.1.5. Peroxycarbamate Macroinitiators
The two-step synthesis of peroxycarbamate-based mac-
roinitiators consists of (i) the reaction of a diisocyanate with
either a hydroxy-terminated PDMS or a hydroxy-terminated
block copolymer containing a PDMS segment and (ii) the
reaction of tert-butyl hydroperoxyde with the unreacted
isocyanate chain ends (Scheme 10). The different structures
reported in the literature are given in Table 1.
The first synthesis of PDMS-containing copolymers using
peroxycarbamate macroinitiators has been performed by
Baysal et al.80,81 They polymerized styrene in bulk or in
toluene solution at 80 °C for 5 days using a peroxycarbamate-
terminated PDMS (Table 1, entry 1). A very high content
of homopoly(styrene) was observed at low macroinitiator
concentration. Unexpectedly, the polymerization yield de-
creased when the macroinitiator concentration increased; no
explanation was given by the authors. The high content of
homopoly(styrene) is essentially due to the presence of an
active small radical created by the decomposition of the
macroinitiator (i.e., t-Bu-O•). The same authors in another
study82,83 used similar peroxycarbamate macroinitiator 1
(Table 1) to prepare PMMA-b-PDMS-b-PMMA, poly(dibu-
tyl itaconate)-b-PDMS-b-poly(dibutyl itaconate), poly(dicy-
clohexyl itaconate)-b-PDMS-b-poly(dicyclohexyl itacon-
ate),83 poly(monobutyl itaconate)-b-PDMS-b-poly(monobutyl
itaconate), and poly(monocyclohexyl itaconate)-b-PDMS-b-
poly(monocyclohexyl itaconate).82 Since the authors could
not achieve the synthesis of high molecular weight block
copolymers, resulting materials exhibited poor mechanical
and physical properties.
More recently, Taskiran84 carried out the synthesis of PS-
b-PDMS-b-PS copolymers to modify the properties of
expandable PS. PDMS-filled expandable PS was synthesized
using siloxane-containing peroxycarbamate macroinitiator 1
(Table 1) by suspension polymerization. The final products
showed better thermal resistance and surface properties than
the non-modified expandable PS.
Uyanik85 prepared a five-block copolymer PVP-b-PCL-
b-PDMS-b-PCL-b-PVP (PVP ) poly(vinyl pyrrolidone) and
PCL ) poly(caprolactone)) using a telechelic peroxycar-
bamate macroinitiator composed of PCL and PDMS (Table
1, entry 2). This macroinitiator was prepared from a
commercial PCL-b-PDMS-b-PCL triblock copolymer and
using the synthesis procedure described by Baysal et al.80
The resulting copolymers showed higher thermal resistance
and increased toughness characteristics than the correspond-
ing homopolymers.
1242 Chemical Reviews, 2010, Vol. 110, No. 3
Table 1. Peroxycarbamate Macroinitiators and Monomers Used to Prepare Hybrid Copolymers
Five block copolymers of acetophenone-formaldehyde
resins (AFR), PDMS, and PVP or PS were synthesized by
Uyanik et al.86 They first prepared a triblock copolymer
PDMS-b-AFR-b-PDMS which was functionalized by reac-
tion of an excess of diisocyanate to the hydroxy chain ends
of the central block followed by the reaction of the
isocyanate-terminated central block with tert-butyl hydro-
peroxide to obtain a macro-peroxycarbamate initiator (Table
1, entry 3). The monomer (styrene or N-vinyl pyrrolidone)
was polymerized using this macroinitiator. The authors
showed that the solubility of the five-block copolymer is
dominated by the middle resin block even though its block
length is much smaller than those of the vinyl polymer
blocks.
Uyanik et al.87 also prepared poly(vinyl pyrrolidone)-
b-PDMS-b-poly(vinyl pyrrolidone) triblock copolymers
using the peroxycarbamate macroinitiator 4 (Table 1). These
copolymers were characterized by differential scanning
calorimetry (DSC) and stress-strain tests (see section 5 of
this review).
In conclusion, the synthesis of peroxycarbamate macro-
initiators is easy and relatively inexpensive. The main
drawback of this method is the undesired formation of
homopolymer concomitant to the copolymerization due to
the presence of free tert-butoxy radicals.
3.1.6. Peroxyester Macroinitiators
Peroxyester PDMS macroinitiators were exclusively used
by the Nippon Oils and Fats (NOF) Corp.88–91 The authors
Pouget et al.
prepared PDMS-containing block copolymers using perox-
yester-terminated PDMS (Scheme 17).
For instance, the polymerization of MMA initiated by a
peroxyester-terminated PDMS (Scheme 17 structure 1)
resulted in the formation of a triblock copolymer which can
be incorporated in a polyurethane resin to form water-
repellent coatings.91 Again, as for the peroxycarbamate
macroinitiators presented in the previous section, tert-butoxy
radicals resulting from the macroinitiator decomposition
initiate the homopolymerization of the vinyl monomers to
yield a mixture of homopolymers and copolymers.
3.1.7. Azo Macroinitiators
Azo-containing polymers are interesting materials for the
synthesis of block and graft copolymers. These compounds were
used in several studies to prepare new materials with enhanced
properties.92 The synthesis of azo macroinitiators based on
PDMS is generally easy. Most often, it is accomplished by
a condensation (esterification, amidification) or a hydrosi-
lylation reaction between a functionalized azo precursor and
a PDMS-containing reactive group (see section 2).
The first synthesis of siloxane-vinyl block copolymers
using an azo macroinitiator of PDMS was reported in a
Japanese patent.93 Inoue et al. condensed 4,4′-azobiscyano-
pentanoyl chloride with R,ω-diaminopropyl PDMS to prepare
an azo macroinitiator 1 (Table 2, entry 1, m ) 3), which
was used for the synthesis of triblock copolymers of
poly(methyl methacrylate) (PMMA), poly(vinyl acetate), or
polystyrene and PDMS.42 The condensation reactions were
Poly(dimethylsiloxane)-Containing Copolymers
Scheme 17. Examples of Peroxyester-Terminated PDMS Used by the NOF Corporation88,90,91
found to proceed almost quantitatively. Depending on the
desired properties, various types of polysiloxanes could be
used: PDMS, poly(methyl-phenyl siloxane), or poly(methyl-
3-trifluoropropyl siloxane). The authors found that the azo
macroinitiators have a high polymerization activity, although
somewhat lower than that of 2,2′-azobisisobutyronitrile
(AIBN), as expected from previous results obtained with
other polymeric azo macroinitiators.94,95
The reactivity of azo macroinitiator 2 (Table 2, entry 2,
m ) 4) was studied by Chang and co-workers in a series of
articles58,111–113 where block copolymers of PDMS and
PMMA were prepared under UV58,113 or thermal112 initiations.
In the case of thermal initiation, the macroinitiator decom-
position followed an expected first-order reaction rate.121
Decomposition rate constants of the macroinitiator were 4.5
× 10-5, 12.5 × 10-5, and 30.5 × 10-5 s-1 at 70, 78, and 85
°C, respectively. In the case of UV-initiated polymerization
of MMA,58,113 a higher polymerization rate was observed
when using the macroinitiator compared to AIBN. The
termination rate constant (kt) in the MMA/PDMS photoini-
tiator system was smaller than in the equivalent MMA/AIBN
system (4.5 × 107 versus 9.1 × 107 mol L-1 s-1). The
propagation rate constants kp for the MMA/PDMS photo-
initiator system and the MMA/AIBN system were, respec-
tively, 495.2 and 445.8 L mol-1 s-1.
Simionescu107 prepared polystyrene, PMMA, poly(acry-
lonitrile), poly(butyl acrylate), or poly(butyl methacrylate)-
silicone block copolymers in bulk and solution (toluene)
using azo macroinitiator 2 (Table 2, entry 2, m ) 3). As
expected, increasing the monomer/azo group ratio resulted
in an increase of molecular weights. PMMA-b-PDMS
copolymers were also synthesized using azo macroinitiator
3 (Table 2, entry 3).108 The influence of the polymerization
temperature, macroinitiator molecular weight, composition
of the initial mixture (vinyl/azo and vinyl/siloxane ratios),
and initiator and monomer concentrations was studied. Due
to the low content of azo-ester groups, the thermal decom-
position of these azo macroinitiators 3 was not studied.
However, their efficiency in radical polymerization of vinyl
monomers was proved.110,114
The thermal decomposition of azo macroinitiator 2 (Table
2, entry 2, m ) 3) in toluene solution was studied by
measuring the decrease of the intrinsic viscosity vs time at
80 °C.108 A plateau was reached after 7-8 h corresponding
to the total decomposition of the azo macroinitiator. De-
composition rate constants kd were found in the range from
8.1 to 9.8 × 10-5 s-1 and decreased with the PDMS chain
Chemical Reviews, 2010, Vol. 110, No. 3 1243
length (between 1700 and 7100 g · mol-1). These values are
similar to those measured by nitrogen release titration for
4,4′-azobis(cyanopentanoic) acid (9.1 × 10-5 s-1).122 High
conversion in solution polymerization demonstrated a higher
efficiency with low-siloxane chain-length initiators (high azo
molar content). Homo-PMMA was not formed with azo
macroinitiator 3 (Table 2, entry 3), which was explained by
the fact that tert-butyl radicals formed through decomposition
of the azo initiator are very reactive and recombine im-
mediately. However, we noticed that this explanation is not
in agreement with the results of Fischer et al.123–125 and
Hammond et al.,126 who established that R1 . Rt (R1 and Rt,
respectively, refer to the rate of the reaction tBu• + MMA
and of the coupling reaction tBu• + tBu•).
The azo macroinitiator 1 (Table 2, entry 1, m ) 3) and
the azo macroinitiator 2 (Table 2, entry 2, m ) 3) were also
used in different processes of polymerization: dispersion
polymerization (organic media, supercritical CO2 (scCO2)),
emulsion polymerization, and precipitation polymerization.
PDMS-PMMA block copolymers using azo macroinitiator
1 were prepared by dispersion polymerization.99 The azo
macroinitiator acts as both a stabilizer and an initiator for
the dispersion polymerization of MMA in a good solvent of
PDMS, typically, n-heptane or cyclohexane, which have
solubility parameters in the range of 7.4-8.2 cal1/2 cm-3/2,
close to the solubility parameter of PDMS (around 7.3 cal1/2
cm-3/2).127 The polymerization produced nanospheres with
diameters in the range of 170-270 nm by controlling the
concentration of the azo macroinitiator.
More recently, Okubo et al.101 produced submicrometer
PMMA particles by dispersion polymerization with azo
macroinitiator 1 (Table 2, entry 1, m ) 3) in scCO2. The
PDMS chains operated as a steric barrier. The number-
average particle diameter measured by scanning electron
microscopy (SEM) was 210 nm.
Noguchi et al.106 prepared latexes by aqueous miniemul-
sion polymerization of azo macroinitiator 1 (Table 2, entry
1, m: unknown) and MMA, butyl acrylate (BuA), methacrylic
acid (MAA), and hydroxyethyl methacrylate (HEMA) in
water. The resulting copolymers were useful for coatings and
as additives for paper, fibers, and films, which showed
improved water repellency, weatherability, sliding property,
and gas permeation.
Precipitation polymerization with azo macroinitiators 1 and
2 (Table 2, m ) 3) in toluene was carried out by Szajdzinska-
Pietek et al.102 and Pinteala et al.109 to prepare PMAA-b-
PDMS copolymers. Pyrene was used as a fluorescent probe
1244 Chemical Reviews, 2010, Vol. 110, No. 3 Pouget et al.

Table 2. Azo Macroinitiators and Monomers Used To Prepare Hybrid Copolymers

to study the conformational changes of this block copolymer
in aqueous solutions (see section 5).
PB-g-PDMS and PDMS-b-PS were synthesized in bulk
using azo macroinitiators 1 (Table 2, entry 1, m ) 4).103,104
In the case of the PB-g-PDMS, the reaction proceeds by
radical addition to pendent vinyl groups of poly(butadiene)
(PB) chains. It was observed that an increase in macroinitiator
concentration up to 52% w/w in the PB solution resulted in
a cross-linked graft copolymer. Molecular weights of soluble
graft copolymers were between 450 000 and 600 000
g · mol-1. In both PB-g-PDMS and PDMS-b-PS cases, the
molecular weights increased when the macroinitiator con-
Poly(dimethylsiloxane)-Containing Copolymers
Scheme 18. Structure of the Macromonomeric Initiator
(macroinimer) Used by Hamurcu104
centration decreased. Hamurcu104 also carried out the cross-
linking of PS using a macroinimer, the structure of which is
given in Scheme 18.
Azo macroinitiator 1 (Table 2, entry 1, m ) 3) was also
used to synthesize poly(vinyl chloride) (PVC)/PDMS block
copolymers105 containing only 4.5% of PDMS. The material
obtained showed better processability properties than pure
PVC and exhibited a contact angle of 100° instead of 85° in
the case of pure PVC.
The synthesis of styrene-, MMA-, butyl methacrylate
(BuMA)-, and BuA-containing block copolymers was per-
formed in toluene using azo macroinitiator 3 (Table 2, entry
3).114 By size exclusion chromatography (SEC) analysis and
careful extraction of unreacted azo macroinitiator, it was
shown that the obtained block copolymers are of ABA type
when polymerizing MMA (due to the termination by
disproportionation)128 and of multiblock copolymers (AB)n
type when styrene was polymerized (due to the termination
by recombination).128,129
More recently, Bertolucci et al.115 carried out the synthesis
of PDMS-b-poly(semifluorinated styrene)-b-PDMS in trif-
luorotoluene using the azo macroinitiator 4 (Table 2, entry
4). A triblock copolymer was formed showing that the
termination occurred by recombination. Even for short PDMS
chains (n ) 8), the authors observed a thermotropic me-
sophase which was attributed to the self-assembly of
fluorinated segments in microphase-separated domains.
Azo macroinitiator 5 (Table 2, entry 5) was synthesized
by a condensation reaction between an R,ω-hydroxy-PDMS
and 4,4′-azobis(4-cyanopentanoyl chloride) with an excess
of PDMS.116 It was used to prepare block copolymers with
styrene and MMA in a mixture of methyl ethyl ketone/
dichloromethane. The major drawback of such an azo
macroinitiator is the sensitivity of the Si-O-C linkage
toward hydrolysis.78,79
In 1990, Inoue’s team published a study117 on the prepara-
tion of fluoroalkylsilicone-poly(methyl methacrylate) block
copolymers and their PMMA blends using the azo macro-
initiator 6 (Table 2, entry 6). The fluoroalkyl silicone
contained three kinds of fluoroalkyl side chains: 3,3,3-
trifluoropropyl (TFP), tridecafluoro-1,1,2,2-tetrahydrooctyl
(TFO), and heptadecafluoro-1,1,2,2-tetrahydrodecyl (HFD)
groups. The formation of triblock copolymers containing
PMMA and fluoroalkylsilicone was observed. The results
concerning the surface properties of this triblock copolymer
are detailed in the section dedicated to the copolymer
properties (section 5).
In a Japanese patent of Wako Pure Chemicals Industries,118
Shiraki described the synthesis of polyorganosiloxane-
polyoxyalkylene block copolymers azo macroinitiators. For
instance, an azo macroinitiator was manufactured by esteri-
fication of R,ω-bis(polyoxyethylene)-polydimethylsiloxane
with 4,4′-azobis(4-cyanopentanoic acid) (Table 2, entry 7).
Then, this azo macroinitiator was reacted with styrene in
toluene to obtain a multiblock copolymer containing PEO,
PDMS, and PS.
Chemical Reviews, 2010, Vol. 110, No. 3 1245
Kollefrath and co-workers119,120 used azo- and triazene-
modified polysiloxanes (Table 2, entry 8) as macroinitiators
for graft copolymerization with MMA, BuMA, and styrene.
Azo1 and Azo2 showed similar thermal behaviors as AIBN.
Two different radicals are formed: an aryl radical connected
to the polymer backbone which acts as the initiator and a
1,1-dicyanoethyl radical which is known to be very stable
and unable to promote initiation of free radical polymeriza-
tion.130 With Azo1 and Azo2, the polymerizations were
performed at 80 °C, giving high monomer conversion.
Nevertheless, the authors also observed the presence of
unreacted PDMS macroinitiator which could be separated
by precipitation. With Tr1, yields were below 50% and
additional degradations of the PDMS backbone were ob-
served at 95 °C.
To conclude this section, the polymerization initiated by
azo macroinitiators is a convenient way to obtain both graft
and block copolymers. The major drawback of azo macro-
initiators is their low initiation efficiency (30-40%). Hence,
the resulting structures are not well controlled, and their
compositions present various proportions of diblock, triblock,
and multiblock copolymers as well as homoPDMS and
homopoly(vinyl monomer). Azo macroinitiators recently
appeared to be of great interest in controlled radical polym-
erization. For example, they could be coupled with a control
agent such as Co(Acac)2 for controlling the polymerization
of VAc as published by Jerôme et al.,131 a cobalt-porphyrin
complex developed by Wayland et al. to control the
polymerization of acrylates,132–134 an organostibine compound
developed by Yamago et al.135–137 to control a wide variety
of monomers, or molecular iodine I2 to perform reverse
iodine-transfer polymerization (RITP) developed by Lacroix-
Desmazes et al.138–143 (see section 4).
3.2. Polysiloxane Macromonomers
In the field of copolymerization, macromonomers represent
an important class of precursors. They are composed of a
macromolecular chain that bears a polymerizable group at
one chain end. Thanks to this functionality, they are able to
copolymerize with common vinylic monomers to generate
graft copolymers via a “grafting through” mechanism. The
key points of this strategy are the synthesis of the mac-
romonomers and their reactivity.
3.2.1. Synthesis of Polysiloxane Macromonomers and
Copolymerization
Some synthesis pathways to silicone macromonomers
bearing usual polymerizable functions such as (meth)acrylic
and styrenic groups were reported a long time ago. To our
knowledge, the first synthesis of a styrene-based silicone
macromonomer was described in the early 1960s.144 The
reaction of a Grignard reagent (p-vinylphenylmagnesium
chloride) with dimethyldichlorosilane followed by hydrolysis
led to a silane macromonomer. This macromonomer was
condensed with R,ω-dichloropolysiloxane, which enabled
Greber et al. to prepare polysiloxane macromonomers with
various chain lengths (Scheme 19).
The same styrenic chlorosilane was later used by Kawaka-
mi et al.145 to prepare polydimethylsiloxane styrenic mac-
romonomers. Instead of a chain elongation by silane con-
densation, they first polymerized D3 starting from lithium
trimethylsilanolate (LTMS) (Scheme 20). Then, this living
polydimethylsiloxane was end-capped with the styrenic
1246 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 19. Silane and Polysiloxane Macromonomers
Synthesized by Greber et al.144
Scheme 20. Preparation of a Living Polydimethylsiloxane by
Kawakami et al.145
chlorosilane to lead to the suitable polysiloxane macromono-
mer. They also used a methacrylated chlorosilane prepared
by hydrosilylation of I and II by dimethylchlorosilane
(Scheme 21).145 It is worth noting that hydrosilylation yields
were, respectively, 27% and 75%.
The styrenic polysiloxane macromonomer (Mn ) 3000
g · mol-1) was involved in a copolymerization with styrene
to prepare a grafted copolymer.145 The initial proportions
were 15.8 wt % of macromonomer and 84.2 wt % of styrene.
The resulting copolymer (Mn ≈ 40 000 g · mol-1) was grafted
Scheme 21. Synthesis of Macromonomers by Kawakami et al.146
Pouget et al.
Scheme 22. Synthesis of a Methacrylate Chlorosilane for
End Capping
with 15 wt % of siloxane component, indicating a similar
reactivity between the styrenic macromonomer and styrene.
They also copolymerized methacrylate siloxane macromono-
mer, issued from a polysiloxane end capped by I′ (see
Scheme 21) (Mn ) 5200 g · mol-1, 21.1 wt %) with MMA
(78.9 wt %). The resulting copolymer (Mn ) 32 000 g · mol-1)
was grafted with 21 wt % of siloxane, indicating a similar
reactivity between the methacrylate macromonomer and
MMA.
Kawakami et al.146 thus developed one of the first methods
to prepare methacrylic polysiloxane macromonomers (Scheme
21). Both styrenic and methacrylic terminating agents led
to macromonomers with a narrow molecular weight distribu-
tion (Mw/Mn e 1.14) and hydrosilylation yield close to 100%.
It is worth noting that the end capping with A (Scheme 22)
led to a macromonomer with a functionality of 0.38.
According to the authors, this low value should be attributed
to the weakness of the C-O-Si bond, which is believed to
cleave during the synthesis.
Considering the initiation step of D3 by an organolithium
reagent, DeSimone et al.147 pointed out that it is better to
control the first addition step of the reaction in a nonpolar
solvent such as cyclohexane. In this manner, the obtained
siloxanolate-lithium pair is not easily separated. The actual
polymerization begins when a polar solvent, such as THF,
Poly(dimethylsiloxane)-Containing Copolymers
Scheme 23. Termination of D3 Polymerization with
3-Methacryloyloxypropyldimethylchlorosilane
Scheme 24. Pyridyl Derivative Silanolate Anion for the
Initiation of D3 Polymerization
is added. Hence, polymers of high and relatively well
controlled molecular weight are obtained.
Later, 3-methacryloyloxypropyldimethylchlorosilane be-
came commercially available, and it is nowadays preferred
as a terminating agent in ROP (Scheme 23).147
Aoyagi et al.148 prepared novel initiators by starting the
polymerization of D3 with the 2-(4-pyridyl ethyl) or 2-(2,6-
di-tert-butyl-4-pyridyl)ethyl silanolate anion (Scheme 24).
After the polymerization, addition of methacryloxypropy-
ldimethylchlorosilane led to the corresponding functionalized
macromonomers. The copolymerization of these macromono-
Scheme 25. Methacrylate Polysiloxane Macromonomer Synthesis by Cameron et al.149
Scheme 26. Termination of a Silanolate Polymers with Chlorodimethylvinylsilane
Chemical Reviews, 2010, Vol. 110, No. 3 1247
mers with MMA or BMA conferred to the obtained materials
an enhanced gas permeability thanks to the pyridyl groups.
Another way to obtain methacrylic polysiloxane mac-
romonomer is to terminate the polymerization of D3 with
chlorodimethylhydrogenosilane and then to run the hydrosi-
lylation of allyl methacrylate onto this hydride-terminated
polysiloxane (Scheme 25).149 This synthetic pathway is close
to the one developed by Kawakami et al.145,146 except that
hydrosilylation is carried out after termination. Nevertheless,
the authors concluded that this method leads to an unsatis-
factory functionality of the macromonomer (less than unity).
Lastly, vinyl polysiloxane macromonomers were also
synthesized by termination with chlorodimethylvinylsilane
(Scheme 26).150–152 Here, the vinylsilane group was used
afterward as a polymerizable function. For instance, Maynard
et al.152 copolymerized this macromonomer with VAc to
reach a PVAc-g-PDMS copolymer (Mn ) 94 800 g · mol-1)
with an average of three PDMS branches per chain. The final
content of PDMS in the copolymer was less than expected
whatever the initial feed in macromonomer. According to
the authors, this low incorporation is due to a segregation
effect during copolymerization. This conclusion is in agree-
ment with previous observations by Tezuka et al.150
3.2.2. Multifunctional Macromonomers
Different ways to obtain multifunctional polysiloxane
macromonomers were developed depending on the targeted
application. Yu et al.153 modified an amino-terminated
polysiloxane with isocyanatoethyl methacrylate (IEM). The
urea linkage formed in the macromonomer backbone made
it more soluble in polar reactive diluents used in the study
described hereafter. This technique of preparation was not
really attractive until recently, when aminopolysiloxanes with
suitable purity allowed reinvestigating this subject.154 Its
reaction with m-isopropenyl-R,R-dimethyl benzyl isocyanate
1248 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 27. Structures of (a) m-TMI, (b) VDMAz, (c) IDMAz, and (d) IEM and Their Corresponding Macromonomers after
Reaction with an Amino-Terminated Polydimethylsiloxane
(m-TMI), vinyldimethyl-azlactone (VDMAz), isopropenyl
dimethyl azlactone (IDMAz), or isocyanatoethylmethacrylate
(IEM) (Scheme 27) led to the corresponding macromonomers
with styrenic, acrylamide, methacrylamide, or methacrylate
functions, respectively.
In order to obtain cross-linked materials, Tenhu and
Heino155 and later O’Shea and George156,157 copolymerized
R,ω-methacrylate-terminated polysiloxane with styrene. The
macromonomer was obtained by hydrosilylation of allyl
methacrylate on hydride-terminated polydimethylsiloxane.
The first team showed that the degree of cross-linking was
a function of the macromonomer’s content. The second team
showed that cross-linking was only possible as long as the
macromonomer chain length was not too high (3700
g · mol-1) and that homopolymerization of the macromono-
Pouget et al.
mers occurred during the formation of the network. However,
more surprisingly, they also demonstrated that when the
PDMS chain was longer (Mn ≈ 21 000 g · mol-1) neither
cross-linking nor homopolymerization of the macromonomer
was taking place. This result was consistent with the finding
that, on average, only one methacrylate group per mac-
romonomer chain reacted, which was attributed to a shielding
of the reactive group due to the adopted coil conformation
of the macromonomer.
Künzler and Ozark158 also prepared R,ω-methacrylate-
terminated polysiloxane by redistribution of III with D4 and
2,4,6,8-tetramethylcyclotetrasiloxane (D4H) (Scheme 28).
This macromonomer was hydrosilylated with fluorinated
allylic compounds to yield fluorinated siloxane macromono-
mers. Methacrylate functions were used to cross-link the
Poly(dimethylsiloxane)-Containing Copolymers Chemical Reviews, 2010, Vol. 110, No. 3 1249

Scheme 28. Methacrylate Siloxane Involved in a Redistribution with D4 and D4H

Scheme 29. Values of r2 vs Length of the Side Chain for the

material using UV cure during the macromonomer copo- Copolymerization of AN (M2) with Various Methacrylatesa
lymerization with hydrophilic dimethylacrylamide or N-vinyl
pyrrolidone to prepare hydrogels with water content ranging
from 18 to 44 wt % and high oxygen permeability.
Fluorinated groups were introduced in order to get a better
resistance to lipid, but it turned out that the CF2H terminal
group improved the solubility of the siloxane in polar
solvents and in hydrophilic monomers. The properties of the
resulting film are discussed in the final section of this review
(section 5).
Thanks to the initiator method, Suzuki et al. prepared
PDMS bearing both ethynylene and polymerizable function-
alities159 or both alkenyl and methacrylate groups.160,161 The
last one was prepared by polymerization of D3 initiated by
a vinylsilanol activated by BuLi. The reaction of the resulting
a
product with a methacrylic chlorosilane gave the desired M1: MMA, BuMA, or methacrylic polysiloxane macromonomers
macromonomer. After its copolymerization with styrene or PDMS-MA1 500 g mol-1 and PDMS-MA4 1110 g mol-1. Reprinted with
permission from ref 166. Copyright 1966 Elsevier Ltd.
MMA, the alkenyl function enabled the authors to cross-
link the obtained copolymer thanks to hydrosilylation or via
photoactivated thiol-ene reaction. The sequential aspect of copolymerization of styrenic macromonomers (Scheme 19)
this preparation renders it particularly elegant and attractive with styrene reactivity ratios r1 ≈ r2 ≈ 1 using the
for applications like photolithography. Mayo-Lewis method.163 Kawakami et al.145 also estimated
that the reactivity ratios of both styrenic and methacrylic
Another UV-cured cross-linked system with acrylate-
macromonomers were quite the same as those of the
terminated polysiloxane and various reactive diluents (ethyl
corresponding monomers but without measuring them
methylacrylate (EMA), 2-hydroxyethylmethacrylate (HEMA),
accurately.
butyl acrylate (BA), acrylic acid (AA), and 4-vinylpyridine
Cameron and Chisholm149 determined the ratios r1 and r2
(4 VP)) was studied.153 Here, the functionalized polysiloxane
for the copolymerization of methacrylic polysiloxane mac-
acts as the cross-linker. An increase in the reactive diluent
romonomers (M1) with styrene (M2). Despite the use of two
quantity decreased the cross-link density, whereas elongation
different methods to reach these coefficients (Fineman-
at break, Young’s modulus, and ultimate tensile strength
Ross164 and Kelen-Tudos165), experimental errors prevented
increased. Moreover, the reactive diluents were classified
them from any meaningful values for macromonomer ratio
according to the increase of the dynamic and tensile moduli
r1. Nevertheless, a tendency was drawn: the value of r2
they generated (4 VP > HEMA > BA-AA > EMA > BA).
increases (from 1.06 to 1.55) as the molecular weight of
Mazurek et al.154 thoroughly described synthesis, morphol-
the macromonomer increases (from about 500 to 10 000
ogies, and mechanical properties of silicone bearing various
g · mol-1). According to the authors, this behavior is due
terminal functions such as methacrylate, styrene, acrylamide,
to steric effects. In another study,166 under similar condi-
or methacrylamide copolymerized with various acrylate
tions except that the comonomer was acrylonitrile (AN),
monomers by UV cure and led to similar conclusions as Yu
the authors reached the same conclusion as illustrated in
et al.153
Scheme 29.
To sum up the macromonomer issue, it can be seen from DeSimone et al.147,167,168 compared the copolymerization
this overview that silicone macromonomers with various of methacrylic polysiloxane macromonomer with MMA by
functionalities have been available for a long time. It emerges different polymerization processes: free radical, group trans-
that siloxane macromonomers are essentially obtained thanks fer (GTP), and anionic polymerizations. They determined
to D3 ROP to reach monofunctional macromonomers. that the molecular weight distributions were logically nar-
Besides, bifunctional macromonomers obtained by a simple rower in GTP and anionic polymerization than in free radical
one-step method allow preparing cross-linked polymeric polymerization. In addition, they nicely obtained the chemical
materials. Both mono- and bifunctionalized PDMS are more composition distribution (CCD) using supercritical fluid
and more extensively used thanks to their commercial fractionation (SCFE) in supercritical chlorodifluoromethane,
availability.162 which is far less laborious and time consuming than other
techniques such as solvent demixing fractionation169 or
3.2.3. Macromonomers Reactivity in Copolymerization reversed-phase high-performance liquid chromatography.170
In general, the copolymerization parameters of a monomer It appeared that CCD is much narrower in GTP and anionic
and a macromonomer, r1 and r2, are assumed to be close polymerization than in free radical polymerization.
when bearing a similar polymerizable group. This was the Vinylsilane has been previously considered as a radically
conclusion of Greber et al.,144 who determined in the polymerizable monomer, although it rather behaves like a
1250 Chemical Reviews, 2010, Vol. 110, No. 3
transfer agent as discussed later in this review (section 3.3.1).
Nevertheless, it is worth mentioning in this part that
vinylsilane-terminated silicone macromonomers were found
to copolymerize with VAc, to our knowledge the only
monomer depicted in the literature able to copolymerize with
vinyl silane.150–152 Vinylsilanes otherwise do not homopo-
lymerize but only form oligomers, which may be ascribed
to the alpha silyl radical being too stable, i.e., unreactive
toward the propagation step, thus only promoting transfer
and/or termination reactions. Considering macromonomers,
their low reactivity was additionally attributed to a mac-
rophase separation during the polymerization process, since
increasing the macromonomers feed did not generate chains
with a higher graft density.
3.3. Polysiloxane Macrotransfer Agents
3.3.1. Transfer to PDMS
Radicals are known to react onto the methyl of the PDMS
backbone (by hydrogen abstraction) from which polymeri-
zation can occur. Nevertheless, preparing copolymers this
way is extremely challenging since the silylmethyl radicals
thus formed can also recombine to yield a cross-linked
material. Therefore, the resulting architectures are typically
not well-defined copolymers but either grafted cross-linked
silicone, when using common bulk and solution processes,
or core-shell particles, when using a seeded emulsion
polymerization process.
Two main radical generator families are generally quoted
in the literature. “Nonvinyl-specific radical generators”
produce radicals onto a PDMS backbone, without the need
for any reactive functions, to cross-link the materials. The
most famous nonvinyl-specific vulcanizing agent is benzoyl
peroxide. “Vinyl-specific radical generators”, such as di-tert-
butyl peroxide, normally only react on PDMS functionalized
with vinyl functions. Depending on the authors, this latter
peroxide can proceed in different ways: either it is unable
to generate a radical on a permethyl PDMS backbone171,172
or it generates a silylmethyl radical on a permethyl PDMS
backbone, which however very quickly recombines with
radicals derived from di-tert-butyl peroxide decomposi-
tion.173–175 Whatever the type of radical generator used,
PDMS cross-linking is induced by the reaction of radicals
generated on the backbone of the silicone polymer, via a
vinyl group or a methyl group, yielding interchain links,
respectively composed of 3 or 2 methylene groups.
Apart from this vulcanization process, it was shown
recently that it is possible to use these generated radicals
for copolymer synthesis thanks to a “grafting from” method.
Vinylic siloxanes, cyclosiloxanes, or polysiloxanes exposed
to a radical generator generally only lead to the formation
of a dimer or more hardly to a trimer; we thus considered
these vinyl-functionalized molecules as transfer agents rather
than (co)polymerizable entities as often quoted in the
literature. Some reports dealt with the radical copolymeri-
zation of vinylic monomers in the presence of vinyl-
functionalized polysiloxane, but none of them focused on
the characterization of the resulting copolymers. For example,
toughened thermoplastics such as PMMA or PS were
synthesized by polymerization initiated by radicals generated
on the vinyl functions of PDMS.176 Dong et al.177 prepared
PS-modified silicone elastomers in a similar manner. In a
first step, a vinyl containing hydroxyl-terminated polysilox-
ane was first reacted with styrene and benzoyl peroxide, and
Pouget et al.
in a second step, cross-linking of the PDMS phase proceeded
thanks to a trialkoxysilane. Using extraction techniques, the
authors showed experimentally that soluble fractions were
less important than the theoretical ones. They noticed that
either formation of PDMS-g-PS during the first step of
styrene polymerization could occur or PS could be so
entangled in the PDMS network that it could not be extracted.
Analyses only focused on the final cured blends, so that no
information is available on the copolymer structures formed
during the first stage of the protocol. Indeed, whereas
abstraction of hydrogen on vinyl-PDMS is possible, it has
not been extensively studied in bulk or in solvent media. As
in the case of transfer to nonvinylic polysiloxane, this may
be due to the difficulty encountered to control the resulting
material architecture.
A series of patents from the same assignee described first
the polymerization from a radical generated on a PDMS
backbone and the generation of cross-linked materials from
hydroxyl groups of the polydimethylsiloxane. Actually, as
described in the examples, hydroxy-terminated PDMS was
mixed with di-tert-butyl peroxide and various monomers such
as styrene, butyl acrylate, or acrylonitrile. Afterward, the
mixture was heated to lead to the corresponding grafted
copolymers prior to cross-linking thanks to trichlorosilane
or tri- or tetra-alkoxysilane and so forth.178,179 This was done
in solvent media180 to obtain coatings by casting, in bulk in
a reactor,181,182 or in extruders183 to give materials. A
surprising fact is the use of di-tert-butyl peroxide, a “vinyl-
specific radical generator”, which is claimed to be preferred
to benzoyl peroxide; the latter however falls in the scope of
the invention. Enhanced mechanical properties such as tensile
strength, elongation, and hardness were claimed, and coatings
were more abrasion resistant. Thanks to this method,
thixotropic polymers such as poly(dimethylsiloxane)-g-
poly(methacrylic acid) were also obtained.184,185
To sum up this part, transfer to PDMS was not an intensive
field of research following the early studies on that topic in
the 1970s, undoubtedly because of the difficulty to control
polymerization reactions especially with nonfunctionalized
PDMS. This process however remains attractive in the latex
field thanks to the ease of process and its versatility as will
be described in the following section.
3.3.2. Silicone Containing Core-Shell Particles via
Radical Polymerization
Core-shell latex particles resulting from seeded emulsion
polymerization are of high interest given their numerous
applications such as coating modification, rubber strengthen-
ing, or thermoplastic and thermosetting polymer toughening.
This last application generally involves a thermoplastic shell
and an elastomeric core, the latter explaining the interest for
the silicone field. In direct emulsion, it is thermodynamically
favored that the most hydrophobic polymer, in most cases
PDMS, will form the core of the structure. However, some
strategies have been developed to obtain core-shell struc-
tures where PDMS surrounds the particles. Note that most
systems involve a polymerization of the shell around the core
without any chemical links between them. Some of them
will be considered as reference samples, but formation of a
shell linked to the core through radical reactions is the
purpose of this section.
We first discuss the formation of copolymers where
unfunctionalized PDMS is used as the seed. Okaniwa and
Ohta186,187 prepared PDMS-g-PS and PDMS-g-PAN copoly-
Poly(dimethylsiloxane)-Containing Copolymers
mers by forming a radical by hydrogen abstraction from the
silylmethyl group of the silicone polymer in emulsion. The
efficiency of the grafting reaction was correlated to the
initiator properties: its ability to generate an oxyradical, its
oil solubility, and the absence of potentially abstractable
protons in its structure which may consume the radical.
Among the studied initiators, tert-butyl perlaurate reduced
by FeSO4 led to the best grafting ratio. In addition,
Okaniwa188 also investigated core-shell materials using
PDMS/polybutadiene as the core and poly(styrene-co-acry-
lonitrile) as the shell. Polybutadiene is widely used for its
elastomeric properties in the core-shell field, and its grafting
with thermoplastic polymers is relatively easy; its major
drawback arises from its sensitivity to oxidation. Okaniwa
overcame this disadvantage from the association of PB with
PDMS which is especially oxidation resistant. The author
showed that the modification of the PDMS with PB remained
efficient (still using tert-butyl perlaurate/FeSO4 redox initia-
tor). Thanks to such PB/PDMS seeds, Okaniwa could graft
SAN more easily than on pure PDMS seeds. As already
mentioned, initiation of polymerization from unfunctionalized
silicone polymers is scarcely used because of the challenging
control of the reaction.
To the best of our knowledge, the first introduction of
vinyl-containing PDMS seed copolymerized with acrylic or
styrenic monomers appeared in patents189–191 which claimed
improved thermal or mechanical properties. An increase of
the vinyl content in the seed led to enhanced elongation at
break and tensile strength of the resulting materials. He
et al.192,193 studied more deeply core-shell particles with
vinyl-PDMS as the starting material (resulting from the
copolymerization of D4 and 2,4,6,8-tetramethyl-2,4,6,8-
tetravinylcyclotetrasiloxane (D4V)). Actually, they compared
two strategies in order to obtain core-shell materials. The
vinyl-functionalized PDMS was either cross-linked before
its use as a seed for the polymerization of the shell or directly
used as the seed. In this latter case, cross-linking and
polymerization occurred simultaneously. The authors ob-
served that the formation of a core-shell structure was
influenced by (i) the reactivity ratios of the monomers and
polysiloxane, (ii) the mobility of the species in the emulsion,
(iii) the hydrophilicity of the monomers (the authors used a
hydrosoluble initiator), and (iv) the compatibility of the vinyl
monomers and their polymers with PDMS. For instance,
using an emulsion polymerization with linear vinyl PDMS
as the seed, they were able to prepare core-shell particles
with BMA and MMA but not with styrene. By contrast,
core-shell particles were obtained with cross-linked PDMS
(i.e., in the absence of vinyl functions) with styrene but not
with MMA. In a more complete study,194 the authors
demonstrated that since styrene and poly(styrene) were
compatible enough with the PDMS seed, they were able to
penetrate the seed as long as PDMS was not cross-linked
and consequently formed a nonsegregated copolymer latex.
When PDMS is cross-linked, it acts as a shield and yields a
PDMS-PS core-shell structure. On the contrary, MMA and
PMMA are much less compatible with PDMS. Consequently,
when the seed consisted of cross-linked PDMS, MMA was
rejected and homopolymerized away from PDMS (Scheme
30).
Kong and Ruckenstein195 prepared PDMS-poly(St-MMA-
AA) core-shell particles thanks to a similar seeded emulsion
polymerization process. Five weight percent of D4V was
added during the preparation of the PDMS seed by anionic
Chemical Reviews, 2010, Vol. 110, No. 3 1251
Scheme 30. Morphologies of Core-Shell Latexes Depending
on the Nature of the PDMS Seed and the Miscibility of the
Comonomers with PDMS
or cationic polymerization of D4. Using transmission electron
microscopy (TEM), they observed that core-shell particles
exhibited a more uniform particle size distribution than
without D4V. 3-(Trimethoxysilyl)propyl methacrylate was
also added to D4 during the polymerization to cross-link the
seed whose diameter decreased. Unfortunately, its effect on
the properties of the final material was not extensively
discussed. Analogous PDMS-PMMA or PDMS-PBuA
core-shell particles using vinyl polysiloxane were obtained
by Dai et al.,196 where polymerization was initiated by 60Co
γ-ray irradiation. Films formed from these materials showed
higher decomposition temperatures, enhanced pendulum
hardnesses, lower water absorptions, but decreased tensile
strengths as the PDMS concentration increased. Zou et
al.197,198 observed that the particle size greatly influences the
film properties such as mechanical performance, water
absorption, and transparency. For instance, increasing the
vinyl content from 1.1 × 10-3 to 2.2 × 10-3 mol per gram
of seed emulsion led to a decrease of particle size and a
20% enhancement of the tensile strength of a PDMS-
P(MMA-co-BA) film. On the other hand, the evolution of
the elongation at break was negligible. Lin et al.199 also
prepared PDMS-poly((meth)acrylate) (MMA and BuA)
core-shell particles, but they observed much more complex
morphologies when 10 wt % of D4V were added to D4 during
ROP of the PDMS seed. Actually, this led to a multiglobular
shell surrounding a shapeless PDMS core. As demonstrated
by Soxhlet extraction of the particles, the structure was
indebted to cross-linking occurring by copolymerization of
acrylate monomers with vinyl groups. Contact angle mea-
surements showed the more hydrophobic character of the
film resulting from the D4V-containing seeds, giving evidence
for a partial shell formation.
In order to obtain a core-shell structure with PDMS as
the shell, Kong et al.200 prepared first a cross-linked PMMA
seed using ethylene glycol dimethacrylate (EGDMA) and
then a D4V/ammonium persulfate mixture was added at the
end of the MMA polymerization. D4V vinyl functions were
supposed to react with residual MMA of the PMMA seed
in formation. Afterward, the shell was obtained by polymer-
izing D4. The authors compared a film made from the
obtained latex with one formed from a similar core-shell
latex which did not involve D4V. TEM observations showed
that PDMS containing vinyl groups formed a continuous
phase in the first film (in presence of D4V). Consequently,
1252 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 31. Chemical Initiation Mechanism of
Telomerizationa
a
YZ is the telogen, and M is an unsaturated monomer that can react by
radical polymerization.
its hardness was lower than the reference sample without
D4V. It should be noticed that 3-methacryloxypropyl tri-
methoxysilane (MATS) was also used to prepare such
core-shell architectures. This molecule can copolymerize
with MMA201,202 or styrene203 when added in the last
moments of the seed polymerization. Afterward, methox-
ysilane groups reacted with D4 to give the core-shell
structure.
To summarize this core-shell section, radical polymeri-
zation allowed researchers to obtain small and fine particles
with well-defined morphologies, which is a key issue for
the properties of the resulting materials.
Table 3. Telogens and Monomers Used To Make Block and Graft PDMS-Based Copolymers by Telomerization
Pouget et al.
3.3.3. Chemical-Initiated Telomerization
The telomerization process consists of reacting, under
polymerization conditions, a molecule YZ, which is called
a telogen, with more than one unit of a compound M
containing an unsaturated group, called a taxogen, to form
products called telomers Y(M)nZ.204,205 The general mech-
anism of telomerization by chemical initiation is given in
Scheme 31. By functionalizing a PDMS chain, it is possible
to transform it into a macrotelogen.
Telomerization differs from polymerization in the follow-
ing issues: the fragments of the initiator mainly induce the
rupture of the telogen, whereas in polymerization they add
onto the monomer; the number of M units in the final product
is larger than 1 but lower than 100; the functional groups
on both chain ends can be easily transformed thanks to the
low molecular weights of the resulting polymers.
Several authors carried out the synthesis of block or graft
copolymers using a PDMS macrotelogen, as summarized in
Table 3. They all carry a thiol function except the one used
by Tezuka that bears a Si-H functionality (Table 3, entry
5).206
The first synthesis of PDMS graft copolymers using
telomerization has been carried out by the Dow Corning
Corp.207 in 1969 using two different emulsion pathways. In
the first step, siloxanes 1 (Table 3, entry 1) were prepared
by emulsion copolymerization of the corresponding tri-
methoxysilane as described by Hyde208 and Findlay.209 The
second step is the addition of the vinyl monomer and
ammonium persulfate to the nitrogen-purged emulsion to
perform the polymerization by heating. In a second approach,
the polysiloxane that has been previously prepared by bulk
Poly(dimethylsiloxane)-Containing Copolymers
or solution techniques is emulsified, the second step being
identical to the one described above.
Falender et al.22,23,210,211 from the Dow Corning Corp. also
performed the telomerization of many different monomers
(acrylic, methacrylic, styrenic monomers) using a PDMS
macrotelogen (Table 3, entry 2). For example, they carried
out the synthesis of PDMS-g-poly(styrene-co-acrylonitrile)
of high impact strength using a mercaptopropyl PDMS as a
telogen in water suspension or in bulk with benzoyl peroxide
as radical initiator.
Saam and Tsai212 carried out the dispersion polymerization
of methyl methacrylate in aliphatic hydrocarbons containing
poly(dimethylsiloxane) modified with mercaptoalkyl side
groups (Table 3, entry 2). They observed the formation of
particles stabilized by a protective layer of solvated poly-
(dimethylsiloxane). The evolution of particle size was studied
while polymerizing MMA. In the very early stage of
polymerization, a large number of very small particles of
broad size distribution was formed. These particles contained
a relatively high amount of silicone stabilizer in the form of
graft copolymers. As the conversion increased, the particles
tended to agglomerate to give some larger particles and a
broad size distribution. Between 5% and 10% conversion, a
critical point was reached: the smaller particles were absorbed
by the larger particles present in the system. From this point,
new particles were no longer formed (end of nucleation) and
polymerization occurred within the swollen monomer par-
ticles. This was evidenced by a steady particle growth and
a final narrow size distribution.
Our group also performed the synthesis of triblock co-
oligomers of poly(methyl methacrylate)-b-PDMS-b-poly(methyl
methacrylate)213 by telomerization. The first step consisted
of the synthesis of a macromonomer containing a PMMA
chain using γ-mercaptopropylmethyldimethoxysilane as te-
logen (Table 3, entry 3), MMA as monomer, and AIBN as
initiator. The second step was the condensation reaction
between the macromonomer and the silanol-terminated PDMS
at 100 °C for 20 h in the presence of a catalyst (2-ethylhexanoic
acid/tetramethylguanidine complex). The presence of small
amounts of diadduct (reaction of the second methoxy silyl group
with a silanol-terminated PDMS) has been observed.
More recently, Fawcett et al.214 carried out the telomeriza-
tion of different monomers (styrene, MMA, chloroprene)
(Table 3, entry 4) using mercaptopropyl-functionalized
PDMS and AIBN as radical initiator. The authors observed
that the thiol/monomer ratio controls the reaction type: for a
low monomer/macrotelogen ratio, only one monomer unit
reacted with the SH group (thiol-ene reaction), whereas a
higher concentration ratio resulted in graft and block copolymers.
Thermal-initiated telomerization was also carried out in
scCO2 to prepare PMMA particles.215 Okubo and co-workers
used a triblock copolymer PMMA-b-PDMS-b-PMMA as
colloidal stabilizer, prepared in situ by telomerization using
a mercaptopropyl telechelic PDMS as telogen with AIBN
as initiator (Table 3, entry 4). The particle diameter could
be controlled by the concentration of the telogen which serves
as steric stabilizer, with sizes ranging from submicrometers
to micrometers. A very small consumption of telogen was
observed, albeit producing sufficient copolymers to stabilize
the PMMA particles.
Tezuka et al.206 used telogen 5 (Table 3, entry 5) to prepare
poly(vinyl alcohol)-b-PDMS-b-poly(vinyl alcohol) (PVA-
b-PDMS-b-PVA) block copolymers. The first step was the
telomerization of VAc in the presence of dimethylchlorosi-
Chemical Reviews, 2010, Vol. 110, No. 3 1253
Scheme 32. General Scheme of Reversible Activation-
Deactivation in Controlled Radical Polymerization217
lane as telogen and AIBN as radical initiator in benzene.
The second step was the coupling reaction of the chlorosilane-
terminated poly(vinyl acetate) (PVAc) prepolymer with a R,ω-
telechelic silanolate-terminated PDMS. Finally, PVA-b-PDMS-
b-PVA block copolymers were obtained by saponification of
the PVAc-b-PDMS-b-PVAc block copolymer using a 10%
aqueous sodium hydroxide solution in methanol. However, due
to the sensitivity of PDMS to basic conditions, a loss of more
than 15 wt % of siloxane was observed during this last
hydrolysis step.
The literature already largely described telomerization of
a wide range of monomers.204,205 To date, mainly thiol-type
PDMS were used as telogens. Using PDMS as a macrote-
logen is an easy way to prepare silicone-containing copoly-
mers, keeping in mind that the final product does not exhibit
high molecular weights. Moreover, chemical-initiated te-
lomerization is compatible with dispersed aqueous or scCO2
media (suspension, emulsion, dispersion polymerizations).
4. Copolymers Obtained by Controlled Radical
Polymerization
The controlled radical polymerization (CRP)216 process
includes a group of radical polymerization techniques that
have attracted much attention over the past decade since they
provide simple and robust routes to the synthesis of well-
defined polymers, low-dispersity polymers, and preparation
of novel functional materials. The general principle of the
reported methods relies on a reversible activation-deactivation
process between dormant chains (or capped chains) and
active chains (or propagating radicals) with different rate
constants kact and kdeact, respectively (Scheme 32).217,218
The past few years have witnessed rapid growth in the
development and understanding of new CRP methods. The
most popular CRP methods treated in the following text are
iniferters,219 nitroxide-mediated polymerization (NMP)220 that
requires alkoxyamines, atom-transfer radical polymerization
(ATRP)221 involving alkyl halides, metallic salts, and ligands,
iodine-transfer polymerization (ITP) using alkyl iodides,141
and reversible addition-fragmentation chain transfer po-
lymerization (RAFT)222 using dithiocarbonyl derivatives.
4.1. Iniferter
Iniferters219 are specific agents that proceed in a radical
polymerization via initiation, propagation, primary radical
termination, and transfer to initiator (Scheme 33). Because
bimolecular termination and other transfer reactions no longer
dominate the polymerization, these polymerizations are
performed by insertion of the monomer molecules into the
iniferter bond, thus generating polymer chains with two
initiator fragments at the chain ends. These thermal iniferters
give controlled molecular weights but do not present a living
chain end (i.e., no possibility of further chain extension).
PDMS-based macroiniferters are given in Table 4 (entries 1
and 2).
1254 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 33. Main Steps Involved in the Mechanism of Iniferter Polymerization
Table 4. PDMS Thermal- and Photo- Macroiniferters
Macroiniferter 1 (Table 4, entry 1) was synthesized by a
hydrosilylation reaction between Si-H-terminated PDMS
and allyl-N-methylamine. Subsequent chain extension of the
macrodiamine by the thiocarbamylation reaction using CS2
and oxidative coupling using I2 led to the formation of the
polymeric iniferters (Scheme 34).
Macroiniferter 1, when used under thermal initiation in
the presence of vinylic monomers, gives a triblock copolymer
with a central vinylic polymer block and a PDMS segment
Pouget et al.
at both ends. This macroiniferter was used for styrene and
MMA polymerization between 60 and 100 °C in bulk or in
toluene solution.223,224 When the polymerization of styrene
or MMA was carried out in concentrated solutions or in bulk,
a slight microsegregation was observed due to the incompat-
ibility of the PDMS segment with the propagating polymer
chain. This polymer demixtion led to a decrease of the
macroiniferter efficiency.223 Moreover, the authors observed
a decrease in chain transfer with increasing PDMS chain
Poly(dimethylsiloxane)-Containing Copolymers
Scheme 34. Synthesis of Macroiniferter 1 (Table 4, entry 1)
length (Mn ) 2800-4200 g · mol-1), which also suggested
the occurrence of some phase segregation.224 The azeotropic
temperatures of the iniferter, where ktr ) kp or Ctr ) ktr/kp )
1, were determined to be 125 °C for MMA and 110 °C for
styrene. When Ctr ) 1, the molecular weight remains constant
during the polymerization. When the temperature is lower
than the azeotropic temperature, the polymerization is not
controlled and Mn decreases as the conversion proceeds
(i.e. Ctr < 1). The authors found that for this macroiniferter
Ctr(styrene) ) 1.5 × Ctr(MMA).
Macroiniferter 2 (Table 4, entry 2) is the equivalent of
macroiniferter 1 but starting from a difunctionnal PDMS
chain. When a vinylic monomer is polymerized in the
presence of this particular transfer agent, a multiblock
copolymer is formed. In the polymerization of MMA, Nair
et al.228 showed that when the concentration of the macro-
iniferter increased, a retardation effect due to the participation
of thiuramyl radicals in termination reactions was observed.
Scheme 35. Main Steps Involved in the Mechanism of Photoiniferter Polymerization
Chemical Reviews, 2010, Vol. 110, No. 3 1255
The molecular weight increased linearly with conversion,
and the final PDIs of the multiblocks were around 2-2.5.
These authors thus reported the successful synthesis of PS-
PDMS and PMMA-PDMS multiblock copolymers, whereas
when polymerizing acrylamide to obtain amphiphilic co-
polymers, a strong phase segregation occurred and led to an
insoluble material due to some physical cross-linking.
Photoiniferters are iniferters that are activated under UV
radiation. These iniferters have the advantage that they lead
to living chains. Indeed, the polymer chain resulting from
termination can be reactivated by cleaving the C-S bond
under UV radiation (Scheme 35). The different PDMS-based
macrophotoiniferters reported here are quoted in Table 4
(entries 3-5).
The first study on photoinitiated polymerization was
carried out by Inoue et al.43 They modified the surface of a
cross-linked poly(dimethylsiloxane-co-methyl chlorometh-
ylsiloxane) by photopolymerizing hydrophilic monomers
1256 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 36. Synthesis of the PDMS Photoiniferter and
Grafted Polymer Obtained
Scheme 37. General Mechanism of Nitroxide-Mediated
Polymerization
(Table 4, entry 3) using diethyldithiocarbamated PDMS as
photoiniferter. The synthesis of this photoiniferter and the
grafted polymer obtained is described in Scheme 36. By this
method, they successfully prepared hydrophilic surfaces
which could find applications in biomedical devices.
Macrophotoiniferter 4 (Table 4, entry 4) was successfully
used for the controlled polymerization of MMA and copo-
lymerization of methyl acrylate and acrylic acid under UV
radiation.230 Moreover, the living character of the chain ends
was demonstrated by further growing two blocks of poly(2-
hydroxyethylacrylate) (PHEA) from the PMMA-b-PDMS-
b-PMMA triblock copolymer. Finally, VAc and 1,1-
dihydroperfluorooctyl acrylate (FOA) were successfully
polymerized by macrophotoiniferter 5 (Table 4, entry 5)
under UV radiation.232 In the case of VAc, diblock copoly-
mers with PDI ) 1.4 (at 30% monomer conversion) were
obtained.
Until now, PDMS-based macroiniferters were used for the
controlled polymerization of styrene, MMA, methyl acrylate
(MeA), acrylamide, AA, MAA, HEMA, HEA, AMPS,
DMAEMA, sodium styrene sulfonate (NaSS), N-vinyl pyr-
rolidone (NVP), and FOA. Iniferters allow the controlled
polymerization of a wide range of monomers,219 but the
control over the molecular weights is poorer than by other
PRC methods and photoactivation is required to achieve
living character.
4.2. NMP: Nitroxide-Mediated Polymerization
In nitroxide-mediated polymerization (Scheme 37),220 the
reversible termination is the key step to keep a low
concentration of propagating radicals in the reaction medium.
All polymer chains should only be initiated by the
alkoxyamine, leading to well-controlled polymer architec-
tures. The nature of the counter radical is essential for the
controlled character of the polymerization.
The alkoxyamine has a rather labile C-O bond which
cleaves homolytically upon heating to liberate the radical
on the polymer chain, which propagates, and the counter
radical, which is only able to end cap the propagating chains.
The equilibrium is overwhelmingly shifted to the left
(dormant chains), and the control is ensured thanks to the
Pouget et al.
persistent radical effect (i.e., accumulation of nitroxides).233,234
Table 5 summarizes the different precursors used in the
synthesis of PDMS-based copolymers by NMP techniques.
Styrene has been the most widely studied monomer for
nitroxide-mediated polymerization from a PDMS precursor. As
for conventional free radical polymerization, azo macroinitiators
can be used to initiate the polymerization. Yoshida et al.235,236
used a bicomponent system consisting of a multiblock azo
macroinitiator 1 (Table 5, entry 1) (Mn ) 30 800 g · mol-1
with PDMS segments Mn ) 5000 g · mol-1) and 4-methoxy-
2,2,6,6-tetramethylpiperidine-1-oxyl (MTEMPO) (Scheme
38). As soon as the azo macroinitiator decomposed, the
formed radicals were trapped by MTEMPO to create the
active macroalkoxyamine in situ. Styrene was polymerized
in bulk at 130 °C for 72 h. Due to the low initiator efficiency,
the resulting copolymer exhibited an AB diblock structure
rather than the expected ABA triblock one. A chain extension
with p-methoxystyrene showed the living character of the
polymer chain end.
Macroinitiator 1 (Table 5, entry 1) was also used by Ryan
et al.237 in the dispersion polymerization of styrene in scCO2
using SG1 (N-tert-butyl-N-(1-diethylphosphono-2,2-dimeth-
ylpropyl)nitroxide) (Scheme 38) as a control agent at 110 °C.
A rather broad bimodal molecular weight distribution (MWD)
was obtained. The use of AIBN as co-initiator and the reduction
of the number of azo moieties in the azo macroinitiator tended
to reduce the bimodality while introducing sufficient siloxane
units to stabilize the dispersion in scCO2. The azo moieties were
partially consumed by heating the macroazo initiator before
adding the monomer to reach a new precursor with Mn) 21 800
g · mol-1 and only 25% remaining azo moieties. A chain
extension showed the high livingness of the polymer chains.
The broad MWD was ascribed to polysiloxane chains bearing
SG1 moieties at one or two chain ends, leading to di- or triblock
copolymers, respectively.
When a bicomponent system is used (azo macroinitiator
+ nitroxide), the low initiator efficiency leads to an ill-
defined architecture and a mixture of homopolymers and
diblock and triblock copolymers.220 To obtain a well-defined
di- or triblock copolymer, it is better to use a monocomponent
system, assuring a better controlled architecture. Therefore,
several organolithium alkoxyamines were used as initiators
for the anionic polymerization of D3, leading to PDMS
alkoxyamines. These macroalkoxyamines were further used
for the nitroxide-mediated polymerization of styrene, leading
to well-defined diblock copolymers. This method was first
applied by using macroalkoxyamine 2 (Table 5, entry 2 and
Scheme 39) for the bulk polymerization of styrene238 at 120
°C. The controlled polymerization proceeded up to 20%
conversion, giving the desired diblock copolymer. However,
above 20% styrene conversion, SEC analysis showed a
bimodality, indicating the formation of homopoly(styrene)
along with low molar mass diblock copolymer.
Macroalkoxyamine 3 (Table 5, entry 3) was used to poly-
merize styrene in bulk at 120 °C.239 The macroalkoxyamine
reacted quantitatively to form diblock copolymers, and a
linear evolution of molecular weight with conversion proved
the controlled character of the polymerization. A final PDI
of 1.3 above 25% conversion was obtained. However,
maximum monomer conversion was limited to 42%.
Macroalkoxyamine 4 (Table 5, entry 4) was used for bulk
and solution polymerizations of styrene240 at 120 °C where
some discrepancies between the molecular weights deter-
mined by SEC and NMR were observed. Moreover, the
Poly(dimethylsiloxane)-Containing Copolymers
Table 5. PDMS Macroalkoxyamines and PDMS Macroinitiators Used in Nitroxide-Mediated Polymerization
Scheme 38. Formulae of TEMPO, MTEMPO, and SG1
resulting copolymer did not have the expected composition
from the monomer feed.
Macroalkoxyamine 5 (Table 5, entry 5) was synthesized by
hydrosilylation,241,242 thus avoiding the presence of the rather
fragile Si-O-C bond.78,79 This macroalkoxyamine was used
Scheme 39. Synthesis of Macroalkoxyamine 2 (Table 5, entry 2)
Chemical Reviews, 2010, Vol. 110, No. 3 1257
for the solution polymerization of styrene,241 leading to a
mixture of homopoly(dimethylsiloxane), homopoly(styrene),
and diblock copolymer. The presence of homoPDMS proved
the incomplete reaction of the macroinitiator, while homoPS
resulted from some thermal autoinitiation. However, by
optimizing the reaction conditions, the amount of homopoly-
mers was reduced, leading to a practically pure diblock
copolymer.242 Nevertheless, some inhomogeneity in SEC
could be attributed to diblocks with varying PS and PDMS
content. Moreover, NMR analysis only showed around 30%
of remaining TEMPO (2,2,6,6-tetramethylpiperidinyl-1-oxy)
units on the chain ends. The difference was attributed to some
hydrogen transfer,220,244 leading to -CHdCH-Ph and
-CH2-CH2-Ph chain ends.
1258 Chemical Reviews, 2010, Vol. 110, No. 3
Because TEMPO has no ability to control the radical
polymerization of vinyl monomers other than styrene and
styrene derivatives, Miura et al.243 used macroalkoxyamine
6 (Table 5, entry 6) to polymerize styrene and other
monomers. The TIPNO (N-tert-butyl 1-phenyl-2-methylpro-
pylaminoxyl)-based alkoxyamine has been reported to act
as an excellent mediator for a wide range of vinyl monomers
including styrene. Styrene was polymerized in bulk at 120
°C. A linear evolution of the molecular weight with conver-
sion was observed, and SEC showed a monomodal MWD
with a PDI of 1.2, a value lower than those obtained at similar
conversion with TEMPO-bearing control agents.238,239 TIP-
NO, like SG1, is a better control agent than the commercially
available TEMPO. The TIPNO-based alkoxyamine 6 (Mn
) 4000 g · mol-1, PDI ) 1.09) was further used for the
polymerization of methyl acrylate (MeA) in bulk at 120
°C.243 A linear evolution of molecular weights with conver-
sion was observed with a final PDI value below 1.2 at 50%
conversion. A good correlation between experimental and
theoretical molecular weights was observed. Moreover, to
show the living character, the PDMS-b-poly(MeA) diblock
copolymer was further used as a macroalkoxyamine in the
polymerization of styrene, leading to an ABC triblock
copolymer with PDI ) 1.4. Alkoxyamine 6 was also used
for the controlled polymerization of isoprene (IP) in bulk at
120 °C.243 The final diblock copolymer presented a PDI of
1.15 and a good agreement was noticed between experi-
mental and theoretical molecular weights. The diblock
copolymer was further used in the polymerization of styrene
to synthesize a PDMS-b-poly(IP)-b-PS triblock copolymer
with a monomodal MWD (PDI ) 1.37).
To date, NMP from PDMS macroalkoxyamines was
essentially used for styrene polymerization. Only Miura
et al.243 polymerized methyl acrylate and isoprene from a
PDMS macroalkoxyamine. It is known that nitroxide-
mediated polymerization allows the controlled polymeriza-
tion of a wide range of monomers.220 Among others, the
controlled polymerization of acrylates (such as tert-butyl
acrylate (tBuA), nBuA, or MeA), styrenics (such as styrene,
p-bromostyrene, p-chlorostyrene, p-epoxystyrene, p-meth-
oxystyrene, p-acetoxystyrene, p-chloromethylstyrene, p-tert-
butoxy styrene, or styrene p-sulfonic acid sodium salt),
isoprene, 4-vinyl pyridine, acrylonitrile, maleic anhydride,
isopropylacrylamide, and N,N-dimethylacrylamide could be
considered in the future. Indeed, all these monomers can be
potentially polymerized using nitroxide-mediated polymer-
ization from a poly(dimethylsiloxane) macroalkoxyamine,
giving a wide range of different properties to the second
block, which could lead to a variety of new applications.
Furthermore, NMP of PDMS macromonomers has not been
reported yet to yield graft copolymers. This is probably due
to the limited range of commercially available PDMS
macromonomers, since those based on methacrylates could
not be easily controlled by NMP. Recent progress in NMP
of methacrylates now makes it possible.245,246
4.3. ATRP: Atom-Transfer Radical Polymerization
4.3.1. Main Features
The mechanism of ATRP221 is described in Scheme 40.
The control is achieved by an equilibrium between active
(or propagating) and dormant species. This equilibrium is
defined by the rate constants kact and kdeact. The lower the
concentration of active species (kact , kdeact), the better the
Pouget et al.
Scheme 40. Mechanism of Atom-Transfer Radical
Polymerization
control and the lower the extent of irreversible termination.
Irreversible termination occurs as in conventional free radical
polymerization by combination or disproportionation.
Several reviews from Matyjaszewski247–251 treat of the
synthesis of organic-inorganic hybrid materials using atom-
transfer radical polymerization. They all deal with the general
synthesis routes used to obtain such hybrid materials, but
none of them gives an extensive overview of the existing
state of the art. Note that a patent from Matyjaszewski et al.
covers the synthesis of poly(dimethylsiloxane)-containing
block and graft copolymers by ATRP.252
4.3.2. Block Copolymers
Various PDMS macroinitiators have been used for the
synthesis of di- and triblock copolymers (Table 6). Most of them
were prepared by hydrosilylation, even though several other
routes have been explored. The most widespread macroinitiators
bear a chlorobenzyl or bromo-isobutyrate end group.
In ATRP, ligands are used to favor the homolytic scission
of the carbon-halogen bond. The formula of the most com-
monly used ligands are given in Scheme 41. Macroinitiator 1
(Table 6, entry 1) (Mn ) 1800-10 000 g · mol-1) was used by
Peng et al. for the controlled polymerization of styrene253,254
and butyl methacrylate.255 PS-b-PDMS-b-PS triblock copoly-
mers were obtained using the CuCl/dNbpy (4,4′-di(5-nonyl-
2,2′-bipyridine)) catalyst system (Scheme 41) in phenyl ether
at 130 °C. The molecular weight increased with conversion,
and the final polydispersity index was 1.33. The influence
of the polysiloxane chain length was also investigated.253
When using higher molecular weight polysiloxane chains
(1800-10 000 g · mol-1), styrene conversion and the rate of
polymerization diminished whereas the polydispersity index
increased. When the catalyst concentration was increased,
the rate of polymerization, the final monomer conversion,
and the polydispersity index also increased.
Macroinitiators 2 and 3 (Table 6, entries 2 and 3) were
prepared in similar ways, macroinitiator 2 from a mono-
functional PDMS and macroinitiator 3 from a difunctional
PDMS, opening the way to di- and triblock copolymers,
respectively. Zhang et al.257 used both macroinitiators (8000
g · mol-1) to synthesize AB and ABA block copolymers in
bulk at 120 °C with NiBr2(PPh3)2 as catalyst. Successfully
polymerized monomers included MMA, n-BuA, t-BuA,
trimethylsilyl methacrylate (TMSMA), trimethylsilyl acrylate
(TMSA), and trimethylsiloxy-ethyl acrylate (TMSEA). t-
BuMA and 1-ethoxyethyl methacrylate (EEMA) were also
polymerized, but a deactivation of the catalyst led to a low
monomer conversion (<20%). The polymerization of t-BuA,
which exhibits a higher reactivity than t-BuMA, continued
up to high monomer conversions. Deprotection of the
hydroxyl group in the case of poly(trimethylsiloxyl-ethyl
acrylate) and of carboxylic acid in the case of poly(trimeth-
ylsilyl methacrylate) and poly(trimethylsilyl acrylate) was
carried out in the presence of aqueous dichloroacetic acid
as catalyst in a THF-acetone-methanol solvent mixture at
room temperature, leading to new and original block
copolymers containing PHEA, poly(acrylic acid) (PAA) or
Poly(dimethylsiloxane)-Containing Copolymers
Table 6. PDMS Macroinitiators Used for the Synthesis of Di- and Triblock Copolymers by Atom-Transfer Radical Polymerization
Scheme 41. Structure of the ATRP Ligands Used in the
Studies Involving PDMS Macroinitiators
poly(methacrylic acid) (PMAA). However, one could wonder
whether the PDMS chain is not degraded in such acidic
media. The tert-butyl ester groups were partially hydrolyzed
(40%) using a NaI/SiMe3Cl two-step technique without
altering the PDMS segment.
Macroinitiator 2 (Table 6, entry 2) (6200 g · mol-1) was
used by Minami et al.258 as IniStab (initiator + stabilizer)
Chemical Reviews, 2010, Vol. 110, No. 3 1259
for the dispersion polymerization of methyl methacrylate in
scCO2. The polymerization was carried out using the CuBr/
HMTETA (1,1,4,7,10,10-hexamethyltriethylenetetramine) cata-
lyst system at 65 °C. Monomer conversions above 70% were
obtained and increased when increasing the initiator con-
centration (i.e., decreasing the targeted molecular weight).
Good initiator efficiency was observed, and although the
molecular weight distributions had a slight tailing of low
molecular weights, a correct polydispersity index of 1.25 was
obtained.259 Huan et al.256 synthesized PMMA and poly(2-
dimethylaminoethyl methacrylate) (PDMAEMA) containing
di- and triblock copolymers by using macroinitiators 2 and
3 (5000 g · mol-1) at 90 °C in toluene and the CuBr/n-propyl-
2-pyridinalmethanimine catalyst system. In both cases, a
linear evolution of the molecular weight with conversion was
observed and monomodal MWDs were obtained. Similarly,
Bes et al. polymerized MMA260,261 and 2-dimethylaminoethyl
methacrylate (DMAEMA)262 from macroinitiator 3 (2100-
5000 g · mol-1) and obtained the expected triblock copolymer
with good control over the molecular weights. The same
group used macroinitiator 8 (Table 6, entry 8) for the
polymerization of MMA and DMAEMA in toluene.269
Macroinitiator 4 (Table 6, entry 4) (8200 g · mol-1) bears
the same bromo-isobutyrate chain ends as macroinitiator 3
but was synthesized differently. Thus, macroinitiator 4 has
a rather fragile Si-O-C78,79 bond between the different
blocks. It was used to polymerize butyl methacrylate at 110
°C in diphenyl ether using the CuCl/dNbpy catalyst sys-
tem.263 Complete initiator efficiency and monomodal MWDs
were obtained. When the macroinitiator concentration was
increased, the polymerization rate and monomer conversion
increased. Moreover, when the catalyst concentration was
1260 Chemical Reviews, 2010, Vol. 110, No. 3
increased, the rate of polymerization as well as the PDI
increased. The rate of polymerization was found to be
proportional to [catalyst]0.82 and to [macroinitiator]0.67.
Macroinitiator 5 (Table 6, entry 5) (R ) n-Bu) exhibits
the same chain end as macroinitiator 2, although it was
synthesized by hydrosilylation. The bulk polymerization of
oligo(ethylene glycol) methyl ether methacrylate (OEG-
MA)264 using macroinitiator 5 (640-3950 g · mol-1) and
CuBr/Me6TREN (tris[2-(dimethylamino)ethyl]amine) or CuBr/
PMDETA (N,N,N′,N′,N′′-pentamethyldiethylenetriamine) at
20 °C or CuBr/N(n-propyl)-2-pyridylmethanimine at 90 °C
showed a poor control of the molecular weights, explained
by an incompatibility between the PDMS macroinitiator and
OEGMA. Upon addition of benzene to compatibilize the
reagents, good control was achieved, especially when using
PMDETA as ligand at 20 °C. The activity of ATRP ligands
progresses in the order N1 < N2 < N3 < N4 (where 1, 2,
3, 4 refer to the number of N atoms in the ligand).221 N-(n-
Propyl)-2-pyridylmethanimine ensured a very good control
of the molecular weights at 90 °C in n-propanol. It was
suggested that a high temperature led to a higher compatibili-
zation of the reaction mixture. The same macroinitiator 5 (R )
nBu) was used for the controlled polymerization of MMA265
at 90 °C in xylene using a CuBr/dNbpy catalyst system. The
molecular weight increased with conversion, and the final
PDI was 1.17. Moreover, BuA was successfully polymerized,
giving the desired diblock copolymer with a monomodal
MWD. Macroinitiator 5 was also used by Pyun et al.268 to
polymerize styrene. Chain extension of the PDMS-b-PS
macroinitiator with 3-(dimethoxymethylsilyl)-propyl acrylate
(DMSA) by ATRP yielded an ABC triblock copolymer. The
latter reactive segment was covalently attached to a silicon
wafer. They used the equivalent difunctional macroinitiator
7 (8200 g · mol-1) to successfully polymerize MMA and
2-trimethylsilyoxyethyl methacrylate (HEMA-TMS). Mac-
roinitiator 5 (R ) PS) was used to synthesize PS-b-PDMS-
b-PBuA and PS-b-PDMS-b-PMMA triblock copolymers.265,267
Hong et al.266 compared the different reactivities of a
2-bromoisobutyrate-terminated PDMS (macroinitiator 5 in
Table 6 (5800 g · mol-1)) and a 2-bromopropionate-termi-
nated PDMS (macroinitiator 6 in Table 6 (8800 g · mol-1))
using the (CuBr/PS8-dMbpy)/(CuBr2/Me6TREN) hybrid
catalyst system. In this catalyst system, 4,4′-dimethyl-2,2′-
bipyridine (dMbpy) was immobilized on a cross-linked PS.
The polymerization of methyl methacrylate with the 2-bro-
mopropionate-terminated PDMS was faster than with the
2-bromoisobutyrate-terminated PDMS; this result was un-
expected, since the isobutyrate group is more reactive than
the 2-bromopropionate group because of the more labile
C-Br bond. It was suggested that the better dissociation
properties of bromoisobutyrate led to a higher amount of
formed radicals and therefore a higher amount of CuBr2
deactivator (persistent radical effect)233,234 resulting in retar-
dation. The bromopropionate-terminated PDMS presented
a low initiator efficiency and produced polymers with a large,
bimodal MWD. Moreover, MMA and BuA were copolymer-
ized using macroinitiator 5 and the previously described
hybrid catalyst system in toluene to avoid a phase segrega-
tion. The rate of polymerization increased with increasing
BuA concentration. Like in conventional free radical po-
lymerization, MMA was incorporated faster than BuA. The
reactivity ratios were specifically calculated for this system
(rMMA ) 2.16 and rBuA ) 0.42).
Pouget et al.
Scheme 42. Structure of the PDMS Methacrylate
Macromonomer
Macroinitiator 9 in Table 6 (9800 g · mol-1) was used by
Nakagawa et al.270,271 to successfully polymerize styrene in
diphenyl ether or in bulk at 130 °C using the CuCl/dNbpy
catalyst system. Styrene conversions around 70% were
obtained, and the linearity of the first-order kinetic plot
confirmed the constant concentration of growing radicals
throughout the polymerization. The whole molecular weight
distribution was shifted toward higher molecular weights.
The same macroinitiator 9 was used for the polymerization
of MMA265 in toluene at 90 °C using the same CuCl/dNbpy
catalyst system. It was shown that the molecular organic
benzyl chloride initiator had a higher efficiency than the
PDMS macromolecular initiator. Moreover, the macroini-
tiator was used for the polymerization of styrene, isobornyl
acrylate, and BuA.265
Macroinitiator 10 in Table 6 was used for the polymeri-
zation of isobornyl acrylate,274 styrene,272–274 and MMA.273
In the polymerization of MMA,273 when using the CuCl/
bpy catalyst system at 130 °C in xylene, a low initiator
efficiency was observed. The living character of the resulting
block copolymer was demonstrated by resuming the polym-
erization with 1-(dimethoxymethylsilyl)propyl acrylate to
give the desired ABC triblock copolymer.272
4.3.3. Graft Copolymers
4.3.3.1. Macromonomer Route. The use of CRP allows
all the chains to grow simultaneously to ensure that they
have the same composition. When a macromonomer is used,
the spacing between the side chains is determined by the
reactivity ratios of the macromonomer and the comonomer.
These reactivity ratios are determined by (i) the reactivity
of the macromonomer and the comonomer based on their
chemical structure, (ii) the rather slower diffusion of the large
macromonomer, and (iii) the possible incompatibility due
to repulsive interactions between the growing polymer chain
and the macromonomer.275,276
When copolymerizing the methacrylate-terminated PDMS
(Mn ) 2200 g · mol-1, PDI ) 1.18) (Scheme 42) and MMA
(5/95 mol %) by ATRP (ethyl 2-bromoisobutyrate as
initiator),275 the proportion of PDMS in the copolymer was
kept close to the feed composition regardless of the monomer
conversion. Therefore, the branches were distributed homo-
geneously all along the polymer backbone. When using ethyl
2-bromoisobutyrate as initiator and CuCl/dNbpy as catalyst
at 90 °C in bulk, a phase segregation was observed276 due
to the incompatibility between the PDMS macromonomer
and the growing PMMA radical. This phase segregation
could be diminished by working in xylene.277 Moreover,
when ethyl 2-bromoisobutyrate was replaced by a PDMS
macroinitiator, a better compatibilization of the growing
polymer chain with the macromonomer was observed.276
Lutz et al.278 used the macromonomer route to synthesize
PMMA gradient graft copolymers containing poly(D-lactic
acid) (PLA) and poly(dimethylsiloxane) side chains. More-
over, they synthesized (PMMA-g-PLA)-b-(PMMA-g-PDMS)
diblock copolymers by using ethyl 2-bromoisobutyrate as
initiator and CuCl/dNbpy as catalyst at 90 °C in a solvent
mixture of p-xylene and diphenyl ether. When using a hybrid
Poly(dimethylsiloxane)-Containing Copolymers
Scheme 43. Preparation of the Macroinitiator for the Graft
Polymerization from Pendant ATRP Initiators
catalyst system (where the catalyst CuBr/PS8-dMbpy was
immobilized onto cross-linked PS) in toluene, MMA was
consumed faster than with the homogeneous catalyst
system.279
Double comb copolymers of poly(ethylene glycol) methyl
ether methacrylate and poly(dimethylsiloxane)-terminated
methacrylate (Scheme 42) were synthesized280 by using ethyl
2-bromoisobutyrate as initiator and CuCl/Me6TREN as
catalyst at 70 °C. The copolymer formed statistical hetero-
grafted brushes due to similar reactivity ratios between the
two macromonomers.
4.3.3.2. Pendant Side Groups. Several publications dealt
with graft copolymerization by ATRP from pendant side
groups. In all cases the macroinitiator was prepared by
hydrosilylation of a pendant vinyl-functionalized PDMS with
a 2-(4-chloromethylphenyl)ethyldimethylsilane (Scheme
43).270,274,281,282 Styrene was polymerized in bulk or in
toluene with the macroinitiator, showing a good efficiency
(>97%). However, the variable number of initiating sites
per PDMS chain was responsible for a rather high PDI
(between 2 and 3).
4.3.3.3. Surface Initiation. In addition to the bulk or
solution graft copolymerization, the vinyl polymerization by
ATRP can also be initiated from the PDMS surface. It is
well known that ATRP initiators are conveniently grafted
onto the surface of silicon wafers by self-assembly of
2-bromoisobutyryl-functionalized trichlorosilane.267 In a simi-
lar manner, ATRP initiator could also be tethered onto PDMS
Scheme 44. Preparation of the Macroinitiator for the Surface-Initiated Graft Polymerization by ATRP
Scheme 45. Preparation of a Four-Arm Siloxane ATRP Initiator
Chemical Reviews, 2010, Vol. 110, No. 3 1261
films by creating a reactive group at the surface of the PDMS
film by UV/ozone treatment (Scheme 44).283 This reactive
group would further be used to graft an ATRP initiator onto
the surface. Thus, acrylamide in water solution was grown
from the PDMS film surface by this technique.283
4.3.3.4. Star Copolymers. Star copolymers are prepared
from multifunctional ATRP initiators. Tetramethylcyclotet-
rasiloxane was first hydrosilylated with p-vinylbenzyl chlo-
ride (Scheme 45). This four-arm siloxane ATRP initiator was
used to polymerize styrene284,285 in bulk using CuCl/dNbpy
as catalyst. To achieve appreciable conversions (72%), the
reaction had to be conducted at 130 °C. An almost pure star
copolymer was obtained, although some thermal initiation
of styrene produced a small fraction of low molecular weight
dead chains. The same macroinitiator was used with Cu(0)/
bpy and CuCl/bpy catalyst systems282 in the polymerization
of styrene. However, since the siloxane content in these star
copolymers is very low, one can assume that the resulting
copolymer does not have properties inherent to PDMS.
In summary, ATRP from PDMS macroinitiators or with
PDMS macromonomers was the most widely studied route
among CRP methods. It has been applied to the polymeri-
zation of a wide range of monomers. ATRP has given
interesting results for the controlled synthesis of complex
architectures from PDMS chains and gave the desired
copolymer architectures with relatively low PDI and accept-
able monomer conversions. However, the insolubility of
certain catalyst systems can present a problem to achieve
good initiator efficiency and correct control over the mo-
lecular weights.
4.4. ITP: Iodine-Transfer Polymerization
Iodine-transfer polymerization141 was developed in the late
1970s by Tatemoto286,287 and relies on the use of alkyl iodides
as transfer agents. The mechanism of ITP with alkyl iodide
is shown in Scheme 46.
The initiating radical, A•, is generated by decomposition
of a radical initiator, such as 2,2′-azobisisobutyronitrile, in
1262 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 46. Reactions in Degenerative Transfer
Polymerization with Alkyl Iodides
step 1. A• adds to monomer in step 2, and the resulting radical
propagates as shown in step 3. The transfer of iodine from
the transfer agent, R-I, to the propagating radical, Pn•, results
in the formation of the dormant chain Pn-I and a new
initiating radical, R• (step 4). Large differences in the stability
of the reactants and products involved in step 4 (i.e., relative
values of k1 and k-1) could result in shifting the equilibrium
overwhelmingly to the right (k1 . k-1) or to the left (k1 ,
k-1). The case where the structure of R• closely resembles
the structure of the propagating radical results in a thermo-
dynamically neutral transfer step (i.e., k1 ≈ k-1). In step 5,
R•, generated from the alkyl iodide, adds to a monomer unit.
The exchange process described in step 6 is thermodynami-
cally neutral (i.e., degenerative transfer: kex ) k-ex, K ) kex/
k-ex ) 1) because the propagating chains and dormant chains
have the same structure on both sides of the equilibrium.
The ratio between transfer rate coefficients and propagation,
namely, the transfer constant, gives the reactivity of the
transfer agent (CT ) k1/kp) and the exchange constant gives
the reactivity of the dormant chains (Cex ) kex/kp). As in
any radical process, conventional termination occurs in ITP
polymerization with alkyl iodides (step 7). Lowering the
contribution of the irreversible termination step remains
essential to keep a good control over the polymerization. In
ITP, the overall concentration of the polymer chains is indeed
equal to the sum of the concentrations of the consumed
transfer agent and of the consumed initiator. Iodine-transfer
polymerization from PDMS macrotransfer agents has been
barely studied. The few PDMS-based macrotransfer agents
studied in ITP are described in Table 7.
Shinya et al.288 used transfer agent 1 (Table 7, entry 1) as
macrotransfer agent or as photoiniferter (under UV activa-
Table 7. PDMS Macrotransfer Agents Used in Iodine-Transfer Polymerization
Pouget et al.
tion) to polymerize tetrafluoroethylene (TFE). Under UV
radiation, the iodinated group cleaved homolytically to create
a radical on the PDMS chain and an iodine radical. Although
the mechanism was not given by the authors, we can assume
that the growing polymer chain was reversibly deactivated
by reaction with an iodine radical, with iodine (created from
the combination of two iodine radicals) or by degenerative
chain transfer. They observed a good incorporation of the
macrotransfer agent into the triblock copolymer and TFE
conversion reached 40%. However, when using vinylidene
fluoride (VDF) or chlorotrifluoroethylene (CTFE) as mono-
mers, low monomer conversion and poor transfer agent
efficiency were observed. Similar results were obtained by
thermal peroxide initiation. The same group used transfer
agent 2 (Table 7, entry 2) to copolymerize VDF and TFE
(VDF/TFE ) 85/15 mol %) under UV radiation. Even
though they observed low conversions of monomer and
transfer agent, the desired triblock structure was confirmed
after extraction.
Our group used macrotransfer agent 3 (Table 7, entry 3)
to synthesize PS-b-PDMS-b-PS triblock copolymers in
miniemulsion polymerization.41 An increase of molecular
weights with conversion showed the controlled character of
the polymerization. A high styrene conversion of 90% was
obtained. NMR and SEC analysis were successfully cor-
related with the theoretical molecular weight. A chain
extension in seeded emulsion polymerization proved the
living character of the iodinated chain end. However, a pH
drop, attributed to chain-end degradation by HI elimination,
was observed during the polymerization. It was demonstrated
that the reaction had to be stopped rapidly once high
monomer conversion has been reached in order to avoid
undesirable chain-end degradation.
More recently our group used the same macrotransfer
agent 3 (Table 7, entry 3) for the controlled synthesis of
PVAc-b-PDMS-b-PVAc triblock copolymers in bulk at 80
°C.290 The molecular weight evolution proved the controlled
character of the synthesis, and high conversions, around 75%,
were obtained. However, the PVAc-I chain end was prone
to degradation (Scheme 47), as previously demonstrated by
Boutevin et al.291 in the case of PVAc-X chain ends (X )
Br or Cl). Iovu et al.292 showed an identical degradation
mechanism by hydrolysis of the PVAc-I chain ends.
A Japanese patent describes the synthesis of poly(dime-
tylsiloxane)-g-poly(ethylacrylate) copolymers.293 The ma-
Poly(dimethylsiloxane)-Containing Copolymers
Scheme 47. Hydrolysis (a) and Decomposition (b) of the
Iodo End Group of Poly(vinyl acetate) after Iodine-Transfer
Polymerization292
Scheme 48. Synthesis of the Macrotransfer Agent Used for
Preparing Poly(dimethylsiloxane) Graft Copolymers by
Iodine-Transfer Polymerization293
crotransfer agent was prepared by reacting 97:3 dime-
thylsiloxane-vinylmethylsiloxane copolymer with ICl
(Scheme 48). The copolymers were prepared by graft
polymerization of iodine-containing silicone polymers with
ethyl acrylate in the presence of a radical polymerization
initiator and then used as compatibilizers for rubber
compounds.
To conclude, the literature extensively describes iodine-
transfer polymerization for a wide range of monomers,141
including styrenics, acrylates, vinyl esters, dienes, fluorinated
monomers (TFE, VDF, hexafluoropropylene (HFP)), and
chlorinated monomers (vinyl chloride, vinylidene chloride).
This leaves a lot of possibilities for ITP from PDMS
macrotransfer agents. It should also be possible to synthesize
graft copolymers by using PDMS macromonomers with
ITP141 or RITP138–141,294,295 techniques. Moreover, ITP41 and
RITP142,143,296–299 are compatible with dispersed aqueous
medium, which is of great interest for implementing the
technique at an industrial scale.
4.5. RAFT: Reversible Addition-Fragmentation
Chain Transfer
RAFT polymerization300,301 involves dithiocarbonyl-
transfer agents of general formula Z-C(S)S-R. The dithio-
carbonyl group reacts in two steps: first, addition of a radical
(formed by initiator decomposition) to the CdS double bond
followed by a beta fragmentation of one of the two C-S
bonds of the intermediate radical. Like iodine-transfer
polymerization, RAFT is based on a degenerative transfer.
When the transfer agent (Scheme 49) involves a xanthate
(Table 8, entry 2), various authors sometimes refer to the
“MADIX” process as claimed by the Rhodia Co.,302,303 which
stands for MAcromolecular Design through the Interchange
of Xanthates. Again, a patent from Matyjaszewski covers
the synthesis of poly(dimethylsiloxane) containing block and
graft copolymers by RAFT.252
Chemical Reviews, 2010, Vol. 110, No. 3 1263
Scheme 49. General Mechanism of RAFT/MADIX
Polymerization
4.5.1. Block Copolymers
Several different macrotransfer agents were used to
synthesize block copolymers by RAFT (Table 8). Pai et al.304
used transfer agent 1 (Table 8, entry 1) to grow two statistical
copolymer blocks from the central PDMS block, comprising
units of N,N-dimethyl acrylamide (N,N-DMA) and 2-(N-butyl
perfluorooctanefluorosulfonamido) ethyl acrylate (BFA) lead-
ing to an amphiphilic triblock copolymer, however containing
fragile Si-O-C bonds.78,79 The polymerization was con-
ducted in trifluorotoluene, which solubilized the macroRAFT
agent and the vinylic copolymer. The molecular weight
increased with conversion as expected, leading to the desired
triblock copolymer. However, the authors observed that the
copolymer composition varied whether a conventional RAFT
agent or the corresponding macroRAFT agent was used.
They assumed that the hydrophobic PDMS chain favored
the incorporation of the hydrophobic BFA, through a
preferred solvation in the PDMS phase.
The RHODIA Co. has developed a series of MADIX
agents exhibiting the general structure of compound 2 (Table
8, entry 2), the R group allowing tuning the reactivity of the
MADIX agent. They used MADIX agent 2 with R ) -C2H5
for the controlled polymerization of VAc,308 ethyl acrylate,308
butyl acrylate,311 and 2-dimethylaminoethyl acrylate,305 giv-
ing well-defined triblock copolymers. One potential drawback
of MADIX lies in the fact that the xanthate terminal group,
if not deactivated, is likely degraded during the lifetime of
the polymer or under some specific application conditions,
leading to some low molecular weight malodorous, poten-
tially toxic sulfur-based byproduct.306,307 Therefore, the
RHODIA Co. has developed some methods to eliminate the
xanthate end group from the copolymer. MADIX agent 2
was oxidized306,310 with dilauryl peroxide at 80 °C in
isopropanol or with di-tert-butyl peroxide at 175 °C in
2-octanol giving the PDMS ethyl ester and S-undecyl-O-
ethyl xanthate which could easily be eliminated by selective
precipitation.312 Moreover, transfer agent 2 with different R
groups (isobutyl, cyclohexyl, and phenyl ethyl) was used to
study elimination of the xanthate group by dethiocarboxyla-
tion,307,309 giving the thiol-terminated PDMS. All were
completely dethiocarboxylated when heated at 180 °C in
o-dichlorobenzene. The phenyl ethyl derivative even under-
went complete dethiocarboxylation when heated for 2 h at
130 °C in chlorobenzene.
4.5.2. Graft Copolymers
Shinoda et al.277,313 used cumyl dithiobenzoate (Scheme
50) as a RAFT transfer agent for the copolymerization of
MMA and PDMS macromonomer (Scheme 42) (Mn ) 2300
g · mol-1). The reaction was conducted at 75 °C with 5 mol
% of PDMS-MA. Xylene was added to the reaction medium
1264 Chemical Reviews, 2010, Vol. 110, No. 3
Table 8. PDMS Macrotransfer Agents Used for the Polymerization by RAFT/MADIX
Scheme 50. Cumyl Dithiobenzoate
Scheme 51. (A) Homogeneously Distributed Branches, (B)
Heterogeneously Distributed Branches (toothbrush
structure), and (C) Intermediate Distribution
to favor homogeneous conditions and to lead to a more
controlled polymerization. Since the used macromonomer bore
a methacrylate end group, the authors expected a similar
reactivity of the macromonomer to MMA (rMMA ) rmacromonomer
) 1) (Scheme 51A). It was shown that MMA was incorporated
slightly faster than the macromonomer into the copolymer.
When using the Jaacks method314 the reactivity ratio rMMA was
found to be equal to 1.49.
In a conventional free radical polymerization, rMMA was
found around 2.98. This means that when a RAFT agent is
used, the incorporation of the macromonomer is more regular
all along the copolymerization, leading to a better defined
architecture. However, at a lower temperature (60 °C), the
incompatibility between PMMA and PDMS increased,
leading to a higher rMMA and worse control of the mac-
romonomer incorporation. Therefore, the reaction tempera-
ture has to be kept relatively high (75 °C) to reach better
control over the copolymerization. Since rMMA > 1, incor-
poration of PDMS-MA is less important in the early stages
of the polymerization, leading to a toothbrush structure
(Scheme 51B) at low temperature (60 °C) and an intermedi-
ate toothbrush-regular distribution structure (Scheme 51C)
at higher temperatures (75 °C).
Pouget et al.
Scheme 52. PDMS with Pendant Xanthate Groups Used To
Synthesize Graft Copolymers308
The RHODIA Co. has claimed the synthesis of PDMS
graft copolymers using a poly(dimethylsiloxane) with pen-
dant xanthate groups (Scheme 52).308 This macro MADIX
agent was synthesized by anionic copolymerization of D4
and D4-containing pendant propanol side groups. The primary
pendant alcohol was esterified with 2-bromopropionyl bro-
mide. However, the esterification reaction was not complete,
and 30% of free OHs were still present according to the 1H
NMR spectrum. The final transfer agent was obtained by
nucleophilic substitution of bromine by the xanthate salt
(K+ -SCSOEt) in acetonitrile.
Until now, RAFT and/or MADIX from PDMS mac-
rotransfer agents were used for the controlled polymerization
of N,N-DMA, BFA, VAc, EtA, BuA, tBuA, and 2-(dim-
ethylamino) ethyl acrylate. RAFT allows the controlled
polymerization of a very wide range of monomers,300,301 and
therefore, many other copolymers could be potentially
prepared. For instance, the controlled polymerization of
acrylates (BuA, MeA, AA, 2-hydroxyethyl-acrylate,...), sty-
renics (styrene, styrene p-sulfonic acid sodium salt,...),
methacrylates (such as MMA and MAA), acrylamides,
N-vinyl monomers (N-vinyl pyrrolidone), vinyl esters (vinyl
acetate, vinyl benzoate), vinyl derivatives of pyridine (4-
vinyl pyridine, 2-vinylpyridine), halogenated monomers
(vinylidene chloride), and acrylonitrile has been described.315
One drawback of RAFT/MADIX is the potential decomposi-
tion of chain ends, giving possible harmful byproducts.
Another limitation is the rather complex synthesis of the
RAFT/MADIX macrotransfer agents since very few precur-
sors (such as KSC(S)OEt) are commercially available.
4.6. Conclusion on CRP Techniques
Controlled radical polymerization has allowed the con-
trolled synthesis of a large number of PDMS-containing
block and graft copolymers with predefined molecular
weights. The use of PDMS azo macroinitiators can be
combined with different CRP methods. For example, a
macroinitiator was used in combination with TEMPO in
Poly(dimethylsiloxane)-Containing Copolymers Chemical Reviews, 2010, Vol. 110, No. 3 1265

Table 9. Advantages and Drawbacks of the Five Main over, the iodinated transfer agents allow the control of
Controlled Radical Polymerization Methods fluorinated and chlorinated monomers. The main drawbacks
Polymerization of ITP are the lack of control of methacrylates (the new RITP
method Advantages Drawbacks method is however able to fill this gap) and the rather large
Iniferter large monomer range large PDI PDIs finally obtained in comparison with other CRP methods
photoactivation is (although medium polydispersity is actually not a real
required to reach
living properties drawback for most industrial applications and is even
NMP low PDI high reaction sometimes helpful).316 The RAFT/MADIX technique enables
temperatures (>100 control of vinylidene chloride, VAc, and functional mono-
°C) mers like AA. However, a potential chain-end degradation
bad control with
TEMPO (except for produces sulfur-containing side products. Therefore, a chain-
styrenics) end elimination or transformation could be necessary de-
ATRP low PDI presence of metallic pending on the desired application.
ions
large monomer range neither control of acidic Controlled radical polymerization allows the synthesis of
monomers nor VAc a wide range of block and graft copolymers from PDMS
catalyst partial precursors. However, it is important to choose the best
insolubility
ITP implemented in relatively large PDI method depending on the type of monomer to be polymerized
dispersed aqueous (especially for functional monomers), on the required level
medium of control of the polymerization (low PDI), and on the type
control of halogenated no control over of application (for instance, sensitivity to metallic species
monomers methacrylates (except
by RITP) from ATRP). Moreover, until now, only ITP from PDMS
RAFT/ control of functional possible chain-end macrotransfer agents has been implemented in dispersed
MADIX monomers (VAc, AA, degradations with aqueous medium. The synthesis of these copolymers is
etc.) harmful byproduct possible in processes where the PDMS chain does not need
release
triblock copolymers to diffuse across the aqueous phase (miniemulsion or
with vinylic central suspension polymerization). Indeed, since the PDMS is too
block hydrophobic to diffuse across the aqueous phase, these
control of vinylidene copolymers cannot be synthesized in ab initio emulsion
chloride
polymerization (unless a reverse emulsion would be con-
sidered). Thus, the use of CRP to synthesize well-defined
NMP. Moreover, it would be possible to polymerize a vinylic
copolymer architectures from PDMS precursors leaves a lot
monomer by RITP and reverse ATRP. When an azo
of unexplored possibilities like the use of functional mono-
macroinitiator (or peroxide macroinitiator) is used, the low
mers (possibly protected monomers), the use of enhanced
initiator efficiency is responsible for an important number
CRP methods like RITP, and the polymerization in dispersed
of dead chains and an ill-defined architecture. The use of
media.
macroiniferters, macroalkoxyamines (NMP), macroinitiators
(ATRP), or macrotransfer agents (ITP, RAFT/MADIX) leads
to a higher initiator efficiency, giving a much better control 5. Some Properties and Some Applications
of the copolymer architecture. The desired block or graft The original architectures obtained by radical polymeri-
copolymers are obtained with good control of the molecular zation of various monomers in the presence of PDMS enable
weights. the preparation of hybrid materials having well-controlled
Among the five major CRP methods, each one has its properties. The presence of PDMS segments in the materials
advantages and drawbacks (Table 9). Iniferters enable the generally creates a phase demixing, which can change the
controlled polymerization of a wide range of monomers, surface properties of a material or its mechanical and solution
although producing polymers with large PDIs. Moreover, properties. The presence of a PDMS block in the material
thermal iniferters permit the synthesis of triblock copolymers, generally enhances the thermal stability compared with the
with the PDMS blocks being at the extremities, as well as reference organic homopolymer. Some interesting applica-
multiblocks. The main drawback of photoiniferters is that tions found in the literature cited in this review will be briefly
irradiation is required to obtain a living character. Nitroxide- detailed in the following discussion.
mediated polymerization leads to polymers with low PDIs,
but the compulsory high temperature of reaction can be 5.1. Phase Segregation Properties
problematic if the recipe involves fragile reactants or when
one would like to work in dispersed aqueous medium. Cameron et al.166,317 studied the variation of the glass
Moreover, TEMPO has a low efficiency for monomers other transition temperature (Tg) by DSC for copolymers of
than styrene and its derivatives. Therefore, other controlling methacrylic polysiloxane macromonomers and styrene (or
agents using more efficient counter radicals, such as SG1, AN) that naturally demix. Their goal was to highlight the
must be used to polymerize other monomers and to work in effect of the forced miscibility due to a copolymer archi-
dispersed aqueous medium. ATRP leads to copolymers with tecture. First, they experimentally measured the Tg’s of the
low PDIs and controls a large range of monomers. However, PS domains versus the size of the macromonomers incor-
acidic monomers or vinyl acetate can not be controlled. porated in the backbone. Following the Gordon-Taylor
Moreover, the presence of metallic ions and the low solubility equation318 for totally miscible blends, the authors concluded
of the catalyst in the reaction medium can be a problem. that an “appreciable degree” of miscibility exists between
Iodine-transfer polymerization is a rather low-cost method the domains of the copolymer. They observed that the
which is applicable in dispersed aqueous medium (mini- plasticizing effect of PDMS with PS is much more important
emulsion and potentially suspension polymerization). More- than in the case of poly(acrylonitrile) (PAN) where the
1266 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 53. Five-Block Copolymers of AFR, PDMS, and
PVP or PS
measured Tg’s were closer to those of the parent homopoly-
mers. It indicates that, in spite of the chemical bonding in
both cases, the forced miscibility is much more efficient with
PS-PDMS copolymers than with PAN-PDMS copolymers.
Blahovici et al.319 investigated the miscibility of PMMA-g-
PDMS copolymers blended with the PMMA parent ho-
mopolymer. By DSC measurements, they observed only one
Tg versus two Tg’s for a blend composed of the two parent
polymers (i.e., PDMS and PMMA). This allowed them to
conclude that such copolymers were miscible with PMMA.
In another study, Hamurcu et al.104 observed that networks
from block copolymers of PDMS and PS showed a Tg of 45
°C, giving evidence for the plasticizing effect of the flexible
PDMS segments (Mn ) 1050 g · mol-1). Fawcett et al.214 also
showed the presence of only one Tg in the silicone graft
copolymers with short chains of poly(chloroprene). Hamurcu
and co-workers103,104 carried out the synthesis of PDMS-b-
PS. Only the Tg of the poly(styrene) (PS) block (90 °C) was
observed, otherwise lower than the one of homo-PS (100
°C) because of the plasticizing effect of the short PDMS
segments.
Uyanik and co-workers86 studied different properties of
PDMS-containing five-block copolymers based on AFR and
PVP or PS (Scheme 53). The Tg of the PS blocks is affected
by their chain length, but they are not significantly altered
by the presence of the middle blocks. In this five-block
copolymer, the domains are well separated and no plasticiz-
ing effect was observed.
Cvetkovska et al.82 studied the morphology of poly-
(monobutyl itaconate)-b-PDMS-b-poly(monobutyl itaconate)
triblock copolymers. The SEM image of the poly(monobutyl
itaconate) homopolymer shows a laminated fracture (Scheme
54a), whereas the triblock copolymer shows a porous
morphology (Scheme 54b). SEM images carried out by
Chang et al.112 in the case of PMMA-b-PDMS copolymers
also show a phase separation (Scheme 55). This phase
segregation is always observed when the length of the blocks
is enough to induce an incompatibility between the two
polymers. The surface of the film shows a PDMS ball-like
structure (confirmed by Si mapping), the average diameter
of the phase domain being around 0.4 µm.
Smith et al.320,321 studied the bulk properties and monitored
the morphology of PMMA-g-PDMS. As expected, they
observed spherical domains of PDMS in a PMMA matrix
for low contents of PDMS. However, for higher contents,
e.g. 45 wt % of PDMS, they observed a cylindrical texture.
Pouget et al.
Scheme 54. SEM Image of Poly(monobutyl itaconate)
Homopolymer ((a) magnification ×3000) and
Poly(monobutyl itaconate)-b-PDMS-b-poly(monobutyl
itaconate) Triblock Copolymer ((b) magnification ×250)a
a
Reproduced with permission from ref 82. Copyright 2000 Wiley-VCH
Verlag GmbH & Co. KGaA.
Scheme 55. SEM Micrograph of PDMS-b-PMMA Filma
a
Reproduced with permission from ref 112. Copyright 1996 Wiley-VCH
Verlag GmbH & Co. KGaA.
The size of the domains increased with the length of the
PDMS side chain (Scheme 56). Smith et al.322 also observed
that the blends of PMMA-g-PDMS copolymers with PMMA
and PVC changed from a spherical to a bicontinuous
morphology when the PDMS content ranged from 25% to
47%, respectively. Incompatible blends with PVC and
PMMA were also obtained for higher PDMS contents.
5.2. Surface Properties
From our literature survey, it was possible to extract
studies dealing with the surface properties either of the
copolymers themselves or of the copolymers embedded in a
matrix. It is well known that PDMS exhibits a low surface
energy (20 mN m-1). Low energies were also observed for
various PDMS-containing copolymers, as investigated by
Kawakami et al.,146,323,324 Smith and McGrath,325 Lee et al.,326
or Taskiran et al.327 As expected, the contact angle measure-
ments revealed that there was a logical increase of the surface
hydrophobicity with an increasing PDMS content or molar
mass (Table 10).
In another study, Bes et al.261 performed the synthesis and
characterization of PMMA-b-PDMS-b-PMMA triblock co-
polymer. DSC measurements revealed a phase segregation
(two Tg’s) even at low PMMA content (23 wt %). A higher
degree of segregation and a high PDMS concentration at the
surface were, respectively, observed by TEM and X-ray
photoelectron spectroscopy (XPS) for the copolymers con-
taining the highest PDMS content. Logically, the surface
tension was higher and the contact angle with water smaller
in the case of a high PMMA-containing copolymer (surface
tension ranging from γS ) 19.15 (41% mol PMMA) to 27.32
mN · m-1 (81% mol PMMA) and a water contact angle
ranging from 112.9° to 105.25°). Nakamura99 also observed
a PDMS accumulation at the surface of PMMA particles
Poly(dimethylsiloxane)-Containing Copolymers
Scheme 56. TEM of PMMA-g-PDMS with an Increase of Side Chain Length and End View of the Cylindrical Morphology of
PMMA-g-PDMS (20 000 g mol-1)a
a
Reprinted with permission from ref 320. Copyright 1992 American Chemical Society.
Table 10. Influence of the Siloxane Content and Molecular
Weight on Water Contact Angle of PMMA-g-PDMS As
Measured by Smith and McGrath325
macromonomer Mn wt % wt %
(g mol-1) (charged) (via NMR)
PMMA homopolymer
control
1000 5 4.0
20 15.0
5000 5 4.3
20 16.8
10 000 5 5.8
20 15.9
Scheme 57. Fluorinated Monomers Used by Bertolucci et
al.: m ) 6 (StyF6) and m ) 8 (StyF8)115
prepared by dispersion polymerization in organic media
using an azo macroinitiator. This result is in accordance
with the fact that the azo macroinitiator is used as a
dispersant (steric stabilizer) during the polymerization.
Smith and co-workers320,321 analyzed the surface character-
istics of PMMA-g-PDMS graft copolymers. Results were
in agreement with those obtained by other groups: the
presence of PDMS resulted in an increase of water contact
angle, i.e., the material was more hydrophobic.
Bertolucci et al.115 studied the wetting behavior of
fluorinated styrene-siloxane block copolymers. The structure
of the fluorinated monomers is given in Scheme 57. The
hydrophobic behavior in water was not influenced by
siloxane or fluoropolymer surface composition. The ole-
ophobic behavior for the copolymer using StyF8 (n ) 8)
with alcohols (θadv ) 64° and θrec ) 45°) was typical of
fluorinated surfaces, while a reference PDMS coating was
wetted (θadv ) θrec ) 0°). The results for the copolymer using
the StyF8 from Wilhelmy plate dynamic contact angle
analysis were characteristic of fluorinated surfaces (both
hydrophobic and oleophobic), in contrast to PDMS surfaces
which are only hydrophobic. In addition, the higher surface
stability of the copolymer using StyF8 compared to that of
the copolymer using StyF6 (n ) 6) was attributed to the
thermodynamic stabilization imparted by the liquid crystal-
line mesophase created by the association of fluorinated
chains with n g 8.
Uyanik et al.86 showed that the solubility of five-block
copolymers PVP-b-PDMS-b-AFR-b-PDMS-b-PVP or PS-
Chemical Reviews, 2010, Vol. 110, No. 3 1267
b-PDMS-b-AFR-b-PDMS-b-PS is affected by the AFR
block, while the surface properties are affected by the PDMS
blocks: as expected, the contact angle increased with PDMS
contact content or molecular weight.
angle
Ohata et al.96,100 found that the excellent ink repellency
74°
and ink resistance of silicone acrylic block copolymers films
97° (UV or thermo curable) resulted from their morphology and
105°
99°
from the silicone domain spacing within the films. These
107° films were found to be hard enough to form a nozzle face
108° and exhibited over 80° of θadv and 60° of θrec when put in
109°
an ink, which are superior values from those observed for
silicone or fluorinated coating films. Moreover, the durability
of the films was outstanding.
Hou et al.328,329 described the surface composition of the
P(HEMA)-g-PDMS, an amphiphilic graft copolymer. Thanks
to XPS analyses, they could determine that an increase of
the length of the grafted PDMS led to an increase of both
the concentration of PDMS at the air interface and the surface
layer thickness. However, even for high PDMS content in
the copolymer, the surface was never saturated with PDMS,
as in the case of block copolymers. By copolymerizing the
two monomers either by anionic polymerization (yielding a
narrow MWD, PDI about 1.1) or by photoinduced polym-
erization (yielding a broad MWD, PDI about 3.0), the authors
did not observe any significant differences under dry condi-
tions. However, under wet conditions, the authors demon-
strated the reorganization of the surface to generate a more
favorable energetic conformation with a decrease in PDMS
content. It is worth mentioning that, in these conditions,
MWD plays an important role owing to the shorter chain
segments ability to reorganize more easily. In summary,
anionic polymerization yielded materials that were more
resistant in wet conditions.
Inoue et al.43 carried out the surface modification of PDMS
using hydrophilic monomers. They noticed a decrease of
the water contact angle versus the extent of surface grafting.
The lowest contact angles were obtained with sodium styrene
sulfonate (NaSS) and methacrylic acid monomers (55°). The
same group98 also observed the surface accumulation of
PDMS in PDMS-b-PMMA/PMMA blends using X-ray
photoelectron spectroscopy (XPS), spectroscopy for chemical
analysis (ESCA), and water contact angle analysis. The water
contact angle increased abruptly with siloxane bulk concen-
tration or siloxane chain length (Mn > 2000 g · mol-1). In
another study from this group,117 the surface characteristics
of fluoroalkylsilicone-PMMA block copolymers and their
PMMA blends were investigated using XPS, water contact
angle analysis, and measurement of the 180° peel strength
against pressure-sensitive adhesives. It was observed from
film casting of the blends that the contact angles were higher
1268 Chemical Reviews, 2010, Vol. 110, No. 3
on the air side of the film than on the glass side. Again, the
higher the siloxane content, the higher the contact angle.
Preferential surface accumulation of siloxane segments was
observed for the block copolymers having long siloxane
chains and also relatively long fluoroalkyl side chains.
Interestingly, the presence of fluoroalkyl groups in the side
chains of silicone decreased significantly the 180° peel
strength on both air and glass sides. The conclusion of their
work was that the fluoroalkyl groups having a relatively long
chain length (tridecafluoro and heptadecafluoro) were phase
separated from PMMA segments and also enhanced the
surface accumulation of methylsiloxane groups.
A very interesting study carried out by Kawakami et al.330
suggested that the incorporation of PMMA-g-PDMS in a
matrix of PMMA changed the contact angle and the
composition at the surface. They showed that the graft
copolymers phase accumulated on both the air- and glass-
side surfaces. Particularly under low surface energy condi-
tions (i.e at the air surface), the siloxane segments caused a
large change in the contact angle value and ESCA spectra.
Finally, according to Taskiran et al.,84,327 microscopy can
give indications on the surface properties. Indeed, comparison
between expandable poly(styrene) (EPS) and EPS modified
by silicone acrylate showed a rough surface for the first one
whereas the modified EPS appeared smooth and glossy.
5.3. Mechanical Properties
PDMS is often used to enhance the mechanical properties
of materials in terms of deformation at break and impact
resistance. The issue that has to be overcome is the
compatibility between PDMS and the host material. The
following examples illustrate the benefit that can be gained
from insertion of PDMS in block copolymers.
The mechanical properties of PDMS-b-PS copolymers
were studied by Baysal et al.80,103 Considering PDMS molar
content ranging from 12% to 18%, the mechanical properties
of the copolymers were quite different from those of PS
homopolymer. It was observed that elongation at break for
PS was around 2%, whereas the elongations for the copoly-
mers were in the range of 7-26%. In addition to these
expected performances, stress at break and Young’s moduli
were only slightly lower than the PS reference. As expected,
the copolymers having the highest PDMS content exhibited
the best elastomeric properties. These results were confirmed
on similar materials103,104 observing that the block copolymers
had higher elongations (6-8%) than PS homopolymer and
that tensile strengths were very close to the values of PS
homopolymer (22-50 MPa).
The enhancement of impact resistance by insertion of
PDMS blocks was described by Falender et al.,22,23 who
prepared PDMS-g-P(AN-co-S) copolymers by telomerization
using a thio-functionalized PDMS as macrotelogen. The
higher the PDMS content, the higher the impact strength.
Uyanik85 studied the mechanical properties of a five-
block copolymer PVP-b-PCL-b-PDMS-b-PCL-b-PVP. The
copolymer showed a tensile strength, a Young’s modulus,
and an ultimate tensile strength lower than the one
observed for homo-PVP. On the contrary, the ultimate
tensile elongation was higher for the copolymer than for
the homo-PVP (Table 11).
Uyanik et al.87 also studied the behavior of water-soluble
PVP-b-PDMS-b-PVP under tensile stress. The authors
surprisingly obtained that the ultimate tensile strength and
ultimate tensile elongation of the films of copolymer were
Pouget et al.
Table 11. Mechanical Properties of
PVP-b-PCL-b-PDMS-b-PCL-b-PVP and Homo-PVP
tensile ultimate tensile Young’s
ultimate tensile strength strength modulus
sample elongation (%) (MPa) (MPa) (GPa)
homo-PVP 10.5 33.2 20.4 0.75
block 13.2 22.6 19.6 0.39
copolymer
lower than PVP’s and that their Young’s moduli values were
higher than PVP’s. They also showed by DSC that the
thermal characteristics of the copolymers were similar to
those of PVP homopolymer.
Neugebauer et al.280 prepared densely grafted PDMS-PEO
copolymers by ATRP using PDMS and PEO macromono-
mers. Morphology and mechanical analyses revealed the
presence of crystalline (PEO) and amorphous (PDMS)
phases. The viscous modulus was found to be higher than
the elastic modulus (G′′ > G′), which is typical of a melt
state. It was also observed that after cross-linking, caused
by annealing at high temperature, the material showed
properties typical of elastomers, more precisely of a soft gel
(G′ < 104 Pa and G′ g G′′).
Shinoda et al.277 carried out the analysis of three graft
PMMA-PDMS copolymers prepared by RAFT, ATRP, and
conventional radical polymerization. The three graft copoly-
mers show different architectures (Scheme 58). At small
deformations, the mechanical properties measured by dy-
namic mechanical analysis (DMA) did not differ for the
different architectures and indicated a thermoplastic-elastomer
behavior. This was confirmed by tensile test measurements
which showed similar moduli. However, at high deforma-
tions, under tensile test, sample (c) exhibited a low elongation
at break (1.3 MPa) and a high strength at break (6.4 MPa),
sample (b) a high elongation at break (3.8 MPa) and a low
strength at break (4.2 MPa), while sample (a) had an
intermediate behavior. These results were linked to the
sample architectures and consequent morphologies, more or
less phase segregated.
5.4. Thermal Stability
Smith et al.331 observed that copolymers of MMA and
silicone methacrylate macromonomers were more thermally
stable than the homo-PMMA. Furthermore, the longer the
PDMS side chains, the more stable the grafted copolymer
(Scheme 59). It was demonstrated332,333 that when MMA is
homopolymerized by free radical polymerization, recombina-
tion and disproportionation both occur, leading, on one hand,
to a sterically hindered linkage which decomposes at 175
°C and, on the other hand, to an unsaturated end group which
degrades at about 225 °C. The other disproportionation
product, that is, the saturated product, degrades beyond 300
°C. In the case of a copolymerization with a macromonomer
such as silicone methacrylate, irreversible termination of the
propagating radical leading to a saturated end group by chain
transfer seems to be more favorable, as long as there is
enough macromonomer in the medium to end cap each chain.
The authors suggested that chain transfer could be favored
by longer radical lifetimes of the macromonomers and
supported this explanation by SEC data which showed that
the higher the macromonomer content, the larger the mo-
lecular weight distribution, i.e., the higher the number of
chains, finally resulting in an increase of the thermal stability.
Chang et al.111 observed the same behavior in the case of
PMMA-b-PDMS block copolymers. The block copolymers
Poly(dimethylsiloxane)-Containing Copolymers
Scheme 58. Schematic Representation of the Three Copolymers Prepared by (a) Conventional Radical Polymerization, (b)
ATRP, and (c) RAFTa
a
Reprinted with permission from ref 277. Copyright 2003 American Chemical Society.
Scheme 59. TGA Analyses of PMMA-g-PDMS with Various
Siloxane Chain Lengthsa
a
Reproduced with permission from ref 331. Copyright 1994 Wiley-VCH
Verlag GmbH & Co. KGaA.
began to degrade at 240 °C, alike PMMA homopolymer,
although at a slower rate as the temperature was raised. The
temperature at 40% weight loss in PDMS-b-PMMA in-
creased with increasing the PDMS content in the copolymers
(from 400 to 420 °C while increasing the Si content from
4.1% to 7.3%).
Uyanik85 studied the thermal properties of a five-block
copolymer PVP-b-PCL-b-PDMS-b-PCL-b-PVP. Thermo-
gravimetric analysis (TGA) clearly showed an increase in
the thermal stability of the copolymer compared to the
homopolymers (PVP and PCL). The decomposition temper-
atures of the homopolymers were all below 440 °C, whereas
for the copolymers, a 5 wt % loss was observed at 170 °C,
and the decomposition temperatures were all around 440 °C.
5.5. Solution Properties
Poly(N-isopropylacrylamide) (PNIPAM) is known to
exhibit a well-defined and reversible lower critical solution
temperature (LCST) in water, around 32-34 °C, which is
close to the body temperature. In a recent study, Uyanik et
al.57 analyzed the solution properties of PNIPAM-b-PDMS
copolymers by measuring the LCST. The LCST was found
to be lower in concentrated solutions (<35 °C at 0.5 wt %)
than in dilute solutions (>35 °C at 0.01 and 0.1 wt %),
presumably a consequence of the presence of the hydropho-
Chemical Reviews, 2010, Vol. 110, No. 3 1269
bic terminal PDMS group. At higher concentrations (>0.5
wt %), a collapse of the macromolecular coils was observed
due to the hydrophobic forces between isopropyl groups.
Thus, colloidally stable and hydrophobically modified
PNIPAMs carrying PDMS chains at either one or two chain
ends exhibited macroscopic solution-phase separation dif-
ferent from homopoly(NIPAM)s.
Bes and co-workers262 evaluated the surfactant properties
of PDMAEMA-b-PDMS-b-PDMAEMA prepared by ATRP
using a PDMS macroinitiator. They carried out surface
tension measurements on aqueous solutions of the triblock
copolymer versus concentrations and pH. It was observed
that the surface tension of the aqueous solution was mostly
reduced at basic pH (around 30-40 mN · m-1 depending on
triblock copolymer concentration). Indeed, protonation of
PDMAEMA at acidic pH induces a lower self-association
of the amphiphilic copolymer: the predominant species in
solution appeared to be unimers, and the surface activity was
very low. On the contrary, at basic pH, the charge density
(degree of protonation) on PDMAEMA is low and the
copolymer behaved as a surfactant. Furthermore, aggregates
with a high hydrodynamic diameter were ascribed to polymer
“hollow vesicles”.
Pinteala et al. studied the conformations of PMAA-b-
PDMS copolymers in water solutions.102,109 In dilute solu-
tions, the prevailing associations were intramolecular asso-
ciations, whereas at higher concentrations, intermolecular
associations also arose, resulting in the formation of complex
assemblies. Moreover, from low to high pH values, the
conformation of the block copolymers also changed from
contracted to expanded assemblies.
Graiver’s44,50,51 PVP-b-PDMS-b-PVP and PDMS-g-PVP
copolymers properties were analyzed in a 5% HCl aqueous
solution containing soap (10 wt %). It was found that the
copolymers have a high antifoaming activity.
5.6. Other Applications
5.6.1. Supercritical CO2
Paisner and DeSimone334 used PS-g-PDMS copolymers
as the starting material to obtain well-ordered mesoporous
low dielectric materials. Actually, a film of copolymer
prepared by spin coating was treated with acetic acid to
1270 Chemical Reviews, 2010, Vol. 110, No. 3
degrade the PDMS chains: it leaves well-defined mesopores
within the PS matrix.
Silicone macromonomers were found to act as steric
stabilizers in MMA,335–339 styrene,339 or VDF340,341 dispersion
polymerization in scCO2. Tai et al.341 showed that the PVDF
molecular weight was much higher in the presence of the
macromonomer than in conventional radical polymerization
without macromonomer (up to 600 000 vs 50 000 g · mol-1,
respectively).
Dinçer et al.342 prepared PDMS-b-PS-b-PDMS triblock
copolymer in scCO2 using a PDMS azo macroinitiator. The
PDMS azo macroinitiator played the role of polymeric
azoinitiator for styrene but not as a steric stabilizer as proven
by the presence of nonstabilized PDMS-b-PS-b-PDMS at the
end of the polymerization.
Okubo et al.258 carried out the ATRP of MMA in dispersion
in scCO2 using a PDMS macroinitiator as an inistab. The
same team observed that the polymerization of styrene237 in
scCO2 was controlled by nitroxide-mediated radical polym-
erization and stabilized using an “inistab” PDMS macroini-
tiator. They also produced PMMA particles by dispersion
polymerization with mercaptopropyl-terminated PDMS
stabilizer.215
5.6.2. Solid Electrolyte
Trapa and co-workers343 prepared a microphase-separated
poly[(oxyethylene) methacrylate]-g-PDMS, a solid electrolyte
material when associated with a lithium salt of high ionic
conductivity. The microphase separation generated solid-like
mechanical properties even though this material was used
at a higher temperature than both components’ Tg’s. In a
previous study, the authors showed that the conductivity was
linked to the Tg of the nonconducting block: the lower the
Tg, the higher the conductivity. Thus, PDMS was found to
be a very good candidate since it is a polymer with one of
the lowest Tg’s. In conclusion, it was outlined that these
copolymers were the best in terms of combination of
properties ever made in the salt-doped materials’ category.
Scheme 60. PDMS Precursors Used for Preparing Block and Direct Graft Copolymersa
a
The routes not reported yet are in italics.
Pouget et al.
5.6.3. Polymer Blends
Kollefrath et al.119,120 used PDMS-g-poly(acrylate) co-
polymers to compatibilize an acrylic and a silicone rubber.
They observed that the compatibilizing effect was higher for
small PDMS grafts and that the size of the acrylic rubber
phases decreased by increasing the percentage of graft
copolymer. A similar strategy was applied by Ochi and
Shimaoka344 for epoxy resin. PDMS-g-PMMA was obtained
by copolymerization of a methacrylate-functionalized PDMS
with MMA. Molecular weights of copolymer segments are
key points of macromolecular compatibilizers for polymer
blends. The authors gained the full benefits of the mac-
romonomer technique by testing various readily tailored
copolymers.
5.6.4. Nanocomposites
Jeong et al.345 synthesized PS-b-PDMS block copolymers
in the presence of organoclay nanocomposites. Increasing
the PDMS content (Mn ) 5000 g · mol-1) resulted in a better
dispersion of organoclay nanocomposites as observed by
X-ray diffraction and TEM. It seems that the organoclay lies
preferentially in the PDMS domains, which may be interest-
ing for the applications.
6. Conclusion
Many synthetic routes have been used for the synthesis
of PDMS-containing copolymers: anionic polymerization,
cationic polymerization, coupling reactions (for instance,
esterification, hydrosilylation), and so on. However, ionic
polymerization requires reagents of high purity and strict
experimental conditions which are problematic for an
industrial process. Even if coupling reactions have given
interesting results, their use with high molecular weight
polymers is challenging due to the low concentration of
reactive groups. The synthesis of PDMS-based copolymers
using radical chemistry has been widely studied in recent
Poly(dimethylsiloxane)-Containing Copolymers
Scheme 61. PDMS Precursor for Inverse Graft Copolymersa
a
The routes not reported yet are in italics.
years. Advantageously, radical polymerization is a versatile
industrial method compatible with a wide range of monomers
and with different media like dispersed aqueous media or
scCO2. Thus, numerous functional PDMS allowed the
synthesis of different architectures such as block, graft,
multiblock, or star copolymers as well as the preparation of
latex particles with a core-shell morphology.
Results described in this review are summarized in
Schemes 60 and 61, where the general structures of the
poly(dimethylsiloxane) precursors and the synthetic pathways
to reach the desired PDMS-containing architectured copoly-
mers are proposed.
Hence, the targeted architecture can be obtained by
selecting the correct polysiloxane macroinitiator or polysi-
loxane macrotransfer agent (monofunctional, difunctional,
or multifunctional to obtain diblock, triblock, or graft
Table 12. Summary of the Architectures of PDMS-containing Copolymers Obtained Using Radical Chemistrya
a
Architectures already obtained using radical chemistry (✓); new conceivable ways for the synthesis of controlled architectures (*); hardly
applicable techniques for obtaining the targeted architectures (dashed compartments); not currently applicable techniques for generating the targeted
architectures (black compartments).
Chemical Reviews, 2010, Vol. 110, No. 3 1271
Scheme 62. General Scheme of the Synthesis of Controlled
Architectures through Atom-Transfer Radical Additiona
a
P(X)n designates a halide-functionnalized polymer chain: n ) 1, m )
1, preparation of diblock copolymers; n ) 1, m ) 2 and n ) 2, m ) 1
preparation of polymer-b-PDMS-b-polymer and PDMS-b-polymer-b-PDMS
triblock copolymers, respectively; if halide is grafted on the P polymer chain,
Polymer-g-PDMS inverse graft copolymers could be obtained with m ) 1
and crosslinked materials could be obtained with m g 2; if double bond is
grafted on the PDMS backbone, PDMS-g-Polymer direct graft copolymers
could be obtained with n ) 1 and crosslinked materials could be obtained
with n g 2. Star copolymers will be obtained with a star precursor.
copolymers, respectively) and polysiloxane macromonomer
(to obtain inverse graft copolymers). Besides, the radical
polymerization technique has to be selected depending on
the desired level of control over the architecture and the cost
of the synthesis. The cheap “radical transfer to PDMS”
method gives a poor control over the number of branches
and the copolymer molecular weights. However, this can be
satisfactory for applications in polymer blends. Core-shell
particles have also been widely prepared by transfer to PDMS
due to their numerous applications in coating modification,
rubber strengthening, or improving the mechanical properties
of thermoplastic polymers. The use of polysiloxane azo
macroinitiators or peroxy macroinitiators allows a better
control of the number of blocks, even if the molecular weight
of the resulting copolymers is seldom controlled. In recent
years, controlled radical polymerization (CRP) techniques
have emerged to allow the precise control of the copolymer
1272 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 63. General Scheme of Macroalkoxyamine Potentially Useable in NMP To Prepare Multiblock Copolymers (example
with hydroxy-TEMPO derivatives)
architecture (number of blocks, branches, and control over
the molecular weight of each block). They all have a certain
number of advantages and drawbacks depending on the
nature of the polymerized monomer, the desired level of
control, or the use of dispersed aqueous media. However,
they allow the synthesis of a wide range of well-defined
copolymer architectures with tuned bulk, solution, surface,
mechanical, and thermal properties.
Table 12 summarizes the architectures already obtained
using radical chemistry (check marks in the table). On the
basis of the progress in radical and controlled radical
polymerization, we also indicated new conceivable ways for
the synthesis of controlled architectures (asterisks in the
table). On the contrary, some techniques are hardly applicable
(dashed compartments in the table) or not currently applicable
for generating some architectures (black compartments in
the table).
Concerning the conventional radical polymerization tech-
niques (polysiloxane macroinitiators, polysiloxane mac-
romonomers, transfer to PDMS, and telomerization), many
architectures cannot be generated. Polysiloxane macroini-
tiators have been widely used, but diblock and star copoly-
mers have not been prepared using this technique yet. The
use of polysiloxane macromonomers is of great interest in
the synthesis of inverse graft copolymers (polymer-g-PDMS)
and cross-linked materials (by using a di- or multifunctionnal
PDMS macromonomer), but the other architectures are hardly
attainable. Starting from the macronomer, a slightly different
route for the synthesis of diblock, triblock, multiblock, and
star copolymers as well as direct PDMS-g-Polymer, inverse
Polymer-g-PDMS and crosslinked materials could lie in the
use of atom-transfer radical addition between halide-func-
tionalized polymers P(X)n (formed by ATRP or ITP/RITP,
for instance) and PDMS macromonomers that would not
homopolymerize (such as allylic derivatives) (Scheme 62).
A similar route based on thiol-ene reactions is also
conceivable. Transfer to PDMS is such an uncontrolled
technique that well-defined architectures could not be gener-
ated. Concerning telomerization, it is quite easy to prepare
macrotelogen and, moreover, some commercial products
exist. This technique permits control of molecular weights
and could advantageously compete with some more expen-
sive techniques such as controlled radical polymerization
methods, but there is still some work to carry out.
The controlled radical polymerization methods have been
very widely studied, and all the architectures could be
generated. Concerning the macroiniferter method, the prin-
cipal issues concern the synthesis of the macroiniferter and
the potential degradation products (CS2 in the case of use of
dithiocarbonyl derivatives). ATRP and ITP cannot be used
to get multiblock copolymers directly during the polymer-
ization. However, a two-step route, for example, through the
atom-transfer radical coupling of triblock copolymers, could
also be a way to reach this architecture. Diblock copolymers
could be easily generated by using a monofunctional mac-
Pouget et al.
Scheme 64. General Structure of the Macrotrithiocarbonate
That Could Be Used in the Synthesis of Multiblock
Copolymers by RAFT
rotransfer agent in RAFT and ITP. Multiblock copolymers
have never been obtained by CRP techniques even if it is
possible by NMP and RAFT. In the first case, it is necessary
to use PDMS chains linked by difunctional alkoxyamine
(Scheme 63). In RAFT, the use of trithiocarbonate (Scheme
64) would give direct access to multiblock copolymers. Star
copolymers could be easily prepared by using a star-
functionnalized PDMS as it has been previously reported in
ATRP. The preparation of a PDMS with grafted alkoxyamine
groups could be a way to obtain direct graft copolymers
(PDMS-g-polymer). Inverse graft copolymers (polymer-g-
PDMS) could be synthesized by copolymerization with
PDMS macromonomers. Finally, cross-linked copolymers
could be prepared by all the techniques provided that a
multifunctional monomer (cross-linker) is used.
To the best of our knowledge, even if many academic and
industrial works have been performed on the synthesis of
well-architectured PDMS-containing copolymers via radical
polymerization and in spite of their obvious potential, to date,
no product is commercially available (unlike polyconden-
sation reactions, e.g., Geniomer). This may be attributed to
the fact that many methods are still in their infancy and need
more investments. Therefore, upcoming developments should
focus on high added value applications such as cosmetics,
biomaterials, membrane technologies or surface modification
of microfluidic devices, and patterning to name a few.346–352
All of these applications require a deeper knowledge of the
structure-property relationship prior to their industrial
developments. Industrial research will also be mandatory to
exploit the properties and extend the range of applications.
7. Abbreviations
VP 4-vinylpyridine
AA acrylic acid
AFR acetophenone formaldehyde resin
AIBN 2,2′-azobisisobutyronitrile
AMPS 2-acrylamido-2-methylpropanesulfonic acid
AN acrylonitrile
ATRP atom-transfer radical polymerization
BDE bond dissociation energy
BFA 2-(N-butyl perfluorooctanefluorosulfonamido) eth-
yl acrylate
bpy bipyridine
BuA butyl acrylate
BuLi butyllithium
BuMA butyl methacrylate
Cex degenerative chain-transfer constant
CRP controlled radical polymerization
Poly(dimethylsiloxane)-Containing Copolymers Chemical Reviews, 2010, Vol. 110, No. 3 1273

CT transfer constant RITP reverse iodine transfer polymerization

CTFE chlorotrifluoroethylene ROP ring-opening polymerization
D3 2,2,4,4,6,6-hexamethylcyclotrisiloxane scCO2 supercritical carbon dioxide
D4 2,2,4,4,6,6,8,8-octamethylcyclotetrasiloxane SEC size exclusion chromatography
D4H 2,4,6,8-tetramethylcyclotetrasiloxane SEM scanning electron microscopy
D4V 2,4,6,8-tetramethyl-2,4,6,8-tetravinylcyclotetrasi- SET single electron transfer
loxane SG1 N-tert-butyl-N-(1-diethylphosphono-2,2-dimethyl-
DMA dynamic mechanical analysis propyl)nitroxide
DMAEMA 2-dimethylaminoethyl methacrylate tBuA tert-butyl acrylate
dMbpy 4,4′-dimethyl-2,2′-bipyridine tBuMA tert-butyl methacrylate
DMSA 3-(dimethoxymethylsilyl)-propyl acrylate TEM transmission electron microscopy
dNbpy 4,4′-di(5-nonyl-2,2′-bipyridine) TEMPO 2,2,6,6-tetramethylpiperidinyl-1-oxy
DPn number-average degree of polymerization TFE tetrafluoroethylene
DSC differential scanning calorimetry Tg glass transition temperature
EA ethyl acrylate TGA thermogravimetric analysis
Ea activation energy TIPNO N-tert-butyl 1-phenyl-2-methylpropylaminoxyl
EEMA 1-ethoxyethyl methacrylate TMSA trimethylsilyl acrylate
EMA ethyl methacrylate TMSEA trimethylsiloxy-ethyl acrylate
EPS expandable poly(styrene) TMSMA trimethylsilyl methacrylate
FOA 1,1-dihydroperfluorooctyl acrylate TPE thermoplastic elastomer
HEA 2-hydroxyethyl acrylate UV ultraviolet
HEMA 2-hydroxyethyl methacrylate VAc vinyl acetate
HEMA-TMS 2-trimethylsilyloxyethyl methacrylate VDF vinylidene fluoride
HFP hexafluoropropylene VDMAz vinyldimethylazlactone
HMTETA 1,1,4,7,10,10-hexamethyltriethylenetetramine XPS X-ray photoelectron spectroscopy
iBnA isobornyl acrylate
IDMAz isopropenyl dimethyl azlactone 8. References
IEM isocyanatoethylmethacrylate (1) Moretto, H.-H.; Schulze, M.; Wagner, G. Ind. Polym. Handbook 2001,
iniferter initiatior transfer termination 3, 1349.
inistab initiator-stabilizer (2) Centre Européen des Silicones; 2002; http://www.silicones-europe.
ITP iodine-transfer polymerization com.
LCST lower critical solution temperature (3) Voronkov, M. G.; Mileshkevich, V. P.; Yuzhelevskii, Y. A. Studies
LTMS lithium trimethylsilanolate in SoViet Science. The Siloxane Bond: Physical Properties and
Chemical Transformations; Plenum Publishing Corporation: New
MAA methacrylic acid York, 1978.
MADIX macromolecular design trough interchange of xan- (4) Hurd, C. B. J. Am. Chem. Soc. 1946, 68, 364.
thates (5) Rochow, E. G.; LeClair, H. G. J. Inorg. Nucl. Chem. 1955, 1, 92.
Me6TREN tris[2-(dimethylamino)ethyl]amine (6) Eaborn, C. Organosilicon Compounds; Academic Press: New York,
MeA methyl acrylate 1960.
MMA methyl methacrylate (7) In The Analytical Chemistry of Silicones; Smith, A. L., Ed.; Chemical
Mn number-average molecular weight Analysis; Wiley Interscience: New York, 1991; Vol. 112.
(8) Noll, W. Chemistry and Technology of Silicones; Academic Press:
MTEMPO 4-methoxy-2,2,6,6-tetramethylpiperidine-1-oxyl New York, 1968.
m-TMI m-isopropenyl-R,R-dimethyl benzyl isocyanate (9) Graiver, D.; Farminer, K. W.; Narayan, R. J. Polym. EnViron. 2003,
Mw weight-average molecular weight 11, 129.
MWD molecular weight distribution (10) Silicone Technology. In Applied Polymer Symposia; Bruins, P. F.,
NaSS sodium styrene sulfonate Ed.; Interscience Publishers: New York, 1970; No. 14.
N,N-DMA N,N-dimethyl acrylamide (11) Lynch, W. Handbook of Silicone Rubber Fabrication; Van Nostrand
NIPAM N-isopropylacrylamide Reinhold Company: New York, 1978.
(12) Polmanteer, K. E. J. Elastoplastics 1970, 2, 165.
NMP nitroxide-mediated polymerization (13) Warrick, E. L.; Pierce, O. R.; Polmanteer, K. E.; Saam, J. C. Rubber
NMR nuclear magnetic resonance Chem. Technol. 1979, 52, 437.
NVP N-vinyl pyrrolidone (14) Lewis, F. M. Rubber Chem. Technol. 1962, 35, 1222.
OEGMA oligo(ethylene glycol) methyl ether methacrylate (15) Bobear, W. J. Rubber Technology, 2nd ed.; Kluwer Academic
PAA poly(acrylic acid) Publishers: Dordrecht, The Netherlands, 1973; p 368.
PAN poly(acrylonitrile) (16) Polmanteer, K. E. Rubber Chem. Technol. 1981, 54, 1051.
PB poly(butadiene) (17) Hunter, M. J. Kautschuk Gummi Kunststoffe 1964, 17, 498.
(18) Lichtenwalner, H. K.; Sprung, M. N. Encycl. Polym. Sci. Technol.
PCL poly(caprolactone) 1970, 12, 464.
PDI polydispersity index (19) Bosman, A. W.; Janssen, H. M.; Van Gemert, G. M. L.; Verstegen,
PDMS poly(dimethylsiloxane) R. M.; Meijer, E. W.; Sijbesma, R. P. (Suprapolix B.V., Netherlands)
PEA poly(ethyl acrylate) WO 2004052963, 2004.
PLA poly(lactic acid) (20) Sijbesma, R. P.; Beijer, F. H.; Brunsveld, L.; Folmer, B. J. B.;
PMAA poly(methacrylic acid) Hirschberg, J. H. K. K.; Lange, R. F. M.; Lowe, J. K. L.; Meijer,
E. W. Science 1997, 278, 1601.
PMDETA N,N,N′,N′,N′′-pentamethyldietylenetriamine (21) Lehn, J. M. Supramolecular Chemistry: Concepts and PerspectiVes;
PMMA poly(methyl methacrylate) VCH: Weinheim, Deutschland, 1995.
PNIPAM poly(N-isopropylacrylamide) (22) Saam, J. C.; Falender, J. R. (Dow Corning Corp., USA) DE 2717192,
PS poly(styrene) 1977.
PtBuMA poly(tert-butyl methacrylate) (23) Falender, J. R.; Saam, J. C.; Mettler, C. M. (Dow Corning Corp.,
PVA poly(vinyl alcohol) USA) DE 2717227, 1977.
PVAc poly(vinyl acetate) (24) Yilgor, I.; McGrath, J. E. AdV. Polym. Sci. 1988, 86, 1.
(25) Belorgey, G.; Sauvet, G. In Silicon-containing polymers; Chojnowski,
PVC poly(vinyl chloride) J., Ed.; Kluwer Academic Publishers: Dordrecht, 2000.
PVP poly(vinyl pyrrolidone) (26) Chojnowski, J.; Cypryk, M.; Fortuniak, W.; Kazmierski, K.; Rozga-
PVPy poly(4-vinylpyridine) Wijas, K.; Scibiorek, M. ACS Symp. Ser. 2003, 838, 12.
RAFT reversible addition-fragmentation chain transfer (27) Abe, Y.; Gunji, T. Prog. Polym. Sci. 2004, 29, 149.
1274 Chemical Reviews, 2010, Vol. 110, No. 3
(28) Embery, C. J.; Matisons, J. G.; Clarke, S. R. ACS Symp. Ser. 2003,
838, 26.
(29) Hay, J. N. Surf. Coat. Int., Part A 2002, 85, 348.
(30) Chojnowski, J.; Cypryk, M. Silicon-Containing Polymers; Kluwer
Academic Publishers: Dordrecht, The Netherlands, 2000.
(31) Yeboah, Y. D. (General Electric Co., USA). US 4423240, 1983.
(32) Graiver, D.; Huebner, D. J.; Saam, J. C. Rubber Chem. Technol. 1983,
56, 918.
(33) Ganachaud, F.; Boileau, S. In Handbook of Ring Opening Polym-
erization; Dubois, P., Coulembier, O., Raquez, J.-M., Eds.; Wiley
VCH: Weinheim, 2009, 65.
(34) Dvornic, P. R.; Govedarica, M. N.; Jovanovic, J. D.; Gerov, V. V.;
Antic, M. P. Polym. Bull. 1995, 35, 539.
(35) Pouget, E. Ph.D. Thesis, University Montpellier 2, 2007.
(36) Kojima, K.; Gore, C. R.; Marvel, C. S. J. Polym. Sci., Part A-1:
Polym. Chem. 1966, 4, 2325.
(37) McGrath, J. E.; Sormani, P. M.; Elsbernd, C. S.; Kilic, S. Makromol.
Chem., Macromol. Symp. 1986, 6, 67.
(38) Yu Xue-Hai, M. R. N.; Grasel, T. G.; Gibson, P. E.; Cooper, S. L.
J. Polym. Sci., Polym. Phys. Ed. 1985, 23, 2319.
(39) Marciniec, B. ComprehensiVe Handbook on Hydrosilylation; Per-
gamon Press: New York, 1992.
(40) Boileau, S.; Bouteiller, L.; Kowalewska, A. Polymer 2003, 44, 6449.
(41) Pouget, E.; Tonnar, J.; Eloy, C.; Lacroix-Desmazes, P.; Boutevin,
B. Macromolecules 2006, 39, 6009.
(42) Inoue, H.; Ueda, A.; Nagai, S. J. Polym. Sci., Part A: Polym. Chem.
1988, 26, 1077.
(43) Inoue, H.; Kohama, S. J. Appl. Polym. Sci. 1984, 29, 877.
(44) Graiver, D.; Khieu, A. Q.; Nguyen, B. T. (Dow Corning Corp., USA).
US 5789516, 1998.
(45) Graiver, D.; Khieu, A. Q.; Nguyen, B. T. (Dow Corning Corp., USA).
US 5739246, 1998.
(46) Graiver, D.; Decker, G. T.; Kim, Y.; Hamilton, F. J.; Harwood, H. J.
Silicon Chem. 2002, 1, 107.
(47) Graiver, D.; Nguyen, B.; Hamilton, F. J.; Kim, Y.; Harwood, H. J.
Polym. Prepr. (Am. Chem. Soc., DiV. Polym. Chem.) 1998, 39, 469.
(48) Graiver, D.; Decker, G. T.; Tselepis, A. J.; Hamilton, F. J.; Harwood,
H. J. Polym. Prepr. (Am. Chem. Soc., DiV. Polym.) Chem. 1999, 40,
146.
(49) Graiver, D.; Nguyen, B.; Hamilton, F. J.; Kim, Y.; Harwood, H. J.
ACS Symp. Ser. 2000, 729, 445.
(50) Kim, Y.; Graiver, D.; Decker, G. T.; Hamilton, F. J.; Harwood, H. J.
ACS Symp. Ser. 2003, 838, 296.
(51) Kim, Y.; Gravier, D.; Decker, G. T.; Hamilton, F. J.; Harwood, H. J.
Polym. Prepr. (Am. Chem. Soc., DiV. Polym. Chem.) 2001, 42, 215.
(52) Jin, S. J.; Arora, P. K.; Sayre, L. M. J. Org. Chem. 1990, 55, 3011.
(53) Sayre, L. M.; Jin, S. J. J. Org. Chem. 1984, 49, 3498.
(54) Holland, T. V.; Goodrich, S. D.; Guo, M.; Harwood, H. J.; Rinaldi,
P. L.; Saito, T. Polym. Prepr. (Am. Chem. Soc., DiV. Polym. Chem.)
1995, 36, 91.
(55) Harwood, H. J.; Christov, L.; Guo, M. M.; Holland, T. V.; Huckstep,
A. Y.; Jones, D. H.; Medsker, R. E.; Rinaldi, P. L.; Saito, T.; Tung,
D. S. Macromol. Symp. 1996, 111, 25.
(56) Harwood, H. J.; Goodrich, S. D. (University of Akron, USA) US
5405913, 1995.
(57) Erbil, C.; Akin, H.; Uyanik, N. J. Appl. Polym. Sci. 2003, 87, 1248.
(58) Chang, T. C.; Chen, H. B.; Ho, S. Y.; Chiu, Y. S. J. Macromol. Sci.,
Chem. 1996, A33, 1263.
(59) Ahn, K. D.; Kim, U. Y.; Kim, C. H. J. Macromol. Sci., Chem. 1986,
A23, 169.
(60) Klemarczyk, P.; Nakos, S. T.; Lin, S. Q. S. (Loctite Corp., USA)
US 4666953, 1987.
(61) Angiolini, L.; Carlini, C. Chim. Ind. 1990, 72, 124.
(62) Pouliquen, L.; Coqueret, X.; Lablache-Combier, A.; Loucheux, C.
Makromol. Chem. 1992, 193, 1273.
(63) Pouliquen, L.; Coqueret, X.; Morlet-Savary, F.; Fouassier, J.-P.
Macromolecules 1995, 28, 8028.
(64) Kinstle, J. F.; Watson, S. L., Jr. J. Rad. Curing 1975, 2, 7.
(65) Mateo, J. L.; Manzarbeitia, J. A.; Sastre, R.; Martinez-Utrilla, R. J.
Photochem. Photobiol., A 1987, 40, 169.
(66) Catalina, F.; Peinado, C.; Sastre, R.; Mateo, J. L.; Allen, N. S. J.
Photochem. Photobiol., A 1989, 47, 365.
(67) Meier, K.; Zweifel, H. J. Photochem. 1986, 35, 353.
(68) Jonsson, S.; Viswanathan, K.; Hoyle, C. E.; Clark, S. C.; Miller, C.;
Morel, F.; Decker, C. Nucl. Instrum. Methods Phys. Res., Sect. B
1999, 151, 268.
(69) Pozos-Vazquez, C. Ph.D. Thesis, University Montpellier 2, 2007.
(70) Graiver, D.; Khieu, A. Q.; Nguyen, B. T. (Dow Corning Corp., USA)
US 5708115, 1998.
(71) Crivello, J. V.; Lee, J. L.; Conlon, D. A. Polym. Prepr. (Am. Chem.
Soc., DiV. Polym. Chem.) 1985, 26, 20.
(72) Crivello, J. V.; Lee, J. L.; Conlon, D. A. AdV. Elastomers Rubber
Elasticity, [Proc. Symp.] 1986, 157.
Pouget et al.
(73) Crivello, J. V.; Lee, J. L.; Conlon, D. A. J. Polym. Sci., Part A:
Polym. Chem. 1986, 24, 1251.
(74) Crivello, J. V.; Conlon, D. A.; Lee, J. L. J. Polym. Sci., Part A:
Polym. Chem. 1986, 24, 1197.
(75) Crivello, J. V. (General Electric Co., USA) US 4675426, 1987.
(76) Crivello, J. V. (General Electric Co., USA) US 4677169, 1987.
(77) Crivello, J. V. (General Electric Co., USA) US 4689289, 1987.
(78) Madec, P.-J.; Marechal, E. J. Polym. Sci.: Polym. Chem. Ed. 1978,
16, 3157.
(79) Madec, P.-J.; Marechal, E. J. Polym. Sci.: Polym. Chem. Ed. 1978,
16, 3165.
(80) Baysal, B. M.; Uyanik, N.; Hamurcu, E. E.; Cvetkovska, M. J. Appl.
Polym. Sci. 1996, 60, 1369.
(81) Cvetkovska, M.; Coseva, S.; Buzarovska, A.; Lazarevic, M.; Baysal,
B. M.; Hamurcu, E. E.; Uyanik, N. Glasnik na Hemicarite i
Tehnolozite na Makedonija (Bulletin of the Chemists and Technolo-
gies of Macedonia) 1996, 15, 15.
(82) Cvetkovska, M.; Koseva, S.; Buzarovska, A.; Baysal, B. M.; Yasar,
B.; Karal-Yilmaz, O.; Popovic, I. G.; Katsikas, L. Macromol. Chem.
Phys. 2000, 201, 685.
(83) Cvetkovska, M.; Koseva, S.; Tasevska, S.; Hamurcu, E. E.; Baysal,
B. M. Glasnik na Hemicarite i Tehnolozite na Makedonija (Bulletin
of the Chemists and Technologies of Macedonia) 1998, 17, 17.
(84) Taskiran, I. J. Appl. Polym. Sci. 2006, 100, 4826.
(85) Uyanik, N. J. Appl. Polym. Sci. 1997, 64, 1961.
(86) Uyanik, N.; Yalcinkaya, H.; Kizilcan, N. Surf. Coat. Int., Part B:
Coat. Trans. 2001, 84, 309.
(87) Uyanik, N.; Koker, B.; Yildiz, Y. J. Appl. Polym. Sci. 1999, 71, 1915.
(88) Saigo, T.; Nakayama, M. (Nippon Oils & Fats Co., Ltd., Japan) JP
63057642, 1988.
(89) Saigo, T.; Nakayama, M. (Nippon Oils & Fats Co., Ltd., Japan) JP
63092628, 1988.
(90) Saigo, T.; Nakayama, M. (Nippon Oils & Fats Co., Ltd., Japan) JP
63057643, 1988.
(91) Hikita, S. (NOF Corp., Japan) JP 2003055419, 2003.
(92) Nuyken, O.; Weidner, R. AdV. Polym. Sci. 1986, 73-74, 145.
(93) Inoue, H.; Ueda, A.; Nagai, S. (Japan) JP 61252230, 1986.
(94) George, M. H.; Ward, J. R. J. Polym. Sci., Polym. Chem. Ed. 1973,
11, 2909.
(95) Ueda, A.; Nagai, S. J. Polym. Sci., Polym. Chem. Ed. 1984, 22, 1611.
(96) Ohata, M. Techno-Cosmos 2002, 15, 8.
(97) Inoue, H.; Ueda, A.; Nagai, S. J. Appl. Polym. Sci. 1988, 35, 2039.
(98) Inoue, H.; Matsumoto, A.; Matsukawa, K.; Ueda, A.; Nagai, S.
J. Appl. Polym. Sci. 1990, 41, 1815.
(99) Nakamura, K.; Fujimoto, K.; Kawaguchi, H. Colloids Surf. A 1999,
153, 195.
(100) Ohata, M.; Mikami, S.; Osugi, H. J. Imaging Sci. Tech. 2001, 45,
24.
(101) Okubo, M.; Fujii, S.; Maenaka, H.; Minami, H. Colloid Polym. Sci.
2002, 280, 183.
(102) Szajdzinska-Pietek, E.; Pinteala, M.; Schlick, S. Polymer 2004, 45,
4113.
(103) Hamurcu, E. E.; Hazer, B.; Misirli, Z.; Baysal, B. M. J. Appl. Polym.
Sci. 1996, 62, 1415.
(104) Hamurcu, E. E.; Hazer, B.; Baysal, B. M. Polymer 1997, 38, 2981.
(105) Murakami, H.; Hirai, S.; Shitani, Y.; Ueda, A.; Inoue, H.; Nagai, S.
(C. I. Kasei Co., Ltd., Japan) JP 04039315, 1992.
(106) Noguchi, T.; Mise, T.; Watanabe, M.; Cho, I. (Showa Highpolymer,
Japan) JP 08003254, 1996.
(107) Simionescu, C. I.; Harabagiu, V.; Comanita, E.; Hamciuc, V.; Giurgiu,
D.; Simionescu, B. C. Eur. Polym. J. 1990, 26, 565.
(108) Simionescu, C. I.; Harabagiu, V.; Cotzur, C.; Simionescu, B. C. ReV.
Roum. Chim. 1994, 39, 319.
(109) Pinteala, M.; Epure, V.; Harabagiu, V.; Simionescu, B. C.; Schlick,
S. Macromolecules 2004, 37, 4623.
(110) Simionescu, C. I.; Harabagiu, V.; Giurgiu, D.; Hamciuc, V.;
Simionescu, B. C. Bull. Soc. Chim. Belg. 1990, 99, 991.
(111) Chang, T. C.; Chen, Y. C.; Ho, S. Y.; Chiu, Y. S. Polymer 1996, 37,
2963.
(112) Chang, T. C.; Chen, H. B.; Chen, Y. C.; Ho, S. Y. J. Polym. Sci.,
Part A: Polym. Chem. 1996, 34, 2613.
(113) Chang, T. C.; Chen, H. B.; Chiu, Y. S.; Ho, S. Y. J. Polym. Sci.,
Part A: Polym. Chem. 1996, 34, 3313.
(114) Harabagiu, V.; Hamciuc, V.; Giurgiu, D.; Simionescu, B. C.;
Simionescu, C. I. Makromol. Chem., Rapid Commun. 1990, 11, 433.
(115) Bertolucci, M.; Galli, G.; Chiellini, E.; Wynne, K. J. Macromolecules
2004, 37, 3666.
(116) Deniz, S.; Baran, N.; Akguen, M.; Akguen, N. A.; Dincer, S. Turkish
J. Chem. 2004, 28, 645.
(117) Inoue, H.; Matsumoto, A.; Matsukawa, K.; Ueda, A.; Nagai, S.
J. Appl. Polym. Sci. 1990, 40, 1917.
(118) Shiraki, K. (Wako Pure Chemical Industries, Ltd., Japan) JP
2000053716, 2000.
Poly(dimethylsiloxane)-Containing Copolymers
(119) Kollefrath, R.; Nuyken, O.; Voit, B.; Dauth, J.; Deubzer, B.;
Hierstetter, T.; Weis, J. Polym. Prepr. (Am. Chem. Soc., DiV. Polym.
Chem.) 1996, 37, 407.
(120) Kollefrath, R.; Nuyken, O.; Voit, B.; Dauth, J.; Deubzer, B.;
Hierstetter, T.; Weis, J. Gummi, Fasern, Kunststoffe 1997, 50, 193.
(121) Engel, P. S. Chem. ReV. 1980, 80, 99.
(122) Laverty, J. J.; Gardlund, Z. G. J. Polym. Sci., Polym. Chem. Ed. 1977,
15, 2001.
(123) Schuh, H. H.; Fischer, H. HelV. Chim. Acta 1978, 61, 2130.
(124) Schuh, H.; Fischer, H. Int. J. Chem. Kinet. 1976, 8, 341.
(125) Walbiner, M.; Wu, J. Q.; Fischer, H. HelV. Chim. Acta 1995, 78,
910.
(126) Weiner, S. A.; Hammond, G. S. J. Am. Chem. Soc. 1969, 91, 986.
(127) Roth, M. J. Polym. Sci., Part B: Polym. Phys. 1990, 28, 2715.
(128) Catalgil-Giz, H.; Giz, A.; Oncul-Koc, A. Polym. Bull. 1999, 43, 215.
(129) Hensley, D. R.; Goodrich, S. D.; Harwood, H. J.; Rinaldi, P. L.
Macromolecules 1994, 27, 2351.
(130) Nuyken, O.; Dyckerhoff, L.; Kerber, R. Angew. Makromol. Chem.
1986, 143, 11.
(131) Debuigne, A.; Caille, J.-R.; Jerome, R. Angew. Chem., Int. Ed. Engl.
2005, 44, 1101.
(132) Wayland, B. B.; Poszmik, G.; Mukerjee, S. L.; Fryd, M. J. Am. Chem.
Soc. 1994, 116, 7943.
(133) Lu, Z.; Fryd, M.; Wayland, B. B. Macromolecules 2004, 37, 2686.
(134) Wayland, B. B.; Fu, X.; Lu, Z.; Fryd, M. Polym. Prepr. (Am. Chem.
Soc., DiV. Polym. Chem.) 2005, 46, 370.
(135) Yamago, S. Proc. Jpn. Acad., Ser. B: Phys. Biol. Sci. 2005, 81, 117.
(136) Yamago, S.; Ray, B.; Kameshima, T.; Kawano, K. (Otsuka Chemical
Co., Ltd., Japan) WO 2006001496, 2006.
(137) Yamago, S.; Ray, B.; Iida, K.; Yoshida, J.; Tada, T.; Yoshizawa,
K.; Kwak, Y.; Goto, A.; Fukuda, T. J. Am. Chem. Soc. 2004, 126,
13908.
(138) Lacroix-Desmazes, P.; Severac, R.; Otazaghine, B.; Boutevin, B.
Polym. Prepr. (Am. Chem. Soc., DiV. Polym. Chem.) 2003, 44, 683.
(139) Lacroix-Desmazes, P.; Severac, R.; Boutevin, B. Macromolecules
2005, 38, 6299.
(140) Boyer, C.; Lacroix-Desmazes, P.; Robin, J.-J.; Boutevin, B. Mac-
romolecules 2006, 39, 4044.
(141) David, G.; Boyer, C.; Tonnar, J.; Ameduri, B.; Lacroix-Desmazes,
P.; Boutevin, B. Chem. ReV. 2006, 106, 3936.
(142) Tonnar, J.; Lacroix-Desmazes, P.; Boutevin, B. ACS Symp. Ser. 2006,
944, 604.
(143) Tonnar, J.; Lacroix-Desmazes, P.; Boutevin, B. Macromol. Rapid
Commun. 2006, 27, 1733.
(144) Greber, G.; Reese, E. Makromol. Chem. 1962, 55, 96.
(145) Kawakami, Y.; Miki, Y.; Tsuda, T.; Murthy, R. A. N.; Yamashita,
Y. Polym. J. 1982, 14, 913.
(146) Kawakami, Y.; Murthy, R. A. N.; Yamashita, Y. Makromol. Chem.
1984, 185, 9.
(147) DeSimone, J. M.; Hellstern, A. M.; Siochi, E. J.; Smith, S. D.; Ward,
T. C.; Gallagher, P. M.; Krukonis, V. J.; McGrath, J. E. Makromol.
Chem., Macromol. Symp. 1990, 32, 21.
(148) Aoyagi, T.; Takamura, Y.; Nakamura, T.; Nagase, Y. Polymer 1992,
33, 1530.
(149) Cameron, G. G.; Chisholm, M. S. Polymer 1985, 26, 437.
(150) Tezuka, Y.; Fukushima, A.; Imai, K. Makromol. Chem. 1985, 186,
685.
(151) Tezuka, Y.; Nobe, S.; Shiomi, T. Macromolecules 1995, 28, 8251.
(152) Maynard, H. D.; Lyu, S. P.; Fredrickson, G. H.; Wudl, F.; Chmelka,
B. F. Polymer 2001, 42, 7567.
(153) Yu, X.; Nagarajan, M. R.; Li, C.; Speckhard, T. A.; Cooper, S. L.
J. Appl. Polym. Sci. 1985, 30, 2115.
(154) Mazurek, M.; Kinning, D. J.; Kinoshita, T. J. Appl. Polym. Sci. 2001,
80, 159.
(155) Tenhu, H.; Heino, E.-L. J. Appl. Polym. Sci. 1992, 44.
(156) O’Shea, M.; George, G. A. Polymer 1994, 35, 4190.
(157) O’Shea, M. S.; George, G. A. Polymer 1994, 35, 4181.
(158) Kunzler, J.; Ozark, R. J. Appl. Polym. Sci. 1997, 65, 1081.
(159) Suzuki, T.; Lo, P. Y. Macromolecules 1991, 24, 460.
(160) Suzuki, T.; Okawa, T. Polym. Comm. 1988, 29, 225.
(161) Suzuki, T.; Okawa, T. Polymer 1988, 29, 2095.
(162) Wang, W.; Irvine, D. J.; Howdle, S. M. Ind. Eng. Chem. Res. 2005,
44, 8654.
(163) Mayo, F. R.; Lewis, F. M. J. Am. Chem. Soc. 1944, 66, 1594.
(164) Fineman, M.; Ross, S. D. J. Polym. Sci. 1950, 5, 259.
(165) Kelen, T.; Tudos, F. J. Macromol. Sci., Chem. 1975, 9, 1.
(166) Cameron, G. G.; Chisholm, M. S. Polymer 1986, 27, 1420.
(167) DeSimone, J. M.; Hellstern, A. M.; Ward, T. C.; McGrath, J. E.;
Smith, S. D.; Gallagher, P. M.; Krukonis, V. J.; Stejskal, J.; Strakova,
D.; Kratochvil, P. Polym. Prepr. (Am. Chem. Soc., DiV. Polym.
Chem.) 1988, 29, 116.
Chemical Reviews, 2010, Vol. 110, No. 3 1275
(168) DeSimone, J. M.; Smith, S. D.; Hellstern, A. M.; Ward, T. C.;
McGrath, J. E.; Gallagher, P. M.; Krukonis, V. J. Polym. Prepr. (Am.
Chem. Soc., DiV. Polym. Chem.) 1988, 29, 361.
(169) Stajeskal, J.; Strakova, D.; Kratochvil, P.; Smith, S. D.; McGrath,
J. E. Macromolecules 1989, 22, 861.
(170) Kawai, T.; Akashima, M.; Teramachi, S. Polymer 1995, 36, 2851.
(171) Dluzneski, P. R. Rubber Chem. Technol. 2001, 74, 451.
(172) Dunham, M. L.; Bailey, D. L.; Mixer, R. Y. Ind. Eng. Chem. 1957,
49, 1373.
(173) Baquey, G. Ph.D. Thesis, Université de Bordeaux I, 2004.
(174) Baquey, G.; Moine, L.; Babot, O.; Degueil, M.; Maillard, B. Polymer
2005, 46, 6283.
(175) Baquey, G.; Moine, L.; Degueil-Castaing, M.; Lartigue, J.-C.;
Maillard, B. Macromolecules 2005, 38, 9571.
(176) Saam, J. C.; Lindsey, S. E. (Dow Corning Corp.) DE 2321904, 1973.
(177) Dong, J.; Liu, Z.; Han, N.; Wang, Q.; Xia, Y. J. Appl. Polym. Sci.
2004, 92, 3542.
(178) Getson, J. C. (SWS Silicones Corp., USA) EP 2744, 1979.
(179) Getson, J. C. (SWS Silicones Corp., USA) US 4172101, 1979.
(180) Neuroth, C. G. (Stauffer-Wacker Silicone Corp.) US 3565851, 1971.
(181) Getson, J. C. (Stauffer-Wacker Silicone Corp.) DE 1957257, 1970.
(182) Getson, J. C.; Adams, P. J. (SWS Silicones Corp., USA) UE 4123472,
1978.
(183) Getson, J. C. (Stauffer Chemical Co.) FR 1579295, 1969.
(184) Getson, J. C. (Stauffer-Wacker Silicone Corp.) DE 2023988, 1971.
(185) Getson, J. C. (Stauffer Chemical Co.) DE 1912671, 1969.
(186) Okaniwa, M.; Ohta, Y. Polym. Prepr. (Am. Chem. Soc., DiV. Polym.
Chem.) 1997, 38, 673.
(187) Okaniwa, M.; Ohta, Y. J. Polym. Sci., Part A: Polym. Chem. 1997,
35, 2607.
(188) Okaniwa, M. Polymer 2000, 41, 453.
(189) Byrzynski, A. J.; Martin, R. E. (Owens) US 3,449,293, 1969.
(190) Hilliard, J. R. (Dow Corning Corp., USA) BE 816210, 1974.
(191) Thomas, R. N. (Dow Corning Corp.) US 3,575,910, 1971.
(192) He, W. D.; Tao, C. T.; Pan, C. Y. J. Appl. Polym. Sci. 1996, 61,
383.
(193) He, W. D.; Tao, C. T.; Pan, C. Y. Polym. Int. 1996, 39, 31.
(194) He, W. D.; Pan, C. Y. J. Appl. Polym. Sci. 2001, 80, 2752.
(195) Kong, X. Z.; Ruckenstein, E. J. Appl. Polym. Sci. 1999, 73, 2235.
(196) Dai, Q.; Zhang, Z.; Wang, F.; Liu, J. J. Appl. Polym. Sci. 2003, 88,
2732.
(197) Zou, M.; Wang, S.; Zhang, Z.; Ge, X. Eur. Polym. J. 2005, 41, 2602.
(198) Zou, M.; Huang, F.; Nie, J.; Zhang, Z.; Ge, X. Polym. Int. 2005, 54,
861.
(199) Lin, M.; Chu, F.; Guyot, A.; Putaux, J.-L.; E., B.-L. Polymer 2005,
46, 1331.
(200) Kong, X. Z.; Kan, C. Y.; Yuan, Q. Polym. AdV. Technol. 1996, 7,
888.
(201) Kan, C. Y.; Kong, X. Z.; Yuan, Q.; Liu, D. S. J. Appl. Polym. Sci.
2001, 80, 2251.
(202) Zou, M.; Zhang, Z.; Shen, X.; Nie, J.; Ge, X. Radiat. Phys. Chem.
2005, 74, 323.
(203) Zou, M.; Zhang, Z.; He, W. D.; Ge, X.; Fan, F. Polym. Int. 2004,
53, 1033.
(204) Boutevin, B. J. Polym. Sci., Part A: Polym. Chem. 2000, 38, 3235.
(205) Ameduri, B.; Boutevin, B.; Gramain, P. AdV. Polym. Sci. 1997, 127,
87.
(206) Tezuka, Y.; Imai, K. Makromol. Chem., Rapid Commun. 1984, 5,
559.
(207) Cekada, J., Jr.; Weyenberg, D. R. (Dow Corning Corp.) FR 1580752,
1969.
(208) Hyde, J. F.; Wehrly, J. R. (Dow Corning Corp.) US 2891920, 1959.
(209) Findlay, D. E.; Weyenberg, D. R. (Dow Corning Corp.) US 3294725,
1966.
(210) Falender, J. R.; Saam, J. C. (Dow Corning Corp., USA) US 4070414,
1978.
(211) Falender, J. R.; Mettler, C. M.; Saam, J. C. (Dow Corning Corp.,
USA) US 4071577, 1978.
(212) John, C.; Saam, C. H. T. J. Appl. Polym. Sci. 1974, 18, 2279.
(213) Pietrasanta, Y.; Fleury, E.; Boutevin, B.; Sarraf, L. Polym. Bull. 1986,
15, 107.
(214) Fawcett, A. H.; Foster, A. B.; Hania, M.; Hohn, M.; McCaffery, G. O.;
Mazebedi, J. L.; Mullen, E. Polym. Prepr. (Am. Chem. Soc., DiV.
Polym. Chem.) 2001, 42, 213.
(215) Fujii, S.; Minami, H.; Okubo, M. Colloid Polym. Sci. 2004, 282,
569.
(216) Braunecker, W. A.; Matyjaszewski, K. Prog. Polym. Sci. 2007, 32,
93.
(217) Matyjaszewski, K. ACS Symp. Ser. 1998, 685, 2.
(218) Goto, A.; Fukuda, T. Prog. Polym. Sci. 2004, 29, 329.
(219) Otsu, T. J. Polym. Sci., Part A: Polym. Chem. 2000, 38, 2121.
(220) Hawker, C. J.; Bosman, A. W.; Harth, E. Chem. ReV. 2001, 101,
3661.
1276 Chemical Reviews, 2010, Vol. 110, No. 3
(221) Matyjaszewski, K.; Xia, J. Chem. ReV. 2001, 101, 2921.
(222) Rizzardo, E.; Chiefari, J.; Mayadunne, R.; Moad, G.; Thang, S.
Macromol. Symp. 2001, 174, 209.
(223) Clouet, G.; Kayser, D. J. Polym. Sci., Part A: Polym. Chem. 1993,
31, 3387.
(224) Clouet, G.; Kayser, D. J. Polym. Sci., Part A: Polym. Chem. 1994,
32, 1423.
(225) Clouet, G.; Kayser, D.; Leising, F. (Rhone-Poulenc Chimie SA,
France) FR 2679912, 1993.
(226) Clouet, G. (Rhone-Poulenc Chimie SA, France) EP 421894, 1991.
(227) Kayser, D.; Chaumont, P.; Nair, C. P. R.; Clouet, G. Polym. Prepr.
(Am. Chem. Soc., DiV. Polym. Chem.) 1993, 34, 534.
(228) Nair, C. P. R.; Clouet, G. Macromolecules 1990, 23, 1361.
(229) Kumar, R. C.; Kantner, S. S. (Minnesota Mining and Manufacturing
Co., USA) EP 413457, 1991.
(230) Kumar, R. C.; Andrus, M. H., Jr.; Dueltgen, R. R.; Mazurek, M. H.
Polym. Prepr. (Am. Chem. Soc., DiV. Polym. Chem. 1994, 35, 786.
(231) Kumar, R. C.; Andrus, M. H., Jr.; Mazurek, M. H. (Minnesota Mining
and Manufacturing Co., USA). EP 413550, 1991.
(232) Betts, D. E.; Johnson, T.; Leroux, D.; Desimone, J. M. ACS Symp.
Ser. 1998, 685, 418.
(233) Fischer, H. Chem. ReV. 2001, 101, 3581.
(234) Lutz, J.-F.; Lacroix-Desmazes, P.; Boutevin, B. Macromol. Rapid
Commun. 2001, 22, 189.
(235) Yoshida, E. Kobunshi Ronbunshu 2000, 57, 484.
(236) Yoshida, E.; Tanimoto, S. Macromolecules 1997, 30, 4018.
(237) Ryan, J.; Aldabbagh, F.; Zetterlund, P. B.; Okubo, M. Polymer 2005,
46, 9769.
(238) Miura, Y.; Hirota, K.; Moto, H.; Yamada, B. Macromolecules 1999,
32, 8356.
(239) Miura, Y.; Sakai, Y.; Taniguchi, I. Polymer 2002, 44, 603.
(240) Singer, D. U.; Morgan, A. M. Polym. Prepr. (Am. Chem. Soc., DiV.
Polym. Chem.) 1999, 40, 370.
(241) Morgan, A. M.; Pollack, S. K.; Beshah, K. Polym. Prepr. (Am. Chem.
Soc., DiV. Polym. Chem.) 2001, 42, 132.
(242) Morgan, A. M.; Pollack, S. K.; Beshah, K. Macromolecules 2002,
35, 4238.
(243) Miura, Y.; Miyake, K. J. Polym. Sci., Part A: Polym. Chem. 2005,
43, 6153.
(244) Adamas, P. J.; Lewis, R. N. (Stauffer-Wacker Silicone Corp.) DE
2000348, 1970.
(245) Charleux, B.; Nicolas, J.; Guerret, O. Macromolecules 2005, 38, 5485.
(246) Guillaneuf, Y.; Gigmes, D.; Marque, S. R. A.; Astolfi, P.; Greci, L.;
Tordo, P.; Bertin, D. Macromolecules 2007, 40, 3108.
(247) Matyjaszewski, K. Nonlinear Opt., Quantum Opt. 2003, 30, 167.
(248) Pyun, J.; Xia, J.; Matyjaszewski, K. ACS Symp. Ser. 2003, 838, 273.
(249) Matyjaszewski, K.; Miller, P. J.; Kickelbick, G.; Nakagawa, Y.;
Diamanti, S.; Pacis, C. ACS Symp. Ser. 2000, 729, 270.
(250) Matyjaszewski, K.; Miller, P. J.; Fossum, E.; Nakagawa, Y. Appl.
Organomet. Chem. 1998, 12, 667.
(251) Pyun, J.; Matyjaszewski, K. Chem. Mater. 2001, 13, 3436.
(252) Matyjaszewski, K.; Lutz, J.-F.; Shinoda, H. (Carnegie Mellon
University, USA) WO 2002068485, 2002.
(253) Peng, H.; Cheng, S.; Fan, Z. J. Appl. Polym. Sci. 2004, 92, 3764.
(254) Peng, H.; Cheng, S.; Feng, L.; Fan, Z. Gaofenzi Xuebao 2003, 480.
(255) Peng, H.; Yan, W.; Wang, J.; Yang, T.-t.; Cheng, S.-y. Youjigui
Cailiao 2005, 19, 9.
(256) Huan, K.; Bes, L.; Haddleton, D. M.; Khoshdel, E. J. Polym. Sci.,
Part A: Polym. Chem. 2001, 39, 1833.
(257) Zhang, J. X.; Varshney, S. K. Des. Monomers Polym. 2002, 5, 79.
(258) Minami, H.; Kagawa, Y.; Kuwahara, S.; Shigematsu, J.; Fujii, S.;
Okubo, M. Des. Monomers Polym. 2004, 7, 553.
(259) Kuwahara, S.; Shigematsu, J.; Kagawa, Y.; Minami, H.; Okubo, M.
Koen Yoshishu-Nippon Setchaku Gakkai Nenji Taikai 2005, 43, 79.
(260) Bes, L.; Huan, K.; Khoshdel, E.; Haddleton, D. M. Polym. Prepr.
(Am. Chem. Soc., DiV. Polym. Chem.) 2001, 42, 134.
(261) Bes, L.; Huan, K.; Khoshdel, E.; Lowe, M. J.; McConville, C. F.;
Haddleton, D. M. Eur. Polym. J. 2002, 39, 5.
(262) Bes, L.; Huan, K.; Haddleton, D. M.; Khoshdel, E. ACS Symp. Ser.
2003, 838, 260.
(263) Peng, H.; Cheng, S.; Fan, Z. J. Appl. Polym. Sci. 2004, 92, 532.
(264) Kurjata, J.; Chojnowski, J.; Yeoh, C.-T.; Rossi, N. A. A.; Holder,
S. J. Polymer 2004, 45, 6111.
(265) Miller, P. J.; Matyjaszewski, K. Macromolecules 1999, 32, 8760.
(266) Hong, S. C.; Lutz, J.-F.; Inoue, Y.; Strissel, C.; Nuyken, O.;
Matyjaszewski, K. Macromolecules 2003, 36, 1075.
(267) Pyun, J.; Miller, P. J.; Matyjaszewski, K. Polym. Prepr. (Am. Chem.
Soc., DiV. Polym. Chem.) 2000, 41, 536.
(268) Pyun, J.; Jia, S.; Kowalewski, T.; Matyjaszewski, K. Macromol.
Chem. Phys. 2004, 205, 411.
(269) Huan, K.; Khoshdel, E.; Haddleton, D. M. Polym. Prepr. (Am. Chem.
Soc., DiV. Polym. Chem.) 2000, 41, 538.
Pouget et al.
(270) Nakagawa, Y.; Miller, P. J.; Matyjaszewski, K. Polymer 1998, 39,
5163.
(271) Nakagawa, Y.; Matyjaszewski, K. Polym. Prepr. (Am. Chem. Soc.,
DiV. Polym. Chem.) 1996, 37, 270.
(272) Pyun, J.; Jia, S.; Kowalewski, T.; Savin, D. A.; Patterson, G. D.;
Liu, T.; Matyjaszewski, K. Polym. Prepr. (Am. Chem. Soc., DiV.
Polym. Chem.) 2002, 43, 328.
(273) Brown, D. A.; Price, G. J. Polymer 2001, 42, 4767.
(274) Nakagawa, Y.; Miller, P.; Pacis, C.; Matyjaszewski, K. Polym. Prepr.
(Am. Chem. Soc., DiV. Polym. Chem.) 1997, 38, 701.
(275) Shinoda, H.; Matyjaszewski, K. Polym. Prepr. (Am. Chem. Soc., DiV.
Polym. Chem.) 2001, 42, 161.
(276) Shinoda, H.; Miller, P. J.; Matyjaszewski, K. Macromolecules 2001,
34, 3186.
(277) Shinoda, H.; Matyjaszewski, K.; Okrasa, L.; Mierzwa, M.; Pakula,
T. Macromolecules 2003, 36, 4772.
(278) Lutz, J.-F.; Jahed, N.; Matyjaszewski, K. J. Polym. Sci., Part A:
Polym. Chem. 2004, 42, 1939.
(279) Hong, S. C.; Neugebauer, D.; Inoue, Y.; Lutz, J.-F.; Matyjaszewski,
K. Macromolecules 2003, 36, 27.
(280) Neugebauer, D.; Zhang, Y.; Pakula, T.; Matyjaszewski, K. Macro-
molecules 2005, 38, 8687.
(281) Matyjaszewski, K.; Beers, K. L.; Coca, S.; Gaynor, S. G.; Miller,
P. J.; Paik, H.-J.; Teodorescu, M. Polym. Prepr. (Am. Chem. Soc.,
DiV. Polym. Chem.) 1999, 40, 95.
(282) Taran, E.; Donose, B.; Higashitani, K.; Asandei, A. D.; Scutaru, D.;
Hurduc, N. Eur. Polym. J. 2005, 42, 119.
(283) Xiao, D.; Zhang, H.; Wirth, M. Langmuir 2002, 18, 9971.
(284) Matyjaszewski, K.; Miller, P. J.; Pyun, J.; Kickelbick, G.; Diamanti,
S. Macromolecules 1999, 32, 6526.
(285) Kickelbick, G.; Miller, P. J.; Matyjaszewski, K. Polym. Prepr. (Am.
Chem. Soc., DiV. Polym. Chem.) 1998, 39, 284.
(286) Tatemoto, M.; Nakagawa, T. (Daikin Kogyo Co., Ltd., Japan) DE
2729671, 1978.
(287) Tatemoto, M.; Suzuki, T.; Tomoda, M.; Furukawa, Y.; Ueta, Y.
(Daikin Kogyo Co., Ltd., Japan) DE 2815187, 1978.
(288) Shinya, S.; Kojima, G. Asahi Garasu Kenkyu Hokoku 1989, 39, 109.
(289) Kojima, G.; Shintani, S. (Asahi Glass Co., Ltd., Japan) JP 62280225,
1987.
(290) Tonnar, J.; Pouget, E.; Lacroix-Desmazes, P.; Boutevin, B. Eur.
Polym. J. 2008, 44, 318.
(291) Boutevin, B.; Hugon, J. P.; Pietrasanta, Y.; Sideris, A. Eur. Polym.
J. 1978, 14, 353.
(292) Iovu, M. C.; Matyjaszewski, K. Macromolecules 2003, 36, 9346.
(293) Sakamoto, M.; Kusumoto, S.; Nishikawa, A.; Sato, H. (Japan
Synthetic Rubber Co Ltd, Japan) JP 07018036, 1995.
(294) Lacroix-Desmazes, P.; Severac, R.; Boutevin, B.; Bodard, V.;
Kurowsky, V. (Solvay SA, Belg.) FR 2839724, 2003.
(295) Lacroix-Desmazes, P.; Severac, R.; Boutevin, B.; Bodart, V.;
Kurowski, V. (Solvay Societe Anonyme, Belg.) WO 2004094356,
2004.
(296) Tonnar, J.; Lacroix-Desmazes, P.; Boutevin, B. Polym. Prepr. (Am.
Chem. Soc., DiV. Polym. Chem.) 2005, 46, 280.
(297) Lacroix-Desmazes, P.; Tonnar, J.; Boutevin, B. Macromol. Symp.
2007, 248, 150.
(298) Tonnar, J.; Lacroix-Desmazes, P. Angew. Chem., Int. Ed. 2008, 47,
1294.
(299) Tonnar, J.; Lacroix-Desmazes, P. Soft Matter 2008, 4, 1255.
(300) Monteiro, M. J. J. Polym. Sci., Part A: Polym. Chem. 2005, 43, 3189.
(301) Favier, A.; Charreyre, M.-T. Macromol. Rapid Commun. 2006, 27,
653.
(302) Destarac, M. (Rhodia Chimie, Fr.) WO 2002022688, 2002.
(303) Destarac, M.; Brochon, C.; Catala, J.-M.; Wilczewska, A.; Zard, S. Z.
Macromol. Chem. Phys. 2002, 203, 2281.
(304) Pai, T. S. C.; Barner-Kowollik, C.; Davis, T. P.; Stenzel, M. H.
Polymer 2004, 45, 4383.
(305) Karagianni, K.; Anthony, O.; Destarac, M. (Rhodia Chimie, France)
FR 2853324, 2004.
(306) Wilczewska, Z. A.; Destarac, M.; Zard, S.; Kalai, C.; Mignani, G.;
Adam, H. (Rhodia Chimie, France) WP 2002090397, 2002.
(307) Wilczewska, Z. A.; Destarac, M.; Zard, S.; Kalai, C.; Mignani, G.;
Adam, H. (Rhodia Chimie, France) WO 2002090424, 2002.
(308) Destarac, M.; Mignani, G.; Zard, S.; Sire, B.; Kalai, C. (Rhodia
Chimie, France) WO 2002008307, 2002.
(309) Destarac, M.; Kalai, C.; Wilczewska, A. Z.; Mignani, G.; Zard, S. Z.
Polym. Prepr. (Am. Chem. Soc., DiV. Polym. Chem.) 2005, 46, 213.
(310) Destarac, M.; Kalai, C.; Petit, L.; Wilczewska, A. Z.; Mignani, G.;
Zard, S. Z. Polym. Prepr. (Am. Chem. Soc., DiV. Polym. Chem.) 2005,
46, 372.
(311) Destarac, M.; Bzducha, W.; Zard, S. Z. Polym. Prepr. (Am. Chem.
Soc., DiV. Polym. Chem.) 2005, 46, 387.
(312) Destarac, M.; Kalai, C.; Wilczewska, A.; Petit, L.; Van Gramberen,
E.; Zard, S. ACS Symp. Ser. 2006, 944, 564.
Poly(dimethylsiloxane)-Containing Copolymers
(313) Shinoda, H.; Matyjaszewski, K. Macromol. Rapid Commun. 2001,
22, 1176.
(314) Jaacks, V. Makromol. Chem. 1972, 161, 161.
(315) Moad, G.; Rizzardo, E.; Thang, S. H. Aust. J. Chem. 2006, 59, 669.
(316) Hillmyer, M. A. J. Polym. Sci.: Part B: Polym. Phys. 2007, 45, 3249.
(317) Cameron, G. G.; Chisholm, M. S. Polymer 1986, 27, 437.
(318) Gordon, M.; Tayor, J. S. J. Appl. Chem. 1952, 2, 493.
(319) Blahovici, T. F.; Brown, G. R.; St. Pierre, L. E. Polym. Eng. Sci.
1982, 22, 1123.
(320) Smith, S. D.; DeSimone, J. M.; Huang, H.; York, G.; Dwight, D. W.;
Wilkes, G. L.; McGrath, J. E. Macromolecules 1992, 25, 2575.
(321) Smith, S. D.; York, G.; Dwight, D. W.; McGrath, J. E.; Stejskal, J.;
Kratochvil, P. Polym. Prepr. (Am. Chem. Soc., DiV. Polym. Chem.)
1987, 28, 458.
(322) Smith, S. D.; DeSimone, J. M.; York, G.; Dwight, D. W.; Wilkes,
G. L.; McGrath, J. E. Polym. Prepr. (Am. Chem. Soc., DiV. Polym.
Chem.) 1987, 28, 150.
(323) Kawakami, Y.; Murthy, R. A.; Yamashita, Y. Polym. Bull. 1983,
10, 368.
(324) Kawakami, Y.; Yamashita, Y. ACS Symp. Ser. 1985, 286, 245.
(325) Smith, S. D.; McGrath, J. E. Polym. Prepr. (Am. Chem. Soc., DiV.
Polym. Chem.) 1986, 27, 31.
(326) Lee, Y.; Akiba, I.; Akiyama, S. J. Appl. Polym. Sci. 2002, 86, 1736.
(327) Taskiran, I.; Ezdesir, A.; Sunal, G.; Uyanik, N. J. Appl. Polym. Sci.
2006, 101, 128.
(328) Hou, Y.; Tulevski, G. S.; Valint, P. L., Jr.; Gardella, J. A., Jr.
Macromolecules 2002, 35, 5953.
(329) Hou, Y.; Valint, P. L., Jr.; Gardella, J. A., Jr. Polym. Mater. Sci.
Eng. 2001, 85, 305.
(330) Kawakami, Y.; Aoki, T.; Yamashita, Y.; Hirose, M.; Ishitani, A.
Macromolecules 1985, 18, 580.
(331) Smith, S. D.; Long, T. E.; McGrath, J. E. J. Polym. Sci., Part A:
Polym. Chem. 1994, 32, 1747.
(332) Giannetti, E.; Mazzocchi, R.; Fiore, L.; Visani, F. J. Polym. Sci.,
Part A: Polym. Chem. 1986, 24, 2517.
Chemical Reviews, 2010, Vol. 110, No. 3 1277
(333) Kashiwagi, T.; Inaba, A.; Brown, J. E.; Hatada, K.; Kitayama, T.;
Masuda, E. Macromolecules 1986, 19, 2160.
(334) Paisner, S. N.; DeSimone, J. M. PMSE Preprints 2002, 87, 446.
(335) O’Neill, M. L.; Yates, M. Z.; Johnston, K. P.; Smith, C. D.;
Wilkinson, S. P. Macromolecules 1998, 31, 2838.
(336) Fehrenbacher, U.; Ballauff, M.; Muth, O.; Hirth, T. Appl. Organomet.
Chem. 2001, 15, 613.
(337) Wang, D.; DeSimone, J. M. Polym. Mater. Sci. Eng. 1999, 80, 526.
(338) O’Neill, M. L.; Yates, M. Z.; Johnston, K. P.; Smith, C. D.;
Wilkinson, S. P. Macromolecules 1998, 31, 2838.
(339) Shaffer, K. A.; Jones, T. A.; Canelas, D. A.; DeSimone, J. M.;
Wilkinson, S. P. Macromolecules 1996, 29, 2704.
(340) Tai, H.; Liu, J.; Howdle, S. M. Eur. Polym. J. 2005, 41, 2544.
(341) Tai, H.; Wang, W.; Martin, R.; Liu, J.; Lester, E.; Licence, P.; Woods,
H. M.; Howdle, S. M. Macromolecules 2005, 38, 355.
(342) Akguen, M.; Deniz, S.; Baran, N.; Uzun, N. I.; Akguen, N. A.; Dincer,
S. Polym. Int. 2005, 54, 374.
(343) Trapa, P. E.; Won, Y.-Y.; Mui, S. C.; Olivetti, E. A.; Huang, B.;
Sadoway, D. R.; Mayes, A. M.; Dallek, S. J. Electrochem. Soc. 2005,
152, A1–A5.
(344) Ochi, M.; Shimaoka, S. Polymer 1999, 40, 1305.
(345) Jeong, H. M.; Choi, J. S.; Ahn, Y. T.; Kwon, K. H. J. Appl. Polym.
Sci. 2006, 99, 2841.
(346) Makamba, H.; Kim, J. H.; Lim, K.; Park, N.; Hahn, J. H. Electro-
phoresis 2003, 24, 3607.
(347) McDonald, J. C.; Whitesides, G. M. Acc. Chem. Res. 2002, 35, 491.
(348) Tsuruta, T. AdV. Polym. Sci. 1996, 126, 1.
(349) Kohn, J. Nat. Mater. 2004, 3, 745.
(350) Salamone, J. C.; Kunzler, J. F. (Bausch & Lomb Inc., USA) US
2007049713, 2007.
(351) Lai, Y.-C.; Bonafini, J. A. (Bausch & Lomb Inc., USA) US
2006276605, 2006.
(352) Kunzler, J.; Ozark, R. ACS Symp. Ser. 2000, 729, 296.
CR8001998
1278 Chem. Rev. 2010, 110, 1278–1319

Raman Microspectrometry Applied to the Study of Electrode Materials for

Lithium Batteries

Rita Baddour-Hadjean* and Jean-Pierre Pereira-Ramos

Institut de Chimie et Matériaux Paris-Est, UMR 7182 CNRS et Université Paris XII, 2 rue Henri Dunant 94320 Thiais, France

Received May 21, 2008

Contents
1. Introduction
2. The Raman Effect
2.1. Principle
2.2. Experimental Considerations
2.3. Advantages and Limitations of Raman
Microspectrometry for the Study of Electrode
Materials for Lithium Batteries
2.4. Advanced Experimental Approaches for
Raman Characterization of Lithium Battery
Components
3. Lithium Metal and Carbon-Based Electrodes
3.1. Lithium Metal Anodes
3.2. Carbonaceous Materials
3.2.1. Raman Spectra of Carbons
3.2.2. Graphite Intercalation Compounds
3.2.3. Carbon-Based Anodes for Lithium Ion
Batteries
4. Transition Metal Oxide-Based Compounds
4.1. Lithium Cobalt Oxide LiCoO2
4.2. Lithium Nickel Oxide LiNiO2 and Its Substitutive
Derivative Compounds LiNi1-yCoyO2 (0 < y < 1)
4.3. Manganese Oxide-Based Compounds
4.3.1. MnO2-Type Compounds
4.3.2. Ternary Lithiated LixMnOy Compounds
4.4. Vanadium Pentoxide V2O5
4.4.1. V2O5 Structure
4.4.2. LixV2O5 Bulk Phases
4.4.3. LixV2O5 Crystallized Thin Films
4.5. Titanium Oxide-Based Compounds
4.5.1. Lithium Titanate Li4Ti5O12
4.5.2. TiO2 Anatase
5. Phospho-olivine LiFePO4 Compound
6. General Conclusion
7. Acknowledgments
8. References
1. Introduction
Raman spectroscopy is a vibrational technique that pro-
vides unique structural information at the atomic scale1,2 on
inorganic and organic compounds. By coupling an optical
microscope to a conventional Raman spectrometer, the
technique becomes a microprobe with a spatial resolution
of less than 1 µm, determined by the wavelength of the
* Phone: 33 1 49 78 11 55. Fax: 33 1 49 78 11 95. E-mail address:
baddour@icmpe.cnrs.fr.
10.1021/cr800344k  2010 American Chemical Society
Published on Web 11/18/2009
1278
1280
1280
1281
1282
1283
1284
1284
1284
Rita Baddour-Hadjean was born in Coulommiers, France, in 1964. She
1284 received a chemical engineering diploma from the National School of
1285 Chemistry of Paris (ENSCP) in 1988. Working on the electrochemical
1286 and structural properties of sol-gel vanadium oxides usable as positive
electrodes in lithium batteries, she obtained her Ph.D. degree in Analytical
Chemistry from Université Pierre et Marie Curie—Paris VI in 1991. Then
1287 she joined the CNRS as researcher to understand the role that proton
1287 dynamics plays on the ionic conduction mechanism of various materials,
1290 using Inelastic Neutron Scattering in combination with other vibrational
techniques, such as infrared and Raman spectroscopies. Since 2000,
1291 using Raman microspectrometry, she has focused on the study of electrode
materials for rechargeable lithium batteries. This kind of investigation of
1291 the local structural changes induced by the lithium insertion process
1293 provides a better understanding of their electrochemical behavior. She
1298 earned a Habilitation qualification (Habilitation à Diriger des Recherches)
1298 from the Université Pierre et Marie Curie—Paris VI in 2004 and then
joined the group of Electrochemistry and Spectroscopy of Materials,
1300 managed by Jean-Pierre Pereira-Ramos at the Institute of Chemistry and
1303 Materials Paris Est (ICMPE), CNRS, Thiais, France. Her current research
1305 interests include the study of various electrode materials (bulk and thin
film oxides), in situ Raman microspectrometry measurements, and a
1306 theoretical approach based upon lattice dynamics simulations to get a
1308 quantitative assignment of the vibrational features.
1312
1314 radiation and the numerical aperture of the microscope
1315 objective. This has prompted several significant Raman
1315 applications in various fields dedicated to the characterization
of a wide range of materials: inorganic solids, ceramics,
semiconductor materials, protective coatings, polymers, and
battery materials.
Raman spectroscopy has been found to be very sensitive
to the state of metal oxide supported catalysts, with the
Raman frequencies being dependent on the metallic oxidation
state.3 As the Raman probe is capable of performing studies
of highly localized volumes with dimensions comparable to
the grain sizes or phase size in microstructures, several
research groups have reported Raman studies on polyphase
ceramics. Microphases and inclusions can be observed
directly under the microscope, then allowing the character-
Electrode Materials for Lithium Batteries
Jean-Pierre Pereira-Ramos was born in Saint-Maur des Fossés, France,
in 1957. He received a Ph.D. in Analytical Chemistry from Université Pierre
et Marie Curie—Paris VI in 1982. He joined the CNRS as researcher in
1983. Working on the electrochemistry of lithium intercalation compounds
in new sulfone-based electrolytes, he earned the grade of Doctor of
Physical Sciences from Université Paris XII—Val de Marne in 1988. For
almost 15 years he mainly has developed research activity on the synthesis
by soft chemistry of new lithium intercalation compounds and their
electrochemical properties. With Prof. Noël Baffier from ENSCP and
University Paris VI, he has investigated the relationships between the
structure, morphology, and electrochemistry of novel sol-gel transition
metal oxides as positive electrodes for rechargeable Li and Li ion batteries.
Director of research in CNRS since 1996, he currently heads the Group
of Electrochemistry and Spectroscopy of Materials at the Institute of
Chemistry and Materials Paris Est (ICMPE), CNRS, Thiais, France. The
areas presently studied in his group include intercalation compounds, thin
film oxides for Li microbatteries, and also lithiated transition metal nitrides
and silicium nanowires as negative electrodes. His main research interest
is to provide a comprehensive view of the Li insertion mechanism into
such electrode materials. He also develops in his group Raman
spectroscopy as a powerful tool to characterize the local structural changes
occurring in electrode materials under operation.
ization of impurity phases, the study of phase transition
mechanisms,4 or the identification of different polymorphic
forms, as reported, for example, in the case of silicon and
boron nitride5 and partially stabilized zirconia.6 In the latter
case, the use of the Raman imaging technique has enabled
us to map phase-transformed zones accompanying crack
propagation. Intense interest in the investigation and char-
acterization of high Tc superconducting oxides by Raman
spectroscopy has also been demonstrated since their discov-
ery in 1986. In addition to fundamental studies of phonon
and related electronic properties,7 Raman spectroscopy has
been extensively used as a controlling technique to evaluate
the quality of variously prepared high-Tc materials (e.g., bulk
powders, thin films, and single crystals). In particular, it has
been found to be an excellent tool for characterizing the
homogeneity and stoichiometry on the micrometer scale of
superconducting films deposited on substrates through the
observation of specific vibrational modes which are very
sensitive to the oxygen content.8 Micro-Raman spectroscopy
has also been recognized as one of the most sensitive tools
for studying the structural properties of carbonaceous materi-
als.9 In addition to its spatial resolution, which permits the
study of individual microcrystals as well as thin films, Raman
spectroscopy can distinguish between the various forms of
carbon. Thus, carbon with sp3-type bonding (diamond),
carbon with sp2-type bonding (graphite and carbonaceous
materials), and carbon in mixtures with these two types of
bonding (diamond-like carbons) can be characterized by their
Raman spectra.10 Protective diamond and silica coatings have
also been characterized using the Raman microprobe tech-
Chemical Reviews, 2010, Vol. 110, No. 3 1279
nique. In the case of diamond films, the deep analysis of
Raman scattering data, including the peak position, intensity,
area, and width, allows us to obtain the sp3- and sp2-bonded
C contents and the nature of internal stresses in the films.11-13
Semiconductors in bulk, thin film, and device configurations
have been thoroughly investigated. For such materials,
Raman microspectrometry is very well suited to collect a
wealth of information on the spatial distribution of physical
quantities such as strains,14 atomic fraction in mixed crystals,
impurity concentrations, free carrier concentrations,15 and
local crystallographic orientation.16 These data are useful not
only to evaluate the quality of a sample but also to infer the
relevant dynamical processes, e.g. growth of crystallites,
atomic diffusion, and reactions at interfaces or surfaces.
Numerous applications of Raman microanalysis can also be
found in the field of art, jewelry, archeology, and forensic
science,17-22 as it constitutes a well adapted technique for
the nondestructive examination of molecular species with
good spatial resolution. Pigment investigation is one of the
most active research domains within art analysis using
Raman spectroscopy,19 but other important research topics
can be found, including the study of corrosion phenomena,20
the analysis of glasses and ceramics,21 and the study of
minerals and gemstones.22
This review is focused on key issues and trends in Raman
research on materials usable as negative or positive elec-
trodes23 in secondary Li and Li ion batteries. Rechargeable
Li metal batteries use metallic Li as the negative electrode
while lithium ion batteries use carbon or a Li intercalation
compound as the negative one. The storage of electrical
energy will be far more important in this century than it was
in the last. The need for clean and efficient energy storage
from renewable sources is important. Rechargeable batteries
are required to power various portable consumer electronic
devices (cell phones, PDAs, laptop computers, etc.), to power
electric vehicles, for implantable medical applications, but
also to store electricity from wind/solar energy. Advances
in rechargeable lithium ion batteries have allowed the success
of mobile electronic equipment, and the worldwide market
for rechargeable lithium batteries is growing every year. But
now an increase in energy and power density to meet these
future challenges is needed. To promote such developments,
it is essential to find new high performance materials but
also to understand the materials properties that are respon-
sible for the electrochemical features, cycle life, and structural
behavior of electrode materials under operation.
The basic process that occurs in a rechargeable lithium battery
is the intercalation of lithium ions into different host materials.
One fundamental problem is the loss of capacity during
successive electrochemical insertion-deinsertion processes. As
is well-known, the electrochemical properties of such materials
(e.g., specific capacity, reversibility, rate capability, cycling
behavior, ...) are strongly dependent on the structural changes
induced by the lithium insertion reaction. Therefore, the
establishment of clear relationships between electrochemical and
structural data seems to be one of the key issues for better
understanding and further improving the electrochemical per-
formance of electrode materials for rechargeable lithium bat-
teries. Considering the experimental techniques carried out to
fulfill this requirement, they allow the obtainment of either long-
range (X-ray, neutron, or electron diffraction, ...) or short-range
data (X-ray absorption, NMR, EPR, XPS, ...). Among the local
probes, Raman spectroscopy is very appropriate, since deter-
mination of frequencies of normal vibrations provides a very
1280 Chemical Reviews, 2010, Vol. 110, No. 3 Baddour-Hadjean and Pereira-Ramos

sensitive tool to detect structural variations at the atomic level.

Indeed, unique molecular and crystalline information is then
accessible, e.g. local disorder, changes in bond lengths, bond
angles, coordination, Li dynamics, and cation ordering. As a
matter of fact, a significant amount of Raman data can be found
in the literature. We present here quantitative results obtained
over the past few years from Raman microspectrometry on
metallic lithium, carbonaceous materials, transition metal oxide
MOy (M ) V, Mn, Ti), lithiated transition metal oxides LixMOy
(M ) Co, Ni, Mn, Ti), and phospho-olivine LiFePO4 com-
pounds used as negative or positive electrode materials in Li
or Li ion batteries. These studies have been selected in order
to highlight the extraordinary potential of Raman microscopy
to obtain structural information on the material under operation
and to provide insight into the mechanisms governing the
electrode performances.

2. The Raman Effect

2.1. Principle
It seems appropriate to review first the Raman effect
principle, which is described in detail by Turrell et al.24 The
Raman effect was named after one of its discoverers, the
Indian scientist Sir C. V. Raman, observed the effect by
means of sunlight in 1928. This phenomenon results from
the interaction of light and matter. When a photon of light
interacts with a molecule, the photon can be absorbed or
scattered. Considering the absorption process, it requires,
first, that the energy of the incident photon is equal to the
energy difference between two states of the molecule (this
is referred to as the resonance condition) and, second, that
the transition between the two states is accompanied by a
change in the dipole moment of the molecule. Such condi-
tions are those corresponding to infrared absorption. Photons
which are not absorbed will be scattered and the incident
photons need not to be resonant with two states of the
molecule for scattering to occur. In quantum mechanics, this
interaction is described as an excitation to a virtual state
lower in energy than a real electronic transition, as shown
in the energy level diagram shown in Figure 1a. When
scattering of the photon occurs without any change in the
atomic coordinates of the molecule, the photon is elastically
scattered. It exhibits the same energy (frequency) and,
therefore, wavelength as the incident photon. This process
is commonly referred to as Rayleigh scattering. However,
for a very small fraction of light (approximately 1 in 107
photons), a vibrational motion occurs. A quantum of
vibrational energy is then transferred between the molecule
and the incident photon and the remaining energy is scattered
inelastically such that the energies of the incident and
scattered photons are no longer equal. The process leading
to this inelastic scattering constitutes the Raman effect. In
this case, the molecule may either gain energy from or lose
energy to the photon. If the transfer of energy in the virtual
state is from the photon to the molecule, the scattered photon
will be lower in energy than the incident photon and the
phenomenon is referred to as Stokes Raman scattering.
Conversely, if the transfer of energy in the virtual state is
from the molecule to the photon, the scattered photon will
be higher in energy than the incident photon, which is
referred to anti-Stokes Raman scattering (Figure 1a). In
classical terms, the scattering process can be viewed as a
perturbation of the molecule’s electric field. Vibrational
Raman scattering occurs because a molecular vibration
Figure 1. (a) Energy level diagram for different processes. (b)
Raman and Rayleigh scattering of excitation at a frequency ν0. The
molecular vibration in the sample is of a frequency νvib.
induces a change in the polarizability. The polarizability, R,
measures the ease with which the electron cloud around a
molecule can be distorted. The change is described by the
polarizability derivative, δR/δQ, where Q is the normal
coordinate of the vibration. The selection rules for Raman-
active vibrations are linked to molecular symmetry and
identify vibrations that change molecular polarizability.
A Raman spectrum is a plot of the intensity of Raman
scattered radiation as a function of its frequency difference from
the incident radiation (Figure 1b). This difference is called the
Raman shift. Note that, because it is a difference value, the
Raman shift is independent of the frequency of the incident
radiation. Figure 1b illustrates the symmetry of Stokes and Anti-
Stokes bands; however, because the ground-state population is
greater than that of the excited state, the Stokes lines are more
intense than the Anti-Stokes lines. For this reason, Raman
spectra are often plotted as a function of intensity versus the
Stokes-shifted frequencies in wavenumbers (cm-1). Frequency
shifts between the incident radiation and the Raman-scattered
radiation correspond to the vibrational energy levels of the
molecule or the crystal. Many parameters, e.g. local environ-
ment, symmetry of the crystal, atomic mass, bond order,
H-bonding, molecular substituents, atomic environment, mo-
lecular geometry, structural disorder, and strains, can affect the
vibrational force constants, which, in turn, dictate the vibrational
energy. Hence, Raman spectroscopy allows the study of
intramolecular vibrations, crystal lattice vibrations, and other
motions of extended solids.
In addition to the Raman shift, which allows the deter-
mination of the vibrational frequencies of the normal modes,
another property can be reached from Raman measurements,
which is the polarization of the diffused light. In the ideal
case of a crystal, the nature of the Raman spectrum depends
Electrode Materials for Lithium Batteries
Figure 2. Schematic description of a confocal Raman microspectrometer system
on the orientation of the crystallographic axes with respect
to the direction and polarization of both the excitation and
the scattered light. Hence, polarized Raman spectroscopic
investigations of single crystals are very useful in the
interpretation of the Raman spectra of crystalline samples,
since they can, by suitable choices of orientation, lead to
the number and symmetry of the active modes.24
Quantum mechanics requires that only concerted atomic
displacements are allowed for a given molecule. These are
known as normal modes, which refer to molecular vibrations
where each atom moves with the same frequency. A linear
molecule with N atoms has 3N - 5 normal modes, and a
nonlinear molecule has 3N - 6 normal modes of vibration.
There are several types of motion that contribute to the
normal modes. Some examples of molecular vibrations are
as follows:
• stretching—a change in the length of a bond;
• bending—a change in the angle between two bonds;
• rocking—a change in the angle between a group of atoms
and the rest of the molecule;
• wagging—a change in the angle between the plane of a
group of atoms and a plane through the rest of the molecule;
• twisting—a change in the angle between the planes of
two groups of atoms;
• out-of-plane—the atom moves in and out of the plane
of the other atoms.
In a rocking, wagging, or twisting coordinate, the angles
and bond lengths within the groups involved do not change.
Rocking is distinguished from wagging by the fact that the
atoms in the group stay in the same plane.
Molecules or crystals can be classified according to
symmetry elements or operations that leave at least one
common point unchanged. This classification gives rise to
the point group representation for the molecule, which is
uniquely defined by a set of symmetry operations—rotations
Cn, reflections (σh, σV, and σd), inversion i, and improper
rotations Sn ) Cnσh—that transform that molecule into
itself.25 The full information of all symmetry transformations
Chemical Reviews, 2010, Vol. 110, No. 3 1281
in a point group is given in the so-called character tables.
Character tables not only give the number and degeneracy
of normal modes, but they also tell us which of the normal
modes will be IR-active, Raman-active, or both. It turns out
that a fundamental transition will be Raman-active if the
normal mode involved belongs to the same symmetry
representation as any one or more of the Cartesian compo-
nents of the polarizability tensor of the molecule.
2.2. Experimental Considerations
The main features of the Raman microspectrometry system
can be described by the basic layout shown in Figure 2 and
described in detail by Delhaye et al.26 In a Raman mi-
crospectrometer, the backscattering geometry is employed:
the laser beam is focused on the sample through the
microscope objective, and then the Raman light is collected
by the same objective in the inverse direction of the incident
light. The use of a microscope operating in a 180° back-
scattering geometry eliminates the need to continually adjust
the laser onto the sample and to focus the scattered light
onto the spectrometer. Raman microspectrometers utilize
research grade microscopes to focus the excitation onto the
sample and to collect and transfer the Raman scattered light
into the Raman spectrometer. High numerical aperture
microscope objectives greatly enhance the spatial resolution
and the optical collection power of the Raman instrument.
These Raman microscopes are easy to use and are capable
of analyzing small areas (∼1 µm2), which is an excellent
spatial resolution to determine the distribution of chemical
species. The concept of confocal scanning microscopy was
introduced by Minsky27 in the early 1960s to overcome some
of the limitations of the conventional optical microscope.
With this technique a significant improvement in both the
contrast and the spatial resolution may be obtained. Unlike
a conventional microscope, where the entire field is il-
luminated, the confocal system measures the intensity of the
light reflected or transmitted by a very small area of sample.
1282 Chemical Reviews, 2010, Vol. 110, No. 3
Detailed theoretical and experimental studies on the proper-
ties of confocal microscopes can be found elsewhere.28 The
application of the confocal principle to Raman spectroscopy
was first described in the 1990s.29 The Raman microscope
focuses the laser beam down to a small volume (on the order
of 1 µm3 in air) and is operated readily in a confocal mode
by placing an aperture at a back focal plane of the
microscope.30 The aperture improves the lateral and axial
spatial resolution of the microscope, allowing nondestructive
depth profiling by acquiring spectra as the laser focus is
moved incrementally deeper into a transparent sample. The
critical importance of a correct interpretation of confocal
Raman data is reported in many works.31-33
Raman microspectrometers are generally composed of
several main parts:
1. The excitation source (laser). Lasers revolutionized Raman
spectroscopy in the 1970s because they give a coherent beam
of monochromatic light and can have very high power. Their
high intensity allows the production of a detectable amount
of Raman scatter.
2. The collection deVice. Monochromatic light from the laser
passes through an interference filter to obtain the desired
bandwidth with an improved rejection of the unwanted
wavelengths. Then the laser beam is reflected by a holo-
graphic notch filter, which is basically a polymeric coating
on a glass substrate. It is specially designed to have low
transmission (high reflectivity) at the laser wavelength and
high transmission at all other wavelengths. Then, the exciting
laser beam is focused on the sample through a microscope
objective. The scattered light is then collected by the same
objective and passed through the notch filter, with the
Rayleigh-scattered light being rejected, which allows an
important gain in sensitivity. The rejection ratio for a single
filter can be as high as 106, which allows a small single-
stage spectrometer to be used in place of the double- and
triple-grating used a few years ago. Indeed, the main
difficulty of Raman spectroscopy is to detect the weak
inelastically Raman scattered light from the intense back-
ground due to the Rayleigh scattered light. Holographic notch
filters typically permit Raman spectral measurements of
frequency shifts greater than ∼100 cm-1.
3. The spectral analysis system and the detector. The
polychromatic light beam collected by the objective of the
Raman microprobe has to be analyzed by a spectrometric
system. When Raman scattered photons enter the spec-
trograph, they propagate through a transmission grating to
separate them by wavelength and are collected by a detector
which records the intensity of the Raman signal at each
wavelength. Significant advances in Raman spectroscopy
have been afforded by the development of detectors, evolving
from photographic plates to photoelectric tubes, photon
counting, and various forms of multichannel detectors:
photodiode arrays (PDA) and, more recently, charge coupled
device (CCD) cameras. These little integrated circuit chips
are extremely sensitive to light and contain thousands of little
picture elements (called pixels) that take the whole spectrum
at once in less than a second. The high detectivity of CCD
detectors allows the use of very low laser power, in order to
prevent thermal or photochemical destruction of the sample.
4. Finally, an acquisition electronics allows the scanning,
the collection, and the processing of the data.
Baddour-Hadjean and Pereira-Ramos
2.3. Advantages and Limitations of Raman
Microspectrometry for the Study of Electrode
Materials for Lithium Batteries
Since Raman microscopy is an optical spectrometry, it has
several characteristics very well suited to the study of
electrode materials. First, under careful excitation conditions,
the analysis is nondestructive. The analysis causes neither
damage nor alteration, so it is possible to continue the
investigation of the same electrode material, even on the
same spot, with another laser wavelength or with another
technique.
Raman experiments can be performed without sample
preparation. Neither special coating nor a controlled atmo-
sphere is necessary. As Raman is relatively unaffected by
strong IR absorbers such as water, CO2, and glass (silica),
no special accessories are needed for aqueous solutions. The
visible excitation source can penetrate transparent container
materials, and thus, Raman measurements can be acquired
through glass vials, envelopes, plastic bags, and several other
packaging materials. This characteristic facilitates the de-
velopment of experimental devices, allowing in situ Raman
investigations of electrode materials under operation. This
is especially desirable when exposure to atmosphere might
change some of a material’s bulk and/or surface properties.
If the low sensitivity of Raman spectroscopy had long been
a major limitation, the detectors have been tremendously
improved and it is now possible to obtain the spectra of
various kinds of electrode materials (composite electrodes,
thin films, single particles, powders, electrode/electrolyte
interfaces, electrolytes, thin films, ...) on a microspot in a
few tens of seconds.
A significant advantage is the capability to focus the
excitation beam on a very small spot, whose diameter
depends on the selected laser wavelength and the aperture
of the objective, typically ∼1 µm. Such a spatial resolution
is required by the heterogeneous character of electrode
materials used in lithium batteries. Because the surfaces of
the composite electrodes typically contain one or two types
of carbon and/or and oxide, a binder, and occasionally an
additive, Raman microspectrometry is a very suitable tech-
nique, as it allows heterogeneous mixtures to be analyzed,
with a lateral resolution at the electrode surface correspond-
ing to a typical particle size of a few micrometers. Informa-
tion about the local structure and chemical composition of
each component may therefore be provided individually.
Furthermore, as confocal Raman microspectrometry is able
to analyze very small volumes on the order of the cubic
micrometer, it is possible to perform Raman imaging from
point by point analysis. Hence, two- or three-dimensional
chemical or structural mapping can be produced with a
micrometric resolution. Another advantage provided by
confocal Raman spectroscopy consists in a true “optical
sectioning” of the sample, because the confocal mode allows
the Raman light coming from the focal plane at the sample
to be selected. Spatial resolution of ∼1 µm3 can be achieved,
which allows the different elements of a working battery to
be investigated (anode, cathode, electrode/electrolyte inter-
face).
Raman spectroscopy is the technique of choice for the
characterization of carbon-based materials used in lithium
ion batteries. This property comes from the fact that the
scattering efficiency gets larger when the laser energy
matches the energy between optically allowed electronic
Electrode Materials for Lithium Batteries
transitions in the material. This intensity enhancement
process is called resonance Raman scattering.
In opaque materials, an interesting point concerns the axial
resolution, determined by the optical skin depth (δ) of the
laser beam, whose value is approximately 30-300 nm. The
skin depth (δ) is directly related to the electronic conductiv-
ity, σ, the magnetic permeability, µ, and the laser wavelength,
λ, through the following equation: δ ) (2λ/µσ)1/2. Hence,
an increase in the electronic conductivity will result in
reduction of the optical skin depth. For example, in the case
of highly oriented pyrolytic graphite, Raman spectroscopy
allows us to examine phenomena within a skin depth of about
50-100 nm with a 514.5 nm green laser beam.34
Raman spectroscopy is a well suited technique for the
characterization of the local structure in transition metal
oxides used as positive (or negative) electrode materials in
lithium and lithium ion batteries. From an analytical point
of view, the Raman spectroscopy technique can solve the
problem of phase identification when various environments
are present. Indeed, the wavenumbers and relative intensities
of the Raman bands are very sensitive to the crystal
symmetry, coordination geometry, and oxidation states. For
example, it is possible to differentiate various kinds of metal
oxides whose atomic arrangements are closely related to one
another, to distinguish between different metal oxide com-
pounds having the same elementary analysis, such as
manganese oxides MnO2, Mn3O4, Mn2O3, or between
compounds with the same stoichiometry but different
crystalline structures, such as anatase and rutile TiO2 or cubic
and hexagonal LiCoO2.
Since it does not need a long-range structural order, Raman
spectroscopy also constitutes an alternative structural tool,
as it allows the study of “amorphous” compounds, thin films,
or cycled cathode materials which exhibit poor XRD
information due either to their low crystallinity, their
preferential orientation, or structural disorder.
From a more fundamental point of view, Raman spec-
troscopy constitutes a local probe of great interest, comple-
mentary to long-range structural techniques such as X-ray
or neutron diffraction, to study the cathodic material under
or after operation. The determination of frequencies of
normal vibrations provides various useful data on the local
structure variations induced by the lithium insertion dein-
sertion process in the lattice host, e.g. changes in metal-oxygen
bond lengths, metal oxidation state, lithium environment,
lattice distortions, disorder, lithium-lithium interactions,
lithium-host lattice interactions, and cation ordering. Fur-
thermore, the use of high resolution Raman microscopy
mapping allows the identification of the local processes that
contribute to the electrode capacity loss, through the analysis
of the surface composition and distribution. This type of
information is of great interest insofar as the processes that
occur on a microscopic scale can be directly linked with the
macroscopic behavior.
The main limitation of Raman spectroscopy is the dif-
ficulty in making this technique quantitative, due to the
experimental effort needed to measure and calibrate the
Raman band intensities, due to the small surface it probes,
and due to the fact that its sensitivity strongly depends on
the polarizability of the analyzed molecules. This yields, for
example, an extremely good capacity to detect minimal
amounts of anatase and rutile (TiO2) and on the other hand
a very low sensitivity toward manganese dioxide (MnO2).
It is also not a suitable technique for the analysis of metallic
Chemical Reviews, 2010, Vol. 110, No. 3 1283
compounds, as nearly all pure metals are Raman silent, and
some materials are quite unstable under the local heating
due to the exposition to the laser light. Fluorescence (much
more intense than the Raman signal) can also limit Raman
usefulness. Background fluorescence occurs often in impuri-
ties or organic materials of the sample (typically fluorescence
of the electrolyte). This problem can be partly solved by
changing the laser wavelength in order to separate the Raman
and the fluorescence spectra, since electronic fluorescence
is excited by a specific wavelength.
2.4. Advanced Experimental Approaches for
Raman Characterization of Lithium Battery
Components
These whole characteristics have prompted the in situ
application of Raman spectroelectrochemistry and stimulated
the emergence of novel experimental approaches during the
past decade. The ability to spatially and temporally probe
working lithium batteries with Raman spectroscopy might
provide another opportunity for researchers to experimentally
verify theoretical models used to simulate the lithium ion
dynamics.
The first in situ Raman spectroscopic studies in an
operating lithium rechargeable battery focused on the
intercalation of lithium ions into various oxide-based cath-
odes (e.g., LixCoO2, LixMn2O4, and LixV2O5)35-37 and
graphitic anodes.38-43 Rey et al. demonstrated the successful
application of confocal Raman microspectrometry for in situ
characterization of a lithium battery that consisted of a Li
metal anode, a P(EO)20LiN(SO2CF3)2 polymer electrolyte,
and a V2O5 cathode.44,45 The authors could detect structural
changes in the V2O5 cathode, concentration gradients in the
polymer electrolyte, and contaminating agents that formed
at the lithium-electrolyte interface. An innovative applica-
tion of Raman microscopy was presented in several studies
reported by Panitz and Novak,43,46-48 who used Raman
surface mapping to generate local surface composition
images of 30 × 35 µm2 areas (at 2 µm lateral resolution) of
LiCoO2 positive and carbon negative electrodes from com-
mercial lithium batteries. The mapping technique application
of Raman microscopy has provided unique insight into the
mechanisms of specific chemical or electrochemical pro-
cesses that may be responsible for the cell degradation, as
demonstrated by the numerous works reported by Kostecki
et al. on positive49-54 and carbon negative55,56 electrodes used
in Li ion and high power Li ion cells. In situ simultaneous
spectroscopic and electrochemical measurements on single
particle electrodes were successfully performed by Scherson
et al.,57-60 who collected high quality, time-resolved Raman
spectra as a function of the applied potential from single
particle graphite electrodes embedded in thermally annealed
Ni foils57 as well as single particles of LiMn2O4 isolated in
a microelectrode.58,59 The same group reported an in situ
space- and time-resolved Raman spectromicrotopography
experiment of an operating lithium ion battery.60 This
arrangement enabled Raman spectra to be collected continu-
ously from a sharply defined edge of the battery exposing
the anode, separator, and cathode, during charge and
discharge. Clear evidence was obtained regarding the state
of charge of graphite particles within the anode and, to a
lesser extent, of LiCoO2, during battery discharge as a
function of both position and time. More recently, Kostecki
et al.61 proposed an experimental approach enabling inves-
1284 Chemical Reviews, 2010, Vol. 110, No. 3
tigation of the in situ spectroscopic behavior of several
individual LiNi0.8Co0.15Al0.05O2 particles in composite cath-
odes during a galvanostatic charge/discharge cycle while
Migge et al.62 performed confocal Raman spectroscopy on
a specially designed button cell to study the intercalation of
lithium into single particle graphite electrodes. Burba et al.63
lately described a modification to industrially available coin
cells to facilitate routine in situ Raman spectroelectrochemi-
cal measurements of electrodes in lithium batteries. The
authors tested their strategy on Li//V2O5 cells. This approach
is of importance for researchers to easily construct cells for
in situ spectroscopy in addition to cells that are used in
normal electrochemical evaluations.
3. Lithium Metal and Carbon-Based Electrodes
3.1. Lithium Metal Anodes
The lithium metal electrode is of great interest to battery
researchers because of its high theoretical specific capacity.
To date, however, problems associated with the reversibility
of the deposition-dissolution process at the interface have
prevented its successful application in secondary batteries
for commercial purposes. The poor reversibility is due to
nonuniform current density across the lithium surface under
electrochemical operation. This limitation has been related
to the formation of an inhomogeneous film on the metal
surface. The film, known as the solid electrolyte interphase
(SEI), is composed of various reduction products, which
result from reaction between lithium and the electrolyte
solution.64 Charge-discharge cycling of the electrode re-
quires the transport of lithium ions through the inhomoge-
neous film, which results in irregular deposition (charging)
or dissolution (discharging) at the electrode surface and
subsequent “dendritic deposition” of Li, which limits the
cyclability of the Li electrode.64
Raman investigation on the lithium metal/electrolyte
interface in an attempt to identify the passive surface film
has so far been unsuccessful. A few Raman studies performed
on the electrolyte itself have nevertheless allowed the
identification of species present in the electrolyte and near
the electrode surface,38,65 the characterization of the local
structure of solvated lithium cations,66,67 and the access to
the transport properties in polymer electrolytes through the
concentration profiles obtained from in situ confocal Raman
microspectrometry experiments.44,45 A sealed lithium optical
cell was specially designed and tested, with the aim to
investigate alternative electrolyte systems for use with lithium
metal electrodes.68 The authors reported optical images of
lithium surface deposits and in situ Raman spectra arising
from both the electrolyte and the deposits formed during
charge-discharge cycling of lithium metal electrodes. Nev-
ertheless, the authors were not able to confidently assign the
Raman peaks corresponding to the deposited species.
Researchers have employed a wide range of techniques
to study the processes which occur on the lithium surface.
The surface chemistry of lithium in organic electrolytes has
been investigated using (FTIR) spectroscopy, EDAX, and
XPS.69 The results showed that solvents, such as propylene
carbonate (PC), were decomposed on lithium and formed
surface films of lithium alkylcarbonate, ROCO2Li, and that
the lithium surface consisted of Li2CO3, LiOH, Li2O, and
lithium halides. Raman studies focusing on the lithium anode
are scarce. Indeed, as far as the characterization of the surface
is concerned, the sensitivity of normal Raman is rather
Baddour-Hadjean and Pereira-Ramos
limited, since Raman signals are only obtained after a
degradation reaction has occurred.70 Because surface-
enhanced Raman scattering (SERS) provides an increase of
sensitivity as a result of the large enhancement effect (up to
a factor 107), this technique has been more helpful in
identifying the SEI on an Ag anode through the formation
of a nanometer-scale islandlike Li-Ag alloy on the electrode
surface.71 These experiments have allowed the detection of
LiOH, H2O, and Li2CO3 as the main stable species of the
SEI film on the surface of the Ag electrode discharged in 1
M LiPF6/EC-DEC electrolyte.72
3.2. Carbonaceous Materials
Carbon-based materials have received considerable atten-
tion as negative electrodes,73 because they constitute good
alternatives to solve the problem of cycle life and safety
raised by the use of metallic lithium. Indeed, graphite serves
as a host structure for lithium intercalation and the structure
is resilient enough to provide reversibility by allowing easy
insertion and deinsertion of lithium. Graphite exists in various
forms ranging from a crystalline state to an almost amorphous
state. The raw materials used to produce carbon anodes
include natural graphite, oil pitch, coal tar, hydrocarbon gas,
benzene, and various resins. In practice, at room temperature,
graphite accepts sufficient lithium to form LiC6, which on
delithiation can deliver 372 (mA h)/g.73 The relation between
the electrochemical features of Li intercalation into various
carbon-based materials and their crystal structure has been
described in detail.74 Before focusing on the Raman contri-
bution to the study of carbon electrode materials, we present
some salient aspects of the vibrational features of carbons.
3.2.1. Raman Spectra of Carbons
From a structural point of view, hexagonal graphite
consists of stacked sheets with the carbons within the layers
arranged in a two-dimensional network of regular hexagons.
It crystallizes in the D46h space group and has two doubly
degenerate Raman-active modes,75 both vibrating in the plane
of the sheets, E2g1, at 42 cm-1, and a strong C-C stretching
mode, E2g2 (G-band), at 1582 cm-1. Figure 3 shows the sharp
intense band at 1580 cm-1 observed for a natural graphite
crystal and HOPG. Because of the weak interlayer bonding,
graphite crystals are subject to disorder along the c axis while,
at the same time, the strong intralayer carbon-carbon
bonding maintains a high degree of order within the
individual carbon sheets. This so-called turbostratic disorder
strongly influences the Raman features. Hence, in addition
to these allowed two lines, many kinds of graphite materials
exhibit a disorder-induced A1g line (D-band) at about
1350-1360 cm-1. Figure 3 illustrates the typical Raman
features observed for polycrystalline graphite, with the 1357
cm-1 band appearing for well-crystallized graphite with small
particle size but not with large grain single crystals. This
mode has been linked to the break of symmetry occurring
at the edges of graphite sheets originating from some kind
of imperfection and disorder, such as defects, discontinuity
in crystallites, and stacking disorder in the crystal structure
of graphite.75 The line width and D/G band intensity ratio
vary depending on the structure of the carbon (Figure 4).
Several authors10,75 have found a linear relationship between
the inverse of the intraplanar microcrystallite dimension La
and the ratio of the intensity of the disorder-induced D-line
to that of the Raman-active E2g2 G-line, ID/IG, denoted R.
Electrode Materials for Lithium Batteries Chemical Reviews, 2010, Vol. 110, No. 3 1285

Table 1. Raman Frequencies (ω) and Widths (Γ) for High

Frequency Graphitic Bounded- and Interior-Layer Phonons for
Stage 1-3 Li-Graphitea
compd temp (K) stage ω (cm-1) Γ (cm-1) (fwhm)
LiC6 300 1 1596 ( 2 35 ( 4
LiC6 300 1 1594 ( 3 48 ( 4
LiC12 300 2 1592 ( 2 33 ( 5
LiC17 300 2 1598 ( 2 28 ( 3
LiC17 300 2 1604 ( 4 33 ( 4
LiC12 240 2 1590 ( 3 33
LiC18 240 3 1602 ( 2 20 (4
1577 ( 2 20 (3
a
Reprinted with permission from ref 93. Copyright 1987 the
American Physical Society.

the Raman excitation. This strong dependence of the R value

Figure 3. Raman spectra of various graphitic carbons, recorded
using a laser excitation line of 514.5 nm. Reprinted with permission
from ref 10. Copyright 1989 Materials Research Society.
upon the laser wavelength has been further observed in many
different graphitic materials, amorphous-C materials,77 pol-
ished glassy carbon samples,78 pyrolytic graphites,79 carbon
nanotubes,80 disordered graphite,81 graphite edges,82 as well
as carbon foams.83 The complete understanding of the
intensity dependence of the D-band in graphitic materials
as a function of Elaser and La is of fundamental importance,
since Raman spectroscopy is one of the preferred tools for
the characterization of graphitic carbons. Vidano et al.84
performed a systematic investigation of the Raman spectra
of different kinds of carbon materials by varying the laser
excitation wavelength, and they observed that the wavenum-
ber value of the disorder-induced D-band also varies linearly
with increasing Elaser, with the slope of this dependence being
∂ωD/∂Elaser = 50 cm-1/eV. The laser energy dependence of
the D-band frequency was observed in many different
graphitic materials, and the slope of this dependence was
observed to be independent of the material type.78 The first
attempt to explain the dispersive behavior in the frequency
of the D-band was presented by Baranov et al.,85 who
proposed that this band originates from a double resonance
(DR) Raman process in graphite. This concept was further
developed by Thomsen and Reich,86 who calculated the
scattering cross section for the double resonance mechanism
that gives rise to the D-band. Despite the fact that the
dispersive behavior of the D-band frequency has been
successfully understood by the DR mechanism, the strong
Elaser dependence of the ratio ID/IG is still an open problem
and there is not yet a theory that explains the dependence of
the ratio ID/IG on the crystallite size La.87

3.2.2. Graphite Intercalation Compounds

Figure 4. Raman spectra of various noncrystalline carbons,
recorded using a laser excitation line of 514.5 nm. Reprinted with
permission from ref 10. Copyright 1989 Materials Research Society.
From the value of R, the crystallite size along the a-axis, La,
can be calculated according to the equation of Tuinstra and
Koenig: La ≈ 4.4/R [nm]; however, this equation, originally
developed for an excitation wavelength of 488 nm, must be
taken cautiously, as Raman intensities of carbonaceous
materials have been found to depend on the excitation
wavelength. As a matter of fact, Mernagh et al.76 have
investigated the Raman spectrum of carbon black and found
that the R value depends on the laser energy Elaser used for
Because Raman spectroscopy is a pertinent tool for the
study of ion insertion into carbonaceous materials, many
groups have studied the Raman spectra of graphite intercala-
tion compounds (GICs).88-93 The most important property
of graphite intercalation compounds is the staging phenom-
enon, which corresponds to intercalate layers that are
periodically arranged in a matrix of graphite layers. The stage
index, n, denotes the number of graphite layers between
adjacent intercalate layers. Li-intercalated graphite is an
example of a donor GIC, with the lithium layers donating
electrons to the graphitic carbon layers, and different phases
have been reported: stage-1 LiC6, stage-2 LiC12 and LiC17,
stage-3 LiC18, and stage-4 LiC27.93 Table 1 summarizes the
main features reported for stage 1-3 Li-graphite.93 Intercala-
tion of guest species into a host structure of carbon strongly
affects the position, shape, and intensity of the E2g2 band at
1582 cm-1. A Raman doublet is observed for GICs with stage
1286 Chemical Reviews, 2010, Vol. 110, No. 3
n > 2, at wavenumbers close to the singlet E2g2 of pristine
graphite.88-93 The lower wavenumber component E2g2(i)
(around 1580 cm-1) corresponds to graphitic layers not
adjacent to intercalate layer planes (named “interior layers”)
whereas the upper wavenumber component E2g2(b) (around
1600-1610 cm-1) corresponds to graphite layers adjacent
to intercalate layer planes (named “bounded layers”). For n
e 2, only bounded carbon layers exist, and a single Raman
line is observed, around 1600 cm-1 for stage-1 LiC6. Whereas
LiCx phases are difficult to detect by X-ray diffraction owing
to the very weak intensity of the (hk0) diffraction peaks,
Raman spectrometry is a pertinent tool to determine the stage
of the Li intercalation process in carbonaceous materials
through the analysis of the shape, the frequency shift, and
the relative intensity of the Raman features in the 1570-1620
cm-1 range.89
3.2.3. Carbon-Based Anodes for Lithium Ion Batteries
Numerous works are devoted to the Raman study of
lithium intercalation into carbon-based anode materials.
Spectroscopy experiments have been conducted either in ex
situ38,55,56,94,95 or in situ39-43,57,60,62,96-98 conditions, on different
kinds of carbons: highly oriented pyrolitic graphite (HOPG),
natural graphite, cokes, carbon fibers, synthetic high tem-
perature graphite, multiwall carbon nanotubes (MWNTs), or
spherical mesocarbon microbeads (MCMBs). Much informa-
tion has been drawn from the observation of the Raman
features: the nature of the lithium intercalated phases, the
local distribution of lithium across the surface of graphite
or carbon particles, the degree of surface structural disorder
on graphitic anodes, the stability of the carbon-based material
during electrochemical cycling and aging, etc.
During an in situ Raman study of the electrochemical
lithium insertion into the HOPG electrode, Inaba et al.39
showed spectral changes associated with phase transitions
corresponding to different staged LiC27 (n ) 4), LiC12 (n )
2), and LiC6 (n ) 1) phases, occurring reversibly during a
charge and discharge cycle. The authors observed that the
electrode potential was determined by the “surface stage”
of graphite intercalation compounds. Recent advances in
Raman spectroscopy in the latest years have made it possible
to study the dynamic aspects of Li intercalation-deintercalation
into single carbon particles embedded in thermally annealed
Ni foils,57 using in situ, time-resolved Raman microscopy.
By analyzing the position of the prominent G band, the
authors were able to determine spectroscopically and in real
time the average concentration of Li+ within the volume of
the particle probed by the laser beam all along the electro-
chemical lithium deintercalation process. More recently,
Migge et al.62 used in situ confocal Raman microspectroscopy
to investigate the first cycle of lithium intercalation-
deintercalation into single graphite particles. During the first
charge of the battery, they found the typical spectroscopic
fingerprints of the LiC27, LiC12, and LiC6 phases. However,
the intercalation process was not homogeneous, even in
single graphite particles, depending on various parameters
of the working battery, such as current density, electrolyte,
and temperature. Raman studies of the electrochemical Li
intercalation into mesocarbon microbeads (MCMBs) heat
treated at different temperatures40 and multiwall carbon
nanotubes (MWNTs)94 revealed the same intercalation mech-
anism in the case of MCMBs heat treated at 2800 °C; that
is, lithium is inserted between graphene layers via stages of
GIC formation. Conversely, for MCMBs heat treated at 1800
Baddour-Hadjean and Pereira-Ramos
and 1000 °C and MWNTs, the authors reported a random
Li insertion without the formation of staged phases.
Several works have demonstrated that Raman micros-
copy constitutes a very convenient diagnostic tool to
estimate the surface structural changes occurring on
graphitic materials.47,55,56,95,99-105 Indeed, a quantitative
characterization of the degree of surface disorder can be
obtained from the analysis of the Raman bands observed
at 1357 and 1580 cm-1. These studies have shown that
several parameters such as sample preparation, surface
modification, heat treatment, and cell operating conditions,
greatly influence the surface crystallinity and, hence, the
electrochemical performances of carbon-based
anodes47,55,56,95,99-101,104,105 and carbon-coated LiFePO4
cathodes102,103 for Li ion batteries. In the course of the
understanding of the graphite disordering mechanism,
Kostecki et al. used Raman microscopy and atomic force
microscopy (AFM) to study the effect of structural changes
which occur in graphitic materials during Li ion cell
cycling at ambient and elevated temperatures.55,56,104 The
authors evaluated the near-surface and surface changes
resulting from the exposure of graphitic electrodes to
stresses associated with elevated temperature and numer-
ous Li intercalation-deintercalation cycles, in terms of
graphite structure disordering and SEI layer morphology,
thickness, and composition change. By applying high
resolution Raman microscopy mapping, a nonuniform
gradual structural degradation process was found in
graphite upon cycling at 60 °C. The authors also detected
the formation of nonhomogeneous electrolyte decomposi-
tion products within the bulk of the anode. The Raman
data were found to be in good correlation with the
corresponding experimental electrochemical data, indicat-
ing an increase of cell impedance and a loss of reversible
capacity. A nonuniform current density distribution was
thought to be responsible for large Li concentration
gradients within the crystalline structure of graphite
eventually surpassing the tensile stress of graphene planes.
The authors suggested that the gradual disordering of the
graphite anode during prolonged cycling leads to the
fragmentation of surface graphene and subsequent con-
tinuous SEI layer reformation reducing the reversible
capacity of the cell.55,56
It is well-known that the electrochemical behavior of the
carbon anode depends not only on the type of carbon
materials but also on the solvent and electrolyte system used
in batteries. A prominent example is the incompatibility of
propylene carbonate (PC) electrolytes with highly crystalline
graphite materials. During the electrochemical insertion of
Li+ in such electrode materials from PC-based electrolytes,
the graphite structure exfoliates, leading to severe battery
failure.69,95,106 An in situ Raman study of the graphite surface
structure revealed drastic Raman spectral changes in the high
potential range during the first discharge process in LiClO4
EC/DME solution, where large irreversible capacity losses
take place.41 Frech et al.41 suggested that this alteration of
the graphite surface structure corresponded to extensive
graphite exfoliation caused by solvent cointercalation with
lithium ions and subsequent decomposition. Hardwick et al.98
recently used in situ Raman microscopy to compare the
stability in PC of two microcrystalline graphites with
different grain sizes (3 and 24 µm) toward exfoliation. A
split of the G-band at higher wavenumber was detected only
for the sample with larger crystallite size and the largest
Electrode Materials for Lithium Batteries Chemical Reviews, 2010, Vol. 110, No. 3 1287

Figure 5. In situ Raman spectra series of the first (a) EMI+ and (b) TFSI- intercalation into graphite. Insets in the figure show Raman
spectra before (bold line) and after (thin line) one cycle. Laser excitation line: 632.8 nm. Reprinted with permission from ref 97. Copyright
2008 Elsevier.

irreversible charge loss, indicating the presence of solvated

ions intercalated into graphite. This spectral change was
associated with the beginning of the exfoliation process. In
situ Raman microscopy was also newly used to demonstrate
large cation (EMI+) and anion (TFSI-) intercalation into
microcrystalline graphite during cyclic voltammetry experi-
ments from the ionic liquid (EMI-TFSI).97 The Raman
spectral series (Figure 5) clearly displays, with the split of
the G-band at 1578 cm-1, the intercalation of these ions into
graphite. From the analysis of the intensity ratio E2g2(i)/
E2g2(b), Hardwick et al.97 show that GICs between stages 2
and 3 were formed at the potential limits of 0.55 V for
(EMI+) and 5.05 V vs Li/Li+ for (TFSI-). Furthermore, a
significant increase in the ID/IG ratio is observed after both
the anodic and cathodic cycles, which indicates graphite Figure 6. Crystal structure of R-NaFeO2-type compounds.
deterioration on anion intercalation. Such a result may have
4.1. Lithium Cobalt Oxide LiCoO2
a negative impact for the use of ionic electrolytes when
carbon additive is used as conductive agent at the high This cathodic material deintercalates lithium at a very high
voltage cathode. voltage, ca. 4 V vs Li/Li+. Three basic lithiated systems (as
well as their substitutive derivative forms) with high operat-
4. Transition Metal Oxide-Based Compounds ing voltage are presently known. These are LiCoO2 and
LiNiO2 with the pseudolayered R-NaFeO2 structure, and the
We intend to present here a review of the most prominent 3D spinel LiMn2O4. To improve their electrochemical
data reported for various metal oxide-based materials, which properties, the mixed compounds resulting from various
have been selected in order to highlight the contribution of substitutions on the transition metal site have been investi-
Raman spectroscopy in the area. The results of these studies gated in these structures.
have provided a better understanding of the different In 1980, Mizushima et al.107 proposed using layered LiCoO2
with the R-NaFeO2 structure as an intercalation cathode. It
processes responsible for the loss of electrochemical per-
took around 10 years to put LiCoO2 to commercial use. This
formance, which constitutes a key step to have an idea about
oxide is now mainly used as the cathode material in
improving the electrode material. Particular attention will commercial lithium ion batteries.108 The layered structure of
be paid to the LixV2O5 and the LixTiO2 systems, for which LiCoO2 is shown in Figure 6. LiNiO2, LiCrO2, and LiVO2
a thorough analysis has been provided using Raman mi- also adopt this structure. These LiMO2 compounds, prepared
crospectrometry thanks to a careful and rigorous experimental at temperature ranges of 700-900 °C, are rock salt-structured
approach combined with a theoretical analysis based upon materials based on a close-packed network of oxygen atoms
lattice dynamics simulations. with Li+ and M3+ ions in octahedral interstices in this
1288 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 7. Atomic displacements of the IR and Raman-active modes
of hexagonal (R3jm) LiCoO2.
packing in alternating (111) planes. This (111) ordering
introduces a slight distortion of the lattice to hexagonal
symmetry.
LiCoO2 crystallizes in the hexagonal (R3jm) space group
5 ) with a unit cell consisting of one formula unit (Z ) 1)
(D3d
and unit cell parameters ahex ) 2.82 Å and chex ) 14.08 Å.
The atoms of the LiCoO2 units are on sites with symmetries
and coordinates given below:
Co: (3a, D3d) 0, 0, 0
Li: (3b, D3d) 0, 0, 1/2
O: (6c, C3V) 0, 0, u 0, 0, uj
By factor group analysis, the total irreducible representa-
tion for the vibrational modes of LiMO2 is obtained as A1g
+ 2A2u + Eg + 2Eu. The gerade modes are Raman-active,
and the ungerade modes are IR-active. The two Raman-
active modes are especially simple. In the A1g mode, two
adjacent oxygen layers move rigidly against each other and
parallel to the c axis, whereas the atomic displacements in
the Eg mode are perpendicular to the c axis (Figure 7).
The Raman spectrum of LiCoO2 was first reported by
Inaba et al.,109 who studied the effects of replacing Co in
LiCoO2 by Ni. The Raman-active lattice modes were later
assigned by the same team from polarized Raman measure-
ments on a c-axis oriented LiCoO2 thin film.110 Two Raman-
active modes have been found, which correspond to oxygen
vibrations involving mainly Co-O stretching, ν1(A1g) at 595
cm-1, and O-Co-O bending, ν2(Eg) at 485 cm-1.
The mechanism of lithium deintercalation and intercalation
in LiCoO2 has been investigated using XRD.108,111,112 The
discharge curve and the lattice parameters changes in bulk
Li1-xCoO2 are shown in Figure 8. The a axis remains
practically constant, whereas the c axis increases from
≈14.08 to 14.45 Å for 0 < x < 0.5, due to the production of
a second hexagonal phase with an expanded c parameter.108
The expansion in the c axis has been ascribed to an increase
in the electrostatic repulsion between adjacent CoO2 layers
because negatively charged oxygen-oxygen interactions
increase with the removal of lithium ions.112 Two monoclinic
phases were also reported, one at about x ) 0.45, and another
for 0.75 < x < 1.108 One single study reports the Raman
spectra of Li1-xCoO2 powder prepared by electrochemical
lithium deintercalation.110 The spectral changes were well
correlated with the structural changes determined by X-ray
diffraction, namely as a series of phase transitions. In
particular, a set of two new bands, located at lower
wavenumbers, are observed for the second hexagonal phase
(Figure 9). These downward shifts of both bands have been
Baddour-Hadjean and Pereira-Ramos
found in good agreement with the increase of the c-axis
length as lithium ions are deinserted. However, the expected
Raman peak splitting due to the distortion in the monoclinic
phase was not observed. In situ Raman spectroscopy has also
been conducted on thin film electrodes of pure LiCoO2.35,113
An interesting feature concerns the reported invariance of
the peak frequency with the electrode potential,35 which is
not discussed by Itoh et al. but can be correlated to the
specific structural response of a LiCoO2 thin film reported
later by Kim et al.114 from XRD measurements on thin films,
that is different from the structural behavior usually known
for the LiCoO2 powder. Indeed, when the cutoff voltage is
limited to the conventional value of 4.2 V corresponding to
the Li0.5CoO2 material, the XRD patterns of charged films
at 4.2 V showed negligible change in the c lattice parameter.
Five or ten cycles are needed to induce the appearance of
the conventional Li0.5CoO2 expanded phase observed from
the first charge for the bulk material. In another work, Itoh
et al.113 reported that applying potentials more positive that
4.7 V leads to a sudden increase of the Raman background
signal which was ascribed to the formation of a film on the
LiCoO2 electrode surface in organic solution (LiClO4/PC or
EC).
Micro-Raman spectrometry constitutes a convenient and
powerful technique for the qualitative microstructural analysis
of LiCoO2 cathodes. Because Raman spectroscopy is capable
of detecting unambiguously the nature of the cobalt oxide
phases present, lithiated or not, it has been extensively used
for the characterization of powders115-117 and thin film
electrodes.118-127 Considering the fabrication of the LiCoO2
compound, especially when the thin film material is prepared
at low temperature, Raman microspectrometry is thoroughly
used as a quality control tool to check the nature and
crystallinity of the LiCoO2 active phase as a function of the
synthesis conditions and to detect the presence of residual
secondary phases, such as Co3O4 or Li2CO3.118-127 An
important point concerns the formation of the layered LiCoO2
structure, which is known to be crucial for obtaining a good
rechargeability of the cell. Indeed, a low temperature spinel
LiCoO2 phase (prepared at 400 °C and denoted as LT-LiCoO2)
has been reported, with different electrochemical features
compared to the high temperature layered LiCoO2 structure
(prepared at 850 °C and denoted as HT-LiCoO2).115,128-130
Comparison of cyclic voltammetric curves for LT- and HT-
LiCoO2 emphasizes a large difference in the electrochemical
behavior (Figure 10). The LT LiCoO2 is characterized by
unusual broad anodic and cathodic peaks with a difference
between the peak potentials greater than 200 mV; these latter
are located at 3.75 and 3.45 V. A faradaic yield around 0.4
F/mol is involved in both cases.
Standard powder neutron diffraction and X-ray diffraction
cannot unambiguously distinguish between layered and spinel
LiCoO2.128-131 Conversely, from a spectroscopic viewpoint,
the ideal spinel type LiCoO2 belongs to the space group
Fd3m, for which the Bravais cell contains four molecules
(Z ) 4), and four Raman-active modes, A1g, Eg, and 2F2g,
are predicted.131 As shown in Figure 11, each LiCoO2 crystal
structure gives rise to a specific Raman fingerprint; that is,
four Raman bands are observed at ca. 605, 590, 484, and
449 cm-1 for LT-LiCoO2, whereas only two Raman bands
at 597 and 487 cm-1 are observed for the layered
HT-LiCoO2.131,132 This result is consistent with the theoretical
prediction given for a spinel (Fd3m) and hexagonal (R3jm)
crystal, respectively. These spectral disparities justify the
Electrode Materials for Lithium Batteries
Figure 8. (a) Continuous charging curve of Li1-xCoO2 at a rate of 0.17 mA · cm-2 at 30 °C. (b) Lattice parameters of the hexagonal unit
cell of Li1-xCoO2. Triangles indicate the converted unit cell parameters from the monoclinic cell. Reprinted with permission from ref 108.
Copyright 1994 The Electrochemical Society.
Figure 9. Variations of the peak wavenumbers of the upper (closed
circles) and lower (open circles) Raman bands with x in Li1-xCoO2.
Symbols in parentheses designate poorly fitted data. Excitation with
514.5 nm radiation. Reprinted with permission from ref 110.
Copyright 1997 Wiley.
Figure 10. Cyclic voltammetric curves at 40 µV/s for HT-LiCoO2
and LT-LiCoO2. Reprinted with permission from ref 130. Copyright
1997 The Electrochemical Society.
systematic use of Raman spectroscopy for the structural
determination of lithiated cobalt oxide phases. When used
as a thin protective layer on the nickel cathode of a molten
carbonate fuel cell (MCFC), LiCoO2 is also of particular
Chemical Reviews, 2010, Vol. 110, No. 3 1289
Figure 11. Raman scattering spectra for (a) spinel LT-LiCoO2 and
(b) layered HT- LiCoO2. Excitation with 514.5 nm radiation.
Reprinted with permission from ref 132. Copyright 2004 Elsevier.
interest due to its good mechanical resistance and low cost.
In this field, Raman spectroscopy is also of interest, as it
has allowed, for instance, obtainment of relevant information
on the nature and stability of cobalt oxides layers when
exposed to a second-generation electrolyte, Li2CO3-Na2CO3,
at 650 °C,132 showing the in situ formation of a cubic LT-
LiCoO2 thin layer, instead of the usually layered HT-
compound reported in molten carbonate media.133,134
Over the past few years, attention has been focused toward
application of Raman spectroscopy as an in situ vibrational
probe of electrode materials in an operating Li ion battery
during discharge-charge cycles. A specially designed in situ
cell was developed by Novak et al.46 with LiCoO2 as cathode
material and graphite as anode (Figure 12). This approach
allowed access to the Raman signature of LiCoO2 particles
randomly selected on the surface of the commercial electrode.
The authors observed that, upon lithium insertion into the
host material, the background intensity rises significantly
whereas, upon lithium deinsertion, the background intensity
is nearly constant. These results were found in accordance
with those reported earlier by Itoh at al. for a three electrode
configuration.35 More recently, in situ Raman measurements
1290 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 12. Example of an electrochemical cell developed for in
situ Raman microscopy. Reprinted with permission from ref 46.
Copyright 2000 Elsevier.
Figure 13. Raman scattering spectra of LiNi1-yCoyO2 powder
samples. Excitation with 514.5 nm radiation. Reprinted with permission
from ref 109. Copyright 1995 The Chemical Society of Japan.
in an operating Li ion battery during discharge were carried
out with the aim to construct time-resolved, two-dimensional
maps of the state of charge within the electrodes.60 However,
if clear evidence was obtained for changes in the amount of
Li+ within particles of graphite during battery discharge, this
effect was observed to a lesser extent for the LiCoO2 cathode.
4.2. Lithium Nickel Oxide LiNiO2 and Its Substitutive
Derivative Compounds LiNi1-yCoyO2 (0 < y < 1)
The limited capacity of LiCoO2, its high cost, and toxicity
are considered as drawbacks, especially for large scale applica-
tions. In order to overcome these problems, another layered
compound of interest, LiNiO2, has been investigated. However,
this material suffers from a tendency to nonstoichiometry in
relation with the presence of an excess of nickel135 and from
poor thermal stability in its highly oxidized state (Ni3+/Ni4+).
Indeed, the metastable layered structure Li0.5NiO2 transforms
into the cubic spinel on heating to 300 °C. The substitution
of nickel by cobalt has been reported to be an easy way to
stabilize the two-dimensional (2D) structure. The use of the
solid solution LiNi1-yCoyO2 (0 < y < 1), has therefore been
explored, and improved electrochemical properties have been
reported.136 The Raman spectra of lithium-cobalt-nickel
oxides have been reported by many authors.109,131,137,138 It
turns out that replacing Co by Ni does not change the space
group, but both bands assigned to the Eg and A1g Raman-
active modes in (R3jm) symmetry are found to decrease
drastically in intensity with increasing Ni content (Figure
13). Hence, the Raman scattering efficiency of LiNiO2
appears to be very weak in comparison to that of other rock-
salt compounds. The origin of these features was ascribed
Baddour-Hadjean and Pereira-Ramos
to a reduction of the rhombohedral distortion by increasing
Ni content or/and an increase in the electrical conductivity
in LiNiO2.109 Conversely, the Raman spectrum of a PLD
deposited LiNi0.8Co0.2O2 film recently reported by Ramana
et al.139 exhibits intense features at 478 and 587 cm-1 which
are not discussed by the authors, whereas they are contradic-
tory toward previous data related with the same composi-
tion.109 Another effect of cobalt substitution in LiNi1-yCoyO2
powdered samples concerns the observed shift of both Raman
bands toward higher frequencies, from 544 cm-1 in LiNiO2
to 595 cm-1 in LiCoO2 for the A1g mode and from 465 cm-1
in LiNiO2 to 485 cm-1 in LiCoO2 for the Eg mode.137 These
frequency shifts are consistent with the observed decrease
of the hexagonal unit cell parameters as the Co content
increases (from 2.86 Å to 2.82 Å for the intralayer
metal-metal distance ahex and from 14.15 Å to 14.08 Å for
the interlayer parameter chex) when y varies from 0.3 to 1.138
It is suggested that the observed increase in vibrational
frequencies with the cobalt content is related to the increase
in the bond covalency inside the layers.
With the aim to better understand Li ion battery cell
performances and the factors that limit battery lifetime, Kostecki
et al. carried out several studies to characterize the degradation
processes occurring on the surface of multisubstituted-based
LiNi1-y-zCoyAlzO2-based electrodes during storage or electro-
chemical cycling under different conditions.49-54,61,140-142 Using
a set of microscale and nanoscale techniques (e.g., Raman
microscopy, current-sensing atomic force microscopy, FTIR,
XPS, EDX, XRD, NMR) coupled to the electrochemical
analysis, the authors obtained maps yielding information about
the chemical composition and structure of the electrode material.
Raman microscopy mapping of LiNi0.8Co0.15Al0.05O2 composite
cathodes harvested from high power Li ion cells showed that
the surface distribution of carbon additives in these cathodes
changes noticeably upon aging and/or cycling at elevated
temperature52 and also upon cycling over different depths of
discharge.53 The noticeable difference in the active material
to carbon concentration ratio at the surface of the cathode
was thought to be at least partially responsible for the
observed power and capacity losses. Very recently, the same
group put forward the nonuniform local kinetic behavior of
individual oxide particles using in situ61 and mapping Raman
microscopy.50 The micro-Raman spectra of the tested
LiNi0.8Co0.15Al0.05O2 cathode50 provided clear evidence of the
inhomogeneous state of charge (SOC) of oxide particles on
the cathode surface despite deep discharge of the Li ion cells
at the end of the test. As shown in Figure 14, the spectra of
oxide particles vary significantly as a function of location
on the cathode surface. The 475/554 band relative intensity
and peak half-width were found to vary noticeably with
lithium content in the lattice from in situ Raman measure-
ments.61 This has allowed the authors to assign the Raman
features of the cycled cathode to fully discharged, partially
charged, and fully charged Li1-xNi0.8Co0.15Al0.05O2 particles.
From these results, the authors proposed that electrode
degradation is due to a deterioration of electronic contact
resistance within the composite cathode. Carbon additive
rearrangement, formation of surface films, as well as the poor
intrinsic properties of the oxide active material would
contribute to this nonuniform local kinetic behavior, which
constitutes a common degradation mode for composite Li
ion cathodes.50,61
Since the interesting observation of high discharge capacity
with Li(Ni1/3Co1/3Mn1/3)O2 material in 2001 by Ohzuku and
Electrode Materials for Lithium Batteries
Figure 14. (a) Average Raman spectrum of the fresh composite
LiNi0.8Co0.15Al0.05O2 electrode; (b) 52 µm × 75 µm Raman image
of the composite LiNi0.8Co0.15Al0.05O2 cathode from the cycled cell.
The image was collected at 0.7 µm resolution. The intensities of
red, blue, and green colors correspond to the integrated band
intensities of LiNi0.8Co0.15Al0.05O2, and D, G carbon bands of each
spectrum, respectively. (c) Raman microscope spectra of three
individual LiNi0.8Co0.15Al0.05O2 particles in the tested cathode.
Reprinted with permission from ref 50. Copyright 2007 Elsevier.
Makimura1,143 enormous research on the layered cathode
materials based on Li(NixCoyMn1-x-y)O2 with various values
of x and y has been reported in the literature.144-148 A latest
ex situ structural study by X-ray diffraction and Raman
scattering at various stages of charging and discharging of
the Li(Ni0.8Co0.1Mn0.1)O2 cathode material showed that the
host layered structure is maintained throughout the electro-
chemical lithiation delithitation process in the 3-4.5 V
range.148 XRD data evidenced a reversible and continuous
change in the lattice parameters, with a decreasing and c
increasing, during the electrochemical extraction, while the
Raman spectra exhibit two peaks at ∼500 and 560 cm-1
characteristic of the rhombohedral layered symmetry, with
slight frequency variations.
4.3. Manganese Oxide-Based Compounds
Manganese oxides (MOs) with three- and two-dimensional
crystal networks constitute a large family of porous materials
that can accept foreign species in their tunnel or interlayer space.
There is a wide variety of natural and synthetic MOs, including
the various allotropic forms of MnO2 and ternary lithiated com-
pounds LixMnOy. Having good electrochemical performance,
they are attractive as positive electrode materials for lithium
cells because manganese has economical and environmental
advantages over compounds based on cobalt or nickel.149,150
4.3.1. MnO2-Type Compounds
MnO2 was originally developed as the positive electrode
for primary alkaline batteries.151 Extensive research was
carried out during the last few decades to improve the
Chemical Reviews, 2010, Vol. 110, No. 3 1291
Figure 15. Schematic representation of the various manganese
dioxide frameworks showing the variation in the chain and tunnel
(m × n) structures. (a) pyrolusite (1 × 1), (b) ramsdellite (1 × 2),
(c) hollandite (2 × 2), (d) birnessite (1 × ∞), (e) romanechite (2 ×
3), and (f) spinel (1 × 1). Reprinted with permission from ref 158.
Copyright 2004 Elsevier.
reversibility of lithium insertion in the manganese dioxide
cathode for rechargeable Li//MnO2 cells.152 Since then, much
effort has been devoted to the study of lithium insertion into
various forms of manganese dioxides, especially synthetic
products prepared by either electrolytic (EMD) or chemical
(CMD) methods that belong to the nsutite (γ-MnO2) group,
for use as cathodes in lithium batteries.153,154 Lithium
accommodation in γ-MnO2 occurs predominantly by inser-
tion into the (2 × 1) tunnels of the ramsdellite (R-MnO2)
domains, while the β-MnO2 domains only accommodate 0.2
Li in the (1 × 1) channels.73 Besides the γ-MnO2 form, which
suffers from moderate capacity, low reversibility, and poor
cycling stability, there are many MO materials under study,
such as LiMn2O4 spinel-like phases,150,155 layered LiMnO2,156
and new layered phases, such as hexagonal RLi0.51Mn0.93O2
and orthorhombic βLi0.52MnO2.157
The common crystallographic unit building the lattice of
MOs and that of their lithiated products is the basal MnO6
octahedron. Table 2158 summarizes the crystallographic data
of various MO compounds. As shown in Figure 15, their
structure can be described as a close packed network of
oxygen atoms consisting of edge- and corner-sharing MnO6
octahedra forming tunnels of various sizes for the insertion
of Li ions, leading to more or less compact structures in
which Mn4+ and/or Mn3+ ions are distributed.
Figure 16 shows the Raman scattering spectra reported
for various manganese dioxide compounds.158,159 Because the
Raman cross section of Mn-oxide is relatively low, the
Raman features of MOs are rather weak. Three major regions
can nevertheless be distinguished: at 200-450, 450-550,
and 550-750 cm-1. They correspond to spectral domains
where the skeletal vibrations, the deformation modes of the
1292 Chemical Reviews, 2010, Vol. 110, No. 3 Baddour-Hadjean and Pereira-Ramos

Table 2. Crystallographic Data of Some MO Compoundsa

compd mineral crystal symmetry lattice parameters (Å) features
MnO manganosite cubic (Fm3m) a ) 4.44 rock-salt
R-MnO2 hollandite tetragonal (I4/m) a ) 9.96; c ) 2.85 (2 × 2) tunnel
R-MnO2 ramsdellite orthorhombic (Pbnm) a ) 4.53; b ) 9.27; c ) 2.87 (1 × 2) tunnel
β-MnO2 pyrolusite tetragonal (P42/mnm) a ) 4.39; c ) 2.87 (1 × 1) tunnel
γ-MnO2 nsutite complex tunnel (hex) a ) 9.65; c ) 4.43 (1 × 1)/(1 × 2)
δ-MnO2 vemadite hexagonal a ) 2.86; c ) 4.7 (1 × ∞) layer
λ-MnO2 spinel cubic (Fd3m) a ) 8.04 (1 × 1) tunnel
MnOx, H2O birnessite tetragonal ahex ) 2.84; chex ) 14.64 (1 × ∞) layer
MnOOH groutite orthorhombic (Pbnm) a ) 4.56; b ) 10.70; c ) 2.87
R-Mn2O3 bixbyite cubic (Ia3) a ) 9.41 C-type
Mn3O4 haussmannite tetragonal (I41/amd) a ) 9.81; c ) 2.85 spinel-like
a
Reprinted with permission from ref.158 Copyright 2004 Elsevier.

Figure 16. Raman scattering spectra reported for various manganese dioxide frameworks. Pr is the intergrowth rate of the pyrolusite into
the ramsdellite matrix. Excitation with 514.5 nm radiation. (a) Reprinted with permission from ref 158. Copyright 2004 Elsevier. (b) Reprinted
with permission from ref 159. Copyright 2006 Elsevier.

metal-oxygen chain of Mn-O-Mn in the MnO2 octahedral
lattice, and the stretching modes of the Mn-O bonds in
MnO6 octahedra occur, respectively.
Few works reported the vibrational spectra of pyrolusite
β-MnO2, ramsdellite R-MnO2, and γ-MnO2 compounds. A
careful examination of the results published up to now shows
discordance regarding Raman spectra reported by different
researchers. Strohmeier at al.160 and Kapteijn et al.161 found
that MnO2 was not Raman-active, whereas Gosztola et al.,162
Bernard et al.,163 Buciuman et al.,164 Widjaja et al.,165 and
Julien et al.158,159 reported Raman spectra of MnO2 with
different spectral features. These disagreements are due to
various factors as follows:
(i) the confusing structural characterization of MnO2
compounds (X-ray diffraction patterns are frequently broad,
indistinct, or absent);
(ii) the wide variety of synthesis routes, which determine
the structural and physicochemical properties (the structural
differences are commonly attributed to variations in oxygen
stoichiometry, Mn oxidation state, particle size, and the type
of defect chemistry);
(iii) the low Raman activity for most of the manganese
oxides;
(iv) the high sensitivity of many manganese compounds
(contrary to nickel compounds) under the laser beam. Most
MOs are so black that they absorb the photon energy, which
results in local heating at the localized region. This point
should be considered with particular attention, since local
heating can cause misleading shifts and broadening of the
Raman modes due to photoinduced or thermal-induced
chemical reactions. Indeed, reduction reactions are easily
produced, leading to degradation compounds such as Mn2O3
and Mn3O4.165-167 It follows that accurate and reliable
determination of the pure Raman components of MO
compounds is not yet elucidated and that particular caution
should be taken toward reported Raman spectra.
The birnessite-type manganese oxides constitute another
class of materials with layered structure, with water mol-
ecules and/or metal cations occupying the interlayer region
(Figure 15d). It has been recently shown that attractive
electrochemical performances, with stable capacities of 170
(mA h)/g after 40 cycles at C/20 (Figure 17) could be reached
for sol-gel prepared birnessite doped with Co (SGCo-Bir)
with chemical formula Co0.15Mn0.85O1.84 · 0.6H2O.168 The
Raman features of several birnessite compounds were first
reported by Julien et al.169 As shown in Figure 18, in spite
of a slight variation in band positions and relative band
intensity, the general similarity of the spectra suggests that
samples are characterized by the same basic structure. In
fact, MnO6 octahedral layers are separated by layers of lower-
valent cations (Li+, Na+, Mn2+, etc.) and by layers of water.
The highest Raman band is assigned to the symmetric
vibration ν(Mn-O) of the MnO6 group, with A1g symmetry
in the Oh7 spectroscopic space group. This mode is observed
Electrode Materials for Lithium Batteries
Figure 17. Comparison of the evolution of the specific capacity
with the number of cyles at C/20 for the sol-gel birnessite (O)
and for the sol-gel Co-Birnessite (b). Reprinted with permission
from ref 168. Copyright 2002 Elsevier.
Figure 18. Raman spectra of birnessite-type manganese oxides (a)
MnO1.84 · 0.6H2O, (b) Co0.15Mn0.85O1.84 · 0.6H2O, (c) Na0.32MnO2 · 0.6H2O,
and (d) Li0.32MnO2 · 0.6H2O. Excitation with 514.5 nm radiation.
Reprinted with permission from ref 169. Copyright 2003 Elsevier.
at 625 and 640 cm-1 for Li-Bir and Na-Bir, respectively,
and at 646 and 638 cm-1 for SG-Bir and SGCo-Bir,
respectively. A correlation between the wavenumber value
of this stretching mode and the interlayer d-spacing was
proposed.169 The band located at 575 cm-1 is attributed to
the ν(Mn-O) stretching vibration with F2g symmetry and is
commonly related to the vibrational stretching frequency
inherent to the presence of Mn4+ ions. Its intensity is
particularly strong in birnessite compound compared with
the literature data related to lithiated spinels due to the high
content of Mn4+ in the birnessite family.168-170 The Raman
spectrum of SGCo-Bir, where cobalt partly substitutes for
manganese, displays similar features to SG-Bir, MnO1.84,
0.6H2O. As a result of the substitution of Mn4+ by Co3+ ions
in MnO2 layers,168 a frequency shift of 10 cm-1, from 575
to 585 cm-1, was observed169 for the ν(Mn-O) stretching
vibration with F2g symmetry (Figure 18). This result was
correlated with a strengthening of the Mn-O bond in the
Table 3. Structure and Raman Activity for Various Lithiated Transition-Metal Oxide Materials
compd type of structure
LiCoO2 layered hexagonal rock-salt
mLiMnO2 layered monoclinic rock-salt
LiMn2O4 normal cubic spinel
λMnO2 modified cubic spinel
Li2Mn2O4 normal tetragonal spinel
Li0.5Mn2O4 ordered cubic spinel
Chemical Reviews, 2010, Vol. 110, No. 3 1293
Figure 19. Li-Mn-O phase diagram showing the compositions
of spinel, defect spinel, and rock salt structures. Reprinted with
permission from ref 171. Copyright 2003 Elsevier.
Co-doped SG-bir. This finding is in good accordance with
the better structural stability of the host lattice for the Co-
doped material during cycling, as illustrated in Figure 17.
4.3.2. Ternary Lithiated LixMnOy Compounds
Lithiated transition manganese oxides involved as positive
electrode materials in high voltage lithium ion batteries
exhibit different crystallographic structures. As a conse-
quence, the number of active bands in the vibrational spectra
of these compounds varies significantly, depending on their
local symmetry (Table 3). Two classes of materials are
currently being studied: the spinel-type and the rock-salt-
type compounds. The Li-Mn-O phase diagram shown in
Figure 19 highlights the positions of spinel and rock salt
compositions within the λ-MnO2, MnO, and Li2MnO3 tie-
triangle. It emphasizes the wide range of spinel and rock
salt compositions that exist in the Li-Mn-O system.
4.3.2.1. Stoichiometric Spinel Phases in the Li-Mn-O
System. Stoichiometric spinels fall on the tie line between
Mn3O4 and Li4Mn5O12. They are defined by a cation/anion
ratio M/O of 3/4. The stoichiometric spinels of importance
for lithium battery applications form a solid solution between
LiMn2O4 and Li4Mn5O12. They are considered currently of
technological interest as insertion electrodes for rechargeable
4 V lithium batteries because of their diffusion pathways
for Li ions, their low cost, low toxicity, and high energy
density (due to the combination of high capacity and high
voltage).150,171-174 The structural, chemical, and electrochemi-
cal properties of the LiMn2O4 system have been extensively
reported.150,171,175 LiMn2O4 has a normal spinel structure,
which is cubic with the space group Fd3m (O7h) containing
8 AB2O4 units per unit cell. It can be represented by the
formula {Li}8a[Mn2]16dO4, in which the subscript 8a indicates
occupancy of tetrahedral sites by Li+ ions, with Mn3+ and
space group Raman activity
D3d5 - R3m
j A1g + Eg
C32h - C2/m 2Ag + Bg
O7h - Fd3m A1g + Eg + 3F2g
O7h - Fd3m A1g + Eg + 2T2g
D194h - I41/amd 2A1g + 2B1g + 6Eg + 4B2g
T2d - F4j3m 3A1 + 3E + 6F2
1294 Chemical Reviews, 2010, Vol. 110, No. 3 Baddour-Hadjean and Pereira-Ramos

a different MnO6 unit and a three-dimensional network of

octahedral 16c holes and tetrahedral 8a sites. When the
lithium ion diffuses in the structure, first it moves from the
8a site to the neighboring empty octahedral 16c site and then
to the next 8a site in such a way that the Li+ takes the
diffusion path (8a-16c-8a) (Figure 20b).
Vibrational features of spinel oxide structures have been
widely studied. Tarte et al. published numerous infrared data
on spinels and solid solutions.176-178 White et al. treated the
vibrational spectra of spinel structures by a factor group
analysis.179 The various types of ordering were considered,
and the predicted changes were compared with literature data.
Early spectroscopic data on electrochemically prepared
Figure 20. Schematic representation of the structure of AB2O4 spinel spinel-type LixMn2O4 composite electrodes were obtained
lattices showing (a) the smallest (primitive) cubic unit cell of a normal
spinel (Fd3m space group) and (b) the diffusion path of lithium. from FTIR measurements.180,181 Observation of LiMn2O4
Reprinted with permission from ref 73. Copyright 2001 Elsevier. Raman features at 593 and 628 cm-1 was first reported from
in situ measurements of lithium ion electroinsertion in a
Mn4+ ions (in 1:1 mixture) being randomly distributed over spinel-type Pt/λ-MnO2 electrode.182 Ammundsen et al.183
the octahedral 16d sites and oxygen anions in 32e Wyckoff calculated the lattice dynamics of spinel-structured lithium
positions (Figure 20a). The approximately cubic close-packed manganese oxides using atomistic modeling methods, which
array of oxide ions incorporates MnO6 octahedra, connected allowed the prediction and the assignment of the Raman
to one another in three dimensions by edge sharing (Figure spectra of lithiated, fully delithiated, and partially delithiated
15f) LiO4 tetrahedra, sharing each of their four corners with LixMn2O4 thin films. In good agreement with the calculations,
Table 4. Calculated and Observed Wavenumbers (in cm-1) for Raman-Active Modes of LiMn2O4, λ- MnO2, and Li0.5Mn2O4 Thin
Films, Assuming a F4j3m Space Group for the Latesta
LiMn2O4 λ-MnO2 Li0.5Mn2O4
calcd obsd calcd obsd symmetry species calcd obsd symmetry species
354 365 382b T2g(1) 313 296 296b T2
434 432 426b 479 463 462b Eg 317 E
455 480 483b 511 498 501b T2g(2) 385 380b T2
597 590 580b 630 647 644b T2g(3) 457 425b E
598 625 625b 592 592 596b A1g 481 493 483b T2
537 522b T2
559 560 563b A1
586 E
593 597 596b A1
615 612 611b T2
646 630 648b T2
665 657 A1
a
Most values are reported from ref 183. b Values reported from ref 171.

Figure 21. Raman spectra of spinel LiMn2O4, Li0.5Mn2O4, λ-MnO2, and Li2Mn2O4. Excitation with 514.5 nm radiation. Reprinted with
permission from ref 171. Copyright 2003 Elsevier.
Electrode Materials for Lithium Batteries
five Raman-active modes are observed for the LiMn2O4
phase, one of A1g symmetry at 625 cm-1, one of Eg symmetry
at 426-432 cm-1, and three of T2g symmetry at 365-380,
480-485, and 580-590 cm-1 (Table 4 and Figure 21). The
slight underestimation of the calculated A1g mode for
LiMn2O4 (598 instead of 625 cm-1) was ascribed to a small
deviation from the model used to describe the electronic
polarizability of oxygen ions in LiMn2O4 (transferred without
any change from the λ-MnO2 model).183 In a localized
vibration analysis, Julien et al.171 attributed the broadness
of the highest frequency A1g mode to the existence of two
types of Mn-O vibrating entities: isotropic Mn+IVO6 octa-
hedra and locally distorted Mn+IIIO6 octahedra. According
to the same author, the shoulder peak around 580 cm-1 of
T2g symmetry would mainly originate from the stretching
vibration of the Mn+IV_O bond. Its intensity would depend
on the Mn+IV concentration in the spinel and reflect the Mn
average oxidation state for this reason. The lowest energy
T2g(1) phonon would predominantly derive from a vibration
of the Li sublattice and can be viewed as a Li-O stretching
motion.
On studying the effect of chromium substitution on the
lattice vibration of spinel lithium manganate by performing
a combinative micro-Raman and X-ray absorption (XAS)
analysis, Hwang et al.184 provided a new interpretation of
the Raman spectrum of LiMn2O4. On the basis of the
dynamic local structural information obtained from XANES/
EXAFS experiments, the authors assigned the two intense
features at ∼580 and ∼620 cm-1 as totally symmetric A1g
modes for regular Mn+IVO6 octahedra and tetragonally
distorted Mn+IIIO6 octahedra, respectively. This new inter-
pretation was further supported by the comparison with the
reference spectra of Li2Mn2O4 and λ-MnO2 and the good
correlation between the ratio of Mn+IV/Mn+III concentration
and the relative intensities of the two Raman peaks.184 An
interesting result provided by this study is that the relative
intensity of the two main Raman peaks of the spinel lithium
manganate can be used as a measure for estimating the local
symmetry and the oxidation state of manganese ion in the
spinel lattice.
Therehavealreadybeenmanyreportsonthecharge-discharge
characteristics of the LixMn2O4 cathode.173-175 The theoretical
specific capacity of LiMn2O4 is 148 (mA h)/g. The open
circuit voltage (OCV) curve of LixMn2O4 (0 < x < 2) is shown
in Figure 22.
The structural changes of LixMn2O4 powder electrodes
during charge and discharge have been mainly characterized
by X-ray diffraction173,185-187 and neutron powder diffrac-
tometry.185 However, due to the high symmetry of the spinel
system, the low X-ray scattering power of lithium, and the
occurrence of partial occupation of the various cation sites,
the XRD patterns are not always easily interpreted. It is often
difficult not only to determine the precise distribution of
cations in a given pure phase but also to quantitatively
analyze multiphase mixtures or to distinguish between
different phases with similar lattice constants. This problem
is especially acute for LixMn2O4 with 0.2 e x e 1, as the
cubic cell parameter changes by only 3% over the entire
composition range. Diffraction data obtained from electrodes
is commonly of poor quality due to degradation of crystal-
linity during cycling, the presence of other phases (carbon,
electrolyte), and compositional inhomogeneity.
While vibrational spectroscopy cannot give the detailed
structural information available from high quality diffraction
Chemical Reviews, 2010, Vol. 110, No. 3 1295
Figure 22. Open circuit voltage (OCV) curve of LixMn2O4 (0 < x
< 2) at 30 °C. Reprinted with permission from ref 173. Copyright
1990 The Electrochemical Society.
data, it is sensitive to the local environments of lithium and
transition metal cations in the oxide lattice. The number,
frequency, and relative intensities of the vibrational bands
depend upon both coordination geometries and bond strengths
(which may, in turn, depend on occupancy factors and
oxidation states). Poorly crystalline or amorphous compo-
nents that make only broad background contributions to XRD
patterns are, however, represented quantitatively in FTIR and
Raman spectra.
Lithium may be electrochemically extracted from
LiMn2O4, at a potential of about 4 V vs metallic lithium,
corresponding to the (I) + (II) region in Figure 22, according
to the oxidation reaction:
LiMn2O4 f LixMn2O4 + (1 - x)e- +
(1 - x)Li+ 0.2 e x e 1
In this composition range, LixMn2O4 retains the cubic
symmetry and appears to be composed of a single phase for
0.5 < x < 1. As lithium is removed from the tetrahedral 8a
site, there is a monotonic decrease in the cubic cell parameter.
The OCV curve is nearly flat, showing two distinct plateaus
above and below x ) 0.5. The flat part of the discharge curve
in the range x < 0.5 is considered to be a two-phase region
in which a new cubic phase is produced and exists up to
full electrochemical delithiation near a composition x ≈
0.015. The Li0.015Mn2O4 phase (termed λ-MnO2) has a
slightly modified cubic symmetry, with the removal of
lithium ions resulting in the loss of the T2g(1) phonon at low
wavenumber 365-382 cm-1 (see Tables 3 and 4). Phonon
lattice calculations predict that the T2g and Eg modes should
increase in wavenumber whereas the A1g mode should
decrease in wavenumber183 (Table 4). However, this predic-
tion is not supported by any structural consideration. The
experimental Raman spectrum of λ-MnO2, reported in Figure
21, is in good accordance with these calculations, in terms
of both number of bands and frequencies.
For intermediate compositions between LiMn2O4 and
λ-MnO2, the local symmetry no longer belongs to the Fd3m
space group. A lower symmetry has been proposed with the
F4j3m (T2d) space group (Table 3). It turns out that a richer
Raman fingerprint is expected for the partially delithiated
1296 Chemical Reviews, 2010, Vol. 110, No. 3
compounds, with 12 expected Raman-active modes instead
of 5 for LiMn2O4. This analysis is supported by reported
experimental Raman data for the Li0.5Mn2O4 phase, sum-
marized in Table 4 and Figure 21. In this structure, every
second lithium tetrahedral site is vacant, producing an
ordered Li configuration, which has been experimentally
observed for the Li0.5Mn2O4 composition.188 This subtle
ordering transition could be responsible for the flat potential
curve observed for 0.5 < x < 1 (Figure 22).
Continuous charge-discharge cycling of LixMn2O4 elec-
trodes in the high voltage region (I + II) results in significant
capacity fading, particularly when they are charged at
potentials greater than 4 V. A FTIR spectroscopy study has
demonstrated that overcharging causes a gradual conversion
to a lithium-poor defect spinel material Via dissolution of
manganese ions in the electrolyte.181
Electrochemical insertion of lithium into LiMn2O4 pro-
ceeds in a very different manner for 1 e x e 2, in the flat
region (III), according to a reduction reaction involving a
phase transition from cubic spinel to an ordered tetragonal,
Li2Mn2O4 phase:
LiMn2O4 + e- + Li+ f Li2Mn2O4
For 1 < x < 2, LixMn2O4 is a mixture of two distinct phases:
the original cubic phase and a tetragonal spinel-type com-
pound. This phase transition is provoked by the Jahn-Teller
distortion due to the increase in Mn3+ (3d4) concentration
during intercalation of lithium in LiMn2O4. As a conse-
quence, the crystal symmetry is lowered from cubic to
tetragonal, which results in a volume expansion of about
6.4%. Mechanical degradation due to the large volume
change during the cubic-to-tetragonal phase transition is the
primary cause of capacity fading in the 3 V region (1 < x <
2).
The structure of Li2Mn2O4 is a distorted spinel, which
belongs to the I41/amd space group (see Table 3). According
to a neutron diffraction study, the added Li ions go into the
previously vacant 16c octahedral sites, and about half of the
lithium ions in the tetrahedral 8a sites move also in the 16c
position.189 The cubic-to-tetragonal transition is detected only
by the splitting of some peaks in the X-ray diffraction
patterns whereas lithium incorporation into LiMn2O4 leads
to an important change in the vibrational features. As shown
in Figure 21, the Raman spectrum of tetragonal Li2Mn2O4
is dominated by four typical lines located at 607, 398, 279,
and 258 cm-1. It should be noticed that a greater number of
bands was expected from the spectroscopic analysis, which
predicts 14 Raman-active modes for the tetragonal Li2Mn2O4
phase with D194h symmetry: 2A1g + 2B1g + 6Eg + 4 B2g (Table
3). A straightforward description of the experimental Raman
spectrum of tetragonal Li2Mn2O4 is not yet provided, even
if some tentative assignments have been speculated from a
localized vibration analysis.190
Raman microspectrometry constitutes a very efficient
probe for the identification of electrochemically produced
spinel-like lithiated manganese oxide materials. In situ
Raman measurements during the charge-discharge of a
LiMn2O4 composite cathode have been performed.36 How-
ever, only poor structural information can be drawn due to
the limited frequency range, low quality Raman spectra and
local inhomogeneity. In situ Raman investigation of pure
Li1-xMn2O4 (0 e x e 1) thin films produced by electrostatic
spray deposition did not allow a straightforward analysis,
Baddour-Hadjean and Pereira-Ramos
Figure 23. (top) Fraction of each of the components of LixMn2O4
as a function of the potential extracted from Raman data using
classical least-squares (CLS) curve resolution. (bottom) Normalized
charge determined by coulometric analysis of the voltammetric
curve (left ordinate, solid curve) and fraction of Li0.5Mn2O4 as a
function of the potential (scattered symbols). Reprinted with
permission from ref 192. Copyright 2005 The Electrochemical
Society.
since the Raman spectra are hampered by the lines of the
sapphire optical window and the electrolyte.191 On the other
hand, in situ spectroelectrochemical Raman measurements
performed by Scherson et al. on a LiMn2O4 single crystal
microelectrode during simultaneous cyclic voltammetry
experiments have afforded the spectroscopic fingerprints of
λ-MnO2, Li0.5Mn2O4, and LiMn2O4, respectively.59,192 From
the analysis of the Raman spectra using classical least-squares
(CLS) curve resolution, the authors provide the fraction of
the different LixMn2O4 components as a function of the
potential (Figure 23). The obtained results allow direct
correlations to be made between the state of charge of single
particle electrodes and their Raman spectroscopic properties,
thereby providing means of monitoring spatiotemporal effects
induced by Li intercalation-deintercalation.
One of the main routes considered to reduce the capacity
fading of LiMn2O4 is partial substitution of manganese by
transition metals M.73,193,194 As a matter of fact, LiMxMn2-xO4
(M ) Cr, Co, Ni, Al, Li, ...) materials are regarded as
attractive cathodes for lithium batteries, allowing the cell
voltage to be increased to 5 V with an improved cycle life.
The stabilization of the octahedral 16d site is achieved by
reducing the amount of Mn3+ causing the Jahn-Teller
distortion. The enthalpy of the cubic-to-tetragonal phase
transition in partially substituted LiMxMn2-xO4 spinels
gradually decreases with increasing amount of substituent
and is completely suppressed with 10-20 mol % of
substitution in the octahedral Mn site.193 The variation of
the specific capacity with cycle number for various LiM1/
6Mn11/6O4 spinels (M ) Cr, Co, Ni, Al) is reported in Figure
24. It can be seen that the cells with substituted LiMn2O4
show better cycling characteristics than the one with undoped
LiMn2O4.
Several studies have shown that the improvement in the
cycleability of the substituted LiMxMn2-xO4 was related to
the enhanced stability of the local structure of the materials
in comparison to that of LiMn2O4.195 Furthermore, when the
cations involved have indistinguishable scattering power,
vibrational spectroscopy has proved to be very powerful in
Electrode Materials for Lithium Batteries
Figure 24. Specific capacity as a function of the cycle number
for LiM1/6Mn11/6O4 (M ) Cr, Co, Ni, Al). Reprinted with permission
from ref 73. Copyright 2001 Elsevier.
Figure 25. Raman spectra of the LiNixMn2-xO4 compounds. Laser
excitation line: 514.5 nm. Reprinted with permission from ref 197.
Copyright 2006 Elsevier.
determining symmetry changes undetectable by XRD. In
spite of this, only limited Raman studies have been devoted
to this topic, presumably because of the difficulties in
interpreting the Raman spectra of such materials.
On investigating the question of cation ordering in
Li-Mn-O spinels with 1/4 substitution on the octahedral
sites, Strobel et al. reported several IR and Raman changes
which have been related to the presence of well-separated,
ordered cationic sites and symmetry lowering, depending on
the nature of the substituent.196 A Raman study of the
evolution of the local structure of spinel LiNixMn2-xO4 with
the Ni substitution (0 < x e 0.5) was performed.197 Particular
attention has been paid to the position of the A1g Raman
band, which reflects the [MO6] octahedron compactness (M
) Ni, Mn). Its value depends on the nature and the rate of
the substituent (Figure 25). The variation of the A1g Raman
wavenumber with the nickel content indicates that there exists
a critical composition level in LiNixMn2-xO4 with respect
to the local structural stability of the materials. Hence, the
most compact [MO6] octahedron with the highest bond
energy of Mn(Ni)-O was formed at the Ni substitution of x
) 0.2.
Raman investigations were conducted for LiNixMn2-xO4
(0 < x < 0.5)198 and LiCoxMn2-xO4 (0 < x < 1)199 thin films
as well as for lithium-rich Li1+δMn2-δO4 powders.200 In
particular, Uchida et al. clearly showed that in situ Raman
spectroscopy is very helpful in clarifying the valence states
Chemical Reviews, 2010, Vol. 110, No. 3 1297
Figure 26. Micro-Raman spectra for (a) monoclinic layered
LiMnO2, (b) cubic spinel LiMn2O4, (c) delithiated LiMnO2, (d)
relithiated LiMnO2, and (e) electrochemically cycled LiMnO2.
Excitation with 514.5 nm radiation. Reprinted with permission from
ref 203. Copyright 2001 The Electrochemical Society.
of the metals in LiNixMn2-xO4198 and LiCoxMn2-xO4199 thin
film electrodes during the electrochemical lithiation/delithia-
tion processes.
4.3.2.2. Layered Rock Salt Phases in the Li-Mn-O
System. Stoichiometric rock salt compounds are located on
the tie line between MnO and Li2MnO3 (Figure 19). These
layered lithiated manganese oxides are also under investiga-
tion for battery applications.156,201 Among them, R-LiMnO2
exhibits the largest initial capacity for the 4 V region, but it
suffers from severe capacity fading after the first charge
process.156 In fact, most layered compounds have structural
features in common with spinels and convert readily upon
cycling.
The structure of R-LiMnO2 is monoclinic (C2/m space
group, C32h spectroscopic symmetry, Z ) 2) because the
coordination polyhedron around the Mn3+ ions is distorted
due to the Jahn-Teller effect. This compound exhibits cation
ordering of the R-NaFeO2 structure, in which Li+ ions are
located between the MnO6 sheets, in the same octahedral
2d interstices as Mn3+ ions, whereas oxygen anions are in
4i sites. According to the group theory, the monoclinic
LiMnO2 oxide is predicted to show three Raman-active
modes with 2Ag + Bg species. As shown in Figure 26a, three
main peaks are detected in the Raman spectrum of the
monoclinic LiMnO2 phase, at 422, 481, and 603 cm-1, with
this latest peak being attributed to the Ag mode involving
the symmetric stretching vibration of equatorial oxygen
atoms, while the shoulder at 575 cm-1 is attributed to the
stretching mode of Mn ions bonded to axial oxygen atoms.190
Indeed, in the case of layered LiMnO2, a manganese ion
possesses six neighbor oxygen ions with two different Mn-O
distances, that is two Mn-Oequatorial ) 1.91 Å and four
Mn-Oaxial ) 2.32 Å.202
Therefore, Raman spectroscopy is very effective in
identifying closely related structures such as rock-salt (Figure
26a) and spinel lithium manganate (Figure 26b) phases. On
studying the local structural changes of lithium manganese
oxide upon electrochemical cycling, an irreversible Mn
migration process into the interlayer lithium site was found
from the first charge process, which results in the creation
of a spinel-like cation ordering203 (Figure 26c-e).
1298 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 27. Crystal structure in V2O5 in xy (a), in xz-projection (b), and in local vanadium environments (c). The V atoms are shown by
the small black circles, and the O2 and O3 atoms are shown by the large black circles. The O1 atoms are shown by the large open circles.
Work is underway by several groups to improve the
metastability by admetal doping or by pillaring the layers.
In fact, less rigid manganese oxide structures, and those not
based on a cubic close packed array of oxygen atoms, are
much more likely to remain phase-stable upon cycling in a
lithium cell configuration. As an example, LixMnO2 com-
pounds derived from Na0.44MnO2 exhibit remarkable stability
in polymer or in carbonate-based electrolytes, attributed to
the double tunnel structure, which cannot easily undergo
rearrangement to spinel.204,205 However, LixMnO2 exhibits
poor cycling behavior in room temperature ionic liquids,
which potentially offer a wider stability window.206 A
vibrational study was recently performed, using Raman
microscopy and FTIR spectroscopy, which showed evidence
for the formation of a thick surface film on the LixMnO2
cathode arising from surface decomposition, phase transfor-
mation of part of the material, and direct oxidation of the
electrolyte at the cathode.207 This complex surface layer,
whose thickness is found to increase during cycling, would
create electronic barriers within the composite cathode and
affect both rate capability and charge capacity.
In conclusion, the wide variety of Mn-O and Li-Mn-O
crystalline phases makes the Raman probe particularly
relevant to identify the local signature of each family of
compounds. Raman research in this field leads to two
different sets of data. Considering the MO system, it is clear
that, apart from general considerations, there is up to now
no clear and relevant trend in the literature data for a
straightforward assignment of the complex and various
vibrational features of MnO2-based compounds. In particular,
there is a lack of experimental approach devoted to the study
of either chemically lithiated samples or cathode materials
under operation. In contrast, for the more attractive spinel
system LiMn2O4 working at 4 V as well as for the layered
LiMnO2 material, detailed spectroscopic data are available
with a relevant interpretation of the Raman spectra for the
charged and discharged electrodes. Raman spectroscopy
constitutes a particularly powerful technique in this field,
since it affords in many cases a clear identification of the
phases whereas X-ray diffraction data analysis is hampered
by their high structural similarity. It remains however that
this complex system is far from being completely investi-
gated. Other spinel compositions such as Li2Mn3O7,
Li2Mn4O9, Li4Mn5O12, etc. require additional efforts, both
Baddour-Hadjean and Pereira-Ramos
to obtain reference Raman spectra and to provide consistent
assignments.
4.4. Vanadium Pentoxide V2O5
Vanadium pentoxide is an attractive material for applica-
tions in electrochromic thin film devices and as cathode in
lithium batteries, due to its capacity to accommodate up to
three lithium ions per mole of oxide, providing a high specific
capacity around 450 (mA h)/g in the voltage range 4/1.5
V.208-213 The electrochemical performances are strongly
related to the nature and the amplitude of the structural
changes induced by the lithium insertion-deinsertion pro-
cess, which has prompted a great number of studies focusing
on the structural features of V2O5 and its lithiated LixV2O5
phases.
4.4.1. V2O5 Structure
V2O5 crystallizes in the Pmmn space group, with lattice
parameters a and c of the orthorhombic cell equal to 11.50
and 4.40 Å, respectively. The point symmetry group of V2O5
is D2h. The structure of the vanadium-oxygen layers in V2O5
is presented in Figure 27. It can be seen (Figure 27c) that
the vanadium atom is located within the oxygen coordination
polyhedron VO5 and is shifted away from the plane formed
by four oxygen atoms at a normal distance of 0.47 Å. Four
types of V-O bonds, characterized by their particular bond
length, can be distinguished:
- the short and strong apical VdO1 bonds, d1 ) 1.577 Å,
- the bridge V-O3 bonds, d2 ) 1.779 Å,
- the “ladder step” V-O2 bonds (giving two equivalent
intra ladder V-O21 and V-O22 bonds), d3 ) 1.878 Å,
- the interchain V-O2 bonds (labeled as V-O2′1 in Figure
27c), d4 ) 2.017 Å.
Polarized Raman214 and IR reflection215 spectra of the V2O5
crystal have been thoroughly studied and interpreted on the
basis of phonon state calculations. More recently, a qualita-
tive characterization of the normal vibrations of the V2O5
lattice has also been performed in terms of atomic displace-
ment.216 With all valence V-O bonds being oriented along
coordinate axes (Figure 27), any bond-stretching mode
involves oscillations of particular oxygen atoms along
particular Cartesians axes. Moreover, due to the difference
in bond lengths, the spectral lines characteristic for the bond-
Electrode Materials for Lithium Batteries Chemical Reviews, 2010, Vol. 110, No. 3 1299

Table 5. Symmetry and Frequency Distribution of Normal

Vibrations of the V2O5 Lattice with Their Assignment to
Particular Atomic Displacementsa
atomic displacement assignment Ag B2g
Z(O1) ν(d1) 994 976b
X(O3) ν(d2) 848b
X(O2′) ν(d4) 526 502b
Z(O3) δ(V-O3-V) 480
X(O1) F(VdO1) 403 350b
Z(O2) 302 310b
X(V) δ(O2-V-O2) 195 195
Z(V) δ(O3-V-O2) 104 143b

atomic displacement assignment B1g B3g

Y(O2) ν(d3) 700 700
Y(O1) F(VdO1) 282 282
Y(O3) δ(O2-V-O2) 220b
Y(V) δ(O3-V-O2) 144 144
a
Reprinted with permission from ref 216. Copyright 2008 American
Chemical Society. b Not observed experimentally.

Figure 28. Symmetric atomic displacement combinations for the
Pmmn space group. The numbers 1, 2, 3, and 4 refer to the four
symmetry equivalent atomic positions for a given atom in the unit
cell. Reprinted with permission from ref 216. Copyright 2008
American Chemical Society.
rise to the B1g and B3g modes. Taking also into account the
infrared B1u and B3u IR-active modes related to x, z-
displacements respectively and the Au and B2u species related
to y-displacements, the vibrational species for V, O1, and
O2 atomic motions can be represented as
Γ(V) ) Γ(O1) ) Γ(O2) ) 2Ag + 2B2g + B1g +
B3g + 2B1u + 2B3u + Au + B2u

Due to the special position of the O3 atoms, the x(O3)

displacements do not contribute to Ag and B1u modes, the
z(O3) displacements do not contribute to B2g and B3u modes,
and the y(O3) displacements do not contribute to B1g and Au
modes. Hence, the vibrational species for the motion of this
atom can be expressed as
Γ(O3) ) Ag + B2g + B3g + B1u + B3u + B2u

In total, the optically vibrational modes of V2O5 are obtained

from the overall contributions of each atom after subtracting
the acoustic modes (Γacoustic ) B1u + B2u + B3u)

Γ(V2O5) ) ∑ Γ(i) - Γacoustic

Figure 29. Raman spectrum of V2O5 (in red) and the deconvolution
shown below (in black). Excitation with 532 nm radiation. Reprinted
with permission from ref 216. Copyright 2008 American Chemical
Society.
stretching oscillations of the different V-O bonds can be
easily assigned. There are four symmetry-equivalent atomic
positions per unit cell for the V, O1, and O2 atoms and only
two for the O3 atoms, because they lay in the mirror plane
perpendicular to the x axis. All these positions correspond
to atoms of the same layer, and all layers, which alternate
with each other in the z-direction, are equivalent. By using
the standard table of characters of irreducible representations
of the D2h group, 12 symmetric combinations can be built
from Cartesian displacements of four equivalent atoms, six
of which are IR-active and six others of which are Raman-
active. The Raman-active combinations are shown in Figure
28, with numbers 1, 2, 3, and 4 referring to the four
symmetry equivalent atomic positions for a given atom in
the unit cell. It is seen that the x- and z-displacements give
rise to the Ag and B2g modes while the y-displacements give
Γ(V2O5) ) [7Ag + 7B2g + 3B1g + 4B3g + 7B1u +
7B3u + 3Au + 4B2u] - [B1u + B2u + B3u]
Γ(V2O5) ) 7Ag + 7B2g + 3B1g + 4B3g + 6B1u +
6B3u + 3Au + 3B2u
It follows then that 21 modes are expected in the Raman
spectrum of V2O5.
The Raman spectrum of V2O5 is shown in Figure 29. The
frequency distribution of the normal vibrations of V2O5, with
their assignments to particular atomic displacements, is
presented in Table 5. The modes originating from the same
atomic displacement are gathered in the rows of Table 5.
The modes of different symmetry are arranged in different
columns in this table.
The bond-stretching modes cover the interval of 500-1000
cm-1. First, z-displacements of O1 atoms give rise to the
highest frequency ν(d1) mode at 994 cm-1. It corresponds
to the in-phase stretching vibration of all apical V-O1 bonds.
The Raman-active ν(d1) mode of B2g symmetry expected
1300 Chemical Reviews, 2010, Vol. 110, No. 3
around 976 cm-1 was not detected in our spectra. According
to the general theory of Raman intensities, the main
contributions to the Raman tensor for the bond-stretching
modes are determined by derivatives of bond polarizability
with respect to the bond lengths. Owing to the symmetry of
the B2g representation, half of the V-O1 bonds stretch,
whereas the other half compress. This could explain that this
mode is hardly observed in spite of the fact that it is Raman
allowed from symmetry considerations. A very weak Raman
line located at 976 cm-1 was assigned the B2g mode by
Abello et al.214
Next in the frequency scale is the ν(d2) mode, which comes
from x(O3) displacements and which corresponds to an
antiphase stretching of the V-O3 bonds forming the
V-O3-V bridges. This mode of B2g symmetry, calculated
at 848 cm-1, was also not detected experimentally.214 Its low
Raman intensity is caused by the fact that the V-O3-V
bridge is pseudocentrosymmetric (V-O3-V angle is 148°).
Then follows the ν(d3) mode (at 700 cm-1) involving the
y(O2) displacements. It corresponds to an antiphase stretching
of the V-O2 bonds. The V-O2 bonds (1.88 Å) are longer
than the V-O3 bonds (1.78 Å). Correspondingly, the
frequencies of the O2-modes are lower than that of the O3-
mode.
The Raman-active mode at around 526 cm-1 originates
from the ν(d4) stretching vibration involving x-displacements
of O2′ atoms. The Ag mode gives rise to the Raman line
observed at 526 cm-1. The B2g line at 502 cm-1 was not
detected because of its low intensity.
The angle-bending modes cover the 200-500 cm-1 range.
It is more difficult to determine the frequency distribution
for these modes because of considerable coupling. The
Raman peak at 480 cm-1 can be assigned to bending
vibrations of the V-O3-V bridge angle (Figure 27). Raman
bands in the frequency region between 400 and 200 cm-1
correspond to the modes which involve the x- and y-
displacements of O1 atoms at 403 and 282 cm-1, respectively,
and the z-displacements of O21 and O22 atoms at 302 cm-1.
These atomic displacements produce the δ(O1-V-O2) and
the δ(O1-V-O3) bending deformations. Corresponding
modes can be characterized as the F(VdO1) bond rocking
oscillations. The F(VdO1) deformation in the xz-plane, which
involves the O1 atom oscillations along x-axis, gives rise to
the Ag Raman peak at 403 cm-1. The Raman features at
around 300 cm-1 correspond to y-oscillations of O1 atoms
accompanied with z-oscillations of O2 atoms. Two peaks in
the low-frequency region are associated with the modes
involving displacements of the V atoms. The line at 195 cm-1
comes from the Ag and B2g modes with the atoms oscillating
along the x-axis, δ(O2-V-O2). The most intense Raman
line at 144 cm-1, δ(O3-V-O2), corresponds to a mixture
of signals coming from B1g and B3g. The B1g mode involves
the shear motion of the ladders, whereas the B2g consists of
rotations of the ladders along their axes. The high intensity
of this line reflects the long-range order in the plane of the
vanadium oxygen layers.
4.4.2. LixV2O5 Bulk Phases
In spite of numerous studies, the Li/V2O5 system is far
from being completely elucidated, and this is mainly due to
the complex nature of the lithiated phases involved during
the lithium insertion-deinsertion process according to the
electrochemical reaction:
Baddour-Hadjean and Pereira-Ramos
V2O5 + xe- + xLi+ a LixV2O5 0<xe3
The typical discharge-charge curve of a composite V2O5
cathode exhibits several voltage plateaus corresponding to
well-known phase transitions reported for the bulk LixV2O5
system (Figure 30). Depending on the amount of lithium (x)
intercalated in V2O5, several structural modifications have
been reported.208,211,213,217-223 The R-, ε-, and δ-LixV2O5 were
identified in the 0 < x e 1 composition range. Above 3 V,
the R-LixV2O5 phase occurs with x < 0.1 whereas the pure
ε-phase exists in the range 0.3 < x < 0.7. Then the pure
δ-phase appears with x between 0.9 and 1. The lithium
contents corresponding to the limiting composition of the
three solid solutions differ slightly from one report to another.
The crystallographic data of the LixV2O5 phases are reported
in Table 6.224
The structure of the vanadium-oxygen layers is rather
similar in pure V2O5 and in the R, ε, and δ phases of LixV2O5
(Figure 31), with however an increased puckering of the
layers revealed by the decrease in the a parameter in the ε
and δ phases. Moreover, the increase of the number of
inserted lithium ions between the layers is responsible for
the increase in the c parameter. All these phase transitions
are fully reversible in the 0 < x e 1 composition range, and
the structure of the pristine V2O5 phase is recovered upon
deintercalation. The situation becomes more problematic for
lithium contents x > 1, with the δ-phase being transformed
into a γ-one on the third voltage plateau at 2.2 V via an
irreversible reconstruction mechanism.211,213 The space group
of the γ-LiV2O5 structure is Pnma. As in the R- and δ-phases,
this structure still contains the V2O5 layers perpendicular to
the z-axis. However, the structure of the γ-phase differs from
the structure of the δ-phase markedly (Figure 31). The Li
atoms are shifted in the x-direction from their symmetric
positions in the δ-phase. This is accompanied by important
deformations of the V2O5 layers leading to irreversible
symmetry loss and bond breaking. It follows that the γ-phase
structure remains stable even after deintercalation of all Li
atoms,213 leading to a new metastable γ′ variety of V2O5.221
Figure 30. First discharge-charge curve of a composite LixV2O5
electrode showing the different phases electrochemically produced
in the 0 < x < 2 composition range. C/10 rate.
Table 6. Lattice Parameters for r-V2O5, ε-Li0.52V2O5, δ-LiV2O5,
and γ-LiV2O5 Phasesa
R-V2O5 ε-phase δ-phase γ-phase
(Pmmn) (Pmmn) (Amma) (Pnma)
a (Å) 11.51 11.38 11.24 9.69
b (Å) 3.56 3.57 3.60 3.60
c (Å) 4.37 4.52 9.91 10.67
a
Reprinted with permission from ref 224. Copyright 2006 American
Chemical Society.
Electrode Materials for Lithium Batteries
Figure 31. Schematic representation of the structures of the
different LixV2O5 phases electrochemically produced in the 0 < x
< 2 composition range. Reprinted with permission from ref 211.
Copyright 1994 Elsevier.
The intercalation of three lithium ions in V2O5 is possible
through the formation of a weakly crystallized phase, namely
ω-Li3V2O5 with a tetragonal222 or cubic223 symmetry. How-
ever, some authors225 put in doubt the Li/V2O5 phase diagram
for x > 1, reporting the appearance of a mixture of LixVO2
and Li3VO4 compounds instead of the so-called “γ-” and
“ω-phase”.
A number of structural investigations on the LixV2O5
system have been carried out using different experimental
techniques such as X-ray diffraction,219,225,226 X-ray ab-
sorption,227 NMR,228 EPR,229 and neutron and electron
diffraction.230-232 Some researchers have also used Raman
spectroscopy to characterize powder37,45,63,224,233 and thin
films216,234-239 LixV2O5 crystalline phases. The first vibrational
ex situ study on chemically prepared lithium vanadium
pentoxides carried out by Frech et al. focused on the δ-, ε-,
and γ-Li0.95V2O5 compounds.233 However, only the γ-Raman
spectrum can be safely considered, with the other spectra
probably corresponding to degradation compounds locally
produced under the laser beam. In situ experiments were
performed by the same authors on a composite γ-Li0.95V2O5
cathode in an operating rechargeable cell with the aim to
study the transformation of γ-LiV2O5 into γ′-V2O5 and
ξ-Li2V2O5.37 Four LixV2O5 samples differing from their
intercalation degree during the second discharge process were
investigated, but only poor spectral changes were detected,
with the Raman spectra recorded for the x ) 0.1, 0.93, and
1.15 compositions being nearly identical, especially in the
high frequency range. In situ Raman microspectrometry was
successfully applied to characterize a lithium battery involv-
ing pure lithium metal at the anode and V2O5 powder at the
cathode.45,63 Rey and Lassègues45 examined a functioning
Li/P(EO)20LiN(SO2CF3)2/V2O5 solid polymer electrolyte
lithium cell with in situ confocal Raman microspectrometry.
The authors provided a quantitative evaluation of salt
concentration gradients in the electrolyte and detected
polluting species (LiOH, Li2CO3, Li2O, Li3N) at the lithium
electrolyte interface. Conversely, Raman changes of the
composite V2O5 positive electrode were hampered by
fluorescence of the polymer electrolyte. A recent spectro-
electrochemical investigation was carried out on a modified
coin cell specially designed to facilitate routine in situ Raman
measurements.63 Raman spectra were nevertheless limited
to the 650-1000 cm-1 range, which prevents consistent
phase identification, especially in the absence of X-ray
diffraction data. However, Burba et al. tentatively assigned
Chemical Reviews, 2010, Vol. 110, No. 3 1301
Figure 32. Raman spectra of V2O5 (a), ε-Li0.52V2O5 (b), and
δ-LiV2O5 (c). Excitation with 532 nm radiation. Reprinted with
permission from ref 224. Copyright 2006 American Chemical
Society.
the Raman spectra observed during the first discharge at 2.5
and 2.1 V to the ε- and γ-phases, respectively.
In fact, the literature data related to the Raman features
in the Li/V2O5 system are very scarce and controversial,
essentially due to inappropriate experimental conditions and
also because the efficiency of Raman spectroscopy depends
on the availability of a relevant band assignment. The
spectroscopic data on the lithiated phases have been recently
enriched by Raman microspectrometry investigations, which
afforded reference Raman spectra for the ε-Li0.5V2O5,
δ-LiV2O5, and γ-LiV2O5 bulk phases chemically prepared224
and to follow the structural changes occurring in a LixV2O5
thin film under operation.216,237,239 We develop in the fol-
lowing the most prominent results extracted from these
studies.
Figure 32 shows the Raman spectra obtained for the
ε-Li0.5V2O5 and δ-LiV2O5 phases prepared through soft
chemistry reduction reactions224 and compared to that of
V2O5. The Raman spectrum of ε-Li0.52V2O5 exhibits several
spectroscopic changes: the intensity of the translational mode
is strongly quenched, and its wavenumber is shifted from
145 to 154 cm-1. Several modes in the 200-700 cm-1 range
are also shifted toward higher wavenumber: 196-218 cm-1,
404-420 cm-1, 480-486 cm-1, 526-535 cm-1, 697-703
cm-1. Conversely, the VdO stretching mode along the c axis
decreases in frequency from 994 to 983 cm-1. This shift in
frequency has been shown to be consistent with the
lengthening of the VdO bond from 1.58 Å for the R-phase
to 1.6 Å for the ε-phase.
The Raman spectrum of δ-LiV2O5 can also be compared
with that of V2O5 (Figure 32 and Table 7). The low intensity
of the bands in the low-frequency part of the Raman
spectrum, as well as the broadening of the bands, indicates
that the δ- phase is less ordered than the R- and ε-lattices. It
is seen also that the same single line, corresponding to the
VdO stretching mode along the c axis, dominates the high-
frequency part of the Raman spectra. In the spectrum of the
δ-phase, this line is narrow and shifted up to 1008 cm-1
with respect to 994 cm-1, typically observed for the R-phase.
This line can serve as a spectral fingerprint of the δ-phase.
Other marked distinctions of the Raman spectrum of this
phase are the presence of the peak at 630 cm-1 and the
disappearance of the peak at 480 cm-1. The Raman features
of the δ-phase reflect the particularity of this structure. As
1302 Chemical Reviews, 2010, Vol. 110, No. 3 Baddour-Hadjean and Pereira-Ramos

Table 7. Wavenumbers (in cm-1) of Raman-Active Modes

Observed for the r-, ε-, δ-, and γ-Phases of LixV2O5 Samplesa
R-V2O5 ε-Li0.52V2O5 δ-LiV2O5 γ-LiV 2O5
145vs 154s 145vw 145
196vw 218w 271vw 171, 197
283s 284m 355vw 209
304s 300m 420s 271, 282
404s 420s 435s 301, 327
480m 486w 533s 376
526s 535m 630m 404
697s 703s 690m 482
994s 983s 837vw 528
1008vs 546
647
703, 737
881
947
966
992, 997
a
vs, very strong; s, strong; m, medium; w, weak; vw, very weak.
Reprinted with permission from ref 224. Copyright 2006 American
Chemical Society.
Figure 34. Raman spectra of heat-treated δ-LiV2O5 at different
temperatures. Excitation with 532 nm radiation. Reprinted with
permission from ref 224. Copyright 2006 American Chemical
Society.

Figure 33. Crystal structures of δ-LiV2O5 (c) and γ-LiV2O5 (d)

in the xz-projection. Reprinted with permission from ref 224.
Copyright 2006 American Chemical Society.

shown in Figure 33, the O1-O1 distances in the δ-phase

(2.77 Å) are markedly shorter than those in the R- and
ε-phases (3 Å). The relatively high wavenumber of the VdO
stretching mode could reflect the steric VdO repulsion,
which becomes very strong in the δ-phase because of
pronounced layer puckering.
Raman spectroscopy has allowed the investigation of the
δ f ε f γ phase transitions during appropriate heat-
treatment. X-ray diffraction measurements224 showed that δ
f ε phase transition is initiated around 100 °C and that ε f
γ transformation proceeds from 135 °C. The δ f ε
transformation ends at 170 °C, with the ε- and γ-phases being
simultaneously present. Finally, at 250 °C, the complete
transformation of ε into γ-LiV2O5 is achieved. These results
are consistent with those recently reported using a synchro-
tron X-ray powder analysis.240
The evolution of the Raman spectrum of δ-LiV2O5 vs
temperature in the 100-250 °C range (Figure 34) shows that
the δ f ε transformation begins from 100 °C, as indicated
by the emergence of the high frequency band at 975 cm-1.
Up to 170 °C, both δ- and ε-phases coexist, with the δ-phase
growing at the expense of the ε-phase. At this temperature,
Figure 35. Raman spectrum of γ-LiV2O5. Excitation with 532 nm
radiation. Reprinted with permission from ref 224. Copyright 2006
American Chemical Society.
the nucleation of the γ-phase begins, as evidenced by the
appearance of a new feature at 945 cm-1. At 200 °C, the
Raman spectrum changes drastically, with the emergence of
intense and well resolved Raman features in the whole
wavenumber range showing the crystallization of the
γ-LiV2O5 phase. The ε f γ transformation ends at 250 °C,
as indicated by the disappearance of the 985 cm-1 band.
Raman data for the resulting γ-LiV2O5 phase are reported
in Figure 35 and Table 7. It follows that the thermal treatment
of the δ-phase at 250 °C allows us to obtain the pure γ-phase
LiV2O5, with lattice parameters a ) 9.69 Å, b ) 3.60 Å,
and c ) 10.67 Å (Table 6). These values are in good
agreement with those previously reported.219-221 The space
group of the γ-LiV2O5 crystal is Pnma (D16 2h). The xz-
projection of this structure is shown in Figure 33d. As in
the R and δ-phases, V2O5 layers are perpendicular to the
z-axis and built up of V-O2 ladders connected by V-O-V
bridges. Likewise, in the δ-phase, the layers are alternatively
Electrode Materials for Lithium Batteries
Figure 36. Raman spectra of (a) δ-LiV2O5 powder illuminated
with a laser power of 0.2 mW. (b) δ- and ε- phases observed under
a laser power of 0.6 mW. (c) ε- and γ- phases observed under a
laser power of 0.8 mW with an integration time of 20 s. (d) γ-phase
observed under a laser power of 0.8 mW with an integration time
of 90 s. (e) γ-phase observed under a laser power of 0.8 mW with
an integration time of 900 s. Excitation with 532 nm radiation.
Reprinted with permission from ref 224. Copyright 2006 American
Chemical Society.
shifted on the vector b/2 and the unit cell of this structure is
doubled in the z-direction; however, the structure of the
γ-phase differs from the structure of the δ-phase markedly.
The Li atoms are shifted in the x-direction from their
symmetric positions in the δ-phase. This is accompanied by
important deformations of the V2O5 layers. Each second
ladder undergoes a strong transformation: it rotates along
the y-axis and the O atoms of neighboring vanadyl groups
exchange their positions. These displacements are shown by
curved arrows. As a result of such a deformation, two
nonequivalent V positions (labeled as V1 and V2) appear
and the V1-O3-V2 bridges become strongly nonsymmetric.
The Raman wavenumbers observed for the γ-LiV2O5
polycrystalline powder are in good agreement with those
previously measured for a LiV2O5 single-crystal.241 Further-
more, on comparing the Raman features of γ-LiV2O5 and
V2O5 (Table 7), the following considerations can be drawn:
first, all the main spectral features observed in V2O5 can be
clearly discriminated in the spectrum of γ-LiV2O5. Second,
some bands, which have a singlet form in the Raman
spectrum of V2O5 and ε- and δ-V2O5, are clearly 2-fold split
in the spectrum of γ-LiV2O5. Third, some new spectral
features not seen in the Raman spectrum of V2O5 can be
detected in the spectrum of γ-LiV2O5 (bands at 209, 376,
546, 647, 881, and 966 cm-1). At least two factors can
account for the distinction between the spectra of these
crystals. First, the nonequivalent character of the ladders in
the lattice of γ-LiV2O5, inducing two kinds of vanadium
environments, must lead to the 2-fold splitting. Second, the
Li atom oscillations may couple with some modes of the
V2O5 lattice.
A very interesting point which has to be outlined is the
possibility for the Raman microprobe to induce in situ phase
transitions, only by increasing the power of the incident laser
beam illuminating the sample. This effect is well illustrated
in Figure 36, where the spectral changes observed are related
to the successive δ f ε f γ phase transitions. It is
noteworthy that the γ-phase formation is achieved at a laser
power less than 1 mW, as a result of the great instability of
the δ-phase. This peculiarity is at the origin of numerous
misinterpretations in the literature, with the Raman response
being strongly affected by this severe experimental artifact.
Chemical Reviews, 2010, Vol. 110, No. 3 1303
Figure 37. First discharge-charge cycles of a V2O5 thin film
(thickness 0.6 µm) at different C rates. Reprinted with permission
from ref 216. Copyright 2008 American Chemical Society.
4.4.3. LixV2O5 Crystallized Thin Films
As for bulk oxides in composite electrodes, the elec-
trochemical properties of pure thin films are strongly
dependent on the structural changes induced by the lithium
insertion reaction. Several studies carried out on poly-
crystalline films242-247 seem to show that the structural
features of the lithiated LixV2O5 phases can be significantly
different from that known for the bulk material. In situ
and ex situ XRD experiments performed on sol-gel,226,246,247
electrodeposited,244,245 and sputtered242,243,248,249 LixV2O5
films have shown the existence of an unexpected elongated
c axis for x ) 0.8 as well as the nonappearance of the
δ-phase usually observed213 in bulk samples from x ≈ 0.6/
0.7.
On the other hand, few Raman spectroscopy studies have
been carried out on polycrystalline V2O5 films.234-239 Fast
amorphization of the sputtered oxide was observed,234 with
the V2O5 lines vanishing nearly completely from the early
lithium content, whereas practically no change seemed to
take place in the PLD film in the 0 < x < 1 composition
range.235 Moreover, only two compositions were reported
by Ramana et al.,235 and the Raman spectra were simply
assigned on the basis of the phase diagram known for the
bulk material. We develop in the following recent results
obtained from ex situ and in situ Raman spectroscopic
analysis of crystallized V2O5 thin films prepared by radio
frequency magnetron reactive sputtering without any heat
treatment,216,239 with particularly attractive electrochemical
properties.248-251 A specific structural response was evidenced
for such films, which strongly contrasts with that known for
bulk samples.
The galvanostatic discharge-charge curves of a 0.6 µm
sputtered thin film (Figure 37) show two well-defined
insertion steps located at about 3.4 and 3.2 V, as expected
for the crystalline form of V2O5. The value of the specific
capacity recovered at C/15 rate, 22 µAh/cm2, corresponds
to the accommodation of 0.94 Li+ ion per mole of oxide.
This insertion reaction is reversible, as shown from the lack
of polarization and the efficiency of 100% in the charge
process. No influence of the current density on the discharge
potential and capacity is seen from C/15 up to C/5 rate, and
the specific capacity is seen to decrease by only ≈18% when
the C rate increases from C/15 to 8C. Moreover, the
polarization observed at high rate never exceeds 150 mV,
which suggests a high kinetics of lithium transport.
Figure 38 pertains to Raman spectra with different x values
in LixV2O5 thin films. Examining the 0 e x e 0.5 composi-
tion range, several comments can be made: (i) The intensity
1304 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 38. Raman spectra of a galvanostatically lithiated LixV2O5 thin film, 0 e x e 1. Excitation with 532 nm radiation. Reprinted with
permission from ref 216. Copyright 2008 American Chemical Society.
of the translational mode at low frequency is progressively
quenched, and its wavenumber is shifted from 145 to 153
cm-1 for x ) 0.5. (ii) A significant loss of intensity is
observed in the 200-700 cm-1 range. (iii) In the vanadyl
stretching frequency range, a new component at 985 cm-1
appears from the first lithium content, which continuously
increases in intensity at the expense of the 997 cm-1 band
and shifts up to 980 cm-1 for x ) 0.5 (Figure 39). (iv) Several
modes in the 400-800 cm-1 range are progressively shifted
toward higher wavenumber (Figures 38 and 40): 404 to 420
cm-1, 527 to 533 cm-1, 702 to 705 cm-1. Conversely, the
482 cm-1 band gradually shifts toward lower wavenumbers,
up to 470 cm-1 for x ) 0.33, and then completely disappears.
(v) The two lines at 282 and 303 cm-1 do not undergo any
frequency shift but progressively broaden to give rise to a
single and broad band centered at 290 cm-1 for x ) 0.5.
In the 0.5 e x e 0.94 composition range, the whole Raman
spectrum of the LixV2O5 film electrode is little affected. The
stretching mode at 980 cm-1 remains unchanged in fre-
quency, while a new line at 958 cm-1 appears from x )
Baddour-Hadjean and Pereira-Ramos
0.75. In addition, the 533 cm-1 band shifts to 527 cm-1
whereas the 705 cm-1 line gives rise to a broad and
asymmetric signal centered around 730 cm-1.
It follows that the structural variations depicted all along
the lithium insertion process from the Raman microprobe
consist in rather moderate local distortions which allow the
thin film material to accommodate up to 0.94 Li/mol of oxide
without breaking of the orthorhombic symmetry. These
structural changes consist mainly in (i) an increase of disorder
within the V2O5 layers, as suggested by the significant
decrease in intensity of the 145 cm-1 mode and the significant
loss of intensity observed in the 200-700 cm-1 range; (ii) a
weakening of the vanadyl stretching mode, illustrated by the
shift in frequency of the apical V-O1 stretching mode toward
lower frequency values; (iii) a stiffening of the lattice along
the a direction, suggested by the shift of the 404 and 527
cm-1 modes, which both come from oxygen displacements
along the a axis. These changes may reflect the puckering
of the V2O5 layers, as indicated by the observed decrease of
the a parameter from 11.51 to 11.35 Å when the lithium
Electrode Materials for Lithium Batteries Chemical Reviews, 2010, Vol. 110, No. 3 1305

Figure 39. Enlarged view in the high-frequency range of the

Raman spectra of a galvanostatically lithiated LixV2O5 thin film,
0.5 < x < 1. Excitation with 532 nm radiation. Reprinted with
permission from ref 216. Copyright 2008 American Chemical
Society.

content increases from 0 to 0.55.249 In the 0.5 e x e 0.94

composition range, the Raman spectra fit well with that
previously reported for εLi0.5V2O5 and εLiV2O5 powders.224
Moreover, the new line observed at 958 cm-1 for the highest
lithium uptakes (x g 0.75) is well consistent with the
presence of an additional kind of vanadium atom in the
epsilon lithium-rich phase. Indeed, although the lithium
positions in the ε-phase remain to be determined accurately,
a unique type of lithium site has been described for the whole
ε phase domain. It consists of elongated cuboctahedra joining
common faces,252 with the symmetry of such sites remaining
unchanged with x. It was however reported that increasing
the amount of lithium in the ε-phase leads to a lithium-rich
phase (called ε′) in which the lithium environment is
modified:219,231,252-254 for lithium amounts below x ) 0.5,
the number of unoccupied cuboctahedra exceeds the number
of occupied octahedral, whereas, above x ) 0.5, neighboring
cuboctahedra have to be occupied. This induces a change in
the interatomic Li-Li distances: the shortest Li-Li distance
decreases from 7.2 Å for x < 0.5 to 3.6 Å for x > 0.5.219
As shown in Figure 41, in situ Raman microspectrometry
experiments performed on sputtered thin films239 confirm the
data obtained from ex situ measurements. The in situ Raman
spectra of thin film electrodes collected along the galvano-
static charge process (Figure 42) show a progressive and
complete recovery of the whole characteristic Raman features
of the pure V2O5 lattice, in terms of both intensities and
frequencies. Such a finding is in good agreement with the
excellent electrochemical reversibility evidenced in the 3.8/
2.8 V potential range.248-251
One of the major findings of this Raman investigation is
that the expected lithiated δ phase, which exhibits a strong
line at 1008 cm-1 for x > 0.5,224 is never identified in the
V2O5 thin film electrode. Instead of that, in the 0.5 e x e
0.94 composition range, both the Raman and XRD experi-
ments provide evidence for the existence of a solid solution
behavior, with the interlayer c parameter of the ε-phase
varying to an unusually large extent, up to 4.68 Å. This lattice
parameter value corresponds to that known for the chemically
Figure 40. Evolution of the Raman band wavenumbers at 403,
526, 700, and 984 cm-1 as a function of x in the LixV2O5 thin film,
0 < x < 1. Reprinted with permission from ref 216. Copyright 2008
American Chemical Society.
formed εLiV2O5.220 This result strongly contrasts with the
more pronounced structural changes described for the
conventional composite V2O5 cathode,213 with the emergence
of one- and two-phase domains as well as the formation of
the δ phase from x ≈ 0.6/0.7. The lower magnitude of the
structural changes evidenced for the thin film electrode can
be ascribed to the nanosize effect of V2O5 planes with a cross
section of 40 nm × 200 nm stacked perpendicular to the
substrate. Hence, a more homogeneous Li insertion reaction
takes place in the present films, avoiding the local over-
lithiation phenomenon. This favors a better relaxation of the
host lattice and a lower structural stress.
4.5. Titanium Oxide-Based Compounds
In the framework of the research for new electrode
materials for lithium batteries, negative electrode materials
have to be developed working at a higher potential than that
of the currently used carbon or graphite electrodes (∼1 V
vs Li), which are to be replaced for safety reasons.255
Titanium oxides with the Ti4+/Ti3+ redox couple working at
approximatively 1.5 V vs Li/Li+ fulfill this requirement. We
will focus on lithium titanate Li4Ti5O12 and anatase TiO2.
1306 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 41. Detailed view of the in situ Raman spectra collected during the first discharge of a LixV2O5 thin film electrode. 0 e x e 0.94.
Excitation with 532 nm radiation. *LiClO4/ACN bands.
4.5.1. Lithium Titanate Li4Ti5O12
Lithium titanates, which are known in the spinel and the
ramsdellite form, are interesting candidates as negative
electrodes in Li ion batteries. Presently, they are being
evaluated for use in the next generation of secondary lithium
cells. The most widely considered metal oxide for low
potential applications is spinel Li4/3Ti5/3O4 (Li4Ti5O12),255-263
reported as a zero strain insertion compound.255-258 This
material shows an extremely flat voltage plateau at 1.55 V
(Figure 43), allowing the reversible insertion of 0.8 to 1 Li
per Li4/3Ti5/3O4 formula unit, leading to a theoretical specific
capacity of 175 (mA h)/g, with an exceptional cycling
stability.257 The demonstration of the use of Li4Ti5O12 as a
negative electrode in a high-power lithium battery with safety
has been done. Excellent cycle life was achieved for
Li4Ti5O12//LiCoO2 cells.262
The constant working potential of this material suggests
a two-phase reaction between the defect spinel Li4/3Ti5/3O4
and the fully lithiated Li7/3Ti5/3O4 according to the general
equation proposed by Ohzuku et al.:258
(Li)8a(Li1/3,Ti5/3)16dO4 + e- + Li+ a
(Li2)16c(Li1/3,Ti5/3)16dO4
Baddour-Hadjean and Pereira-Ramos
Both lithium titanium oxides crystallize in the Fd3jm space
group (Figure 44). In the structure of the spinel Li4/3Ti5/3O4,
the octahedral 16d sites are randomly occupied by Li+ and Ti4+
whereas the tetrahedral 8a positions are occupied by Li+.
Insertion of additional lithium transforms the spinel
Li4/3Ti5/3O4 into a structure of the NaCl type, Li7/3Ti5/3O4, by
incorporation of the inserted lithium into octahedral 16c
positions. Simultaneously, the tetrahedral Li+ ions of
Li4/3Ti5/3O4 are shifted to 16c positions. Hence, in the rock-
salt Li7/3Ti5/3O4 phase, the octahedral 16c positions are
occupied by lithium, and the occupancy of the octahedral
16d positions is the same as in Li4/3Ti5/3O4, but in contrast,
the tetrahedral 8a positions are not occupied.
The difference in the unit cell volume of Li4/3Ti5/3O4 and
lithium-rich Li7/3Ti5/3O4 is extremely small. For instance, on
studying the structural changes as lithium accommodation
proceeds in Li4/3+xTi5/3O4 using XRD, Bach et al.260 reported
only a negligible modification in the cubic cell parameter,
from 8.344 Å for x ) 0 to 8.357 Å for x ) 0.9. These results
have been confirmed by Scharner et al.,263 which show the
ordered rock-salt structure of lithiated Li7/3Ti5/3O4 using high
angle X-rays scans.
Raman data on lithium titanium oxide spinels are very
scarce.264-268 The vibrational analysis of Li4Ti5O12 can be
Electrode Materials for Lithium Batteries
Figure 42. In situ Raman spectra series of the first discharge-charge process of a LixV2O5 thin film electrode. 0 e x e 0.94. Excitation
with 532 nm radiation. *LiClO4/ACN bands.
Figure 43. Typical charge and discharge curve of a Li4/3Ti5/3O4
pellet. C rate: 2.1 mA/g. Reprinted with permission from ref 263.
Copyright 1999 The Electrochemical Society.
done using the factor-group theoretical treatment of normal
A[B2]O4 spinel-type compounds (O7h), which predicts five
optic mode Raman-active vibrations A1g + Eg + 3F2g (see
Table 3). While studying the superconducting Li1+xTi2-xO4
Chemical Reviews, 2010, Vol. 110, No. 3 1307
system,266 Liu et al. reported five phonon bands for poly-
crystalline Li4/3Ti5/3O4, with slightly different frequencies
compared to those of Leonidov et al.,265 that is three main
lines centered at 239, 427, and 675 cm-1 and weak features
at 274 and 367 cm-1 (Figure 45).
Proskuryakova et al.264 and Leonidov et al.265 analyzed
the Raman spectrum of the Li4Ti5O12 spinel. Six Raman
bands were observed at room temperature, at 160, 235, 335,
430, 675, and 740 cm-1, with the most intense lines being
located at 235, 430, and 675 cm-1. The authors propose a
qualitative band assignment in the following terms: the higher
frequency bands at 675 and 740 cm-1 were assigned to
vibrations of Ti-O bonds in TiO6 octahedra whereas the
440 and 335 cm-1 Raman lines were assigned to the
stretching vibrations of the Li-O bonds in LiO4 and LiO6
polyhedra, respectively. The Raman bands below 300 cm-1
were attributed to the bending vibrations of O-Ti-O bonds
(235 cm-1) and O-Li-O bonds (160 cm-1).
A Raman spectrum of a pristine Li4Ti5O12 electrode
reported by Julien et al. was found to display bands at 671,
430, 347, 271, and 232 cm-1 which were tentatively assigned
to the five expected A1g, F2g, and Eg symmetry species for
the spinel O7h spectroscopic point symmetry group.267 It is
1308 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 44. (a) Crystal structure of Li4/3Ti5/3O4 (space group Fd3jm). (b) Crystal structure of Li7/3Ti5/3O4 with the space group Fd3jm. Reprinted
with permission from ref 263. Copyright 1999 The Electrochemical Society.
Figure 45. Raman spectra of polycrystalline Li1+xTi2-xO4 powders
showing the end members (a) LiTi2O4 and (f) Li4/3Ti5/3O4. Excitation
with 514.5 nm radiation. Reprinted with permission from ref 266.
Copyright 1994 Elsevier.
noteworthy to outline the very weak Raman intensity of this
spectrum in the low frequency region compared to other
reported Raman data.265,266
In contrast to the Li-Mn-O system, there is practically
no Raman data related to electrochemical lithium insertion
in titanate spinels. This is probably due to the lack of
theoretical analysis based on lattice dynamics simulation,
which prevents a consistent analysis of the vibrational
features exhibited by the rock-salt lithium-rich compound.
Raman spectra have been reported267 for a Li4+xTi5O12
electrode material electrochemically reduced up to x ) 2.7
(Figure 46). The main effect of lithium insertion is the shift
of the strongest component from 675 to 642 cm-1 for the
fully lithiated material and the decrease of the Raman activity
in the low frequency region. However, the authors did not
provide any explanation for the observed evolution and
proposed instead a tetragonal structure for the fully lithiated
Li7Ti5O12 compound without structural validation. This
assertion is rather surprising because factor group theory
predicts 14 Raman active modes for the tetragonal D19 4h
spectroscopic point group. Aldon et al. reported the chemical
lithium insertion into the Li4Ti5O12 spinel.268 Using XRD,
neutron diffraction, NMR, and Raman spectroscopy, the
authors proposed a different mechanism for chemical lithium
insertion; that is, the spinel to the rock-salt transition would
not occur in the chemically inserted phase.The Raman
spectrum of the chemically inserted Li5.9Ti5O12 shows an
Baddour-Hadjean and Pereira-Ramos
Figure 46. Raman spectra of a Li4+xTi5O12 electrode at different
Li uptake. Excitation with 514.5 nm radiation. Reprinted with
permission from ref 267. Copyright 2004 Elsevier.
overall broadening of the lines, changes in wavenumber
values, and new features, but these variations are not
discussed.
4.5.2. TiO2 Anatase
Titanium dioxide TiO2 is an interesting candidate for
electrode applications in photoelectrochemical solar cells269
and rechargeable lithium batteries.270 The tetragonal anatase
polymorph of titanium dioxide is of special interest, due to
its ability to store a significant amount of Li as well as its
insertion potential around 1.5 V. The first ex situ XRD
investigation of the anatase LixTiO2 system reported the
existence of a tetragonal to cubic phase transition,271 which
was not validated by further neutron272,273 and in situ powder
X-ray diffraction274 investigations, which rather invoked a
phase transformation toward a Li-rich phase, here referred
to as lithium titanate (LT), with orthorhombic symmetry.
As far is known from the literature, spectroscopic data
for the anatase bulk have been limited because of the poor
availability of single crystals.275-277 Nevertheless, two Raman
spectroscopic studies275,276 and one IR study277 made it
possible to determine all the zone-center vibrational frequen-
cies. Considering the anatase LixTiO2 system, Raman data
are limited to a few studies.278-282 The main contributions
were recently reported on composite powder electrodes.280-282
We report here recent results obtained from a Raman study
on LixTiO2 composite electrodes (0 e x e 0.6), with the
Electrode Materials for Lithium Batteries
Figure 47. Schematic structure of TiO2 anatase.
Figure 48. First discharge charge curve at the C/20 rate of a TiO2
anatase electrode as a function of the depth of discharge in 1 M
LiClO4/PC: (a) x ) 0.56; (b) x ) 0.45; (c) x ) 0.3. Reprinted with
permission from ref 280. Copyright 2004 Wiley.
pure anatase powder being synthesized via the sol-gel
process.280,281 Structural changes and reversibility will be
discussed in relation with structural data drawn from XRD
experiments. We will also illustrate through this example
the interesting contribution of lattice dynamics simulation
to perform a quantitative and thorough analysis of the Raman
features.
TiO2 anatase has a tetragonal symmetry283 (space group
D4h19 I41/amd, number 141). The unit cell parameters are a
) b ) 3.8 Å; c ) 9.61 Å. It is known that zigzag chains are
formed in the close packed planes due to anion stacking and
the half occupancy of the octahedral sites by titanium ions.
Each [TiO6] octahedron shares two adjacent edges with other
[TiO6] units in the (b, c) plane to form infinite, planar double
chains parallel to a and b (Figure 47). These double chains
share corners with identical chains above and below and are
shifted along the c axis with respect to each others. Finally,
all [TiO6] octahedra share four edges, and all oxygen ions
are bonded to three titanium ions.
The typical discharge curve of TiO2 anatase is shown in
Figure 48. At a C/20 rate, one main discharge process located
at 1.75 V is evidenced and corresponds to the insertion of
0.56 Li ion per mole of oxide according to the reaction:
TiO2 + xe- + xLi+ + S LixTiO2
The charge process occurs at a higher voltage around 1.9
V. The efficiency (Qox/Qred) of the intercalation process in
Chemical Reviews, 2010, Vol. 110, No. 3 1309
Table 8. Interatomic distance (in Å) in tetragonal anatase TiO2
and in orthorhombic lithium titanate Li0.5TiO2 at 10 K.
Subscripts eq and ax denote the oxygen coordination in the (a,b)
plane and in c direction, respectively. Reprinted from ref 280.
Copyright 2004 with permission from Wiley
anatase TiO2 (I41/amd) lithium titanate (Imma)
4Ti-Ti 3.0384 2Ti-Ti 3.1344
4Ti-Oeq 1.9367 2Ti-Ti 2.8914
2Ti-Oax 1.9792 2Ti-Oeq 1.9416
2Ti-Oeq 2.0438
1Ti-Oax 1.9639
1Ti-Oax 2.1331
2Li1-Oeq 2.0487
2Li1-Oeq 1.9301
1Li1-Oax 2.3697
1Li1-Oax 2.5991
2Li2-Oeq 2.1135
2Li2-Oeq 2.1465
1Li2-Oax 1.6717
1Li2-Oax 3.2972
2Li3-Li3 2.5473
2Li3-Li3 3.5033
the anatase is around 80%. These electrochemical data show
that the extraction of lithium is only partially quantitative
whatever the depth of discharge is.
Several authors have shown that chemical272,273 or
electrochemical271,274 lithium insertion in the anatase structure
induces from the first lithium contents a phase transformation
leading to lithiated titanate Li0.5TiO2. Ohzuku et al. reported
from XRD powder experiments a cubic symmetry for the
Li0.5TiO2 compound.271 However, recent works based on
X-ray274 and neutron272,273 diffraction found an orthorhombic
structure for lithiated titanate Li0.5TiO2, indexed with the
Imma space group.
In fact, the initial anatase structure with tetragonal sym-
metry and the new orthorhombic lithiated titanate structure
have very close diffraction patterns. The orthorhombic
lithiated titanate consists of edge-sharing TiO6 octahedra
giving rise to an open structural framework. The lithium ions
are located between the TiO6 octahedra within the octahedral
voids. The overall orthorhombic distortion of the atomic
positions in the change from anatase to lithium titanate is
small. XRD studies272,273,280 showed the following changes
in the unit cell dimensions of the anatase structure upon
lithiation: from a ) b ) 3.8 Å, c ) 9.61 Å to a ) 3.81 Å,
b ) 4.07 Å, c ) 9.04 Å. In lithium titanate, the a and b
axes become different in length by about 7%, mainly
increasing in the b direction and decreasing in the c direction
by 9%. Examination of the interatomic distances (Table 8)
reveals that the anatase structure is deformed upon lithiation.
This entails a lowering of symmetry by loss of the C4 axis
on going from the pristine TiO2 tetragonal I41/amd to the
orthorhombic Li0.5TiO2 Imma structure.
According to Cava et al.,272 the Li ions in the orthorhombic
phase are randomly located in about half of the available
interstitial octahedral 4e sites. This suggestion was also
deduced from theoretical calculations.284 Hence, only a
maximum uptake of 0.5-0.6 Li+ ions per mole of TiO2 can
be inserted in the lithiated titanate structure at room tem-
perature. This result agrees with recent quantitative NMR
findings285 and with the composition value of the electro-
chemically formed Li0.56TiO2 orthorhombic phase.280
The Raman spectra observed for different LixTiO2 (0 e x
e 0.6) compositions are shown in Figure 49. Corresponding
wavenumbers of the Raman bands are reported in Table 9.
At low Li concentration (up to x ) 0.15), the Raman
1310 Chemical Reviews, 2010, Vol. 110, No. 3 Baddour-Hadjean and Pereira-Ramos

Figure 49. Raman spectra of electrochemically lithiated LixTiO2 (x ) 0; 0.15; 0.3; 0.5; 0.56). * indicates PC/LiClO4 vibrational bands.
Excitation with 514.5 nm radiation. Reprinted with permission from ref 280. Copyright 2004 Wiley.
Table 9. Raman Wavenumbers (in cm-1) Observed for a tion, the modification of the original anatase Raman spectrum
Discharged LixTiO2 Electrodea indicates a break in the local symmetry, probably due to the
x)0 x ) 0.15 x ) 0.3 x ) 0.5 appearance of the lithiated phase Li0.5TiO2 with an orthor-
144s 144s 144s 144s hombic symmetry (Imma). This new phase exhibits several
166w 166m typical Raman bands with wavenumbers listed in Table 9.
198 198 233vw 233vw A good agreement exists with the experimental data recently
320-335b 320-335b
357 357
reported from an in situ Raman study on nanosized TiO2
398 397 397s 397s anatase powders.282 A Raman study on TiO2 anatase thin
450w 450b 450b films278 reported several broad Raman bands, located at 176,
518 518 525 525 224, 316, 405, 531, and 634 cm-1. However, the vibrations
559 559 of the propylene carbonate/LiClO4 electrolyte and the ITO
639 637 637 634
848 847 847 substrate hinder most of the Raman spectrum from being
897 894 observed. Moreover, the authors did not assign the observed
934 930 new features to the Li0.5TiO2 phase.
a
Reprinted with permission from ref 280. Copyright 2004 Wiley. s,
The tetragonal-to-orthorhombic structural transition implies
strong; m, medium; w, weak; b, broad. a lowering of symmetry of the TiO2 lattice from D4h to D2h.
As a consequence, the A1g and B1g modes of the tetragonal
anatase lattice transform into Ag modes, and Eg modes split
spectrum is very close to that of the pure titanium oxide. into B2g (x) and B3g (y) modes of the orthorhombic LT lattice.
Only a complex band structure in the high frequency region It has been determined that nine lattice modes are predicted
seems to indicate the presence of a new system. Along with for orthorhombic Li0.5TiO2:
increasing x up to x ) 0.3 and x ) 0.5, several spectroscopic
changes are observed: (i) The main bands of pure anatase
(144, 398, 518, and 639 cm-1) are clearly observed; however,
ΓRaman(LT-lattice) ) 3Ag + 3B2g + 3B3g
the band at 518 cm-1 shifts to 525 cm-1 and the band at
639 cm-1 slightly shifts to 634 cm-1. (ii) Several new bands Due to the small extent of the structural distortion, a quite
appear: a complex band structure in the high wavenumber moderate perturbation of the phonon states of the TiO2 lattice
region, a band at 559 cm-1 whose intensity increases with can be expected. However, the coupling between TiO2 lattice
x, another broad-band structure around 320-330 cm-1, and modes and vibrations of the Li atoms will cause considerable
two bands in the low wavenumber region at 166 and 233 changes in the Raman spectrum, giving rise to the emergence
cm-1. of new modes assigned to lithium vibrations and band
According to the factor group analysis,275 the 15 normal splitting due to perturbed lattice modes.281
modes of TiO2 anatase have the irreducible representation The interpretation of the complex Raman spectrum of
1A1g + 1A2u + 2B1g + 1B2u + 3Eg + 2Eu. The A1g, B1g, lithiated anatase requires the knowledge of Li positions. The
and Eg modes are Raman active, and thus, six fundamental NMR285 and quasi-elastic neutron scattering data273 suggest
transitions are expected in the Raman spectrum of anatase: that the Li ions occupy several positions within the octahedral
interstices and that dynamic hopping between them occurs.
ΓRaman ) 1A1g + 3B1g + 3Eg Neutron scattering data273 were interpreted with triply split
Li positions, namely Li1, Li2, and Li3, one of which (Li2)
The typical Raman fingerprint of anatase, shown in Figure was found considerably displaced from the center of the
49 curve a, exhibits five Raman bands at 144, 198, 398, 518, interstitial site with the shortest Li-Oax distance of 1.67 Å
and 639 cm-1. The band at 518 cm-1 splits into two (Table 8).
components at 507 cm-1 and 519 cm-1 only at 73 K.275 Lattice dynamics simulations were performed in order to
Upon increasing the Li concentration in the TiO2 host determine the frequency region of vibrations of Li atoms
lattice, the evolution of the Raman spectra clearly indicates within the TiO2 host lattice.281 Within these simulations, the
the emergence of a new phase. For low Li content (x ) 0.15), potential model of the TiO2 lattice275 was supplemented by
the local symmetry of pure anatase is conserved, as shown the Li-O, Li-Ti, and Li-Li potentials. Taking into account
by curve b in Figure 49. Upon increasing the Li concentra- the variety of possible Li atom positions, the Li-O force
Electrode Materials for Lithium Batteries
Figure 50. Multiple Li environments in LT-Li0.5TiO2, with the
different Li-O bond lengths indicated (in Å). Reprinted with
permission from ref 281. Copyright 2004 American Institute of
Physics.
constant K was represented as a function of Li-O distance
R. The dependence K(R) was retrieved from the available
references on the Li-O force constant values in different
molecules and crystals286 and justified by the quantum
mechanical simulations.281 Spectral calculations have been
carried out for each Li position, with R(Li-O) values shown
in Figure 50. Taking into account the Li-O stretching modes
along each of the three orientations, 12 Raman active modes
are expected for each of the three Li positions in lithiated
titanate Li0.5TiO2:
ΓRaman(LT) ) (3Ag + 3B2g + 3B3g)lattice + (Ag +
B2g + B3g)Li
This leads to a complex pattern of simulated Raman
spectra which enabled the assignment of all the experimental
observed features (Figure 51).
The Raman spectrum and X-ray diffraction patterns of the
reoxidized material (Figure 52) are typical of anatase, which
indicates the recovery of the tetragonal framework upon
oxidation. However, some extra-Raman bands are still
observed in the high wavenumber region. This result suggests
the presence of strong Li-O bonds which are not broken in
the course of the reoxidation process and which are probably
responsible for the partial loss of rechargeability of the
material evidenced in Figure 48.
In conclusion, Raman experimental data280 on the
lithiated TiO2 anatase system, supported by lattice dynamics
simulation,281 allowed access to the following straightforward
information: (i) Li ions in the lithiated titanate phase have
multiple positions in the octahedral interstices of this
orthorhombic structure. The multiplicity of possible Li
positions manifests itself in the Raman scattering spectra as
splitting of the bands originating from Li atom oscillations
along different orientations. This leads to the occurrence of
multiple Raman bands originated from the Li atoms vibra-
Chemical Reviews, 2010, Vol. 110, No. 3 1311
Figure 51. Simulated Raman spectrum of TiO2 anatase and lithium
titanate with Li atoms in the Li2 position (b), in the Li3 position
(c), and in the Li1 position (d) in comparison with the experimental
Raman spectra of anatase (e) and Li0.5TiO2 (f). The Li modes are
labeled by bold letters in parts b-d. The TiO2 lattice modes are
labeled in italic in parts a and d. Assignment of the observed spectral
features induced by lithium intercalation is shown by dotted lines.
Reprinted with permission from ref 281. Copyright 2004 American
Institute of Physics.
Figure 52. Raman spectra of a LixTiO2 electrode, x ) 0 (TiO2),
x ) 0.5 (lithiated titanate phase), and a reoxidized electrode
Li0.04TiO2. Excitation with 514.5 nm radiation. The X-ray diffraction
pattern of Li0.04TiO2 is shown in the inset. Reprinted with permission
from ref 280. Copyright 2004 Wiley.
tions, covering a wide wavenumber range from 450 up to
950 cm-1. Simulated spectra show six Li-bands lying above
450 cm-1 and six lattice modes below 650 cm-1. (ii) Some
Li interactions take place with the TiO2 lattice, with this
perturbation leading to the splitting of the low frequency
lattice modes. (iii) Particular Li positions with rather short
Li-O distances were found to give rise to high wavenumber
features above 800 cm-1. (vi) Partial electrochemical re-
chargeability could be explained by the existence of several
modes in the high frequency range observed for the charged
electrode material and assigned to rather strong Li-O bonds
which are not broken in the course of the reoxidation.
1312 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 53. Discharges of Li/1 M LiPF6, EC-DMC/LiFePO4 cells
at 0.055 mA/cm2, showing the carbon coating improvement
achieved by incorporating small amounts of a polyaromatic additive
during the sol-gel synthesis. (a) Sol gel synthesis with additive,
1.15% total carbon; (b) solid state synthesis from iron acetate, 1.5%
carbon; (c) sol gel synthesis with no additive, 0.69% total carbon;
(d) sol gel synthesis with no additive, 0.4% total carbon. Reprinted
with permission from ref 103. Copyright 2003 The Electrochemical
Society.
5. Phospho-olivine LiFePO4 Compound
Over the past few years, phosphate-based materials LiM-
PO4 (M ) Fe, Mn, Co, or Ni) isostructural with olivine have
generated considerable interest as a new class of cathodes
for lithium ion rechargeable batteries.287-290 Indeed, in
contrast to LixCoO2 materials, which decompose at elevated
temperature, leading to oxygen evolution, LiMPO4 com-
pounds have a highly stable three-dimensional framework
due to strong P-O covalent bonds in (PO4)3- polyanion,
which prohibits the liberation of oxygen.287 These charac-
teristics provide an excellent safety and a high stability of
the battery under operation.
Much attention has been focused on the low cost and low
toxicity LiFePO4 compound because of its attractive theoreti-
cal capacity (170 (mA h)/g), based on the two-phase reaction
occurring at a 3.4 V vs Li+/Li working potential:287-290
LiFePO4 a FePO4 + e- + Li+
Its excellent stability during normal cycling and storage
conditions make this material particularly attractive for large
scale applications such as hybrid vehicles (HEVs) and electric
vehicles (EVs). However, the major drawbacks are the low
capacities achieved at even moderate discharge and the poor
rate capability, associated with limited electronic and/or ionic
conductivity. To address these issues, researchers have
optimized the synthesis techniques to minimize particle size,
which proved successful in achieving long-term cyclability
of nanosized LiFePO4,291 or have incorporated additives to
increase conductivity. The latter approach includes coating
olivine particles with carbon by incorporating an organic or
polymeric component with the precursors before firing,103,292
adding metal particles to the mix,293 or solid-solution
doping294 by metals having higher valence than Li+. All have
met success in improving performance. An example of
optimization is illustrated in Figure 53, where the quality of
the carbon coating, achieved by additive incorporation,
accounts mainly for the performance variation.
There have been several investigations into the structure
of LiMPO4 compounds. The majority have used Mössbauer
spectroscopy and X-ray absorption spectroscopy to focus on
Baddour-Hadjean and Pereira-Ramos
Figure 54. Schematic representation of the polyhedral structure
of LiFePO4, showing the arrangement of tetrahedral PO4 and
octahedral FeO6 entities. Reprinted with permission from ref 287.
Copyright 1997 The Electrochemical Society.
the local structure around transition metal ions.290,295,296
Tucker et al. used 7Li and 31P magic angle spinning nuclear
magnetic resonance (MAS NMR) to investigate the environ-
ment of the Li and P atoms in some LiMPO4 materials.297,298
The vibrational features of olivine-type compounds are well
documented. Early work by Paques-Ledent and Tarte focused
on identifying the atomic contributions of vibrational modes
in a series of LiMPO4 olivine compounds.299,300 Using a
computational simulation employing a normal coordinate
analysis based on the Wilson’s FG matrix method,301
Paraguassu et al.302 provided for the first time a complete
set of Raman wavenumbers for LiMPO4 (M ) Fe, Co, Ni)
with their assignments to vibrational motions and symmetry
species of the crystal group. Their predicted values are in
good agreement with the experimental Raman features
reported for single crystals.303,304
LiFePO4 is characterized by a true olivine structure (space
group D2h16 Pnma). The primitive unit cell is centrosym-
metric, with four formula units in the cell. There are two
types of octahedral sites; the divalent cations Fe2+ are
randomly located on the largest 4c site with Cs symmetry,
the monovalent cations Li+ are located on the smaller 4a
site with Ci symmetry. The pentavalent cations P5+ are
located in “isolated” tetrahedra PO4 in the 4c site with Cs
symmetry. They bridge the FeO6 chains to form an intercon-
nected three-dimensional structure (Figure 54). The arrange-
ment of the different coordinated groups was previously
provided.305
Group theory303,304 showed that the normal vibrational
modes of LiMPO4 are distributed on the irreducible repre-
sentation of the D2h point group as
Γvibr ) 11Ag + 7B1g + 11B2g + 7B3g + 10Au +
14B1u + 10B2u + 14B3u
The number of predicted active fundamentals is rather large:
36 for the Raman spectrum, but the number of observed
bands is smaller than the number of predicted active
fundamentals. It is usual, although not strictly correct, to
classify those vibrations into internal and external modes.300
Internal modes refer to vibrations occurring in the PO43-
tetrahedra and external to pseudorotations and translations
of the units. The translations include motions of the center-
of-mass PO43- and M2+. It should be mentioned that this
separation is a guide to discussion only, because the
vibrations may be coupled. Assignments of the LiMPO4
internal modes are made using the simplified Herzberg’s
Electrode Materials for Lithium Batteries Chemical Reviews, 2010, Vol. 110, No. 3 1313

Table 10. Raman Wavenumbers and Assignments for LiFePO4

notation for the vibrational modes of a free PO4 tetrahe- changes in the spectral features were observed in the first
dron.306 According to this formalism, the free PO4 group delithiation step; they are the presence of new lines with
representations decompose into the irreducible representa- increased intensities upon LixFePO4 delithiation, several band
tions of the Td symmetry group as A1 + E + F1 + 3F2. splittings, and frequency shifts (Figure 56). In particular, new
Among these, the internal vibrations in terms of symmetry Raman bands were observed in the ν1 and ν3 stretching region
species are A1(ν1) symmetric P-O stretching, E(ν2) sym- at 911, 962, 1064, 1080, and 1124 cm-1. In spite of a lack
metric O-P-O bond bending, F2(ν3) antisymmetric P-O of thorough vibrational analysis and symmetry assignments,
stretching, and F2(ν4) antisymmetric O-P-O bond bending.
The Raman wavenumbers and assignments observed for
LiFePO4 are summarized in Table 10. The Raman spectrum
of LiFePO4 is shown in Figure 55.
Raman spectroscopy was used to follow the chemical
delithiation process in LixFePO4 (0 e x e 1).307 Important

Figure 56. Raman spectra in the 800-1200 cm-1 frequency range

Figure 55. Raman spectrum of LiFePO4 microcrystals. Excitation obtained for chemically delithiated LixFePO4 samples. Excitation
with 514.5 nm radiation. Reprinted with permission from ref 302. with 532 nm radiation. Reprinted with permission from ref 307.
Copyright 2005 Wiley. Copyright 2004 The Electrochemical Society.
1314 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 57. Micro-Raman spectra for different processed LiFePO4
powders. The corresponding electrochemical discharge capacities
are (from top to bottom) as follows: 58, 77, 106, 120, 126 (mA
h)/g. Excitation with 632 nm radiation. Reprinted with permission
from ref 103. Copyright 2003 The Electrochemical Society.
Figure 58. Raman spectrum of the residual carbon of a LiFePO4
powder, with the deconvolution shown below. Excitation with 632
nm radiation. Reprinted with permission from ref 102. Copyright
2007 The Electrochemical Society.
the authors assigned these new features to the FePO4
compound. The Raman band positions were found to be
independent of the lithium content, and these results were
consistent with the two-phase process usually described.287
Conversely, a recent vibrational study reported a contrast-
ing finding which failed to support the well accepted two-
crystalline-phase mixture model for lithium extraction.308 In
this work, the main Raman bands observed for delithiated
Li0.11FePO4 were found to be in good agreement with those
previously reported.307 However, a deconvolution analysis
of the vibrational spectra provided evidence for an additional
significant broad-band contribution, in both the Raman and
FTIR spectra of Li0.11FePO4. The authors claimed that the
two-phase process for transformation of LiFePO4 into FePO4
upon lithium extraction probably includes a pathway where
a highly disordered phase is generated.
The carbon coating of the LiFePO4 matrix required for
the electrochemical application of this material hinders any
bulk Raman data on the LixFePO4 cathode material to be
obtained. This is illustrated in Figure 57, where the typical
Raman spectrum of a LiFePO4 powder exhibits only a sharp
band at 953 cm-1 and two weaker bands at 997 and 1098
cm-1 assigned to the internal modes of the (PO4)3- anion.103
These weak features are explained by the fact that Raman
spectroscopy allows examination of phenomena on a carbon
surface within a limited penetration depth (∼50-100 nm
Baddour-Hadjean and Pereira-Ramos
Figure 59. Electrochemical discharge capacity of LiFePO4 elec-
trodes in lithium cells vs structure of residual carbon. Reprinted
with permission from ref 103. Copyright 2003 The Electrochemical
Society.
with a 514.5 nm green laser beam34). In fact, two intense
broad bands located at ∼1590 and 1350 cm-1 are observed,
which are assigned to carbon vibrations on the basis of the
corresponding features in the spectrum of graphite. The
structure at 1590 cm-1 mainly corresponds to the G-line
associated with the optically allowed E2g zone center mode
of crystalline graphite. The structure at 1350 cm-1 mainly
corresponds to the D-line associated with the disorder-
allowed zone-edge mode of graphite. In so far as the
structural properties of carbons are known to be strongly
related to the shape of the G and D bands,9-13,38,47,75,95,99-101
Raman spectroscopy in the field of LiFePO4-based cathode
materials is a very convenient diagnostic tool to evaluate
the quality of the carbon coating on LiFePO4 powders and
films.102,103,309-313 This property is of utmost importance, since
these studies indicated that the structure of carbon coating
influences greatly the electrochemical performance. As a
matter of fact, the shapes of the D and G carbon bands were
found to change substantially with the pyrolysis temperature
and the nature of the precursor material. This effect is
illustrated in Figure 57, where the Raman spectra of the
residual carbon are resolved into four components located
at ∼1190, 1350, 1518, and 1590 cm-1 (Figure 58). The bands
at 1190 and 1518 cm-1 have an uncertain origin but were
assigned to short-range vibrations of sp3-coordinated carbons,
as already observed in disordered carbon blacks and diamond-
like carbons. Several authors102,103,309-313 proposed a relation-
ship between the electrochemical capacity and the integrated
Raman intensity ratios of D/G bands and sp3/sp2 coordinated
carbons. As shown in Figure 59, the lower both ratios, the
higher the performance exhibited by the LiFePO4 composite
electrode. This trend was interpreted in terms of the
increasing amount of larger graphene clusters in the very
disordered carbon structure and, consequently, an improved
electronic conductivity of the carbon deposit.
6. General Conclusion
The application of Li batteries is becoming more prevalent,
and demand for the next generation of energy storage for
transportation and renewable energy is becoming very high.
Through the examples developed in this review, we expect
to have provided background and context for applying a
powerful analytical tool to further understand battery materi-
als, making advances more likely.
Electrode Materials for Lithium Batteries
We hope to have convinced the reader of the considerable
interest there is in using Raman spectroscopy for better
understanding of the relationships between the structure and
electrochemistry of electrode materials for lithium batteries.
Indeed, the research reviewed in this work clearly demon-
strates that it is now possible to obtain a wealth of structural
information from the Raman study of various Li intercalated
host lattices, from carbon to transition metal oxides and
phosphates.
The beneficial aspects of Raman microspectrometry within
the field of rechargeable Li batteries and Li ion batteries
highlighted in this review are the following:
1. Most of the electrode materials, including carbonaceous
compounds, transition metal oxides, and phospho-olivines,
are Raman active compounds.
2. From an experimental point of view, Raman spectros-
copy is a nondestructive and noninvasive analysis technique
under appropriate experimental conditions and does not
require any specific conditioning of the sample, which makes
its use possible under various environments.
3. Over the past few years, several groups have demon-
strated the application of Raman spectroscopy as a powerful
in situ vibrational probe of electrode materials for Li-
batteries.
4. The ability to spatially and temporally explore working
lithium batteries has also been confirmed in several works,
which might provide another opportunity for researchers to
experimentally verify theoretical models used to simulate the
movement of Li+ ions inside batteries.
5. Raman spectroscopy constitutes a very pertinent local
tool to enrich the knowledge of the structure of Li intercala-
tion compounds at the scale of interatomic bonds by
providing information regarding local disorder, changes in
bond lengths, bond angles, coordination, Li dynamics, metal
oxidation state, and cation ordering. For instance, several
authors have shown that, in the case of carbon, LixTiO2
anatase and LixV2O5 systems, Raman spectroscopy has
allowed access to a dynamic picture of the structural changes
induced by the electrochemical reaction as a function of
various parameters such as Li content, grain size, electrode
morphology, or temperature. In other respects, it has also
been possible to answer specific electrochemical questions,
such as (i) the loss of electrochemical reversibility for LixTiO2
related to the existence of strong Li-O interactions and (ii)
the large irreversible capacity of carbon-based electrodes due
to a solvent/Li+ cointercalation leading to graphite exfolia-
tion. For the LixV2O5 system, Raman investigations revealed
that the structural response was strongly influenced by the
electrode morphology (bulk/thin film) and the nanosize effect.
In the case of LiFePO4, the electrochemical performance of
the carbon coated cathode has been found to be strongly
related to the amount of graphene clusters evaluated from
the Raman analysis of the carbon bands.
6. New and pertinent data allowing the identification of
the local processes that contribute to the mechanism of Li-
battery degradation on cycling or aging have been recently
obtained from the great possibilities offered by the confocal
Raman imaging technique. Through comparison of peak
positions and peak intensities, quantitative and qualitative
information concerning the composition of electrode materi-
als can be extracted, such as the spatial distribution of the
different electrode components at the surface of the electrode,
especially useful for composite electrodes, or the repartition
of the local mechanical stresses induced by the electrochemi-
Chemical Reviews, 2010, Vol. 110, No. 3 1315
cal reaction. For instance, the Raman mapping technique
allowed in several cases demonstration that the microscopic
information can be directly linked to the macroscopic
behavior. Indeed, through access to Raman images showing
a marked change in the material’s structure as well as its
surface composition and distribution after prolonged charge/
discharge cycling and/or storage, the searchers get a better
understanding of the nature of the processes that affect the
electrochemical performance of the electrode and the whole
battery.
The ultimate step for a significantly improved and wider
application of Raman microspectrometry will consist in
systematically combining a rigorous experimental approach
with appropriate lattice dynamics simulations. This is the
price to benefit from well founded fingerprints for the various
lithiated host lattices involved in the electrochemical reac-
tions. Of course, this requires challenging the scientific
community for developing reliable theoretical analysis,
allowing a comprehensive view of the local structure and
Li-induced distortions and then a relevant interpretation of
Raman bands. This should make the Raman spectroscopy
more efficient, in terms of both analytical and mechanistic
analysis, and it prompts the interest of this technique for
many researchers implicated in materials science applied to
the field of energy storage.
7. Acknowledgments
The authors would like to warmly acknowledge all the
collaborators who have participated in the experiments, data
treatments, and interpretations of some results described in
this review, in particular S. Bach, M. Smirnov, C. Navone,
and E. Rackelboom. Also, many thanks to the reviewers,
who suggested relevant corrections to improve the quality
of the whole manuscript.
8. References
(1) Delhaye, M.; Dhamelincourt, P. J. Raman Spectrosc. 1975, 87, 33.
(2) Delhaye, M.; Bridoux, M.; Wallart, F. J. Mol. Struct. 1982, 79, 51.
(3) Edward, L. L.; Wachs, I. E. J. Phys. Chem. 2007, C 11, 14410.
(4) Loridant, S.; Abello, L.; Siebert, E.; Lucazeau, G. Solid State Ionics
1995, 78, 249.
(5) Arutyunyan, N. R.; Obraztsova, E. D.; Silly, M.; Jaffrennou, P.; Attal-
Tretout, B.; Loiseau, A.; Chuvilin, A. L. Phys. Status Solidi 2006, B
243, 3316.
(6) Veirs, D. K.; Ager, J. W.; Loucks, E. T.; Rosenblatt, G. M. Appl.
Opt. 1990, 29, 4969.
(7) Ferraro, J. R.; Maroni, V. A. Appl. Spectrosc. 1990, 44, 351.
(8) Burns, G.; Dacol, F. H.; Feild, C.; Holtzberg, F. Solid State Commun.
1991, 77, 367.
(9) Lespade, P.; Marchand, A.; Couzi, M.; Cruege, F. Carbon 1984, 22,
375.
(10) Knight, D. S.; White, W. B. J. Mater. Res. 1989, 4, 385.
(11) Ghodbane, S.; Deneuville, A.; Tromson, D.; Bergonzo, P.; Bustarret,
E.; Ballutaud, D. Phys. Status Solidi 2006, A203, 2397.
(12) Haouni, A.; Mermoux, M.; Marcus, B.; Abello, L.; Lucazeau, G.
Diamond Relat. Mater. 1999, 8, 657.
(13) Ramamurti, R.; Shanov, V.; Singh, R. N.; Mamedov, S.; Boolchand,
P. J. Vac. Sci. Technol. A 2006, 24 (2), 179.
(14) Huang, Y.; Yu, P.; Charasse, M. N.; Lo, Y.; Wang, S. Appl. Phys.
Lett. 1987, 51, 192.
(15) Nakashima, S.; Yugami, H.; Fujii, A.; Hangyo, M.; Yamanaka, H.
J. Appl. Phys. 1988, 64, 3067.
(16) Yoshikawa, M.; Ishida, H.; Ishitani, A.; Koisumi, S.; Inuzuka, T.
Appl. Phys. Lett. 1991, 58, 1387.
(17) Vandenabeele, P.; Edwards, H. G. M.; Moens, L. Chem. ReV. 2007,
107 (3), 675.
(18) Coupry, C.; Brissaud, D. In Raman Microscopy, DeVelopments and
Applications; Turrel, G., Corset, J., Eds.; Elsevier Academic Press:
1996; pp 421-453.
(19) Clarke, R. J. H. C. R. Chim. 2002, 5 (1), 7.
(20) Frost, R. L. Spectrochim. Acta 2003, A59, 1195.
1316 Chemical Reviews, 2010, Vol. 110, No. 3
(21) Bellot-Gurlet, L.; Le Bourdonnec, F. X.; Poupeau, G.; Dubernet, S.
J. Raman Spectrosc. 2004, 35, 671.
(22) Reiche, I.; Pages-Camagna, S.; Lambacher, L. J. Raman Spectrosc.
2004, 35, 719.
(23) Whitthingham, M. S. Chem. ReV. 2004, 104, 4271.
(24) Turrell, G. In Raman Microscopy, DeVelopments and Applications;
Turrel, G.; Corset, J. Elsevier Academic Press: 1996; pp 1-25.
(25) Cotton, F. A. Chemical Applications of Group Theory; Wiley
Interscience: New York, NY, 1963.
(26) Delhaye, M.; Barbillat, J.; Aubard, J.; Bridoux, M.; Da Silva, E. In
Raman Microscopy, DeVelopments and Applications; Turrel, G.,
Corset, J., Eds.; Elsevier Academic Press: 1996; pp 51-173.
(27) Minsky, M. Scanning 1988, 10, 128.
(28) Wilson, T. Confocal Microscopy; Academic Press: London, 1990.
(29) Dhamelincourt, P.; Barbillat, J.; Delhaye, M. Spectrosc. Eur. 1993,
5, 16.
(30) Turrel, G.; Delhaye, M.; Dhamelincourt, P. In Raman Microscopy,
DeVelopments and Applications; Turrel, G., Corset, J., Eds.; Elsevier
Academic Press: 1996; pp 27-49.
(31) Everall, N. Spectroscopy 2004, 19, 22.
(32) Everall, N. Appl. Spectrosc. 2000, 54, 1515.
(33) Everall, N. Appl. Spectrosc. 2000, 54, 773.
(34) Solin, S. A. In Graphite Intercalation Compounds; Zabel, I. H., Solin,
S. A., Eds.; Springer-Verlag: Berlin, 1990; p 165.
(35) Itoh, T.; Sato, H.; Nishina, T.; Matue, T.; Uchida, I. J. Power Sources
1997, 68, 333.
(36) Huang, W.; Frech, R. J. Power Sources 1999, 82, 616.
(37) Zhang, X.; Frech, R. J. Electrochem. Soc. 1998, 145, 847.
(38) Irish, D. E.; Deng, Z.; Odziemkowski, M. J. Power Sources 1995,
54, 28.
(39) Inaba, M.; Yoshida, H.; Ogumi, Z.; Abe, T.; Mizutani, Y.; Asano,
M. J. Electrochem. Soc. 1995, 142, 20.
(40) Inaba, M.; Yoshida, H.; Ogumi, Z. J. Electrochem. Soc. 1996, 143,
2572.
(41) Huang, W.; Frech, R. J. Electrochem. Soc. 1998, 145, 765.
(42) Novak, P.; Joho, F.; Imhof, R.; Panitz, J.-C.; Haas, O. J. Power
Sources 1999, 81-82, 212.
(43) Panitz, J. C.; Joho, F.; Novak, P. Appl. Spectrosc. 1999, 53, 1188.
(44) Rey, I.; Bruneel, J. L.; Grondin, J.; Servant, L.; Lassègues, J. C. J.
Electrochem. Soc. 1998, 145, 3034.
(45) Rey, I.; Lassègues, J. C.; Baudry, P.; Majastre, H. Electrochim. Acta
1998, 43, 1539.
(46) Novak, P.; Panitz, J.-C.; Joho, F.; Lanz, M.; Imhof, R.; Coluccia,
M. J. Power Sources 2000, 90, 52.
(47) Panitz, J. C.; Novak, P. J. Power Sources 2001, 97-98, 174.
(48) Panitz, J. C.; Novak, P.; Haas, O. Appl. Spectrosc. 2001, 55, 1131.
(49) Kerlau, M.; Marcinek, M.; Kostecki, R. J. Power Sources 2007, 174,
1046.
(50) Kerlau, M.; Marcinek, M.; Srinivasan, V.; Kostecki, R. M. Electro-
chim. Acta 2007, 52, 5422.
(51) Kostecki, R.; Lei, J.; McLarnon, F.; Shim, J.; Striebel, K. J.
Electrochem. Soc. 2006, 153, A669.
(52) Kostecki, R.; McLarnon, F. Electrochem. Solid-State Lett. 2004, 7,
A380.
(53) Striebel, K. A.; Shim, J.; Cairns, E. J.; Kostecki, R.; Lee, Y.-J.;
Reimer, J.; Richardson, T. J.; Ross, P. N.; Song, X.; Zhuang, G. V.
J. Electrochem. Soc. 2004, 151, A857.
(54) Kostecki, R.; McLarnon, F. Electrochem. Solid-State Lett. 2002, 5,
A164.
(55) Kostecki, R.; McLarnon, F. J. Power Sources 2003, 119-121, 550.
(56) Hardwick, L. J.; Marcinek, M.; Beer, L.; Kerr, J. B.; Kostecki, R. J.
Electrochem. Soc. 2008, 155, A442.
(57) (a) Totir, D. A.; Scherson, D. A. Electrochem. Solid-State Lett. 2000,
3, 263. (b) Luo, Y.; Cai, W.-B.; Scherson, D. A. J. Electrochem.
Soc. 2002, 149, A1100.
(58) Luo, Y.; Cai, W.-B.; Scherson, D. A. Electrochem. Solid-State Lett.
2001, 4, A101.
(59) Dokko, K.; Shi, Q.; Stefan, I.-C.; Scherson, D.-A. J. Phys. Chem B
2003, 107, 12549.
(60) Luo, Y.; Cai, W. B.; Xing, X. K.; Scherson, D. A. Electrochem.
Solid-State Lett. 2004, 7, 1.
(61) Lei, J.; McLarnon, F.; Kostecki, R. J. Phys. Chem. B 2005, 109, 952.
(62) Migge, S.; Sandmann, G.; Rahner, D.; Dietz, H.; Plieth, W. J. Solid
State Electrochem. 2005, 9, 132.
(63) Burba, M.; Frech, R. Appl. Spectrosc. 2006, 60, 490.
(64) Peled, E. J. Electrochem. Soc. 1979, 126, 2047.
(65) Deng, Z.; Irish, D. E. Langmuir 1994, 10, 586.
(66) Hardwick, L. J.; Holzapfel, M.; Wokaun, A.; Novak, P. J. Raman
Spectrosc. 2007, 38, 110.
(67) Hyodo, S. A.; Okabayashi, K. Electrochim. Acta 1989, 34, 1557.
(68) Howlett, P. C.; MacFarlane, D. R.; Hollenkamp, A. F. J. Power
Sources 2003, 114, 277.
(69) Aurbach, D. J. Power Sources 2000, 89, 206.
Baddour-Hadjean and Pereira-Ramos
(70) Naudin, C.; Bruneel, J. L.; Chami, M.; Desbat, B.; Grondin, J.;
Lassègues, J. C.; Servant, L. J. Power Sources 2003, 124, 518.
(71) Li, H.; Mo, Y.; Pei, N.; Xu, X.; Huang, X.; Chen, L. J. Phys. Chem.
B 2000, 104, 8477.
(72) Li, G.; Li, H.; Mo, Y.; Chen, L.; Huang, X. J. Power Sources 2002,
104, 190.
(73) Wakihara, M. Mater. Sci. Eng. 2001, R33, 109.
(74) Dahn, J. R.; Sleigh, A. K.; Shi, H.; Reimers, J. N.; Zhong, Q.; Way,
B. M. Electrochim. Acta 1993, 38, 1179.
(75) Tuinstra, F.; Koenig, J. L. J. Chem. Phys. 1976, 53, 1126.
(76) Mernagh, T. P.; Cooney, R. P.; Johnson, R. A. Carbon 1984, 22, 39.
(77) Ramsteiner, M.; Wagner, J. Appl. Phys. Lett. 1987, 51, 1355.
(78) Wang, Y.; Aolsmeyer, D. C.; McCreery, R. L. Chem. Mater. 1990,
2, 557.
(79) Matthews, M. J.; Pimenta, M. A.; Dresselhaus, G.; Dresselhaus, M. S.;
Endo, M. Phys. ReV. B 1999, 59, 6585.
(80) Kastner, J.; Pichler, T.; Kuzmany, H.; Curran, S.; Blau, W.; Weldon,
D. N.; Dlamesiere, M.; Draper, S.; Zandbergen, H. Chem. Phys. Lett.
1994, 221, 53.
(81) Pócsik, I.; Hundhausen, M.; Koós, M.; Ley, L. J. Non-Cryst. Solids
1998, 227-230, 1083.
(82) Tan, P.; Dimovski, S.; Gogotsi, Y. Philos. Trans. R. Soc. A 2004,
362, 2289.
(83) Barros, E. B.; Demir, N. S.; Souza Filho, A. G.; Mendes Filho, J.;
Jorio, A.; Dresselhaus, G.; Dresselhaus, M. S. Phys. ReV. B 2005,
71, 165422.
(84) Vidano, R. P.; Fishbach, D. B.; Willis, L. J.; Loehr, T. M. Solid
State Commun. 1981, 39, 341.
(85) Baranov, A. V.; Bekhterev, A. N.; Bobovich, Ya. S.; Petrov, V. I.
Opt. Spectrosk. 1987, 62, 612.
(86) Thomsen, C.; Reich, S. Phys. ReV. Lett. 2000, 85, 5214.
(87) Sato, K; Saito, R.; Oyama, Y.; Jiang, J.; Cançado, L. G.; Pimenta,
M. A.; Jorio, A.; Samsonidze, Ge. G.; Dresselhaus, G.; Dresselhaus,
M. S. Chem. Phys. Lett. 2006, 427, 117.
(88) Dresselhaus, M. S.; Dresselhaus, G.; Eklund, P. J.; Chung, D. D.
Mater. Sci. Eng. 1977, 31, 141.
(89) Underhill, C.; Leung, S. Y.; Dresselhaus, G.; Dresselhaus, M. S. Solid
State Commun. 1979, 29, 769.
(90) Dresselhaus, M. S.; Dresselhaus, G. AdV. Phys. 1981, 30, 139.
(91) Abe, T.; Inaba, M.; Ogumi, Z.; Yokota, Y.; Mizutani, Y. Phys. ReV.
B 2000, 61, 11344.
(92) Chan, C. T.; Ho, K. M.; Kamitakahara, W. A. Phys. ReV. B 1987,
36, 3499.
(93) Doll, G. L.; Ecklund, P.; Fischer, J. E. Phys. ReV. B 1987, 36, 4940.
(94) Maurin, G.; Bousquet, Ch.; Henn, F.; Bernier, P.; Almairac, R.;
Simon, B. Solid State Ionics 2000, 136, 1295.
(95) Spahr, M. E.; Paladino, T.; Wilhelm, H.; Würsig, A.; Goers, D.; Buqa,
H.; Holzapfel, M.; Novak, P. J. Electrochem. Soc. 2004, 151, 1383.
(96) Hardwick, L. J.; Buqa, H.; Novak, P. Solid State Ionics 2006, 177,
2801.
(97) Hardwick, L. J.; Ruch, P. W.; Hahn, M.; Scheifele, W.; Kötz, R.;
Novak, P. J. Phys. Chem. Solids 2008, 69, 1232.
(98) Hardwick, L. J.; Buqa, H.; Holzapfel, M.; Scheifele, W.; Krumeich,
F.; Novak, P. Electrochim. Acta 2007, 52, 4884.
(99) Goers, D.; Buqua, H.; Hardwick, L.; Würsig, A.; Novak, P. Ionics
2003, 9, 258.
(100) Nakajima, T. J. Fluorine Chem. 2007, 128, 277.
(101) Reynier, Y.; Yazami, R.; Fultz, B.; Barsukov, I. J. Power Sources
2007, 165, 552.
(102) Wilcox, J. D.; Doeff, M. M.; Marcinek, M.; Kostecki, R. J.
Electrochem. Soc. 2007, 154A, 389.
(103) Doeff, M. M.; Hu, Y.; McLarnon, F.; Kostecki, R. Electrochem. Solid-
State Lett. 2003, 6A, 207.
(104) Striebel, K.; Shim, J.; Sierra, A.; Yang, H.; Song, X.; Kostecki, R.;
McCarthy, K. J. Power Sources 2005, 146, 33.
(105) Kostecki, R.; Tran, T.; Song, X.; Kinoshita, K.; McLarnon, F. J.
Electrochem. Soc. 1997, 144, 3111.
(106) Fong, R.; Von Sacken, U.; Dahn, J. R. J. Electrochem. Soc. 1990,
137, 2009.
(107) Mizushima, K.; Jones, P. C.; Wiseman, P. J.; Goodenough, J. B.
Mater. Res. Bull. 1980, 15, 783.
(108) Ohzuku, T.; Ueda, A. J. Electrochem. Soc. 1994, 141, 2972.
(109) Inaba, M.; Todzuka, Y.; Yoshida, H.; Grincourt, Y.; Tasaka, A.;
Tomida, Y.; Ogumi, Z. Chem. Lett. 1995, 889.
(110) Inaba, M.; Iriyama, Y.; Ogumi, Z.; Todsuka, Y.; Tasaka, A. J. Raman
Spectrosc. 1997, 28, 613.
(111) Reimers, J. N.; Dahn, J. R. J. Electrochem. Soc. 1992, 139, 2091.
(112) Amatucci, G. G.; Tarascon, J. M.; Klein, L. C. J. Electrochem. Soc.
1996, 143, 1114.
(113) Itoh, T.; Anzue, N.; Mohamedi, M.; Hisamitsu, Y.; Umeda, M.;
Uchida, I. Electrochem. Commun. 2000, 2, 743.
(114) Kim, Y. J; Lee, E. K.; Kim, H.; Cho, J.; Cho, Y. W.; Park, B.; Oh,
S. M.; Yoon, J. K. J. Electrochem. Soc. 2004, 151, 1063.
Electrode Materials for Lithium Batteries
(115) Kang, S. G.; Kang, S. Y.; Ryu, K. S.; Chang, S. H. Solid State Ionics
1999, 120, 155.
(116) Santiago, E. I.; Andrade, A. V. C.; Paiva-Santos, C. O.; Bulhoes,
L. O. S. Solid State Ionics 2003, 158, 91.
(117) Ra, D. I.; Han, K. S. J. Power Sources 2006, 163, 284.
(118) Julien, C.; Camacho-Lopez, M. A.; Escobar-Alarcon, L.; Haro-
Poniatowski, E. Mater. Chem. Phys. 2001, 68, 210.
(119) Tang, S. B.; Lai, M. O.; Lu, L. J. Alloys Compd. 2006, 424, 342.
(120) Park, H. Y.; Lee, S. R.; Lee, Y. J.; Cho, B. W.; Cho, W. I. Mater.
Chem. Phys. 2005, 93, 70.
(121) Jeon, S. W.; Lim, J. K.; Lim, S. H.; Lee, S. M. Electrochim. Acta
2005, 51, 268.
(122) Pracharova, J.; Pridal, J.; Bludska, J.; Jakubec, I.; Vorlicek, V.;
Malkova, Z.; Dikonimos Makris, Th.; Giorgi, R.; Jastrabik, L. J.
Power Sources 2002, 108, 204.
(123) Liao, C. L.; Wu, M. T.; Chen, J. H.; Leu, I. C.; Fung, K. Z. J. Alloys
Compd. 2006, 414, 302.
(124) Lee, T.; Cho, K.; Oh, J.; Shin, D. J. Power Sources 2007, 174, 394.
(125) Koike, S.; Tatsumi, K. J. Power Sources 2007, 174, 976.
(126) Kushida, K.; Kuriyama, K. J. Cryst. Growth 2002, 237-239, 612.
(127) Lee, J. H.; Han, K. S.; Lee, B. J.; Seo, S. I.; Yoshimura, M.
Electrochim. Acta 2004, 50, 467.
(128) Rossen, E.; Reimers, J. N.; Dahn, J. R. Solid State Ionics 1993, 62,
53.
(129) Gummow, R. J.; Liles, D. C.; Tackeray, M. M.; David, W. I. F. Mater.
Res. Bull. 1993, 28, 1177.
(130) Garcia, B.; Farcy, J.; Pereira-Ramos, J. P.; Baffier, N. J. Electrochem.
Soc. 1997, 144, 1179.
(131) Huang, W.; Frech, R. Solid State Ionics 1996, 86-88, 395.
(132) Mendoza, L.; Baddour-Hadjean, R.; Cassir, M.; Pereira-Ramos, J. P.
Appl. Surf. Sci. 2004, 225, 356.
(133) Kuk, S. T.; Song, Y. S.; Suh, S. I.; Kim, J. Y.; Kim, K. J. Mater.
Chem. 2001, 11, 630.
(134) Uchida, I.; Sato, H. J. Electrochem. Soc. 1995, 142, L139.
(135) Goodenough, J. B.; Wickham, D. G.; Croft, W. J. J. Phys. Chem.
Solids 1958, 5, 107.
(136) Delmas, C.; Saadoune, I. Solid State Ionics 1992, 53-56, 370.
(137) Julien, C. Solid State Ionics 2000, 887, 136-137.
(138) Rougier, A.; Nazri, G. A.; Julien, C. Ionics 1997, 3, 170.
(139) Ramana, C. V.; Zaghib, K.; Julien, C. M. J. Power Sources 2006,
159, 1310.
(140) Shim, J.; Kostecki, R.; Richardson, T.; Song, X.; Striebel, K. A. J.
Power Sources 2002, 112, 222.
(141) Norin, L.; Kostecki, R.; McLarnon, F. Electrochem. Solid-State Lett.
2002, 5, A67.
(142) Kerlau, M.; Kostecki, R. J. Electrochem. Soc. 2006, 153, A1644.
(143) Ohzuku, T.; Makimura, Y. Chem. Lett. 2001, 642.
(144) Hwang, B. J.; Tsai, Y. W.; Carlier, D.; Ceder, G. Chem. Mater. 2003,
15, 3676.
(145) Tran, N.; Croguennec, L.; Jordy, C.; Biensan, P.; Delmas, C. Solid
State Ionics 2005, 176, 1539.
(146) Wang, Z. X.; Sun, Y. C.; Chen, L. Q.; Huang, X. J. J. Electrochem.
Soc. 2004, 151, A914.
(147) Tran, N.; Croguennec, L.; Labrugere, C.; Jordy, C.; Biensan, Ph.;
Delmas, C. J. Electrochem. Soc. 2006, 153, A251.
(148) Saavedra-Arias, J. J.; Karan, N.; Pradhan, D. K.; Kumar, A.; Nieto,
S.; Thomas, R.; Katiyar, R. S. J. Power Sources 2008, 138, 761.
(149) Le Goff, P.; Baffier, N.; Bach, S.; Pereira-Ramos, J. P. Mater. Res.
Bull. 1996, 31, 63.
(150) Thackeray, M. M. Prog. Solid State Chem. 1997, 25, 1.
(151) McBreen, J. Electrochim. Acta 1975, 20, 221.
(152) Nardi, J. C. J. Electrochem. Soc. 1985, 132, 1787.
(153) Ohzuku, T.; Kitagawa, M.; Hirai, T. J. Electrochem. Soc. 1989, 136,
3169.
(154) Hill, L. I.; Portal, R.; La Salle, A. L.; Verbaere, A.; Guyomard, D.
Electrochem. Solid-State Lett. 2001, 4, 1.
(155) Thackeray, M. M.; De Picciotto, L. A.; De Kock, A.; Johnson, P. J.;
Nicholas, V. A.; Andendorf, K. T. J. Power Sources 1987, 21, 1.
(156) Armstrong, A. R.; Bruce, P. G. Nature 1996, 381, 499.
(157) Bordet-Le Guenne, L.; Deniard, P.; Biensant, P.; Siret, C.; Brec, R.
J. Mater. Chem. 2000, 10, 2201.
(158) Julien, C. M.; Massot, M.; Poinsignon, C. Spectrochim. Acta, Part
A 2004, 60, 689.
(159) Julien, C. M. Solid State Ionics 2006, 177, 11.
(160) Strohmeier, B. R.; Hercules, D. M. J. Phys. Chem. 1984, 88, 4923.
(161) Kapteijn, F.; Van Langeveld, A. D.; Moulijn, J. A.; Andreini, A.;
Vuurman, M. A.; Turek, A. M.; Jehng, J.-M.; Wachs, I. E. J. Catal.
1994, 150, 94.
(162) Gosztola, D.; Weaver, M. J. J. Electroanal. Chem. Interfacial
Electrochem. 1989, 271, 141.
(163) Bernard, M. C.; Hugot-le Goff, A.; Vu, B. Thi; Cordoba de Torresi,
S. J. Electrochem. Soc. 1993, 140, 3065.
Chemical Reviews, 2010, Vol. 110, No. 3 1317
(164) Buciuman, F.; Patcas, F.; Craciun, R.; Zahn, D. R. T. Phys. Chem.
Chem. Phys. 1999, 1, 185.
(165) Widjaja, E.; Sampanthar, J. T. Anal. Chim. Acta 2007, 585, 241.
(166) Paolone, A.; Sacchetti, A.; Corridoni, T.; Postorino, P.; Cantelli, R.;
Rousse, G.; Masquelier, C. Solid State Ionics 2004, 170, 135.
(167) Paolone, A.; Sachetti, A.; Postorino, P.; Cantelli, R. J. Phys. Chem.
B 2005, 109, 7587.
(168) Franger, S.; Bach, S.; Farcy, J.; Pereira-Ramos, J. P.; Baffier, N. J.
Power Sources 2002, 109, 262.
(169) Julien, C.; Massot, M.; Baddour-Hadjean, R.; Franger, S.; Bach, S.;
Pereira-Ramos, J. P. Solid State Ionics 2003, 159, 345.
(170) Franger, S.; Bach, S.; Pereira-Ramos, J. P.; Baffier, N. J. Electrochem.
Soc. 2000, 147, 3226.
(171) Julien, C. M.; Massot, M. Mater. Sci. Eng. B 2003, 97, 217.
(172) Thackeray, M. M.; Johnson, P.; De Piciotto, L.; Bruce, P. G.;
Goodenough, J. B. Mater. Res. Bull. 1984, 19, 179.
(173) Ohzuku, T.; Kitagawa, M.; Hirai, T. J. Electrochem. Soc. 1990, 137,
769.
(174) Guyomard, D.; Tarascon, J. M. J. Electrochem. Soc. 1992, 139, 937.
(175) Yamada, A.; Miura, K.; Hinokuma, K.; Tanaka, M. J. Electrochem.
Soc. 1995, 142, 2149.
(176) Tarte, P. J. Inorg. Nucl. Chem. 1967, 29, 915.
(177) Tarte, P.; Preudhomme, J. Spectrochim. Acta 1970, 26A, 747.
(178) Preudhomme, J.; Tarte, P. Spectrochim. Acta 1971, 27A, 845.
(179) White, W. B.; De Angelis, B. A. Spectrochim. Acta 1967, 23A, 985.
(180) Richardson, J. T.; Wen, S. J.; Striebel, K. A.; Ross, P. N.; Cairns,
E. J. Mater. Res. Bull. 1997, 32, 609.
(181) Wen, S. J.; Richardson, J. T.; Ma, L.; Striebel, K. A.; Ross, P. N.;
Cairns, E. J. J. Electrochem. Soc. 1996, 143, L136.
(182) Kanoh, H.; Tang, W.; Ooi, K. Electrochem. Solid-State Lett. 1998,
1, 17.
(183) Ammundsen, B.; Burns, G. R.; Islam, M. S.; Kanoh, H.; Roziere, J.
J. Phys. Chem. B 1999, 103, 5175.
(184) Hwang, S.-J.; Park, D.-H.; Choy, J.-H.; Campet, G. J. Phys. Chem.
B 2004, 108, 12713.
(185) Liu, W.; Kowal, K.; Farrington, G. C. J. Electrochem. Soc. 1998,
145, 459.
(186) Mukerjee, S.; Thurston, T. R.; Jisrawi, N. M.; Yang, X. Q.; McBreen,
J.; Daroux, M. L.; Xing, X. K. J. Electrochem. Soc. 1998, 145, 466.
(187) Xia, Y.; Yoshio, M. J. Electrochem. Soc. 1996, 143, 825.
(188) Gao, Y.; Reimers, J. N.; Dahn, J. R. Phys. ReV. 1996, B54, 3878.
(189) David, W. I. F.; Thackeray, M. M.; De Picciotto, L. A.; Goodenough,
J. B. J. Solid State Chem. 1987, 67, 316.
(190) Julien, C. M.; Massot, D. Mater. Sci. Eng. B 2003, 100, 69.
(191) Anzue, N.; Itoh, T.; Mohamedi, M.; Umeda, M.; Uchida, I. Solid
State Ionics 2003, 156, 301.
(192) Shi, Q.; Takahashi, Y.; Akimoto, J.; Stefan, I. C.; Scherson, D. A.
Electrochem. Solid-State Lett. 2005, 8, A521.
(193) Song, D.; Ikuta, H.; Uchida, T.; Wakihara, M. Solid State Ionics 1999,
117, 151.
(194) Li, G. H.; Ikuta, H.; Uchida, T.; Wakihara, M. J. Electrochem. Soc.
1996, 143, 178.
(195) Kim, K. W.; Lee, S. W.; Han, K. S.; Chung, H. J.; Woo, S. I.
Electrochim. Acta 2003, 48, 4223.
(196) Strobel, P.; Ibarra-Palos, A.; Anne, M.; Poinsignon, C.; Crisci, A.
Solid State Sci. 2003, 5, 1009.
(197) Wei, Y.; Kim, K.-B.; Chen, G. Electrochim. Acta 2006, 51, 3365.
(198) Dokko, K.; Mohamedi, M.; Anzue, N.; Itoh, T.; Uchida, I. J. Mater.
Chem. 2002, 12, 3688.
(199) Dokko, K.; Anzue, N.; Mohamedi, M.; Itoh, T.; Uchida, I. Electro-
chem. Commun. 2004, 6, 384.
(200) Ramana, C. V.; Massot, M.; Julien, C. M. Surf. Interface Anal. 2005,
37, 412.
(201) Zhang, F.; Whittingham, M. S. Electrochem. Solid-State Lett. 2000,
3, 7.
(202) Park, H. S.; Hwang, S. J.; Choy, J. H. J. Phys. Chem. 2001, B105,
4860.
(203) Hwang, S. J.; Park, H. S.; Choy, J. H.; Campet, G.; Portier, J.; Kwon,
C. W.; Etourneau, J. Electrochem. Solid-State Lett. 2001, 4, A213.
(204) Doeff, M. M.; Anapolsky, A.; Edman, L.; Richardson, T. J.; De
Jonghe, L. C. J. Electrochem. Soc. 2001, 148, A230.
(205) Saint, J. A.; Doeff, M. M.; Wilcox, J. Chem. Mater. 2008, 20, 3403.
(206) Saint, J.; Best, A. S.; Hollenkamp, A. F.; Kerr, J.; Shin, J.-H.; Doeff,
M. M. J. Electrochem. Soc. 2008, 155, A127.
(207) Hardwick, L. J.; Saint, J. A.; Lucas, I. T.; Doeff, M. M.; Kostecki,
R. J. Electrochem. Soc. 2009, 156, A120.
(208) Murphy, D. W.; Christian, P. A.; Disalvo, F. J.; Waszczak, J. V.
Inorg. Chem. 1979, 18, 2800.
(209) Whittingham, M. S. J. Electrochem. Soc. 1976, 126, 315.
(210) Wiesener, K.; Schneider, W.; Ilic, D.; Steger, E.; Hallmeir, K. H.;
Brackmann, E. J. Power Sources 1978, 20, 157.
(211) Delmas, C.; Cognac-Auradou, H.; Cocciantelli, J. M.; Ménétrier, M.;
Doumerc, J. P. Solid State Ionics 1994, 69, 257.
1318 Chemical Reviews, 2010, Vol. 110, No. 3
(212) Bates, J. B.; Gruzalski, G. R.; Dudney, N. J.; Luck, C. F.; Xiaohua,
Y. Solid State Ionics 1994, 70/71, 619.
(213) Cocciantelli, J. M.; Doumerc, J. P.; Pouchard, M.; Broussely, M.;
Labat, J. J. Power Sources 1991, 34, 103.
(214) Abello, L.; Husson, E.; Repelin, E.; Lucazeau, G. Spectrochim. Acta
1983, 39A, 641.
(215) Clauws, P.; Vennik, J. Phys. Status Solidi 1980, B59, 469.
(216) Baddour-Hadjean, R.; Pereira-Ramos, J. P.; Navone, C.; Smirnov,
M. Chem. Mater. 2008, 20, 1916.
(217) Dickens, P. G.; French, S. J.; Hight, A. T.; Pye, M. F. Mater. Res.
Bull. 1979, 14, 1295.
(218) Galy, J.; Darriet, J.; Hagenmuller, P. ReV. Chim. Miner. 1971, 8, 509.
(219) Rozier, P.; Savariault, J. M.; Galy, J.; Marichal, C.; Horschinger, J.;
Granger, P. Eur. J. Solid State Inorg. Chem. 1996, 33, 1.
(220) Galy, J. J. Solid State Chem. 1992, 100, 229.
(221) Cocciantelli, J. M.; Gravereau, J. M.; Doumerc, J. P.; Pouchard, M.;
Hagenmuller, P. J. Solid State Chem. 1991, 93, 497.
(222) Leger, C.; Bach, S.; Soudan, P.; Pereira-Ramos, J. P. J. Electrochem.
Soc. 2005, 152A, 236.
(223) Delmas, C.; Brethes, S.; Ménétrier, M. J. Power Sources 1991, 34, 113.
(224) Baddour-Hadjean, R.; Rackelboom, E.; Pereira-Ramos, J. P. Chem.
Mater. 2006, 18, 3548.
(225) Rozier, P.; Savariaut, J. M.; Galy, J. Solid State Ionics 1997, 98, 133.
(226) Meulenkamp, E. A.; van Klinken, W.; Schlatmann, A. R. Solid State
Ionics 1999, 126, 235.
(227) Prouzet, E.; Cartier, C.; Vilain, F.; Tranchant, A. J. Chem. Soc.,
Faraday Trans. 1996, 92, 103.
(228) Hirschinger, J.; Mongrelet, T.; Marichal, C.; Granger, P.; Savariault,
J. M.; Déramond, E.; Galy, J. J. Phys. Chem. 1993, 97, 10301.
(229) Pecquenard, B.; Gourier, D.; Baffier, N. Solid State Ionics 1995, 78, 287.
(230) Cava, R. J.; Santoro, A.; Murphy, D. W.; Zahurak, S. M.; Fleming,
R. M.; Marsh, P.; Roth, R. S. J. Solid State Chem. 1986, 65, 63.
(231) Savariault, J. M.; Rozier, P. Physica B 1997, 234-236, 97.
(232) Katzke, H.; Czank, M.; Depmeier, W.; van Smaalen, S. Philos. Mag.
1997, B75, 757.
(233) Zhang, X.; Frech, R. Electrochim. Acta 1997, 42, 475.
(234) Cazzanelli, E.; Mariotto, G.; Passerini, S.; Decker, F. Solid State
Ionics 1994, 70-71, 412.
(235) Ramana, C. V.; Smith, R. J.; Hussain, O. M.; Massot, M.; Julien,
C. M. Surf. Interface Anal. 2005, 37, 406.
(236) Julien, C.; Ivanov, I.; Gorenstein, A. Mater. Sci. Eng. 1995, B33, 168.
(237) Baddour-Hadjean, R.; Golabkan, V.; Pereira-Ramos, J. P.; Mantoux,
A.; Lincot, D. J. Raman Spectrosc. 2002, 33, 631.
(238) McGraw, J. M.; Perkins, J. D.; Zhang, J. G.; Liu, P.; Parilla, P. A.;
Turner, J.; Schulz, D. L.; Curtis, C. J.; Ginley, D. S. Solid State Ionics
1998, 113-115, 407.
(239) Baddour-Hadjean, R.; Navone, C.; Pereira-Ramos, J. P. Electrochim.
Acta 2009, 54, 6674.
(240) Satto, C.; Sciau, P.; Dooryhee, P. E.; Galy, J.; Millet, P. J. Solid
State Chem. 1999, 146, 103.
(241) Popovic, Z. V.; Gajic, R.; Konstantinovic, M. J.; Provoost, R.;
Moshchalkov, V. V.; Vasil’ev, A. N.; Isobe, M.; Ueda, Y. Phys. ReV.
2000, B61, 11454.
(242) Talledo, A.; Granqvist, C. G. J. Appl. Phys. 1995, 77, 4655.
(243) Scarminio, J.; Talledo, A.; Andersson, A.; Passerini, S.; Decker, F.
Electrochim. Acta 1993, 38, 1637.
(244) Ptitsyn, M. V.; Tikhonov, K. I.; Rotinyan, A. L. SoV. Electrochem.
1981, 17, 1297.
(245) Andrukaitis, E.; Jacobs, P. W. M.; Lorimer, J. W. Solid State Ionics
1988, 27, 19.
(246) Meulenkamp, E. A. J. Electrochem. Soc. 1998, 145, 2759.
(247) Vivier, V.; Farcy, J.; Pereira-Ramos, J. P. Electrochim. Acta 1998,
44, 831.
(248) Navone, C.; Pereira-Ramos, J. P.; Baddour-Hadjean, R.; Salot, R. J.
Electrochem. Soc. 2006, 153A, 2287.
(249) Navone, C.; Baddour-Hadjean, R.; Pereira-Ramos, J. P.; Salot, R. J.
Electrochem. Soc. 2005, 152A, 1790.
(250) Navone, C.; Baddour-Hadjean, R.; Pereira-Ramos, J. P.; Salot, R.
Electrochim. Acta 2008, 53, 3329.
(251) Navone, C.; Pereira-Ramos, J. P.; Baddour-Hadjean, R.; Salot, R. J.
Power Sources 2005, 146, 327.
(252) Katzke, H.; Depmeier, W. Phase Transitions 1996, 59, 91.
(253) Katzke, H.; Czank, M.; Depmeier, M.; Van Smaalen, S. J. Phys.:
Condens. Matter 1997, 9, 6231.
(254) Galy, J.; Satto, C.; Sciau, P.; Millet, P. J. Solid State Chem. 1999,
146, 129.
(255) Ferg, E.; Grummov, R. J.; De Kock, A.; Thackeray, M. M. J.
Electrochem. Soc. 1994, 141, 147.
(256) Colbow, K. M.; Dahn, J. R.; Haering, R. R. J. Power Sources 1989,
26, 397.
(257) Rossen, E.; Reimers, J. N.; Dahn, J. R. Solid State Ionics 1993, 62, 53.
(258) Ohzuku, T.; Ueda, A.; Yamamoto, N. J. Electrochem. Soc. 1995,
142, 1431.
Baddour-Hadjean and Pereira-Ramos
(259) Bach, S.; Pereira-Ramos, J. P.; Baffier, N. J. Mater. Chem. 1998, 8,
251.
(260) Bach, S.; Pereira-Ramos, J. P.; Baffier, N. J. Power Sources 1999,
81-82, 273.
(261) Zaghib, K.; Simoneau, M.; Armand, M.; Gauthier, M. J. Power
Sources 1999, 81-82, 300.
(262) Jansen, A. N.; Kahaian, A. J.; Kepler, K. D.; Nelson, P. A.; Amine,
K.; Dees, D. W.; Vissers, D. R.; Thackeray, M. M. J. Power Sources
1999, 81-82, 902.
(263) Scharner, S.; Weppner, W.; Schmid-Beurmann, P. J. Electrochem.
Soc. 1999, 146, 857.
(264) Proskuryakova, E. V.; Kondratov, O. I.; Porotnikov, N. V.; Petrov,
K. I. Zh. Neorg. Khim. 1983, 28, 1402.
(265) Leonidov, I. A.; Leonidova, O. N.; Perelyaeva, L. A.; Samigullina,
R. F.; Kovyazina, S. A.; Patrakeev, M. V. Phys. Solid State 2003,
45, 2183.
(266) Liu, D. Z.; Hayes, W.; Kurmoo, M.; Dalton, M.; Chen, C. Physica
C 1994, 235, 1203.
(267) Julien, C. M.; Massot, M.; Zaghib, K. J. Power Sources 2004, 136, 72.
(268) Aldon, L.; Kubiac, P.; Womes, M.; Jumas, J. C.; Olivier-Fourcade,
J.; Tirado, J. L.; Corredor, J. I.; Pérez Vicente, C. Chem. Mater. 2004,
16, 5721.
(269) O’Regan, B.; Gratzel, M. Nature 1991, 353, 737.
(270) Bonino, F.; Busani, L.; Manstretta, M.; Rivolta, B.; Scrosati, B. J.
Power Sources 1981, 6, 261.
(271) Ohzuku, T.; Takehara, Z.; Yoshizawa, S. Electrochim. Acta 1979,
24, 219.
(272) Cava, R. J.; Murphy, D. W.; Zahurak, S.; Santoro, A.; Roth, R. S. J.
Solid State Chem. 1984, 53, 64.
(273) Wagemaker, M.; Kearley, G. J.; Van Well, A.; Mutka, H.; Mulder,
F. M. J. Am. Chem. Soc. 2003, 125, 840.
(274) De Krol, R. V.; Goossens, A.; Meulenkamp, E. J. Electrochem. Soc.
2003, 146, 3150.
(275) Ohsaka, T.; Izumi, F.; Fujiki, Y. J. Raman Spectrosc. 1978, 7, 321.
(276) Sekiya, T.; Ohta, S.; Kamei, S.; Hanakawa, M.; Kurita, S. J. Phys.
Chem. Solids 2001, 62, 717.
(277) Gonzales, R. J.; Zallen, R.; Berger, H. Phys. ReV. 1997, B55, 7014.
(278) Dinh, N. N.; Oanh, N. Th. T.; Long, P. D.; Bernard, M. C.; Hugot-
Le Goff, A. Thin Solid Films 2003, 423, 70.
(279) Lindsay, M. J.; Blackford, M. G.; Attard, D. J.; Luca, V.; Skyllas-
Kasacos, M.; Griffith, C. S. Electrochim. Acta 2007, 52, 6401.
(280) Baddour-Hadjean, R.; Bach, S.; Smirnov, M.; Pereira-Ramos, J. P.
J. Raman Spectrosc. 2004, 35, 577.
(281) Smirnov, M.; Baddour-Hadjean, R. J. Chem. Phys. 2004, 121, 2348.
(282) Hardwick, L. J.; Holzapfel, M.; Novak, P.; Dupont, L.; Baudrin, E.
Electrochim. Acta 2007, 52, 5357.
(283) Murphy, D. W.; Greenblatt, M.; Zahurak, S. M.; Cava, R. J.;
Waszczak, J. V.; Hull, G. W.; Hutton, R. S. ReV. Chim. Miner. 1982,
19, 441.
(284) Lunell, S.; Stashans, A.; Ojamae, L.; Lindstrom, H.; Hagfeldt, A.
J. Am. Chem. Soc. 1997, 119, 7374.
(285) Wagemaker, M.; van de Krol, R.; Kentgens, A. P. M.; van Well,
A. A.; Mulder, F. M. J. Am. Chem. Soc. 2001, 123, 11454.
(286) Lazarev, A. N.; Mirgorodsky, A. P.; Ignatiev, I. S. Vibrational Spectra
of Complex Oxides; Nauka: Leningrad, 1975.
(287) Padhi, A. K.; Nanjundaswamy, K. S.; Goodenough, J. B. J.
Electrochem. Soc. 1997, 144, 1188.
(288) Amine, K.; Yasuda, K.; Yamachi, M. Electrochem. Solid-State Lett.
2000, 3, 178.
(289) Okada, S.; Sawa, S.; Egashira, M.; Yamaki, J.; Tabuchi, M.;
Kageyama, H.; Konishi, T.; Yoshino, A. J. Power Sources 2001,
97-98, 432.
(290) Yamada, A.; Chung, S. C.; Hinokuma, K. J. Electrochem. Soc. 2001,
148A, 224.
(291) Prosini, P. P.; Karewska, M.; Scaccia, S.; Wisniewski, P.; Pasquali,
M. Electrochim. Acta 2003, 48, 4205.
(292) Huang, H.; Yin, S. C.; Nazar, L. Electrochem. Solid-State Lett. 2001,
4A, 170.
(293) Croce, F.; D’Epifanio, A.; Haussoun, J.; Deptula, A.; Olczac, A.;
Scrosati, B. Electrochem. Solid-State Lett. 2002, 5A, 47.
(294) Chung, S. Y.; Bloking, J. T.; Chiang, Y. M. Nat. Mater. 2002, 2,
123.
(295) Anderson, A. S.; Kalska, B.; Häggström, L.; Thomas, J. O. Solid
State Ionics 2000, 130, 41.
(296) Yamada, A.; Takei, Y.; Koizumi, H.; Sonoyama, N.; Kanno, R. Chem.
Mater. 2006, 18, 804.
(297) Tucker, M. C.; Doeff, M. M.; Richardson, T. J.; Finones, R.; Reimers,
J. A.; Cairnes, E. J. Electrochem. Solid-State Lett. 2002, 5A, 95.
(298) Tucker, M. C.; Doeff, M. M.; Richardson, T. J.; Finones, R.; Cairnes,
E. J.; Reimers, J. A. J. Am. Chem. Soc. 2002, 124, 3832.
(299) Paques-Ledent, M. T.; Tarte, P. Spectrochim. Acta 1973, 29A, 1007.
(300) Paques-Ledent, M. T.; Tarte, P. Spectrochim. Acta 1974, 30A, 673.
Electrode Materials for Lithium Batteries
(301) Wilson, E. B.; Decius, J. C.; Cross, P. C. Molecular Vibrations. The
Theory of Infrared and Raman Vibrational Spectra; Dover: New
York, 1955.
(302) Paraguassu, W.; Freire, P. T. C.; Lemos, V.; Lala, S. M.; Montoro,
L. A.; Rosolen, J. M. J. Raman Spectrosc. 2005, 36, 213.
(303) Fomin, V. I.; Gnezdilov, V. P.; Kumosov, V. S.; Peschanskii, A. V.;
Yeremenko, V. V. Low Temp. Phys. 1999, 25, 829.
(304) Fomin, V. I.; Gnezdilov, V. P.; Kumosov, V. S.; Peschanskii, A. V.;
Yeremenko, V. Low Temp. Phys. 2002, 28, 203.
(305) Geller, S.; Durand, J. L. Acta Crystallogr. 1960, 13, 325.
(306) Herzberg. G. Infrared and Raman Spectra of Polyatomic Molecules;
Van Nostrand: New York, 1975.
(307) Burba, C. M.; Frech, R. J. Electrochem. Soc. 2004, 151A, 1032.
Chemical Reviews, 2010, Vol. 110, No. 3 1319
(308) Lemos, V.; Guerini, S.; Mendes Filho, J.; Lala, S. M.; Montoro, L. A.;
Rosolen, J. M. Solid State Ionics 2006, 177, 1021.
(309) Hong, J.; Wang, C.; Dudney, N. J.; Lance, M. J. J. Electrochem.
Soc. 2007, 154A, 805.
(310) Julien, C. M.; Zaghib, K.; Mauger, A.; Massot, M.; Ait-Salah, A.;
Selmane, M.; Gendron, F. J. Appl. Phys. 2006, 100, 63511.
(311) Doeff, M. M.; Wilcox, J. D.; Kostecki, R.; Lau, G. J. Power Sources
2006, 163, 180.
(312) Doeff, M. M.; Wilcox, J. D.; Yu, R.; Aumentado, A.; Marcinek, M.;
Kostecki, R. J. Solid State Electrochem. 2008, 12, 995.
(313) Hu, Y.; Doeff, M. M.; Kostecki, R.; Finones, R. J. Electrochem.
Soc. 2004, 151, A1279.
CR800344K
1320 Chem. Rev. 2010, 110, 1320–1347

EPR Characterization and Reactivity of Surface-Localized Inorganic Radicals

and Radical Ions

Mario Chiesa,† Elio Giamello,*,† and Michel Che‡

Dipartimento di Chimica IFM and NIS, Università di Torino, 10125 Torino, Italy, and Laboratoire de Réactivité de Surface, UMR 7197-CNRS,
Université Pierre et Marie CuriesParis 6 and Institut Universitaire de France, 75005 Paris, France

Received June 9, 2008

Contents 9. Sulfur- and Chlorine-Containing Radicals 1343

10. Conclusions 1345
1. Introduction 1320 11. Acknowledgments 1345
2. Solid Surfaces and Paramagnetic Species 1322 12. Note Added after ASAP Publication 1345
3. EPR Techniques and Surface Chemistry 1323 13. References 1345
3.1. Spin Hamiltonian Formalism 1323
3.2. Single-Crystal Systems 1324 1. Introduction
3.3. Disordered Systems 1325
4. Formation of Radical Species on Solid Surfaces 1326 Radicals are chemical species containing one or more
unpaired electrons which generally react via electron pairing
4.1. Mechanisms of Radical Generation at the 1326
Surface or electron transfer mechanisms. A charged radical is called
a radical ion. Radicals and radical ions are key species in a
4.2. Chemical Bonds between Radicals and 1327
Surfaces number of important chemical and biochemical processes
ranging from atmospheric chemistry to polymerization and
4.3. Identification and Reactivity of Radical 1328
Intermediates in Heterogeneous Catalysis biochemistry.
4.4. Radical Intermediates in Heterogeneous 1329 For our purpose, radicals can be conveniently classified into
Photocatalysis two classes, organic and inorganic. While the former have been
5. Experimental Approaches in EPR of Surface 1330 much investigated as intermediates in the context of reaction
Radical Species mechanisms, inorganic radicals have attracted the attention of
5.1. Paramagnetic Probes of Surface Adsorption 1330 the scientific community since the early days of radiation
Sites chemistry, because high-energy irradiation often leads to the
5.2. Molecular Motion of Adsorbates on Solid 1331 formation of radicals. In solid materials, the latter are often
Surfaces trapped in the bulk, which makes their experimental investiga-
6. Main Families of Surface Inorganic Radicals: 1332 tion relatively easy. The book by Atkins and Symons,1
Electronic Configurations and Geometry published in 1967, represents the first thorough analysis of
6.1. Monoatomic Species 1332 inorganic radicals available in the literature at the time.
6.2. Diatomic Radical Species 1333 Attention to inorganic radicals further increased with the
6.3. Triatomic, Tetraatomic, and Pentaatomic 1333 development of the matrix isolation technique, for trapping,
Radical Species in basically inert matrixes and often at low temperatures, a
7. Carbon-Containing Inorganic Radicals 1334 large number of inorganic radicals. A later book by Weltner,
7.1. CO2- Radical Anion 1334 Magnetic Atoms and Molecules, discusses this fascinating
7.2. Radicals Generated from Carbon Monoxide 1336 domain.2
7.3. Tetraatomic and Pentaatomic 1336 Although the field of surface-stabilized radicals is certainly
Carbon-Containing Radicals more limited, in terms of the number of species isolated and
8. Nitrogen-Containing Radicals 1337 characterized, it is however far richer from the chemical point
8.1. g Tensor of 11-Electron Π Radicals 1337 of view. As a matter of fact, radicals can be stabilized by
8.2. Diatomic Species: Radical Anion of Dinitrogen 1338 the surface via specific interactions on one hand and possess
a different reactivity toward incoming adsorbates on the other
8.3. Diatomic Species: Nitric Oxide, a Multipurpose 1339
Surface Probe hand. The importance of surface radicals owes much to that
of surface phenomena which are involved in numerous fields
8.3.1. Cationic Sites 1339
such as heterogeneous catalysis, photochemistry, electro-
8.3.2. Anionic Sites 1340 chemistry and corrosion phenomena, microelectronics, op-
8.3.3. NO as a Probe of Surface Transition-Metal 1341 toelectronics, and, in general terms, nanosciences and
Ions nanotechnology. Solids structured on a nanoscopic or me-
8.4. Nitrogen-Containing Triatomic Surface Radicals 1342 soscopic scale bridge the gap between the atomic and
8.5. Nitrogen-Containing Tetraatomic and 1342 macroscopic forms of matter and often possess very special
Pentaatomic Surface Radicals chemical and physical properties caused by size and bound-
†
Università di Torino.
‡
Université Pierre et Marie CuriesParis 6 and Institut Universitaire de ary effects. These very same effects are often responsible
France. for enhanced chemical reactivity and formation of uncommon
10.1021/cr800366v  2010 American Chemical Society
Published on Web 12/14/2009
Surface-Localized Inorganic Radicals and Radical Ions
Mario Chiesa received his chemistry degree from the University of Torino
(Italy). He then moved to the University of Cardiff (U.K) with the support
of a Marie Curie fellowship and in 2001 obtained his Ph.D. in chemistry
under the guidance of Dr. Damien Murphy. Currently he is Assistant
Professor (Ricercatore) at the Department of Chemistry of the University
of Torino. His main scientific interests are concerned with the application
of electron paramagnetic resonance techniques to the study of inorganic
solids and interfacial phenomena occurring at their surfaces.
Elio Giamello, born in 1950, has been Full Professor of Inorganic Chemistry
at the University of Torino since 1999. At the same university he is
presently director of the Ph.D. School in Sciences and High Technology.
He has been active since the early 1980s in the field of surface chemistry
of metal oxides, devoting particular attention to the applications of electron
paramagnetic resonance spectroscopy to this field. The focus of his interest
is the paramagnetic centers (defects, reactive intermediates, transition-
metal centers) at the surface and their reactivity. He has three times
covered the position of Invited Professor at the Université Pierre et Marie
CuriesParis 6. In 2007 he was elected the recipient of a Humboldt
Research Award by the Alexander von Humboldt Stiftung, Germany, for
his scientific activity in the field of surface chemistry.
radical species which are also important in structured solids
with micro- and/or mesopores.3 The elegant statement made
by Wolfgang Pauli, “God created the solids, the devil their
surfaces”, nicely summarizes the problems concerned with
sample preparation, low symmetry, and lack of experimental
sensitivity which are commonly faced by researchers dealing
with the characterization of surfaces. While there are many
methods to characterize an adsorbate, there are fewer which
can give information at the molecular level on the adsorbate-
surface system, in terms of interaction and structure. Electron
paramagnetic resonance (EPR) techniques have turned out
to be most powerful to achieve this goal, providing of course
that the adsorbate-surface system is paramagnetic.
In the present review we describe and classify the variety
of inorganic radicals which are formed at oxide surfaces,
emphasizing the importance of the structure-reactivity
Chemical Reviews, 2010, Vol. 110, No. 3 1321
Michel Che was born in Lyon, France, completed his doctorate (EPR
study of titanium dioxide) in 1968 at the University of Lyon, and studied
as a postdoctoral fellow at Princeton University (1969-1971). He was
appointed Professor at the Université Pierre et Marie CuriesParis 6 in
1975 and Senior Member of the Institut Universitaire de France in 1995.
His work has led to around 400 publications in international journals. He
has been very active in the promotion and organization of catalysis, being
the President-Founder of the EFCATS (European Federation of Catalysis
Societies) with creation in 1993 of the cycle of the now famous biennial
“EuropaCat” congresses and President of the IACS (International As-
sociation of Catalysis Societies) in 2000-2004, culminating with the
organization and opening of the 13th International Congress on Catalysis
held in Paris in 2004. His research activity has been largely devoted to
catalysis processes involving gas-solid, liquid-solid, and solid-solid
interfaces. He has pioneered a molecular approach to heterogeneous
catalysis, based on transition elements taken as probes, specific isotopes,
and physical techniques, which endows him with an original position in
the field. His work has led to the emergence of interfacial coordination
chemistry at the junction of colloidal chemistry, electrochemistry, su-
pramolecular chemistry, geochemistry, and solid-state chemistry.
relationship and adsorbate-surface interaction. The surface
can be that of small crystallites or the internal surface of a
porous system accessible to molecules. Radical formation
is not uncommon during chemical processes taking place at
surfaces. Radicals are usually intermediates of complex
processes such as an electrochemical process or a catalytic
reaction. However, under particular experimental conditions
(e.g., at temperatures lower than that of the reaction) they
can be stable and directly observed. The quality of investiga-
tions on inorganic radicals has followed the development of
EPR techniques which are most suitable to detect paramag-
netic species and to unravel, often in great detail, their nature
and their geometrical and electronic structure. It is therefore
understandable that, rather often, experimental work on
inorganic radical species relies exclusively on these experi-
mental techniques.
A number of topics are not included in the present review.
First are surface organic radicals because their large number
and particular features require a specific analysis. For an
exhaustive account on this specific topic we refer the reader
to a review by Garcia and Roth4 and to that of J. K. Thomas5
illustrating radicals formed by photolysis in porous materials
and clays. Second are transition-element compounds involved
in interfacial coordination chemistry which have been the
object of some specific recent reviews.6,7 The solid-solid
interface is also not discussed, because the corresponding
radicals are essentially not reactive. Furthermore, we have
to mention the family of surface oxygen radicals. Because
of their importance in oxidation catalysis, these radicals have
been the subject of extensive reviews.8,9 We therefore avoid,
in the present paper, a systematic review of this field, limiting
ourselves to use recent examples to clarify important concepts
1322 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 1. Illustration of typical situations present at a gas-solid interface: (A) before reaction, (B) after reaction. Yellow circles identify
paramagnetic species (see the text). Notice that a paramagnetic center in (A) can be found “in” the bulk (a, d), “at” the surface when it
belongs to the solid but is localized at the surface (b, e, f), or “on” the surface when it derives from chemical addition (c, g).
concerning structure-reactivity relationships or to discuss
adsorbate-surface interactions. Finally, we have to mention
the particular case of radicals generated at a solid-liquid
interface and released in the liquid phase. In this case, the
direct detection of the radicals is often impossible and an
indirect technique called spin trapping is used10 based on
the addition of the reactive radical to a diamagnetic molecule
(the trap). The reaction is the origin of a stable paramagnetic
adduct whose EPR spectrum bears information on the starting
radical.11 Classification of this kind of spin-trapped radical
in the liquid phase is not among the purposes of the present
review.
The present review is organized as follows. In section 2,
the importance of paramagnetic species in phenomena
occurring at solid surfaces is outlined. In section 3 we give
a brief survey of EPR techniques used to characterize the
radicals. A description of the nature of surface inorganic
radicals is reported in section 4, with particular attention to
their classification, to the nature of their interaction with the
surface, and to the chemical mechanisms leading to their
formation, while the role of EPR in surface chemistry and
catalysis studies is dealt with in section 5. In section 6,
section 7, section 8, and section 9, the main families of
surface inorganic radicals are reviewed using a criterion
based on their composition and number of valence electrons.1
2. Solid Surfaces and Paramagnetic Species
In the field of surface science and heterogeneous catalysis,
EPR techniques may be used to investigate (i) either directly
a radical or (ii) indirectly a diamagnetic system via a radical,
often called a spin probe. The latter is a molecule which,
Chiesa et al.
interacting with the surface, allows one of its properties
(section 8.3) to be monitored.
In the bulk of a solid or at its surface, different types of
paramagnetic species can be present before and/or after
adsorption of gas-phase molecules (Figure 1). These can be
related to the composition of the solid or derive from
chemical reactivity in heterogeneous processes. Among the
various elementary steps which initiate a surface chemical
reaction, in fact, there are several cases involving the
formation of species with unpaired electrons.12 This occurs,
for example, in the case of electron transfer between the solid
and adsorbed molecules (oxidation catalysis, electrochemical
processes), in that of hydrogen abstraction from a molecule
(activation of alkanes), and in photocatalytic processes which
are based on photoinduced charge separation and successive
surface reaction of adsorbed molecules with either an electron
or a hole.
In a catalytic system paramagnetic species can be found,
prior to any surface interaction with incoming molecules, in
the bulk or at the surface of the solid. In both cases they are
observed by EPR, but only those located at the surface are
able to react with molecules in the gas phase or, alternatively,
to broaden their EPR lines in the presence of a physisorbed
paramagnetic gas such as molecular oxygen, thus providing
an easy method for distinguishing surface from bulk para-
magnetic species.
Far from being exhaustive and with an obvious degree of
arbitrariness, Figure 1 illustrates some of the possible cases
available at a gas-solid interface, adopting as an example
an oxide catalyst composed by Mx+ and O2- ions and
containing a variety of paramagnetic centers in interaction
Surface-Localized Inorganic Radicals and Radical Ions
Table 1. Characteristics of the Microwaves Employed in EPR Experiments
L-band S-band
frequency, ν/GHz 1 3 9.7
wavelength, λ/cm 30 10 3 1.25
energy, E/kJ mol-1 4 × 10-4 1.2 × 10-3 3.9 × 10-3
with diamagnetic molecules of carbon monoxide from the
gas phase. Prior to reaction paramagnetic centers (evidenced
in yellow) are (a) bulk transition-metal ions, (b) surface
transition-metal ions, (c) grafted transition-metal ions, (d)
bulk paramagnetic defects (for example, trapped electrons
or trapped holes), (e, f) surface-localized defect centers
(trapped electrons and holes, respectively), and (g) metal
clusters.
Reactivity involves surface centers only, which, in the
examples shown in the figure, maintain upon reaction their
paramagnetic character. Paramagnetic carbonyls (b′) form
after CO addition, while the two surface defects are
transformed into CO- (e′) and CO2- (f′) radical ions,
respectively. The subscript “(ads)” is hereafter employed in
chemical equations to indicate a species adsorbed on the
surface which does not contain atoms of the solid (CO(ads),
Figure 1b′), while the subscript “(surf)” is adopted in the
case of a species formed at the surface, containing one or
more atoms belonging to the solid (Figure 1f′, CO2-(surf)).
Although very schematic, Figure 1 conveys the idea that
there exist a large variety of cases which can be investigated
by EPR. In all these cases EPR is an instrumental tool as it
can deeply probe those features of the molecular structure
related to the unpaired electron charge distribution, thus
providing detailed answers relative to the nature, structure,
spatial distribution, and dynamics of surface paramagnetic
species. Modern surface chemistry is driven by the aim of
understanding at the molecular level the chemical processes
occurring at the surface. In this respect relevant fundamental
questions concern the structure and reactivity of surfaces by
themselves and of surfaces with atoms or molecules on them.
In the same way the discovery of the basic mechanistic steps
involved in surface reactivity and catalysis represents one
of the most critical issues. These are the very questions that
can be addressed (and answered) by EPR when paramagnetic
species are involved.
3. EPR Techniques and Surface Chemistry
EPR techniques are, beyond any doubt, the reference
techniques for detection of radicals and, in general, of
paramagnetic species.13 In the field of surface radical species
an advantage of EPR, besides its specificity, is its high
sensitivity, which allows the detection of species in concen-
trations well below a monolayer coverage. Surface species,
in fact, are often formed in very small amounts during surface
chemical processes and catalytic reactions. Review papers
devoted to the applications of EPR to surface chemistry, and
mainly oriented toward catalytic phenomena, have already
been published.14-17
Conventional CW-EPR (continuous-wave electron para-
magnetic resonance) has been until now largely dominant
in surface studies. In a CW-EPR experiment the substance
under investigation interacts with a homogeneous magnetic
field which is allowed to vary in a selected range and is
irradiated by a continuous flow of microwaves at fixed
frequency which, when resonance conditions are fulfilled,
entails a transition between two spin states. The microwave
Chemical Reviews, 2010, Vol. 110, No. 3 1323
X-band K-band Q-band W-band
24 34 94
0.88 0.32
9.5 × 10-3 1.35 × 10-2 3.74 × 10-2
frequencies are usually classified in terms of bands. Com-
mercially available bands range from 1 GHz (L-band) to 94
GHz (W-band) as shown in Table 1, which also gives,
besides the frequency, the wavelength and photon energy.
The X-band is the most commonly used, and the CW-EPR
spectra reported in the following are, unless otherwise
specified, recorded in the X-band. A CW-EPR spectrum
usually reports the first derivative of microwave absorption
as a function of the magnetic field B.
Following what was done in nuclear magnetic resonance
about 40 years ago, pulsed methods were introduced in EPR
recently. With the advent of commercial instruments, these
techniques, based on irradiation by controlled pulses of
microwaves, started to undergo rapid development and found
wider applications. Continuous-wave techniques include CW-
EPR, ENDOR (electron nuclear double resonance), and HF-
EPR (high-field electron paramagnetic resonance), while,
among the many pulsed methods,18-20 we can find FS-ESE
(field-swept electron spin echo), ESEEM (electron spin echo
envelope modulation), or FT-ESR (Fourier transform electron
spin resonance), and pulsed ENDOR. Pulsed EPR has been
applied to surface chemistry mainly in the case of systems
containing transition-metal ions.21-23
3.1. Spin Hamiltonian Formalism
The EPR spectrum of a paramagnetic species can be
described by the spin Hamiltonian, which defines the main
energy terms as follows:
H ) HEZ + HF + HHFS + HNZ + HQ (1)
The first term is the electronic Zeeman operator (HEZ )
βeS · g · B), which accounts for the interaction of the electron
spin vector S with the external magnetic field B, βe being
the Bohr magneton. The interaction is gauged by the g tensor,
a 3 × 3 matrix which, in general, can be reduced to its
diagonal form. In this form, all extradiagonal elements are
zero and the principal elements gxx, gyy, and gzz are put into
evidence, their values depending on the electronic structure
(ground and excited states) of the paramagnetic species. The
fine structure term (HF ) S · D · S) describes the interaction
between two or more unpaired electrons through the D tensor.
This term is zero in the case of S ) 1/2, i.e., for the vast
majority of surface inorganic radicals. The third term (HHFS
) S · A · I) is much more important in our case because it
represents the hyperfine interaction between the electron spin
and nuclear spins. A is the hyperfine tensor, and I is the
nuclear spin vector. In CW-EPR the hyperfine interactions
give rise to lines splitting in the spectrum (the hyperfine
structure). 2I +1 lines are expected for the interaction of
the electron spin with a nucleus having nuclear spin quantum
number I. A is composed of two main contributions, i.e.,
the isotropic Fermi contact term (a scalar arising from the
finite probability of the electron being located at the position
of the nucleus) and the anisotropic electron-nucleus dipolar
coupling expressed by matrix T. The term hyperfine interac-
tion is usually associated with the magnetic interaction
1324 Chemical Reviews, 2010, Vol. 110, No. 3 Chiesa et al.

Table 2. Relationship between the EPR Symmetry and Molecular Point Symmetry of Paramagnetic Centersa
relation between coincidence of molecular point
EPR symmetry
g and A tensors tensor axes symmetry
isotropic gxx ) gyy ) gzz all coincident Oh, Td, O, Th, T
Axx ) Ayy ) Azz
axial gxx ) gyy * gzz all coincident D4h, C4V, D4, D2d, D6h, C6V, D6, D3h, D3d, C3V, D3
Axx ) Ayy * Azz
rhombic gxx * gyy * gzz all coincident D2h, C2V, D2
Axx * Ayy * Azz
axial noncollinear gxx ) gyy * gzz only gzz and Azz coincident C3, S6, C4, S4, C4h, C6, C3h, C6h
Axx ) Ayy * Azz
monoclinic gxx * gyy * gzz one axis of g and A coincident C2h, Cs, C2
Axx * Ayy * Azz
triclinic gxx * gyy * gzz all axes noncoincident C1, Ci
Axx * Ayy * Azz
a
Adapted with permission from ref 24. Copyright 1992 Elsevier.

between the unpaired electron of a given species and the
nuclei belonging to the species itself. The same type of
interaction which involves magnetic nuclei of the surrounding
lattice is usually called the superhyperfine (shf) interaction.
Its physical origin is of course the same as the hyperfine
interaction, and the elements of the shf tensors may give
important information about the composition of the surface
adsorption site, the local symmetry, and the nature of
the chemical bond. The last two terms in eq 1 (representing
the nuclear Zeeman energy and the nuclear quadrupolar
energy, respectively) are less important than the previous
ones as they very weakly affect the EPR spectra.
The EPR tensors can be classified on the basis of their
symmetry (for instance, an isotropic tensor has three equal
principal components, and an axial tensor has two equal
components differing from the third), which in turn depends
on the point symmetry of the paramagnetic center. Radicals
and, in general, paramagnetic species located at solid surfaces
are submitted to some symmetry restrictions (some of the
point groups possible for molecular species are not permitted
at the surface7), and often radicals with very reduced
symmetry are observed. In the three most symmetric cases
(isotropic, axial, and rhombic tensors) the axes of the g and
A tensors coincide, while in the other cases they do not. The
structure of the EPR tensors and the relationships between
the EPR symmetry and point symmetry of the paramagnetic
center are gathered in Table 2.24
3.2. Single-Crystal Systems
The presence of tensors in all terms of eq 1 reflects the
anisotropy of magnetic interactions. In pragmatic terms, this
means that, when the paramagnetic system is located in a single
crystal, the EPR signal changes according to the orientation of
the crystal in the external magnetic field B. In a classic CW-
EPR spectrum (performed at constant microwave frequency and
by sweeping the magnetic field), the values of the g tensor
elements determine the position of the resonant field Bres while
those of A determine the amplitude A of the hyperfine splitting.
Both Bres and A depend on the orientations. In the following
we consider a simple system of a doublet state (S ) 1/2) with
no magnetic nuclei (I ) 0) and an axial g matrix with g⊥ > g|
(gxx ) gyy ) g⊥; gzz ) g|).
In this case, the spin Hamiltonian is limited to the HEZ )
βeS · g · B term and the effective g factor can be expressed as
g ) [g|2 cos2 ϑ + g⊥2 sin2 ϑ]1/2 (2)
where ϑ is the angle between the applied field and the
principal symmetry axis. The experimental procedure is

Figure 2. Angular dependence of the EPR signal of an S ) 1/2 species with uniaxial symmetry (yellow) on the axial g tensor in a single
crystal. (a) Orientation of the paramagnet with respect to the applied magnetic field. For the sake of simplicity the molecular axes of the
radical center are assumed to be coincident with the crystal axes. (b) Computer simulation of the spectrum taken at different orientations.
(c) Roadmap of the resonance signal (g and Bres as a function of the orientation).
Surface-Localized Inorganic Radicals and Radical Ions
Figure 3. EPR powder pattern for the same system analyzed in Figure 2.
exemplified in Figure 2 with the aid of computer simulation.
For the sake of simplicity, the principal g axes of the radical
are assumed to coincide with the crystal symmetry axes.
Single-crystal EPR spectra are usually recorded with the
paramagnetic crystal mounted so that it can be rotated in
the resonant cavity about one of the reference axes which is
normal to the applied external field. With reference to Figure
2a, by rotating the crystal in the z′x′ plane along the y′ axis,
a plot of the observed spectrum as a function of the
orientation of the crystal in that plane is obtained, which is
shown in Figure 2b. After repeating this procedure in each
of the coordinate planes, one then seeks to fit the observed
curves with the theoretical formulas by adjusting the relevant
parameters (in this case, g⊥ and g|), obtaining plots analogous
to the one reported in Figure 2c. EPR measurements of
single-crystal surfaces or thin films are now available in the
literature.25-27 Although the number of EPR experiments on
single-crystal surfaces will probably grow in the near future,
the large majority of published work refers to polycrystalline
systems or powders.
3.3. Disordered Systems
Disordered systemsspolycrystalline powders or amor-
phous frozen solutionssare composed of many small
crystallites or species randomly oriented in space. To
illustrate the effect of a disordered system on the EPR
spectrum, let us consider the same radical in the single
crystal (Figure 2) or in the corresponding powder (Figure
3). This situation is actually observed for the O- radical
ion which was studied in KCl single crystals and at the
surface of MgO powders.28
The resultant EPR spectrum, called a powder spectrum,29
is the envelope of spectra corresponding to all possible
orientations of the paramagnetic species with respect to the
applied magnetic field. The simultaneous presence of reso-
nance in the whole range between Bmin and Bmax does not
create a uniform envelope, because the resonant lines are
not uniformly distributed over this range.
Chemical Reviews, 2010, Vol. 110, No. 3 1325
An analytical treatment shows in fact that the microwave
absorption reported as a function of the resonant field Bres
has turning points for orientations corresponding to the
principal components of the g tensor (Figure 3) which show
up in the first-derivative spectral trace. In other words, the
turning points in a powder spectrum can be individuated on
the Bres(θ,φ) surface in correspondence with the points for
which
∂Bres(ϑ,φ)
)0 (3)
∂ϑ
∂Bres(ϑ,φ)
)0 (4)
∂φ
These equations have, in the most general case, three
different solutions: ϑ ) 0; ϑ ) 90°, φ ) 0; ϑ ) φ ) 90°.
Since ϑ and φ are defined in the g matrix reference axis
system, observable features correspond to orientations of the
applied field along the principal axes of the g matrix.
Referring to the simple case of Figure 3, two turning points
are found in the xz plane at ϑ ) 0 (corresponding to g|) and
ϑ ) π/2 (g⊥). Between these two extremes a continuous
absorption is observed (red line in Figure 3), the derivative
leading to the typical shape of an axial powder spectrum
with g⊥ > g| (blue line in Figure 3). In the figure the angular
dependence for the resonance is also shown (black line),
which indicates that, in a powder spectrum, at a given value
of the magnetic field only certain radicals with a given
orientation with respect to the magnetic field fall “in
resonance”. In other words, specific orientations of the
radicals are successively selected by the magnetic field. With
reference to the axial case of Figure 3 at the g| position,
only one orientation, with radicals aligned along the z axis,
contributes to the spectrum, as shown by the plot of
orientation selections on the unit sphere. On the other extreme
at the g⊥ position the turning point in the EPR spectrum
corresponds to all the radicals oriented perpendicularly to
the applied magnetic field, in the xy (⊥) plane. These
1326 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 4. (a) Spin density plot of an O- ion localized at a corner site of MgO. (b) Corresponding experimental (77 K) and simulated EPR
spectra obtained for a 17O-enriched MgO sample. Adapted with permission from ref 28. Copyright 2005 Elsevier.
considerations are important as by setting the magnetic field
at one such turning point either single-crystal-like or powder-
like nuclear frequency spectra are obtained in the presence
of magnetically active nuclei, by ENDOR or ESEEM
techniques.20,30
In real cases, CW-EPR powder spectra of surface radicals
often provide challenging problems, such as heterogeneity
of surface sites, the presence of different paramagnetic
species with overlapping signals, line broadening due to
interaction between nearby paramagnetic species, etc. Some
of the reported complications can be overcome using
advanced EPR techniques such as ENDOR or pulsed EPR.
However, even in the case of classic CW-EPR, as nicely
pointed out by Rieger, one frequently encounters powder
spectra “sufficiently well resolved to give information
rivalling that from dilute single crystal spectra”.31 As is
shown later, systematic use of tools such as multifrequency
EPR, isotopic substitution, and computer simulation is
instrumental to achieve such valuable information.17
4. Formation of Radical Species on Solid
Surfaces
Because of the difficulty of applying EPR techniques to
metals, the radical chemistry is far less advanced for metal
surfaces than for insulators or semiconductors. Among the
latter, metal oxides, because of their dominant role in
heterogeneous catalysis, either as catalysts or as supports,
have been much more investigated than other compounds
such as sulfides, nitrides, halides, etc.
The formation and stabilization of a radical species on a
solid surface have three main origins.
(a) Radical species formed by irradiation of insulating or
semiconducting oxides: In these cases an electromagnetic
radiation of suitable frequency causes charge separation and
formation of two distinct carriers, an electron and a positive
hole. The positive hole may be localized by an O2- ion,
leading to a center describable in chemical terms as an O-
radical ion, with the odd electron confined in an oxygen p
orbital as shown in Figure 4a for the classic case of MgO.
Its EPR spectrum obtained with 17O-enriched MgO exhibits
a hyperfine structure due to 17O (I ) 5/2, Figure 4b). Because
the O- radical ion is actually part of the solid, it can be seen
as a surface point defect. The features of surface centers
generated by irradiation are examined in more detail in
section 4.4.
Chiesa et al.
(b) Adsorption of a stable radical: Important examples are
nitrogen oxides (NO and NO2), which are relatively stable
molecular radicals and can be adsorbed onto surfaces. As
shown later their radical character may be retained when they
weakly interact with surfaces. In other circumstances they
react with the surface, giving rise to more complex adsorbed
species (e.g., NO22-; cf. section 8.3.2).
(c) Chemical interactions at solid surfaces: Atoms and
molecules often undergo dramatic changes upon chemical
interactions with solid surfaces to form either diamagnetic
or paramagnetic (radical or radical ion) species. Radicals can
form either upon direct interaction of a molecule with the
surface or, in other cases, upon interaction between two or
more molecules assisted by the surface. Chemical reactivity
represents the most frequent source for the formation of
surface radical intermediates. For this reason it is convenient
to schematize the basic chemical mechanisms of formation
of surface radicals and to describe the types of chemical
bonds which are commonly formed between radicals and
surfaces.
4.1. Mechanisms of Radical Generation at the
Surface
In section 2 we have briefly mentioned some of the basic
steps in heterogeneous processes which lead to the formation
of surface paramagnetic intermediates. In this section the
different mechanisms leading to the formation of a surface
radical are systematically dealt with considering, for the sake
of simplicity, the surface of a metal oxide containing Mn+
cations and O2- anions and a generic AB or AH molecule.
The principal mechanisms leading to the formation of
surface-adsorbed radical species are the following.
(a) nondissociative electron transfer
-
n+
M(surf) + AB f M(surf)
(n+1)+
+ AB(ads) (5)
+
n+
M(surf) + AB f M(surf)
(n-1)+
+ AB(ads) (6)
(b) surface-induced homolytic splitting
AB f A• + B• (7)
(c) atom transfer32
-
A-H + O(surf) f A• + OH(surf)
-
(8)
Surface-Localized Inorganic Radicals and Radical Ions Chemical Reviews, 2010, Vol. 110, No. 3 1327

Figure 5. Possible models of chemical bonding developing between a surface and an adsorbed radical species: (a) ionic bonding, (b)
covalent bonding, (c) coordinative bonding, (d) van der Waals interaction.

(d) heterolytic splitting followed by surface intermolecular

electron transfer (SIET)33
- -
A-H + Mn+ - O(surf)
2-
f A(ads)Mn+ + OH(surf)
(9)
-
A(ads) + X f A• + X(ads)
-
(10)

(e) addition or coordination of a radical AB• molecule

2-
O(surf) + AB• f ABO(surf)
2•-
(11)

n+
M(surf) + AB• f Mn+ - AB(ads)
•
(12)

AB, in this case a radical molecule, reacts with a surface

oxide ion to form a paramagnetic charged adduct (eq 11) or
coordinates onto a surface cation (eq 12).
In the present review, we only consider adsorbed radicals,
keeping in mind that, at catalytic reaction temperatures,
radicals can become mobile and even desorb into the gas
phase. Readers interested in the behavior of radicals in
heterogeneous catalytic reactions are referred to specific
papers.34,35
4.2. Chemical Bonds between Radicals and
Surfaces
The different types of chemical bonds which exist between
the surface and a radical formed by either one of the reactions
given above are schematically illustrated in Figure 5.
(a) Ionic bonding (Figure 5a): Nondissociative electron
transfer (eqs 6 and 7) leads to radical ions in electrostatic
interaction with the surface. This can be illustrated by the
case of the superoxide ion (O2-) (Figure 6), which has been
widely investigated in relation to the activation of dioxygen.
Its formation at ionic surfaces can occur not only by electron
transfer (eq 6) but also via different pathways reviewed
elsewhere.33 The bond between O2- and the surface of oxides
such as MgO is essentially ionic, O2- being stabilized on
Figure 6. Schematic representation and CW-EPR spectra of the
superoxide radical anion stabilized on MgO. Spectra obtained with
different 17O levels of enrichment are shown. Adapted with
permission from ref 36. Copyright 2002 American Institute of
Physics.
top of a surface Mg2+ cation. Analysis of the complete
hyperfine structure36 of adsorbed 17O2- (Figure 6) leads to a
spin density on O2- of 0.964, i.e., close to unity, showing
that the electron transfer from the surface to dioxygen is
practically complete. The energy of ionic bonding is rather
high and depends on the type of surface cation. In the case
of O2- stabilized on low coordinated Mg2+ cations, values
ranging between 2.5 and 3.0 eV are obtained depending on
the coordination of the Mg2+ ions.37
(b) Covalent bonding (Figure 5b,c): This is the case of
surface radicals formed via overlap of orbitals of the adsorbed
species and a surface atom or ion. If the bonding involves a
transition-metal ion (Figure 5c), it can be described in the
context of interfacial coordination chemistry.6 A simple
example of a radical covalently bound to the surface is the
O3- ion formed by addition of dioxygen to a surface O-
center:38
1328 Chemical Reviews, 2010, Vol. 110, No. 3 Chiesa et al.

- -
O2(g) + O(surf) f O3(surf) (13)

The ozonide radical ion is bent with a single covalent bond

between the oxygen belonging to the surface and the
molecule (Figure 7). At room temperature rapid rotation
around this bond averages two of the rhombic g tensor
components.
(c) van der Waals interactions (Figure 5d): These are at
play in physical adsorption with no electron exchange or
significant orbital overlap with the surface. The physisorbed
radical forms by adsorption of a molecular radical. As
mentioned above the most relevant case is that of nitrogen
oxides (NO, NO2) used as spin probes of surface properties
(section 8.3).
Figure 7. Schematic representation and CW-EPR spectra of the
ozonide radical anion stabilized on the surface of CaO (experimental
4.3. Identification and Reactivity of Radical spectrum recorded at 77 K). Adapted from ref 38. Copyright 2006
Intermediates in Heterogeneous Catalysis American Chemical Society.

Surface science and surface chemistry investigations are

necessary (though not sufficient) requirements to unravel the
essential steps of catalytic reactions. The role of EPR is not
limited however to preliminary characterizations of catalytic
systems but can be easily extended toward the investigation
of the catalytic mechanisms itself. EPR actually gave
important contributions to the understanding of several
catalytic processes including carbon monoxide oxidation,39
dehydrogenation of ethane,40 oxidation of methane to metha-
nol and formaldehyde,41 and alkene oligomerization.42 In two
other recent cases, the role of EPR has been essential.
The first case is the oxidative coupling of methane on Li-
doped MgO first investigated by Lunsford using an elegant
matrix isolation approach. Methyl radicals (•CH3)present in
the gas phase over the catalyst at 773 K were frozen on a
sapphire rod kept in solid argon at 14 K43 (Figure 8). The
methyl radical, observed at low oxygen partial pressure,
exhibits a quartet of hyperfine lines, with 1:3:3:1 relative
intensities. The state of the surface of Li/MgO during the
catalytic reaction was explored by quenching a working
catalyst from the reaction temperature to that of liquid oxygen
and by comparing its EPR spectrum with that of a solid
slowly cooled to room temperature.
In the rapidly quenched solid various paramagnetic species
are present44 including Li+O- pairs (Figure 8). The O- site
(a positive hole trapped at an oxide O2- ion) is able to Figure 8. Schematic representation of the activation of methane
abstract hydrogen from methane to form •CH3 radicals on Li-doped MgO. The solid contains surface-exposed O- sites
according to the heterogeneous reaction which react with methane via H transfer. The O- singly occupied
molecular orbital (SOMO), in yellow, and the corresponding
- simulated EPR spectrum are at the bottom of the figure. Methyl
CH4(g) + O(surf) f •CH3(g) + OH(surf)
-
(14) radicals are formed in the gas phase (SOMO orbital, yellow, and
simulated EPR spectrum at the top of the figure) where they undergo
The formation of ethane occurs via •CH3 radical coupling, radical coupling.
while ethene forms via successive reactions involving ethane
and again surface O- ions, for further H abstraction. The presence of a cocatalyst, an alkylaluminum compound.
above reaction is a true example of a heterogeneous-ho- Despite the huge number of papers devoted to this field, a
mogeneous catalytic reaction,45 i.e., initiated on the surface detailed, atomistic understanding of the reaction mechanism
for the production of methyl radicals and continuing in the remains somewhat elusive.
gas (homogeneous) phase for radical coupling. The fact that In particular, the critical process of the catalyst activation
methyl radicals are neutral explains that they can easily leave with alkylaluminum compounds has been claimed to involve
the ionic surface at the reaction temperature. the formation of alkyl radicals. The latter have been
The second case concerns olefin polymerization investi- evidenced recently by in situ EPR on a catalyst carefully
gated by Freund and co-workers46,47 on a model Ziegler-Natta prepared by means of surface science techniques (Figure 9).
catalyst. Since their discovery, Ziegler-Natta catalysts have A MgCl2 film was grown layer by layer on Pd(111). Defects
undergone a number of improvements and modifications, (F centers) were then created in the film to favor the grafting
which led to the so-called third-generation catalysts, made of TiCl4, the active component. The EPR signal of the F
of TiCl4 supported on MgCl2 crystals, working in the centers was observed to decrease after reaction with TiCl4,
Surface-Localized Inorganic Radicals and Radical Ions Chemical Reviews, 2010, Vol. 110, No. 3 1329

Figure 9. Preparation of the model Ziegler-Natta catalyst. (a) Top: preparation of a well-ordered defect-free MgCl2 film. Center: defect
creation by different techniques. Color centers are symbolized by yellow spheres. Bottom: anchoring of TiCl4.; TiCl4* (red) denotes TiCl4
moieties adsorbed to color centers. (b) Top: see the bottom of part a. Center: adsorption of AlR3 on the surface at 40 K. Alkyl radicals are
symbolized by yellow spheres. The experimental EPR spectrum recorded upon adsorption of Al(CH3)3 is also shown. Bottom: catalyst after
polymerization of ethane. Adapted with permission from ref 47. Copyright 2002 Wiley.
revealing the importance of such centers in the grafting step migration of the carriers to the surface of the solid is of vital
of the catalyst onto the support. Finally, the system is importance because chemical interactions (reduction and
activated by exposing it to alkylaluminum, which acts as a oxidation for the electron and hole, respectively) become
cocatalyst. Upon addition of trimethylaluminum, the spec- possible at the surface with adsorbed chemical entities.
trum of Figure 9, unambiguously due to ethyl radicals, was Monitoring the presence, stabilization, and reactivity of
observed thus showing that alkyl radicals are indeed involved charge carriers at oxide surfaces, which often lead to EPR
in the activation of the catalyst. active-species, is a crucial problem for photochemical
applications of inorganic compounds and has been the object
4.4. Radical Intermediates in Heterogeneous of intense EPR research, starting with the seminal work of
Photocatalysis Howe and Grätzel on colloidal TiO2.52,53 Under such condi-
Photocatalysis makes use of the energy of photons to tions, the electron tends to localize in the d orbital of a cation
initiate a chemical process at the surface of a solid catalyst, while the positive hole moves to a nearby O2- ion, forming
usually a semiconducting oxide.48,49 Major applications of paramagnetic Ti3+ and O- ions, respectively:
photocatalysis include the mineralization of organic pollutants
and the splitting of water into H2 and O2,50 following the
discovery of Fujishima and Honda in 1972,51 who observed
this electrochemical photolysis at a semiconductor (TiO2)
electrode. These photochemical applications have a common
starting point, which consists in the charge separation,
induced by irradiation of a semiconductor (often an oxide,
MO) with photons having energy higher than the semicon-
ductor band gap. Photon absorption promotes electrons from
the valence band (VB; constituted by the O2- orbitals) to
the conduction band (CB), leaving in the former a positive Figure 10. Typical ESEEM spectra observed upon 355 nm
excitation of TiO2 nanoparticles measured at a resonant field of
hole: trapped holes (brown) and trapped electrons (blue). The peak at
+ - the 1H frequency (14.9 MHz) demonstrates that hole centers (O-)
MO + hν f hVB + eCB (15) are subjected to weak dipolar coupling with surrounding 1H nuclei
from adsorbed water and are thus localized at the surface of TiO2
After formation the two carriers can either recombine particles. Adapted from ref 64. Copyright 2007 American Chemical
(dissipation of photon energy) or move in the crystal. The Society.
1330 Chemical Reviews, 2010, Vol. 110, No. 3
Ti4+ + e- f Ti3+ (16)
2-
O(s) + h+ f O(s)
-
(17)
Great efforts were made to improve our understanding of
the state and localization of surface-stabilized carriers on
TiO254-58 and other oxides.59,60 Knözinger and Diwald and
co-workers, in particular, investigated the dynamics of charge
separation by EPR following time-resolved responses of the
photoexcited O- species. They also monitored the chemical
activation of small molecules such as H2 or O2 at the surface
of UV-excited nanocrystals.61,62 While numerous studies dealt
with the nature of the photogenerated charges, few considered
the influence of the size/shape of the crystals on the
electron-hole separation processes.63
Using CW and pulsed EPR techniques, Dimitrijevic et al.64
confirmed the trapping of electrons at bulk titanium ions
while the holes are trapped at surface oxygen ions. T1 and
T2 relaxation times were measured using two-pulse echo and
three-pulse inversion-recovery techniques, which revealed
that the trapped holes were interacting by weak dipolar
couplings with surrounding hydrogen nuclei from adsorbed
water64 (Figure 10). The authors also demonstrated that
distortions to the Ti-O bond, resulting from reconstruction
Figure 11. Surface reduction of oxygen to superoxide ion, a probe of adsorption sites: (a, b) surface-trapped electrons on MgO (a)
and superoxide stabilization on distinct surface Mg2+ sites (b), (c, d) surface-deposited Cs atoms on MgO (c) and superoxide stabilization
on Cs+ and Mg2+ sites (d). In the central part of the figure the SOMO orbital of O2- and the corresponding energy level scheme are
reported.
Chiesa et al.
of the surface of nanoparticles, caused a variation in the g
tensor values and spin-lattice relaxation processes of trapped
electrons. In a field largely dominated by CW-EPR, the
authors in this paper show how pulsed EPR can help improve
our understanding of TiO2 surface chemistry.
Another approach to photocatalytic investigations is to
monitor surface radical species formed after reaction of a
molecule with a photogenerated charge carrier. The subse-
quent photoformed radical species include superoxide and
ozonide ions (O2-, O3-) on titanium dioxide65,66 or zirconium
dioxide,67 CO2- from carbon dioxide on MgO and ZrO2,68-70
and N2O2- from nitrous oxide on TiO2.71 In section 7 and
section 8 we provide a systematic analysis of each type of
radical.
5. Experimental Approaches in EPR of Surface
Radical Species
5.1. Paramagnetic Probes of Surface Adsorption
Sites
A problem common to surface chemistry and catalysis is
identification of the adsorption sites. The same problem exists
in studies dealing with surface-stabilized radicals. For such
Surface-Localized Inorganic Radicals and Radical Ions
investigations, EPR may help identify the adsorption site via
the shf structure on the basis of the direct observation of the
adsorbate-surface interaction (section 3.1) and, sometimes,
via the g tensor. The shf structure is observed only when
nonzero nuclear spins are present.
When the shf interaction is too small, better results can
be achieved using pulsed techniques. The example of
superoxide radical ions (O2-) detected by CW-EPR illustrates
how surface/adsorption sites can be identified (Figure 11).
O2- is a 13-electron π radical with 3 electrons in the two
π* antibonding orbitals (MO scheme of Figure 11). For ionic
systems, O2- is usually adsorbed on a cationic site. The two
limiting values of the g tensor for this species are to first
order:
gxx ) ge, gzz ) ge + 2λ/∆ (18)
where λ is the spin-orbit coupling constant of oxygen and
∆ the energy between the π* orbitals, caused by the surface
crystal field, i.e., the charge of the cation onto which O2- is
adsorbed. z is the direction of the internuclear axis and x
that perpendicular to the surface. The gzz component, which
depends on ∆, can be used to determine, in particular in
mixed oxides containing more than one type of cation, the
nature of the adsorption site, because ∆ is roughly related
to the electric charge of the adsorbing cation.7 The sensitivity
of O2- to the characteristics of the adsorption goes however
beyond the simple identification of the site itself. The gzz
component in the spectra of O2- ions adsorbed on binary
oxides such as MgO or TiO2 is often resolved in various
components monitoring distinct families of surface cations
having the same nominal charge (e.g., 2+ in the case of
MgO) but different coordinative environments (e.g., ions at
planar faces, edges, and corners).72,73 Figure 11 illustrates
two cases of oxygen reduction to superoxide on MgO
containing electron donor centers. In one case the centers
are surface-trapped electrons, or (e-)(H+) centers,74,75 and
adsorbed cesium atoms in the other.76 Figure 11 shows the
EPR spectra of the starting materials. Figure 11a gives the
typical signal of trapped electrons (the hyperfine doublet is
due to the weak interaction of the electron with the nearby
proton) and Figure 11c that of adsorbed Cs atoms with eight
hyperfine lines from 111Cs (I ) 7/2).76 Both signals disappear
upon interaction with oxygen, and new signals due to O2-
appear. The first one (Figure 11b) is characterized by three
gzz values (2.091, 2.077, 2.065) corresponding to three distinct
Mg2+ ions. No superhyperfine structure due to 25Mg (I )
5/2) is observed, probably because of its low natural
abundance (10%). The site discrimination is based, in this
case, on the g tensor only.72 The second O2- spectrum (Figure
11d) is more complex with an eight-line shf structure due to
111
Cs associated with the three g components. Additionally,
two unstructured lines at g ) 2.077 and g ) 2.065 correspond
to O2- adsorbed on Mg2+. This indicates that superoxide ions
are formed by electron transfer at the expense of Cs atoms
and that a fraction of the radical ions are stabilized on the
resulting Cs+ cations while a second fraction undergoes
spillover toward the matrix sites.
In summary, EPR can provide information not only on
the nature (e.g., O2-) and structure (e.g., side-on) of the
adsorbed species but also on the nature (e.g., Cs+) and
structure (e.g., Mg2+ at corners, edges, and faces) of the sites
present at the surface.
Chemical Reviews, 2010, Vol. 110, No. 3 1331
5.2. Molecular Motion of Adsorbates on Solid
Surfaces
As mentioned in section 3, EPR can help in the investiga-
tion of the dynamics of paramagnetic adsorbates, which is
of major interest in surface science and catalysis. The time
scale typical of most spectroscopies, in particular electron
spectroscopic methods, is too short to detect molecular
motions, which are in the range 10-6-10-10 s. This is the
operating regime of EPR (Table 1), which is thus very useful
in studying molecular dynamics at the gas-solid interface.
The EPR spectrum dramatically depends on the rate of
the tumbling, which results in a loss of the spin density
anisotropy. Two motional regimes can be distinguished: (1)
fast regime, where the correlation time is much smaller than
the microwave frequency, which results in a complete loss
of spectral anisotropy, leading to line narrowing); (2) slow
regime,77 where the interactions of the unpaired spins,
determining the shape of the EPR spectrum, are modulated
through molecular dynamics, giving rise to distinct EPR
spectral features, which enable different forms of molecular
motion and adsorbate orientations to be distinguished. This
regime is particularly informative on the local structure and
dynamics of adsorbates.
EPR has been employed to investigate the local dynamics
of NO in different molecular sieves,78 as well as O2 on
different surfaces.79 The adsorption of NO2 on copper,80
Vycor glass,81 silica gel,82 and a number of other matrixes
has been studied in detail. EPR studies are also available on
NO2 in rare gas matrixes83-85 and NO2/N2O4 thick films,86
where preferential orientations were obtained by deposition
of the gases on flat surfaces. Importantly, EPR investigations
of NO2 on single-crystal surfaces of clean Au(111)87 and
Al2O3(111)88 under UHV conditions have also been reported,
which give insights into the geometrical arrangement and
local dynamics of the adsorbate.
To exemplify the potential of EPR in studying the
orientation and dynamics of adsorbates on surfaces, in the
following we illustrate the case of NO2, which can be
considered as an effective surface “spin probe” and for which
detailed data are available on powders and single-crystal
surfaces. We discuss here the results reported by Shiotani
and Freed81 on NO2 adsorbed on Vycor glass. In Figure 12a,
the computer simulation for the rigid limit (the adsorbed
molecule is immobile) is reported and the different compo-
nents (x, y, and z) for the three MI quantum numbers are
shown. The features of the spectrum are due to two main
contributions. The first is the electron Zeeman interaction,
giving an anisotropic g tensor (gx * gy * gz) whose
components correspond to three distinct resonant fields. Each
g component is then split into three lines due to the hyperfine
interaction of the unpaired electron with the nuclear spin of
N (I ) 1).
The EPR spectra shown in Figure 12b-e following the
work by Shiotani and Freed81 correspond to the effect of
rotation of NO2 about the molecular axes x, y, and z indicated.
The simulation was performed using the Easyspin package,89
imposing a diffusion rate about a given axis of 1 × 108 s-1
while a value of 5 × 105 s-1 was used for the perpendicular
axis. In this way, the A and g tensor components perpen-
dicular to the rotational axis are averaged out while the other
components are only slightly broadened. For example, for
rotation about the x axis (Figure 12b), it can be seen that
the x components remain at their rigid limit position and are
only slightly broadened while the y and z components are
1332 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 12. Theoretical spectra illustrating the change of the EPR spectrum of NO2 occurring as a function of rotation about different
molecular axes: (a) rigid limit, (b) rotation about the x axis, (c) rotation about the y axis, (d) rotation about the z axis, (e) isotropic rotation.
The rotational correlation time about a given axis was set to 1 × 10-8 s, while rotation about the perpendicular axis was 5 × 10-5 s.
averaged out. The same arguments apply to the other cases,
clearly showing how information on motion can be revealed
by the spectrum. Finally, in Figure 12e the case of isotropic
motional averaging is shown, which leads to the collapse of
all anisotropies, which are reflected by the line width
dependency on MI. Analysis of the experimental spectra in
the range 4.2-220 K led Shiotani and Freed81 to conclude
that the NO2 molecule adsorbed on Vycor glass undergoes
rotation about the y axis (suggested to be parallel to the
surface) at temperatures below 77 K, while an isotropic
motion likely due to translational diffusion is observed at T
> 77 K.
The EPR spectra of NO2 at coverages below the monolayer
on an oxide single crystal under UHV conditions have been
reported by Schlienz et al.88 These studies are particularly
interesting as they allow in principle establishment of the
orientation of the molecular radical on the surface, i.e., the
molecular axis about which the motion occurs. In the case
of NO2 on the 111 surface of Al2O3 no rotational motion
was found below 100 K, indicating that the rotational
correlation time is longer than 10-7 s. The EPR spectra were
found to be drastically influenced by translational diffusion
of the NO2 monomers on the surface, which leads to
dimerization (2NO2 T N2O4) and loss of the EPR spectrum.
6. Main Families of Surface Inorganic Radicals:
Electronic Configurations and Geometry
For the sake of clarity surface radical species can be
classified as is done in gas phase or in various matrixes.2 It
is however convenient, first, to introduce their most common
structural and electronic features. In some cases, it happens
that radicals and radical ions observed at surfaces have their
bulk analogues (in rare gas matrixes or in solids) often with
a different structure (section 4.2). Other radicals have never
been isolated in a matrix and are typical of surfaces.
The influence on EPR parameters of the electronic
configuration and the nature of the SOMO have been
discussed by Atkins and Symons1 for inorganic radicals on
the basis of Walsh diagrams. These diagrams, giving the
energy of molecular orbitals as a function of bond angles,
are based on the rules formulated by Walsh90 which relate
the shape of a molecule in a given structural class to the
number of valence electrons. The most important rule states
that a molecule adopts the structure that best stabilizes its
Chiesa et al.
highest occupied molecular orbital (HOMO). If the HOMO
is unperturbed by the structural change under consideration,
the occupied MO lying closest to it determines the preferred
geometry. The first-principles foundations for these empirical
rules have been clarified by March.91 In the following we
refer to this systematic classification. In some cases, the
structural features and the SOMO orbitals calculated for the
gas-phase radicals using the Gaussian 03 code are also
reported.92
6.1. Monoatomic Species
Among single metal atoms, the most widely investigated
in EPR are those having a 2S1/2 nondegenerate electronic state
because they give rise to spectra easily observable. This is
the case of hydrogen and alkaline metals whose spectra are
characterized by a hyperfine (hf) structure because of their
nonzero nuclear spin.
Hydrogen atoms trapped in the bulk of different matrixes
exhibit spectra characterized by large hf coupling constants
rather close to the gas-phase value.14 A remarkably different
situation is found when H atoms are stabilized by the surface
of ionic oxides. Hydrogen atoms spontaneously dissociate
on the surface of alkaline-earth-metal oxides to give blue-
colored samples which exhibit intense and complex EPR
spectra. The nature of these surface paramagnetic species,
which has been recently elucidated by some of us,74 can be
described in terms of unusual (the simplest) (H+)(e-) ion
pairs stabilized on low coordinated sites of the ionic surface
(Figure 11a). It should be pointed out that this unusual
hydrogen chemistry, only observed on insulating ionic
oxides, is due to the strong proton affinity of the very basic
O2- surface ions coupled with the energy gain obtained by
trapping the electron near low coordinated M2+ cations.
Alkali-metal atoms have been studied in a variety of solid
matrixes93 and liquid solvents2 and, only recently, at sur-
faces.94 In all cases the matrix influences the electronic
properties of the metal atom (matrix effect) as reflected by
the hyperfine coupling constant of the EPR spectra.95
Typically, the EPR spectra of trapped alkali-metal atoms in
frozen rare gases and hydrocarbons exhibit hyperfine interac-
tions which depart by only a few percent from the gas-phase
values.93
A drastically different situation occurs when alkali-metal
atoms are deposited on basic oxides such as alkaline-earth-
Surface-Localized Inorganic Radicals and Radical Ions
Figure 13. Experimental and simulated EPR spectrum of Au atoms
adsorbed on MgO(001)/Mo(100). The angular dependence of the
reonance position and the orientation of the g components with
respect to the applied magnetic field are also shown. Reprinted with
permission from ref 100. Copyright 2006 American Physical
Society.
metal oxides.94,96 A consistent reduction (about 50%) of the
hf coupling constant with respect to the gas-phase value is
observed in this case, which implies a strong perturbation
of the unpaired electron wave function. A very similar effect
is observed when alkali metals are dissolved in solvents such
as methylamines or crown ethers,97,98 the common fea-
ture with alkaline-earth-metal oxides being the ability of the
matrix to donate a lone pair. Upon interaction with the lone
pair localized on a surface O2- ion, the ns orbital of the
alkali-metal atom containing the unpaired electron is desta-
bilized, leading to an “expanded” atom structure reminiscent
of a Rydberg state.94 This intriguing situation, which seems
to be a general feature of alkali-metal atoms interacting with
basic solvents, is fully supported by theoretical calculations,94,99
which allow rationalization of the matrix effect within the
framework of Pauli repulsions.
The same arguments were adopted by Yulikov et al.100 to
explain the reduction of the Au hf coupling constant observed
when Au atoms are deposited on a MgO thin film (Figure
13).
It is interesting to compare these results with those
obtained by Mile and co-workers101 when depositing coinage
metals (Cu, Ag, Au) on alkali-metal chlorides by means of
the rotating-cryostat technique. Although this technique leads
typically to bulklike species, in this particular case the authors
suggest that the metal atoms may be trapped at asymmetric
sites located at the interface between layers of the alkali-
metal chloride, resembling species trapped at surfaces or
grain boundaries rather than in the bulk. Indeed the EPR
spectra show much lower hyperfine interactions than for
atoms prepared in alkali-metal chloride single crystals from
metal ions in cationic substitutional sites by electron
capture102,103 and are in line with the spectra reported by
Yulikov et al.100 for Au deposited on MgO. Finally, it is
interesting to note that the trend observed by Mile and co-
workers101 for coinage metals on alkali-metal chlorides (LiCl,
NaCl, KCl) parallels the trend reported by some of us94,104
on alkali metals deposited on alkaline-earth-metal oxides
(MgO, CaO, SrO). In both cases a nearly linear relation can
be found between the metal hf coupling constant and the
lattice parameter. This trend has been discussed previously104
in terms of increasing lone pair donaticity (i.e., Lewis
basicity) brought about by the reduced Madelung potential
around the basic O2- ion as the lattice parameter increases.
Chemical Reviews, 2010, Vol. 110, No. 3 1333
Figure 14. SOMO orbitals of representative 11-electron (NO) and
13-electron (ClO) radicals in the gas phase. The energy level
schemes on the right-hand side are referred to the same species in
the adsorbed state.
Systems having orbitally degenerate ground state are
hardly observed in EPR in a inert matrix. Their orbital
angular momentum however is sensitive to asymmetric
environments so that they can be observed, as “quenched”
species, in crystals or at surfaces, the field acting at the
surface often being asymmetric. This fact, which is widely
discussed in the following for diatomic molecules and radical
ions, is also true for some monatomic species such as O-
(Figure 4) and S- which have been identified in a number
of cases.
6.2. Diatomic Radical Species
While 2Σ systems prevail in the field of matrix isolation,
the most important biatomic surface radical species are 2Π
radicals. Hardly observed in an inert matrix, because their
spectrum, spreading over a large domain of the magnetic
field, becomes undetectably weak, 2Π radicals have been,
in some cases, individuated in an ionic matrix whose electric
fields can more efficiently perturb the degenerate orbital
states. In this case they are referred to as quenched species.
The asymmetric field typical of a surface is also a powerful
quenching agent so that EPR spectra of surface-adsorbed 2Π
radicals are easily observed. Examples of both possible
substates of a 2Π molecule (2Π1/2 and the 2Π3/2) have been
observed. Eleven-electron radicals such as NO, N2-, and CO-
have one electron in 2π antibonding orbitals and a 2Π1/2
ground state, while thirteen-electron radicals such as O2- or
ClO have three electrons (or one positive hole) in these
orbitals and a 2Π3/2 ground state (Figure 14).
The diatomic radicals mentioned above are often crucial
intermediates in the activation of small molecules. Some of
them are important also to probe surface crystal fields and
to identify adsorption sites (sections 5, Figure 11). Nitric
oxide (NO) and the superoxide radical ion (O2-) are the two
most widely used probes.
6.3. Triatomic, Tetraatomic, and Pentaatomic
Radical Species
A variety of polyatomic radicals and radical ions have been
observed at surfaces. All of them have a nondegenerate
ground state with well-separated energy levels so that they
are insensitive to the influence of local crystal fields. The
1334 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 15. SOMO orbitals and Walsh diagram of the molecular
orbital energy of the CO2- radical ion in the gas phase.
Figure 16. SOMO orbitals of O3- (a) and CO3- (b) radical ions
in the gas phase.
observed g tensors are close to those observed for their
matrix-isolated analogues (if any) and can be considered
fingerprints of the species. We describe in the following two
families of triatomic surface radical species both having
nonlinear structure.
The former family includes radicals with 17 valence
electrons such as CO2- and NO2. The electronic structure
of these species can be easily understood starting from that
of the linear, 16-electron CO2 molecule. The 17th electron
causes the molecule to bend, and the SOMO (2A1 symmetry)
has the features of a π antibonding orbital built up by three
parallel p orbitals on C and O atoms with a contribution of
the carbon 2s orbital1 (Figure 15). This involves a consider-
able isotropic splitting in the EPR spectrum from the central
atom (13C and 14N or 15N in the examples proposed) which
adds to the anisotropic one, typical of p orbitals.
By contrast, for bent 19-electron species (O3-, NO22-,
ClO2) the isotropic coupling is instead not expected (except
for a small contribution due to spin polarization) because
the odd electron is now in the orbital which is formed purely
by antibonding interaction of the other three parallel π
orbitals on the three atoms (2B1 symmetry, Figure 16a).
Among the tetraatomic radical species we consider cases
of 21, 23, and 25 valence electrons, respectively, the latter
two having matrix-isolated analogues.
Twenty-three-electron radicals such as CO3- or NO3
generally have planar D3h symmetry. The SOMO, built up
by the out-of-phase combination of in-plane p orbitals of
the outer atoms (oxygen in our case), is essentially non-
bonding (2A2 symmetry) so that no or a very small contribu-
tion from the central atom to the hyperfine structure is
expected, and these species can be actually defined as
tetraatomic planar σ radicals (Figure 16b).
Twenty-five-electron radicals such as NO32- and ClO3
generally have pyramidal C3V symmetry. The electronic
structure recalls that of 17-electron radicals because the 25th
unpaired electron upon entering the antibonding 2a2 orbital
causes bending of the 24-electron planar structure. The
Chiesa et al.
distortion introduces some s character (from the s orbital of
the central atom) to this orbital, reducing its antibonding
character. The SOMO has thus a 2A1 symmetry, and both
isotropic and anisotropic hpf contributions from the central
atom are expected.
The 29-electron radicals observed at surfaces are low-
symmetry molecules without analogues among matrix-
isolated species so that no detailed analysis of the electronic
structure comes from previous studies.
7. Carbon-Containing Inorganic Radicals
The interest of this family of surface radicals comes from
the importance of the chemistry of carbon oxides. The
activation of carbon oxides may be achieved via heteroge-
neous catalysis, by surface photochemistry, and also via
mechanochemistry. While the catalytic conversion of CO has
always been a basic topic in heterogeneous catalysis, the
interest in the artificial conversion of CO2 is more recently
triggered by its impact on global warming and sustainabil-
ity.105 CO2 is thermodynamically very stable, and its conver-
sion requires high free energy chemicals or an external supply
of thermal, electrical, or photochemical energy. Radicals are
often intermediates of surface chemical processes involving
carbon oxides. The relevant members of the family are
illustrated in the following.
7.1. CO2- Radical Anion
The CO2- carboxylate radical anion was first observed by
Lunsford and Jayne106 upon adsorption of CO2 on UV-
irradiated MgO containing trapped electrons, following the
direct electron transfer reaction
- -
CO2(g) + e(surf) f CO2(ads) (19)
The same radical, with slightly different EPR parameters,
was also observed107,108 after adsorption of CO on MgO
containing O- ions, following the reaction
- -
CO(g) + O(surf) f CO2(surf) (20)
A detailed X- and Q-band EPR study of the CO2- radical
using 13C and 17O labeling was undertaken by Vedrine and
co-workers109 on irradiated MgO. The formation of the
carboxylate radical anion on “electron-rich” MgO was
recently revisited by some of us in light of the advanced
understanding of excess electron centers on the surface of
alkaline-earth-metal oxides.110,111 The results confirm those
reported earlier106 leading to a detailed mapping of the
unpaired electron spin density distribution over the entire
radical anion. The EPR spectrum of CO2- displays a rhombic
g tensor (g1 ) 2.0026, g2 ) 2.0009, g3 ) 1.9965) very close
to that observed for its analogues, observed in irradiated
carbonates.112 CO2- is a 17-electron radical with its unpaired
electron in the first πu* orbital of the CO2 molecule, built
from three parallel p atomic orbitals. The extra electron
causes the molecule to bend, thus introducing a partial
admixture of the 2s carbon orbital to the SOMO orbital
(Figures 15 and 17). This fact explains the structure of the
nearly axial 13C hyperfine tensor (I(13C) ) 1/2) with A1 )
181 G, A2 ) 224 G, and A3 ) 177 G (Table 3), in good
agreement with the parameters of the same species observed
in irradiated sodium formate.112 As discussed in section 4,
the nondegenerate ground state makes the radical rather
Surface-Localized Inorganic Radicals and Radical Ions Chemical Reviews, 2010, Vol. 110, No. 3 1335

Figure 17. Calculated structure and EPR spectra of the CO2- radical ion stabilized on MgO obtained by CO2 adsorption on the surface
containing trapped electrons: (a) CW-EPR spectrum of 13C-enriched CO2-, (b) hyperfine sublevel correlation spectroscopy (HYSCORE)
spectrum of the same radical stabilized on a 17O-enriched MgO matrix. Adapted from ref 111. Copyright 2008 American Chemical Society.

Table 3. Spin Hamiltonian Parameters of Carbon-Containing Surface Paramagnetic Species

g tensor A tensor(13C)/G A tensor(17O)/G
radical
species solid g1 g2 g3 A1 A2 A3 A1′ A2′ A3′ refs
CO2- MgO a
2.0026 2.0009 1.9965 181 224 177 24 60 25 110
CO2- ZrO2 EPR spectra reported without detailed parameters 69, 70
CO2- SiO2 2.0031 2.0012 1.9977 unres 236 unres 114
CO2- apatite 2.0030 2.0015 1.9970 164 199 159 127, 128

COδ- (CO-) MgOb 2.0041 2.0014 1.9917 25.3 44.5 120

C2O2- MgOc 2.0060 2.0050 2.0021 7.8 8.4 19.0 117-120

C2O2- CaOd 2.0063 2.0048 2.0023 Aav ) 11.8 G 118

CO3- apatite 2.0170 2.0084 2.0060 unres 13.0 13.0 127

CO4- e
MgO 2.0040 2.0072 2.0015 unres 3.5 2.8 unres unres 100 129
a b
The reported values refer to the most abundant of three slightly different species. The reported species is the most abundant of four slightly
different species. The average value for δ is about 0.6. c Two equivalent C atoms (ethylenedione anion). Assigned to cyclic C6O63- in ref 7.
d
Assigned to cyclic C6O63-. e Peroxy radical type (O2COO). Two inequivalent terminal O atoms.

insensitive to the local crystal field, explaining the similarity
between the parameters of the radical on the surface or in
the bulk of irradiated crystals.
The 13C hyperfine structure indicates a spin density mainly
associated with the 2pz carbon orbital (z being the direction
perpendicular to the molecular plane) with c2(2pz) ) 0.45 and
c2(2s) ) 0.18. 17O-labeled CO2 allows determination of the 17O
(I ) 5/2) nearly axial hyperfine tensor of CO2- adsorbed on
MgO:110 A1 ) 24 G, A2 ) 60 G, and A3 ) 25 G, in agreement
with earlier work on the same system109 and with results on
bulk radicals in irradiated sodium formate.113 The O nuclei
are found to be equivalent, suggesting a side-on structure
on low coordinated Mg2+ cations which is shown in Figure
17 together with the CW-EPR spectrum of the 13C-enriched
radical.110 As a complement to the CW results, HYSCORE
experiments allowed the observation of the coupling of the
carboxylate radical with a nearby proton belonging to a
surface OH- group, resulting in the following superhyperfine
tensor: A(H) ) -3.5 - 4.3 + 3.9 MHz. These results,
together with DFT calculations, enabled establishment of the
location and geometry of the adsorbed CO2- radical anion
(Figure 17).111
Quite recently it has been shown that photoreduction of
CO2 to CO takes place on MgO under UV irradiation in the
presence of hydrogen or methane as a reducing agent.68 The
process starts with the photoinduced formation of a CO2-
radical ion whose spectrum is similar to those described
before. The radical ion is reduced in the dark to form a
formate or an acetate by hydrogen and methane, respectively.
A very similar mechanism was proposed by the same authors
for the photoreduction of CO2 to CO on ZrO2 (zirconia).
The photoformed CO2- radical, observed for the first time
1336 Chemical Reviews, 2010, Vol. 110, No. 3
on ZrO2, reacts with hydrogen in the dark68-70 during the
successive steps of the reaction.
The formation of CO2- was also observed on silica gel,
UV-irradiated under an atmosphere of O2 and CO.114 The
authors suggested that a photogenerated hole center of O-
reacts with CO to form CO2- with the scheme reported in
eq 20.
7.2. Radicals Generated from Carbon Monoxide
While CO2 does not form paramagnetic species by reaction
with the surface of bare MgO, carbon monoxide leads to a
complex chemistry which involves both radical and diamag-
netic species. The first observation of a paramagnetic species
by adsorption of CO on thermally activated MgO was
reported by Lunsford and Jayne.115 The axial spectrum (g⊥
) 2.0055 and g| ) 2.0021) was assigned to a surface-bonded
monomeric carbonyl species. The formation of the radicals
is very slow, and the EPR spectrum takes weeks to reach its
maximum intensity. A similar spectrum, observed later by
Klabunde on MgO116 and Cordischi et al. on Co2+/MgO,117
was independently assigned to the dimer C2O2- radical anion.
This assignment was based on 13C labeling experiments and
computer simulation showing two equivalent carbon atoms.
It is not surprising that Lunsford and Jayne115 were unable
to draw the same conclusion because their experiment with
13
CO was affected by the presence of a large percentage of
12
CO, leading to spectral features obscuring the central line
of the 13CO signal. Further work, also by Klabunde’s118 and
Cordischi’s119 groups, questioned the earlier hypothesis of
the dimer radical ion proposing that the radical was in fact
C6O63- adsorbed with only two carbons interacting with the
unpaired electron. The work was extended by Morris and
Klabunde118 to CaO, leading to spectra very similar to those
recorded with MgO and to the same assignment.
The chemistry of paramagnetic CO species was revisited
in the 1990s by Giamello et al.,120 who investigated two
different surfaces of MgO: one bare and the other, called
electron-rich, obtained by ionization of alkali metals on MgO
and containing surface-trapped electrons. Upon interaction
with CO the latter surface gives an EPR signal which is the
same as the one slowly developing on thermally activated
bare MgO. Spectra obtained using 13CO confirmed the
presence of two magnetically equivalent carbon atoms in the
species. On the basis of both theoretical calculations121 and
low-temperature experiments,120 it was proposed to reassign
the EPR spectra to the dimeric C2O2- ethylenedione radical
ion originally proposed by Klabunde’s116 and Cordischi’s117
groups.
Contacting the electron-rich solid with CO at 77 K and
carefully raising the temperature, the early steps of CO
reduction were followed. Using 13C- and 17O-enriched CO,
it was possible to assign the spectra obtained at 100 K to
the monomer radical anion CO- and to evaluate the spin
density distribution on this previously unreported species.
The g tensor measured on the simpler 12CO spectrum showed
two components to be close to the free spin value and a third
one at g = 1.98 which is the unambiguous fingerprint (section
8.1) of an 11-electron π radical, such as NO.
The new species was thus assigned to the CO- radical
anion (Figure 18). Actually, various similar species were
detected in the experiment, the electron transfer for none of
which is complete, the total spin density on the molecule (C
+ O) ranging from 0.46 to 0.62. The observed monomer
CO radicals are more correctly described by the general
Chiesa et al.
Figure 18. EPR spectra observed after adsorption of CO at 78 K
on surface electron-rich MgO and heating at increasing temperature.
Spectrum a was observed after heating at about 100 K and is due
to the CO- radical anion. All spectra were recorded, after a
temperature increase, at 77 K. Adapted with permission from ref
120. Copyright 1993 Royal Society of Chemistry.
formula COδ-. As confirmed by theoretical calculations122
the CO- radical anion is unstable and reacts further with
CO to give the more stable dimer ethylenedione anion radical
(C2O2-). The reactions observed120 can be described as
follows:
- -
CO(g) + e(surf) f CO(ads) (about 100 K) (21)
- -
CO(ads) + CO(g) f C2O2(ads) (120-150 K) (22)
Some ethylenedione anions are stabilized at the surface,
where they are observed by EPR, while others react with
electrons to give the diamagnetic C2O22- ion, which, in turn,
undergoes a series of condensation reactions, producing
cyclic and linear oxycarbon dianions (CnOn2-) which have
been observed by vibrational spectroscopy.123,124 Polymeric
radical species derived from CO, and similar to those
described here for MgO, have been reported on scandium,
yttrium, and lanthanum oxides125 and on zirconium diox-
ide.126
7.3. Tetraatomic and Pentaatomic
Carbon-Containing Radicals
A number of studies have reported results on C-containing
biological calcified tissues and synthetic apatites. Both bulk and
surface-stabilized radicals generated by high-energy irradiation
were identified.127,128 Surface radicals are formed by interaction
of the solid with atmospheric agents. There are two slightly
different CO2- radical anions, characterized by large 13C
hyperfine constants not very different from those reported in
section 7.1, and a CO3- radical anion which, differently from
CO2-, has a rather small 13C constant associated with the central
atom (A1 is unresolved, A2 ) A3 ) 13 G). This species has in
fact a ground electronic state of A2′ symmetry with the unpaired
electron in an antibonding orbital built up by three oxygen p
orbitals. A fourth radical formed by exposure of the solid to
the atmosphere and characterized by g1 ) 2.0030, g2 ) 2.0015,
Surface-Localized Inorganic Radicals and Radical Ions
and g3 ) 2.0023 was tentatively assigned127 to CO-. This is
however in contrast with the expected structure of the g
tensor of this radical species, which has been previously
discussed in this section.
An interesting pentaatomic carbon radical ion, isolated at
the surface of various systems, is composed by one carbon
and four oxygen atoms and bears one negative charge. This
radical ion is hereafter referred to as CO4-. However, in the
first report on this species, formed by oxygen addition at 77
K to a CO2- radical adsorbed on MgO, the authors described
the species as the association of CO2 to a superoxide ion
(O2-).129 A later report130 described the species as a peroxy
radical, because of its motional behavior and the gzz value
of 2.040, typical of true peroxy radicals. The hyperfine
coupling constants of 13C (very small for all components as
in the case of CO3-) and 17O derived from 17O-enriched
dioxygen (very asymmetric distribution of spin density) are
in good agreement with the peroxy nature of CO4-. Carbon
is surrounded by three oxygen atoms (small 13C hyperfine),
one of which is bound to the fourth oxygen to form the
peroxy group. The same radical species was observed, with
similar parameters, by Kazanskii and co-workers131 on both
Cr/SiO2 and Mo/SiO2 catalysts.
The mechanochemical approach has been successfully em-
ployed for the activation of molecules and for catalytic reactions.
The mechanical energy is provided to a solid or a gas solid
system via milling in a given atmosphere, leading to some
radical intermediates. The most interesting case is that of
crystalline SiO2 (quartz). In the early 1970s Hochstrasser and
Antonini showed that, upon crushing quartz, radical defect
centers were formed on the cleaved surfaces which they
identified as Es′ centers or Si•, dangling bonds on the Si atom
bearing the unpaired electron.132 This defect and other defects
produced by irradiation in both crystalline and amorphous
SiO2, were widely investigated by Griscom.133,134 Es′ centers
are stable under ultrahigh vacuum and can react, via partial
charge transfer, with CO2, forming a Si+CO2- complex,
characterized by an hf coupling constant of A(13C) ) 217
G. The mechanochemistry of quartz surfaces was much
investigated by Radtzig, who proposed the existence of,
besides the Es′ center of SiO•, an EPR-silent counterpart
formed upon cleavage of a Si-O-Si.135 This could explain
the formation of a SiCO2• surface center upon adsorbing CO
on the surface of quartz crushed under helium:136,137
SiO•(surf) + CO(g) f SiCO2•(surf) (23)
In eq 23, we have used the author’s covalent vision of the
chemical bond in SiO2, but there is no doubt that the SiCO2•
center and the Si+CO2- reported by Hochstrasser and
Antonini132 are basically the same, as confirmed by the
similar 13C hf coupling constants (236 G obtained from a
badly resolved structure for the latter case, compared to 217
G for the former). Addition of oxygen at 77 K on the SiCO2•
radical center leads to the SiCO4• species, which is the
covalent form of the species discussed above as CO4-. This
species is unstable and can be thermally decomposed,
producing CO2 in the gas phase. The overall reaction is the
oxidation of CO to CO2 at the surface of quartz performed
via mechanochemical activation.
Table 3 gathers the main features of the radicals discussed
in this section.
Chemical Reviews, 2010, Vol. 110, No. 3 1337
8. Nitrogen-Containing Radicals
The chemistry of nitrogen inorganic compounds, being
intimately connected to biogeochemical nitrogen cycles,138
atmospheric and environmental processes, large-scale indus-
trial processes,139 and anthropogenic pollutants abatement,140
is a cornerstone of chemical sciences. It is not therefore
surprising that nitrogen chemical species are of great
importance also in interfacial chemistry in relation to the
molecular activation of dinitrogen, often by heterogeneous
catalysis, and to atmospheric pollution control. We focus here
on surface radicals generated along processes involving
essentially dinitrogen as well as nitrogen oxides. Because
N2- and NO are important radicals in this context, it is
convenient to preliminarily introduce the electronic and g
tensor structures of these 11-electron radicals in some detail.
8.1. g Tensor of 11-Electron Π Radicals
Eleven-electron radicals, which include CO- discussed in
the previous section, are characterized by the presence of
an unpaired electron in the π antibonding orbitals as shown
in Figure 14 for NO.
The degeneracy of the two antibonding orbitals is lifted
by the spin-orbit coupling and by the asymmetric electric
field, such as that exerted by a surface cation. As discussed
in section 5, this quenching effect makes EPR-active a
species otherwise hardly visible. The electronic configuration
leads to an orthorhombic g tensor whose elements were
derived by Brailsford et al. as follows:141
gzz ) ge - 2λ sin 2R (24a)
gyy ) ge cos 2R + (λ/E)(1 + cos 2R + sin 2R)
(24b)
gxx ) ge cos 2R + (λ/E)(cos 2R - sin 2R - 1)
(24c)
where λ is the spin-orbit coupling constant, E and ∆ are
the energy differences shown in Figure 14a, and tan 2R is
defined as λ/∆. It follows from the above equations that gyy
> gxx > gzz, and an important shift from ge is expected for
one component (gzz) as indeed experimentally observed
(Table 4). The previous equations become first to order:
gzz ) ge - 2λ/∆ (25a)
gyy ) ge + 2λ/E (25b)
gxx ) ge (25c)
To illustrate the influence of the g and A tensors on the
EPR line shape, Figure 19 gives the simulated spectra for
three hypothetical cases related to radicals with the same g
tensor and zero, one, or two equivalent nuclei with I ) 1.
The hyperfine coupling constant values (Ayy > Axx, Azz) are
similar to those found for radicals described in the following.
One nitrogen nucleus gives a triplet of hf lines with 1:1:1
relative intensities for all directions of the g tensor, while
two equivalent nuclei produce a quintuplet of hf lines with
1:2:3:2:1 relative intensities.
1338 Chemical Reviews, 2010, Vol. 110, No. 3 Chiesa et al.

Table 4. Spin Hamiltonian Parameters of Nitrogen-Containing Surface Paramagnetic Species

radical g tensor A tensor(14N)/G A tensor(15N)/G

species solid g1 g2 g3 A1 A2 A3 A1 A2 A3 refs
N2- MgO 2.0018 2.0042 1.9719 2.90 21.50 4.20 4.06 30.10 5.88 147-149
N2- CaO 2.0006 2.0048 1.9677 2.50 20.00 4.30 3.50 28.00 6.02 147-149

NO MgO 1.9948 1.9980 1.8900 32.0 2.57 8.96 44.8 3.60 12.4 154, 164, 165
ZnO, TiO2, ZnS, Al2O3, 154-163
SnO2, CeO2, Y zeolitesa

NO2- TiO2 2.0054 2.0036 2.0030 2.3 4.4 32.3 176

NO22- MgO 2.010 2.0044 2.0078 unres 40.6 unres unres 56.8 unres 164
CaO 2.008 2.0075 2.0080 0.5 33.0 0.5 174

NO2 Na/mordenite 2.0051 1.9910 2.0017 51.0 48.0 66.0 194

- b
N2O2 TiO2/PVG 2.0034 2.0029 2.0029 25.5 19.0 19.0 193

NO32- CaO 2.0057 2.0055 2.0055 61.7 41.0 41.0 38

a
All spectra have a structure similar to that of the spectrum reported in detail for MgO. b Hyperfine values are related to the external N atom.
The central atom has A < 2G.

8.2. Diatomic Species: Radical Anion of
Dinitrogen
The nitrogen molecule is extremely stable, and its chemical
activation is a challenging task for both inorganic142,143 and
bioinorganic144 chemists. A recent achievement in this field
is the reversible reduction of dinitrogen to form the N2-
radical anion on MgO and on CaO containing surface-trapped
electrons. Besides interesting optical and magnetic proper-
ties,145,146 these systems show a tremendous reducing capac-
ity. The EPR spectrum of the N2- radical ion is observed
after nitrogen adsorption at low temperature (T < 120 K) on
the oxide electron-rich surface147 and exhibits a hyperfine
structure (Figure 20) using either 14N2 or 15N2 molecules.
The features of the spectrum were computer simulated and
assigned to a biatomic 11-electron 2Π1/2 species with two
magnetic and structurally equivalent nitrogen nuclei, namely,
the N2- radical ion,148,149 adsorbed on a positive ion at the
oxide surface and formed as follows:
- -
N2(g) + e(surf) h N2(ads)
Analysis of the hyperfine tensor148 indicates that the
electron-nucleus interaction is essentially dipolar, the un-
paired electron being mainly confined in the πy antibonding
orbital with some admixture of the πx orbital. The spin
density calculated with the classic procedure of comparison
between the experimental and the atomic values is about 0.92
for N2- on MgO and 0.89 in the case of CaO. The charge
transfer degree from the surface to the molecule is thus very
high, and the interaction is essentially ionic. It is indeed the
large stabilizing contribution of the ionic interaction between
a surface cation and the adsorbed radical anion to the
chemical bonding that explains the electron transfer onto the
N2 molecule. The ionization energy of the trapped electron
and the electron affinity of the molecule are in fact both
(26)

The adsorption is reversible and, by lowering the gas pressure

or raising the temperature, molecular nitrogen is desorbed
and the electron trapped back by the surface.

Figure 19. Simulated powder EPR spectra of 11-electron Π Figure 20. Experimental (77 K) and computer-simulated EPR
radicals containing respectively zero (a), one (b), and two (c) spectra of 14N2- (a, b) and 15N2- (c, d) on MgO. Adapted from ref
equivalent nuclei with I ) 1. 148. Copyright 2001 American Chemical Society.
Surface-Localized Inorganic Radicals and Radical Ions
endothermic. The reason for the reversibility of the process
is more subtle. Calculations at the DFT level148,150 show that
the formation of the charge transfer complex is slightly
exothermic (fraction of an electronvolt according to the type
of adsorption site)151 and that there is a low activation energy
at the crossing between the unbound (e-(surf)/N2) and the
bound (N2-(surf)) states. The calculations also indicate that
the adsorbed N2- radical ion is elongated by about 8% with
respect to the gas-phase molecule, consistent with the
occupation of an antibonding orbital by the unpaired electron
and with the view of the described process in terms of a
true activation of the nitrogen molecule.
8.3. Diatomic Species: Nitric Oxide, a
Multipurpose Surface Probe
Nitrogen oxides (NOx) are ubiquitous byproducts of high-
temperature combustions and very noxius pollutants of the
atmosphere. Their removal from burnt gases is required in
stationary installation (power plants) and in mobile sources
(automotive vehicles). In the former case, nitrogen oxide
abatement is achieved either by selective catalytic reduction
(SCR) using various reducing agents (NH3, hydrocarbons,
or alcohols) or by direct decomposition of NO into N2 and
O2. This latter reaction is extremely difficult due to the
enormous kinetic limitations, and only recently it was
discovered by Iwamoto and co-workers152 that Cu/ZSM5
zeolites are active catalysts for this process. The interaction
of NO with surfaces has therefore become an important
field of research and has been investigated by means of a
plethora of experimental techniques.
There is another reason that explains the interest in NO:
because of its radical nature, this molecule is also often used
to probe the properties of oxide surfaces and its dynamics,
particularly in complex environments such as micro- and
mesoporous systems. In what follows, we first describe the
interaction of NO with cationic sites (section 8.3.1) as
followed by EPR, then we discuss the less common case of
NO as a probe of basic sites (section 8.3.2), and in section
8.3.3 we show how NO has been used to monitor the
properties of certain transition-metal ions (TMIs).
8.3.1. Cationic Sites
Although NO is a most stable radical, it has remained
unidentified by EPR for many years. In the gas phase the
molecule has a complete coupling between the electron spin
S and the angular momentum L, leading to a 2P1/2 ground
state which is not split by a magnetic field (the components
of S and L are antiparallel so that gJ ) 0).153 The EPR
spectrum recorded for gas-phase NO is due to the partially
populated high-energy 2P3/2 state and is observed at the
resonant field corresponding to gJ ) 4/5 (0.77). When NO
is in an adsorbed state, two effects allow the observation of
an EPR spectrum. The former one is the quenching of the
orbital momentum by the electric field of the adsorption site,
which causes the 2P1/2 state to become a paramagnetic, pure
spin state. The second is the lifting of the degeneracy of the
two π* orbitals, producing an energy splitting ∆ which
determines (Figure 14, eqs 25a-25c) the g values, close to
2, of adsorbed NO.
The first EPR spectrum, dealing with MgO and recorded
at 93 K by Lunsford,154 was assigned to “an adsorbed NO
molecule in a strong field at the surface such that the
spin-orbit coupling is partially quenched”. The signal
Chemical Reviews, 2010, Vol. 110, No. 3 1339
Figure 21. Experimental and computer-simulated EPR spectra of
adsorbed 14NO (a, b) and 15NO (c, d) on MgO. Spectra were
recorded at 77 K. Adapted from ref 165. Copyright 2002 American
Chemical Society.
exhibits a structure expected for an 11-electron Π radical
with gxx ) gyy ) 1.995, gzz ) 1.90, Axx ) 30 G, and Azz < 10
G. The surface-molecule bond however is rather weak, and
NO desorbs, outgassing the sample at room temperature. The
NO molecule is thought to weakly interact with the polarizing
field of a surface Mg2+ cation. A number of papers followed,
reporting spectra of NO adsorbed on cationic sites of ZnO
and ZnS,155 Y156-158 and A159,160 zeolites, Al2O3,161 TiO2 and
SnO2,162 and CeO2.163 The energy splitting parameter ∆ was
calculated via eq 25a for various systems, and the values
obtained show that the gzz component is indeed very sensitive
to the charge of the positive adsorption site, substantiating
its use as a probe of cationic sites.164
Recently the interaction of NO with MgO was revisited
employing highly crystalline nanostructured materials pre-
pared by chemical vapor deposition (CVD).165 The EPR
spectrum, better resolved than in earlier studies,154 exhibits
three independent lines at high field (Figure 21a) corre-
sponding to the gzz values of NO radicals having very similar
gxx and gyy and adsorbed on three distinct Mg2+ cations which,
because of their different coordination states, exert different
polarizing powers on the molecule. The lower the coordina-
tion, the higher the strength of the electric field exerted by
the cation.
This result outlines the remarkable property of NO to
discriminate surface cationic sites. In the spectrum recorded
using 15NO (I(15N) ) 1/2) instead of 14NO (Figure 21c) each
hyperfine triplet is substituted by a doublet (see, for instance,
the component at g ) 1.8900). Analysis of the hyperfine
structure (Table 4) indicates that the total spin density on
the N atom of the best resolved species is about 0.803. This
remarkable localization of the spin density on the N atom is
not surprising, if the prevalent nitrogen character of the
antibonding orbitals in the heteronuclear NO molecule is
taken into account. The spin density on the adsorbed
molecule is very close to that of free NO, confirming that
the perturbation of the molecule associated with the interac-
tion with the surface is rather weak. The nature of the
“surface-NO” interaction originally proposed by Lunsford
was confirmed by theoretical calculations. The NO molecule
interacts with low coordinated Mg2+ cations at edges and
corners. The interaction energies, in agreement with the
1340 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 22. (a) EPR spectrum of the NO22- radical ion formed on MgO. (b) Computed structures of NO chemisorbed at O2- anions at
different surface sites. Adapted from ref 165. Copyright 2002 American Chemical Society.
experimental results, are weak (0.10-0.20 eV) but increase
with decreasing surface coordination of Mg2+. In this bonding
mode the molecule is virtually unperturbed with respect to
the free molecule. On five-coordinated ions at the surface
the interaction energy of a single molecule is negligible, and
physisorbed diamagnetic (NO)2 pairs are preferentially
formed.
A further application of the properties of NO as a surface
probe was found by investigating the dissociative chemi-
sorption of H2 on MgO. It was found that two out of the
three described NO sites are involved in H2 dissociation. As
a matter of fact, they no longer lead to EPR spectra typical
of adsorbed NO radicals upon contact of NO with a surface
precovered by hydrogen.
Pulsed techniques and high-field EPR have led to new
investigations on NO complexes in zeolites aiming to
describe the nature of Lewis acid sites. Though the general
description of the NO weak interaction on Lewis sites is
confirmed, the use of the W-band,166 ENDOR,167 and pulsed
ENDOR168 on Na-containing ZSM-5 and A zeolites allowed
the determination of previously unknown structural param-
eters for the Na+NO and Al3+NO complexes. The NO adduct
on Na+ ions, for instance, has a bent structure with a
Na-N-O angle of 142°. Desorption of the NO complex
has also been investigated in terms of the equilibrium
between the adsorbed molecule and the gaseous one. The
latter was monitored via the 2P3/2 state resonating at g )
0.77.169
NO coordination can lead also to dinitrosyl complexes
especially in the case of transition-metal ion compounds. In
some cases the presence of NO diadducts with S ) 1 (triplet
state) has been observed on surfaces such as in the case of
zirconium dioxide170 or in that of the widely studied A
zeolites.157,171,172
8.3.2. Anionic Sites
The cationic acidic sites described above are not the only
ones involved in NO adsorption; basic sites also are, as
reported by Lunsford.154 This finding was described in the
paper which first illustrated the NO interaction with MgO,
and it was revisited recently.165,173 NO can be strongly
chemisorbed on particularly basic O2- sites of MgO, i.e.,
Chiesa et al.
via covalent bonding and formation of the radical anion
NO22- as follows:
NO(g) + O(surf)
2- 2-
f NO2(surf) (27)
The above is an example of covalent bonding between a
molecule and a solid surface (Figure 5b of section 4.2).
Though a notation such as that adopted here (NO22-) is
compact and easy to use, one has to realize that one of
the two oxygens belongs to the solid and that, in principle,
the spin density and the charge are not restricted to the sole
triatomic radical. NO22- has 19 electrons and the same
electronic structure as that of the ozonide ion (O3-). Its EPR
spectrum (Figure 22a) is characterized by three g values
rather close to one another (Table 4) and a nitrogen hyperfine
coupling of about 40 G centered on the smaller g component.
The main contribution to the hf interaction is dipolar because
the odd electron is in an antibonding orbital formed from
three parallel p orbitals centered on the three atoms. The
non-negligible contribution of the 2s nitrogen orbital is
essentially due to spin polarization. Theoretical calculations165
have shown that the two N-O bonds are not equivalent but
both have the features typical of covalent bonding (Figure
22b). In the case of MgO, the formation of NO22- concerns
very few low coordinated O2- ions, while the large majority
of NO radicals are formed on surface cations, as described
earlier.
The situation is completely reversed in the case of CaO,174
which, being more basic than MgO, interacts with NO with
its basic oxide ions only and not with its cations. This
interaction, however, is rather complex, and four distinct
radical species, all formed according to eq 27, have been
observed showing different stability and progressive variation
of the spin density on the nitrogen atom.174
The surface O2- ions involved represent around 0.5% of
the total number of surface sites and are the strongest basic
sites, i.e., those with low coordination number (four- and
three-coordinated O2- ions at edges, corners, and other
morphological defects). The more abundant five-coordinated
sites present at the planar (100) faces of the cubic oxide are
not concerned. The concentration of such basic sites is 25
times higher on CaO than on the less basic MgO.
Well-resolved X- and Q-band EPR spectra assigned to
NO22- were also reported for TiO2 upon calcination in air
(300-450 °C) of samples prepared by addition of TiCl4 to
water solutions containing either sodium hydroxide or
ammonia followed by filtration and drying.175 The formation
Surface-Localized Inorganic Radicals and Radical Ions
Figure 23. Three-dimensional representation of relevant MO
orbitals of NO/Ni2+MgO: (a) HOMO, (b) SOMO, (c) lowest
unoccupied molecular orbital (LUMO). Reprinted with permission
from ref 178. Copyright 2003 Elsevier.
of NO22- was attributed to the reaction of NO, formed by
oxidation of ammonia (4NH3 + 5O2 f 4NO + 6H2O) with
oxygen ions according to the reaction in eq 27. NO22- ions
are located in the bulk because their EPR spectra were not
broadened by excess gas-phase oxygen. This explanation was
substantiated by the fact that the thermal treatment under
Vacuum (instead of calcinations) did not generate the spectra
assigned to NO22-. The results show that the samples keep
the memory of the ammonia preparation route leading to
N-doped TiO2.
The same system has been revisited to prepare such
samples used as photocatalysts, and the spectra were as-
signed176 to a nitrogen species in interstitial sites of the bulk
of TiO2 which binds to an oxygen ion to form a NO2- center
with EPR parameters close to those of NO22-. This result
was corroborated by ab initio calculations which proved the
consistency of the experimental parameters with those
expected for a N interstitial species but also suggested that
reaction 27, occurring at the surface of alkaline-earth-metal
oxides, cannot take place in the bulk of TiO2, whose O2-
ions are not basic enough to bind NO. This example
illustrates the importance of theoretical modeling in assigning
EPR spectra.
8.3.3. NO as a Probe of Surface Transition-Metal Ions
The weak bonding of inorganic radicals described in
section 8.3.1 with surface cations becomes more involved
when the latter are TMIs. The interaction between NO and
TMIs leads to a family of adducts called nitrosyls.177 Surface
analogues of molecular nitrosyls are frequently observed at
oxide surfaces or in zeolites in relation to catalysis work
aiming at eliminating nitrogen oxides from various pollutants
emission. Review papers on this subject are available in the
literature.140
The formation of a nitrosyl adduct with a TMI (in most
cases two open-shell systems) often involves spin pairing
with important variation of the electronic structure of the
system. An example of such behavior is found in Ni2+-doped
MgO.178 The EPR spectrum of the solid, dominated by the
broad signal of bulk Ni2+ ions in octahedral coordination
(3d8, high spin), is modified by adsorption of NO. Besides
the already discussed NO-Mg2+ signal, a novel axial signal
appears (g| ) 2.274 and g⊥ ) 2.131, values typical of 3d9
Ni+ ions) related to the NO-Ni2+ interaction occurring with
the ions located at the planar (100) face of the cubic MgO
crystals (C4V symmetry).
The chemical bonding is determined by the interaction of
the dz2 Ni2+ orbital with the π*xz orbital of NO (both singly
occupied), which form a doubly occupied MO, the HOMO
(Figure 23a). One unpaired electron is left in the dx2-y2 orbital,
Chemical Reviews, 2010, Vol. 110, No. 3 1341
Figure 24. X-band, Q-band, and W-band EPR spectra of the NO/
Cu+ZSM-5 system. The magnetic axes of the complex are also
shown. Adapted with permission from ref 191. Copyright 2005
American Chemical Society.
which becomes the SOMO and is responsible for the axial
EPR signal. The Ni2+ ion has therefore a formal d9
configuration with one hole in the d shell, which explains
the nature of the observed EPR spectrum even though one
of the d electrons is shared with the NO molecule and no
effective electron transfer from the adsorbate to the solid
has occurred. A similar interaction has been monitored in
Fe2+-containing zeolites which form different nitrosyls
according to the NO pressure.179,180
In other cases NO-TMI interaction occurs without spin
pairing and is therefore more similar to the weak adsorption
described in section 4.2 (Figure 5d) and section 8.3.1 as in
the case of NO on Cu+ ions in various zeolites181-183 or on
silica.184 Cu+ (3d10) is a closed-shell diamagnetic ion. The
adsorption of NO gives the EPR-visible Cu+NO complex,
which is particularly important because it is a crucial
intermediate in the decomposition of NO into N2 and O2
catalyzed by Cu+/ZSM5 zeolites.152,185 The presence of the
copper nucleus having nuclear spin I ) 3/2 leads to a NO
spectrum with a superhyperfine structure in addition to the
hyperfine one due to N with I ) 1. The analysis of the
complex X-band EPR spectrum was performed by Sojka et
al.186 on the basis of simulation of the spectrum. The
determination of the spin Hamiltonian parameters is par-
ticularly difficult because the axes of the g and A tensors do
not coincide due to the low symmetry (Cs) of the complex
(Table 2). The Cu+NO complex has an η1 bent structure
(Figure 24) with the unpaired electron residing mainly on
the coordinated NO moiety. The large copper hf constant is
actually due to the partial delocalization of this electron on
the Cu 4s orbital. The total spin density on copper is found
to be equal to 0.2 (0.8 on NO) and is shared among 4s (0.1),
3dz2 (0.08), and 3dxz (0.02) orbitals. Analysis of the EPR
parameters allowed a thorough description of the bonding
interaction, which essentially involves the overlap of the two
π* orbitals of the NO molecule with the copper d orbitals
of suitable symmetry.186 This picture is fully supported by
multifrequency EPR studies187-189 (Figure 24), theoretical
calculations,190 and ENDOR191 experiments.
1342 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 25. EPR spectrum of the N2O2- radical ion obtained with
14 15
N NO on TiO2 dispersed on a porous glass. Reprinted from ref
193. Copyright 1985 American Chemical Society.
8.4. Nitrogen-Containing Triatomic Surface
Radicals
Triatomic radical ions derived from nitrous oxide (N2O),
namely, N2O+ and N2O-, have been in some cases
proposed,192,193 but their existence has never been unambigu-
ously demonstrated.
Nitrogen dioxide, NO2, is a stable radical whose EPR
spectrum in solid matrixes has been widely investigated. The
bent 17-electron NO2 molecule, isoelectronic of CO2-, has
a large nitrogen isotropic hyperfine constant (aiso) due to the
contribution of the N(2s) orbital to the SOMO (see section
6.3). Nitrogen dioxide is very reactive, and its use as a surface
probe is limited to the case of relatively inert oxides. In such
a case, NO2 is used essentially as a probe for mobility
investigation on oxide surfaces (section 5.2) and in the
channels of zeolites194 because it does not possess the
sensitivity of NO to the surface crystal field. In many other
cases, the NO2 molecule reacts with oxide surfaces, forming
diamagnetic species (nitrates, nitrites, etc.), and in a few cases
only, such as that of MgO,195 a fraction of the molecules are
adsorbed nearly unperturbed on less reactive sites.
NO2 was employed in the first EPR investigation of a
submonolayer adsorbate on a single-crystal surface88 pub-
lished in 1995.
8.5. Nitrogen-Containing Tetraatomic and
Pentaatomic Surface Radicals
Two tetraatomic N-containing radical ions adsorbed on
oxide surfaces have been reported, with 23 and 25 electrons,
respectively. The former one is the N2O2- radical ion
observed after UV irradiation at 77 K of titanium dioxide
dispersed on a porous glass in a N2O atmosphere71 (Figure
25).
The N2O2- radical ion forms by reaction of N2O with a
photogenerated hole (eq 28) localized at the surface (section
4.4). Surprisingly, the EPR signal of N2O2- (Figure 25)
exhibits a hyperfine structure suggesting the presence of one
nitrogen atom only (Table 4). However, the spectra obtained
using either 14N15NO or 14N14NO are very similar to one
another, indicating that the spin density on the central N atom
is negligible. The addition of N2O on the O- ion takes place
at the central nitrogen atom of the molecule.
-
N2O(g) + O(surf) -
T N2O2(surf) (28)
Chiesa et al.
Figure 26. SOMO orbitals of N2O2- (a) and NO32- (b) radical
ions in the gas phase.
The N2O2- radical anion (Figure 26a), an analogue of the
more common CO3-, is expected to be planar, the unpaired
electron being confined in the π antibonding orbital formed
by the three outermost atoms (1N and 2O) around the central
nitrogen. The spin density on this latter one is very small as
indicated also by the SOMO orbital reported in Figure 26a,
which, though neglecting the perturbation from the solid, is
consistent with the isotopic labeling result mentioned above.
While the observed hyperfine structure is in line with the
assignment, the g tensor of the radical should have a
component greater than ge in the direction perpendicular to
the molecular plane. This seems not to be the case, making
the assignment open to discussion.
The formation of the 25-electron, tetratomic NO32- radical
ion, observed in the early steps of the complex interaction
of NO with the surface of activated CaO,38 is believed to
involve peroxy groups (or at least O- pairs) capable of
binding a single NO molecule following the reaction
NO(g) + O2(surf)
2- 2-
T NO3(surf) (29)
The species is rather unstable, and its spectrum disappears
after an increase of the NO pressure due to the formation of
nitrate groups. The EPR spectrum of NO32- is characterized
by axial g and A tensors with rather high nitrogen hyperfine
coupling in both directions (Table 4) similar to those
observed for the same species in irradiated KCl or KNO3.
This tetraatomic radical ion is pyramidal (section 6.3), the
25th electron being in an antibonding orbital (Figure 26)
which acquires some s character and becomes less antibond-
ing with progressive bending of the structure.
The last case considered in this section is that of a 25-
electron, pentaatomic radical ion containing both carbon and
nitrogen first assigned196 to CNO22- and finally identified as
CNO32-.197
In environmental catalysis, CO and NO components found
in automotive exhaust gases are made to react to form N2
and CO2. The important feature of the CNO32- radical anion
is that it is formed by coadsorption of these two gases on
MgO and can thus be regarded as a reaction intermediate.
Once again, the basic O2- sites of MgO play a role in binding
one of the molecules (CO), forming an initial surface
complex (CO22-) capable of NO addition, as follows:
NO(g) + CO(g) + O(surf)
2- 2-
T CNO3(surf) (30)
The assignment of this radical ion was done on the basis of
both a systematic isotopic substitution (spectra with 12C, 13C,
14
N, and 15N were obtained) and comparison with the infrared
spectra of the same radical.196 The structure of CNO32- is
Surface-Localized Inorganic Radicals and Radical Ions
Figure 27. EPR spectrum (77 K) and gas-phase SOMO orbital of
the CNO32- radical anion adsorbed on MgO.
similar to that of CO4- (section 7.3) with a carboxy group
in contact with the surface and a NO group on top of it with
a bent geometry198 (Figure 27). The spin density is mainly
confined on the three terminal (CNO) atoms with a smaller
contribution of the two carboxylic oxygens.
9. Sulfur- and Chlorine-Containing Radicals
The chemistry of sulfur compounds adsorbed on oxide
surfaces plays an important role in hydrodesulfurization
catalytic reactions. Significant information on the chemical
nature of sulfur-containing radical intermediates and on their
structure have been derived, particularly in the few studies
performed using molecules enriched in the 33S isotope (I )
3/2). The natural abundance of this isotope (0.76%) is not
sufficient to give detectable hyperfine structures using
nonenriched materials.
The only report available concerning the S2- dimer radical
(an analogue of O2-) is due to Kazansky and co-workers,199
who observed well-resolved X- and K-band EPR spectra
upon treatment of MoO3/SiO2 and MoO3/Al2O3 catalysts with
33
S-enriched H2S at 573 K. The g tensor of the species is
rhombic with g1 ) 2.048, g2 ) 2.029, and g3 ) 1.998. The
hyperfine structure is due to the overlap of the quartet related
to the 33S-32S- isotopomer with the seven-line structure of
33
S-33S-. The assignment of the species is mainly done on
the basis of the g tensor, while the electron spin density
distribution was not discussed.
Triatomic negative sulfur and sulfur-oxygen monoanions
were observed at the surface of MgO by Lunsford and co-
workers for all the possible compositions (SO2-, S2O-, S3-).
All these radical ions bear 19 valence electrons. In spite of
their similarity, they were obtained by different chemical
routes.
SO2- can be formed by electron transfer from trapped
electrons to SO2 on MgO.200 The radical ion (previously
investigated in γ-irradiated solid dithionite201 or generated
by electrochemical methods202) was investigated in its
adsorbed form using 33S- and 17O-enriched reactants. This
has allowed the electron density distribution and the radical
structure to be derived. The unpaired electron occupies a
2b1′′ molecular orbital formed by the three parallel p orbitals
of the sulfur and of the two oxygen atoms. The radical ion
is similar to the ozonide anion (section 4.2, Figure 7), but
Chemical Reviews, 2010, Vol. 110, No. 3 1343
Figure 28. X-band EPR spectrum of S3- on MgO at different
enrichment levels of 33S. Reprinted with permission from ref 206.
Copyright 1973 American Institute of Physics.
due to the different energy levels of the involved orbitals
(3p and 2p for S and O, respectively), the electron spin
density is mainly concentrated on the S atom (which bears
from 0.71 to 0.75 of the total spin density in the case of the
two main species observed). An earlier observation of the
SO2- adsorbed radical is due to Kazanskii203 and co-workers
to explain the EPR spectrum obtained by adsorption of SO2
on prereduced TiO2. In another case SO2- was produced by
electron transfer from adsorbed superoxide ions (O2-) to SO2
in the channels of Y zeolites.204
The S2O- radical anion was observed after surface reaction
of H2S with SO2. The assignment of the rhombic spectrum,
very similar to that of SO2-, was made using, in distinct
experiments, 33S-enriched H2S and SO2. The two sets of
hyperfine lines indicate the presence of two nonequivalent
S atoms in the species. The chemical process is based on
the reaction of H2S and SO2 producing water and S2. The
diatomic species then reacts with a surface oxygen ion,
yielding S2O- and releasing an electron:205
4H2S(ads) + 2SO2(ads) f 4H2O(ads) + 3S2(ads)
(31)
-
S2(ads) + O(surf)
2-
f S2O(surf) + e- (32)
The S3- radical ion206 originates by reaction of elemental
sulfur with partially hydroxylated MgO at 400 K. A sharp
axial spectrum is observed using nonenriched sulfur whose
principal g values are g⊥ ) 2.043 and g| ) 2.004. Around
these features, a complex hyperfine structure (with primary,
secondary, and tertiary splittings) can be analyzed on the
basis of two distinct coupling constants (Figure 28). The 2:1
ratio for the two quartets indicates that there are twice as
1344 Chemical Reviews, 2010, Vol. 110, No. 3 Chiesa et al.

Table 5. Molecular Radical Species Adsorbed at Solid Surfaces

ground state of
the free radical
no. of no. of formula of or symmetry of
atoms electrons the species structure the SOMO notes
2 11 CO- 2
Π1/2 Quenched diatomic molecules
2
2 11 NO Π1/2 crystal field sensitive
2 11 N2- 2
Π1/2

2 13 O2- 2
Π3/2 Quenched diatomic molecules
2 13 NO2- 2
Π3/2 crystal field sensitive
2
2 13 ClO Π3/2

2 15 Cl2- 2
Σ1/2 Crystal field insensitive

3 17 CO2- bent, C2V symmetry 2

A1 Contribution of 2s to the SOMO. Isotropic hyperfine expected.
Crystal field insensitive
2
3 17 NO2 A1
- 2
3 19 O3 bent B1 Π system, weak isotropic coupling by spin polarization.
3 19 SO2- bent 2
B1 Crystal field insensitive
3 19 NO22- bent 2
B1
2
3 19 ClO2 bent B1
2
3 19 ClOO bent B1 Peroxy-type species

4 21 C2O2- linear 2
B1
4 23 N2O22- planar 2
A′2 Π system
4 23 CO3- planar D3h 2
A′2 Small spin density on the central atom.
4 25 NO32- pyramidal 2
A1 Isotropic and anisotropic couplings from the central atom.

5 29 CO4- Pseudo-peroxy radical (O2COO)

5 29 CNO32- Pseudo-nitroxide (O2CNO)2-

many atoms of one type with respect to the other, thus
allowing an unambiguous assignment of the spectrum to a
trimeric entity, the S3- radical ion. A complex reaction was
proposed to explain the formation of the radical on the basis
of the reaction of S with surface OH groups.206
Other adsorbed sulfur-containing radicals, which include
CS2-, COS-, and H2S2-, were formed on MgO by transfer
of surface-trapped electrons to the diamagnetic parent
molecules. The first and second types of radical ions,
analogues of CO2-, were formed from either CS2 or COS.207
The third type of radical ion obtained from H2S is formed
by a complex reaction.208 The first intermediate (H2S-) in
fact further reacts with H2S to form H2S2- with hydrogen
desorption:
- -
H2S(ads) + H2S(g) f H2S2(ads) + H2(g) (33)
The unpaired electron in this unusual radical species is
thought to be localized in a σ3p* antibonding orbital between
the two sulfur atoms.
In a recent paper Livraghi et al.209 reinvestigated the
electron transfer reaction leading to CS2- on the electron-
rich MgO surface, following the process at temperatures
above those used by Lin et al.207 The authors reported a series
of spectra due to the fragmentation of carbon sulfide at the
surface. The mechanism, starting from the formation of CS2-,
continues with its cleavage and formation of CS and S-. The
latter radical ion entails an oligomerization reaction with
formation of a mixture of Sn- radical ions with n g 3 quite
similar to species already observed in frozen solutions of
sulfur.
While halogen-containing radicals play an important role
in the field of solid-state defects, very few examples are
reported of such radicals in an adsorbed state, and they
concern chlorine species in zeolites. Gardner and co-workers
reported the EPR spectra of ClO2 and of Cl2- adsorbed in
the framework of faujasite and mordenite.210 ClO2 is
paramagnetic, and its spectrum is simply observed after
adsorption in the zeolite channels. The spectrum of the 19-
electron radical molecule (section 6.3), characterized by a
rhombic g tensor and by a hyperfine coupling due to the Cl
atom (the largest component has a constant of about 75 G),
is perturbed by the inner electric fields of the zeolitic
framework which cause an increase of the Cl coupling
constant.
Cl2- radical ions are observed after Cl2 adsorption in the
same microporous solids and subsequent γ irradiation. Cl2-,
a σ radical having the unpaired electron in a σ* antibonding
orbital, has been widely investigated in irradiated alkali-metal
chlorides. Cl2- forms upon localization of an electron hole
on an anion pair (Cl--Cl-) in correspondence with a cation
vacancy.211 The complex hyperfine structure of Cl2- is due
to the presence of two equivalent Cl atoms (I ) 3/2). This
radical is an interesting example of a hyperfine pattern whose
features are determined by the presence of different isoto-
pomers. Chlorine has in fact two magnetic nuclei, 35Cl and
37
Cl, whose natural abundances are 75% and 25%, respec-
tively. Both have nuclear spin I ) 3/2 but different nuclear
magnetic moments and thus are the origin of distinct
hyperfine structures.
An interesting process involving chlorine dioxide was
reported by Raghunathan and Sur212 who elegantly showed,
by EPR, the photoisomerization of ClO2 obtained by UV
irradiation after adsorption in cancrinite. The chlorine dioxide
molecule (OClO) in these conditions transforms into ClOO,
the chloroperoxy radical. The transformation dramatically
changes the hyperfine structure of the molecule because most
spin density in ClO2 is on the central O atom. The main
chlorine coupling constant in the chloroperoxy isomer is
Surface-Localized Inorganic Radicals and Radical Ions Chemical Reviews, 2010, Vol. 110, No. 3 1345

therefore about 1 order of magnitude lower with respect to possess an unpaired electron (S ) 1/2), but CH3• is only a radical. A
limited exception to this general behavior is, however, represented
the value observed for chlorine dioxide. by those radicalic molecules showing, for electronic reasons, a certain
stability. This is the case, for instance, of the adducts formed by
10. Conclusions (14)
spin trapping of reactive radicals.
Adrian, F. J. Colloid Interface Sci. 1968, 26, 317.
The role of surface inorganic radicals and radical ions (15) Lunsford, J. H. AdV. Catal. 1972, 22, 265.
(16) Howe, R. F. Colloids Surf. 1993, 7, 353.
formed at the interface between a solid and a gas phase has (17) Che, M.; Giamello, E. In Catalyst Characterization: Physical
been discussed in the present review. The formation of such Techniques for Solid Materials; Imelik, B., Vedrine, J. C., Eds.;
species is relevant to a number of fields, including electro- Plenum Press: New York, 1994.
chemistry, heterogeneous catalysis, photocatalysis, and cor- (18) Schweiger, A. Angew. Chem., Int. Ed. 1991, 30, 265.
(19) Schweiger, A. Appl. Magn. Reson. 1993, 5, 229.
rosion phenomena. Identification and characterization of (20) Schweiger, A.; Jeschke, G. Principles of Pulse Electron Paramagnetic
surface radicals are essentially conducted using EPR tech- Resonance; Oxford University Press: Oxford, U.K., 2001.
niques which provide, on the basis of g, hyperfine, and (21) Xu, J.; Kevan, L. Appl. Magn. Reson. 2001, 20, 3.
(22) Goldfarb, D.; Bernardo, M.; Strohmaier, K. G.; Vaughan, D. E. W.;
superhyperfine tensors, a detailed understanding of the radical Thomann, H. J. Am. Chem. Soc. 1994, 117, 6344.
composition and structure and, in some cases, the nature of (23) Goldfarb D. In Electron Paramagnetic Resonance. A Practitioners
the surface adsorption sites. The first part of the review has Toolkit; Brustolon, M., Giamello, E., Eds.; Wiley: Hoboken, NJ, 2009.
been devoted to a methodological approach aiming both at (24) Mabbs, F. E.; Collison, D. Electron Paramagnetic Resonance of d
Transition Metal Compounds; Elsevier: Amsterdam, 1992.
clarifying the genesis of radical intermediates in surface (25) Nilges, M.; Freed, J. H. Chem. Phys. Lett. 1981, 82, 203.
processes and at evidencing their importance in typical (26) Futako, W.; Umeda, T.; Nishizawa, M.; Yasuda, T.; Isoya, J.;
catalytic and photocatalytic phenomena. Furthermore, the two Yamasaki, S. J. Non-Cryst. Solids 2002, 299, 575.
(27) Sterrer, M.; Fischbach, E.; Risse, T.; Freund, H. J. Phys. ReV. Lett.
experimental approaches typical of EPR (direct monitoring 2005, 94, 186101.
of paramagnetic sites and use of paramagnetic molecules to (28) Chiesa, M.; Giamello, E.; Di Valentin, C.; Pacchioni, G. Chem. Phys.
probe the properties of diamagnetic systems) have been Lett. 2005, 403, 124.
described with particular attention to the unsurpassed ability (29) The spectroscopic behavior of powders is also shown by glasses
containing paramagnetic centers.
of EPR in describing motional behaviors. The second part (30) Rist, G. H.; Hyde, J. S. J. Chem. Phys. 1970, 52, 4633.
is devoted to a systematic analysis of inorganic radicals and (31) Rieger, P. H. In Specialist Periodic Reports. Electron Spin Resonance;
radical ions reported in the literature, providing careful Royal Society of Chemistry: Cambridge, U.K., 1993; Vol. 13B, pp
178-199.
guidelines for deriving structural information from the (32) The presence of an O- ion at the oxide surface can derive from
experiments. The main types of systems discussed, classified photogeneration or, alternatively, from charge compensation of an
on the basis of the number of valence electrons, are gathered impurity (e.g., Na+ in CaO).
in Table 5 with essential information on their structure and (33) Anpo, M; Che, M; Fubini, B.; Garrone, E.; Giamello, E.; Paganini,
M. C. Top. Catal. 1999, 8, 189.
ground state. (34) Lunsford, J. H. EPR Methods in Heterogeneous Catalysis. In
Catalysis. Science and Technology; Anderson, J. R., Boudart, M.,
Eds.; Springer-Verlag: Berlin, 1987; Vol. 8, pp 227-256.
11. Acknowledgments (35) Sojka, Z.; Giamello, E.; Paganini, M. C. Radicals in CatalysissHetero-
geneous. In Encyclopedia of Catalysis; John Wiley & Sons: Chich-
The present review is dedicated to Prof. Jack H. ester, U.K., 2002.
Lunsford, a pioneer in the scientific area described here. (36) Chiesa, M.; Giamello, E.; Paganini, M. C.; Sojka, Z. J. Chem. Phys.
We gratefully acknowledge Prof. Piero Ugliengo for 2002, 116, 4266.
assistance in the computation of the SOMO orbitals. E.G. (37) Ricci, D.; Pacchioni, G.; Sushko, P. V.; Shluger, A. L. Surf. Sci.
2003, 542, 293.
thanks the Université Pierre et Marie CuriesParis 6 for (38) Paganini, M. C.; Chiesa, M.; Dolci, F; Martino, P.; Giamello, E. J.
granting a position of invited professor during the prep- Phys. Chem. B 2006, 110, 11918.
aration of this work. (39) Kon, M. Ya.; Shvets, V. A.; Kazansky, V. B. Kinet. Katal. 1973,
14, 403.
(40) Aika, K. I.; Lunsford, J. H. J. Phys. Chem. 1978, 82, 1794.
12. Note Added after ASAP Publication (41) Liu, II. F.; Liu, R. S.; Johnson, R. E.; Lunsford, J. H. J. Am. Chem.
Soc. 1984, 106, 4117.
This paper was published on the Web on December 14, (42) Bonneviot, L.; Olivier, D.; Che, M. J. Mol. Catal. 1983, 21, 415.
2009, with an error reference 13. The corrected version was (43) Driscoll, D. J.; Martir, W.; Wang, J. X.; Lunsford, J. H. J. Am. Chem.
Soc. 1985, 107, 58.
reposted on December 18, 2009. (44) Wang, J. X.; Lunsford, J. H. J. Phys. Chem. 1986, 90, 5883.
(45) Driscoll, D. J.; Campbell, K. D.; Lunsford, J. H. AdV. Catal. 1987,
13. References 35, 139.
(46) Schmidt, J.; Risse, T.; Hamann, H.; Freund, H.-J. J. Chem. Phys.
(1) Atkins, P. W.; Symons, M. R. C. The Structure of Inorganic Radical; 2002, 116, 10861.
Elsevier: Amsterdam, 1967. (47) Risse, T.; Schmidt, J.; Hamann, H.; Freund, H.-J. Angew. Chem.,
(2) Weltner, W. J. Magnetic Atoms and Molecules; Dover Publications Int. Ed. 2002, 41, 1517.
Inc.: New York, 1989. (48) Photocatalysis: Fundamentals and Applications; Serpone, N., Pel-
(3) Corma, A. Chem. ReV. 1997, 97, 2373. izzetti, E., Eds.; Wiley & Sons: Chichester, U.K., 1989.
(4) Garcia, H.; Roth, H. D. Chem. ReV. 2002, 102, 3947. (49) Thompson, T. L.; Yates, J. T. Chem. ReV. 2006, 106, 4428.
(5) Thomas, J. K. Chem. ReV. 2005, 105, 1683. (50) Kudo, A.; Miseki, Y. Chem. Soc. ReV. 2009, 38, 253.
(6) Dyrek, K.; Che, M. Chem. ReV. 1997, 97, 305. (51) Honda, K.; Fujishima, A. Nature 1972, 238, 37.
(7) Sojka, Z.; Che, M. Colloids Surf. 1999, 158, 165. (52) Howe, R. F.; Grätzel, M. J. Phys. Chem. 1987, 91, 3906.
(8) Che, M.; Tench, A. J. AdV. Catal. 1982, 31, 77. (53) Grätzel, M.; Howe, R. F. J. Phys. Chem. 1990, 94, 2566.
(9) Che, M.; Tench, A. J. AdV. Catal. 1983, 32, 1. (54) Nakaoka, Y.; Nosaka, Y. J. Photochem. Photobiol., A 1997, 110,
(10) Fubini, B.; Hubbard, A. Free Radical Biol. Med. 2003, 34, 1507. 299.
(11) Janzen, E. G. In Free Radicals in Biology; Pryor, W. A., Ed.; (55) Hurum, D. C.; Gray, K. A.; Rajh, T.; Thurnauer, M. C. J. Phys.
Academic Press: New York, 1980. Chem. B 2005, 109, 977.
(12) Giamello, E. Catal. Today 1998, 41, 239. (56) Ke, S. C.; Wang, T. C.; Wong, M. S.; Gopal, N. O. J. Phys. Chem.
(13) Both radicals and paramagnetic species possess unpaired electrons. B 2006, 110, 11628.
The difference between the two terms depends on their reactivity. (57) Hurum, D. C.; Agrios, A. G.; Crist, S. E.; Gray, K. A.; Rajh, T.;
Radicals are typically highly reactive species. The same does not Thurnauer, M. C. J. Electron Spectrosc. Relat. Phenom. 2005, 109,
necessarily apply to paramagnetic species. CH3• and Cu2+ both 977.
1346 Chemical Reviews, 2010, Vol. 110, No. 3
(58) Soria, J.; Sanz, J.; Sobrados, I.; Coronado, J. M.; Fresno, F.;
Hernandez-Alonso, M. D. Catal. Today 2007, 129, 240.
(59) Berger, T.; Sterrer, M.; Diwald, O.; Knözinger, E.; Panayotov, D.;
Thompson, T. L.; Yates, J. T. J. Phys. Chem. B 2005, 109, 6061.
(60) Berger, T.; Sterrer, M.; Diwald, O.; Knözinger, E. ChemPhysChem
2005, 6, 2104.
(61) Sterrer, M.; Berger, T.; Diwald, O.; Knözinger, E. J. Am. Chem. Soc.
2003, 125, 195.
(62) Berger, T.; Diwald, O.; Knözinger, E.; Napoli, F.; Chiesa, M.;
Giamello, E. Chem. Phys. 2007, 339, 138.
(63) Anpo, M.; Shima, T.; Kodama, S.; Kubokawa, Y. J. Phys. Chem.
1987, 91, 4305.
(64) Dimitrijevic, N. M.; Saponjic, N. C.; Rabatic, B. M.; Poluektov, O. G.;
Rajh, T. J. Phys. Chem. C 2007, 111, 14597.
(65) Coronado, J. M.; Soria, J. Catal. Today 2007, 123, 37.
(66) Murata, C.; Hattori, T.; Yoshida, H. J. Catal. 2005, 231, 292.
(67) Bedilo, A. F.; Plotnikov, M. A.; Mezentseva, N. V.; Volodin, A. M.;
Zhidomirov, G. M.; Rybkin, I. M.; Klabunde, K. J. Phys. Chem.
Chem. Phys. 2005, 7, 3059.
(68) Teramura, K.; Tanaka, T.; Ishikawa, H.; Kohno, Y.; Funabiki, T. J.
Phys. Chem. B 2004, 108, 346.
(69) Kohno, Y.; Tanaka, T.; Funabiki, T.; Yoshida, S. Phys. Chem. Chem.
Phys. 2000, 2, 2635.
(70) Kohno, Y.; Tanaka, T.; Funabiki, T.; Yoshida, S. Phys. Chem. Chem.
Phys. 2000, 2, 5771.
(71) Anpo, M.; Aikawa, N.; Kubokawa, Y.; Che, M.; Louis, C.; Giamello,
E. J. Phys. Chem. 1985, 89, 5689.
(72) Giamello, E.; Garrone, E.; Ugliengo, P. J. Chem. Soc., Faraday Trans.
1 1989, 85, 1373.
(73) Giamello, E; Garrone, E.; Ugliengo, P.; Che, M.; Tench, A. J.
J. Chem. Soc., Faraday Trans. 1 1989, 85, 3987.
(74) Chiesa, M.; Paganini, M. C.; Giamello, E.; Di Valentin, C.; Pacchioni,
G. Acc. Chem. Res. 2006, 39, 861.
(75) Chiesa, M.; Paganini, M. C.; Spoto, G.; Giamello, E.; Di Valentin,
C.; Del Vitto, A.; Pacchioni, G. J. Phys. Chem. B 2005, 109, 7314.
(76) Chiesa, M.; Paganini, M. C.; Giamello, E.; Murphy, D. M. J. Phys.
Chem. B 2001, 105, 10457.
(77) Freed, J. H. In Spin Labelling: Theory and Applications; Berliner,
L. J., Ed.; Academic Press: New York, 1976; Vol. 1.
(78) Yahiro, H.; Lund, A.; Shiotani, M. Spectrochim. Acta. A 2004, 60,
1267, and references therein.
(79) Shiotani, M.; Moro, G.; Freed, J. H. J. Chem. Phys. 1981, 74, 2616.
(80) Nilges, M.; Shiotani, M.; Yu, C. T.; Barkley, G.; Kera, Y.; Freed, J.
J. Chem. Phys. 1980, 73, 588.
(81) Shiotani, M.; Freed, J. H. J. Phys. Chem. 1981, 85, 3873.
(82) Iwaizumi, M.; Kubota, S.; Isobe, T. Bull. Chem. Soc. Jpn. 1971, 44,
3227.
(83) Kasai, P. H.; Weltner, Jr, W.; Whipple, E. B. J. Chem. Phys. 1965,
42, 1120.
(84) McDowell, C. A.; Nakajima, H.; Raghunathan, P. Can. J. Chem.
1970, 48, 805.
(85) Myers, G. H.; Easley, W. C.; Zilles, B. A. J. Chem. Phys. 1970, 53,
1181.
(86) Beckendorf, M.; Katter, U. J.; Schlienz, H.; Freund, H.-J. J. Phys.:
Condens. Matter 1993, 5, 5471.
(87) Beckendorf, M.; Katter, U. J.; Risse, T.; Schlienz, H.; Freund, H.-J.
J. Phys. Chem. 1996, 100, 9242.
(88) Schlienz, H.; Beckendorf, M.; Katter, U. J.; Risse, T.; Freund, H.-J.
Phys. ReV. Lett. 1995, 74, 761.
(89) Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178, 42.
(90) Walsh, A. D. J. Chem. Soc. 1953, 2260.
(91) March, N. H. J. Chem. Phys. 1981, 74, 2973.
(92) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin,
K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone,
V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson,
G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.;
Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai,
H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.;
Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.;
Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.;
Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg,
J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.;
Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.;
Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;
Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A.
Gaussian 03, revision C.02; Gaussian, Inc.: Wallingford, CT, 2004.
(93) Goldsborough, J. P.; Koehler, T. R. Phys. ReV. 1964, 133A, 135.
(94) Chiesa, M.; Giamello, E.; Di Valentin, C.; Pacchioni, G.; Sojka, Z.;
Van Doorslaer, S. J. Am. Chem. Soc. 2005, 127, 16935.
Chiesa et al.
(95) Catterall, R.; Edwards, P. P. J. Phys. Chem. 1980, 84, 1196.
(96) Lian, J. C.; Finazzi, E.; Di Valentin, C.; Gao, H.-J.; Risse, T.;
Pacchioni, G.; Freund, H.-J. Chem. Phys. Lett. 2008, 450, 308.
(97) Catterall, R.; Edwards, P. P. Chem. Phys. Lett. 1976, 42, 540.
(98) Catterall, R.; Edwards, P. P. Chem. Phys. Lett. 1976, 43, 122.
(99) Finazzi, E.; Di Valentin, C.; Pacchioni, G.; Chiesa, M.; Giamello,
E.; Gao, H.; Lian, J.; Risse, T.; Freund, H.-J. Chem.sEur. J. 2008,
14, 4404.
(100) Yulikov, M.; Sterrer, M.; Heyde, M.; Rust, H. P.; Risse, T.; Freund,
H.-J.; Pacchioni, G.; Scagnelli, A. Phys. ReV. Lett. 2006, 96, 146804–
4.
(101) Chenier, J. H. B.; Howard, J. A.; Joly, H. A.; Mile, B. J. Chem.
Soc., Faraday Trans. 1990, 86, 2169.
(102) Olm, M. T.; Symons, M. C. R.; Eachus, R. S. Proc. R. Soc. London,
Ser. A 1984, 392, 227.
(103) Holmberg, G. E.; Unruh, W. P.; Friauf, R. J. Phys. ReV. B 1976, 13,
983.
(104) Chiesa, M.; Napoli, F.; Giamello, E. J. Phys. Chem. C 2007, 111,
5481.
(105) CO2 Utilisation for Global Sustainability; Park, E. S., Clay, J. S.,
Lee, K. W., Eds.; Elsevier: Amsterdam, 2004.
(106) Lunsford, J. H.; Jayne, J. P. J. Phys. Chem. 1965, 69, 2182.
(107) Naccache, C. Chem. Phys. Lett. 1971, 11, 323.
(108) Huzimura, R.; Kurisu, H.; Okuda, T. Surf. Sci. 1988, 197, 444.
(109) Meriaudeau, P.; Vedrine, J. C.; Ben Taarit, Y.; Naccache, C. J. Chem.
Soc., Faraday Trans. 2 1975, 71, 736.
(110) Chiesa, M.; Giamello, E. Chem.sEur. J. 2007, 13, 1261.
(111) Preda, G.; Pacchioni, G.; Chiesa, M.; Giamello, E. J. Phys. Chem. C
2008, 112, 19568.
(112) Ovenall, D. W.; Whiffen, D. H. Mol. Phys. 1961, 4, 135.
(113) Schlick, S.; Silver, B. L.; Luz, Z. J. Chem. Phys. 1971, 54, 867.
(114) Ogata, A.; Kazusaka, A.; Enyo, M. J. Phys. Chem. 1986, 90, 5201.
(115) Lunsford, J. H.; Jayne, J. P. J. Chem. Phys. 1966, 44, 1492.
(116) Morris, R. M.; Kaba, K. A.; Groshens, T. G.; Klabunde, K. J.;
Baltisberger, R. J.; Woolsey, N. F.; Stenberg, V. I. J. Am. Chem.
Soc. 1980, 102, 3419.
(117) Cordischi, D.; Indovina, V.; Occhiuzzi, M. J. Chem. Soc., Faraday
Trans. 1 1980, 76, 1147.
(118) Morris, R. M.; Klabunde, K. J. J. Am. Chem. Soc. 1983, 105, 2633.
(119) Cordischi, D.; Indovina, V. Stud. Surf. Sci. Catal. 1980, 76, 1147.
(120) Giamello, E.; Murphy, D.; Marchese, L.; Martra, G.; Zecchina, A.
J. Chem. Soc., Faraday Trans. 1 1993, 89, 3715.
(121) Olivella, S.; Pericas, M. A.; Serratosa, F.; Messegner, A. J. Mol.
Struct. 1983, 105, 91.
(122) Ferrari, A. M.; Pacchioni, G. J. Chem. Phys. 1997, 107, 2066.
(123) Zecchina, A.; Coluccia, S.; Spoto, G.; Scarano, D.; Marchese, L.
J. Chem. Soc., Faraday Trans. 1 1990, 86, 703.
(124) Marchese, L.; Coluccia, S.; Martra, G.; Giamello, E.; Zecchina, A.
Mater. Chem. Phys. 1991, 29, 437.
(125) Topchieva, K. V.; Spiridonov, S. E.; Loginov, A. Yu. Zh. Fiz. Khim.
1986, 60, 411.
(126) Moon, S. C; Yamashita, H.; Anpo, M. Stud. Surf. Sci. Catal. 1994,
90, 479.
(127) Moens, P.; Callens, F.; Matthys, P.; Maes, F.; Verbeeck, R.; Naessens,
D. J. Chem. Soc., Faraday Trans. 1991, 87, 3137.
(128) Moens, P.; Callens, F. J.; Verbeeck, R. M. H.; Naessens, D. E. Appl.
Radiat. Isot. 1993, 44, 279.
(129) Ben Taarit, Y.; Vedrine, J. C.; Naccache, C.; de Montgolfier, P.;
Meriaudeau, P. J. Chem. Phys. 1977, 67, 2880.
(130) Schlick, S.; Kevan, L. J. Chem. Phys. 1980, 72, 784.
(131) Lipatkina, N. I.; Shubin, V. E.; Shvets, V. A.; Chuvylkin, N. D.;
Kazanskii, V. B. Kinet. Katal. 1982, 23, 670.
(132) Hochstrasser, G.; Antonini, J. F. Surf. Sci. 1972, 32, 644.
(133) Griscom, D. L. In Defects in SiO2 and Related Dielectrics: Science
and Technology; Pacchioni, G., Skuja, L., Griscom, D. L., Eds.;
NATO Science Series; Kluwer Academic Publishers: Dordrecht, The
Netherlands, 2000.
(134) Griscom, D. L. ReV. Solid State. Sci. 1990, 4, 565.
(135) Radtsig, V. A.; Bystykov, A. V. Kinet. Katal. 1978, 19, 713.
(136) Radtsig, V. A. Kinet. Katal. 1979, 20, 448.
(137) Radtsig, V. A. Kinet. Katal. 1979, 20, 1203.
(138) Biogeochemical Cycles in Elements; Sigel, A., Sigel, H., Sigel,
R. K. O., Eds.; Marcel Dekker: New York, 2005.
(139) Catalytic Ammonia Synthesis. Fundamentals and Practice; Jennings,
R. L., Ed.; Plenum Press: New York, 1991.
(140) Shelef, M. Chem. ReV. 1995, 95, 209.
(141) Brailsford, J. R.; Morton, J. R.; Vanotti, L. E. J. Chem. Phys. 1969,
50, 1051.
(142) Henderson, R. A.; Leigh, G. J.; Pickett, C. J. AdV. Inorg. Chem.
Radiochem. 1983, 27, 197.
(143) Bazhenova, T. A.; Shilov, A. E. Coord. Chem. ReV. 1995, 144, 64.
(144) Kim, J.; Rees, D. C. Biochemistry 1994, 33, 389.
Surface-Localized Inorganic Radicals and Radical Ions
(145) Paganini, M. C.; Chiesa, M.; Giamello, E.; Coluccia, S.; Martra, G.;
Murphy, D. M.; Pacchioni, G. Surf. Sci. 1999, 421, 246.
(146) Giamello, E.; Murphy, D. M.; Paganini, M. C.; Ferrari, A. M.;
Pacchioni, G. J. Phys. Chem. B 1997, 101, 971.
(147) Giamello, E.; Chiesa, M.; Murphy, D. M.; Paganini, M. C.; Pacchioni,
G.; Soave, R.; Rockenbauer, A. J. Phys. Chem. B 2000, 104, 1887.
(148) Chiesa, M.; Giamello, E.; Murphy, D. M.; Pacchioni, G.; Paganini,
M. C.; Soave, R.; Sojka, Z. J. Phys. Chem. B 2001, 105, 497.
(149) Sojka, Z.; Chiesa, M.; Paganini, M. C.; Giamello, E. Stud. Surf. Sci.
Catal. 2001, 140, 413.
(150) Ferrari, A. M.; Soave, R.; D’Ercole, A.; Pisani, C.; Giamello, E.;
Pacchioni, G. Surf. Sci. 2001, 479, 83.
(151) Ricci, D.; Pacchioni, G.; Sushko, P.; Shluger, A. Surf. Sci. 2003,
542, 293.
(152) Iwamoto, M.; Furukawa, H.; Mine, Y.; Uemura, F.; Mikuriya, S.;
Kagawa, S. J. Chem. Soc., Chem. Commun. 1986, 1272.
(153) Whittaker, J. W. J. Chem. Educ. 1991, 68, 421.
(154) Lunsford, J. H. J. Chem. Phys. 1967, 46, 4347.
(155) Lunsford, J. H. J. Phys. Chem. 1968, 72, 2141.
(156) Lunsford, J. H. J. Phys. Chem. 1968, 72, 4163.
(157) Kasai, P.; Bishop, R. J. J. Am. Chem. Soc. 1972, 94, 5560.
(158) Kasai, P. H.; Bishop, R. J. In Zeolite Chemistry and Catalysis; Rabo,
J. A., Ed.; ACS Monograph 171; American Chemical Society:
Washington, DC, 1976.
(159) Hoffman, B. M.; Nelson, N. J. J. Chem. Phys. 1969, 50, 2598.
(160) Yahiro, H.; Lund, A.; Benetis, N. P.; Shiotani, M. Chem. Lett. 2000,
7, 736.
(161) Lunsford, J. H. J. Catal. 1969, 14, 379.
(162) Primet, M.; Che, M.; Naccache, C.; Mathieu, M. V.; Imelik, B.
J. Chim. Phys. Phys.-Chim. Biol. 1970, 67, 1629.
(163) Martinez-Arias, A.; Soria, J.; Conesa, J. C.; Seoane, X. L.; Arcoya,
A.; Cataluna, R. J. Chem. Soc., Faraday Trans. 1995, 91, 1679.
(164) Che, M.; Giamello, E. Stud. Surf. Sci. Catal. 1990, 57B, 265.
(165) Di Valentin, C.; Pacchioni, G.; Chiesa, M.; Giamello, E.; Abbet, S.;
Heiz, U. J. Phys. Chem. B 2002, 106, 1637.
(166) Rudolf, T.; Pöppl, A.; Hofbauer, W.; Michel, D. Phys. Chem. Chem.
Phys. 2001, 3, 2167.
(167) Pöppl, A.; Rudolf, T.; Michel, D. J. Am. Chem. Soc. 1998, 120, 4879.
(168) Pöppl, A.; Rudolf, T.; Manikandan, P.; Goldfarb, D. J. Am. Chem.
Soc. 2000, 122, 10194.
(169) Rudolf, T.; Bohlmann, W.; Pöppl, A. J. Magn. Reson. 2002, 155,
45.
(170) Volodin, A.; Biglino, D.; Itagaki, I.; Shiotani, M.; Lund, A. Chem.
Phys. Lett. 2000, 327, 165.
(171) Yahiro, H.; Lund, A.; Aasa, R.; Benetis, N. P.; Shiotani, M. J. Phys.
Chem. A 2000, 104, 7950.
(172) Biglino, D.; Bonora, M.; Volodin, A.; Lund, A. Chem. Phys. Lett.
2001, 349, 511.
(173) Zhang, G.; Tanaka, T.; Yamaguchi, T.; Hattori, H.; Tanabe, K. J.
Phys. Chem. 1990, 94, 506.
(174) Paganini, M. C.; Chiesa, M.; Martino, P.; Giamello, E. J. Phys. Chem.
B 2002, 106, 12531.
(175) Che, M.; Naccache, C. Chem. Phys. Lett. 1971, 8, 45.
(176) Di Valentin, C.; Pacchioni, G.; Selloni, A.; Livraghi, S.; Giamello,
E. J. Phys. Chem. B 2005, 109, 11414.
(177) Enemark, J. H.; Feltham, R. D. Coord. Chem. ReV. 1974, 13, 339.
(178) Chiesa, M.; Paganini, M. C.; Giamello, E.; Di Valentin, C.; Pacchioni,
G J. Mol. Catal. A 2003, 204-205, 779.
Chemical Reviews, 2010, Vol. 110, No. 3 1347
(179) Berlier, G.; Spoto, G.; Bordiga, S.; Ricchiardi, G.; Fisicaro, P.;
Zecchina, A.; Rossetti, I.; Selli, E.; Forni, L.; Giamello, E.; Lamberti,
C. J. Catal. 2002, 208, 64.
(180) Fisicaro, P.; Giamello, E.; Berlier, G.; Lamberti, C. Res. Chem.
Intermed. 2003, 29, 805.
(181) Chao, C. C.; Lunsford, J. H. J. Phys. Chem. 1972, 76, 1546.
(182) Naccache, C.; Che, M.; Ben Taarit, Y. Chem. Phys. Lett. 1972, 13,
109.
(183) Giamello, E.; Murphy, D. M.; Magnacca, G.; Morterra, C.; Shioya,
Y.; Nomura, T.; Anpo, M. J. Catal. 1992, 136, 510.
(184) Anpo, M.; Nomura, T.; Kitao, T.; Giamello, E.; Murphy, D. M.; Che,
M.; Fox, M. A. Res. Chem. Intermed. 1991, 15, 225.
(185) Iwamoto, M.; Hamada, H. Catal. Today 1991, 10, 57.
(186) Sojka, Z.; Che, M.; Giamello, E. J. Phys. Chem. B 1997, 101, 4831.
(187) Poppl, A.; Hartmann, M. Stud. Surf. Sci. Catal. 2002, 142, 375.
(188) Umamaheswari, V.; Poppl, A.; Hartmann, M. J. Mol. Catal. A 2004,
223, 123.
(189) Umamaheswari, V.; Hartmann, M.; Poppl, A. J. Phys. Chem. B 2005,
109, 19723.
(190) Neyman, K. M.; Ganyushin, D. I.; Nasluzov, V. A.; Rosch, N.; Poppl,
A.; Hartmann, M. Phys. Chem. Chem. Phys. 2003, 5, 2429.
(191) Umamaheswari, V.; Hartmann, M.; Poppl, A. J. Phys. Chem. B 2005,
109, 10842.
(192) Naccache, C.; Che, M. In Proceedings of the 5th International
Congress on Cataysis; Hightower, J. W., Ed.; North Holland:
Amsterdam, 1973.
(193) Anpo, M.; Aikawa, N.; Kubokawa, Y.; Che, M.; Louis, C.; Giamello,
E. J. Phys. Chem. 1985, 89, 5017.
(194) Nagata, M.; Yahiro, H.; Shiotani, M.; Lindgren, M.; Lund, A. Chem.
Phys. Lett. 1996, 256, 27.
(195) Lunsford, J. H. J. Colloid Interface Sci. 1968, 26, 355.
(196) Garrone, E.; Guglielminotti, E.; Zecchina, A.; Giamello, E. J. Chem.
Soc., Faraday Trans. 1984, 80, 2723.
(197) Garrone, E.; Giamello, E. Stud. Surf. Sci. Catal. 1985, 21, 225.
(198) Ugliengo, P.; Garrone, E; Giamello, E. Z. Phys. Chem. 1987, 152,
31.
(199) Kolosov, A. K.; Shvets, V. A.; Chuvylki, N. D.; Kazansky, V. B. J.
Catal. 1977, 47, 190.
(200) Schoonheydt, R. A.; Lunsford, J. H. J. Phys. Chem. 1972, 76, 323.
(201) Atkins, P. W.; Horsfield, A.; Symons, M. C. R. J. Chem. Soc. 1964,
5220.
(202) Dinse, K. P.; Möbius, K. Z. Naturforsch., A 1968, 23, 695.
(203) Mashchenko, A. I.; Pariiskii, G. B.; Kazanskii, V. B. Kinet. Katal.
1968, 9, 151.
(204) Ben Taarit, Y.; Naccache, C.; Che, M.; Tench, A. J. Chem. Phys.
Lett. 1974, 24, 41.
(205) Lin, M. J.; Lunsford, J. H. J. Phys. Chem. 1976, 80, 635.
(206) Lunsford, J. H.; Johnson, D. P. J. Chem. Phys. 1973, 58, 2079.
(207) Lin, M. J.; Johnson, D. P.; Lunsford, J. H. Chem. Phys. Lett. 1972,
15, 412.
(208) Lin, M. J.; Lunsford, J. H. J. Phys. Chem. 1976, 80, 2015.
(209) Livraghi, S.; Paganini, M. C.; Chiesa, M.; Giamello, E. Res. Chem.
Intermed. 2006, 32, 777.
(210) Coope, J. A. R.; Gardner, C. L.; McDowell, C. A.; Pelman, A. I.
Mol. Phys. 1971, 21, 1043.
(211) Castner, T. G.; Känzig, W. J. Phys. Chem. Solids 1957, 3, 178.
(212) Raghunathan, P.; Sur, S. K. J. Am. Chem. Soc. 1984, 106, 8014.
CR800366V
1348 Chem. Rev. 2010, 110, 1348–1385

Carbon Nanotube-Inorganic Hybrids

Dominik Eder*
Department of Materials Science and Metallurgy, University of Cambridge, New Museums Site, Pembroke Street,
Cambridge CB2 3QZ, United Kingdom

Received July 15, 2008

Contents 1. Motivation and General Aspects

1. Motivation and General Aspects
1.1. Introduction
1.2. Hybrid Materials
2. Carbon Nanotubes in Hybrid Materials
2.1. Why Carbon Nanotubes?
2.2. Preparation of CNTs for Use in Hybrid
Materials
2.2.1. Purification of CNTs
2.2.2. Separation of Metallic and Semiconducting
CNTs
2.2.3. Functionalization
2.3. Filling CNTs
3. Synthesis of CNT-Inorganic Hybrids
3.1. Ex Situ Approach: Attaching Nanoparticles to
CNTs
3.1.1. Covalent Interactions
3.1.2. Noncovalent Interactions
3.1.3. π-π Stacking
3.1.4. Electrostatic Interactions
3.2. In Situ Synthesis Directly on the CNT Surface
3.2.1. Electrochemical Techniques
3.2.2. Sol-Gel Process
3.2.3. Hydrothermal and Aerosol Techniques
3.2.4. Gas-Phase Deposition
3.3. Comparison of Synthesis Techniques
4. Properties and Potential Applications of
CNT-Inorganic Hybrids
4.1. Example 1: Photochemical and
Photoelectrochemical Applications
4.1.1. Photocatalytic Production of Hydrogen
4.1.2. Photocatalytic Decomposition of Organic
Compounds
4.1.3. Photovoltaic Devices
4.1.4. Synergistic Effects of CNTs
4.2. Example 2: Heterogeneous Catalysis and
Electrocatalysis
4.3. Example 3: Gas Sensors, Chemical Sensors
4.4. Example 4: Supercapacitors and Batteries
4.5. Example 5: Photonics and Field Emission
Devices
4.6. Example 6: Oxidation Resistance
4.7. Example 7: Synthesis of Inorganic Nanotubes
5. Challenges and Outlook
6. Abbreviations
7. Acknowledgments
8. References
* E-mail: de235@cam.ac.uk.
10.1021/cr800433k  2010 American Chemical Society
Published on Web 01/28/2010
1348 1.1. Introduction
1348
1349 The world of nanomaterials provides exciting challenges
and opportunities for chemists, physicists, biologists, and
1351
material scientists. The use of nanomaterials as building
1351 blocks for devices not only helps downscale conventional
1352 technologies by at least an order of magnitude but also offers
a cheaper and more environmentally friendly production
1352 route, due to a drastic reduction in the necessary amount of
1352 raw materials, thus mimicking nature’s efficient ways of
managing with less when it comes to chemical and physical
1353 processing.1 Moreover, chemical and physical properties of
1353 substances can be considerably altered and fine-tuned when
1355 they exist on a nanoscopic scale, introducing the particle size
1355 as a powerful new parameter.2 The quantum dot (QD), a
nanosized cluster displaying atom-like behavior,3 is a familiar
1355 example of this effect.
1356 Scaling down the particle size to nanometer dimensions
1356 also increases the specific surface area of the material; thus,
1357 applications with reactions at the gas-solid or liquid-solid
1358 interface will also benefit from this “striving to the smaller”.
1358 Typical applications include catalysis, energy conversion,
1361 electrochemistry, and environmental chemistry, where the
1363 use of nanomaterials increases response time, efficiency, and
sensitivity.4 One of the major challenges in nanoscale science
1364
is the synthesis of nanomaterials with monodisperse sizes,
1368 uniform morphologies, and functionalized surfaces. Regard-
1368 ing potential electronic applications, such as nanocapacitors
or nanotransistors, addressing and assembling arrays of
1368 particles become necessary. For alignment and functional-
ization procedures, nanoparticles with an anisotropic mor-
1369 phology are advantageous because of their ability to shear
1369 align, form porous structures, and percolate at low concentra-
tions.
1370
Many conventional materials, such as metals, ceramics,
1371 and plastics, cannot fulfill all requirements for new technolo-
1371 gies seeking to solve the world’s most immediate problems
related to energy and the environment. In many cases,
1373 however, the combination of two materials can show
1374 properties superior to those of their individual constituents.
1376 A well-known example is inorganic fiber-reinforced polymer
composites, which in recent years have become ubiquitous
1377 in applications where lightweight, tough materials are
1378 required, including aerospace, high-end automotive, and
1379 sporting equipment. Decreasing the size of the inorganic units
1380 to the same level as the organic building blocks has led to
1380 more homogeneous materials, which allowed a further
1380 refinement of material properties on the molecular level. This
has resulted in the introduction of organic-inorganic hybrids,
such as metal organic frameworks. In contrast to nanocom-
Carbon Nanotube-Inorganic Hybrids
Dominik Eder received his Ph.D. in Physical Chemistry in 2003 from the
University of Innsbruck, Austria, working on the in situ characterization of
metal oxide-supported noble metal catalysts with electrochemical imped-
ance spectroscopy. For his thesis, he was awarded the Sosnovski Medal
2004 from the University of Innsbruck. After his civil service at the Red
Cross in Innsbruck, he joined the Macromolecular Materials Laboratory
(Prof. A. H. Windle) in the Department of Materials Science and Metallurgy
at the University of Cambridge in 2005, where he worked as an Erwin
Schroedinger Research Fellow on the synthesis of oxide nanotubes. In
2006, he was awarded an APART Advanced Research fellowship from
the Austrian Academy of Science. He is currently associated with the
Functional Inorganic and Hybrid Materials group (Prof. K. Cheetham) in
the same Department and his research interests include the synthesis of
functional oxide nanostructures and CNT-inorganic hybrids and their
application in photocatalysis, environmental catalysis, and batteries.
posites, which simply combine the individual properties of
the components, these hybrid materials merge the properties
of the components in a way that creates new properties
distinct from those of either building block.5,6
This review focuses on a new class of hybrid materials
made from carbon nanotubes (CNTs) and inorganic glasses
or ceramics, CNT-inorganic hybrids. The many advantages
of CNTs in hybrid materials include their high aspect ratio
(>1000) and tubular geometry, which provides ready gas
access to a large specific surface area and percolation at very
low volume fractions. Their excellent mechanical, electrical,
and optical properties support CNTs as an ideal building
block in hybrid materials. The high thermal conductivity of
CNTs enable them to behave as a heat sink during calcination
and activation treatments, thereby stabilizing small inorganic
moieties that can decorate the sidewalls of the CNTs. This
results in materials with higher specific surface areas that
allow the use of less material, reducing cost and toxicity.
Over the past few years, CNTs have been combined with a
variety of inorganic compounds, including oxides, nitrides,
carbides, chalcogenides, and ceramics. Out of these, the
oxides are by far the most commonly explored species.
Figure 1 shows the evolution of publications since the year
2000 that deal with either the synthesis or application of new
CNT-oxide hybrids (Sources: ISI Web of Knowledge
(Thomson Reuters) and Scopus (Elsevier)).
Although still in a very early stage of research, CNT-
inorganic hybrids have shown exceptional performance in
applications such as gas sensors (SnO2) and photovoltaics
(ZnO), exhibiting an enhanced ability to trap electrons and
reduce the electron-hole recombination rate compared to the
bulk materials. CNTs can also serve as additional photosen-
sitizers in photocatalysts (TiO2) or intrinsic capacitors in
supercapacitors and batteries (MnO2, RuO2). These few
examples prove the great potential of these hybrids and show
that the next technological frontiers will be opened by
Chemical Reviews, 2010, Vol. 110, No. 3 1349
Figure 1. Number of publications dedicated to the synthesis
and application of CNT-oxide hybrid materials since the year
2000.
understanding and optimizing material combinations and their
synergistic functions, rather than through a better understand-
ing and application of a particular material.
This review offers a comprehensive critical evaluation of
CNT-inorganic hybrids and their potential applications and
is structured in the following way. Starting with an attempt
to distinguish between various hybrid materials, such as
metal-organic frameworks and nanocomposites (section
1.2), I will then introduce CNTs as building blocks for hybrid
materials (section 2.1) and discuss the challenges involved,
including purification and functionalization issues (section
2.2). This will be followed by an overview of CNTs whose
hollow central cavities have been filled with various metals and
inorganic compounds (known as meta-CNTs), which can be
seen as a first attempt to produce CNT-inorganic hybrid
structures (section 2.3). The second part of the review provides
a comprehensive summary of various chemical and physical
synthesis techniques, which I categorize into ex situ (section
3.1) and in situ (section 3.2) routes. Finally, the last part of the
review demonstrates the potential of CNT-inorganic hybrids
for a wide range of applications (section 4) and discusses the
remaining challenges and future prospects (section 5).
1.2. Hybrid Materials
The last two decades have seen the development of some
new types of hybrid materials, such as nanocomposites and
metal oxide frameworks, and their implementation in various
applications that require, for example, light structural,
bioactive, or smart materials.5
(1) Nanocomposites are multiphase materials, in which one
phase (filler) is dispersed in a second phase (matrix), resulting
in a combination of the individual properties of the compo-
nent materials. Filler materials are typically inorganic build-
ing blocks in the nanosize regime and in the form of particles,
whiskers, fibers, lamellae, or a mesh. The matrix can be either
organic (e.g., polymer) or inorganic (e.g., ceramic or metal).
The materials are typically synthesized ex situ by simple
mixing techniques (e.g., ball milling or shear mixing), which
often results in a heterogeneous distribution of the filler and
consequently in nonuniform properties. As an example, the
result of the incorporation of inorganic nanoparticles, nano-
rods, carbon fibers, CNTs and galleries of clay minerals into
1350 Chemical Reviews, 2010, Vol. 110, No. 3
Table 1. Definitions for Hybrid Various Materials
nanocomposites
composition two phases: a filler is distributed
within organic or inorganic matrix
formation both phases are discrete building
blocks
properties use of individual properties,
known structure-property
relationship
an organic polymer or a ceramic matrix is considered to be
a nanocomposite material. In contrast to some other hybrid
materials, nanocomposites simply use the intrinsic properties
of their individual components, whose discrete and control-
lable size, shape, and structure enable a good structure-
property relationship. In this respect, the purpose of the filler
is often to mechanically reinforce the matrix or to alter its
thermal, electrical, or barrier properties, and as such, these
nanomaterials have already had a considerable impact in
applications such as lightweight structural materials in
aerospace, electrically conducting plastics, and packaging
materials with reduced gas permeability.
(2) Other hybrid materials, such as metal-organic frame-
works, consist of a single phase containing both inorganic
and organic building blocks and merge the properties of the
components on the molecular scale in a way that creates
novel properties distinct from those of either building block.
For instance, the organic component can provide biocom-
patibility and chemical functionality and allows easy shaping
and better processing of the materials. The inorganic
components provide mechanical and thermal stability but
also new functionalities, which depend on the chemical
nature, the structure, the size, and the crystallinity of the
inorganic phase. They can implement or improve electronic,
magnetic, and redox properties, density, refraction index, etc.
So in this case, the final materials are not merely the sum of
the primary components, but rather completely new materials
with new properties. A synergistic effect often occurs from
the close proximity of the two phases through size domain
effects and the nature of the interfaces. Considering these
effects, the ultimate aim is to create so-called smart materials
that can react to environmental changes or switchable
systems, as these will pave the way to novel technologies,
such as electroactive and electrochromic materials, chemical
and biosensors, smart coatings, and biohybrid materials.7
(3) This review article focuses on a new type of hybrid
materials, CNT-inorganic hybrids, which replace the organic
compound with CNTs. In contrast to nanocomposites, the
CNTs are coaxially coated with the inorganic compound. A
significant synergistic effect in CNT-hybrids is expected
through size domain effects and charge transfer processes
through the CNT-inorganic interface, which will be dis-
cussed in section 4 in more detail. Consequently, this new
class of functional materials combines the multiphase
characteristics of nanocomposites with functions of hybrid
frameworks (Table 1).
In general, hybrid materials can be further classified into
Class I and II materials. This distinction is based on the
strength of interaction between the two components, which
also affect the hybrids’ properties. The synthesis strategies
and properties of Class I and II hybrids have been intensively
Eder
metal-organic frameworks CNT-inorganic hybrids
single phase: inorganic units two phases: the CNTs
are linked by organic groups are coaxially coated
on the molecular scale with inorganic second
component
often formed in situ by inorganic phase can be
molecular precursors formed either in situ
(molecular precursor) or
ex situ (building blocks)
properties are combined to synergistic effects
create multifunctional similar to hybrid
materials with yet unknown frameworks create new
properties properties
described in several books5,6,8 and reviews;7,9,10 hence, this
review will only provide a short summary.
• Class I hybrids typically show weak interactions between
the two phases, such as van der Waals, hydrogen bonding,
or weak electrostatic interactions. Examples of these hybrid
materials include organic dyes or monomers that are attached
to a sol-gel matrix or inorganic nanoparticles that are
embedded in a polymer blend. Interpenetrating networks
(IPNs) are another example of Class 1 hybrid materials and
are formed when an inorganic and an organic compound
form networks that interpenetrate each other without strong
chemical interactions.
• Class II hybrids are formed via strong chemical interac-
tions between the components, for example, when the
discrete inorganic building blocks are covalently bonded to
organic polymers, when inorganic sol-gel networks are
modified with organic functional groups (e.g., ORCOMER),
or when inorganic and organic polymers are covalently
connected to each other by capping or linking agents. The
building blocks can be produced in a variety of shapes,
structures, and size distributions and with the desired material
properties. Hence, this step-by-step approach generally
enables a good structure-property prediction.
Both approaches can be combined in a powerful template-
assisted synthesis process for designing novel complex
materials. The template can be either the organic or the
inorganic compound in the form of one-dimensional struc-
tures (e.g., CNTs, inorganic wires, polymer fibers, biomol-
ecules, or DNA) or multidimensional networks (e.g., gyroid-
phase block copolymers and polymer blends, zeolites, and
membranes). With this technique, a new generation of
crystalline microporous hybrid solids have recently been
discovered by Cheetham and Rao,11 Stein,12,13 Ferey et al.,14,15
Wiesner et al.,16,17 Yang and Stucky,18 and Yaghi et al.,19,20
just to mention a few. Among other characteristics, these
materials combine exciting magnetic and electronic properties
with very high surface areas (from 1000 to 4500 m2/g21,22).
An important advantage of hybrid materials is the
diversity in suitable synthesis routes. In contrast with pure
solid-state inorganic materials that often require a high-
temperature treatment for their processing, hybrid materi-
als may benefit from the convenience of traditional
polymer-processing techniques (e.g., extrusion, compres-
sion, molding, etc.). This is either because of their large
organic content or because of the formation of cross-linked
inorganic networks from small molecular precursors, as
in polymerization reactions. Hence, these materials can
be produced at low temperatures (often using sol-gel and
hydrothermal reactions) in various morphologies, such as
3D networks, thin films, or nanoparticles.
Carbon Nanotube-Inorganic Hybrids Chemical Reviews, 2010, Vol. 110, No. 3 1351

Table 2. Advantages and Disadvantages of Various CNT Synthesis Techniques

synthesis technique description product advantages disadvantages improvements
DC arc discharge 1991: Iijima381 1992: MWCNTs and SWCNTs easy design, few short tubes and low magnetic field, rotating
Ebbesen and (with catalyst) structural defects yield, low purity, electrode, liquid N2
Ajayan382 random diameters
laser ablation 1995: Smalley383 primarily SWCNTs few defects, good most costly method, ultrafast electron laser
control over poor scalability,
diameter requires Class 4
lasers
molten salt 1995: Terrones and primarily MWCNTs simple process, low yield and
Kroto384 used for filling crystallinity; poor
CNTs controllability
chemical vapor 1998: Endo385 both types high yields, easy some defects, plasma enhanced, aerosol
deposition scalability, long medium purity process, laser assisted,
tubes, alignment CoMoCat, HiPCo,
and pattern growth

Table 3. Summary of Physical Properties of CNTs in Comparison with Graphite; Values Take from Ref 20
property SWCNTs MWCNTs graphite
specific gravity 0.8 g/cm3 <1.8 g/cm3 2.26 g/cm3
elastic modulus ∼1.4 TPa ∼0.3-1 TPa 1 TPa (in plane)
strength 50-500 GPa 10-60 GPa
resistivity 5-50 µΩ cm 50 µΩ cm (in plane)
thermal conductivity 3000 W m-1 K-1 3000 W m-1 K-1 (in plane)
6 W m-1 K-1 (c-axis)
magnetic susceptibility 22 × 106 EMU/g (perpendicular) 0.5 × 106 EMU/g (parallel)
0.5 × 106 EMU/g (parallel)
thermal expansion negligible -1 × 10-6 K-1 (in plane)
29 × 10-6 K-1 (c-axis)
thermal stability 600-800 °C in air 450-650 °C in air
2800 °C in vacuum

The choice of synthesis techniques for CNT-inorganic
hybrids and the extent of their synergistic function depend
on the type and purity of CNTs and the modification of their
surface chemistry. Therefore, the next section will summarize
the characteristics and properties of CNTs and their synthesis
routes.
2. Carbon Nanotubes in Hybrid Materials
2.1. Why Carbon Nanotubes?
Carbon nanotubes and fullerenes have defined the research
field of nanotechnology like no other type of material. With
applications as diverse as integrated circuits and memory
devices, sensors and filters, solar cells, and field emission
displays,23 as well as artificial muscles24 and the space
elevator,25 CNTs have been popularized in science fiction
novels and television documentaries throughout the world.
A vast number of review articles and textbooks have been
dedicated to the unique and fascinating properties of CNTs;
hence, this review will only provide the very basics required
for understanding their role in CNT-inorganic hybrids.
Many textbooks describe CNTs in simple terms as tubular
structures made entirely of rolled-up layers of interconnected
carbon atoms,26,27 with diameters ranging from about one
nanometer to tens of nanometers and lengths up to centi-
meters. CNTs can be open-ended or closed by a hemispheri-
cal fullerene-type cap, depending on their synthesis method.28
Along with structures related to those of fullerenes,29 CNTs
are considered a third allotropic form of carbon, with the
others being diamond and graphite. They are classified as
either (a) “single-walled” tubes (SWCNTs, 0.7 < d < 2 nm),
which consist of a single layer of graphene sheet seamlessly
rolled into a cylindrical tube, or (b) multiwalled CNTs
(MWCNT, 1.4 < d < 150 nm), which comprise multiple
concentric tubes separated by about 0.34 nm.
CNTs are generally produced by four main techniques (arc
discharge, laser ablation, molten salt intercalation, and
chemical vapor deposition), which are frequently discussed
in detail in several reviews.23,30-32 Depending on the resulting
purity, defect concentration, yield, and other parameters, each
technique has several advantages and disadvantages, which
are summarized in Table 2.
In general, CNTs possess large specific surface areas due
to their hollow geometry, while their structural integrity and
chemical inertness support relatively high oxidation stability.
Other advantages include their exceptional physical proper-
ties, which have been extensively discussed in several
books28,33,34 and reviews32,35,36 and are summarized in Table
3 (taken from refs 28 and 36) in comparison with graphite.
• Electrical properties: The electrical resistivity of CNTs
is determined by the unique structure of graphite and the
quantum mechanical properties associated with their 1D
character and small size, which results in the near-total
elimination of electron collisions (scattering). Hence, CNTs
are ballistic conductors, whose resistance is independent of
the nanotube length. Furthermore, they can carry the highest
current density of any known material, with reported
measurements as high as 109 A/cm2. For comparison, copper
wires burn out at 106 A/cm2.37 Depending on their helicity
and diameter, CNTs can be either metallic or semiconducting.
This is of importance as semiconducting SWCNTs may
perform better in applications that involve charge transfer
processes, including sensors, field emission devices, and
photocatalytic applications, while metallic CNTs are pre-
ferred as interconnects in electronic devices or as conductive
filler in CNT-composites. Consequently, this requires stan-
dardization of synthesis and improvement in separation and
purification techniques, which will be discussed in the next
section.
1352 Chemical Reviews, 2010, Vol. 110, No. 3
Table 4. Summary of Typical Purification Techniques for CNTs
metalsa carbonb
heating in air/O2 X X
heating in wet O2 X
HClconc Xd
HNO3/H2SO4 X X
microwave Xe
Ar @ 2000 °C X X
a
Treatment can remove metal catalyst residues. b Carbon residues (e.g., amorphous or organic aromatic debris). c Purification introduces covalently
bonded functional groups. d Only if not covered with carbon or encapsulated within CNT. e Only amorphous carbon around metal particles.
• Mechanical properties: The mechanical properties of
CNTs originate from the strong CdC double bonds, which
yield a very large Young’s modulus in their axial direction
(1.4 TPa for single-walled CNTs).38 CNTs have an expected
elongation-to-failure of 20-30%, which, combined with their
stiffness, projects to a tensile strength well above 100 GPa
(e.g., steel: 1-2 GPa)sby far the highest known.39 However,
both Young modulus and tensile strength are strongly
reduced by the presence of defects in the graphitic walls of
CNTs, such as Stone-Wales defects. Hence, experimental
values tend to be considerably smaller than theoretical
predictions.40 Because of their high aspect ratio, CNTs are
also very flexible and thus potentially suitable for applications
in composite materials that require anisotropic properties.41
• Thermal properties: Theoretical works predicted room-
temperature thermal conductivities of individual single-
walled CNTs of up to 6600 W/(m K).42 CNTs would
therefore transmit nearly twice as much heat as isotopically
pure diamond. Experimental studies give somewhat lower
values, with Hone et al. reporting room-temperature thermal
conductivities of 300 W/(m K) for bulk single-walled CNTs
and of 3000 W/(m K) for individual multiwalled CNTs.43
In summary, due to their remarkable mechanical, electrical,
biological, optical, and thermal properties, CNTs promise
enormous potential for various technological areas in the
energy, information, aerospace, medicine, and chemical
industries, where they can be used as gas adsorbents,
templates, actuators, composite reinforcements, catalyst
supports, or chemical sensors, among other things. For the
same reasons, they are promising building blocks for hybrid
materials. The next section describes some requirements for
CNTs to be successfully implemented in hybrid materials.
2.2. Preparation of CNTs for Use in Hybrid
Materials
In this section, we will consider the purification, the
separation, and the functionalization of CNTs as preparation
for their use in hybrid materials. As-produced CNTs contain
a variety of impurities such as fragments of wrapped-up
graphene sheets, amorphous carbon, fullerenes, and metal
catalyst particles.44 As these impurities interfere with most
of the desired properties as well as with the biocompatibility
of the CNTs, it is desirable to remove them, preferably
without affecting the performance of the CNTs themselves
(section 2.2.1).
Furthermore, many applications require uniform and stable
dispersions of CNTs. However, pristine single-walled CNTs
are insoluble in most organic solvents and aqueous solutions
and tend to aggregate as a result of van der Waals interactions
between individual tubes. It is also desirable to select CNTs
by structure and size and to separate semiconducting from
metallic CNTs, as the electronic properties of SWCNTs
depend on their chirality and diameter (section 2.2.2). Finally,
Eder
functionalc comments
opens tips
assisted with microwave or magnetic field
X open tips, shortens CNTs
aided by acids (e.g., HCl)
anneals crystal structure
the CNT surface can be functionalized in order to improve
the chemical reactivity, to recognize specific target molecules,
or to enhance the interactions with other compounds in hybrid
materials (section 2.2.3).
2.2.1. Purification of CNTs
Purification of CNTs has been a matter of intensive
study.45-49 Table 4 shows the advantages and disadvantages
of various common purification techniques. The efficiency
and yield of the purification procedure depends on a variety
of factors, such as metal content, oxidation time, environ-
ment, oxidizing agent, and temperature.46
For instance, oxidative treatment of CNTs is a simple way
to remove carbonaceous impurities. A variety of oxidizing
atmospheres have been tested, including air,50 a mixture of
H2S and O2,51 and steam.52 However, oxidation often
damages the CNTs’ surface, especially when combined with
ultrasonication or high-temperature treatment. Generally, the
suitable oxidation temperature should remain well below that
for CNT combustion, which can range from 550 to 850 °C,
depending on the number of structural defects. Additionally,
the metal residues can act as an oxidation catalyst, decreasing
the oxidation temperature even further. Oxidation with strong
acids, such as HNO3/H2SO4,53 also creates structural defects
and the formation of various organic groups, hence altering
the surface chemistry of the CNTs.
To remove metal residues without affecting the carbon-
aceous species, the CNTs are typically treated in strong
nonoxidizing acids such as HCl. Assisted by a magnetic field,
these acids predominantly dissolve the metal particles that
are not covered by amorphous carbon or encapsulated within
CNTs. In contrast, microwave-assisted purification heats up
the metal and thus primarily removes the amorphous carbon
attached to it.47
Annealing at high temperatures in vacuum or inert gas is
a powerful alternative to the oxidizing techniques above.54,55
With this technique, amorphous and graphitic defects can
be removed selectively simply by adjusting the temperature
(600-2000 °C). The metal residues are typically removed
at temperatures close to the melting point of the nanosized
metal particles (e.g., above 1600 °C for Fe). Finally, at very
high temperatures (1900-2000 °C) the carbon atoms in the
CNT walls rearrange, thus decreasing the number of struc-
tural defects and so enhancing the degree of graphitization.
2.2.2. Separation of Metallic and Semiconducting CNTs
In contrast to the purely metallic MWCNTs, SWCNTs
can be either metallic or semiconducting, depending on
their diameter and chirality. For many applications,
including nanoscale electronic, optoelectronic, and sensing
devices, it is desirable to use SWCNTs that are either
purely metallic or purely semiconducting. For instance,
Carbon Nanotube-Inorganic Hybrids
metallic SWCNTs find application as leads in a nanoscale
circuit, whereas nanoscale Schottky-type field-effect tran-
sistors would require semiconducting SWCNTs. As the
synthesis of electronically pure monochiral nanotubes has
yet to be accomplished, the current strategy is to separate
the two CNT types. Much progress was achieved in 2003,
with the introduction of four separation routessthree
chemical and one physical.
The chemical approach utilizes the stronger interactions
of various chemical molecules with one type of SWCNT
over the other. For instance, surfactants, such as octadecy-
lamine,56 have a strong affinity toward semiconducting CNTs,
while diazonium reagents,57 biomolecules, and DNA58,59
favor the metallic type. Once attached to the CNTs, they
form larger aggregates, which can be separated easily by
standard separation techniques such as ion-exchange chro-
matography or microfiltration. In a similar way, CNTs have
been separated according to their length,60 diameter,59 and
chirality.61
The physical method of Krupke et al. extracts metallic
SWCNTs based on AC dielectrophoresis.62,63 When exposed
to an alternating electric field (10 MHz), the SWCNTs
develop an induced dipole moment, which causes the two
types of CNTs to migrate along the electric field gradient in
opposite directions. The metallic SWCNTs deposit onto the
electrodes, while the semiconducting SWCNTs remain in
suspension. This process is limited by the tendency of
SWCNTs to agglomerate into mixed bundles, causing a
coadsorption of both CNT types. These separation techniques
can be improved by using better dispersions (e.g., with
surfactants).64 However, much improvement is still required,
particularly in scaling-up the separated CNT amounts.
2.2.3. Functionalization
Functionalization of CNTs remains one of the most studied
areas in the CNT research field, and many review articles
have been dedicated to this topic.65-70 In general, CNTs can
be functionalized by (a) covalent attachment of chemical
groups through bonding to the π-conjugated skeleton of the
CNT or (b) noncovalent adsorption or wrapping of various
functional molecules.
The most common covalent functionalization involves the
addition of carbonyl and carboxyl groups via an aggressive
treatment with a mixture of HNO3/H2SO4 or by plasma
etching.69 The latter technique can also be used to introduce
basic functionalities when applied in a nitrogen atmosphere.
Carboxyl groups may then be acylated with thionyl chloride
to make a basis for various amine compounds71 or to attach
to various proteins and DNA.72-74 Other commonly used
chemical reactions to attach organic groups include cycload-
ditions (e.g., Bingel, Diels-Alder), electrophilic and nu-
cleophilic additions, ozonolysis, halogenation, or radical
reactions (oxidative and reductive).65,68,70
Covalent functionalization of CNTs has been shown to
be an efficient method for increasing their solubility and
chemical reactivity.75,76 However, it also introduces additional
structural defects and disrupts the delocalized electron system
in the CNT sidewalls, and consequently alters the electronic
and mechanical properties to a degree that would significantly
affect the performance in hybrid materials.44
In contrast, noncovalent functionalization utilizes van der
Waals interactions and hydrogen bonding. Various surfac-
tants and polymers have been applied to enhance the CNTs’
solubility in hydrophilic solvents and to increase their
Chemical Reviews, 2010, Vol. 110, No. 3 1353
dispersion in polymer or ceramic matrices.41 Currently, there
is much interest in functionalizing CNTs with biomolecules
such as peptides or DNA.66,77,78 Another approach uses the
well-known π-π interactions between aromatic compound
(e.g., porphyrins, pyrens) and the delocalized electron system
in CNTs.79,80
The CNT reactivity is directly related to the π-orbital
mismatch caused by an increased curvature. Therefore, a
distinction must be made between the sidewall and the end-
caps of a nanotube. The sidewalls can be considered as
curved graphite, while the tips are reminiscent of the structure
of a fullerene hemisphere and are thus relatively reactive.
Hence, most reactions will occur first at the tips and then on
the surface, especially where structural defects are present.
The difference in reactivity can be used to selectively open
and functionalize (covalently) the tips of CNTs, while the
sidewalls remain inert.65 This route has frequently been used
to fill CNTs with inorganic compounds (section 2.3) and to
attach metal particles onto the tips of CNTs (nanoplug81).
In contrast, by using nonconvalent interactions, surfactants,
aromatic compounds, and biomolecules, the inorganic nano-
particles can be attached equally on both tips and sidewalls.
Thus, this route is ideal for the deposition of uniform and
complete inorganic coatings.
2.3. Filling CNTs
The earliest attempt toward CNT-inorganic hybrids (in
1993) was the filling of MWCNTs with metal oxides
(PbO82 and Bi2O583), carbides (Y3C and TiC84), and metals
(Ni85). Because of their larger inner diameter (5-50 nm)
compared with SWCNTs (1-1.5 nm), most efforts had been
spent on filling MWCNTs, and it was not until five years
later that Sloan et al. reported the filling of SWCNTs with
RuCl3.86 Since then, many atoms, molecules, and compounds
have been incorporated into both MWCNTs and SWCNT,87
including such fascinating molecules as fullerenes
(C60@CNT, “peapods”88).
Initially, the motivation arose mainly from the prospect
of forming encapsulated or (upon oxidation) freestanding
inorganic nanowires. Soon, the possibility was considered
of using the CNTs as nanocapsules, to protect the encapsu-
lated material from reaction/oxidation due to contact with
the atmosphere, and even as nanoreactors, whose confined
reaction volume was expected to yield nanomaterials with
new crystal structures and chemical compositions and with
novel properties.
There have been various methods of filling, mainly
depending on the cavity diameter to be filled and on physical
properties of the material being inserted (e.g., viscosity,
surface tension). The inserted material can be solid, liquid,
or vapor; hence many physical properties have to be
considered, including solubility, melting point, surface ten-
sion, boiling point, viscosity, and decomposition tempera-
ture.87 Furthermore, the inner diameter of the CNTs will
determine the maximum size of the inserted molecules or
compounds and, hence, the filling efficiency.89
CNTs can be filled in situ during the synthesis process
(either arc-discharge90 or CVD91), with the advantages that
the CNT integrity remains intact and the encapsulated
material is completely protected from the postsynthesis
atmosphere. Furthermore, this allows filling with compounds
with surface tensions too high for other processes. However,
the choice of materials is restricted to those that can be used
as catalysts for the CNT growth (Fe, Co, Ni, etc). In contrast,
1354 Chemical Reviews, 2010, Vol. 110, No. 3 Eder

Table 5. Comprehensive List of Inorganic Compounds Used in CNT Hybrid Materials Including Applied Synthesis Techniques and
Tested and Potential Applications
inorganic compound synthesis routes applications Nr
Al2O3 ex situ - noncovalent 149,386,387
field emission >5
in situ - hydrothermal233 oxidation resistance389
in situ - sol-gel218
chemical vapor deposition259,388
physical vapor deposition247,248
BaSrO3 physical vapor deposition245 field emission351,390 1-2
CeO2 in situ - sol-gel200 heterogeneous catalysis 1-2
in-situ - hydrothermal199,233,391 gas sensors
Co3O4 in situ - sol-gel210 magnetics 3-5
chemical vapor deposition253 batteries392
Cu2O in situ - hydrothermal229,230,393,394 photocatalysis 1-2
Eu2O3 ex situ -non covalent135 diodes, lasers 2
in situ - hydrothermal232
FexOy ex situ - covalent395 magnetics 3-5
in situ - hydrothermal226,231,396 biosensors324
heterogeneous catalysis397
HfO2 physical vapor deposition247,248 oxidation resistance 1
MgO physical vapor deposition242 field emission242,342,350,353,398 3-5
electrocatalysis111
MnO2 in situ - electrochemical167-169 electrocatalysis399 3-5
in situ - electrodeposition179,334 heterogeneous catalysis
in situ - hydrolysis399 supercapacitors110,167,179,188,332,333,400,335
chemical vapor deposition400 oxidation resistance168
MoO2 in situ - hydrolysis292 electrocatalysis292 1
NiO in situ - sol-gel209,253 supercapacitors181,209,330,331 1
chemical vapor deposition253
RuO2 in situ - electrodeposition182,183 supercapacitors182,207,208,237,246,251,338 >5
in situ - sol-gel207,208,216 biosensors
in situ - hydrothermal237 heterogeneous catalysis216,294,298
physical vapor deposition158,246
chemical vapor deposition183,251,261
SiO2 ex situ - covalent114,401 field emission346 >5
in situ - sol-gel118,215,218,219,402 oxidation resistance215,405
physical vapor deposition243,247,248,403,404
SnO2 in situ - sol-gel194,198,199,201,202,406 gas sensors201,203,210,319,321,410-413 >5
in situ - hydrothermal199,234,235,407,408 electrocatalyis302
chemical vapor deposition250,409 nanofluids414
batteries202,408,415
TiO2 ex situ - covalent114 photocatalysis196,205,206,224,278,280,282,421,422,281 >5
ex situ - noncovalent149 optoelectronics112,423
in situ - electrodeposition180,416 biosensors
in situ - microemulsion177 electrocatalysis197,293,424
in situ - sol-gel55,191-193,195,204,218,417-419 supercapacitors180,415
in situ - hydrothermal224,225 batteries177,425
chemical vapor deposition259,420 oxidation resistance195
VO2, V2O5 in situ - electrochemical170 batteries336,337,426 2
heterogeneous catalysis170
WO3 in situ - electrochemical300 gas sensors314,427 2
heterogeneous catalysis300
ZnO in situ - electrochemical428 photocatalysis221,224 >5
in situ - microemulsion178 optoelectronics220,255
in situ - sol-gel220,221,429 diodes, lasers244
in situ - hydrothermal224,227 field emission240,348
physical vapor deposition244,247,248,254,348
chemical vapor deposition256-258,430
ZrO2 ex situ - noncovalent318 oxidation resistance 3-5
in situ - hydrothermal431,432 dielectric devices260
physical vapor deposition247,248 heterogeneous catalysis
chemical vapor deposition260 chemical sensors318
carbides (Fe, W, Ta, Ti) in situ - electrochemical175 heterogeneous catalysis 3-5
in situ - sol-gel214 electrochemistry211
chemical vapor deposition213,433 field emission213,434
chalcogenides (Zn, Cd, Hg; X ) S, Se, Te) ex situ - covalent134,435 optoelectronics165 >15
ex situ - noncovalent166,436
in situ - electrochemical164-166
in situ - hydrothermal228
in situ - sol-gel217
nitrides (Ti,Fe) in situ - sol-gel212 field emission252 2-3
physical vapor deposition211
chemical vapor deposition252

ex situ filling enables the use of a variety of compounds but to open the CNT tips (heating in air or treatment in oxidizing
typically requires a multistep process including preoxidation acids)83,92 and post-treatment to remove any excess material
Carbon Nanotube-Inorganic Hybrids Chemical Reviews, 2010, Vol. 110, No. 3 1355

via washing or vacuum annealing.93 The filling is either

carried out from the gas phase (sublimation or capillary
condensation)94,95 or from the liquid phase (capillary wetting
from suspension, solution, or melt).96,97 The filling strategies
with their various advantages and problems were summarized
and discussed recently by Monthioux et al.,87 as well as in
earlier reviews.98,99
Although both MWCNTs and SWCNTs have been filled
with a vast number of compounds, little is known about their
properties and potential in applications. This has been
frequently blamed on high impurity levels in the synthesized
hybrids, the lack of bulk quantities, and the need for
specifically designed measurement devices.87 In general, the
distinction should be made between the intrinsic properties
of the filler, the altered properties of the encapsulated material
due to the confined space, and novel properties arising from
interactions between the filler and CNTs (a requirement for
CNT-inorganic hybrids). For example, theoretical calcula-
tions predict a change from semimetallic to semiconducting
properties of a HgTe crystal upon its encapsulation in
SWCNTs.100 Although this system had shown early promise
of demonstrating the modification of electronic properties
in filled CNTs due to the geometric constraints, experimental
results have so far been rather disappointing. To date, there
has been no evidence for charge transfer between the filler
and the CNTs, band gap modulation, or any change in
thermal or electric properties.

3. Synthesis of CNT-Inorganic Hybrids

Instead of filling CNTs (see previous section), the inor-
ganic component can be attached to the outer surface of
CNTs to produce novel hybrid materials. The various
synthesis strategies for CNT-inorganic hybrids can be
categorized as ex situ and in situ techniques. The ex situ
(building block) approach first produces the inorganic
component in the desired dimensions and morphology
(typically spherical nanoparticles), then modifies and attaches
this component to the surface of CNTs via covalent,
noncovalent, or electrostatic interactions. In contrast, the in
situ approach carries out the synthesis of the inorganic
component in the presence of pristine or functionalized
CNTs, onto which the inorganic material grows as particles,
nanowires, or thin films.
As the ex situ process has been discussed in recent
reviews,101,102 this review summarizes this approach with a
few significant examples and then focuses on the in situ
approach for the formation of oxides, nitrides, and carbides.
Table 5 summarizes the inorganic materials hitherto used in
CNT-inorganic hybrids, along with their synthesis technique
and potential applications. The references in the application
column represent initial studies on CNT-inorganic hybrids
tested for the desired application and will be discussed in
section 4.
3.1. Ex Situ Approach: Attaching Nanoparticles
to CNTs
In this ex situ or building block approach, nanoparticles
(mostly metals or semiconductor QDs) are attached to the
CNTs via linking agents that utilize covalent, noncovalent,
or electrostatic interactions. In this approach, either the
inorganic nanoparticles or the CNTs (or both) require
modification with functional groups. The type of function-
alization and, thus, the strength of interaction determine the
Figure 2. Common ex situ process for the covalent attachment of
Au nanoparticles to amino- or mercapto-terminated CNTs. Redrawn
from ref 103.
distribution and concentration of the inorganic nanoparticles
on the CNT surface.
3.1.1. Covalent Interactions
Carboxyl groups on the surface of acid-treated CNTs (see
section 2.2.3) are often used to attach amine-terminated or
mercapto-terminated inorganic nanoparticles via amide
bonds.103 This can be achieved either by directly linking
amine-terminated or mercapto-terminated nanoparticles with
the carboxyl groups or by modifying these carboxyl groups
into thiol groups, which then anchor to colloidal nanoparticles
(Figure 2). In a similar way, QDs have been linked by first
stabilizing them with a mixed monolayer of trioctylphosphine
oxide (TOPO) and 2-aminoethanethiol.104 The resulting
amino-functionalized QDs then reacted with the carboxylic
groups of the acid-treated CNTs to form amide bonds.
Banerjee et al. used a multistep procedure to attach
modified CdSe QDs to functionalized CNTs105 (Figure 3):
(1) SWCNTs were first oxidized in a H2SO4/KMnO4
solution, and the resulting carboxylic groups were reacted
with ethylenediamine to form amine-terminated amide
groups. (2) CdSe nanoparticles were stabilized with TOPO,
which was substituted by a thiol ligand, using p-mercapto-
benzoic acid, to form acid-terminated CdSe nanocrystals. (3)
The modified CdSe nanoparticles were linked to the func-
tionalized CNTs in the presence of 1-ethyl-3-(3-dimeth-
ylaminopropyl) carbodiimide hydrochloride (EDC). In a
similar way, thiol-stabilized ZnS-capped CdSe QDs were
protected with 2-aminoethanethiol and linked to the acid-
terminated CNTs in the presence of EDC.106
1356 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 3. Multistep process for linking modified CdSe QDs to
CNTs via amide bonds, involving (1) the oxidation of CNTs in
KMnO4 and functionalization with ethylenediamine and (2) the
thiolization of CdSe and termination with acid groups. Redrawn
from ref 105.
Among the various metal nanoparticles, gold has been used
most frequently due to its excellent potential in biosensing
and other medical applications.107 The Au nanoparticles have
been linked to the acid-terminated CNTs using aminothiols,
bifunctional thiols, or thioether bonds.108,109
In contrast, metal oxides can be attached to the carboxyl
groups without any linking agent due to their hydrophilic
nature, as recently demonstrated for MnO2,110 MgO,111
TiO2,112 and Zr(SO4)2.113 However, the authors observed only
weak interactions between the oxides and the acid-terminated
CNTs, resulting in rather nonuniform distributions of the
nanoparticles. Better adhesion was observed when capping
agents were used. For instance, Sainsbury and Fitzmaurice
produced capped TiO2 and SiO2 nanoparticles (d ≈ 4-5 nm)
via a standard sol-gel process using titanium tetraisopro-
poxide (TTIP) and tetraethyl orthosilicate (TEOS) as precur-
sors with cetyltrimethyl ammoniumbromide (CTAB) as the
capping agent.114 MWCNTs were modified with 2-amino-
ethylphosphoric acid and then mixed with the oxide nano-
particles. The authors showed that the phosphonic acid
groups on the CNTs were well-distributed and provided an
excellent driving force for the attachment of TiO2 nanopar-
ticles (Figure 4).
3.1.2. Noncovalent Interactions
Alternatively, the nanoparticles can be attached to pristine
CNTs via van der Waals interactions, hydrogen bonding,
π-π stacking, or electrostatic interactions. For instance,
surfactants such as sodium dodecylsulfate (SDS) were used
to decorate MWCNTs with nanoparticles of pure Pt,115 EuF3,
and TbF3116 particles. Similarly, Whitsitt et al. evaluated
various surfactants for their ability to facilitate the deposition
of SiO2 NPs onto SWCNTs.117,118 In the presence of SDS,
the SiO2 nanoparticles were deposited around bundles of
SWCNTs, while dodecyl dimethyl ammoniumbromide
(DTAB) enabled a significantly better dispersion and de-
bundling of the SWCNTs so that individual nanotubes were
coated.
The functionalization of the inorganic nanoparticles with
hydrophobic capping agents provides another route. As an
Eder
Figure 4. Example of an ex situ attachment of SiO2 and TiO2
nanoparticles to functionalized CNTs via silane and phosphonic
acid bonds, respectively. Redrawn from ref 114.
Figure 5. Mixed nonconvalent attachment of modified Au nano-
particles to CNTs via hydrogen bonds and hydrophobic interactions.
Redrawn from ref 121.
example, Au clusters were first modified with a monolayer
of octanethiols119 or dodecanethiols120,121 and then simply
dispersed in a suspension of pristine SWCNTs in dichlo-
romethane. The advantage of this approach is that both
coverage and morphology of the hybrid materials can be
controlled by modifying the length and functionality of the
chains in the capping agents.122
Han et al. used both hydrophobic and hydrogen-bond
interactions in their work.121 The authors modified Au
nanoparticles (2-5 nm) with a mixed monolayer of de-
canethiol and mercaptoundecanoic acid, providing mixed
chemical functionality, and immersed them in a suspension
of acid-treated CNTs. The alkyl chains of decanethiol offered
hydrophobic interactions with the unmodified part of the
CNT surface, while the acid-modified nanoparticles formed
strong hydrogen bonds with the carboxylic groups of CNTs
that could not even be dissembled upon heat treatment at
300 °C (Figure 5). Furthermore, as the acid groups are
typically concentrated around the tip of the CNTs (especially
when using weak oxidizing conditions), this approach has
the potential for selective immobilization of one type of
nanoparticle at the tips of the CNTs via H-bonds, while
adsorbing another type hydrophobically on the sidewalls.
3.1.3. π-π Stacking
Similar to noncovalent attachment, this approach uses the
moderately strong interactions between delocalized π-elec-
trons of CNTs and those in aromatic organic compounds,
such as derivatives of pyrene,123-125 porphyrins,126-128 ph-
Carbon Nanotube-Inorganic Hybrids
Figure 6. Examples of linking agents and ligands used to attach inorganic nanoparticles to pristine CNTs via π-π interactions: (a) pyrene
derivatives, (b) porphyrins and phthalocyanines, (c) triphenyl phosphines, and (d) benzyl alcohol.
thalocyanines,129 or combinations thereof,130 as well as benzyl
alcohol55 or triphenylphosphine (Figure 6).131 These mol-
ecules are often modified with long alkyl chains that are
terminated with thiol, amine, or acid groups, which can then
connect to the inorganic nanoparticles and enable their
attachment to pristine CNTs via π-π stacking.
For example, Au nanoparticles have been anchored with
thiol-terminated132 or amine-terminated133 pyrene molecules.
Li et al. used 1-aminopyrene to immobilize CdS nanoparticles
as large as 20 nm,134 while pyrene derivatives with a
carboxylic termination were used to anchor magnetic nano-
particles such as Co or Fe3O4.135 In contrast with pyrene
derivatives, compounds with phenyl groups typically support
weaker π-π interactions but are smaller in size and thus
can attract more nanoparticles. For instance, very dense layers
of metal nanoparticles close to the surface of CNTs have
been achieved for Pt nanoparticles using triphenylphosphine
(PPh3) as a linking agent.131
One of the major advantages of this approach is that
the pyrene compounds remain strongly adsorbed on the
CNT surface after workup steps (e.g., washing, filtration)
and thus provide enhanced solubility and allow continuous
redispersing of the modified CNTs in various aqueous and
organic solvents. Furthermore, spectroscopic experiments
on CNT-Pt136 and CNT-porphyrin hybrids126,137 revealed
an enhanced charge transfer from the inorganic nanoparticles
to the CNTs, mediated by the aromatic compound. This was
also observed for attached Co and Fe3O4 nanoparticles,135
whose magnetic and electronic properties were altered due
to a strong electron transfer. Moreover, this effect is tunable
by the length of the chain.
3.1.4. Electrostatic Interactions
The fourth approach utilizes electrostatic interactions
between modified CNTs and inorganic nanoparticles. Among
the known examples, the deposition of ionic polyelectrolytes
to attract charged nanoparticles is the most common
route.138-142
The choice of polyelectrolyte determines whether the
CNTs are positively or negatively charged. For instance,
oxidized CNTs have been coated with a thin film of
poly(diallyldimethylammonium chloride) (PDDA), a cationic
Chemical Reviews, 2010, Vol. 110, No. 3 1357
polyelectrolyte, which served as the real template for
negatively charged Au nanoparticles.141 Alternatively, by
depositing a second layer of a negatively charged polymer,
such as poly(sodium 4-styrenesulfonate) (PSS), on top of
the PDDA layer, the surface of the CNT could be negatively
charged and would then accept positively charged nanopar-
ticles such as SiO2.142,143 The order of polymer deposition
can also be inverted as shown by the group of Liz-Marzan,
who first adsorbed PSS to provide a stable dispersion of
negatively charged CNTs.142,144,145 After the deposition of a
thin layer of SiO2 via hydrolysis of TEOS, a second layer
of the positive charged PDDA was added, which provided
an excellent anchor for CdTe nanoparticles (Figure 7). The
SiO2 spacer in this rather complex hybrid material minimized
the photoluminescence quenching by the CNTs and preserved
the quantum confinement effects.144
These polyelectrolytes typically bond covalently to the
functional groups on the CNT, in contrast to polyethylene-
imine (PEI), which interacts with CNTs via physisorption.146
Another example is the use of Nafion to make the surface
of CNTs negatively charged.147 Nafion is a biocompatible
perfluorosulfonated polymer with a polar side chain and with
unique ion-exchange properties. Luong et al. immobilized
Pt nanoparticles on the modified CNT surface by strong
electrostatic interactions and tested this hybrid material for
use in fuel cells.147 Continuous metal coatings were obtained
in a similar way, as shown for Au.148 The metal nanoparticles
were stabilized with tetraoctylammonium cations and dis-
persed in chloroform, into which the CNTs were immersed.
Subsequent heating caused the nanoparticles to melt and
amalgamate to form a metal nanowire with the CNTs as the
cores. Sun et al. deposited Al2O3, ZrO2, and TiO2 nanopar-
ticles on charged CNTs in a slightly modified way.149 CNTs
were pretreated in NH3 at 600 °C to induce a positive surface
charge. The addition of PEI increased the positive charges
even further and enabled a better dispersion. Commercially
available R-Al2O3 and 3Y-TZP were then dispersed in
poly(acrylic acid) (PAA), which provided a negative surface
potential over a wide range of pH values. Upon mixing, the
Al2O3 and ZrO2 nanoparticles formed strong electrostatic
attractive interactions and covered the CNT surface completely.
Sun et al. attached nanocrystals of TiO2 to acid-treated
SWCNTs, also using PEI as a modifier (Figure 8).150 In the
1358 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 7. Example of the deposition of inorganic nanoparticles on CNTs via electrostatic interactions using polyelectrolytes: (1) polymer
wrapping around a CNT using PSS, (2) self-assembly of PDDA on top of PSS, and (3) dense layer of inorganic nanoparticles around the
modified CNT. Reprinted with permission from ref 145. Copyright 2006 Royal Society of Chemistry.
Figure 8. Two-step process for the attachment of TiO2 nanopar-
ticles to CNTs: (1) modification of the nanoparticles with positively
charged PEI, containing primary, secondary, and tertiary amines,
and (2) attachment via amide linkage or electrostatic interactions.
Redrawn from ref 150.
first step, TiCl4 was mixed with PEI, whose charged amino
groups (25% primary, 50% secondary, and 25% tertiary
amine) were protonated at a pH of 8. The positively charged
amino groups of PEI accelerated the hydrolysis of TiCl4 into
TiO2-nanoparticles and stabilized them electrosterically.
These amine-terminated TiO2-nanoparticles with positive
charge then attached to acid-treated SWCNTs, either via
amide linkage (Ieft-hand side) or through electrostatic
interaction (right-hand side). Another route has been sug-
gested by Jerome et al.,151 who grafted MWCNTs with poly-
2-vinylpyridine (P2 VP) to form carboxylate terminated alkyl
chains, onto which positively charged magnetic Fe3O4
nanoparticles were anchored. The main advantage of the
polymer route is that the polymers provide a very dense,
uniform distribution of either negative or positive charges
over the entire CNT surface, which enables very dense
assemblies of inorganic nanoparticles.
In summary, these few examples demonstrate the simplic-
ity and feasibility of the ex situ approach, which remains
the method of choice for the deposition of metal nanoparticles
and quantum dots. The main advantage is the possibility of
Eder
using prepared nanoparticles with controlled morphology,
structure, shape, and size, and, therefore, a good structure-
property relationship. The downside of this route, however,
is the need to chemically modify either the CNTs or the
inorganic component. This process is often work-intensive,
and functionalization alters both the surface chemistry of the
CNTs and also, particularly for SWCNTs, their physical
properties. Furthermore, the use of predefined building blocks
restricts the synthesis of novel hybrid materials and, thus,
the development of new physical properties.
3.2. In Situ Synthesis Directly on the CNT
Surface
The inorganic compound can also be formed directly on
the surface of pristine or modified CNTs. The main advantage
of this route is that the inorganic compound can be deposited
as a continuous amorphous or single-crystalline film with
controlled thickness, or as discrete units in the form of
nanoparticles, nanorods, or nanobeads. Furthermore, the
CNTs may act as a support to stabilize uncommon or even
novel crystal phases or prevent crystal growth during
crystallization and phase-transformation processes. Finally,
a variety of chemical and physical synthesis techniques can
be applied. The deposition can be carried out either in
solution via electrochemical reduction of metal salts, electro-
or electroless deposition, sol-gel processing, and hydro-
thermal treatment with supercritical solvents, or from the gas
phase using chemical deposition (CVD or ALD) or physical
deposition (laser ablation, electron beam deposition, thermal
evaporation, or sputtering).
3.2.1. Electrochemical Techniques
Electrochemistry is a powerful technique for the deposition
of various nanoparticles (especially metal particles), as it
enables effective control over nucleation and growth.152,153
Most research has been conducted on the deposition of noble
metals and alloys such as Pd,152,154 Pt,153-155 Au,154 Ag,153
and bimetallic Pt-Ru,156 as they are the metals of choice
for applications like heterogeneous catalysis and electroca-
talysis,157 supercapacitors,158 gas sensors,159 and biosensors.160
In general, metal nanoparticles are obtained via reduction
of metal complexes, such as H[AuCl4], K2[PtCl4], and
(NH4)2[PdCl4], by chemical agents (chemical reduction), or
by electrons (electrodeposition). The size of the metal
nanoparticles and their coverage on the sidewalls of CNTs
can be controlled by the concentration of the metal salt and
various electrochemical deposition parameters, including
nucleation potential and deposition time.152,153,161
Carbon Nanotube-Inorganic Hybrids Chemical Reviews, 2010, Vol. 110, No. 3 1359

Figure 9. Example of the deposition of MnO2 on CNTs via chemical reduction. The multistep process involves (1) the oxidation of
MWCNTs with KMnO4 to form hydroxyl groups, (2) the precipitation of permanganate ions, and (3) their reduction with citric acid to
MnO2. Reprinted with permission from ref 168. Copyright 2007 Elsevier Publishing.

3.2.1.1. Chemical Reduction and Oxidation. These tech-
niques involve reactions, in which the reduction of the
precursor is carried out with liquid or gaseous reducing
agents with the aid of heat, light, ultrasound, microwave, or
supercritical CO2.162,163 As an example, Cao et al. mixed
MWCNTs with sulfur powder and CdCl2 with the aid of
ultrasonication.164 The sulfur was first reduced with KBH4
to form S2- ions, which then reacted with the Cd2+ ions to
form CdS nanoparticles on the surface of the CNTs. The
authors observed a uniform distribution of the fibers, which
caused a significant charge transfer to the CNTs and
enhanced photovoltaic response. Robel et al. used a similar
approach with CdI2 and Na2S.165 A very uniform dispersion
of metal nanoparticles was obtained by Li et al., who chose
noncovalent functionalization with 1-aminopyrene and used
electrostatic interactions for the deposition of Pt and CdS
nanoparticles.166 Under reaction conditions, the amino groups
were slightly positively charged and thus attracted (PtCl6)2-
ions, which were then reduced with NaBH4 to give Pt
nanoparticles.
Sivvakkumar et al. deposited MnO2 via chemical reduction
of KMnO4.167 The authors suspended the CNTs in Na-p-
toluene sulfonate and pyrrole, which polymerized with the
aid of ultrasonication. KMnO4 was then slowly added and
reduced with acetonitrile to form hydrous MnO2. A very
elegant variation of this process is shown in Figure 9 and

Figure 10. Example of a water-in-oil microemulsion technique. Formation of hollow Zn(OH)2 spheres on CNTs and their subsequent
calcination to dense ZnO nanoparticles. Redrawn from ref 178.
1360 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 11. Electrodeposition of Pd nanoparticles on MWCNTs, via (1) covalent functionalization of CNTs with aminobenzyl groups via
electrochemical reduction, (2) potentiostatic reduction of PdCl62- ions to Pd nanoparticles. Reprinted with permission from ref 152. Copyright
2004 Elsevier Publishing.
uses KMnO4 as both the oxidizer and reactant.168 In detail,
pristine MWCNTs were first oxidized under reflux with
KMnO4, which introduced exclusively hydroxyl groups on
the sidewalls of CNTs (step 1), in contrast to other oxidation
treatments, e.g., with oxidizing acids. These hydroxyl groups
then acted as anchors for Mn7+ ions (step 2), which
subsequently were reduced to Mn4+ with citric acid (step 3)
to form a coating of
-MnO2. For comparison, a pretreatment
of the CNTs in strong acids, which induces the formation
of carboxyl groups, resulted instead in γ-MnO2.169 Therefore,
this work provides an interesting example of the effect of
the CNT surface chemistry on the crystal structure of the
inorganic coating.
In contrast to MnO2, the deposition of other metal oxides
typically requires oxidizing rather than reducing processes.
For instance, Huang et al. added acid-treated CNTs to a
solution containing ammonium metavanadate.170 The ad-
sorbed VO2+ ions were then oxidized with oxalic acid to
V2O5.
Another common method involves the reduction of the
adsorbed precursor at high temperatures under a hydrogen
atmosphere. This approach has been applied to a variety of
metal nanoparticles such as Pt,171,172 Pd,173 Rh, and Ru,163,174
as well as to various carbides such as TiC, TaC, and NbC.175
A different approach combined the reduction of cationic
precursors by hydrogen with a water-in-oil microemulsion
technique for the deposition of metal157,176 or oxide nano-
particles.177 Sun et al. used a slightly modified process, shown
in Figure 10.178 The authors mixed the CNTs first with an
aqueous solution of sodium dodecylbenzene sulfonate (NaD-
DBS) and then with a mixture of Triton-X and cyclohexane,
which resulted in very small water droplets on the CNT
surface. Upon adding zinc acetate as the metal precursor,
the Zn2+ ions concentrated in the aqueous phase and then
reacted with NH3 or LiOH to form spherical, hollow ZnOH
nanoparticles. Subsequent calcinations oxidized them to
create small and dense ZnO nanoparticles. In all cases, the
microemulsion technique produced fine dispersions of small
nanoparticles.
Eder
3.2.1.2. Electrodeposition. Many of the above-mentioned
reduction/oxidation techniques are very time-consuming and
as such allow impurities in the bath solution to be incorpo-
rated into the inorganic phase. In contrast, electrodeposited
nanoparticles, especially noble metals such as Pt, Pd, or Au,
are formed faster upon reduction of the corresponding metal
salts under an applied potential and, thus, exhibit higher
purities as well as a good adhesion to the CNT surface.156,161
In most cases, simple van der Waals interactions between
the CNTs and the nanoparticles seem to be sufficient to
provide strong enough adhesion.
However, functionalization of CNTs can still enhance the
dispersion and decrease the size of nanoparticles, as dem-
onstrated by Guo et al. (Figure 11).152 The authors function-
alized MWCNTs covalently with 4-aminobenzene via the
direct electrochemical reduction of the diazonium salt of
nitrobenzene. These groups provided good electrostatic
attractions for [PdCl6]2- ions, which were then reduced to
very small Pd nanoparticles (2.5 nm in diameter) via
potentiostatic reduction in 0.1 M H2SO4.
Although most research currently concentrates on the
electrodeposition of metal nanoparticles, there have also been
a few reports on electrodeposited metal oxides. As an
example, Lee et al. drop-casted acid-treated CNTs on a Ni
working electrode and used Pt wires and saturated calomel
as counter and reference electrodes, respectively.179 By
applying a potentiostatic method with MnSO4 at pH 5.6 and
a potential of 0.4 V, the authors deposited rather large
aggregates of MnO2, predominantly around the tips of the
CNTs. Similarly, Frank et al. used pristine SWCNT bucky
paper as a working electrode, using Pt wires as both reference
electrodes and counterelectrodes.180 TiCl3 was used as a
precursor and electrolyte and was kept at pH 2.5 with HCl/
Na2CO3. The deposition was then carried out via galvano-
static oxidation with 1 mA/cm2 and resulted in a rather
irregular and partial coating of a mixture of anatase and
TiO2-B.
The galvanostatic approach (3 mA/cm2) has also been
applied to codeposit Ni and Co oxides from their nitrates,
Carbon Nanotube-Inorganic Hybrids
Figure 12. (a) Scheme of the substrate-enhanced electroless
deposition (SEED) of nanoparticles using the CNT as cathode. (b)
SEM image of Pt nanoparticles deposited on SWCNTs. Reprinted
with permission from ref 185. Copyright 2005 American Chemical
Society.
with Pt and saturated calomel as the counterelectrodes and
reference electrodes, respectively.181 By cycling the applied
potential from 0 to 0.5 V, the authors could obtain a more
uniform and stabilized coating. This was also demonstrated
by Kim et al., who produced a continuous 3 nm thick coating
of RuO2 via the potential cycling method while varying the
potential from -200 to 1000 mV with a scan rate of 50 mV/
s.182 They observed that the gas atmosphere during the
postannealing process had a significant effect on the mor-
phology of RuOx. Furthermore, heating in argon produced
more uniform, spherical nanoparticles, while annealing in
O2 resulted in elliptical RuO2 nanoparticles.183
One advantage of this technique is that the electrodepo-
sition occurs to the same extent on both the sidewalls of the
tubes and the tips.184 Consequently, the presence of carboxyl
or hydroxyl groups as activators is not required. Furthermore,
He et al. showed that the high purity is also of great
advantage for the codeposition of bimetallic nanoparticles,
which also exhibited enhanced electrocatalytic activity
toward the oxidation of methanol.156 The authors deposited
Pt-Ru nanoparticles potentiostatically (-0.25 V) with
diameters of 60-80 nm onto oxidized CNTs, starting from
a solution containing ruthenium chloride and chloroplatinic
acid in 0.5 M H2SO4. However, despite the simplicity and
elegance of this technique for the deposition of metal
nanoparticles especially, the major drawback of electrodepo-
sition is that it is difficult to produce bulk quantities.
3.2.1.3. Electroless Metal Deposition. Electroless metal
deposition was first mentioned by Qu et al.,185 who observed
that Au and Pt nanoparticles formed spontaneously on
Chemical Reviews, 2010, Vol. 110, No. 3 1361
SWCNTs immediately after immersing the CNTs in a
solution of HAuCl4 or Na2PtCl2. Because of the absence of
any reducing agent, the authors suggested that a direct redox
reaction via electron transfer occurred between the metal ions
and the CNTs, which acted as the cathode.186,187 With the
exception of MnO2,188,189 this substrate-enhanced electroless
metal deposition (SEED) could only be successfully applied
to noble metals, while Ag+, Ni2+, and Cu2+ had redox
potentials too low to be reduced in the same way.
3.2.2. Sol-Gel Process
The sol-gel process is a versatile, solution-based process
for producing various ceramic and glass materials in the form
of nanoparticles, thin-film coatings, fibers, or aerogels and
involves the transition of a liquid, colloidal “sol” into a solid
“gel” phase.190 Typical starting materials for the preparation
of the sol include metal salts or metal organic compounds,
such as metal alkoxides, which undergo a series of hydrolysis
and condensation reactions to form a colloidal or polymeric
sol. Upon aging, the sol forms a wet inorganic continuous
network with oxo (M-O-M) or hydroxo (M-OH-M)
bridges, creating a gel. Subsequent drying under supercritical
conditions converts the gel into a low-density, highly porous
aerogel, while drying induced by heating typically results
in a xerogel (low temperature) or a dense ceramic (high
temperature). The sol-gel process can be used to produce
ultrafine and uniform ceramic powders, thin-film coatings,
monolithic glasses and ceramics, membranes, and, with a
suitable viscosity, ceramic fibers. It is, furthermore, a cheap
and low-temperature technique that allows fine control of
chemical composition and the introduction of lowest con-
centrations of finely dispersed dopants, such as organic dyes
and rare earth metals. One of the major drawbacks is that
the product typically consists of an amorphous phase rather
than defined crystals and, thus, requires crystallization and
postannealing steps.
In general, the sol-gel process has emerged as the most
common technique to synthesize CNT-inorganic hybrids.
Early attempts concentrated on the dispersion of CNTs
within a matrix of inorganic nanoparticles. Vincent et al.
synthesized TiO2 nanoparticles using metal organic pre-
cursors and acetic acid as a gelator.191 They observed that
the dispersion of pristine CNTs was more stable when the
TiO2 nanoparticles had been produced in the presence of the
CNTs (in situ) compared with the simple mixing of the two
materials. Upon reducing the amount of TiO2 with respect
to CNTs, Jitianu et al. obtained a thin but rather irregular
and partial coating of TiO2 on CNTs.192,193 Typically, the
thickness of the coating can be controlled by various
parameters, such as the reaction time,194 the reaction com-
position, and the choice of metal precursor.195 For instance,
in the case of TiO2, the use of titanium tetraisopropoxide
(TTIP) produced irregular coatings,192,196 while tetraethoxy
orthotitanate (TEOT)192 or tetrabutoxy orthotitanate
(TBOT)195,197 enabled a more uniform deposition. The
sol-gel process was sometimes carried out under reflux,194
or with the aid of ultrasonication,198 microwave,199 or
magnetic agitation,200 in order to enable faster and simulta-
neous nucleation resulting in a more homogeneous coating.
3.2.2.1. Covalent. These early works used pristine CNTs,
whose hydrophobic nature provides little attractive interaction
with the inorganic compound and thus limits the quality of
the coating. Similar to the ex situ approach, the most common
approach to change the surface chemistry of CNTs is to treat
1362 Chemical Reviews, 2010, Vol. 110, No. 3 Eder

them in strong oxidizing acids (H2SO4-HNO3). This process

introduces a variety of organic groups, with limited control
over their number, type, and location, and causes surface
etching and shortening of the tubes. Consequently, the
inorganic coatings on acid-treated CNTs were often non-
uniform, although they provided better interaction in com-
parison with pristine CNTs.
Despite these drawbacks, most researchers have used such
acid-treated CNTs for various inorganic coatings, including
SnO2,198,199,201-203 TiO2,197,204-206 RuO2,207,208 CeO2,200 NiO,209
and mixed oxides (e.g., Co3O4-SnO2).210 Some oxide
coatings were then heated at 800-900 °C in a carbon- or
nitrogen-containing atmosphere (e.g., CH4, NH3) to produce
metal nitride and carbide coatings such as TiN and Fe2N,211,212
TiC,213 and WC.214
3.2.2.2. Noncovalent. In a similar way to the ex situ
approach, noncovalent attractions and π-π interactions can
be used to grow the inorganic component on the surface of
CNTs. For instance, Bourlinos et al. wetted pristine CNTs
with vinyl silane molecules via noncovalent interactions
between the vinyl groups and CNT surface.215 After con-
densation to an oligomeric siloxane network and subsequent
calcinations, the authors obtained small SiO2 nanoparticles
(5-12 nm), which were well-dispersed on the CNT surface.
Another approach was used by Cao et al., who modified
CNTs with surfactants such as sodium dodecylsulfate
(SDS).216 The hydrophobic aliphatic chain interacted with
the surface of the CNTs, while the hydrophilic end attracted
the metal ions of the RuCl3 precursor, which then reacted to
form RuO2. Similarly, Wei et al. used metal nitrates and Na2S
to attach fibers such as CdS, Ag2S, and HgS to the surface
of pristine CNTs.217
Recently, this author has developed a nondestructive,
simple process to coat pristine CNTs with TiO2 by using
benzyl alcohol as a surfactant.55,195 Benzyl alcohol adsorbs
onto the CNTs’ surface via π-π interactions with the
alcohol’s benzene ring (similar to pyrene derivatives), while
simultaneously providing hydrophilic hydroxyl groups for
the hydrolysis of the titanium precursor (TBOT) (Figure 13).
In contrast to the sample without benzyl alcohol (Figure 13a),
the addition of small amounts of benzyl alcohol resulted in
a very uniform coating that covered the whole CNT surface.
The work further showed that benzyl alcohol strongly
affected the phase transition from anatase to rutile, providing
very small and uniform rutile nanocrystals with very high
specific surface areas (60-100 m2/g) without too great a
hindrance of the anatase to rutile transformation.
3.2.2.3. Electrostatic. In contrast to the ex situ approach,
the use of electrostatic interactions for the in situ sol-gel
route has been demonstrated only for a few metal oxides.
As an example, Hernadi et al. used CNTs that had been Figure 13. (Top) Scheme of the beneficial role of benzyl alcohol
pretreated with SDS, dried, and redispersed in 2-propanol.218 in the in situ coating of pristine CNTs with TiO2. The benzene
Using metal halides as precursors, the authors could suc- rings of the alcohol adsorb onto the CNT surface via π-π
cessfully produce coatings of Al2O3, SiO2, and TiO2. On the interactions and at the same time provide a high density of hydroxyl
groups for the hydrolysis of the titanium precursor directly on the
other hand, the use of metal organic precursors (aluminum CNT surface. (Bottom) SEM images of pristine MWCNTs (A)
isopropoxide (AIIP), TEOS, TEOT) did not produce any without the use of benzyl alcohol and (B) with a titanium-to-benzyl
coating, but rather nanoparticles in solution. It appears that, alcohol molar ratio of 5. Reprinted with permission from ref 55.
in the latter example, the applied surfactant had a repressive Copyright 2008 Wiley-VCH.
role in the adhesion of the inorganic precursor due to different
polarities between surfactant-treated CNTs (ionic) and the
alkene groups of the precursor. This work provides yet surface.219 Using the same metal organic precursor (TEOS)
another example of the importance of surface chemistry. as the previous example, the negatively charged SiOx colloids
Seeger et al. modified the MWCNT with the polyelectro- could then easily attach via electrostatic interactions to form
lyte PEI, which provided positive charges on the CNT a coating of amorphous SiO2 (thickness ≈ 3-10 nm).
Carbon Nanotube-Inorganic Hybrids Chemical Reviews, 2010, Vol. 110, No. 3 1363

However, since the reaction was carried out at room

temperature, it was very time-consuming and exceeded 100 h.
In contrast, Vietmeyer et al. used acid-treated CNTs to
attract Zn2+ ions (using zinc acetate), which then reacted
with LiOH in an ice-water bath to form ZnO nanoparticles.220
While some nanoparticles were attached to the CNT surface
and tips, most of the nanoparticles formed large clusters
rather than a continuous coating. For comparison, Jiang et
al. enhanced the negative charge on the surface of acid-
treated CNTs by simply depositing a thin layer of SDS and
carried out the same reaction with zinc acetate and LiOH.221
This surface modification significantly enhanced the attractive
interactions with the metal ions and led to a more uniform
coating with ZnO crystals.

3.2.3. Hydrothermal and Aerosol Techniques

In recent years, many organic-inorganic hybrids have
been produced by hydrothermal techniques.5,222 In contrast
to standard sol-gel routes, the hydrothermal method typically
enables the formation of crystalline particles or films without
the need for postannealing and calcinations. Furthermore,
the forced crystallization enables the formation of inorganic
nanowires and nanorods.223
3.2.3.1. Vapor-Assisted, Polyol-Assisted Process. In this
simplest case of hydrothermal synthesis, pristine or acid-
treated CNTs were added to an aqueous solution of the
precursor and treated in an autoclave at temperatures between
100 and 240 °C to produce crystalline films of ZnO,224
TiO2,225 or Fe2O3.226 These works consistently produced
dense coatings of spherical or slightly elongated nanopar-
ticles. Zhang et al. used aligned CNTs, which they precoated
with a thin, amorphous layer of ZnO via magnetron sput-
tering227 (see section 3.2.4.1). Then they dissolved a fine ZnO
powder in NaOH at a pH of 10-12, which provided a
saturated solution of Zn(OH)42-. The modified CNT carpet
was then placed top-down in an autoclave, floating on the
precursor solution. By keeping the reaction at a temperature
of 100 °C for several hours, the Zn precursor nucleated on
the CNT-ZnO film to grow ZnO nanowires perpendicular Figure 14. Example of the hydrothermal synthesis: growth of ZnO
to the CNTs, with a thickness of 30-70 nm and lengths of nanowires perpendicular to MWCNTs. Reprinted with permission
up to 0.5 µm (Figure 14). from ref 227. Copyright 2006 Institute Of Physics.
The crystallinity of the inorganic compound typically
depends on the reaction temperature and time. For instance, CNTs and acted as a glue to form heterojunctions between
Du et al. showed that autoclaving a mixture of MWCNTs the CNTs (Figure 15).
with ZnCl2, ethanol, and thiourea228 at 180 °C produced well- 3.2.3.2. Supercritical Solvent. A very important hydro-
defined ZnS particles (diameter ≈ 40 nm) on the surface of thermal method involves the use of supercritical CO2 as an
the CNTs, while a reaction temperature of 80 °C only antisolvent that reduces the solvent strength of ethanol,
resulted in an amorphous coating. resulting in the precipitation of the oxide due to high
In general, the use of specific molecules, which hinder saturation. Using metal nitrates or halides, this method has
the crystal growth by steric configuration (capping agents), been applied to deposit Eu2O3,232 CeO2,233 La2O3,233 Al2O3,233
provides efficient control of the crystal size of the nanopar- SnO2,234,235 and Fe2O3236 onto pristine CNTs. Sun et al. used
ticles. For example, Yu et al. dissolved copper acetate in supercritical ethylenediamine as a solvent to produce thin
water and diethylene glycol and added acid-treated CNTs.229,230 coatings of RuO2.237 They also observed various morphol-
Upon heating in an autoclave at 180 °C, the diethylene ogies and structures of cerium oxide by simply changing the
glycol-capped copper species condensed to CuOx and nucle- reaction temperature.233 For instance, the authors could alter
ated to form small Cu2O crystals. After a reaction time of the composition of cerium oxide from preferentially Ce2O3
2 h, the crystals were 5-10 nm in diameter and covered by at 120 °C to CeO2 at 150 °C. In contrast, a reaction
an amorphous layer. Similarly, small crystals of Fe3O4 were temperature of 120 °C was needed to obtain a coating of
produced by Jia et al., who used polyethylene glycol (PEG) SnO2, while at 35 °C the oxide was only encapsulated.235
and FeCl3 and a reaction temperature of 200 °C.231 However, 3.2.3.3. Liquid-Source Misted Chemical Deposition
the 5 nm crystals agglomerated to about 180 nm aggregates (LSMCD). In the LSMCD process, the liquid precursor is
(nanobeads), which attached preferably to the carboxyl converted into submicrometer droplets with a monodisperse
groups on the surface of acid-treated CNTs. Consequently, aerosol generator (atomizer). Electrostatic and fluid forces
the magnetite nanobeads were concentrated at the tips of the transport the droplets to the surface of CNTs, while preserv-
1364 Chemical Reviews, 2010, Vol. 110, No. 3 Eder

Figure 15. Example of a polyol-assisted hydrothermal deposition of Fe3O4 on acid-treated MWCNTs. Polyethylene glycol (PEG) was
used to reduce FeCl2 and to control the size of the magnetite nanoparticles, which formed large aggregates (nanobeads) near the carboxyl
groups on the CNT surface. Reprinted with permission from ref 231. Copyright 2007 Elsevier Pubishing.

ing the precursor stoichiometry. After curing, conformal thin
films with high uniformity can be achieved, as demonstrated
by Kawasaki et al. for lead zirconium titanate (PZT) on
vertically aligned individual CNTs.238 A chemical deposition
solution containing a Pb-Zr-Ti mixture of 1.1:0.4:0.6 was
dissolved in methylethylketone, converted into mist with an
atomizer, and sprayed at 140 °C vertically onto the CNTs,
kept at 120 °C. The precursor reacted preferably with Stone-
Walls defects on the surface of the CNTs and formed a
uniform, still wet layer, which was then pyrolyzed at 300 °C
to give an amorphous coating and subsequently heated to
650 °C to crystallize into PZT. The advantages of this
technique include the use of commercially available chemical
deposition solutions and the low reaction temperatures.
3.2.4. Gas-Phase Deposition
Chemical and physical vapor deposition techniques are
among the most common methods to produce inorganic
nanomaterials, as they provide excellent control over the size,
shape, and uniformity of the inorganic material. Furthermore,
it is possible to deposit thin, continuous films on carbon
substrates, without altering the 3D integrity of vertically
aligned CNTs. This section provides examples of the
synthesis of various CNT-inorganic hybrids using physical
techniques, such as evaporation, sputtering, and pulsed laser
deposition (PLD), and chemical methods, including chemical
vapor deposition (CVD) and atomic layer deposition (ALD).
3.2.4.1. Evaporation and Sputtering. Physical vapor
deposition involves the evaporation of material in a crucible
under high vacuum, using either resistive heating (thermal
evaporation) or electron bombardment (electron beam depo-
sition), which is generated from a hot filament and focused
with a magnetic field. In contrast, sputtering (magnetron and
radio frequency, RF) relies on plasma (typically argon) to
bombard the target material, which can be kept at a relatively
low temperature. Reactive sputtering involves a small amount
of oxygen or nitrogen, which reacts with the sputtered
material to deposit oxides or nitrides.
The deposition of metal oxides via thermal evaporation
has been demonstrated by Kim et al., who mixed annealed
CNTs with Zn powder in a ratio of 1:12.239 Depending on
the reaction temperature, the Zn particles reacted with oxygen
impurities in argon to form a coating on the CNTs consisting
either of spherical particles (450 °C), nanowires (800 °C),
or short nanorods (900 °C). Similar results were obtained
by Yu et al., who tested the hybrid materials in field emission
devices and observed a significant improvement in emission
spot density due to the one-dimensional shape of ZnO.240
Zhang et al. used an electron beam to deposit various
metals on SWCNTs and observed that Ti, Ni, and Pd attached
Carbon Nanotube-Inorganic Hybrids Chemical Reviews, 2010, Vol. 110, No. 3 1365

Figure 16. Examples of electron beam deposition showing the different morphologies of electron beam-deposited metals on CNTs. Reprinted
with permission from ref 241. Copyright 2000 Elsevier Publishing.

strongly and formed metal nanowires, while Au, Al, and Fe

interacted weakly with SWCNTs via van der Waals forces
and formed isolated particles instead (Figure 16).241 This
technique can also be used to deposit oxides and nitrides,
when using O2 and NH3 atmospheres, respectively, as
demonstrated for MgO242 and TiN.211
RF or magnetron sputtering has been used to deposit
RuOx,158 SiO2,243 and ZnO.227,244 Furthermore, Jin et al. have
cosputtered Ba and Sr in an oxygen atmosphere to obtain a
BaO/SrO coating.245 In most of these works, the coating
around the CNTs was generally conformal, although in the
case of vertically aligned CNTs (“carpet”), the inorganic
material was deposited predominantly along the top of the
carpet. This is in contrast to other techniques such as electron
beam deposition, where the thickness of the coating was
found to increase from top to bottom.242 The size of the
particles can be controlled by adjusting the sputtering time,
as demonstrated in Figure 17 by Zhu et al. for ZnO.244
Interestingly, Fang et al. observed that the distribution of
RuO2 particles on vertically aligned CNTs was significantly
better on nitrogen-doped CNTs compared with pure CNTs.246
It is well-known that the incorporation of nitrogen into the
sp2-type walls of CNTs causes many structural defects
(pyridine-like bonding) due to the different valence of
nitrogen ions. It seems that these defects are more reactive
toward the adsorption of RuO2 than the graphitic walls of
pure CNTs and allow a uniform dispersion along the Figure 17. Example of the size control of ZnO nanoparticles
sidewalls of nitrogen-doped CNTs (Figure 18). deposited via magnetron sputtering: the average particle size
increased from roughly 20 to 37 nm upon increasing the sputtering
In summary, sputtered films typically have a better time from 1 to 3 min. Reprinted with permission from ref 244.
adhesion to the substrate than evaporated films and a Copyright 2006 Wiley-VCH.
composition closer to that of the source material. In contrast
to evaporation techniques, sputtering also enables the deposi- beam. The directed laser pulse is absorbed by the target,
tion of materials with very high melting points and can be vaporizes the material, and creates a plasma plume containing
performed “top-down”, while evaporation must be operated various energetic species, such as atoms, molecules, elec-
“bottom-up”. On the other hand, evaporation techniques trons, ions, clusters, particulates, and molten globules, which
typically offer better structural and morphological control then expand into the vacuum and deposit on a typically hot
and more flexible deposition rates. substrate to nucleate and grow as a thin film. This process
3.2.4.2. Pulsed Laser Deposition. Pulsed laser deposition can occur in ultrahigh vacuum or in the presence of a
(PLD) is related to the evaporation techniques described in background gas, such as oxygen, which is commonly used
the previous section but utilizes a high-power pulsed laser when depositing oxides.
1366 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 18. Scheme of the deposition of RuO2 nanoparticles on
aligned CNTs (top) and nitrogen-doped CNTs (N-CNT, bottom)
via magnetron sputtering. The pyridine-like defects in the N-CNTs
provide a stronger interaction with RuO2 and a more uniform
dispersion along the sidewalls. Redrawn from ref 246.
The work of Ikuno et al. provides an excellent example
of the potential of PLD in coating CNTs.247,248 The authors
used individual MWCNTs (grown by arc discharge), which
were attached to a molybdenum plate via electrophoresis. A
pulsed Nd:YAG (yttrium aluminum garnet) laser with a
wavelength of 355 nm (laser energy of 140 mJ and a pulse
duration of 5 ns) was focused onto a target at a repetition
rate of 10 Hz. A single-crystalline Si wafer and pellets of
Zr, Hf, Al, ZnO, Au, and Ti were used as targets for the
coating of SiOx, ZrOx, HfOx, AlOx, and ZnOx. The experiment
was carried out at room temperature andsfor the oxide
materialsswith an oxygen pressure of 1.3 × 10-2 Pa.
The authors observed that all oxides covered the CNTs in
a very uniform, dense, and continuous layer with a thickness
between 6 and 11 nm, while Au adsorbed on the CNTs as
nanoparticles (diameter ≈ 7 nm) (Figure 19). With the
exception of the polycrystalline ZnO coating, the oxides of
Zr, Hf, and Al were all amorphous. The differences in
morphology were attributed to the different surface adhesions
and diffusivities of the studied materials. The low mobility
of the oxides on the CNT surface enabled a uniform wetting
layer and prevented crystallization, while Au could diffuse
easily and crystallized into nanoparticles. The authors further
demonstrated the deposition of complex multilayers with
nanometer accuracy, such as the SiOx-Ti-SiOx-CNT
hybrid shown in Figure 19.
The advantages of this technique over other thin-film
deposition methods include the relatively simple basic setup
and the operation at room temperature. On the other hand,
PLD has a lower average deposition rate than other deposi-
tion techniques, such as CVD or evaporation/sputtering
techniques, but is faster than ALD. However, this enables a
stable nucleation of smaller nucei (critical nuclei radius in
the range of 1-2 atoms) and consequently higher nucleation
rates and, hence, smoother films. Further information on PLD
as a coating technique for CNTs can be found in a review
by the same authors,249 in which they also describe the
Eder
possibility of etching the tips to leave CNTs free for
nanoprobe and nanotip applications.
3.2.4.3. Chemical Vapor Deposition. Chemical vapor
deposition (CVD) is a versatile technique often used in the
semiconductor industry that involves the growth of a solid
material from the gas phase via chemical reaction at the
surface of a substrate. Volatile reaction byproducts are simply
exhausted from the coating zone and sent to waste process-
ing. In contrast to high-pressure/high-temperature synthesis,
the CVD technique typically operates at medium tempera-
tures (600-800 °C) and at slightly reduced atmospheric
pressures. Because CVD utilizes reagents of very high purity,
the technique is capable of synthesizing crystals with
controlled purity and composition. Other advantages of CVD
include the high deposition rate, good degree of control
(purity and composition), and easy scalability. Tuning the
process is simply a matter of manipulating the vapor flows
in the coating zone. However, due to the fast deposition, it
is difficult to achieve uniform and defect-free coatings when
scaling down to a few nanometers.
Several groups have used CVD to synthesize CNT-inorganic
hybrids with SnO2,250 RuO2,251 Si3N4,252 and TiC.213 These
attempts can be separated into two categories: (a) coating
of prepared CNTs and (b) in situ growth of CNTs onto or
around inorganic nanoparticles, which act as the required
catalyst.
(a) Kuang et al. deposited acid-treated MWCNTs on a Si
waver and heated them to 550 °C in a SnH4-containing N2
atmosphere.250 At this temperature, the precursor decom-
posed, attached to the functional groups of the CNTs, and
reacted with the oxygen impurities to produce SnO2. By
altering the flow rate and the reaction time, the thickness of
the layer could be increased from 3 nm (for 0.5 min) to 80
nm (for 3 min). In a similar way, CNTs have been coated
with Si3N4 (T ) 720 °C, precursor ) H2SiCl2 and NH3),252
RuO2 (T ) 350 °C, precursor ) Ru(acac)3),183 NiO, and
Co3O4 (using metal nitrates).253 In the case of RuO2, the
authors could also produce various morphologies, ranging
from 1D single crystal nanorods with [001] growth direction
to spherical nanoparticles of uniform size (10 nm), depending
on the postannealing conditions.183
Chrissanthopoulos et al. used the carbothermal reduction
of ZnO to coat MWCNTs.254 The CNTs were placed a few
centimeters downstream from the reaction zone (1000 °C)
and close to the end of tube (850-950 °C) in order to achieve
the desired temperature gradient. Following the reduction
of ZnO, Zn vapor was transferred through the gaseous phase,
attached to the defects on the CNT surface via Zn-C bonds,
and reoxidized to ZnO. Depending on the temperature
gradient, the authors observed either thick rods of ZnO
(850 °C) or polypods of thin wires (950 °C).
A fascinating approach has been reported by Lazareck et
al.255,256 Aligned CNTs were synthesized within an anodized
aluminum oxide (AAO) template, which was removed
afterward with concentrated HNO3. This acid treatment also
produced carboxyl groups, predominantly on the tips of the
CNTs, which were subsequently linked to a single-strand
amine-terminated DNA. A complementary DNA strand,
connected with a 20 nm Au nanoparticle, was then hybridized
with the first strand to link Au nanoparticles to the tip of
CNTs. Using CVD at 600 °C, these Au nanoparticles then
acted as catalysts for the growth of ZnO via a vapor-liquid-
solid (VLS) mechanism.
Carbon Nanotube-Inorganic Hybrids Chemical Reviews, 2010, Vol. 110, No. 3 1367

Figure 19. Examples of the deposition of various metals and metal oxides with different morphologies on free-standing CNTs via PVD.
The far-right image shows a complex CNT-inorganic hybrid with a layer of Ti between two layers of amorphous SiO2. Reprinted with
permission from ref 247. Copyright 2007 Elsevier Publishing.

(b) An example of the in situ growth of CNTs is the work

of Liu et al., who used a water-assisted CVD process to
produce CNTs with a ZnO coating.257 In detail, MWCNTs
were synthesized on Zn foil via pyrolysis of C2H2 at 850 °C
in a gas mixture of argon, hydrogen, and water. As the CNTs
grew via a tip-growth mechanism, they carried up part of
the Zn, which at the same time was oxidized by water vapor.
Most of the ZnO coating was observed at the tip of the CNTs,
which would suggest the use of this CNT-hybrid as an
electron emitter (see section 4.4).
An interesting alternative to ZnO was demonstrated by
Ho et al.258 The authors used Ni catalyst particles, supported
on MgO, which were deposited on a Si wafer and placed
inside a tube furnace. ZnO powder was put inside a thin
ceramic tube, whose ends were covered with Al foil, and
placed inside the reactor tube, close to the catalyst powder.
Upon reaching a certain temperature (300-350 °C), assisted
via a plasma-enhanced CVD technique, CNTs began to grow
on the Ni particles and kept growing as the temperature was
steadily increased to 700 °C over a period of about 6-7
min. At this temperature, the Al foil melted and exposed
the ZnO powder, which was instantaneously transformed to
Zn vapor via carbothermal reduction and deposited on the Figure 20. (a) Scheme of the deposition of amorphous layers of
still-growing CNTs as a thin coating of ZnO. TiO2 and SiO2 on acid-treated CNTs via ALD and (b) corresponding
3.2.4.4. Atomic Layer Deposition. In contrast to CVD, SEM image for the case of SiO2. (c) TEM image of vertically grown
the precursors for atomic layer deposition (ALD) are kept CNT coated with RuO2 both outside and inside. Reprinted with
permission from refs 259 and 261. Copyright 2003 and 2005 Wiley-
separate and exposed sequentially. Ideally, each precursor VCH.
forms a monolayer on the substrate, and the excess vapor is
removed before the next precursor is introduced. This process bonds. This two-step process (shown in Figure 20) was
is then repeated until the deposited film reaches the desired repeated until the thickness of the coating reached 1, 3, or
thickness. Hence, ALD film growth is self-limiting and based 10 nm. As the coating at this stage consisted of amorphous,
on surface reactions, which enables deposition control on predominantly hydroxide species, the samples needed sub-
the atomic scale. sequent calcination at 350 °C to form the oxide 350 °C.
As an example, Gomathi et al. used metal chloride Javey et al. demonstrated coatings of ZrO2 as thin as 8
precursors to coat acid-treated MWCNTs with SiO2, TiO2, nm covering the top of horizontally attached SWCNTs for
and Al2O3.259 Their experimental setup consisted of three device applications.260 In contrast, Min et al. used vertically
glass chambers, for the metal halide, reaction sample, and aligned arrays of MWCNTs, grown inside anodized alumi-
water, interconnected using high-vacuum stopcocks and num oxide (AAO) via CVD using C2H2.261 These CNTs were
maintained at 80 or 150 °C. The sample chamber was first then treated in phosphoric acid, which not only etched away
evacuated with a diaphragm pump, then filled with metal the AAO template but also opened the CNTs tips and
halide for 5 min and evacuated again to remove the unreacted introduced carboxyl groups that enabled attractive interaction
metal halide. The surface hydroxyl groups on the CNTs with the ruthenium precursor. As shown in Figure 20c, the
reacted with the metal precursor and formed metal-oxygen CNTs were coated both on the inside and on the outside.
bonds. In the second step, water vapor was passed through The authors also showed that the structure of the coating
the chamber, which hydrolyzed the residual metal-chlorine depends on the type of precursor. Although the reaction was
1368 Chemical Reviews, 2010, Vol. 110, No. 3
Table 6. List of Advantages and Disadvantages of the Various Synthesis Approaches for CNT-Inorganic Hybrids
building blocks
morphology spherical particles
structure, crystallinity good control of structure and crystallinity
size control capping agents, narrow size distribution
coating monolayer
confocal
interface needs anchor molecules
carried out in an oxygen-containing atmosphere, the use of
Ru(od)3 (od ) octane-2,4-dionate) resulted in a metallic Ru
film, while Ru-β-diketonate produced RuOx coatings.261
Furthermore, upon oxidation at 500 °C, the CNTs were
removed and the two coatings combined to form RuO2 NTs.
ALD has unique advantages over other thin-film deposition
techniques, as it can be operated at low temperatures (e.g.,
80 °C) and allows exact control over the thickness of the
deposited coating. Furthermore, ALD-grown films are very
uniform, pinhole-free, and chemically bonded to the sub-
strate. However, because of the sequential exposure of the
precursors, the technique has the lowest deposition rate
compared with CVD and PLD. As demonstrated, it is also
possible to deposit various ceramics, from insulators to
conductors, deep inside porous materials as well as around
spherical particles and 3D architectures, such as aligned
CNTs.
3.3. Comparison of Synthesis Techniques
Table 6 summarizes the advantages and disadvantages of
the various synthesis techniques regarding the morphology,
structure, and interface of CNT–inorganic hybrid materials.
The building block approach synthesizes inorganic building
blocks of defined size and shape and attaches them to the
CNTs via linking agents that utilize covalent, noncovalent,
or electrostatic interactions. Consequently, either the inor-
ganic nanoparticles or the CNTs (or both) have to be
modified with functional groups. The type of functionaliza-
tion and, thus, the strength of interaction determine the
distribution of the inorganic nanoparticles on the CNT
surface. This wet-chemical technique is typically limited to
the formation of monolayers of nanoparticles. Excess nano-
particles, not anchored to the CNTs, can easily be removed
by filtration or centrifugation. Another advantage of this
method is the control of particle size and distribution, which
can be achieved by capping agents or by the addition of salts,
which support heterogeneous nucleation and thus depress the
growth of larger aggregates caused by homogeneous nucle-
ation. The control of particle size and shape enables a better
structure-property prediction.
In the in situ approach, the inorganic compound is directly
formed on the surface of pristine or modified CNTs. The
main advantage of this route is that the inorganic compound
can be deposited as continuous amorphous or single-
crystalline films with controlled thickness or as discrete units
in the form of nanoparticles, nanorods, or nanobeads. In most
cases, the inorganic particles are remarkably smaller when
deposited on CNTs compared with amorphous carbon or
graphite. For instance, CNTs can support TiO2 nanoparticles
during the reconstructive stresses of phase transformation
from anatase to rutile at high temperatures, keeping the size
of rutile particles small.195,262 Hence, the CNTs can prevent
crystal growth during crystallization and phase-transforma-
Eder
in situ wet chemical in situ gas phase
spheres, wires, films spheres, wires, films
often amorphous, requires heat treatment nonstoichiometry possible
concentration and reaction time reaction time
good control in film thickness
range in crystal sizes
multilayer multilayer
confocal confocal (chemical)
directed (electrodeposition) directed (physical)
pristine or functionalized CNTs possible good coverage in pristine CNTs
tion processes and may even stabilize uncommon or even
novel crystal phases. Therefore, it is very possible with this
synthesis route to develop novel materials with new properties.
Another advantage of the in situ approach is that a variety
of chemical and physical synthesis techniques can be applied.
The deposition can be carried out either in solution, via
electrochemical reduction of metal salts, electro- or electro-
less deposition, sol-gel processing and hydrothermal treat-
ment with supercritical solvents, or from the gas or vapor
phase using chemical deposition (CVD, ALD) or physical
deposition (laser ablation, electron beam deposition, thermal
evaporation, sputtering).
In general, the wet chemical techniques are simple and
cheap and can be performed at low temperatures but may
need post-treatments to remove residues and to transform
the amorphous product into a crystalline phase. Again, the
size and shape of the particles can be controlled using
capping agents. In contrast, the physical techniques enable
excellent control of composition and size distribution but
are limited by the availability of suitable precursors, the
required high temperatures, the slow deposition rates, and
the complex instrumentation.
4. Properties and Potential Applications of
CNT-Inorganic Hybrids
This section presents several exciting examples of the
improved performance of CNT-inorganic hybrids in ap-
plications, such as photocatalysis, electrocatalysis, and
environmental catalysis, gas sensors, supercapacitors, and
field emission devices. Furthermore, it describes how inor-
ganic compounds can significantly increase or decrease the
oxidation stability of CNTs, which is useful for the applica-
tion of CNTs as a template for inorganic nanotubes.
4.1. Example 1: Photochemical and
Photoelectrochemical Applications
Beginning in Japan in the early 1970s263 with photoelec-
trochemical electrolysis, or “splitting”, of water to produce
hydrogen (a source of abundant clean energy), a tremendous
amount of research has been carried out in the two closely
related fields of semiconductor photoelectrochemistry and
photocatalysis, both considered among the most important
research areas.264 Current research aims primarily at direct
solar energy conversion as an alternative approach to solid-
state junction photovoltaic cell (e.g., Graetzel dye-sensitized
solar cell),265-267 as well as the use of heterogeneous
photocatalytic oxidation for water and air purification,
decomposition of organic compounds, and self-cleaning
surfaces.264 The past few years saw a renewed interest in
photocatalytic “water splitting”, which was recently identified
by the European Science Foundation as one of the world’s
emerging key research fields. The aim is to find new material
Carbon Nanotube-Inorganic Hybrids
systems that enable conversion efficiencies beyond 10%, the
U.S. Department of Energy’s target for a commercially viable
catalyst.
The majority of research currently focuses on the design
of suitable materials (typically a semiconductor) that are able
to efficiently harvest sunlight for the generation of electrons.
In general, when the semiconductor is illuminated with
photons having energies greater than that of the material’s
band gap, EG, electron-hole pairs are generated and sepa-
rated in a space-charge layer. In the case of water splitting,
the electrons and holes are separated and the photogenerated
electrons flow to the counterelectrode (usually platinum) and
reduce protons, resulting in hydrogen gas evolution without
an applied potential. This requires a semiconductor with a
conduction band energy (ECB) of at least -0.4 V vs SHE
in acid solution or 1.2 V in alkali solution.264 In contrast,
for the purpose of photocatalytic oxidation of harmful
compounds, the reduction reaction is not necessarily hydro-
gen production and no counterelectrode is needed. In these
cases, both the reduction and oxidation sites are located on
the semiconductor surface. The holes (h+) generated in TiO2
are highly oxidizing, and most compounds are typically
oxidized completely (to their final oxidation state). In
addition, oxygen can act as an electron acceptor to form
superoxide radical ions (O2-), while OH- and H2O are
available as electron donors to yield the hydroxyl radical,
which is very reactive, strongly oxidizing, and therefore
capable of totally mineralizing organic pollutants.
TiO2 is the most suitable material for industrial use in
photoelectrochemical and photocatalytic applications due to
its efficient photoactivity, chemical and biological inertness,
nontoxicity, high photostability, cost effectiveness, and easy
production.268,269 The utilization of the strong photocatalytic
oxidation potential of TiO2 for the destruction of pollutants
was shown in 1977 by Frank and Bard, who described the
decomposition of cyanide in the presence of aqueous TiO2
suspensions.270 Using supported TiO2 powders, detoxification
of various harmful compounds in both water and air was
demonstrated actively as a potential purification method of
wastewater and polluted air.271
However, with band gaps of 3.25 and 3.0 eV for anatase
and rutile, respectively,268 TiO2 needs photons with wave-
lengths smaller than 385 and 410 nm for efficient excitation.
Consequently, TiO2 only absorbs about 3% of the solar light
it is exposed to and currently requires near-UV light to
operate as an efficient photocatalyst.272 Ideally, the band gap
of the semiconductor is close to about 1.35 eV, which is
required for optimum utilization of solar energy. However,
semiconductors with such small band gaps, including CdS
and CdSe, have shown considerably lower efficiencies and
stabilities than TiO2.273
Therefore, it is desirable to extend the absorption of TiO2
into the visible region and thus improve its photocatalytic
efficiency. So far, strategies have concentrated on the
generation of defect structures, by doping with light ions
(N, S, or halides)274 or transition metal ions,275-278 to induce
space-charge separation and the modification of TiO2 with
noble metals or other semiconductors.279 The use of organic
dyes as photosensitizers is another approach, which is at
present limited by the very low stabilities of the dye
molecules.264 Because of their electronic properties, chemical
inertness, and stability, CNTs are considered promising
candidates to improve the photoefficiency of semiconductors,
and, as shown in the next section, some work has recently
Chemical Reviews, 2010, Vol. 110, No. 3 1369
Figure 21. Kinetics of phenol removal under UV illumination in
the presence of various photocatalysts. Reprinted with permission
from ref 206. Copyright 2005 Elsevier Publishing.
been directed toward investigating CNT-inorganic hybrids
for photocatalytic applications.
4.1.1. Photocatalytic Production of Hydrogen
Ou et al. impregnated anatase-TiO2 particles with small
Ni clusters and used them as a catalyst to grow MWCNTs
via CVD at 550 °C.280 The hybrid materials were tested for
water splitting under visible light and a methanol/water
solution. The addition of organic alcohols as reducing agents
considerably enhanced the reaction efficiency by preventing
the H2 and O2 gases from recombining on the surface of
TiO2.
In contrast to the completely inactive TiO2-Ni catalyst,
the addition of 4.4 wt % CNTs produced significant amounts
of hydrogen, with a reaction rate of 38 µmol g-1 h-1 (5 µmol
g-1 h-1 in pure water). Increasing the amount of CNTs above
4.4 wt %, however, led to a decrease in the absorption of
light and, thus, in the activity of water splitting. These initial
results are especially promising, if one considers that the
tested hybrids merely consisted of a network of CNTs, grown
on and connected by TiO2 particles, resulting in only a
relatively small interfacial area. In this respect, this author
believes that a uniform continuous coating of TiO2 on CNTs,
and hence an optimized interface, will lead to considerably
higher hydrogen evolution rates.
4.1.2. Photocatalytic Decomposition of Organic
Compounds
There have been quite a few studies concerning the
photocatalytic activity of CNT-TiO2 hybrids for the oxida-
tion degradation of organic compounds, such as acetone,196
propene,281 and phenol205,206,282 as well as the inactivation
of bacterium anthraxium.283 CNT-ZnO has been investigated
for the decomposition of methylene blue,221 while the
decomposition of indigo carmine with both CNT-ZnO and
CNT-TiO2 has also been tested.224 All of these studies
observed superior photocatalytic performance by the hybrid
over the individual component.
For instance, Wang et al. coated acid-treated MWCNTs
with TiO2 via sol-gel and investigated the degradation of
phenol under UV light206 and visible light.205 In both cases,
they observed a considerable acceleration of the oxidation
reaction, with most of the phenol being transformed after
4 h (6 h in visible), while pure TiO2 needed 6 h (9 h in
visible) and pure MWCNTs only converted up to 5% (Figure
21). The maximum activity was observed with a CNT
concentration of 20 wt % and was considerably higher than
1370 Chemical Reviews, 2010, Vol. 110, No. 3 Eder

the simple mechanical mixture of the same composition and

surface area. A similar behavior was shown by Yu et al. for
the oxidation of acetone (maximum activity at 5 wt %
CNTs),196 as well as by Lee et al. for the decomposition of
methylene blue,221 who observed a superior performance of
CNT-ZnO hybrids over the mechanically mixed composite.
The importance of the interface was further supported by
the work of Yao et al., who compared MWCNT-TiO2
hybrids with SWCNT-TiO2 hybrids.282 In contrast to
MWCNTs, the higher flexibility and resilience of SWCNTs
caused them to wrap around the TiO2 particles, thus
increasing the interfacial area. The authors believe that this
enhanced interface was the reason for the higher photocata-
lytic activity observed in SWCNT-TiO2 hybrids for the
decomposition of phenol. All these works clearly demon-
strated a synergistic effect of CNTs in enhancing the
performance of TiO2 or ZnO.

4.1.3. Photovoltaic Devices

The photocatalytic activity of semiconductors in photo-
voltaics has been successfully demonstrated in the so-called
dye-sensitized solar cell (DSSC),265,266 which utilizes nano-
structured TiO2 films modified with sensitizing dyes. The
reported solar conversion efficiencies of up to 10% are indeed
very encouraging; however, further improvement in the
performance of DSSC has yet to be achieved. Currently, the
photoconversion efficiency is mainly limited by the transport
of photogenerated electrons across grain boundaries in the
semiconductor particle network. Such a random transit path
for the electrons also increases the probability of their
recombination with oxidized sensitizer. To counter these
drawbacks, networks of TiO2 or ZnO nanotubes and nanow-
ires have been used to direct the flow of photogenerated
charge carriers, while the use of a redox couple such as I3-/I
has been shown to facilitate the electron transport by rapid
regeneration of the oxidized sensitizer.264,284,285
Another very promising approach has been demon-
strated by Kamat et al., which involves the use of CNTs Figure 22. (Top) SEM images of various electrodes tested in a
as support for light-harvesting semiconductor particles, photelectrochemical cell: (a) pure carbon fiber electrode (CFE), (b)
such as CdS, CdSe, or CdTe QDs.165,286 Upon irradiation TiO2 nanoparticles spray-coated on CFEs, (c) SWCNTs electro-
with visible light, the authors observed induced charge phoretically deposited on CFEs, and (d) the final hybrid. (Bottom)
Incident photoconversion efficiency (IPCE) for various electrodes.
transfer processes of the photogenerated electrons from Reprinted with permission from ref 112. Copyright 2007 American
excited semiconductors into semiconducting SWCNTs, Chemical Society.
whose electron-accepting ability further facilitated electron
transport to the collecting electrode and thus increased In view of these results, Kamat et al. also investigated
the photoconversion efficiency of the solar cell. Guldi et SWCNT-TiO2 and SWCNT-ZnO hybrids for use in
al. used pyrene derivatives to attach CdTe QDs and photovoltaic devices.112,220 The hybrid materials’ SWCNTs
investigated the electron-transfer chemistry in both the were first refluxed in 5 M HNO3 to remove any metal
ground and excited states.287 Using an excitation at 350 and organic impurities and to open the end of the tubes,
nm, the authors observed strong blue-shifts and luminescence then solubilized in tetrahydrofuran (THF) containing
quenching, which they attributed to strong electronic interac- tetraoctylammonium bromide (TOAB) with the aid of
tions between the two electroactive components. A charge ultrasonication, and assembled as a thin film on a carbon
transfer from the photoexcited electron donor (CdTe) to the fiber electrode (CFE) using an electrophoretic deposition
electron acceptor (SWCNTs and MWCNTs) was also technique. A suspension of commercial TiO2 powder
credited for the strong increase in charge lifetimes. Using (Degussa P25) in 1% acetic acid/methanol was slowly
femtosecond transient absorption and nanosecond laser dropped onto the CFE and CFE/SWCNT electrodes and
flash photolysis, the decay rates decreased considerably dried in air (Figure 22).
from 3.5 × 107 s-1 for CdTe to 4.7 × 104 s-1 for In the case of SWCNT-TiO2, the authors observed that
CNT-pyrene+-CdTe. The hybrid was then deposited on the photon conversion efficiency for the hybrid material was
indium-tin oxide (ITO) to make a photoelectrochemical almost doubled compared with that of pure TiO2. Further
device, and very promising internal photoconversion experiments revealed a positive shift in the flat-band potential
efficiencies (IPCEs) of up to 2.3% were observed, which of SWCNT-TiO2, which indicates electron transfer between
was about 6 times more efficient than a red CdTe cell. TiO2 and SWCNTs. Although not directly participating in
Carbon Nanotube-Inorganic Hybrids
Figure 23. Schemes of CNTs as photosensitizers: (a) electron injection into the conduction band of TiO2, (b) electron back-transfer to
CNTs with the formation of a hole in the valence band of TiO2 and reduction of the hole by oxidation of adsorbed OH- species.
the photocurrent generation due to their low photocurrent
efficiencies in visible light, the SWCNTs play a crucial role
in the charge collection and charge transport processes and,
thus, significantly enhance the photoconversion efficiency
of the photovoltaic cell.
However, it can be expected that an improvement in
morphology of the TiO2 particles and control of the interface
between CNT and TiO2 will further enhance the performance
of the hybrid. Additionally, acid treatment of the SWCNTs
should be avoided in order to preserve their superior
electronic properties, though separation of semiconducting
and metallic CNTs may prove advantageous.
4.1.4. Synergistic Effects of CNTs
There are several possible explanations for the beneficial
role of CNTs, depending on the inorganic material and choice
of application.
(1) The first explanation is purely physical and can be
described in terms of MWCNT acting as a dispersing agent
that prevents TiO2 from agglomerating, thus providing a
higher active surface area for the resulting catalyst compared
with the single-phase TiO2. It is well-known that CNTs can
affect the morphology of the deposited oxide, reducing its
particle size and consequently increasing the specific surface
area, even affecting the crystal size of several layers of TiO2
nanoparticles.195 Wang showed that, in first approximation,
the decrease in TiO2 particle size with increasing
MWCNT-TiO2 ratio correlated to the observed increase in
photocatalytic activity, up to 20 wt % CNTs.206 At higher
CNT concentrations, however, the activities dropped, al-
though the particle sizes decreased further. This suggests that
the contribution of the particle size was not the only
important factor. This was supported by Lee et al., who
compared the hybrid and the pure TiO2 with equal effective
surface areas, using a comparably lower amount of the high
specific surface area hybrid, and observed highly improved
efficiencies in destroying bacteria anthraxium for the hybrid
material compared with pure TiO2.283
(2) CNTs are well-known for their enhanced adsorption
properties for some gases32,288 and thus may act as
additional adsorbents for the organic compounds, which
then diffuse to the TiO2-CNT phase boundary to undergo
degradation. This possibility was disproved by Wang et
al.,206 who observed similar adsorption capacities of phenol
for the hybrid, the single compounds, and the mechanical
mixture.
(3) The CNTs may act as photosensitizers for n-type
semiconductors like TiO2 (Figure 23). In this model, pho-
Chemical Reviews, 2010, Vol. 110, No. 3 1371
toinduced electrons are easily transferred to the CNT-TiO2
interface and injected into the TiO2 conduction band. A
similar electron transfer was observed between various other
carbon materials and the TiO2 semiconductor.289 Simulta-
neously, a positively charged hole (h+) can be formed by an
electron migrating from the TiO2 valence band to a MWCNT.
With this understanding, the role played by CNTs can be
illustrated by injecting electrons into the TiO2 conduction
band under visible light irradiation, triggering the formation
of very reactive radicals such as superoxide radical ions
(O2•-) and hydroxyl radicals (HO•), which are then respon-
sible for the degradation of the organic compound.
(4) Finally, CNTs could suppress the recombination of
electrons and holes (Figure 23). Upon light irradiation,
valence band electrons of the semiconductor are excited to
its conduction band, leading to the formation of electron-hole
pairs. CNTs then act as electron acceptors, promoting
interfacial electron-transfer processes from the attached oxide
to the nanotube. This separation of charges retards the
recombination of photoinduced electrons and holes and hence
improves the photocatalytic activity of the semiconductor.
Wang et al. investigated this possibility by photolumines-
cence (PL) emission spectroscopy, which is generally used
to measure the efficiency of charge carrier trapping, im-
migration, and transfer, and to understand the recombination
mechanisms of e-/h+ pairs in semiconductor particles.206 In
the case of CNT-TiO2, the intensity of the emission band
at 525 nm decreased considerably with increasing CNT
concentration. Since the emission in TiO2 can be attributed
to the radiative recombination process of self-trapped excita-
tions,264 the reduction of the PL intensity suggests a strong
inhibition of e-/h+ recombination due to electron transfer.
In summary, the role of CNTs is presumably a combina-
tion of these effects. CNTs do not only photoinduce electrons
and prevent e-/h+ pairs from recombining but also provide
a large surface area for smaller nanoparticles.
4.2. Example 2: Heterogeneous Catalysis and
Electrocatalysis
The major obstacle for the successful commercialization
of direct alcohol fuel cells (DAFCs) is the slow kinetics of
the electro-oxidation of alcohols (e.g., methanol or ethanol),
which is agreed to be caused by poisoning of the surface of
the platinum electrode with CO.290 The most promising
solution seems to be the addition of a second component,
such as Ru or a transition metal oxide (RuO2, TiO2, SnO2),
which assists in the oxidation of CO to CO2 by the
1372 Chemical Reviews, 2010, Vol. 110, No. 3 Eder

Figure 24. (a) Cyclic voltammograms for (a) Pt/RuO2 · 0.56H2O-CNT, (b) PtRu-C, (c) Pt-CNT, and (d) Pt-C (d), taken at 25 °C in 1
M CH3OH + 1 M HClO4 with a scan rate of 50 mVs-1. The CNT-RuO2 hybrid showed the lowest onset and peak potentials of all tested
catalysts, indicating the highest activity for the electrooxidation of methanol. Reprinted with permission from ref 216. Copyright 2006
Wiley-VCH. (b) Partial oxidation of n-butane over (a) VxOy catalyst without CNTs and (b) the CNT-VxOy hybrid. The online proton
transfer reaction mass spectrometry (PTR-MS) response for masses 71 and 99 amu, heated sequentially to 200, 300, and 400 °C, shows the
activity of the CNT-VxOy hybrid at lower temperatures. Reprinted with permission from ref 299. Copyright 2007 Materials Research
Society.

dissociation of water. Still, the electrocatalytic activities and
CO tolerances of the catalysts need significant improvement,
which may be achieved by the use of CNTs.
Indeed, noble metal particles show significantly higher
power densities at higher backpressures, lower onset
potentials, and higher anodic currents, when supported on
CNTs rather than on activated carbon (AC) or other
substrates. For instance, one of the current highest values
of anodic currents for Pt-C has been reported by Motorola
with 40 mA/cm2 for 2.5 mg Pt/cm2, while Pt-CNT
catalysts exhibited currents as high as 54 mA/cm2 for 0.43
mg Pt/cm2 and 25 mA/cm2 for 10 µg Pt/cm2.291 These high
values suggest that the use of CNTs can significantly reduce
the amount of expensive metal catalysts. Even higher currents
and thus activities have been reported for the systems Pt/
MoOx-CNT,292 Pt/TiO2-CNT,197,293 and Pt/RuO2 · H2O-
CNT.294
The role of CNTs is often attributed to their ability to
stabilize highly dispersed oxide nanoclusters, resulting in
higher specific surface areas. However, Cao et al. investigated
the role of CNTs in more detail and chose the bifunctional
system of Pt/RuO2 · H2O.216 The authors observed that, with
almost identical Pt loadings and surface areas, Pt/
RuO2 · 0.56H2O-CNT showed a 50% higher current density
and, thus, a higher activity for methanol electrooxidation,
compared with the PtRu-C catalyst (Figure 24a). Further
analysis revealed that the CNT-containing sample required
only eight scanning cycles for activation, compared with 32
cycles needed for the PtRu-C. Thus, CNTs can dramatically
reduce the induction period of the electrocatalyst. Further-
more, the onset potential for CO oxidation was considerably
lower for the CNT-hybrid (223 mV) than for the PtRu-C
catalyst (338 mV). This is attributed to the exceptional
electronic properties of CNTs causing electron transfer to
the RuO2 crystals.
The beneficial role of CNTs has also been demonstrated
for application in heterogeneous catalysis. For instance, Pd
showed significantly higher catalytic activities toward hy-
drogenation of nitrobenzene when supported on CNTs than
on activated carbon,295 with nearly complete conversion after
5 h, while at the same time Pd/C showed a conversion of
only 10%. While not all materials showed such enhanced
catalytic activities, the hybrids typically developed excellent
selectivities toward a certain product. For example, in the
case of the hydrogenation of cinnamaldehyde, Pd and Ru
showed high selectivities toward hydrocinnamaldehyde296 and
cinnamyl aldehyde,297 respectively, when supported on
CNTs, while metals supported on activated carbon produced
mixtures with phenyl propanol.
There have also been a few reports on CNT hybrids with
transition metal oxides for heterogeneous catalysis, such as
RuO2,298 VxO5,170,299 WOx,300 ZnO,301 and sulphated ZrO2,113
as well as for electrocatalysis, such as SnO2.302 For instance,
Fu et al. synthesized hydrous RuO2 nanoparticles supported
on CNTs via a homogeneous oxidation-precipitation (HOP)
method using H2O2 as both the oxidant and precipitant at
room temperature.298 The hybrid was very active in the
aerobic oxidation of various aromatic, saturated, and cyclic
alcohols and quite selective toward the corresponding alde-
hydes or ketones in liquid phase under mild conditions. The
Carbon Nanotube-Inorganic Hybrids
improvement was attributed to a better dispersion of RuO2
nanoparticles on the CNT surface compared with other
substrates, like γ-Al2O3 and activated carbon. Huang et al.
investigated the activity of V2O5 catalysts toward the selective
catalytic reduction (SCR) of NO with NH3 in the presence
of excess oxygen.170 In supporting V2O5 on CNTs, the
authors could decrease the required reaction temperature
significantly to temperatures well below 200 °C, which is
important for increasing the lifetime of the catalyst and
saving energy.
Other examples involve solid acid catalysts such as WO3
and ZrO2, which play an important role in the petrochemical
industry for the conversion of hydrocarbon, including crack-
ing, isomerization, and alkylation, which are important in
the chemical and petrol industries. Using WO3, Pietruszka
et al. demonstrated the use of CNT as an active support for
skeletal isomerization reactions.300 Furthermore, the
CNT-WO3 hybrids exhibited very high skeletal isomeriza-
tion selectivity exclusively toward olefin reactions. In this
example, the improved performance of the CNT-WO3
hybrid was not only attributed to a more uniform dispersion
of WO3; CNTs also prevented the complete reduction of WO3
to metallic tungsten, a process that typically causes complete
deactivation of the catalyst.
Another example of a successful application of CNT
hybrids has been demonstrated by Chen et al., who inves-
tigated the catalytic performance of CNT-VOx in the partial
oxidation of n-butane.303 Such dehydrogenation reactions are
often highly exothermic, and the heat is typically released
to the catalyst surface, causing sintering of the particles and
decreased catalytic activities. Because of their high thermal
conductivity, the CNTs can act as a heat sink, thus keeping
the crystal size small (see section 4.1.4). Indeed, the authors
observed higher activities at lower reaction temperatures in
the CNT-VOx hybrids compared with unsupported VOx as
well as a very high selectivity toward the formation of maleic
anhydride.
In summary, the role of CNTs in heterogeneous catalysis
is predominantly a support for small catalyst particles with
excellent dispersion and high surface areas, which can
dramatically increase the activity and affect the selectivity
of the catalysts. Recently, Schlögl, Su, and co-workers
demonstrated that, by modifying their surface chemistry, the
CNTs alone can be active for the oxidative dehydrogenations
(ODH) of 1-butene,304,305 ethylbenzene,306 and styrene.307 In
all cases, quinine groups were identified as the active
functional groups, which can be introduced by gradually
heating in oxygen. A contribution of this catalytic behavior
of functionalized CNTs should be considered in order to
understand the role of CNTs in heterogeneous catalysis.
4.3. Example 3: Gas Sensors, Chemical Sensors
Metal oxide semiconductors (MOS) are prominent ex-
amples of sensing materials in gas sensors, as their electrical
properties are highly affected by the surrounding gas
environment. For instance, tungsten trioxide (WO3) shows
sensitivity to pollutants such as SO2, H2S, NO, and NH,308,309
while SnO2 is sensitive to NOx, CO, ethanol, and C2H4.310
The surface reactivity of the metal oxides toward reducing
and oxidizing gases is associated with the formation of
oxygen vacancies and the transfer of d-electrons into
adsorbates.311 Hence, the mechanism for gas detection is
controlled by the change in surface conductivity caused by
the release/trapping of electrons during gas interaction with
Chemical Reviews, 2010, Vol. 110, No. 3 1373
the surface. In general, the specific surface area, morphology,
crystallinity, and chemical composition (e.g., dopants, im-
purities) have an impact on the sensing properties. Conse-
quently, gas sensors based on MOS nanoparticles and
structures would benefit from a small grain size, high surface-
to-volume ratio, and increased surface activities. When their
size becomes comparable to that of the space-charge layer,
the electron transport properties of nanoparticles can be
strongly modulated by adsorption and desorption processes,
resulting in high gas sensitivities to ambient gases. However,
their sensing properties often suffer degradation due to
growth and aggregation.
Improvements in metal oxide based sensors, such as
enhanced sensitivity and selectivity to target gases, reduced
response and recovery times, and lowered operating tem-
perature, have been achieved by adding small amounts of
catalytically active metals onto the oxide surface, by produc-
ing core-shell materials, by doping with n-type or p-type
ions, and by adding a sacrificial component (mostly liquid)
as an electron donor.308,312-314 However, it is desirable for
gas-detection purposes to operate the metal oxide-based
sensors at room temperature, in order to reduce the power
consumption of the device and to enable safer detection of
flammable gases.297
In contrast with metal oxide sensors, CNTs exhibit
excellent adsorption properties due to their high specific
surface area, which provides a large number of active surface
sites.288,315-317 As the CNTs’ electric properties are effectively
altered by very small amounts of adsorbed gas molecules,
the CNT gas sensor can be operated at temperatures close
to room temperature. However, as MWCNTs are not very
sensitive to ambient gas, the use of CNTs as gas sensors is
mainly restricted to SWCNTs. Furthermore, due to their long
recovery times, CNT-based sensors typically need reactiva-
tion, e.g., by UV light irradiation, which purges the surface
from adsorbed gas molecules.
Similar to core/shell heterostructures like MOS/CdS,
MOS–CNT hybrids have shown improved photoluminescent
quantum efficiencies and enhanced gas-sensing properties
including reduced response and recovery times.201,318 So far,
CNT-SnO2 has been tested for detection of CO,210 NO2,201,203
NH3,319 formaldehyde,320 and ethanol,321 while glucose was
detected with CNT hybrids containing SiO2,322 clay,323 and
Fe3O4.324
Wei et al. used a sol-gel process to coat pristine SWCNTs
with SnO2 and investigated the gas-sensing performance for
NO2 at room temperature.203 They observed considerably
enhanced sensitivities (∆R/∆C, Figure 25) compared with
the pure SnO2 sensor. Because the morphology and surface
area of the hybrid sensors were similar to those of the pure
SnO2, and the observed sensitivities increased with increasing
CNT loading, the authors concluded that the advanced
sensing behavior originated from a common interface with
CNTs. In contrast to conventional SnO2 sensors, which
typically operate at temperatures between 200 and 500 °C,
the SWCNT/SnO2 hybrid gas sensors could indeed be
operated at room temperature.
When the NO2 gas molecules adsorb on the surface of
pure SnO2, they extract electrons, leaving the oxide surface
positively charged. This leads to the formation of a depletion
zone and to an increase in the sensor resistance. In the
CNT-SnO2 hybrid sensor, the electric properties of the oxide
are strongly enhanced by the highly conducting CNTs.
Consequently, the sensor resistance is dominated by the
1374 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 25. Example of CNT-SnO2 hybrid in gas sensors. (a) Relative resistances versus NO2/air gas concentrations of SnO2 and hybrids
with low (A) and high (B) concentrations of CNTs at room temperature. The sensitivities (∆R/∆C) increase considerably with increasing
CNT concentration. (b) Scheme of the presence of depletion zones near the CNT-oxide interface. Reprinted with permission from ref 203.
Copyright 2004 Elsevier Publishing.
Schottky barrier at the interface between the n-type SnO2
grains and the p-type CNTs, causing the formation of
additional depletion layers, which then amplifies the increase
in resistance upon NO2 adsorption and enables the operation
of the gas sensor at room temperature.
The barrier height between the SnO2 grains and the
MWCNTs seems to vary for different gases. In contrast with
NO2,201 exposure to ethanol,321 NH3,319 or acetylene201 has
been shown to release electrons and consequently reduce the
sensor resistance. This effect was observed only for the
hybrid sensors; the pure SnO2 sensor did not show any
response to ethanol or acetylene at room temperature. This
underlines the importance of the CNT–SnO2 interface.
Furthermore, Chen et al. showed that the response and
recovery times of the CNT–SnO2 hybrid sensor were
considerably improved over pure SnO2 sensors.321
In summary, hybrid SWCNTs–SnO2 gas sensors can
successfully combine the advantages of both components to
produce significantly higher sensitivities in detecting both
oxidizing (e.g., NO2) and reducing (e.g., ethanol, C2H4) gases.
Furthermore, the sensors exhibit shorter response times,
enhanced recovery properties, and better stability and
reproducibility, and can indeed be operated at room tem-
perature. There is still plenty of room for improvement
toward successful application as gas sensors, including the
control of morphology and the maximization of the surface
area and interfacial area. It would also be useful to investigate
other MOS in CNT hybrid materials.
4.4. Example 4: Supercapacitors and Batteries
Reliable and affordable electricity storage is a prerequisite
for optimizing the integration of renewable energy systems.
Energy storage, therefore, has a pivotal role to play in the
effort to combine a future, sustainable energy supply with
the standards of technical services and products. For both
stationary and transport applications, energy storage is of
growing importance as it enables the smoothing of transient
and/or intermittent loads and the downsizing of base-load
capacity with substantial potential for energy and cost
savings. The extended lifetime of batteries in handheld
devices is credited not only to higher energy densities but
also to a simultaneous reduction of energy consumption of
the portable devices. In contrast, the electric vehicle has failed
Eder
to become the accepted mode of transportation mainly
because of the battery. Short distances between recharging
and a limited service life of the battery are to blame, but
also the incredible weight and volume of the batteries.
Electrochemical capacitors (ECs) are energy-storage de-
vices that possess higher energy and power density than
conventional dielectric capacitors and batteries and are used
in applications including electric vehicles, noninterruptible
power supplies, dc power systems, lightweight electronic
fuses, memory backups, and solar batteries.325 The challenges
for these applications concern limitations in volumetric/
gravimetric power densities and RC time, long life, safety,
simplicity of design, cost, and the possibility of recharging,
and can only be accomplished by specially designed materials.
According to the mechanism of energy storage, ECs can
be categorized into two classes:325 (a) electrochemical double
layer capacitors (EDLC), based on double-layer capacitance
due to charge separation at the electrode/electrolyte interface,
which thereby need materials with high specific surface area
(e.g., activated carbon, CNTs), and (b) pseudocapacitors or
supercapacitors, based on the pseudocapacitance of faradaic
processes in active electrode materials such as transition
metal oxides and conducting polymers.
Because of their exceptional electronic properties, which
allow ballistic transport of electrons over long nanotube
lengths, CNTs have been considered a most promising
candidate for electrochemical capacitors.326,327 However, pure
CNTs possess a rather low specific capacitance, typically
about 10-40 F/g, which depends on the microtexture, purity,
and electrolyte.327 A considerable enhancement can be
expected from the combination of CNTs with an electroactive
material, which provides the additional pseudocapacitance
while each tube acts as a minute electrode. For instance,
capacitance values up to 170 F/g have been obtained by
coating CNTs with conducting polymers, such as polypyr-
role.328,329 However, these materials do not withstand the high
cycle life (>100 000) due to the strong propensity of
conducting polymers toward degradation, which may be
further decreased by overcharge and overdischarge misuse.
In this context, the electrochemical stability of transition
metal oxides makes them a better choice, provided they are
highly conducting. A wide range of oxides has been
investigated for use in CNT hybrids, including NiO,209,330,331
Carbon Nanotube-Inorganic Hybrids
Figure 26. Example of the supercapacitive performance of
CNT-MnO2 hybrids. Nyquist (a) and Bode plots (b) of pure
MWCNT and the hybrid electrode in 1 M KOH aqueous solution
(inset is the amplification of the high-frequency region), recorded
between 105 and 0.01 Hz with a perturbation of 5 mV. Reprinted
with permission from ref 168. Copyright 2007 Elsevier Publishing.
MnO2,110,188,332-335 V2O5,336,337 and RuO2.207,237,251,338 In all
of these studies, the synergistic effects of CNT-oxide
hybrids have shown a considerable improvement of the
otherwise poor electric properties and deficient charge
transfer channels of the pure oxide.
Among these transition metal oxides, ruthenium oxides
are regarded as the most promising electrode material for
supercapacitor applications due to their high specific capaci-
tance, highly reversible redox reactions, wide potential
window, and very long cycle life.237 Park et al. reported a
specific capacitance for CNT-RuO2 (13 wt %) of 450 F/g
at a potential scan rate of 20 mV s-1, which could be further
enhanced, by using hydrous ruthenium oxide and function-
alized CNTs, to 800 F/g.208 Reducing the thickness of the
RuO2 layer to 3 nm led to an even higher capacitance (1170
F/g at a potential scan rate of 10 mV s-1).182 All of these
studies also demonstrated longer lifetimes and greater
stabilities and rate capabilities, while the specific capacitances
were in general considerably higher than those observed for
pure MWCNTs (<80 F/g) or pure RuO2 (<400 F/g).
In view of the rather high costs and environmental issues
concerning RuO2, and despite the lower specific capacitances,
there has been a clear trend toward the use of hydrous
manganese oxides in the past few years.188 Xie et al. produced
a CNT-MnO2 hybrid by in situ reduction of KMnO4 on
MWCNTs with citric acid and observed that the specific
capacitance increased from 18 to 190 F/g (Figure 26a).168
while Ma et al. reported 250 F/g188 at a large current density
of 1 Ag1-, both for scan rates of 10 mV/s. Zhang et al.
Chemical Reviews, 2010, Vol. 110, No. 3 1375
electrodeposited MnO2 “nanoflowers” onto the surface of
vertical CNT arrays and obtained capacitances of about 200
F/g, as well as excellent rate capabilities (50.8%, capacity
retention at 77 A/g) and a long cycle life (3% capacity loss
after 20 000 charge/discharge cycles).334 The authors pro-
posed a model for the improved electrochemical storage
characteriztics based on high electronic conductivity in CNTs
(Figure 26b). In this model, the geometry of the MnO2
nanoflowers and their hierarchical localization at CNT
junctions limits the diffusion length of ions within the MnO2
phase during the charge/discharge process and enhances the
ionic conductivity of the hybrid.
To maximize the electrochemical utilization of the super-
capacitor, it is desirable to (a) use a low loading of metal
oxide and (b) increase the interfacial area between CNTs
and metal oxide. Consequently, the preparation of a thin,
uniform, connected coating on CNTs is expected to improve
the specific capacitance significantly. This was confirmed
by Nam et al., who observed increasing capacitances with
decreasing NiO content and thus coating thickness.331
A promising approach to enhance the capacitive perfor-
mance has recently been reported by Sivakkumar et al., who
synthesized a ternary hybrid of CNT/polypyrrole/hydrous
MnO2, which showed a significantly higher specific capaci-
tance (280 F/g) compared to the binary CNT/MnO2 (150
F/g).167 The increase in capacitance is believed to be due to
the better dispersion of hydrous MnO2 in the composite
electrode. However, the presence of the polymer also
accounts for the poor cyclability behavior.
Fang et al. chose a different approach by depositing RuO2
on N-doped CNT,246 for which they measured considerably
higher specific capacitances. The authors showed that the
RuO2 particles were considerably better dispersed on N-
doped CNTs than on pure CNTs (Figure 18), due to the
presence of structural defects in N-CNTs. Consequently, the
hybrids with N-doped CNTs had an enhanced interface,
which resulted in their better performance as a supercapacitor
(Figure 27).
A good cyclability is crucial to a successful application
as an anode material for lithium-ion batteries, as demon-
strated by Chen et al. for CNT-SnOx hybrids.202 Pure CNTs
typically show good cycle stability, but also a large and
irreversible capacity loss in the first cycle. The authors
showed that well-dispersed SnOx on CNTs enhanced the
electrochemical performance significantly as the capacity
fade of the hybrid electrode was strongly reduced compared
to unsupported SnOx or pure CNT. The authors attributed
the improvement in performance to the enhanced conductiv-
ity and ductility of the hybrid. This is in line with Zheng et
al., who observed a 5-fold increase in electronic conductivity
and thus enhanced cycling rate and stability in CNT-CuO
hybrids compared to pure CuO.
In summary, electrochemical devices have been among
the first applications for which CNT-inorganic hybrids were
tested. In the long run, CNT hybrids are expected to strongly
enhance the batteries’ intrinsic capacities, stabilities, and
lifetimes, as well as the charging/discharging characteristics.
Furthermore, the substitution of part of the heavy oxide
material (e.g., in car batteries) with the lightweight CNTs
will significantly reduce the weight of the batteries.
1376 Chemical Reviews, 2010, Vol. 110, No. 3 Eder

Figure 27. Capacitance-voltage diagrams of hybrids with RuO2 using undoped and nitrogen-doped CNTs (Figure 18); (a) light (RF
power ) 100 W), (b) moderate (RF power ) 140 W), and (c) heavy (RF power ) 180 W) loading of RuO2. (d) Capacitance of hybrids
as a function of RF sputtering power. In contrast to the undoped CNTs, the N-doped CNTs are coated completely with RuO2, resulting in
a larger interfacial area and an improved capacitance. Reprinted with permission from ref 437. Copyright 2007 Institute of Physics.
4.5. Example 5: Photonics and Field Emission and cost-efficient solution for passive optical regeneration,
Devices long error-free optical transmission distances, and high noise-
suppression capabilities.
Since the early days of CNT research, the prospect of
new CNT-based electronic and photonic devices has Most other works on CNT-inorganic hybrids have
attracted tremendous interest for both fundamental and concentrated on field emission properties (secondary electron
applied science. For instance, as ideal one-dimensional emission, thermionic emission), which have attracted exten-
systems, semiconducting CNTs are direct-bandgap materi- sive attention because of their potential application in flat-
als, which can be used to both generate and detect light panel displays and compact X-ray tubes, high-power micro-
simply by changing the applied voltage. In this respect, wave devices, and fluorescence lamps.342-344 For these
extensive research has been conducted on studying the applications, CNTs are typically grown vertically via a CVD
electronic and optical properties of pure CNTs and technique, which allows both positional control and the
applying them in transistors, field emission devices, and growth of well-aligned CNTs (carpets). This unique geometry
nonlinear optics, and has been summarized recently by and the high aspect ratio of CNTs enable the creation of
Avouris et al.339,340 On the other hand, only a few works very high, localized electric fields, characterized by the
have been directed toward CNT-inorganic hybrids. electric field enhancement factor β. Most of the studies have
Recently, the importance of the oxide substrate for so far focused on improving the β factor of CNTs.344,345
measuring the electrical and optical properties of CNTs However, CNTs also have a relatively high work function
has been highlighted by Lin et al.341 The authors observed (4.5 eV), which typically increases the barrier for electron
that the electrical noise (1/f) was one magnitude higher in tunnelling and thus limits the current density. Ideally, one
supported CNTs than in suspended CNT, thus demonstrating would want to have an emitter that has both a high β factor
that the noise amplitude 1/f is dominated by the trapped and a low work function. This has been achieved by
charges at the CNT-oxide interface. Another example has coating CNTs with a thin layer of a low work function
been shown by Zhu et al., who observed that the optical material, such as SiOx,346 ZnO,347,348 Cr2O3,349 MgO,242,350
properties of CNTs can be affected by a charge transfer BaSrOx,245,351,352 and TiC.213
between CNTs and the ZnO coating, which enables an As an example, Yi et al. employed electron beam
ultrafast nonlinear optical switching behavior.244 Furthermore, deposition to coat well-aligned CNTs with a uniform layer
the combination of the three-photon absorption properties of MgO with thickness ranging between 100 and 500
of ZnO with the saturable absorption properties of CNTs nm.342 Following the creation of secondary electrons, they
creates a multifunctional nanohybrid with promising ap- observed sensationally high secondary electron emission (>1
plications in saturable absorber devices that offer a simple × 104),353,354 compared to ∼103 for pure CNT,355 which
Carbon Nanotube-Inorganic Hybrids
Figure 28. (Left) TEM image of a CNT-ZnO heterojunction,
composed of vertically aligned CNTs and perpendicularly grown
ZnO nanowires. (Right) Field emission current density as a function
of the applied electric field for the randomly oriented ZnO
nanowires, the pristine CNT arrays, and the ordered CNT-ZnO
heterojunction arrays, respectively. The inset corresponds to
Fowler-Nordheim (F-N) plots. The CNT-ZnO hybrid has the
best field emission properties with the lowest turn-on field of 1.8
V/µm and the lowest threshold field of 2.7 V/µm. Reprinted with
permission from ref 347. Copyright 2008 Elsevier Publishing.
strongly depended on the thickness of the coating.350 When
the layer was too thick, the electrons experienced difficulties
in escaping the oxide, while in a layer that was too thin the
few created secondary electrons caused only a small potential
difference. The highest emission values were observed for
a thickness of 200 nm.350 Hence, the authors concluded that
both the geometry of the hybrid and the intrinsic properties
of MgO are crucial to the design of electron emission devices.
Huang et al. deposited a 200 nm thick ZnO layer around
the tips of aligned CNTs via the filtered cathodic vacuum
arc technique,348 while other researchers used the vapor-phase
transport method to produce either a thick film of ZnO240 or
a heterojunction array of CNTs with ZnO nanowires (Figure
28a).347 In all these works, the improved geometries resulted
in higher and stable emission currents of secondary electrons,
as well as lower turn-on and threshold fields and stable
emission currents compared with pure CNTs and pure ZnO.
The thermionic emission of an oxide-modified CNT device
has been investigated by Jin et al.,245,351,352 who used
magnetron sputtering to deposit a 100 nm thick layer of
barium strontium oxide onto CNTs. Because of the decreased
work function of the hybrid (2.1 eV), the current densities
were about 50 times higher than for the pure CNTs (Figure
28b).
Most of these studies have in common that they deposited
rather thick layers of inorganic materials (100-200 nm),
which typically reduced the field enhancement factor β and,
hence, the performance of CNTs as an emitter. An alternative
hybrid material for field emission has been demonstrated by
Pan et al., who used the electron beam technique to deposit
layers of MgO or TiC, which were considerably thinner
(1-10 nm).213,242 Consequently, these thin films did not alter
the geometry and, hence, the β factor of the CNT emitter.
Instead, the electrons were transferred from the CNTs into
the TiC coating (due to the low work functions2.8 eVsof
TiC) and emitted easily without barrier. Therefore, the
coating increased the current density of the CNTs consider-
ably, while also reducing the turn-on voltage from 411 V
for pure CNTs to 267 V in the hybrid.
Another problem for CNT emitters arises from the fact
that emitted electrons tend to ionize surrounding gas
molecules, which then bombard the CNT surface and slowly
degrade their structure and peel them off the substrate. As a
result, the field emission characteristics and emission stability
Chemical Reviews, 2010, Vol. 110, No. 3 1377
Figure 29. Differential scanning calorimetry (DSC) and thermo-
gravimetric analysis (TGA) of uncoated CNTs and CNT-TiO2
hybrid materials, taken in air in a temperature range from 25 to
1000 °C (10 °C/min). The exothermic peak at around 550-600 °C
indicates the oxidation of CNTs (675 °C) and shifts to lower
temperature when coated with anatase (565 °C) or rutile (520 °C),
indicating a reduced oxidation resistance in the hybrid materials.
Reprinted with permission from ref 195. Copyright 2008 Royal
Society of Chemistry.
could be affected.356 Coating the CNTs with materials such
as SiOx or MgO was shown to protect the CNTs from
reactive sputter etching. Moon et al. showed that even a very
thin layer of amorphous SiOx (1-2 nm) could improve the
lifetime of the CNT emitter by a factor of 2.5.346 Although
SiOx improved the current density by a factor of 3, it is this
author’s opinion that a slightly thicker but uniform coating,
covering the whole CNT surface, would improve the
performance even more.
4.6. Example 6: Oxidation Resistance
For many applications, it is important to know about
the oxidation resistance of CNTs in CNT-inorganic
hybrids, since a reduced stability can limit the operation
temperature. In general, CNTs oxidize in air at temper-
atures above 700 °C.357 However, the presence of defects,
catalyst residues, and amorphous carbon can considerably
reduce their oxidation temperature. In contrast, the oxidation
resistance of CNTs can be significantly increased by anneal-
ing in an inert atmosphere at 2000 °C, which produces very
pure and crystalline CNTs with low defect concentration.
A conformal coating of a dense layer of inorganic
material around the CNTs, such as in CNT-inorganic
hybrids, can dramatically affect the oxidation resistance
of CNTs. For instance, the present author observed that a
coating of TiO2-anatase decreased the oxidation temper-
ature to 565 °C (Figure 29), which was about 100 °C lower
than that required for pristine, uncoated CNTs (675 °C).195
Interestingly, the observed oxidation temperature was even
lower when the CNTs were coated with TiO2 in the rutile
phase (520 °C).195
A similar behavior was reported for RuO2 (570 °C)294
and MnO2, for which an oxidation temperature as low as
350 °C was reported.150 So far, the lowest oxidation
temperature, 320 °C, was observed by the present author
1378 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 30. Scheme of Mars-van Krevelen mechanism, involving (a) the reaction of a TiO2-lattice oxygen with carbon to CO or CO2,
(b) the diffusion of an oxygen vacancy to the surface of TiO2, and (c) the reoxidation of the vacancy with gaseous oxygen.
for Bi2O3, a promising material for electrochemical applica-
tions. On the other hand, the deposition of a thin layer of
SiO2 was shown to hinder the combustion of CNTs,
increasing their oxidation temperature by at least 60 °C,215
while Al2O3 and ZrO2 increased the oxidation temperature
to about 800 °C.358
Reducible oxides like TiO2, CeO2, V2O5, and MnO2 can
provide lattice oxygen in promoting the C-O bond formation
to produce CO or CO2, similar to the Mars-van Krevelen
mechanism (Figure 30).359 Upon reaction with CNTs, the
lattice oxygen is consumed and produces an oxygen vacancy,
a reduced Ti3+ species, and an electron, via360,361
C + Ti4+ + O2- f CO + VO + Ti3+ + e-
The oxygen vacancy then diffuses to the surface or triple
phase boundary to be reoxidized by the gaseous oxygen. The
oxygen diffusion is believed to be the rate-limiting step and
is, in the case of TiO2, faster in rutile than in anatase.362
Consequently, the oxidation temperature and heat evolution
in rutile-coated CNTs are lower than those in the CNT-
anatase hybrid (Figure 29). This effect obviously depends
on the reducibility of the oxide. Therefore, the nonreducible
SiO2 and ZrO2 do not catalyze the CNT oxidation; on the
contrary, a complete coverage of CNTs hinders the access
of gaseous oxygen and consequently their oxidation.
4.7. Example 7: Synthesis of Inorganic
Nanotubes
A reduced oxidation stability can be advantageous for
another application of CNT-inorganic hybrids: the fabrica-
tion of inorganic nanotubes, for which CNTs can be used as
a template.363-365,195,262,366 In general, a thin film of the desired
material is deposited on the CNTs, which are subsequently
removed by oxidation in air at the temperature required for
the specific hybrid. This approach was first demonstrated by
Ajayan and co-workers for the synthesis of V2O5 nanotubes357
and has since been used to generate tubular forms of a variety
of oxides, including SiO2 and Al2O3,367,368 MoO3 and ZrO2,369
and RuO2370 and TiO2.195,262,366 Recent examples include
nanotubes of In2O3,371 Co3O4,372 Fe2O3,236 and Eu2O3.373 The
advantage of oxides in nanotubular morphology is their three-
dimensional mechanically coherent architecture, which al-
lows ready gas access to a very high specific surface
area.195,262,366 Consequently, most of the above-mentioned
nanotubes showed excellent performance in gas sensing or
photocatalysis.374-376
Besides their role as a template, the CNTs can further act
as a support during crystallization and phase transformation,
Eder
which led to the first TiO2 nanotubes in rutile phase, as
demonstrated recently by this author.195,262 Out of the
polymorphs of TiO2, anatase is the kinetically favored phase,
while rutile has to be transformed from anatase via heat
treatment at temperatures between 600 and 900 °C.377 These
high temperatures generally induce grain growth and cause
collapse of the tubes’ structure, thus reducing the specific
surface area considerably. However, when the phase trans-
formation was conducted in an inert atmosphere, the
preserved CNTs could act as a support so that the stresses
associated with the reconstruction of the phase did not
destroy the nanotube morphology (Figure 31). Furthermore,
the CNTs prevented the grain growth typical for the anatase
to rutile transition, affecting several layers of rutile crystals
(diameter ≈ 10 nm) within a distance of about 50 nm.
Consequently, the subsequent oxidation of the CNT template
produced phase-pure rutile nanotubes with high surface
area.195,262
This template-based synthesis route has yet another, very
promising potential: the synthesis of electronically modified
inorganic nanotubes.278 The addition of benzyl alcohol (see
section 3.2.2.2) has enabled the use of pristine CNTs as
templates without the need for covalent functionalization.55
Pristine, and thus unpurified, CNTs still contain metal (e.g.,
Fe) catalyst residues (see section 2.2.1), which are often
encapsulated within the tubes’ walls. Recently, the present
author demonstrated that, upon oxidative removal of the CNT
template, these iron residues can be used to produce iron-
doped anatase and rutile nanotubes.378-380 In contrast to other
doping routes, this unique process does not require postan-
nealing, it enables a uniform distribution of iron at high
concentrations as a solute in anatase and rutile nanotubes
without the formation of secondary phases. This iron-doping
has enabled a considerable extension of light absorption into
the visible region (red-shift) as well as a reduced electron-hole
recombination rate.379 These advantages, combined with the
high surface and illumination area of the anatase and rutile
nanotubes, have led to an improved performance in photo-
catalytic applications, such as water splitting.
In summary, the template-directed synthesis provides a
simple, versatile, and cost-effective procedure for pure and
modified inorganic nanotubes. Furthermore, since the CNTs
can be produced with lengths of millimeters, this approach
can be adopted for fabricating very long nanotubes from a
rich pool of materials.
Carbon Nanotube-Inorganic Hybrids
Figure 31. Scheme for the synthesis of iron-doped anatase and rutile nanotubes, using the catalyst residues within CNTs as iron source
(A, B) involving coating of pristine CNTs with TiO2 via sol-gel process (C), heat treatment in argon to either anatase or rutile, and
oxidation of the CNT template in air at 550 °C, during which the iron oxidizes and enters the TiO2 lattice as substituents (Fe3+) for Ti4+
ions.
5. Challenges and Outlook
This review introduces CNT-inorganic hybrids as a new
and very promising class of functional materials that combine
the multiphase characteristics of nanocomposites with the
synergistic functions of hybrid frameworks. In these materi-
als, a synergistic effect in CNT-hybrids arises through size
domain effects and charge transfer processes through the
CNT-inorganic interface. The beneficial role of CNTs is
manifold, ranging from photosensitizers and electron traps
to magnetic shields, intrinsic capacitors, and electrodes.
Because of their high surface area and their excellent thermal
conductivity, CNTs can also support very small nanopar-
ticles, hinder their growth during postannealing treatments,
and even stabilize new phases, so creating new functional
materials with new properties.
This review provides a comprehensive description of the
various synthesis approaches of CNT-inorganic hybrids and
also demonstrates their outstanding potential for a wide range
of applications concerning energy and the environment.
Despite being at a very early stage of research, CNT-inorganic
hybrids have demonstrated considerably enhanced activities
and selectivity in photo- and environmental catalysts, mark-
edly improved sensitivities, and lower operation temperatures
in gas and biosensors, as well as increased capacitances and
extended recyclability, stability, and lifetime in batteries.
Among the many applications described in this review, gas
sensors and batteries have so far enjoyed the greatest attention
and are very likely to become commercially feasible.
Considering the outstanding results of some preliminary
studies, I see the greatest potential of CNT-inorganic hybrids
in their use in photocatalytic applications, including water
and air purification, the production of hydrogen from water,
and the direct energy conversion of sunlight.
The development of these and other applications from
laboratory devices to industrial prototypes, however, will
Chemical Reviews, 2010, Vol. 110, No. 3 1379
require further optimization toward designing the ideal hybrid
material. Among the most important challenges is the control
of the interface, especially for applications that involve
charge transfer processes. However, most hybrid materials
tested for applications have been produced simply by
mechanically mixing the compounds or by growing the
inorganic material on pristine CNTs, both of which often
resulted in irregular and partial coatings and thus a poor
interface. An increased interfacial area and a uniform coating
can considerably improve the performance of the hybrid
materials. This can be achieved with a variety of synthesis
techniques by modifying the surface chemistry by covalent
(functional groups) or noncovalent (surfactant, biomolecules)
means or by using ionic polymers as a glue. Each of these
techniques has advantages and disadvantages that will have
to be considered for designing a new hybrid material. For
instance, covalent functionalization often involves harsh
chemical reactions that, especially when involving ultrasoni-
cation, typically shorten the tubes, roughen the surface, and
even alter the intrinsic (mostly electronic) properties of CNTs
due to the formation of structural defects. In contrast,
noncovalent attraction is nondestructive but may not be
sufficient to guarantee a mechanically stable coating. The
length of the surfactants, capping agents, and other linkers
can be used to control the extent of charge transfer, while
the choice of CNTs (metallic vs semiconducting) may cause
additional barriers at the interface (e.g., metal-insulator
junctions).
The second challenge involves the control of the morphol-
ogy and phase composition of the inorganic compounds,
which can be deposited either as an amorphous or single
crystalline thin film or as a layer of small polycrystalline
nanoparticles. Consequently, the material can be designed
with either high specific surface area or a decreased number
of grain boundaries. Photovoltaics is a good example of the
importance of these properties, as a high surface area enables
1380 Chemical Reviews, 2010, Vol. 110, No. 3 Eder

the adsorption of more dye molecules, while grain boundaries DSSC dye-sensitized solar cell
limit the charge transfer in the oxide. As mentioned above, DTAB dodecyl dimethylammonium bromide
CNTs can support small particle sizes, while also directing EC electrochemical capacitors
their growth orientation. Most importantly, CNTs can even EDLC electrochemical double-layer capacitor
ECB conduction band energy
stabilize new, otherwise unstable phases, leading to novel EG ethylene glycol
CNT-hybrid materials with yet unknown properties. Fur- FED field emission display
thermore, the design of the functionality of the inorganic FET field effect transistor
component may involve layer-by-layer or core-shell struc- ITO indium tin oxide
tures of mixed inorganic compounds as well as complex IPCE internal photoconversion efficiencies
structures (e.g., mixed oxide solutions, perovskites) and LSMCD liquid-source misted chemical deposition
electronically modified materials (e.g., using dopants). In this NaDDBS sodium dodecylbenzene sulfonate
respect, exploration and control of the hybrid’s phase NPs,NWs nanoparticles, nanowires
composition and morphology will likely extend its impact P2 VP poly-2-vinylpyridine
in various applications. PAA polyacrylic acid
PDDA poly(diallyl dimethyl ammonium chloride)
There are also challenges concerning the type and quality PECVD plasma-enhanced chemical vapor deposition
of the CNTs, which must be controlled to further understand PEG poly(ethylene glycol)
the beneficial role of CNTs in the hybrid materials. For PEI poly(ethylene imine)
instance, some applications like photovoltaics or gas sensors PLD pulsed laser deposition
may favor thin SWCNTs with high aspect ratio over PV photovoltaics
MWCNTs. Semiconducting SWCNTs may perform better PVD physical vapor deposition
in charge transfer processes than metallic ones, and conse- PPh3 triphenylphosphine
quently the separation techniques need further improvements PSS poly(sodium 4-styrenesulfonate)
in yield, selectivity, and total conversion. Other important SDS sodium dodecylsulfate
SEE secondary electron emission
issues are the purification of the CNTs from metal residues SEED substrate-enhanced electroless metal deposition
and carbonaceous species (e.g., amorphous carbon, fullerenes, SET single electron transistor
aromatic debris) and the reproducibility of CNT synthesis SHE standard hydrogen electrode
with respect to graphitization and defect concentration. This SPS scanning probe spectroscopy
requires the standardization of synthesis and purification TAA thioacetamide
techniques. I further recommend the use of other carbon- THF tetrahydrofuran
aceous nanomaterials, such as graphene or fullerenes, as well TBOT tetrabutyl orthotitanate/titanium(iv) butoxide
as fibers and transparent thin films made from CNTs, which TEOS tetraethyl orthosilicate/silicon(iv) ethoxide
are very likely to enhance the diversity and adaptability of TEOT tetraethyl orthotitanate/titanium(iv) ethoxide
CNT-inorganic hybrids and also facilitate new applications. TOAB tetraoctylammonium bromide
TOPO trioctylphosphine oxide
Finally, a successful application of CNT-inorganic hy- TTIP titanium(iv) isopropoxide
brids and their implementation into the market requires a VLS vapor-liquid-solid mechanism
strong improvement in methodology to ensure reproducibility QDs quantum dots
and better understanding of the structure-property relation-
ship. Equally important is addressing the biocompatibility 7. Acknowledgments
of the hybrid materials for use in biological and medical
applications. Health and safety regulations further require I would like to thank Profs. A. K. Cheetham, U. Schubert,
detailed studies on toxicology and exposure, and research U. Steiner, and A. H. Windle for helpful discussions and A.
directed toward those issues has just started. White for proofreading. I am further grateful to the Austrian
Although at the early stage of research, this new class of Academy of Science for financial support via the Austrian
materials offers great potential for use in a wide variety of Programme for Advanced Research and Technology (APART).
applications, including sustainable energy, environmental,
and medical applications. As a multidisciplinary research 8. References
field, the further development and successful implementation
(1) Roco, M. C.; Williams, R. S.; Alivisatos, A. P. Nanotechnology
strongly requires the expertise of researchers from chemistry, Research Directions; Kluwer Academic Publishers: Boston, 2000.
physics, and nanotechnology, but also from various areas in (2) Rao, C. N. R.; Cheetham, A. K. J. Mater. Chem. 2001, 11, 2887–
engineering and medicine. In this respect, the purpose of this 2894.
(3) Wolf, S. A.; Awschalom, D. D.; Buhrman, R. A.; Daughton, J. M.;
review is to increase the visibility of CNT-inorganic hybrids Molnar, S. v.; Roukes, M. L.; Chtchelkanova, A. Y.; Treger, D. M.
and to offer a comprehensive overview of recent advances Science 2001, 294, 1488.
for both academia and industry. (4) On, D. T.; Desplantier-Giscard, D.; Danumah, C.; Kaliaguine, S. Appl.
Catal. A: Gen. 2001, 222, 299.
(5) Kickelbick, G. Hybrid Materials: Synthesis, Characterization and
Application; Wiley-VCH: Weinheim, Germany, 2007.
6. Abbreviations (6) Sanchez, C.; Gomez-Romero, P. Functional Hybrid Materials; Wiley
VCH: Weinheim, Germany, 2004.
3Y-TZP tetragonal ZrO2, stabilized with 3 mol % Y2O3 (7) Sanchez, C.; Julian, B.; Belleville, P.; Popall, M. J. Mater. Chem.
AAO anodized aluminum oxide 2005, 15, 3559–3592.
AIIP aluminum isopropoxide (8) Organic-Inorganic Hybrids forPhotonics; Hubert, L., Najafi, S. I.,
ALD atomic layer deposition Eds.; SPIE: Bellingham, WA, 1998; Vol. 3469.
CFE carbon fiber electrode (9) Schubert, U.; Huesing, N.; Lorenz, A. Chem. Mater. 1995, 7, 2010.
CTAB cetyltrimethyl ammoniumbromide (10) Judenstein, P.; Sanchez, C. J. Mater. Chem. 1996, 6, 511.
(11) Cheetham, A. K.; Rao, C. N. R. MRS Bull. 2005, 30, 93.
CVD chemical vapor deposition (12) Holland, B. T.; Blanford, C. F.; Stein, A. Science 1998, 281, 538.
DAFC direct alcohol fuel cell (13) Stein, A.; Schroden, R. C. Curr. Opin. Solid State Mater. Sci. 2001,
DMFC direct methanol fuel cell 5, 553.
Carbon Nanotube-Inorganic Hybrids
(14) Ferey, G. Chem. Mater. 2001, 13, 3084.
(15) Ferey, G.; Mellot-Draznieks, C.; Serre, C.; Millange, F. Acc. Chem.
Res. 2005, 38, 217.
(16) Templin, M.; Franck, A.; Chesne, A. D.; Leist, H.; Zhang, Y.; Ulrich,
R.; Schädler, V.; Wiesner, U. Science 1997, 278, 1795.
(17) Warren, S. C.; Messina, L. C.; Slaughter, L. S.; Kamperman, M.;
Zhou, Q.; Gruner, S. M.; DiSalvo, F. J.; Wiesner, U. Science 2008,
320, 1748.
(18) Yang, P.; Zhao, D.; Margolese, D. I.; Chmelka, B. F.; Stucky, G. D.
Nature 1998, 396, 152.
(19) Yaghi, O. M.; O’Keeffe, M.; Ockwig, N. W.; Chae, H. K.; Eddaoudi,
M.; Kim, J. Nature 2003, 423, 705.
(20) Yaghi, O. M.; Li, G.; Li, H. Nature 1995, 378, 703.
(21) Mellot-Draznieks, C.; Dutour, J.; Ferey, G. Angew. Chem., Int. Ed.
2004, 43, 6290.
(22) Loiseau, T.; Serre, C.; Huguenard, C.; Fink, G.; Taulelle, F.; Henry,
M.; Bataille, T.; Ferey, G. Chem.sEur. J. 2004, 10, 1373.
(23) Terrones, M. Annu. ReV. Mater. Sci. 2003, 33, 419.
(24) Aliev, A. E.; Oh, J.; Kozlov, M. E.; Kuznetsov, A. A.; Fang, S.;
Fonseca, A. F.; Ovalle, R.; Lima, M. D.; Haque, M. H.; Gartstein,
Y. N.; Zhang, M.; Zakhidov, A. A.; Baughman, R. H. Science 2009,
323, 1575.
(25) Pugno, N. M. J. Phys.: Condens. Matter 2006, 18, S1971.
(26) Iijima, S. Nature 1991, 354, 56.
(27) Iijima, S. Nature 1993, 363, 603.
(28) Harris, P. J. F. Carbon Nanotubes and Related Structures; Cambridge
University Press: Cambridge, U.K., 2003.
(29) Kroto, H. W.; Heath, J. R.; O’Brien, S. C.; Curl, R. F.; Smalley,
R. E. Nature 1985, 318, 162.
(30) Dai, H. Acc. Chem. Res. 2002, 35, 1035.
(31) Moisala, A.; Nasibulin, A. G.; Kauppinen, E. I. J. Phys.: Condens.
Matter 2003, 15.
(32) Ajayan, P. M. Chem. ReV. 1999, 99, 1787.
(33) Carbon Nanotubes: Synthesis, Structure, Properties, and Applications;
Dresselhaus, M. S., Dresselhaus, G., Avouris, P., Eds.; Springer: New
York, 2000.
(34) Saito, R.; Dresselhaus, G.; Dresselhaus, M. S. Physical Properties
of Carbon Nanotubes; Imperial College Press: London, 1998.
(35) Terrones, M. Annu. ReV. Mater. Res. 2003, 33, 419.
(36) Dresselhaus, M. S.; Dresselhaus, G.; Charlier, J. C.; Hernandez, E.
Philos. Trans. R. Soc. London, A 2004, 362, 2065.
(37) Wei, B. Q. Appl. Phys. Lett. 2001, 79, 1172.
(38) Robertson, D. H.; Brenner, D. W.; Mintmire, J. W. Phys. ReV. B
1992, 45, 12592.
(39) Yu, M. F. Phys. ReV. Lett. 2000, 84, 5552.
(40) Pugno, N. Acta Materialia 2007, 55, 5269.
(41) Xie, X. L.; Mai, Y. W.; Zhou, X. P. Mater. Sci. Eng. R: Reports
2005, 49, 89.
(42) Berber, S.; Kwon, Y. K.; Tomanek, D. Phys. ReV. Lett. 2000, 84,
4613.
(43) Hone, J.; Batlogg, B.; Benes, Z.; Johnson, A. T.; Fischer, J. E. Science
2000, 289, 1730.
(44) Salzmann, C. G.; Llewellyn, S. A.; Tobias, G.; Ward, M. A. H.;
Huh, Y.; Green, M. L. H. AdV. Mater. 2007, 19, 883.
(45) Banerjee, S.; Hemraj-Benny, T.; Balasubramanian, M.; Fischer, D. A.;
Misewich, J. A.; Wong, S. S. ChemPhysChem 2004, 5, 1416.
(46) Pillai, S. K.; Ray, S. S.; Moodley, M. J. Nanosci. Nanotechnol. 2007,
7, 3011.
(47) Park, T. J.; Banerjee, S.; Hemraj-Benny, T.; Wong, S. S. J. Mater.
Chem. 2006, 16, 141.
(48) Moon, J. M.; An, K. H.; Lee, Y. H.; Park, Y. S.; Bae, D. J.; Park,
G. S. J. Phys. Chem. B 2001, 105, 5677.
(49) Fei, B.; Lu, H.; Hu, Z.; Xin, J. H. Nanotechnology 2006, 17, 1589.
(50) Hou, P.-X.; Liu, C.; Cheng, H.-M. Carbon 2008, 46, 2003.
(51) Jeong, T.; Kim, W. Y.; Haha, Y. B. Chem. Phys. Lett. 2001, 344,
18.
(52) Tobias, G.; Shao, L. D.; Salzmann, C. G.; Huh, Y.; Green, M. L. H.
J. Phys. Chem. B 2006, 110 (45), 22318.
(53) Li, Y.; Zhang, X. B.; Luo, J. H.; Huang, W. Z.; Chen, J. P.; Luo,
Z. Q. Nanotechnology 2004, 15, 1645.
(54) Andrews, R.; Jacques, D.; Qian, D.; Dickey, E. C. Carbon 2001, 39,
1681.
(55) Eder, D.; Windle, A. H. AdV. Mater. 2008, 20, 1787.
(56) Chattopadhyay, D.; Galeska, L.; Papadimitrakopoulos, F. J. Am.
Chem. Soc. 2003, 125, 3370.
(57) Strano, M. S.; Dyke, C. A.; Usrey, M. L.; Barone, P. W.; Allen,
M. J.; Shan, H. W.; Kittrell, C.; Hauge, R. H.; Tour, J. M.; Smalley,
R. E. Science 2003, 301, 1519.
(58) Zheng, M.; Jagota, A.; Strano, M. S.; Santos, A. P.; Barone, P.; Chou,
S. G.; Diner, B. A.; Dresselhaus, M. S.; McLean, R. S.; Onoa, G. B.;
Samsonidze, G. G.; Semke, E. D.; Usrey, M.; Walls, D. J. Science
2003, 302, 1545.
(59) Arnold, M. S.; Stupp, S. I.; Hersam, M. C. Nano Lett. 2005, 5, 713.
Chemical Reviews, 2010, Vol. 110, No. 3 1381
(60) Feng, Q.-P.; Xie, X. M.; Liu, Y.-T.; Gao, Y.-F.; Wang, X.-H.; Ye,
X.-Y. Carbon 2001, 45, 11.
(61) Arnold, M. S.; Green, A. A.; Hulvat, J. F.; Stupp, S. I.; Hersam,
M. C. Nat. Nanotechnol. 2006, 1, 60.
(62) Krupke, R.; Hennrich, F.; Lohneysen, H. v.; Kappes, M. M. Science
2003, 301, 344.
(63) Krupke, R.; Hennrich, F.; Weber, H. B.; Kappes, M. M.; Lohneysen,
H. v. Nano Lett. 2003, 3, 1019.
(64) Krupke, R.; Hennrich, F. AdV. Eng. Mater. 2005, 7, 111.
(65) Tasis, D.; Tagmatarchis, N.; Bianco, A.; Prato, M. Chem. ReV. 2006,
106, 1105.
(66) Prato, M.; Kostarelos, K.; Bianco, A. Acc. Chem. Res. 2008, 41, 60.
(67) Kuzmany, H.; Kukovecz, A.; Simona, F.; Holzweber, M.; Kram-
berger, C.; Pichler, T. Synth. Met. 2004, 41, 113.
(68) Hirsch, A.; Vostrowsky, O. Funct. Mol. Nanostruct. 2005, 245, 193.
(69) Klein, K. L.; Melechko, A. V.; McKnight, T. E.; Retterer, S. T.;
Rack, P. D.; Fowlkes, J. D.; Joy, D. C.; Simpson, M. L. J. Appl.
Phys. 2008, 103, 061301.
(70) Hirsch, A. Angew. Chem., Int. Ed. 2002, 41, 1853.
(71) Li, W. W.; Gao, C.; Qian, H. F.; Ren, J. C.; Yan, D. Y. J. Mater.
Chem. 2006, 16, 1852.
(72) He, P. G.; Dai, L. M. Chem. Commun. 2004, 3, 348.
(73) Lee, C. S.; Baker, S. E.; Marcus, M. S.; Hamers, W. S. J. Nano Lett.
2004, 4, 1713.
(74) Daniel, S.; Raoa, T. P.; Raob, K. S.; Rani, S. U.; Naidu, G. R. K.;
Lee, H.-Y.; Kawai, T. Sens. Actuators, B 2007, 122, 672.
(75) Fu, K. F.; Sun, Y. P. J. Nanosci. Nanotechnol. 2005, 3, 351.
(76) Tasis, D.; Tagmatarchis, N.; Georgakilas, V.; Prato, M. Chem.sEur.
J. 2003, 9, 4000.
(77) Cui, D. X. J. Nanosci. Nanotechnol. 2007, 7, 1298.
(78) Katz, E.; Willner, I. Chem. Phys. Chem. 2004, 5, 1085.
(79) Giordani, S.; Bergin, S. D.; Nicolosi, V.; Lebedkin, S.; Kappes,
M. M.; Blau, W. J.; Coleman, J. N. J. Phys. Chem. B 2006, 110,
15708.
(80) Cambre, S.; Wenseleers, W.; Culin, J.; Doorslaer, S. V.; Fonseca,
A.; Nagy, J.; Goovaerts, E. Chem. Phys. Chem. 2008, 9 (13), 1930.
(81) Jurkschat, K.; Wilkins, S. J.; Salter, C. J.; Leventis, H. C.; Wildgoose,
G. G.; Jiang, L.; Jones, T. G. J.; Crossley, A.; Compton, R. G. Small
2005, 2, 95.
(82) Ajayan, P. M.; lijima, S. Nature 1993, 361, 333.
(83) Ajayan, P. M.; Ebbesen, T. W.; Ichihashi, T.; Iijima, S.; Tanigaki,
K.; Hiura, H. Nature 1993, 362, 522.
(84) Seraphin, S.; Zhou, D.; Jiao, J.; Withers, J. C.; Loufty, R. Nature
1993, 362, 503.
(85) Saito, Y.; Yoshikawa, T. J. Cryst. Growth 1993, 134, 154.
(86) Sloan, J.; Hammer, J.; Zwiefka-Sibley, M.; Green, M. L. H. Chem.
Commun. 1998, 3, 347.
(87) Monthioux, M.; Flahaut, E.; Cleuziou, J. P. J. Mater. Res. 2006, 21,
2774.
(88) Smith, B. W.; Monthioux, M.; Luzzi, D. E. Nature 1998, 296, 323.
(89) Ugarte, D.; Chatelain, A.; Heer, W. A. d. Science 1996, 274, 1897.
(90) Thamavaranukup, N.; Höppe, H. A.; Ruiz-Gonzalez, L.; Costa,
P. M. F. J.; Sloan, J.; Kirkland, A.; Green, M. L. H. Chem. Commun.
2004, 15, 1686.
(91) Watts, P. C. P.; Hsu, W. K.; Kotzeva, V.; Chen, G. Z. Chem. Phys.
Lett. 2002, 366, 42.
(92) Satishkumar, B. C.; Govindaraj, A.; Mofokeng, J.; Subbanna, G. N.;
Rao, C. N. R. J. Phys. Ber. Bunsenges. Phys. Chem. 1996, 29, 4925.
(93) Ugarte, D.; Stöckli, T.; Bonard, J. M.; Châtelain, A.; Heer, W. A. d.
Appl. Phys. A: Mater. Sci. Process. 1998, 67, 101.
(94) Chancolon, J.; Archaimbault, F.; Pineau, A.; Bonnamy, S. J. Nanosci.
Nanotechnol. 2006, 6, 1.
(95) Brown, G.; Bailey, S.; Sloan, J.; Xu, C.; Friedrichs, S.; Flahaut, E.;
Coleman, K. S.; Hutchison, J. L.; Dunin-Borkowski, R. E.; Green,
M. L. H. Chem. Commun. 2001, 845.
(96) Dujardin, E.; Ebbesen, T. W.; Hiura, H.; Taginaki, K. Science 1994,
265, 1850.
(97) Kim, B. M.; Qian, S.; Bau, H. H. Nano Lett. 2005, 5, 873.
(98) Monthioux, M. Carbon 2002, 40, 1809.
(99) Zhang, B.; Liu, C.; Cheng, H. M.; Cai, Q. K. New Carbon Mater.
2003, 18, 174.
(100) Carter, R.; Sloan, J.; Kirkland, A. I.; Meyer, R. R.; Lindan, P. J. D.;
Lin, G.; Green, M. L. H.; Vlandas, A.; Hutchison, J. L.; Harding, J.
Phys. ReV. Lett. 2006, 96, 215501.
(101) Georgakilas, V.; Gournis, D.; Tzitzios, V.; Pasquato, L.; Guldi, D. M.;
Prato, M. J. Mater. Chem. 2007, 17, 2679.
(102) Wildgoose, G. G.; Banks, C. E.; Compton, R. G. Small 2006, 2, 182.
(103) Chen, J.; Hamon, M. A.; Hu, H.; Chen, Y. S.; Rao, A. M.; Ecklund,
P. C.; Haddon, R. C. Science 1998, 282, 95.
(104) Haremza, J. M.; Hahn, M. A.; Krauss, T. D.; Chen, S.; Calcines, J.
Nano Lett. 2002, 2, 1253.
(105) Banerjee, S.; Wong, S. S. Nano Lett. 2002, 2, 195.
1382 Chemical Reviews, 2010, Vol. 110, No. 3
(106) Ravindran, S.; Chaudhary, S.; Colburn, B.; Ozkan, M.; Ozkan, C. S.
Nano Lett. 2003, 3, 447.
(107) Daniel, M. C.; Astruc, D. Chem. ReV. 2004, 104, 293.
(108) Azamian, B. R.; Coleman, K. S.; Davis, J. J.; Hanson, N.; Green,
M. L. H. Chem. Commun. 2002, 366.
(109) Zanella, R.; Basiuk, E. V.; Santiago, P.; Basiuk, V. A.; Mireles, E.;
Puente-Lee, I.; Saniger, J. M. J. Phys. Chem. B 2005, 109, 16290.
(110) Wang, G.-X.; Zhang, B.-L.; Yu, Z.-L.; Qu, M.-Z. Solid State Ionics
2005, 176, 1169.
(111) Liu, B.; Chen, J. H.; Xiao, C. H.; Cui, K. Z.; Yang, L.; Pang, H. L.;
Kuang, Y. F. Energy Fuels 2007, 21, 1365.
(112) Kongkanand, A.; Dominguez, R. M.; Kamat, P. V. Nano Lett. 2007,
7, 676.
(113) Juan, J. C.; Jiang, Y.; Meng, X.; Cao, W.; Yarmo, M. A.; Zhang, J.
Mater. Res. Bull. 2007, 42, 1278.
(114) Sainsbury, T.; Fitzmaurice, D. Chem. Mater. 2004, 16, 3780.
(115) Lee, C.-L.; Ju, Y.-C.; Chou, P.-T.; Huang, Y.-C.; Kuo, L.-C.; Oung,
J.-C. Electrochem. Commun. 2004, 7, 453.
(116) Wei, X.-W.; Xu, J.; Song, X.-J.; Ni, Y.-H. Zhongguo Youse Jinshu
Xuebao 2004, 14, 236.
(117) Whitsitt, E. A.; Moore, V. C.; Smalley, R. E.; Barron, A. R. J. Mater.
Chem. 2005, 15, 4678.
(118) Whitsitt, E. A.; Barron, A. R. Nano Lett. 2003, 3, 775.
(119) Ellis, A. V.; Vijayamohanan, K.; Goswami, R.; Chakrapani, N.;
Ramanathan, L. S.; Ajayan, P. M.; Ramanath, G. Nano Lett. 2003,
3, 279.
(120) Rahman, G. M.; Guldi, D. M.; Zambon, E.; Pasquato, L.; Tagma-
tarchis, N.; Prato, M. Small 2005, 1, 527.
(121) Han, L.; Wu, W.; Kirk, F. L.; Luo, J.; Maye, M. M.; Kariuki, N. N.;
Lin, Y.; Wang, C.; Zhong, C.-J. Langmuir 2004, 20.
(122) Sainsbury, T.; Fitzmaurice, D. Chem. Mater. 2004, 16, 2174.
(123) Yang, D. Q.; Hennequin, B.; Sacher, E. Chem. Mater. 2006, 18, 5033.
(124) Guldi, D. M.; Rahman, G. M. A.; Jux, N.; Tagmatarchis, N.; Prato,
M. Angew. Chem., Int. Ed. 2004, 43, 5526.
(125) Chen, R. J.; Zhang, Y.; Wang, D.; Dai, H. J. Am. Chem. Soc. 2001,
123, 3838.
(126) Murakami, H.; Nomura, T.; Nakashima, N. Chem. Phys. Lett. 2003,
378, 481.
(127) Satake, A.; Miyajima, Y.; Kobuke, Y. Chem. Mater. 2005, 17, 716.
(128) Rahman, G. M. A.; Guldi, D. M.; Campidelli, S.; Prato, M. J. Mater.
Chem. 2006, 16, 62.
(129) Wang, X.; Liu, Y.; Qiu, W.; Zhu, D. J. Mater. Chem. 2002, 12, 1636.
(130) D’Souza, F.; Chitta, R.; Sandanayaka, A. S. D.; Subbaiyan, N. K.;
D’Souza, L.; Araki, Y.; Ito, O. Chem.sEur. J. 2007, 13, 8277.
(131) Mu, Y.; Liang, H.; Hu, J.; Jiang, L.; Wan, L. J. Phys. Chem. B 2005,
109, 22212.
(132) Liu, L.; Wang, T.; Li, J.; Guo, Z.-X.; Dai, L.; Zhang, D.; Zhu, D.
Chem. Phys. Lett. 2003, 367, 747.
(133) Ou, Y. Y.; Huang, M. H. J. Phys. Chem. B 2006, 110, 2031.
(134) Li, C. S.; Tang, Y. P.; Yao, K. F.; Zhou, F.; Ma, Q.; Lin, H.; Tao,
M. S.; Liang, J. Carbon 2006, 44, 2021.
(135) Georgakilas, V.; Tzitzios, V.; Gournis, D.; Petridis, D. Chem. Mater.
2005, 17, 1613.
(136) Zhou, J.; Zhou, X.; Sun, X.; Lib, R.; Murphy, M.; Ding, Z.; Sun,
X.; Sham, T.-K. Chem. Phys. Lett. 2007, 437, 229.
(137) Guldi, D. M.; Taieb, H.; Rahman, G. M. A.; Tagmatarchis, N.; Prato,
M. AdV. Mater. 2005, 17, 871.
(138) Correa-Duarte, M. A.; Liz-Marzán, L. M. J. Mater. Chem. 2006,
16, 22.
(139) Chen, H.; Donga, S. Talanta 2006, 71, 1752.
(140) Ostranger, J. W.; Mamedov, A. A.; Kotov, N. A. J. Am. Chem. Soc.
2001, 123, 1101.
(141) Jiang, K.; Eitan, A.; Schadler, L. S.; Ajayan, P. M.; Siegel, R. W.;
Grobert, N.; Mayne, M.; Reyes-Reyes, M.; Terrones, H.; Terrones,
M. Nano Lett. 2003, 3, 275.
(142) Correa-Duarte, M. A.; Perez-Juste, J.; Sanchez-Iglesias, A.; Giersig,
M. Angew. Chem., Int. Ed. 2005, 44, 4375.
(143) Kim, B.; Sigmund, W. M. Langmuir 2004, 20, 8239.
(144) Grzelczak, M.; Correa-Duarte, M. A.; Salgueirino-Maceira, V.;
Giersig, M.; Diaz, R. AdV. Mater. 2006, 18, 415.
(145) Correa-Duarte, M. A.; Liz-Marzan, L. M. J. Mater. Chem. 2006,
16, 22.
(146) Hu, X.; Wang, T.; Qu, X.; Dong, S. J. Phys. Chem. B 2006, 110,
853.
(147) Wang, J.; Musameh, M.; Lin, Y. J. Am. Chem. Soc. 2003, 125, 2408.
(148) Fullam, S.; Cottell, D.; Rensmo, H.; Fitzmaurice, D. AdV. Mater.
Chem. Phys. 2000, 12, 1430.
(149) Sun, J.; Gao, L. J. Electroceram. 2006, 17, 91.
(150) Sun, J.; Iwasa, M.; Gao, L.; Zhang, Q. Carbon 2004, 42, 895.
(151) Stoffelbach, F.; Aqil, A.; Jerome, C.; Jerome, R.; Detrembleur, C.
Chem. Commun. 2005, 4532.
(152) Guo, D. J.; Li, H. L. Electrochem. Commun. 2004, 6, 999.
Eder
(153) Day, T. M.; Unwin, P. R.; Wilson, N. R.; Macpherson, J. V. J. Am.
Chem. Soc. 2005, 127, 10639.
(154) Quinn, B. M.; Dekker, C.; Lemay, S. G. J. Am. Chem. Soc. 2005,
127, 6147.
(155) Qu, J.; Shen, Y.; Qu, X.; Dong, S. Chem. Commun. 2004, 34.
(156) He, Z.; Chen, J.; Liu, D.; Zhou, H.; Kuang, Y. Diamond Relat. Mater.
2004, 13, 1764.
(157) Yoon, B.; Wai, C. M. J. Am. Chem. Soc. 2005, 127, 17174.
(158) Ye, J. S.; Cui, H. F.; Liu, X.; Lim, T. M.; Zhang, W. D.; Sheu, F. S.
Small 2005, 1, 560.
(159) Kong, J.; Chapline, M. G.; Dai, H. AdV. Mater. 2001, 13, 1384.
(160) Kim, H. S.; Lee, H.; Han, K. S.; Kim, J. H.; Song, M. S.; Park,
M. S.; Lee, J. Y.; Kang, J. K. J. Phys. Chem. B 2005, 109, 8983.
(161) Guo, D. J.; Li, H. L. J. Electroanal. Chem. 2004, 573, 197.
(162) Xing, Y. J. Phys. Chem. B 2004, 108, 19255.
(163) Lin, Y.; Cui, X.; Yen, C.; Wai, C. M. J. Phys. Chem. B 2005, 109,
14410.
(164) Cao, J.; Sun, J. Z.; Hong, J.; Li, H. Y.; Chen, H. Z.; Wang, M. AdV.
Mater. 2004, 16, 84.
(165) Robel, I.; Bunker, B. A.; Kamat, P. V. AdV. Mater. 2005, 17, 2458.
(166) Li, X. L.; Liu, Y. Q.; Fu, L.; Cao, L. C.; Wei, D. C.; Wang, Y. AdV.
Funct. Mater. 2006, 16, 2431.
(167) Sivakkumar, S. R.; Ko, J. M.; Kim, D. Y.; Kim, B. C.; Wallace,
G. G. Electrochim. Acta 2007, 52, 7377.
(168) Xie, X.; Gao, L. Carbon 2007, 45, 2365.
(169) Huang, X. P.; Pan, C. X.; Huang, X. T. Mater. Lett. 2007, 61, 934.
(170) Huang, B.; Huang, R.; Jin, D.; Ye, D. Catal. Today 2007, 126, 279.
(171) Huang, H.; Zhang, W.; Li, M.; Gan, Y.; Chen, J.; Kuang, Y. J. Colloid
Interface Sci. 2005, 284, 593.
(172) Kyotani, T.; Tsai, L.-f.; Tomita, A. Chem. Commun. 1997, 701.
(173) Carmo, M.; Paganin, V. A.; Rosolen, J. M.; Gonzalez, E. R. J. Power
Sources 2005, 142, 169.
(174) Ye, X. R.; Lin, Y.; Wai, C. M. Chem. Commun. 2003, 9, 642.
(175) Fukunaga, A.; Chu, S.; McHenry, M. E. J. Mater. Sci. Lett. 1999,
18, 431.
(176) Giordano, R.; Serp, P.; Kalck, P.; Kihn, Y.; Schreiber, J.; Marhic,
C.; Duvail, J.-L. Eur. J. Inorg. Chem. 2003, 610.
(177) Moriguchi, I.; Hidaka, R.; Yamada, H.; Kudo, T.; Murakami, H.;
Nakashima, N. AdV. Mater. 2006, 18, 69.
(178) Sun, J.; Gao, L.; Iwasa, M. Chem. Commun. 2004, 832.
(179) Lee, C. Y.; Tsai, H. M.; Chuang, H. J.; Li, S. Y.; Lin, P.; Tseng,
T. Y. J. Electrochem. Soc. 2005, 152, A716.
(180) Frank, O.; Kalbac, M.; Kavan, L.; Zukalova, M.; Prochazka, J.;
Klementova, M.; Dunsch, L. Phys. Status Solidi 2007, 244, 4040.
(181) He, K.-X.; Wu, Q.-F.; Zhang, X.-G.; Wang, X.-L. J. Electrochem.
Soc. 2006, 153, A1568.
(182) Kim, I.-H.; Kim, J.-H.; Lee, Y.-H.; Kim, K.-B. J. Electrochem. Soc.
2005, 152, A2170.
(183) Kim, J. D.; Kang, B. S.; Noh, T. W.; Yoon, J.-G.; Baik, S. I.; Kim,
Y.-W. J. Electrochem. Soc. 2005, 152, D23.
(184) Wang, C.; Waje, M.; Wang, X.; Tang, J. M.; Haddon, R. C.; Yan,
Y. Nano Lett. 2004, 4, 345.
(185) Qu, L.; Dai, L. J. Am. Chem. Soc. 2005, 127, 10806.
(186) Choi, H. C.; Shim, M.; Bangsaruntip, S.; Dai, H. J. Am. Chem. Soc.
2002, 124, 9058.
(187) Qu, L.; Dai, L.; Osawa, E. J. Am. Chem. Soc. 2006, 128, 5523.
(188) Ma, S.-B.; Nam, K.-W.; Yoon, W.-S.; Yang, X.-Q.; Ahn, K.-Y.; Oh,
K.-H.; Kim, K.-B. J. Power Sources 2008, 178, 483.
(189) Ma, S.-B.; Nam, K.-W.; Yoon, W.-S.; Yang, X.-Q.; Ahn, K.-Y.; Oh,
K.-H.; Kim, K.-B. Electrochem. Commun. 2007, 9, 2807.
(190) Brinker, C. J.; Scherer, G. W. Sol-Gel SciencesThe Physics and
Chemistry of Sol-Gel Processing; Academic Press: New York, 1990.
(191) Vincent, P.; Brioude, A.; Journet, C.; Rabaste, S.; Purcell, S. T.;
Brusq, J. L.; Plenet, J. C. J. Non-Cryst. Solids 2002, 311, 130.
(192) Jitianu, A.; Cacciaguerra, T.; Benoit, R.; Delpeux, S.; Beguin, F.;
Bonnamy, S. Carbon 2004, 42, 1147.
(193) Jitianu, A.; Cacciaguerra, T.; Berger, M. H.; Benoit, R.; Beguin, F.;
Bonnamy, S. J. Non-Cryst. Solids 2004, 345-46, 596.
(194) Zhao, L. P.; Gao, L. Carbon 2004, 42, 1858.
(195) Eder, D.; Windle, A. H. J. Mater. Chem. 2008, 18, 2036.
(196) Yu, Y.; Yu, J. C.; Yu, J.-G.; Kwok, Y.-C.; Che, Y.-K.; Zhao, J.-C.;
Ding, L.; Ge, W.-K.; Wong, P.-K. Appl. Catal., A 2005, 289, 186–
196.
(197) Song, H.; Qiu, X.; Li, F.; Zhu, W.; Chen, L. Electrochem. Commun.
2007, 9, 1416.
(198) Han, W.-Q.; Zettl, A. Nano Lett. 2003, 3, 681.
(199) Bai, J. Y.; Xu, Z. D.; Zheng, Y. F. Chem. Lett. 2006, 35, 96.
(200) Li, Y.; Ding, J.; Chen, J.; Xu, C.; Wei, B.; Liang, J.; Wu, D. Mater.
Res. Bull. 2002, 37, 313.
(201) Liang, Y. X.; Chen, Y. J.; Wanga, T. H. Appl. Phys. Lett. 2004, 85,
666.
(202) Chen, M. H.; Huang, Z. C.; Wu, G. T.; Zhu, G. M.; You, J. K.; Lin,
Z. G. Mater. Res. Bull. 2003, 38, 831.
Carbon Nanotube-Inorganic Hybrids
(203) Wei, B.-Y.; Hsu, M.-C.; Su, P.-G.; Lin, H.-M.; Wu, R.-J.; Lai, H.-J.
Sens. Actuators, B 2004, 101, 81.
(204) Wang, W. D.; Serp, P.; Kalck, P.; Silva, C. G.; Faria, J. L. Mater.
Res. Bull. 2008, 43, 958.
(205) Wang, W. D.; Serp, P.; Kalck, P.; Faria, J. L. J. Mol. Catal. A: Chem.
2005, 235, 194.
(206) Wang, W. D.; Serp, P.; Kalck, P.; Faria, J. L. Appl. Catal., B 2005,
56, 305.
(207) Arabale, G.; Wagh, D.; Kulkarni, M.; Mulla, I. S.; Vernekar, S. P.;
Vijayamohanan, K.; Rao, A. M. Chem. Phys. Lett. 2003, 376, 207.
(208) Park, J. H.; Ko, J. M.; Parka, O. O. J. Electrochem. Soc. 2003, 150,
A864.
(209) Lee, J. Y.; Liang, K.; Ana, K. H.; Lee, Y. H. Synth. Met. 2005, 150,
153.
(210) Wu, R.-J.; Wu, J.-G.; Yu, M.-R.; T, T.-K; Yeh, C.-T. Sens. Actuators,
B 2008, 131, 306.
(211) Jiang, L. Q.; Gao, L. J. Am. Ceram. Soc. 2006, 89, 156.
(212) Jiang, L. Q.; Gao, L. J. Mater. Chem. 2005, 15, 260.
(213) Pan, L.; Shoji, T.; Nagataki, A.; Nakayama, Y. AdV. Eng. Mater.
2007, 9, 584.
(214) Shi, X.; Yang, H.; Sun, P.; Shao, G.; Duan, X.; Zhen, X. Carbon
2007, 45, 1735.
(215) Bourlinos, A. B.; Georgakilas, V.; Zboril, R.; Dallas, P. Carbon 2007,
45, 2136.
(216) Cao, L.; Scheiba, F.; Roth, C.; Schweiger, F.; Cremers, C.; Stimming,
U.; Fuess, H.; Chen, L.; Zhu, W.; Qiu, X. Angew. Chem., Int. Ed.
2006, 45 (32), 5315.
(217) Wei, X. W.; Song, X. H.; Xu, J.; Ni, Y. H.; Zhang, P. Mater. Chem.
Phys. 2005, 92, 159.
(218) Hernadi, K.; Ljubovic, E.; Seo, J. W.; Forro, L. Acta Mater. 2003,
51, 1447.
(219) Seeger, T.; Kohler, T.; Frauenheim, T.; Grobert, N.; Ruhle, M.;
Terrones, M.; Seifert, G. Chem. Commun. 2002, 34.
(220) Vietmeyer, F.; Seger, B.; Kamat, P. V. AdV. Mater. 2007, 19, 2935–
2940.
(221) Jiang, L.; Gao, L. Mater. Chem. Phys. 2005, 91, 313.
(222) Yoshimura, M.; Byrappa, K. J. Mater. Sci. 2008, 43, 2085.
(223) Menzel, R.; Peiró, A. M.; Durrant, J. R.; Shaffer, M. S. P. Chem.
Mater. 2006, 18, 6059.
(224) Byrappa, K.; Dayananda, A. S.; Sajan, C. P.; Basavalingu, B.; Shayan,
M. B.; Soga, K.; Yoshimura, M. J. Mater. Sci. 2008, 43, 2348.
(225) Lee, Y.; Song, H. J.; Shin, H. S.; Shin, H. J.; Choi, H. C. Small
2005, 1, 975.
(226) Huang, Y.; Lin, J.; Ding, X. X.; Tang, C.; Gu, C. Z.; Qi, S. R. Mater.
Lett. 2007, 61, 697.
(227) Zhang, W. D. Nanotechnology 2006, 17, 1036–1040.
(228) Du, J.; Fu, L.; Liu, Z.; Han, B.; Li, Z.; Liu, Y.; Sun, Z.; Zhu, D. J.
Phys. Chem. B 2005, 109, 12772.
(229) Yu, Y.; Ma, L. L.; Huang, W. Y.; Li, J. L.; Wong, P. K.; Yu, J. C.
J. Solid State Chem. 2005, 178, 1488.
(230) Yu, Y.; Ma, L. L.; Huang, W. Y.; Du, F. P.; Yu, J. C.; Yu, J. G.;
Wang, J. B.; Wong, P. K. Carbon 2005, 43, 670.
(231) Jia, B. P.; Gao, L.; Sun, J. Carbon 2007, 45, 1476.
(232) Fu, L.; Liu, Z. M.; Liu, Y. Q.; Han, B. X.; Wang, J. Q.; Hu, P. A.;
Cao, L. C.; Zhu, D. B. AdV. Mater. 2004, 16, 350.
(233) Sun, Z. Y.; Zhang, X. R.; Han, B. X.; Wu, Y. Y.; An, G. M.; Liu,
Z. M.; Miao, S. D.; Miao, Z. J. Carbon 2007, 45, 2589.
(234) Sun, X.; Chu, Y.; Wang, D. W.; Du, J. H.; Zhang, B. Y.; Wang,
F. P. Rare Metals 2007, 26, 191.
(235) Sun, Z.; Liu, Z.; Han, B.; An, G. Mater. Lett. 2007, 61, 4565.
(236) Sun, Z.; Yuan, H.; Liu, Z.; Han, B.; Zhang, X. AdV. Mater. 2005,
17, 2993.
(237) Sun, Z.; Liu, Z.; Han, B.; Miao, S.; Du, J.; Miao, Z. Carbon 2006,
44, 888.
(238) Kawasaki, S.; Scott, J. F.; Fan, H.; Catalan, G.; Saad, M. M.; Gregg,
J. M.; Correa, M. A.; Morrison, F. D.; Tatsuta, T.; Tsuji, O. Integr.
Ferroelectr. 2007, 95, 180.
(239) Kim, H.; Sigmund, W. Appl. Phys. Lett. 2002, 81, 2085.
(240) Yu, K.; Zhang, Y. S.; Xu, F.; Li, Q.; Zhua, Z. Q. Appl. Phys. Lett.
2006, 88, 153123.
(241) Zhang, Y.; Franklin, N. W.; Chen, R. J.; Dai, H. Chem. Phys. Lett.
2000, 331, 35.
(242) Pan, L.; Konishi, Y.; Tanaka, H. J. Vac. Sci. Technol., B 2007, 25,
1581.
(243) Kim, H. W.; Shim, S. H.; Lee, J. W. Carbon 2007, 45, 2695.
(244) Zhu, Y. W.; Elim, H. I.; Foo, Y. L.; Yu, T.; Liu, Y. J.; Ji, W.; Lee,
J. Y.; Shen, Z. X.; Wee, A. T. S.; Thong, J. T. L.; Sow, C. H. AdV.
Mater. 2006, 18, 587.
(245) Jin, F.; Liu, Y.; Day, C. M. Appl. Phys. Lett. 2007, 90, 143114.
(246) Fang, W.-C.; Chyan, O.; Sun, C.-L.; Wu, C.-T.; Chen, C.-P.; Chen,
K.-H.; Chen, L.-C.; Huang, J.-H. Electrochem. Commun. 2007, 9,
239.
Chemical Reviews, 2010, Vol. 110, No. 3 1383
(247) Ikuno, T.; Yasuda, T.; Honda, S. I.; Oura, K.; Katayama, M.; Lee,
J. G.; Mori, H. J. Appl. Phys. 2005, 98, 114305.
(248) Ikuno, T.; Katayama, M.; Kamada, K.; Honda, S.; Lee, J. G.; Mori,
H.; Oura, K. Jpn. J. Appl. Phys. Lett. 2003, 42, L1356.
(249) Kawasaki, S.; Catalan, G.; Fan, H.; Saad, M. M.; Gregg, J. M.;
Correa-Duarte, M. A.; Rybczynski, J.; Morrison, F. D.; Tatsuta, T.;
Tsuji, O.; Scott, J. F. Appl. Phys. Lett. 2008, 92, 053109.
(250) Kuang, Q.; Li, S. F.; Xie, Z. X.; Lin, S. C.; Zhang, X. H.; Xie, S. Y.;
Huang, R. B.; Zheng, L. S. Carbon 2006, 44, 1166.
(251) Deng, G. H.; Xiao, X.; Chen, J. H.; Zeng, X. B.; He, D. L.; Kuang,
Y. F. Carbon 2005, 43, 1557.
(252) Peng, H. B.; Golovchenko, J. A. Appl. Phys. Lett. 2004, 84, 5428.
(253) Fan, Z.; Chen, J.; Cui, K.; Sun, F.; Xu, Y.; Kuang, Y. Electrochim.
Acta 2007, 2007, 2959.
(254) Chrissanthopoulos, A.; Baskoutas, S.; Bouropoulos, N.; Dracopoulos,
V.; Tasis, D.; Yannopoulos, S. N. Thin Solid Films 2007, 515, 8524.
(255) Lazareck, A. D.; Kuo, T.-F.; Xu, J. M.; Taft, B. J.; Kelley, S. O.;
Cloutier, S. G. Appl. Phys. Lett. 2006, 89, 103109.
(256) Lazareck, A. D.; Cloutier, S. G.; Kuo, T.-F.; Taft, B. J.; Kelley, S. O.;
Xu, J. M. Nanotechnology 2006, 17, 2661.
(257) Liu, J.; Li, X.; Dai, L. AdV. Mater. 2006, 18, 1740.
(258) Ho, Y. M.; Liu, J. W.; Qi, J. L.; Zheng, W. T. J. Phys. D: Appl.
Phys. 2008, 41, 665308.
(259) Gomathi, A.; Vivekchand, S. R. C.; Govindaraj, A.; Rao, C. N. R.
AdV. Mater. 2005, 17, 2757.
(260) Javey, A.; Kim, H.; Brink, M.; Wang, Q.; Ural, A.; Guo, J.; Mcintyre,
P.; McEuen, P.; Lundstrom, M.; Dai, H. Nat. Mater. 2002, 1, 241.
(261) Min, Y.-S.; Bae, E. J.; Jeong, K. S.; Cho, Y. J.; Lee, J.-H.; Choi,
W. B.; Park, G.-S. AdV. Mater. 2003, 15, 1019.
(262) Eder, D.; Kinloch, I. A.; Windle, A. H. Chem. Commun. 2006, 13,
1448.
(263) Fujishima, A.; Honda, K. Nature 1972, 238, 37.
(264) Tryk, D. A.; Fujishima, A.; Honda, K. Electrochim. Acta 2000, 45,
2363.
(265) Graetzel, M.; Rotzinger, F. P. Chem. Phys. Lett. 1985, 118, 474.
(266) Graetzel, M. Nature 2001, 414, 338.
(267) O’Regan, B.; Graetzel, M. Nature 1991, 353, 737.
(268) Carp, O.; Huisman, C. L.; Reller, A. Prog. Solid State Chem. 2004,
32, 33.
(269) Diebold, U. Surf. Sci. Rep. 2003, 48, 53.
(270) Frank, S. N.; Bard, A. J. J. Am. Chem. Soc. 1977, 99, 303.
(271) Aguado, M. A.; Anderson, M. A.; Hill, C. G. J. Mol. Catal. 1994,
89 (1–2), 165.
(272) Hoffmann, M. R.; Martin, S. T.; Choi, W. Y.; Bahnemann, D. W.
Chem. ReV. 1995, 95, 69.
(273) Zhang, Z. B.; Wang, C. C.; Zakaria, R.; Ying, J. Y. J. Phys. Chem.
B 1998, 102, 10871.
(274) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Science
2001, 293.
(275) Zhang, Y. H.; Reller, A. J. Mater. Chem. 2004, 11, 2537.
(276) Wang, J.; Una, S.; Klabunde, K. J. Appl. Catal., B 2004, 48, 151.
(277) Eder, D.; Motta, M. S.; Windle, A. H. Nanotechnology 2009, 20,
055602.
(278) Eder, D.; Motta, M. S.; Windle, A. H. Acta Mater. submitted for
publication.
(279) Anpo, M.; Ichihashi, Y.; Takeuchi, M.; Yamashita, H. Res. Chem.
Intermed. 1998, 24, 143.
(280) Ou, Y.; Lin, J.; Fang, S.; Liao, D. Chem. Phys. Lett. 2006, 429, 199.
(281) Bouazza, N.; Ouzzine, M.; Lillo-Ródenas, M. A.; Eder, D.; Linares-
Solano, A. Appl. Catal. B: EnViron. 2009, 92 (3–4), 377.
(282) Yao, Y.; Li, G.; Ciston, S.; Lueptow, R. M.; Gray, K. A. EnViron.
Sci. Technol. 2008, 42, 4952.
(283) Lee, S.-H.; Pumprueg, S.; Moudgil, B.; Sigmund, W. Colloids Surf.,
B 2005, 40, 93.
(284) Jayadevan, K. P.; Tseng, T. Y. J. Nanosci. Nanotechnol. 2005, 5,
1768.
(285) Baxter, J. B.; Walker, A. M.; Ommering, K. v.; Aydil, E. S.
Nanotechnology 2006, 17, S304.
(286) Kamat, P. V. J. Phys. Chem. C 2007, 111, 2834.
(287) Guldi, D. M.; Rahman, G. M. A.; Sgobba, V.; Kotov, N. A.; Bonifazi,
D.; Prato, M. J. Am. Chem. Soc. 2006, 128, 2315.
(288) Varghese, O. K.; Kichambre, P. D.; Gong, D.; Ong, K. G.; Dickey,
E. C.; Grimes, C. A. Sens. Actuators, B 2001, 81, 32.
(289) Huang, H.-C.; Huang, G.-L.; Chen, H.-L.; Lee, Y.-D. Thin Solid Films
2006, 511-512, 203.
(290) Jusys, Z.; Behm, R. J. J. Phys. Chem. B 2001, 105, 10874.
(291) Rajesh, B.; Thampi, K. R.; Bonard, J. M.; Viswanathan, B. J. Mater.
Chem. 2000, 10, 1757.
(292) Wang, M.-Y.; Chen, J.-H.; Cui, K.-Z.; Liu, B.; Zhou, H.-H.; Kuang,
Y.-Z. Chin. J. Chem. 2006, 24, 881.
(293) Song, H.; Qiu, X.; Li, F. Electrochim. Acta 2008, 53, 3708.
(294) Yu, H.; Zeng, K.; Fu, X.; Zhang, Y.; Peng, F.; Wang, H.; Yang, J.
J. Phys. Chem. C 2008, 112, 11875.
1384 Chemical Reviews, 2010, Vol. 110, No. 3
(295) Li, C.-H.; Yu, Z.-X.; Yao, K.-F.; Ji, S.-F.; Liang, J. J. Mol. Catal.
A: Chem. 2005, 226, 101.
(296) Tessonnier, J. P.; Pesant, L.; Ehret, G.; Ledoux, M. J.; Huu, C. P.
Appl. Catal., A 2005, 288, 203.
(297) Planeix, J. M.; Coustel, N.; Coq, B.; Brotons, V.; Kumbhar, P. S.;
Dutartre, R.; Geneste, P.; Bernier, P.; Ajayan, P. M. J. Am. Chem.
Soc. 1994, 116, 7935.
(298) Fu, X.; Yu, H.; Peng, F.; Wang, H.; Qian, Y. Appl. Catal., A 2007,
321, 190.
(299) Chen, X.-W.; Zhu, Z.; Haevecker, M.; Su, D. S.; Schloegl, R. Mater.
Res. Bull. 2007, 42, 354.
(300) Pietruszka, B.; DiGregorio, F.; Keller, N.; Keller, V. Catal. Today
2005, 102-103, 94.
(301) Yang, H.-M.; Liao, P.-H. Appl. Catal., A 2007, 317, 226.
(302) Zhang, D.; Pan, C.; Shi, L.; Mai, H.; Gao, X. Appl. Surf. Sci. 2009,
255, 4907.
(303) Chen, X. W.; Zhu, Z.; Hävecker, M.; Su, D. S.; Schlögl, R. Mater.
Res. Bull. 2007, 42, 354.
(304) Liu, X.; Su, D. S.; Schloegl, R. Carbon 2008, 46, 547.
(305) Zhang, J.; Liu, X.; Blume, R.; Zhang, A. H.; Schlogl, R.; Su, D. S.
Science 2008, 322, 73.
(306) Zhang, J.; Su, D. S.; Zhang, A.; Wang, D.; Schlögl, R.; Hebert, C.
Angew. Chem., Int. Ed. 2007, 46, 7319.
(307) Mestl, G.; Maksimova, N.; Keller, N.; Roddatis, V. V.; Schlögl, R.
Angew. Chem., Int. Ed. 2001, 40, 2066.
(308) Jimenez, I.; Centeno, M. A.; Scotti, R.; Morazzoni, F.; Cornet, A.;
Morante, J. R. J. Electrochem. Soc. 2003, 150, H72.
(309) Frühberger, B.; Grunze, M.; Dwyer, D. J. Sens. Actuators, B 1996,
31, 167.
(310) Wu, N. L.; Wang, S. Y.; Rusakova, I. A. Science 1999, 285, 1375.
(311) Ottaviano, L.; Bussolotti, F.; Lozzi, L.; Passacantando, M.; Rosa,
S. L.; Santucci, S. Thin Solid Films 2003, 436, 9.
(312) Gao, T.; Wang, T. H. Chem. Commun. 2004, 2558.
(313) Cabot, A.; Marsal, A.; Arbiol, J.; Morante, J. R. Sens. Actuators, B
2004, 99, 74.
(314) Chen, L.; Tsang, S. C. Sens. Actuators, B 2003, 89, 68.
(315) Collins, P. G.; Bradley, K.; Ishigami, M.; Zettl, A. Science 2000,
287, 1801.
(316) Sinha, N.; Ma, J. Z.; Yeow, J. T. W. J. Nanosci. Nanotechnol. 2006,
6, 573.
(317) Kong, J.; Franklin, N. R.; Zhou, C.; Chapline, M. G.; Peng, S.; Cho,
K.; Dai, H. Science 2000, 287, 622.
(318) Sun, Z.; Zhang, X.; Na, N.; Liu, Z.; Han, B.; An, G. J. Phys. Chem.
B 2006, 110, 13410.
(319) Van Hieu, N.; Thuy, L. T. B.; Chien, N. D. Sens. Actuators, B 2008,
129, 888.
(320) Liu, J.; Guo, Z.; Meng, F.; Jia, Y.; Liu, J. J. Phys. Chem. C 2008,
112, 6119.
(321) Chen, Y.; Zhu, C.; Wang, T. Nanotechnology 2006, 17, 3012.
(322) Kang, X.; Maia, Z.; Zou, X.; Cai, P.; Moa, J. Talanta 2008, 74, 879.
(323) Hsu, H.-L.; Jehng, J.-M. Mat. Chem. Phys. 2009, 113 (1), 166.
(324) Liu, Z.; Wang, J.; Xie, D.; Chen, G. Small 2008, 4, 462.
(325) Conway, B. E. J. Electrochem. Soc. 1991, 138, 1539.
(326) Niu, C.; Sichel, E. K.; Hoch, R.; Moy, D.; Tennet, H. Appl. Phys.
Lett. 1997, 70, 1480.
(327) Lota, G.; Lota, K.; Frackowiak, E. Electrochem. Commun. 2007, 9,
1828.
(328) Hughes, M.; Shaffer, M. S. P.; Renouf, A. C.; Singh, C.; Chen, G. Z.;
Fray, D. J.; Windle, A. H. AdV. Mater. 2002, 14, 382.
(329) Hughes, M.; Chen, G. Z.; Shaffer, M. S. P.; Fray, D. J.; Windle,
A. H. Chem. Mater. 2002, 14, 1610.
(330) Yi, W.; Yu, S. G.; Lee, W. T.; Han, I. T.; Jeong, J. T.; Woo, Y. S.;
Lee, J. H.; Jin, S. W.; Choi, W. B.; Heo, J. N.; Jeon, D.; Kim, J. M.
J. Appl. Phys. 2001, 89, 4091.
(331) Nam, K. W.; Kim, K. H.; Lee, E. S.; Yoon, W. S.; Yang, X. Q.;
Kim, K. B. J. Power Sources 2008, 182, 642.
(332) Subramanian, V.; Zhu, H. W.; Wei, B. Q. Electrochem. Commun.
2006, 8, 827.
(333) Deng, M. G.; Yang, B. C.; Zhang, Z. A.; Hu, Y. D. J. Mater. Sci.
2005, 40, 1017.
(334) Zhang, H.; Cao, G.; Wang, Z.; Yang, Y.; Shi, Z. Nano Lett. 2008, 8,
2664.
(335) Reddy, A. L. M.; Shaijumon, M. M.; Gowda, S. R.; Ajayan, P. M.
Nano Lett. 2009, 9, 1002.
(336) Wu, G.-M.; Wang, A.-R.; Zhang, M.-X.; Yang, H.-Y.; Zhou, B.;
Shen, J. J. Sol-Gel Sci. Technol. 2008, 46, 79.
(337) Fang, W.-C. J. Phys. Chem. C 2008, 112, 11552.
(338) Kim, Y.-T.; Tadai, K.; Mitani, T. J. Mater. Chem. 2005, 15, 4914.
(339) Avouris, P.; Chen, J. Mater. Today 2006, 9, 46.
(340) Avouris, P.; Freitag, M.; Perebeinos, V. Nat. Photonics 2008, 2, 341.
(341) Lin, Y.-M.; Tsang, J. C.; Freitag, M.; Avouris, P. Nanotechnology
2007, 18, 295202.
Eder
(342) Yi, W. K.; Jeong, T.; Yu, S. G.; Heo, J.; Lee, C.; Lee, J.; Kim, W.;
Yoo, J. B.; Kim, J. AdV. Mater. 2002, 14, 1464.
(343) Zhang, Y. B.; Lau, S. P.; Huang, L.; Tanemura, M. Appl. Phys. Lett.
205, 86, 123115.
(344) Shih, A.; Yater, J.; Hor, C.; Abrams, R. Appl. Surf. Sci. 1997, 111,
251.
(345) Besteman, K.; Lee, J. O.; Wiertz, F. G. M.; Heering, H. A.; Dekker,
C. Nano Lett. 2003, 3, 727.
(346) Moon, J. S.; Alegaonkar, P. S.; Han, J. H.; Lee, T. Y.; Yoo, J. B.;
Kim, J. M. J. Appl. Phys. 2006, 100, 104303.
(347) Yan, X.; Tay, B.-K.; Miele, P. Carbon 2008, 46, 753.
(348) Huang, L.; Lau, S. P.; Yang, H. Y.; Yu, S. F. Nanotechnology 2006,
17, 1564.
(349) Huo, K.; Outlaw, R. A.; Wang, S.; Zhu, M.; Quinlan, R. A.; Manos,
D. M.; Kordesch, M. E.; Arp, U.; Holloway, B. C. Appl. Phys. Lett.
2008, 92, 133112.
(350) Heo, J. N.; Kim, W. S.; Jeong, T. W.; Yu, S.; Lee, J. H.; Lee, C. S.;
Yi, W. K; Lee, Y. H.; Yoo, J. B.; Kim, J. M. Physica B 2002, 323,
174.
(351) Jin, F.; Liu, Y.; Day, C. M.; Little, S. A. Carbon 2007, 45, 587.
(352) Jin, F.; Liu, Y.; Day, C. M. Appl. Phys. Lett. 2006, 88, 163116.
(353) Nam, K. W.; Kim, K. B. J. Electrochem. Soc. 2002, 149, A346.
(354) Kim, W. S.; Lee, J.; Jeong, T. W.; Heo, J. N.; Kong, B. Y.; Jin,
Y. W.; Kim, J. M. Appl. Phys. Lett. 2005, 87, 163112.
(355) Collins, P. G.; Zettl, A. Phys. ReV. B 1997, 55, 9391.
(356) Milne, W. I.; Teo, K. B. K.; Chhowalla, M.; Amaratunga, G. A. J.;
Lee, S. B.; Hasko, D. G.; Ahmed, H.; Groening, O.; Legagneux, P.;
Gangloff, L.; Schnell, J. P.; Pirio, G.; Pribat, D.; Castignolles, M.;
Loiseau, A.; Semet, V.; Binh, V. T. Diamond Relat. Mater. 2003,
12, 422.
(357) Ajayan, P. M.; Stephan, O.; Redlich, P.; Colliex, C. Nature 1995,
375, 564.
(358) Eder, D. 2009, submitted for publication.
(359) Mars, P.; Krevelen, D. W. v. Chem. Eng. Sci. (Special issue) 1954,
3, 41.
(360) Haerudin, H.; Bertel, S.; Kramer, R. J. Chem. Soc., Faraday Trans.
1998, 94, 1481.
(361) Eder, D.; Kramer, R. Phys. Chem. Chem. Phys. 2003, 5, 1314.
(362) Eder, D.; Kramer, R. J. Phys. Chem. B 2004, 108, 14823.
(363) Caruso, R. A.; Antonietti, M. Chem. Mater. 2001, 13, 3272.
(364) Patzke, G. R.; Krumeich, F.; Nesper, R. Angew. Chem., Int. Ed. 2002,
41, 2446.
(365) Rao, C. N. R.; Nash, M. Dalton Trans. 2003, 1, 1.
(366) Eder, D.; Motta, M. S.; Kinloch, I. A.; Windle, A. H. Physica E
2007, 37, 245.
(367) Satishkumar, B. C.; Govindaraj, A.; Vogl, E. M.; Baumallick, L.;
Rao, C. N. R. J. Mater. Res. 1997, 12, 604.
(368) Kim, M.; Hong, J.; Lee, J.; Hong, C. K.; Shim, S. E. J. Colloid
Interface Sci. 2008, 322, 321.
(369) Rao, C. N. R.; Satishkumar, B. C.; Govindaraj, A. Chem. Commun.
1997, 1581.
(370) Satishkumar, B. C.; Govindaraj, A.; Nath, M.; Rao, C. N. R. J. Mater.
Chem. 2000, 10, 2115.
(371) Du, N.; Zhang, H.; Chen, B.; Ma, X.; Liu, Z.; Wu, J.; Yang, D. AdV.
Mater. 2007, 19, 1641.
(372) Du, N.; Zhang, H.; Chen, B.; Wu, J. B.; Ma, X. Y.; Liu, Z. H.; Zhang,
Y. Q.; Yang, D.; Huang, X. H.; Tu, J. P. AdV. Mater. 2007, 19, 4505.
(373) Yang, H.; Zhang, D.; Shi, L.; Fang, J. Acta Mater. 2008, 56, 955.
(374) Mor, G. K.; Varghese, O. K.; Paulose, M.; Grimes, C. A. Sens. Lett.
2003, 1, 42.
(375) Adachi, M.; Murata, Y.; Okada, I.; Yoshikawa, S. J. Electrochem.
Soc. 2003, 150, G488.
(376) Lin, C. H.; Lee, C. H.; Chao, J. H.; Kuo, C. Y.; Cheng, Y. C.; Huang,
W. N.; Chang, H. W.; Huang, Y. M.; Skih, M. K. Catal. Lett. 2004,
98, 61.
(377) Yoganarasinhan, S. R.; Rao, C. N. R. Trans. Faraday Soc. 1962,
58, 1579.
(378) Deleted in proof.
(379) Eder, D.; Motta, M. S.; Windle, A. H. Nanotechnology 2009, 20,
055602.
(380) Eder, D.; Sandeman, K. G.; Windle, A. H. Phys. ReV. B 2009, in
press.
(381) Ijima, S. Nature 1991, 354, 56.
(382) Ebbesen, T. W.; Ajayan, P. M. Nature 1992, 358, 220.
(383) Guo, T.; Nikolaev, P.; Thess, A.; Colbert, D. T.; Smalley, R. E. Chem.
Phys. Lett. 1995, 1,2, 243.
(384) Hsu, W. K.; Hare, J. P.; Terrones, M.; Kroto, H. W.; Walton, D. R.;
Harris, P. J. F. Nature 1995, 377, 687.
(385) Endo, M.; Takeuchi, K.; Igarashi, S.; Susumu, K.; Kiyoharu, S. M.;
Kroto, H. W. J. Phys. Chem. Solids 1993, 12, 54.
(386) Hernadi, M.; Couteau, E.; Seo, J. W.; Forro, L. Langmuir 2003, 19,
7026.
(387) Balani, K.; Agarwal, A. Nanotechnology 2008, 19, 165701.
Carbon Nanotube-Inorganic Hybrids
(388) Park, C. D.; Jeon, H. J.; Wang, H. J.; Choa, Y. H.; Oh, S. T.; Kang,
K. M.; Kang, S. G. Proceedings of the 3rd International Symposium
on Designing, Processing and Properties of Advanced Engineering
Materials, ISAEM-2003, Jeju Island, Korea, November 5-8, 2003.
In Materials Science Forum; Kang, S.-G., Kobayashi, T., Eds.; Trans
Tech Publications Inc.: Zurich, Switzerland, 2004; Vol. 449.
(389) Jiang, D. T.; Thomson, K.; Kuntz, J. D.; Ager, J. W.; Mukherjee,
A. K. Scr. Mater. 2007, 56, 959.
(390) Jin, F.; Liu, Y.; Day, C. M.; Little, S. A. J. Vac. Sci. Technol., B
2007, 25, 1785.
(391) Zhang, D.; Pan, C.; Zhang, J.; Shi, L. Mater. Lett. 2008, 62, 3821.
(392) Shan, Y.; Gao, L. Chem. Lett. 2004, 33, 1560.
(393) Ma, L. L.; Yu, Y.; Huang, W. Y.; Zhu, L. P.; Li, J. L.; Zhuang,
Y. Y.; Qi, X. H. Acta Chim. Sinica 2005, 63, 1641.
(394) Reddy, K. R.; Sin, B. C.; Yoo, C. H.; Park, W. J.; Ryu, K. S.; Lee,
J. S.; Sohn, D. W.; Lee, Y. I. Scr. Mater. 2008, 58, 1010.
(395) Mahajan, S. V.; Hasan, S. A.; Cho, J.; Shaffer, M. S. P.; Boccaccini,
A. R.; Dickerson, J. H. Nanotechnology 2008, 19, 195301.
(396) Jiang, L. Q.; Gao, L. Chem. Mater. 2003, 15, 2848.
(397) Peng, F.; Fu, X. B.; Yu, H.; Wang, H. J. New Carbon Materials
2007, 22, 213.
(398) Lee, W. T.; Im, S. J.; Han, J. G.; Yu, S.; Kim, H. Y.; Han, I. T.;
Yoo, J. B.; Kim, K. Y.; Lee, J. W.; Kim, J. M.; Choi, E. H. Jpn.
J. Appl. Phys. 2002, 41, 6550.
(399) Gong, K. P.; Yu, P.; Su, L.; Xiong, S. X.; Mao, L. Q. J. Phys. Chem.
C 2007, 111, 1882.
(400) Fan, Z.; Chen, J.; Wang, M.; Cui, K.; Zhou, H.; Kuang, Y. Diamond.
Relat. Mater. 2006, 15, 1478.
(401) Bottini, M.; Tautz, L.; Huynh, H.; Monosov, E.; Bottini, N.; Dawson,
M. I.; Bellucci, S.; Mustelin, T. Chem. Commun. 2005, 758.
(402) Fu, Q.; Lu, C.; Liu, J. Nano Lett. 2002, 2, 329.
(403) Seeger, T.; de la Fuente, G.; Maser, W. K.; Benito, A. M.; Callejas,
M. A.; Martinez, M. T. Nanotechnology 2003, 14, 184.
(404) Fan, W. G.; Gao, L. Chem. Lett. 2005, 34, 954.
(405) Seeger, T.; Kohler, T.; Frauenheim, T.; Grobert, N.; Terrones, M.;
Seifert, G.; Ruhle, M. Z. Metallkd. 2002, 93, 455.
(406) Fang, H. T.; Sun, X.; Qian, L. H.; Wang, D. W.; Li, F.; Chu, Y.;
Wang, F. P.; Cheng, H. M. J. Phys. Chem. C 2008, 112, 5790.
(407) An, G. M.; Na, N.; Zhang, X. R.; Miao, Z. J.; Miao, S. D.; Ding,
K. L.; Liu, Z. M. Nanotechnology 2007, 18, 435707.
(408) Wen, Z. H.; Wang, Q.; Zhang, Q.; Li, J. H. AdV. Funct. Mater. 2007,
17, 2772.
(409) Fan, W. G.; Gao, L.; Sun, J. J. Am. Ceram. Soc. 2006, 89, 2671.
(410) Wu, N.; Wang, S.; Rusakova, I. A. Science 1999, 285, 1375.
(411) Zhao, L.; Choi, M.; Kim, H. S.; Hong, S. H. Nanotechnology 2007,
18, 445501.
(412) Liu, Y. L.; Yang, H. F.; Yang, Y.; Liu, Z. M.; Shen, G. L.; Yu,
R. Q. Thin Solid Films 2006, 497, 355.
Chemical Reviews, 2010, Vol. 110, No. 3 1385
(413) Lu, J.; Ma, A.; Yang, S. H.; Ng, K. M. J. Nanosci. Nanotechnol.
2007, 7, 1589.
(414) Han, Z. H.; Yang, B.; Kim, S. H.; Zachariah, M. R. Nanotechnology
2007, 18, 105701.
(415) Reddy, A. L. M.; Ramaprabhu, S. J. Phys. Chem. C 2007, 111, 7727.
(416) Yang, Y.; Qu, L.; Dai, L.; Kang, T.-S.; Durstock, M. AdV. Mater.
2007, 19, 1239.
(417) Oh, W. C.; Chen, M. L. Bull. Korean Chem. Soc. 2008, 29, 159.
(418) Huang, Q.; Gao, L. J. Mater. Chem. 2003, 13, 1517.
(419) Liu, B.; Zeng, H. C. Chem. Mater. 2008, 20, 2711.
(420) Fan, W.; Gao, L.; Sun, J. J. Am. Ceram. Soc. 2006, 89, 731.
(421) Wang, G.-J.; Lee, M.-W.; Chen, Y.-H. Photochem. Photobiol. 2008,
20.
(422) Xia, X.-H.; Jia, Z.-J.; Yu, Y.; Liang, Y.; Wang, Z.; Ma, L.-L. Carbon
2007, 45, 717.
(423) Jung, K. H.; Jang, S. R.; Vittal, R.; Kim, V. D.; Kim, K. J. Bull.
Korean Chem. Soc. 2003, 24, 1501.
(424) He, D.; Yang, L.; S. Kuang, Q. C. Electrochem. Commun. 2007, 9,
2467.
(425) Kalbac, M.; Frank, O.; Kavan, L.; Zukalova, M.; Prochazka, J.;
Klementova, M.; Dunsch, L. J. Electrochem. Soc. 2007, 154, K19.
(426) Kim, I.-H.; Kim, J.-H.; Cho, B.-W.; Kim, K.-B. J. Electrochem. Soc.
2006, 153, A1451.
(427) Bittencourt, C.; Felten, A.; Espinosa, E. H.; Ionescu, R.; Moreau,
N.; Heszler, P.; Granqvist, C. G.; Pireaux, J.-J.; Llobet, E. Smart
Mater. Struct. 2006, 15, 1555.
(428) Zhang, R.; Fan, L.; Fang, Y.; Yang, S. J. Mater. Chem. 2008, 18,
4964.
(429) Chen, C. S.; Chen, X. H.; Yi, B.; Liu, T. G.; Li, W. H.; Xu, L. S.;
Yang, Z.; Zhang, H.; Wang, Y. G. Acta Mater. 2006, 54, 5401.
(430) Guo, G.; Guo, J.; Tao, D.; Choy, W. C. H.; Zhao, L.; Qian, W.;
Wang, Z. Appl. Phys.-Mater. Sci. Proc. 2007, 89, 525.
(431) Lupo, F.; Kamalakaran, R.; Scheu, C.; Grobert, N.; Ruhle, M. Carbon
2004, 42, 1995.
(432) Shan, Y.; Gao, L. Nanotechnology 2005, 16, 625.
(433) Ci, L.; Ryu, Z.; Jin-Phillipp, N. Y.; Rühle, M. Diamond. Relat. Mater.
2007, 16, 531.
(434) Pang, L. X.; Sun, K. N.; Ren, S.; Sun, C.; Bi, J. Q. J. Comp. Mater.
2007, 41, 2025.
(435) Correa-Duarte, M. A.; Liz-Marzan, L. M. J. Mater. Chem. 2006,
15, 22.
(436) Engtrakul, C.; Kim, Y. H.; Nedeljkovic, J. M.; Ahrenkiel, S. P.;
Gilbert, K. E. H.; Alleman, J. L.; Zhang, S. B.; Micic, O. I.; Nozik,
A. J.; Heben, M. J. J. Phys. Chem. B 2006, 110, 25153.
(437) Fang, W.-C.; Chen, K.-H.; Chen, L.-C. Nanotechnology 2007, 18,
485716.
CR800433K
1386 Chem. Rev. 2010, 110, 1386–1434

N,N-Dialkylhydrazones in Organic Synthesis. From Simple

N,N-Dimethylhydrazones to Supported Chiral Auxiliaries

Ryszard Lazny* and Aneta Nodzewska

Institute of Chemistry, University of Bialystok, ul. Hurtowa 15-399, Bialystok, Poland

Received February 20, 2009

Contents 12.1.2. N,N-Dialkylhydrazones as Ligands for 1419

Pd-Catalyzed Allylation
1. Introduction 1386 12.1.3. N,N-Dialkylhydrazones as Ligands for 1420
2. General Reactivity Characteristics of 1387 Pd-Catalyzed Mizoroki-Heck,
N,N-Dialkylhydrazones and Their Azaenolates Sonogashira, and Hiyama Reactions
3. Preparation and Cleavage of 1389 12.1.4. N,N-Dialkylhydrazones as Ligands for 1420
N,N-Dialkylhydrazones Pd-Catalyzed Asymmetric Suzuki-Miyaura
4. Short History of Synthetic Applications 1390 Cross-Coupling
5. Formation of Azaenolates 1391 12.1.5. N,N-Dialkylhydrazones as Ligands for 1421
6. Reactions of Azaenolates with Electrophiles: 1391 Pd-Catalyzed Allyl Cross-Coupling
Carbon-Carbon Bond Formation Reactions and Cu-Catalyzed N-Arylation of
Amides and Azoles
6.1. Typical Electrophilic Reactions of Azaenolates 1391
(Prior to 2000) 12.1.6. N,N-Dialkylhydrazones as Ligands in Lewis 1421
Acid-Catalyzed Diels-Alder Reaction
6.2. Alkylations 1392
12.1.7. Lewis Acid-Catalyzed Addition of 1421
6.2.1. Alkylation of N,N-Dimethylhydrazones 1392 Formaldehyde N,N-Dialkylhydrazones to
6.2.2. Alkylation of Zincated 1393 Hydroxy Enones
N,N-Dialkylhydrazones 12.2. Organocatalytic Reactions (Brønsted Acid 1421
6.2.3. Alkylations of Chiral N,N-Dialkylhydrazones 1393 Catalysis) of N,N-Dialkylhydrazones
6.2.4. Alkylation of Dioxanone Hydrazones 1396 12.2.1. Addition of Aldehyde Hydrazone to Imines 1422
6.3. Aldol Reactions 1401 12.2.2. Addition of Formaldehyde Hydrazone to 1422
6.4. Michael Reactions 1402 β,γ-Unsaturated Keto Esters
7. Reactions of Azaenolates with Electrophiles: 1402 12.2.3. Addition of Aldehyde Hydrazone to 1422
Carbon-Heteroatom Bond Formation R,β-Unsaturated Nitroalkenes
8. Reactions at the Azomethine Carbon with 1404 12.3. Lewis Acid Catalysis 1423
Electrophiles 13. N,N-Dialkylhydrazones in Solid-Phase Organic 1423
8.1. Typical Reactions of Formaldehyde and Higher 1404 Synthesis
Aldehyde Hydrazones (Prior to 2000) 13.1. 1,2-Addition of Organolithiums to 1424
8.2. 1,2-Addition of N,N-Dialkylhydrazones to 1404 Polymer-Supported Hydrazones
Aldehydes 13.2. R-Alkylation of Polymer-Supported Hydrazones 1424
8.3. Michael Reactions 1405 13.3. Other Polymer-Supported Reactions 1426
9. Nucleophilic Additions to N,N-Dialkylhydrazones 1406 13.3.1. Polymer-Supported Reactions of 1426
9.1. Reactions of Hydrazones with Organometallics 1406 Azomethine Carbon Nucleophiles
9.2. R-Alkylation Combined with 1,2-Addition to the 1410 13.3.2. Polymer-Supported Multicomponent 1427
Hydrazone Reactions
10. Formation of Heterocycles 1411 14. Hydrazones as Protecting Groups and 1427
10.1. Hetero-Diels-Alder Cycloaddition 1412 Miscellaneous Applications
10.2. Staudinger-like [2 + 2] Cycloaddition 1413 15. Conclusions and Outlook 1429
10.3. Miscellaneous Cyclization Reactions 1414 16. List of Abbreviations 1430
11. Free Radical Reactions 1416 17. Acknowledgments 1430
11.1. Radical Cyclization Reactions 1416 18. References 1430
11.2. Radical Addition of Silicon-Tethered Vinyl and 1417
Acetylene Groups 1. Introduction
12. Catalytic Reactions 1418
The well-known properties of the carbonyl group render
12.1. Catalytic Reactions of Hydrazone Catalysts in 1418
Organometallic Chemistry aldehydes and ketones prominent substrates in both C-C
and C-heteroatom bond forming methodologies of organic
12.1.1. N,N-Dialkylhydrazone Catalysts for Addition 1418
of Organometallics to Aldehydes synthesis. In practical syntheses, nitrogen analogues of
aldehydes and ketones such as enamines, imines, or hydra-
* To whom correspondence should be addressed. E-mail: lazny@uwb.edu.pl. zones, acting as synthetic equivalents of the carbonyl
10.1021/cr900067y  2010 American Chemical Society
Published on Web 12/14/2009
N,N-Dialkylhydrazones in Organic Synthesis
Ryszard Lazny was born in Olsztyn, Poland. He received his M.Sc. degree
in chemistry in 1990 from the University of Warsaw, Bialystok Branch.
His thesis supervisor was Prof. J. W. Morzycki. After one year as a research
assistant at the University, he began postgraduate studies at the University
of Saskatchewan, Canada, working under the supervision of Prof. M.
Majewski. He completed his Ph.D. in 1996. Following graduation, he
worked with Prof. D. E. Ward at the University of Saskatchewan and then
with Prof. D. Enders at the Institute of Organic Chemistry, RWTH-Aachen
University, Germany. After returning to Poland, he researched and lectured
at the University of Bialystok and received his habilitation (D.Sc.) in 2005
from the Institute of Organic Chemistry, Polish Academy of Sciences in
Warsaw. He is presently a Professor at the University of Bialystok. Professor
Lazny has wide-ranging research experience that includes asymmetric
synthesis with chiral reagents (enantioselective deprotonation) and chiral
auxiliaries (SAMP/RAMP-methods), synthesis of cyclic depsipeptides
(destruxin B), synthesis of chiral phosphinoalcohol ligands (asymmetric
phosphinylation), synthesis of tropane alkaloids, and development of solid-
phase synthesis methodologies based on triazene and hydrazone linkers.
His current research interests extend to polymer-supported reagents and
catalysts, reactions in the presence of water, and organocatalytic processes.
In addition to research papers, he is the author/coauthor of several review
articles and has made a contribution to Science of Synthesis and a chapter
in Linker Strategies in Solid-Phase Organic Synthesis. His most rewarding
role, thus far, has been that of father to his six year old son, Patryk.
Aneta Nodzewska was born in Monki near Bialystok, Poland. She
graduated from the University of Bialystok in 2002 (M.Sc. thesis under
the supervision of Prof. R. Lazny). Her graduate thesis concerned
polymer supported reactions and the preparation of polymers with
triazene linkers. After graduation, she started work as an assistant at
the Institute of Chemistry, University of Bialystok. Currently, she is
carrying out research on the development and synthetic applications
of N,N-dialkylhydrazone linkers and reactions of polymer supported
N,N-dialkylhydrazones. Her time is skillfully divided between research,
teaching, and raising her two daughters: 6 year old Emilia and 1 year
old Izabela.
compounds, are often preferred. From a synthesis point of
view, N,N-dialkylhydrazones offer many advantages: high
nucleophilicity of their metalated species, regioselectivity,
controlled R-monoalkylation, and the possibility of using the
hydrazone moiety as a chiral auxiliary or a multifunctional
linker in solid-phase synthesis.
Chemical Reviews, 2010, Vol. 110, No. 3 1387
Even though N,N-dialkylhydrazones show very diverse
reactivity and can participate in polar, free radical, pericyclic,
and organometallic catalytic reactions, it is their reactivity
as aldehyde and ketone synthetic equivalents, where they
participate in nucleophile-electrophile interactions, that has
found major uses in synthesis. The goal of this review is to
bring the reader up to date regarding the synthetic utility of
these important and versatile reagents.
In this review, both early achiral and later chiral hydrazone
methods prominent in synthesis will be recalled. Since
excellent reviews1-4 have already been published on the
subject, the reactivity, cleavage, and formation of hydrazones
will be presented only in brief. The short overview of
reactivity and typical (classical) synthetic applications is
intended to help an unfamiliar reader gain an appreciation
of synthetic applications of hydrazones prior to 2000. The
20th century scientific literature is host to a number of
reviews on the hydrazones, and thus, only the most important
concepts that set the stage for later development will be
presented. Reference to the published reviews, as well as
selected landmark primary literature, will be made for the
convenience of the reader. In the main part of the paper,
recent synthetic applications (2000-2009) of aldehyde and
ketone N,N-dialkylhydrazones, including novel catalytic and
solid-phase methodologies, will be described comprehen-
sively. Reactivity and applications of N-acylhydrazones have
been reviewed5,6 and are outside the scope of this review,
although they parallel to some extent those of N,N-dialkyl-
hydrazones.7
2. General Reactivity Characteristics of
N,N-Dialkylhydrazones and Their Azaenolates
The usefulness of N,N-dialkylhydrazones in synthetic
organic chemistry arises mostly from the rich reactivity of
the hydrazones and the high reactivity of hydrazone-derived
organometallic species, particularly organolithium deriva-
tives. Hydrazones have two bonds susceptible to cleavage,
yet usually are stable enough to allow problem-free trans-
formations in other parts of the hydrazone molecule. The
CdN bond is susceptible to hydrolytic, oxidative, and
reductive cleavage, restoring the carbonyl group,8 and the
N-N bond is predisposed to reductive cleavage to produce
primary amines (Figure 1). The carbon atom of the CdN
bond (the azomethine carbon atom) is prone to nucleophilic
addition of organometallic nucleophiles such as organolithi-
ums, organomagnesiums, and organoceriums or organoyt-
terbiums. On the other hand, the same carbon atom of
formaldehyde hydrazone and some aromatic hydrazones,
demonstrating the azaenamine character of the hydrazones,
may be attacked under suitable conditions by electrophiles
such as Michael acceptors.9 These reactions may take place
on either the nitrogen or carbon atom of the azaenamine
ambident nucleophile. However, for the N,N-disubstituted
hydrazones the reaction on the nitrogen atom despite higher
nitrogen nucleophilicity is a reversible process, contrary to
the reaction on the azomethine carbon, which, in effect, is
the prevailing pathway for the azaenamine hydrazone
reactivity.
The acidity of hydrogens at the R-carbon atom of the
hydrazone group is deemed to be lowered by ca. 10 orders
of magnitude compared to the C-H acidity of the parent
carbonyl compound; the evaluated pKa of the hydrazone is
ca. 30 compared to the pKa of the corresponding ketone of
ca. 20. The lower C-H acidity grants higher reactivity of
1388 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 1. Typical reactivity sites of N,N-dialkylhydrazones.
the hydrazone conjugate bases toward electrophiles, an
advantage compared to synthetically equivalent carbonyl
compounds. The hydrazone C-H acidity and the stability
of the metalated hydrazones (owing to coordination of the
metal with the hydrazone nitrogen atoms) are high enough
to allow, usually, R-metalation of the hydrazones with alkali-
metal amides such as LDA,10,11 LTMP12 (at 0 °C in 2-4 h),
and KDA13 or with alkyllithium bases such as n-BuLi and
t-BuLi (at -78 °C in 15-120 min).11,14 Other possible
metalating bases for the formation of stabilized azaenolates
are NaH in the presence of HMPA, potassium amide, KH,
and NaHMDS.4 On the other hand, the C-H acidity of
hydrazones is low enough to prevent any racemization of
stereogenic R-centers of chiral hydrazones by typical bases
(basic glass, carbonates, hydroxides, alkoxides), which is in
sharp contrast to high racemization rates of analogous ketones
and especially aldehydes (silylation of laboratory glassware
is often needed to prevent enolizable R-stereocenters of
aldehydes or ketones from racemization caused by the basic
glass surface).15 The deprotonation regioselectivity of hy-
drazones is typically high and predictable.16,17 It takes place
at the less substituted carbon atom unless there is an anion-
stabilizing group present at the competing site.
Subsequent alkylation of the formed azaenolates gives, as
a result, regioselectively functionalized or branched hydra-
zones. It is worth noting that alkylation of hydrazones occurs
selectively at the R-carbon unlike ketones or aldehydes,
where O-alkylation often competes with C-alkylation. N,N-
Dialkylhydrazones offer other advantages: greater nucleo-
philicity of the corresponding metalated species, selective
monoalkylation (no problems with polyalkylation), and the
possibility of using the hydrazone moiety as a chiral
auxiliary. In addition, alkylation of the aldehyde hydrazones
has a major advantage over alkylation of the corresponding
aldehydes, since aldehydes cannot be easily R-alkylated with
strong bases and alkylating agents. It is not widely known
that the problem of aldehyde deprotonation/alkylation arises
not so much from self-aldolization but from aldehyde
complexation and reduction by lithium amides, e.g., LDA.18,19
Moreover, alkylations (and likely other electrophilic reac-
tions) of azaenolates of simple cyclic ketones (e.g., substi-
tuted cyclohexanones) are diastereoselective and show
preference for the product with the new alkyl substituent in
Lazny and Nodzewska
the R-axial position. The diastereoselectivity for axial me-
thylation (LDA, MeI) was reported as high as 97:3 (trans:
cis) for 2-methylcyclohexanone N,N-dimethylhydrazone10
and quantitative for 4-tert-butylcyclohexanone DMH-hydra-
zone.11 The CdN bond of the N,N-dialkylhydrazone group
can act as a radical acceptor, typically in intramolecular
reactions, leading to cyclized hydrazine products.20,21 Besides
radical additions and consecutive fragmentation reactions,
the Eschenmoser hydrazones (N-aziridinylimines such as (2-
phenylaziridinyl)- and (trans-2,3-diphenylaziridinyl)imine)22,23
can undergo thermal and photochemical decomposition to
diazo compounds and alkenes. Owing to their special
structural features, reactivity of the N-aziridinylimines
provides entry into carbene or carbenoid chemistry and
cabanionic Shapiro-type reactions.24,25
By the virtue of vinylogy, the R,β-unsaturated aldehyde
or ketone hydrazones can react with nucleophiles in the β
position or be metalated and reacted with electrophiles in
the γ position of the hydrazone group (Figure 1). By analogy
to enamines, hydrazones (azaenamines) have the azadiene
structure enriched in electrons through electron-donating
properties of the tertiary amine nitrogen atom and as such
can react with electron-poor dienophiles in Diels-Alder and
related cycloaddition reactions (Figure 1).
Lithiated chiral N,N-dialkylhydrazones developed and
propagated by the Enders research team (corresponding
hydrazines SAMP, RAMP, SADP, SAEP, SAPP, and
RAMBO, Figure 2) undergo metalation and react in many
useful asymmetric transformations.1 Stereoselectivities of the
reactions in the case of the most often used species, i.e.,
SAMP/RAMP-hydrazones, were thoroughly studied.2,26 Cryo-
scopic27 and X-ray studies28 of lithiated SAMP-hydrazone
of methyl 2-naphthyl ketone (Scheme 1) suggest that the
lithiohydrazones form monomers in THF solution. However,
simple lithio-N,N-dimethylhydrazones show aggregation in
solution (suggested as high as tetramer for lithiated cyclo-
hexanone DMH-hydrazone).29 Fast aggregate dissociation
may precede slow alkylation reaction. In general, complex
homonuclear and heteronuclear (with a lithium amide such
as LDA) aggregation and metal coordination may be
expected for metalo-N,N-dimethylhydrazones and related
simple hydrazones.30 On the basis of the investigation of
configurations of lithiated SAMP-hydrazone species in
N,N-Dialkylhydrazones in Organic Synthesis
Figure 2. Chiral hydrazine auxiliaries used for the formation of
hydrazones for asymmetric synthesis.
solution (NMR of DMH- and SAMP-hydrazone)26,31 and in
crystals,28 it was concluded that they form four possible
rotational isomers.16,26 The preferred configuration is 7
(ECCZCN, Figure 3). The configuration of the C-C and C-N
bonds of lithiated N,N-dialkylhydrazones in the presence of
HMPA is changed to ZCCECN (10).26 The absolute configu-
ration of the products produced by cleavage of the SAMP
chiral auxiliary results from the diastereoselective (more
precisely diastereotopic face selective) attack of the electro-
philic reagent on one of the sides of the SAMP-hydrazone
azaenolate (Scheme 1). The diastereoselectivity in the
reaction with an electrophilic reagent (typically an alkyl
halide) is rationalized by the so-called metalloretentive, SE2′-
front mechanism of electrophilic substitution (Scheme 2,
electrophilic attack from the chelated metal side). As a
consequence the hydrazone products, analogous to 12, have
the usually less thermodynamically favored ZCN configuration
and isomerize spontaneously to the more stable ECN config-
uration. The reversal of configuration of C-C and C-N
bonds in Li-SAMP-hydrazone, in the presence of HMPA,
results in reversal of the configuration of the stereocenter
formed newly by alkylation reaction, albeit with low ste-
reoselectivity.26
The stereoselectivity of asymmetric reactions with chiral
hydrazones has been thoroughly studied since the late 1970s.
Generalized conclusions, based on observations of electro-
philic alkylations by Enders, and other reactions (e.g.,
nucleophilic 1,2-addition by Denmark)32 with chiral SAMP-
hydrazones, suggest that the highest diastereoselectivity is
observed in ethereal solvents such as Et2O and THF, at very
low temperatures, without chelating cosolvents or additives,
and with judiciously chosen electrophilic reagents (e.g.,
Me2SO4 is better than MeI in some cases although iodides
Scheme 1
Chemical Reviews, 2010, Vol. 110, No. 3 1389
Figure 3. Configurations of lithiated N,N-dialkylhydrazones.
are usually preferred). The methoxy group in the SAMP-
like chiral auxiliaries is essential because a change to other
more bulky groups, such as TMS, trityl, or TBDMS,33 or a
change to NEt2 or removal of O (change of OMe to Me)
leads to a medium to very high drop in diastereoselectivity.2
Substitution of the methyl group in SAMP with CH2OMe
or CH2OCH2CH2OMe retains virtually the same level of
selectivity tested on methylation of cyclohexanone.2 Interest-
ingly, slightly higher diastereoselectivity than for SAMP-
hydrazone was reported for the SAMP analogue hydrazone
with the methyl group exchanged for CH2OCH2CH2OMe in
the 1,2-addition of organoceriums.32
3. Preparation and Cleavage of
N,N-Dialkylhydrazones
Preparation of N,N-dialkylhydrazones 13 from simple
aldehydes and ketones 14 (Scheme 2) is often a spontaneous
reaction, where fast separation of water from immiscible
organics is observed when reagents are mixed neat or in
benzene, dichloromethane, or hexane solutions. Hindered,
or less reactive, aldehydes and especially ketones may require
acidic catalysts (AcOH, TFA, p-TsOH), heating, and removal
of forming water (azeotropic, by molecular sieves or other
water scavengers) to achieve high conversions or reasonable
reaction times.1,3,4 In difficult cases of hydrazone formation,
special reagents can be used such as trimethylsilyl chloride34
or trimethylaluminum (forming reactive [Me2AlNHNR2]n
with N,N-dialkylhydrazines).35 Trimethylsilyl chloride was
used for the formation of a N,N-dimethylhydrazone of a
hindered ketone with an R-quaternary carbon atom,34 while
the trimethylaluminum reagent was successful for planar
ferrocenyl ketones.36 Besides direct formation from hydra-
zines, the ketone and aldehyde hydrazones can be synthesized
from alkyl azides 17 (NH2NMe2/FeCl3 · 6H2O, MeCN/re-
flux),37 from ketimines by transamination,38 or from other
hydrazones 18 by R-alkylation and other hydrazone reactions.
The polyfunctional hydrazones, in particular, can be formed
1390 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 2
by many hydrazone reactions (vide infra) at the R-carbon
atom and by reactions of aldehyde hydrazones at the
azomethine carbon (Michael-type addition to the activated
CdC bond, electrophilic Vilsmeier-type formylation, or
acylation with reactive acylating reagents). In addition to the
rather typical methods mentioned above, some special routes
can also provide hydrazones. For example, alkynes can be
hydrohydrazinated with N,N-disubstituted hydrazines using
the titanium catalyst Ti(dap)2(NMe2)2 to give hydrazones.39
A special family of hydrazones, the N-aziridinylimines,
known as Eschenmoser hydrazones, can be prepared by direct
reaction of 1-aminoaziridines (1-amino-2-phenylaziridine and
1-amino-trans-2,3-diphenylaziridine) with reactive carbonyl
compounds at temperatures below 40 °C or with imines
(substitutes for less reactive carbonyl compounds).24,40
Cleavage of the N,N-dialkylhydrazones (Scheme 2) can
be effected by a multitude of methods.41 Methods for
recovery of carbonyl compounds from N,N-dimethyl- and
SAMP/RAMP-hydrazones were thoroughly reviewed in the
Enders’ account.8 General uses in preparative practice have
a few methods, including oxidative, hydrolytic, and reductive
protocols. In the typical laboratory practice, carbonyl com-
pounds 14 are obtained from their N,N-dialkylhydrazones
13 by oxidative cleavage with magnesium monoperoxyph-
thalate hexahydrate (MMPP · 6H2O; ketones only),42,43 m-
chloroperbenzoic acid (ketones only),44a recently reported
peroxyselenous acid (ketones only),44b and ozonolysis (O3,
-78 °C).2,45 In contrast to ketone N,N-dialkylhydrazones,
the aldehyde hydrazones give nitriles 16 when oxidized with
MMPP or m-CPBA.46 A one-pot ozonolytic cleavage com-
bined with reductive workup is a practical method that
provides alcohols 19 and may be desired when racemization-
susceptible R-chiral carbonyl compounds47 (e.g., aldehydes,
R-phosphino ketones, etc.) produced by ozonolysis cannot
be isolated without loss of enantiomeric purity. Owing to
its mild conditions and low temperature, ozonolysis is the
oxidative protocol most broadly used for regeneration of
carbonyl compounds from the corresponding hydrazones.
One should however be conscious of a toxic N-nitrosoamine
byproduct forming during ozonolytic cleavage of hydrazones
and take proper care during the reaction workup.
The hydrolytic cleavage methods are broadly used for
polyfunctional hydrazones possessing other functionalities
that react with oxidants. Typical hydrolytic cleavage is often
conducted through methylation and acidic hydrolysis of
SAMP/RAMP-hydrazones in a two-phase system (MeI, 3
M HCl, n-pentane),48,49 which prevents racemization of
Lazny and Nodzewska
R-chiral carbonyl products 14. The use of an aqueous oxalic
acid solution in a two-phase system50 is currently quite a
popular method for the cleavage of sensitive products. This
also allows for the recovery of the precious SAMP chiral
auxiliary. Copper(II)-induced hydrolytic cleavage (aqueous
cupric acetate or dichloride) may also provide good results
in most cases51,52 including R,β-unsaturated aldehyde N,N-
dimethylhydrazones.53 Hydrolysis of N,N-dimethylhydra-
zones, and RAMP-hydrazones containing an olefin or acetal
group, to ketones was reported using ammonium dihydrogen
phosphate buffer solutions.54 Interestingly, enzymatic hy-
drolytic cleavage to the carbonyl group is possible with
porcine pancreatic lipase (PPL).55 An attractive, racemization-
free cleavage using BF3 · Et2O, paraformaldehyde, and acetone/
water was recently reported,56 as was a study on the
hydrolytic stability of N,N-dimethylhydrazones and related
compounds.57
The most commonly used reductive cleavage methods
involve N-N bond cleavage and as a result give primary
amines 15. The hydrazone may be reduced (e.g., with
catecholborane,58 DIBAL,59 or DMAB/p-TsOH60) to hydra-
zine, cleavage of which is effected by Raney nickel58 or
borane/THF.61 The whole hydrazone reduction-cleavage is
also possible in one step with borane/THF itself.62
4. Short History of Synthetic Applications
Although simple N,N-dialkylhydrazones such as dimethyl-
and diethylhydrazones of benzaldehyde, p-nitrobenzalde-
hyde,63 aromatic aldehydes, and cyclohexanone64 were known
in the early times of modern organic chemistry, the record
of synthetic applications of N,N-dialkylhydrazones starts in
the 1960s, when piperazine-derived hydrazones of type 20
(Figure 4) and dimethylhydrazones of aldehydes were
prepared and used in sodium borohydride reductions to
hydrazines.65 In the same decade, oxidation of hydrazones
of aliphatic and aromatic aldehydes to nitriles with hydrogen
peroxide was reported.66 Oxidation of hydrazones to parent
carbonyl compounds was also observed by Horner.67 N,N-
Figure 4. The earliest N,N-dialkylhydrazones.
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 3
Dialkylhydrazones were described in 1969 as possible
protecting groups for ketones, furnishing under hydrolytic
or oxidative cleavage the parent carbonyl compounds.68 The
1,2-addition of phenyl- and butyllithiums to CdN bonds of
N,N-dimethylhydrazones of 4-pyridinealdehyde and 4-acetylpy-
ridine was described in 1964 (the reaction with cyclohex-
anone or nicotinaldehyde was not successful).69 Nucleophilic,
azaenamine character of N-aminopyrrolidine-derived ben-
zaldehyde hydrazones 21 was recognized and demonstrated
through Vilsmeier formylation70 and reactions with electro-
philic sulfonyl isocyanates71 as early as 1968. N,N-Dialky-
lhydrazones of R,β-unsaturated aldehydes and ketones can
be viewed as aza analogues of electron-rich dienes (Figure
1). The synthetic applications of such hydrazones in
hetero-Diels-Alder72-74 and [2 + 2] cycloaddition75 reactions
have been known at least since the 1980s.
Despite the early reports, N,N-dialkylhydrazones did not
gain widespread application in synthesis until the landmark
papers on alkylation of metalated N,N-dimethylhydrazones
(described also as metal azaenolates) were published by
Corey and Enders in the years 1976-1978.10,11,51,76-80 It was
shown that hydrazones, acting like aza analogues of the
parent carbonyl compounds, could be R-deprotonated (lithi-
ated with LDA or n-BuLi) and reacted with a number of
electrophiles (in particular alkyl halides). The pioneering
studies have shown that N,N-dialkylhydrazones (mostly N,N-
dimethylhydrazones) could be important intermediates in the
synthesis of R-substituted aldehydes and ketones, and also
in other C-C bond forming reactions. Shortly thereafter,
Enders developed a very successful approach to the synthesis
of enantiomerically pure compounds based on the reactions
of chiral N,N-dialkylhydrazones made from carbonyl com-
pounds and prolinol-derived chiral hydrazines known as
SAMP and its enantiomer RAMP (Figure 2).1,2,48,49,81-84
Other chiral hydrazines shown in Figure 2 were developed
later for specific purposes.
The free radical acceptor character of the hydrazone
functional group was demonstrated in a cyclization of
N-aziridinylimines by Kim in 1991.85
High nucleophilic reactivity of formaldehyde N,N-dim-
ethylhydrazones (FDMHs) 22 toward nitroolefins 23 was
exploited in a Michael addition for the first time by
Lassaletta86,87 and Enders88 (Scheme 3) in both achiral and
chiral versions based on formaldehyde SAMP-hydrazone
(FSAMPH).
5. Formation of Azaenolates
Metalated hydrazones can be prepared via several methods.
The most commonly used metalated hydrazones 25 (Scheme
5) are lithium and potassium azaenolates (also known as
azaallyllithiums and -potassiums)4,89 which are made via
deprotonation of hydrazones with LDA (sometimes with
Chemical Reviews, 2010, Vol. 110, No. 3 1391
Scheme 4
additives such as HMPA, LiBr, and TMEDA), n-BuLi
(sometimes with additives such as HMPA, TMEDA, and
DMPU), t-BuLi, LTMP, t-BuOK, NDA,2,90 KDA (potassium
diisopropylamide made from diisopropylamine, t-BuOK, and
n-BuLi),91 and the Lochmann-Schlosser superbase, i.e.,
t-BuOK/n-BuLi.92
Transition-metal azaenolates such as zinc and copper are
typically made by transmetalation of the lithium enolate. Zinc
azaenolates, made through a reaction of a hydrazone with
LDA, ZnCl2, and 1 equiv of BuLi, were successful reagents
in the addition to unactivated olefins.93 Lithium-copper
mixed azaenolates, which are presumably formed through
Li-Cu transmetalation reaction of lithiated hydrazones (n-
BuLi), with copper(I) iodide in i-Pr2S solution, undergo
conjugate addition to R,β-unsaturated ketones or esters
(adducts 26).77 Titanium azaenolates (titanated hydrazone)
made by the action of TiCl4 and NEt-i-Pr2 on a SAMP-
hydrazone have been used in syn-selective aldol reaction
(products 27).94
An interesting case constitutes the formation of a lithium
azaenolate of a β-stannane (29) through 1,4-addition of
R3SnLi (R3 ) Me, n-Bu) or Li2[CuMeCN(SnBu3)] to the
hydrazone of an R,β-unsaturated ketone (28) (Scheme 4),
which after typical R-alkylation provides access to β-stan-
nylketones.95
6. Reactions of Azaenolates with Electrophiles:
Carbon-Carbon Bond Formation
The typical electrophilic reactions on the R-carbon of a
hydrazone require prior deprotonation (metalation) of the
hydrazone C-H acid. As a result of the above-discussed
reactivity of hydrazones and their metalated derivatives, several
synthetically useful transformations based on electrophilic
reactions are known, many of them belonging already to the
classical organic synthesis literature (Scheme 5).
Once the azaenolates are formed they can react with a
number of electrophiles. The known synthetically useful
C-C bond forming reactions (Scheme 5) of R-metalated
hydrazones include Michael-type additions (26), aldol-type
reactions (27), R-alkylation (31), and Claisen-type acylations
(32).4
6.1. Typical Electrophilic Reactions of
Azaenolates (Prior to 2000)
r-Alkylation. The reaction having the most synthetic uses
is R-alkylation (Scheme 5). The reaction was used, for
1392 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 5
example, in the synthesis of R-branched aldehydes including
intermediates in the epothilone synthesis,96 R-branched
acyclic and cyclic ketones,2 dioxanones,97 lactones,98 and
β-keto esters.2 The alkylating agents used are alkyl, allyl,
and benzyl halides (Br, I), bromo or iodo ethers (halide-
substituted ethers) especially BOM-Cl, bromo or iodo esters
(halide-substituted esters), protected bromo or iodo alcohols,
oxiranes, tosylaziridines, methyl sulfate, and even trialkylsilyl
triflate-activated THF.99 The asymmetric syntheses of branched
ketones and aldehydes through alkylation of SAMP/RAMP-
hydrazones,48,49,81,100 including many natural products, pub-
lished up to the year 2000 have been reviewed by Enders.1,2
Aldol-Type, Michael-Type, and Other Addition Reac-
tions.4 Typical 1,2-addition of azaenolates to the carbonyl
group of ketones and aldehydes provides access to azaaldol-
type products (Scheme 5) which can further be hydrolyzed
to carbonyl derivatives.76,77 The hydroxyls of the azaaldols
are often silylated in “one pot” to protect the aldols from
hydrazone cleavage conditions. Related addition to carbon
disulfide gives lithium hydrazonoalkanedithiolates that can
be further S-alkylated.101 1,4-Addition of hydrazone anions
to R,β-unsaturated carbonyl compounds, esters, nitriles,
alkenyl sulfones, and alkenyl phosphonates results in various
Michael-type adducts (Scheme 5).77 Claisen-type substitution
on sp2 carbon atoms of acid chlorides by azaenolates
followed by hydrazone hydrolysis provides an access to
1,3-diketones.80,102
Various applications of hydrazone azaenolates (1-azaallylic
anions) toward the synthesis of heterocyclic structures have
previously been reviewed.103
6.2. Alkylations
Electrophilic alkylations were among the first synthetically
useful transformations of hydrazones and remain still the
most frequently used reactions of hydrazones. The alkylations
of DMH-hydrazones and SAMP/RAMP-hydrazones are often
employed as key steps in the synthesis of complex targets
which are typically natural products.
Lazny and Nodzewska
6.2.1. Alkylation of N,N-Dimethylhydrazones
Alkylation of DMH-hydrazones was used in recent years
as the major C-C bond forming reaction in the synthesis of
spiroketals (Scheme 6).
Spiroketal 35 was synthesized starting with alkylation of
cyclopentanone N,N-dimethylhydrazone 33 (R1, R2 ) (CH2)3)
with the chiral iodide 34 (either the (R)- or (S)-enantiomer),
derived from enantiomerically pure ethyl β-hydroxybutyrate.
The alkylation served as a method for introducing the
stereocenter at C2 of the target molecule 35. The alkylation
products were transformed into a diastereomeric mixture
(2:1) of compound (2R,5S)-35 or compound (2S,5S)-35 in
five synthetic steps including stereoselective spiroketalization
in acidic media.104
A very efficient alkylation of a lithiated N,N-dimethylhydra-
zone of a chiral ketone (33, R2 ) Me, R1 ) CH2CH(Me)-
CH(O(TBS))CH2O(PMB)) followed by spiroketal formation
under acidic conditions was also used in a short synthesis of
the 6,6-spiroketal fragments 38 and 39 useful for synthesis of
the antifungal antibiotics spirofungins A and B.105
Another alkylation reaction of the lithiated N,N-dimeth-
ylhydrazones of elaborate methyl ketones 33 (R1 ) H, R2
) chiral O-protected substituted trihydroxynonyl) with
suitable, protected iodide 40 afforded linear spiroketal
intermediates. After functional group adjustment, these
advanced intermediates were cyclized to their respective
spiroketals 41 (R ) H, R ) Me), which constituted subunits
in the asymmetric synthesis of the macrolide antibiotics (+)-
rutamycin B and (+)-oligomycin.106
Several chiral spiroketal skeletons were obtained by
sequential R- and R′-alkylations of acetone N,N-dimeth-
ylhydrazone with acetonide-protected dihydroxy iodides
42. A one-pot acidic deprotection/spirocyclization se-
quence (Amberlyst 15, MeOH) provided spiroketal 44.107
The same convergent approach via a double alkylation of
acetone N,N-dimethylhydrazone was also used for synthesis
of a spiroheterocycle, 1-oxa-7-azaspiro[5.5]undecane (45).108
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 6
Scheme 7
6.2.2. Alkylation of Zincated N,N-Dialkylhydrazones
Zincated N,N-dimethylhydrazones of ketones, prepared by
transmetalation of lithium azaenolates with zinc chloride,
underwent stereospecific syn-addition to both (E)- and (Z)-
alkenylboronates to give zinc intermediates, which reacted
with subsequently added electrophiles (Scheme 7) to give
γ-borylhydrazones 47 in yields from 48% to 93% and with
good to excellent diastereoselectivities (up to 99%).109 It is
noteworthy that the method (three-component coupling)
allowed the generation of two to four contiguous stereogenic
centers in a one-pot reaction. Recently, an asymmetric
version of the process with zincated SAMP-hydrazones,
along with thorough experimental and theoretical study, was
reported.110 The SAMP-hydrazone adducts corresponding to
47 were obtained with a diastereoselectivity of 95% (after
recrystallization the reported ds was >99% and the ee was
>99%).
6.2.3. Alkylations of Chiral N,N-Dialkylhydrazones
The well-known alkylation of chiral hydrazones, notably
SAMP/RAMP-hydrazones (Scheme 8), is often used as a
key step in the asymmetric synthesis of natural products.
The alkylation of the diethyl ketone SAMP-hydrazone (S)-
48 with p-MePhO(CH2)3I (yield 80%, de > 98%, Scheme
Chemical Reviews, 2010, Vol. 110, No. 3 1393
Scheme 8
8), followed by titanium-mediated syn-diastereoselective
aldol reaction and other transformations, allowed for the first
highly diastereo- and enantioselective total synthesis of
stigmatellin A (53; Figure 5), a powerful inhibitor of electron
transport in mitochondria and chloroplasts isolated from the
myxobacterium Stigmatella aurantiaca.94
Similar asymmetric R-alkylation of an O-protected deriva-
tive of 4-hydroxybutanal with methyl iodide via the SAMP/
RAMP-hydrazone method (de 72% to >95%) was exploited
for the preparation of intermediates in the synthesis of the
potent cytotoxic marine natural product (-)-callystatin A (54;
Figure 5)111,112 and its 20-epi analogue (overall yield 18%
and 17%, respectively). The synthesis also required a
diastereoselective syn-aldol reaction of an R-substituted chiral
ethyl ketone and an R-substituted chiral aldehyde, both
prepared in enantiomerically pure form, again by means of
the asymmetric alkylation of corresponding RAMP-hydra-
1394 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 5. Products synthesized via hydrazone alkylation as key steps (bonds formed are shown in red).
zones (reactions of diethyl ketone with BOM-Cl and bu-
tyraldehyde with MeI, respectively).112
Both enantiomers of (Z)-1,1,1-trifluoro-3-methyl-4-thia-
13-hexadecen-2-one (55a; Figure 5) and (Z)-1,1,1-trifluoro-
3-methyl-4-thia-13-octadecen-2-one (55b; Figure 5), potential
inhibitors of the pheromone action of two major maize pests,
Sesamia nonagrioides and Ostrinia nubilalis, were synthe-
sized via alkylation of SAMP/RAMP-hydrazones with meth-
yl iodide as the key step (79-86% yield, de 92-94%).
Specific hydrazone cleavage under racemization-free condi-
tions (BF3 · Et2O, paraformaldehyde, acetone/water) gave the
target ketones 55 (ee g 90%).56
The synthesis of C4-alkylated pyrrolidin-3-ones of type
56 (Figure 5) by regio- and diastereoselective (de 38% to
>95%) alkylation of SAMP-, SAPP-, and SADP-hydrazones
of N-protected pyrrolidine-3-ones was also reported (LDA
or LHDMS, THF, MeI or BnBr, -78 or -100 °C).113 The
pyrrolidin-3-one 56 was prepared by the alkylation of SADP-
hydrazone (63% yield, de > 95%) and subsequent cleavage
with aqueous CuCl2 (67% yield, ee > 95%).
An interesting approach to asymmetric synthesis of tosyl-
protected γ-amino ketones 57 (Figure 5) and γ-amino nitriles
58 through reactions of lithiated SAMP- and RAMP-
hydrazones with tosylaziridine is worth noting. The key step
was the highly diastereoselective (de g 98%) R-aminoet-
hylation of the hydrazones. Cleavage of the hydrazones with
magnesium monoperoxyphthalate (MMPP) provided γ-ami-
no nitriles 58 in good yields and excellent enantiomeric
excesses (ee g 98%). Likewise, γ-amino ketones 57 were
obtained in good overall yields by cleavage of the hydrazones
with aqueous copper(II) chloride (ee g 98%).114
The first asymmetric synthesis of both enantiomers of
Tropional 59 (Figure 5) was also realized by alkylation
Lazny and Nodzewska
(LDA, THF, -100 °C) of the SAMP- and RAMP-hydra-
zones of propanal with 5-(bromomethyl)-1,3-benzodioxole
(50; Scheme 8). The resulting alkylated hydrazones had to
be oxidatively cleaved with MMPP to nitriles, and via
subsequent reduction with diisobutylaluminum hydride
(DIBAL-H), transformed to the desired enantiomers of
Tropional (overall yields 52-53%, ee 90%).115
The asymmetric synthesis of all stereoisomers of 7,11-
dimethylheptadecane (60a) and 7-methylheptadecane (60b;
Figure 5) was possible owing to highly diastereoselective
SAMP/RAMP-hydrazone alkylation. A mixture of (7S,11R)-
60a and its demethylated analogue (S)-60b is the female sex
pheromone of the spring hemlock looper moth (Lambdina
athasaria) and the pitch pine looper moth (Lambdina
pellucidaria); both are forest pests of northeastern North
America.116 The two stereogenic centers of the first compo-
nent of the pheromone mixture were assembled by R-alky-
lation of propanal SAMP-hydrazone with n-hexyl iodide
(lithiation with LiTMP and THF at 0 °C, alkylation at -100
°C) in an effective and highly diastereoselective way (de g
96%). Under standard conditions at -78 °C lower selectivity
was observed (de 90%). The other component, (S)-60b, was
put together employing alkylation of n-octanal SAMP-
hydrazone with decyl iodide under somewhat unusual
conditions (LiTMP and THF at 0 °C, alkylation at -100 °C
followed by warming to rt and refluxing), necessary to
increase the yield to 68% (de g 96%).
The RAMP-hydrazone auxiliary was essential for the
asymmetric synthesis (67-87% ee) of (R)-2-benzyl-, (S)-2-
octyl-, and (S)-2-tetradec-5′-enylcyclobutanones 61 by alky-
lation of cyclobutanone respectively with benzyl bromide,
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 9
n-octyl bromide, or tetradec-5-enyl bromide. The hydrazone
hydrolysis to the ketone products was effected by oxalic acid
cleavage.117
The SAMP-hydrazone alkylation of 2,2,6,6-tetramethyltet-
rahydropyran-4-one with functionalized benzylic bromide 52
(72% yield, single diastereomer) was again a key step in a
synthesis of the heptacyclic core of (-)-nodulisporic acid D
(62), a member of a class of architecturally complex,
ectoparasiticidal indole alkaloids.118
The R-alkylation of aldehyde SAMP-hydrazones 48
(Scheme 8, R1 ) H, R2 ) Me, Et, i-Pr, t-Bu, Bn, 4-ClPh)
with alkyl (methyl or tert-butyl) bromoacetates, combined
with MMPP oxidation to nitriles and a reductive cycliza-
tion with Raney Ni or nickel boride, provided the pyrrolidin-
2-ones 63. They were prepared in good overall yields
(27-78%) and ee (93-99%). The method was applicable
for an asymmetric synthesis of GABAs (γ-aminobutyric
acids) as was demonstrated for (R)-(-)-baclofen hydrochlo-
ride (64; four steps, 55% yield and 94% ee).119
The same SAMP/RAMP-hydrazone alkylation method also
allowed for the first asymmetric synthesis of potential HIV
protease inhibitors of types II, III, and IV starting from a
derivative of pyrimidine-2,5-dione which is a readily avail-
able building block.120,121 The R- or R,R′-alkylation of 1,3-
dibenzyltetrahydropyrimidine-2,5-dione hydrazones (meta-
lation, LiTMP; alkylation, i-PrI, n-BuI, BnBr, PhCH2CH2I,
allyl halide) followed by cleavage (ozone, dimethyldioxirane,
or aqueous CuCl2,) gave carbonyl compounds 65. Monoalky-
lated tetrahydropyrimidine-2,5-diones were obtained with
excellent de’s (>96%) and good ee’s (76% to >96%), while
the bisalkylated products were prepared with good de’s (80%
to >96%) and ee’s (76% to >96%). The potential HIV
protease inhibitors (5-hydroxy derivative isomers of 65) were
prepared by stereoselective reduction of 65 (LAH, NaBH4,
L-Selectride, Super-Hydride, catecholborane, or BH3/dimeth-
yl sulfide) in good yields (71-92%) and with good diastereo-
and enantiomeric purities.
The SAMP/RAMP-hydrazone alkylation method was
investigated as one of the approaches to the unusual [5.3.2]-
bicyclic structure of the insecticidal Amaryllidaceae alkaloids
cripowellin A and B.122
Alkylation of diethyl ketone via the SAMP-hydrazone
method also allowed the construction of three stereogenic
centers of serricornin, a sex pheromone of the cigarette beetle
(Lasioderma serricorne), and provided material for the
determination of its absolute configuration by CD spectro-
scopy.123,124
Three prolinol-derived hydrazones, analogues of SAMP-
hydrazone with alternative O-substituents, 66 (R1 ) TBDMS,
TMS, and trityl, Scheme 9) were produced, and their
effectiveness in directing alkylation of diethyl ketone hy-
drazone azaenolate with benzyl bromide was studied.33 The
Chemical Reviews, 2010, Vol. 110, No. 3 1395
Scheme 10
substituent on the oxygen of the prolinol had a significant
effect on both the stereoselectivity and the configuration of
the asymmetric alkylation. The tert-butyldimethylsilyl hy-
drazones, lithiated with LDA in THF, gave the expected (S)-
enantiomer of the benzylation product in good yield and
moderate enantiomeric purity (61% yield, 61% ee). The sense
of diastereoselection was in agreement with the known
SAMP-hydrazone (R1 ) Me) performance where the (S)-
enantiomer is also preferred. Changing the substituent on
the oxygen atom to the trityl group caused a reversal of
configuration of the major alkylation products and rather low
stereoselectivity (7-64% depending on the solvent, (R)-
enantiomer). The change was rationalized by the large -CPh3
group hindering chelation of the oxygen to the lithium, thus
disrupting the configuration of the transition state. Solvents
also had a substantial effect on the reaction. Toluene gave
the best yields and ee’s, while both pentane and hexane were
inferior. Additives (HMPA and TMEDA) caused a lowering
of yields and dramatic decrease of the ee, contrary to lithium
chloride, which increased the yield to 90%. The best
performance was observed for the hydrazone with the least
sterically demanding TMS substituent (R1 ) TMS) in toluene
(65% yield, ee 86%, (S)-enantiomer). The effect of the
oxygen substituents on the described asymmetric benzylation
of SAMP-type hydrazones matches observations previously
reported by Enders.2 An interesting observation was made
while testing chiral lithium amides (S,S)-67 and (R,R)-67 as
deprotonating/lithiating bases. They have the potential to
form a matched and mismatched pair of diastereomeric
transition states; however, both enantiomers gave a dimin-
ished enantiomeric purity of 68 (R1 ) TMS, reaction in
hexane 55% and 74% ee compared to 86% ee with LDA).
The authors suggest that the reaction diastereoselectivity was
governed entirely by the chiral auxiliary but do not comment
on possible causes of observed ee reduction.
An interesting approach to the asymmetric synthesis of
planarly chiral 2-mono- and 2,2′-disubstituted 1,1′-bisben-
zoylferrocenes is based on the highly diastereoselective
ortho-metalation of 1,1′-bisbenzoylferrocene via the corre-
sponding bis(SAMP-hydrazone) 69 (Scheme 10, yield
85-100%, de g 96%). This illustrates yet another extension
of the SAMP-hydrazone methodology to new reactions. The
lithiated ferrocenes were trapped with several carbon, silicon,
phosphorus, and sulfur electrophiles. Cleavage of the mono-
substituted hydrazones led to monosubstituted ketones (ee
g 98%). Further ortho-substitution of the monohydrazone
1396 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 11
Scheme 12
afforded 2,2′-disubstituted hydrazones 70 (yield 45-89%,
de 92% to g96%), cleavage of which gave disubstituted
ferrocenyl diketones (ee g 99%). The cleavage of the
auxiliary was effected by ozonolysis or reduction using TiCl3
or SnCl2 (yield 58-99%, de 98% to g99%).125
The same diastereoselective ortho-metalation of diferro-
cenyl ketone SAMP-hydrazone 71 was applied to the
preparation of planar chiral 2-monosubstituted diferrocenyl
ketones (yields 88-99%, de g 96%, Scheme 11). The
racemization-free cleavage of the resulting hydrazones 72
afforded the corresponding ketones in good to very good
overall yields (21-97%, depending on the cleavage method)
and high enantiomeric excesses (97-99%).126 The mono-
substituted diferrocenyl ketones 72 (E ) SMe, PPh2/BH3,
C(OH)Ph2) have been subjected to ortho-metalation and
reactions with electrophiles MeI, (i-PrS)2, (MeS)2, Me3SiCl.
The reactions gave chiral disubstituted bisferrocenes in low
to moderate yields (20-54%) and excellent stereoselectivities
(97% to >99% ee, >96% de), however with poor regiose-
lectivities in several cases.127
Phosphorus-controlled regioselective R-alkylation of R-phos-
phorylated hydrazones derived from phosphine oxides or
phosphonates allowed for synthesis of 1-aminopyrrol-2-ones,
1-amino-3,4-dihydropyridin-2-ones 75 (n ) 1, 2) or 1-ami-
nopyrroles containing phosphinyl or phosphoryl substituents,
76 (Scheme 12). The azaenolates of 73, of DMH-, SAMP-,
and RAMP-hydrazones (obtained with LDA and THF at -78
°C or n-Bu2CuLi and THF/Et2O at -50 °C), were alkylated
with alkyl halides (yield 42-89%). Diastereoselectivity of
SAMP/RAMP-hydrazone alkylation was low (de 0-50%).128
Alkylation product 74 (R4 ) (CH2)nCOOEt) underwent
cyclocondensation on addition of LDA to form 75. A
Lazny and Nodzewska
palladium-catalyzed cyclization to form 76 was effective for
the propargylic product 74 (R4 ) propargyl).
6.2.4. Alkylation of Dioxanone Hydrazones
The N,N-dialkylhydrazone alkylation reaction is certainly
the hydrazone transformation, most widely used in the
synthesis of complex natural products. 1,3-Dioxanones
(cyclic acetals of 1,3-dihydroxyacetone) are a group of
synthetically versatile building blocks. N,N-Dialkylhydra-
zones of these synthetic equivalents of dihydroxyacetone
(DHA) can be used by synthetic chemists in the same way
that dihydroxyacetone phosphate (DHAP) is used by nature,
i.e., as donors in important C-C bond forming reactions.
The power of dioxanones as C3 donor reagents has been
demonstrated convincingly in many asymmetric electrophilic
substitution reactions (including stereoselective alkylations
and aldol reactions) in target-oriented syntheses. Synthetic
applications of 1,3-dioxanones have been summarized in
excellent reviews.97,129
Currently reliable procedures for the preparation of chiral
and achiral 2,2-dimethyl-1,3-dioxan-5-one hydrazones,130
their R- and R,R′-alkylation130 and mild, selective hydrazone
cleavage97,130 to afford substituted dioxanones, are available.
Dioxanone hydrolysis to 1,3-diols has been effected with
PPTS/acetone/water,131 Dowex 50/ethanol,132 TFA/water/
THF,133 and aqueous HCl.134,135 In recent years, a number
of synthetic applications of dioxanone alkylation that deserve
particular mention have been reported.
r-Alkylation of Dioxanone Hydrazones. The R-alkyla-
tion of 2,2-dimethyl-1,3-dioxan-5-one SAMP- or RAMP-
hydrazones, followed by trifluoromethylation of the resulting
enantiomerically pure, alkylated dioxanones, was used for
the asymmetric synthesis of the 2-trifluoromethylated 1,2,3-
triols 79 (Scheme 13). The nucleophilic trifluoromethylation
of the monoalkylated dioxanones with (trifluoromethyl)tri-
methylsilane in the presence of TBAF gave the 2-trifluo-
romethylated acetonide-protected triols 78 in high diastereo-
and enantiomeric excesses (de 96%, ee 92-98%) and good
overall yields (52-97%, depending on R1 ) Me, Et, i-Pr,
n-Bu, and n-Hex). Deprotection of the acetonide-protected
triol under acidic conditions afforded trifluorotriol 79 and
ent-79 (de 96%, ee 95%, 96%, R1 ) n-Hex) chosen as typical
examples. Extension of the methodology to the trialkylated
dioxanone led to the analogous R-trifluoromethylated alcohol
with tertiary and quaternary R-stereocenters in good yield
(77%) and very good diastereo- and enantiomeric excesses
(de 96%, ee 98%).132
An analogous approach was used for the asymmetric
synthesis of (S,S)- and (R,R)-2-methylthreitol (82 and ent-
82; Scheme 13). Starting from the alkylation of the SAMP-
or RAMP-hydrazone of 2,2-dimethyl-1,3-dioxan-5-one with
BOM-Cl, as a key step, a benzyl-protected R-hydroxymeth-
yldioxanone (81, R1 ) BnOCH2) was prepared. The second
stereogenic center was assembled by nucleophilic 1,2-
addition of methyllithium or alternatively by the diastereo-
selective, bis(acetylacetonato)oxovanadium(IV) [VO(acac)2]
catalyzed epoxidation of the exocyclic methylene group
(made from dioxanone carbonyl by Wittig reaction). After
acidic methanolysis of the acetonide protection (Dowex
50X2-200) the enantiomeric methylthreitols 82 were obtained
in excellent diastereo- and enantiomeric excesses (g98% de,
98% ee) and in good overall yields (40-61% depending on
the method used).134
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 13
The SAMP/RAMP-hydrazone methodology was also
employed in the enantioselective synthesis of mono-TBS-
protected, allylic diols 84 (R2 ) TBS, ee 90-94%).
Alkylation (R1 ) Bn, i-Pr, n-Bu, (CH2)2Ph, 4-t-BuC6H4CH2)
of dioxanone SAMP-hydrazone followed by hydrazone
ozonolysis gave the acetonide-protected, R-substituted ke-
todiols (90-94% ee), which in turn were converted to
exocyclic olefins 83 by Wittig reaction without racemization.
Acidic acetal cleavage with TFA/water (room temperature)
to give 84 (R2 ) H) and subsequent, selective, TBS
protection of the primary hydroxyl group, furnished 84 (R2
) TBS) in very good overall yields (54-99%) and enan-
tiomeric excesses.133
The asymmetric synthesis of protected 2-amino 1,3-diols
(S,R)-85 starting from 2,2-dimethyl-1,3-dioxan-5-one via
SAMP/RAMP-hydrazone methodology could be extended
to the synthesis of 86. The two adjacent stereogenic centers
were built up by R-alkylation using the SAMP-hydrazone
method followed by oxalic acid hydrazone cleavage and
diastereoselective reduction of the resulting ketones with
L-Selectride. The resulting (S,S)-alcohol products were
further transformed into the amines (S,R)-85 by nucleophilic
substitution with an azide and subsequent reduction of the
azide with lithium aluminum hydride to the amine. The
amine products were obtained in high diastereomeric and
enantiomeric excesses (de g 96%, ee 90-94%). By employ-
Chemical Reviews, 2010, Vol. 110, No. 3 1397
ing this approach and starting from RAMP-hydrazone, the
ammonium salt of D-erythro-sphinganine, (R,S)-95 (Scheme
14), was synthesized in 47% overall yield and with diaster-
eomeric and enantiomeric excesses of g96%.136
The 11-step asymmetric synthesis of (+)-2-epi-deoxo-
prosopinine [(S,S,R)-89] (Scheme 13) based on the R-alky-
lation of 2,2-dimethyl-1,3-dioxan-5-one SAMP-hydrazone
with a Cbz-protected alkylation reagent is also noteworthy.137
The alkylation reagent 92 was also obtained via the SAMP-
hydrazone methodology (Scheme 13). First, the 1,2-addition
of a dodecyl nucleophile (YbCl3 · H2O/C12H25Li) to TBS-
protected 3-hydroxypropanal SAMP-hydrazone (90) gave a
hydrazine, 91, which in turn was cleaved with borane/THF
to the amine. Further protection-deprotection and functional
group manipulation gave the protected 3-amino-1-pentadecyl
iodide (92). Oxalic acid cleavage of the alkylated hydrazone
80, followed by deprotection of the amine 87 (hydrogenolysis
on Pd/C) combined in one step with cyclization to the imine,
and imine hydrogenation, provided a mixture of acetonide-
protected epimers S,S,R/S,S,S-88 (98:2). Chromatographic
purification and hydrolytic cleavage (acidic ion-exchange
resin Lewatit S 100) of the acetonide protecting group
afforded (S,S,R)-89 in excellent diastereomeric and enantio-
meric purity (de, ee g 96%). It is noteworthy that the SAMP-
hydrazone methodology was employed in two key steps of
the synthesis, generating two of the three stereogenic centers.
1398 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 14
Scheme 15
The third stereogenic center was built in a domino depro-
tection/cyclization/reduction sequence.137
The R-alkylation of the RAMP-hydrazone of 2,2-dimethyl-
1,3-dioxan-5-one, acting as a 1,3-dihydroxyacetone equiva-
lent, was also the key step in the asymmetric total synthesis
of (+)-altholactone (97; Scheme 14).138 The hydrazone
alkylation with 3-(benzyloxy)-1-bromopropane and cleavage
with oxalic acid gave product 96 in 82% yield (two steps)
and with ee > 98%. The other reactions used in the synthesis
included a boron-mediated aldol reaction, a six- to five-
membered ring acetonide transacetalization, an oxidation of
1,5-diol to δ-lactone, and an Amberlist 15-catalyzed ac-
etonide removal, concomitant with stereoselective ring
closure to generate the annulated tetrahydrofuran structure
of (+)-altholactone. Overall, the natural product target 97
was synthesized enantioselectively in 18 steps and 13.7%
overall yield.
An elegant diastereo- and enantioselective approach to 4′-
quaternary 2′-deoxy-3′-epi-β-C-nucleosides 99 (Scheme 14)
through construction of the first stereocenter by the RAMP-
hydrazone method is another example of a total synthesis
built upon hydrazone alkylation. After alkylation of diox-
anone RAMP-hydrazone with tert-butyl bromoacetate, and
ozonolysis of the hydrazone, the resulting dioxanones were
subjected to diastereoselective nucleophilic 1,2-additions of
Grignard reagents (addition to the ketone group of the
dioxanone). The adducts 98 were further transformed via
Lazny and Nodzewska
deprotection, cyclization, organocerium addition, and reduc-
tion/cyclization to nucleosides 99 (R2 ) Naph, Ph).139
Recently, the approach has been extended to the preparation
of both enantiomers of 4′-quaternary 2′-deoxy-3′- and 4′-
epi-β-C- and -N-nucleosides 101.140 The extended syntheses
also involved the SAMP/RAMP-hydrazone R-alkylation and
the same transformations to get the intermediate 98.140
Further transformations of the lactone 100, including ma-
nipulation of the substituents in the anomeric position,
allowed access to the thermodynamically more stable β-ano-
mers of nucleosides (N-nucleoside 101) with both diastereo-
and enantiomeric purity of >99%.
r,r′-Alkylation of Dioxanone Hydrazones. The asym-
metric R,R′-bisalkylation of dioxanone SAMP-hydrazones
is another valuable synthetic approach. The method was used
as a key asymmetric transformation in the diastereo- and
enantioselective synthesis of pseudo-C2-symmetric, 2-methyl-
substituted, acetonide-protected diols 104 (Scheme 15).
Hydrazone 102 was made by iterative diastereoselective R,R′-
bisalkylation. Ozonolytic cleavage, followed by an epimer-
ization-free Wittig olefination of the resulting ketone, and
subsequent hydrogenation of the exocyclic methylene group
in 103 using either the Adams (PtO2 · H2O) or Wilkinson
catalyst afforded the acetonide-protected 1,3-diols 104 in very
good overall yields and with virtually complete stereoiso-
meric purity (de g 96%, ee g 98-99%). Quantitative
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 16
removal of the acetonide protection with trifluoroacetic acid
in THF/water led to the free pseudo-C2-symmetric diols
105.141
A diastereoselective SAMP-hydrazone R,R′-bisalkyla-
tion/deoxygenation protocol was effective for making the
anti-1,3-diol moiety (intermediate 106a, R3 ) Ph, X )
CH2SO2(CN4)Ph) which was used in the asymmetric
synthesis and also in a formal asymmetric synthesis of
(+)-strictifolione (107; Scheme 15). A Julia-Kocienski
olefination was used as the key step to create the
E-configured alkene.131
The R,R′-bisalkylation of SAMP-hydrazones, combined
with Sharpless asymmetric dihydroxylation, was applied as
a key stereoselective step in the convergent, asymmetric total
synthesis of attenol A (108; from intermediate 106b, R3 )
CH2CHdCH2, X ) CH2I; Scheme 15) and attenol B. The
attenols, which possess challenging structures and interesting
biological activities, were prepared, as a 6.3:1 mixture, in
15 steps, with good overall yield (19%) and high stereose-
lectivity (de, ee g 96%).142
The above-described methodology for the synthesis of
2-methylenated 1,3-diols, and a homogeneous hydrogenation,
was used for the asymmetric synthesis of δ-lactones such
as prelactone B (112a, R3 ) i-Pr) and V (112b, R3 ) Me;
Scheme 16). This time the synthesis was based on consecu-
tive R,R′-bisalkylation of RAMP-hydrazones of dioxanones
using alkyl iodides and tert-butyl bromoacetate as the key
steps. The ketone group resulting from the hydrazone
cleavage with oxalic acid (under two-phase conditions) was
converted via Wittig reaction into a methylene group, giving
a bisalkylation product, 111 (R4 ) CH2COO-t-Bu), analogous
to 103. Acidic acetonide hydrolysis concomitant with lac-
tonization (TFA/DCM/water) provided, after diastereoselec-
tive iridium-catalyzed hydrogenation (Crabtree’s catalyst
Chemical Reviews, 2010, Vol. 110, No. 3 1399
[Ir(cod)(PCy3)(py)PF6]), the lactones 112 in moderate yields
and excellent diastereo- and enantiomeric purities (>98%).143
All stereoisomers of the 2-deoxypentoses and the 2,6-
dideoxyhexoses can be synthesized from 2-phenyl-1,3-
dioxan-5-one RAMP-hydrazone by R-alkylation with allyl
bromide or R,R′-bisalkylation with allyl bromide and methyl
iodide, respectively (Scheme 16).144 Interestingly, the RAMP-
hydrazone of the Cs-symmetric ketone formed one diaste-
reomer, 109, on equilibration by prolonged storage. The
diastereomer 109 was monoalkylated, giving the product with
the new R-alkyl group in the equatorial position with high
selectivity. The second alkylation introduced the R′-alkyl
group in the axial position, giving 110 (R3 ) Me, allyl; R4
) allyl, Me). Aqueous ammonium dihydrogen phosphate
hydrolysis of the hydrazone, followed by stereoselective
carbonyl reduction, and ozonolysis of the allyl CdC, gave
benzylidene-protected aldose or furanose 114.144 The deox-
ysugar diastereomers 114 were synthesized in the racemic
form via the corresponding N,N-dimethylhydrazones.
The R,R′-bisalkylations of the RAMP-hydrazone of 2,2-
dimethyl-1,3-dioxan-5-one 93 with elaborate alkyl halides
(Scheme 16) were used also as strategic transformations in
synthetic approaches to the total synthesis of the potent
antitumor macrolides amphidinolide N (115) and caribenolide
I (116).145,146 The coupling of the dioxanone with alkyl iodide
118 and allylic bromide 117 through hydrazone alkylation
processes (LDA, THF, -78 °C, alkylation time <1 h)
generated the complete C6-C29 carbon framework of the
target amphidinolide N. Subsequent hydrazone cleavage
(saturated aqueous oxalic acid/Et2O) produced the ketone
intermediate in high stereoisomeric purity (dr >95:5 by 1H
NMR). Nonetheless, the intended fusion of the remaining
C1-C5 part of the target onto the carbon framework of the
obtained intermediate by metathesis-based methods was
1400 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 17
Scheme 18
fruitless. The C6-C29 skeleton of caribenolide I was
prepared similarly through the sequential alkylation of
hydrazone 93 with allylic bromide 120 (91% yield) and
iodide 119, giving, after hydrolysis (saturated aqueous oxalic
acid/Et2O), the highly functionalized ketone intermediate in
70% overall yield. Unfortunately, the transformed intermedi-
ate failed, at a later stage in the synthesis, to engage in the
designed, cross-coupling reaction with other building blocks.
r,r′-Alkylation and r-Quaternization (r,r′,r-Alkyla-
tion and r,r′,r,r′-Alkylation). The triple alkylation of
hydrazones is a further extension of the described bisalky-
lation, furnishing a challenging quaternary carbon unit. The
trisalkylation sequence of 2,2-dimethyl-1,3-dioxan-5-one
RAMP-hydrazone (93) was successful for the asymmetric
synthesis of the acetonide-protected 2-keto-1,3-diols and
1,2,3-triols 122 (Scheme 17) bearing a quaternary stereogenic
center. The three stereogenic centers were generated by
sequential R-alkylation, ozonolytic hydrazone cleavage, and
stereoselective reduction of the resulting ketones with
L-Selectride. The products were obtained in good yields and
high diastereomeric and enantiomeric excesses (de, ee g
96%).147
Another triple R-alkylation of the 2,2-dimethyl-1,3-dioxan-
5-one RAMP-hydrazone allowed for the construction of the
two stereogenic centers of (1S,3R,5R,7S)-(+)-sordidin (123;
Scheme 17) and 7-epi-(1S,3R,5R,7R)-(-)-sordidin, both
components of the natural male-produced aggregation phero-
mone of the banana weevil (Cosmopolites sordidus (Ger-
mar)). Another key step of the synthesis, diastereoselective
Lazny and Nodzewska
alkylation via epoxide ring-opening of 124, employed the
azaenolate of 3-pentanone SAEP-hydrazone (125) as the
nucleophile. Subsequent acidic intramolecular acetalization
provided the sordidin C7-epimers (separable by preparative
GC) in good overall yield (39%) as a 1.5:1 diastereomeric
mixture. Each of the epimers could be obtained in high
diastereomeric and enantiomeric purity (de g 97%, ee g
98% by preparative GC).148
The trisalkylation methodology was extended to the
reaction of hydrazone 77 with methyl iodide and hexyl
bromide, leading to 126 (R3 ) Hex). This was transformed
to the trifluoromethylated alcohol 129, with two neighboring
quaternary stereocenters, in good yield (77%) and very good
diastereo- and enantiomeric excesses (de 96%, ee 98%,
Scheme 18).132
Starting from 2,2-dimethyl-1,3-dioxan-5-one, two nucleo-
sides (potential adenosine receptor agonists), 4′-epi-trachy-
cladines A (128) and B, were synthesized in 14 steps
employing the triple R,R′,R-alkylation (R,R′-alkylation and
R-quaternization) of SAMP-hydrazones (Scheme 18). Re-
moval of the chiral auxiliary from the trisubstituted diox-
anone SAMP-hydrazone, and subsequent reduction, gave the
corresponding alcohol, which could be transformed over four
steps into TBS-protected 2′-C-methyl-5′-deoxy-L-lyxose. The
trachycladines were then obtained in an overall yield of
18-21%.149
The ultimate extension of the alkylation methodology
toward the assembly of two quaternary R-carbon atoms of a
hydrazone has been demonstrated in the asymmetric synthesis
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 19
Scheme 20
of 1,3-dihydroxy-2-ketones 131 (Scheme 19).150 The ketones,
bearing two quaternary stereocenters in the R- and R′-
positions, were made starting from 2,2-dimethyl-1,3-dioxan-
5-one SAMP-hydrazone (77) by four consecutive R- and R′-
lithiation/alkylation procedures. The last lithiation was carried
out in the presence of DMPU additive. Ozonolysis of 130
gave the acetonide-protected 1,3-dihydroxy-2-ketones 132,
albeit in low yields (25-31%) with high diastereoselectivity
(de g 96% by 13C NMR). The low yields were attributed to
the high steric hindrance caused by the two quaternary carbon
atoms. Acidic cleavage of 130 with 6 M aqueous HCl/
pentane in a two-phase system (typically requiring a few
days) removed both the chiral auxiliary and the acetal
function (acetonide protection), providing 1,3-dihydroxy-2-
ketones 131 in moderate to very good overall yields
(14-61%) and with high stereomeric purity (de g 91-97%,
ee g 96%).150
6.3. Aldol Reactions
The conditions optimized for the generation and alkylation
of azaenolates of 66 (see the section on alkylation),33 which
gave rise to the highest level of stereoselection (LDA,
toluene, -78 °C, 86% ee), were also applied for the aldol
addition reactions of azaenolates with benzaldehyde and
propionaldehyde. The diastereoselectivities of the processes
were determined by proton NMR of the syn/anti-133 (R2 )
Ph, Et, Scheme 20) mixtures, after in situ hydrazone
hydrolysis (Amberlyst, acetone/water). The optimized hy-
drazone (R1 ) TMS) for the aldol reaction between diethyl
ketone and propionaldehyde gave an anti-isomer (de 37%).
Both isomers had an ee of 83-84%.33 The same reaction
with benzaldehyde gave predominantly the anti-product 133
(de 24%). In this case the anti-isomer had an ee of 77% and
the syn-isomer an ee of 69%. Overall the reaction results
Chemical Reviews, 2010, Vol. 110, No. 3 1401
Scheme 21
Scheme 22
(yields 37% for propionaldehyde and 77% for benzaldehyde)
were moderate but should be treated as a starting point for
further investigations.
A titanium tetrachloride-mediated azaenolate aldol reac-
tion, combined with subsequent 1,2-addition of organocerium
to aldehyde hydrazone, was effective for a diastereo- and
enantioselective synthesis of all-syn-configured, N,O-pro-
tected 1,3-amino alcohols 137 (Scheme 21).151 The titanated
SAMP-hydrazone was obtained as a deeply red DCM
solution by reaction with titanium tetrachloride and diiso-
propylethylamine (DIPEA, Hünig’s base) at -78 °C. Sub-
sequent reaction with aldehydes afforded preferentially the
syn-aldol-like products (β-hydroxyhydrazones) in good yields
and with good diastereoselectivities (de 72-77% by 13C
NMR). The protection of hydroxyl with TBS allowed for
enrichment of the major diastereomer (de 83% to g96%).
Overall, the four-step synthesis gave the protected amino
alcohols 137 with moderate to good overall yields (18-58%)
and high diastereomeric (de 78% to g96%) and enantiomeric
(ee g96%) excesses.151
The syn-selective asymmetric aldol reaction of titanium
azaenolates of SAMP/RAMP-hydrazones 134 with benzyl-
protected 3-hydroxypropanal was also used in an approach
to the convergent, asymmetric, total synthesis of pironetin
(139), a polyketide with immunosuppressive, antitumor, and
plant-growth-regulating activities.152 The TBS-protected prod-
uct 138 (Scheme 22) was obtained in good yield (80% over
two steps). The diastereoselectivity was moderate (de 55%)
but could be boosted to de g 96% through HPLC enrich-
ment. The synthesis of the natural product (ent-138 was used)
required, in addition to the azaenolate aldol reaction, a
Mukaiyama aldol reaction, an asymmetric R-alkylation
(diethyl ketone SAMP-hydrazone with (E)-crotyl bromide,
1402 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 23
LiTMP, -110 °C, THF), and a ring-closing metathesis as
other key steps.
6.4. Michael Reactions
The asymmetric Michael addition of lithiated trifluoroac-
etone SAMP-hydrazone (140) to alkylidenemalonates 141
allowed the synthesis of indolizidinones 145 (Scheme 23).153a
The addition gave products 142 in good yields (64-84%)
and with variable diastereoselectivity (dr from 56:44 to
92:8). Acidolytic cleavage of the hydrazone (H2SO4 in formic
acid), concomitant with ester hydrolysis and decarboxylation,
gave enantiomerically pure keto acids 143, which were
cyclized to dihydropyridinones. Further N-iodopropylation,
followed by radical cyclization, gave optically active trif-
luoromethylated indolizidinones.153a Some R,β-unsaturated
hydrazones reacted unexpectedly (likely via Michael-type
addition) with dimethyl oxoglutaconate (MeOOCsCHd
CHsCO-COOMe) giving functionalized dihydropyrans.153b
7. Reactions of Azaenolates with Electrophiles:
Carbon-Heteroatom Bond Formation
Known, synthetically useful, C-heteroatom bond form-
ing reactions of R-metalated hydrazones 25 (Scheme 24)
include R-sulfenylation with disulfides154 or benzenesulfe-
Scheme 24
Lazny and Nodzewska
nyl chloride (product 146), selenylation with benzenesele-
nenyl bromide (product 147),155 R-phosphinylation156 with
borane-protected chloro-disubstituted phosphane (product
149), R-amination with TrisN3,157 di-tert-butyl azodicarboxy-
late (DBAD),157 or (diphenylphosphinyl)(N,N-dimethylami-
no)hydroxylamine (product 150), R-hydroxylation with
oxaziridines (product 151),158 and R-silylation with silyl
triflates or chlorosilanes to form 148.159 R′-Silyl enol ethers
of R-silyl ketones obtained from 148 are useful reagents for
regio- and enantioselective Mannich reactions (electrophile
Bn2NCH2OCH3 with BF3 · Et2O).160
In recent years, hydrazones have been used, among other
things, for the introduction of heteroatoms in the R-position
of ketones and aldehydes. The R-silyl hydrazones, and the
corresponding carbonyl compounds, were synthesized in
ways analogous to R-alkylation (lithiation with lithium
amides or alkyllithiums with subsequent addition of 2,2,6,6-
teramethylpiperidine) and reaction with silyl electrophiles
such as chlorides or triflates. The introduction of a removable
silicon substituent, which expresses stereodirecting or acti-
vating properties (“traceless directing group”), is often
advantagous from a synthetic strategy point of view (so-
called “silyl trick”). Earlier applications (before the year
2000) of silyl ketones made by the SAMP/RAMP-hydrazone
methodology to stereocontrolled synthesis have been sum-
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 25
Scheme 26
marized in a paper.161 The R-heterosubstituted aldehydes
made by the SAMP/RAMP-hydrazone methodology could
be used in further asymmetric reactions, e.g., in the Mu-
kaiyama aldol addition, giving excellent anti-selectivities.162
The temporary introduction of a thexyldimethylsilyl block-
ing group was used in an elegant asymmetric synthesis of
the diketotriterpenoid 157 (Scheme 25),163 which was isolated
from the Indonesian marine sponge (Hyrtios erectus). Starting
with silylation and alkylation of butanone SAMP-hydrazone
(152) with the appropriate allylic bromide 154, the enanti-
omer of the natural product (S,S)-157 was obtained in low
overall yield (4%) due to inefficiency of one of the last steps
(reductive coupling of allylic alcohols by TiCl3/LiAlH4).
Alternatively, generation of both stereogenic centers at the
C6 and C19 positions of the C2-symmetrical molecule 157
and dimerization with dibromide 155 were effected via
R-alkylation of SAMP/RAMP-hydrazones. Finally, the ra-
cemization-free, hydrazone cleavage, combined with desi-
lylation (CuCl2 or oxalic acid), allowed for the synthesis of
both enantiomers in good overall yields (56%) and with high
asymmetric inductions (de, ee g 96%). The absolute con-
figuration of the natural material was determined as R,R
owing to the known absolute stereocontrol by the SAMP/
RAMP-hydrazone alkylation.163
The R-silylation (Me2(Thex)SiCl), followed by R-alkyla-
tion (MeI) of SAMP-, RAMP-, and DMH-hydrazones of
acetaldehyde or acetone, allowed for the synthesis of both
enantiomers and the racemic form of organosilicon odor-
ants.164
Analogous, temporary R-silylation of SAMP-hydrazones
of simple cyclic and acyclic ketones 159 was used to obtain
enantiopure R-silyl ketones 161 (ee > 98%, Scheme 26).
Subsequent regio- and diastereoselective electrophilic fluo-
Chemical Reviews, 2010, Vol. 110, No. 3 1403
rination of the R-silylketones 161, using N-fluorobenzosul-
fonimide as the fluorinating agent, gave R′-fluorinated
products 162 (de 37% to g98%). An almost quantitative,
racemization-free cleavage of the silyl directing group with
a buffered mixture of fluorides (HF, Bu4NF, NH4F, KH2PO4
in THF/water) gave R-fluoroketones in good yields (88-99%)
and high enantiomeric excesses (ee 87-96%).165
Some enantiomerically pure R-silyl ketones 161, prepared
via the above-described SAMP/RAMP-hydrazone methodol-
ogy, were used in regio- and enantioselective Mannich
reactions for the preparation of R-substituted β-amino
ketones.166
R-Silyl dioxanones, made by silylation of dioxanone
RAMP-hydrazones, have found similar synthetic applica-
tions.167,168
Interestingly, R-phenylselenenylation of lithium N,N-
dimethylhydrazone azaenolates, formed in the reaction of
LDA with linear aliphatic aldehyde hydrazones 163, led to
R-phenylselenenyl hydrazones 164 (R2 ) H) when 1.2 equiv
of phenylselenenyl bromide was used (Scheme 27). However,
when excesses (2.5 equiv) of the base and of PhSeX (X )
Cl, Br) were used, phenylselenenyl nitriles 165 were formed.
Hydrazones of R-branched aldehydes (R2 * H) also gave
nitriles 165. SAMP-hydrazones reacted analogously, albeit
providing the corresponding nitriles in racemic form. The
reactions of linear aldehyde hydrazones with PhSCl in place
of PhSeBr led to R-phenylsulfanyl hydrazones 166, but
R-branched aldehyde hydrazones again gave nitriles 167.155
The asymmetric formation of the carbon-heteroatom (Si,
P, S) bonds was also possible via ortho-lithiation in the
enantioselective synthesis of planarly chiral 1,1′-bisbenzoyl-
ferrocene and diferrocenyl ketone derivatives (Schemes 10
and 11).125-127
1404 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 27
8. Reactions at the Azomethine Carbon with
Electrophiles
8.1. Typical Reactions of Formaldehyde and
Higher Aldehyde Hydrazones (Prior to 2000)
Direct reactions on the second nucleophilic site of hydra-
zones 22 (the azomethine carbon) with active electrophiles
(Scheme 28) can be combined with further hydrazone
transformations. The overall process, in which hydrazones
(especially formaldehyde hydrazones) act as nucleophilic acyl
(and nitrile) equivalents, can be used for the preparation of
aldehydes, nitriles, and 1,2-dicarbonyl compounds. The
applications of this useful synthetic methodology have been
recently reviewed.9,169
Hydrazones (derived from DMH, diisopropylhydrazine,
N-aminopyrrolidine, SAMP, or SAMP analogues) of form-
aldehyde and some higher aldehydes (saturated, unsaturated,
or aromatic aldehydes) react with a multitude of reagents
including Vilsmeier-type formylating reagent (product 168),
Scheme 28
Lazny and Nodzewska
Mannich aminomethylating reagent (CH2dNR2+X-) or Bö-
hme reagent ((CH2)5NdCH2+Cl-) (product 169), bromine
(giving geminal dibromo hydrazones), Ph2PCl/Py (phospho-
rylation), p-nitrophenylsulfenyl chloride (sulfenylation) to
form corresponding products 170, TFAA (in the presence
of 2,6-lutidine or pyridine, trifluoroacetylation, product 171),
and sulfonyl isocyanates (e.g., TsSO2NdCdO, product
172).9,169 Reactions known for formaldehyde hydrazone only
include acylation with milder reagents (to form 173), e.g.,
acetyl chloride, ethyl chloroglyoxylate, and trichloroacetyl
chloride, and reactions with other electrophilic reagents, e.g.,
chloral and trifluoromethyl substituted phthalazine. Other
known reactions of formaldehyde hydrazones are additions
to Michael acceptors (to form 174), alkylidenemalonates,
R,β-unsaturated ketones, R,β-unsaturated lactones, nitroalk-
enes, or activated cyclic alkenes with two electron-withdraw-
ing groups. The reactions on the nucleophilic azomethine
carbon of formaldehyde hydrazones with aldehydes (e.g.,
n-BuCHO, CyCHO, and PhCHO) are also possible but
require the Lewis acid promoters (product 175). The same
reactions with more active aldehydes, Cl3CCHO,
(TBS)OCH2CHO, may be run without the promoters. Other
reactions with R-amino aldehydes, with carbohydrate-derived
aldehydes (to give R-hydroxy hydrazones); with trifluoro-
methyl ketones (to give R-hydroxy-R-(trifluoromethyl) hy-
drazones) are also known. In some of the additions, the
electrophiles also required activation by Brønsted acid or
Lewis acid catalysts.169
8.2. 1,2-Addition of N,N-Dialkylhydrazones to
Aldehydes
The known reactions on the azomethine carbon of alde-
hyde hydrazones 176b,c and 176e,f have been extended to
aldehydic electrophiles. The ZnCl2- or Et2AlCl-promoted 1,2-
addition of achiral and chiral formaldehyde N,N-dialkylhy-
drazones to simple aldehydes afforded the corresponding
R-hydroxy hydrazones 177 (Scheme 29).170 More reactive
aldehydes reacted without the addition of promoters. Reac-
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 29
tions of SAMP-hydrazone (176c) and the D-mannitol-derived
hydrazone afforded the corresponding adducts with low
stereoselectivity and as inseparable mixtures of diastereo-
mers, albeit in high yields. Application of the formaldehyde
hydrazone of (S)-1-amino-2-(methoxydiphenylmethyl)pyr-
rolidine (SAPP, 176e) gave chromatographically separable
mixtures of diastereoisomers, also in good yields. Thus, the
chiral hydrazone acted in this process as a resolving agent
allowing for the preparation of enantiomerically pure adducts.
Scheme 30
Chemical Reviews, 2010, Vol. 110, No. 3 1405
Similarly, 1,2-addition of formaldehyde N,N-dialkylhy-
drazones such as DMH-hydrazone (176a), N-aminopyrroli-
dine hydrazone (176b), and SAMP/RAMP-hydrazone
(176c,d) to carbohydrate-derived R-alkoxy aldehydes took
place under neutral conditions and in the absence of
promoters (Scheme 29).171 The corresponding R-hydroxy
hydrazone products were obtained in fair to good yields
(38-80%) and with high anti-diastereoselectivities (dr from
79:21 to >98:2). The same addition reaction of formaldehyde
N-aminopyrrolidine hydrazone (176b) to N-Boc-protected
R-amino aldehydes (Boc-leucinal and Boc-phenylalaninal)
provided the corresponding adducts in good yields and
diastereoselectivity (75% and 82%, dr 78:22 and 85:15,
respectively). A variety of interesting molecules including
R-homologated carbohydrates, cyanohydrins, or hydroxy-β-
(aminocarbonyl) compounds could be accessed upon sub-
sequent manipulation of the hydrazone functionality follow-
ing known procedures.
8.3. Michael Reactions
The Michael addition of formaldehyde hydrazones 176c-i
(Schemes 29 and 30) to aliphatic and aromatic alkylidene
malonates was investigated.172 The addition of SAMP- and
SAPP-hydrazones (176c and 176e) in the presence of MgI2
(mild Lewis acid) afforded Michael adducts 178 correspond-
ing to 179-181 (Scheme 30) in excellent yields (70-98%)
and good diastereoselectivities (68-79%). Subsequent ozo-
nolysis, or direct BF3 · OEt2-catalyzed thiolysis of the hy-
drazone CdN bond, gave the corresponding aldehyde 181
or dithioketals 179 in enantiomerically enriched form (ee
68% to >98%).173 Additional dithioketal desulfurization
(Raney Ni, ultrasound) to give 180 or decarboxylation was
also possible. In the series of aromatic aldehyde hydrazones,
diastereomerically pure (de > 98%) major Michael addition
products (S,S)-178 were obtained in good yields (77-93%)
after chromatographic purifications.172
1406 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 31
Scheme 32
The Michael addition of formaldehyde SAMP-hydrazone
(176c) to 5,6-dihydro-2H-pyran-2-one (R,β-unsaturated lac-
tone) under neutral conditions, followed by trans-selective
R-alkylation, was used in the asymmetric synthesis of
R-substituted β-formyl δ-lactones 183 (de g 98%, ee
80-95%) and 4-substituted furofuran lactones 182 (de g
98%, ee 80% to >98%). The final products were obtained in
acceptable overall yields by cleavage of the auxiliary, using
ozonolysis or a hydrolytic domino reaction.174
The same nucleophilic conjugate addition of chiral SAMP-
and SAPP-hydrazones of formaldehyde (176c and 176e) to
doubly activated cyclic alkenes was reported to afford the
corresponding Michael adducts 178 (Scheme 30) in variable
yields (71-97%) and selectivities (dr from 1.3:1 to 5:1). The
reactions of achiral hydrazone 176b were also tested. Major
isomers could be chromatographically purified. The reactions
took place either with or without the presence of MgI2 (mild
Lewis acid) depending on the type of substrate.175 Further
reactions, i.e., hydrazone cleavage and carbonyl transforma-
tions, provided access to ketones, acetals, thioacetals, and
nitriles.
Related reactions of 1,3-disubstituted 1H-pyrazole-4-
carbaldehyde N,N-dimethylhydrazones with the Vilsmeier-
Haack reagent have been investigated.176 The electrophilic
substitution reactions of the hydrazone azaenamine at the
azomethine C-atom yielded the 1,4,5-triazapentadienium
salts, hydrolysis of which gave 2-hydrazono-2-(1H-pyrazol-
4-yl)ethanals 185 (Scheme 31). The same reactions of
substituted benzaldehyde N,N-dialkylhydrazones (DMH-
hydrazones and N-aminopyrrolidine hydrazones) gave 1,4,5-
triazapentadienium salts that reacted with N,N-dimethylhy-
drazine to give (1E,2Z)-2-(N,N-dialkylhydrazono)-2-phenyl-
ethanal N,N-dimethylhydrazones 186.177
Reaction on the azomethine carbon was also used for the
synthesis of a heterocyclic nucleus. A key reaction in a
sequence leading to 2,3-bis(trifluoromethyl)quinoxaline (188;
Scheme 32) was the reaction of trifluoroacetaldehyde N,N-
dimethylhydrazone (187) with TFAA followed by reaction
with o-phenylenediamine and cyclization.178
Lazny and Nodzewska
Scheme 33
Scheme 34
9. Nucleophilic Additions to
N,N-Dialkylhydrazones
9.1. Reactions of Hydrazones with
Organometallics
The nucleophilic 1,2-addition of various organometallics,
including organocerium, organomagnesium, organolithium,
organoytterbium (e.g., RLi/YbCl3), and organobarium com-
pounds, and (TMS)CN to the CdN group1,179,180 can be used
for the preparation of many hydrazines 189 and amines 190
(Scheme 33). This methodology, combined with other
reactions of hydrazones, mostly chiral SAMP-hydrazones,
is successful for the preparation of chiral amine natural
products.1 The diastereoselectivity of the 1,2-addition to
SAMP analogue hydrazones have been studied by Denmark
(see reactivity characteristics).32 Syntheses of indolizidine,
pyrrolidine, and piperidine alkaloids based on stereoselective
1,2-addition of organometallics to the CdN bond of SAMP/
RAMP-hydrazones have been described by Enders.181 The
nucleophilic addition of organometallics to CdN double
bonds combined with Raney Ni or samarium iodide cleavage
of the hydrazine bond is widely used in the synthesis of
amines, including bioactive chiral amines.
The addition reaction of dilithiated ketones 192 (ketone
dilithio R,β-dianions prepared from β-(dichlorobutyl)stannyl
ketone 191 with n-BuLi) to the DMH-hydrazone of benzal-
dehyde to give lithium (Z)-enolates containing a lithium
hydrazone amide, 193 (Scheme 34), was recently reported.182
The addition to the CdN bond took place selectively at the
β-position of the dianion. Quenching (water) or trapping
((TMS)Cl) of the enolate resulted in racemic γ-hydrazino
ketone 194 or the silyl enol ether of 194, respectively (74%
yield for both reactions).
The asymmetric version of the 1,2-addition was almost
exclusively realized with the SAMP-hydrazones. For ex-
ample, efficient asymmetric synthesis of R-trifluoromethyl-
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 35
Scheme 36
substituted primary amines (usually isolated as amides) was
possible via 1,2-addition of alkyllithium reagents (including
allyllithium) to trifluoroacetaldehyde SAMP- or RAMP-
hydrazones 195.183,184 The addition gave hydrazines 196,
which in turn were benzoylated and transformed into chiral
amines via SmI2-promoted N-N bond cleavage (Scheme 35).
Reaction of trifluoroacetaldehyde SAMP- or RAMP-hydra-
zone 195 with alkyllithiums resulted in the hydrazines in
good yields (48-79%) and good to excellent diastereose-
lectivities (72-98%). However, the reaction with phenyl-
lithium gave low yield (15%) but good de (86%, after
chromatography 88%). The subsequent benzoylation (Ph-
COCl, DMAP, Et3N or n-BuLi, PhCOCl) and cleavage
(SmI2, THF/DMPU) produced the N-benzoyl R-trifluorom-
ethylated amides (R ) alkyl) usually in good yields
(71-95%) and enantioselectivities (97-99%).183 The ally-
lation reaction with an achiral morpholine-derived hydrazone
was also investigated; yields varied from 0% to 62%.184 The
approach via the oxidation of chiral R-trifluoromethylated
homoallylamines was also effective for the synthesis of
β-amino aldehydes, carboxylic acids, and esters.185
Very high diastereoselectivity (>98% de) was also reported
in an analogous addition reaction of organolithium and
Grignard reagents to trifluoroacetaldehyde hydrazone 200
derived from (R)-N-benzylphenylglycinol (Scheme 36). The
addition proceeded on the Re face of the chelated hydroxy-
hydrazone, providing, after hydrogenolysis, the (R)-R-
trifluoromethylated amines 202.186
An ingenious approach to the construction of chiral
benzazepine derivatives187 and cyclic alkaloid skeletons188
was realized by the combination of a diastereoselective 1,2-
addition and the ring-closing metathesis reaction. For
example, the highly stereoselective 1,2-addition to benzyl-
Chemical Reviews, 2010, Vol. 110, No. 3 1407
Scheme 37
Scheme 38
protected 4-hydroxybutyraldehyde or 5-hydroxypentanal
SAMP-hydrazones (Scheme 37), combined in a one-pot
reaction of the directly resulting lithium hydrazide with
acryloyl chloride, gave hydrazine derivatives 204. A ring
closure metathesis of 204 with the first-generation Grubbs
catalyst, Cl2Ru(dCHPh)(PCy3)2 (or the second-generation
catalysts189) gave lactams 205. The intermediates 205 could
be transformed to chiral bicyclic lactams and further to (-)-
coniceine188b (207, n ) 1) in high enantiomeric purity. The
analogous 1,2-addition, acryoylation, and metathesis on
benzyl-protected 2-hydroxybutyraldehyde SAMP-hydrazone
as the starting material provided intermediate 205, which was
used for synthesis of the enantiomerically pure piperidine
alkaloid (+)-β-conhydrine188a (206). In these syntheses the
N-N bonds were cleaved either reductively (BH3/THF) with
simultaneous lactam reduction188a or oxidatively (MMPP).188b
The planar chirality of ferrocenes bearing hydrazone
groups (SAMP-hydrazone 208 or analogous RAMP-hydra-
zones and N-aminopyrrolidine-derived hydrazones, Scheme
38) competes with the typically excellent stereocontrol of
diastereoselectivity afforded by the chiral auxiliaries (SAMP,
RAMP) in the nucleophilic addition to the CdN hydrazone
bond (dr 24.3:1 for matched pair, 2.1:1 for mismatched pair,
and 19.3:1 for the achiral N-aminopyrrolidine-derived hy-
drazone),190 despite the fact that the chiral hydrazone
auxiliary remained the dominant stereocontrolling element.
1408 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 39
Scheme 40
The nucleophilic 1,2-addition of alkyllithium reagents to
SAMP/RAMP- or RAMBO/SAMBO-hydrazones of 4-formyl-
[2.2]paracyclophane was a key reaction in the synthesis of
all four possible stereoisomers of enantiopure and diaste-
reomerically highly enriched, R-branched [2.2]paracyclo-
phanylalkylamines (Scheme 39). The chiral SAMP/RAMP
or RAMBO/SAMBO auxiliaries allowed chromatographic
separation of the epimers 211. Addition of alkyllithiums gave
hydrazines 212. Reductive N-N hydrazine bond cleavage,
followed by treatment with benzyloxycarbonyl chloride
afforded the N-Cbz-protected, virtually diastereo- and enan-
tiopure (de, ee 99%), or diastereomerically enriched (de
89-96%), amines 213.191 A series of analogous reactions,
but with DMH-hydrazones, which were prepared in enan-
tiomerically pure form via chiral phase HPLC separation,
was also reported.192
Similarly, an efficient enantioselective synthesis of R-(het-
eroaryl)alkylamines 216 by nucleophilic 1,2-addition of
lithiated heteroarenes to aldehyde SAMP-hydrazones 214
was also described (Scheme 40).193 As before, the hydrazine
N-N single bonds were cleaved by SmI2 or BH3/THF, giving
amines which were isolated in the form of the corresponding
amide or urethane (Cbz or benzoyl) derivatives in good
overall yields (40-78%) and again excellent enantiomeric
excesses (ee 88-99%). Analogous addition reactions of
2-lithiobenzo[b]thiophene, followed by oxone oxidation and
L-Selectride reduction, were used for the asymmetric five-
Lazny and Nodzewska
Scheme 41
step synthesis of heterocyclic β-aminosulfonessthe R-(1,1-
dioxo-2,3-dihydro-1H-1λ6-benzo[b]thiophen-2-yl)-substi-
tuted amines 217 (overall yields 23-49%, de > 96%, ee
88-99%).194
A related reaction, the asymmetric 1,2-addition of trim-
ethylsilyl cyanide to aldehyde SAMP-hydrazones 214,
provided access to amino acids 220 (Scheme 41). The
addition reaction required 2 equiv of titanium tetrachloride
for activation of the imine bond. Methoxycarbonylation of
the resulting hydrazine, and removal of the chiral auxiliary
by action of MMPP (1 week at room temperature) followed
by acid hydrolysis (6 M aqueous HCl), afforded the amino
acids in high enantiomeric excesses (yields 60-98%, ee
94-97%).195 The analogous cyanosilylation reaction of
(2S,5S)-1-amino-2,5-diphenylpyrrolidine-derived aliphatic
hydrazones 221, promoted by Et2AlCl, afforded the corre-
sponding hydrazinonitriles 222 with high diastereoselectivity
(dr 91:9 to >99:1). The resolving properties of the auxiliary
allowed chromatographic isolation of the adduct products
in diastereomerically pure form (dr >99:1) and in good yields
(80-84%).196
The classical, diastereoselective, nucleophilic 1,2-addition
of various organocerium compounds to the CdN double
bond of ω-SAMP-hydrazonosulfonates 223 was used as the
key step in a flexible asymmetric synthesis of various
3-substituted γ- and δ-sultams 226 (Scheme 42).197 The
resulting hydrazines were obtained in good to excellent
diastereomeric excesses (de 78-96%). Removal of the chiral
auxiliary by reductive (BH3/THF) N-N bond cleavage and
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 42
Scheme 43
reaction with chloroformate afforded the corresponding
N-protected aminosulfonates 225 without racemization. Hy-
drolysis of the sulfonic acid ester, and subsequent treatment
of the free sulfonic acids with phosgene in toluene, led to
the aminosulfonyl chlorides, which were cyclized to 3-sub-
stituted γ- and δ-sultams 226 in good to excellent overall
yields (39-51%) and high enantiomeric excesses (ee
78-99%).
The 1,2-addition to CdN was combined with other
reactions in a spontaneous reaction sequence. A tandem
Scheme 44
Chemical Reviews, 2010, Vol. 110, No. 3 1409
nucleophilic 1,2-addition and ring closure of SAMP- or
RAMP-hydrazones, followed by oxidative cleavage of the
auxiliary, was used for the asymmetric synthesis of 3-aryl-
substituted 2,3-dihydro-1H-isoindol-1-ones 228 and 229
(overall yields 31-69%, ee g 97%, Scheme 43)198 and
3-substituted 3,4-dihydro-2H-isoquinolin-1-one 233 (Scheme
44). Addition of the ring-lithiated or methyl-lithiated dieth-
ylbenzamides (prepared by directed ortho-metalation) pro-
duced the lithiated hydrazines, which in turn reacted, in situ,
in a substitution reaction with the amide group closing the
five-membered (228) or six-membered (230) rings (yield
36-70%, de up to g96%). The hydrazine cleavage (oxida-
tive with MMPP) gave the corresponding five-membered
isoindolinones 229 (yield 83-97%, ee up to g99%), the six-
membered dihydroisoquinolines 233 (Scheme 44, yield
23-67%, ee 0% to g96%), the dihydro-2H-isoquinolin-1-
ones 231 (yield 14-93%, ee 0% to g96%), or the tetrahy-
droisoquinolines 232 (yield 35-85%, ee g 97%). Depending
on the specific hydrazine product, three different cleavage
methods (Zn/AcOH, MMPP, or SmI2) were used to give
racemization-free N-N bond cleavage to amides 231 with
varied degrees of success.199 Other methods used were: BH3/
THF to give amine 232 and Li or Ca in liquid NH3 to give
imine 233. The method based on borane cleavage was
applied to the enantioselective synthesis of both enantiomers
of a natural alkaloid belonging to the tetrahydroprotober-
berine familystetrahydropalmatine (234; ee 98%). The
natural product (R)-(+)-234 was obtained in seven steps in
9% overall yield, while the (S)-(-)-234 isomer was prepared
in 17% overall yield.200
An interesting addition of R-lithiomethoxyallene (235;
prepared in situ by reaction of n-BuLi with methoxyallene
in THF at -40 °C) to aromatic hydrazones 227 and 237 in
THF, leading to N-(dialkylamino)-3-pyrrolines 236, was also
described (Scheme 45).201 In the case of SAMP-hydrazones
227, this reaction occurred with excellent stereoselectivity
(de > 99%) and good yields (69-88%). The reaction with
other hydrazones (237, dimethyl, piperidinyl, morpholinyl)
gave, besides 3-pyrrolines 239, azetidine byproducts 240 and
the uncyclized intermediates R-allenyl hydrazines 238. It was
shown that formation of azetidines could, however, be
reversible: at higher temperatures and longer reaction times,
the azetidines transformed to 3-pyrrolines. The reaction in
diethyl ether, instead of THF, stopped at the stage of addition
and provided R-allenic hydrazines 238.202 In this case,
cleavage of the N-N hydrazine bond could be effected under
1410 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 45
Scheme 46
a variety of conditions, giving various products: chlorofor-
mate followed by Ni/H2 gave a mixture of pyrrolines and
pyrrolidines, m-CPBA gave pyrroles, and aqueous HCl gave
aminopyrroles.
A double addition of organometallics to bishydrazones
derived from SAMP/RAMP-hydrazines and dialdehydes is
an interesting expansion of the methodology. Bis(SAMP-
hydrazone)s (C2-symmetric) of dialdehydes 241 also under-
went nucleophilic 1,2-addition of organocerium reagents to
the CdN double bond and, according to previously described
typical procedures, gave amide-protected C2-symmetric 1,n-
diamines 242 in high diastereo- and enantiomeric purity (de
72-98%, ee 96-98%).203 The addition of PhMgCl to the
glyoxal bis(RAMP-hydrazone) 241 (Scheme 46) was used
for the synthesis of stable N-heterocyclic carbenes 243.204
9.2. r-Alkylation Combined with 1,2-Addition to
the Hydrazone
The combination of R-alkylation of chiral aldehyde
hydrazones (usually SAMP/RAMP-hydrazones) with sub-
sequent diastereoselective 1,2-addition is an effective method
for assembling two adjacent stereocenters. The proce-
dure, which is typically followed by cleavage of the N-N
bond of the resulting hydrazine, offers access to chiral
Lazny and Nodzewska
Scheme 47
amines including tetrahydrobenzazepines,205 disubstituted
azetidines,206 and piperidines.
The combination of R-alkylation with 1,2-addition allowed
for the synthesis of 2,3-trans-disubstituted azetidines (Scheme
47)206a including 3-substituted azetidine-2-carboxylic acids
(Scheme 47, R3 ) COOH).206b Alkylation of aldehyde
SAMP-hydrazones 214 with SEM- or BOM-Cl was followed
by nucleophilic 1,2-addition of various organocerium re-
agents to the hydrazone CdN double bond to give 1,3-
hydrazino alcohols 245. Epimerization-free reductive cleav-
age of the auxiliary with borane/THF, N-tosylation of the
resulting amine, and cyclization under Mitsunobu conditions
provided the corresponding N-tosylazetidines 247 in good
yields (77-97%). Detosylation of the azetidine could be
accomplished with sodium/naphthalene. The products were
obtained with excellent diastereomeric (de 93-96%) and
enantiomeric (ee 96%) excesses. The 2-phenylhydrazino
alcohols 245 (R2 ) SEM, R3 ) Ph), obtained by addition of
PhLi/CeCl3, gave an azetidine that could be oxidatively
transformed (oxidation of the phenyl group to the carboxyl
group with RuCl3, H5IO6, and MeCN/CCl4/water at room
temperature) into R-amino acid-type azetidine 247 (R3 )
COOH) with excellent stereochemical purity (de, ee 96%).206b
In this case the phenyl moiety served as a synthetic
equivalent of the carboxylic acid function.
The diastereoselective R-alkylation of R-tert-butyl- or
R-benzylsulfanylated acetaldehyde SAMP-hydrazones with
various electrophiles, and subsequent nucleophilic 1,2-
addition of organocerium reagents to the hydrazone CdN
double bond, was also used for the asymmetric synthesis of
various protected anti-1,2-sulfanyl amines 251 bearing two
adjacent stereogenic centers (Scheme 48). The resulting
hydrazines 250 were again converted, by reductive N-N
bond cleavage, to amines which were protected with the Cbz
group before isolation or isolated directly. The products,
again, had excellent diastereomeric and enantiomeric ex-
cesses (de, ee 96%).207,208
The unnatural enantiomer of the alkaloid R-conhydrine
(256) was synthesized starting from a protected glycol
aldehyde SAMP-hydrazone, 252, via R-alkylation and 1,2-
addition of alkyllithium reagent 254 (Scheme 49). After
borane cleavage of the auxiliary and simultaneous acidic
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 48
hydrolysis of the acetal, hydrolysis of the tert-butyldimeth-
ylsilyl ether and imine formation took place during acidic
workup. The final alkaloid product 256 was obtained by
imine reduction in moderate overall yield (21%) and with
excellent diastereo- and enantiomeric excesses (de, ee >
96%).209 Analogously, starting from the same protected
glycol aldehyde hydrazone, the 2-substituted piperidin-3-ols
260 were also obtained in good overall yields (51-76%)
and excellent diastereomeric and enantiomeric excesses (de,
ee > 96%, Scheme 49).210
Similarly, the 1,2-addition of an organocerium reagent
(made from [3-(tert-butyldimethylsiloxy)propyl]lithium) to
an R-alkylated (alkylating reagent 257, Scheme 49) n-
pentanal RAMP-hydrazone, 261, was used successfully for
the synthesis of indolizidine alkaloids (-)-209I and (-)-
223J and their C5-epimers with high stereoselectivity (de
96-99%, ee > 99%, Scheme 50).211
Scheme 49
Scheme 50
Chemical Reviews, 2010, Vol. 110, No. 3 1411
Scheme 51
Following a four-step reaction sequence of R-alkylation,
organolithium 1,2-addition (with later carbamate protection
of the formed hydrazine), TBS-deprotective cyclization, and
cleavage of the auxiliary, several cis-4,5-disubstituted ox-
azolidin-2-ones 267 (Scheme 51) were obtained in fair to
good yields (27-78%) and in high diastereo- and enantio-
meric purities (de 88% to >96%, ee > 96%).212
10. Formation of Heterocycles
The application of electrophilic and nucleophilic reactions
toward the synthesis of heterocyclic structures is covered in
the relevant sections of the review. Cycloaddition reactions,
and carbene insertion reactions giving heterocycles, will be
covered in this section.
1412 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 52
10.1. Hetero-Diels-Alder Cycloaddition
R,β-Unsaturated N,N-dimethylhydrazones (azadienes) react
with electron-deficient dienophiles in [4 + 2] cycloadditions,
building six-membered N-heterocyclic products (Scheme
52).74,213 The reactions provide, depending on the reactants
and reaction conditions, access to substituted pyridines and
tetrahydropyridines, important azaheterocyclic frameworks
useful for the synthesis of polycyclic alkaloids.214,215 Depend-
ing on the reaction conditions, the cycloaddition (to give 269,
271, and 273) may, or may not, be accompanied by
spontaneous dimethylamine elimination (to form dihydro-
pyridine structure), usually followed by aromatization, to
form the pyridine nucleus (270, 272, 274). In the intermo-
lecular version of the Diels-Alder reaction, active dieno-
philes, e.g., acrylate, isothiazole, maleimide, oxazoles, and
quinone derivatives, were used with varying degrees of
success (yields varied from <10% to >90%). The intramo-
lecular version of the Diels-Alder cycloaddition reaction
usually gives low to moderate yields (12-87%) of the
cyclized products (273, 274). The Diels-Alder reaction of
a boronate-functionalized diene (4-boronato-1-azadiene), e.g.,
DMH-hydrazone or chiral hydrazone (SAMP analogues), can
be combined with allylboration to form tandem aza [4 +
2]/allylboration reactions.216
Figure 6. Selected heterocyclic products synthesized by aza-Diels-Alder reaction.
Lazny and Nodzewska
More recently, the well-known [4 + 2] cycloadditions of
N,N-dimethylhydrazones of R,β-unsaturated aldehydes acting
as azadienes to dienophiles, described in the literature as aza-
Diels-Alder (ADA reaction) or hetero-Diels-Alder (HDA
reaction), were used as key steps in the syntheses of many
marine alkaloids and other products with piperidine and
pyridine skeletons. Figure 6 shows the main heterocyclic
products synthesized by this approach, 9-(methoxycarbonyl)-
7H-[1]benzothieno[4,5,6-ij][2,7]naphthyridin-7-one (275), an
analogue of kuanoniamine A (marine alkaloid),217 a regioi-
somer of the marine alkaloid meridine 9-hydroxybenzo[b]-
pyrido[4,3,2-de](1,10)-phenanthrolin-8-one (276),218 diaza-
quinomycin A (277),219 dihydrofuroquinolinedione derivative
278,220 substituted 1,8-diaza-9,10-anthraquinone 279,221 1,5-
diazaanthraquinone derivatives 280,222,223 and tetracyclic
compound 281.224
The Diels-Alder reaction of 282 (Scheme 53) was the
key transformation in the synthesis of analogues of the
marine pyridoacridines meridine and ascididemin. Mixtures
of diazaanthraquinones 284 and 285 (R3 ) Me, Scheme 53;
the major isomer was 284) were obtained by the reaction of
different quinoline-5,8-diones and substituted N,N-dimeth-
ylhydrazones.225 A series of tetracyclic compounds 286 and
287 were subsequently made by the action of dimethylfor-
N,N-Dialkylhydrazones in Organic Synthesis Chemical Reviews, 2010, Vol. 110, No. 3 1413

Scheme 53 Scheme 55

Scheme 56
mamide diethylacetate and cyclization. The alkaloid sebas-
tianine A (288) and its regioisomer were prepared in
analogous reactions of N-protected indole-4,7-diones with
(trifluoroacetamido)cinnamaldehyde N,N-dimethylhydra-
zone226 followed by cyclization. The β-unsubstituted R,β-
unsaturated N,N-dimethylhydrazone (acroleine DMH-hydra-
zone) was also tested and typically gave a mixture of
products 284 and 285 (R3 ) H).227
Notable examples of these new applications are the
reactions of phosphorus, sulfur, and silicon substrates as
azadienes. The hetero-Diels-Alder cycloaddition of the
azadiene 289 with N-phenylmaleimide or an excess of diethyl
acetylenedicarboxylate (Scheme 54) gave the phosphorylated
pyridine derivatives 290 (yield 37%) and 291 (yield 63%),
respectively. The substituted pyridine 291 resulted from two
subsequent [4 + 2] additions of two substituted acetylene give heterocyclic compound 295 or 296 with good yields
molecules to 289.228 The substrate for the reaction, 3-phos- (75-80%, Scheme 56).231,232
phinyl-1-aza-1,3-butadiene 289, was obtained in the reaction An implementation of the HDA reaction that demonstrates
of N,N-dimethylformamide dimethyl acetal with (hydra- its powerful synthetic utility is the intramolecular formation
zonoalkyl)phosphine oxide. of multiple cyclic structures. For example, a double intramo-
The Wittig-type reactions of saturated β-phosphorylated lecular hetero-Diels-Alder reaction of R,β-unsaturated hy-
hydrazones can provide azadiene substrates for the hetero- drazones 297 was applied to the synthesis of 2,2-bipyridines
Diels-Alder reactions.229 298 (Scheme 57).233
Despite constraints of unfavorable diene configuration,
(RS,E,E)-2-[(1S)-isoborneol-10-sulfinyl]-2-butenal N,N-dim-
ethylhydrazone (292; Scheme 55) acted as 1-azabuta-1,3-
10.2. Staudinger-like [2 + 2] Cycloaddition
diene in a reaction with N-methylmaleimide, giving a The [2 + 2] Staudinger-like addition of ketenes to
cycloadduct (293) in 20% yield and with total endo- formaldehyde,234 alkyl aldehyde,235 or benzaldehyde236 N,N-
selectivity and facial selectivity.230 Enantiopure, sulfinyl R,β- dialkylhydrazones (regarded as N-amino-substituted imines)
unsaturated hydrazones were prepared by addition of can be used to synthesize β-lactam rings (azetidin-2-one
isoborneolsulfenic acidsthe chiral auxiliarysto the corre- rings). The ketene substrates (benzyloxy)ketenes 299236 and
sponding alkynyl hydrazones, or by reaction of the chiral R-aminoketenes (glycine derivative 300237 or oxazolidinone
sulfenic acid with an alkynyl aldehyde or ketone, and derivative 301238) were made in situ through the reaction of
subsequent hydrazone formation with H2NNMe2. acid chlorides with tertiary amines or via the corresponding
A cascade of two subsequent Diels-Alder cycloaddition acids, with 2-chloro-N-methylpyridinium iodide as the
reactions of the N,N-dimethylhydrazone of R-[(trimethylsi- activating agent in the presence of Et3N or DIPEA (Scheme
lyl)oxy]alkylacrolein (294) with an excess of N-phenylma- 58). Simple achiral hydrazones, DMH- and N-aminopyrro-
leimide or methyl acrylate as the dienophile was reported to lidine-derived, typically gave good yields (57-84%). Chiral
Scheme 54
1414 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 57
hydrazones such as SAMP-hydrazones, their analogues
SAEP (302d) and 302e, chiral hydrazones from C2-sym-
metrical hydrazines, e.g., (2S,5S)-2,5-dimethylpyrrolidine
(302f), (2S,5S)-2,5-diphenylpyrrolidine (302g), and 302h
were all used to produce stereoselectively β-lactam hy-
drazides 303-305. There was no diastereoselectivity for the
addition of the chiral SAMP-hydrazone of formaldehyde;
however, the diastereoselectivity was good to excellent for
more sterically demanding auxiliaries (e.g., for SAEP de
64-98% and for C2-symmetrical hydrazone 302h de 98%).234
Yields of the additions ranged from 57% to 96%. The
addition of formaldehyde hydrazones to ketenes formed only
C3-substituted β-lactams (4-unsubstituted azetidin-2-ones).
Higher aldehyde hydrazones gave disubstituted (substituted
in positions C3 and C4) β-lactam rings. In general, the best
stereoselectivities were observed for C2-symmetric auxilia-
ries, in particular for the (2S,5S)-N-amino-2,5-dimethylpyr-
rolidine hydrazone of different aldehydes (Scheme 58, R3
) i-Pr, i-Bu, PhCH2CH2, Ph) and (benzyloxy)ketene in the
presence of Et3N in toluene at 60-100 °C, where cis-3R,4S-
products 303 were obtained with excellent diastereoselec-
tivity (typically cis:trans ) 99:1, yield 67-97%).236 Inter-
Scheme 58
Lazny and Nodzewska
Figure 7. Adducts prepared in cascade cycloadditions.
estingly, the use of the same aldehyde hydrazones and
glycine derivative 300 with DIPEA base in toluene at 80 °C
resulted in the trans-(3R,4R)-product in good yield (59-74%)
and diastereoselectivity (cis:trans ) 99:1).235 The use of
(benzyloxy)acetaldehyde hydrazone under the same condi-
tions produced a 54:46 mixture of trans and cis products. It
was found that the cis-product dominated (99:1) at both room
temperature and 40 °C. The trans-product predominated at
higher temperatures (at 120 °C, cis:trans ) 3:97).237 An
extension of this method, using a subsequent oxidative N-N
bond cleavage reaction by MMPP239 for selected cycload-
ducts (hydrazides), gave access to β-lactams.234,235,238
10.3. Miscellaneous Cyclization Reactions
Some cycloaddition reactions could be successfully com-
bined in cascades, resulting in even more powerful synthetic
transformations. For example, a [4 + 2]/[3 + 2] cascade
cycloaddition of R-(diethoxymethyl)acrolein N,N-dimethyl-
hydrazone to acrylonitrile gave bicyclic adduct 306 (Figure
7).231,232 In this case, the product of the first [4 + 2]
cycloaddition subsequently reacted with a second molecule
of acrylonitrile, affording the [3 + 2] cycloaddition adduct
306 with good yield (80%).
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 59
Scheme 60
The analogous [4 + 2]/[3 + 2] cascade reaction of
crotonaldehyde N,N-dimethylhydrazone and p-quinone
monoimide gave the bicyclic adduct 307.240 Alongside
cycloadduct 307 (yield 24%) the reaction produced 5-sul-
fonamido-2,3-dihydrobenzofuran 308 according to [3 + 2]
cycloaddition only (yield of the major product 32%).
A three-component cascade of reactions, [4 + 2] cycload-
dition, allylboration, and retro-sulfinyl-ene reaction of bor-
onate-substituted unsaturated N,N-dialkylhydrazones with
chiral sulfinimide 310 and an aldehyde, was reported to give
2,6-disubstituted N-(alkylamino)piperidines 312 (Scheme 59).
This transformation was used in the total synthesis of
dihydropalustramate 313.241-243
Boronate-substituted, achiral or chiral, unsaturated N,N-
dialkylhydrazones 314 (Scheme 60) were also used as dienes
in tandem, multicomponent [4 + 2] cycloaddition/allylbo-
ration reaction with maleimides 315 and aldehydes 311.244
Polysubstituted R-(hydroxyalkyl)piperidines 316 were pre-
pared in this way with good yields (39-80%) and good to
excellent diastereoselectivities (80% to >95%). SAMP and
its dimethyl analogue SADP were used for the enantiose-
lective approach. The hydrazone substrates for the reaction
were prepared from the pinacol ester of 3-acroleinboronic
acid and achiral or chiral hydrazines.
Related syntheses use hydrazone azadienes and carbene
complexes: the coupling reaction of enyne N,N-dimethyl-
hydrazones 317 with Fischer carbene complexes 318 pro-
vided alkenylpyrrole derivatives 319 (Scheme 61). An alkyne
insertion, followed by nucleophilic attack of the azomethine
nitrogen at the intermediate alkenylcarbene complex, was
suggested as a mechanism for the reaction.245 An analogous
reaction of Fischer alkenylcarbene complexes 321 with
alkenyl N,N-dimethylhydrazones (1-amino-1-azadienes) 320
furnished substituted cyclopentenes 323 (45-55%) in a
regio- and diastereoselective way, along with minor amounts
of pyrroles 322 (25-28%). The diastereoselective modifica-
Chemical Reviews, 2010, Vol. 110, No. 3 1415
Scheme 61
tion of the reaction using enantiopure carbene complexes
derived from (-)-8-phenylmenthol and (-)-8-(2-naphthyl)-
menthol was investigated and showed preference for products
of type 323 with face selectivity up to dr 96:4 (yields 25%
and 43%).246
The carbene-generating thermolysis of the Eschenmoser
hydrazones 326 (prepared in the reaction of aldehydes with
1-amino-trans-2,3-diphenylaziridine 325, Scheme 62) was
used for the preparation of cyclopropane rings fused to
benzimidazole derivatives (327) and benzimidazolquinones,
albeit in varied yields (11-85%).247-249 The effectiveness
of the process could be affected adversely by the formation
of a carbene insertion byproduct.249
Similarly, carbenes generated thermolytically in refluxing
toluene, from Eschenmoser hydrazones such as R-oxacyclo-
N-aziridinylimines 328 and R-azacyclo-N-aziridinylimines
331 underwent ring expansions via the insertion of alkyl
carbenes into carbon-carbon bonds and intramolecular
ammonium ylide formation, respectively (Scheme 63).
Competition with 1,2-hydrogen shift was observed and
investigated. The ring expansion reaction of R-oxetanyl-N-
aziridinylimines 328 (n ) 0) involved alkylidenecarbene
intermediates, whereas the reaction of R-azetidinyl-N-aziri-
dinylimines 331 (n ) 0) gave R-aminoacetylenes 332 (81%
and 78% yield) via 1,2-migration of the hydrogen atom in
alkylidenecarbene intermediates.250 The thermal ring expan-
sion reaction applied to five-membered rings 331 (n ) 1)
gave the six-membered ring structures 337 (n ) 1, R1 ) H,
0-71% yield, Scheme 63) along with 1,2-H migration
products 336 (0-60%). On the contrary, the thermal reaction
of the six-membered substrates 331 (n ) 2) afforded
preferentially the migration products 335 (54-72% yield)
with small amounts (0-19% yield) of the seven-membered
ring expansion product 337 (n ) 2, R1 ) H). The product
distribution depended on the ring size and the nitrogen
substituent R2.
In addition, N,N-dialkylhydrazones may react to form
cyclic products such as 338 and 339 (Figure 8). A β-keto
ester N,N-dimethylhydrazone reacted in the presence of NaH
with a cyclic, sugar-derived, epoxy triflate (C-alkylation) to
form, after intramolecular in situ cyclization (N-alkylation)
of a sodium azaenolate, the polyfunctionalized dihydropy-
1416 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 62
Scheme 63
rrolidine 338.251 The slow reaction (storage over 6 months)
of benzaldehyde N,N-dimethylhydrazone with 4-oxo-2-
(pentafluorophenoxy)-5,6-benzo-1,3,2-dioxaphosphorinane in
DCM solution yielded a product of the phosphorus hetero-
cycle expansion. In this way, the hydrolytically labile product
4-(dimethylamino)-2,5-dioxo-2-(pentafluorophenoxy)-3-phe-
nyldihydro-6,7-benzo-1,4,2-oxazaphosphepine (339) was
formed with high stereoselectivity, in the form of large
yellow crystals.252
11. Free Radical Reactions
Radical reactions of N,N-dialkylhydrazones reported before
year 2000 embraced mostly additions to N-aziridinylimines
and consecutive transformations of so-formed radicals. The
consecutive cascade reactions typically involved diazo
compounds that decomposed to nitrogen and styrene, giving
products of type 343 (Scheme 64). The early synthetic
applications of free radical chemistry of N-aziridinylimines
came from the research group of Sunggak Kim.85,253,254 The
radical reactions of N,N-dimethylhydrazones255-257 and
SAMP-hydrazones257-259 of suitably functionalized radical
Figure 8. Cyclic products accessible from N,N-dimethylhydra-
zones.
Lazny and Nodzewska
Scheme 64
precursors, including β-allenic hydrazones 344, were limited
to intramolecular cyclizations, leading typically to hydrazines
341 (Scheme 64). This type of hydrazone reaction has been
covered in several reviews.20,21,24,25,260 Radical reaction of
N,N-diphenylhydrazones and acyl hydrazones, which are
outside the scope of the present review, have also been
reviewed.20,21
11.1. Radical Cyclization Reactions
More recently, the free radical cyclization reaction of
N-aziridynylimine (hydrazone 346) was used as a key
transformation in the synthesis of a hispidospermidin ana-
logue starting from m-toluic acid (Scheme 65).261 The radical
cyclization, followed by consecutive cascade reactions, gave
the trans-fused perhydroindan 347 in 80% yield.
An appropriately placed N,N-dimethylhydrazone acted as
a suitable radical acceptor in an exo-hept-6-enyl- or exo-
hex-5-enyl-type cyclization (Scheme 66). The addition of
the first radical intermediate of the thiourethane-mediated
alcohol deoxygenation (the Barton-McCombie reaction) to
the hydrazone group in 349 was moderately efficient (yield
of 56%) with triphenyltin hydride as the radical source
Scheme 65
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 66
(product 350, isomer ratio altro-R:altro-β:allo-R:allo-β )
4:9:0:100). On the other hand, use of tributyltin hydride in
this reaction resulted in no cyclization but deoxygenation,
giving product 348 in 41% yield. When 352 was used as a
substrate in the cyclization, a 10:3 mixture of anomers of
the cyclized N-furanosides 353 was formed with preference
for the R-anomer.262,263
11.2. Radical Addition of Silicon-Tethered Vinyl
and Acetylene Groups
Silicon-tethered vinyl or acetylene (ethynyl) groups are
advantageous for C-C bond construction via radical addition
to hydrazones. The temporary incorporation of the silicone
group avoids unfavorable intermolecular reactions and may
provide control of diastereoselectivity of the addition via a
chairlike Beckwith-Houk transition state.264,265 The thiyl
radical-mediated transfer of a vinylsilane group to N,N-
dibenzylhydrazone (or related N,N-diphenylhydrazone)
(Scheme 67) was skillfully used as one of the synthetic steps
in an asymmetric synthesis of an aminosugar, L-daunosamine,
derivative.266 Addition of thiyl radical (generated from PhSH
and AIBN) to the vinyl group in 357 generated an intermedi-
ate alkyl radical which underwent cyclization with the CdN
bond of the hydrazone group. The one-pot treatment of the
resulting intermediate with fluoride removed silicon and
regenerated the vinyl group via elimination of benzenethi-
olate. The resulting allylic hydrazine 358 was obtained in
77% yield and high diastereomeric purity (dr 91:9). The yield
and diastereoselectivity of this transformation were higher
compared to those of the analogous method based on N,N-
diphenylhydrazone. Compared to dibenzyl, the diphenyl
Scheme 67
Chemical Reviews, 2010, Vol. 110, No. 3 1417
analogue was less stable during storage and handling. Diol
protection and trifluoroacetylation of the hydrazine 358
followed by SmI2 N-N bond cleavage provided a suitable
substrate for the Wacker oxidation reaction. The unusually
regioselective Wacker oxidation transformed the vinyl group
in the aldehyde function of the protected form of L-
daunosamine trifluoroacetamide 360. Acidolitic deprotection
of the acetal in methanolic solution provided the methyl
pyranoside of the (trifluoroacetyl)-L-daunosamine 361 in
quantitative yield (3:1 mixture of anomers, Scheme 67).266
A similar approach, based on the radical addition to a
silicon-tethered ethynyl group, obviated the need for the
Wacker oxidation (Scheme 68).267 However, contrary to the
vinylsilane case, the desilylation of the cyclized intermediates
gave no analogous thiolate elimination and consequently
resulted in (phenylthio)vinyl adduct 364, which may be
considered a masked β-aminoaldehyde. Accordingly, the
daunosamine derivative 366 was prepared via diol protection,
hydrazine bond cleavage, and conversion of the vinyl sulfide
to aldehyde under Grieco’s conditions ((TMS)Cl, NaI, HgCl2,
and moist MeCN). One of the anomers of 366, separated by
crystallization, was found to be N-(trifluoroacetyl)-L-daun-
osamine, ultimately confirming the stereochemical course of
the thiyl radical-initiated cyclization of the N,N-dialkylhy-
drazone with the silicon-tethered ethynyl group. Overall, the
one-pot, tin-free method for radical addition-cyclization with
thiophenol and treatment with fluoride leads to diastereose-
lective group transfer from a silicon-tethered ethynyl group
to the CdN bond of N,N-dibenzylhydrazones, affording anti-
hydrazino alcohols with a trans-2-(phenylthio)vinyl sub-
stituent. Combined with methods for the conversion of vinyl
1418 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 68
Scheme 69
sulfides to carbonyls, the described method for radical
cyclization of silicon-tethered ethynyl may constitute a free
radical equivalent of the Mannich reaction of acetaldehyde.267
Synthetic applications of radical reactions of N,N-diphe-
nylhydrazones, which are not the subject of the present
review, are similar to their N,N-dibenzyl counterparts.267 It
was noted, however, that the electron-rich N,N-dibenzylhy-
drazones are rather susceptible to elimination of β-hydroxy
and β-silyloxy groups.268
The thiyl radical-initiated addition-cyclization under
previously described conditions provided a product of vinyl
transfer from silicon to the hydrazone azomethine carbon in
368 (6-exo-cyclization) with almost complete diastereose-
lectivity (98:2, Scheme 69). The yield was dependent upon
the choice of initiator and reaction conditions (10-55%).
The resulting hydrazine 369 was transformed further to the
dihydroxylation substrate 370, a key intermediate for ami-
nosugar synthesis. The stereochemical outcome of the
reaction was in disagreement with the usual prediction of
the Beckwith-Houk model. This observation was explained
on the grounds of a hypothetic dipole repulsion modification
to the Beckwith-Houk model. It was assumed that minimiz-
ing the dipole repulsion between neighboring CdN and C-O
bonds favored a CRsC(dN) dihedral angle, placing the CdN
bond axial within a chairlike transition state. The hypothesis
was substantiated by experimentally observed lowering of
diastereoselectivity for diastereomeric substrate 368a (from
dr 98:2 for product 368 to dr 70:30 for product 369a).269
Two key structural features in the substrates were proposed
to play a crucial role in successful, silicone-tethered 6-exo-
cyclizations: (i) the presence of R-alkoxy substituents and
Lazny and Nodzewska
(ii) conformational constraints such as the benzylidene acetal
ring. In analogy to silanes, a carbon-tethered acetylene
(propargyl ether) 6-exo-cyclization, induced by tributyltin
radicals under conventional heating or microwave irradiation,
produced 372. It is noteworthy that a protodestannylation
of crude 372 (filtration through silica, n-BuLi followed by
propionic acid) provided the functionalized product 373 with
complete diastereoselectivity and in 48% yield from D-
galactose.269
12. Catalytic Reactions
In general, catalytic reactions of N,N-dialkylhydrazones
can be divided into organometallic reactions catalyzed with
hydrazones acting as ligands or metal-hydrazone complexes
and reactions of hydrazone substrates catalyzed with organic
Brønsted acid (BA) catalysts (organocatalysts) or metal-
containing Lewis acids (LAs).
12.1. Catalytic Reactions of Hydrazone Catalysts
in Organometallic Chemistry
12.1.1. N,N-Dialkylhydrazone Catalysts for Addition of
Organometallics to Aldehydes
The enantioselective addition of diethylzinc to aromatic
(e.g., benzaldehyde) and aliphatic (e.g., 3-phenylpropional-
dehyde) aldehydes 374 can be catalyzed with chiral N,N-
dialkylhydrazones (Scheme 70) such as binaphthyl-derived
salicylhydrazones of type 376.270
The best of the hydrazone catalysts, in terms of achievable
yield (48%) and enantioselectivity (58%) of the addition, was
N,N-Dialkylhydrazones in Organic Synthesis Chemical Reviews, 2010, Vol. 110, No. 3 1419

Scheme 70

Scheme 71
the hydrazone 376a (R1 ) H). Catalysis of the same reaction
with binaphthyl hydroxy hydrazones 379271 gave better yields
(75-94%) but low selectivity (20-29% ee). The use of the
bis(RAMP-hydrazone) of type 378271 resulted in the best ee
(60% at -35 °C at 5 mol % catalyst loading) of the entire
series (16-60%); unfortunately, this was usually at the
expense of the yield (36-58%). The terpene-related ketopinic
acid-derived hydrazones 381,271 used with an equimolar
amount of n-BuLi, were also moderately stereoselective
(34-58% ee) but effective catalysts (74-98% yield at 4 mol
% loading). The best yield of 98% was achieved with the
SAMP-hydrazone 381c and the best ee with its DMH
analogue 381a (58%). The β-binaphthol-derived DMH-
hydrazone 378a270 (R ) Me) gave a quantitative yield of
the diethylzinc addition to 4-methoxybenzaldehyde with a
disappointing ee of 31%. On the other hand, the best ee of
80% was obtained in the addition to benzaldehyde. The
N-aminopyrrolidine analogue hydrazone 378b (R ) N(CH2)4)
gave yields of 51-80% with 72-76% ee. The most effective
catalyst from another group of hydrazones, the (1R,2S)-
ephedrine- and (1R,2S)-N-isopropylnorephedrine-derived β-hy-
droxysalicyl hydrazones,272 377 was the ephedrine-based
hydrazone 377h (R1 ) Me, R2 ) 2-hydroxy-1-naphthyl).
This gave the (S)-product of addition with ee from 78% for
3-phenylpropenal to 92% for chlorobenzaldehyde. The yields
ranged from 50% to 84% depending on the aldehyde used.
No clear winner emerges from the gathered data. It seems
necessary to optimize the structure of the hydrazone catalysts
for each specific reaction.

12.1.2. N,N-Dialkylhydrazones as Ligands for
Pd-Catalyzed Allylation
Several SAMP- and SAMP analogue-derived chiral phos-
phino hydrazones, 385,273 386,274 387,275-277and 388,278 have
been developed as ligands for asymmetric palladium-
catalyzed allylation of dimethyl malonate with racemic 1,3-
diphenyl-2-propenyl acetate (Scheme 71).
From all the investigated hydrazones, i.e., derivatives of
SAMP/RAMP (R1 ) H), SADP (R1 ) Me), or SAEP (R1 )
Et), the best performance was reported for the SAMP-
hydrazones regardless of the aldehyde part of the hydrazones.
1420 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 72
In general, the allylation products were obtained with
moderate to good enantiomeric excesses (48-98%) and with
yields of 26-99%. Typical conditions (2 mol % [Pd(η3-
C3H5)Cl]2 and 6 mol % 387a,276 N,O-bis(trimethylsilyl)ac-
etamide, LiOAc, room temperature, 24 h, and THF as a
solvent) gave a product with 98% ee. Substitution of LiOAc
base with KOAc lowered the ee to 87% and that with NaOAc
lowered the ee to 72%. A change of the solvent from THF
to DCM, or a change from SAMP to another hydrazine, also
lowered the enantioselectivity.
Asymmetric allylic amination (N-allylation of benzylamine
with racemic allylic acetate) was also described under similar
conditions with SAMP-hydrazone 386a as the palladium
complex ligand.274
12.1.3. N,N-Dialkylhydrazones as Ligands for
Pd-Catalyzed Mizoroki-Heck, Sonogashira, and Hiyama
Reactions
Some N,N-dialkylhydrazones have been developed as air-
stable ligands for phosphine-free palladium catalysts for C-C
bond forming reactions such as Mizoroki-Heck, Sonogash-
ira, and Hiyama reactions (Scheme 72). Two types of
hydrazones 392, prepared from 2-pyridinecarboxaldehyde,
and bishydrazone 391, prepared from glyoxal and various
dialkyl hydrazines (DMH and differently sized cyclic hy-
drazines), were tested for the Heck palladium-catalyzed
Scheme 73
Lazny and Nodzewska
cross-coupling reaction of alkenes with aryl halides.279a
During optimization of the reaction conditions (different
ligands, 392 or 391; palladium sources, PdCl2(MeCN)2,
PdCl2, Pd(OAc)2; solvents, DMF, DMAc, DMSO; bases,
K3PO4, NaOAc, Cs2CO3 with or without tetrabutylammo-
nium bromide) the best conditions were found to be
PdCl2(MeCN)2, K3PO4, DMF, and 391d as the best hydra-
zone ligand. Reactions of conjugated olefins gave conjugated
vinyl aromatic products in good yields for aryl iodides
(73-99%) and aryl bromides (30-93%) as substrates.
Significantly worse performance was observed for aryl
chlorides (yields from trace to 69%). Hydrazones of type
391 (in particular, 391d) were also used as effective ligands
for coupling of allyl aryl ethers with aryl iodides.279b
The optimized conditions (DMF, K3PO4, 80 °C, 5 h) using
PdCl2(MeCN)2 with added CuI as a cocatalyst and the best
hydrazone ligand for the Heck reaction, i.e., 391d, were
applied for the Sonogashira cross-coupling reaction of
terminal alkynes with aryl halides. In the case of the
Sonogashira reaction, however, the bishydrazone ligand with
a seven-membered ring (391d) gave the arylated alkyne280
in low yield (37%). A very good yield (93%) was obtained
with the aryl alkyl hydrazone ligand derived from N-phenyl-
N-methylhydrazine and 2-pyridinecarboxaldehyde.
Analogous results were reported for the Hiyama cross-
coupling reaction of aryltrialkoxysilanes with aryl bromides.
Use of the bishydrazone ligand 391d (dioxane, Pd(OAc)2,
TBAF, 80 °C, 20 h) gave the biphenyl280 with a moderate
yield of 49%. Again, application of an aryl alkyl hydrazone
ligand improved the yield to 83%.
12.1.4. N,N-Dialkylhydrazones as Ligands for
Pd-Catalyzed Asymmetric Suzuki-Miyaura
Cross-Coupling
Very recently, the C2-symmetric bishydrazone of glyoxal
and chiral (S,S)-N-amino-2,5-diphenylpyrrolidine 395 (Scheme
73) was very successfully used as a ligand for phosphine-
free, asymmetric Suzuki-Miyaura cross-couplings. The
catalyzed reactions afforded a variety of enantiomerically
enriched biaryls with different substitution patterns. The
catalytically active complex [PdCl2(bishydrazone)] allowed
for room temperature reactions (typical catalyst loading 5
mol %, Cs2CO3, toluene, at 20 °C, 7 days, 36-98% yield)
and showed excellent enantioselectivities (ee 77% to >98%)
in all tested cases.281 At higher temperature (80 °C), the
reactions required shorter times (from 3.5 to 16 h) and the
yields were slightly higher (66-99%); however, this was at
the expense of the enantioselectivity (70-90% ee).
N,N-Dialkylhydrazones in Organic Synthesis Chemical Reviews, 2010, Vol. 110, No. 3 1421

Scheme 74 Scheme 75

Scheme 76

12.1.5. N,N-Dialkylhydrazones as Ligands for

Pd-Catalyzed Allyl Cross-Coupling Reactions and
Cu-Catalyzed N-Arylation of Amides and Azoles
N,N-Dialkylhydrazones were also used as phosphine-free
ligands in the palladium-catalyzed cross-coupling reaction
of allylic acetates with a range of boronic acids. The room
temperature reaction required Pd(OAc)2 as a palladium
source (2 mol % Pd(OAc)2, 2 mol % of the optimal ligand
391c, K2CO3, DMF/H2O) and gave the allylbenzene deriva-
tives 399 in good yields of 52-96% (Scheme 74).282
Hydrazones 391a and 391d were also used in related,
copper-catalyzed transformations, the C-N bond forming
Ullmann-type N-arylation of azoles and Goldberg-type
N-arylation of amides with aryl iodides 400 (Scheme 74).283 12.1.7. Lewis Acid-Catalyzed Addition of Formaldehyde
In both cases, the best yields of N-arylated products 402 were N,N-Dialkylhydrazones to Hydroxy Enones
observed for hydrazone ligand 391d. Optimal conditions for
N-arylation of azoles (5 mol % CuI, 10 mol % 391d, Cs2CO3, Two classes of ligands for metal-containing Lewis acid
DMF, 110 °C, 24 h) resulted in yields of up to 98%, and catalysts, BINOL-derived ligands and bis(oxazoline)s such
those for N-arylation of amides (5 mol % CuI, 10 mol % as PhBOX, PyBOX, t-BuBOX, were tested in formaldehyde
391d, K3PO4, DMSO, 110 °C, 5 h) resulted in yields of hydrazone additions to hydroxy enones 408 and related
25-92%. compounds (Scheme 76).285 From several tested acyclic and
cyclic N,N-dialkylhydrazone substrates, the best was the
12.1.6. N,N-Dialkylhydrazones as Ligands in Lewis N-aminopyrrolidine-derived cyclic hydrazone of formalde-
hyde (409a). The other hydrazones, acyclic (409c-e) and
Acid-Catalyzed Diels-Alder Reaction
piperidine-derived (409b), were poor substrates in these
A new class of effective Lewis acid metal catalysts with addition reactions. From the tested Lewis acids (triflates,
monohydrazone and bishydrazone ligands was developed for halides, hexafluoroantimonates, and tetrafluoroboranes of Mg,
enantioselective, metal-catalyzed, Diels-Alder reactions Zn, and Cu), the zinc triflate showed the most promising
(Scheme 75).284 The C2-symmetrical hydrazone ligands 405a activity. The optimal catalyst was found to be the complex
and 405b have been synthesized. Use of the Lewis acid of bis(oxazoline) t-BuBOX (410) with Zn(OTf)2. Optimal
Cu(OTf)2 on the model reaction of N-acryloyloxazolidinone conditions for N-aminohydrazone addition to aliphatic R-hy-
with cyclopentadiene revealed the best bis(2,5-diphenyl droxy enones were as follows: 10 mol % Zn(OTf)2, 11 mol
hydrazone) ligand 405a (90% yield of the (R,R)-endo-product % 410, toluene, 5 or 10 °C, 17-168 h. Depending on the
with 96% de and 95% ee). The cycloaddition of N- hydroxy enone 408 used, the yields ranged from 50% to 95%
and enantioselectivities from 66% to 84% (er 83:17 to 92:
acryloyloxazolidinone to different dienes (80-91% yield),
8). Cleavage of the resulting hydrazones with MMPP could
in the presence of a 1:1 complex of Cu(OTf)2 with 405a,
provide a route to the corresponding nitriles.
gave the (R,R)-endo-adduct of furan with the best ee of 96%
and diastereoselectivity of 74% de (88% yield). The least
enantioselective reaction (ee 84%) was observed for cyclo-
12.2. Organocatalytic Reactions (Brønsted Acid
hexadiene, albeit with high (R,R)-endo-selectivity (>99% de). Catalysis) of N,N-Dialkylhydrazones
The other mono- and bishydrazone ligands 406 and 407 gave Recent dynamic developments in the field of organoca-
low ee’s (0-18%) with the following Lewis acids: MgI2, talysis are also relevant to the chemistry of N,N-dialkylhy-
EtMgBr, Sc(OTf)3, and Cu(OTf)2. drazones. One of the approaches to catalysis with small
1422 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 77
organic molecules uses activation of reactants by organic
Brønsted acids and hydrogen bonding. This approach found
implementation in aldehyde hydrazone addition to imines286
and Michael acceptors such as β,γ-unsaturated R-keto
esters287 and nitroalkenes.288
12.2.1. Addition of Aldehyde Hydrazone to Imines
The first organo-catalyzed addition of formaldehyde hy-
drazone to N-protected (N-SO2R or N-CO2R) imines286 in
the presence of BINOL and binaphthol-derived bisalcohol
BIMBOL 415 (Scheme 77) was reported in 2005. Although
the uncatalyzed reaction of benzaldehyde N-protected imines
with formaldehyde N-aminopyrrolidine hydrazone 412a took
place with 0-15% conversion in 22 h, the addition of chiral
BINOL increased the conversions to 30% for the N-Boc
derivative, to 66% for the N-CO2Et derivative, and to 71%
for the N-tosylimine derivative. The reactions, however,
showed no enantioselectivity whatsoever. The application of
a novel class of (S)-BINOL-derived, 3,3′-bismethanol-2,2′-
bisnaphthol catalysts resulted in a moderately enantioselective
(3-67% ee, 4-48% conversion) addition reaction for
benzaldehyde N-Boc-imine. The optimal catalyst structure
for this reaction was 415286 (specifically 415a, R1 ) Ph).
Other aromatic N-Boc aldehyde imines gave, with the
optimized catalyst, products in 44-87% yields and with ee’s
from 17% to 75% (best ee 75% for 2-methylbenzaldehyde
in CDCl3). Subsequent MMPP hydrazone cleavage was
shown to proceed without racemization to give the corre-
sponding N-Boc amino nitriles.
Other Brønsted acids, synthesized and tried as catalysts
in the same formaldehyde addition to N-protected imines,
were chiral phosphoric acid derivatives of BINOL289 (416
and 417, Scheme 77). The N-aminopyrrolidine-derived
hydrazone and N-Boc-imine were again found to be the best
substrates in terms of achievable enantioselectivity. The best
enantioselectivities of 74-90% ee (48-82% yields) were
observed for the catalyst 417a (90% ee in the case of
piperonal aldimine). The best yield was obtained for the
benzaldehyde aldimine reaction catalyzed by 416a (92%
yield and 61% ee).
Recently, the application of axially chiral, dicarboxylic
acid derivatives of BINOL (418)290 as effective catalysts in
addition reactions of both formaldehyde and aromatic alde-
Lazny and Nodzewska
hyde hydrazones was reported (Scheme 77). The N-ami-
nopyrrolidine-derived N,N-dialkylhydrazones of formalde-
hyde (412a) reacted with aromatic N-Boc-aldimines 413 in
chloroform with good to excellent enantioselectivities
(89-96% ee, 70-89% yields). The highest enantiomeric
excess of 96% was obtained in the reaction of benzaldehyde
imine. The work showed the first successful application of
arylaldehyde hydrazones in the addition to imines. The
dicarboxylic acid-catalyzed (5 mol % (R)-418, -20 °C, 96 h,
MS 4A, CHCl3) reactions of arylaldehyde tetramethylene-
hydrazones (N-aminopyrrolidine-derived 412b) with aromatic
N-Boc-aldimines provided R-aminohydrazones in 35-77%
yields and with 84-95% ee. Ozonolysis of the products gave
the corresponding R-amino ketones in good yields.
12.2.2. Addition of Formaldehyde Hydrazone to
β,γ-Unsaturated Keto Esters
In analogy to the above-discussed addition of formalde-
hyde hydrazone nucleophiles to electrophilic acceptors
catalyzed with metal complexes,285 addition to Michael
acceptors can be catalyzed with thiourea-derived organo-
catalysts 421-425 (Scheme 78).287 Catalysis of the addition
of the formaldehyde hydrazone 412a to β,γ-unsaturated
R-keto esters with potential catalysts such as BINOL, BINOL
phosphate, mandelic acid, and ureas 421-425 was recently
tested. The only catalysts showing enantioselectivity were
thioureas 421 and 424a and urea 424c (28-32% ee) with
424a selected as the best performer. The optimal conditions
for this catalyst (10 mol % catalyst loading, DCM, low
temperature of -60 °C) gave the addition product for the
keto ester (R ) Me) with 80% ee. In the reactions with other
unsaturated keto esters the catalyst 424a gave yields from
60% to 80% and ee’s of 58-80%. As before oxidative
cleavage (MMPP or ozonolysis) of the hydrazones could be
used to gain access to nitriles and esters.287
12.2.3. Addition of Aldehyde Hydrazone to
R,β-Unsaturated Nitroalkenes
Strictly related to the aforementioned applications of
organocatalysts is the addition of different aldehyde hydra-
zones to electrophilic nitroalkenes.291 The conjugate addition
of formaldehyde or aliphatic aldehyde hydrazones to β-ni-
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 78
Scheme 79
trostyrene or other nitroalkenes catalyzed by thiourea 428288
(catalyst loading 20 mol %, DCM, 24 h) gave γ-nitro
hydrazones (Scheme 79), as a mixture of diastereoisomers,
in good yields (58-90%). The uncatalyzed reaction was also
observed, albeit with lower conversion at the same reaction
time (yield 61% in 18 h vs catalyzed yield 90%). Interest-
ingly, the use of ionic liquids (BMImBF4 or BMImPF6) as
the solvent for the uncatalyzed reaction accelerated the
completion of the conversion, but gave products in low
isolated yields, and was not feasible with the use of the
thiourea catalyst 428.
12.3. Lewis Acid Catalysis
The indium trichloride-catalyzed (10 mol %) reaction
between aromatic imines and the N,N-dimethylhydrazone of
methacrolein afforded 1,2,3,4-tetrahydroquinolines bearing
Figure 9. 1,2,3,4-Tetrahydroquinoline products of the reaction of
imines with N,N-dimethylhydrazones.
Chemical Reviews, 2010, Vol. 110, No. 3 1423
a hydrazone function at C4. The one-pot process involving
diastereoselective formation of two stereocenters, one of them
quaternary, gave the products 430 (Figure 9) in good to
excellent yields (70-93%). According to the authors, the
process illustrated the first example of an R,β-unsaturated
N,N-dimethylhydrazone behaving not as a diene, but as a
dienophile in a hetero-Diels-Alder reaction.292
The N-amino-2,6-diphenylpiperidine moiety was used as
a chiral hydrazone auxiliary in a scandium triflate-catalyzed
Mannich-type addition of aldehyde hydrazones 431 to ketene
silyl acetal (ester silyl enol ether, 432) or thioacetals (thioester
silyl enol ether, Scheme 80). The reaction proceeded in
aqueous THF (THF/water, 9:1), giving the expected adducts
in high yields (88-98%) and with good to high diastereo-
selectivities (dr from 4:1 to 99:1). The N-N bond cleavage
in 433 removed the chiral auxiliary, affording enantiomeri-
cally pure β-amino esters.293
13. N,N-Dialkylhydrazones in Solid-Phase Organic
Synthesis
Solid-phase synthesis of small molecules has been devel-
oping rapidly for over a decade. It is one of the workhorses
used in automated synthesis and combinatorial chemistry.294,295
The reversibility of hydrazone formation and the known
reactivity3 of N,N-dialkylhydrazones make the hydrazones
attractive reagents for multifunctional linkers296,297 and for
construction of varied carbon frameworks of small molecules.
Methodologies which allow for construction of C-C bonds
on the solid phase are especially important from the synthetic
point of view.298
Scheme 80
1424 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 81
Scheme 82
Hydrazones serving as linkers can in principle be used
either for reversible binding of carbonyl compounds to solid
supports functionalized with hydrazine groups or for binding
of hydrazines to supports exhibiting ketone or aldehyde
carbonyls. It is noteworthy that all hydrazones used to date
on a solid support functioned as linkers for carbonyl
compounds but also took part in the supported reactions.296
The polymers with N,N-dialkylhydrazine functional groups
that could be used for binding carbonyl compounds as
hydrazones are not available commercially, and therefore,
synthesis of polymer-supported hydrazines is a precondition.
The supported N,N-dialkylhydrazines have been prepared on
Merrifield-type polymers by typical reduction of previously
produced N-nitrosoamines,299-304 anchoring protected N,N-
dialkylhydrazines,62,305-307 or the direct functionalization of
the polymer with N-alkylhydrazine.46 The last method is the
simplest but may be limited by poor regioselectivity of
monalkyl hydrazine alkylation. Nonetheless, the direct func-
tionalization of Merrifield polymer with N-methylhydrazine
was used to prepare a polymer-supported reagent for the
synthesis of nitriles from aldehydes.46 Despite many synthetic
applications in solution chemistry, N,N-dialkylhydrazones
were applied so far only in four groups of C-C bond forming
reactions on a solid support, 1,2-addition, R-alkylation,
cycloaddition combined with allylboration, and in a very
limited way (one example) reaction on the azomethine carbon
with an aldehyde.
13.1. 1,2-Addition of Organolithiums to
Polymer-Supported Hydrazones
Polymer-supported hydrazones could be used in the same
way as hydrazones in solution for the synthesis of amines
by 1,2-addition of organometallic reagents. For the imple-
mentation of the idea a synthesis of polymer-supported
hydrazines was crucial.
Polymer-supported hydrazine 435 was synthesized on
Merrifield polymer through reaction of n-butylamine with
(chloromethyl)polystyrene, nitrosation of the resulting
secondary amine 434 with tert-butyl nitrite, and reduction
Lazny and Nodzewska
with diisobutylaluminum hydride solution (Scheme 81).299
The hydrazine 435 was subsequently reacted with several
aldehydes. The resulting hydrazones 436 (i.e., aldehydes
immobilized via the hydrazone linker) underwent the known-
from-solution synthesis, 1,2-nucleophilic addition of org-
anolithium reagents to give the corresponding supported
hydrazines 437 (Scheme 81). The reductive cleavage of the
hydrazine N-N bonds with borane/THF complex released
secondary amines. To facilitate isolation, the amines were
transformed to the amides 438.
Solid-phase asymmetric synthesis (SPAS) is a recently
introduced concept.308 The chiral analogue of the described
supported hydrazine and the analogous, diastereoselective
1,2-addition was used for the synthesis of enantiomerically
enriched chiral secondary amines.300 Two polymer-supported
chiral hydrazines, a SAMP-hydrazine analogue (441) and a,
so-called, RAML analogue (444), were synthesized from
readily available chiral building blocks, trans-4-hydroxy-L-
proline and N,N-dibenzyl-L-leucinol, respectively (Scheme
82). The supported chiral hydrazines 441 and 444 reacted
with 3-phenylpropionaldehyde or p-methylbenzaldehyde, and
the resulting hydrazones 445 and 447 acted as acceptors in
the nucleophilic 1,2-addition of n-BuLi, t-BuLi, n-HexLi,
and PhLi (Scheme 83). Subsequent application of the
previously described cleavage provided primary amines
protected as amides 449 in yields from 24% to 53% and
moderate to high ee’s (50-86%).
13.2. r-Alkylation of Polymer-Supported
Hydrazones
As in the case of the above-presented 1,2-addition, the
R-alkylation of solid-phase-supported achiral and chiral
hydrazones derived from immobilized hydrazines was in-
vestigated. Two approaches to the synthesis of polymer-
supported hydrazines, which bind both aldehydes and ketones
via hydrazones and allow for their alkylation, were used.
One strategy that allowed for the first hydrazone alkylation
on a solid phase303 was based on the preparation of supported
N-nitrosoamines and their reduction to N,N-dialkylhydra-
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 83
Scheme 84
zines,304 while the other way was based on the use of
hydrazones as masking groups for N,N-dialkylhydrazine
during anchoring on Merrifield-type polymers (Scheme 84).62
Thus, supported N,N-dialkylhydrazines 456 were prepared
on Merrifield-type polymers either by loading of nitrosoam-
ines and reduction of the N-nitrosoamines on the polymer
or by loading of 4-tert-butylcyclohexanone hydrazones
(Scheme 84)62 acting as protected hydrazines and giving 452
or 455. The latter method furnished typically cleaner
polymeric materials. The hydrazone-protected hydrazines
could be stored for prolonged periods of time. The protected,
polymeric hydrazines needed activation by washing with
10% TFA solution in wet THF prior to loading of a new
ketone or aldehyde. The hydrazones of the immobilized
ketones and aldehydes could be lithiated with LDA and
alkylated with alkyl halides.62,303,304,309 The alkylated products
were isolated in good yields (47-94%) and typically in
acceptable purity (59-91%).62 This alkylation method al-
Scheme 85
Chemical Reviews, 2010, Vol. 110, No. 3 1425
lowed for the expansion of the carbon skeleton of carbonyl
compounds. The availability of various methods for the
cleavage of N,N-dialkylhydrazones provided the opportunity
for multifunctional cleavage of the hydrazone linkers.
Depending on the cleavage method, different final products
could be isolated: (i) aldehydes and ketones under acidic
conditions (TFA/THF/water), (ii) acids or nitriles (from
aldehyde hydrazones) with oxidative workup (peroxide) or
oxidative cleavage (m-CPBA), (iii) alcohols or amines with
reductive workup (NaBH4) or reductive cleavage with a
borane/THF complex (Scheme 85).62 It has also been shown
that aldehydes (and presumably ketones) immobilized via
N-monoalkylhydrazone could be released from the polymeric
support in the form of N,N-dialkylhydrazone via a hydrazone
exchange reaction, i.e., treatment with a solution of N,N-
dialkylhydrazine in the presence of concentrated hydrochloric
acid (1-amino-4-methylpiperazine, 37% aqueous HCl/THF,
1:100).310
Asymmetric alkylation of ketones or aldehydes is waiting
to be widely implemented in SPAS. Proof of concept studies
on the asymmetric (enantioselective) alkylation of ketones
via supported chiral hydrazone auxiliaries have already been
conducted independently by some research groups (Figure
10). The polymer-supported SAMP analogue derived from
trans-4-hydroxy-L-proline 441, previously described as an
auxiliary for 1,2-addition301 when used for hydrazone alky-
lation, gave alkylated ketones and aldehydes with enantio-
meric excesses in the range of 55-79% at -78 °C (Scheme
86). Improvement of the stereoselectivity was observed by
lowering the reaction temperature to -100 °C (alkylation
of cyclohexanone with n-HexI and 3-pentanone with n-PrI,
giving 464c, R3 ) n-Hex, 86% ee, and 464a, R3 ) n-Pr,
85% ee). This was still lower than that achievable for the
same alkylations with SAMP-hydrazone in solution (91%
and 99% ee, respectively). Using a different anchoring site
1426 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 10. Polymer-supported chiral hydrazones utilized in
alkylation reactions.
for the SAMP auxiliary, i.e., SAMP analogue anchored to a
solid support via an oxymethylene group (460 and 461),306,307
resulted in enantioselectivities ranging from 10% to 73%.
The best performance was shown by alkylation of 3-pen-
tanone with n-propyl iodide (at -78 °C, product 464a, R3
) n-Pr, 73% ee) and the chiral auxiliary connected to the
polymer matrix through a six-carbon atom spacer (Scheme
87). Interestingly, somewhat increased enantiomeric excesses
of alkylation products (by 25-47%) were obtained for the
alkylation of a Cs-symmetrical ketone, anchored via the chiral
SAMP analogue hydrazone, when a chiral bidentate lithium
amide was used for lithiation of the polymeric hydrazone.307
In an independent study, the polymer-supported hydra-
zones 460 and 462, analogues of known-in-solution SAMP-
and RAMBO-hydrazones (Figure 10), immobilized via the
oxymethylene bridge, were also used for immobilization of
Scheme 86
Scheme 87
Lazny and Nodzewska
aldehydes (PhCH2CH2CHO, p-methoxybenzaldehyde) and
ketones (3-pentanone, cyclohexanone).302 The ketone hy-
drazones were alkylated with MeI or PrI and cleaved with
ozonolysis or acidolytically with TFA. Generally, overall
enantioselectivities of the process ranged from 58% to 66%
for ozonolysis and from 54% to 73% for TFA cleavage. The
best enantioselectivity (73%) was obtained for the alkylation
of 3-pentanone with PrI and the TFA cleavage.302
The investigations on hydrazone alkylation by three
different groups, and with hydrazone auxiliaries anchored
via an oxymethylene group or oxygen atom on C4, reached
similar levels of stereoselection. Overall, the performance
of the asymmetric alkylation on the studied solid supports
in terms of enantioselectivities was moderate. The low to
moderate enantioselectivities could be a result of a detri-
mental effect of the polymer matrix on this type of reaction.
13.3. Other Polymer-Supported Reactions
13.3.1. Polymer-Supported Reactions of Azomethine
Carbon Nucleophiles
To date, there is only one example of the reaction of
polymer-bound formaldehyde hydrazone at the azomethine
carbon. A polymer-supported, chiral N-aminopyrrolidine-
derived hydrazine was prepared through anchoring (KH,
THF) of the hydroxy group of N-trityl-protected (3R)-1-
amino-3-pyrrolidinol (synthesized in solution from com-
mercial (3R)-3-pyrrolidinol by N-nitrosation and LiAlH4
reduction of the resulting N-nitrosoamine) on the Merrifield
polymer. The immobilized formaldehyde hydrazone 471 was
obtained by the reaction of the free hydrazine polymer with
N,N-Dialkylhydrazones in Organic Synthesis Chemical Reviews, 2010, Vol. 110, No. 3 1427

Scheme 88

Scheme 89

an excess of formalin in THF or paraformaldehyde in THF/ nate N,N-dimethylhydrazone, containing protected aldehyde
methanol. The addition of the nucleophilic azomethine groups were used as Horner-Wadsworth-Emmons reagents
carbon of the N-aminopyrrolidine-derived hydrazone 471 to in a reaction with (2E)-2-methyl-2-butenal. The reaction was
Fmoc-phenylalaninal 472 was attempted and gave after a key step in the synthesis of (2E,4E,6E,8E)-3,5-dimethyl-
hydroxyl protection and ozonolytic cleavage a mixture of 7-ethyl-2,4,6,8-undecatetraene, a pheromone of the beetle
diastereomeric products 475 (Scheme 88).305 Carpophilus lugubris. Removal of the hydrazone protective
group was effected with a biphasic mixture of dilute HCl
13.3.2. Polymer-Supported Multicomponent Reactions and petroleum ether.312
N,N-Dimethylhydrazone was also used as a protecting
The previously discussed (vide cyclizations), multicom- group for aldehydes (hydrazone 486, 487) in the synthesis
ponent aza [4 + 2] cycloaddition/allylboration reaction of furyl-1,4-naphthoquinones and furyl-1,4-benzoquinones.313
between 1-aza-4-boronobutadienes, maleimides, and alde- Ketone (cyclohexanone, tert-butylcyclohexanone) hydra-
hydes was also performed on a solid support (Scheme 89). zones 488 served as protection for the hydrazine group during
The supported tandem reaction was realized through either
immobilization of maleimide on a Sasrin resin (476) or the
anchoring of m-hydroxybenzaldehyde via a silyloxy linker
(482). The loaded resins were reacted with other components,
giving the immobilized products, which, in turn, were cleaved
from the Sasrin supports by washing with 2% TFA in DCM
or from the silyloxy resin by treatment with HF · Py in THF.
The resulting products 480 and 484 were obtained in overall
yields of 50% and 75%, respectively.311

14. Hydrazones as Protecting Groups and

Miscellaneous Applications
N,N-Dimethylhydrazone was used as the protecting group
for an aldehyde in phosphonate reagents. Two phosphonates
485 (Figure 11), diethyl (ethylformyl)-2-phosphonate N,N- Figure 11. N,N-Dimethylhydrazone-protected aldehydes used in
dimethylhydrazone and diethyl (1-propylformyl)-2-phospho- synthesis.
1428 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 90
Scheme 91
O-alkylation of hydroxy hydrazines 451 with benzyl chloride
and its polymeric analogue, the Merrifield resin (Scheme
90).62
A highly enantiomerically enriched [6,7]-cis-fused system,
493, was obtained by addition of vinyllithium to the RAMP-
hydrazone of (η6-arene)tricarbonylchromium(0) complex 490
(Scheme 91) followed by acetylation (MeI, CO), methylation
with methyl iodide, and allylation.314
The N,N-dialkylhydrazone group served simultaneously
as a heteroaryl activating group and a masked formyl and
nitrile group. Hydroxyalkylation of furfural and pyrrole-2-
carbaldehyde N,N-dimethylhydrazones with or without con-
comitant isomerization gave the heteroaryl analogues of
R-benzoins 495 or 496 in 10-92% yields (Scheme 92).315,316
Subsequent N-methylation followed by acidic (giving alde-
hyde) or aqueous (giving nitrile) hydrolysis provided 497
and 498.
Chlorodifluoroacetylated N,N-dialkylhydrazones 499 re-
acted with aromatic aldehydes, ethyl pyruvate, and benzal-
dehyde N-tosylimine in a reductive coupling mediated by
tetrakis(dimethylamino)ethylene (TDAE; Scheme 93). The
method gave access to a series of new, R-diketone-derived,
gem-difluorinated monohydrazone derivatives 501.317
Scheme 92
Lazny and Nodzewska
Figure 12. Racemic aldehydes and ketones resolved via diaster-
eomeric SAMP/RAMP-hydrazones.
Scheme 93
Resolution of racemic aldehydes or ketones 502,318 503,319
504,320 and 505321 (Figure 12), through formation of a mixture
of diastereomeric hydrazones and their separation by column
chromatography320,321 or crystallization,318,319 was another
preparative application of SAMP/RAMP-hydrazones. The
resolved diastereomeric hydrazones were typically hydro-
lyzed to optically enriched carbonyl compounds using an
aqueous oxalic acid/organic solvent (ether, hexane) two-phase
system.
N-Aziridinylimines (Eschenmoser hydrazones 506) are a
specific type of N,N-dialkylhydrazones that were used for
the generation of non-carbonyl-stabilized diazo compounds
(possibly of type 507). Thermal decomposition of the
Eschenmoser hydrazones gave a mixture of decomposition
products, presumably by a carbene pathway. The likely
products of the same reaction in the presence of rhodium(II)
catalyst were rhodium carbenoid intermediates.322 Combina-
tion of the process with stereoselective 1,2-hydride migration
and thermal Claisen rearrangement resulted in a highly
stereoselective tandem Rh-catalyzed Bamford-Stevens/
Claisen (or even Cope) reaction (508, Scheme 94, yields
62-88%, dr from 3:1 to >20:1).322
The reaction of N-(trans-2,3-diphenylaziridin-1-yl)imines
509 with alkyllithiums R2Li (Scheme 95) afforded the
products 510 (59-70% yield) via anionic addition followed
by fragmentation with the liberation of stilbene and nitrogen
gas. The reaction was also effective for the aziridines 509
N,N-Dialkylhydrazones in Organic Synthesis
Scheme 94
Scheme 95
Scheme 96
where R1 ) PhO(CH2)4 using organocuprates where R2 )
n-Bu (68%), Grignard reagent R2 ) n-Bu, allyl (59%, 84%),
and organolithiums n-BuLi (52%) and PhCC-Li (75%).323
The analogous reaction of monosubstituted Eschenmoser
hydrazones, N-(2-phenylaziridin-1-yl)imines, gave an anoma-
lous byproduct. The consecutive trapping reactions of the
putative anionic intermediates with benzaldehyde and an-
hydrides were probed with varying degrees of success. On
the other hand, anionic cyclizations of N-(trans-2,3-diphe-
nylaziridin-1-yl)imines using unactivated alkenes and alkynes
as acceptors proceeded smoothly, yielding various cyclized
products 513 in good yields (60-91%, Scheme 96) and
preference for the trans-isomers (X ) Me or -CHdCH2).323
15. Conclusions and Outlook
Since the 1970s, we have witnessed major developments
in both the synthetic methodology and the target-oriented
synthetic applications of N,N-dialkylhydrazones. Methods of
asymmetric synthesis and C-C bond formation have espe-
cially benefited from the blossoming of hydrazone chemistry
prior to 2000.1,3 Today, C-heteroatom (Si, Se, S, O, P, N,
F) bond formation and C-C bond formation using both
achiral and chiral hydrazone methods are well-known
synthetic transformations. In the past few years we have seen
a steady growth of applications of the known reactions of
N,N-dialkylhydrazones such as R-alkylation, 1,2-addition to
Chemical Reviews, 2010, Vol. 110, No. 3 1429
CdN bonds, and hetero-Diels-Alder reactions as key steps
in the syntheses of natural products and other complex
molecules.
Alkylation is the most prominently used transformation
of hydrazones (ca. 60 cited references report on R-alkylation
in solution and 6 on a solid phase out of ca. 190 published
after 2000). The second most popular N,N-dialkylhydrazone
reaction in the years covered by the review was the
nucleophilic 1,2-addition (ca. 35 cited papers) followed by
the hetero-Diels-Alder reactions of unsaturated hydrazones
(17 citations).
The 1,3-dihydroxyacetone equivalent 1,3-dioxanone97 is
a synthetically important building block, for which N,N-
dialkylhydrazone transformations are well suited (over 20
cited papers). The sophisticated buildup on 1,3-dioxanone
molecules by sequential R-alkylation, R,R′-alkylation, or even
triple R,R′,R-alkylation (quaternization) has been used as the
central process in multistep syntheses. The ultimate quater-
nization of both R-carbons of a 1,3-dioxanone was also
possible by the SAMP-hydrazone method. Diastereoselec-
tivities and enantioselectivities of chiral hydrazone reactions
utilized in the syntheses were typically excellent (90% to
>99%) or could be improved by separation of the diastere-
omeric hydrazone mixtures. Although several chiral hydra-
zines are available in both enantiomeric forms, the SAMP
and RAMP auxiliaries dominate the asymmetric transforma-
tions based on chiral hydrazones. Aside from the notable
case of titanated hydrazones, the aldol reaction and, similarly,
the acylation and Michael reactions of metalated hydrazones
have found little application in target-oriented synthesis and
relatively little attention in methodological studies.
On the other hand, the dynamically growing methodologi-
cal research in areas of N,N-dialkylhydrazones as catalysts
and catalyst ligands, organocatalytic reactions of the hydra-
zones, and electrophilic reactions of aldehyde hydrazones
(mostly formaldehyde) on the azomethine carbon (e.g.,
Michael reaction) have so far found little synthetic utility in
actual multistep syntheses of specific complex targets. It
seems that, contrary to R-alkylation and 1,2-addition, these
areas of hydrazone chemistry are not yet mature. The
majority of the reported reactions at the azomethine carbon
involved formaldehyde hydrazone substrates. In the modern
area of catalysis, N,N-dialkylhydrazones (ca. 25 references
on catalysis) have been used as catalysts for asymmetric
Et2Zn addition, as ligands for Pd catalysts (allylation, Heck-,
Sonogashira-, and Hiyama-type reactions), and in Lewis acid-
catalyzed Diels-Alder reactions. Although relatively less
developed, the free radical chemistry of N,N-dimethylhy-
drazones, N,N-dibenzylhydrazones, SAMP-hydrazones, and
N-aziridinylimines (Eschenmoser hydrazones) have also
found applications in synthesis. The radical transformations
were pioneered by Sunggak Kim and more recently devel-
oped by Gregory K. Friestad. Organocatalytic reactions of
N,N-dialkylhydrazones are confined to several notable ex-
amples (BA organocatalyzed) of hydrazone substrates react-
1430 Chemical Reviews, 2010, Vol. 110, No. 3 Lazny and Nodzewska

ing with electrophiles such as imines, R,β-unsaturated keto HAD hetero-Diels-Alder

esters, and nitroalkenes. In addition, Lewis acid-catalyzed LA Lewis acid
reactions of hydrazones have also been reported. The MMPP magnesium monoperoxyphthalate
recyclability of hydrazone-based catalysts (especially chiral) NFSI N-fluorobenzenesulfonimide
PPL porcine pancreatic lipase
has, so far, not been addressed. RAMBO (R,R,R)-2-amino-3-(methoxymethyl)-2-azabicyclo-
Applications of hydrazone chemistry in supported syn- [3.3.0]octane
thesis are also limited to reports of methodological studies RAML (R)-N-amino-O-methylleucinol
and to the classical reactions (R-alkylation, 1,2-addition). The RAMP (R)-1-amino-2-(methoxymethyl)pyrrolidine
hydrazone units fulfill a dual role in supported syntheses, SADP (S)-1-amino-2-(1-methoxy-1-methylethyl)pyrroli-
the linker and the reacting functional group. The N,N- dine
dialkylhydrazone linkers are well suited for, and already have SAEP (S)-1-amino-2-(1-methoxy-1-ethylpropyl)pyrroli-
been used as, multifunctional linkers. There are no reports dine
on reactions of solid-phase-supported hydrazones not acting SAMP (S)-1-amino-2-(methoxymethyl)pyrrolidine
SAPP (S)-1-amino-2-(methoxydiphenylmethyl)pyrroli-
as linkers, i.e., reactions of hydrazone molecules anchored
dine
via other functionalities. SEM-Cl [2-(trimethylsilyl)ethoxy]methyl chloride
The majority of the hydrazone methodology and total SPAS solid-phase asymmetric synthesis
syntheses using N,N-dialkylhydrazone chemistry comes from TBAI tetrabutylammonium iodide
the research group of Dieter Enders of RWTH-Aachen (ca. TBDMS tert-butyldimethylsilyl
120 references in this review). Other groups leading in TBDPS tert-butyldiphenylsilyl
specific areas are the teams of José M. Lassaletta (24 TDAE tetrakis(dimethylamino)ethylene
references, mostly reactions on azomethine carbon and
catalysis) and Takashi Mino (14 references, mostly catalysis 17. Acknowledgments
with hydrazones).
It is rather hard to judge whether the current state of We thank Professor Dieter Enders, Professor Marek
hydrazone chemistry is beginning to show the limits of Majewski, and Dr. D. Mark Gleave for helpful suggestions
synthetic hydrazone applicability or, on the contrary, whether during the preparation of the manuscript and the University
we can expect to see a dynamic expansion of the frontiers of Bialystok for financial support (BST 125).
of hydrazone chemistry, to the benefit of preparative
technologies. Clearly, some issues of hydrazone methodolo- 18. References
gies may still need improvement: Chiral auxiliary recycling (1) Job, A.; Janeck, C. F.; Bettray, W.; Peters, R.; Enders, D. Tetrahedron
and hydrazone catalyst recovery may be addressed by solid- 2002, 58, 2253–2329.
phase immobilization approaches. However, it seems that (2) Enders, D. In Asymmetric Synthesis; Morrison, J. D., Ed.; Academic
better, less stereoselectivity attenuating, polymeric supports Press: Orlando, FL, 1984; Vol. 3.
(3) Kim, S.; Yoon, J.-Y. In Science of Synthesis; Padwa, A., Ed.; Georg
will be needed before truly high levels of enantioselectivity Thieme: Stuttgart, Germany, New York, 2004; Vol. 27.
with polymer-supported, recyclable chiral hydrazine auxil- (4) Bergbreiter, D. E.; Momongan, M. In ComprehensiVe Organic
iaries can be achieved. So far, the SAMP analogues anchored Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon Press: Oxford,
U.K., 1991; Vol. 2.
on (chloromethyl)polystyrene (Merrifield polymer) consis- (5) Sugiura, M.; Kobayashi, S. Angew. Chem., Int. Ed. 2005, 44, 5176–
tently provided products with ee’s up to 86%. 5186.
The accelerated progress in synthetic methodology and (6) Friestad, G. K. Eur. J. Org. Chem. 2005, 3157–3172.
significant position of N,N-dialkylhydrazone methods in (7) Lim, D.; Coltart, D. M. Angew. Chem., Int. Ed. 2008, 47, 5207–
5210.
organic synthesis ensures that we will continue to see new (8) Enders, D.; Wortmann, L.; Peters, R. Acc. Chem. Res. 2000, 33, 157–
hydrazone methods, reactions, and catalysts and their novel 169.
and inventive application toward the syntheses of complex (9) Fernández, R.; Lassaletta, J. M. Synlett 2000, 1228–1240.
(10) Corey, E. J.; Enders, D. Tetrahedron Lett. 1976, 17, 3–6.
molecules in the coming years. (11) Corey, E. J.; Enders, D. Chem. Ber. 1978, 111, 1337–1361.
(12) Enders, D.; Schüβeler, T. New J. Chem. 2000, 24, 973–975.
(13) Sainsbury, M.; Mahon, M. F.; Williams, C. S. Tetrahedron 1991,
16. List of Abbreviations 47, 4195–4210.
(14) André, C.; Bolte, J.; Demuynck, C. Tetrahedron: Asymmetry 1998,
ADA aza-Diels-Alder 9, 3737–3739.
BA Brønsted acid (15) Enders, D.; Eichenauer, H. Angew. Chem., Int. Ed. Engl. 1979, 18,
BIMBOL 3,3′-bismethanol-2,2′-binaphthol 397–399.
(16) Bergbreiter, D. E.; Newcomb, M. Tetrahedron Lett. 1979, 4145–
BINOL 1,1′-bi-2-naphthol 4148.
BOM-Cl benzyloxymethyl chloride (17) Ludwig, J. W.; Newcomb, M.; Bergbreiter, D. E. J. Org. Chem. 1980,
BOX bis(oxazoline) 45, 4666–4669.
m-CPBA 3-chloroperoxybenzoic acid (18) Majewski, M. Tetrahedron Lett. 1988, 29, 4057–4060.
dap 2,6-diaminopyridine (19) Majewski, M.; Gleave, D. M. J. Organomet. Chem. 1994, 470, 1–
DBAD di-tert-butyl azodicarboxylate 16.
(20) Fallis, A. G.; Brinza, I. M. Tetrahedron 1997, 53, 17543–17594.
DHA dihydroxyacetone (21) Friestad, G. K. Tetrahedron 2001, 57, 5461–5496.
DHAP dihydroxyacetone phosphate (22) Felix, D.; Schreiber, J.; Piers, K.; Horn, U.; Eschenmoser, A. HelV.
DIPEA diisopropylethylamine Chim. Acta 1968, 51, 1461–1465.
DMAB dimethylamine-borane complex (23) Müller, R. K.; Joos, R.; Felix, D.; Schreiber, J.; Wintner, C.;
DMH N,N-dimethylhydrazine Eschenmoser, A. Organic Syntheses; Wiley: New York, 1988;
DMPU N,N′-(dimethylpropylene)urea Collect. Vol. VI, pp 56-61.
(24) Kirmse, W. Eur. J. Org. Chem. 1998, 201–212.
EWG electron-withdrawing group (25) Hashmi, A. S. K. J. Prakt. Chem. 1999, 341, 600–604.
FDMH formaldehyde N,N-dimethylhydrazone (26) Davenport, K. G.; Eichenauer, H.; Enders, D.; Newcomb, M.;
FSAMPH formaldehyde SAMP-hydrazone Bergbreiter, D. E. J. Am. Chem. Soc. 1979, 101, 5654–5659.
GABA γ-aminobutyric acid (27) Bauer, W.; Seebach, D. HelV. Chim. Acta 1984, 67, 1972–1988.
N,N-Dialkylhydrazones in Organic Synthesis Chemical Reviews, 2010, Vol. 110, No. 3 1431

(28) Enders, D.; Bachstädter, G.; Kremer, K. A. M.; Marsch, M.; Harms, (75) Sharma, S. D.; Pandhi, S. B. J. Org. Chem. 1990, 55, 2196–2200.
K.; Boche, G. Angew. Chem., Int. Ed. Engl. 1988, 27, 1522–1524. (76) Corey, E. J.; Enders, D. Tetrahedron Lett. 1976, 17, 11–14.
(29) Collum, D. B.; Kahne, D.; Gut, S. A.; DePue, R. T.; Mohamadi, F.; (77) Corey, E. J.; Enders, D. Chem. Ber. 1978, 111, 1362–1383.
Wanat, R. A.; Clardy, J.; Duyne, G. V. J. Am. Chem. Soc. 1984, (78) Corey, E. J.; Enders, D.; Bock, M. G. Tetrahedron Lett. 1976, 17,
106, 4865–4869. 7–10.
(30) Galiano-Roth, A. S.; Collum, D. B. J. Am. Chem. Soc. 1989, 111, (79) Corey, E. J.; Knapp, S. Tetrahedron Lett. 1976, 17, 4687–4690.
6772–6778. (80) Enders, D.; Weuster, P. Tetrahedron Lett. 1978, 19, 2853–2856.
(31) Ahlbrecht, H.; Düber, E. O.; Enders, D.; Eichenauer, H.; Weuster, (81) Enders, D.; Eichenauer, H. Angew. Chem., Int. Ed. Engl. 1976, 15,
P. Tetrahedron Lett. 1978, 19, 3691–3694. 549–551.
(32) Weber, T.; Edwards, J. P.; Denmark, S. E. Synlett 1989, 20–22. (82) Enders, D.; Eichenauer, H.; Baus, U.; Schubert, H.; Kremer, K. A. M.
(33) McGlacken, G. P.; Breeden, S. W. Tetrahedron: Asymmetry 2005, Tetrahedron 1984, 40, 1345–1359.
16, 3615–3618. (83) Enders, D.; Kipphardt, H.; Fey, P. Org. Synth. 1987, 65, 183–202.
(34) Evans, D. A.; Polniaszek, R. P.; DeVries, K. M.; Guinn, D. E.; (84) Fey, P. In StereoselectiVe Synthesis; Helmchen, G., Hoffmann, R. W.,
Mathre, D. J. J. Am. Chem. Soc. 1991, 113, 7613–7630. Mulzer, J., Schaumann, E., Eds.; Houben-Weyl, Methods of Organic
(35) Bildstein, B.; Denifl, P. Synthesis 1994, 158–160. Chemistry, Vol. E21a; Georg Thieme: Stuttgart, Germany, 1995.
(36) Enders, D.; Peters, R.; Lochtman, R.; Runsink, J. Synlett 1997, 1462– (85) Kim, S.; Kee, I. S.; Lee, S. J. Am. Chem. Soc. 1991, 113, 9882–
1464. 9883.
(37) Barrett, I. C.; Kerr, M. A. Synlett 2000, 1673–1675. (86) Lassaletta, J.-M.; Fernández, R. Tetrahedron Lett. 1992, 33, 3691–
(38) Gmouh, S.; Jamal-Eddine, J.; Valnot, J. Y. Tetrahedron 2000, 56, 3694.
8361–8366. (87) Fernández, R.; Gasch, C.; Lassaletta, J.-M.; Llera, J.-M. Tetrahedron
(39) Banerjee, S.; Shi, Y.; Cao, C.; Odom, A. L. J. Organomet. Chem. Lett. 1994, 35, 471–472.
2005, 690, 5066–5077. (88) Enders, D.; Syrig, R.; Raabe, G.; Fernández, R.; Gasch, C.; Lassaletta,
(40) Kirmse, W.; Krzossa, B.; Steenken, S. J. Am. Chem. Soc. 1996, 118, J.-M.; Llera, J.-M. Synthesis 1996, 48–52.
7473–7477. (89) Whitesell, J. K.; Whitesell, M. A. Synthesis 1983, 517–536.
(41) Greene, T. W.; Wuts, P. G. M. Green’s ProtectiVe Groups in Organic (90) Baus, U. Diploma, University at Giessen, Giessen, Germany, 1982.
Synthesis, 4th ed.; Wiley-Interscience: Hoboken, NJ, 2007. (91) Gawley, R. E.; Termine, E. J.; Aube, J. Tetrahedron Lett. 1980, 21,
(42) Enders, D.; Plant, A. Synlett 1990, 725–726. 3115–3118.
(43) Enders, D.; Plant, A.; Backhaus, D.; Reinhold, U. Tetrahedron 1995, (92) Enders, D.; Dyker, H.; Raabe, G.; Runsink, J. Synlett 1992, 901–
51, 10699–10714. 903.
(44) (a) Duraisamy, M.; Walborsky, H. M. J. Org. Chem. 1984, 49, 3410– (93) Nakamura, E.; Kubota, K. J. Org. Chem. 1997, 62, 792–793.
3411. (b) Smith III, A. B.; Liu, Z.; Simov, V. Synlett 2009, 3131– (94) Enders, D.; Geibel, G.; Osborne, S. Chem.sEur. J. 2000, 6, 1302–
3134. 1309.
(45) Erickson, R. E.; Andrulis Jr, P. J.; Collins, J. C.; Lungle, M. L.; (95) Enders, D.; Heider, K.-J.; Raabe, G. Angew. Chem., Int. Ed. Engl.
Mercer, G. D. J. Org. Chem. 1969, 34, 2961–2966. 1993, 32, 598–601.
(46) Baxendale, I. R.; Ley, S. V.; Sneddon, H. F. Synlett 2002, 775–777. (96) Nicolaou, K. C.; Sarabia, F.; Ninkovic, S.; Finlay, M. R. V.; Boddy,
(47) Enders, D.; Berg, T. Synlett 1996, 796–798. C. N. C. Angew. Chem., Int. Ed. 1998, 37, 81–84.
(48) Enders, D.; Eichenauer, H. Chem. Ber. 1979, 112, 2933–2960. (97) Enders, D.; Voith, M.; Lenzen, A. Angew. Chem., Int. Ed. 2005, 44,
(49) Enders, D.; Eichenauer, H. Tetrahedron Lett. 1977, 18, 191–194. 1304–1325.
(50) Enders, D.; Hundertmark, T.; Lazny, R. Synlett 1998, 721–722. (98) Enders, D.; Brauer-Scheib, S.; Fey, P. Synthesis 1985, 393–396.
(51) Corey, E. J.; Knapp, S. Tetrahedron Lett. 1976, 17, 3667–3668. (99) Lohray, B.; Enders, D. Synthesis 1993, 1092–1094.
(52) Enders, D.; Hundertmark, T.; Lazny, R. Synth. Commun. 1999, 29, (100) Enders, D.; Eichenauer, H.; Pieter, R. Chem. Ber. 1979, 112, 3703–
27–33. 3714.
(53) Mino, T.; Fukui, S.; Yamashita, M. J. Org. Chem. 1997, 62, 734– (101) Oliva, A.; Delgado, P. Synthesis 1986, 865–866.
735. (102) Enders, D.; Pathak, V. N.; Weuster, P. Chem. Ber. 1992, 125, 515–
(54) Ulven, T.; Carlsen, P. H. J. Eur. J. Org. Chem. 2000, 3971–3972. 524.
(55) Mino, T.; Matsuda, T.; Hiramatsu, D.; Yamashita, M. Tetrahedron (103) Mangelinckx, S.; Giubellina, N.; De Kimpe, N. Chem. ReV. 2004,
Lett. 2000, 41, 1461–1463. 104, 2353–2400.
(56) Muñoz, L.; Bosch, M. P.; Guerrero, A. Tetrahedron: Asymmetry 2007, (104) Zarbin, P. H. G.; de Oliveira, A. R. M.; Delay, C. E. Tetrahedron
18, 651–658. Lett. 2003, 44, 6849–6851.
(57) Kalia, J.; Raines, R. T. Angew. Chem., Int. Ed. 2008, 47, 7523– (105) Dias, L. C.; de Oliveira, L. G. Org. Lett. 2004, 6, 2587–2590.
7526. (106) Panek, J. S.; Jain, N. F. J. Org. Chem. 2001, 66, 2747–2756.
(58) Enders, D.; Schubert, H. Angew. Chem., Int. Ed. Engl. 1984, 23, (107) Tursun, A.; Canet, I.; Aboaba, B.; Sinibaldi, M.-E. Tetrahedron Lett.
365–366. 2005, 46, 2291–2294.
(59) Enders, D.; Tiebes, J.; De Kimpe, N.; Keppens, M.; Stevens, C.; (108) Tursun, A.; Aboab, B.; Martin, A.-S.; Sinibaldi, M.-E.; Canet, I.
Smagghe, G.; Betz, O. J. Org. Chem. 1993, 58, 4881–4884. Synlett 2005, 2397–2399.
(60) Casarini, M. E.; Ghelfi, F.; Libertini, E.; Pagnoni, U. M.; Parsons, (109) Nakamura, M.; Hatakeyama, T.; Hara, K.; Fukudome, H.; Nakamura,
A. F. Tetrahedron 2002, 58, 7925–7932. E. J. Am. Chem. Soc. 2004, 126, 14344–14345.
(61) Enders, D.; Lochtman, R.; Meiers, M.; Müller, S.; Lazny, R. Synlett (110) Hatakeyama, T.; Nakamura, M.; Nakamura, E. J. Am. Chem. Soc.
1998, 1182–1184. 2008, 130, 15688–15701.
(62) Lazny, R.; Nodzewska, A.; Sienkiewicz, M.; Wolosewicz, K. (111) Vicario, J. L.; Job, A.; Wolberg, M.; Müller, M.; Enders, D. Org.
J. Comb. Chem. 2005, 7, 109–116. Lett. 2002, 4, 1023–1026.
(63) Brady, O. L.; McHugh, G. P. J. Chem. Soc., Trans. 1922, 121, 1648– (112) Enders, D.; Vicario, J. L.; Job, A.; Wolberg, M.; Müller, M.
1652. Chem.sEur. J. 2002, 8, 4272–4284.
(64) Todd, D. J. Am. Chem. Soc. 1949, 71, 1353–1355. (113) Schneider, U.; Pannecoucke, X.; Quirion, J.-C. Synlett 2002, 1669–
(65) Walker, G. N.; Moore, M. A.; Weaver, B. N. J. Org. Chem. 1961, 1672.
26, 2740–2747. (114) Enders, D.; Janeck, C. F.; Runsink, J. Synlett 2000, 641–643.
(66) Smith, R. F.; Albright, J. A.; Waring, A. M. J. Org. Chem. 1966, (115) Enders, D.; Backes, M. Tetrahedron: Asymmetry 2004, 15, 1813–
31, 4100–4102. 1817.
(67) Horner, L.; Fernekess, H. Chem. Ber. 1961, 94, 712–724. (116) Enders, D.; Schüβeler, T. Tetrahedron Lett. 2002, 43, 3467–3470.
(68) Avaro, M.; Levisalles, J.; Rudler, H. J. Chem. Soc. D 1969, 445b– (117) Hazelard, D.; Fadel, A. Tetrahedron: Asymmetry 2005, 16, 2067–
446. 2070.
(69) Marxer, A.; Horvath, M. HelV. Chim. Acta 1964, 47, 1101–1113. (118) Smith III, A. B.; Davulcu, A. H.; Kürti, L. Org. Lett. 2006, 8, 1669–
(70) Mitteil, I.; Brehme, R.; Nikolajewski, H. E. Z. Chem. 1968, 8, 226– 1672.
227. (119) Enders, D.; Niemeier, O. Heterocycles 2005, 66, 385–403.
(71) Brehme, R.; Nikolajewski, H. E. Tetrahedron 1969, 25, 1159–1163. (120) Enders, D.; Wortmann, L.; Raabe, G.; Dücker, B. Heterocycles 2004,
(72) Serckx-Poncin, B.; Hesbain-Frisque, A.-M.; Ghosez, L. Tetrahedron 62, 559–581.
Lett. 1982, 23, 3261–3264. (121) Enders, D.; Wortmann, L. Heterocycles 2002, 58, 293–299.
(73) Dı́az-Guerra, L. M.; Ocaña, B.; Pérez, J. M.; Avendaño, C.; Espada, (122) Enders, D.; Lenzen, A.; Backes, M.; Janeck, C.; Catlin, K.; Lannou,
M.; Menéndez, J. C.; Ramos, M. T.; Ruiz, M. A.; Pingarrón, J. M.; M.-I.; Runsink, J.; Raabe, G. J. Org. Chem. 2005, 70, 10538–10551.
Salvatierra, D.; Jaime, C. Bull. Soc. Chim. Belg. 1995, 104, 683– (123) Fleischhauer, J.; Gabriel, S.; Job, A.; Enders, D.; Wollmer, A. Z.
690. Naturforsch. 2001, 56b, 1344–1348.
(74) Pautet, F.; Nebois, P.; Bouaziz, Z.; Fillion, H. Heterocycles 2001, (124) Job, A.; Nagelsdiek, R.; Enders, D. Collect. Czech. Chem. Commun.
54, 1095–1138. 2000, 65, 524–538.
1432 Chemical Reviews, 2010, Vol. 110, No. 3
(125) Enders, D.; Klumpen, T. J. Organomet. Chem. 2001, 637-639, 698–
709.
(126) Enders, D.; Klumpen, T.; Raabe, G. Synlett 2003, 1198–1200.
(127) Enders, D.; Jonas, E. A.; Klumpen, T. Eur. J. Org. Chem. 2009,
2149–2162.
(128) Palacios, F.; Aparicio, D.; de los Santos, J. M.; Vicario, J. Tetrahedron
2001, 57, 1961–1972.
(129) Enders, D.; Narine, A. A. J. Org. Chem. 2008, 73, 7857–7870.
(130) Enders, D.; Voith, M.; Ince, S. J. Synthesis 2002, 1775–1779.
(131) Enders, D.; Lenzen, A.; Müller, M. Synthesis 2004, 1486–1496.
(132) Enders, D.; Herriger, C. Eur. J. Org. Chem. 2007, 1085–1090.
(133) Enders, D.; Voith, M. Synthesis 2002, 1571–1577.
(134) Enders, D.; Peiffer, E.; Raabe, G. Synthesis 2007, 1021–1026.
(135) Enders, D.; Whitehouse, D. L.; Runsink, J. Chem.sEur. J. 1995, 1,
382–388.
(136) Enders, D.; Müller-Hüwen, A. Eur. J. Org. Chem. 2004, 1732–1739.
(137) Enders, D.; Kirchhoff, J. H. Synthesis 2000, 2099–2105.
(138) Enders, D.; Barbion, J. Chem.sEur. J. 2008, 14, 2842–2849.
(139) Enders, D.; Hieronymi, A.; Ridder, A. Synlett 2005, 2391–2393.
(140) Enders, D.; Hieronymi, A.; Raabe, G. Synthesis 2008, 1545–1558.
(141) Enders, D.; Voith, M. Synlett 2002, 29–32.
(142) Enders, D.; Lenzen, A. Synlett 2003, 2185–2187.
(143) Enders, D.; Haas, M. Synlett 2003, 2182–2184.
(144) Ulven, T.; Carlsen, P. H. J. Eur. J. Org. Chem. 2001, 3367–3374.
(145) Nicolaou, K. C.; Brenzovich, W. E.; Bulger, P. G.; Francis, T. M.
Org. Biomol. Chem. 2006, 4, 2119–2157.
(146) Nicolaou, K. C.; Bulger, P. G.; Brenzovich, W. E. Org. Biomol. Chem.
2006, 4, 2158–2183.
(147) Enders, D.; Nühring, A.; Runsink, J.; Raabe, G. Synthesis 2001, 1406–
1414.
(148) Enders, D.; Breuer, I.; Nühring, A. Eur. J. Org. Chem. 2005, 2677–
2683.
(149) Enders, D.; Breuer, I.; Drosdow, E. Synthesis 2005, 3239–3244.
(150) Enders, D.; Breuer, I.; Raabe, G. Synthesis 2005, 3517–3530.
(151) Enders, D.; Moser, M.; Geibel, G.; Laufer, M. C. Synthesis 2004,
2040–2046.
(152) Enders, D.; Dhulut, S.; Steinbusch, D.; Herrbach, A. Chem.sEur. J.
2007, 13, 3942–3949.
(153) (a) Okano, T.; Fumoto, M.; Kusukawa, T.; Fujita, M. Org. Lett. 2002,
4, 1571–1573. (b) Mullins, J. E.; Etoga, J.-L. G.; Gajewski, M.;
DeGraw, J. I.; Thompson, C. M. Tetrahedron Lett. 2009, 50, 2298–
2300.
(154) Enders, D.; Schäfer, T.; Mies, W. Tetrahedron 1998, 54, 10239–
10252.
(155) Ternon, M.; Pannecoucke, X.; Outurquin, F.; Paulmier, C. Tetrahe-
dron 2002, 58, 3275–3279.
(156) Enders, D.; Berg, T.; Raabe, G.; Runsink, J. HelV. Chim. Acta 1996,
79, 118–122.
(157) Enders, D.; Joseph, R.; Poiesz, C. Tetrahedron 1998, 54, 10069–
10078.
(158) Enders, D.; Bhushan, V. Tetrahedron Lett. 1988, 29, 2437–2440.
(159) Enders, D.; Lohray, B. B. Angew. Chem., Int. Ed. Engl. 1987, 26,
351–352.
(160) Enders, D.; Ward, D.; Adam, J.; Raabe, G. Angew. Chem., Int. Ed.
Engl. 1996, 35, 981–984.
(161) Enders, D.; Adam, J.; Klein, D.; Otten, T. Synlett 2000, 1371–1384.
(162) Enders, D.; Burkamp, F. Collect. Czech. Chem. Commun. 2003, 68,
975–1006.
(163) Enders, D.; Schüβeler, T. Synthesis 2002, 2280–2288.
(164) Doszczak, L.; Gasperi, T.; Saint-Dizier, A.; Loreto, M. A.; Enders,
D. Chem. BiodiVersity 2004, 1, 1921–1935.
(165) Enders, D.; Faure, S.; Potthoff, M.; Runsink, J. Synthesis 2001, 2307–
2319.
(166) Enders, D.; Adam, J.; Oberbörsch, S.; Ward, D. Synthesis 2002, 2737–
2748.
(167) Enders, D.; Ince, S. J. Synthesis 2002, 619–624.
(168) Enders, D.; Ince, S. J.; Bonnekessel, M.; Runsink, J.; Raabe, G. Synlett
2002, 962–966.
(169) Brehme, R.; Enders, D.; Fernández, R.; Lassaletta, J. M. Eur. J. Org.
Chem. 2007, 5629–5660.
(170) Fernández, R.; Martı́n-Zamora, E.; Pareja, C.; Alcarazo, M.; Martı́n,
J.; Lassaletta, J. M. Synlett 2001, 1158–1160.
(171) Fernández, R.; Martı́n-Zamora, E.; Pareja, C.; Lassaletta, J. M. J.
Org. Chem. 2001, 66, 5201–5207.
(172) Lassaletta, J. M.; Vázquez, J.; Prieto, A.; Fernández, R.; Raabe, G.;
Enders, D. J. Org. Chem. 2003, 68, 2698–2703.
(173) Vázquez, J.; Prieto, A.; Fernández, R.; Enders, D.; Lassaletta, J. M.
Chem. Commun. 2002, 498–499.
(174) Enders, D.; Vázquez, J.; Raabe, G. Eur. J. Org. Chem. 2000, 893–
901.
(175) Vázquez, J.; Cristea, E.; Dı́ez, E.; Lassaletta, J. M.; Prieto, A.;
Fernández, R. Tetrahedron 2005, 61, 4115–4128.
Lazny and Nodzewska
(176) Brehme, R.; Gründemann, E.; Schneider, M.; Radeglia, R.; Reck,
G.; Schulz, B. Synthesis 2003, 1615–1619.
(177) Brehme, R.; Reck, G.; Schulz, B.; Radeglia, R. Synthesis 2003, 1620–
1625.
(178) Kamitori, Y. Tetrahedron Lett. 2000, 41, 9267–9270.
(179) Enders, D.; Reinhold, U. Tetrahedron: Asymmetry 1997, 8, 1895–
1946.
(180) Bloch, R. Chem. ReV. 1998, 98, 1407–1438.
(181) Enders, D.; Thiebes, C. Pure Appl. Chem. 2001, 73, 573–578.
(182) Ryu, I.; Yamamura, G.-h.; Omura, S.; Minakata, S.; Komatsu, M.
Tetrahedron Lett. 2006, 47, 2283–2286.
(183) Enders, D.; Funabiki, K. Org. Lett. 2001, 3, 1575–1577.
(184) Funabiki, K.; Nagamori, M.; Matsui, M.; Enders, D. Synthesis 2002,
2585–2588.
(185) Funabiki, K.; Nagamori, M.; Matsui, M.; Raabe, G.; Enders, D. ACS
Symp. Ser. 2007, 949, 447–461.
(186) Fries, S.; Pytkowicz, J.; Brigaud, T. Tetrahedron Lett. 2005, 46, 4761–
4764.
(187) Dumoulin, D.; Lebrun, S.; Deniau, E.; Couture, A.; Grandclaudon,
P. Eur. J. Org. Chem. 2009, 3741–3752.
(188) (a) Lebrun, S.; Couture, A.; Deniau, E.; Grandclaudon, P. Tetrahe-
dron: Asymmetry 2008, 19, 1245–1249. (b) Lebrun, S.; Couture, A.;
Deniau, E.; Grandclaudon, P. Synthesis 2008, 2771–2775.
(189) Lebrun, S.; Couture, A.; Denlau, E.; Grandclaudon, P. Org. Lett.
2007, 9, 2473–2476.
(190) Joly, K. M.; Wilson, C.; Blake, A. J.; Tucker, J. H. R.; Moody, C. J.
Chem. Commun. 2008, 5191–5193.
(191) Enders, D.; Noll, S.; Raabe, G.; Runsink, J. Synthesis 2008, 1288–
1296.
(192) Enders, D.; Noll, S.; Bats, J. Synlett 2005, 2679–2681.
(193) Enders, D.; Del Signore, G. Tetrahedron: Asymmetry 2004, 15, 747–
751.
(194) Enders, D.; Del Signore, G. Heterocycles 2004, 64, 101–120.
(195) Enders, D.; Moser, M. Tetrahedron Lett. 2003, 44, 8479–8481.
(196) Ros, A.; Dı́ez, E.; Marqués-López, E.; Martı́n-Zamora, E.; Vázquez,
J.; Iglesias-Sigüenza, J.; Pappalardo, R. R.; Álvarez, E.; Lassaletta,
J. M.; Fernández, R. Tetrahedron: Asymmetry 2008, 19, 998–1004.
(197) Enders, D.; Moll, A.; Bats, J. W. Eur. J. Org. Chem. 2006, 1271–
1284.
(198) Enders, D.; Braig, V.; Raabe, G. Can. J. Chem. 2001, 79, 1528–
1535.
(199) Enders, D.; Braig, V.; Boudou, M.; Raabe, G. Synthesis 2004, 2980–
2990.
(200) Boudou, M.; Enders, D. J. Org. Chem. 2005, 70, 9486–9494.
(201) Breuil-Desvergnes, V.; Goré, J. Tetrahedron 2001, 57, 1951–1960.
(202) Breuil-Desvergnes, V.; Goré, J. Tetrahedron 2001, 57, 1939–1950.
(203) Enders, D.; Meiers, M. Synthesis 2002, 2542–2560.
(204) Alcarazo, M.; Roseblade, S. J.; Alonso, E.; Fernández, R.; Alvarez,
E.; Lahoz, F. J.; Lassaletta, J. M. J. Am. Chem. Soc. 2004, 126,
13242–13243.
(205) Dumoulin, D.; Lebrun, S.; Couture, A.; Deniau, E.; Grandclaudon,
P. Tetrahedron: Asymmetry 2009, 20, 1903–1911.
(206) (a) Enders, D.; Gries, J.; Kim, Z.-S. Eur. J. Org. Chem. 2004, 4471–
4482. (b) Enders, D.; Gries, J. Synthesis 2005, 3508–3516.
(207) Enders, D.; Moll, A.; Schaadt, A.; Raabe, G.; Runsink, J. Eur. J.
Org. Chem. 2003, 3923–3938.
(208) Enders, D.; Schaadt, A. Synlett 2002, 498–500.
(209) Enders, D.; Nolte, B.; Raabe, G.; Runsink, J. Tetrahedron: Asymmetry
2002, 13, 285–291.
(210) Enders, D.; Nolte, B.; Runsink, J. Tetrahedron: Asymmetry 2002,
13, 587–593.
(211) Enders, D.; Thiebes, C. Synlett 2000, 1745–1748.
(212) Enders, D.; Kallfass, U.; Nolte, B. Synlett 2002, 33–36.
(213) Behforouz, M.; Ahmadian, M. Tetrahedron 2000, 56, 5259–5288.
(214) Valderrama, J. A.; González, M. F.; Valderrama, C. Tetrahedron
1999, 55, 6039–6050.
(215) Lyon, M. A.; Lawrence, S.; Williams, D. J.; Jackson, Y. A. J. Chem.
Soc., Perkin Trans. 1 1999, 437–442.
(216) Tailor, J.; Hall, D. G. Org. Lett. 2000, 2, 3715–3718.
(217) Jackson, Y. A.; Hepburn, S. A.; Reynolds, W. F. J. Chem. Soc., Perkin
Trans. 1 2001, 2237–2239.
(218) de la Fuente, J. Á.; Martı́n, M. J.; del Mar Blanco, M.; Pascual-
Alfonso, E.; Avendaño, C.; Menéndez, J. C. Bioorg. Med. Chem.
2001, 9, 1807–1814.
(219) Pérez, J. M.; López-Alvarado, P.; Pascual-Alfonso, E.; Avendaño,
C.; Menéndez, J. C. Tetrahedron 2000, 56, 4575–4583.
(220) Tapia, R. A.; Salas, C.; Morello, A.; Maya, J. D.; Toro-Labbé, A.
Bioorg. Med. Chem. 2004, 12, 2451–2458.
(221) Pérez, J. M.; López-Alvarado, P.; Avendaño, C.; Menéndez, J. C.
Tetrahedron 2000, 56, 1561–1567.
(222) Manzanaro, S.; Vicent, M. J.; Martı́n, M. J.; Salvador-Tormo, N.;
Pérez, J. M.; del Mar Blanco, M.; Avendaño, C.; Menéndez, J. C.;
de la Fuente, J. A. Bioorg. Med. Chem. 2004, 12, 6505–6515.
N,N-Dialkylhydrazones in Organic Synthesis
(223) Avendaño, C.; Pérez, J. M.; del Mar Blanco, M.; de la Fuente, J. Á.;
Manzanaro, S.; Vicent, M. J.; Martı́n, M. J.; Salvador-Tormo, N.;
Menéndez, J. C. Bioorg. Med. Chem. Lett. 2004, 14, 3929–3932.
(224) Cuellar, M. A.; Alegra, L. K.; Prieto, Y. A.; Cortes, M. J.; Tapia,
R. A.; Preite, M. D. Tetrahedron Lett. 2002, 43, 2127–2131.
(225) Delfourne, E.; Kiss, R.; Corre, L. L.; Dujols, F.; Bastide, J.; Collignon,
F.; Lesur, B.; Frydman, A.; Darro, F. J. Med. Chem. 2003, 46, 3536–
3545.
(226) Legentil, L.; Bastide, J.; Delfourne, E. Tetrahedron Lett. 2003, 44,
2473–2475.
(227) Delfourne, E.; Kiss, R.; Corre, L. L.; Dujols, F.; Bastide, J.; Collignon,
F.; Lesur, B.; Frydman, A.; Darro, F. Bioorg. Med. Chem. 2004, 12,
3987–3994.
(228) Palacios, F.; Aparicio, D.; López, Y.; de los Santos, J. M.; Ezpeleta,
J. M. Tetrahedron 2006, 62, 1095–1101.
(229) Palacios, F.; Aparicio, D.; Lópes, Y.; de los Santos, J. M. Heterocycles
2006, 67, 815–822.
(230) Aversa, M. C.; Barattucci, A.; Bilardo, M. C.; Bonaccorsi, P.;
Giannetto, P. Synthesis 2003, 2241–2248.
(231) Tsvetkov, N. P.; Vakhmistrov, V. E.; Koldobsky, A. B.; Korlyukov,
A. A.; Kalinin, V. N. Russ. Chem. Bull. 2002, 51, 326–331.
(232) Vahmistrov, V. E.; Tsvetkov, N. P.; Koldobsky, A. B.; Vorontsov,
E. V.; Kalinin, V. N. Russ. Chem. Bull. 2004, 53, 233–235.
(233) Bushby, N.; Moody, C. J.; Riddick, D. A.; Waldron, I. R. J. Chem.
Soc., Perkin Trans. 1 2001, 2183–2193.
(234) Fernández, R.; Ferrete, A.; Lassaletta, J. M.; Llera, J. M.; Monge,
A. Angew. Chem., Int. Ed. 2000, 39, 2893–2897.
(235) Marqués-López, E.; Martı́n-Zamora, E.; Dı́ez, E.; Fernández, R.;
Lassaletta, J. M. Eur. J. Org. Chem. 2008, 2960–2972.
(236) Fernández, R.; Ferrete, A.; Lassaletta, J. M.; Llera, J. M.; Martı́n-
Zamora, E. Angew. Chem., Int. Ed. 2002, 41, 831–833.
(237) Dı́ez, E.; Fernández, R.; Marqués-López, E.; Martı́n-Zamora, E.;
Lassaletta, J. M. Org. Lett. 2004, 6, 2749–2752.
(238) Martı́n-Zamora, E.; Ferrete, A.; Llera, J. M.; Muñoz, J. M.;
Pappalardo, R. R.; Fernández, R.; Lassaletta, J. M. Chem.sEur. J.
2004, 10, 6111–6129.
(239) Fernández, R.; Ferrete, A.; Llera, J. M.; Magriz, A.; Martı́n-Zamora,
E.; Dı́ez, E.; Lassaletta, J. M. Chem.sEur. J. 2004, 10, 737–745.
(240) Lomberget, T.; Baragona, F.; Fenet, B.; Barret, R. Org. Lett. 2006,
8, 3919–3922.
(241) Touré, B. B.; Hall, D. G. Angew. Chem., Int. Ed. 2004, 43, 2001–
2004.
(242) Touré, B. B.; Hall, D. G. J. Org. Chem. 2004, 69, 8429–8436.
(243) Touré, B. B.; Hall, D. G. Chem. ReV. 2009, 109, 4439–4486.
(244) Tour, B. B.; Hoveyda, H. R.; Tailor, J.; Ulaczyk-Lesanko, A.; Hall,
D. G. Chem.sEur. J. 2003, 9, 466–474.
(245) Zhang, Y.; Herndon, J. W. Org. Lett. 2003, 5, 2043–2045.
(246) Barluenga, J.; Ballesteros, A.; Santamarı́a, J.; Tomás, M. J. Orga-
nomet. Chem. 2002, 643-644, 363–368.
(247) Shaughnessy, J.; Cunningham, D.; Kavanagh, P.; Leech, D.; McArdle,
P.; Aldabbagh, F. Synlett 2004, 2382–2384.
(248) Shaughnessy, J.; Aldabbagh, F. Synthesis 2005, 1069–1076.
(249) Hehir, S.; O’Donovan, L.; Carty, M. P.; Aldabbagh, F. Tetrahedron
2008, 64, 4196–4203.
(250) Kim, S.; Yoon, J.-Y. Synthesis 2000, 1622–1630.
(251) Al-Qawasmeh, R. A.; Al-Telb, T. H.; Khan, K. M.; Perveen, S.;
Voelter, W. ARKIVOC [Online] 2007, part vii, 310-317. http://
www.arkat-usa.org/get-file/19669/ (accessed 10/11/2009).
(252) Mironov, V. F.; Gubaidullin, A. T.; Ivkova, G. A.; Litvinov, I. A.;
Buzykin, B. I.; Burnaeva, L. M.; Konovalova, I. V. Russ. J. Gen.
Chem. 2001, 71, 420–428.
(253) Kim, S.; In, S. K. Tetrahedron Lett. 1993, 34, 4213–4214.
(254) Kim, S.; Cheong, J. H.; Yoon, K. S. Tetrahedron Lett. 1995, 36,
6069–6072.
(255) Marco-Contelles, J.; Rodrı́guez, M. Tetrahedron Lett. 1998, 39, 6749–
6750.
(256) Iserloh, U.; Curran, D. P. J. Org. Chem. 1998, 63, 4711–4716.
(257) Marco-Contelles, J.; Balme, G.; Bouyssi, D.; Destabel, C.; Henriet-
Bernard, C. D.; Grimaldi, J.; Hatem, J. M. J. Org. Chem. 1997, 62,
1202–1209.
(258) Bowman, W. R.; Stephenson, P. T.; Terrett, N. K.; Young, A. R.
Tetrahedron 1995, 51, 7959–7980.
(259) Bernard-Henriet, C. D.; Grimaldi, J. R.; Hatem, J. M. Tetrahedron
Lett. 1994, 35, 3699–3702.
(260) Li, J. J. Tetrahedron 2001, 57, 1–24.
(261) Ellis, D. A.; Hart, D. J.; Zhao, L. Tetrahedron Lett. 2000, 41, 9357–
9360.
(262) Rhee, J. U.; Bliss, B. I.; RajanBabu, T. V. Tetrahedron: Asymmetry
2003, 14, 2939–2959.
(263) Rhee, J. U.; Bliss, B. I.; RajanBabu, T. V. J. Am. Chem. Soc. 2003,
125, 1492–1493.
(264) Spellmeyer, D. C.; Houk, K. N. J. Org. Chem. 1987, 52, 959–974.
Chemical Reviews, 2010, Vol. 110, No. 3 1433
(265) Beckwith, A. L. J.; Schiesser, C. H. Tetrahedron 1985, 41, 3925–
3941.
(266) Friestad, G. K.; Jiang, T.; Mathies, A. K. Org. Lett. 2007, 9, 777–
780.
(267) Friestad, G. K.; Jiang, T.; Fioroni, G. M. Tetrahedron 2008, 64,
11549–11557.
(268) Friestad, G. K.; Jiang, T.; Mathies, A. K. Tetrahedron 2007, 63, 3964–
3972.
(269) Friestad, G. K.; Mathies, A. K. Tetrahedron 2007, 63, 9373–9381.
(270) Arai, T.; Endo, Y.; Yanagisawa, A. Tetrahedron: Asymmetry 2007,
18, 165–169.
(271) Mino, T.; Suzuki, A.; Yamashita, M.; Narita, S.; Shirae, Y.; Sakamoto,
M.; Fujita, T. J. Organomet. Chem. 2006, 691, 4297–4303.
(272) Parrott II, R. W.; Dore, D. D.; Chandrashekar, S. P.; Bentley, J. T.;
Morgan, B. S.; Hitchcock, S. R. Tetrahedron: Asymmetry 2008, 19,
607–611.
(273) Mino, T.; Segawa, H.; Yamashita, M. J. Organomet. Chem. 2004,
689, 2833–2836.
(274) Mino, T.; Ogawa, T.; Yamashita, M. J. Organomet. Chem. 2003,
665, 122–126.
(275) Mino, T.; Komatsumoto, E.; Nakadai, S.; Toyoda, H.; Sakamoto,
M.; Fujita, T. J. Mol. Catal. A 2003, 196, 13–20.
(276) Mino, T.; Shiotsuki, M.; Yamamoto, N.; Suenaga, T.; Sakamoto, M.;
Fujita, T.; Yamashita, M. J. Org. Chem. 2001, 66, 1795–1797.
(277) Mino, T.; Ogawa, T.; Yamashita, M. Heterocycles 2001, 55, 453–
456.
(278) Mino, T.; Shirae, Y.; Yajima, T.; Sakamoto, M.; Fujita, T. Hetero-
cycles 2006, 68, 1233–1240.
(279) (a) Mino, T.; Shirae, Y.; Sasai, Y.; Sakamoto, M.; Fujita, T. J. Org.
Chem. 2006, 71, 6834–6839. (b) Mino, T.; Shindo, H.; Kaneda, T.;
Koizuma, T.; Kasashima, Y.; Sakamoto, M.; Fujita, T. Tetrahedron
Lett. 2009, 50, 5358–5360.
(280) Mino, T.; Shirae, Y.; Saito, T.; Sakamoto, M.; Fujita, T. J. Org. Chem.
2006, 71, 9499–9502.
(281) Bermejo, A.; Ros, A.; Fernández, R.; Lassaletta, J. M. J. Am. Chem.
Soc. 2008, 130, 15798–15799.
(282) Mino, T.; Kajiwara, K.; Shirae, Y.; Sakamoto, M.; Fujita, T. Synlett
2008, 2711–2715.
(283) Mino, T.; Harada, Y.; Shindo, H.; Sakamoto, M.; Fujita, T. Synlett
2008, 614–620.
(284) Lassaletta, J. M.; Alcarazo, M.; Fernández, R. Chem. Commun. 2004,
298–299.
(285) Monge, D.; Martı́n-Zamora, E.; Vázquez, J.; Alcarazo, M.; Álvarez,
E.; Fernández, R.; Lassaletta, J. M. Org. Lett. 2007, 9, 2867–2870.
(286) Dixon, D. J.; Tillman, A. L. Synlett 2005, 2635–2638.
(287) Herrera, R. P.; Monge, D.; Martı́n-Zamora, E.; Fernández, R.;
Lassaletta, J. M. Org. Lett. 2007, 9, 3303–3306.
(288) Pettersen, D.; Herrera, R. P.; Bernardi, L.; Fini, F.; Sgarzani, V.;
Fernández, R.; Lassaletta, J. M.; Ricci, A. Synlett 2006, 239–242.
(289) Rueping, M.; Sugiono, E.; Theissmann, T.; Kuenkel, A.; Kockritz,
A.; Pews-Davtyan, A.; Nemati, N.; Beller, M. Org. Lett. 2007, 9,
1065–1068.
(290) Hashimoto, T.; Hirose, M.; Maruoka, K. J. Am. Chem. Soc. 2008,
130, 7556–7557.
(291) Berner, O. M.; Tedeschi, L.; Enders, D. Eur. J. Org. Chem. 2002,
1877–1894.
(292) Sridharan, V.; Perumal, P. T.; Avendaño, C.; Menéndez, J. C. Org.
Biomol. Chem. 2007, 5, 1351–1353.
(293) Dı́ez, E.; Prieto, A.; Simon, M.; Vázquez, J.; Álvarez, E.; Fernández,
R.; Lassaletta, J. M. Synthesis 2006, 540–550.
(294) Handbook of Combinatorial Chemistry; Nicolaou, K. C., Hanko, R.,
Hartwig, W., Eds.; Wiley: New York, 2002.
(295) Seneci, P. Solid-Phase Synthesis and Combinatorial Technologies;
Wiley: New York, 2000.
(296) Jung, N.; Wiehn, M.; Bräse, S. Top. Curr. Chem. 2007, 278, 1–88.
(297) Scott, P. J. H.; Steel, P. G. Eur. J. Org. Chem. 2006, 2251–2268.
(298) Dörwald, F. Z. Organic Synthesis on Solid Phase, 2nd ed.; Wiley-
VCH: Weinheim, Germany, 2002.
(299) Kirchhoff, J. H.; Bräse, S.; Enders, D. J. Comb. Chem. 2001, 3, 71–
77.
(300) Enders, D.; Kirchhoff, J. H.; Köbberling, J.; Peiffer, T. H. Org. Lett.
2001, 3, 1241–1244.
(301) Köbberling, J. Ph.D. Thesis [Online], RWTH-Aachen, Aachen,
Germany, 2001. http://syluester.bth.rwth-aachen.de/dissertationen/
2001/061/01_061.pdf (accessed 10/11/2009).
(302) Schooren, J. Ph.D. Thesis [Online], Darmstadt Technical Univer-
sity, Darmstadt, Germany, 2003. http://tuprints.ulb.tu.darmstadt.de/
361/1/Dissertation_Schooren.pdf (accessed 10/11/2009).
(303) Lazny, R.; Michalak, M. Synlett 2002, 1931–1934.
(304) Lazny, R.; Nodzewska, A.; Wolosewicz, K. Synthesis 2003, 2858–
2864.
1434 Chemical Reviews, 2010, Vol. 110, No. 3
(305) Weik, S. Ph.D. Thesis [Online], Eberhard-Karls University, Tübingen,
Germany, 2004. http://w210.ub.uni-tuebingen.de/dbt/volltexte/2004/
1444/pdf/Dissertation_Steffen_Weik_2004.pdf (accessed 10/11/2009).
(306) Żabicka, B. M.Sc. Thesis, University of Bialystok, Bialystok, Poland,
2005.
(307) Lazny, R.; Nodzewska, A.; Zabicka, B. J. Comb. Chem. 2008, 10,
986–991.
(308) Leβmann, T.; Waldmann, H. Chem. Commun. 2006, 3380–3389.
(309) Lazny, R.; Nodzewska, A.; Sienkiewicz, M. Pol. J. Chem. 2006, 80,
655–658.
(310) Zhu, M.; Ruijter, E.; Wessjohann, L. A. Org. Lett. 2004, 6, 3921–
3924.
(311) Ulaczyk-Lesanko, A.; Pelletier, E.; Lee, M.; Prinz, H.; Waldmann,
H.; Hall, D. G. J. Comb. Chem. 2007, 9, 695–703.
(312) Petroski, R. J.; Bartelt, R. J. J. Agric. Food Chem. 2007, 55, 2282–
2287.
(313) Benites, J.; Valderrama, J. A.; Rivera, F.; Rojo, L.; Campos, N.; Pedro,
M.; Nascimento, M. S. J. Bioorg. Med. Chem. 2008, 16, 862–868.
(314) Kündig, E. P.; Bellido, A.; Kaliappan, K. P.; Pape, A. R.; Radix, S.
Org. Biomol. Chem. 2006, 4, 342–351.
Lazny and Nodzewska
(315) Ivonin, S. P.; Lapandin, A. V.; Shtamburg, V. G. Chem. Heterocycl.
Compd. 2005, 41, 1484–1493.
(316) Ivonin, S. P.; Lapandin, A. V.; Anishchenko, A. A.; Shtamburg, V. G.
Eur. J. Org. Chem. 2004, 4688–4693.
(317) Médebielle, M.; Kato, K.; Dolbier, W. R. J. Tetrahedron Lett. 2003,
44, 7871–7873.
(318) Minuti, L.; Taticchi, A.; Marrocchi, A. Tetrahedron: Asymmetry 2000,
11, 4221–4225.
(319) Lanari, D.; Marrocchi, A.; Minuti, L.; Rosini, C.; Superchi, S.;
Taticchi, A. Tetrahedron: Asymmetry 2002, 13, 1257–1263.
(320) Nyangulu, J. M.; Nelson, K. M.; Rose, P. A.; Gai, Y.; Loewen, M.;
Lougheed, B.; Quail, J. W.; Cutler, A. J.; Abrams, S. R. Org. Biomol.
Chem. 2006, 4, 1400–1412.
(321) Caro, Y.; Masaguer, C. F.; Raviña, E. Tetrahedron: Asymmetry 2003,
14, 381–387.
(322) May, J. A.; Stoltz, B. M. J. Am. Chem. Soc. 2002, 124, 12426–12427.
(323) Hwang, J.-l.; Hong, Y.-T.; Kim, S. AdV. Synth. Catal. 2005, 347,
1513–1516.
CR900067Y
 Chem. Rev. 2010, 110, 1435–1462 1435

Cobalt-Catalyzed Cross-Coupling Reactions

Gérard Cahiez* and Alban Moyeux
Department of Chemistry (FRE 3043), CNRS-Université de Paris 13, 74 Rue Marcel Cachin, F-93017 Bobigny, France

Received February 28, 2009

Contents 9.2. Allylation of 1,3-Dicarbonyl Compounds 1455

9.3. Activation of C-H Bonds: C-N Coupling 1456
1. Introduction 1435 Reactions
2. Homocoupling Reactions 1435 9.4. Three-Component Reactions 1456
3. Csp2-Csp2 Cross-coupling Reactions 1436 9.4.1. Synthesis of β-Acetamido Ketones 1456
3.1. Alkenylation 1436 9.4.2. Synthesis of Homoallylsilanes 1457
3.1.1. From Aromatic Organomagnesium 1436 9.5. Coupling of Alkenes and Alkynes 1457
Reagents 9.6. Various Reactions 1458
3.1.2. From Vinylic Grignard Reagents 1437 10. Conclusion 1459
3.1.3. From Aryl Halides and Alkenyl Acetates 1438 11. Acknowledgments 1460
3.2. Aryl-Aryl Cross-coupling 1438 12. References 1460
3.2.1. From Aromatic Organometallic Reagents 1438
3.2.2. From Two Aromatic Halides 1440 1. Introduction
4. Csp2-Csp3 Cross-coupling Reactions 1440
4.1. Alkenylation 1440 Since the pioneering work of Kharasch1 on the metal-
4.1.1. From Aliphatic Organometallic Reagents 1440 catalyzed homocoupling reaction of aromatic Grignard
reagents (see 2) in the middle of the 20th century, cobalt-
4.1.2. From Aliphatic Halides 1442
catalyzed carbon-carbon bond-forming reactions have re-
4.2. Arylation 1442 ceived particular attention.2 The scope of these reactions is
4.2.1. From Aliphatic Halides 1442 different from that of the palladium- and nickel-mediated
4.2.2. From Aromatic Halides 1445 procedures. Thus, cobalt-catalyzed cross-coupling reactions
4.3. Allylation of Aromatic Organometallics 1446 are very efficient for the elaboration of Csp2-Csp2 bonds (see
5. Alkynylation 1446 section 2), but they are especially interesting for couplings
5.1. Pioneering Works 1446 involving alkyl halides, since the decomposition by β-hy-
5.2. Benzylation of Acetylenic Grignard Reagents 1446 drogen elimination of alkyl-cobalt intermediates is not a
5.3. Alkylation of Acetylenic Grignard Reagents 1447 limitation as in the case of palladium or nickel catalyzed
5.4. Alkenylation of Acetylenic Grignard Reagents 1447 reactions (see sections 3, 4, and 5). Cobalt-mediated acyla-
6. Csp3-Csp3 Cross-coupling 1448 tions, radical cyclizations, and Heck-type reactions have also
been described (see sections 6 and 7). All these reactions
6.1. Allylation 1448
will be discussed hereafter.
6.1.1. Allylation of Aliphatic Organozinc 1448
Compounds It is worthy of note that cobalt salts have also been
6.1.2. Alkylation of Allylic Grignard Reagents 1448 extensively used as catalysts for Pauson-Khand3 and hydro-
6.1.3. Allylation of Trimethylsilylmethylmagnesium 1449 formylation4 reactions but also for the cyclopropanation5 of
Chloride olefins and for [2 + 2 + 2]6 and [2 + 2]7 cycloaddition
6.1.4. Allylation of Allylic Grignard Reagents 1449 reactions. These reactions will not be discussed herein.
6.2. Benzylation 1449
6.3. Alkylation 1450 2. Homocoupling Reactions
7. Acylation 1451 The first reports on cobalt-catalyzed cross-coupling reac-
7.1. From Organometallic Compounds 1451 tions describe the homocoupling reaction of Grignard
7.2. Cobalt-Mediated Acylation of Aryl Bromides 1451 reagents. In 1939, Gilman and Lichtenwalter8 obtained nearly
8. Reductive Cyclization and Heck-Type Reactions 1451 quantitative yields of homocoupling product by treating
8.1. Radical Cyclization 1451 aromatic Grignard reagents with a stoichiometric amount of
8.1.1. Intramolecular Radical Cyclization 1451 cobalt chloride. In 1941, Kharasch1 discovered that good
yields of homocoupling products are obtained by using only
8.1.2. Cyclization via an Intermolecular Radical 1453
Addition catalytic amounts (i.e., 3 mol %) of cobalt chloride and a
stoichiometric amount of an oxidant such as aromatic or
8.2. Cobalt-Catalyzed Heck-Type Reactions 1453
aliphatic halides (Scheme 1). The role of this latter is to
9. Miscellaneous Reactions 1455 oxidize the Cobalt(0) species into a cobalt(II) species after
9.1. Cobalt-Mediated Michael Addition 1455 the reductive elimination step.
Twenty years later, in 1962, Morizur applied the Kharasch
* E-mail: gerard.cahiez@univ-paris13.fr. reaction to the synthesis of various biaryls.9a The yields are
10.1021/cr9000786  2010 American Chemical Society
Published on Web 02/11/2010
1436 Chemical Reviews, 2010, Vol. 110, No. 3
Gérard Cahiez received his Ph.D. in 1973, at the University Pierre &
Marie Curie (Paris VI), under the supervision of Professor Jean François
Normant, on the carbocupration of terminal alkynes (vinyl copper reagents).
Then, he joined the CNRS. After a postdoctoral year in the Roussel Uclaf
Laboratories (now Sanofi Aventis) on the chemistry of steroid, he came
back to the University Pierre and Marie Curie, and in 1980, he became
Director of Research at the CNRS. Then, he moved to the Ecole
Supérieure de Chimie Organique et Minérale (ESCOM, Cergy-Pontoise)
in 1993. From 1993 to 2008 he was Director of Research at the CNRS
and Professor of Chemistry at ESCOM. From 2000 to 2009, he was also
Director of the UMR 8123, a joint research unit CNRS-University of Cergy
Pontoise-ESCOM. In 2009, he moved to the University of Paris 13, as
Director of Research at the CNRS, to create a new research group in
organometallic chemistry. The research developed since 1973 dealt with
the use of organometallic reagents in organic synthesis and especially
with the development of the chemistry of organomanganese reagents.
His current interest is always focused on organomanganese chemistry
but more generally on the search for new highly selective organometallic
reactions, i.e., Mn-, Co-, and Fe-catalyzed cross-coupling reactions,
involving no toxic and expensive metal or additive.
Alban Moyeux was born in La Bassée (France) in 1982. He obtained his
B.Sc. Degree from the University of Lens in 2004. He obtained his M.Sc.
Degree from the University of Cergy-Pontoise in 2005. He then worked
on iron- and cobalt-catalyzed cross-coupling reactions under the supervi-
sion of Professor Gérard Cahiez and was awarded his Ph.D. Degree in
November 2008. He finally joined the group of Professor Alois Fürstner
as a postdoctoral fellow at the Max-Planck Institute of Mülheim (Germany),
where he is currently working on total synthesis.
significantly lower than those previously reported by Khara-
sch. Almost identical results were obtained by using butyl
bromide (Scheme 2) or bromobenzene (Table 1) as an
oxidant.
It should be underlined that aryllithium and arylmagnesium
compounds give similar yields9b (Table 1).
Interestingly, benzylic or homobenzylic Grignard reagents
can also be used (Table 1, entries 6-10).
Cahiez and Moyeux
Scheme 1. The Kharasch Reaction
Scheme 2. Cobalt-Catalyzed Homocoupling Reaction of
Aromatic Grignard Reagents
3. Csp2-Csp2 Cross-coupling Reactions
3.1. Alkenylation
3.1.1. From Aromatic Organomagnesium Reagents
In 1943, Kharasch described the first cobalt-catalyzed
alkenylation of aromatic Grignard reagents10 (Table 2).
The very reactive vinyl or 1-propenyl halides give moder-
ate yields (Table 2, entries 1-3) whereas sterically hindered
or more substituted alkenyl halides only afford very poor
results. Large amounts of homocoupling products are then
obtained from both alkenyl halide and aryl Grignard reagent
(Table 2, entries 4-6). The synthetic interest of this reaction
is limited, since, even in the best cases, 2 equiv of alkenyl
halides are used.
In 1982, Uemura reported the cross-coupling reaction
between aryl Grignard reagents and alkenyl tellurides under
cobalt catalysis11 (Scheme 3).
Yields are quantitative, but a large excess of Grignard
reagent is used (2.5 equiv). It should be underlined that a
substantial amount of homocoupling product (biaryl) is
formed as a side product. This is a clear drawback, since
the purification of the final product is often tricky.
In 1998, Cahiez reinvestigated the cobalt-catalyzed alk-
enylation of aromatic Grignard reagents.12 A clear improve-
ment was observed by using a THF/NMP mixture as a
solvent (Table 3). Thus, various alkenyl halides have been
coupled stereospecifically in good to excellent yields (Table
3, entries 1-5).
Very recently, Hayashi13 described the cobalt-catalyzed
coupling between alkenyl triflates and aryl Grignard reagents.
Cobalt-Catalyzed Cross-Coupling Reactions
Table 1. Cobalt-Catalyzed Homocoupling Reaction of
Aryllithium and Arylmagnesium Halides
Table 2. Cobalt-Catalyzed Alkenylation of Aromatic Grignard
Reagents
The coupling takes place under mild conditions (Scheme 4).
The presence of NMP is not necessary, but PPh3 is used as
a ligand.
This method can be complementary to the precedent
method using alkenyl bromides, since vinyl triflates can be
easily prepared from the corresponding ketones.
Chemical Reviews, 2010, Vol. 110, No. 3 1437
Scheme 3. Cross-coupling between Alkenyl Tellurides and
Aromatic Grignard Reagents
Table 3. Cobalt-Catalyzed Alkenylation of Aromatic Grignard
Reagents in the Presence of NMP
Scheme 4. Cobalt-Catalyzed Cross-coupling between
Alkenyl Triflates and Aromatic Grignard Reagents
3.1.2. From Vinylic Grignard Reagents
Until now, there have only been very few cobalt-catalyzed
alkenyl-alkenyl cross-coupling reactions. In 1998, Cahiez12
described one example of cobalt-catalyzed cross-coupling
between an alkenyl halide and an alkenyl Grignard reagent
(Scheme 5) in moderate yield. The reaction is stereospecific
and gives only the pure trans-diene.
Recently, Hayashi13 described the cobalt-catalyzed cross-
coupling between alkenyl triflates and alkenyl Grignard
reagents. Good yields are obtained from R-monosubstituted
1438 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 5. Cobalt-Catalyzed Coupling between
Styrylmagnesium Bromide and 2-Methyl-1-bromopropane
Scheme 6. Cobalt-Catalyzed Coupling between
Alkenylmagnesium Halides and Alkenyl Triflates
Table 4. Cobalt-Catalyzed Coupling between Functionalized
Aryl Halides and Alkenyl Acetates
or R,β- and β,β-disubstituted alkenyl Grignard reagents
(Scheme 6).
3.1.3. From Aryl Halides and Alkenyl Acetates
In 2005, Périchon reported the cobalt-catalyzed coupling
between functionalized aryl halides and alkenyl acetates. It
is the only example of a cobalt-catalyzed arylation of enol
esters. The reaction is performed by using a catalytic amount
of cobalt bromide (5%) in the presence of 10 equiv of
manganese (Table 4).14
Moderate to good yields are obtained. As a rule, aryl
chlorides or bromides give similar results (Table 4, entries
Cahiez and Moyeux
Scheme 7. Cross-coupling between Chloropyridines and
Isopropenyl Acetate
Scheme 8. Mechanism Proposal for the Cobalt-Catalyzed
Alkenylation of Aryl Halides by Vinyl Acetates
1-6). However, in some cases, aryl chlorides lead to better
yields (2- and 4-halobenzonitriles, entries 7-8 and 13-14).
Aryl iodides such as 4-iodoanisole and ethyl 4-iodoben-
zoate are not suitable substrates, since they lead essentially
to a mixture of reduced and homocoupling products.
The reaction was extended to halopyridines (Scheme 7).
A mechanistic pathway was proposed (Scheme 8). In the
first step, CoII is reduced by manganese powder in the
presence of pyridine either in CoI or in Co0. In a precedent
work, Périchon demonstrated that CoI is stabilized by vinylic
acetate in an acetonitrile/pyridine mixture.15 The low-valent
cobalt species formed is probably coordinated by the vinyl
acetate. It then undergoes an oxidative addition of the aryl
halide to give an “arylcobalt” species. This species could
then react with vinylic acetate according to a six-membered
transition state [A]. After reductive elimination, cobalt is
reduced by the excess of Mn0 to regenerate the active cobalt
species.
3.2. Aryl-Aryl Cross-coupling
3.2.1. From Aromatic Organometallic Reagents
In 1983, Uemura described the cross-coupling between
diaryltellurides (Ar2Te) and aromatic organomagnesium
reagents in the presence of cobalt salts11 (Scheme 9).
The reaction affords poor yields, and a substantial amount of
homocoupling product accruing from the Grignard reagent is
produced. The purification of the final product is thus often delicate.
In 2003, Knochel and Cahiez reported the cross-coupling
between heteroaryl chlorides and aryl- or heteroarylmagne-
sium halides16 (Scheme 10).
Cobalt-Catalyzed Cross-Coupling Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1439

Scheme 9. Cross-coupling between Diaryltellurides and Scheme 12. Cobalt-Catalyzed Cross-coupling between
Aromatic Grignard Reagents Nonactivated Aryl Halides or Heteroaryl Bromides and
Aromatic Grignard Reagents

Scheme 10. CoCl2-Catalyzed Cross-coupling between

Heteroaryl Chlorides and Aromatic or Heteroaromatic
Grignard Reagents

Scheme 13. Cobalt-Catalyzed Cross-coupling between

Functionalized Arylcyanocuprates and ortho-Bromo or
ortho-Chloro Aromatic Ketones, Esters, or Aldehydes

Scheme 11. Cobalt or Iron Powder-Mediated Aryl-Aryl

Coupling Reactions

Yields are excellent in most cases. Noteworthy, the

influence of steric hindrance is not very important. Thus,
the reaction between mesitylmagnesium chloride and 1-chlor-
oisoquinoline leads to 85% of coupling product. Two
equivalents of Grignard reagent are required.
Interestingly, cobalt or iron powders are suitable catalysts
for the reaction16 (Scheme 11).
One year later, Oshima16b reported a very similar reaction.
Then, Knochel has shown that the chemoselective coupling
The coupling is achieved in dioxane in the presence of
between polyfunctional aryl- or heteroarylcyanocuprates
cobalt(II) acetylacetonate whereas diethyl ether and cobalt(II)
ArCu(CN)MgCl and ortho-bromo or ortho-chloro aromatic
chloride were used in the previous method16a (Scheme 10).
ketones, esters, or aldehydes can be performed under cobalt
Yields are comparable, but 3 equiv of phenyl magnesium
bromide are used instead of 2 and the amount of catalyst is catalysis.18 Numerous polyfunctional biaryls were success-
twice higher. fully synthesized according to this methodology. Selected
The above examples were performed by using heteroaromatic examples are presented in Scheme 13.
or actived aromatic halides. Very recently, Nakamura17 de- To react the starting halide completely, it is essential to
scribed a cobalt-catalyzed cross-coupling reaction from perform the coupling in the presence of Bu4NI (1 equiv) and
nonactivated aryl chlorides or heteroaryl bromides and 4-fluorostyrene (20 mol %) by using a large excess of
aromatic Grignard reagents (Scheme 12). The reaction is cyanoarylcuprate (3 equiv).19 The use of a THF/DME/DMPU
performed in the presence of cobalt(II) fluoride and NHC mixture instead of THF as a solvent clearly accelerates the
ligand. reaction (15 min instead of 21 h).
1440 Chemical Reviews, 2010, Vol. 110, No. 3 Cahiez and Moyeux

Scheme 14. Cobalt-Catalyzed Reaction of Scheme 17. Cobalt-Catalyzed Coupling Reaction between
p-Anisylcyanocuprates with o-, m-, and Arylcyanocuprates and Pentafluorobenzophenone
p-Bromobenzophenones

Scheme 18. Synthesis of Unsymmetrical Biaryl Compounds

by Coupling Two Aryl Halides under Cobalt Catalysis

Scheme 15. Cobalt-Catalyzed Cross-coupling between

Heteroarylcyanocuprates and Activated Aryl Bromides

Table 5. Cobalt-Catalyzed Coupling of Two Aryl Halides:

Influence of the Nature of the Aryl Halides

Scheme 16. Cobalt-Catalyzed Cross-coupling between

Activated Aryl Fluorides or Tosylates and
Arylcyanocuprates
a
Isolated yield. b Yields in parentheses refer to GC yield.

Scheme 19. Cobalt-Catalyzed Coupling Reaction between

β-Bromostyrene and Methylmagnesium Iodide

The yields are significantly lower when the activating

carbonyl group is in the meta or the para position (Scheme
14). Scheme 20. Influence of NMP on the Cobalt-Catalyzed
Heteroarylcyanocuprates can also be used successfully Alkenylation of Grignard Reagents
(Scheme 15).
The reaction has been extended to ortho-fluoro aromatic
ketones, esters, or aldehydes20 (Scheme 16). The correspond-
ing tosylates also react successfully. This is worthy of note,
since the use of aryl fluorides21 and tosylates22 in metal-
catalyzed cross-coupling reactions is not usual. As a rule,
good to excellent yields are obtained. successfully (Table 5, entry 7). It should be underlined that
With pentafluorobenzophenone, the two fluorine atoms in a significant amount of homocoupling product is produced
the ortho-positions can be selectively substituted (Scheme
during the reaction. This is a practical drawback, since the
17). However, yields are moderate and 6 equiv of arylcy-
anocuprate are required. expected product could then be very tricky to isolate.

3.2.2. From Two Aromatic Halides 4. Csp2-Csp3 Cross-coupling Reactions

Very recently, Gosmini reported the synthesis of unsym-
metrical biaryl compounds by coupling two aromatic halides
ArX and Ar′X under cobalt catalysis23 (Scheme 18). The
use of 2 equiv of the more reactive aryl halide is generally
required to obtain satisfactory yields of cross-coupling
products.
Similar results were obtained from aryl iodides, bromides,
or chlorides (Table 5). Interestingly, aryl triflates also react
4.1. Alkenylation
4.1.1. From Aliphatic Organometallic Reagents
In 1945, Kharasch24 showed that the coupling between
methyl magnesium iodide and β-bromostyrene takes place
in good yield in the presence of cobalt chloride (Scheme
19).
Cobalt-Catalyzed Cross-Coupling Reactions
Scheme 21. Cobalt-Catalyzed Alkenylation of Grignard
Reagents: Stereospecificity of the Reaction
Table 6. Cobalt-Catalyzed Coupling between Alkenyl Halides
and Aliphatic Grignard Reagents
However, the method is limited to the very reactive
β-bromostyrene, and 2 equiv of Grignard reagent are
necessary.
In 1998, Cahiez discovered that the use of N-methylpyr-
rolidone (NMP) as a cosolvent allows a dramatic improve-
ment (Scheme 20; see also section 3.1). The yields are clearly
better, and only 1.1 equiv of Grignard reagent is required.12
Alkenyl iodides, bromides, and chlorides give good yields
of cross-coupling product (Scheme 21).
It is important to note that the reaction is stereospecific
(Scheme 21).
The procedure has been applied to the synthesis of various
olefins (Table 6).
Good yields are obtained by coupling primary (Table 6,
entries 1-3) or secondary (entries 4-5) aliphatic Grignard
reagents with various alkenyl bromides. It is noteworthy that
tertiary aliphatic Grignard reagents give poor results (entry
6).
The reaction is highly chemoselective. Thus, the presence
of an ester or even a keto group is tolerated (Scheme 22).
Finally, it should be underlined that (E)-1,2-dichloroeth-
ylene reacts to give only the corresponding (E)-alkenyl
chlorides. The formation of the product resulting from a
substitution of both chlorine atoms is not observed (Scheme
23). This result is in sharp contrast with that obtained under
Chemical Reviews, 2010, Vol. 110, No. 3 1441
Scheme 22. Chemoselective Cobalt-Catalyzed Alkenylation
of Grignard Reagents
Scheme 23. Synthesis of (E)-1-Alkenyl Chlorides from
(E)-Dichloroethylene
Scheme 24. Cobalt-Catalyzed Alkenylation of Functionalized
Organozinc Halides
Scheme 25. Cobalt-Catalyzed Alkenylation of Benzylzinc
Bromide
iron catalysis, since, in this case, the double substitution
product is formed exclusively.25
Cahiez and Knochel then extended the reaction to alkylzinc
halides.26 These reagents are less reactive than the corre-
sponding organomagnesium halides, and the coupling takes
place slowly at a higher temperature (55 °C, 4-8 h). In
addition, 20-30 mol % of catalyst and a large excess of
organozinc halide (3 equiv) are necessary. As with Grignard
reagents,12 the reaction is chemoselective and stereospecific.
Good yields of cross-coupling product are obtained (Scheme
24).
It is worthy of note that benzylic organozinc halides can
be coupled with alkenyl iodides (Scheme 25).
The coupling of various silylated Grignard reagents with
1,2-dihalogenoethylenes was also reported (Table 7).27
Excellent yields are obtained from trimethylsilylmethyl-
magnesium chloride (Table 7, entries 1 and 3) or phenyldim-
ethylsilylmethylmagnesium chloride (Table 7, entry 2). As
previously reported by Cahiez,12 the formation of the
disubstituted products was not observed. It is noteworthy that
stereodifferentiation between (E)- and (Z)-1,2-dihaloethylenes
is possible. Thus, by reacting a mixture of (E)- and (Z)-1,2-
dibromoethylenes (5 equiv) with trimethylsilylmethyl mag-
nesium chloride, the only product is the (E)-1-bromoalkene
(Table 7, entry 3).
1442 Chemical Reviews, 2010, Vol. 110, No. 3
Table 7. Cobalt-Catalyzed Selective Coupling between
1,2-Dihalogenoethylenes and PhMe2SiCH2MgCl or
Me3SiCH2MgCl
a
5 equiv of 1,2-dibromoethylene was used.
Table 8. Cobalt-Catalyzed Cross-coupling between Alkenyl
Halides and Silylated Grignard Reagents
a
The reaction was performed at 35 °C.
Scheme 26. Cobalt-Catalyzed Alkylation of
Styrylmagnesium Bromide
Later, the cross-coupling between alkenyl halides and
dimethylalkylsilylmethylmagnesium chlorides was reported
by Oshima.28 Good to excellent yields are obtained (Table
8).
R- or β-monosubstituted as well as β,β-bisubstituted
alkenyl iodides, bromides, or chlorides afford good yields
(entries 1-6). On the other hand, R,β-bisubstituted alkenyl
halides only lead to poor results (entry 7).
4.1.2. From Aliphatic Halides
29
In 2006, Oshima described the reaction between octyl
bromide and styrylmagnesium bromide in the presence of 5
mol % cobalt chloride and N,N,N′,N′-tetramethyl-1,2-cyclo-
hexanediamine as a ligand (Scheme 26).
The same year, he reported the cobalt-catalyzed cross-
coupling between various alkyl halides and 1-(trimethylsil-
yl)ethenylmagnesium bromide by using TMEDA as a sol-
Cahiez and Moyeux
Table 9. Cross-coupling between Primary or Secondary Alkyl
Halides and 1-(Trimethylsilyl)ethenylmagnesium Bromide
a
The trans isomer was formed exclusively.
Table 10. Tandem Radical Cyclization and Cross-coupling
Reaction
a
Diastereomeric ratio. b The product was isolated as a lactone after
oxidation of the cyclic acetal.
vent.30 Good to excellent yields are obtained from either
primary or secondary alkyl halides, but a large excess of
organometallic reagent (4 equiv) has to be used and the
reaction seems limited to this specific Grignard reagent
(Table 9).
From ε-unsaturated alkyl iodides 1, a cyclic product is
formed via a radical cyclization followed by a cross-coupling
reaction (Table 10). The mechanism proposed by Oshima is
depicted in Scheme 27. Thus, various heterocyclic com-
pounds have been synthesized in good to excellent yields.
4.2. Arylation
4.2.1. From Aliphatic Halides
The first example of cobalt-catalyzed arylation of a Csp3
center was described in 1969. In this report, Hey31 noted
that the yield of the reaction between 2-bromopyridine and
methylmagnesium iodide is clearly improved in the presence
of catalytic amounts of cobalt(II) chloride (Scheme 28).
Cobalt-Catalyzed Cross-Coupling Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1443

Scheme 27. Mechanism for Tandem Radical Cyclization and Table 11. Tandem Radical Cyclization and Cross-coupling
Cross-coupling Reaction Reaction from Unsaturated Haloacetal 2

Scheme 28. First Cobalt-Catalyzed Arylation of Aliphatic a

Isolated yield. The diastereomeric ratio is given in parentheses.
Grignard Reagents
Scheme 29. Heck-Type Reaction from Unsaturated
Haloacetals 3 and 4

However, the reaction is not general and very poor yields

are obtained in most other cases.
In 2001, Oshima32 described the cobalt-catalyzed tandem
radical cyclization and arylation reaction from ethylenic
haloacetal 2 (Table 11). This methodology has been applied
to the synthesis of heterocyclic (entries 1-7) and carbocyclic
(entry 8) compounds. It should be noted that the yield clearly
depends on both the position and the number of substituents Scheme 30. Synthesis of Oxasilacyclopentanes
on the substrate (entries 4 and 6). The mechanism proposed
in this case is similar to the one described in Scheme 27.
It is noteworthy that when the double bond of the starting
product is not terminal (haloacetals 3 and 4, Scheme 29)
the arylation product is not obtained. In this case, the reaction
leads to a Heck-type product (see section 9).
In 2006, this reaction was successfully applied to the
synthesis of oxasilacyclopentanes (Scheme 30).33
These compounds can easily be converted into 4-aryl-1,3-
diols by Tamao-Fleming oxidation.34 This approach has
been used in a synthesis of an antagonist of human CCR5
receptor (Scheme 31).
Heteroaromatic organomagnesium compounds can be used
The same year, Oshima described the cross-coupling successfully (entry 5). The reaction is sensitive to steric
between primary alkyl halides and aromatic Grignard
hindrance; thus, ortho-substituted aryl Grignard reagents do
reagents in the presence of CoCl2(dppp).35 Selected examples
are presented in Table 12. It must be underlined that a large not afford the expected cross-coupling product (entry 6).
excess of Grignard reagent is required (3 equiv). As a rule, Interestingly, some functional groups such as an ester or an
moderate yields are obtained from primary alkyl bromides acetal are tolerated (entries 7 and 8).
whereas the corresponding iodides give low yields (entries Oshima has finally shown29 that better results are obtained
1 and 2). Moreover, secondary cyclic alkyl bromides such by using a diamine (N,N,N′,N′-tetramethyl-1,2-cyclohex-
as cyclohexyl bromide lead to poor results (entry 4) and alkyl anediamine) as a ligand instead of dppp (Table 13).
chlorides do not react. Moreover, only a slight excess (1.2 equiv) of Grignard
1444 Chemical Reviews, 2010, Vol. 110, No. 3 Cahiez and Moyeux

Scheme 31. Synthesis of an Antagonist of Human CCR5 Table 13. Cobalt-Catalyzed Cross-coupling between Primary or
Receptor Secondary Alkyl Halides and Aromatic Grignard Reagents

Table 12. Cobalt-Catalyzed Cross-coupling between Primary

Alkyl Halides and Aromatic Grignard Reagents

Scheme 32. Synthesis of AH13205

a
Only 3-chloropropylbenzene is formed.

reagent is then used. With this catalytic system, the yields

are significantly higher from primary alkyl iodides than from
the corresponding bromides (entries 3 and 4). In addition,
secondary cyclic and acyclic alkyl halides (entries 1 and 2)
can also be used successfully. Finally, the reaction is less
sensitive to steric hindrance; as an example, ortho-tolylmag-
nesium bromide leads to an excellent yield (entry 7).
The reaction is highly chemoselective, and various halo-
genoesters were coupled successfully (entries 9-11).
This methodology has been applied to the synthesis of
AH13205, a synthetic prostaglandin used as an EP2-receptor and aromatic organomagnesium compounds. A cheap and
agonist that lowers intraocular pressure (Scheme 32).36 simple catalytic system, cobalt(III) acetylacetonate/TMEDA
The reaction of five- and six-membered cyclic halo acetals (1:1), is used. This is very interesting, since commercial
5 having a stereogenic center in the R position of the C-Br TMEDA is a very simple and inexpensive starting material
bond has been examined (Scheme 33). The stereoselectivity compared to the N,N,N′,N′-tetramethyl-1,2-cyclohexanedi-
of this reaction is highly dependent on the size of the ring. amine previously employed by Oshima.29
Thus, good enantioselectivity was only obtained with the It is possible to couple secondary alkyl bromides with
five-membered cyclic acetal. excellent yields, whereas, under the conditions described by
Very recently, Cahiez37 described the cobalt-catalyzed Oshima,29 alkyl bromides lead to clearly lower yields than
cross-coupling between primary or secondary alkyl bromides alkyl iodides. This point is very interesting, since alkyl
Cobalt-Catalyzed Cross-Coupling Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1445

Scheme 33. Cobalt-Catalyzed Stereoselective Arylation of Table 15. Cross-coupling of Aryl Grignard Reagents with
Cyclic Haloacetals 5 Functionalized Primary Alkyl Bromides

Table 14. Cobalt-Catalyzed Cross-coupling between Aliphatic

Bromides and Aromatic Grignard Reagents

Scheme 34. Cobalt-Catalyzed Chemoselective Arylation of

Functionalized Secondary Alkyl Bromides

Scheme 35. Cobalt-Catalyzed Cross-coupling between Aryl

Bromides and Alkyl Grignard Reagents

bromides are cheaper and more stable than the corresponding

iodides. Results are summarized in Table 14.
As a rule, excellent yields are obtained from primary and
secondary alkyl Grignard reagents (entries 1-8). On the other
hand, tertiary alkyl bromides cannot be used (entry 9). It
should be underlined that only a slight excess (1.1 equiv) of
Grignard reagent was used. Alkyl iodides and bromides lead
to similar results (entries 3-4) whereas alkyl chlorides or
tosylates only afford poor yields (entry 5).
The reaction is not very sensitive to steric hindrance, and
it is possible to couple an ortho-substituted aromatic Grignard
reagent with a secondary alkyl bromide in high yields (entry
8).
The reaction is highly chemoselective (Table 15). Thus,
1-bromo-5-chloropentane couples selectively to give the 4.2.2. From Aromatic Halides
corresponding alkyl chloride in good yield (entry 1). In In 2008, Oshima38 described the cobalt-catalyzed coupling
addition, the presence of various sensitive groups such as between aryl bromides and aliphatic Grignard reagents
an acetate (entry 2), an ester (entry 3), an amide (entry 4), (Scheme 35). It is the only report in this field. This reaction
or even a ketone (entry 5) is tolerated. is an interesting alternative to the related palladium or nickel
The reaction is also highly chemoselective in the case of procedures.
secondary alkyl bromides (Scheme 34). It is the first report
of a metal-catalyzed cross-coupling between an aromatic Yields are generally good to excellent; however, this
Grignard reagent and a functionalized secondary alkyl reaction can only be applied to primary alkyl Grignard
bromide. reagents.
1446 Chemical Reviews, 2010, Vol. 110, No. 3 Cahiez and Moyeux

Scheme 36. Cobalt-Catalyzed Coupling between Allyl Scheme 38. Cobalt-Catalyzed Cross-coupling between 1- and
Acetate and Functionalized Aryl Bromides 3-Phenyl-2-propenyl Methyl Ethers and Phenylmagnesium
Bromide

Scheme 39. Cobalt-Catalyzed Arylation of Allylic Acetals

Scheme 37. Cobalt-Catalyzed Coupling between Allyl

Acetate and Functionalized Aryl Chlorides

Scheme 40. Cobalt-Catalyzed Coupling between

2-Bromo-1-phenylacetylene and Methylmagnesium Bromide

Scheme 41. Cobalt-Catalyzed Coupling between

Phenylethynylmagnesium Bromide and Phenylethynyl Iodide

4.3. Allylation of Aromatic Organometallics

In 2003, Périchon reported the coupling of aryl halides with allyl
acetates in the presence of a catalytic amount of cobalt bromide;
zinc is used as a reductant.39
Moderate yields were obtained from various aryl bromides
(Scheme 36). It should be noted that the presence of an ester Table 16. Cobalt-Catalyzed Alkynylation of Various Grignard
Reagents
or a nitrile is tolerated. This method is complementary to
the cobalt-catalyzed electrochemical methods previously
reported.40
The reaction can also be achieved with aryl chlorides
(Scheme 37), but the reaction conditions are very different.
It is necessary to increase the amount of cobalt bromide (40%
instead of 20%), zinc dust has to be replaced by manganese
dust as a reductant, and the presence of pyridine as well as
a stoichiometric amount of iron bromide is required. More-
over, the coupling is performed at a higher temperature
(50 °C), and the reaction time is longer (24 h). However,
yields are generally better than those previously obtained the presence of cobalt chloride (Scheme 40). It is the first
from the corresponding aryl bromides. example of cobalt-catalyzed alkynylation of Grignard reagents.
In 2004, Oshima41 described the reaction between allylic A few years later, the cobalt-catalyzed coupling between
ethers and phenylmagnesium bromide in the presence of phenylethynylmagnesium bromide and phenylethynyl iodide
cobalt salts. Good yields were obtained, but the reaction is was reported (Scheme 41).43 A satisfactory yield of diphe-
not regioselective, excepted with 1- and 3-phenyl-2-propenyl nylbutadiyne was obtained.
methyl ethers (Scheme 38). In 1954, several examples of coupling between aliphatic,
Under similar conditions, allylic acetals lead to the aromatic, or acetylenic Grignard reagents and 1-haloalkynes
monosubstitution product (Scheme 39). were reported by Weedon.44 As a rule, only poor yields were
obtained (Table 16).
5. Alkynylation
5.2. Benzylation of Acetylenic Grignard Reagents
5.1. Pioneering Works
In 2006, Oshima reported the benzylation of acetylenic
In 1945, Kharasch42 reported the coupling between me- Grignard reagents under cobalt catalysis.45 Moderate to
thylmagnesium bromide and 2-bromo-1-phenylacetylene in excellent yields are obtained (Table 17).
Cobalt-Catalyzed Cross-Coupling Reactions
Table 17. Cobalt-Catalyzed Benzylation of Alkynylmagnesium
Halides
Scheme 42. Preparation of Diynes from
1,2-Bis(bromomethyl)- or 1,3-Bis(chloromethyl)benzene
Scheme 43. Cobalt-Catalyzed Cross-coupling between
Primary or Secondary Alkyl Bromides and
1-Trimethylsilylethynylmagnesium Bromide
Trimethylsilylethynylmagnesium bromide reacts with ben-
zylic chlorides or bromides to give good yields of cross-
coupling products (entries 1-4, R ) Me3Si). On the other
hand, only benzylic bromides lead to good results from
1-hexynylmagnesium bromide (entry 1, R ) n-Bu).
With 1,2-bis(bromomethyl)- or 1,3-bis(chloromethyl)ben-
zene, the double substitution is possible and the reaction
affords diynes (Scheme 42).
5.3. Alkylation of Acetylenic Grignard Reagents
Recently, Oshima reported the reaction between alkyl
halides and trimethylsilylethynylmagnesium bromide in the
presence of cobalt chloride.29 Satisfactory yields are obtained
from primary or secondary alkyl halides (Scheme 43), but a
large excess of Grignard reagent is used.
ε-Haloalkenes 6 react with trimethylsilylethynylmagnesium
bromide according to a tandem radical cyclization and cross-
Chemical Reviews, 2010, Vol. 110, No. 3 1447
Table 18. Cobalt-Catalyzed Tandem Radical Cyclization and
Alkynylation Reaction
a
Diastereomeric ratio. b The product was isolated as a lactone after
oxidation of the cyclic acetal.
Scheme 44. Cobalt-Catalyzed Tandem Radical Cyclization
and Alkynylation Reaction
coupling reaction. Various heterocyclic compounds have thus
been synthesized in moderate to good yields (Table 18).
Oshima46 showed that simple acetylenic Grignard reagents
can also be used (Scheme 44).
However, it is important to note that 5 equiv of Grignard
reagent have to be used in this case.
5.4. Alkenylation of Acetylenic Grignard Reagents
The cobalt-catalyzed coupling between alkynylmagnesium
halides and alkenyl triflates was described by Hayashi in
2007.47 Good to excellent yields are obtained from a wide
range of substrates (Scheme 45).
Some examples of chemoselective couplings in the pres-
ence of an alkenyl, an alkyl, or an aryl bromide are described
(Table 19). It is noteworthy that no additional ligand was
used in this case.
1448 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 45. Cobalt-Catalyzed Coupling between Alkenyl
Triflates and Alkynyl Grignard Reagents
Table 19. Chemoselective Cobalt-Catalyzed Alkenylation of
Acetylenic Grignard Reagents
Scheme 46. Cobalt-Catalyzed Allylation of Diorganozinc
Compounds
6. Csp3-Csp3 Cross-coupling
6.1. Allylation
6.1.1. Allylation of Aliphatic Organozinc Compounds
In 1996, Knochel48 described the allylation of organozinc
compounds (R2Zn or RZnX) under cobalt catalysis (Scheme
46).
It should be pointed out that both R groups from R2Zn
are transferred. Moreover, only the SN2 substitution
products are obtained. Dialkylzincs and alkylzinc halides
give similar yields, but the latter require longer reaction
times (about 5 h at -10 °C).
Allylic phosphates can also be used successfully (Scheme
47).
6.1.2. Alkylation of Allylic Grignard Reagents
The coupling between unactivated alkyl halides and
allylic Grignard reagents in the presence of cobalt chloride
Cahiez and Moyeux
Scheme 47. Cobalt-Catalyzed Coupling between Diethyl
Cinnamyl Phosphate and Dipentylzinc
Table 20. Cobalt-Catalyzed Coupling between Unactivated
Alkyl Halides and Allylmagnesium Chloride
a
trans/cis ) 86:14. b dppe was used as a ligand.
Scheme 48. Cobalt-Catalyzed Alkylation of Crotyl- and
Prenylmagnesium Chlorides
and dppp was described by Oshima in 2002 (Table 20).49
Good yields are obtained under mild conditions from primary
(entry 1), secondary (entries 2 and 3), or tertiary (entries 4
and 5) alkyl halides. Benzylic halides can also be used (entry
6).
With substituted allylic Grignard reagents, the regiose-
lectivity is very dependent on the structure of the allylic
moiety. A mixture of products resulting from an alkylation
on the less and the more substituted sides of the allylic
system is generally obtained (Scheme 48).
A tandem radical cyclization and cross-coupling reaction
was used to prepare various cyclic compounds in good to
excellent yields (Scheme 49). For the mechanism of the
reaction, see Scheme 27.
Unexpectedly, by using ether as a solvent, a cyclopropane
derivative is formed (Scheme 50).
The following mechanism is proposed. At first, cobalt
chloride is converted to an active reduced species [Con]
by action of the Grignard reagent. After a single electron
transfer to the halide (formation of a radical anion) and
then cleavage of the C-I bond, a radical is obtained.
Sequential 5-exo radical cyclization and trapping of the
resulting radical by the catalytic cobalt species then afford
the cobalt intermediate 10. In diethyl ether, 10 undergoes
Cobalt-Catalyzed Cross-Coupling Reactions
Scheme 49. Cobalt-Catalyzed Tandem Radical Cyclization
and Alkylation of Allylic Grignard Reagents
Scheme 50. Mechanism of the Cobalt-Catalyzed Reaction of
δ-Iodoallylic Acetal 9 with Allylmagnesium Chloride
Scheme 51. Cobalt-Catalyzed Reaction of δ-Iodoallylic
Acetal 9 with Crotyl- and Prenylmagnesium Chlorides
a rapid elimination of BuO-Metal (Metal ) Co or Mg) to
give the cyclopropane derivative 11. Such an elimination
probably involves the complexation of the metallic center
of 10 (Co or Mg) with the oxygen atom of the butoxy
group (in THF, this complexation is much more difficult
and the elimination does not occur). The resulting alde-
hyde 11 then reacts with allylmagnesium chloride to afford
the alcohol 12.
The reaction was extended to prenyl- or crotylmagnesium
chlorides (Scheme 51).
6.1.3. Allylation of Trimethylsilylmethylmagnesium
Chloride
The coupling between allylic ethers and trimethylsilylm-
ethylmagnesium chloride under cobalt catalysis was reported
in 2004.41 As with phenylmagnesium bromide (see section
4.3), good yields and selectivity are obtained from 1- and
3-phenyl-2-propenyl methyl ethers (Scheme 52). However,
the reaction is generally not regioselective with the other
allylic ethers.
It is noteworthy that, from allylic acetals 13 (Scheme 53),
it is possible to substitute one or two methoxy groups by
using respectively 1.5 or 3 equiv of Grignard reagent.
Chemical Reviews, 2010, Vol. 110, No. 3 1449
Scheme 52. Cobalt-Catalyzed Cross-coupling between 1- and
3-Phenyl-2-propenyl Methyl Ethers and
Trimethylsilylmethylmagnesium Chloride
Scheme 53. Cobalt-Catalyzed Coupling of Allylic Acetals
with Trimethylsilylmethylmagnesium Chloride
Scheme 54. Cobalt-Catalyzed Coupling between
Allylmagnesium Bromide and Cinnamyl Methyl Ether
Scheme 55. Cobalt-Catalyzed Benzylation of
4-Acetoxybutylzinc Iodide
6.1.4. Allylation of Allylic Grignard Reagents
A cobalt-catalyzed coupling reaction between allylic
Grignard reagents and allylic ethers has been described by
Oshima (Scheme 54).41
Good yields are obtained. Interestingly, the regioselectivity
of the reaction can be easily controlled by changing the
nature of the catalyst. Thus, only the linear substitution
product 14 is formed in the presence of uncomplexed cobalt
chloride whereas the branched substitution product 15 is the
main isomer when the complex CoCl2•dppp is used.
6.2. Benzylation
An example of cobalt-catalyzed benzylation of an orga-
nozinc compound (Scheme 55) was reported by Knochel
in1996.48
In 2004, Oshima achieved the coupling between secondary
alkyl halides and benzylic Grignard reagents in the presence
of CoCl2•dppp. As a rule, yields are moderate (Table 21).49
Tertiary alkyl iodides also react but lead to poor yields (entry
7).
1450 Chemical Reviews, 2010, Vol. 110, No. 3
Table 21. Cobalt-Catalyzed Alkylation of Benzylic Grignard
Reagents
a
3 equiv of Grignard reagent was used.
Table 22. Cobalt-Catalyzed Tandem Radical Cyclization and
Alkylation Reaction of Various δ-Iodo Ethylenic Compounds
with RMe2SiCH2MgCl
a
Diastereomeric ratio.
6.3. Alkylation
In 2007, Oshima reported the tandem radical cyclization
and cross-coupling reaction of various δ-halogeno ethylenic
compounds with RMe2SiCH2MgCl under cobalt catalysis.46
N-Heterocyclic carbenes are employed as ligands. As a rule,
good yields are obtained under mild conditions. Selected
examples are shown in Table 22.
Very recently, Cahiez described the cross-coupling be-
tween functionalized alkyl halides and aliphatic Grignard
Cahiez and Moyeux
Table 23. Cobalt-Catalyzed Alkyl-Alkyl Coupling Reaction:
Effect of Iodide Anion on the Course of the Reaction
Entry Catalytic system (5 mol %) Yield (%)
1 CoCl2 traces
2 CoCl2; 4TMEDA 35
3 CoCl2•2LiCl; 4TMEDA 27
4 CoCl2•2LiBr; 4TMEDA 30
5 CoCl2•2LiI; 4TMEDA 79
Table 24. Cobalt-Catalyzed Coupling between Unactivated
Alkyl Halides and Aliphatic Grignard Reagents
a
A mixture of decane and 1-decene was mainly obtained.
reagents in the presence of the ate complex CoCl2•2LiI.50
TMEDA is used as a ligand.
By using cobalt chloride as a catalyst, only traces of cross-
coupling product are obtained (Table 23, entry 1). In the
presence of TMEDA, the yield is better but remains
unsatisfactory (entry 2). Finally, the best results are obtained
by adding both TMEDA and lithium iodide (entry 5). It is
the first example of such a beneficial influence of iodide
anion on the course of a cobalt-catalyzed cross-coupling
reaction.
Various alkyl-alkyl couplings were performed under these
conditions (Table 24). Acyclic or cyclic secondary alkyl
bromides (entries 1-8) gave moderate to good yields of
cross-coupling product. It is important to note that the
influence of steric effects is determinant. Thus, 2-bromopen-
tane gave a better yield than 3-bromopentane (entries 5 and
6). Similarly, the results obtained with various 2-bromoal-
kanes clearly depend on the length of the alkyl chain (entries
1, 3, and 7). It is possible to use alkyl iodides in place of
bromides, but the corresponding chlorides do not react
(entries 2-4).
The reaction has been successfully extended to primary
alkyl bromides (entries 9 and 10). On the other hand, tertiary
alkyl bromides do not react (entry 11). Finally, it should be
noted that secondary or tertiary alkyl Grignard reagents
cannot be used (entries 12 and 13).
The reaction is chemoselective (Table 25), thus, an ester
(entries 1 and 2), a nitrile (entry 3), or even a keto group
(entry 4) are tolerated. It should be underlined that func-
Cobalt-Catalyzed Cross-Coupling Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1451

Table 25. Cobalt-Catalyzed Coupling between Functionalized Scheme 58. Cobalt-Catalyzed Acylation of Diorganozincs
Alkyl Halides and Aliphatic Grignard Reagents

Scheme 56. Cobalt-Catalyzed Chemoselective Alkyl-Alkyl

Coupling Reaction from Functionalized Secondary Alkyl
Halides Table 26. Cobalt-Mediated Reaction between Aryl Bromides
and Carboxylic Acid Anhydrides

Scheme 57. Cobalt-Catalyzed Acylation of Methyl- and

Phenylmagnesium Bromides Entry Aryl bromide R Yield (%)
1 p-MeO-C6H4Br Me 67
2 p-MeO-C6H4Br n-Bu 79
3 p-MeO-C6H4Br Ph 46
4 p-NC-C6H4Br Me 54
5 p-NC-C6H4Br n-Bu 76
6 p-NC-C6H4Br Ph 30a
7 p-EtO2C-C6H4Br Me 62
8 p-EtO2C-C6H4Br n-Bu 78
9 p-EtO2C-C6H4Br Ph 33b
10 p-F-C6H4Br Me 69a
11 p-F3C-C6H4Br Me 61a
12 m-MeO-C6H4Br Me 72
tionalized secondary alkyl halides can also be used success- 13 m-NC-C6H4Br Me 34b
14 m-F3C-C6H4Br Me 71a
fully (Scheme 56).
a
It is the first example of chemoselective alkyl-alkyl 10 mol % CoBr2 was used. b 7.5 mol % CoBr2 was used.
coupling from a functionalized secondary alkyl bromide.
Various para- (entries 1-11) or meta-substituted (entries
7. Acylation 12-14) functionalized aryl bromides react successfully. As
a rule, aliphatic acid anhydrides give good results whereas
7.1. From Organometallic Compounds aromatic acid anhydrides only lead to moderate yields (entries
In 1943, Kharasch showed that methyl- or phenylmagne- 3, 6, and 9).
sium bromides react with mesitoyl chloride in the presence
of cobalt chloride to give the expected ketones in moderate 8. Reductive Cyclization and Heck-Type
yields (Scheme 57).51 It is the first report of a cobalt-catalyzed Reactions
acylation of organometallic compounds.
Fifty years later, Knochel showed that diorganozincs react 8.1. Radical Cyclization
with carboxylic acid chlorides in the presence of cobalt It is well-known that various cobalt(II) complexes react
bromide in a THF/NMP mixture as a solvent (Scheme 58).48 with alkyl or aryl halides via a halogen atom transfer to lead
Excellent yields are obtained with aromatic and aliphatic to the corresponding alkyl or aryl radicals.53,54 Cyclization
carboxylic acid chlorides as well as with oxalyl chloride and via an inter- or an intramolecular trapping of the radical thus
trifluoroacetic anhydride. It is however important to note that generated provides an easy access to various cyclic
three equivalents of diorganozinc compound have to be used compounds.
(i.e., 6 equiv of organometallic reagent).
8.1.1. Intramolecular Radical Cyclization
7.2. Cobalt-Mediated Acylation of Aryl Bromides 8.1.1.1. Radical-Mediated Aryl-Aryl Coupling. Cobalt-
The cobalt-mediated synthesis of aromatic ketones from catalyzed reductive cyclization has been reported for the first
aryl bromides and carboxylic acid anhydrides was reported time by Tiecco55 in 1965 (Scheme 59).
by Périchon52 in 2004 (Table 26). The reaction is performed 8.1.1.2. From δ-Halogeno Acetylenic Compounds. In
in the presence of a catalytic amount of cobalt bromide by 1982, Tada56 reported the reductive cyclization of 2-prop-
using a stoichiometric amount of zinc as a reductant. argyloxyalkyl bromides in the presence of bis(dimethylgly-
1452 Chemical Reviews, 2010, Vol. 110, No. 3 Cahiez and Moyeux

Scheme 59. Cobalt-Catalyzed Cyclization via a Scheme 61. Cobaloxime(I)- or Vitamin B12-Catalyzed
Radical-Mediated Aryl-Aryl Coupling Radical Cyclization of 2-Allyloxyalkyl Bromides

Table 27. Cobaloxime(I)-Catalyzed Reductive Cyclization of

2-Propargyloxyalkyl Bromides Scheme 62. Radical Cyclization Catalyzed by
Cobalt(I)-Salen

Entry R R1 R2 Yield (%)

1 Ph Ph H 85
2 Ph Me H 73
3 Ph H H 78 Table 28. Radical Cyclization Catalyzed by Cobalt(I)-Salen:
4 H -(CH2)4- 64 Influence of the Nature of the Reducing System on the Outcome
5 H -(CH2)3- 48 of the Reaction
* Cobaloxime(I) ) [bis(dimethylglyoximato)(pyridine)cobalt(I)].

Scheme 60. Cobaloxime(I)-Catalyzed Reductive Cyclization

of Halogeno Propargyl Acetals 16

Ratio 19/20 (yield %)

Reaction conditions
20 mol % catalyst, 1 mol % catalyst,
Entry R Zn, DMF, hυ Zn/NH4Cl, DMF
1 H 81/19(80) 1/99(80)
oximato)(pyridine)cobalt(I), generally called cobaloxime(I), 2 Ph 99/1 (90) 3/97(88)
by using a stoichiometric amount of sodium borohydride as
a reductant. Results are reported in Table 27. Scheme 63. Tandem Radical Cyclization and Michael
Addition Reaction
Good yields of exomethylenefuranes are obtained from
primary alkyl halides (entries 1-3). Cyclic alkyl halides lead
to bicyclic compounds in satisfactory yields (entries 4 and
5).
In 1989, Hoffmann57 applied the reaction to the synthesis
of various heterocyclic compounds (Scheme 60).
8.1.1.3. From δ-Halogeno Ethylenic Compounds. In
1984, Pattenden58 described the radical cyclization of 2-al-
lyloxyalkyl bromides using cobaloxime or Vitamin B1259
as a catalyst under the conditions described by Tada (Table
27) for 2-propargyloxyalkyl bromides.56 The organocobalt
species resulting from the cyclization undergoes a β-hydro-
gen elimination to give a bicyclic product in good yields Interestingly, the radical (or the organocobalt species)
(Scheme 61). obtained after the cyclization step can be trapped by using a
A few years later, Jones60 reported a similar reaction reactive Michael acceptor such as an acrylic ester (Scheme
catalyzed by a cobalt-salen complex. A mixture of saturated 63).61
and unsaturated products 17 and 18 is obtained in all cases In 1994, Jones62 performed the reductive cyclization of
(Scheme 62). ortho-amidoiodobenzenes 21 in moderate to good yields in
Thereafter, Giese61 showed that it is possible to form the presence of cobalt chloride by using methylmagnesium
selectively the saturated or the unsaturated products 19 and iodide as a reductant (Scheme 64). A large amount of catalyst
20 by changing the reaction conditions (Table 28). is required (50 mol %).
Cobalt-Catalyzed Cross-Coupling Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1453

Scheme 64. Cobalt-Catalyzed Radical Cyclization of Table 29. Cobalt-Catalyzed Carbocyclization of ortho-Iodoaryl
ortho-Amidoiodobenzenes 21 Ketones and Aldehydes with Alkynes

Scheme 65. CoCl2•dppb-Catalyzed Radical Cyclization of

δ-Halogeno Ethylenic Compounds

It is possible to lower the amount of catalyst by using a

complex cobalt CoCl2(dppb). Thus, Oshima63 achieved the
radical cyclization of δ-halogeno ethylenic compounds in
the presence of CoCl2(dppb) by using trimethylsilylmethyl-
magnesium chloride as a reductant (Scheme 65). Various
cyclopentane derivatives were prepared in good yields. a
Ratio 22/23.
8.1.2. Cyclization via an Intermolecular Radical Addition
Table 30. Intermolecular Cyclization of ortho-Iodoaryl Ketones
In 2003, Cheng64 reported the reaction of various o- and Aldehydes with Acrylic Esters
iodoaryl ketones and aldehydes with alkynes or conjugated
alkenes in the presence of CoI2•dppe and a stoichiometric
amount of zinc as a reductant. The cyclization reaction
proceeds via a tandem carbometalation and intramolecular
1,2-addition to the ketone.
Such a reaction had been previously described under
palladium65 or nickel66 catalysis. Nevertheless, the cobalt-
catalyzed reaction gives higher yields and better regioselec-
tivities under milder conditions. In addition, the reaction
times are shorter (Table 29).
Good to excellent yields are obtained from o-iodoaryl
ketones (Table 29, entries 1-5) or aldehydes (entries 6-9).
When unsymmetrical alkynes are used, a mixture of two
regioisomers 22 and 23 is obtained (entries 2 and 7).
Interestingly, with trimethylsilylpropyne or ethyl trimethyl-
silylpropiolate, the regioselectivity is excellent, since only
compound 22 is formed (entries 3-5 and 8).
A similar reaction takes place by using various acrylic
esters instead of alkynes (Table 30).
It is noteworthy that it is possible to perform a three-
component reaction between o-iodobenzaldehyde, p-tolui-
dine, and alkynes to form heteroatom-substituted derivatives67
(Scheme 66).
β-hydrogen elimination. Interestingly, the cobalt-catalyzed
Heck type reaction is complementary to the classical
8.2. Cobalt-Catalyzed Heck-Type Reactions palladium-catalyzed reaction, since it allows avoidance
The Heck reaction68,69 is a powerful tool for the of this drawback.
elaboration of carbon-carbon bonds. However, the reac- In 1991, Branchaud70 reported the first cobalt-catalyzed
tion still suffers some limitations. Thus, the use of alkyl Heck-type reaction between styrene and various alkyl halides
halides having hydrogen(s) atom(s) in the β-position to (Table 31) in the presence of cobaloxime(II) and a stoichio-
the halide atom leads to poor results, since the intermediate metric amount of zinc as a reductant. It should be noted that
catalytic alkylpalladium species undergoes a very fast exposure to daylight is necessary.
1454 Chemical Reviews, 2010, Vol. 110, No. 3 Cahiez and Moyeux

Scheme 66. Cobalt-Catalyzed Three-Component Cyclization Table 32. CoCl2(dpph)-Catalyzed Heck-Type Reaction
Reaction

Scheme 67. Cobalt-Catalyzed Diastereoselective Heck-Type

Table 31. Cobaloxime-Catalyzed Heck-Type Reaction Reaction

Scheme 68. Mechanism of the CoCl2(dpph)-Catalyzed

Heck-Type Reaction

Primary (Table 31, entries 1-3 and 6) and secondary

(entry 4) alkyl bromides lead to moderate yields whereas
the reaction failed with tertiary alkyl bromides (entry 5).
Later, Oshima reported that excellent yields were obtained
by using CoCl2•dpph as a catalyst and trimethylsilylmethyl-
magnesium chloride as a reducing agent (Table 32).71a
Primary (Table 32, entries 1-3), secondary (entries 4-5
and 8-10), and even tertiary (entry 6) alkyl halides react to
give good to excellent yields. As a rule, alkyl bromides afford
better yields than the corresponding iodides or chlorides
(entries 1-3). Functionalized styrene derivatives can be
coupled efficiently (entries 7-10). It is interesting to note
that the reaction can be achieved stereoselectively (Scheme [E]. The Heck product RCHdCHPh is then obtain from
67). [E] by β-hydrogen elimination. Finally, Co0•dpph [B] is
A detailed mechanistic study was performed (Scheme regenerated from the hydridocobalt species [F] by reduc-
68). tive elimination.
At first, CoCl2•dpph [A] reacts with Me3SiCH2MgCl It is important to note that a mechanism involving a
to give Co0•dpph [B], which is the real catalyst. A single CoI/CoIII couple like this one proposed for the vitamin
monoelectronic transfer to the alkyl halide leads to the B12- or the cobaloxime-mediated reaction60,61 cannot be
CoI species [C] that reacts with Me3SiCH2MgCl to give discarded.
the intermediate [D]. This one reacts with the benzylic The cobalt-catalyzed Heck-type reaction between an
radical accruing from RX to produce the diorganocobalt epoxide and styrene has also been described (Table 33).72
Cobalt-Catalyzed Cross-Coupling Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1455

Table 33. Cobalt-Catalyzed Heck-Type Reaction between Scheme 71. Enantioselective Cobalt-Catalyzed Michael
Epoxides and Styrene Addition to Chalcone

Entry R Yield (%) 24/25

1 Me 57 79:21
2 n-Bu 58 57:43
3 c-C6H11 48 63:37a
4 t-Bu 26 >99:1
Scheme 72. Cobalt-Catalyzed Michael Addition to Various
a
Toluene is used as solvent. Activated Alkenes

Scheme 69. Vitamin B12-Catalyzed Michael-like Addition

Scheme 70. Enantioselective Cobalt-Mediated Michael

Addition to Chalcone

Table 34. Cobalt-Mediated Michael Reaction between Aryl

Halides or Triflates and Activated Alkenes

Entry FG X EWG Yield (%)

1 p-CO2Et Br CO2Et 80
2 p-CO2Et Br CN 85
3 p-CO2Et OTf CO2Et 72a
The reaction leads to moderate yields, and a mixture of 4 p-CN Br CO2Et 70a
isomers is obtained. Thus, except in the case of a hindered 5 p-CN Br CN 77a
epoxide, the stereoselectivity of the reaction is very poor 6 p-CN Cl CO2Et 66
7 p-CN OTf CO2Et 69a
(Table 33, entry 4).
a
20 mol % of catalyst was used.
9. Miscellaneous Reactions
In 2006, a cobalt-mediated Michael addition was per-
formed with alkyl bromides and activated alkenes in the
9.1. Cobalt-Mediated Michael Addition presence of CoI2(dppe) and zinc as a reductant.76 Good to
As mentioned above, alkyl and aryl radicals are generated excellent yields are obtained (Scheme 72).
by reacting an alkyl or an aryl halide with various cobalt(II) Finally, a cobalt-mediated Michael addition from aryl
complexes (see section 8). halides or triflates and activated alkenes has also been
Diastereoselective conjugate addition of the radicals thus described by Périchon (Table 34).77 Aryl bromides (entries
obtained to activated alkenes has been studied for the first 1, 2, 4, and 5), chlorides (entry 6), or triflates (entries 3 and
time by Giese73 in 1992 (Scheme 69). The reaction is 7) give similar yields. The reaction is chemoselective; thus,
performed by using a catalytic amount of vitamin B12 and ester and cyano groups are tolerated.
zinc as a reductant.
In 1998, Pfaltz74 obtained a poor yield, but a promising 9.2. Allylation of 1,3-Dicarbonyl Compounds
high enantiomeric excess, by reacting tert-butyl malonate The allylation of 1,3-dicarbonyl compounds under cobalt
with chalcone in the presence of cobalt acetate and the chiral catalysis has been reported by Iqbal.78 The reaction can be
macroheterocyclic ligand 28 (Scheme 70). performed under neutral conditions, contrary to the pal-
A better yield, but a moderate enantiomeric excess, was ladium-79 or the molybdenum-catalyzed80 reactions. There-
obtained by Wu and Zhou75 by reacting ethyl malonate in fore, the use of base-sensitive substrates like 30 is possible.
the presence of a similar catalytic system (Scheme 71). Selected examples are given in Table 35.
1456 Chemical Reviews, 2010, Vol. 110, No. 3 Cahiez and Moyeux

Table 35. Cobalt-Catalyzed Allylation of 1,3-Dicarbonyl Scheme 73. Cobalt-Catalyzed Allylic C-H Bond Amination
Compounds with Allylic Acetates

Scheme 74. Cobalt-Catalyzed Benzylic C-H Bond

Amination

Table 36. Cobalt-Catalyzed Allylation of 1,3-Dicarbonyl

Compounds with Allylic Alcohol

Entry Z R R1 R2 Catalyst 33/34 Yield (%)

1 H Me Me Pr CoIIIDMG 1/8 61
2 Ac Me Me Pr CoCl2 1/3 68
3 H OMe Ph Et CoIIIDMG 1/9 64
4 Ac OMe Ph Et CoCl2 1/3 55
5 H OMe Me Bu CoIIIDMG 100/0 53
6 Ac OMe Me Bu CoCl2 1/1.5 24 Table 37. Benzylic C-H Amination (see Scheme 74)
Yield (%)

Moderate to good yields are obtained. Regio- and stereo- Entry Substrate Catalysta Amine Imine
selectivity heavily depend on the nature of the substrate; a 1 Toluene A 25
2 Isopropylbenzene A 53
mixture of isomers is often obtained. 3 Cyclohexylbenzene B 28
This reaction has then been extended to allylic alcohols81 4 Ethylbenzene A 25 19
(Table 36). 5 Diphenylmethane B 55 11
6 Fluorene C 41
9.3. Activation of C-H Bonds: C-N Coupling a
CoII porphyrin A: R2 ) H, R3 ) p-Cl-C6H4; B: R2 ) H, R3 ) Ph;
Reactions C: R2 ) Et, R3 ) H.

Transition metal-mediated activation and functionalization

of C-H bonds is currently a very challenging research area.82
Cobalt-porphyrin-catalyzed C-H bond aminations have
been widely investigated by Cenini.83 In 1999, the amination
of an allylic C-H bond was described84a (Scheme 73).
However, only moderate yields are obtained and the scope
of the reaction is limited to cyclohexene.
The reaction has then been extended to benzylic C-H
bonds.84b,c According to the nature of the substrate, an amine
or an imine can be obtained in moderate yields (Scheme 74
and Table 37).
9.4. Three-Component Reactions
Multicomponent reactions have emerged as a useful
synthetic tool in the past decade.84 As illustrated above, cobalt
salts can be employed as catalysts to achieve such reactions.
9.4.1. Synthesis of β-Acetamido Ketones
In 1994, Iqbal85 described a three component reaction
between a ketone, an aldehyde, and acetonitrile. Various
β-acetamido ketones were prepared in moderate to good
yields (Scheme 75).
Similarly, the reaction between a 1,3-dicarbonyl com-
pound, an aldehyde, and acetonitrile86 affords β-acetami-
dodiketones in moderate to good yields (Scheme 76).
It should be noted that aromatic aldehydes lead to
acetamidodiketones when the reaction is performed under
an inert atmosphere whereas the Knoevenagel adduct is
formed in the presence of atmospheric oxygen (Scheme 77).
On the other hand, aliphatic aldehydes behave differently,
since the acetamidodiketones are obtained only in the
presence of atmospheric oxygen. A complex mixture of
products is formed when the reaction takes place under an
inert atmosphere (Scheme 77).
Cobalt-Catalyzed Cross-Coupling Reactions
Scheme 75. Synthesis of β-Acetamido Ketones via a
Cobalt-Catalyzed Three-Component Reaction
Scheme 76. Synthesis of β-Acetamido Diketones via a
Cobalt-Catalyzed Three-Component Reaction
Scheme 77. Cobalt-Catalyzed Reaction between
1,3-Dicarbonyl Compounds, Aldehydes, and Acetonitrile:
Influence of the Presence of Oxygen on the Course of the
Reaction
This methodology has been applied to the synthesis of
furans and γ-lactams (Scheme 78).
9.4.2. Synthesis of Homoallylsilanes
In 2003, Oshima87 described a cobalt-catalyzed three-
component reaction between an alkyl halide, a 1,3-diene, and
trimethylsilylmethylmagnesium chloride. Various homoal-
lylsilanes have been synthesized in good to excellent yields
from primary, secondary, or tertiary alkyl halides (Scheme
79).
The following catalytic cycle was proposed (Scheme 80).
Chemical Reviews, 2010, Vol. 110, No. 3 1457
Scheme 78. Synthesis of Furans and γ-Lactams via a
Cobalt-Catalyzed Three-Component Reaction
Scheme 79. Synthesis of Homoallylsilanes via a
Cobalt-Catalyzed Three-Component Reaction
Scheme 80. Tentative Mechanism for the Formation of
Homoallylsilanes via a Cobalt-Catalyzed Three-Component
Reaction
At first, Me3SiCH2MgCl reduces cobalt(II) chloride into
an active Co0 species [A]. This one reacts with the alkyl
halide via a single electron transfer to form the corresponding
alkyl radical and the CoI complex [B]. Afterward, the alkyl
radical adds to the diene to give a new radical, which
combines with [B] to form the CoII complex [C].
Me3SiCH2MgCl then reacts with [C] to afford the diorga-
nocobalt [D], which undergoes a reductive elimination,
leading to the homosilane [E] and to the complex [A].
9.5. Coupling of Alkenes and Alkynes
Copper-mediated reactions have been extensively used for
the preparation of conjugated diynes.88 The Glaser89 and
Eglington90 procedures for the homocoupling of terminal
alkynes, or the Cadiot-Chodkiewicz91 procedure for the
1458 Chemical Reviews, 2010, Vol. 110, No. 3 Cahiez and Moyeux

Scheme 81. Cobalt-Catalyzed Dimerization of Terminal Scheme 83. Cobalt-Catalyzed Reductive Coupling of
Alkynes Internal Alkynes with Conjugated Alkenes

Scheme 84. Cobalt-Catalyzed Alder-Ene Reaction

Scheme 82. Cobalt-Catalyzed Reductive Dimerization of
Activated Alkenes

Scheme 85. Cobalt-Catalyzed Intramolecular Reductive

Coupling of Activated Alkenes and Alkynes

coupling of a terminal alkyne with a 1-halo-1-alkyne, are

probably the most popular coupling reactions for the
synthesis of 1,3-diynes. A cobalt-catalyzed reaction can also
be used to prepare diynes.92 Thus, in 2001, Krafft93 reported
the cobalt-catalyzed dimerization of terminal alkynes in
excellent yields (Scheme 81). It is noteworthy that, in the
case of conjugated enynes, no trace of the Pauson-Khand94
product is detected.
The cobalt-catalyzed reductive dimerization of conjugated
alkenes was described by Cheng in 2004.95 Excellent results
are obtained from various activated alkenes or styrenes
(Scheme 82). It should be noted that such reactions have
been previously performed in the presence of stoichiometric96
or catalytic97 amounts of cobalt salts, but the yields are very
poor. various exomethylenecyclopentane derivatives in good to
excellent yields.
The same methodology can be successfully applied to the It is noteworthy that it is possible to perform a cobalt-
reductive coupling of internal alkynes with conjugated catalyzed intramolecular reductive coupling/lactonization of
alkenes.98 Interestingly, the reaction can be highly regio- and acrylates with propargyl alcohols (Scheme 86).
stereoselective. Thus, only one product is obtained from When propargyl amine derivatives are used, the reaction
conjugated acetylenic esters and phenylacetylene derivatives leads to the formation of six-membered cyclic lactams
(Scheme 83). (Scheme 87).
Nonactivated terminal olefins were also used.99 As a rule,
good yields and satisfactory stereoselectivities can be
obtained under mild conditions (Scheme 84).
9.6. Various Reactions
Very recently, Cheng100 extended the reaction to the In 2005, Oshima101 reported the syn-hydrophosphination
intramolecular reductive coupling of activated alkenes with of alkynes under cobalt catalysis. It should be noted that
alkynes (Scheme 85). This reaction allows the synthesis of lanthanide-,102 palladium-, or nickel-catalyzed103 reactions
Cobalt-Catalyzed Cross-Coupling Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1459

Scheme 86. Cobalt-Catalyzed Intramolecular Reductive Scheme 88. Cobalt-Catalyzed syn-Hydrophosphination of

Coupling/Lactonization of Acrylates with Propargyl Alcohols Alkynes: Application to the Wittig Reaction

Scheme 89. Cobalt-Catalyzed Synthesis of 1,2-Di- and

Scheme 87. Cobalt-Catalyzed Intramolecular Reductive 1,1,2-Trisilylethylenes
Coupling/Lactonization of Acrylates with Propargyl Amines
Derivatives

Scheme 90. Synthesis of Thioethers via Cobalt-Catalyzed

C-S Coupling Reaction
Table 38. Cobalt-Catalyzed syn-Hydrophosphination of
Alkynes

Entry R R1 Yield (%) 35/36

1 pent Me 82 66/34
2 Ph Me 74 82/18
3 o-anisyl Me 41 80/20
4 t-Bu H 81 100/0
5 Ph H 70 89/11
6 Et3Si H 62 94/6

have also been described. However, the stereoselectivity of

these reactions is highly dependent on the nature of the
substrate. On the contrary, the cobalt-catalyzed reaction
mainly leads to the syn-addition product 35 (Table 38).
This reaction is of particular interest, since the resulting
alkenylphosphine can be used for performing Wittig reactions
(Scheme 88).
The same year, Oshima104 reported the preparation of 1,2- reaction conditions and low catalyst loading are necessary.
di- and 1,1,2-trisilylethylenes via a cobalt-mediated reaction Moreover, the use of sophisticated phosphine ligand is
of dibromomethylsilanes with trialkylsilylmethylmagnesium avoided.
reagents. Disilylethylenes are usually synthesized by hy- Under these conditions, ethyl 2-iodoacrylate can also be
drosilylation of silylacetylenes105 or disilylation of acety- coupled successfully (Scheme 91).
lenes.106 Contrary to these reactions, the cobalt-mediated
procedure is highly regioselective and stereoselective (Scheme 10. Conclusion
89).
In 2006, Cheng107 reported the cobalt-mediated synthesis Sustainable development prompted organic chemists to
of thioethers from thiols and aryl halides (Scheme 90). This look for more ecocompatible and more economical transition
coupling method is advantageous compared to copper-, metal-catalyzed procedures. Thus, in the last ten years, a
palladium-, and nickel-mediated reactions,108 since mild growing number of iron- or manganese-catalyzed reactions
1460 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 91. Cobalt-Catalyzed Cross-coupling between
Thiophenol and 2-Iodoethyl Acrylate
were proposed to replace the older palladium and nickel-
catalyzed cross-coupling procedures. Of course, in the
framework of sustainable development, cobalt is less inter-
esting than iron or manganese; however, it compares
favorably to nickel and palladium. Thus, cobalt can be an
interesting alternative when iron or manganese cannot be
used. On the other hand, it should be underlined that several
reactions performed under cobalt catalysis are specific to this
metal. For instance, iron-catalyzed cross coupling between
aryl Grignard reagents and functionalized secondary alkyl
bromides generally failed whereas excellent yields are
obtained under cobalt catalysis. Thus, in the future, it should
be possible to develop more distinct reactions using cobalt-
based catalytic systems.
In spite of the numerous interesting reports mentioned in
this review, cobalt-catalyzed reactions are still in their
infancy. In the future, they should be more extensively
developed and could take a significant place in the renewal
of transition metal-catalyzed reactions.
11. Acknowledgments
We thank the CNRS for financial support. Gérard Cahiez
thanks current and former members of his laboratory for their
contribution to the development of cobalt-catalyzed coupling
reactions. Their names appear in the list of references. We
also thank Julien Buendia and Dr. Olivier Gager for
proofreading.
12. References
(1) Kharasch, M. S.; Fuchs, C. F. J. Am. Chem. Soc. 1941, 63, 2316.
(2) For reviews, see: (a) Yorimitsu, H.; Oshima, K. Pure Appl. Chem.
2006, 78, 441. (b) Shinokubo, H.; Oshima, K. Eur. J. Org. Chem.
2004, 2081. (c) Iqbal, J.; Mukhopadhyay, M.; Mandal, A. K. Synlett
1997, 876. (d) Hess, W.; Treutwein, J.; Hilt, G. Synthesis 2008, 3537.
(e) Omae, I. Appl. Organomet. Chem. 2007, 21, 318. (f) Gosmini,
C.; Begouin, J.-M.; Moncomble, A. Chem. Commun. 2008, 28, 3221.
(3) (a) Khand, I. U.; Knox, G. R.; Pauson, P. L.; Watts, W. E.; Foreman,
M. I. J. Chem. Soc. Perkin Trans. I 1973, 977. (b) Pauson, P. L.
Tetrahedron 1985, 41, 5855. (c) Khand, I. U.; Pauson, P. L. J. Chem.
Soc., Chem. Commun. 1974, 379. For a review, see: (d) Struebing,
D.; Beller, M. Top. Organomet. Chem. 2006, 18, 165.
(4) (a) Pino, P.; Piacenti, F.; Bianchi, M. In Organic Synthesis Via Metal
Carbonyls; Wender, I., Pino, P., Eds.; Wiley: New York, 1977; Vol.
2, Chapter 2. (b) Pruett, R. L. In AdVances in Organometallic
Chemistry; Stone, F. G. A., West, R., Eds.; Academic Press: New
York, 1979; Vol. 17, p 1. (c) Paulik, F. E. Catal. ReV. 1972, 6, 49.
(d) Orchin, M.; Replius, W. Catal. ReV. 1972, 6, 85. (e) Heck, R. F.;
Breslow, D. S. J. Am. Chem. Soc. 1961, 83, 4023. (f) Mirback, M. F.
J. Organomet. Chem. 1984, 265, 205. (g) Klinger, R. J.; Rathke,
J. W. J. Am. Chem. Soc. 1994, 116, 4772. (h) Ojima, I.; Tsai, C.-Y.;
Tzamarioudaki, M.; Bonafoux, D. Org. React. 2000, 56, 1. (i) Wiese,
K.-D.; Obst, D. Top. Organomet. Chem. 2006, 18, 1.
(5) Leading references: (a) Zhu, S.; Ruppel, J. V.; Lu, H.; Wojtas, L.;
Zhang, X. P. J. Am. Chem. Soc. 2008, 130, 5042. (b) Penoni, A.;
Wanke, R.; Tollari, S.; Gallo, E.; Musella, D.; Ragaini, F.; Demartin,
F.; Cenini, S. Eur. J. Inorg. Chem. 2003, 1452. (c) Ikeno, T.; Iwakura,
I.; Yabushita, S.; Yamada, T. Org. Lett. 2002, 4, 517. (d) Uchida,
T.; Saha, B.; Katsuki, T. Tetrahedron Lett. 2001, 42, 2521. (e) Kanai,
H.; Matsuda, H. J. Mol. Catal. 1985, 29, 157. (f) Nakamura, A.;
Konishi, A.; Tatsuno, Y.; Otsuka, S. J. Am. Chem. Soc. 1978, 100,
3443. (g) Fantauzzi, S.; Gallo, E.; Rose, E.; Raoul, N.; Caselli, A.;
Issa, S.; Ragaini, F.; Cenini, S. Organometallics 2008, 27, 6143, and
references cited therein.
Cahiez and Moyeux
(6) (a) Vollhardt, K. P. C. Angew. Chem., Int. Ed. 1984, 23, 539. (b)
Grotjahn, D. B.; Vollhardt, K. P. C. J. Am. Chem. Soc. 1986, 108,
2091. (c) Butenschon, H.; Winkler, M.; Vollhardt, K. P. C. Chem.
Commun. 1986, 388. (d) Johnson, E. P.; Vollhardt, K. P. C. J. Am.
Chem. Soc. 1991, 113, 381. For reviews, see: (e) Wu, M. S.;
Shanmugansundaram, M.; Cheng, C.-H. Chem. Commun. 2003, 718.
(f) Aubert, C.; Buisine, O.; Petit, M.; Slowinski, F.; Malacria, M.
Pure Appl. Chem. 1999, 71, 1463. (g) Bonãga, L. V. R.; Zhang, H.-
C.; Moretto, A. F.; Ye, H.; Gauthier, D. A.; Li, J.; Leo, G. C.;
Maryanoff, B. E. J. Am. Chem. Soc. 2005, 127, 3473. (h) Agenet,
N.; Gandon, V.; Vollhardt, K. P. C.; Malacria, M.; Aubertand, C.
J. Am. Chem. Soc. 2007, 129, 8860. (i) Scheuermann née Taylor,
C. J.; Ward, B. D. New J. Chem. 2008, 32, 1850. For a recent
communication, see: (j) Geny, A.; Agenet, N.; Iannazzo, L.; Malacria,
M.; Aubert, C.; Gandon, V. Angew. Chem., Int. Ed. 2009, 48, 1810,
and references cited therein.
(7) Chen, K. C.; Rayabarapu, D. K.; Wang, C. C.; Cheng, C.-H. J. Org.
Chem. 2001, 66, 8804, and references cited therein.
(8) Gilman, H.; Lichtenwalter, M. J. Am. Chem. Soc. 1939, 61, 957.
(9) (a) Morizur, J. P.; Pallaud, R. C. R. Acad. Sci. Paris 1962, 254, 1093.
(b) Morizur, J. P. Bull. Soc. Chim. Fr. 1964, 1331.
(10) Kharasch, M. S.; Fuchs, C. F. J. Am. Chem. Soc. 1943, 65, 504.
(11) (a) Uemura, S.; Fukuzawa, S.-I. Tetrahedron Lett. 1982, 23, 1181.
(b) Uemura, S.; Fukuzawa, S.-I.; Path, S. R. J. Organomet. Chem.
1983, 243, 9.
(12) (a) Cahiez, G.; Avedissian, H. Tetrahedron Lett. 1998, 39, 6159.
The beneficial influence of NMP on iron-catalyzed alkenylation
reactions was also described: (b) Cahiez, G.; Marquais, S. Tetrahe-
dron Lett. 1996, 37, 1773. (c) Cahiez, G.; Marquais, S. Pure Appl.
Chem. 1996, 68, 53. (d) Cahiez, G.; Avedissian, H. Synthesis 1998,
1199. NMP has also a favorable effect on the Cu-catalyzed alkylation
of organomagnesium and organomanganese halides. (e) Cahiez, G.;
Marquais, S. Synlett 1993, 45. See also ref.12c
(13) Shirakawa, E.; Imazaki, Y.; Hayashi, T. Chem. Lett. 2008, 37, 654.
(14) Amatore, M.; Gosmini, C.; Périchon, J. Eur. J. Org. Chem. 2005,
989.
(15) Buriez, O.; Nédélec, J.-Y.; Périchon, J. J. Electroanal. Chem. 2001,
506, 162.
(16) (a) Korn, T. J.; Cahiez, G.; Knochel, P. Synlett 2003, 12, 1892. (b)
Ohmiya, H.; Yorimitu, H.; Oshima, K. Chem. Lett. 2004, 33, 1240.
(17) Hatakeyama, T.; Hashimoto, S.; Ishizuka, K.; Nakamura, M. J. Am.
Chem. Soc. 2009, 131, 11949.
(18) Korn, T. J.; Knochel, P. Angew. Chem., Int. Ed. 2005, 44, 2947.
(19) 4-Fluorostyrene has already been used as an additive to perform cross-
coupling reactions:Jensen, A. E.; Knochel, P. J. Org. Chem. 1988,
53, 2390. See also: Johnson, J. B.; Rovis, T. Angew. Chem., Int. Ed.
2008, 47, 840.
(20) (a) Korn, T.; Schade, M.; Schade, S.; Knochel, P. Org. Lett. 2006,
8, 725. (b) Korn, T. J.; Schade, M. A.; Cheemala, M. N.; Wirth, S.;
Guevara, S. A.; Cahiez, G.; Knochel, P. Synthesis 2006, 3547.
(21) For the use of aryl fluorides, see: (a) Cahiez, G.; Lepifre, F.;
Ramiandrasoa, P. Synthesis 1999, 2138. (b) Dankwardt, J. W. J. J.
Organomet. Chem. 2005, 690, 932, and references cited therein. (c)
Bahmanyar, S.; Borer, B. C.; Kim, Y. M.; Kurtz, D. M.; Yu, S. Org.
Lett. 2005, 7, 1011. (d) Saeki, T.; Takashima, Y.; Tamao, K. Synlett
2005, 1771. (e) Ackermann, L.; Born, R.; Spatz, J. H.; Meyer, D.
Angew. Chem., Int. Ed. 2005, 44, 7216.
(22) For the use of aryl tosylates, see: (a) Terao, J.; Watanabe, H.; Ikumi,
A.; Kuniyasu, H.; Kambe, N. J. Am. Chem. Soc. 2002, 124, 4222.
(b) Netherton, M. R.; Fu, G. C. Angew. Chem., Int. Ed. 2002, 41,
3910. (c) Huang, X.; Anderson, K. W.; Zim, D.; Jiang, L.; Klapars,
A.; Buchwald, S. L. J. Am. Chem. Soc. 2003, 125, 6653. (d) Tang,
Z.-Y.; Hu, Q.-S. J. Am. Chem. Soc. 2004, 126, 3058, and references
cited therein. (e) Limmert, M. E.; Roy, A. H.; Hartwig, J. F. J. Org.
Chem. 2005, 70, 9364, and references cited therein.
(23) (a) Amatore, M.; Gosmini, C. Angew. Chem., Int. Ed. 2008, 47, 2089.
Very recently, Gosmini also described the formation of symmetrical
biaryls under similar conditions. See: (b) Moncomble, A.; Le Floch,
P.; Gosmini, C. Chem.sEur. J. 2009, 15, 4770.
(24) Kharasch, M. S.; Fuchs, C. F. J. Am. Chem. Soc. 1945, 10, 292.
(25) Cahiez, G.; Avedisssian, H. Unpublished results.
(26) Avedissian, H.; Bérillon, L.; Cahiez, G.; Knochel, P. Tetrahedron
Lett. 1998, 39, 6163.
(27) Kamachi, T.; Kuno, A.; Matsuno, C.; Okamoto, S. Tetrahedron Lett.
2004, 45, 4677.
(28) Affo, W.; Ohmiya, H.; Fujioka, T.; Ikeda, Y.; Nakamura, T.;
Yorimitsu, H.; Oshima, K.; Inamura, Y.; Mizuta, T.; Miyoshi, K.
J. Am. Chem. Soc. 2006, 128, 8068.
(29) Tsuji, T.; Yorimitsu, H.; Oshima, K. J. Am. Chem. Soc. 2006, 128,
1886.
(30) Ohmiya, H.; Yorimitsu, H.; Oshima, K. Org. Lett. 2006, 8, 3093.
(31) Davies, D. I.; Done, J. N.; Hey, D. H. J. Chem. Soc. C 1969, 2019.
Cobalt-Catalyzed Cross-Coupling Reactions
(32) Wakabayashi, K.; Yorimitsu, H.; Oshima, K. J. Am. Chem. Soc. 2001,
123, 5374.
(33) Someya, H.; Kondoh, A.; Sato, A.; Ohmiya, H.; Yorimitsu, H.;
Oshima, K. Synlett 2006, 18, 3061.
(34) (a) Tamao, K.; Nakajima, T.; Kumada, M. Organometallics 1984,
3, 1655. (b) Fleming, I.; Henning, R.; Plaut, H. J. Chem. Soc., Chem.
Commun. 1984, 29.
(35) Ohmiya, H.; Wakabayashi, K.; Yorimitsu, H.; Oshima, K. Tetrahe-
dron 2006, 62, 2207.
(36) Kobayashi, Y.; Nakata, K.; Ainai, T. Org. Lett. 2005, 7, 183.
(37) Cahiez, G.; Chaboche, C.; Duplais, C.; Moyeux, A. Org. Lett. 2009,
11, 277.
(38) Hamaguchi, H.; Uemura, M.; Yasui, H.; Yorimitsu, H.; Oshima, K.
Chem. Lett. 2008, 37, 1178.
(39) Gomes, P.; Gosmini, C.; Périchon, J. Org. Lett. 2003, 5, 1043.
(40) (a) Durandetti, M.; Nédélec, J.-Y.; Périchon, J. J. Org. Chem. 1996,
61, 1748. (b) Gomes, P.; Gosmini, C.; Périchon, J. J. Org. Chem.
2003, 68, 1142. (c) Gomes, P.; Gosmini, C.; Périchon, J. Fr. Pat.
Appl. 01/08808, July 3, 2001.
(41) (a) Mizutani, K.; Yorimitsu, H.; Oshima, K. Chem. Lett. 2004, 33,
832. (b) Yasui, H.; Mizutani, K.; Yorimitsu, H.; Oshima, K.
Tetrahedron 2006, 62, 1410.
(42) Kharasch, M. S.; Lambert, F. L.; Urry, W. H. J. Org. Chem. 1945,
10, 298.
(43) Schlubach, H. H.; Franzen, V. I. Liebigs Ann. Chem. 1951, 572, 116.
(44) Black, H. K.; Horn, D. H. S.; Weedon, B. C. L. J. Chem. Soc. C
1954, 1704.
(45) Kuno, A.; Saino, N.; Kamachi, T.; Okamoto, S. Tetrahedron Lett.
2006, 47, 2591.
(46) Someya, H.; Ohmiya, H.; Yorimitsu, H.; Oshima, K. Org. Lett. 2007,
9, 1565.
(47) Shirakawa, E.; Sato, T.; Imazaki, Y.; Kimura, T.; Hayashi, T. Chem.
Commun. 2007, 4513.
(48) Reddy, C. K.; Knochel, P. Angew. Chem., Int. Ed. 1996, 35, 1700.
(49) (a) Tsuji, T.; Yorimitsu, H.; Oshima, K. Angew. Chem., Int. Ed. 2002,
41, 21. (b) Ohmiya, H.; Tsuji, T.; Yorimitsu, H.; Oshima, K.
Chem.sEur. J. 2004, 10, 5640.
(50) Cahiez, G.; Chaboche, C.; Giulliani, A.; Duplais, C.; Moyeux, A.
AdV. Synth. Catal. 2008, 350, 1484.
(51) Kharasch, M. S.; Morrison, R.; Urry, W. H. J. Am. Chem. Soc. 1944,
66, 368.
(52) (a) Kazmierski, I.; Bastienne, M.; Gosmini, C.; Paris, J.-M.; Périchon,
J. J. Org. Chem. 2004, 69, 936. A similar reaction has been previously
reported from arylzinc halides prepared Via an electrochemical
procedure. (b) Fillon, H.; Gosmini, C.; Périchon, J. Tetrahedron 2003,
59, 8199.
(53) For a general review of transition metal-promoted free-radical
reactions in organic synthesis, see: Iqbal, J.; Bhatia, B.; Nayyar, N. K.
Chem. ReV. 1994, 94, 519.
(54) For reviews on cobalt-mediated radical reactions in organic synthesis,
see: (a) Bhandal, H.; Patel, V. F.; Pattenden, G.; Russel, J. J. J. Chem.
Soc., Perkin Trans. I 1990, 2691. (b) Pattenden, G. Chem. Soc. ReV.
1988, 17, 361.
(55) Tiecco, M. Chem. Commun. 1965, 555.
(56) (a) Okabe, M.; Abe, M.; Tada, M. J. Org. Chem. 1982, 47, 1775.
(b) Okabe, M.; Tada, M. J. Org. Chem. 1982, 47, 5382.
(57) Last, K.; Hoffmann, H. M. R. Synthesis 1989, 901.
(58) (a) Ladlow, M.; Pattenden, G. Tetrahedron Lett. 1984, 25, 4317. (b)
Bhandhal, H.; Pattenden, G.; Russel, J. J. Tetrahedron Lett. 1986,
27, 2299. (c) Begley, M. J.; Ladlow, M.; Pattenden, G. J. Chem.
Soc., Perkin Trans. I 1988, 1095.
(59) For a review on the use of Vitamin B12 as a catalyst, see: Schrauzer,
G. N. Angew. Chem., Int. Ed. 1976, 15, 417.
(60) Clark, A. J.; Jones, K. Tetrahedron Lett. 1989, 30, 1989.
(61) Giese, B.; Erdmann, P.; Göbel, T.; Springer, R. Tetrahedron Lett.
1989, 30, 1989.
(62) (a) Clark, A. J.; Davies, D. I.; Jones, K. J.; Millbanks, C. J. Chem.
Soc., Chem. Commun. 1994, 41. A single example of a similar
reaction has been described by Knochel: (b) Kneisel, F. F.; Monguchi,
Y.; Knapp, K. M.; Zipse, H.; Knochel, P. Tetrahedron Lett. 2002,
43, 4875.
(63) Fujioka, T.; Nakamura, T.; Yorimitsu, H.; Oshima, K. Org. Lett. 2002,
4, 2257.
(64) (a) Chang, K.-J.; Rayabarapu, D. K.; Cheng, C.-H. Org. Lett. 2003,
5, 3963. (b) Chang, K.-J.; Rayabarapu, D. K.; Cheng, C.-H. J. Org.
Chem. 2004, 69, 4781.
(65) (a) Quan, L. G.; Gevorgyan, V.; Yamamoto, Y. J. Am. Chem. Soc.
1999, 121, 3545. (b) Quan, L. G.; Gevorgyan, V.; Yamamoto, Y.
J. Am. Chem. Soc. 1999, 121, 9485. (c) Quan, L. G.; Gevorgyan,
V.; Yamamoto, Y. J. Tetrahedron Lett. 1999, 40, 4089.
(66) (a) Rayabarapu, D. K.; Cheng, C.-H. Chem. Commun. 2002, 942.
(b) Rayabarapu, D. K.; Yang, C.-H.; Cheng, C.-H. J. Org. Chem.
2003, 68, 6726.
Chemical Reviews, 2010, Vol. 110, No. 3 1461
(67) Liu, C.-C.; Korivi, R. P.; Cheng, C.-C. Chem.sEur. J. 2008, 14,
9503.
(68) (a) Mizorocki, T.; Mori, K.; Ozaki, A. Bull. Chem. Soc. Jpn. 1971,
44, 581. (b) Heck, R. F.; Nolley, J. P., Jr. J. Org. Chem. 1972, 37,
2320.
(69) For reviews see: (a) Bräse, S.; De Meijere, A. In Metal Catalyzed
Cross-Coupling Reactions; Diederich, F., Stang, P. J., Eds.; Wiley-
VCH: Weinheim, 1998; Chapter 3. (b) Link, J. T.; Overman, L. E.
In Metal Catalyzed Cross-Coupling Reactions; Diederich, F., Stang,
P. J., Eds.; Wiley-VCH: Weinheim, 1998; Chapter 6. (c) Bräse, S.;
De Meijere, A. In Metal Catalyzed Cross-Coupling Reactions, 2nd
ed.; Diederich, F., Stang, P. J., Eds.; Wiley-VCH: Weinheim, 2004;
Chapter 5. (d) Heck, R. F. Org. React. 1982, 27, 345. (e) Beletskaya,
I. P.; Cheprakov, A. V. Chem. ReV. 2000, 100, 3009. (f) Crips, G. T.
Chem. Soc. ReV. 1998, 27, 427. (g) de Meijere, A.; Meyer, F. E.
Angew. Chem., Int. Ed. 1994, 33, 2379. (h) Cabri, W.; Candiani, I.
Acc. Chem. Res. 1995, 28, 2. (i) Beller, M.; Riermeier, T. H.; Stark,
G. In Transition Metals for Organic Synthesis; Beller, M., Bolm,
C., Eds.; Wiley-VCH: Weinheim, 1998; Vol. 1, Chapter 2.13. (j)
Handbook of Organopalladium Chemistry for Organic Synthesis;
Negishi, E., Ed.; Wiley: New York, 2002; Vol. 1, Chapter IV.2.
(70) Branchaud, B. P.; Detlefsen, W. D. Tetrahedron Lett. 1991, 32, 6273.
(71) (a) Ikeda, Y.; Nakamura, T.; Yorimitsu, H.; Oshima, K. J. Am. Chem.
Soc. 2002, 124, 6514. (b) Recently, Bao reported that a Heck-type
reaction can be performed in the presence of cobalt nanospheres: Zhou,
P.; Li, Y.; Sun, P.; Zhou, J.; Bao, J. Chem. Commun. 2007, 1418.
(72) Ikeda, Y.; Yorimitsu, H.; Shinokubo, H.; Oshima, K. AdV. Synth.
Catal. 2004, 346, 1631.
(73) Erdmann, P.; Schäfer, J.; Springer, R.; Zeitz, H.-G.; Giese, B. HelV.
Chim. Acta 1992, 75, 638.
(74) End, N.; Macko, L.; Zehnder, M.; Pfaltz, A. Chem.sEur. J. 1998,
4, 818.
(75) Chen, C.; Zhu, S.-F.; Wu, X.-Y.; Zhou, Q.-L. Tetrahedron: Asym-
metry 2006, 17, 2761.
(76) Shukla, P.; Hsu, Y.-C.; Cheng, C.-H. J. Org. Chem. 2006, 71, 655.
(77) Amatore, M.; Gosmini, C.; Périchon, J. J. Org. Chem. 2006, 71, 6130.
(78) (a) Bhatia, B.; Reddy, M. M.; Iqbal, J. Tetrahedron Lett. 1993, 34,
6301. (b) Maikap, G. C.; Reddy, M. M.; Mukhopadhyay, M.; Bhatia,
B.; Iqbal, J. Tetrahedron 1994, 50, 9145.
(79) (a) Trost, B. M. Tetrahedron 1977, 33, 2615. (b) Trost, B. M. Acc.
Chem. Res. 1980, 13, 385. (c) Heck, R. F. Palladium reagents in
organic synthesis; Academic Press: London, 1985. (d) Godleski, S. A.
ComprehensiVe Organic Synthesis; Trost, B. M., Ed.; Pergamon: New
York, 1991; Vol. 4, p 585.
(80) (a) Trost, B. M.; Lautens, M. J. Am. Chem. Soc. 1992, 104, 5543.
(b) Trost, B. M.; Lautens, M. J. Am. Chem. Soc. 1983, 105, 3343.
(c) Trost, B. M.; Merlic, C. A. J. Am. Chem. Soc. 1990, 112, 9590.
(81) Mukhopadhyay, M.; Iqbal, J. Tetrahedron Lett. 1995, 36, 6761.
(82) For a review on C-H activation by metal complexes, see: Shilov,
A. E.; Shul’pin, G. B. Chem. Rev. 1997, 97, 2879.
(83) (a) Cenini, S.; Tollari, S.; Penoni, A.; Cereda, C. J. Mol. Catal. A
1999, 135. (b) Cenini, S.; Gallo, E.; Penoni, A.; Ragaini, F.; Tollari,
S. Chem. Commun. 2000, 2265. (c) Ragaini, F.; Penoni, A.; Gallo,
E.; Tollari, S.; Li Gotti, C.; Lapadula, M.; Mangioni, E.; Cenini, S.
Chem.sEur. J. 2003, 9, 249.
(84) For reviews, see: (a) Domling, A.; Ugi, I. Angew. Chem., Int. Ed.
2000, 39, 3168. (b) Terret, N. K.; Gardner, M.; Gordon, D. W.;
Kobylecki, R. J.; Steele, J. Tetrahedron 1995, 51, 8135. (c) Thomson,
L. A.; Ellman, J. A. Chem. ReV. 1996, 96, 555. (d) Ellman, J. A.
Acc. Chem. Res. 1996, 29, 132.
(85) (a) Bhatia, B.; Reddy, M. M.; Iqbal, J. J. Chem. Soc., Chem. Commun.
1994, 713. (b) Mukhopadhyay, M.; Bhatia, B.; Iqbal, J. Tetrahedron
Lett. 1997, 38, 1083.
(86) (a) Reddy, M. M.; Bhatia, B.; Iqbal, J. Tetrahedron Lett. 1995, 36,
4877. (b) Nageshwar Rao, I.; Prabhakaran, E. N.; Das, S. K.; Iqbal,
J. J. Org. Chem. 2003, 68, 4079. (c) This reaction can be achieved
using a polyaniline-supported cobalt catalyst: Prabhakaran, E. N.;
Iqbal, J. J. Org. Chem. 1999, 64, 3339.
(87) Mizutani, K.; Shinokubo, H.; Oshima, K. Org. Lett. 2003, 5, 3959.
(88) (a) For a comprehensive review on alkynyl-alkynyl coupling, see:
Siemsen, P.; Livngston, R. C.; Diederich, F. Angew. Chem., Int. Ed.
2000, 39, 2632. (b) For a general review on coupling reactions
between sp-carbon centers, see: Sonogashira, K. In ComprehensiVe
Organic Synthesis; Trost, B. M., Ed.; Pergamon Press: New York,
1991; Vol. 3, p 551.
(89) (a) Hay, A. S. J. Org. Chem. 1962, 27, 3320. (b) Nicolaou, K. C.;
Petasis, N. A.; Zipkin, R. E.; Uenishi, J. J. Am. Chem. Soc. 1982,
104, 5555. For a recent publication, see, for example: (c) Hilt, G.;
Hengst, C.; Arndt, M. Synthesis 2009, 395.
(90) (a) Berscheid, R.; Vögtle, F. Synthesis 1992, 58. (b) Nicolaou, K. C.;
Zipkin, R. E.; Petasis, N. A. J. Am. Chem. Soc. 1982, 104, 5558. (c)
Nomoto, T.; Fukui, K.; Nakagawa, M. Bull. Chem. Soc. Jpn. 1976,
49, 305. (d) Eglington, G.; Galbraith, A. R. J. Chem. Soc. 1959, 889.
1462 Chemical Reviews, 2010, Vol. 110, No. 3
(91) (a) Ando, T.; Vu, M. H.; Yoshida, S.; Takahashi, N. Agric. Biol.
Chem. 1982, 46, 717. (b) Crombie, L.; Hobbs, A. J. W.; Horsham,
M. A. Tetrahedron Lett. 1987, 28, 4875. (c) Miller, J. A.; Zweifel,
G. Synthesis 1983, 128.
(92) For a review, see: Jeganmohan, M.; Cheng, C.-C. Chem.sEur. J.
2008, 14, 10876.
(93) Krafft, M. E.; Hirosawa, C.; Dalal, N.; Ramsey, C.; Stiegman, A.
Tetrahedron Lett. 2001, 42, 7733.
(94) Dicobalt octacarbonyl is known to be an efficient catalyst for the
Pauson-Khand reaction; for instance, see: Krafft, M. E.; Bonãga,
L. V.; Hirosawa, C. J. Org. Chem. 2001, 66, 3004, and references
cited therein.
(95) Wang, C.-C.; Lin, P.-S.; Cheng, C.-H. Tetrahedron Lett. 2004, 45, 6203.
(96) Agnès, G.; Chiusoli, G. P.; Cometti, G. Chem. Commun. 1968, 1515.
(97) (a) Kanai, H.; Okada, M. Chem. Lett. 1975, 167. (b) Kanai, H.; Ishii,
K. Bull. Chem. Soc. Jpn. 1981, 54, 1015.
(98) Wang, C.-C.; Lin, P.-S.; Cheng, C.-H. J. Am. Chem. Soc. 2002, 124, 9696.
(99) Hilt, G.; Treutwein, J. Angew. Chem., Int. Ed. 2007, 46, 8500.
(100) Chang, H.-T.; Jayanth, T. T.; Wang, C.-C.; Cheng, C.-C. J. Am. Chem.
Soc. 2007, 129, 12032.
(101) Ohmiya, H.; Yorimitsu, H.; Oshima, K. Angew. Chem., Int. Ed. 2005,
44, 2368.
(102) (a) Douglass, M. R.; Marks, T. J. J. Am. Chem. Soc. 2000, 122, 1824. (b)
Douglass, M. R.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 2001, 123,
10221. (c) Kawaoka, A. M.; Douglass, M. R.; Marks, T. J. Organometallics
2003, 22, 4630. (d) Takaki, K.; Takeda, M.; Koshoji, G.; Shishido, T.;
Takehira, K. Tetrahedron Lett. 2001, 42, 6357. (e) Takaki, K.; Koshoji, G.;
Komeyama, K.; Takeda, M.; Shishido, T.; Kitani, A.; Takehira, K. J. Org.
Chem. 2003, 68, 6554. (f) Takaki, K.; Komeyama, K.; Takehira, K.
Tetrahedron 2003, 59, 10381.
(103) For the addition of phosphane-borane to alkynes under palladium
catalysis, see: (a) Mimeau, D.; Gaumont, A.-C. J. Org. Chem. 2003,
68, 7016. For the transition metal-catalyzed addition of tertiary
phosphines to alkynes, see: (b) Arisawa, M.; Yamaguchi, M. J. Am.
Chem. Soc. 2000, 122, 2387.
Cahiez and Moyeux
(104) Ohmiya, H.; Yorimitsu, H.; Oshima, K. Angew. Chem., Int. Ed. 2005,
44, 3488.
(105) (a) Suzuki, T.; Lo, P. Y. J. Organomet. Chem. 1990, 391, 19. (b)
Lukevics, E.; Belyakova, Z. V.; Pomerantseva, M. G.; Voronkov,
M. G. J. Organomet. Chem. Libr. 1977, 5, 1. (c) Ojima, I. In The
Chemistry of Organic Silicon Compounds; Patai, S., Rappoport, Z.,
Eds.; Wiley: Chichester, 1989; p 1479. (d) ComprehensiVe Handbook
on Hydrosilylation; Marciniec, B., Ed.; Pergamon: Oxford, 1992. (e)
Hiyama, T.; Kusumoto, T. In ComprehensiVe Organic Chemistry;
Trost, B. M., Fleming, I., Eds.; Pergamon: Oxford, 1991; Vol. 8,
Chapter 3.12. (f) Horn, K. A. Chem. ReV. 1995, 95, 1317.
(106) (a) Suginome, M.; Ito, Y. Chem. ReV. 2000, 100, 3221. (b) Hibino,
J.; Nakatsukasa, S.; Fugami, K.; Matsubara, S.; Oshima, K.; Nozaki,
H. J. Am. Chem. Soc. 1985, 107, 6416. (c) Watanabe, H.; Kobayashi,
M.; Saito, M.; Nagai, Y. J. Organomet. Chem. 1981, 216, 149.
(107) Wong, Y.-C.; Jayanth, T. T.; Cheng, C.-H. Org. Lett. 2006, 8, 5613.
(108) For the Cu-catalyzed coupling of aryl halides with thiols, see: (a)
Bates, C. G.; Gujadhur, R. K.; Venkatamaran, D. Org. Lett. 2002, 4,
2803. (b) Kwong, F. Y.; Buchwald, S. L. Org. Lett. 2002, 4, 3517.
(c) Wu, Y.-J.; He, H. Synlett 2003, 1789. (d) Bates, C. G.; Saejueng,
P.; Doherty, M. Q.; Venkatamaran, D. Org. Lett. 2004, 6, 5005. (e)
Deng, W.; Zou, Y.; Wang, Y.-F.; Liu, L.; Guo, Q.-X. Synlett 2004,
1254. For Pd-catalyzed reactions, see: (f) Mispelaere-Canivet, C.;
Spindler, J.-F.; Perrio, S.; Beslin, P. Tetrahedron 2005, 61, 5253.
(g) Itoh, T.; Mase, T. Org. Lett. 2004, 6, 4587. (h) Fernàndez
Rodriguez, M. A.; Shen, Q.; Hartwig, J. F. J. Am. Chem. Soc. 2006,
128, 2180. (i) Murata, M.; Buchwald, S. L. Tetrahedron 2004, 60,
7397. (j) Schopfer, U.; Schlapbach, A. Tetrahedron 2001, 57, 3069.
(k) Li, G. Y. Angew. Chem., Int. Ed. 2001, 40, 1513. For Ni-catalyzed
reactions, see: (l) Cristau, H. J.; Chabaud, B.; Chêne, A.; Christol,
H. Synthesis 1981, 892. (m) Millois, C.; Diaz, P. Org. Lett. 2000, 2,
1705. (n) Percec, V.; Bae, J. Y.; Hill, D. H. J. Org. Chem. 1995, 60,
6895. (o) Takagi, K. Chem. Lett. 1987, 2221.
CR9000786
 Chem. Rev. 2010, 110, 1463–1497 1463

Normal Mode Analysis of Biomolecular Structures: Functional Mechanisms of

Membrane Proteins

Ivet Bahar,* Timothy R. Lezon,† Ahmet Bakan,† and Indira H. Shrivastava

Department of Computational Biology, School of Medicine, University of Pittsburgh, 3064 BST3, 3501 Fifth Avenue, Pittsburgh, Pennsylvania 15213

Received March 5, 2009

Contents 3.1. Ion Channels 1477

3.1.1. Gramicidin A 1477
1. Introduction 1463 3.1.2. Potassium Channels 1478
1.1. Protein Dynamics and Allostery 1463 3.1.3. Mechanosensitive Channels 1479
1.1.1. Dynamic Equilibrium between Pre-existing 1463 3.2. Receptors 1480
Conformations
3.2.1. Nicotinic Acetylcholine Receptor 1480
1.1.2. Functional Significance of Collective 1465
Motions 3.2.2. Rhodopsin 1483
1.1.3. Normal Mode Analysis: An Old Technique 1466 3.3. Transporters 1484
That Recently Found a Revival in 3.3.1. Glutamate Transporters 1485
Molecular Biology 3.3.2. ATP Binding Cassette Transporter BtuCD 1486
1.1.4. Elastic Network Models Inspired by the 1466 4. Conclusion 1488
Robustness of Global Modes 4.1. Robustness and Functionality of Global Modes 1488
1.2. Structural Dynamics of Membrane Proteins: 1467 4.1.1. Robustness of Global Modes: A 1488
Significance and Challenges Requirement in Evolutionary Selection of
1.2.1. Classification, Biological Role, and 1467 Structures
Pharmacological Importance 4.1.2. Toward Gaining Insights into Functional 1488
1.2.2. Increasing Structural Data on Membrane 1468 Dynamics of Membrane Proteins
Proteins 4.1.3. Many Functional Motions of Membrane 1489
1.2.3. Multiscale Dynamics of Membrane Proteins 1468 Proteins Are Intrinsic to Their 3D Structure,
Probed by Experiments Independent of Membrane Environment
1.2.4. Structure-Based Models, Theory, and 1469 4.1.4. Entropic vs Enthalpic Effects, or Geometry 1489
Computations vs Specificity
2. Theory 1470 4.2. Extensions of Coarse-Grained NMA and 1490
2.1. Principal Component Analysis of 1470 Future Directions
Experimentally Resolved Conformations 4.2.1. Hybrid Methods That Integrate CG NMA 1490
2.1.1. Definitions of System and Parameters 1470 and MD
2.1.2. Covariance Matrix: A Measure of 1471 4.2.2. Docking and NMA in Drug Discovery 1491
Correlations between Residue Motions 4.2.3. Normal Modes for Structural Refinement 1492
2.2. Normal Mode Analysis 1471 4.2.4. Exploring Allosteric Transitions in Large 1492
2.2.1. Assumptions and Limitations 1471 Biomolecular Systems
2.2.2. Underlying Potential and Hessian Matrix 1472 5. Acknowledgments 1493
2.2.3. Equation of Motion and Its Solution 1472 6. References 1493
2.2.4. Significance of Normal Modes and 1472
Dominance of Slow Modes 1. Introduction
2.2.5. Covariance Computed from NMA: Bridging 1473
with PCA of Structural Ensembles 1.1. Protein Dynamics and Allostery
2.2.6. Using Normal Modes for Exploring the 1473 1.1.1. Dynamic Equilibrium between Pre-existing
Potential Energy Surface
Conformations
2.3. Elastic Network Models 1473
2.3.1. Gaussian Network Model 1474 The ability of macromolecules to sample an ensemble of
2.3.2. Anisotropic Network Model 1474 conformations has been evident for decades, starting from
2.3.3. Rotating-Translating Blocks Model 1475 the statistical mechanical theory and simulations of polymers.1-3
2.3.4. Extensions for Treating Environmental 1476 A polymer chain of N atoms enjoys 3N - 6 internal degrees
Effects of freedom, which gives rise to infinitely many conforma-
3. Intrinsic Dynamics of Membrane Proteins and 1477 tions. Even a simple model of N ) 100 atoms where bond
Their Functional Significance lengths and bond angles are fixed, and dihedral angles are
restricted to discrete isomeric statesssay three states per
* To whom correspondence should be addressed. Phone: 412 648 3332.
Fax: 412 648 3163. E-mail: bahar@pitt.edu. bondshas access to 3N-3 ) 1.9 × 1046 conformations.
†
These authors contributed equally to this work. Proteins, too, are polymers, and have access to ensembles
10.1021/cr900095e  2010 American Chemical Society
Published on Web 09/28/2009
1464 Chemical Reviews, 2010, Vol. 110, No. 3
Ivet Bahar is the John K. Vries Founding Chair of the Department of
Computational Biology, School of Medicine, University of Pittsburgh, and
Director of the Joint Ph.D. Program in Computational Biology between
Carnegie Mellon University and University of Pittsburgh. She served as
an Assistant (1986-87), Associate (1987-93), and Full Professor
(1993-2001) at the Chemical Engineering Department of Bogazici
University, Istanbul, Turkey, before joining the University of Pittsburgh.
Her research areas are biomolecular systems structure, dynamics, and
interactions. She authored in over 170 papers in the areas of polymer
chemistry, molecular biophysics, and computational structural biology. She
has been an elected member of EMBO since 2000, a principal member
of the Turkish Academy of Sciences, an elected member of the Biophysical
Society Council, and an Executive Committee Member of the International
Society of Quantum Biology and Pharmacology.
Timothy Lezon received his B.S. in physics from the University of Illinois
at Urbana-Champaign in 1997 and his Ph.D. in physics from the
Pennsylvania State University in 2007. His research interests include the
application of statistical physics techniques to biological systems,
particularly in the development of coarse-grained models for exploring
protein structure and dynamics.
of conformations. The main structural difference between
proteins and other chain molecules is that, under physiologi-
cal conditions, proteins sample a significantly narrower
distribution of conformations compared to disordered poly-
mers. Their conformational variations are confined to the
neighborhood of a global energy minimum that defines their
“native state”.
While the native state has been traditionally viewed as a
“unique structure” selected or encoded by the particular
amino acid sequence, it is now established by theory,
computations, and experiments, after the work of pioneering
scientists in the field,4-15 that the native state actually
represents an ensemble of microstates: these microstates
maintain the overall “fold” and usually share common
secondary structure, but they differ in their detailed atomic
coordinates. Differences are manifested by variations in bond
Bahar et al.
Ahmet Bakan received his undergraduate education at Koç University,
Istanbul. He is currently pursuing his Ph.D. degree in computational biology
at the University of Pittsburgh. His interests focus on protein-ligand
interactions, protein dynamics, and the role of dynamics in recognition
and binding events.
Dr. Indira Shrivastava is currently a Research Associate at the Department
of Computational Biology at the University of Pittsburgh. She received
her Ph.D. in Quantum Chemistry from the Department of Chemistry at
University of Pune, India, in 1993. She did postdoctoral work at the Indian
Institute of Science, Bangalore, EMBL, Heidelberg, and the Structural
Biology Unit, at the University of Oxford, and was a visiting fellow at the
National Institutes of Health, Maryland. Her research interests include
molecular modeling and simulations of membrane proteins to explore
functional dynamics of these proteins by an integrated approach of
macroscopic, microscopic, and submicroscopic simulations. She dedicates
this review article to her father, Mr. Harigovind Shrivastava, upon his 75th
birthday.
lengths, bond angles, dihedral angles, loop conformations,
substructure packing, or even entire domain or subunit
positions and orientations.
Importantly, these microstates are not static: there is a
dynamic equilibrium which allows for continual intercon-
versions between them while maintaining their probability
distribution. These “jigglings and wigglings of atoms” as
expressed by Feynman,16 and clearly observed in molecular
dynamics (MD) simulations, were originally viewed as
random events, or stochastic properties, hardly relevant to
biological function. They essentially account for local
relaxation phenomena in the nanoseconds regime, which may
facilitate, for example, the diffusion of oxygen into the heme
cavity of myoglobin17 or the permeation of ions across
selectivity filters in ion channels.18-20 However, recent studies
indicate that these thermal fluctuations may not only pas-
sively facilitate but also actively drive concerted domain
movements and/or allosteric interactions, such as those
required for substrate binding, ion channel gating, or catalytic
Functional Mechanisms of Membrane Proteins
Figure 1. Equilibrium motions of proteins. Motions accessible near
native state conditions range from femtoseconds (bond length
vibrations) to milliseconds or slower (concerted movements of
multiple subunits; passages between equilibrium substates). X-ray
crystallographic structures span length scales up to several hundreds
of angstroms. Fluctuations in the subnanosecond regime are
indicated by X-ray crystallographic B-factors. NMR spectroscopy
is restricted to relatively smaller structures, but NMR relaxation
experiments can probe a broad range of motions, from picoseconds
to seconds, including the microseconds time range of interest for
several allosteric changes in conformation. Also indicated along
the abscissa are the time scales of processes that can be explored
by MD simulations and coarse-grained computations. Molecular
diagrams here and in the following figures have been generated
using Jmol (http://www.jmol.org/), PyMol (http://www.pymol.org/
), or VMD35 visualization software.
function.15,21-34 Figure 1 provides an overview of the broad
range of equilibrium motions accessible under native state
conditions, ranging from bond length vibrations, of the order
of femtoseconds, to coupled movements of multimeric
substructures, of the order of milliseconds or seconds.
1.1.2. Functional Significance of Collective Motions
In the last two decades, there has been a surge in the
number of studies based on principal components analysis
(PCA)36 of biomolecular structures and dynamics. These
studies have proven useful in unraveling the collectiVe modes,
and in particular those at the low frequency end of the mode
spectrum, that underlie the equilibrium dynamics of pro-
teins.37 Normal mode analysis (NMA) of equilibrium
structures,38,39 essential dynamics analysis (EDA) of the
covariance matrices retrieved from MD runs,40 and singular
value decomposition (SVD) of MD or Monte Carlo (MC)
trajectories41-43 all fall in this category of PCA-based
methods. Recently, a server has been developed to efficiently
perform such calculations using a variety of input structures.44
PCA-based studies provide increasing support to the view
that the apparently random fluctuations of proteins under
native state conditions conceal contributions from highly
cooperative movements (e.g., concerted opening and closing
of domains) that are directly relevant to biological function.
Functional movements indeed involve passages between
collections of microstates or substates that coexist in a
Chemical Reviews, 2010, Vol. 110, No. 3 1465
Figure 2. Energy profile of the native state modeled at different
resolutions. N denotes the native state, modeled at a coarse-grained
scale as a single energy minimum. A more detailed examination
of the structure and energetics reveals two or more substates (S1,
S2, etc.), which in turn contain multiple microstates (m1, m2, etc.).
Structural models corresponding to different hierarchical levels of
resolution are shown: an elastic network model representation where
the global energy minimum on a coarse-grained scale (N) is
approximated by a harmonic potential along each mode direction;
two substates S1 and S2 sampled by global motions near native
state conditions; and an ensemble of conformers sampled by small
fluctuations in the neighborhood of each substate. The diagrams
have been constructed using the following rhodopsin structures
deposited in the Protein Data Bank: 1U19 (N); 1U19 and 3CAP
(S1 and S2); and 1F88, 1GZM, 1HZX, 1L9H, 1U19, 2G87, 2HPY,
2I35, 2I36, 2I37, 2J4Y, 2PED, 3C9L, and 3C9M (microstates).
dynamic equilibrium (Figure 2). The most cooperative
motions usually occur at the low frequency end of the mode
spectrum. These modes engage large substructures, if not
the entire structure, hence their designation as global or
essential modes. They are intrinsically accessible to bio-
molecules, arising solely from structure. In a sense, in the
same way as sequence encodes structure, structure encodes
the equilibrium dynamics. We refer to these global move-
ments that are collectively encoded by the 3-dimensional
(3D) structure as intrinsic motions of the examined protein,
intrinsic to the protein fold or topology of native contacts.
Biomolecular structures conceivably evolved to favor the
global modes that help them achieve their biological or
allosteric functions.21 Briefly, the emerging paradigm is
structure-encodes-dynamics-encodes-function, and an evo-
lutionary pressure originating from functional dynamics
requirements may have selected the relatively small space
of functional structures.
The predisposition of proteins to undergo functional
changes in structure is now supported by numerous experi-
mental and computational studies, and an increasing amount
of data demonstrates that allosteric responses are driven by
intrinsically accessible motions.15,23,24,45-51 These studies have
brought a new understanding to the role of collective
dynamics in protein functions, demonstrating in particular
how the functions of membrane proteins such as signal
transduction, pore opening, ion gating, or substrate translo-
cation are enabled by the cooperative movements of sym-
metrically arranged subunits. These findings are in support
of the original Monod-Wyman-Changeux (MWC) view
of allosteric effects,52,53 the main tenets of which are
predisposition of the structure to access alternative confor-
mations via cooperative changes in structure (simultaneously
engaging all subunits) and selection from this pool of
1466 Chemical Reviews, 2010, Vol. 110, No. 3
accessible conformation to achieve biological function in the
presence of ligand/substrate binding. Recent findings on the
relevance of global modes to functional dynamics are
presented below for select, widely studied membrane pro-
teins. The goal here is to review NMA-based computational
methods and their applications to membrane proteins. We
will also discuss recent developments for improving the
methodology and its implementation in structure refinement
and drug discovery methods.
1.1.3. Normal Mode Analysis: An Old Technique That
Recently Found a Revival in Molecular Biology
Normal mode analysis provides information on the equi-
librium modes accessible to a system, assuming that the
system is stabilized by harmonic potentials. It has been used
for several decades in studying classical physical phenomena
such as atomic vibrational spectra and transport in the solid
state. Its application to proteins dates back to the early
1980s.54-57 However, only in the past decade has it become
a tool widely used for exploring functional motions. A major
reason behind its broader use is the observation that global
modes elucidated by NMA bear functional significance. This
feature became even more evident with the use of simplified
models in coarse-grained (CG) NMA.38
From a physical perspective, the global modes simply
represent reconfigurations along directions (principal axes)
that are most easily accessible (require the least energy ascent
for a given size deformation) on the multidimensional energy
landscape. Mode frequency (squared) provides a measure
of the curvature (or stiffness) of the energy landscape along
a given mode direction, with lower frequency modes being
softer motions. Given that the energy landscape is, in turn,
defined by the molecular structure, the global modes are
structure-encoded by definition.
A striking feature of NMA and other PCAs of biomo-
lecular structures is the observed robustness of the global
modes to details in atomic coordinates or specific interatomic
interactions. The global modes are defined by the entire
structure (or architecture). Their insensitivity to local interac-
tions, or to the specific energy functions and parameters that
define the force field, presumably results from their systemic
nature. As evidenced by the pioneering study of Tirion,58 a
hypothetical force field with uniform (single-parameter)
harmonic potentials yields global modes almost indistin-
guishable from those obtained from a detailed force field
with specific nonlinear terms. The property that apparently
dominates the shape of the global modes is the network of
inter-residue contacts, which is a purely geometric quantity
defined by the overall shape, form, or native contact topology
of the protein.59,60
1.1.4. Elastic Network Models Inspired by the Robustness
of Global Modes
The insensitivity of global modes to structural and
energetic details has now been confirmed in many studies61-66
and led to the following question: If these modes are not
sensitive to structural and energetic details, why not use CG
models to elucidate such collective movements? This line
of thought opened the way to the adoption of elastic network
models (ENMs) for exploring protein dynamics.38,39 ENMs
take rigorous account of the topology of contacts. In this
respect, they may be viewed as Go models which are also
based on native contact maps.4-6 The major difference is,
Bahar et al.
however, that the network representation adopted in ENMs
permits us to take advantage of methods of NMA or spectral
graph theory to obtain analytical solutions for equilibrium
dynamics that can be readily implemented in efficient
computational algorithms. The main advantages of ENM-
based approaches are indeed (i) their ability to provide an
exact solution for the unique dynamics of each structure and
(ii) their applicability to large biomolecular complexes and
assemblies beyond the range of atomic simulations.
The first such simplified model, the Gaussian network
model (GNM), was introduced a decade ago,67,68 inspired
by the work of Tirion.58 GNM is based on the theory of
elasticity set forth by Flory and co-workers69-73 for polymer
networks. The structure is represented as a network of nodes
(R-carbons) and elastic springs. The springs connect the
R-carbon pairs that are closer than a cutoff distance, Rc, in
the native structure. A Kirchhoff matrix of inter-residue
contacts, Γ, is the sole determinant of equilibrium dynamics.
The accessible spectrum of relaxation modes is computed
using statistical mechanical theories of solid state physics
and/or graph-theoretic methods that have found wide utility
in other applications. The GNM has been shortly followed
by other ENMs, including, in particular, Hinsen’s model with
distance-dependent force constants74,75 and the anisotropic
network model (ANM),76-79 which will be described in some
detail in section 2.3.2.
A major reason behind the broadening recognition of NMA
as a tool for exploring functional motions of proteins is the
observation that global modes elucidated by NMA bear
functional significance. For example, the allosteric change
in conformation undergone by hemoglobin from its tense
(T) form to its relaxed (R) form has been shown by both
full atomic NMA80 and ANM31 to match closely the
collective motions along the second lowest frequency mode
intrinsically accessible to the original structure. The ratchet-
like motion of the two subunits of the ribosome is enabled
by the third slowest mode;81,82 or in general, the opening/
closing of domains/subunits in many enzymes conforms to
their low frequency modes.32,79 It is now broadly recognized
that ligand binding cooperatively triggers collective move-
ments and stabilizes conformers that are already favored by,
or accessible to, the unbound protein structure.30
In recent years, ENM-based NMAs have contributed to
improving our understanding of the collective dynamics of
membrane proteins, among other allosteric systems. Under-
standing the functional motions of membrane proteins is
significant from both biological and pharmaceutical points
of view, as described in section 1.2.1. Progress in this field
has been slow, however, due to the scarcity of structural data
and the complexity of the involved multiscale interactions.
The majority of structure-based computations performed for
membrane proteins in the past decade focused on localized
events, such as passage of ions across a selectivity filter,
which are observable by MD simulations of tens of nano-
seconds. The computational study of events such as ion/
substrate gating, on the other hand, has been limited by the
more cooperative nature of associated changes in structure.
Models and methods designed to observe longer time or
larger size windows, not obscured by atomic details or
random noise, are needed in this case. CG NMAs and their
extensions and combinations with atomic simulations44,83-88
are beginning to fill this gap. The applications summarized
in section 3 permit us to observe for the first time a new
regime of motions at residue-level detail, providing insight
Functional Mechanisms of Membrane Proteins
Figure 3. Schematic representation of different types of integral
membrane proteins. (A) Single helical TM protein (a bitopic
membrane protein), (B) a polytopic TM protein composed of
multiple TM elements (here two helices), and (C) an integral
monotopic membrane protein.
into mechanisms of pore opening, ion gating, or allosteric
signal transduction by membrane proteins.
1.2. Structural Dynamics of Membrane Proteins:
Significance and Challenges
1.2.1. Classification, Biological Role, and Pharmacological
Importance
Membrane proteins are classified into two broad groups:
integral membrane proteins (IMPs) and peripheral membrane
proteins. IMPs are permanently located at the membrane;
peripheral membrane proteins are temporarily attached, either
to IMPs or to the peripheral regions of the membrane. IMPs
include channels, receptors, transporters, and enzymes, in
addition to cell adhesion and energy transduction proteins.
They are divided into two broad groups depending on the
degree to which they span the lipid bilayer: transmembrane
(TM) (or polytopic) and integral monotopic. Integral mono-
topic proteins are attached to the membrane from one side,
while TM proteins are typically composed of three domains:
extracellular (EC), intracellular/cytoplasmic (CP), and TM
domains (Figure 3). The present review focuses on the
equilibrium dynamics of selected TM proteins that have been
explored by NMA-based approaches in recent years (section
3).
The continual flow of ions and metabolites across the
membranes is essential for many of life’s processes. The
membrane acts as an insurmountable barrier for the passage
of ions and/or solutes into or out of the cell under equilibrium
conditions, thus maintaining a net voltage difference between
the cell interior and exterior, known as the resting membrane
potential. TM proteins maintain the equilibrium concentra-
tions of ions/substrates in the EC and CP regions, elicit or
regulate cell signaling and energy transduction processes,
regulate cell volume, or secrete electrolytes.89,90 In particular,
ion pumps and ion exchangers “actively” transport ions: they
pump ions against their gradient by coupling the “uphill”
process to an energy source such as ATP hydrolysis or the
“downhill” movement of an ion or substrate. Likewise,
carriers transport substrates, against their concentration
gradient in many cases, assisted by proton or ion counter-
or cotransport. Ion channels, on the contrary, are usually
viewed to be “passive” membrane proteins: they allow for
“downhill” permeation of ions and may exhibit very high
conduction rates.91 The electrochemical gradients built by
Chemical Reviews, 2010, Vol. 110, No. 3 1467
Figure 4. Distribution of small molecule drugs based on the
targeted molecular function. The distribution is shown for the top-
ranking ten functional categories targeted by 965 FDA-approved
small molecule drugs, excluding biotechnology drugs, nutraceuticals
such as vitamins and amino acids, and those with uncertain targets.
The top ten categories shown in the pie chart are associated with
more than 75% of the drugs in the data set. The distribution is
based on 1008 drug-protein associations. A given category is
counted once if a given drug targets multiple proteins in that
category.
ion pumps serve as a driving source for ion channels and
other transporters.89
With their locations at cell boundaries, membrane proteins
are involved not only in the transport of endogenous
substrates/ions but also in drug uptake92 and drug action.
While approximately 30% of sequenced genes encode
membrane proteins, the fraction of membrane proteins among
drug targets has been estimated to be 70% in 2001.93 An
updated distribution of drug targets is presented in Figure 4.
The pie chart refers to 965 U.S. Food and Drug Administra-
tion (FDA) approved small-molecule drugs, obtained from
DrugBank (http://www.drugbank.ca)94 as primary source and
refined using Therapeutic Target Database (DB),95 Super-
Target DB,96 and the literature.97 A total of 380 proteins are
targeted by these drugs, most of which belong to the human
genome. The corresponding molecular functions, retrieved
from the PANTHER Classification System,98 are grouped
into 71 functional categories. Figure 4 displays the most
frequently targeted ten such categories. The top-ranking four
categories are G-protein coupled receptors (GPCRs), nuclear
hormone receptors, ligand-gated ion channels, and voltage-
gated ion channels. These constitute targets for more than
half of the drugs. These results are consistent with those
recently compiled by Hopkins and co-workers,99 apart from
minor differences, presumably due to differences in the data
set.
The membrane proteins that are most frequently targeted
by small molecule drugs are histamine H1 receptors, R1-
adrenergic receptors, and D2 dopamine receptors. All three
are members of the GPCR family of proteins. These are
succeeded by γ-aminobutyric-acid (GABA) A receptor R1,
a ligand-gated ion channel. These proteins are still being
investigated in relation to a broad spectrum of diseases
1468 Chemical Reviews, 2010, Vol. 110, No. 3 Bahar et al.

including central nervous system disorders, allergies, inflam-

mation, respiratory disorders, headache, and sleep disor-
ders.100 Notably, most of the drugs currently in use were not
initially developed to interact with a specific target protein
but to induce certain phenotypes in cultured cell or animal
assays.101 The identification of the targets followed the
completion of the drug discovery process (a trial-and-error
process using combinatorial chemistry rules). The importance
of assessing drug targets and understanding the mechanistic
aspects of drug-target associations became clear only in
recent years. Knowledge of structure and dynamics of target
proteins is now recognized to be a crucial element in making
progress in rational drug discovery.102

1.2.2. Increasing Structural Data on Membrane Proteins

Although atomic resolution crystal structures of soluble
proteins have historically been reported with an exponentially
increasing frequency, similar progress has not been made
for membrane proteins. Determination of membrane protein
structures is difficult for a number of reasons. Their crystal-
lization requires disruption of the bilayer, usually with
detergents, which renders many of the physical methods of
crystallization difficult or impractical. The amphipathic nature
of membrane proteins and their inherent conformational
flexibility also poses a problem for crystallization.103 Yet
another reason for the limited structural data on membrane
proteins is their low concentrations in tissues, i.e., the
difficulties in overexpressing and purifying them at the
milligram level.
In spite of these limitations, important progress has been
made in recent years in determining membrane protein
structures.104 With the advances in high-throughput tech-
niques in structural biology, multiple protein targets are being
cloned, expressed, and purified in parallel, with clones being
generated on multiwell plates and crystallization trials being
completed at the rate of 100 plates per day.105,106 These Figure 5. PDB statistics for membrane proteins. (A) Structures
techniques are now being advantageously used in membrane of 182 unique membrane proteins are available in the PDB, as of
protein structure determination. Membrane proteins are January 2009. The pie chart displays the distribution of these
unstable in detergent micelles, and finding conditions that structures among different functional/structural categories. (B)
stabilize them helps in protein crystallization. Often, a series Growth of released membrane protein structures and other protein
of detergents are tested, and the one that extracts the structures starting from 1990. Note that the number of “other”
maximum quantity of soluble, active, homogeneous protein proteins is reduced by a factor of 54 in the curve, for display
purposes. We also show the breakdown of membrane proteins into
is used. Dodecyl maltoside is a detergent that provides such different structural groups: R-helical TM, β-barrel TM, and mono-
conditions.107 The recently solved full-length KcsA structure topic. An exponential growth with an R2 value of 0.99 is observed
is a nice example of engineering an enhanced stability at in the last ten years for both membrane proteins and all other
the C-terminal domain of the membrane protein by the use proteins. The corresponding growth rates are 0.23 and 0.18,
of synthetic antigen-binding fragments as crystallographic respectively; that is, they are higher for membrane proteins due to
chaperones. We refer our readers to comprehensive re- initiatives in that direction.
views103,104,108,109 for more information on advances in NMR
spectroscopy,108-116 X-ray crystallography,117-122 electron
www.mpdb.ul.ie). The rapidly growing data in the PDB now
crystallography of 2D crystals in the presence of lipids,93 holds promise for exploring the structure-based dynamics
and infrared spectroscopy (IR).123
of membrane proteins.91 The NMA-based studies presented
These advances resulted in a remarkable increase in the in section 3 provide some examples of such explorations.
number of structurally known membrane proteins: while at
the end of 2003, there were about 326 resolved membrane 1.2.3. Multiscale Dynamics of Membrane Proteins Probed
proteins (75 of them being unique), this number jumped to by Experiments
859, including 182 unique structures by the end of 2008
(Figure 5). A comprehensive summary of membrane protein The biological function of many membrane proteins
structures available to date is provided by the site (http:// involves a transient change in structure, with the associated
blanco.biomol.uci.edu/Membrane_Proteins_xtal.html) main- processes usually spanning a broad range of time scales
tained by Stephen White’s lab. Access to all membrane (Figure 1). Since the early days of electrophysiology, theories
protein structures in the Protein Data Bank (PDB) is provided and models have been developed to gain a better understand-
by databases such as the PDB of Transmembrane Proteins124 ing of the structural basis of biological function and
(http://pdbtm.enzim.hu/) and the Membrane PDB125 (http:// biochemical data.89 Spectroscopic methods for examining
Functional Mechanisms of Membrane Proteins
conformational vibrations have been applied to proteins as
early as 1952, long before the availability of detailed X-ray
structures.126 Their successful application to membrane
proteins has been possible only after the development of
sensitive instrumentation and sophisticated analysis tech-
niques.123
Here, we briefly point to two spectroscopic methods, NMR
and electron paramagnetic resonance (EPR), that emerge as
powerful tools for probing protein dynamics over broad time
scales, and we refer our readers to excellent reviews127-136
for more information.
Structural dynamics probed by NMR. Many of the
experimental data showing that structural dynamics play
a central role in protein function currently originate from
NMR spectroscopy.22,131,135,137 While traditionally NMR
studies have been limited to relatively small proteins
(typically <50 kDa), the dynamics of significantly larger
structures are now being explored with advances in
labeling technology and solution NMR,132 and motions in
both short (pico- to nanoseconds) and long (micro- to
milliseconds) time scales are being probed.116,131,138-140 Using
13
C- and 2H-NMR relaxation rates, Kay and co-workers
showed that site-specific quantitative data could be collected
for the correlation times of methyl groups for proteins of
the order of 100 kDa;141,142 for example, picoseconds-to-
nanoseconds cross-correlation rates for intramethyl 1H-1H
dipolar spin relaxation have been measured for a half
proteasome complex.133 Not all methyl sites undergo such
fast dynamics, however. Those embedded in regions that
undergo highly concerted subunit motions exhibit relaxation
times of the order of milliseconds, as observed143 in the
R-rings of the 20S core particle proteasomesa molecular
machine of 670 kDa. Finally, the ability of NMR experiments
to separate local and global changes in conformation is
noteworthy. A classical example is the T f R transition of
aspartate transcarbamoylase,144 the 2D 1H-13C spectra of
which clearly evidenced the pre-existence of a dynamic
equilibrium between the two forms and the stabilization of
one (R) upon ligand binding.145 Residual dipolar couplings
(RDCs) observed in NMR spectra are increasingly providing
information on collective motions in the nano- to microsec-
ond regime.26,130,146 The PCA-based comparison of an
ensemble of NMR models refined against RDCs26 for
ubiquitin with the X-ray structures of the same protein in
the presence of different substrates showed (i) the accord
between the conformations sampled in solution by the
unbound protein and those assumed upon complexation with
different substrates and (ii) the fact that conformational
variations could be explained to a large extent by a few low
frequency modes intrinsically accessible to the structures.
Structural dynamics probed by site-directed spin labeling
(SDSL) combined with EPR.127-129,147 In this technique,
residues at selected sites are substituted, mostly by cysteines,
followed by the selective modification of the sulfhydryl group
with a nitroxide radical that serves as an EPR spin label/
probe. A set of spin-labeled proteins is thus prepared,
differing by the position of spin-labeled cysteines (Figure
6). The experiments allow for characterizing the mobilities
of the spin-labeled residues and the changes in the distances
between them. Applications of time-resolved SDSL-EPR to
membrane proteins (e.g., to bacteriorhodopsin148,149 and
rhodopsin134,150-155 by Hubbell, Khorana, and co-workers,
to the K+ channel from Streptomyces liVidans (KcsA),156-159
and to a prokaryotic mechanosensitive channel (MscL)160 by
Chemical Reviews, 2010, Vol. 110, No. 3 1469
Figure 6. Site-directed spin-labeling coupled with EPR illustrated
for a potassium channel. (A) Molecular model of KcsA (omitting
two of the four subunits for clarity). The green discs indicate the
positions of the spin-labeled residues (probes) on TM1 (yellow),
TM2 (blue), and the selectivity filter (red). (B) Measurement of
the structural parameter from the spectral line shape of an EPR-
spectrum. The amplitude (Aj) of the normalized central resonance
line M ) 0 and the mobility parameter ∆Ho (the peak-to-peak width
at M ) 0) are shown. Changes in two structural parameters are
usually examined: (i) probe mobility (∆Ho) and (ii) spin-spin
interaction parameter W. Changes in the probe mobility, ∆∆Ho,
indicate rearrangements in tertiary or quaternary contacts, while
the W parameter obtained from the ratio of the normalized amplitude
spectra (Aj) in different forms reports changes in the intersubunit
probe-to-probe proximity. Such measurements performed by Perozo
and co-workers for the open and closed conformations of KcsA as
a function of pH157 revealed the coupled rigid-body rotations of
TM helices TM1 and TM2 of the four subunits and the opening of
the permeation pore (gating) induced by the concerted rotations of
the TM2 helices while the selectivity filter remained practically
immobile.
Perozo and co-workers) have successfully provided informa-
tion on interhelical movements or changes in tertiary contacts
accompanying their functional rearrangements. Notably,
fluctuations and correlations over wide time scales, from
nanoseconds to milliseconds, can be examined by this
technique, and changes in distance between pairs of labels
separated by 20-60 Å can be probed by double electron
resonance (DEER) with a resolution better than 1 Å.161 A
recent application to rhodopsin clearly demonstrated, for
example, that the activation of the molecule is accompanied
by an outward movement of TM helix 6 by about 5 Å.161
1.2.4. Structure-Based Models, Theory, and Computations
According to Moore’s law, the cost of computing halves
roughly every 1.5 to 2 years.162 Looking back at MD
simulations of biomolecules at atomic resolution, we have
indeed progressed from tens of picoseconds in the early
1980s to tens of nanoseconds at present, roughly consistent
with Moore’s law. The progress in recent years in MD
simulations of membrane proteins163-166 has been enabled
by advances in computing technology, algorithms, and
methods.167,168 Classical examples include the simulations of
aquaporin,33,169-172 gramicidin,20,173-175 and KcsA.18,176-179
Yet, the time scales of tens of nanoseconds, or even tenths
of microseconds,180 being accessed in advanced MD simula-
tions still fall short of providing an accurate sampling of the
complete conformational space that many multimeric mem-
brane proteins explore under physiological conditions.
Indeed, multimeric structure is recognized as essential for
enabling highly cooperative structural rearrangements.52
Physics-based CG models such as ENMs emerge as ap-
proximate structural models that provide analytical solutions
1470 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 7. Transmembrane proteins studied by NMA, considered in the present review: (A) gramicidin A, (B) KcsA, (C) MscL, (D)
nAChR, (E) rhodopsin, (F) glutamate transporter (Gltph), and (G) BtuCD. The bilayer is indicated by the dashed lines. The ribbon diagrams
were constructed using the respective structures 1NRU, 1K4C, 2OAR, 2BG9, 1L9H, 1XFH, and 1L7V available in the PDB.
for the collective dynamics of such complex systems, which
cannot be determined with MD.
Here we present recent progress in exploring the
dynamics of membrane proteins using NMA with ENM-
based models. The proteins considered are grouped into
three categories: ion channels, receptors, and transporters,
presented in the respective sections 3.1-3.3, and among
them we present NMA-based studies on gramicidin A
(GA), potassium channels (KcsA and others), a MscL, the
nicotinic acetylcholine receptor (nAChR), rhodopsin (as
a prototypic GPCR), a glutamate transporter (GltPh), and
an ATP binding cassette (ABC) transporter (BtuCD)
(Figure 7). Note that, in many cases, there is no clear-cut
distinction between these categories; some receptors or
transporters may also function as ion channels (e.g.,
nAChR is a ligand-gated ion channel, or glutamate
transporter also functions as a chloride channel).
Membrane proteins are one of the most challenging
biomolecular systems from theoretical and computational
aspects, because their functional mechanisms involve not
only the global dynamics of multiple subunits but also the
highly specific local interactions (e.g., ionic interactions,
solvation) in the presence of lipid bilayer and water
molecules. The former group of motions is conveniently
explored by CG models such as ENMs, but the latter requires
detailed full atomic geometry and energy considerations and
cannot be studied by ENMs. This broad range of events also
spans a hierarchy of functionally significant time scales,
which in turn requires adopting multiscale approaches. The
combination of atomic-level resolution with high-level (low
resolution) computational approaches is deemed as a promis-
ing path for interpreting experimental observations and
determining structure-function relations.20,181 Attempts at
developing such integrated approaches by steering MD
simulations along ENM directions85 or sampling transition
pathways using adaptive ENM methods have recently been
undertaken. These will be presented in section 4.2. Other
recent studies support the utility of ENM-based approaches,
not only for assessing functional dynamics but also for
flexible docking of substrates and refining low resolution
structures, and will be presented in section 4.2.
Bahar et al.
2. Theory
2.1. Principal Component Analysis of
Experimentally Resolved Conformations
2.1.1. Definitions of System and Parameters
Principal component analysis of ensembles of structures
is an orthogonal linear transformation that transforms data
from the Cartesian coordinate system into a new system of
collective coordinates.36 The goal is to gain a simplified view
of the structural variability in the examined data set by
identifying the dominant directions of structural changes. The
new coordinate system is such that the greatest variance in
the data set lies along the first principal component (PC)
axis, followed by the second PC axis, and so on. Here we
will focus on the application of PCA methods to extract
information on equilibrium dynamics. PCA is performed in
this case for an ensemble of conformers such as the PDB
structures for the same protein determined in the presence
of different substrates, or NMR models for the same
protein.44 Alternatively, an ensemble of MD snapshots may
be used. The utility of PCA for understanding functional
dynamics is clearly demonstrated by a recent study of
ubiquitin structures in which a single mode of motion was
found to largely account for ubiquitin’s recognition ability.26
Let us consider an ensemble of M conformations, for a
protein of N interaction sites (N atoms or residues, or any
CG representation of an interaction site). Each conformation,
k, is described by a 3N-dimensional vector consisting of the
position vectors Ri(k) ) (xi(k) yi(k) zi(k))T of the N sites (1 e i
e N) in that particular conformation, organized as
6q(k) ) ((R1(k))T,...,(RN(k))T)T )
(x1(k),y1(k),...,xN(k),yN(k),zN(k))T (1)
Likewise, we define a 3N-dimensional fluctuation vector
∆q(k) ) q(k) - q0 for each conformation, describing the
departure ∆Ri(k) ) Ri(k) - Ri0 in the position vectors of the
N sites from their equilibrium position Ri0 ) (xi0 yi0 zi0)T
(Figure 8). The equilibrium positions may be identified by
the average over all snapshots from MD trajectories or over
all optimally superimposed PDB structures. The PDB
coordinates are usually assumed to be the equilibrium

Functional Mechanisms of Membrane Proteins
Figure 8. Schematic view of interaction sites and their displace-
ments. In the initial conformation, CG sites i and j are located
respectively at Ri0 and Rj0, and the vector Rij0 ) Rj0 - Ri0 defines
the distance vector between these sites. Upon displacement along
mode k, the CG sites move to Ri0 + ∆Ri(k) and Rj0 + ∆Rj(k), and
the distance vector becomes Rij(k). The solid gray line represents
the structural details of the initial-state protein that are above the
resolution of the coarse graining, and the broken gray line indicates
the structure after a displacement along mode k.
positions when performing NMA of a given structure using
the ENM, as will be described below.
2.1.2. Covariance Matrix: A Measure of Correlations
between Residue Motions
In many applications, it is of interest to understand the
type and strength of coupling between the variations in
different degrees of freedom. The cross-correlations between
the components of the fluctuations vectors are given by the
averages 〈∆qi∆qj〉 over all conformations, which are con-
veniently organized in a 3N × 3N covariance matrix C,
C ) M-1 ∑ ∆q(k)∆q(k)T (2)
k
A detailed description of equilibrium motions, including the
mean-square fluctuations of individual sites and their cross-
correlations, is provided by the covariance matrix C. The
elements of C may alternatively be viewed as N × N blocks
(or submatrices of size 3 × 3), Cij, each of the form
[ ]
〈(∆xi∆xj)〉 〈(∆xi∆yj)〉 〈(∆xi∆zj)〉
Cij ) 〈(∆yi∆xj)〉 〈(∆yi∆yj)〉 〈(∆yi∆zj)〉 (3)
〈(∆zi∆xj)〉 〈(∆zi∆yj)〉 〈(∆zi∆zj)〉
Here we use boldface subscripts to designate a (sub)matrix
or vector and lightface subscripts for scalars (e.g., elements
of vectors or matrices). The sum of the diagonal elements
of Cij,
tr{Cij} ) 〈∆xi∆xj〉 + 〈∆yi∆yj〉 + 〈∆zi∆zj〉 ) 〈∆Ri·∆Rj〉
(4)
provides a measure of the cross-correlation between the
fluctuations ∆Ri and ∆Rj of sites i and j; similarly, the mean-
square fluctuations in the positions of individual sites are
given by the trace of the diagonal submatrices, i.e., tr{Cii}
) 〈(∆Ri · ∆Ri)〉 ) 〈(∆Ri)2〉 using i ) j in eq 4. In many
applications, it proves useful to analyze the N × N covariance
matrix, Cj , composed of the correlations between the fluctua-
tion vectors ∆Ri, themselves,
[ ]
Chemical Reviews, 2010, Vol. 110, No. 3 1471
〈(∆R1)2〉 〈∆R1 · ∆R2〉 ... ...
C̄ )
〈∆R2 · ∆R1〉 〈(∆R2)2〉 ... ...
... ... ... ...
〈∆RN · ∆R1〉 ... ... 〈(∆RN)2〉
(5)
The fluctuations in the Cartesian space are mapped onto
the space spanned by the 3N (or N) principal axes upon
diagonalizing the covariance matrix C (or C j ) as
3N
C ) PSPT ) ∑ σkpkpkT (6)
k)1
where P is the unitary matrix of normalized displacements
along the principal axes, also called the principal modes
of structural changes, each given by a column pk, (1 e k
e 3N), and S is the diagonal matrix of eigenvalues σ1,
σ2, ..., σN, usually ordered in descending order. The
eigenvalues are directly proportional to the variance along
the principal axes such that the fractional contribution of
pk to the structural variability in the data set is fk ) σk/
∑Iσi, where the summation is performed over all 3N
modes. Equation 6 permits us to assess the contribution
of each mode or subset of modes to the observed
covariance. For example, the square displacements in the
position of the ith interaction site induced by the top
ranking m PC modes are
|
m
(∆Ri)2 ) tr{[ ∑ σkpkpkT]ii} (7)
1ekem k)1
2.2. Normal Mode Analysis
2.2.1. Assumptions and Limitations
The mathematical theory of NMA is detailed in any
classical mechanics text;182 hence, here we will present only
an outline of the theory as it pertains to its recent applications
to proteins and their complexes. The essence of NMA is
again the diagonalization of a 3N × 3N matrix H, called the
Hessian, the inverse of which yields the covariance matrix
C.
The underlying assumption in NMA is that any given
equilibrium system fluctuates about a single well-defined
conformation and that the nature of these thermally induced
fluctuations can be calculated assuming a simple harmonic
form for the potential. This directly leads to a basic limitation
of NMA: it is only valid in proximity to the equilibrium. As
the system is displaced from its equilibrium conformation,
the extent to which the harmonic approximation holds grows
increasingly uncertain. The excursions from equilibrium
along the normal modes must be closely monitored, lest one
propose a conformational change in excess of the model’s
capabilities. In situations where the potential is calculated
using exact units, for example when an atomistic force field
is used as the kernel for the potential energy surface, then
the magnitude of the excursions along the normal modes
might be approximated by the absolute temperature of the
system. When further coarse-graining is used, for example
in the ENMs, then the absolute magnitudes of the interactions
are unknown and even greater care must be used.
A second caveat to NMA is that the normal modes
represent instantaneous displacements and are tangent to the
1472 Chemical Reviews, 2010, Vol. 110, No. 3
direction of motion at equilibrium. The molecule often
contains additional internal distance constraints that are not
explicitly included in the NMA, such as fixed bond lengths
or bond angles. All but the smallest distortions along the
normal modes will violate these constraints unless measures
are taken to satisfy them.183,184 As a result, (i) NMA results
are “accurate” in the immediate vicinity of the energy
minimum, where the “direction” of motion is accurately
predicted; large excursions in the conformational space may
necessitate the use of an adaptive scheme to re-evaluate
normal modes at a minimum, and (ii) NMA with ENMs is
most useful for predicting the large-scale motions, which are
insensitive to structural and energetic details; localized highly
specific interactions, including in particular electrostatic
interactions, cannot be precisely accounted for. Increasingly,
NMA is used in conjunction with traditional MD or other
simulation methods to explore large-scale motions in the
presence of detailed atomic interactions.85,183,185-189
2.2.2. Underlying Potential and Hessian Matrix
For our purposes, the physical system under consideration
is a molecular system containing N interaction sites, the
Cartesian coordinates of which are given by eq 1. We will
omit the superscript k here, since NMA is performed for a
single structure (M ) 1). Near the equilibrium conformation,
q0, the potential energy can be expanded as a power series
in q as
V(q) ) V(q0) + ∑ ( )
∂V 0
∂qi
(qi - qi0) +
( )
i
∂2V 0
1
∑
2 i,j ∂qi ∂qj
(qi - qi0)(qj - qj0) + ... (8)
where superscripts of zero indicate the equilibrium confor-
mation. The first term is the minimum value of the potential,
which may be set to zero. The second term is identically
zero at any local minimum of the potential. To second order,
the potential is then a sum of pairwise potentials
( )
∂2V 0
V(q) )
1
∑
2 ij ∂qi ∂qj
(qi - qi0)(qj - qj0)
) ∑ (qi - qi0) Hij(qj - qj0) ) ∆qTH∆q
1 1
2 i,j 2 represented by eq 12 in compact notation as
(9)
where H is the Hessian matrix obtained from the second
derivatives of the potential with respect to the components
of q (or ∆q):
( )
∂2V
0
Hij ) (10)
∂qi ∂qj
In the same way as the covariance matrix C is organized, H
may be thought of as an N × N matrix of 3 × 3 submatrices,
each of which describes the energetic contribution from the
interaction of two CG sites. Two important properties of the
Hessian arise from eq 10. First, H is real and symmetric by
construction and is therefore diagonalized by an orthogonal
transformation. Where H not symmetric, its eigenvectors
would not form an orthonormal basis over the full space of
molecular motions and NMA could not be performed.
Second, none of the eigenvalues of H can be negative if H
Bahar et al.
is constructed at a local potential energy minimum. The sign
of a given eigenvalue indicates the local curvature of the
potential along the corresponding mode directional vector
or eigenvector: Positive eigenvalues indicate local minima,
and negative eigenvalues, local maxima. The local potential
energy landscape for a system in a potential energy minimum
will have only positive or zero curvature in all directions.
Eigenvalues that are identically zero indicate conformational
changes that have no effect on the system’s (internal)
potential energy. Typically, H has six zero eigenvalues,
corresponding to the rigid-body rotations and translations of
the molecule. Obtaining the proper form of the Hessian can
be a difficult process that must be handled delicately and
will be discussed later.
2.2.3. Equation of Motion and Its Solution
The Hessian does not contain the full story of NMA.
Because NMA is applied to the study of dynamics, it is
necessary to account for kinetic energy as well as potential
energy. In doing so, the form of the matrix that is to be
diagonalized changes slightly, but the physical interpretation
of the results is more transparent. By considering the system
to be a collection of classically behaving particles, the
equation of motion can be written as
d2∆q
M + H∆q ) 0 (11)
dt2
Here the diagonal matrix M contains the masses of the
particles. Each mass is repeated three times, once for each
of the particle’s three Cartesian coordinates. A solution to
eq 11 is the 3N-dimensional vector uk(t) ) ak exp{-iωkt},
where ak is a complex vector containing both amplitude and
phase factor, and ωk is the frequency of the mode of motion
represented by uk(t). Substituting this solution into eq 11,
the equation of motion becomes
Huk ) ωk2Μuk (12)
which is a generalized eigenvalue equation. The complete
set of solutions uk(t), 1 e k e 3N, and the corresponding
squared frequencies ωk2 may be organized as the respective
columns of the matrix U and the elements λk ) ωk2 of the
diagonal matrix Λ to rewrite the set of 3N equations
HU ) MUΛ (13)
Equation 13 is usually solved by transforming it to a
standard eigenvalue equation H̃Ũ) ŨΛ in mass-weighted
coordinates through the transformations Ũ ) M1/2Uand H̃
) M-1/2HM-1/2. The mass-weighted Hessian, H̃, retains the
symmetry of the original Hessian, and its eigenvectors ũk )
M1/2uk form an orthonormal basis set (i.e., ŨTŨ ) 1). These
are the normal modes of the system. Their Cartesian
counterparts are found through the inverse transformation
U ) M-1/2Ũ and satisfy the orthonormality condition UTMU
) 1. If the interaction sites have all equal mass m, M reduces
to the identity matrix multiplied by m, Ũ ) m1/2U, and H̃ )
m-1 H.
2.2.4. Significance of Normal Modes and Dominance of
Slow Modes
The energy associated with a given normal mode is directly
proportional to the square of its frequency (or its eigenvalue
Functional Mechanisms of Membrane Proteins
λk ) ωk2). This can be seen by rewriting eq 9 for a single
mode k:
1 ωk2
V(uk) ) ukTHuk ) (14)
2 2
Displacements along high-frequency modes are therefore
energetically more expensive than those of equal magnitude
along low-frequency modes. The vibrational energy is, on
average, equally partitioned among all the modes, such that
the average amplitude of oscillation along mode k scales with
1/ωk2. Thus, the molecule experiences the greatest displace-
ment along the lowest frequency, or “slowest”, modes.
Conceptually, the energy landscape slopes most gently along
the slow modes, and these are consequently the most
accessible. These modes are also of highest interest when
seeking to determine the most probable global fluctuations
of a molecule. Large eigenvalues, on the other hand, indicate
directions of steep energetic ascent, and excursions along
these modes will quickly raise the system’s energy.
The cross-correlations 〈∆qi∆qj〉 between the displacements
of the interaction sites along different coordinates are
calculated as statistical mechanical averages of the form
1
∫ d3Nqe-1/(2kBT)∆q H∆q∆qi∆qj )
T
〈∆qi∆qj〉 )
Z we wish to explore structures that are far away from this
kBT(H-1)ij (15)
using the configurational integral
∫ d3Nqe-1/(2k T)∆q H∆q ) (2πkBT)3N/2[det(H-1)]1/2
T
Z) B mode while remaining at a minimum for all the other modes.
(16)
Here the integrations are performed over the complete space
of equilibrium fluctuations, kB is the Boltzmann constant, T
is the absolute temperature, and (H-1)ij designates the ijth
element of the inverse of H. Because only internal motions
affect the system’s potential energy, H has exactly six
eigenvalues that are identically zero, corresponding to the
three translational and three rotational degrees of freedom.
The inverse of H is therefore replaced by the pseudoinVerse,
which is the inverse evaluated only in the space correspond-
ing to the nonzero eigenvalues,
3N-6
ũkũkT
H̃ -1
) ∑ ωk2
(17)
k)1
The importance of the slow modes is again highlighted in
these equations: The lowest frequency modes contribute
most to the spatial partition function because det(H̃-1) is the
product of the reciprocal nonzero eigenvalues of H̃.
2.2.5. Covariance Computed from NMA: Bridging with
PCA of Structural Ensembles
The cross-correlation 〈∆qi∆qj〉 on the left-hand side of eq
15 is simply the ijth element of the covariance matrix C;
therefore, eq 15 may be rewritten in compact notation as
C ) kBTH-1 (18)
This equation establishes the bridge between the PCA of
ensembles of conformations and NMA of a given structure.
In the former case, the top ranking (principal) modes of
Chemical Reviews, 2010, Vol. 110, No. 3 1473
structural changes are extracted from experimental data (or
sets of known structures for a given protein). In the latter,
the same such structural changes are predicted by the theory
using one structure to construct H.
The top-ranking modes obtained by PCA should, in
principle, be comparable to the lowest frequency modes
derived by NMA (i.e., λi ∼ 1/σi, and pi ∼ ui), provided that
(i) the data set of conformations subjected to PCA represents
an equilibrium distribution and (ii) the Hessian in NMA
provides an accurate description of dominant interactions.
Recent PCAs performed for ensembles of PDB structures
exhibit good agreement with the global modes predicted by
CG NMAs.44,190 Notably, ENMs have been adopted in those
NMAs. The relevance of ENM predictions for a given protein
to PC modes derived from sets of structures experimentally
resolved for the same protein (under different conditions, in
the presence of different ligands) lends support to the use of
ENMs for assessing functional changes in structure. Similar
results will be presented below for rhodopsin.
2.2.6. Using Normal Modes for Exploring the Potential
Energy Surface
The harmonic approximation only holds in the immediate
vicinity of a local potential energy minimum, but what if
minimum? One method for exploring remote regions of the
potential energy surface is the normal mode following (NMF)
technique.191 In this method, one starts at an energy minimum
and iteratively moves the structure along its slowest eigen-
Eventually one of the eigenvalues will become negative,
indicating the neighborhood of a saddle point or a transition
state. From that point, other local minima can be found by
iteratively distorting the structure along the potential energy
gradient.
The NMF method has recently been enhanced through
using the Metropolis MC criterion to control the size of the
steps taken.192 As discussed in section 3.1.1, this technique
has proven useful in revealing the gating mechanism of
gramicidin A. Similar techniques that take advantage of
movements along the mode coordinates have been exploited
for investigating transition pathways between known minima,
as described in section 4.4.
2.3. Elastic Network Models
NMA requires knowledge of a symmetric and nonnegative
Hessian. An energy minimization is required prior to
performing NMA on a protein crystal structure to ensure that
the first derivative of the total potential is zero with respect
to all degrees of freedom and to evaluate the second
derivatives (elements of H). Energy minimization is a
computationally expensive task and generally distorts the
initial conformation, resulting in NMA being performed on
a structure altered from the original. Lu and Ma have
demonstrated that the problem of initial energy minimization
can be overcome by mathematically moving the minimum
to the initial structure.193 Their technique involves decompos-
ing the Hessian into submatrices, replacing each submatrix
with its nearest symmetric positive semidefinite matrix, and
reconstructing the Hessian. Far easier is adopting an ENM
that by design accepts the initial structure, usually taken from
the PDB, to be an energy minimum.
1474 Chemical Reviews, 2010, Vol. 110, No. 3
The ENM representation is readily scalable to any level
of coarse-graining and requires very few parameters. The
ENM approximates the protein’s potential energy as that of
a classical network of masses coupled by springs: each node
in the network is a CG site, and each edge is a spring. The
network topology is defined by the native structure, with
edges placed between nodes/sites that lie within a prespeci-
fied cutoff distance from each other. Comparisons of
predicted rms fluctuations to motions inferred from crystal-
lographic B-factors have identified optimal cutoff distances
of 7.3 Å for the GNM and 18 Å for the ANM, provided that
nodes are identified by the CR-atoms.78,194 As to the spring
constants, the simplest ENMs use a uniform force constant
for all interactions; Hinsen proposed using a force constant
that decays rapidly with distance.74 Sen and Jernigan empiri-
cally investigated how the force constants should vary with
the residues’ coordination numbers.195 The adoption of stiffer
springs for sequentially neighboring residues196 or amino
acid-specific force constants197,198 has been shown to improve
the agreement with experiments.
The choice of the specific spring constants has little, if
any, effect on the global modes. The global modes of motion
are widely recognized to be intrinsic properties of the 3D
shape of the protein and have been verified in several studies
to be insensitive to model parameters61-66 and almost
identically reproduced at various hierarchical levels of
resolutions.45,61,199 The robustness of global modes permits
us to utilize ENMs in the study of membrane proteins. One
might conceivably adopt different force constants for the
internal and interfacial regions of membrane proteins and
even differentiate between the interactions with lipid mol-
ecules and those with water molecules in the EC or CP
regions. However, as will also be illustrated below, com-
parisons of ENM results with those obtained from full atomic
NMAs conducted in the presence of explicit water/lipid
molecules have shown that the global modes of membrane
proteins are essentially dictated by the protein architecture/
fold/shape, similar to the cases for other proteins, and are
robust to small variations in the EN topology and environ-
mental effects. Furthermore, the structural changes along the
global modes are observed to correlate well with those
experimentally observed for particular membrane proteins
that are structurally characterized in different states (e.g., apo
vs ligand-bound forms).
Several ENM servers have been developed to date, which
permit users to readily retrieve results based on the ENMs
and their extensions to several applications.78,186,200-205 Below
we present the theory and assumptions underlying commonly
used ENMs, and sections 3 and 4 will illustrate their
applications and extensions.
2.3.1. Gaussian Network Model
The GNM is based on the assumption that all residue
fluctuations (and inter-residue distances) are Gaussianly
distributed around their equilibrium coordinates,67,68 similar
to the statistical mechanical behavior of polymer networks.69-73
The equilibrium coordinates are identified by the position
vectors Ri0 of CR-atoms in the PDB structures. Residue pairs
are connected by a spring of force constant γij, provided that
they are located within a cutoff distance Rc. The fluctuations
in residue positions and their cross-correlations are fully
controlled by the N × N Kirchhoff matrix, Γ, defined in terms
of the spring constants as
Bahar et al.
Γij ) -γij (19)
for all i * j, and
Γii ) Σγij
j
where the summation is performed over all off diagonal terms
in the row i (or column j). Γ is the N × N counterpart of H.
Its pseudoinverse, Γ-1, scales with the covariance matrix
j 67,68,206 (see eq 5)
C
C̄ ) 3kBTΓ-1 (20)
The above equation is obtained from a statistical mechanical
average similar to eqs 15 and 16, where the overall
interaction potential is replaced by207,208
2∑
γij(Rij - Rij0 )(Rij - Rij0 ) ) ∑ γij(∆Rij)2
1 1
VGNM )
ij
2 ij
(21)
The above summation is performed over all connected
pairs. A major simplification in the GNM is the adoption of
a uniform spring constant γij ) γ for all residue pairs (i, j)
that are connected. As a result, Γ reduces to the adjacency
matrix, or Lagrangian, multiplied by γ. Note that the absolute
value of γ does not affect the mode shapes (or eigenvectors)
but uniformly scales their squared frequencies (eigenvalues).
To date, the GNM has been tested in numerous applica-
tions and proven to yield results in reasonable agreement
with a wealth of experimental data, including X-ray crystal-
lographic B- factors for amino acids194,204 and nucleotides,205
root-mean-square deviations in residue coordinates for NMR
models,209 H/D exchange free energy costs,210 hinge sites in
many enzymes and their spatial proximity to catalytic sites,32
NMR order parameters211,212 and changes in NMR parameters
upon ligation,213 highly conserved core amino acids,206
unfolding pathways214 and folding nuclei215 in proteins (e.g.,
rhodopsin),216 or the common dynamics of families of
proteins applied to globins,217 and potassium channels.218 The
good correlation between GNM predictions and experimental
data observed in numerous applications despite the simplicity
of the model highlights the important role of native contact
topology in defining the collective dynamics.
The eigenvalue decomposition of Γ permits us to assess
the contribution of different modes to equilibrium dynamics.
Γ has N - 1 nonzero eigenvalues, with the lowest corre-
sponding to the first (global) mode. Typical outputs from
GNM mode decomposition include the displacement of
residues along each mode axis (global hinge sites being
located between sequence segments that undergo opposite
direction movements along slowest modes), cross-correla-
tions between residues in individual modes, and square
displacement profiles of residues driven by individual modes
or subsets of modes. No information on the 3-dimensional
directions of motions can be obtained with the GNM, because
the main ingredient of the theory is an N × N matrix (as
opposed to the 3N × 3N Hessian in NMA). The anisotropic
network model, described next, is the simplest ENM that
provides information on directionalities.
2.3.2. Anisotropic Network Model
The most broadly used ENM is the anisotropic network
model.76-79 The positions of the nodes in the ANM are
Functional Mechanisms of Membrane Proteins
identified by the coordinates of CR-atoms for amino acids,
and P, C4*, and C2-atoms for nucleotides. ANM analysis is
simply a CG NMA, subject to the potential74,75
2∑
1 modes. Similar to eq 7, the change in the square fluctuations
V) γij(Rij - Rij0 )2
ij
Note that there is a central difference between VGNM and the
above potential. Here V ) 0 if Rij ) Rij0, irrespective of the
direction of the corresponding distance vectors. In the case
of the GNM, on the other hand, changes in the distance
Vector incur a potential energy increase, even if the inter-
residue distance is maintained (see eq 21). The mean-square
fluctuations and cross-correlations predicted by GNM have
been shown in comparative studies to yield better agreement
with experimental data than the ANM predictions.38,39,207,208
Using the ANM, it is possible to readily write a closed
form expression for H using eq 22 in eq 10. The second
derivatives of the potential in this case are simply given by
∂2V γij(xj - xi)(yj - yi)
)-
∂xi ∂yj Rij2
Using the notation xij0 ) (xj0 - xi0) and similarly for yij0 and
zij0 for the three components of the instantaneous distance
vector Rij0, the off-diagonal 3 × 3 submatrices of H take in
the ANM the form
[ ]
(xij0 )2 xij0 yij0 xij0 zij0
γij
Hij ) - 0 2 xij0 yij0 (yij0 )2 yij0 zij0
(Rij) 0 0
xijzij yij0 zij0 (zij0 )2
and the diagonal submatrices satisfy the identity
Hii ) - ∑ Hij
j;j*i
This simple expression for H is readily used in NMA to
determine the collective dynamics. We note that the amino
acid specificity can be included in ENM-based studies by
adopting residue-specific force constants, and indeed we have
deliberately presented the GNM Kirchhoff matrix and ANM
Hessian (respective eqs 19 and 24) in terms of force
constants, γij, that are dependent on the identity of the amino
acids i and j connected in the network. However, in most
applications, γij is taken as a constant, γ, for all pairs of
residues connected in the network. Equation 22 with a single
parameter γij ) γ has been originally used by Tirion for
representing interatomic interactions (as opposed to inter-
residue interactions considered in all succeeding ENM
studies, starting from the GNM) and demonstrating the
reproducibility of global modes obtained by detailed atomic
force fields.58 As mentioned above, the absolute value of γ
for a given level representation does not affect the mode
shapes (i.e., the eigenvectors, uk, (1 e k e 3N - 6) of H)
but their frequencies, because the eigenvalues of H, λk, are
proportional to γ. Likewise, the global modes are insensitive
to the adoption of residue-specific force constants. A more
detailed assessment of the specific role of particular residues
in these global modes and the redistribution of interactions
(e.g., salt bridges) resulting from global movements, and their
effect on allosteric pathways will be given below.
Chemical Reviews, 2010, Vol. 110, No. 3 1475
A major utility of the ANM is its ability to generate
alternative conformations (substates or microstates) in the
close neighborhood of a given structure upon deforming the
original structures along the dominant (lowest frequency)
(22) of residue i contributed by the movement along a given mode
k is given in terms of the kth eigenvector (uk) and eigenvalue
(λk) of H as
|
(∆Ri)2 ) tr{[λk-1ukukT]ii} (26)
k
Or the alternative conformations induced upon moving along
a given mode are simply31
q(k) ) ((R1(k))T,...,(RN(k))T)T )
((R10)T,...,(RN0)T)T ( sλk-1/2uk (27)
where the coefficient s scales with (kBT)1/2. In principle, given
the uncertainty in the absolute value of γ, which is reflected
on the eigenvalues, a range of s values giving rise to
(23) movements comparable in size to those experimentally
observed may be generated and used for further calculations
(such as generating an ensemble of conformations to be used
in docking simulations; see section 4.2). Alternatively, the
choice of s may be based on the correlation cosine or
oVerlap219
Ik ) (∆qAB·uk)/|∆qAB | (28)
between the normalized directional vector uk and the targeted
(24) direction of deformation ∆qAB ) q(B) - q(A), provided that
the goal is to explore the transition from substate A to B.
The potential contribution of subsets of modes to such
a transition may be deduced from the cumulatiVe oVerlap
[∑kIk2 ]1/2, where the summation is performed over the subset
of modes of interest, usually starting from the lowest-lying
(25) modes. Note that this summation is identically equal to unity
if it is performed over all 3N - 6 modes/eigenvectors, which
form a complete orthonormal basis set for the 3N - 6
dimensional space of conformational changes. Another
quantity of interest is the degree of collectiVity, κk, for mode
k, defined as220
| |
N
κk ) N exp{- ∑ R(∆Ri)2 log(R∆Ri)2 }
-1
(29)
i)1 k k
where R is the normalization constant ∑i R(∆Ri)2|k ) 1. The
form of eq 29 suggests that the degree of collectivity has an
entropic significance. The mode with the highest degree of
collectivity has the highest entropy: it is distributed over a
larger number of residues rather than being orderly confined
to a few residues. Lower frequency modes are usually more
collective; their high degree of collectivity is indeed needed
for triggering cooperative (allosteric) responses. Of interest
is to identify the most collective modes toward disclosing
potentially functional movements intrinsically favored by the
overall structure. Sections 3.1, 3.2, and 3.3 will present
applications of the ANM to ion channels, receptors, and
transporters, respectively.
2.3.3. Rotating-Translating Blocks Model
A key strength of ENMs is their scalability. Because the
interactions are all pairwise and harmonic, once the CG sites
1476 Chemical Reviews, 2010, Vol. 110, No. 3
are defined, the ENM can be constructed and its Hessian
determined. Scalability is particularly useful when modeling
very large systems, as it is often the case that the memory
required for diagonalizing H exceeds that currently acces-
sible. The slow modes predicted by NMA are robust to the
level of coarse-graining, and bundling 20 or more residues
into a single CG site still produces slow modes that overlap
well with the global modes of the full CR representation.61,199
The disadvantage of excessive coarse-graining is the loss of
information on the detailed local movements. Although
global motions are accurately reproduced with high levels
of coarse-graining, reconstructing their details can be daunt-
ing. Mixed models221-223 that use detailed descriptions only
for specific regions of interest and CG descriptions for most
of the molecule are helpful in retaining desired detail while
discarding unnecessary information.
In special cases, the size of the Hessian can be reduced
by exploiting the symmetry of the system. Group theoretical
calculations were used to represent the Hessian of icosahedral
viral capsids in reduced forms.224-226 Alternatively, by
making the assumption that all repeat units in a symmetric
system behave identically, one can construct a reduced
Hessian that has only symmetrical modes.227 A more general
method for reducing the complexity of H without eliminating
any structural detail is the rotations and translations of blocks
(RTB)228 or the block normal mode229 (BNM) method. This
method assumes that the system is constructed of nb rigid
blocks and that the normal modes can be expressed as rigid
body rotations and translations of its constituent blocks. Each
block has six degrees of freedom (three translational, three
rotational). The number of degrees of freedom thus reduces
from 3N to 6nb. The blocks are defined as seen fit for the
application at hand: An all-atom protein model might be
simplified by assuming that each residue forms a rigid block,
or a CR-only model might be simplified into blocks of
secondary structure. Furthermore, the size of the blocks is
not restricted: If some domain is known to be particularly
rigid, it might be modeled as a block, whereas a small but
flexible loop may consist of several blocks. The limitation
of the RTB method is that it does not reproduce internal
motions of the blocks, so that a great deal of information
can be lost if flexible regions with high internal mobility
are assumed to be rigid.
Consider a system of N particles that can be collected into
nb < N rigid blocks connected by elastic springs. Define the
3N × 6nb projection matrix, P, from the 3N-dimensional
space of all particles into the 6nb-dimensional space of
rotations and translations of the rigid blocks. The original
Hessian is projected into the space of rigid blocks with the
transformation
HBLK ) PTHP (30)
HBLK is diagonalized with VBLKTHVBLK ) ΛBLK, and the
resulting eigenvectors are projected back into the full 3N-
dimensional space with the inverse projection V ) PTVBLK.
Thus, 6nb - 6 normal modes result from the rigid block
approximation. Each mode is 3N dimensional.
This method was first applied to small proteins by Durand
et al.,230 who used it to simplify conventional MD force fields
by grouping atoms into rigid amino acids. It has since been
used to investigate the role of intrinsic dynamics in confor-
mational changes in molecular motors,229,231 to study the
motion of the ribosome,81 the maturation of cowpea chlorotic
Bahar et al.
mottle virus,232 and the mechanical properties of icosahedral
viral capsids.233
2.3.4. Extensions for Treating Environmental Effects
Methods based on ENMs have been enhanced to include
the effect of viscous drag,234-236 and altered to include the
environmental perturbations.28,237 In order to systematically
assess the effect of environment on protein dynamics, Ming
and Wall28,238 and Zheng and Brooks237 proposed a method
that relies on separating the problem into a system that
contains all relevant degrees of freedom and an enVironment
that contains all other degrees of freedom. The Hessian
matrix is then composed of four blocks that relate the system
with itself (Hss), the environment with itself (Hee), and the
system with the environment (Hse),
( )
Hss Hse
H) T (31)
Hse Hee
At a minimum of the potential energy, the pseudo-Hessian,
j , is found as
H
-1 T
H̄ ) Hss - HseHee Hse (32)
j has the same dimension as Hss but includes the effects
H
of the environment. Its eigenmodes can be directly compared
to those of any system of equal size. This technique has been
used to study a range of phenomena, including the coupling
of motor protein binding pocket dynamics to global protein
structure,237 substrate induced conformational changes,239 and
allostery in membrane proteins.29,240
Another method for introducing viscous damping into a
vibrational system is to use the Langevin equation,
d2∆q d∆q
M +Z + H∆q + ξ(t) ) 0 (33)
dt2 dt
Here the elements of the friction matrix, Z, provide velocity-
dependent damping, and the white noise vector ξ accounts
for thermal energy transferred to the molecule from the
solvent. The elements of this vector obey the properties
〈ξi(t)〉 ) 0 (34)
〈ξi(t) ξj(t′)〉 ) 2ZijkBT δ(t - t′) (35)
From eq 34 it is seen that the net external force incident on
each CG center averages to zero. Equation 35 indicates that
the external force is random in time and provides as much
energy as is lost due to damping. The solution to eq 33 for
macromolecules was given by Lamm and Szabo241 and has
further been modified to incorporate the use of rigid blocks.83
When compared with MD, the Langevin models provide
insight into the role of friction in protein dynamics.234,236,242
This technique has been used in conjunction with ENMs to
calculate scattering functions of proteins,243 to investigate
the sources behind damping in global protein motions,244 and
to estimate the fractional free energy loss in the myosin
power stroke.235
We note that, in a related study,83 the response of
membrane-embedded gramicidin A dimer to a sudden
velocity kick near one end was explored by examining the
time evolution of the molecule, modeled as a collection of
harmonic oscillators, under the Langevin equation. Calcula-
Functional Mechanisms of Membrane Proteins
tions shortly referred to as Langevin dynamics (conceptually
similar to NMA, but in the phase space of displacements
and momenta) were repeated twice, for the fully atomic
model and for a mixed model where the eight indole groups
on the molecule were represented by the RTB model.
Detailed comparison of the results from the two sets of
calculations showed that the results from RTB-Langevin
dynamics closely agreed with those from full atomic Lan-
gevin dynamics, in support of the adoption of the RTB model
for structural elements that are known to be nearly rigid.
3. Intrinsic Dynamics of Membrane Proteins and
Their Functional Significance
3.1. Ion Channels
Ion channels are usually accepted to be passive transport
proteins: they allow for the conduction of ions when the
electrochemical gradients are shifted away from the equi-
librium membrane potential. Their functions include estab-
lishing a resting membrane potential, controlling cell volume,
and regulating the flow of ions across the epithelial cell.89
Their role of “facilitator” is achieved by undergoing changes
between open and closed conformations. Fundamental ques-
tions associated with ion channel functions concern the
location of the activation gate as well as the conformational
changes that ensure the reversible occlusion of the channel.
Here, we will present the NMA results for gramicidin A,
potassium channels, and MscL.
We note that, among the global modes predicted for
symmetric structures (e.g., homotetrameric potassium chan-
nels and homopentameric MscL), a group of modes maintain
the structural symmetry; that is, they induce the same type
of deformation in all monomers. These modes are nonde-
generate; that is, they have unique eigenvalues. Nondegen-
erate modes play a dominant role in enabling the cooperative
transitions of large multimeric, structurally symmetric pro-
teins (or assemblies), by simultaneously exploiting the
intrinsic preferences of individual subunits. It will be shown
below that such nondegenerate modes are instrumental in
initiating the gating process in ion channels.
3.1.1. Gramicidin A
Gramicidin A is the first membrane protein that has been
examined by NMA.245 It is also the smallest ion channel
known to date: it is a dimer, with each monomer being
simply a left-handed helix of 16 amino acids. The two helices
are stacked head-to-head and allow for the selective perme-
ation of small cations (e.g., Cs+, K+) and water molecules
through a narrow opening along the helical axes (Figure 7A).
The original NMA of GA dimer by Roux and Karplus was
one of the earliest studies of membrane proteins, if not the
first, that overruled the use of rigid, or nearly rigid, models
as appropriate approximations to simulate ion permeation.245
The GA was demonstrated therein to enjoy considerable
structural flexibility. Another interesting observation, which
probably was one of the early indications of the robustness
of low frequency modes, was the observed insensitivity of
the low frequency modes to changes in the strength of
hydrogen bond interactions. Computations performed by
varying the partial charges assigned to carbonyl dipoles
showed that the frequencies below 75 cm-1 remained
virtually unchanged; and since these low frequency modes
had a dominant effect on the fluctuation behavior, the overall
Chemical Reviews, 2010, Vol. 110, No. 3 1477
Figure 9. Counter-rotations of the two helical dimers of gramicidin
A, viewed from the EC side along the channel axis. This is the
lowest frequency ANM mode of motion of the dimer. It is
accompanied by a lateral expansion at the helical termini near the
CP and EC regions. This mode was found to be crucially important
for the initiation of the dissociation of the monomers needed for
ion channel gating. Calculations were performed on an ANM
server78 (http://ignmtest.ccbb.pitt.edu/cgi-bin/anm/anm1.cgi), using
the PDB structure 1JNO. Panel A displays the PDB structure, and
panels B and C show two conformations fluctuating in opposite
directions along the lowest frequency mode. Water molecules were
placed inside the pore using Sybyl 8.3. (figure inspired by ref 192).
GA dynamics exhibited little dependence on the strength of
hydrogen bond interactions.
Extensive computational studies have been performed for
gramicidin since then.20 The gating mechanism and the slow
conformational transitions undergone by GA have now been
identified to be a counter-rotation of the two helices around
the pore axis, accompanied by a slight expansion of the
channel mouths at the EC and CP ends. To elucidate this
mechanism, Miloshevsky and Jordan performed192 a series
of computations, including MC-NMF (see section 2.2.6)
coupled with a Metropolis algorithm, full atomic NMA,
NMA with the one-residue-per-block RTB approach, and
NMA with the ANM (using the elNémo server202). Notably,
the global mode predicted by these models of various
complexities was invariably observed to be the counter-
rotation of the two helices around the pore axis (Figure 9).
The global mode shape obtained with the ANM was reported
therein to look “as accurate as that obtained with the all-
atom CHARMM22 force field”. It is worth noting, however,
that the global mode frequency predicted by the RTB and
ANM is significantly larger than that (∼6.5 cm-1) observed
in full atomic NMA. Therefore, while the functionally
relevant mechanism of motion can be accurately predicted
by these CG models, the time scales are not. This is
understandable, as the slow modes are dampened by solvent
and these CG models do not take account of the viscous
drag effect.
The conservation of the mode shape between ANM- and
CHARMM-based NMAs is in accord with compelling
evidence that the directionality of the structural changes
natively accessible to proteins is not altered by the solvent
effect or the use of highly simplified potential functions.66
Miloshevsky and Jordan concluded that the predicted gating
mechanism is an inherent property of GA architecture and
should not be changed by the surrounding lipid and water
molecules.192
It is also worth noting that in contrast to the generally
accepted gating model of two states, closed and open,
involved in GA’s functional transitions, the simulations show
that the GA may exist in multiple intermediate states,
consistent with experiments.246,247 Finally, the effective
sampling of the transition state using the MC-NMF supports
the view that the movements along the global mode predicted
by NMA (either full atomic or using ANM) form the crucial
1478 Chemical Reviews, 2010, Vol. 110, No. 3 Bahar et al.

Figure 10. Sequence and structure of the pore region of five structurally known K+ channels. (A) Alignment of the pore region sequences.
The regions corresponding to the helices TM1 and TM2 and the P-loop are indicated by the boxed green, blue, and red letters, respectively.
The alignment was performed using ClustalW.248 Fully or highly conserved amino acids are written in boldface. Two regions of interest are
the signature motif GYG (highlighted) at the selectivity filter and the conserved glycine on TM2 (e.g., G83 in MthK) enclosed in a magenta
box. (B and C) Structural comparison of the pore forming regions aligned in panel A. These are all tetrameric structures. The monomers
contain either two TM helices (KcsA, MthK, and KirBac, with PDB ID’s 1K4C, 1LNQ, and 1P7B, respectively) colored yellow (TM1) and
blue (TM2) or six TM helices (KvAP and Shaker with PDB ID’s, 1ORQ and 2A79, respectively). The helices S5 and S6 of KvAP and
Shaker are equivalent to the respective helices TM1 and TM2 of the other K+ channels and are displayed here, along with the P-loop
region, colored red. The channels are viewed from side (B) and from the top (EC region) (B) (see ref 218 for more details).
step for initiating pore opening. Motions along the lowest The KcsA structure is proposed to be in the closed form.156
eigenvalue modes encountered at the later stages of the The structure of a calcium-dependent K+ channel from
transition pathway via adaptive NMF exhibited some de- Methanobacterium thermoautrophicum (MthK)120 is consid-
partures, however, from those predicted by RTB and ANM ered to be the open form, containing a wide open intracellular
for the initial state. pore of ∼16 Å. Since the determination of the KcsA
structure,118 many more K+ channel structures have been
3.1.2. Potassium Channels resolved in either open or closed forms (Figure 10).
Potassium channels are tetramers, cylindrically arranged The availability of these structures allows NMA-based
to form a bundle of TM helices, enclosing a central pore, or studies exploring the collective movements of the potassium
a channel, through which ions are conducted. The pore channels and assessing, in particular, the pore opening
regions of most K+ channels are considered to have similar mechanism. One of the first studies in that direction was a
structure, despite significant differences in sequence (Figure NMA of KcsA by Ma and co-workers, in 2002.249 The study
10). They all contain two TM helices, TM1 (yellow) and pointed for the first time to the concerted rotational motion
TM2 (blue), per monomer, connected by a stretch of 30 of all four TM2 helices as a collective mode favored by the
residues, known as the P-loop region (red). The P-loop tetrameric structure. More recent examination of a series of
contains three structural elements: a narrow selectivity filter potassium channels using the ANM demonstrated that the
of ∼10 Å length near the EC entrance of the pore region; core domains favor exactly the same mechanism of global
the P-helix, which spans only the upper half of the bilayer; motion in all cases, which allows for pore opening.218 This
and the exposed loops, also known as the turret, at the EC global mode of motion is a counter-rotation of the two halves
side. The selectivity filter is followed by a large cavity in of the molecule about the cylindrical axis of symmetry, akin
the middle of the core region, which ends in a CP gating to a concerted twisting-and-torsion motion of all TM helices.
region, as illustrated in Figure 11A for KcsA, the first K+ This nondegenerate mode equally distorts all four subunits
channel that has been crystallized and structurally resolved.118 and confers a remarkable expansion at the gate region (Figure
The outer helices (TM1) are exposed to the lipid environ- 11B) by swinging the M2 helices away from the cylindrical
ment; the inner helices (TM2) line the pore. The four P-loops axis, while the selectivity filter region remains fairly rigid.
together form the EC vestibule, which opens up into a large A striking observation is the appearance of a kink in the
central aqueous cavity (of ∼10 Å diameter in KcsA). TM2 helices which further enhances the pore opening.218
Functional Mechanisms of Membrane Proteins Chemical Reviews, 2010, Vol. 110, No. 3 1479

Figure 11. Opening up of the potassium channel pore by the global twisting mode. (A) Ribbon diagram of KcsA illustrating the basic
structural regions and the motion along the second slowest ANM mode. This is a global twisting (nondegenerate) mode that induces
counter-rotations at the EC and CP regions, indicated by the white/gray arrows. (B) Top panel: The pore-radius profile as a function of the
pore axis (Z-axis), calculated for the X-ray structure (black curve) and for two conformations visited by fluctuations in opposite directions
along the global twisting mode (red curves). The inset shows the backbone trace of two of the monomers in the X-ray structure (blue) and
the ANM-predicted conformation (red). The separation between the inner (TM2) helices at the gate is enlarged by about 1.5 Å. Bottom
panel: A mesh-wire representation of the channel pore before (left) and after (right) reconfiguration along the second ANM mode. Color
code: blue, radius > 3 Å; green, 3 Å > radius > 2 Å; red, radius < 2 Å. For visual clarity, only two monomers of the tetramer are shown
(see ref 218 for more details).

Notably, the movement of the M2 helices is consistent
with the displacement observed in the MthK crystal struc-
ture.120 Furthermore, the change in the relative positions of
the four TM2 helices at the gate is also consistent with the
spin-labeling experiments of Perozo et al.157,158 and single
molecule techniques250 which point to an increase in the
distances between the M2 helices (or diameter) at the pore
region and a kink at G83 in MthK TM2 (counterpart of G99
in KcsA and G134 in KirBac1.1; see Figure 10A), in accord
with ANM results. The comparison of the pore-radius profiles
for the wild type protein and its “deformed” form predicted
by the ANM in Figure 11B clearly illustrates the increase in
the pore-radius at the CP gate region.
The recently resolved X-ray structure of a nonselective cation
channel NaK (PDB ID: 3E86) in the open form251 provides an
elegant example of the role of kink-formation in opening
up the cation channel. This structure, when superimposed
onto the closed form (PDB ID: 2AHY)252 indicates the
selectivity filter to be static during gating,251 in agreement
with ANM predictions218 and SDSL-EPR measurements157,158
described above. The major conformational change is a helix
bending at the highly conserved G87, which acts as a hinge.
The counterpart of this glycine, conserved in potassium
channels’ TM2 helices (Figure 10), has been pointed out to
act as a hinge site,218 which also exhibits a kink during the
gating motion. Furthermore, a comparison of the open and
closed forms of NaK also shows a global-twisting motion
around the helical axis of the inner helix,251 in agreement
with experiments250 as well as ANM predictions.218
Miloshevski and Jordan applied their MC-NMF method
(successfully used in their earlier examination of GA channel
gating; see Figure 9) to KcsA. Their study also confirmed
that the gating mechanism of KcsA involves a rotation and
unwinding of the TM2 bundle away from the channel axis,
leading to an open state with an inner vestibule of ∼5-7 Å
radius, in agreement with the computational models described
above.189 Haliloglu and Ben-Tal253 also analyzed the transi-
tion between the closed and open structures (KcsA and
MthK, respectively) using the ANM and in silico alanine-
scanning mutagenesis data. Their ANM study again con-
firmed the global torsion motion as the dominant mechanism
of pore opening, while the alanine-scanning mutagenesis
study identified a network of energetically and dynamically
coupled residues between the selectivity filter and the CP
region, consistent with experimental data.254
3.1.3. Mechanosensitive Channels
One of the most basic demands of primitive cells is to
tolerate changes in the environment, such as tonicity, without
bursting. This function involves regulation of cell volume
by ion flow.255 MscL is an ion channel that is able to detect
and relieve such tensions in the membrane.256-259
The crystal structure of the homopentameric MscL from
Myobacterium tuberculosis has been resolved at 3.5 Å by
Chang and co-workers.117 Each subunit contains three
R-helices: two TM (TM1 and TM2) and one CP (Figure
12A). An aqueous cavity opening, approximately 18 Å in
diameter, leads from the EC side, through a pore lined with
hydrophilic residues narrowing down at the CP side, to an
occluded hydrophobic apex, which is proposed to be the
gate.117 The five subunits are organized into two domains,
the TM and CP domains, both exhibiting a 5-fold symmetry.
Their respective diameters are 50 and 15 Å. The pore-lining
helix of each subunit, TM1 (yellow in Figure 12A), is
connected to the outer helix, TM2 (blue), by an EC loop of
44-68 residues, forming a flap at the EC surface. TM2 is
connected to the CP helix by a shorter loop of 10-12
1480 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 12. Global dynamics of M. tuberculosis MscL predicted
by the ANM. (A) Side view of the pentameric structure, and (B)
views from the EC (top) and CP (bottom) regions. The TM helices
are colored yellow (TM1) and blue (TM2), which are the inner
and outer helices, respectively. CP helices are colored red. The TM
and CP helices rotate in opposite directions in the slowest ANM
mode. The directions of the arrows in panel B refer to the rotations
as viewed from the EC and CP regions, hence their “apparent”
rotation in the same direction. We also note that the structure
fluctuates between two conformers where the TM helices and CP
helices undergo counter-rotations, in either direction; that is, the
arrows displayed in the figure represent one of the two opposite
direction movements along this mode axis. The ribbon diagrams
are generated using the structure (PDB ID: 2OAR) resolved by
Chang et al.117
residues. The TM1 and TM2 helical axes are tilted by about
28° with respect to the 5-fold axis while the CP helix is tilted
by 15°. The radius of the pore varies between 2 and 18 Å
and is partially occluded at the CP region. In the open form,
the cross-sectional diameter of the TM domain is estimated
to be between 30 and 40 Å,260 suggesting a substantial
increase in the size of the channel pore.
The dynamical properties of MscL from E. coli have been
explored by NMA261 using the homology models for the
closed form based on the M. tuberculosis MscL structure258,261
and other structural models proposed by Sukharev and co-
workers.258 The study identified two major kinds of motions:
type I, a symmetrical motion that corresponds to an overall
twisting and tilting of all TM helices around the cylindrical
axis, exhibited by the first nondegenerate ANM mode, and
type II, a global bending, via modes 2 and 3. The “twist to
open” motions261 are consistent with the iris-like mechanism
proposed by Sukharev and co-workers258 to be implicated
in the gating process.
Figure 12 displays the structural changes driven by the
lowest nondegenerate ANM mode (mode 1), calculated for
the M. tuberculosis MscL. As indicated by the arrows, the
TM domain undergoes a global twist in this mode, with the
EC and CP regions undergoing counter-rotations about the
cylindrical axis. The amplitude of the motion is higher at
the CP ends than at the EC ends. Calculations performed
with and without the CP domain show that the same
mechanism of motion is maintained at the TM domain,
irrespective of the presence or the excision of the CP domain,
except for shifting the location of the hinge region by no
more than three or four residues.
Bahar et al.
A small subset of low frequency modes accessible to
the initial substate have been shown in previous work to
account for 65% of the conformational change observed
between the closed and open states.261 The contribution of
the individual modes to the transition between two states
may be assessed by examining the overlap (eq 28) between
the eigenvectors predicted for a given starting conformation
and the difference vector ∆qAB ) q(B) - q(A) between the
two end points. Here we focus on the difference vector ∆qOC
between an open form (O) modeled by Sukharev et al.258
and the closed form (C). Figure 13 displays the cumulative
overlap between ∆qOC and the modes predicted by the ANM.
ANM calculations were repeated with the open (red dashed
curve) and the closed (blue, solid curve) structures. It is
clearly seen that although the pentameric structure has access
to 3N - 6 modes (i.e., 1539 modes in the present case, with
N ) 515), only a small number of modes make a distinctive
contribution (see the jumps in the curves). The fact that the
transition is achieved by moving in such a small subspace
is remarkable. The modes that make the largest contributions
are the nondegenerate slow modes. These modes maintain
the pentameric symmetry of the channel. In panel B of Figure
13, the motions induced in the second lowest nondegenerate
mode (mode 6) are illustrated by the color-coded ribbon
diagrams (red, most mobile; blue, most rigid) and arrows
along the deformation directions. It can be clearly seen that
this motion tends to contract/expand the channel along the
cylindrical axis, with the strongest effect exerted on the
partially disordered segments exposed to the EC region.
3.2. Receptors
3.2.1. Nicotinic Acetylcholine Receptor
Communication between nerve cells takes place at junc-
tions called synapses. The presynaptic cells release, upon
activation, neurotransmitters into the synapse, which bind
to ligand-gated ion channels (LGICs) on the surface of the
postsynaptic cells. Binding of neurotransmitter causes the
channels to open, allowing the ions to flow across the
postsynaptic cell membrane. The opening and closing of
LGICs rapidly convert chemical signals into an electrical
output, regulating the flow of information. Mutations in
LGICs lead to a number of “channelopathies”, such as
congenital myasthenic syndromes, epileptic disorders, and
hereditary hyperekplexia.262 Approximately 8.3% of small-
molecule drugs target LGICs (Figure 4).
The nicotinic acetylcholine receptor is a member of a
superfamily of pentameric transmitter-gated ion channels,
also called Cys-loop receptors, which include the serotonin
5-HT3, GABA A and GABA C, and glycine receptors.
Members of this superfamily contain a signature loop of 13
residues closed by a disulfide bridge, called the Cys-loop,
at the interface between the EC and TM domains of their
respective monomers.263 The nAChR activity is triggered by
binding of acetylcholine (ACh) or nicotine.
The structure of nAChR in the closed state has been
determined by cryo-EM of tubular crystals grown from the
electric organ of Torpedo marmorata.264,265 The structure
consists of five subunits (R, β, δ, R, and γ), two of which
(R-subunits) have a slightly distorted (closed or tense, T)
conformation compared to the other three (open, relaxed,
R), hence the pseudosymmetric organization of the quater-
nary structure. The receptor is organized into three domains
(Figure 7D): a large N-terminal EC domain involved in
Functional Mechanisms of Membrane Proteins Chemical Reviews, 2010, Vol. 110, No. 3 1481

Figure 13. Cumulative contribution of ANM modes to the structural change between the open and closed forms of MscL. The ordinate
displays the cumulative overlap between the ANM modes (eigenvectors) predicted for the starting conformation and the targeted direction
of structural change. ANM calculations were performed using either the closed (C) form (blue, solid curve) or open (O) form (red, dashed)
as the starting substate. In either case, a cumulative overlap of about 0.8 is achieved by the top-ranking ∼120 modes (less than 1/10th of
accessible modes). Concrete (stepwise) contributions are made by the nondegenerate modes. The 2nd lowest nondegenerate mode accessible
to the closed form (mode 6) is illustrated in panel B. This mode induces a contraction/expansion along the pentameric axis, mainly the
portion close to the EC region, as seen from the side (top) and EC (bottom) views of the channel.

Figure 14. Ligand-gated ion channel nAChR structure and dynamics. (A) Structure of the EC and TM domains of nAChR265 (PDB ID:
2BG9). The secondary structure of one of the monomers (R) is colored to display the β-sandwich fold (red) of the EC domain and the four
TM helices (M1-M4; blue) of the TM domain; and the remaining four monomers are shown in gray. The lowest frequency ANM mode
induces a quaternary symmetric twist, as indicated by the arrows shown for monomer R. (B) CP end of TM domain (bottom) and close up
view of one of the monomers (monomer R, colored) (top). Red dashed circle indicates the channel pore. Arrows indicate the collective
movements of M2 helices along ANM mode 1. Green circles represent the CP end of the M2 helices after deformation along ANM mode
1. (C) Comparison of bacterial homopentameric LGICs ELIC (2VL0) and GLIC (3EAM) shows the contribution of this quaternary twist
mode to the conformational changes involved in activation. One subunit (closest to the viewer) is omitted to display the channel pore.

binding the neurotransmitter, a TM pore domain, and a
smaller CP domain. The N-terminal domain of each subunit
is composed of an N-terminal R-helix and two β-sheets
arranged in a curled β-sandwich connected by the Cys-loop
(Figure 14A). The same fold is exhibited by the soluble ACh
binding protein (AChBP).266 There are two ACh binding sites
at the interfaces between the R-δ and R-γ subunits’ EC
domains. The TM domains of individual subunits are
composed of four helices, M1-M4, overall forming a cluster
of 20 TM helices. The pore lining helix, M2, is tilted radially
inward toward the central axis up to the middle of the
membrane. The outer helices (M1, M3, and M4) tilt both
radially toward and tangentially around the central 5-fold
axis.264 Comparison of the ligand-free nAChR and ligand-
bound AChBP structures suggests that ACh binding induces
a local structural rearrangement (closure of two loops around
1482 Chemical Reviews, 2010, Vol. 110, No. 3
ACh) to convert the R-subunits to their open (relaxed, R)
state, which cooperatively triggers a transient opening of the
channel pore at a distance of about 40 Å, thus allowing
cations, particularly Na+ and K+, to pass through.
Several models have been proposed for elucidating the
gating mechanism of nAChR.267-272 NMAs performed by
different groups for the complete structure of nAChR270 and
for the EC-TM domains of the homopentameric R7 nAChR
models based on the nAChR and AChBP structures271,272
invariably showed that the lowest frequency mode is a
concerted quaternary twist with counter-rotations of the EC
and CP domains around the 5-fold symmetry axis (Figure
14A). Like all vibrational modes, this global mode gives rise
to two sets of conformers, corresponding to positive and
negative movements along the mode axis, manifested as
opposite torsions in this case. Of these two sets, one is found
to induce an opening in the TM channel of nAChR: the
counterclockwise torsional rotation of the TM domain
accompanied by clockwise rotation of the EC domain when
viewed from the CP region. As can be seen in Figure 14B,
the five M2 helices lining the pore are displaced slightly away
from the center during this particular quaternary twisting.
The calculation of the pore size profile along the TM channel
(using HOLE273) shows that a relatively small (up to ∼3 Å)
increase in diameter is induced in the constriction zone, the
original value of which is 5.7 Å in the known structure. The
diameter of the first hydration shell of a monovalent cation
is typically around 8 Å. This small opening of the pore
induced by the global mode is thus expected to enable the
passage of hydrated cations.271
An increase in the pore radius by ∼1.5 Å has indeed been
suggested by MD and Brownian dynamics simulations to
be sufficient to raise the computed conductance to ∼22 pSsa
value comparable to the experimental measurements for the
open channel.274 The above results from NMAs (including
those obtained with ENMs) support the view that small but
concerted rearrangements of the M2 helices lining the pore
readily allow for an expansion of this size in the pore, thus
providing an efficient gating mechanism. Concerted rigid-
body motions of M2 helices were inferred by Unwin from
early comparisons of the original structures at various
resolutions.275 Grosman and co-workers made extensive
single-channel electrophysiological measurements to analyze
the change in the microenvironment of the helices M1, M2,
and M3 between the open and closed forms of the
channel.276,277 Mainly, they examined the position-dependent
proton transfers (or pKa shifts) for ionizable residues that
have been engineered in the inner faces of these helices.
These experiments led them to conclude that nAChR pore
dilation involved subtle rearrangements, if any, of these three
helices.276,277 Notably, the twisting mode predicted by the
NMA does not necessarily implicate any significant change
in the orientation of the M2 helix side chains with respect
to the channel lumen but small rotations of about ∼15° that
presumably induce minimal changes in the exposure of side
chains, which may explain the experimental observations.
The changes induced by the NMA-predicted quaternary
twisting mode, in the exposure of M2 residues’ side chains
to the central pore, were indeed pointed out by Changeux
and co-workers to be compatible with the experimental data
from Grosman and co-workers.271
The global twisting-to-open motion of nAChR resembles
those observed in other multimeric ion channels, discussed
above. The collective modes of the M2 bundle (pore-lining
Bahar et al.
helices) predicted by NMA are also observed in PCA of MD
simulation trajectories.268 Conventional MD simulations of
30 ns for nAChR embedded in an explicit lipid bilayer also
indicate269 the concerted rotations of M1 and M2 helices
accompanying the shrinking of the ACh binding pocket, and
the open-close transition of the structure can be driven by
introducing a torsional rotation around the pore axis in
steered MD. The accord between NMA results for the
nAChR, in the absence of a lipid environment, and MD
trajectories conducted in explicit water and lipid bilayer
corroborates the dominant role of the membrane proteins’
intrinsic features in defining the movements that facilitate
essential functions such as gating.
In addition to gating, the mechanism of signal transduction
from the ACh binding site to the pore, which presumably
triggers the channel gating, has been a topic widely studied
by both experiments and computations. In particular, the
allosteric roles of individual residues and loops potentially
involved in communicating agonist binding have been
examined. Sine and co-workers identified, for example, the
loops at the interface of the EC and TM domains that are
required to couple the ligand-binding and pore domains in
the serotonin type-3A receptor278 and identified the key
residues within these loops, which signal agonist binding.279
Single-molecule measurements of open-like vs closed-like
propensities (in terms of Φ-values) of individual residues at
the transition state of the receptor280,281 suggested a Brownian
cascade of domain motions, whereby the transmitter binding
domain assumes an open state and the M2 helices move
toward the open state in discrete steps. Such a sequential
cascade of discrete changes is not compatible with the all-
or-none MWC-type allosteric motions predicted by NMA.
The normal mode motions in the low frequency regime are
smooth and concerted movements that simultaneously engage
both the ACh binding and pore domains, rather than
gradually progressing from one site to another. More recently,
Sine and co-workers showed that the closed-to-open transi-
tion of the receptor involves two primed closed states
independent of agonist binding.282 The primed closed states
elicit short- or long-lived openings. Structural mapping of
these states eludes computational studies due to the limita-
tions in the resolution of the structures and those of the
computational methods themselves.
The recently resolved X-ray structures of two bacterial
homopentameric ligand-gated ion channels shed further light
into pore opening/closing mechanisms. These are the closed
state structure of the Erwinia chrysanthemi ligand-gated ion
channel (ELIC)283 and two open-state structures of the
Gloebacter Violaceus ligand-gated ion channel (GLIC).284,285
These structures do not include the CP helical bundle but
bear EC and TM domains comparable in size and fold to
their counterparts in nAChR. In particular, their EC domain
superimposes closely with AChBP and with the EC domain
of nAChR, except for a missing R-helix. The most striking
difference between the ELIC and nAChR structures is at their
pore domain: the EC half of the ELIC pore is occluded with
Phe246 and Leu239 side chains that narrow down the pore
diameter to less than 1 Å, while the remaining CP half is
wide open (diameter of 6 Å). The two GLIC structures, on
the other hand, are in the open state, being crystallized in
the presence of an activating ligand proton. Figure 14C
compares the ELIC and GLIC structures after their optimal
superimposition. In addition to a symmetric tilt of the pore
forming helices, the most visible difference is a quaternary
Functional Mechanisms of Membrane Proteins
Figure 15. Rhodopsin structure and its ERY motif at the CP region.
(A) Ribbon diagram of the first rhodopsin structure determined by
Palczewski and co-workers,289 shown in a lipid bilayer. This is a
seven TM helix structure, enclosing a chromophore (cis-retinal,
shown in space filling, magenta). The C- and N-termini are labeled
as CT and NT, along with some of the TM helices that can be
distinguished clearly. Note that there is an eighth helix, at the CP
region, that runs parallel to the membrane surface. (B) Enlarged
view of the CP region containing the ERY motif (E134-R135-Y137)
on the TM helix 3 (or H3), involved in G-protein recognition. (C)
Reconfiguration of the ERY-motif-containing domain upon
cis-trans isomerization of the retinal induced by light activation,
suggested by an ANM analysis85 of rhodopsin dynamics.
twist similar to that observed in nAChR. Bocquet et al.284
reported that the lowest ENM mode, a quaternary twist of
the two domains, explains 29% of the structural difference
between the cores of the structures. Overall, these structural
data are consistent with a model of pore opening involving
a quaternary twist and tertiary deformation.284
3.2.2. Rhodopsin
G-protein coupled receptors constitute one of the largest
protein superfamilies in the human genome, with more than
800 members. Among the five families that form this
superfamily, the rhodopsin family is the largest, with 701
members.286 All GPCRs share a common architecture of
seven TM R-helices (H1-H7) (Figure 7E). They transmit
EC signals to the CP region via allosteric movements of TM
helices. The resulting changes in the CP surface facilitate
G-protein binding and activation, which, in turn, triggers a
cascade of cellular responses.287,288
The vast majority of the structure-based computations for
GPCRs have been done using the bovine rhodopsin structure,
originally resolved by Palczewski and co-workers.289 In
addition to the bundle of seven TM helices, referred to as
opsin, the structure contains an 11-cis-retinal (chromophore)
deeply embedded in the core (Figure 15A). The EC domain
consists of the N-terminus and three interhelical loops
EC1-EC3; the CP domain contains three interhelical loops
CL1-CL3 connecting respective pairs of helices H1-H2,
H3-H4, and H5-H6, and a C-terminal helix H8 that runs
parallel to the membrane. The EC domain contains a β-sheet,
which serves as a lid to the chromophore binding pocket,
stabilized by a highly conserved disulfide bond between
Cys110 and Cys187. The retinal, covalently bound Lys296
on H7, undergoes a cis/trans isomerization upon light
activation. This gives rise to a local conformational strain
Chemical Reviews, 2010, Vol. 110, No. 3 1483
that propagates through the concerted rearrangement of the
TM helical bundle to the CP domain, inducing an opening
at the conserved D(E)RY motif, which is recognized by the
G-protein (Figure 15B). The active form, metarhodopsin II,
is reached after a series of photointermediates. It binds the
heterotrimeric G-protein, transducin, and interacts with
several other signaling proteins.
Recent years have witnessed a remarkable progress in the
number of newly solved GPCR structures.290 Comparison
of the structures of bovine opsin in its G-protein-interacting
form (referred to as opsin*)291 and rhodopsin shows, for
example, an outward tilt of 6 Å in TM6, and pairing of TM5
to TM6, in agreement with aforementioned experimental data
and the above computational model proposed for metar-
hodopsin II. Comparison of the ligand-free opsin292 and
opsin*, on the other hand, shows little structural difference,
suggesting that the opsin conformational population is shifted
toward the activated state in the absence of retinal and
G-protein. In addition to opsin structures, the structures of
four other unique GPCRs were recently determined,290
including the β2 adrenergic receptor (β2AR) structure by the
Kobilka lab, an additional β2AR structure that pointed to
the stabilizing role of cholesterol,290 the structures of an
antagonist-bound A2A adenosine receptor,293 turkey β1-
adrenergic receptor,294 and squid rhodopsin.295 All together,
these structures provide important information on the diver-
gent EC regions, differences in the ligand binding sites, and
convergent features of the TM domains.
The type and extent of conformational changes under-
gone upon activation of rhodopsin have been extensively
examined by various experiments128,134,150-155,161,296-298 and
computations.85,180,216,299-306 GNM and ANM studies216,300
show that the global mode is controlled by a broad hinge-
bending region that includes the chromophore binding pocket
and a number of highly constrained conserved residues in
the close neighborhood (e.g., E113, T118, E122 on H3, F261
and W265 on H5, Y268 on H6, and C187 on β4) such that
the structural changes locally induced upon the isomeric
transition of the cis-retinal are efficiently propagated through
cooperative rigid-body movements of the TM helices, toward
both the CP and EC regions. An effect of these cooperative
movements is opening the CP ends of the TM helices 3, 4,
and 6, thus exposing the ERY motif at the G-protein binding
site (Figure 15C). A model for the Meta II state has been
proposed300 by analyzing the lowest ANM modes in con-
junction with experimental data. The model was shown to
correctly predict 93% of the experimentally observed effects
in 119 rhodopsin mutants for which the decay rates and
misfolding data have been reported, including a systematic
analysis of Cys f Ser replacements.300
With the elucidation of a large number of structures, we
are now in a position to examine more closely the correlation
between the experimentally observed structural differences
and theoretically predicted conformational changes. We
performed a PCA of currently available rhodopsin and opsin
structures and compared the resulting PC modes to ANM
modes. Our data set includes 16 structures, comprised of 14
rhodopsin and two opsin X-ray structures. Out of N ) 348
residues, 312 are commonly resolved in the data set of
examined structures, excluding the segments 230-240 on
CL3, 311-313 between H7 and H8, and 327-348 at the
C-terminus. The distribution of the structures along the first
two principal modes is shown in Figure 16 panel A. These
two modes contribute about 62% and 12%, respectively, to
1484 Chemical Reviews, 2010, Vol. 110, No. 3 Bahar et al.

Figure 16. PCA and ANM calculations for rhodopsin. (A) Distribution of 16 X-ray structures in the subspace spanned by the PCA mode
directions 1 and 2. These respective modes account for 62% and 12% of the structural variability in the data set. The principal axis 1
differentiates the inactive and (putative) activated structures which are clustered in two distinctive groups, and the PCA axis 2 further
differentiates between the structures in the cluster of inactive rhodopsins (B). Superimposition of experimentally determined rhodopsin and
opsin structures, indicated by the labels on panel A. (C) Rhodopsin structure generated by deforming the opsin structure along the first
principal mode, p1. (D) Rhodopsin conformation predicted by deforming the opsin structure along the 20 lowest frequency ANM modes.
The 14 rhodopsin structures in the analyzed set include, in addition to the ground state289,307-311 and various photoactivated states,
lumirhodopsin,312 bathorhodopsin,313 9-cis-rhodopsin,314 photoacivated deprotonated intermediate,309 and thermostabilized mutants.310,315
These microstates are dispersed along the second principal axis. These calculations have been performed for the CR atoms only; the remaining
backbone atoms were reconstructed with the BioPolymer module of Sybyl 8.3 (Tripos). ANM calculations were performed using the relatively
short cutoff distance of Rc ) 8 Å, so as to release interhelical constraints.

the structural variability in the data set. The PCA clearly
separates the structures into two clusters along the first
principal axis. These two clusters may be viewed, in a sense,
as the two substates illustrated in Figure 2. Notably, the first
cluster includes all the 14 rhodopsin structures in the inactive
(sub)state and the second, two opsin conformations in the
putative active (sub)state. Mode 1 therefore unambiguously
distinguishes between these two substates, representative
conformations of which are displayed in Figure 16 panel B.
The second principal mode, on the other hand, further
disperses the structures within the first cluster. This mode
essentially refers to the changes in loop conformations and
termini orientations. These can be viewed as the microstates
in the inactive substate.
Rhodopsin thus provides an excellent example of how
functional modes can be determined through PCA and
NMA. One utility of PCA is to provide us with a simple
organization of the ensemble of conformations accessible
to a given protein, and this use will become increasingly
valuable with the growth in PDB structures for the same
protein, hence the development of PCA servers to perform
such tasks.44 The second utility is to assist in our assessment
of the dominant changes in structure, which is usually
described by the 3N-dimensional PCA mode 1, p1, and the
corresponding amplitude of motion scales with σ11/2 (see eq
7). Panel C in Figure 16 illustrates how the rhodopsin
conformation (red) is closely reproduced upon reconfiguring
the opsin structure along p1. Comparison of the range of the
principal axes 1 and 2 in Figure 16 shows that the size of
motions along p1 is at least twice as large as that along p2.
Third, and most importantly, the principal modes may be
directly compared with those predicted by NMA. The PCA
modes are exclusively based on experimental data for an
ensemble of structures, while ANM modes are predicted by
the theory for a single structure. Comparison of the two sets
can help benchmark the computational predictions, provided
that the experimental data set represents a more or less
complete ensemble (see, for example, the study performed
by the Jernigan lab for HIV-1 protease190), or consolidate
the results, given that both sets involve approximations. In
the present case, the set of PDB structures is far from
complete. Yet, ANM calculations performed for the two
representative structures (labeled) from each cluster showed
that p1 exhibits a cumulative overlap of 0.79, with the first
20 ANM modes intrinsically accessible to opsin, and a
cumulative overlap of 0.74, with the first 20 ANM modes
accessible to rhodopsin. Thus, 2% of ANM modes in the
low frequency regime provide a reasonable description of
the change observed experimentally. The reconfiguration
predicted by moving the opsin structure along these ANM
modes is shown in panel D. These results again confirm the
view that the relative movements of the TM helices 5 and 6
observed upon light activation are intrinsic properties
encoded in the rhodopsin architecture.
3.3. Transporters
Transporters are generally active carriers. They require an
energy-producing process to translocate a substrate against
its concentration gradient. Secondary active transporters take
advantage of the movement of a solute down a concentration
gradient, so as to translocate another substrate across the
membrane. Glutamate transporter, discussed below, is an
example of such a transporter, where the uphill translocation
of glutamate is coupled to downhill Na+ transport.
Functional Mechanisms of Membrane Proteins
Figure 17. Release, uptake, and reuptake of glutamate at an
excitatory synapse. Upon arrival of an action potential at the
presynaptic axon terminus, voltage-sensitive Ca2+ channels trigger
the fusion of vesicles with the cell membrane to release glutamate
molecules in the synaptic cleft. Glutamates bind and activate
receptors on the postsynaptic cell membrane. Excess glutamate is
cleared by glutamate transporters, which are more abundant and
efficacious in the glia in the vicinity of the synapse.
3.3.1. Glutamate Transporters
The concentration of glutamate in the EC space may
increase by 103-104 fold during periods of synaptic activa-
tion, and it is critical to have a mechanism in place to clear
the excess glutamate, which, otherwise, may have neurotoxic
effects. Glutamate transporters clear the excess glutamate
molecules. They are located on neurons and glia (Figure 17).
Their precise functioning (glutamate uptake and reuptake)
is essential not only to protect against excitotoxicity316,317
but also to regulate glutamatergic signal transmission by
preventing sustained activation and desensitization of iono-
tropic receptors and modulating the activation of metabo-
tropic receptors.318 Members of this family, including the
human excitatory amino acid transporters (hEAATs), utilize
Na+-derived electrochemical gradients to transport glutamate,
hence their classification as the family of dicarboxylate/amino
acid: cation symporters,319 also referred to as the glutamate
transporter family.320 Notably, these transporters also function
as chloride channels.321,322
The first member structurally resolved in this family is an
archaeal aspartate transporter, GltPh, from Pyrococcus
horikoshii.323,324 GltPh provided for the first time a structural
model for gaining insights into the molecular basis of
glutamate transport by the human orthologs, hEAAT1-5.325
The top view (from the EC side) and side view of GltPh can
be viewed in the respective Figures 18A and 19B. As can
be seen, GltPh is a homotrimer, the three monomers of which
are arranged cylindrically to form a bowl-shaped basin of
∼50 Å diameter and ∼30 Å depth toward the EC surface at
the center of the protein. Each subunit is composed of two
domains: the N-terminal domain consists of TM1-TM6
(gray in Figure 18A), and the C-terminal core is comprised
of the helices TM7 and TM8, and helical hairpins HP1 and
HP2 (colored in Figure 18A).324
We note that TM7 has an unusual structure, with two
helical segments connected by a partially unwound conserved
motif N310MDGT314.324 This motif is in a sense “frustrated”,
containing potentially hydrogen-bond-forming groups that
Chemical Reviews, 2010, Vol. 110, No. 3 1485
lack partners, hence the role of this motif in binding the
substrate (confirmed in simulations326) and the clustering of
several conserved residues near this region (Figure 18A). In
other words, the irregularity of this helix at its central portion
is actually functional. We note that this region is located
exactly near the tips of the two loops HP1 and HP2. Another
region important in substrate recognition is indeed the serine-
rich tip of the HP1 loop, which is in van der Waals contact
with the tip of HP2. The amphipathic TM8 helix also shows
irregularities, which have been observed in simulations326
to be involved in substrate binding and channeling.328 HP8
has indeed been proposed to form a portion of the transport
pathway.320 Intersubunit contacts are confined to the N-
terminal domain, suggesting that the three core domains
function independently of each other, as also suggested by
experiments.327
We recently examined the substrate recognition and
binding events of GltPh by MD runs of tens of nanosec-
onds.326 Our simulations clearly showed that the HP2 hairpin
acts as an “EC gate”, in accord with the mechanisms
inferred324 from structural data. The fluctuations of this gate
between its open and closed substates occur within the time
scale of nanoseconds. Figure 18B illustrates the sequence
of events observed326 in a typical run, starting from diffusion
of the substrate toward the partially open “gate” and
continuing with the recognition of conserved glycines (G354
and G357) at the HP2 tip, gradual insertion into the region
between HP1 and HP2, near a conserved 3-Ser motif on HP1,
and subsequent stabilization at a site near the unwound
portion of TM7. The latter site and geometry are in close
agreement with the binding pose experimentally observed323
for aspartate in the high resolution structure of GltPh.
These conventional MD simulations thus provided ample
information on the early recognition and binding events.
However, no substrate translocation could be observed in
these simulations. Instead, we performed nonequilibrium
steered MD (SMD) simulations,328 which helped elucidate
two substrate translocation pathways, one of which is more
readily accessible, along with the key interactions and energy
barriers encountered during the translocation. The SMD thus
provided information on “slower” events, which would
otherwise be inaccessible via classical MD.
Notably, both substrate binding and substrate translocation
events examined in these two sets of simulations are local
events that involve the core regions in each subunit, and they
appeared to occur independently in the three subunits,
consistent with experimental observations. There is, however,
yet another dimension: the movements that cooperatively
engage all three subunits, which may perhaps explain the
raison d’être for the functioning of glutamate transporter as
a trimer, rather than as three monomeric proteins. ANM
analysis of the global dynamics of GltPh yields the move-
ments depicted in Figure 18C as the first nondegenerate mode
(mode 3), along with a doubly degenerate asymmetric
stretching/contraction mode accessible to the transporter. In
mode 3, the three subunits undergo concerted opening/closing
movements to alternately expose and cover the central basin.
The diagrams in the figure are color coded, with the red
portions corresponding to the most mobile regions. The
largest movements are undergone in this case by the
N-terminal domain residues that are exposed to the EC region
and the aqueous basin. It is interesting to note the possible
occurrence of intersubunit contacts between the EC-exposed
portions of TM8 and HP2 during the concerted movements
1486 Chemical Reviews, 2010, Vol. 110, No. 3 Bahar et al.

Figure 18. Structure and dynamics of glutamate transporter. (A) The homotrimer, viewed from the EC region. The N-terminal region
(TM1-TM6) is displayed in gray; the C-terminal core HP1-TM7-HP2-TM8 is colored yellow-orange-red-violet and labeled in the
figure on the right. (B) Snapshots from MD simulations, illustrating the time-resolved recognition and binding events, starting from t ) 0,
where the substrate is in the aqueous cavity, up to t ) 7.5 ns, where the substrate is sequestered at the binding site and remains therein for
the remaining duration of the simulation, of ∼20 ns.326 (C) Symmetric opening/closing mode of GltPh, as observed in ANM. The middle
diagram displays the GltPh structure viewed from the EC side (PDB: 1xfh); the top and bottom diagrams display the ANM-predicted open
and closed conformations, respectively. In the X-ray structure, the basin is exposed to the EC aqueous environment, while in the closed
form contacts between neighboring subunits occur (see, for example, the L34 loops colored red).

of the three subunits in this mode. The diagram and arrows
in Figure 19 panel B illustrate the same movement, viewed
from the side; and Figure 19A displays the corresponding
mobility profile (normalized distribution of residue square
displacements driven by this mode). G144 and T182 exhibit
minimal mobilities. These two residues are located at the
base of the aqueous basin on the EC and IC sides,
respectively, and are proposed to play a role in modulating
the concerted motion of the subunits. The base of the EC
vestibule indeed remains rigid and immobile during these
movements, which may be a requirement to maintain the
integrity of the trimer. As to the peaks in the mobility profile,
we note a number of histidines exposed to the EC region. It
is interesting to note, in this context, that Vandenberg and
co-workers observed that Zn2+ ions inhibited the anion
conductance of EAAT4 and attributed this inhibition to the
binding of Zn2+ to His146 and His154 conserved in EAAT1,
-2, -4, and -5.329 Interestingly, these two histidines lie very
close, both sequentially and spatially, to the top-ranking
residue (His114 in GltPh or Lys152 in EAAT1) in the global
mode profile, suggesting that the peak residues observed here
could possibly serve as “sensors” for capturing negatively
charged substrates.
3.3.2. ATP Binding Cassette Transporter BtuCD
ATP binding cassette transporters mediate the transport
of various substrates, including ions, drugs, lipid molecules,
and small proteins, across the membranes via an ATP-
dependent mechanism.330,331 BtuCD is a member of the
family of ABC transporters that transports vitamin B12 in E.
coli, and it has been investigated by both structural and
computational studies.
The BtuCD complex, like most ABC transporters, consists
of four subunits, arranged as two homodimers: the TM dimer,
BtuC, that forms the specific substrate translocation pathway,
consisting of 20 TM helices; and the CP dimer, BtuD,
composed of two nucleotide-binding domains (NBDs) where
the ATP binding and hydrolysis activities take place.332-334
The TM dimers of ABC transporters usually exhibit little
sequence similarity, with their sequence being specific to the
particular substrate that they recognize and translocate. The
Functional Mechanisms of Membrane Proteins Chemical Reviews, 2010, Vol. 110, No. 3 1487

Figure 19. Global dynamics of the aspartate transporter GltPh predicted by the ANM. (A) Distribution of square displacements of residues,
(∆Ri)2|1 + (∆Ri)2|2 (see eq 26), induced in the asymmetric stretching-contraction mode (a 2-fold degenerate mode). The same profile is
induced in all three subunits upon superposition of these two lowest frequency modes, leading to a cylindrically symmetric reconfiguration.
Peaks refer to the most mobile residues, and minima to the hinge centers (e.g., Gly144 and T182) controlling the concerted movements of
the subunits. The large amplitude swinging movements of the extracellular histidines suggest a possible role in facilitating the attraction of
the anions or engulfing them into the central basin. (B) Mechanism of motion in the first nondegenerate ANM mode (see also panel C in
Figure 18). The arrows indicate the direction of the concerted movements of the three subunits (note that the third subunit in the back is
lightly visible).

NBDs, on the other hand, are highly conserved, sequentially

and structurally. They share the common motifs Walker A
and Walker B typical of many ATP-binding domains, as well
as a signature sequence, also called the C motif, that
completes the ATP binding site at the interface between the
two NBDs.
Examination of the crystal structure of BtuCD332 and MD
simulations performed for BtuCD335 and MalK (an ABC
transporter with similar architecture336) suggests that ATP
binding to the BtuD dimer triggers conformational changes
that propagate to BtuC. ATP binding induces a closing at
the interface between the two NBDs of the BtuD dimer and
stabilizes the interactions (or reduces the fluctuations) at the
interface, as confirmed by MD simulations337 of the dimer.
However, the extent to which these structural changes occur
in the presence of the BtuC dimer has been questioned, given
that the MD simulations for the tetramer did not exhibit the
decrease in the inter-NBD distance observed upon ATP
Figure 20. Global dynamics of the ABC transporter BtuCD. Top
binding in the BtuD dimer.337 A related question of interest and bottom panels display the collective motions of the tetramer
is the role of BtuF, the protein that delivers the vitamin B12 in the ANM modes 1 and 3, respectively, recently examined by
to the periplasmic entrance of the BtuCD, on the dynamics Weng et al. (2008). The color-coded diagrams on the left in both
of the transporter. panels display the size of motions (red, most mobile; blue, almost
rigid) induced in these modes. The other diagrams display the
NMA of BtuCD dynamics was recently performed by Ma relative motions of the two TM domains (1, red; 2, yellow) of the
and co-workers toward elucidating the intrinsic dynamics of BtuC dimer, and the two NBDs (3, blue; 4, green) of the BtuD
the transporter.338 In this study, the lowest frequency modes dimer, that compose the BtuCD tetramer, viewed from the side
accessible to the two dimers, BtuC and BtuD, were analyzed (middle diagram in top panel) or from the EC region (right diagrams
separately with ANM, as were those accessible to the in both panel). The two gates (EC and CP gates) of the substrate
(vitamin B12) translocation pore are indicated by the orange arrows
tetrameric BtuCD structure (of ∼1100 residues). This in the left diagram of the top panel.
analysis demonstrated that the lowest mode accessible to the
tetramer, ANM mode 1, is a highly cooperative motion that
involves the reverse rotations of the two halves of the BtuCD The mechanism of BtuCD mode 1 is more clearly
transporter, as illustrated in Figure 20, top panel. Interest- visualized from the side (middle diagram in Figure 20 top
ingly, the relative movements of the two TM domains in panel, obtained by rotating the diagram on the left by 90°
this mode correlate (with an overlap of 0.83, after removing around the vertical 2-fold symmetry axis). This view shows
the rigid-body contributions) with the mode 1 accessible to that monomers 2 and 4 (belonging to the respective dimers
the BtuC dimer alone,338 suggesting that the intrinsic dynam- BtuC and BtuD; see labels in panel A) rotate almost rigidly
ics of BtuC strongly affects this global mode of the tetramer. together, while monomers 1 and 3 undergo an opposite
This mode was also shown to be insensitive to BtuF rotation with the pivot located near the CP gate of the
capping.338 translocation pathway.338 As a result, this movement induces
1488 Chemical Reviews, 2010, Vol. 110, No. 3
a simultaneous shear opening between the pairs of NBDs (3
and 4) or their ATP binding sites and between the pairs of
TM domains (1 and 2) near the EC ends of the TM helices,
while the CP gate undergoes minimal, if any, displacement
(see also the color-coded diagram on the left). Thus, closing
of NBDs upon ATP binding is accompanied by a simulta-
neous closing of the EC pore, which prior to this allosteric
effect was sufficiently large to accommodate vitamin B12
entry to the translocation pathway. Thus, Weng et al.
proposed that the vitamin B12 molecule is trapped into the
periplasmic cavity rather than transported to the CP region,
upon ATP binding.338 Weng et al. further proposed that the
conformational change required for substrate translocation
and the opening of the CP end of the pore is associated with
ATP hydrolysis (rather than ATP binding). In particular,
mode 3 has been proposed to contribute largely to the
conformational change powered upon ATP hydrolysis. As
can be seen in the bottom panel of Figure 20, this mode
induces an opening at the pore region of the BtuC dimer.
BtuF capping appears to restrict these movements, while
another mode (mode 7, not shown) has been pointed out to
effectively enable substrate transport. Modes 3 and 7 of the
tetramer have been shown to be similar in shape to the lowest
two modes favored by the BtuD dimer structure,338 which
supports the significance of these intrinsically accessible
modes in mediating the ATP-dependent coupling of the
dimers.
4. Conclusion
Over the years, many techniques have been developed to
tackle an ever broadening range of problems using NMA in
general and ENMs in particular. The linearity of the theory
endows it with considerable flexibility, and the clever
applications of matrix algebra to NMA have expanded its
utility. In particular, ANM and its extensions have been of
great use in studying dynamic phenomena that exceed the
time or length scales of MD, such as investigations of
Megadalton-scale structures’ dynamics,64,81,207,232,233,339 ex-
ploring the CG transition pathways,340-346 and studying the
effects of crystal packing on protein dynamics.194,347-350 Other
studies have taken advantage of the computational efficiency
of the GNM/ANM to perform serial analyses of large data
sets and gain insights into design principles. An example is
the colocalization of global hinge sites and catalytic sites in
enzymes, which appears to be a mechanism of efficiently
coordinating the mechanical and chemical activities of the
protein.32 Another example is the intrinsic ability of the
proteins in the unbound form to undergo structural changes
that are stabilized upon substrate binding.21,30,351 We pre-
sented several applications to membrane proteins in section
3. Below, we present an overview of insights into mecha-
nisms and principles of functional dynamics gained from
ENM-based studies (section 4.1) and recent extensions that
are anticipated to be exploited and further developed in future
studies (section 4.2).
4.1. Robustness and Functionality of Global
Modes
4.1.1. Robustness of Global Modes: A Requirement in
Evolutionary Selection of Structures
Designable protein structures are usually referred to as
structures that are the lowest energy conformer for a
Bahar et al.
multitude of sequences; that is, these structures can usually
tolerate sequence substitutions with minimal change in their
overall fold. This type of insensitivity to sequence variations
is what makes a stable structure, in a sense. But stability
does not necessarily imply functional aptitude.
Function, on the contrary, requires a well-defined flexibility
and conformational malleability, within a coarse-grained
view of the global energy minimum, perhaps evidenced by
substates that are accessible via small energy barriers. In the
same way that stable structures are those which are insensi-
tive to sequence variations, one can think of functional
proteins as those whose dynamics are insensitive to structural
details. Indeed, the success of ENM-based NMAs presum-
ably originates from the insensitivity of global modes to
structural and energetic details.
The observed robustness of global modes may reflect an
evolutionary pressure. Stable structures are those mapped
onto by many sequences, according to the designability
principle set forth by Wingreen and co-workers.352 Functional
structures, on the other hand, are proposed to be those that
intrinsically favor the global modes that facilitate/accom-
modate biological functions such as substrate binding,
translocation, or gating by membrane proteins; the global
modes are in this case favored, or mapped onto, by the
overall architecture despite minor changes/perturbations in
structure.
4.1.2. Toward Gaining Insights into Functional Dynamics
of Membrane Proteins
Biomolecular dynamics is a complex process. In
particular, the transitions between conformational states
separated by high energy barriers, such as the folding of
proteins, continue to pose a challenging problemsexcept
for small proteins where some success has been recently
achieved. The transitions between microstates within a
global energy well, or between substates separated by
relatively low energy barriers, on the other hand, appear
to be a more tractable problem, with the development and
applicability of elastic network models and PCA-based
methods. The rapidly increasing structural data now permit
us to test and improve these coarse-grained models and
methods. In the present review, we illustrated the recent
applications to membrane proteins, as a group of proteins
that are extremely important from biological and phar-
maceutical points of view.
These studies provide us with insights into the collective
mechanisms of motions preferentially accessed by mem-
brane proteins. A striking observation is the occurrence
of a global twisting as a mechanism of pore opening or
ligand gating in many membrane proteins. The “twist-to-
open” mechanism instrumental in the gating function of
most of the membrane proteins discussed here suggests a
common mechanism of pore-opening when the pore
architecture exhibits a cylindrical symmetry with funnel-
like organization of a bundle of helices. Another observa-
tion is the high cooperativity of the motions, which
becomes even more pronounced by the structural sym-
metry or multimerization. In this respect, nAchR presents
a unique structure, being a heteropentamer. Yet, the
dominant mechanism conducive to channel opening is
again observed to be a global twist. It is also interesting
to note that the transition between the closed and open
forms has been observed in many applications, to be
realized by a small subset of modes at the low frequency
Functional Mechanisms of Membrane Proteins
regime, and among them, nondegenerate modes usually
provide the most productive paths leading to functional
substates in the case of structurally symmetric multimers.
The passage between the open and closed forms of the
MscL achieved by a few nondegenerate modes is a typical
example (Figure 13).
ENM-based NMAs not only provide insights into the most
easily accessible movements of quaternary structures but also
point to sites that may play a critical role in mediating or
propagating allosteric signals. In the same way as there are
particular amino acids whose substitution may be deleterious
(conserved residues) to stability, there are particular sites on
the structure whose perturbation could impair the global
dynamics (e.g., hinge sites in the global modes). We have
learned that these sites are utilized by proteins to elicit
cooperative responses, e.g., ligand binding pockets that
efficiently transmit allosteric signals, especially if fueled by
the energy released by an exothermic reaction (e.g., ATP
hydrolysis) in the vicinity. These sites are referred to as
mechanically critical sites32 or sites with a high allosteric
potential.238,239 Not surprisingly, more and more structures
show us that active sites, the drug binding sites, or residues
that are known to mediate allosteric effects, or ATP binding
sites, coincide with, or closely neighbor, such mechanically
critical sites. It is clear that an improved understanding of
the structural basis for allosteric and chemical communica-
tions in these proteins will assist in the rational discovery of
drugs against the various channelopathies or signaling
diseases.
4.1.3. Many Functional Motions of Membrane Proteins
Are Intrinsic to Their 3D Structure, Independent of
Membrane Environment
Evidence for the dominance of intrinsic dynamics in
defining certain collective motions and/or allosteric mech-
anisms of membrane proteins (such as gating or signaling),
independent of the membrane environment, is provided by
the applications presented in section 3.
We began with gramicidin A, for example, in section 3.1.1.
The calculations performed by Miloshevsky and Jordan
unambiguously demonstrated the equivalence of the NMA
results from ENMs and those from full atomic models in
the presence of explicit lipid and water molecules subject to
the CHARMM22 force field.192 In particular, the gating
mechanism (counter-rotation of the two helices) was con-
cluded to be an inherent property of the GA architecture,
independent of surrounding lipid and water molecules.
Likewise, the ANM calculations performed for a series of
potassium channels by Shrivastava and Bahar218 yielded
results (cooperative rotational/twisting motions of M2 helices
to induce pore opening) in excellent agreement with SDSL
EPR data from Perozo’s lab157,158 and confirmed by recent
structural data (kink region) determined for a newly resolved
cation channel (NaK),251 again showing that the rigorous
consideration of the native contact topology permits us to
predict global movements relevant to function regardless of
the potential perturbation of the membrane environment. In
the case of nAChR, the quaternary twist model from ANM
studies270-272 not only agrees with the mechanisms inferred
from MD and Brownian dynamics simulations268,269,274 but
also compares favorably with the newly elucidated closed
and open structures of the ligand-gated ion channels GLIC284,285
and ELIC.283 The global movements of the archaeal aspartate
transporter GltPh, on the other hand, drive the cooperative
Chemical Reviews, 2010, Vol. 110, No. 3 1489
opening/closing of the three subunits around the central
aqueous basin, which were not observed in tens-of-
nanoseconds simulations. The computations with GltPh using
MD,326 steered MD,328 and ANM indeed provide a nice
example of the complementarity of results and the utility of
exploiting multiple scale computations. An even better
approach is to develop integrated MD-ANM approaches, like
the ANM-steered MD applied to rhodopsin,85 which simul-
taneously provided access to global movements while
viewing atomic interactions and rearrangements. A striking
observation in this case is the close correspondence between
ANM modes and the dominant modes derived from the PCA
of 16 X-ray structures resolved for rhodopsin in different
forms. The fact that a small subset (2%) of ANM modes in
the low frequency regime yields an overlap of ∼0.75 with
the principal modes of deformations derived from experi-
mental data again lends strong support to the physical and
biological significance of ANM modes, which are based
exclusively on the protein structure, independent of mem-
brane environment.
The mechanisms of collective movements essential to
certain functions such as gating or allosteric signaling thus
appear to be intrinsic to protein structure, in accord with
ENM-based predictions exclusively based on the inter-residue
contact topology of the membrane protein. However, the
function of membrane proteins involves many other specific
and subtle interactions that cannot be studied by CG models
and NMA, such as the selection of particular ions at the
selectivity filter, recognition and binding of substrate by
specific interactions, and the assistance of substrate trans-
location by cotransported ions. Moreover, ENM-based NMA
cannot provide information on the absolute time scales of
the movements either, due to lack of a proper consideration
of the frictional drag or other environmental factors that may
affect the relative frequencies or probabilistic occurrence of
different modes of motions. Essentially, the ENM-based
studies provide information on the “accessible” most coop-
erative movements that are selected/recruited functions.
However, there also exist several accessible, energetically
favorable movements at local scales, including side chain
isomerizations or specific reorientations of polar groups,
which require full atomistic and, in some cases, even
quantum mechanical calculations. Adding to the complexity
is the potential coupling between local events and global
movements, hence the need for developing multiscale
methodologies that take advantage of the capabilities of both
MD and NMA.
4.1.4. Entropic vs Enthalpic Effects, or Geometry vs
Specificity
It is important to note that ENM-based approaches are
based on purely geometric considerations such as inter-
residue contact topology or overall shape/architecture of the
examined structure. As a result, the predicted movements
are those which are entropically favored, as originally
proposed for polymeric networks,69,70 and do not contain
contributions from specific (enthalpic) interactions. ENM-
based approaches are therefore useful to the extent that
geometry or topology plays a dominant role in the process
being explored. The machinery of supramolecular systems
could be a prime example for processes dominated by
collective mechanics, rather than specific/local chemical
events. However, in many applications it may be important,
if not indispensable, to invoke both effects and adopt hybrid
1490 Chemical Reviews, 2010, Vol. 110, No. 3 Bahar et al.

Figure 21. Critical interactions near the chromophore binding pocket and CP ends of TM1, TM2, and TM3 in rhodopsin. The ribbon
diagram on the left is color-coded (from red, least mobile, to blue, most mobile) by the RMSDs observed in the positions of residues during
ANM-steered MD simulations of rhodopsin activation. Two regions enlarged on the right are distinguished by their highly constrained
dynamics: the chromophore binding pocket and the CP end of helices 1, 2, and 7. The tight packing in the former region ensures efficient
propagation of the local conformational strains (induced upon cis f trans isomerization of the retinal) to distant regions, including in
particular the ERY-binding motif at the CP end of helices H3 and H6 (note the enhanced mobility at this region). Water molecules play an
important role in stabilizing the CP ends of TM1, TM2, and TM7. For more details, see ref 85.

or multiscale approaches. Recent years have indeed seen a
large number of studies in that direction, which have been
partly reviewed here. Overall, it is important to interpret the
ENM-predicted dynamics as one aspect of the complex
mechanochemical process, mainly that intrinsically preferred
by the particular architecture, in the absence of perturbations
from specific intra- and intermolecular effects.
4.2. Extensions of Coarse-Grained NMA and
Future Directions
4.2.1. Hybrid Methods That Integrate CG NMA and MD
As discussed above, MD simulations realistically ex-
plore, in the presence of explicit solvent and/or membrane,
events on the nanosecond time scale for biomolecules of
a few hundreds of residues, while their application to
larger systems (e.g., multimeric proteins) and longer
processes (e.g., of the order of microseconds or slower)
suffers from sampling inaccuracies. NMA with ENMs
provide an understanding of the global movements of
Megadaltons systems, but at the cost of loosing accuracy
and specificity at the local scale. The above-described
applications of MD and NMA- methods to membrane
proteins nicely illustrate the capabilities and limitations
of the two sets of computations. For example, the CG
modeling and the atomistic simulations of inward rectify-
ing potassium channel Kir3.4.1353 provided similar pictures
of the overall dynamics of the ligand-binding domain,
suggesting dimer-of-dimers motion as an intrinsic property
of the CP domain of this K+ channel. Thus, combining the
data from these alternative computational approaches may
help consolidate the inferred mechanisms, if a consistent
behavior is captured. Likewise, the PCA of MD trajectories
generated for BtuCD in a lipid bilayer supports the hypothesis
that ATP-binding drives closure of the nucleotide binding
domains in BtuD, while the apo state of BtuD randomly
switches between open and closed substates,337 consistent
with ANM analysis of the same dimer. It is not generally
sufficient, however, to perform and compare two independent
sets of computations, such as MD and CG NMA, to make
inferences on multiscale dynamics. Instead, there is a need
to develop hybrid methodologies or more integrated ap-
proaches that exploit the complementary utilities of the two
methods and take account of possible couplings between
these different scale events.
A new protocol that steers MD along ANM modes has
been recently developed to this aim and used to study of the
conformational changes associated with photoactivation
signal transmission in rhodopsin.85 In this method, global
conformational changes that are not accessible via conven-
tional MD trajectories can be sampled, while motions and
interactions at atomic scale can be observed in the presence
of explicit solvent and lipid bilayer. Two stable regions were
identified by this method for rhodopsin, one clustered at the
chromophore and the second at the CP end of the TMs 1, 2,
and 7 (Figure 21). These simulations elucidate the redistribu-
tion of the interactions between the retinal and its neighboring
residues on H3-H6, induced upon cis f trans isomerization
of retinal. Eleven of the 16 residues identified to participate
in the central hinge region near the retinal have been tested
by experiments and confirmed to play a critical role in
stabilizing the activated state. Furthermore, these simulations
draw attention to the possible role of water molecules in
coordinating the interactions between conserved residues at
the CP ends of the helices H1, H2, and H7, illustrated in
Figure 21.
Another study in the same spirit is the examination of the
allosteric changes in the conformation of BtuCD TMDs.
These movements that appear to be induced upon ATP
binding to the BtuD dimer have been examined by Tieleman
Functional Mechanisms of Membrane Proteins
Figure 22. Use of NMA in modeling protein-ligand interactions. Alternative conformations for the target protein were generated for (A)
cAMP-dependent protein kinase382 (PDB ID, 1JLU; inhibitor PDB ID, 1REK), (B) matrix metalloproteinase-3383 (PDB ID, 1UEA; inhibitor
PDB ID, 1HY7), and (C) cyclin-dependent kinase 2384 (complex PDB ID, 1G5S), by reconfiguring these target proteins along the global
modes of motions indicated by the arrows. See text for details.
and co-workers using the perturbed ENM (see subsection
2.3.4) and biased MD simulations.240 The results support the
MalK model for the transport mechanism; that is, a closure
of the nucleotide binding domain upon ATP binding is
predicted, which results in closing of the TMD toward the
CP side while inducing an opening toward the periplasmic
side.
4.2.2. Docking and NMA in Drug Discovery
Understanding the mechanism of interactions between the
target protein and a small molecule inhibitor is of crucial
importance in drug discovery.102 Molecular docking is the
primary computational tool to model these interactions362 and
screen compound libraries of small molecules with potential
inhibitory/agonistic/antagonistic activities.363 There are nu-
merous successful applications of docking to membrane
proteins. Predix Pharmaceuticals, for example, targeted five
different GPCRs in in silico screens of commercially
available libraries and identified 11 compounds per target,
with an average hit rate of 17%.364 In another study, Wang
and co-workers targeted dopamine (D3) receptors and identi-
fied four compounds that bind at 100 nM levels, with 60%
hit rates.365
The ligand-selective conformational heterogeneity of
GPCRs has been recognized, however, as a limiting factor
in in silico efforts.287,366 The binding site geometries of
GPCRs differ, depending on the functionality and the potency
of bound ligands.367 Kinetic measurements and single
molecule spectroscopy both reveal that the 7TM helix bundle
samples distinguishable conformational states in the absence
or presence of ligand, and the populations of these confor-
mational states shift upon ligand binding.368,369 State-of-art
docking programs usually allow for only partial binding site
flexibility limited to optimizing a small number of side-chain
rotations or short loop conformers. Overlooking such con-
formational flexibilities hampers the success of in silico drug
discovery.
Abagyan and co-workers made prominent contributions
to developing algorithms and tools that take account of target
protein conformational flexibility,370,371 which have been
successfully applied to GPCRs.372 In particular, a ligand-
steered homology modeling approach was developed, which
uses existing ligands to shape and optimize the GPCRs
binding site.373 The idea therein is to start with hundreds of
crude homology models as probable conformations of the
target protein and then filter them based on their interaction
energy with known ligands probed by flexible docking and
Chemical Reviews, 2010, Vol. 110, No. 3 1491
on their ability to detect known ligands in virtual screening
tests. The utility of this approach was demonstrated by its
application to melanin-concentrating hormone receptor 1,
where a 10-fold improvement over random high-throughput-
screening was achieved and six novel antagonists were
identified. In a similar recently published study,374 β2AR
interactions with agonist/antagonist were examined upon
generating multiple conformations of β2AR. The models were
reduced and further refined by flexible docking of selected
agonists in the light of mutagenesis data to obtain models
that outperformed rhodopsin-based models. In accord with
these findings, Kobilka and co-workers reported that rhodop-
sin-based homology models of β2AR developed prior to
β2AR structure resolution were more similar to rhodopsin
rather to β2AR,375 stipulating the need to consider more
distinctive target conformations.
The generation of multiple conformations for the target
protein emerges from the above and other studies376-381 as
an important component of computational tasks for modeling
and simulating protein-inhibitor interactions. NMA with
ENMs appears to be particularly suitable for generating
backbone rearrangements. It suffices to have but one structure
to generate a distribution of energetically favorable confor-
mations in its neighborhood. Likewise, the method can be
used to refine/broaden an existing collection of conformations.
Figure 22 illustrates three cases where such NMA-based
generation of alternative conformers improved the perfor-
mance of docking simulations. Panel A shows the results
for cyclic AMP-dependent protein kinase.382 A ligand binding
loop in this protein is known to assume different conforma-
tions in the presence of different ligands. Alternative loop
conformations favored by the structure were determined in
this case by selecting from the ensemble of low-to-medium
frequency modes those that specifically induce reorientations
in this particular loop. The use of this ensemble in docking
simulations was shown to improve the discrimination rate
between binders and nonbinders.382 Panel B refers to the
study of matrix metalloproteinases inhibitors by Perahia and
co-workers.383 The global mode that directly affects the
opening/closing of the ligand binding cavity was identified
in this case to be the second lowest mode, and a set of
conformations was generated by gradually reconfiguring
the protein along that mode. Docking of inhibitors to the
resulting target ensemble was shown to improve docking
in all cases compared to docking to a single energy-
minimized structure. Finally, May and Zacharias used
NMA to improve protein-protein and ligand-DNA
1492 Chemical Reviews, 2010, Vol. 110, No. 3
docking in a number of studies.379,384,385 Panel C illustrates
the application to cyclin-dependent kinase 2.384 An
ensemble of ligands were randomly placed in the binding
site of the protein in this case, and the protein-ligand
interactions were optimized by deforming the protein along
the lowest ten modes accessible to the structure.384 The
approach improved all poorly docked cases, at only a modest
increase in computational cost. Two more studies from
Perahia’s lab support the utility of NMA-based modeling,
one involving a protein that interacts with a membrane
protein (CD47 receptor), and the other showing how col-
lective motions (of a synthase) relate to its catalytic
activity.386,387
4.2.3. Normal Modes for Structural Refinement
An obstacle to understanding the chemistry of biological
molecules has always been the determination of their
structures to atomic precision. Unless a protein readily
crystallizes or is sufficiently small to produce a clean NMR
signal, its structure cannot be determined to high resolution.
In many cases, the structure can be predicted by homology
modeling and then refined to locate the optimal conformation
for the particular sequence. Feig and co-workers showed388
that a good set of decoy structures against which to refine
can be generated by distorting a homologous template along
its slowest normal modes. Indeed, because these slowest
modes indicate the easiest directions of motion, refinement
using normal modes produces higher resolution structures
than does refinement using other CG methods or MD.388
Normal modes are also used in refining electron micros-
copy (EM) structures. The idea is to use structural data
available from X-ray crystallography or homology modeling
for substructures (e.g., individual domains, subunits, etc.) and
exploit their NMA-predicted alternative conformers to
optimally fit cryo-EM data for the intact structure. Several
methods have been developed for structural refinement using
ENMs,389-395 some of which are available as software
packages.201,396 The basic technique in these approaches is
to start with a known high-resolution structure and iteratively
alter it along its normal modes, preferably the slow modes,
until its structure agrees with the EM density map.
Vector quantization-based techniques have also been used
to predict the collective motions of macromolecules from
low-resolution structures.397,398 The underlying idea therein
is that an ENM constructed around the EM density map
produces the same dynamics as an ENM constructed using
the detailed structure that is represented by the density map.
The map is divided into a set of discrete points that act as
nodes in the ENM, and the global dynamics are calculated
with NMA. This technique has been shown to predict
motions in accord with experimentally observed fluctua-
tions.399
Recently, the combination of MD and NMA results for
DHFR complexed with nicotinamide adenine dinucleotide
phosphate by the Perahia lab demonstrated that the inelastic
neutron scattering spectrum may reflect proteins trapped in
different conformations (at 120 K), in addition to the
vibrational modes of different conformations, leading to
inhomogeneous broadening of the spectrum.400
Bahar et al.
Figure 23. Schematic representation of the energy landscape for
two substates. The cartoon shows the putative free energy landscape
around a conformational transition for a two-state system. Both
conformations, A and B, are contained within a global free energy
well, represented here as the outermost oval. The slowest mode of
the well, indicated by the broken blue line, is expected to overlap
with the transition between states A and B. Each stable conforma-
tion lies at the bottom of its own local well. The transition between
states (red dotted line) is expected to proceed along the slowest
local mode in the vicinity of each end point. The slow modes
accessible to the metastable intermediate conformation between the
end points provide further information on the pathway near the
transition point.
4.2.4. Exploring Allosteric Transitions in Large
Biomolecular Systems
Proteins usually sample multiple substates, prompted by
an external event, such as ligand binding or assembly with
another protein, especially if an allosteric change in confor-
mation is triggered. The details of the transition from one
state to the other are in most cases only marginally
understood; the transition likely does not follow a single
linear trajectory but instead winds through a complex energy
landscape. Nonetheless, these global transitions appear to
proceed, or are at least initiated, via the collective global
mode directions that are studied with NMA, and it has been
shown in many applications that biomolecular structural
transitions between functional substates are largely accounted
for by a few slow modes. This observation puts ENMs
among the primary tools for theoretical studies of transition
pathways.
Since NMA is valid only in the local region surrounding
a potential energy minimum, its application to nonequilibrium
events such as conformational changes must be handled
delicately. When studying the simplest case of a system with
two stable conformationsscall them “A” and “B”sit is
assumed that each conformation resides at the bottom of a
harmonic potential energy well and that the transition state
is sufficiently close to both end points as to be within the
range of the harmonic approximation about each conforma-
tion. The system can then transition smoothly from the
harmonic surface surrounding A to the harmonic surface
surrounding B.
An early technique for studying transition pathways using
ENMs340 involves constructing two ENMs for the initial
conformation (Figure 23). The topology of the first ENM is
determined by the contacts in state A, and the topology for
the second is determined by the native contacts in state B.
The system is initially modeled with the EN for state A only,
and the transition proceeds by gradually reducing the effects
of the state A EN while increasing the contribution of the
Functional Mechanisms of Membrane Proteins
state B EN. Throughout the transition, intermediate confor-
mations are calculated by minimizing a cost function based
on the instantaneous EN. This method has since been adapted
to use rigid clusters or combinations of rigid clusters and
pointlike beads.341 A related method342,343 involves construct-
ing an ENM for the initial state only and then slowly
perturbing the structure to satisfy known distance constraints
from the final structure. A similar “nonlinear” elastic model,
in which the modes are continuously modified in a series of
small steps through the transition, was used to study the open
to closed transition of adenylate kinase.345 All three methods
are capable of generating putative transition pathways with
low computational cost.
Another technique, the plastic network model,344 combines
the potentials about states A and B into an analytical double-
well potential. The transition states are defined as conforma-
tions on the cusp between states A and B. This model was
also used to investigate the allosteric transition in adenylate
kinase. A similar “mixed ENM” was used to investigate the
helix-to-sheet transition of the Arc repressor,401 as well as
transitions of kinesin and myosin.237 The plastic network
model was modified402 to make the potential a double-well
for all interactions. The resulting potential energy landscape
is unlike the original but has multiple local extrema. This
model, too, was used to investigate the open-to-closed
transition of adenylate kinase, revealing an alternative
transition pathway. Yet another method, the adaptive ANM
of Yang et al.346 gradually moves the structures from both
end points along their respective slow modes until a common
structure is reached at the assumed transition state. It is also
worth noting that not all NMA-based methods of generating
transition pathways are strictly analytical. The MC normal
mode following method of Miloshevsky and Jordan192 utilizes
normal modes to guide simulations. An advantage of this
technique is that it does not require two conformations but
generates an approximate potential energy landscape from
the initial conformation alone.
Besides the proteins discussed in this review article, there
are also other membrane proteins where the normal modes
predicted by CG NMA have been observed to correlate
closely with experimental observations on gating mecha-
nisms, for example acid-sensing ion channel (Asc1).403 In
this very recent study, modes 1 and 3 induce a twisting of
the entire TM domain coupled to the motions of a β-turn,
which in turn induces an opening of the channel pore. The
predicted mechanism is consistent with mutagenesis and
electrophysiological experiments. To quote the authors, “This
result indicates that the structure of the closed, desensitized
state of Asc1intrinsically tends to undergo a twisting motion
to open the gate”. This study reinforces the point made in
this review, namely, the “twisting-to-open” motion is a
common mechanism for gating membrane proteins with an
inherent cylindrical symmetry.
5. Acknowledgments
The authors benefited from Stephen White’s Web site on
membrane protein structure statistics, Dr. Yang Zheng’s
assistance in generating some of the ANM figures, and Dr.
Basak Isin’s useful comments on rhodopsin dynamics.
Support by NIH Grants 5R01 GM086238-02 and 5R01
LM007994-06 and by PA Department of Health (No.
0317401) is gratefully acknowledged by I.B.
Chemical Reviews, 2010, Vol. 110, No. 3 1493
6. References
(1) Birshtein, T. M.; Ptitsyn, O. B. Conformations of Macromolecules;
Interscience: New York, 1966.
(2) Flory, P. J. Statistical Mechanics of Chain Molecules; Wiley,
Interscience (2nd Edition published in 1989 by Butterworth-Heine-
mann Ltd): New York, 1969; pp 1-432.
(3) Volkenstein, M. V. Configurational Statistics of Polymer Chains;
Interscience: New York, 1963.
(4) Taketomi, H.; Ueda, Y.; Go, N. Int. J. Pept. Protein Res. 1975, 7,
445.
(5) Noguti, T.; Go, N.; Ooi, T.; Nishikawa, K. Biochim. Biophys. Acta
1981, 671, 93.
(6) Go, N. Annu. ReV. Biophys. Bioeng. 1983, 12, 183.
(7) Levitt, M.; Warshel, A. Nature 1975, 253, 694.
(8) Cooper, A. Proc. Natl. Acad. Sci. U.S.A 1976, 73, 2740.
(9) Frauenfelder, H.; Sligar, S. G.; Wolynes, P. G. Science 1991, 254,
1598.
(10) Frauenfelder, H.; Parak, F.; Young, R. D. Annu. ReV. Biophys.
Biophys. Chem. 1988, 17, 451.
(11) Frauenfelder, H.; Petsko, G. A.; Tsernoglou, D. Nature 1979, 280,
558.
(12) Gelin, B. R.; Karplus, M. Proc. Natl. Acad. Sci. U.S.A 1975, 72,
2002.
(13) Ansari, A.; Berendzen, J.; Bowne, S. F.; Frauenfelder, H.; Iben, I. E.;
Sauke, T. B.; Shyamsunder, E.; Young, R. D. Proc. Natl. Acad. Sci.
U.S.A 1985, 82, 5000.
(14) McCammon, J. A.; Gelin, B. R.; Karplus, M.; Wolynes, P. G. Nature
1976, 262, 325.
(15) Frauenfelder, H.; McMahon, B. H.; Austin, R. H.; Chu, K.; Groves,
J. T. Proc. Natl. Acad. Sci. U.S.A 2001, 98, 2370.
(16) Feynman, R. The Feynman lectures on physics; Addison-Wesley
Publishing Company: Reading, MA, 1963.
(17) Wittenberg, J. B.; Wittenberg, B. A. Annu. ReV. Biophys. Biophys.
Chem. 1990, 19, 217.
(18) Shrivastava, I. H.; Sansom, M. S. Biophys. J. 2000, 78, 557.
(19) Berneche, S.; Roux, B. Nature 2001, 414, 73.
(20) Roux, B. Acc. Chem. Res. 2002, 35, 366.
(21) Bahar, I.; Chennubhotla, C.; Tobi, D. Curr. Opin. Struct. Biol. 2007,
17, 633.
(22) Eisenmesser, E. Z.; Millet, O.; Labeikovsky, W.; Korzhnev, D. M.;
Wolf-Watz, M.; Bosco, D. A.; Skalicky, J. J.; Kay, L. E.; Kern, D.
Nature 2005, 438, 117.
(23) Henzler-Wildman, K.; Kern, D. Nature 2007, 450, 964.
(24) Henzler-Wildman, K. A.; Thai, V.; Lei, M.; Ott, M.; Wolf-Watz,
M.; Fenn, T.; Pozharski, E.; Wilson, M. A.; Petsko, G. A.; Karplus,
M.; Hubner, C. G.; Kern, D. Nature 2007, 450, 838.
(25) Kern, D.; Zuiderweg, E. R. Curr. Opin. Struct. Biol. 2003, 13, 748.
(26) Lange, O. F.; Lakomek, N. A.; Fares, C.; Schroder, G. F.; Walter,
K. F.; Becker, S.; Meiler, J.; Grubmuller, H.; Griesinger, C.; de Groot,
B. L. Science 2008, 320, 1471.
(27) Lena, C.; Changeux, J. P. Trends Neurosci. 1993, 16, 181.
(28) Ming, D.; Wall, M. E. Proteins 2005, 59, 697.
(29) Taly, A.; Corringer, P. J.; Grutter, T.; Prado de, C. L.; Karplus, M.;
Changeux, J. P. Proc. Natl. Acad. Sci. U.S.A 2006, 103, 16965.
(30) Tobi, D.; Bahar, I. Proc. Natl. Acad. Sci. U.S.A 2005, 102, 18908.
(31) Xu, C.; Tobi, D.; Bahar, I. J. Mol. Biol. 2003, 333, 153.
(32) Yang, L. W.; Bahar, I. Structure 2005, 13, 893.
(33) Tajkhorshid, E.; Nollert, P.; Jensen, M. O.; Miercke, L. J.; O’Connell,
J.; Stroud, R. M.; Schulten, K. Science 2002, 296, 525.
(34) Noskov, S. Y.; Berneche, S.; Roux, B. Nature 2004, 431, 830.
(35) Humprey, W.; Dalket, A.; Schulten, K. J. Mol. Graphics 1996, 14,
33.
(36) Jolliffe, I. T. Principal Component Analysis; Springer: New York,
2002.
(37) Kitao, A.; Go, N. Curr. Opin. Struct. Biol. 1999, 9, 164.
(38) Bahar, I.; Rader, A. J. Curr. Opin. Struct. Biol. 2005, 15, 586.
(39) Cui, Q.; Bahar, I. E. Normal Mode Analysis. Theory and Applications
to Biological and Chemical Systems; Taylor & Francis Group: Boca
Raton, FL, 2006.
(40) Amadei, A.; Linssen, A. B.; Berendsen, H. J. Proteins 1993, 17, 412.
(41) Bahar, I.; Erman, B.; Haliloglu, T.; Jernigan, R. L. Biochemistry 1997,
36, 13512.
(42) Garcia, A. E.; Harman, J. G. Protein Sci. 1996, 5, 62.
(43) Romo, T. D.; Clarage, J. B.; Sorensen, D. C.; Phillips, G. N., Jr.
Proteins 1995, 22, 311.
(44) Yang, L. W.; Eyal, E.; Bahar, I.; Kitao, A. Bioinformatics 2009, 25,
606.
(45) Chennubhotla, C.; Yang, Z.; Bahar, I. Mol. BioSyst. 2008, 4, 287.
(46) Henzler-Wildman, K. A.; Lei, M.; Thai, V.; Kerns, S. J.; Karplus,
M.; Kern, D. Nature 2007, 450, 913.
(47) Swain, J. F.; Gierasch, L. M. Curr. Opin. Struct. Biol. 2006, 16,
102.
1494 Chemical Reviews, 2010, Vol. 110, No. 3
(48) Hawkins, R. J.; McLeish, T. C. Biophys. J. 2006, 91, 2055.
(49) Popovych, N.; Sun, S.; Ebright, R. H.; Kalodimos, C. G. Nat. Struct.
Mol. Biol. 2006, 13, 831.
(50) Brath, U.; Akke, M.; Yang, D.; Kay, L. E.; Mulder, F. A. J. Am.
Chem. Soc. 2006, 128, 5718.
(51) Velyvis, A.; Yang, Y. R.; Schachman, H. K.; Kay, L. E. Proc. Natl.
Acad. Sci. U.S.A. 2007, 104, 8815.
(52) Changeux, J. P.; Edelstein, S. J. Science 2005, 308, 1424.
(53) Monod, J.; Wyman, J.; Changeux, J. P. J. Mol. Biol. 1965, 12, 88.
(54) Go, N.; Noguti, T.; Nishikawa, T. Proc. Natl. Acad. Sci. U.S.A. 1983,
80, 3696.
(55) Brooks, B.; Karplus, M. Proc. Natl. Acad. Sci. U.S.A 1983, 80, 6571.
(56) Levitt, M.; Sander, C.; Stern, P. S. J. Mol. Biol. 1985, 181, 423.
(57) Naik, V. M.; Krimm, S.; Denton, J. B.; Nemethy, G.; Scheraga, H. A.
Int. J. Pept. Protein Res. 1984, 24, 613.
(58) Tirion, M. M. Phys. ReV. Lett. 1996, 77, 1905.
(59) Taylor, W. R.; Aszodi, A. Protein Geometry, Classification, Topology
and Symmetry: A Computational Analysis of Structure; Taylor &
Francis Group: New York, 2005.
(60) Lezon, T. R.; Banavar, J. R.; Lesk, A. M.; Maritan, A. Proteins 2006,
63, 273.
(61) Doruker, P.; Jernigan, R. L.; Bahar, I. J. Comput. Chem. 2002, 23,
119.
(62) Lu, M.; Ma, J. Biophys. J. 2005, 89, 2395.
(63) Nicolay, S.; Sanejouand, Y. H. Phys. ReV. Lett. 2006, 96, 078104.
(64) Tama, F.; Brooks, C. L. Annu. ReV. Biophys. Biomol. Struct. 2006,
35, 115.
(65) Zheng, W.; Brooks, B. R.; Thirumalai, D. Proc. Natl. Acad. Sci. U.S.A
2006, 103, 7664.
(66) Ma, J. Structure 2005, 13, 373.
(67) Bahar, I.; Atilgan, A. R.; Erman, B. Fold. Des. 1997, 2, 173.
(68) Haliloglu, T.; Bahar, I.; Erman, B. Phys. ReV. Lett. 1997, 79, 3090.
(69) Eichinger, B. E. Macromolecules 1972, 5, 496.
(70) Flory, P. J. Proc. R. Soc. London, A 1976, 351, 351.
(71) Pearson, D. S. Macromolecules 1977, 10, 696.
(72) Kloczkowski, A.; Mark, J. E.; Erman, B. Macromolecules 1998, 22,
1423.
(73) Erman, B.; Mark, J. E. Structure and Properties of Rubberlike
Networks; Oxford University Press: New York, 1997.
(74) Hinsen, K. Proteins 1998, 33, 417.
(75) Hinsen, K.; Thomas, A.; Field, M. J. Proteins 1999, 34, 369.
(76) Atilgan, A. R.; Durell, S. R.; Jernigan, R. L.; Demirel, M. C.; Keskin,
O.; Bahar, I. Biophys. J. 2001, 80, 505.
(77) Doruker, P.; Atilgan, A. R.; Bahar, I. Proteins 2000, 40, 512.
(78) Eyal, E.; Yang, L. W.; Bahar, I. Bioinformatics 2006, 22, 2619.
(79) Tama, F.; Sanejouand, Y. H. Protein Eng. 2001, 14, 1.
(80) Mouawad, L.; Perahia, D. J. Mol. Biol. 1996, 258, 393.
(81) Tama, F.; Valle, M.; Frank, J.; Brooks, C. L., III. Proc. Natl. Acad.
Sci. U.S.A. 2003, 100, 9319.
(82) Wang, Y.; Rader, A. J.; Bahar, I.; Jernigan, R. L. J. Struct. Biol.
2004, 147, 302.
(83) Essiz, S. G.; Coalson, R. D. J. Chem. Phys. 2007, 127, 104109.
(84) Haddadian, E. J.; Cheng, M. H.; Coalson, R. D.; Xu, Y.; Tang, P. J.
Phys. Chem. B 2008, 112, 13981.
(85) Isin, B.; Schulten, K.; Tajkhorshid, E.; Bahar, I. Biophys. J. 2008,
95, 789.
(86) Miller, B. T.; Zheng, W.; Venable, R. M.; Pastor, R. W.; Brooks,
B. R. J. Phys. Chem. B 2008, 112, 6274.
(87) Moritsugu, K.; Smith, J. C. Biophys. J. 2007, 93, 3460.
(88) Sherwood, P.; Brooks, B. R.; Sansom, M. S. Curr. Opin. Struct. Biol.
2008, 18, 630.
(89) Hille, B. Ion channels of excitable membranes; Sinauer Associates:
Sunderland, MA, 2001.
(90) Staros, J. V. Acc. Chem. Res. 1988, 21, 435.
(91) Gouaux, E.; MacKinnon, R. Science 2005, 310, 1461.
(92) Dobson, P. D.; Kell, D. B. Nat. ReV. Drug DiscoV. 2008, 7, 205.
(93) Stahlberg, H.; Fotiadis, D.; Scheuring, S.; Remigy, H.; Braun, T.;
Mitsuoka, K.; Fujiyoshi, Y.; Engel, A. FEBS Lett. 2001, 504, 166.
(94) Wishart, D. S.; Knox, C.; Guo, A. C.; Shrivastava, S.; Hassanali,
M.; Stothard, P.; Chang, Z.; Woolsey, J. Nucleic Acids Res. 2006,
34, D668–D672.
(95) Chen, X.; Ji, Z. L.; Chen, Y. Z. Nucleic Acids Res. 2002, 30, 412.
(96) Gunther, S.; Kuhn, M.; Dunkel, M.; Campillos, M.; Senger, C.;
Petsalaki, E.; Ahmed, J.; Urdiales, E. G.; Gewiess, A.; Jensen, L. J.;
Schneider, R.; Skoblo, R.; Russell, R. B.; Bourne, P. E.; Bork, P.;
Preissner, R. Nucleic Acids Res. 2008, 36, D919–D922.
(97) Imming, P.; Sinning, C.; Meyer, A. Nat. ReV. Drug DiscoV. 2006, 5,
821.
(98) Thomas, P. D.; Campbell, M. J.; Kejariwal, A.; Mi, H.; Karlak, B.;
Daverman, R.; Diemer, K.; Muruganujan, A.; Narechania, A. Genome
Res. 2003, 13, 2129.
(99) Overington, J. P.; Al-Lazikani, B.; Hopkins, A. L. Nat. ReV. Drug
DiscoV. 2006, 5, 993.
Bahar et al.
(100) Zheng, C. J.; Han, L. Y.; Yap, C. W.; Ji, Z. L.; Cao, Z. W.; Chen,
Y. Z. Pharmacol. ReV. 2006, 58, 259.
(101) Filmore, D. Modern Drug DiscoVery 2004, 17, 24.
(102) Congreve, M.; Murray, C. W.; Blundell, T. L. Drug DiscoVery Today
2005, 10, 895.
(103) Ostermeier, C.; Michel, H. Curr. Opin. Struct. Biol. 1997, 7, 697.
(104) White, S. H. Protein Sci. 2004, 13, 1948.
(105) Blow, N. Nat. Methods 2008, 5, 203.
(106) Service, R. F. Science 2008, 319, 1612.
(107) Carpenter, E. P.; Beis, K.; Cameron, A. D.; Iwata, S. Curr. Opin.
Struct. Biol. 2008, 18, 581.
(108) Arora, A.; Tamm, L. K. Curr. Opin. Struct. Biol. 2001, 11, 540.
(109) Marassi, F. M.; Opella, S. J. Curr. Opin. Struct. Biol. 1998, 8, 640.
(110) Creuzet, F.; McDermott, A.; Gebhard, R.; van der Hoef, K.; Spijker-
Assink, M. B.; Herzfeld, J.; Lugtenburg, J.; Levitt, M. H.; Griffin,
R. G. Science 1991, 251, 783.
(111) Ma, D.; Tillman, T. S.; Tang, P.; Meirovitch, E.; Eckenhoff, R.;
Carnini, A.; Xu, Y. Proc. Natl. Acad. Sci. U.S.A 2008, 105, 16537.
(112) Marassi, F. M.; Ramamoorthy, A.; Opella, S. J. Proc. Natl. Acad.
Sci. U.S.A 1997, 94, 8551.
(113) Marassi, F. M.; Ma, C.; Gratkowski, H.; Straus, S. K.; Strebel, K.;
Oblatt-Montal, M.; Montal, M.; Opella, S. J. Proc. Natl. Acad. Sci.
U.S.A 1999, 96, 14336.
(114) Marassi, F. M.; Opella, S. J. J. Magn. Reson. 2000, 144, 150.
(115) Marassi, F. M.; Opella, S. J. Protein Sci. 2003, 12, 403.
(116) Opella, S. J.; Nevzorov, A.; Mesleb, M. F.; Marassi, F. M. Biochem.
Cell Biol. 2002, 80, 597.
(117) Chang, G.; Spencer, R. H.; Lee, A. T.; Barclay, M. T.; Rees, D. C.
Science 1998, 282, 2220.
(118) Doyle, D. A.; Morais, C. J.; Pfuetzner, R. A.; Kuo, A.; Gulbis, J. M.;
Cohen, S. L.; Chait, B. T.; MacKinnon, R. Science 1998, 280, 69.
(119) Jiang, Y.; Lee, A.; Chen, J.; Cadene, M.; Chait, B. T.; MacKinnon,
R. Nature 2002, 417, 523.
(120) Jiang, Y.; Lee, A.; Chen, J.; Cadene, M.; Chait, B. T.; MacKinnon,
R. Nature 2002, 417, 515.
(121) Yeh, J. I.; Biemann, H. P.; Prive, G. G.; Pandit, J.; Koshland, D. E.,
Jr.; Kim, S. H. J. Mol. Biol. 1996, 262, 186.
(122) Yeh, J. I. Acta Crystallogr., D: Biol. Crystallogr. 2003, 59, 1408.
(123) Arrondo, J. L.; Goni, F. M. Prog. Biophys. Mol. Biol. 1999, 72, 367.
(124) Tusnady, G. E.; Dosztanyi, Z.; Simon, I. Nucleic Acids Res. 2005,
33, D275–D278.
(125) Raman, P.; Cherezov, V.; Caffrey, M. Cell. Mol. Life Sci. 2006, 63,
36.
(126) Sutherland, G. B. AdV. Protein Chem. 1952, 7, 291.
(127) Fanucci, G. E.; Cafiso, D. S. Curr. Opin. Struct. Biol. 2006, 16, 644.
(128) Hubbell, W. L.; Cafiso, D. S.; Altenbach, C. Nat. Struct. Biol. 2000,
7, 735.
(129) Hubbell, W. L.; Altenbach, C. Membrane Protein Structure: Experi-
mental Approaches; Oxford University Press: New York, 1994; pp
224-248.
(130) Lakomek, N. A.; Walter, K. F.; Fares, C.; Lange, O. F.; de Groot,
B. L.; Grubmuller, H.; Bruschweiler, R.; Munk, A.; Becker, S.;
Meiler, J.; Griesinger, C. J. Biomol. NMR 2008, 41, 139.
(131) Mittermaier, A.; Kay, L. E. Science 2006, 312, 224.
(132) Sprangers, R.; Velyvis, A.; Kay, L. E. Nat. Methods 2007, 4, 697.
(133) Tugarinov, V.; Sprangers, R.; Kay, L. E. J. Am. Chem. Soc. 2007,
129, 1743.
(134) Hubbell, W. L.; Altenbach, C.; Hubbell, C. M.; Khorana, H. G. AdV.
Protein Chem. 2003, 63, 243.
(135) Ishima, R.; Torchia, D. A. Nat. Struct. Biol. 2000, 7, 740.
(136) Kay, L. E. J. Magn. Reson. 2005, 173, 193.
(137) Labeikovsky, W.; Eisenmesser, E. Z.; Bosco, D. A.; Kern, D. J. Mol.
Biol. 2007, 367, 1370.
(138) Akke, M.; Skelton, N. J.; Kordel, J.; Palmer, A. G., III; Chazin, W. J.
Biochemistry 1993, 32, 9832.
(139) Akke, M.; Fiala, R.; Jiang, F.; Patel, D.; Palmer, A. G., III. RNA
1997, 3, 702.
(140) Akke, M.; Liu, J.; Cavanagh, J.; Erickson, H. P.; Palmer, A. G., III.
Nat. Struct. Biol. 1998, 5, 55.
(141) Tugarinov, V.; Kay, L. E. Biochemistry 2005, 44, 15970.
(142) Tugarinov, V.; Kay, L. E. J. Am. Chem. Soc. 2006, 128, 12484.
(143) Sprangers, R.; Kay, L. E. Nature 2007, 445, 618.
(144) Thomas, A.; Field, M. J.; Mouawad, L.; Perahia, D. J. Mol. Biol.
1996, 257, 1070.
(145) Velyvis, A.; Yang, Y. R.; Schachman, H. K.; Kay, L. E. Proc. Natl.
Acad. Sci. U.S.A. 2007, 104, 8815.
(146) Tolman, J. R.; Flanagan, J. M.; Kennedy, M. A.; Prestegard, J. H.
Nat. Struct. Biol. 1997, 4, 292.
(147) Hubbell, W. L.; Mchaourab, H. S.; Altenbach, C.; Lietzow, M. A.
Structure 1996, 4, 779.
(148) Altenbach, C.; Marti, T.; Khorana, H. G.; Hubbell, W. L. Science
1990, 248, 1088.
Functional Mechanisms of Membrane Proteins
(149) Steinhoff, H. J.; Mollaaghababa, R.; Altenbach, C.; Hideg, K.; Krebs,
M.; Khorana, H. G.; Hubbell, W. L. Science 1994, 266, 105.
(150) Altenbach, C.; Yang, K.; Farrens, D. L.; Farahbakhsh, Z. T.; Khorana,
H. G.; Hubbell, W. L. Biochemistry 1996, 35, 12470.
(151) Altenbach, C.; Klein-Seetharaman, J.; Cai, K.; Khorana, H. G.;
Hubbell, W. L. Biochemistry 2001, 40, 15493.
(152) Altenbach, C.; Cai, K.; Klein-Seetharaman, J.; Khorana, H. G.;
Hubbell, W. L. Biochemistry 2001, 40, 15483.
(153) Lewis, J. W.; van Kuijk, F. J.; Thorgeirsson, T. E.; Kliger, D. S.
Biochemistry 1991, 30, 11372.
(154) Resek, J. F.; Farahbakhsh, Z. T.; Hubbell, W. L.; Khorana, H. G.
Biochemistry 1993, 32, 12025.
(155) Steinhoff, H. J.; Schwemer, J. J. Photochem. Photobiol. B 1996, 35,
1.
(156) Liu, Y. S.; Sompornpisut, P.; Perozo, E. Nat. Struct. Biol. 2001, 8,
883.
(157) Perozo, E.; Cortes, D. M.; Cuello, L. G. Science 1999, 285, 73.
(158) Perozo, E.; Cuello, L. G.; Cortes, D. M.; Liu, Y. S.; Sompornpisut,
P. NoVartis Found. Symp. 2002, 245, 146.
(159) Sompornpisut, P.; Liu, Y. S.; Perozo, E. Biophys. J. 2001, 81, 2530.
(160) Perozo, E.; Cortes, D. M.; Sompornpisut, P.; Kloda, A.; Martinac,
B. Nature 2002, 418, 942.
(161) Altenbach, C.; Kusnetzow, A. K.; Ernst, O. P.; Hofmann, K. P.;
Hubbell, W. L. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 7439.
(162) Doniach, S.; Eastman, P. Curr. Opin. Struct. Biol. 1999, 9, 157.
(163) Khalili-Araghi, F.; Gumbart, J.; Wen, P. C.; Sotomayor, M.;
Tajkhorshid, E.; Schulten, K. Curr. Opin. Struct. Biol. 2009, 19, 128.
(164) Tieleman, D. P.; Biggin, P. C.; Smith, G. R.; Sansom, M. S. Q. ReV.
Biophys. 2001, 34, 473.
(165) Ash, W. L.; Zlomislic, M. R.; Oloo, E. O.; Tieleman, D. P. Biochim.
Biophys. Acta 2004, 1666, 158.
(166) Tieleman, D. P. Clin. Exp. Pharmacol. Physiol. 2006, 33, 893.
(167) Stone, J. E.; Phillips, J. C.; Freddolino, P. L.; Hardy, D. J.; Trabuco,
L. G.; Schulten, K. J. Comput. Chem. 2007, 28, 2618.
(168) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.;
Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. J. Comput.
Chem. 2005, 26, 1781.
(169) Hub, J. S.; de Groot, B. L. Proc. Natl. Acad. Sci. U.S.A. 2008, 105,
1198.
(170) Wang, Y.; Schulten, K.; Tajkhorshid, E. Structure 2005, 13, 1107.
(171) Tornroth-Horsefield, S.; Wang, Y.; Hedfalk, K.; Johanson, U.;
Karlsson, M.; Tajkhorshid, E.; Neutze, R.; Kjellbom, P. Nature 2006,
439, 688.
(172) de Groot, B. L.; Grubmuller, H. Science 2001, 294, 2353.
(173) Allen, T. W.; Bastug, T.; Kuyucak, S.; Chung, S. H. Biophys. J.
2003, 84, 2159.
(174) Allen, T. W.; Andersen, O. S.; Roux, B. Biophys. J. 2006, 90, 3447.
(175) Bastug, T.; Patra, S. M.; Kuyucak, S. Chem. Phys. Lipids 2006, 141,
197.
(176) Roux, B.; MacKinnon, R. Science 1999, 285, 100.
(177) Shrivastava, I. H.; Sansom, M. S. Eur. Biophys. J. 2002, 31, 207.
(178) Shrivastava, I. H.; Tieleman, D. P.; Biggin, P. C.; Sansom, M. S.
Biophys. J. 2002, 83, 633.
(179) Allen, T. W.; Kuyucak, S.; Chung, S. H. Biophys. J. 1999, 77, 2502.
(180) Grossfield, A.; Feller, S. E.; Pitman, M. C. Proteins 2007, 67, 31.
(181) Roux, B.; Allen, T.; Berneche, S.; Im, W. Q. ReV. Biophys. 2004,
37, 15.
(182) Goldstein, H. Classical Mechanics; Addison-Wesley: Cambridge,
1953.
(183) Song, G.; Jernigan, R. L. Proteins 2006, 63, 197.
(184) Yang, L.; Song, G.; Jernigan, R. L. Biophys. J. 2007, 93, 920.
(185) Chu, J. W.; Voth, G. A. Biophys. J. 2006, 90, 1572.
(186) Jang, Y.; Jeong, J. I.; Kim, M. K. Nucleic Acids Res. 2006, 34, W57–
W62.
(187) Kim, M. K.; Li, W.; Shapiro, B. A.; Chirikjian, G. S. J. Biomol.
Struct. Dyn. 2003, 21, 395.
(188) Liu, X.; Xu, Y.; Li, H.; Wang, X.; Jiang, H.; Barrantes, F. J. PLoS
Comput. Biol. 2008, 4, e19.
(189) Miloshevsky, G. V.; Jordan, P. C. Structure 2007, 15, 1654.
(190) Yang, L.; Song, G.; Carriquiry, A.; Jernigan, R. L. Structure 2008,
16, 321.
(191) Nichols, J.; Taylor, H.; Schmidt, P.; Simons, J. J. Chem. Phys. 1990,
92, 340.
(192) Miloshevsky, G. V.; Jordan, P. C. Structure 2006, 14, 1241.
(193) Lu, M.; Ma, J. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 15358.
(194) Kundu, S.; Melton, J. S.; Sorensen, D. C.; Phillips, G. N., Jr. Biophys.
J. 2002, 83, 723.
(195) Sen, T. Z.; Jernigan, R. L. Normal Mode Analysis: Theory and
Applications to Biological and Chemical Systems; Chapman & Hall/
CRC: Boca Raton, FL, 2006; pp 171-186.
(196) Kondrashov, D. A.; Cui, Q.; Phillips, G. N., Jr. Biophys. J. 2006,
91, 2760.
Chemical Reviews, 2010, Vol. 110, No. 3 1495
(197) Hamacher, K.; McCammon, J. A. J. Chem. Theory Comput. 2006,
2, 873.
(198) Lyman, E.; Pfaendtner, J.; Voth, G. A. Biophys. J. 2008, 95, 4183.
(199) Chennubhotla, C.; Bahar, I. Mol. Syst. Biol. 2006, 2, 36.
(200) Franklin, J.; Koehl, P.; Doniach, S.; Delarue, M. Nucleic Acids Res.
2007, 35, W477–W482.
(201) Lindahl, E.; Azuara, C.; Koehl, P.; Delarue, M. Nucleic Acids Res.
2006, 34, W52–W56.
(202) Suhre, K.; Sanejouand, Y. H. Nucleic Acids Res. 2004, 32, W610–
W614.
(203) Wako, H.; Kato, M.; Endo, S. Bioinformatics 2004, 20, 2035.
(204) Yang, L. W.; Liu, X.; Jursa, C. J.; Holliman, M.; Rader, A.; Karimi,
H. A.; Bahar, I. Bioinformatics 2005, 21, 2978.
(205) Yang, L. W.; Rader, A. J.; Liu, X.; Jursa, C. J.; Chen, S. C.; Karimi,
H. A.; Bahar, I. Nucleic Acids Res. 2006, 34, W24–W31.
(206) Bahar, I.; Atilgan, A. R.; Demirel, M. C.; Erman, B. Phys. ReV. Lett.
1998, 80, 2733.
(207) Chennubhotla, C.; Rader, A. J.; Yang, L. W.; Bahar, I. Phys. Biol.
2005, 2, S173–S180.
(208) Rader, A. J.; Chennubhotla, C.; Yang, L. W.; Bahar, I. Normal Mode
Analysis. Theory and Applications to Biological and Chemical
Systems; Taylor & Francis Group: New York, 2006; Chapter 3,
pp41-64.
(209) Yang, L. W.; Eyal, E.; Chennubhotla, C.; Jee, J.; Gronenborn, A. M.;
Bahar, I. Structure 2007, 15, 741.
(210) Bahar, I.; Wallqvist, A.; Covell, D. G.; Jernigan, R. L. Biochemistry
1998, 37, 1067.
(211) Haliloglu, T.; Bahar, I. Proteins 1999, 37, 654.
(212) Temiz, N. A.; Meirovitch, E.; Bahar, I. Proteins 2004, 57, 468.
(213) Zhuravleva, A.; Korzhnev, D. M.; Nolde, S. B.; Kay, L. E.; Arseniev,
A. S.; Billeter, M.; Orekhov, V. Y. J. Mol. Biol. 2007, 367, 1079.
(214) Su, J. G.; Li, C. H.; Hao, R.; Chen, W. Z.; Wang, C. X. Biophys. J.
2008, 94, 4586.
(215) Rader, A. J.; Bahar, I. Polymer 2004, 45, 659.
(216) Rader, A. J.; Anderson, G.; Isin, B.; Khorana, H. G.; Bahar, I.; Klein-
Seetharaman, J. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 7246.
(217) Maguid, S.; Fernandez-Alberti, S.; Ferrelli, L.; Echave, J. Biophys.
J. 2005, 89, 3.
(218) Shrivastava, I. H.; Bahar, I. Biophys. J. 2006, 90, 3929.
(219) Marques, O.; Sanejouand, Y. H. Proteins 1995, 23, 557.
(220) Bruschweiler, R. J. Chem. Phys. 1995, 102, 3396.
(221) Ayton, G. S.; Noid, W. G.; Voth, G. A. Curr. Opin. Struct. Biol.
2007, 17, 192.
(222) Kurkcuoglu, O.; Jernigan, R. L.; Doruker, P. Biochemistry 2006, 45,
1173.
(223) Noid, W. G.; Chu, J. W.; Ayton, G. S.; Krishna, V.; Izvekov, S.;
Voth, G. A.; Das, A.; Andersen, H. C. J. Chem. Phys. 2008, 128,
244114.
(224) van Vlijmen, H. W.; Karplus, M. J. Chem. Phys. 2001, 115, 698.
(225) van Vlijmen, H. W.; Karplus, M. J. Mol. Biol. 2005, 350, 528.
(226) Yang, Z.; Bahar, I.; Widom, M. Biophys. J. 2009, 96, 4438.
(227) Kim, M. K.; Jernigan, R. L.; Chirikjian, G. S. J. Struct. Biol. 2003,
143, 107.
(228) Tama, F.; Gadea, F. X.; Marques, O.; Sanejouand, Y. H. Proteins
2000, 41, 1.
(229) Li, G.; Cui, Q. Biophys. J. 2002, 83, 2457.
(230) Durand, P.; Trinquier, G.; Sanejouand, Y. H. Biopolymers 1994, 34,
759.
(231) Li, G.; Cui, Q. Biophys. J. 2004, 86, 743.
(232) Tama, F.; Brooks, C. L., III. J. Mol. Biol. 2002, 318, 733.
(233) Tama, F.; Brooks, C. L., III. J. Mol. Biol. 2005, 345, 299.
(234) Hinsen, K.; Kneller, G. R. Proteins 2008, 70, 1235.
(235) Miller, B. T.; Zheng, W.; Venable, R. M.; Pastor, R. W.; Brooks,
B. R. J. Phys. Chem. B 2008, 112, 6274.
(236) Moritsugu, K.; Smith, J. C. J. Phys. Chem. B 2005, 109, 12182.
(237) Zheng, W.; Brooks, B. R. Biophys. J. 2005, 89, 167.
(238) Ming, D.; Wall, M. E. Phys. ReV. Lett. 2005, 95, 198103.
(239) Ming, D.; Wall, M. E. J. Mol. Biol. 2006, 358, 213.
(240) Sonne, J.; Kandt, C.; Peters, G. H.; Hansen, F. Y.; Jensen, M. O.;
Tieleman, D. P. Biophys. J. 2007, 92, 2727.
(241) Lamm, G.; Szabo, A. J. Chem. Phys. 1986, 85, 7334.
(242) Moritsugu, K.; Smith, J. C. J. Phys. Chem. B 2006, 110, 5807.
(243) Kneller, G. R. Chem. Phys. 2000, 261, 1.
(244) Hinsen, K.; Petrescu, A. J.; Dellerue, S.; Bellissent-Funel, M. C.;
Kneller, G. R. J. Chem. Phys. 2000, 261, 25.
(245) Roux, B.; Karplus, M. Biophys. J. 1988, 53, 297.
(246) Lu, H. P. Acc. Chem. Res. 2005, 38, 557.
(247) Harms, G. S.; Orr, G.; Montal, M.; Thrall, B. D.; Colson, S. D.; Lu,
H. P. Biophys. J. 2003, 85, 1826.
(248) Thompson, J. D.; Higgins, D. G.; Gibson, T. J. Nucleic Acids Res.
1994, 22, 4673.
(249) Shen, Y.; Kong, Y.; Ma, J. Proc. Natl. Acad. Sci. U.S.A. 2002, 99,
1949.
1496 Chemical Reviews, 2010, Vol. 110, No. 3
(250) Shimizu, H.; Iwamoto, M.; Konno, T.; Nihei, A.; Sasaki, Y. C.; Oiki,
S. Cell 2008, 132, 67.
(251) Alam, A.; Jiang, Y. Nat. Struct. Mol. Biol. 2009, 16, 35.
(252) Shi, N.; Ye, S.; Alam, A.; Chen, L.; Jiang, Y. Nature 2006, 440,
570.
(253) Haliloglu, T.; Ben-Tal, N. PLoS Comput. Biol. 2008, 4, e1000164.
(254) Yifrach, O.; MacKinnon, R. Cell 2002, 111, 231.
(255) Jan, L. Y.; Jan, Y. N. Nature 1994, 371, 119.
(256) Anishkin, A.; Chiang, C. S.; Sukharev, S. J. Gen. Physiol. 2005,
125, 155.
(257) Hamill, O. P.; Martinac, B. Physiol. ReV. 2001, 81, 685.
(258) Sukharev, S.; Durell, S. R.; Guy, H. R. Biophys. J. 2001, 81, 917.
(259) Poolman, B.; Blount, P.; Folgering, J. H.; Friesen, R. H.; Moe, P. C.;
van der Heide, T. Mol. Microbiol. 2002, 44, 889.
(260) Cruickshank, C. C.; Minchin, R. F.; Le Dain, A. C.; Martinac, B.
Biophys. J. 1997, 73, 1925.
(261) Valadie, H.; Lacapere, J. J.; Sanejouand, Y. H.; Etchebest, C. J. Mol.
Biol. 2003, 332, 657.
(262) Czajkowski, C. Nature 2005, 438, 167.
(263) Sine, S. M.; Engel, A. G. Nature 2006, 440, 448.
(264) Miyazawa, A.; Fujiyoshi, Y.; Unwin, N. Nature 2003, 423, 949.
(265) Unwin, N. J. Mol. Biol. 2005, 346, 967.
(266) Brejc, K.; van Dijk, W. J.; Klaassen, R. V.; Schuurmans, M.; van
der Oost, J.; Smit, A. B.; Sixma, T. K. Nature 2001, 411, 269.
(267) Changeux, J. P.; Edelstein, S. J. Neuron 1998, 21, 959.
(268) Hung, A.; Tai, K.; Sansom, M. S. Biophys. J. 2005, 88, 3321.
(269) Liu, X.; Xu, Y.; Li, H.; Wang, X.; Jiang, H.; Barrantes, F. J. PLoS
Comput. Biol. 2008, 4, e19.
(270) Szarecka, A.; Xu, Y.; Tang, P. Proteins 2007, 68, 948.
(271) Taly, A.; Delarue, M.; Grutter, T.; Nilges, M.; Le, N. N.; Corringer,
P. J.; Changeux, J. P. Biophys. J. 2005, 88, 3954.
(272) Cheng, X.; Lu, B.; Grant, B.; Law, R. J.; McCammon, J. A. J. Mol.
Biol. 2006, 355, 310.
(273) Smart, O. S.; Neduvelil, J. G.; Wang, X.; Wallace, B. A.; Sansom,
M. S. J. Mol. Graph. 1996, 14, 354.
(274) Corry, B. Biophys. J. 2006, 90, 799.
(275) Unwin, N. Nature 1995, 373, 37.
(276) Cymes, G. D.; Ni, Y.; Grosman, C. Nature 2005, 438, 975.
(277) Cymes, G. D.; Grosman, C. Nat. Struct. Mol. Biol. 2008, 15, 389.
(278) Bouzat, C.; Gumilar, F.; Spitzmaul, G.; Wang, H. L.; Rayes, D.;
Hansen, S. B.; Taylor, P.; Sine, S. M. Nature 2004, 430, 896.
(279) Lee, W. Y.; Sine, S. M. Nature 2005, 438, 243.
(280) Auerbach, A. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 1408.
(281) Purohit, P.; Mitra, A.; Auerbach, A. Nature 2007, 446, 930.
(282) Mukhtasimova, N.; Lee, W. Y.; Wang, H. L.; Sine, S. M. Nature
2009, 459, 451.
(283) Hilf, R. J.; Dutzler, R. Nature 2008, 452, 375.
(284) Bocquet, N.; Nury, H.; Baaden, M.; Le, P. C.; Changeux, J. P.;
Delarue, M.; Corringer, P. J. Nature 2009, 457, 111.
(285) Hilf, R. J.; Dutzler, R. Nature 2009, 457, 115.
(286) Fredriksson, R.; Lagerstrom, M. C.; Lundin, L. G.; Schioth, H. B.
Mol. Pharmacol. 2003, 63, 1256.
(287) Kobilka, B. K. Biochim. Biophys. Acta 2007, 1768, 794.
(288) Oldham, W. M.; Hamm, H. E. Nat. ReV. Mol. Cell Biol. 2008, 9, 60.
(289) Palczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.; Motoshima,
H.; Fox, B. A.; Le, T. I.; Teller, D. C.; Okada, T.; Stenkamp, R. E.;
Yamamoto, M.; Miyano, M. Science 2000, 289, 739.
(290) Hanson, M. A.; Stevens, R. C. Structure 2009, 17, 8.
(291) Scheerer, P.; Park, J. H.; Hildebrand, P. W.; Kim, Y. J.; Krauss, N.;
Choe, H. W.; Hofmann, K. P.; Ernst, O. P. Nature 2008, 455, 497.
(292) Park, J. H.; Scheerer, P.; Hofmann, K. P.; Choe, H. W.; Ernst, O. P.
Nature 2008, 454, 183.
(293) Jaakola, V. P.; Griffith, M. T.; Hanson, M. A.; Cherezov, V.; Chien,
E. Y.; Lane, J. R.; Ijzerman, A. P.; Stevens, R. C. Science 2008,
322, 1211.
(294) Warne, T.; Serrano-Vega, M. J.; Baker, J. G.; Moukhametzianov,
R.; Edwards, P. C.; Henderson, R.; Leslie, A. G.; Tate, C. G.;
Schertler, G. F. Nature 2008, 454, 486.
(295) Murakami, M.; Kouyama, T. Nature 2008, 453, 363.
(296) Kusnetzow, A. K.; Altenbach, C.; Hubbell, W. L. Biochemistry 2006,
45, 5538.
(297) Farrens, D. L.; Altenbach, C.; Yang, K.; Hubbell, W. L.; Khorana,
H. G. Science 1996, 274, 768.
(298) Farahbakhsh, Z. T.; Altenbach, C.; Hubbell, W. L. Photochem.
Photobiol. 1992, 56, 1019.
(299) Bhattacharya, S.; Hall, S. E.; Vaidehi, N. J. Mol. Biol. 2008, 382,
539.
(300) Isin, B.; Rader, A. J.; Dhiman, H. K.; Klein-Seetharaman, J.; Bahar,
I. Proteins 2006, 65, 970.
(301) Shacham, S.; Marantz, Y.; Bar-Haim, S.; Kalid, O.; Warshaviak, D.;
Avisar, N.; Inbal, B.; Heifetz, A.; Fichman, M.; Topf, M.; Naor, Z.;
Noiman, S.; Becker, O. M. Proteins 2004, 57, 51.
(302) Lemaitre, V.; Yeagle, P.; Watts, A. Biochemistry 2005, 44, 12667.
Bahar et al.
(303) Huber, T.; Botelho, A. V.; Beyer, K.; Brown, M. F. Biophys. J. 2004,
86, 2078.
(304) Crozier, P. S.; Stevens, M. J.; Forrest, L. R.; Woolf, T. B. J. Mol.
Biol. 2003, 333, 493.
(305) Rohrig, U. F.; Guidoni, L.; Rothlisberger, U. Biochemistry 2002, 41,
10799.
(306) Saam, J.; Tajkhorshid, E.; Hayashi, S.; Schulten, K. Biophys. J. 2002,
83, 3097.
(307) Okada, T.; Fujiyoshi, Y.; Silow, M.; Navarro, J.; Landau, E. M.;
Shichida, Y. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5982.
(308) Okada, T.; Sugihara, M.; Bondar, A. N.; Elstner, M.; Entel, P.; Buss,
V. J. Mol. Biol. 2004, 342, 571.
(309) Salom, D.; Lodowski, D. T.; Stenkamp, R. E.; Le, T. I.; Golczak,
M.; Jastrzebska, B.; Harris, T.; Ballesteros, J. A.; Palczewski, K.
Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 16123.
(310) Stenkamp, R. E. Acta Crystallogr., D: Biol. Crystallogr. 2008, D64,
902.
(311) Teller, D. C.; Okada, T.; Behnke, C. A.; Palczewski, K.; Stenkamp,
R. E. Biochemistry 2001, 40, 7761.
(312) Nakamichi, H.; Okada, T. Proc. Natl. Acad. Sci. U.S.A. 2006, 103,
12729.
(313) Nakamichi, H.; Okada, T. Angew. Chem., Int. Ed. 2006, 45, 4270.
(314) Nakamichi, H.; Buss, V.; Okada, T. Biophys. J. 2007, 92, L106–
L108.
(315) Standfuss, J.; Xie, G.; Edwards, P. C.; Burghammer, M.; Oprian,
D. D.; Schertler, G. F. J. Mol. Biol. 2007, 372, 1179.
(316) Zerangue, N.; Kavanaugh, M. P. Nature 1996, 383, 634.
(317) Tanaka, K.; Watase, K.; Manabe, T.; Yamada, K.; Watanabe, M.;
Takahashi, K.; Iwama, H.; Nishikawa, T.; Ichihara, N.; Kikuchi, T.;
Okuyama, S.; Kawashima, N.; Hori, S.; Takimoto, M.; Wada, K.
Science 1997, 276, 1699.
(318) Amara, S. G.; Fontana, A. C. Neurochem. Int. 2002, 41, 313.
(319) Saier, M. H., Jr.; Tran, C. V.; Barabote, R. D. Nucleic Acids Res.
2006, 34, D181–D186.
(320) Slotboom, D. J.; Konings, W. N.; Lolkema, J. S. Microbiol. Mol.
Biol. ReV. 1999, 63, 293.
(321) Amara, S. G. Nat. Struct. Mol. Biol 2007, 14, 792.
(322) Kanner, B. I.; Borre, L. Biochim. Biophys. Acta 2002, 1555, 92.
(323) Boudker, O.; Ryan, R. M.; Yernool, D.; Shimamoto, K.; Gouaux, E.
Nature 2007, 445, 387.
(324) Yernool, D.; Boudker, O.; Jin, Y.; Gouaux, E. Nature 2004, 431,
811.
(325) Torres, G. E.; Amara, S. G. Curr. Opin. Neurobiol. 2007, 17, 304.
(326) Shrivastava, I. H.; Jiang, J.; Amara, S. G.; Bahar, I. J. Biol. Chem.
2008, 283, 28680.
(327) Larsson, H. P.; Tzingounis, A. V.; Koch, H. P.; Kavanaugh, M. P.
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 3951.
(328) Gu, Y.; Shrivastava, I. H.; Amara, S. G.; Bahar, I. Proc. Natl. Acad.
Sci. U.S.A. 2009, 106, 2589.
(329) Mitrovic, A. D.; Plesko, F.; Vandenberg, R. J. J. Biol. Chem. 2001,
276, 26071.
(330) Higgins, C. F. Annu. ReV. Cell Biol. 1992, 8, 67.
(331) Higgins, C. F. Res. Microbiol. 2001, 152, 205.
(332) Locher, K. P.; Lee, A. T.; Rees, D. C. Science 2002, 296, 1091.
(333) Locher, K. P. Curr. Opin. Struct. Biol. 2004, 14, 426.
(334) Locher, K. P.; Borths, E. FEBS Lett. 2004, 564, 264.
(335) Oloo, E. O.; Tieleman, D. P. J. Biol. Chem. 2004, 279, 45013.
(336) Oloo, E. O.; Fung, E. Y.; Tieleman, D. P. J. Biol. Chem. 2006, 281,
28397.
(337) Ivetac, A.; Campbell, J. D.; Sansom, M. S. Biochemistry 2007, 46,
2767.
(338) Weng, J.; Ma, J.; Fan, K.; Wang, W. Biophys. J. 2008, 94, 612.
(339) Rader, A. J.; Vlad, D. H.; Bahar, I. Structure 2005, 13, 413.
(340) Kim, M. K.; Jernigan, R. L.; Chirikjian, G. S. Biophys. J. 2002, 83,
1620.
(341) Kim, M. K.; Jernigan, R. L.; Chirikjian, G. S. Biophys. J. 2005, 89,
43.
(342) Zheng, W.; Brooks, B. R. Biophys. J. 2005, 88, 3109.
(343) Zheng, W.; Brooks, B. R. Biophys. J. 2006, 90, 4327.
(344) Maragakis, P.; Karplus, M. J. Mol. Biol. 2005, 352, 807.
(345) Miyashita, O.; Onuchic, J. N.; Wolynes, P. G. Proc. Natl. Acad. Sci.
U.S.A. 2003, 100, 12570.
(346) Yang, Z.; Májek, P.; Bahar, I. PLoS Comput. Biol. 2009, 5, e1000360.
(347) Eyal, E.; Chennubhotla, C.; Yang, L. W.; Bahar, I. Bioinformatics
2007, 23, i175–i184.
(348) Hinsen, K. Bioinformatics 2008, 24, 521.
(349) Song, G.; Jernigan, R. L. J. Mol. Biol. 2007, 369, 880.
(350) Liu, L.; Koharudin, L. M. I.; Gronenborn, A. M.; Bahar, I. Proteins,
in press.
(351) Bakan, A.; Bahar, I. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 14349.
(352) Li, H.; Helling, R.; Tang, C.; Wingreen, N. Science 1996, 273, 666.
(353) Haider, S.; Grottesi, A.; Hall, B. A.; Ashcroft, F. M.; Sansom, M. S.
Biophys. J. 2005, 88, 3310.
Functional Mechanisms of Membrane Proteins
(354) de Vries, A. H.; Yefimov, S.; Mark, A. E.; Marrink, S. J. Proc. Natl.
Acad. Sci. U.S.A. 2005, 102, 5392.
(355) Faller, R.; Marrink, S. J. Langmuir 2004, 20, 7686.
(356) Marrink, S. J.; de Vries, A. H.; Mark, A. E. J. Phys. Chem. B 2004,
108, 750.
(357) Bond, P. J.; Holyoake, J.; Ivetac, A.; Khalid, S.; Sansom, M. S. J.
Struct. Biol. 2007, 157, 593.
(358) Scott, K. A.; Bond, P. J.; Ivetac, A.; Chetwynd, A. P.; Khalid, S.;
Sansom, M. S. Structure 2008, 16, 621.
(359) Lee, A. G. Mol. BioSyst. 2005, 1, 203.
(360) Powl, A. M.; East, J. M.; Lee, A. G. Biochemistry 2003, 42, 14306.
(361) Powl, A. M.; Wright, J. N.; East, J. M.; Lee, A. G. Biochemistry
2005, 44, 5713.
(362) Brooijmans, N.; Kuntz, I. D. Annu. ReV. Biophys. Biomol. Struct.
2003, 32, 335.
(363) Shoichet, B. K.; McGovern, S. L.; Wei, B.; Irwin, J. J. Curr. Opin.
Chem. Biol. 2002, 6, 439.
(364) Becker, O. M.; Marantz, Y.; Shacham, S.; Inbal, B.; Heifetz, A.;
Kalid, O.; Bar-Haim, S.; Warshaviak, D.; Fichman, M.; Noiman, S.
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 11304.
(365) Varady, J.; Wu, X.; Fang, X.; Min, J.; Hu, Z.; Levant, B.; Wang, S.
J. Med. Chem. 2003, 46, 4377.
(366) Kenakin, T. Trends Pharmacol. Sci. 2003, 24, 346.
(367) Ghanouni, P.; Steenhuis, J. J.; Farrens, D. L.; Kobilka, B. K. Proc.
Natl. Acad. Sci. U.S.A. 2001, 98, 5997.
(368) Peleg, G.; Ghanouni, P.; Kobilka, B. K.; Zare, R. N. Proc. Natl.
Acad. Sci. U.S.A. 2001, 98, 8469.
(369) Swaminath, G.; Xiang, Y.; Lee, T. W.; Steenhuis, J.; Parnot, C.;
Kobilka, B. K. J. Biol. Chem. 2004, 279, 686.
(370) Bottegoni, G.; Kufareva, I.; Totrov, M.; Abagyan, R. J. Med. Chem.
2009, 52, 397.
(371) Totrov, M.; Abagyan, R. Curr. Opin. Struct. Biol 2008, 18, 178.
(372) Cavasotto, C. N.; Orry, A. J.; Abagyan, R. A. Proteins 2003, 51,
423.
(373) Cavasotto, C. N.; Orry, A. J.; Murgolo, N. J.; Czarniecki, M. F.;
Kocsi, S. A.; Hawes, B. E.; O’Neill, K. A.; Hine, H.; Burton, M. S.;
Voigt, J. H.; Abagyan, R. A.; Bayne, M. L.; Monsma, F. J., Jr. J. Med.
Chem. 2008, 51, 581.
(374) Reynolds, K. A.; Katritch, V.; Abagyan, R. J. Comput. Aided Mol.
Des. 2009, 23, 273.
(375) Cherezov, V.; Rosenbaum, D. M.; Hanson, M. A.; Rasmussen, S. G.;
Thian, F. S.; Kobilka, T. S.; Choi, H. J.; Kuhn, P.; Weis, W. I.;
Kobilka, B. K.; Stevens, R. C. Science 2007, 318, 1258.
(376) Carlson, H. A.; McCammon, J. A. Mol. Pharmacol. 2000, 57, 213.
(377) Cavasotto, C. N.; Abagyan, R. A. J. Mol. Biol. 2004, 337, 209.
Chemical Reviews, 2010, Vol. 110, No. 3 1497
(378) Ferrari, A. M.; Wei, B. Q.; Costantino, L.; Shoichet, B. K. J. Med.
Chem. 2004, 47, 5076.
(379) May, A.; Zacharias, M. Biochim. Biophys. Acta 2005, 1754, 225.
(380) Ota, N.; Agard, D. A. J. Mol. Biol. 2001, 314, 607.
(381) Wei, B. Q.; Weaver, L. H.; Ferrari, A. M.; Matthews, B. W.; Shoichet,
B. K. J. Mol. Biol. 2004, 337, 1161.
(382) Cavasotto, C. N.; Kovacs, J. A.; Abagyan, R. A. J. Am. Chem. Soc.
2005, 127, 9632.
(383) Floquet, N.; Marechal, J. D.; Badet-Denisot, M. A.; Robert, C. H.;
Dauchez, M.; Perahia, D. FEBS Lett. 2006, 580, 5130.
(384) May, A.; Zacharias, M. J. Med. Chem. 2008, 51, 3499.
(385) Zacharias, M.; Sklenar, H. J. Comput. Chem. 1999, 20, 287.
(386) Floquet, N.; Dedieu, S.; Martiny, L.; Dauchez, M.; Perahia, D. Arch.
Biochem. Biophys. 2008, 478, 103.
(387) Floquet, N.; Durand, P.; Maigret, B.; Badet, B.; Badet-Denisot, M. A.;
Perahia, D. J. Mol. Biol. 2009, 385, 653.
(388) Stumpff-Kane, A. W.; Maksimiak, K.; Lee, M. S.; Feig, M. Proteins
2008, 70, 1345.
(389) Delarue, M.; Dumas, P. Proc. Natl. Acad. Sci. U.S.A. 2004, 101,
6957.
(390) Gorba, C.; Miyashita, O.; Tama, F. Biophys. J. 2008, 94, 1589.
(391) Hinsen, K.; Reuter, N.; Navaza, J.; Stokes, D. L.; Lacapere, J. J.
Biophys. J. 2005, 88, 818.
(392) Schroder, G. F.; Brunger, A. T.; Levitt, M. Structure 2007, 15, 1630.
(393) Tama, F.; Miyashita, O.; Brooks, C. L., III. J. Struct. Biol. 2004,
147, 315.
(394) Tama, F.; Miyashita, O.; Brooks, C. L., III. J. Mol. Biol. 2004, 337,
985.
(395) Marechal, J. D.; Perahia, D. Eur. Biophys. J. 2008, 37, 1157.
(396) Suhre, K.; Navaza, J.; Sanejouand, Y. H. Acta Crystallogr., D: Biol.
Crystallogr. 2006, 62, 1098.
(397) Ming, D.; Kong, Y.; Lambert, M. A.; Huang, Z.; Ma, J. Proc. Natl.
Acad. Sci. U.S.A. 2002, 99, 8620.
(398) Tama, F.; Wriggers, W.; Brooks, C. L., III. J. Mol. Biol. 2002, 321,
297.
(399) Brink, J.; Ludtke, S. J.; Kong, Y.; Wakil, S. J.; Ma, J.; Chiu, W.
Structure. 2004, 12, 185.
(400) Balog, E.; Smith, J. C.; Perahia, D. Phys. Chem. Chem. Phys. 2006,
8, 5543.
(401) Best, R. B.; Chen, Y. G.; Hummer, G. Structure 2005, 13, 1755.
(402) Chu, J. W.; Voth, G. A. Biophys. J. 2007, 93, 3860.
(403) Yang, H.; Yu, Y.; Li, W. G.; Yu, F.; Cao, H.; Xu, T. L.; Jiang, H.
PLoS Biol. 2009, 7, e1000151.
CR900095E
1498 Chem. Rev. 2010, 110, 1498–1517

Vibrational Spectroscopy as a Probe of Structure and Dynamics in

Liquid Water

H. J. Bakker*,† and J. L. Skinner*,‡

FOM Institute for Atomic and Molecular Physics, Kruislaan 407, 1098 SJ Amsterdam, The Netherlands, and Theoretical Chemistry Institute and
Department of Chemistry, University of Wisconsin, Madison, Wisconsin 53706

Received May 10, 2009

Contents From a theoretical point of view, this can probably be

1. Introduction
2. Vibrational Spectroscopy as a Probe of Structure
2.1. Experimental IR and Raman Line Shapes
2.2. Theoretical Approaches to Vibrational Line
Shapes
3. Vibrational Spectroscopy as a Probe of Spectral
Diffusion
3.1. Experimental Techniques for Studying Spectral
Diffusion
3.1.1. Spectral Hole Burning
3.1.2. Photon-Echo Peak Shift Spectroscopy
3.1.3. Other Photon-Echo Experiments
3.1.4. Two-Dimensional Vibrational Spectroscopy
3.2. Theoretical Approaches to Spectral Diffusion
Observables
4. Vibrational Spectroscopy as a Probe of Rotational
Dynamics
4.1. Polarization-Resolved Pump-Probe
Spectroscopy
4.1.1. Isotropic Absorption Changes
4.1.2. Long-Time Results: Collective Orientational
Dynamics
4.1.3. Ultrashort-Time Results: Librational Motions
4.1.4. Short-Time Results: Jumping Molecules
4.2. Theoretical Approaches to Rotational
Anisotropy Observables
5. Conclusions and Outlook
6. Acknowledgments
7. References
1. Introduction
Water is, of course, a fascinating and important substance.
For such a simple molecule, its condensed phase properties
are surprisingly complex. Here we might mention the many
different solid phases, the higher density of the liquid as
compared to ice Ih, and the density maximum (as a function
of temperature) in the liquid phase. Moreover, for such a
light molecule, many of the liquid-state properties are
anomalous: the boiling point, freezing point, heat capacity,
surface tension, and viscosity are all unusually high. Even
so, it is perhaps surprising that we still do not fully
understand the properties of the liquid state.1-3
* To whom correspondence should be addressed.
†
FOM Institute for Atomic and Molecular Physics.
‡
University of Wisconsin, Madison.
10.1021/cr9001879  2010 American Chemical Society
Published on Web 11/16/2009
attributed to two features of liquid water: cooperative
1498 hydrogen bonding (H-bonding) and nuclear quantum effects.
1500 The former refers to the fact that the binding energy of two
1500 H-bonded molecules is modified by the presence of a third
1500 molecule.4-9 In terms of simulating the liquid, then, it follows
that the potential energy cannot be written as a sum of two-
1502 molecule terms. This means that simple two-body simulation
models cannot completely describe reality, and attempts to
1502 capture the effects of these many-body interactions with
polarizable models are not fully satisfactory either.8,10 Nuclear
1503 quantum effects occur because the hydrogen nucleus is
1503 sufficiently light that classical mechanics for the nuclear
1504 motion is simply not adequate. Thus, classical mechanics
1505 cannot describe such important properties as spatial disper-
1506 sion of the hydrogen positions, nuclear tunneling, zero-point
energy, and quantization of nuclear motions. Much energy
1507 has recently been expended toward the simulation of liquid
water using ab initio electronic-structure methods (which,
1508 in priniciple, will produce the correct Born-Oppenheimer
potential surface, including the effects of many-body
1508 interactions),11-19 together with methods for quantum
dynamics,19-25 but still more work needs to be done before
1509
we have a complete and accurate description.
1510 On the experimental side, one is limited by the existing
1511 experimental techniques that have been brought to bear on the
1512 problemsthese techniques include nuclear magnetic resonance
(NMR);26-31 dielectric relaxation;32 X-ray scattering;33 X-ray
1514 absorption, emission, and Raman spectroscopies;34,35 elastic
1515 and inelastic neutron scattering;36-38 and THz spectroscopy.39,40
For example, NMR is very useful for measuring the
1515
correlation times for rotation of various molecule-fixed axes,
but it cannot provide the time dependence of the rotational
correlation functions. X-ray scattering is primarily sensitive
to oxygen positions and can provide the oxygen-oxygen
radial distribution function but cannot say much about higher-
order oxygen-position correlations or about hydrogen posi-
tions. Through the use of isotope-substitution experiments,36
neutron scattering does provide information about both
hydrogen and oxygen positions but, again, not about higher-
order correlations. The new techniques of X-ray absorption,
emission, and Raman scattering are sensitive to the electronic
excited states of the liquid, which are difficult to calculate,
perhaps explaining why interpretation of these experiments
has not been unambiguous.34,35,41
In this review, we discuss the experimental technique of
vibrational spectroscopy and the role that it has played in
elucidating the structure and dynamics of liquid water in
thermal equilibrium. These techniques begin with the straight-
Structure and Dynamics in Liquid Water
Huib Johan Bakker was born on March 2, 1965, in Haarlem, The
Netherlands. He did his Ph.D. studies in the group of Prof. dr. Ad Lagendijk,
at the FOM Institute for Atomic and Molecular Physics (AMOLF). From
1991 to 1994 he worked as a postdoc in the group of Prof. dr. Heinz
Kurz at the Institute of Semiconductor Electronics at the Technical
University of Aachen, Germany. In 1995 he became a group leader at
AMOLF, heading the group “Ultrafast Spectroscopy”. The research work
of the group includes the spectroscopic study of the structure and ultrafast
dynamics of water interacting with ions and (bio)molecular systems and
the study of the mechanism of proton transfer in aqueous media. In 2001
he became a full professor of Physical Chemistry at the University of
Amsterdam, The Netherlands. In 2004, he received the Gold Medal of
the Royal Netherlands Chemical Society for his work on the ultrafast
dynamics of aqueous systems.
James Lauriston Skinner received his Ph.D. from Harvard under the
direction of Prof. Peter Wolynes in 1979, and he did postdoctoral work
with Hans Andersen at Stanford. He then spent nine years on the faculty
of Columbia University, becoming full professor in 1986. In 1990 he moved
to the University of Wisconsin as the Joseph O. Hirschfelder Professor of
Chemistry and Director of the Theoretical Chemistry Institute, positions
he currently occupies. His research interests involve the structure,
dynamics, relaxation, and spectroscopy of condensed-phase systems.
forward infrared (IR) absorption experiments of many years
ago, followed by polarized and depolarized Raman scattering
experiments within the last 40 years and ultrafast IR and
Raman experiments of the past decade.42-45 Vibrational
spectroscopy is a useful technique for studying the structure
and dynamics of water for two reasons. First, vibrational
frequencies, especially in the OH stretch region, are very
sensitive to molecular environment, including H-bonding.
Thus, a simple absorption experiment, which in large part
measures the ensemble of vibrational frequencies in the
liquid, reflects the distributions of local environments, which,
in turn, is intimately related to the structure of the liquid.
Local structure, of course, evolves in time due to molecular
dynamics. As such, the frequency of any one OH stretch
also fluctuates in timesa process known as spectral diffusion.
Chemical Reviews, 2010, Vol. 110, No. 3 1499
A time-resolved experiment with laser-pulse durations on
the time scale of molecular motion in the liquid, or shorter,
in principle can measure the statistics of spectral diffusion,
which, in turn, provides an excellent probe of molecular
dynamics. The second fundamental reason that vibrational
spectroscopy is so useful is that the transition dipole vector
and polarizability tensor of a molecule, relevant for IR and
Raman spectroscopy, respectively, are tied to its molecule-
fixed frame. Thus, for example, ultrafast IR spectroscopy
with polarized lasers can provide a detailed probe of the
dynamics of molecular rotation.43
For an isolated gas-phase water molecule, the two local-
mode OH stretches are coupled, which leads to a splitting
between the symmetric and antisymmetric stretch normal-
mode frequencies of some 100 cm-1. In condensed phases,
the degeneracy of the two local-mode frequencies is broken,
very strongly (in comparison with the coupling), due to
different local environments, diminishing the effect of the
coupling.46 On the other hand, in condensed phases, there is
also the possibility of vibrational coupling between OH
stretches on different molecules. Our current understanding
is that this coupling has a pronounced effect on instantaneous
vibrational eigenstates for single excitations, which can
extend over ten or more molecules.47,48 Concomitantly, this
delocalization has a pronounced effect on both the steady-
state and ultrafast vibrational spectroscopy, causing, for
example, the appearance of a collective vibrational peak in
the polarized Raman spectrum48 and the very rapid anisotropy
decay (due to vibrational energy transfer) in pump-probe
experiments.49-54 In terms of the study of vibrational
dynamics,55-71 these are important and fascinating effects.
They are, however, less relevant to the structure and
dynamics of liquid water in thermal equilibrium (where all
OH stretch vibrations are in their ground states).
Therefore, in order to study the structure and dynamics
of liquid water in thermal equilibrium, it is more useful to
employ isotopes, for example, by considering dilute HDO
in D2O or dilute HDO in H2O (and, actually, Nicodemus
and Tomakoff72 have also looked at dilute HTO in H2O). In
these two situations, the OH or OD stretches, respectively,
function as localized chromophores, since, because of
substantial frequency mismatch, they are effectively un-
coupled from all other stretches in the sample. This means
that they can act as effective reporters of local structure and
dynamics, unfettered by the complications of delocalization
and energy transfer described above in neat H2O. For these
reasons, we will focus only on these two systems in this
review.
Ultrafast vibrational spectroscopy has also been used to
study vibrational energy relaxation in HDO:H2O or HDO:
D2O, typically measuring the OD or OH stretch lifetime,
respectively,73-82 although some recent experiments have
focused on bend lifetimes.83,84 This process is fundamentally
important for understanding water’s role as a solvent for
chemical and electron transfer reactions. However, knowl-
edge of these rates provides only indirect information about
the structure and dynamics of water in thermal equilibrium
(and even this relies heavily on theoretical modeling).
Therefore, except for the inescapable role vibrational relax-
ation plays in all vibrational spectroscopy, we will not focus
on it herein.
Thus, the topic of this review is conventional and ultrafast
vibrational spectroscopysespecially those techniques that
shed light on H-bonding structure and dynamics, spectral
1500 Chemical Reviews, 2010, Vol. 110, No. 3
diffusion, and rotational dynamicssof dilute HDO in either
liquid H2O or D2O, in the OD or OH stretch region,
respectively. Earlier experiments focused on the latter, since
it was easier to coax lasers to perform well at these shorter
wavelengths, whereas with later developments it has become
just as easy to perform experiments on the former. At this
point, it might be worth mentioning that there is a slight
preference for working on the HDO:H2O system, for several
different reasons. First, this system probes the structure and
dynamics of H2O (rather than D2O), which is inherently more
important. Second, the OD stretch lifetime is longer than
the OH stretch lifetime, providing more dynamic range in
the ultrafast experiments. Third, theoretical simulation models
have been parametrized for H2O (and usually not for D2O),
which makes modeling H2O more reliable. On the other hand,
in ultrafast experiments, one is often worried about local
heating from vibrational relaxation, and these effects are more
problematic for HDO:H2O (since in this case the HDO
concentration needs to be higher, as described below). Note,
also, that nuclear quantum effects, not taken into account
explicitly in classical simulations, are more important for
H2O than for D2O. Interpretation of vibrational spectroscopic
observables, especially for the most modern ultrafast tech-
niques, is greatly facilitated by molecular dynamics simula-
tions and associated theoretical spectroscopy. Thus, this
review also focuses on these complementary computational
efforts.
In the next section we discuss how conventional IR and
Raman experiments, together with computer simulation and
theoretical interpretation, shed some light on the problems
of liquid structure and H-bonding. In section 3 we consider
ultrafast experiments that can probe spectral diffusion, most
commonly and conveniently characterized statistically by the
frequency time-correlation function, and we discuss attempts
to understand experimental results from computer simulation
and theoretical modeling. In section 4 we consider pump-probe
experiments that measure rotational anisotropy decay, dis-
cussing both short-time (inertial) and long-time (collective)
dynamics, and comparing them with simulation and theory.
In section 5 we conclude.
2. Vibrational Spectroscopy as a Probe of
Structure
2.1. Experimental IR and Raman Line Shapes
IR and Raman (polarized, depolarized, and unpolarized)
line shapes for dilute HDO in D2O at room temperature are
all rather similar and are characterized by a large red shift
of roughly 300 cm-1 from the gas-phase OH stretch
frequency (for HDO) of 3707 cm-1 and a breadth of between
250 and 300 cm-1. Both the red shift and the spectral breadth
are unusually large for a vibrational transition. The shift is
generally believed to result from H-bonds to the H atom in
HDO. Such H-bonds weaken and lengthen the covalent OH
bond, leading to a red shift in the average frequency. Liquid
water, being disordered, has a wide range of H-bond
distances and angles, and, in addition, has some broken
H-bonds. Since the OH stretch frequency is so sensitive to
these H-bonds, the thinking is that the distribution of H-bond
configurations leads to a distribution of OH stretch frequen-
cies, which is reflected in the large spectral breadth. There
are some subtle differences between the IR and Raman
spectra, namely that the Raman line shape is shifted to the
Bakker and Skinner
blue by about 30 cm-1, and it also has a shoulder on the
blue side that is absent in the IR spectrum.
The situation for HDO:H2O is similar, although all (OD)
frequencies are smaller (because of the heavier D atom).
Thus, the gas-phase OD stretch frequency is 2724 cm-1, the
liquid-state red-shift is about 225 cm-1, and the spectral
breadths are between 160 and 180 cm-1. Note that, as
discussed below, the ratios of gas-phase frequencies and red
shifts for the OD and OH stretches are roughly the same,
while the ratio of the line widths is somewhat smaller. The
slight blue shift and blue shoulder in the Raman spectra are
also seen for this (HDO:H2O) system.
From a theoretical perspective, the goal is to reproduce
the line shapes as closely as possible and then use the
simulation results and other theoretical analysis to address a
number of questions. For example, are the general thoughts
about the shift and broadening in these systems correct? Can
we understand the differences between IR and Raman line
shapes and between the HDO:D2O and HDO:H2O systems?
What, if anything, are the line shapes telling us about
H-bonding in liquid water? The idea has been pervasive in
the literature, as evidenced by multiple Gaussian fits to line
shapes85,86 and the ability to burn narrow transient spectral
holes,87-90 that different H-bonding configurations correspond
to relatively narrow frequency distributions, which together
make up the observed spectra. Is this idea correct?
2.2. Theoretical Approaches to Vibrational Line
Shapes
The modern starting point for the calculation of spectral
line shapes comes from quantum time-correlation functions
(TCFs).91,92 Thus, in principle, one could simulate liquid
water, treating all electrons and nuclei quantum mechanically,
and use existing TCF expressions to obtain IR and Raman
line shapes. For a variety of technical and computational
reasons, we are not yet close to being able to proceed in
this fashion. One approach is to treat the nuclear motion,
which occurs either on an empirical93-95 or ab initio96
potential energy surface, classically. Notwithstanding issues
having to do with the accuracy of either of these surfaces,
especially for high-frequency (OH or OD stretch) motion,
there is reason to be concerned with the adequacy of the
classical approximation for high frequency (compared to kT/
p) and very anharmonic modes.97 Therefore, a number of
workers have instead considered a mixed quantum/classical
approach, where (in the case, for example, of HDO:D2O)
the OH stretch is treated quantum mechanically, the water
bends and OD stretches are neglected, and the other nuclear
degrees of freedom are treated with classical mechanics. Even
within this approach there are several critical issues to be
addressed. First, one needs to define the potential surface
for the classical degrees of freedom, which is typically taken
to be an empirical surface for rigid water. And second, and
perhaps more importantly, one needs to define the potential
surface for the OH stretch motion. There are really two
choices here: either one can use the empirical surface for
rigid water,98-109 together with a reference potential for the
gas-phase stretch, or one can, independent of the empirical
potential, use ab initio calculations to determine this
surface.110-112
The efforts Skinner and co-workers have undertaken in
recent years involve the mixed quantum/classical approach,
using the SPC/E model113 for a molecular dynamics (MD)
simulation of rigid water and electronic structure (ES)
Structure and Dynamics in Liquid Water
Figure 1. Theoretical97 and experimental124,125 IR and (unpolarized)
Raman line shapes for dilute HDO in H2O under ambient conditions.
Reproduced with permission from ref 97. Copyright 2009 John
Wiley and Sons.
calculations to determine the OH/OD stretch fundamental
transition frequency for each simulation configuration.97,112,114-118
In the latest implementation,97,117,118 these ES calculations
lead to a simple quadratic map between the vibrational
frequency and the electric field, produced by the point
charges of the surrounding molecules, at the H/D atom of
interest, and in the direction of the OH/OD bond. With this
map it is a simple matter to produce a frequency trajectory
from the MD simulation. Another important spectral quantity
is the transition dipole, which lies more or less along the
OH/OD bond vector.114,119 In liquid water the magnitude of this
transition dipole depends strongly on the local environment,120,121
which, in analogy with electronic spectroscopy, is called a
non-Condon effect. It is important to take this non-Condon
effect into account,114,115 and so they developed an ES-based
electric-field map for this quantity as well.114,117 A similar
(although much less important) non-Condon effect can be
employed for the components of the transition polarizability
tensor, which they implemented within the bond-polariz-
ability model.97,114,117 Armed with frequency, transition
dipole, and transition polarizability trajectories, one can,
within the mixed quantum/classical formulation, calculate
IR and Raman spectra.122 The results include the effects of
motional narrowing,123 and the broadening effects of mo-
lecular rotation, and vibrational relaxation (which is described
phenomenologically).
In Figure 1 we show experimental124,125 and theoretical97
IR and (unpolarized) Raman line shapes for dilute HDO in
H2O at 298 K. There are no adjustable parameters in the
theoretical calculation, except a scale factor that matches
experimental and theoretical peak heights. In both cases, one
sees quite reasonable agreement between experiment and
theory. The slight blue shift of the Raman spectrum (with
respect to the IR spectrum) and the weak shoulder on the
blue side of the Raman spectrum are both captured by theory.
We can begin to answer some of the questions posed
above, by considering the theoretical distribution of frequen-
cies sampled by the ensemble of OD chromophores, the
weighted (by the square of the transition dipole) frequency
distribution (or spectral density) appropriate for the IR
spectrum,126 and the IR line shape, as shown in Figure 2.
Chemical Reviews, 2010, Vol. 110, No. 3 1501
Figure 2. Theoretical frequency distribution, IR spectral density,
and IR line shape, for dilute HDO in H2O.
Figure 3. Theoretical frequency distribution, IR spectral density,
and IR line shape for dilute HDO in D2O.
All three curves have the same (area) normalization. The
frequency distribution itself is quite asymmetrical, while
the spectral density is more symmetrical and red-shifted. These
differences are both due to non-Condon effects, since the
transition dipole is much larger for molecules on the red
side of the line.114,115,117,127 Since non-Condon effects are much
smaller for Raman spectroscopy, this immediately explains
why the Raman spectrum is blue-shifted relative to the IR
and why the shoulder (on the blue side) in the Raman
spectrum is suppressed in the IR. The IR line shape is
narrower than the spectral density due to motional narrowing
(despite the weak broadening effects of rotations and
vibrational relaxation). Still, the effects of motional narrow-
ing are not dominant, and so the qualitative picture that the
breadth of the line is determined by the distribution of
frequencies is correct. In addition, the overall red shift clearly
comes from H-bonding interactions with other molecules,
as discussed above and as borne out by the ES calculations.
In Figure 3 we show the same quantities for HDO:D2O,
with similar conclusions, although in this case the amount
of motional narrowing (relative to the spectral density) is
smaller. This is easy to understand within the context of the
simple Kubo model.123 In this model the frequency fluctua-
tions describe a Gaussian process and, hence, are character-
ized completely by the two-point frequency-frequency
correlation function (FFCF) 〈δω(t) δω(0)〉, where δω(t) )
ω(t) - 〈ω〉. Thus, ω(t) is the time-dependent transition
frequency of a given OH oscillator and 〈ω〉 is the average
1502 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 4. IR spectral density subdistributions for different H-
bonding classes, for dilute HDO in D2O.
frequency of the ensemble. Within the Kubo model, the
frequency fluctuations describe a Gauss-Markov process,
meaning that the FFCF decays exponentially:
〈δω(t) δω(0)〉 ) ∆2e-|t|/τ (1)
∆2 is the mean-squared frequency fluctuation, and τ is the
correlation time. The FFCF for water is somewhat more
complicated (see below); still, one can understand the
motional narrowing trend by considering this model. The
time scale for frequency fluctuations (τ) is not very different
for the HDO:D2O and HDO:H2O systems, while ∆ is
significantly larger in the case of HDO:D2O. According to
the Kubo model, the dimensionless product ∆τ determines
the extent of motional narrowing. Since this product for
HDO:D2O is larger, there is less motional narrowing. This
explains why the observed line width for HDO:H2O is
smaller than one might expect, given the scaling, discussed
above, of the gas-phase frequencies and the red-shift.
Similarly, there is even more motional narrowing in the HTO:
H2O system.72
Finally, we can attempt to address the question about the
relationship between H-bonding and vibrational line shapes.
To do so, we need a definition of the H-bond. Although this
topic is fraught with ambiguities,128 herein we will use a new
ES-based definition developed recently in the Skinner
group.128 Within this approach, each HDO molecule is
assigned to one of eight H-bonding classes, according to the
total number of H-bonds, the number of donors (H-bonds to
the H or D atoms), and, if it is a single donor, whether the
donor is the H or D atom. Thus, for example, an HDO
molecule in class 2SD has a total of two H-bonds, it is a single
(S) donor, and the single donor is the D atom. For the HDO:
D2O problem, Auer et al. decomposed the frequency
distribution into subdistributions for each of the eight
H-bonding classes.117 In Figure 4 we do something analo-
gous, but instead we decompose the spectral density (since
that is more relevant for the IR spectrum).126 One sees that
the eight distributions fall into two groups, of four classes
each. Those lacking an H-bond to the H (1N, 2N, 2SD, and
3SD) all absorb between 3450 and 3750 cm-1. On the other
hand, those four classes possessing an H-bond to the H (2SH,
3SH, 3D, 4D) all absorb in the broad range between 2900 and
3700 cm-1. From this analysis we can say that (within the
model!) molecules with a free-OH group do absorb in a
relatively narrow frequency range (which, by the way, are
Bakker and Skinner
the molecules producing the blue shoulder in the Raman
spectrum114), while molecules without a free-OH (that is,
those in the H-bonding region) do not. To reiterate, then,
it is not correct to say that molecules in specific H-bonding
environments (for example, those in the 4D, sometimes called
“ice-like”, class) absorb in relatively narrow frequency
ranges.97,99,117,126 This conclusion seems to be robust to
different H-bond definitions126 (although at present we have
no information as to robustness with respect to different water
simulation models).
3. Vibrational Spectroscopy as a Probe of
Spectral Diffusion
The H-bond dynamics of liquid water play an important
role in chemical processes occurring in aqueous media.120,129
For instance, the formation and breaking of H-bonds between
water and reacting molecules can strongly affect the kinetics
and energetics of chemical reactions. The dynamics of the
H-bond network occur over many time scales, ranging from
femtosecond fluctuations that involve a few molecules to
picosecond diffusive motions that involve the breaking and
formation of H-bonds.99,103,130,131
Time-resolved IR spectroscopy is an ideal technique to
investigate the dynamics of H-bonds in water, because, as
we discussed in the previous section, the frequencies of the
intramolecular OH stretching vibrations of a water molecule
are particularly sensitive to the molecule’s H-bond environ-
ment. Generally speaking, the strengthening of the bond will
lead to a red shift of the OH frequency, and a weakening
will lead to a blue shift of this frequency.132 Therefore, time-
dependent shifts in the OH vibrational frequency, or spectral
diffusion, can be used to characterize changes in H-bonding
and intermolecular configuration.
Spectral diffusion is most succinctly described by the
FFCF, which, within the Kubo model, is a single exponential,
as in eq 1. More generally, the time dependence of the FFCF
is often modeled as a sum of exponentials,
〈δω(t) δω(0)〉 ) ∑ ∆i2e-|t|/τi
(2)
i
each with an amplitude ∆i2 and a time constant τi.
3.1. Experimental Techniques for Studying
Spectral Diffusion
In the recent past, the frequency fluctuations of the OH
stretch vibrations have been investigated with nonlinear
vibrational spectroscopy. In the following, we will briefly
describe and compare the nonlinear techniques that have been
used. In all experiments, the femtosecond mid-IR pump and
probe pulses are generated via a sequence of nonlinear
frequency-conversion processes that are pumped by the
output of Ti:sapphire multipass and/or regenerative amplifiers
(800 nm, 1 kHz, pulse energy g1 mJ). The mid-IR pump
has a wavelength of ∼3 µm (=3300 cm-1) or ∼4 µm (=2500
cm-1), depending on whether the OH or the OD stretch
vibration is being studied. In most studies, the pulses are
generated via an optical parametric amplification process in
a potassium titanyl phosphate (KTP) or KNbO3 crystal that
is pumped by a part of the 800 nm pulses of the Ti:sapphire
laser. The energy of the generated pump pulses varies
between 2 µJ133 and 10 µJ.134-139 In some studies, the KTP
and KNbO3 crystals are quite short (1 mm) and the 800 nm
Structure and Dynamics in Liquid Water Chemical Reviews, 2010, Vol. 110, No. 3 1503

pulses have a pulse duration of 30105,140 or 40 fs.133,141 The

resulting mid-IR pulses are ultrashort, having a pulse duration
of ∼50 fs.105,133,140,141 The bandwidth of these pulses is ∼400
cm-1, thereby covering completely the absorption band. In
other studies, longer crystals (4-5 mm) and longer 800 nm
pulses (∼100 fs) have been used, leading to the generation
of longer (∼150 fs) mid-IR pump pulses with a much
narrower spectral bandwidth of ∼100 cm-1.134-137,139
The pulses are used to excite and probe OH/OD stretch
vibrations. As discussed above, isotopically diluted water is
studied, probing either the OH stretch vibration of HDO
dissolved in D2O or the OD vibration of HDO dissolved in
H2O. The use of isotopically diluted water has as advantages
that the heating effects are limited and, as pointed out in the
Introduction, that the signals are not affected by the strong Figure 5. Transient spectra of an aqueous solution of 0.5% HDO
intra- and intermolecular couplings of the OH/OD stretch dissolved in D2O obtained with three different pump pulses. The
vibrations of pure H2O/D2O.49 In cases where the stretch spectra are measured at a delay of 300 fs after excitation by the
vibration of HDO:D2O is studied, the concentration of HDO pump pulse.
is ∼1%. When the stretch vibration of HDO:H2O is studied,
the concentration of HDO is somewhat higher, ranging from
2.5 to 4%. In the latter experiments, a higher concentration
has to be used, because the H2O solvent has a non-negligible
absorption in the frequency region of the stretch vibration.
For either system, the excitation induces a transfer of a few
percent of the population of the V ) 0 ground state of the
OD/OH stretch vibration to the V ) 1 excited state. This
transfer leads to a bleaching effect at the fundamental
transition frequencies due to a decrease of the V ) 0 f 1
absorption and the presence of the V ) 1 f 0 stimulated
emission. The population transfer to V ) 1 also leads to
induced absorption at frequencies corresponding to the V )
1 f 2 transition. The latter absorption is red-shifted with
respect to the fundamental transition by ∼220/150 cm-1 due
to the anharmonicity of the OH/OD stretch vibration. Figure 6. Transient spectra of an aqueous solution of 0.5% HDO
dissolved in D2O at different time delays after excitation with a
pump pulse centered at 3580 cm-1.
3.1.1. Spectral Hole Burning
a different central frequency for the pump pulse. For probe
In spectral hole burning, the OH/OD stretch absorption
frequencies >3300 cm-1, the signal is positive and results
band is excited with a relatively narrow-band mid-IR from the bleaching of the fundamental V ) 0 f 1 transition.
excitation pulse. If the time scale for spectral diffusion is For frequencies <3300 cm-1, the signal is negative and
long compared to the laser pulse duration, the pulse will only results from the induced V ) 1 f 2 transition absorption.
excite water molecules in a limited frequency range, leading For all three pump frequencies, the bleaching is narrower
to a so-called spectral hole in the absorption band. Due to
than the linear absorption band, which means that the
spectral diffusion, the spectral hole will broaden and the excitation leads to the formation of a spectral hole in an
central frequency of the hole will shift toward the peak of inhomogeneously broadened absorption band. In Figure
the absorption band. If the spectral modulation forms a 6, transient spectra are shown at different delays after
Gauss-Markov process, the shift of the central frequency excitation in the blue wing of the absorption band. It is
of the spectral hole shows the same dynamics as the FFCF; seen that the spectrum broadens and that its central
that is, this frequency shifts toward the peak with an
frequency shifts toward the maximum of the absorption
exponential time dependence, characterized by a time
band. From these spectral dynamics, it was deduced that
constant τ. The width of the spectral hole broadens with a the FFCF shows a dominant exponential decay component
time constant half as long. with a time constant of ∼1 ps.142-144
The first studies of spectral diffusion of the OH stretch
vibrations of water were performed via spectral hole-burning 3.1.2. Photon-Echo Peak Shift Spectroscopy
spectroscopy of the OH stretch of HDO dissolved in D2O
using relatively long mid-IR pulses with pulse durations > In photon-echo spectroscopy, the OH stretch absorption
1 ps. From the measurements, it was concluded that the band is excited with broad-band IR laser pulses that cover
stretch absorption band contains several sub-bands.87,88 Later the whole width of the absorption band. The excitation is
studies employing femtosecond pulses did not find evidence performed with two pulses, with wavevectors k1 and k2, that
for the existence of sub-bands and rather found that the enter the sample at different angles. As a result, a population
spectral dynamics can be modeled well as a Gauss-Markov grating is excited, with a wavevector corresponding to the
process.142,143 difference wavevector of the two excitation pulses. If the
Figure 5 presents transient absorption spectra of the OH absorption band is inhomogeneously broadened, there will
stretch vibration of HDO:D2O measured with IR pulses with be a population grating for each different frequency in the
a pulse duration of ∼150 fs. Each spectrum is obtained with absorption band. The relative phases of these different
1504 Chemical Reviews, 2010, Vol. 110, No. 3
frequency gratings will change with increasing delay between
the two excitation pulses, because of the differences in the
rate of phase evolution of the OH oscillators in the time
interval between the two excitation pulses. In the population
grating phase (after the second pulse), there is no further
phase accumulation. The frequency gratings are read out by
a third pulse that has a direction k3. This third pulse generates
a polarization in the directions k3 + k2 - k1 and k3 + k1 -
k2. In the so-called rephasing direction, which is k3 + k2 -
k1, if the pulse with k1 enters first, the phase accumulation
is opposite to the phase accumulation between the first and
second pulses. Hence, the polarizations generated by the third
pulse get into phase again after a time delay that corresponds
to the time difference between the first two excitation pulses.
At that moment in time, all polarizations add up construc-
tively, and a light pulse is emitted from the sample in the
rephasing direction. Because the emission is delayed with
respect to the entrance time of the third pulse, the emitted
light pulse is denoted as a photon echo.
In an echo-peak shift measurement, the delay between the
first two excitation pulses that gives the maximum echo
signal is measured as a function of the so-called waiting time
between the excitation pulses and the third interrogation
pulse. In the absence of spectral diffusion and for short
waiting times, the maximum photon-echo signal is attained
if the first two excitation pulses are somewhat delayed with
respect to each other, because this configuration leads to the
largest amplitude of the population gratings. However, if
there is spectral diffusion and the waiting time is comparable
to the characteristic time scale of these processes, the optimal
delay between the two excitation pulses will become smaller.
This can be understood as follows. If the excitation pulses
enter the sample simultaneously, all frequency gratings are
excited in phase, and the echo signal is generated directly
with the entrance of the third pulse. In this case, changes in
the frequencies of the oscillators will not affect their relative
phases and there is no rephasing process required to let the
polarizations add up constructively. Hence, the echo signal
becomes insensitive to the frequency fluctuations. The decay
of the optimal time delay between the two excitation pulses
as a function of the waiting time constitutes the so-called
echo peak signal. For long waiting times, it can be shown
that the waiting-time dependence directly represents the time
dependence of the FFCF.145,146
The spectral diffusion of HDO dissolved in D2O has been
measured with several echo-peak shift spectroscopic studies.105,140,147
These studies confirmed the presence of a spectral fluctuation
process with a time constant of ∼1 ps, as was also observed
with spectral hole burning. In one study, very slow additional
dynamics were observed occurring on a time scale of 5-15
ps.147 However, in a later study, it was argued that these
dynamics likely result from an interference effect with signal
generated from the D2O solvent.148 In the other echo-peak
shift study,105,140 by Tokmakoff and co-workers, in which
40 fs mid-IR pulses were used, it was observed that there
were two different time scales for the spectral modulation.
The echo-peak shift resulting from this study is presented
in Figure 7. Analysis of the peak shift shows a slow spectral
diffusion process with a time constant of 1.4 ps and a fast
spectral diffusion process with a time constant of ∼50 fs.
Moreover, an increase in the FFCF extracted from experi-
ment was observed at a waiting time of ∼180 fs. This
“recurrence” was assigned to an oscillation associated with
the stretching vibration of the bond.98,103 It follows from this
Bakker and Skinner
Figure 7. Echo intensity and echo-peak shift as a function of the
waiting time τ2 for HDO:D2O.105 Reproduced with permission from
ref 105. Copyright 2003 American Association for the Advancement
of Science.
observation that the bond between HDO and a D2O molecule
is underdamped.
3.1.3. Other Photon-Echo Experiments
The first photon-echo studies on water were performed
with only two pulses.149 In these experiments the second
pulse serves both to excite and to probe the population
gratings. In the experiment the time-integrated echo signal
was measured as a function of the delay between the two
excitation pulses. This experiment does not give direct
information on the time scales of the frequency fluctuation
processes, but it does give information on the dephasing time
of the OH oscillators. Using certain assumptions, this
dephasing time can be related to the time scale of the
frequency fluctuations. If the oscillators lose their phase
relation in the time interval between the first and second
pulses, the generated population grating will not have a well-
defined relative phase and there will be no echo signal.
Hence, the echo signal will decrease with increasing delay
between the two excitation pulses. In such a two-pulse
photon-echo experiment, the dephasing time T2 was measured
to be 66 fs < T2 < 132 fs. This experiment shows that the
absorption line of the OH stretch vibrations contains a
significant homogeneous component, which indicates the
presence of very rapid spectral fluctuations. The time scale
of these very rapid spectral fluctuations was estimated to be
∼30 fs. The observation of a significant homogeneous
broadening component agrees with the findings of the
spectral hole-burning studies, in which it was found that the
spectral width of the hole is already substantial (>100 cm-1)
directly after the excitation (Figures 5 and 6).
In another two-pulse photon echo experiment, the echo
signal was time-resolved by interfering this signal with a
reference pulse (local oscillator).150 The delay of the maxi-
mum of the echo signal with respect to the second pulse
was measured as a function of the delay between the first
and second pulses. For a purely inhomogeneously broadened
system, the delay of the echo signal would always correspond
to the delay between the first and second pulses. In the
presence of frequency fluctuations, the delay of the maximum
of the echo signal will not exactly follow the increase of the
delay between the two excitation pulses, because the
frequency fluctuations prevent the full rephasing of the echo
signal from occurring. Hence, the maximum echo signal will
Structure and Dynamics in Liquid Water
occur at a shorter delay after the second pulse. From these
experiments it followed that the spectral modulation includes
two processes with time scales of 130 and 900 fs. This
observation thus agrees reasonably well with the results of
the echo-peak shift studies.
3.1.4. Two-Dimensional Vibrational Spectroscopy
Two-dimensional (2D) vibrational spectroscopy is another
variation of third-order nonlinear spectroscopy. The name
2D spectroscopy refers to the fact that the signals measured
with this technique can be presented in a contour plot as a
function of two frequency dimensions, corresponding to the
excitation and the probing frequencies. There are two
variations of 2D spectroscopy: double-resonance or dynamic
hole-burning spectroscopy and pulsed Fourier-transform or
heterodyne-detected photon-echo spectroscopy.
Double-resonance spectroscopy is a conventional spectral
hole-burning experiment in which a narrow-band pump pulse
is scanned through the absorption band. The frequency
spectrum of the pump pulse is often produced by passing a
short, broad-band pump pulse through a piezocontrolled
Fabry-Perot etalon. With the etalon, both the central
frequency and the bandwidth can be varied. The probe pulse
is a fraction of the broad-band input pulse, and after its
transmission through the sample, it is spectrally dispersed
and frequency-resolved detected. The main difference with
a conventional spectral hole-burning experiment is formed
by the adjustable Fabry-Perot filter used to control the pump
spectrum before the pump enters the sample.
Heterodyne-detected photon-echo spectroscopy is a three-
pulse photon-echo experiment. The generated photon-echo
signal is Fourier transformed with respect to two time
variables, with the first being the delay time between the
two excitation pulses and the second being the time between
the echo signal and the third pulse. This technique can be
best explained as a multiple hole-burning experiment. Two
ultrashort pulses separated by a certain time show a
frequency-modulated spectrum with a modulation frequency
that is inversely proportional to the pulse delay. This spectral
modulation imprints a corresponding population frequency
grating in the sample absorption band or, in other words,
burns multiple holes in the absorption line. Increasing the
time separation between the pulses in the first pair leads to
a finer frequency modulation and to more severe smearing
of the holes due to spectral diffusion. The dependence of
the signal on the excitation frequency is obtained by
performing many experiments in which the delay between
the two excitation pulses is varied. Fourier transformation
of all these signals gives the dependence of the signal on
the excitation frequency. An important condition for this
procedure is that the phase relation between the two
excitation pulses is well-defined. The probing frequency axis
is obtained by interferometric superposition of the echo signal
with a fourth laser pulse. This fourth laser pulse acts as a
local oscillator in a heterodyning experiment. The frequency
of the echo signal can be determined in the time domain by
scanning the time delay of the local oscillator and Fourier
transforming the thus obtained signal with respect to this
time variable or by spectrally dispersing both beams in a
spectrograph. The outcome of the experiment is dependent
on the phase difference between the two excitation pulses
and the phase difference between the echo signal and the
local oscillator field. Hence, the experiment requires a high
mechanical stability.
Chemical Reviews, 2010, Vol. 110, No. 3 1505
The two types of 2D spectroscopic experiments give the
same type of information. The double-resonance technique
has as an advantage that it is relatively simple and that it
does not require phase stability. This technique has as a
disadvantage that the measured spectral shapes are always
convoluted with the prechosen bandwidth of the pump pulse.
As a result, fine spectral details (narrow homogeneous lines)
could be missed if the prechosen pump bandwidth is too
large. The heterodyne-detected echo has as an advantage that
it automatically provides the optimal frequency resolution
for studying the spectral dynamics. If the spectrum contains
very narrow homogeneous lines, the signals obtained with
long delay times between the excitation pulses will strongly
contribute to the 2D spectrum, meaning that the 2D spectrum
indeed will show these fine spectral details.
Both 2D techniques can be used to monitor spectral
diffusion of the OH/OD stretch vibrations of water by
measuring the signal as a function of the delay between the
second and third excitation pulses. For short delays, there is
a strong correlation between the frequency at which the OH/
OD stretch vibration is excited and the frequency at which
it is probed. This means that the spectral contour in the 2D
spectrum will be strongly elongated along the diagonal (for
which the excitation and the probing frequency are the same).
With increasing delay time between the excitation and the
probing, the frequency correlation decreases, which leads to
a broadening of the spectral contour along the off-diagonal
direction. Eventually this contour acquires a circular shape,
indicating that the correlation between the excitation and
probing frequencies is completely lost. This change in shape
is directly connected to the time dependence of the FFCF.
Different methods have been developed to relate the time-
dependent spectral contour to the FFCF, involving the time
dependences of the nodal slope,151 the dynamic line width,152
the ellipticity,153 and the inverse of the center line slope.154
2DIR experiments on HDO:H2O and HDO:D2O have been
performed by the Fayer152,155,156 and Tokmakoff151,157,158
groups, respectively. The former experiments were analyzed
within the context of the dynamic line width. This is the
width of the spectral hole as a function of the time delay
between the excitation and the probing process. The dynamic
line width can in principle be obtained by taking a cross
section of the spectral contour along the probing frequency
for a fixed excitation frequency. However, the 2D spectrum
will contain not only a spectral contour due to the bleaching
of the fundamental V ) 0 f 1 transition but also a contour
associated with the excited state V ) 1 f 2 absorption. This
latter contour is red-shifted along the probing frequency axis
by the anharmonicity of the stretch vibration. As the two
contours overlap, the projection of the bleaching signal along
the probing frequency axis has to be corrected for the
contribution of the excited state absorption. In Figure 8 the
dynamic line width is presented, for HDO:H2O, as a function
of the delay between the excitation and probing pulses.156
The dynamical line width is clearly observed to broaden with
increasing time delay, closely following the dynamics of the
FFCF. It is also observed that the initial line width is quite
large, indicating that the OD absorption line contains a
significant homogeneous line broadening component, in
agreement with the results of two-pulse photon echo and
spectral hole-burning studies. The FFCF extracted from these
experiments shows a long-time decay of 1.4 ps.156 These
experiments155 also showed an interesting frequency depen-
dence: at short times (∼100 fs) the dynamic line width
1506 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 8. Dynamic line width as a function of the waiting time
Tw for HDO:H2O.156 Experimental results are the points with error
bars, and the solid line going through the points results from a
triexponential fit to the FFCF. Also shown are simulation results
from the TIP4P (top line), SPC/E (middle line), and SPC/FQ
(bottom line) models. The dashed line at 147 cm-1 is the long-
time asymptotic line width. Reproduced with permission from ref
156. Copyright 2004 American Institute of Physics.
depends on excitation frequency, being wider on the blue
side of the line (than at the line center). This demonstrates
that molecules on the blue side undergo more rapid spectral
diffusion. This was interpreted as resulting from the less
constrained environments for water molecules with weaker
H-bonds.
An example of the 2DIR line shapes for HDO:D2O151,157,158
is shown in Figure 9.158 These experiments were analyzed
with nodal slope151 and spectral contour ellipticity157 metrics.
The spectral shape changes with increasing waiting time were
consistent with the FFCF, as deduced from the earlier peak-
shift experiments from the same group.105,140 These 2DIR
line shapes also show rapid and large frequency changes in
the blue wing. From the experimental results and computer
simulation and theoretical modeling, Tokmakoff and co-
workers argue that the rapid spectral diffusion on the blue
side indicates that molecules with broken H-bonds are
intrinsically unstable; thus, their existence is only “fleeting”,
lasting less than 150 fs. Most of these broken H-bonds result
from excursions from local H-bonded minima, while others
occur at the transition state of rapid H-bond switching events.
This then has broader implications for the nature of liquid
water itself; the authors conclude that “continuum” rather
than “mixture” models are more appropriate.151
Figure 9. Experimental (top) and simulated (bottom) absorptive 2DIR heterodyne-detected photon-echo spectra of the stretch vibration of
HDO:D2O for several waiting times τ2.158 Reproduced with permission from ref 158. Copyright 2006 American Institute of Physics.
Bakker and Skinner
3.2. Theoretical Approaches to Spectral Diffusion
Observables
A number of calculations related to spectral diffusion and
associated observables in HDO:D2O or HDO:H2O have
appeared in recent years.98-106,111,112,115-117,140,151,152,155,156,158-160
All of these studies proceed with the approach outlined in
section 2.2: a classical MD simulation of rigid water
molecules is run, and one method or another (see section
2.2) is used to determine the OH/OD stretch transition
frequency for a given configuration of the classical variables
(rotations and translations of all the molecules). This
produces a trajectory of the fluctuating transition frequency.
This information can then be used at different levels of
theory to compare with ultrafast observables. At the simplest
level one assumes the frequency fluctuations are Gaussian,
which is equivalent to making a second-cumulant trunca-
tion,122 in which case all information from the frequency
trajectory is contained within the FFCF. In addition, it is
then possible to obtain the FFCF from different nonlin-
ear experiments. Several such calculations of the
FFCF98-106,111,112,115-117,140,151,158-160 give results that are in
qualitative agreement with that extracted from experiment,
showing a rapid ∼50-100 fs initial decay, a small oscillation
with a period of about 150 fs, and a slower decay of about
0.5-1 ps. The long-time decay is somewhat faster than in
experiment (∼1.4 ps), which has been attributed to deficien-
cies of the nonpolarizable simulation models used for the
MD simulation. A follow-up study160 using the fluctuating-
charge simulation model of Rick et al.161 found a slower long-
time decay of the FFCF of ∼1.5 ps, in closer agreement with
experiment. This model, however, does not show the
experimentally observed oscillation. These calculations
confirm the interpretations described earlier, that the short-
time decay corresponds to inertial motion along the H-bond
stretch coordinate and the oscillation corresponds to under-
damped motion near the minimum of the same coordinate.
The decay of the FFCF at times on the order of 1 ps occurs
on the same time scale as the making and breaking of
H-bonds and has been attributed to this process,99,100,103,104
although Tokmakoff and co-workers have argued that the
long-time decay of the FFCF is due to more collective
structural relaxation.105,106,140
Within the Gaussian approximation one can then use the
FFCF to calculate nonlinear response functions and hence
ultrafast observables.122 However, it has come to be appreci-
Structure and Dynamics in Liquid Water
Figure 10. Calculated116 (for six different simulation models) and
experimental105 echo-peak shifts as a function of waiting time T
for HDO:D2O. Reproduced with permission from ref 116. Copyright
2007 Elsevier.
ated theoretically that for water the frequency fluctuations
are not particularly Gaussian.98,101-104,162 This was confirmed
experimentally by the asymmetric 2DIR line shapes.151,152,155-158
One consequence of this is that it then becomes impossible
to extract the FFCF from nonlinear experiments. One can
avoid this approximation theoretically by calculating the
nonlinear response functions directly from the frequency
trajectory.101,102,115-117,151,158,159 One such calculation102 for the
echo-peak shift of HDO:D2O actually preceded the experi-
ments of Fecko et al.,105 showing qualitative agreement,
including the underdamped oscillation. Similar calculations
for 2DIR spectra have also appeared115-117,151,158 and have
been used, for example, to support arguments for the
“fleeting” nature of non-H-bonded configurations (see Figure
9, bottom panel).151,158
At the next level of theory one can include non-Condon
effects,114-116 by determining the trajectory of the fluctuating
dipole moment as in section 2.2 and including that in the
calculation of nonlinear response functions. These non-
Condon effects (occurring because the magnitude of the
vibrational transition dipole depends on bath coordinates such
as H-bonding distances) are moderately important for the
IR and Raman line shapes in water114 and somewhat more
important for nonlinear experiments.115 In particular, for
pump-probe, spectral hole-burning, and heterodyned photon-
echo experiments, the signal goes like the fourth power of
the transition dipole, whereas for homodyned photon echoes
it goes like the eighth power.
With the large variety of linear and nonlinear vibrational
spectroscopy experimental results available for liquid water,
and the presence of presumably reasonably reliable frequency
and transition dipole maps, one wonders whether experiment
can be used to assess the reliability of classical simulation
models for water. Thus, in a combined theoretical and
experimental study,116 spectroscopic observables were cal-
culated for six different popular simulation models; we
included non-Condon effects, non-Gaussian frequency fluc-
tuations, vibrational relaxation, rotations, and convolutions
over the experimental pulse envelopes, and the results were
compared with line shape, echo-peak shift, and 2DIR
experiments (nodal slope and dynamic line width metrics).
In Figure 10 we show experimental105 and calculated (for
the six different simulation models) echo-peak shifts for
HDO:D2O. One sees that some models produce the correct
initial peak shift, some produce the short-time oscillation,
and some produce the correct time constant for the long-
Chemical Reviews, 2010, Vol. 110, No. 3 1507
Figure 11. Calculated116 (for six different simulation models) and
experimental156 dynamic line widths as a function of waiting time
T for HDO:H2O. Reproduced with permission from ref 116.
Copyright 2007 Elsevier.
time decay, but none do all three! In Figure 11 we show
experimental156 and calculated dynamic line widths for HDO:
H2O. In this case one sees that one simulation model (SPC/
FQ161) is in good agreement with experiment, while the rest
are not. Thus, the conclusions are, in fact, not very
conclusive: that all of the simulation models gave results in
qualitative agreement with the experimental results; for
individual experiments some models are definitely better than
others; and no model performed well for all observables.116
Overall, the nonpolarizable SPC/E model113 gave perhaps
the best agreement with experiment, although the authors’
final conclusion is that “the endorsement of the SPC/E model
is also not meant to diminish the importance of polarizability,
which is certainly expected to be essential when dealing with
inhomogeneous systems (such as the liquid/vapor interface)
or aqueous solution”.116
4. Vibrational Spectroscopy as a Probe of
Rotational Dynamics
Molecular reorientation in liquid water has been studied
with several experimental techniques, such as NMR,26-31
dielectric relaxation,32 THz absorption spectroscopy,39,40
optical Kerr effect,163 and quasi-elastic neutron scattering.38
These experiments are related to the first- and second-rank
Legendre polynomial orientational TCFs Cn(t) )
〈Pn(û(t) · û(0))〉, with n ) 1 and 2, respectively, for different
molecule-fixed unit vectors û. With NMR it was found that
the correlation time (total integral) of the second-rank TCF
(hereafter called the anisotropy correlation function or ACF)
for the intramolecular H-H vector is approximately 2.5
ps.27-30,164,165 Experiments on isotopic mixtures allow for the
measurement of correlation times of other (OH and out-of-
plane) molecule-fixed unit vectors, the latter of which turns
out to be about 25% shorter, showing that the rotational
motion of water is not isotropic.26 With dielectric relaxation
and THz absorption, fluctuations of the total dipole moment
of the sample are probed, yielding a Debye time of ∼8 ps
for room temperature liquid water.32,39,40 It is difficult to make
a straightforward comparison between these numbers for
several reasons: (1) the NMR values are correlation times,
as described above, while the Debye value is a “relaxation”
time (characterizing long-time exponential decay). If the
decay of a TCF is nonexponential at short times, these two
times will be different. (2) NMR times describe rotational
properties of single molecules, while Debye relaxation is a
collective process. (3) NMR times are for the second-rank
1508 Chemical Reviews, 2010, Vol. 110, No. 3
TCF, while Debye relaxation is for the first-rank TCF. (4)
NMR times are for three different molecule-fixed unit
vectors, none of which is in the direction of the molecule’s
dipole (which is probed by Debye relaxation).
Ultrafast IR spectroscopy can measure the second-rank
orientational dynamics of the OH bond vector of individual
molecules.133-141 One advantage of the IR experiment is that
the full ACF can be measured (whereas the NMR experiment
measures only the integrated correlation time). In addition,
through excitation and probing with different light frequen-
cies, IR experiments have the possibility to measure the
reorientation of different subensembles.
The measurement of the orientational dynamics of water
via the excitation of the OH/OD stretch vibrations relies on
the fact that the rotation of the molecule changes the direction
of the vibrational transition dipole moment. This method does
not work if there are other processes contributing to the
change of the direction of the transition dipole moment, such
as intramolecular and/or intermolecular resonant (Förster)
energy transfer between the vibrations. As a result, it is
unfortunately not possible to measure the orientational
dynamics of the molecules in pure liquid H2O.49 Conse-
quently, the reported femtosecond mid-IR measurements of
the orientational dynamics of water always refer to HDO
molecules dissolved in either normal or heavy water.133-144
In both cases the dynamics of a hydroxyl group (either OH
or OD) embedded in a network of isotopically distinct
hydroxyl groups are studied. Studying the orientational
dynamics of the stretch vibration of HDO:D2O has the
advantage that it is easier to resolve dynamical inhomoge-
neities, as the OH stretch vibration of HDO:D2O is more
strongly inhomogeneously broadened than the OD vibration
of HDO:H2O.147,152 On the other hand, studying the OD
vibration of HDO:H2O has the advantage that the lifetime
of the OD vibration is more than two times longer than that
of the OH vibration, thereby allowing the measurement of
the orientational dynamics of the OD group over a signifi-
cantly longer time interval.
4.1. Polarization-Resolved Pump-Probe
Spectroscopy
The orientational dynamics of the OH/OD stretch vibra-
tions of HDO dissolved in D2O/H2O have been studied with
polarization-resolved pump-probe spectroscopy.43,133-141 The
required femtosecond mid-IR laser pulses at wavelengths of
∼3 µm (=3300 cm-1) or ∼4 µm (=2500 cm-1) are obtained
via similar nonlinear frequency conversion processes, as were
briefly described in the previous section. The probe pulses
are often obtained by splitting off a small fraction of the
pump pulse.133,134,138-140 In some experiments, independently
tunable probe pulses are generated by a separate sequence
of nonlinear frequency-conversion processes.43,135-137
The rate of molecular reorientation of the water molecules
is studied by measuring the time dependence of the anisot-
ropy of the excitation of the OH/OD stretch vibration. In
most experiments, the polarization of the pump is rotated
before the sample at 45° with respect to the probe polariza-
tion with a λ/2 plate. After the sample, the polarization
components of the probe parallel and perpendicular to the pump
polarization are alternatingly chosen using a polarizer.134,135,137-141
In one study133 the experiment is performed with a probe
pulse that has a polarization that is parallel to the pump
polarization, and a probe pulse with a polarization at the
magic angle (54.7°) with respect to that of the pump.
Bakker and Skinner
The signals measured are the ratios of the transmitted
probe and reference beams: T| ) I|/Iref, T⊥ ) I⊥/Iref, and Tiso
) Iiso/Iref, where Iiso represents the signal measured by a probe
pulse with its polarization at the magic angle. To determine
the pump-induced changes T|, T ⊥, and Tiso, the signals
are alternatingly measured with and without the pump pulse
present using a 500 Hz chopper in the pump beam. The
signals are used to construct the absorption changes ∆R| )
ln[T|/T0], ∆R⊥ ) ln[T⊥/T0], and ∆Riso ) ln[Tiso/T0], where
T0 represents the transmission signal in the absence of the
pump. From ∆R| and ∆R⊥, the rotational anisotropy can be
constructed:
∆R| - ∆R⊥ ∆R| - ∆R⊥
R(t) ) ) (3)
∆R| + 2∆R⊥ 3∆Riso
The anisotropy can also be obtained from ∆R| and ∆Riso:133
∆R| - ∆Riso
R(t) ) (4)
2∆Riso
The denominator of eqs 3 and 4 is not affected by the
reorientation. Hence, isotropic effects such as vibrational
relaxation are divided out. Thus, R(t) is directly related to
the ACF:166
2 2
R(t) ) C2(t) ) 〈P2(cos θ(t))〉 (5)
5 5
where P2(x) is the second Legendre polynomial and θ(t) is
the angle between the OH/OD bond vectors at time 0 and
time t.
4.1.1. Isotropic Absorption Changes
In Figure 12 the isotropic absorption changes are shown
as a function of the delay between pump and probe for three
different probe frequencies.43 At probe frequencies of 3400/
2500 cm-1, the signal is dominated by a decaying bleach,
whereas at 3170/2380 cm-1 the signal consists of a decaying
induced absorption. In between these two frequencies, more
Figure 12. Absorption change as a function of delay between pump
and probe for a pump pulse at 3400/2500 cm-1 (upper/lower panel)
exciting the OH/OD stretch vibration of HDO dissolved in D2O/
H2O.43 Shown are transients for probe frequencies of 3170/2380
cm-1 (circles), 3250/2430 cm-1 (squares), and 3400/2480 cm-1
(triangles). Reproduced with permission from ref 43. Copyright
2008 American Chemical Society.
Structure and Dynamics in Liquid Water Chemical Reviews, 2010, Vol. 110, No. 3 1509

complicated dynamics are observed, especially for the OD

vibration. For a probe frequency of 2430 cm-1, a bleaching
signal is observed that shows a decay followed by a rerise
to a constant (bleached) end level. The latter ingrowing part
of the signal reflects the temperature increase of the sample
in the focus of the pump beam. The fact that the signal turns
over shows that the temperature rise takes place on a slower
time scale than the relaxation of the stretch vibration, which
implies that the vibrational relaxation proceeds via a non-
thermal intermediate state. With MD simulations it was found
that the relaxation of the vibration of HDO:D2O is dominated
by energy transfer to the overtone of the bending mode.78,80,167
This overtone subsequently rapidly relaxes to the first excited
state of the bending mode. Hence, the intermediate level is Figure 13. Anisotropy of the OD stretch vibration of HDO
likely formed by the first and/or second excited state of the dissolved in H2O as a function of delay between pump and probe
bending mode of the HDO molecule. for a pump frequency of 2500 cm-1 and probe frequencies of 2500
At longer delays (>3 ps) transmission changes are observed cm-1 (circles), 2550 cm-1 (squares), and 2600 cm-1 (triangles).43
Reproduced with permission from ref 43. Copyright 2008 American
that persist over the experimentally accessible time range Chemical Society.
(500 ps) and that have the character of a bleach on the red
side of the spectrum and of an induced absorption on the
that of the OH group, as its moment of inertia is almost twice
blue side. These persistent transient spectra reflect the
as large. However, there exist two regimes for the reorienta-
temperature rise in the sample that results from the absorption
tion. On very short time scales (<200 fs), the OH/OD group
and thermalization of the energy of the pump pulse. A rise
can show a limited free rotation (libration) while keeping
in temperature induces a blue shift and a decrease in cross
the donated H-bond intact. This regime will be discussed in
section of the OH/OD vibrations.133,138,139 The thermal signal
the following subsection. On longer time scales, which are
observed at large delays is isotropic, meaning that the
of relevance here, the orientational dynamics of the groups
magnitude of this signal is the same for ∆R| and ∆R⊥. Hence,
require rearrangements of the water network.108,170 In this
to obtain the anisotropy of the excited OH/OD vibration,
regime the orientational dynamics are no longer determined
the time-dependent thermalization signal has to be subtracted
by the moment of inertia but instead are governed by the
from ∆R| and ∆R⊥. To determine the delay dependence of
relative motions of the water molecules and, in particular,
the thermalization signal, the spectral responses of the OH
by the dynamics of H-bond breaking and reformation. A
and the OD vibration are fitted to a relaxation model.133,139,168
good measure for the translational mobility of the water
In this model the relaxation proceeds with time constant τ1
molecules is provided by the value of the viscosity. Using
) 1/k1 to an intermediate state. This latter state relaxes with
the viscosities of H2O and D2O (0.9 mPa s and 1.1 mPa s,
time constant τ/ ) 1/k/ to the final state, in which the energy
respectively), the reorientation times of the OH and OD are
has become thermal. The two time constants τ1 and τ/ are
estimated to show a ratio of 0.8, which is very similar to the
treated as global fit parameters, and the cross sections of
measured ratio of the reorientation time constants. It should
the transitions between the vibrational levels are allowed to
be noted that this finding does not imply that the orientational
vary over the spectrum. The resulting fits are represented
motion of the water molecules is diffusive. In fact, in a recent
by the solid lines in Figure 12. It is seen that the model
molecular dynamics study it was shown that the reorientation
provides an accurate description of the data, which implies
involves fast and large angular jumps between different
that spectral diffusion effects do not play an important role
H-bond configurations.170 The scaling with the viscosity thus
after a delay of 0.4 ps for pump pulses that are resonant
likely follows from the fact that the rate-limiting steps of
with the center of the absorption bands. For the OH stretch
the reorientation are formed by (translational) molecular
vibration of HDO dissolved in D2O, τ1 ) 0.7 ( 0.1 ps and
motions, of which the rate is well characterized by the
τ/ ) 0.6 ( 0.1 ps. For the OD stretch vibration of HDO
viscosity.171
dissolved in H2O, τ1 ) 1.8 ( 0.2 ps and τ/ ) 0.9 ( 0.1 ps.
The measured molecular reorientation times of 2.5/3 ps
4.1.2. Long-Time Results: Collective Orientational in H2O/D2O agree reasonably well with the results obtained
Dynamics with other techniques. NMR studies arrive at a reorientation
time of the water molecule of about 2.35-2.5 ps in liquid
Figure 13 shows the anisotropy of the OD stretch vibration H2O at 298 K27,30,31 and 2.4-2.9 ps in liquid D2O at 298
of HDO dissolved in H2O measured as a function of delay K.26,30 Note that the values from Jonas et al.30 are obtained
for a pump frequency of 2500 cm-1 and three different probe by interpolation using an Arrhenius expression, from the
frequencies. The anisotropy signals shown are obtained by measured results at 283 K and 303 K. Here again one should
correcting the measured ∆R| and ∆R⊥ for the thermalization remember that, for the IR pump-probe experiments and the
signal as determined from the isotropic data. An exponential deuterium quadrupolar relaxation experiments in D2O,26 the
fit to the data shown in Figure 13 yields a time constant of relevant molecule-fixed unit vector is the OH/OD bond
2.5 ( 0.2 ps at all probe frequencies, in agreement with vector, while in the proton NMR experiments on H2O it is
earlier results.139,141,169 the H-H intramolecular unit vector.27,30 Therefore, in this
The reorientation time constant of the OD vibration of case we can make a direct comparison between the two D2O
HDO in H2O (2.5 ( 0.2 ps) is observed to be somewhat experiments (and even here one should keep in mind that
shorter than that of the OH vibration of HDO in D2O (3 ( the IR experiments actually measure the reorientation of
0.3 ps).135-140 At first sight it may seem surprising that the HOD in D2O). One must also emphasize again that the NMR
orientational relaxation of the OD group proceeds faster than experiments measure the integrated correlation time, whereas
1510 Chemical Reviews, 2010, Vol. 110, No. 3
the IR experiments measure the long-time decay. If the ACF
is not exponential, these two are not the same. Both
experiments and simulation (see below) show, in fact, that
the ACF has a very fast drop at short times, whose amplitude
is about 10-20% of the initial value, meaning that the NMR
time (2.4-2.9 ps) should be about 10-20% shorter than the
IR time (3 ps); this is indeed observed to be the case for
D2O.
In comparing the results of femtosecond pump-probe
and NMR experiments with the results of dielectric
relaxation studies and THz absorption, it should be
emphasized again that these techniques measure different
orientational TCFs. Femtosecond pump-probe and NMR
experiments probe the (second-rank) ACF, whereas τD as
measured in dielectric relaxation and THz absorption
spectroscopy is related to the first-rank TCF. And again
we mention that for dielectric and THz spectroscopy the
relevant unit vector is different from that in both the NMR
and IR experiments (in this case it is along the molecular
dipole moment). In dielectric relaxation studies of liquid
water, a main relaxation component with a time constant
τD of 8.3 ps was found.32 Similar values were found in THz
spectroscopic studies of H2O and D2O:39,40 at room temper-
ature the Debye times τD of the slow component were
determined to be 8.5 ps for H2O and 10 ps for D2O. To arrive
at the time constant τ1 of the decay of the first-rank TCF,
the values of τD have to be corrected for collective effects.
This correction is not without ambiguity, but using the
approach proposed by Wallqvist and Berne,172 one arrives
at values of τ1 of 7.6 and 9 ps for H2O and D2O, respectively.
When comparing these values with the second-rank time
constants (2.5 and 3 ps, respectively), one finds ratios that
are close to 3. The ratio between the first- and second-rank
decay times τ1 and τ2 is determined by the nature of the
reorientation mechanism. In the case of pure (small step)
rotational diffusion τ1 ) 3τ2, but if the reorientation takes
places via, for example, jump diffusion, this ratio can be
lower.170 In addition to the main component with τD,
dielectric relaxation and THz absorption studies report a
weaker and much faster component with a time constant on
the order of 100 fs.39,40 The origin of this fast component
will be discussed in the following subsection.
4.1.3. Ultrashort-Time Results: Librational Motions
In Figure 14 the anisotropy of the OH stretch vibration
of HDO:D2O is shown as a function of delay time up to
1.5 ps.140,173 This measurement is performed with mid-IR
pulses with a pulse duration of 45 fs. The figure clearly
Figure 14. Anisotropy of the OH stretch vibration of HDO
dissolved in D2O as a function of delay,173 measured with ultrashort
(45 fs), broad-band (400 cm-1) pump and probe pulses with a
central frequency of 3400 cm-1. The anisotropy shows a biexpo-
nential decay with a fast component in the first 200 fs. Reproduced
with permission from ref 173. Copyright 2004 American Physical
Society.
Bakker and Skinner
shows that the anisotropy decay takes place on two
distinctly different time scales. The anisotropy shows a
rapid partial decay in the first 100 fs. This decay is
interpreted as being due to librational (hindered rotational)
motion. Here it should be noted that in these experiments140,173
the signals measured show little contribution from coherent
artifacts because the sample used is a free-flowing water
jet. The fast decay is followed by the much slower
dominant decay component with a time constant of ∼3
ps.135-137
The rotational motion observed at short times is,
presumably, hindered by the much more slowly rearrang-
ing H-bond network. Earlier in the review we discussed
how molecules in different H-bonding environments
absorb at different frequencies and, in particular, that
molecules on the red side of the band have weaker
H-bonds than molecules on the blue side. It follows, then,
that the short-time dynamics might well be frequency-
dependent. Such experiments can be performed with
relatively long pulses by pumping and probing at different
frequencies in the band.43,134,136,137 Alternatively, one can
use a shorter broad-band probe pulse and then frequency-
resolve the signal.43 Finally, one can use shorter pump and
probe pulses, whose bandwidths span the absorption line and
then frequency-resolve the transient absorption.141
The first experiments showing a frequency-dependent
anisotropy134,136 used the narrow-band pump-probe tech-
nique and were performed on HDO:D2O. In these two
studies, pump and probe frequencies were chosen to be
the same, but the frequency of the pair of pulses was
varied throughout the band. The anisotropy decay was
distinctly frequency-dependent for times up to 1.5 ps, with
faster decays for bluer frequencies, but the decay times
for longer times were identical.136 An interpretation100 of
the experiment was that on the blue side of the line the
restraining potential due to H-bonding is weaker, and so the
angular excursions at short times will be greater (see below).
At times longer than the spectral diffusion time, frequency
memory is lost, and so all members of the ensemble decay
at the same rate.100 Related two-color (different pump and
probe frequencies) experiments on the same system by Gallot
et al.137 showed similar results.
Figure 15. C2(100 fs) versus OD stretch (probe) frequency for
HDO:H2O.141 The dashed line is a linear fit to the data (filled
circles). The open circles represent calculations from MD simula-
tions (see section 4.2). Reproduced with permission from ref 141.
Copyright 2008 National Academy of Sciences.
Structure and Dynamics in Liquid Water
Figure 16. Anisotropy of the OD stretch vibration of HDO
dissolved in H2O as a function of delay,43 for the same pump and
probe frequency of 2500 cm-1 (circles), 2550 cm-1 (squares), 2600
cm-1 (triangles), and 2650 cm-1 (diamonds). Reproduced with
permission from ref 43. Copyright 2008 American Chemical
Society.
As mentioned above, better time and frequency resolu-
tion can be achieved with shorter pump and probe pulses,
and then frequency-resolving the transient absorption.141
Fayer and co-workers studied the HDO:H2O system using
this approach, finding that the value of R(t) at t ) 200 fs
varied by about 10%, depending on probe frequency.
Considering that the ultrafast inertial component is over by
about 100 fs,140,173 they then extrapolated their results
backward to 100 fs and plotted the value of C2(100 fs) as a
function of probe frequency; the results are shown in Figure
15. One sees that the initial drop of the ACF varies from
about 10% to about 20% as one goes from the center of the
band to the blue edge. One interesting aspect of this work
Figure 17. Anisotropy as a function of frequency at delays of 0.1 ps (circles), 0.2 ps (squares), 0.5 ps (triangles), 1 ps (diamonds), and
2 ps (stars). Shown are results obtained with a pump frequency of 2450 cm-1 (upper panel), 2550 cm-1 (middle panel), and 2650 cm-1
(lower panel). Part a shows the experimental results (solid lines are guides to the eye), and part b presents calculated results.43
Chemical Reviews, 2010, Vol. 110, No. 3 1511
involves the temperature dependence of this frequency-
dependent initial drop. At 65 °C the drop varies from 8% to
32%, while at 1 °C the drop is almost constant at about 15%.
This last feature is interpreted as indicating that motion just
above the freezing point is more collective, and so the
strength of the angular potential is not a single-molecule
property, explaining the lack of a frequency dependence.141
4.1.4. Short-Time Results: Jumping Molecules
Narrow-band pump/broad-band probe experiments, re-
cently performed by Bakker et al.,43 provide separate pump
and probe frequency resolution, albeit at the expense of some
time resolution (due to the relatively long pump pulse). These
experiments can shed light on the mechanism of molecular
reorientation. For example, results43 for identical pump and
probe frequencies, on the HDO:H2O system, are shown in
Figure 16. One sees that the measured anisotropy shows an
additional relatively fast component if the OD absorption
band is both excited and probed in the blue wing. It should
be noted that this additional faster decay cannot be caused
by librations. The partial decay of the anisotropy due to
librations is nearly complete after 100 fs,108,140 whereas the
dynamics observed in the blue wing of the stretch vibration
are much slower, having a time constant of ∼700 fs.
Results for different pump and probe frequencies43 are
shown in Figure 17. When the OD stretch vibration is
pumped close to its central frequency (upper panels), the
anisotropy is nearly the same at all probe frequencies, except
in the frequency region where the bleaching changes into
an induced absorption. In this frequency region the signal
results from the competition of these two signals, which leads
1512 Chemical Reviews, 2010, Vol. 110, No. 3
to an erratic behavior of the anisotropy. With increasing
delay, the anisotropy decay shows the same decay dynamics
at all probe frequencies. When the pump frequency is tuned
to the blue wing of the absorption spectrum (lower panels
of Figure 17), the anisotropy is observed to become strongly
frequency dependent in the first few picoseconds. An
interesting observation is that the anisotropy in the center
and the red wing is significantly lower than 0.4, already at
a delay of 0.2 ps. This observation shows that, directly after
the excitation, the signals in the center and in the red wing
contain a significant contribution of water molecules that
have reoriented.
The low value of the anisotropy in the center and the red
wing following excitation in the blue wing is not due to a high
intrinsic reorientation rate of molecules absorbing in the center
and the red wing. The upper panels of Figure 17 clearly show
that water molecules absorbing in the center and the red wing
in fact show a slow reorientation. Therefore, the large fraction
of reoriented molecules in the center and the red wing following
excitation in the blue wing must be due to excited molecules
that reorient while rapidly changing their frequencies from the
excited blue wing to the center and the red wing of the
absorption band. This finding closely agrees with the molecular
jump model for reorientation that was developed by Laage and
Hynes based on MD simulations.170 In this model, the
reorientation involves the breaking of the old H-bond and
the formation of a new H-bond, which leads to a large and
abrupt change in the frequency of the OD vibration. If the
pump pulse is tuned to the blue wing of the absorption band,
there will be few molecules directly excited in the center
and the red wing of the absorption band. As a result, the
relative contribution to the signal of molecules that have
reoriented and undergone a large frequency change will be
relatively large at these frequencies, leading to a low
anisotropy already at early delays. A significant part of the
frequency changes takes place within 100 fs, as no initial
fast anisotropy decay could be resolved in the center and
the red wing of the absorption band. In the blue wing of the
absorption band, the number of directly excited molecules
is large and the relatiVe contribution of molecules that have
reoriented and sampled all possible frequencies in the
absorption band will be small. Hence, in the blue wing the
initial anisotropy is high. Descriptions of model calculations,
shown in the right panels of Figure 17, as well as more
detailed comparison between these pump-probe results and
the predictions of the Laage/Hynes picture,108,170,171 can be
found in the recent Feature Article by Bakker et al.43
4.2. Theoretical Approaches to Rotational
Anisotropy Observables
The above experimental results show a number of intrigu-
ing features for which molecular interpretations are desirable.
These include the nonexponential decay of C2(t), especially
the very rapid initial drop and the frequency dependence of
its amplitude, and the eventual frequency-independent long-
time decay.
As discussed above, small-amplitude librations can occur
without large-scale rearrangement of the H-bond network, which
leads naturally to the wobbling-in-a-cone model.43,108,166,169,171
The idea is that a restraining angular potential due to
H-bonding restricts the angular motion at short times to
explore angles within a cone of semiangle θo. At such time
τ when the cone is fully explored, the value of the ACF drops
to166
Bakker and Skinner
[ 21 cos θ (1 + cos θ )]
2
C2(τ) ) o o (6)
Thus, within this model, from the initial drop of the ACF
one can determine the cone semiangle θo.43,108,169 For
example, if one assumes that the cone is fully explored at a
time τ such that C2(τ) ) 0.9, θo ) 15°, while if C2(τ) )
0.75, θo ) 24°, etc. Estimates for τ range between about
100 and 200 fs,108,140,141 and estimates for the cone semiangle
range from about 15° to 35°.43,108 In passing, we note that
experimental estimates of cone semiangles appear to be
somewhat larger for the OH ACF than for the OD ACF.43
This is perhaps due to larger librational quantum effects19-24
in the former case (where the relevant reduced mass is
presumably lighter, and so the angular extent of the wave
function is greater).
Moilanen et al. have generalized the wobbling-in-a-cone
model to a harmonic cone model.141 Thus, instead of free
angular motion within a cone, there is motion within an
angular harmonic potential characterized by frequency ω.
Within this model the angular drop in the ACF, at such time
τ that the distribution within the angular potential has attained
thermal equilibrium, can be related to the cone frequency.
Taking τ to be 100 fs, this analysis yields frequencies ω
from 320 to 450 cm-1 as the probe frequency goes from the
blue edge to the middle of the band (for HDO:H2O). Smaller
(larger) estimates for τ would lead to smaller (larger) initial
drops, which would lead to higher (lower) frequencies. In
any case, the estimated cone frequency at the middle of the
band is in qualitative agreement with the estimated librational
frequency (585 cm-1) for HDO:H2O.141 Note that the lowest
frequency of 320 cm-1 has a period of about 100 fs, which
is only consistent with equilibration within 100 fs if the
libration is strongly damped.
C2(t) for various molecule-fixed unit vectors has been
calculated for many simulation models of liquid water.
Relevant to this section are recent calculations using the
TIP4P174 or SPC/E113 simulation models for the OH bond
unit vector.26,97,100,108,140,141 All of the results show a fast decay
within 50 fs, followed by a recurrence at between 60 and 80
fs and then followed by an approximately exponential decay
with a time constant on the order of 1.5 ps (for the TIP4P
model)100 or about 2.5 ps (for the SPC/E model).26,97,108,140,141
The recurrence has been interpreted as arising from under-
damped librational motion. This recurrence has not been
observed experimentally, possibly due to difficulties arising
from coherent artifacts at such short times.
The frequency dependence of the initial drop was first
studied theoretically by Lawrence and Skinner.100 They
calculated C2(t) for HDO:D2O using the TIP4P water
simulation model, but only those molecules whose OH
stretch frequency was in a specified frequency window at t
) 0 contributed to the average. This was intended to
represent the frequency selection in a narrow-band pump-
probe experiment. The results showed a frequency-dependent
initial drop, followed by the above-described recurrence and
then followed by a frequency-independent exponential decay
with a time constant of about 1.4 ps. The agreement with
the experimental data available at the time136 was only very
qualitative. Lawrence and Skinner argued that at longer
timessthose longer than the spectral diffusion timesthe
decay rate of C2(t) becomes frequency independent, as
molecules have lost memory of their initial frequency.100
In the earlier experiments136 the OH stretches were pumped
and probed at the same frequency. Thus, the calculation of
Structure and Dynamics in Liquid Water Chemical Reviews, 2010, Vol. 110, No. 3 1513

C2(t) should have included only those molecules whose OH

stretch frequency was in a specified window at time 0 and
time t. This shortcoming was corrected in a more recent paper
by Laage and Hynes,108 where they calculated C2(t) for
different frequencies for HDO:D2O using the SPC/E simula-
tion model. Their results are similar to the earlier results,100
showing a frequency-dependent initial drop, a recurrence,
and frequency-independent exponential decay at longer times
(now with a time constant of about 2.5 ps,171 due to the more
realistic and slower SPC/E simulation model). Laage and
Hynes108,171 extended the separation-of-time-scale argument
between spectral diffusion and rotation invoked earlier.100
They first showed (from simulations of the SPC/E model)
that the transition state for molecular rotation involves a
bifurcated H-bond (described more fully below) and that the
rate-limiting step for the rotation process involves the
concerted arrival and departure of H-bond partners.170 They Figure 18. Calculated C2(t) for HDO:H2O, as a function of probe
then showed that the bifurcated transition state is extremely frequency.97 Reproduced with permission from ref 97. Copyright
2009 John Wiley and Sons.
short-lived and will contribute negligibly to the absorption,
even in the blue wing of the absorption band.108 Therefore,
exciting water molecules on the blue edge of the band does
not specifically select molecules close to the transition state.
However, the probability to evolve to the transition state
within 125 fs is about five times larger in the blue wing of
the absorption band than in the center.171 This finding is
consistent with the experimental observation of a dependence
of the anisotropy decay on excitation frequency for delay
times up to ∼1 ps.43 After spectral diffusion is complete (∼1
ps), this frequency dependence will vanish, and the further
decay of the ACF will be independent of pump and probe
frequencies.
In the more recent studies in HDO:H2O by Moilanen et
al.141 as described above, where the broad-band pump at t
) 0 excites all molecules, the transient absorption probed at
a time t later is frequency-resolved. These authors calculated
C2(t) as done by Lawrence and Skinner,100 averaging over
only those molecules whose OH stretches are in (in this case
very narrow) frequency windows at t ) 0. Note that since
trajectories from equilibrium simulations were used in this
analysis and the classical equations of motion are fully time-
reversible, this is equivalent to averaging over only those Figure 19. Experimental141 and theoretical97 initial drops of the
molecules whose OH stretches are in the same frequency rotational anisotropy correlation function, for HDO:H2O, as a
windows at time t, which is consistent with the (broad-band function of probe frequency, for three different temperatures.
pump/frequency-resolved probe) experiments.141 Theoretical Reproduced with permission from ref 97. Copyright 2009 John
results97 for a range of probe frequencies from the center of Wiley and Sons.
the band to the blue edge are shown in Figure 18, for the compared with experiment141 in Figure 19. In this case the
SPC/E simulation model and using a recently developed theoretical results are in relatively poor agreement with
frequency map,97,117 as described in section 2.2. One sees experiment, showing neither the increased probe frequency
clearly the probe frequency dependence of the initial drop, dependence of the initial drop at higher temperature (65 °C)
followed by the recurrence described earlier. The values at nor the nearly flat frequency dependence at 1 °C. Presumably
100 fs were reported in the paper by Moilanen et al.141 and this is due, at least in part, to the inability of the SPC/E
are also shown in Figure 15. Note that the recurrence time model, which was parametrized at room temperature, to
(and hence the librational frequency) is dependent on the perform well at other temperatures. In particular, the freezing
probe frequency, ranging from about 56 fs at the center of point of the SPC/E model is 215 K.175 Moilanen et al. argue
the band to about 68 fs at the blue edge. These recurrence that the flat frequency dependence is a signature of collective
times correspond to librational frequencies of 595-490 cm-1. dynamics very close to the freezing point, which would not
Note that the former is in good agreement with the estimated be expected to be demonstrated by the SPC/E model at 1
librational frequency (585 cm-1) of HDO:H2O141 and that °C (since this is so far above the freezing point for the
the values of these librational frequencies and their trend model).
with probe frequency are in qualitative agreement with those The last topic for this theoretical discussion involves the
deduced from the harmonic cone model described above.141 molecular mechanism of water rotation at long times, which,
The temperature dependences of the probe-frequency- as discussed above at length, is characterized by a decay
dependent initial drops of C2(100 fs) have also been time of C2(t) for the OH bond unit vector of HDO:D2O of
calculated theoretically for HDO:H2O,97 and the results are about 3 ps and for the OD bond unit vector in HDO:H2O of
1514 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 20. Schematic picture of the reorientation of water.108 The rotating water molecule has oxygen and hydrogen atoms labeled with
O* and H*, respectively. Reproduced with permission from ref 108. Copyright 2006 Elsevier.
about 2.5 ps. In two recent papers,170,171 Laage and Hynes
analyze trajectories from a simulation of SPC/E water,
showing first of all that the picture based on small-step
rotational diffusion is not correct. Rather, they find that
rotations occur through large-amplitude angular jumps,
involving a concerted H-bond switching event with a
transition state characterized by a bifurcated H-bond. A
schematic picture108 of the molecular mechanism is shown
in Figure 20. This picture is consistent with the interpretation
of 2DIR spectra and associated calculations by Loparo et
al.158 As mentioned above, analysis of trajectories within the
context of this extended jump model sheds further light on
the frequency-dependent anisotropy decay at short times and
its frequency independence at later times.108 In addition, as
described above, the 2DIR experiments of Tokmakoff and
co-workers151,158 and the two-color pump-probe anisotropy
experiments of Bakker et al.43 provide experimental support
for this rotational mechanism.
5. Conclusions and Outlook
We have reviewed the contributions that vibrational
spectroscopy in the OH/OD stretch regions of HDO:D2O/
HDO:H2O has made to our understanding of the structure
and dynamics of liquid water. We have focused on three
classes of experimental observables: line shapes (IR and
Raman), observables that probe spectral diffusion (hole-
burning, integrated echoes, and 2DIR), and observables that
probe molecular rotation (polarization-resolved pump-probe).
In each case we have summarized the relevant experimental
and theoretical studies.
Considering first the Raman line shapes, theoretical
analysis indicates that the shoulder on the blue side is due
to HOD molecules lacking an H-bond to the H/D atom for
HDO:D2O/HDO:H2O, respectively.114,117 Most of these bro-
ken H-bonded configurations are inherently unstable and,
hence, are transient, and a small fraction of them correspond
to bifurcated H-bond transition states.151,158,171 In any case,
most simulation models (together with an H-bond definition)
indicate that 10-15% of the time a given H atom is not
H-bonded. Thus, these broken H-bond configurations make
an important contribution to the equilibrium structure of
water. Theoretical analysis indicates that, in the H-bonding
region of the line shapes, information about structure is not
readily available, since, apparently, there is a poor correlation
between OH/OD stretch frequency and H-bonding class.117,126
The spectral diffusion observables provide quite direct
information about the frequency-frequency time-correlation
function. Integrated three-pulse echo-peak shift experi-
ments140 show that the correlation function has an initial
inertial time decay within 50 fs, a recurrence indicative of
an underdamped oscillation at about 180 fs, and a long-time
decay with a time constant of about 1.4 ps. Molecular
dynamics simulations and other experiments support the idea
that the underdamped oscillation is due to the intermolecular
Bakker and Skinner
H-bond stretch. The initial decay on the 50 fs or shorter time
scale is due to small-amplitude local rearrangements. The
long-time decay time indicates the time scale over which
the transition frequency becomes uncorrelated. Since the
frequency is a functional of the nuclear positions of the
surrounding molecules, this time scale is related to that for
structural relaxation. Some have advanced the notion that
this structural relaxation comes from making and breaking
H-bonds with the HDO molecule,99,100,103,104 while others have
argued that the relaxation has a more long-range collective
origin.106 It is certainly possible that both are true;104 for
example, much of the relaxation on the 1 or 2 ps time scale
may be due to making and breaking local hydrogen bonds,
while the residual relaxation, possibly on slightly longer time
scales, may be due to more collective processes. 2DIR
experiments permit a more nuanced picture of spectral
diffusion.155,158 In particular, the finding that spectral diffusion
is faster on the blue side is consistent with the intuitive idea
that molecules with weak or broken H-bonds sample con-
figuration space, and therefore frequency space, faster.
The pump-probe anisotropy experiments provide an
equally intimate view of molecular rotation. The second-
rank rotational time-correlation function also has an ultrafast
decay within 100 fs, followed by a long-time exponential
decay of 2.5-3 ps.43,169,173 The former is found to depend
on the frequencies of the pumping and probing light.43,136,141
In general, on the blue side of the line the initial drop is
largest, indicating more extensive orientational excursions.
The amplitude of the drop can be related to the angle of an
accessible cone or to the librational frequency of a harmonic
cone potential. Pumping or probing on the blue side leads
to larger cone angles or smaller cone librational frequencies.
The subsequent exponential decay is frequency independent
and characterizes the rotational relaxation time of the entire
ensemble.100,108 Since many rotational “attempts” are made
before an actual rotational jump occurs, and during these
attempts local, possibly including H-bond, rearrangements
occur, which produce spectral diffusion, it is clear that the
spectral diffusion time is faster than the rotational time. To
rephrase this slightly differently, many configurational
changes can cause spectral diffusion, but only a specific
configuration involving the concerted departure and arrival
of H-bond partners can lead to rotation. In this sense, spectral
diffusion is a necessary but not sufficient condition for
rotation.
Both sets of experiments lead to the following picture of
motion in liquid water. Short-time dynamics is restricted,
on the average, by both an angular potential and a transla-
tional potential, both due to H-bonding. For times less than
50 fs, a molecule can make small-amplitude angular and
translational excursions near the bottom of this restraining
potential, which lead to rapid but only partial loss of
frequency and angular correlation. Theory and simulation
predict that on slightly longer time scales these excursions
lead to underdamped oscillations in both angular and
Structure and Dynamics in Liquid Water
translational coordinatessthe former is characterized by a
librational frequency, and the latter is characterized by the
hydrogen-bond stretch. Only the latter has been observed
experimentally. At longer times molecules rotate and break
hydrogen bonds, and as a result of recent 2DIR155,158 and
frequency-dependent pump-probe experiments141,43 and
simulations and theory,100,106,108,141,170 the precise way in
which this happens is now much clearer. Molecules rotate
by way of a concerted H-bond switching event, going
through a transition state with a weak bifurcated H-bond.170
Thus, the molecule makes relatively infrequent but fast and
large-amplitude angular jumps. The rate-limiting step is the
approach of an H-bond accepting molecule. We anticipate
that further frequency-dependent pump-probe experiments
will undoubtedly refine the above picture of dynamics in
liquid water.
Where do we go from here? Many of these same
techniques can be, and in fact have already been, used to
explore the structure and dynamics of water in other phases
and in heterogeneous situations. For example, one can use
vibrational spectroscopy to probe structure and dynamics in
the many crystalline ice phases, in amorphous solid water,
at the surfaces of solid and liquid water, and in the near-
critical regime. One can use vibrational spectroscopy to probe
the structure and dynamics of water in aqueous solutions of
ionic and nonpolar species and in more heterogeneous
situations such as in reverse micelles, in or near polymer or
bilayer membranes, and in concentrated solutions of bio-
molecules. As in the case of liquid water, the exquisite
sensitivity of OH stretch vibrational frequencies to local
environments, coupled with the excellent time and frequency
resolution of modern ultrafast vibrational spectroscopy, make
this an excellent technique for unravelling complicated
structural and dynamical issues in these fascinating and
important systems.
6. Acknowledgments
The authors thank their research groups and collaborators
for essential discussions about water over the past decade.
H.J.B. thanks the Stichting Fundamenteel Onderzoek der
Materie (Foundation for Fundamental Research on Matter)
and the Nederlandse Organisatie voor Wetenschappelijk
Onderzoek (Netherlands Organization for the Advancement
of Research) for financial support. J.L.S. thanks Yu-Shan
Lin for preparing several figures and for a critical reading
of the manuscript and the National Science Foundation for
support of this work through Grants CHE-0132538, CHE-
0446666, and CHE-0750307.
7. References
(1) Ball, P. Life’s Matrix: A Biography of Water; Farrar, Straus, and
Giroux: New York, 1999.
(2) Ball, P. Nature 2008, 452, 291.
(3) Ball, P. Chem. ReV. 2008, 108, 74.
(4) Reed, A. E.; Weinhold, F.; Curtiss, L. A.; Pochatko, D. J. J. Chem.
Phys. 1986, 84, 5687.
(5) Ojamäe, L.; Hermansson, K. J. Phys. Chem. 1994, 98, 4271.
(6) Pedulla, J. M.; Vila, F.; Jordan, K. D. J. Chem. Phys. 1996, 105,
11091.
(7) Xantheas, S. S. Chem. Phys. 2000, 258, 225.
(8) Kumar, R.; Skinner, J. L. J. Phys. Chem. B 2008, 112, 8311.
(9) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88, 899.
(10) Guillot, B. J. Mol. Liq. 2002, 101, 219.
(11) Car, R.; Parrinello, M. Phys. ReV. Lett. 1985, 5, 2471.
(12) Izvekov, S.; Voth, G. A. J. Chem. Phys. 2002, 116, 10372.
(13) Lee, H.-S.; Tuckerman, M. E. J. Chem. Phys. 2007, 126, 164501.
Chemical Reviews, 2010, Vol. 110, No. 3 1515
(14) Schwegler, E.; Grossman, J. C.; Gygi, F.; Galli, G. J. Chem. Phys.
2004, 121, 5400.
(15) Mantz, Y. A.; Chen, B.; Martyna, G. J. J. Phys. Chem. B 2006, 110,
3540.
(16) Todorova, T.; Seitsonen, A. P.; Hutter, J.; Kuo, I.-F. W.; Mundy,
C. J. J. Phys. Chem. B 2006, 110, 3685.
(17) VandeVondele, J.; Mohamed, F.; Krack, M.; Hutter, J.; Sprik, M.;
Parrinello, M. J. Chem. Phys. 2005, 122, 014515.
(18) Bukowski, R.; Szalewicz, K.; Groenenboom, G. C.; van der Avoird,
A. Science 2007, 315, 1249.
(19) Fanourgakis, G. S.; Xantheas, S. S. J. Phys. Chem. A 2006, 110,
4100.
(20) Kuharski, R. A.; Rossky, P. J. J. Chem. Phys. 1985, 82, 5164.
(21) Poulsen, J. A.; Nyman, G.; Rossky, P. J. J. Chem. Theory Comput.
2006, 2, 1482.
(22) Stern, H. A.; Rittner, F.; Berne, B. J.; Friesner, R. A. J. Chem. Phys.
2001, 115, 2237.
(23) Paesani, F.; Iuchi, S.; Voth, G. A. J. Chem. Phys. 2007, 127, 074506.
(24) de la Peña, L. H.; Kusalik, P. G. J. Chem. Phys. 2004, 121, 5992.
(25) Paesani, F.; Voth, G. A. J. Phys. Chem. B 2009, 113, 5702.
(26) Ropp, J.; Lawrence, C.; Farrar, T. C.; Skinner, J. L. J. Am. Chem.
Soc. 2001, 123, 8047.
(27) Smith, D. W. G.; Powles, J. G. Mol. Phys. 1966, 10, 451.
(28) Godralla, B. C.; Zeidler, M. D. Mol. Phys. 1986, 59, 817.
(29) Hardy, E. H.; Zygar, A.; Zeidler, M. D.; Holz, M.; Sacher, F. D.
J. Chem. Phys. 2001, 114, 3174.
(30) Jonas, J.; DeFries, T.; Wilbur, D. J. J. Chem. Phys. 1976, 65, 582.
(31) Lang, E.; Lüdemann, H.-D. J. Chem. Phys. 1977, 67, 718.
(32) Barthel, J.; Bachhuber, K.; Buchner, R.; Hetzenauer, H. Chem. Phys.
Lett. 1990, 165, 369.
(33) Head-Gordon, T.; Hura, G. Chem. ReV. 2002, 102, 2651.
(34) Wernet, P.; Nordlund, D.; Bergmann, U.; Cavalleri, M.; Odelius, M.;
Ogasawara, H.; Näslund, L. Å.; Hirsch, T. K.; Ojamäe, L.; Glatzel,
P.; Pettersson, L. G. M.; Nilsson, A. Science 2004, 304, 995.
(35) Smith, J. D.; Cappa, C. D.; Wilson, K. R.; Messer, B. M.; Cohen,
R. C.; Saykally, R. J. Science 2004, 306, 851.
(36) Soper, A. K. Chem. Phys. 2000, 258, 121.
(37) Laage, D. J. Phys. Chem. B 2009, 113, 2684.
(38) Teixeira, J.; Bellissent-Funnel, M. C.; Chen, S. H.; Dianoux, A. J.
Phys. ReV. A 1985, 31, 1913.
(39) Rønne, C.; Thrane, L.; Åstrand, P.-O.; Wallqvist, A.; Mikkelsen,
K. V.; Keiding, S. R. J. Chem. Phys. 1997, 107, 5319.
(40) Rønne, C.; Åstrand, P.-O.; Keiding, S. R. Phys. ReV. Lett. 1999, 82,
2888.
(41) Prendergast, D.; Galli, G. Phys. ReV. Lett. 2006, 96, 215502.
(42) Nibbering, E. T. J.; Elsaesser, T. Chem. ReV. 2004, 104, 1887.
(43) Bakker, H. J.; Rezus, Y. L. A.; Timmer, R. L. A. J. Phys. Chem. A
2008, 112, 11523.
(44) Zheng, J.; Kwak, K.; Fayer, M. D. Acc. Chem. Res. 2007, 40, 75.
(45) Cho, M. Chem. ReV. 2008, 108, 1331.
(46) Auer, B.; Skinner, J. L. J. Chem. Phys. 2007, 127, 104105.
(47) Buch, V.; Tarbuck, T.; Richmond, G. L.; Groenzin, H.; Li, I.; Schultz,
M. J. J. Chem. Phys. 2007, 127, 204710.
(48) Auer, B.; Skinner, J. L. J. Chem. Phys. 2008, 128, 224511.
(49) Woutersen, S.; Bakker, H. J. Nature 1999, 402, 507.
(50) Cowan, M. L.; Bruner, B. D.; Huse, N.; Dwyer, J. R.; Chugh, B.;
Nibbering, E. T. J.; Elsaesser, T.; Miller, R. D. J. Nature 2005, 434,
199.
(51) Kraemer, D.; Cowan, M. L.; Paarmann, A.; Huse, N.; Nibbering,
E. T. J.; Elsaesser, T.; Miller, R. J. D. Proc. Natl. Acad. Sci. U.S.A.
2008, 105, 437.
(52) Torii, H. J. Phys. Chem. A 2006, 110, 9469.
(53) Paarmann, A.; Hayashi, T.; Mukamel, S.; Miller, R. J. D. J. Chem.
Phys. 2008, 128, 191103.
(54) Paarmann, A.; Hayashi, T.; Mukamel, S.; Miller, R. J. D. J. Chem.
Phys. 2009, 130, 204110.
(55) Lock, A. J.; Bakker, H. J. J. Chem. Phys. 2002, 117, 1708.
(56) Pakoulev, A.; Wang, Z.; Dlott, D. D. Chem. Phys. Lett. 2003, 371,
594.
(57) Pakoulev, A.; Wang, Z.; Pang, Y.; Dlott, D. D. Chem. Phys. Lett.
2003, 380, 404.
(58) Wang, Z.; Pakoulev, A.; Pang, Y.; Dlott, D. D. Chem. Phys. Lett.
2003, 378, 281.
(59) Wang, Z.; Pakoulev, A.; Pang, Y.; Dlott, D. D. J. Phys. Chem. A
2004, 108, 9054.
(60) Wang, Z.; Pang, Y.; Dlott, D. D. Chem. Phys. Lett. 2004, 397, 40.
(61) Bakker, H. J.; Lock, A. J.; Madsen, D. Chem. Phys. Lett. 2004, 384,
236.
(62) Bakker, H. J.; Lock, A. J.; Madsen, D. Chem. Phys. Lett. 2004, 385,
329.
(63) Pakoulev, A.; Wang, Z.; Pang, Y.; Dlott, D. D. Chem. Phys. Lett.
2004, 385, 332.
1516 Chemical Reviews, 2010, Vol. 110, No. 3
(64) Huse, N.; Ashihara, S.; Nibbering, E. T. J.; Elsaesser, T. Chem. Phys.
Lett. 2005, 404, 389.
(65) Lindner, J.; Vöhringer, P.; Pshenichnikov, M. S.; Cringus, D.;
Wiersma, D. A.; Mostovoy, M. Chem. Phys. Lett. 2006, 421, 329.
(66) Ashihara, S.; Huse, N.; Espagne, A.; Nibbering, E. T. J.; Elsaesser,
T. Chem. Phys. Lett. 2006, 424, 66.
(67) Ashihara, S.; Huse, N.; Espagne, A.; Nibbering, E. T. J.; Elsaesser,
T. J. Phys. Chem. A 2007, 111, 743.
(68) Wang, Z.; Pang, Y.; Dlott, D. D. J. Phys. Chem. A 2007, 111, 3196.
(69) Roychowdhury, S.; Bagchi, B. Chem. Phys. 2008, 343, 76.
(70) Mallik, B. S.; Semparithi, A.; Chandra, A. J. Phys. Chem. A 2008,
112, 5104.
(71) Ingrosso, F.; Rey, R.; Elsaesser, T.; Hynes, J. T. J. Phys. Chem. A
2009, 113, 6657.
(72) Nicodemus, R. A.; Tokmakoff, A. Chem. Phys. Lett. 2007, 449, 130.
(73) Rey, R.; Møller, K. B.; Hynes, J. T. Chem. ReV. 2004, 104, 1915.
(74) Woutersen, S.; Emmerichs, U.; Nienhuys, H.-K.; Bakker, H. J. Phys.
ReV. Lett. 1998, 81, 1106.
(75) Gale, G. M.; Gallot, G.; Lascoux, N. Chem. Phys. Lett. 1999, 311,
123.
(76) Deàk, J. C.; Rhea, S. T.; Iwaki, L. K.; Dlott, D. D. J. Phys. Chem.
A 2000, 104, 4866.
(77) Schwarzer, D.; Lindner, J.; Vöhringer, P. J. Chem. Phys. 2005, 123,
161105.
(78) Rey, R.; Hynes, J. T. J. Chem. Phys. 1996, 104 (6), 2356.
(79) Lawrence, C. P.; Skinner, J. L. J. Chem. Phys. 2002, 117, 5827.
(80) Lawrence, C. P.; Skinner, J. L. J. Chem. Phys. 2003, 119, 1623.
(81) Lawrence, C. P.; Skinner, J. L. J. Chem. Phys. 2003, 119, 3840.
(82) Kropman, M. F.; Nienhuys, H.-K.; Woutersen, S.; Bakker, H. J. J.
Phys. Chem. A 2001, 105, 4622.
(83) Larsen, O. F. A.; Woutersen, S. J. Chem. Phys. 2004, 121, 12143.
(84) Bodis, P.; Larsen, O. F. A.; Woutersen, S. J. Phys. Chem. A 2005,
109, 5303.
(85) Brubach, J. B.; Mermet, A.; Filabozzi, A.; Gerschel, A.; Roy, R.
J. Chem. Phys. 2005, 122, 184509.
(86) Schmidt, D. A.; Miki, K. J. Phys. Chem. A 2007, 111, 10119.
(87) Graener, H.; Seifert, G.; Laubereau, A. Phys. ReV. Lett. 1991, 66,
2092.
(88) Laenen, R.; Rauscher, C.; Laubereau, A. Phys. ReV. Lett. 1998, 80,
2622.
(89) Laenen, R.; Rauscher, C.; Laubereau, A. J. Phys. Chem. B 1998,
102, 9304.
(90) Laenen, R.; Simeonidis, K.; Laubereau, A. J. Phys. Chem. B 2002,
106, 408.
(91) Gordon, R. G. AdV. Magn. Reson. 1968, 3, 1.
(92) Skinner, J. L. Mol. Phys. 2008, 106, 2245.
(93) Ahlborn, H.; Ji, X.; Space, B.; Moore, P. B. J. Chem. Phys. 1999,
111, 10622.
(94) Martı́, J.; Guàrdia, E.; Padró, J. A. J. Chem. Phys. 1994, 101, 1083.
(95) Mankoo, P. K.; Keyes, T. J. Chem. Phys. 2008, 129, 034504.
(96) Silvestrelli, P. L.; Bernasconi, M.; Parrinello, M. Chem. Phys. Lett.
1997, 277, 478.
(97) Skinner, J. L.; Auer, B. M.; Lin, Y.-S. AdV. Chem. Phys. 2009, 142,
59.
(98) Lawrence, C. P.; Skinner, J. L. J. Chem. Phys. 2002, 117, 8847.
(99) Lawrence, C. P.; Skinner, J. L. Chem. Phys. Lett. 2003, 369, 472.
(100) Lawrence, C. P.; Skinner, J. L. J. Chem. Phys. 2003, 118, 264.
(101) Piryatinski, A.; Lawrence, C. P.; Skinner, J. L. J. Chem. Phys. 2003,
118, 9664.
(102) Piryatinski, A.; Lawrence, C. P.; Skinner, J. L. J. Chem. Phys. 2003,
118, 9672.
(103) Rey, R.; Møller, K. B.; Hynes, J. T. J. Phys. Chem. A 2002, 106,
11993.
(104) Møller, K. B.; Rey, R.; Hynes, J. T. J. Phys. Chem. A 2004, 108,
1275.
(105) Fecko, C. J.; Eaves, J. D.; Loparo, J. J.; Tokmakoff, A.; Geissler,
P. L. Science 2003, 301, 1698.
(106) Eaves, J. D.; Tokmakoff, A.; Geissler, P. L. J. Phys. Chem. A 2005,
109, 9424.
(107) Harder, E.; Eaves, J. D.; Tokmakoff, A.; Berne, B. J. Proc. Natl.
Acad. Sci. U.S.A. 2005, 102, 11611.
(108) Laage, D.; Hynes, J. T. Chem. Phys. Lett. 2006, 433, 80.
(109) Diraison, M.; Guissani, Y.; Leicknam, J.-C.; Bratos, S. Chem. Phys.
Lett. 1996, 258, 348.
(110) Hermansson, K.; Knuts, S.; Lindgren, J. J. Chem. Phys. 1991, 95,
7486.
(111) Hayashi, T.; l. C. Jansen, T.; Zhuang, W.; Mukamel, S. J. Phys. Chem.
A 2005, 109, 64.
(112) Corcelli, S. A.; Lawrence, C. P.; Skinner, J. L. J. Chem. Phys. 2004,
120, 8107.
(113) Berendsen, H. J. C.; Grigera, J. R.; Straatsma, T. P. J. Phys. Chem.
1987, 91, 6269.
(114) Corcelli, S. A.; Skinner, J. L. J. Phys. Chem. A 2005, 109, 6154.
Bakker and Skinner
(115) Schmidt, J. R.; Corcelli, S. A.; Skinner, J. L. J. Chem. Phys. 2005,
123, 044513.
(116) Schmidt, J. R.; Roberts, S. T.; Loparo, J. J.; Tokmakoff, A.; Fayer,
M. D.; Skinner, J. L. Chem. Phys. 2007, 341, 143.
(117) Auer, B.; Kumar, R.; Schmidt, J. R.; Skinner, J. L. Proc. Natl. Acad.
Sci. U.S.A. 2007, 104, 14215.
(118) Auer, B.; Skinner, J. L. J. Chem. Phys. 2008, 129, 214705.
(119) Whalley, E.; Klug, D. D. J. Chem. Phys. 1986, 84, 78.
(120) Hadz̆i, D.; Bratos, S. In The Hydrogen Bond; Schuster, P., Zundel,
G., Sandorfy, C., Eds.; Elsevier: Amsterdam, 1976; Vol. 2, Chapter
12.
(121) Pimentel, G. C.; McClellan, A. L. The Hydrogen Bond; W. H.
Freeman and Company: San Francisco, 1960.
(122) Mukamel, S. Principles of Nonlinear Optical Spectroscopy; Oxford:
New York, 1995.
(123) Kubo, R. AdV. Chem. Phys. 1969, 15, 101.
(124) Smith, J. D.; Cappa, C. D.; Wilson, K. R.; Cohen, R. C.; Geissler,
P. L.; Saykally, R. J. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 14171.
(125) Lin, Y.-S.; Auer, B. M.; Skinner, J. L. J. Chem. Phys. 2009, 131,
144511.
(126) Auer, B.; Skinner, J. L. Chem. Phys. Lett. 2009, 470, 13.
(127) Loparo, J. J.; Roberts, S. T.; Nicodemus, R. A.; Tokmakoff, A. Chem.
Phys. 2007, 341, 218.
(128) Kumar, R.; Schmidt, J. R.; Skinner, J. L. J. Chem. Phys. 2007, 126,
204107.
(129) Eisenberg, D.; Kauzmann, W. The Structure and Properties of Water;
Oxford University Press: New York, 1969.
(130) Jimenez, R.; Fleming, G. R.; Kumar, P. V.; Maroncelli, M. Nature
1994, 369, 471.
(131) Ohmine, I.; Saito, S. Acc. Chem. Res. 1999, 32, 741.
(132) Mikenda, W. J. Mol. Struct. 1986, 147, 1.
(133) Steinel, T.; Asbury, J. B.; Zheng, J.; Fayer, M. D. J. Phys. Chem. A
2004, 108, 10958.
(134) Woutersen, S.; Emmerichs, U.; Bakker, H. J. Science 1997, 278, 658.
(135) Nienhuis, H.-K.; van Santen, R. A.; Bakker, H. J. J. Chem. Phys.
2000, 112, 8487.
(136) Bakker, H. J.; Woutersen, S.; Nienhuys, H.-K. Chem. Phys. 2000,
258, 233.
(137) Gallot, G.; Bratos, S.; Pommeret, S.; Lascoux, N.; Leicknam, J.-C.;
Koziński, M.; Amir, W.; Gale, G. M. J. Chem. Phys. 2002, 117,
11301.
(138) Rezus, Y. L. A.; Bakker, H. J. J. Chem. Phys. 2006, 125, 144512.
(139) Rezus, Y. L. A.; Bakker, H. J. J. Chem. Phys. 2005, 123, 114502.
(140) Fecko, C. J.; Loparo, J. J.; Roberts, S. T.; Tokmakoff, A. J. Chem.
Phys. 2005, 122, 054506.
(141) Moilanen, D. E.; Fenn, E. E.; Lin, Y.-S.; Skinner, J. L.; Bagchi, B.;
Fayer, M. D. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 5295.
(142) Gale, G. M.; Gallot, G.; Hache, F.; Lascoux, N.; Bratos, S.; Leicknam,
J.-C. Phys. ReV. Lett. 1999, 82, 1068.
(143) Woutersen, S.; Bakker, H. J. Phys. ReV. Lett. 1999, 83, 2077.
(144) Bakker, H. J.; Nienhuys, H.-K.; Gallot, G.; Lascoux, N.; Gale, G. M.;
Leicknam, J.-C.; Bratos, S. J. Chem. Phys. 2002, 116, 2592.
(145) Cho, M.; Yu, J.-Y.; Joo, T.; Nagasawa, Y.; Passino, S. A.; Fleming,
G. R. J. Phys. Chem. 1996, 100, 11944.
(146) Piryatinski, A.; Skinner, J. L. J. Phys. Chem. B 2002, 106, 8055.
(147) Stenger, J.; Madsen, D.; Hamm, P.; Nibbering, E. T. J.; Elsaesser,
T. J. Phys. Chem. A 2002, 106, 2341.
(148) Yeremenko, S.; Pshenichnikov, M. S.; Wiersma, D. A. Phys. ReV. A
2006, 73, 021804.
(149) Stenger, J.; Madsen, D.; Hamm, P.; Nibbering, E. T. J.; Elsaesser,
T. Phys. ReV. Lett. 2001, 87, 027401.
(150) Yeremenko, S.; Pshenichnikov, M. S.; Wiersma, D. A. Chem. Phys.
Lett. 2003, 369, 107.
(151) Eaves, J. D.; Loparo, J. J.; Fecko, C. J.; Roberts, S. T.; Tokmakoff,
A.; Geissler, P. L. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 13019.
(152) Asbury, J. B.; Steinel, T.; Stromberg, C.; Corcelli, S. A.; Lawrence,
C. P.; Skinner, J. L.; Fayer, M. D. J. Phys. Chem. A 2004, 108, 1107.
(153) Roberts, S. T.; Loparo, J. J.; Tokmakoff, A. J. Chem. Phys. 2006,
125, 084502.
(154) Kwak, K.; Park, S.; Finkelstein, I. J.; Fayer, M. D. J. Chem. Phys.
2007, 127, 124503.
(155) Steinel, T.; Asbury, J. B.; Corcelli, S. A.; Lawrence, C. P.; Skinner,
J. L.; Fayer, M. D. Chem. Phys. Lett. 2004, 386, 295.
(156) Asbury, J. B.; Steinel, T.; Kwak, K.; Corcelli, S. A.; Lawrence, C. P.;
Skinner, J. L.; Fayer, M. D. J. Chem. Phys. 2004, 121, 12431.
(157) Loparo, J. J.; Roberts, S. T.; Tokmakoff, A. J. Chem. Phys. 2006,
125, 194521.
(158) Loparo, J. J.; Roberts, S. T.; Tokmakoff, A. J. Chem. Phys. 2006,
125, 194522.
(159) l. C. Jansen, T.; Hayashi, T.; Zhuang, W.; Mukamel, S. J. Chem.
Phys. 2005, 123, 114504.
(160) Corcelli, S. A.; Lawrence, C. P.; Asbury, J. B.; Steinel, T.; Fayer,
M. D.; Skinner, J. L. J. Chem. Phys. 2004, 121, 8897.
Structure and Dynamics in Liquid Water
(161) Rick, S. W.; Stuart, S. J.; Berne, B. J. J. Chem. Phys. 1994, 101,
6141.
(162) Garrett-Roe, S.; Hamm, P. J. Chem. Phys. 2008, 128, 104507.
(163) Winkler, K.; Lindner, J.; Bürsing, H.; Vöhringer, P. J. Chem. Phys.
2000, 113, 4674.
(164) Lankhorst, D.; Schriever, J.; Leyte, J. C. Ber. Bunsen-Ges. Phys.Chem.
1982, 86, 215.
(165) van der Maarel, J. R. C.; Lankhorst, D.; de Bleijser, J.; Leyte, J. C.
Chem. Phys. Lett. 1985, 122, 541.
(166) Lipari, G.; Szabo, A. Biophys. J. 1980, 30, 489.
(167) Kandratsenka, A.; Schroeder, J.; Schwarzer, D.; Vikhrenko, V. S.
J. Chem. Phys. 2009, 130, 174507.
(168) Timmer, R. L. A.; Bakker, H. J. J. Chem. Phys. 2007, 126, 154507.
Chemical Reviews, 2010, Vol. 110, No. 3 1517
(169) Piletic, I. R.; Moilanen, D. E.; Spry, D. B.; Levinger, N. E.; Fayer,
M. D. J. Phys. Chem. A 2006, 110, 4985.
(170) Laage, D.; Hynes, J. T. Science 2006, 311, 832.
(171) Laage, D.; Hynes, J. T. J. Phys. Chem. B 2008, 112, 14230.
(172) Wallqvist, A.; Berne, B. J. J. Phys. Chem. 1993, 97, 13841.
(173) Loparo, J. J.; Fecko, C. J.; Eaves, J. D.; Roberts, S. T.; Tokmakoff,
A. Phys. ReV. B 2004, 70, 180201.
(174) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.;
Klein, M. L. J. Chem. Phys. 1983, 79, 926.
(175) Vega, C.; Sanz, E.; Abascal, J. L. F. J. Chem. Phys. 2005, 122,
114507.
CR9001879
1518 Chem. Rev. 2010, 110, 1518–1563

Microfluidic Applications of Magnetic Particles for Biological Analysis and

Catalysis

Martin A. M. Gijs,* Frédéric Lacharme, and Ulrike Lehmann

Laboratory of Microsystems, Ecole Polytechnique Fédérale de Lausanne, 1015 Lausanne EPFL, Switzerland

Received May 14, 2009

Contents 7.1.2. Enzyme Reaction-Based Detection 1549

7.2. Magnetic Beads as Label for Detection 1551
1. Introduction 1518 7.2.1. Surface Coverage Measurement 1551
1.1. Microfluidic Systems 1518 7.2.2. Magnetic Properties Measurement 1552
1.2. Basic Characteristics of Magnetic Beads 1520 7.3. Agglutination Tests 1553
1.3. Scope of This Review 1521 8. Catalytic Applications 1554
2. Magnetic Beads 1521 8.1. Homogenizing Heterogeneous Catalysis Using 1554
2.1. Synthesis 1521 Magnetic Particles
2.2. Protection and Stabilization 1522 8.2. Transition Metal Catalysts 1554
2.3. Functionalization 1523 8.3. Biocatalysts 1555
3. Forces on Magnetic Beads 1524 8.4. Magnetic Particle-Based Protein Digestion 1556
3.1. Magnetic Force 1524 9. Conclusions and Outlook 1556
3.2. Viscous Drag Force 1525 10. Acknowledgments 1557
3.3. Electrical Interaction with the Microchannel 1525 11. References 1557
Wall
3.4. Other Forces 1526 1. Introduction
4. Magnetic Bead Manipulation 1526
4.1. Retention and Separation 1527 1.1. Microfluidic Systems
4.1.1. Systems with Macroscopic Magnets 1527
Since the introduction of the concept of Micro Total
4.1.2. Systems with Microscopic Magnets: 1528
Magnetic Templates Analysis System or µTAS in 1990,1 the interest in and impact
4.1.3. Systems Using Microelectromagnets 1529 of such devices, also called lab-on-a-chip or miniaturized
analysis systems, has grown explosively. Clinical diagnosis,
4.2. Magnetic Transport 1529
as practiced in blood screening laboratories, is for economic
4.3. Magnetic Beads as Labels for Detection 1531
reasons dominated by high-throughput, highly reliable, and
4.4. Enhancing the Bead-Flow Interaction and 1532 robust modular automated systems. Such systems generate
Mixing
a minimum of consumables per test indeed. However,
4.5. Magnetic Droplets 1534
microfluidics has shown to provide attractive solutions for
5. Magnetic Cell Manipulation 1534 many problems in chemical and biological analysis, espe-
5.1. Cell Types 1534 cially for in-field use or point-of-care testing. A detailed
5.1.1. Cells with Magnetic Character 1534 overview of the various fluidic operations in microfluidic
5.1.2. Nonmagnetic Cells Labeled with Magnetic 1535 systems, such as sample preparation, sample injection, sample
Beads manipulation, reaction, separation, and detection, published
5.2. Magnetophoretic Mobility of a Cell 1536 in the 1990-2008 period, was presented in a series of review
5.3. Cell Separation Methods 1536 articles by the group of A. Manz.2-4 Three of the most
5.4. Magnetic Cell Separation and Purification 1537 important advantages of using fluidic systems of reduced
5.5. Cell Biophysics 1542 dimension for analytical applications are (i) the possibility
6. Magnetic Nucleic Acid Assays 1543 of using only minute quantities of sample and reagents (down
6.1. DNA Capture and Purification 1543 to picoliters), as problems of fluidic connectors with large
6.2. DNA Processing 1544 dead volumes can be avoided for an integrated lab-on-a-
6.3. DNA Detection 1545 chip, (ii) comparatively fast reaction times, when molecular
6.4. Integrated DNA Analysis Systems Starting 1545 diffusion lengths are of the order of the microchannel
from Cells dimension, and (iii) a large surface-to-volume ratio offering
7. Magnetic Immunoassays 1547 an intrinsic compatibility between the use of microfluidic
7.1. Magnetic Beads as Substrate 1548 systems and surface-based assays. The long-range nature of
7.1.1. Direct Fluorescent Detection 1548 viscous flows and the reduced device dimensions imply that
the influence of boundaries or “channel walls” is very
* To whom correspondence should be addressed. Tel.: +41 21 693 67 34. significant for a microfluidic system. A variety of strategies
Fax: +41 21 693 59 50. E-mail: martin.gijs@epfl.ch. have been implemented to manipulate fluids by exploiting
10.1021/cr9001929  2010 American Chemical Society
Published on Web 12/04/2009
Microfluidic Applications of Magnetic Particles
Martin A. M. Gijs received his degree in physics in 1981 from the
Katholieke Universiteit Leuven, Belgium, and his Ph.D. degree in physics
at the same university in 1986. He joined the Philips Research Laboratories
in Eindhoven, The Netherlands, in 1987. Subsequently, he has worked
there on micro- and nanofabrication processes of high critical temperature
superconducting Josephson and tunnel junctions, the microfabrication of
microstructures in magnetic multilayers showing the giant magnetoresis-
tance effect, the design and realization of miniaturized motors for hard
disk applications, and the design and realization of planar transformers
for miniaturized power applications. He joined the Ecole Polytechnique
Fédérale de Lausanne (EPFL) in 1997. He presently is a professor in the
Institute of Microengineering, where he is responsible for the Microsystems
Technology Group. His main interests are in developing technologies for
novel magnetic devices, new microfabrication technologies for Microsys-
tems fabrication in general, and the development and use of microfluidics
for biomedical applications in particular. He is on the editorial board of
Microfluidics and Nanofluidics and the Journal of Micromechanics and
Microengineering. He has published over 160 papers in peer-reviewed
journals and holds over 20 patents.
Frédéric Lacharme was born in Bordeaux, Gironde, France in 1978. He
received his Master’s degree in Physical-Chemistry from the University
of Bordeaux 1 “Sciences and Technologies” in 2002. In 2003, he joined
the laboratory of Professor Martin Gijs at the Ecole Polytechnique Fédérale
de Lausanne, where he received his Ph.D. degree in 2008. His research
was focused on the development of analytical microfluidic systems using
magnetic nanoparticles. He is currently working on the development of
microchips for analytical and diagnostic applications.
boundary effects, among them electrokinetic effects, droplet
generation, acoustic streaming, and fluid-structure interac-
tions, topics that were also reviewed extensively.5-8
Multiple technologies for the realization of fluidic micro-
systems have been developed, as shown for example in the
book of Madou9 and the review of Reyes et al.10 The basic
microfabrication sequence of a fluidic chip usually involves
the patterning of a microchannel structure, a bonding
operation to seal the open channel, followed by surface
coating or functionalization steps. Glass and silicon have
Chemical Reviews, 2010, Vol. 110, No. 3 1519
Ulrike Lehmann was born in Rostock, Germany in 1978. She obtained
her Dipl-Ing. degree in electrical engineering, with a specialization on
MEMS, in 2003, at the Chemnitz Technical University, Chemnitz, Germany.
She subsequently joined the group of Prof. Gijs in the Laboratory for
Microsystems at the Ecole Polytechnique Fédérale de Lausanne to work
on her Ph.D. In 2008, she obtained her Ph.D. for her work on the
manipulation of magnetic microparticles in liquid phases for application in
biomedical systems, which included the magnetic manipulation of droplets
as well as the direct manipulation of single magnetic microparticles. Ulrike
Lehmann is currently working at Microsens S.A., where she is involved
in the development of environmental sensor probes based on electro-
chemical and resistive microsensors.
been the dominating material for the realization of microf-
luidic chips in the early years,11 but have been more and
more replaced by polymers. Still, borosilicate glass forms
an interesting option for more demanding microreactor
applications in harsh environments, for example, during
catalytic reactions at high temperatures. Fused silica is a
material with low autofluorescence in the ultraviolet region,
making it a preferred choice for high-resolution fluorescent
detection methods. However, microfabrication of glass9 can
be expensive due to the requirement of clean room-based
technologies such as deep plasma etching or hydrofluoric
acid-based wet etching. This explains the increasing interest
in the replication of microchannel master structures in an
elastomeric material like poly dimethylsiloxane (PDMS),12,13
which has become a preferred microfluidic technology for
the lab-on-a-chip research community. Also, high-throughput
polymer microfabrication techniques such as microinjection
molding and hot embossing14 have emerged for affordable
and disposable microfluidic analytical applications.
Parallel to the boom of microfluidic systems, nanomaterials
and nanoparticles have become a hot topic in research. When
brought into a microfluidic channel, nano- and microparticles
offer a relatively large specific surface for chemical binding.
Such particles are also called in general “beads” in literature,
independent of what their size is. A polymer colloid or
microsphere solution has a low viscosity as compared to
solutions having the same amount of solid, giving it special
properties. Also, such small particles can be advantageously
used as a “mobile substrate” in catalysis, for bioassays or
even for in vivo applications; they can be easily recovered
from a dispersion, reversibly redispersed, etc. Several reviews
on the preparation and use of polymer particles and polymer
colloids for medical, biological, and optical applications
exist.15,16 Moreover, the technologies for the realization of
nano- and microparticle handling systems, with a focus on
the integration of the latter with microfluidic systems, were
reviewed.17 With respect to open or “empty” microchannels,
microfluidic structures with packed beds of functionalized
beads or containing bead suspensions profit from an even
1520 Chemical Reviews, 2010, Vol. 110, No. 3 Gijs et al.

larger surface-to-volume ratio, an enhanced interaction of

reactive surfaces with fluids passing by, and an improved
recuperation of reaction products.

1.2. Basic Characteristics of Magnetic Beads

If a magnetic material is placed in a magnetic field of
strength H, the individual atomic moments in the material
contribute to its overall response, resulting in the magnetic
induction B ) µ0(H + M),where µ0 is the permeability of
free space, and the magnetization M ) m/V is the magnetic
moment per unit volume, where m is the magnetic moment
of a volume V of the material. The magnetic character of a
material depends on its atomic structure and temperature.
Materials may be conveniently classified in terms of their
volumetric magnetic susceptibility, χ, where M ) χH
Figure 1. (a) Schematic of a ferromagnetic microparticle contain-
describes the magnetization induced in a material by H. In ing either a single magnetic core or a core composed of an ensemble
SI units, χ is dimensionless, B is expressed in tesla (T), and of magnetic nanoparticles of size s. (b) Schematic magnetization
both M and H are expressed in A/m. curve of ferromagnetic particles with indication of the saturation
Magnetic beads have additional advantages beyond those magnetization Msat and the remnant magnetization Mrem. (c)
mentioned for nonmagnetic particles: having embedded Schematic of a superparamagnetic nanoparticle having a single
magnetic entities, they can be magnetically manipulated using magnetic nanoparticle of size s as a core. (d) Schematic magnetiza-
permanent magnets or electromagnets, independent of normal tion curve of superparamagnetic particles with indication of the
magnetic susceptibility χ and the saturation magnetization Msat.
microfluidic or biological processes. This additional degree
of freedom is the basis of a still improved exposure of the
functionalized bead surface to the surrounding liquid, due
to the increased relative motion of the bead with respect to
the fluid. It is important to note that a magnetic field gradient
is required to exert a translation force, because a uniform
field solely gives rise to a torque, but no translational action.
The magnetic force acting on a point-like magnetic dipole
or “magnetic moment” m in a magnetic induction B can be
written as a function of the derivative of the magnetic
induction:18-23 Figure 2. Schematic diagram of the dipole field of a magnetic
particle. When a superparamagnetic particle is placed in a magnetic
1 field H, the resulting moment m leads to a stray magnetic induction
Fm ) (m·∇)B (1) at each position vector r on the sensor surface, given by eq 3.
µ0
In case of superparamagnetic particles in a biological
As an example, for a constant moment m in the x-direction, medium, one can describe the moment at small fields by the
leading to m · ∇ ) mx((∂)/(∂x)), a force will be exerted on linear relation m ) Vµ0 M ) Vµ0∆χH, with M the
the moment, provided there is a magnetic field gradient in magnetization of the particle and ∆χ the difference in
the x-direction. Figure 1a is a schematic diagram of a magnetic susceptibility between the magnetic particle and
ferromagnetic microparticle, which can either have a homo- the surrounding liquid medium. Using the relation B ) µ0H,
geneous magnetic core or be composed of an ensemble of eq 1 for a superparamagnetic particle in the linear suscep-
primary magnetic nanoparticles that each have a dimension tibility regime becomes
s that is in the nanometer range. The magnetization measure-
ment of an ensemble of such microparticles results in a V∆χ
magnetization loop as shown in Figure 1b, which is Fm ) (B·∇)B (2)
µ0
characterized by a remnant magnetization Mrem close to the
saturation magnetization Msat. Such particles will be subjected
The magnetic moment of a superparamagnetic nanoparticle
to substantial forces when brought into a magnetic field
generally is smaller than that of a larger ferromagnetic
gradient, but suffer from magnetic clustering and agglomera-
microparticle. Hence, the magnetic force on a superpara-
tion effects due to the magnetic dipole interaction. Super-
magnetic particle will be smaller, which will result in slower
paramagnetic nanoparticles, as schematically shown in Figure
magnetic separation processes. However, the advantage of
1c, have zero magnetization in the absence of a magnetic
superparamagnetic particles is the possibility to simply
field, as their magnetic anisotropy energy (proportional to
“switch off” the magnetic effects by removing the magnetic
their magnetic volume V) is typically smaller than the thermal
induction field.
energy. They therefore essentially behave as nonmagnetic
Besides actuation of magnetic beads, the latter can also
particles in the absence of a magnetic field and are character-
be used as labels for detection by a magnetic field sensor.
ized by a magnetization curve as shown in Figure 1d. The
The magnetic dipole stray field of a point-like magnetic bead
slope at low fields is characterized by the magnetic suscep-
is schematically shown in Figure 2 and is given by
tibility χ. The susceptibility of a particle is also called

[ ]
effective susceptibility and is related to the intrinsic suscep- 1 m 3(m·r)r
tibility of the material χmat by χ ) χmat/(1 + Ndχmat), with Nd B(r) ) - + (3)
the demagnetization factor (1/3 for a spherical particle).24 4π r3 r5
Microfluidic Applications of Magnetic Particles
Typically, the sensor output will be proportional to the total
magnetic induction integrated over the sensor area. Another
consequence of eq 3 is that, when a magnetic bead is in the
magnetic induction field of other beads, the particles will
interact via the magnetic dipole interaction, and chain-like
or more complicated magnetic supraparticle structures (SPS)
can be formed.
1.3. Scope of This Review
Pankhurst et al.,25 Gijs,26 and Pamme27 reviewed the
applications of magnetic beads, with focus on the underlying
physics25 and microfluidic aspects.26,27 Here, we intend to
present a comprehensive overview of recent research on the
manipulation of magnetic beads in microfluidic systems and
their utilization for biological analysis and chemical catalysis.
We start with a discussion on the production of various types
of magnetic beads and their chemical/biological functional-
ization. We then discuss the main forces that act on magnetic
particles, that is, magnetic forces that provide the primary
means of actuation, viscous drag forces in the liquid that
oppose the magnetically induced bead motion, and electro-
static forces that can be responsible for agglomeration or
sticking of the particles to microchannel walls. Thereafter
we discuss the basic manipulation steps of magnetic particles:
retention, separation, mixing, and transport. We also discuss
the use of beads as magnetic detection labels, or contained
as actuators/substrates in discrete droplets for digital mi-
crofluidics. We then focus on the applications of magnetic
beads: the handling, detection, and separation of biological
cells, on-chip nucleic acid assays, and immunoassays.
Finally, the promising application area of magnetic micro-
and nanoparticles for chemical catalysis is discussed. Through-
out this Review, we try to compare the use of magnetic beads
in microfluidic systems with alternative solutions and point
out the advantages. We end this Review with an outlook
section on possible future developments.
2. Magnetic Beads
Magnetic bead suspensions or magnetic fluids are stable
dispersions of magnetic beads or encapsulated magnetic
nanoparticles in an organic or aqueous carrier medium.
Requirements of the bead matrix such as biocompatibility,
biodegradability, and stability in different media must be
combined with a preferable uniform size distribution and a
correct shape. The magnetic material content determines the
magnetic behavior of the beads and must be associated with
suitable measures of protection against corrosion. For
example, the primary magnetic nanoparticles (schematically
shown as the black dots on the left of Figure 1a and in Figure
1c) can be coated with polymers or surfactants, with silica
or carbon, or can be embedded in a chemically inert
protective matrix. Moreover, for use in specific target
applications, the bead surface needs to be functionalized
(hydrophilicity versus hydrophobicity, chemical functionality
at the surface, etc.) to allow covalent bonding or simple
aspecific adsorption of biomolecules (proteins, antibodies,
nucleic acids) or cells. Many types of magnetic beads are
commercially available nowadays and are usually tailor-made
for a specific final application. The synthesis, protection, and
functionalization of magnetic nanoparticles were reviewed
by Lu et al.28 and Horák et al.29 Also, the articles of
Landfester and Ramı́rez,30 Bergemann et al.,31 Grüttner et
al.,32,33 and Latham et al.34 present specific short review
Chemical Reviews, 2010, Vol. 110, No. 3 1521
sections on the synthesis and chemical modifications of
magnetic beads. Therefore, we only provide representative
examples of bead synthesis, protection, and functionalization
in this Review, rather than covering this vast field of literature
in a comprehensive way.
2.1. Synthesis
Reviews of the synthesis of inorganic nanoparticles in the
liquid phase (not limited to magnetic materials) were
presented by Grieve et al.,35 Trindade et al.,36 Murray et al.,37
and Lu et al.,28 while a review of the synthesis of such
particles from the vapor phase was presented by Swihart.38
The synthesis and applications of (nonmagnetic) polymer
micoparticles were reviewed by Kawaguchi.15 The synthesis
of magnetic beads is also a well-covered subject in patent
literature. Magnetic nanoparticles have been synthesized with
a number of different compositions that include iron oxides,
such as magnetite (Fe3O4) and maghemite (γ-Fe2O3),39,40 pure
metals,41,42 such as Fe and Co, or alloys,43,44 such as CoPt3
and FePt. Among the practiced magnetic nanoparticle
synthesis methods, coprecipitation, thermal decomposition,
microemulsion, and hydrothermal technologies are mentioned
here. To date, magnetic nanoparticles prepared from copre-
cipitation and thermal decomposition are the most widely
studied. Synthesis in microfluidic chips is an alternative
technique for very controlled realization of magnetic
particles.45-47
Coprecipitation
In early publications, magnetic bead solutions (“ferroflu-
ids”) were produced by grinding magnetite with long-chain
hydrocarbons and a grinding agent.48 Later, magnetic fluids
were produced by precipitating an aqueous Fe3+/Fe2+ solution
with a base, coating these particles with an adsorbed layer
of oleic acid, and then dispersing them in a nonaqueous
fluid.49 Both types of processes result in very small magnetic
particles with a surfactant coating in a nonaqueous liquid
carrier, in which the hydrophobic magnetite particles are
dispersed. However, a lot of applications of magnetic beads
rely on water as the continuous phase. A water-based
magnetic fluid was realized by conversion of iron products
to magnetic iron oxide in an aqueous medium under
controlled pH conditions.50 The experimental challenge in
the synthesis of Fe3O4 by coprecipitation is in the control of
the particle size and the achievement of a narrow particle
size distribution. To produce monodisperse iron oxide
magnetic nanoparticles in a coprecipitation reaction, a process
with a short burst of nucleation followed by a slow controlled
growth is necessary.
Thermal Decomposition
Monodisperse magnetic nanoparticles can be synthesized
through the thermal decomposition of organometallic com-
pounds in boiling organic solvents that contain stabilizing
surfactants.51 The ratios of the starting reagents, including
the organometallic compounds, surfactant, and solvent, are,
together with the reaction temperature and reaction time, the
main parameters for controlling the nanoparticle synthesis.
For example, nearly monodisperse iron oxide crystals with
sizes adjustable over a wide size range (3-50 nm) have been
prepared.52,53 The reaction system was composed of iron fatty
acid salts, fatty acids, hydrocarbon solvents, and activation
1522 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 3. (a) Three-dimensional superlattice of iron nanocubes
as deposited on a carbon-coated copper grid from a toluene
dispersion. Bar: 10 nm. (b) TEM photograph of dextran nanopar-
ticles; the dextran coating is not visible in the TEM picture, and
the particles have an irregular shape. (c) More regular polystyrene
particles with encapsulated magnetite nanoparticles. (d) SEM
photograph of irregular magnetic glass particles [obtained from the
MagNA Pure LC DNA isolation Kit from Roche Diagnostics
(Rotkreuz, Switzerland)]. (a) Reprinted with permission from ref
54. Copyright 2004 American Association for the Advancement
of Science. (b) Reprinted with permission from ref 33. Copyright
1999 Elsevier. (c) Reprinted with permission from ref 30. Copyright
2003 The Institute of Physics Publishing Ltd.
reagents. The thermal-decomposition method can also be
used to prepare metallic nanoparticles. A metallic ferromag-
net has a larger magnetization as compared to a magnetic
oxide; therefore, the magnetic force (eq 1) on such particles
will be larger. As an example, the reaction at 150 °C of the
metal-organic precursor Fe[N(SiMe3)2]2 with H2 in the
presence of a long-chain acid and a long-chain amine
produces monodisperse cubic nanoparticles that have edges
of 7 nm and are incorporated into extended superlattices54
(see Figure 3a).
Microemulsion Techniques
A microemulsion is a stable dispersion of two immiscible
liquids in which small-size droplets are stabilized by an
interfacial film of surfactant molecules. In water-in-oil
microemulsions, the aqueous phase forms droplets (1-50
nm diameter) in a continuous hydrocarbon phase. By mixing
two identical water-in-oil emulsions containing the desired
reactants, the droplets will collide, coalesce, and split and
induce the formation of precipitates. Adding a solvent like
ethanol to the microemulsion allows extraction of the
precipitate by filtering or centrifuging the mixture. Also, the
preparation of magnetizable polymer particles from aqueous
dispersions is known. The particles can be prepared by
dispersing magnetic elements in an organic phase containing
an organosoluble initiator and/or monomer(s), mixing the
dispersion with an aqueous solution made of water and
emulsifier, homogenizing the mixture to give droplets of
organic phase with (sub)micrometer size, and finally poly-
merizing the homogenized mixture after the addition of
monomer(s), if necessary.55 The distribution of the droplets
(and hence of the polymer particles) is a function of the
proportion of emulsifier present in the aqueous solution, and
the ratio of the organic to the aqueous phase. The size
distribution of the polymer particles obtained by this process
is generally wide. By applying a purification process to the
initial polydisperse crude emulsion, it is possible to obtain a
Gijs et al.
set of highly monodisperse samples with high magnetic
content (up to 70%).56 The purification method is analogous
in principle to a fractionated crystallization process.33,56,57
Hydrothermal Synthesis
Hydrothermal reactions with mixed oxides or hydroxides
of iron and other metals can be carried out in water under
supercritical conditions, that is, at temperatures around or
higher than 200 °C under a pressure higher than 14 MPa.
Under these conditions, water plays the role of a hydrolytic
reactant. The size and morphology of the reaction products
are controlled by the reaction time and temperature. How-
ever, hydrothermal reaction procedures are experimentally
demanding.
Synthesis in Microfluidic Chips
Microfluidic devices provide an alternative for the con-
trolled generation of monodisperse emulsion droplets by
coflowing two immiscible fluids to induce droplet formation.
The size of the monodiperse emulsion droplets can be
controlled by tuning the relative flow rates of the two fluids.
In these devices, spherical emulsion droplets are usually
generated because the disperse phase seeks to minimize its
interfacial free energy.45,46 However, also nonspherical
hydrogel microparticles with embedded magnetic nanopar-
ticles have been synthesized using ultraviolet-initiated pho-
topolymerization.47
2.2. Protection and Stabilization
Maintaining the stability of magnetic beads for a long time
without agglomeration or precipitation problems is a pre-
requisite for applications. Especially pure metallic particles
are subjected to oxidation or degradation, but also magnetite
nanoparticles are not very stable under ambient conditions,
and can be easily oxidized to maghemite or dissolved in an
acid medium. Particle protection results in magnetic beads
having a core-shell structure, where the role of the shell is
to protect the magnetic core against environmental influence.
Also, some particle compounds, in particular cobalt, are
considered extremely toxic, and thus the leakage of this into
the liquid medium must be avoided. Several coating strategies
exist, ranging from coating the magnetic nanoparticles with
organic shells containing surfactants and polymers, with
inorganic compounds like silica, with carbon, to coating
with precious metals. Besides simply coating individual
magnetic nanoparticles, it is also possible to embed a number
of magnetic nanoparticles or magnetic material in a polymer
or silica matrix to form composites. This way, one can
synthesize microbeads that, due to a higher magnetic content,
are more easy and rapid to manipulate magnetically than the
primary magnetic nanoparticles. Such types of beads (size
around 1 µm) are therefore often used in microfluidic
systems.
Surfactant and Polymer Coating
To stabilize magnetic nanoparticles after synthesis, elec-
trostatic or steric repulsion can be used to keep the nano-
particles dispersed in a nonagglomerated colloidal state. For
the so-called ferrofluids,48 the stability results from a control
of surface charges and the use of specific surfactants; both
water- and oil-based ferrofluids are available. Surfactants or
polymers can be chemically linked to or physically adsorbed
Microfluidic Applications of Magnetic Particles
to the magnetic nanoparticles, creating repulsive forces (due
to steric hindrance) that balance the attractive magnetic,
electrostatic, and van der Waals force. Polymers containing
functional groups, such as carboxylic acids, phosphates, and
sulfates, can bind to the surface of iron oxide,58 while surface-
modified magnetic nanoparticles with certain biocompatible
polymers are intensively studied for magnetic-field-directed
drug targeting59 and as contrast agents for magnetic resonance
imaging.60,61 For metallic magnetic nanoparticles, a thin
polymer coating may not be sufficient to prevent oxidation
of the highly reactive metal particles, while, more generally,
such coating will also be less stable at higher temperatures.
Precious-Metal and Carbon Coating
Several precious-metal coating procedures have been
proposed. Gold seems to be an ideal coating due to its low
reactivity. Coating magnetic particles directly with gold is
difficult due to the dissimilar nature of the two surfaces.
However, gold-coated iron nanoparticles with 11 nm core
size and 2.5 nm gold shell thickness have been prepared by
a partial replacement reaction in a polar aprotic solvent.62
Gold-coated iron nanoparticles were also prepared by a
reverse microemulsion method.63 Coating with gold is
interesting, because the gold surface can be easily chemically
functionalized with ligands via thiol groups.
Carbon-based coatings can be advantageous over polymers
or silica, due to their higher chemical and thermal stability,
as well as their biocompatibility. For example, a sonochemi-
cal coating procedure was developed that leads to air-stable
cobalt nanoparticles.64
Silica Coating
Silica coatings are attractive due to their relatively good
stability under aqueous conditions, their easy surface modi-
fication, and control of interparticle interactions via the
variation of shell thickness. The Stöber process65 and other
sol-gel techniques66,67 are the preferred methods for coating
the magnetic nanoparticles with silica. The coating thickness
can be varied by tuning the concentration of ammonium and
the ratio of tetraethoxysilane (TEOS) to water. Silica-coated
magnetic nanoparticles are hydrophilic and can be easily
functionalized with several groups.68 On pure metal nano-
particles, direct silica deposition is complicated because of
the lack of hydroxyl groups on the metal surface. Another
problem can be the oxidation of a metal like iron or cobalt
in the presence of dissolved oxygen in the reaction medium.
Although silica coatings are in general robust, silica is
unstable under basic conditions and in addition may contain
pores through which oxygen can diffuse.
Embedding Nanoparticles in a Protective Matrix
Matrix-dispersed magnetic nanoparticles can be created
in different ways. The nanoparticles can be homogeneously
dispersed in a continuous matrix, they can be dispersed as a
coating on other larger particles, or they can form agglomer-
ates that are connected through their protective shells. The
preparation of magnetic polymer microbeads was reviewed
in detail by Horák et al.29 For example, polymer-coated
magnetic beads can be produced by in situ precipitation of
magnetic materials in the presence of a polymer. In this way,
single or multiple magnetic core beads surrounded by a
hydrophilic polymer shell have been made, choosing for the
Chemical Reviews, 2010, Vol. 110, No. 3 1523
polymer the water-soluble dextran,69 poly-(ethylene imine),70
poly-(vinyl alcohol),71 poly-(ethylene glycol),72 etc. The well-
known Dynabeads73 are magnetic monodisperse microbeads
developed in a multistep procedure. They contain iron oxides
that are formed in situ by precipitation in preformed porous
and monodisperse polymer microspheres, followed by a
coating step.74,75 Such types of polystyrene-coated magnetic
particles are known for their excellent size distribution and
spherical shape.73,75 However, their hydrophobic surface
results in a high amount of aspecific protein and antibody
adsorption on the particle surface, so that it often needs to
be modified chemically. Magnetic silica particles are very
efficient in adsorbing proteins and DNA on their surface,
but are hardly available with a small size distribution and
an ideal spherical shape.76,77 Magnetic polysaccharide par-
ticles are important for many in vivo applications. They
combine biocompatibility with availability in a size range
below 300 nm,33,78 but the particles are irregular in shape,
and the soft particle matrix causes them to be sensitive to
mechanical stress. Magnetic poly-(lactic acid) particles also
play an important role in the in vivo applications:79,80 they
are biodegradable, and their degradation time in the blood
can be adjusted by their molecular weight and exact chemical
composition. However, because of their hydrophobic sur-
faces, these particles stick to plastic surfaces found in
microfluidic systems, resulting in problems with particle
handling and analytical errors. In other work, nonspherical
hydrogel microparticles with embedded magnetic nanopar-
ticles have been synthesized in microfluidic chips using
ultraviolet-initiated photopolymerisation.47 As an example
of matrix-dispersed magnetic nanoparticles, Figure 3b is a
transmission electron microscopy (TEM) micrograph of
dextran nanoparticles; note that the dextran is not visible in
the TEM picture and that the particles have an irregular
(nonspherical) shape. Figure 3c is a TEM photograph of more
regular polystyrene particles with encapsulated magnetite
nanoparticles. Figure 3d is a scanning electron microscopy
(SEM) micrograph of irregular magnetic glass microparticles.
2.3. Functionalization
As was already indicated in the previous section, a
protective shell around a magnetic nanoparticle or a protec-
tive matrix not only protects the particle against degradation,
but can also be used to functionalize the bead surface with
specific molecules, like those used in catalytic applications,
or for use in experiments with proteins, cells, etc. Proteins
may bind/adsorb to hydrophobic surfaces, such as those
found in polymer-coated beads, forming a monolayer that
is resistant to washing. It may also be desirable to have a
strong covalent binding between the particle surface and the
protein. This is achievable through specific groups at the
particle surface (-COOH-NH2, -CONH2, -OH groups),
which via an activating reagent bind to -NH2 or -SH groups
on the proteins.81 Also, streptavidin, biotin, histidine, protein
A, protein G, etc., can be grafted onto the bead surface for
specific biorecognition reactions.29 Small antibody-binding
peptides have also been proposed for surface functionaliza-
tion.82 Silica-coated beads can be used as-prepared to recover
and purify the total DNA content of a lysed cell solution.
Sample DNA is for example bound to the silica surface after
chemical cell lysis in a guanidine thiocyanate binding buffer.
It binds preferentially to the silica surface of the magnetic
beads due to the presence of chaotropic salts like guanidine
or sodium iodine in the buffer solution.83
1524 Chemical Reviews, 2010, Vol. 110, No. 3
In many catalytic applications, transition metals are used
as catalyst to transform or bond reagents. The good catalytic
activity of these elements is due to their ability to have
various oxidation states helping electron transfers or, in the
case of metals, to adsorb other substances onto their surface
and activate them in the process. When used with magnetic
beads, different strategies are developed to link the metal
atoms to the bead surface. The metal can be either complexed
using ligands, chemically bonded to the surface, or directly
immobilized on or incorporated into the magnetic bead
surface using a supporting matrix (see section 8.2).
3. Forces on Magnetic Beads
3.1. Magnetic Force
The magnetic force (eq 1 or 2) is responsible for the unique
possibilities offered by magnetic beads. Particularly interest-
ing are small magnetic monodomain nanoparticles, because
they do not express hysteretic magnetization curves and have
zero remnance. Leslie-Pelecky and Rieke84 have reviewed
the relation between the morphology of nanostructured
materials and their magnetic properties. Magnetic particles
are single-domain, when they have a dimension that is
typically of the order or smaller than the thickness of a
magnetic domain wall δ ) (JS2π2/Ka)1/2, with J the magnetic
exchange constant, S the total spin quantum number of each
atom, a the interatomic spacing, and K the magnetic
anisotropy constant of the magnetic material.19 For iron,
assuming that S ) 1, and taking J ) 2.16 × 10-21 J, a )
2.86 × 10-10 m, and K ) 4.2 × 104 J/m3, one calculates a
domain wall width of 42 nm. The time over which the
magnetization of a single-domain particle is stable and will
remain in a certain state is of importance for probing the
moment and magnetization. The relaxation time of the
magnetic moment of a particle can be written as25
( )
KeffV
τ ) τ0 exp (4)
kBT
with τ0 ≈ 10-9s, Keff the effective anisotropy constant of
the particle, kB the Boltzmann constant, and kBT the thermal
energy. Equation 4 shows that monodomain magnetic
particles become superparamagnetic, that is, their time-
averaged magnetization without a magnetic field is zero, over
a typically supposed experimental time scale τ ) 100 s, when
their magnetic energy Keff × (4/3)πr3 is lower than about
20 times the thermal energy kBT, with r the particle radius
of a supposed spherical particle.84 When the particle magnetic
moment does not reverse its magnetic moment over the
experimental time scale, such a particle is in the “blocked”
state. The temperature that separates the two regimes is the
blocking temperature and very sensitively depends on the
particle size. For example, at room temperature, one finds a
maximum radius r ) 6 nm for a superparamagnetic spherical
particle of iron. In the superparamagnetic state, such a
nanoparticle behaves similar to a paramagnetic spin but with
a much larger moment. The induced magnetization of an
ensemble of such particles is described by the Langevin
function, which at low fields (mH , kBT) behaves as mH/
kBT, and at high fields (mH . kBT), as (1 - (kBT)/(mH)),
with m the size of the individual particle moment.19
An additional advantage of using small nanoparticles,
besides the possibility to easily switch off the magnetic state
Gijs et al.
and particle interactions by removing the external magnetic
induction, is the minimum disturbance such a particle has
on reactions, implying the biomolecules attached to the
particle85 and the large surface-to-volume ratio, which is of
high interest for chemical binding. A disadvantage, however,
of such a particle may be that the magnetic force is small
due to the small volume, so that viscous forces dominate
and magnetic separations can take a long time (tens of
minutes).
This explains the interest of using larger magnetic particles
(typically 0.2-5 µm in diameter) in microfluidic applications.
These can either have a single magnetic core or have a core
composed of multiple (non)interacting nanoparticles in a
nonmagnetic matrix (see Figure 1a). Such microparticles
often have a multidomain structure and are characterized by
a hysteretic magnetization characteristic (see Figure 1b). For
example, spherical particles with radius r ) 1.5 µm, a
saturation magnetization µ0Msat ) 0.2 T, and a remnant
magnetization µ0Mrem ) 1 mT will have the remnant
magnetic moment size mrem ) Vµ0Mrem ) (4/3)πr3µ0Mrem )
1.4 × 10-20 T m3. One notes here that the magnetic moment
of a magnetic bead not only depends on its volume or size,
but also on its magnetic load, that is, the kind and amount
(up to 70%) of magnetic material that is present in the
particle. Two of such particles will have the maximum
magnetic attraction energy |Umax| ) (mrem2/2πµ0)(1/(2r)3) )
9.2 × 10-19 J, much larger than the thermal energy of ∼4 ×
10-21 J at room temperature, resulting in strong magnetic
dipolar forces between the particles.19 As a consequence,
when exposed to an external magnetic induction, such
magnetic microparticles coalesce under influence of the
magnetic dipole interaction into a SPS consisting of chain-
like “columnar” structures along the field direction. The exact
shape of this SPS depends on parameters such as the particle
concentration and applied magnetic field. Even after removal
of the external magnetic induction, such particle structures
can stay agglomerated, hindering separation and recovery
of the magnetic beads.
While the magnetic character of a magnetic bead is
essential for delivering the magnetic force, it is interesting
to note that also diamagnetic objects brought in solutions of
paramagnetic ions can be displaced on the basis of their
repulsive force from a high magnetic field.86,87 The magnetic
force on a magnetic bead can be induced using a magnetic
field generated either by permanent magnets or by coils. A
permanent magnet typically is characterized by a magnetic
induction Bm ) 0.5-1 T and a field gradient ∇B ≈ Bm/w,
with w the typical geometrical dimension of the permanent
magnet.88 For a cylindrical permanent magnet with a diameter
Ø ) 5 mm, one induces on a spherical particle with radius
r ) 500 nm and ∆χ ) 1 a magnetic moment m ) 2.6 ×
10-19 T m3, resulting in a magnetic force of about 40 pN.
For a current-fed coil, the generated field is much smaller:
a flat millimeter-size coil with 10 windings and a current of
0.5-1 A generates typically a magnetic induction of 1-10
mT, at least 100 times smaller than the permanent magnet.
Consequently, the gradient is also a factor 100 lower, so that
the force of eq 2 is a factor 104 larger, when using a
permanent magnet rather than a coil. On the other hand, a
coil offers more flexibility in switching the field on and off
by simply controlling the current (and heat dissipation!),
whereas a permanent magnet requires a mechanical action
to move it away from the microfluidic system containing
the magnetic beads.
Microfluidic Applications of Magnetic Particles
3.2. Viscous Drag Force
In many applications, a magnetically labeled material is
separated from a liquid solution by passing the fluid mixture
through a region with a magnetic field gradient that can
immobilize the tagged material via magnetic forces. For in
vivo applications, magnetic particles can be transported by
the blood flow and locally retained with the help of an
external magnet. In other applications, there is a magnetic
translational driving force, and the liquid solution is static.
In the equilibrium state, the magnetic force Fm is opposed
to the hydrodynamic drag force Fd acting on the magnetic
particle. The hydrodynamic drag force is shown in Figure 4
and is a consequence of the velocity difference between the
magnetic particle and the liquid ∆v and, for a spherical
particle with radius r, is given by89
Fd ) 6πηr∆vfD (5)
where η is the viscosity of the medium surrounding the
particle (for water, η ) 8.9 × 10 -4 N s/m2). fD is the drag
coefficient of the particle and incorporates the influence of
a solid wall in the vicinity of the moving particle and is
calculated as90,91
9 r 1 r 3 45 r 4
[
fD ) 1 - (
16 r + z
+ ) (
8 r+z ) - (
256 r + z
-)
1 r 5 -1
(
16 r + z )] (6)
Here, z is the distance of the particle to the solid wall; in the
limit of small z, fD ≈ 3, so that the drag will be a factor 3
higher than when no solid wall is in the vicinity. In a
microfluidic system, a particle may be moving at different z
from the microchannel wall, under influence of the applied
magnetic forces, so that the viscous force of eq 5 will vary
along the particle trajectory.
Equalizing eqs 2 and 5 permits one to determine the
maximum flow rate a particle can withstand when exposed
to a magnetic immobilization force, or the maximum particle
flow rate that can be generated by a magnetic force in a
surrounding static liquid:
2r2∆χ(B·∇)B 1 (ζs2 + ζp2) exp(-2κz)]ez (10)
∆v ) ≡ ξ(B·∇)B (7)
9µ0ηfD µ0fD
with
Figure 4. Simplified schematic diagram of the main directions of
the most important forces acting on a magnetic particle with radius
r and positioned at a distance z from a surface. The particle is in
a fluid and is subjected to a magnetic force Fm, a viscous drag
force Fd, a gravitational force Fg, and an electrical force Fel
originating from the surface.
Chemical Reviews, 2010, Vol. 110, No. 3 1525
2r2∆χ V∆χ
ξ≡ ) (8)
9η 6πrη
the “magnetophoretic mobility” of the particle, a parameter
describing how magnetically manipulable the particle is. The
quantity (B · ∇)B can be a strongly varying function in space,
which implies a similar spatial variation for ∆v.
3.3. Electrical Interaction with the Microchannel
Wall
Both electrostatic and electrodynamic (van der Waals)
forces act on a particle in solution. The van der Waals force
originates from attractive electromagnetic interactions be-
tween permanent electrical dipoles and/or induced dipoles.
The van der Waals force between a particle and a substrate
or the wall of a fluidic microchannel is expressed in the
Hamaker approach as91,92
[ ]
A123r 1
FVdW ) - e (9)
6z 2 1 + 14(z/λret) z
with A123 the Hamaker constant of particle of material 1 on
a substrate of material 2 in a medium of material 3, and ez
the unit vector in the vertical direction; λret is a characteristic
length scale of the interaction, often assumed to be 100 nm.
For example, one finds for an acrylate-coated particle close
to a SiO2 substrate in water a value of A123 ) 3.4 × 10-21
J.91,93
When immersed in an ion-containing aqueous solution,
the substrate and particle acquire a surface charge through
the adsorption of ions present in the solution and/or the
presence of charged surface groups. This surface charge is
neutralized by mobile ions of opposite charge also present
in the solution, thus forming the well-known double layer.
When the double layers of two surfaces overlap, an elec-
trostatic interaction occurs, resulting in either a repulsive or
an attractive force. This electrical force Fel is given, for a
constant surface charge density on substrate and particle, by94
2πεκr
Fel ) [2ζ ζ exp(-κz) +
1 - exp(-2κz) s p
with ε the permittivity of the medium (7 × 10-10 F/m for
water), and ζs and ζp the zeta-potential of the substrate and
particle, respectively.95 The electrical force is acting over a
spatial range given by the Debye-Hückel double layer
thickness κ-1:

kBT
κ-1 ) (11)
2000NAe2Ic
with NA Avogardo’s number, and Ic the ionic strength in
mol/L of the solution. κ-1 is typically in the 10 nm range.
Wirix-Speetjens et al.91 showed how, by changing the pH
and ionic strength of the solution, it is possible to obtain
repulsive electrical force conditions between the particle and
the substrate surface, considering combined van der Waals
and electrostatic forces. This is important, if one wants to
avoid the issue of unwanted particle sticking to the micro-
fluidic channel. Higher pH values (6-10) and low ionic
strength (1-10 mM) are shown to favor repulsive electrical
1526 Chemical Reviews, 2010, Vol. 110, No. 3
forces for the acrylate-coated particle/SiO2 substrate system.91
Alternatively, at the cost of more complex technology, it is
possible to coat the microchannel wall with proteins or
polymers to avoid sticking of beads or molecules.96
3.4. Other Forces
In addition to the magnetic, viscous, and electrical forces
on magnetic beads, several other forces exist. As indicated
in Figure 4, a gravitational force acts on a magnetic bead
and is given by
Fg ) -mbgez (12)
with g the gravitation constant, and mb ) V(F - Ffl) the
“buoyant mass” that represents the effective mass of the
magnetic bead by taking into account the average density
of the bead F and the surrounding fluid Ffl. For example, 1.0
µm diameter MyOne Dynabeads73 have a volume of 0.52
µm3, 26% iron content, and a density of 1.8 g/cm3, resulting
in a gravitational force of 0.004 pN, which is much smaller
than the typical magnetic force generated by a permanent
magnet, but of the order of the force generated by a simple
coil. When only gravitation is taken into account, these
magnetic beads will sink to the bottom of the microchannel,
hindered by a viscous force (eq 5) in the z-direction, resulting
in a “sinking speed” of 0.5 µm/s. As the ratio of gravitational
to viscous forces is proportional to r2, smaller particles will
sink much slower.
When the size of the magnetic bead is in the submicrome-
ter range, stochastic Brownian forces become relatively more
important, as the magnetic (and gravitational) forces, being
proportional to r3, are fastly decreasing. In a long-time
dynamics, the friction coefficient 6πηr of eq 5 is related to
the diffusion coefficient D of the particle by the well-known
Stokes-Einstein relation:97
kBT
D) (13)
6πηr
For example, a MyOne Dynabead has a diffusion coefficient
in water of 4.8 × 10-13 m2/s at room temperature, resulting
in the time-dependent diffusion length (Dt); for example,
after 1 s of diffusion, the particle will have moved an average
distance of 0.7 µm. Taking a magnetic nanoparticle with
radius of 100 nm instead, the diffusion coefficient becomes
2.4 × 10-12 m2/s, resulting in a more than 2 times increased
diffusion. Analysis by optical microscopy of the Brownian
motion of magnetic nanoparticles in a magnetic field gradient
was used to characterize the basic magnetic parameters of
the particles.98,99 A rotating magnetic field was used to apply
a controlled torque on superparamagnetic beads, leading to
a tunable bead rotation frequency in fluid, as observed via
optical microscopy.100 It is suggested that control of torque
and measurement of the rotation will enable future torsion-
based protein assays as well as nucleic acid assays on a single
bead.
A combination of magnetic and dielectrophoretic forces
to discriminate specific and nonspecific molecular bindings
for on-chip magnetic bioassays was reported.101 Conjugated
to the analytes, magnetic particles are used as the agents for
dielectrophoretic force generation. Because of a weaker
binding strength, the nonspecifically bound particles are
removed, while specific bindings remained intact. This
technique can not only be used to improve the specificity of
Gijs et al.
on-chip bioassays, but also be developed as a tool of force
spectroscopy for the study of biomolecular binding physics.
Additional forces act on magnetic beads, such as hydro-
dynamic, magnetic, and electrostatic interaction forces orig-
inating from other beads,102 especially when bead concen-
trations are high and the particles come close to each other.
Magnetic interactions between beads lead to the formation
of bead clusters or SPS, which have their own dynamics, as
well as magnetic and viscous forces.98,103,104 Finally, to
describe the motion of a magnetic bead, one should solve
Newton’s law involving all forces present, which becomes
extremely complex and leads to cumbersome numerical
procedures. Therefore, a practical approach is to consider
only the most important forces (for example, the magnetic
and viscous forces for magnetic microparticles), when
designing magnetic bead manipulation systems.
4. Magnetic Bead Manipulation
The spatial and temporal combination of the forces
introduced in the previous section allows designing proce-
dures or manipulation protocols, which are at the basis of
the applications of magnetic beads in bioanalysis or catalysis.
Figure 5 shows diagrams representing basic manipulations
of magnetic beads in microfluidic systems. In separation
(Figure 5a), magnetic beads are retained from a flow by
focusing a magnetic field over the channel using, electro-
magnets, coils, or permanent magnets. Also, systems with
miniaturized soft magnetic microstructures exist, which can
be coupled to a macroscopic electromagnet or placed in the
magnetic induction field of a macroscopic permanent magnet.
Magnetic transport (Figure 5b) is more difficult, as it requires
stronger and long-range magnetic forces to move the
magnetic beads within the liquid, without the need of a
microfluidic flow. The use of magnetic beads as labels for
detection is shown in Figure 5c: here, a magnetic bead is
bound to the surface of the microchannel, and a magnetic
field sensor monitors its stray field, when the particle is
placed in an external magnetic induction. A particularly
interesting property of magnetic beads is that they can be
magnetically suspended in a microfluidic channel using
magnetic forces, without the requirement of having a
supporting substrate. Such trapping of beads or SPS can be
advantageous, if one wants to have a high exposure of the
beads to a liquid flow. Also, when locally alternating
Figure 5. Schematics of basic manipulations of magnetic beads
in a microfluidic channel. (a) Separation of magnetic beads from a
flow by actuation of electromagnets or positioning of magnets. (b)
Transport of magnetic beads using long-range magnetic forces
provided by an electromagnet or magnet. (c) Detection of the stray
field of a bead by a magnetic field sensor, after specific binding of
the labeled bead on the sensor surface. (d) Mixing of two laminar
streams by dynamic agitation of a supraparticle structure using a
locally applied alternating magnetic field between two soft magnetic
tips.
Microfluidic Applications of Magnetic Particles Chemical Reviews, 2010, Vol. 110, No. 3 1527

magnetic fields are applied, for example, via soft magnetic
field focusing structures, a dynamic agitation of the magnetic
beads is possible that can be used to mix the essentially
laminar flow patterns within a microfluidic channel (Figure
5d).
4.1. Retention and Separation
4.1.1. Systems with Macroscopic Magnets
Separation of magnetic bead-labeled biomolecules or cells
from a liquid solution is a well-documented and a widely
practiced application today. Many types of magnetic particles
have been developed for use in separation processes,
including purification processes and immuno-assays.105-109
Magnetic separator design can be as simple as the application
and removal of a permanent magnet to the wall of a test
tube to cause particle aggregation, followed by removal of
the supernatant. However, it is preferable to increase the
separator efficiency by producing regions with a high
magnetic field gradient to capture the magnetic beads as they
float or flow by in the carrier medium. Magnetic cell sorting
(MACS) is a commonly used magnetic cell sorting techno-
logy,78,110 in which cells bound to superparamagnetic particles
are retained in a high-gradient magnetic field, generated by
placing a porous magnetic column in the field of external
magnets. After flowing the sample solution through the
column, the latter is removed from the separator, and the
retained particles can be analyzed using flow cytometry or
microscopy. For such application, it is necessary to loosely
pack a flow column with a stationary magnetizable matrix
of thin wires or beads.111,112 Such approach has the drawback
that close monitoring of the separation process is difficult
due to the nature of the filter geometry. Continuous flow
separation in microfluidic devices was reviewed by Pamme
(not restricted to magnetic beads).113 Considering magnetic
beads, separation in a microfluidic capillary without magnetic
stationary phase has been demonstrated.114 In this separation
technique, also known under the name magnetic field-flow
fractionation,115 a magnetic force acts on the magnetic beads,
forcing them toward the capillary wall. The magnetic
retention force is opposed by diffusion back into the flow
stream, so that a steady-state particle distribution and an
ensemble of beads of specific height is formed at the capillary
surface. Hydrodynamic forces act perpendicular to the
applied field and laterally displace the magnetic plugs along
the capillary wall. Less-magnetic species will be less closely
packed, extend more to the center of the capillar, and will
therefore experience larger hydrodynamic forces. Hence, less-
magnetic species will move faster and “elute” first from the
capillary.
Continuous flow magnetic particle and cell separation
systems have been intensively studied by groups at The Ohio
State University and The Cleveland Clinic Foundation.116
These systems have a macroscopic dimension from the
magnet point of view, but, from a liquid transport point of
view, they may be called “microfluidic”, as their functioning
is based on the presence of laminar flow patterns. Figure 6a
is a schematic diagram of a so-called quadrupole magnet
configuration, where four magnets are arranged to induce a
maximum magnetic field gradient toward the outer side of a
liquid carrying tube, inserted in the free space between the
magnets.117 Williams et al. described how such a quadrupole
magnet can be combined with an annular fluidic circuit.118
The separation takes place within a laminar flow of the carrier
fluid along a thin annular channel. A magnetic field gradient
is imposed across the thin dimension of the channel,
perpendicular to the direction of the flow. The sample
mixture is arranged to enter the system close to one of the
channel walls, and, as the sample is carried along the channel
by the flow of the fluid, those components that interact more
strongly with the field gradient are carried transverse across
the channel thickness. A division of the flow at the channel
outlet using a stream splitter completes the separation into
two fractions. The radial particle separation velocity is
induced by the field gradient and can be calculated using eq
7. The technique was given the name split-flow thin

Figure 6. (a) Schematic diagram of a magnetic quadrupole separator. The magnetic force increases linearly in the radial direction. A
magnetic particle suspension is injected from the top in the inner annular flow channel, while a buffer solution flows within the outer
annular flow channel. As the magnetic particle suspension flows through the separator, the magnetic particles are deflected in the radial
direction. If sufficiently deflected, the particles are caught in the lower outer annular flow channel. Undeflected or weakly deflected particles
follow the inlet streamlines and are collected in the inner annular waste flow channel. (b) Schematic diagram of a magnetic dipole separator.
A magnetic force directed to the right is created, and a magnetic particle suspension injected in the bottom port migrates to the right.
Different types of magnetic particles will be deflected into different outlet ports. (c) Distribution of the magnetically induced velocity and
magnetophoretic mobility of four different types of magnetic beads in a magnetic dipole separator. (c) Reprinted with permission from ref
124. Copyright 2001 Springer.
1528 Chemical Reviews, 2010, Vol. 110, No. 3
(SPLITT) fractionation, a derivative of field flow fraction-
ation, as separation is obtained within a mobile phase without
the use of a stationary phase.119,120 This separation is
essentially a binary one, in the sense that the input sample
stream is split into a first exit stream of highly magnetic and
a second stream of less or nonmagnetic particles. Besides
annular liquid flow channels, SPLITT fractionation devices
with rectangular geometry121,122 and a magnetic field-flow
fractionation device with helical microfluidic channel placed
in a quadrupole magnetic field123 have been proposed.
Permanent magnets arranged in a dipole configuration have
been used for separation of magnetic beads as well. Figure
6b is a schematic diagram of a magnetic dipole separator.
Magnetic beads migrate perpendicular to the direction of the
flow at a rate proportional to their magnetic content. This
permits separation of the injected sample into multiple outlet
streams on the basis of the magnetic properties of each bead
species. By combining a separator design with a constant
magnetic energy gradient, an optical monitoring of the beads
(or bead-labeled cells), and a theoretical model for interpreta-
tion, it is possible to trace back the linear bead trajectories
to basic magnetic properties of the magnetic beads. This
technique was coined “cell tracking velocimetry”.124-126
Figure 6c is a histogram of the distribution of four different
types of magnetic beads with different magnetically induced
velocities, as measured by cell tracking velocimetry, together
with the derived magnetophoretic mobilities. In this case,
the magnetic energy gradient of the dipole separator was
designed to be (1/µ0)(B · ∇)B ) 198 T A/mm2 in the region
of interest (see eq 7). In other work, the combination of
magnetic dipole separation with microfluidics has been
proposed,127 experimentally demonstrated on-chip128 and
analytically modeled.129 Other permanent magnet configura-
tions have been proposed for separation as well.130-133
4.1.2. Systems with Microscopic Magnets: Magnetic
Templates
While the separation systems discussed so far have a
macroscopic dimension from the magnet point of view, an
interesting option for generating a local and strong magnetic
field gradient is to bring micropatterned soft magnetic
material into the field generated by a larger permanent
magnet (as done in a MACS column134). Because of the high
magnetic permeability of the magnetic material, the magnetic
flux from the permanent magnet is focused into the magnetic
microstructure and will create strong gradients at certain
magnetic interfaces, at which magnetic beads can be
separated from a passing solution. Magnetic induction
gradients of 0.1-1 T/mm have typically been realized in
microfluidic channels using soft magnetic structures com-
bined with permanent magnets.135,136 In principle, using
arrangements of bulk permanent magnets, 10 T/mm gradients
are possible.137 A first reported magnetic separation structure
was based on electrodeposited nickel posts of 7 µm height
and 15 µm diameter. Once magnetized by an external
neodymium-iron-boron permanent magnet, these nickel
post generated the strong magnetic field gradient that
efficiently trapped 4.5 µm superparamagnetic beads sus-
pended in water flowing past the posts.138 This experiment
was later confirmed by other authors.139 Using a similar idea,
magnetic separation was demonstrated by evaporating a thin
layer of cobalt on a preformed microstructure realized via
soft lithography or classical lithography.140,141 In one case,
liquid PDMS polymer mixed with magnetic particles was
Gijs et al.
Figure 7. (a) Schematic diagram of the microfluidic chip with
magnetic wires at an angle with respect to the hydrodynamic flow
direction. (b) Because of the vector sum of hydrodynamic and
magnetic force, a magnetic bead or magnetically labeled cell will
move along the direction of the magnetic stripe, while nonmagnetic
objects will move along the hydrodynamic force direction. Reprinted
with permission from ref 136. Copyright 2001 Wiley-VCH,
Weinheim.
allowed to polymerize in the presence of a magnetic field, a
process during which the magnetic particles can be trapped
in the polymerized PDMS microchannel wall. Hereafter,
these embedded magnetic beads act as “anchors” that can
trap magnetic beads that are flowing in the microchannel.
Also, the packing of Ni particles within a channel for the
generation of high magnetic field gradients in a nearby
microfluidic channel was reported.142 Several papers exist
on the analytical and numerical modeling of the magnetic
trapping by a micropatterned magnetic structure.140,143-146 A
magnetic separator consisting of micropatterned permalloy
elements adjacent to a microfluidic channel was presented
by Smistrup et al.147 When an external magnetic field is
applied transverse to the channel, the elements are magne-
tized and generate magnetic field gradients within the
channel, such that the magnetic beads are attracted to and
captured at the channel side walls. Two types of magnetic
beads could be selectively placed at either side of the channel,
with negligible cross-talk, by use of hydrodynamic focusing.
This system also enabled the simultaneous exposure of a
DNA sample to two types of DNA probes.148 A variant of
this separation system was presented, in which a staggered
herringbone microfluidic mixer was integrated in the micro-
channel.149 This proved to increase the capture-and-release
efficiency by up to a factor of 2. Simulations of bead
trajectories, capture efficiencies, and capture distributions
were also presented. Following a similar idea, a system with
stainless steel wires placed close to a capillary in the
magnetic field of permanent magnets showed magnetic flux
focusing at the wire ends and was able to separate magnetic
beads from a solution flowing through the capillary.150
An original idea for continuous magnetic separation
(without accumulation of magnetic beads at certain
positions in a microchannel) was presented by Berger et
al.136 It is based on the magnetic force generated by an array
of magnetized stripes that are positioned at an angle with
respect to a hydrodynamic flow direction. These stripes create
a series of magnetic field gradients that trap the magnetic
beads and alter their flow direction. Figure 7a is a schematic
diagram of the microfluidic chip with indication of the
oblique magnetic stripes or wires. The sum of the hydrody-
namic force and the magnetic force acting on a magnetic
bead creates a resultant force vector, which moves the
Microfluidic Applications of Magnetic Particles
magnetic bead or cell labeled with beads out of the main
sample stream,151 as illustrated in Figure 7b. A similar
separation principle was later presented by others.152-154
4.1.3. Systems Using Microelectromagnets
An important difference between the use of a permanent
magnet and an electromagnet is the much lower magnetic
induction generated by the latter. A permanent magnet easily
generates a magnetic induction of 0.5-1 T, while the
magnetic induction of a simple planar coil is typically in
the mT range. Following the discussion in section 3.1, one
understands that a microelectromagnet will produce a much
smaller magnetic force (in the fN range for a simple spiral
coil and the chosen parameters for the magnetic beads), so
that fluidic flow in the microchannels needs to be strongly
limited or that magnetic beads need to pass at an extremely
close distance to the planar coil. On the other hand, coils
offer flexibility, as the magnetic field can be simply switched
off by setting the coil current to zero. The group of C. H.
Ahn has pioneered the use of microelectromagnets for the
separation of magnetic beads.155-157
Figure 8a presents a schematic diagram of a microelec-
tromagnet consisting of a planar coil and a soft magnetic
yoke structure. Copper spiral coils were electroplated into
photoresist molds on a Si substrate, and, when actuated by
currents in the order of 0.3 A, could effectively retain and
separate micrometer-size superparamagnetic beads from a
flow. The integrated fluidic microchannel was realized by
anodically bonding a microstructured Pyrex wafer to the Si
substrate. For enhancing the generated magnetic field (by
about a factor of 2 with respect to a simple coil, as half of
the space with µr ) 1 around the coil is replaced by a soft
magnetic material with high magnetic permeability), the
spiral coil is combined with an electroplated permalloy
magnetic yoke microstructure (Figure 8b), leading to effec-
tive magnetic bead separation from a flow (Figure 8c). Such
soft magnetic microstructures have also been used in
combination with micropatterned current-carrying wires for
magnetic field enhancement.158 However, the spiral design
results in stronger magnetic fields and forces as compared
Figure 8. (a) Schematic diagram of a microelectromagnet used for separation of magnetic beads. (b) Micrograph of the fabricated
electromagnet (coil width ) 50 µm). (c) Separation of 1 µm diameter magnetic beads out of a flow by actuating the second coil with a
current of 0.3 A. Reprinted with permission from ref 156. Copyright 2001 Elsevier.
Chemical Reviews, 2010, Vol. 110, No. 3 1529
to other current-carrying geometries.159 Similar work exists
on the separation of magnetic beads by microelectromagnets
made of copper spiral coils and various wire microstructures
in combination with nickel soft magnetic yokes.160-163
Magnetic bead capture by current-carrying wires also has
been studied theoretically, and optimum separation condi-
tions, as influenced by parameters like microchannel dimen-
sion, magnetic field magnitude, and flow speed, were
determined.164,165 Besides separation via electromagnets
realized directly underneath a microfluidic channel, also
“external” microelectromagnets or magnetic tweezers have
been proposed to trap magnetic beads in a liquid solution.
The tweezers were made by winding 25 µm thick copper
wire around a 50 µm diameter soft magnetic needle and can
be scanned to effectively displace the magnetic beads in a
fluidic reservoir.166
Coils offer flexibility, but only relatively small magnetic
forces can be generated. Combining a coil with a soft
magnetic yoke structure can, at maximum, increase the force
by a factor of 2 with respect to the simple coil. Also, ohmic
heating of a coil can pose problems and require active cooling
of the microfluidic system.167 An interesting option to
generate larger forces is to combine a current-carrying coil
or wire for local magnetic field gradient generation with a
uniform external magnetic field for imposing a large
magnetic moment to the superparamagnetic beads (see eq
1).168-170 Such setup allowed for the application of relatively
strong pN forces (100× enhancement) for magnetic beads
in the micrometer range and on a spatial scale corresponding
to the microfluidic channel width.
4.2. Magnetic Transport
Magnetic separation is different from magnetic transport
in the sense that, in separation, the beads are retained
(separated) by action of a magnetic field, but transported
using a liquid flow. In magnetic transport, magnetic forces
effectively transport the particle, which is a bigger challenge:
it requires magnetic fields and magnetic forces that act on a
longer range than necessary for separation, where magnetic
1530 Chemical Reviews, 2010, Vol. 110, No. 3 Gijs et al.

beads approach very closely the magnetic actuation region

by the fluid motion. Transport is a difficult task, as the
magnetic susceptibility ∆χ of the magnetic beads is rather
weak (typically e1), due to small magnetic core volumes
and demagnetization effects of the particles (see eq 2). This
explains why mostly the large field of permanent magnets
has been used for the separation and positioning of magnetic
beads.78
In an approach toward miniaturization and automation of
analytical applications, a system has been proposed in which
liquid movement is substituted by magnetically induced
movement of magnetic particles.171 Fluidic channels were
realized on a plastic cartridge of centimeter size, and
magnetic transport was induced by mechanically moving
external permanent magnets. In another approach, magnetic
particles have been transported over millimeter distances in
a microfluidic channel using an array of electromagnets
actuated in a four-phase scheme.172 Each electromagnet
consisted of a 0.3 mm diameter magnetic needle core with
a wire-wound coil of 300 turns. For coil currents of the order
of 0.5 A, forces of 0.1 pN were possible. Also, alginate
microdroplets with encapsulated magnetic particles have been
transported using this method.173
Besides these still “macroscopic” fluidic transport systems,
also miniaturized solutions have been proposed for bead
transport, thereby taking advantage of batch microfabrication Figure 9. Sequence of schematics and corresponding microscopic
technologies. Typically, the size of the (micropatterned) images, showing the stepwise transport of magnetic field maxima
magnets determines the spatial range, where appreciable and magnetic beads, by arranging the relative positions and currents
i of two serpentine wires. Reprinted with permission from ref 170.
magnetic forces act on the microbeads. Serpentine gold wires Copyright 2001 American Institute of Physics.
micropatterned on silicon substrates have been combined
with microfluidic structures realized in PDMS to transport
dimensional stepwise transport was realized by applying low
4.5 µm polystyrene-coated magnetic beads.170 By engineering
frequency nonoverlapping clock pulses to both conductors
the magnetic field generated by different current-carrying
alternatively. The average speed of a magnetic particle was
wires, a microsystem was realized that could generate local
defined as the distance over which the particle is transported
magnetic field maxima able to trap the magnetic beads (see
parallel to the conductor edge divided by the minimal time
Figure 9). When the field maxima change location, the beads
followed those maxima. The device allowed precise position- needed to reach the next minimal cross-section and is of the
ing and transport over a distance of 100 µm in a single order of several 10 µm/s. In another approach, a planar coil
actuation event, which is partly due to the presence of a array-based magnetic transport system has been proposed,
permanent magnet placed in proximity of the microfluidic in which an individual coil is capable of displacing beads
chip. The main role of this magnet was to enhance the over millimeter distances in a liquid-containing capillary.176
magnetic force by inducing a magnetic moment in the A drastic increase of the magnetic energy and magnetic
magnetic beads. A microelectromagnet wire matrix, based forces acting on the beads was obtained by placing the
on two layers of mutually orthogonal arrays of linear wires, complete coil array in a uniform static magnetic field that
has demonstrated magnetic transport of 1-2 µm size imposes a permanent magnetic moment onto the beads. The
magnetic particles over 20 µm distances in a single actuation very small magnetic field (gradient) of a simple planar coil
event.174 This is a typical working range for the magnetic proved then to be sufficient to displace 1 µm diameter beads
force, generated by a current-carrying conductor, when no over a distance of the order of the coil size. The coils were
external permanent magnet is used to induce a magnetic realized using simple Printed Circuit Board (PCB) technology
moment in the beads. A transporting device consisting of a (100 µm Cu winding width, 35 µm winding height, 200 µm
set of two tapered conductors shifted linearly over half a winding pitch) and had a small number of windings (N )
period was demonstrated by Wirix-Speetjens et al.91,175 This 4-10). Arranging adjacent coils with spatial overlap over
device was fabricated using standard semiconductor technol- two layers of the PCB circuit and actuating them in a specific
ogy and photolithography techniques. Current conductors three-phase scheme assured the long-range displacement of
(TiW 10 nm/Au 150 nm/TiW 10 nm) were evaporated and the beads. Moreover, it was found that these polarized beads
patterned using a lift-off process on an oxidized Si wafer. A formed cylindrical columns with a length of the order of the
250 nm thick SiO2 layer was then sputtered onto the wafer microfluidic channel size, due to magnetic dipole interactions.
for passivation. By sending a dc current through one of the This column formation helped to generate the strong
current conductors, a magnetic field is generated that magnetic force. On a smaller length scale, arrays of 50-100
magnetizes the particles. Once the particles are magnetized, µm-size microcoils were made on silicon substrates by
they moved toward the edge of the conductor, where the electroplating soft magnetic NiCoP pillars at the center of
field is highest. Subsequently, they were attracted to the electroplated Cu coils.163,177 These microelectromagnets
nearest narrow cross section, driven by the field gradient that generated sufficient magnetic force to transport 1 µm
is created by tapering the conductor. This way, one- diameter beads from coil to coil.
Microfluidic Applications of Magnetic Particles
Also, a transport system was proposed, which employed
a translating periodic potential energy landscape to move
magnetic beads horizontally over a substrate. The potential
energy landscape was created by superimposing an external
rotating magnetic field on top of the local magnetic field
distribution of a periodic arrangement of Co micromagnets.
At low driving frequencies of the external field rotation (a
few Hz), the magnetic beads become locked into the potential
energy landscape and move at the same velocity as the
traveling magnetic field wave. Therefore, this technique was
coined traveling wave magnetophoresis.178 Also, complex
patterns of soft magnetic microstructures placed in a rotating
magnetic field were able to trap, transport, and release
magnetic particles. The translatable local magnetic field
maxima were created here by variations in local radii of
curvature at structural edges of the microstructures.179
4.3. Magnetic Beads as Labels for Detection
A common approach to detecting biological molecules is
to attach a label that produces an externally observable signal.
The label may be a radio-isotope, enzyme, fluorescent
molecule, or charged molecule, but also magnetic beads can
be used as labels for biosensing. Magnetic labeling, detection,
and the combination of a sensor chip with microfluidic
system integration aspects were reviewed by several
authors.180-182 Magnetic labels have several advantages over
other labels. The magnetic properties of the beads are stable
over time, especially because the magnetism is not affected
by reagent chemistry or subject to photobleaching (a problem
with fluorescent labeling). There is also no significant
magnetic background present in a biological sample, and
magnetic fields are not screened by aqueous reagents or
biomaterials. In addition, magnetism can be used to remotely
manipulate the magnetic particles. Finally, a number of very
sensitive magnetic field detection devices have been devel-
oped during recent years, like giant magnetoresistance
(GMR)183 and spin-valve184,185 magnetic sensors that enable
the measurement of extremely weak magnetic fields, such
as the magnetic stray field generated by the magnetization
of a single magnetic bead. A basic GMR or spin-valve device
consists of a pair of magnetic films separated by a nonmag-
netic conducting layer.186 When an external magnetic field
rotates the magnetizations of the magnetic layers toward
alignment, spin-dependent electron scattering is reduced at
the interfaces within the device, thus decreasing its electrical
resistance. GMR sensors can be microscopic in size and
sensitive to the presence of micrometer and smaller size
magnetic particles, when in close proximity; even single
micrometer-size particles immobilized down to a few
hundred nanometer above the sensor could be detected.182
Also, magnetic particles passing by in a flow, rather than
being immobilized above the sensor area, have been de-
tected.187 Besides GMR sensors, measurements of single
magnetic beads have been demonstrated using miniaturized
Hall sensors188,189 and planar Hall effect sensors based on
permalloy thin films.190 A possible inconvenience of the
magnetic sensor-based detection systems for commercial
applications, however, is that the cost of such a system can
be an issue, due to sensor fabrication (often on a silicon
substrate) and integration of the sensor in a microfluidic
package.
A research group at the Naval Research Laboratory,
Washington, DC,191-194 followed by others,195,196 has devel-
oped a microsystem for the capture and detection of
Chemical Reviews, 2010, Vol. 110, No. 3 1531
Figure 10. Optical micrographs of a third-generation bead array
counter (BARC) sensor chip. (a) Photograph of the 68 pin-out
chip, including a central sensing area with 64 sensors and two
reference sensors, and a number of test structures. (b) Closer view
of the central sensing area. (c) Close-up of one serpentine GMR
sensor trace encompassing a 200 µm diameter sensing zone.
Reprinted with permission from ref 193. Copyright 2000 Elsevier.
micrometer-sized, paramagnetic beads on a chip containing
an array of GMR sensors, the so-called bead array counter
(BARC). Bound beads are detected by the GMR sensors by
applying a uniform magnetic field perpendicular to the
substrate, and therefore imposing a magnetic moment to the
superparamagnetic beads. These induced moments generate
an in-plane magnetic field component that is measured by
the GMR sensor. Applying the uniform field normal to the
plane of the GMR sensor rather than in-plane has the
advantage that, due to demagnetization effects, a much larger
magnetizing field can be applied to the beads without
saturating the sensor.197 Figure 10 shows optical micrographs
that represent a sensor chip with 64 individually addressable
GMR sensors and two reference sensors. Each sensor is a
serpentine resistor trace that is 1.6 µm wide and with a 4.0
µm pitch (see Figure 10c), with a total length of 8 mm within
a 200 µm diameter circular zone. The zone was matched to
an arraying system for probe deposition (250 µm spots) to
selectively capture beads from a flow. The sensor chips are
covered with a silicon nitride passivation layer of about 1
µm thick to protect the circuitry from the corrosive and
conductive media and biochemical reagents. This nitride layer
was etched down to a final thickness of 250 nm over each
sensor zone leading to a larger signal due to the smaller
immobilized bead-sensor distance. Using Dynal M-280 2.8
µm diameter beads, the threshold for detection was ap-
proximately 10 beads per 200 µm diameter sensing area. The
GMR sensor sensitivity increases with decreasing surface
area of the sensor; however, the chemical sensitivity, or
number of analyte molecules that can bind to the surface,
increases with increasing surface area. Theoretical modeling
showed that a GMR sensor can detect a single superpara-
magnetic bead of any size,197-199 as long as all system
dimensions (bead size, sensor size, distance between bead
and sensor) scale down proportionally. When the sensor size
approaches the size of the bead, it should be possible to detect
beads with a radius down to 100 nm or smaller.
A group at Philips Research has proposed a compact
biosensor platform with GMR sensors suited for the detection
1532 Chemical Reviews, 2010, Vol. 110, No. 3
of superparamagnetic nanoparticle labels.200,201 The platform
consists of a disposable microfluidic cartridge, a GMR chip,
and an electronic reader. Wires integrated in the silicon of
the sensor chip are used to generate a well-defined magnetic
field on the sensor surface, thus removing the need for
mechanical alignment with an external apparatus. To opti-
mize the signal-to-noise ratio, the magnetic labels are excited
at high frequency (1 MHz). These wires also apply forces
onto the beads to attract the latter toward the sensor surface.
A signal modulation scheme is applied to obtain optimal
detection accuracy. Experimental results indicated that three
beads of 300 nm can be detected on a sensor surface of 1500
µm2 for a measurement time of 1 s. In an effort of cost
optimization, the GMR sensors were deposited on a passive
silicon substrate, with the electronic circuitry placed on an
external chip.202 Despite this, the use of comparatively
expensive silicon as substrate and the microfluidic packaging
of the chip203 might raise cost issues for a (single-use)
diagnostic test, unless this particular approach would prove
to outperform others, by, for example, a much higher
sensitivity.
A lot of bead-sensing experiments today have been done
with micrometer-size beads composed of magnetic γ-Fe2O3
and Fe3O4 nanoparticles dispersed in a polystyrene matrix.
To maximize the sensor response, the magnetic beads should
have a magnetization as high as possible and yet remain
nonremnant to avoid clustering when suspended in solution.
With the goal of achieving larger signals from the magnetic
labels, one has developed soft ferromagnetic beads with
100% magnetic content and much higher saturation magne-
tization.192 One micrometer diameter NiFe beads showed a
susceptibility of ∼3, the maximum obtainable value for a
uniformly magnetized sphere. Because of this property,
smaller solid ferromagnetic beads could effectively be used
as biomagnetic labels, which would also increase the dynamic
range of biosensor assays by allowing more labels per unit
area.
4.4. Enhancing the Bead-Flow Interaction and
Mixing
Most of the magnetic bead-related phenomena discussed
so far were based on the interaction of individual beads with
the magnetic field. An interesting property of magnetic beads
is that they can form linear-like chain structures or more
complex SPS, when placed in a magnetic field due to the
magnetic dipole interaction between beads. When the
magnetic induction can be very locally applied, it is pos-
sible to have a small-size magnetic SPS or bead “plug”
spanning the cross-section of a microchannel, which will be
characterized by a strong perfusion by a liquid flow. This is
of importance for both biomolecule capture and catalysis
applications. Hayes et al.204 used a varying magnetic field
generated by a mechanically moved permanent magnet to
create a relatively large magnetic SPS plug composed of 1-2
µm diameter superparamagnetic beads (see Figure 11a). The
plug could be rotated through all axes in a microfluidic
channel, without loss of structural form, allowing dynamic
micrometer-scale movement without direct mechanical,
electrical, or photonic interactions.204 A number of potential
applications of this phenomenon were described, such as
binding biomolecules on the magnetic particles for immuno-
assays,207 studying subcellular biomechanics, and microflu-
idic mixing in pico- and femtoliter volumes.
Gijs et al.
Figure 11. (a) Superparamagnetic particles in a cylindrical
microchannel (20 µm diameter), as observed by optical microscopy:
(i) High-volume fraction 1-2 µm diameter particles without applied
external magnetic field. (ii) Induced column-like structures that form
immediately upon application of an external magnetic field in the
plane of the page. (iii) Upon rotation of the applied magnetic field
perpendicular to the page, the top of the column-like structures is
visible. (b) Video microscopy images of (i) a dense SPS kept over
the microchannel cross-section in an alternating magnetic field by
two soft magnetic tips, at a field oscillation frequency f ) 5 Hz, a
liquid flow velocity V0 ) 0 cm/s, and a magnetic field amplitude
B0 ) 100 mT, (ii) the expanded SPS of (i) for f ) 5 Hz, V0 ) 0
cm/s, and B0 ) 5 mT, (iii) a snapshot image of the oscillating SPS
at the most downstream position in the channel for f ) 5 Hz, V0 )
0.4 cm/s, and B0 ) 25 mT, and (iv) a snapshot image of the
oscillating SPS at the most downstream position in the channel for
f ) 5 Hz, V0 ) 0.4 cm/s, and B0 ) 20 mT. (c) Optical image of a
microchannel showing the retention of 18 self-assembled nanopar-
ticle magnetic chains that are perfectly positioned in the geometrical
traps of the microchannel, while applying a flow from the left to
the right. (a) Reprinted with permission from ref 204. Copyright
2001 American Chemical Society. (b) Reprinted with permission
from ref 205. Copyright 2004 American Institute of Physics. (c)
Reprinted with permission from ref 206. Copyright 2008 American
Chemical Society.
The shape of a SPS depends on different parameters, like
the size of the magnetic moment of the beads and the
magnetic dipolar interaction between different beads. These
properties are linked to the amplitude and frequency of the
applied magnetic field, the shape and magnetic content of
the beads, the concentration of the magnetic particles in
the fluid, the temperature, etc. Despite the complexity of the
aggregation process of a magnetic fluid into a SPS, the
physical effects of a magnetic field on such a structure are
now very well understood.208,209 When exposed to a strong,
continuous magnetic field, the magnetic fluid will rapidly
form a cross-linked network. The continuous-field structure
is determined by the kinetics of bead aggregation, and the
particles are prohibited from rearranging to minimize energy,
as long as the field persists. In contrast, it was found that
the application of a pulsed field (square wave alternating
between field-on and field-off states) to a magnetic fluid did
produce an energetically determined suspension structure.210,211
By allowing particle diffusion during the field-off state, a
pulsed field enables minimization of energy through struc-
tural rearrangements, and the SPS consists of one-dimen-
sional periodic patterns composed of high-concentration
regions of magnetic particles (“columns”), aligned in the field
direction and sharply separated from low-concentration
regions.208,212,213 Qualitatively, the dependence of the structure
complexity on the tuning parameters can be understood from
the time it takes for the particles to aggregate, which depends
on the ratio of the magnetic interaction energy between
particles relative to thermal energy. Applying a biaxial
rotating magnetic field, produced by two orthogonal pairs
of Helmholtz coils in quadrature, induces a rotation of dipolar
Microfluidic Applications of Magnetic Particles
chains of superparamagnetic particles and subjects the
aggregates to magnetic forces causing them to rotate within
the suspending fluid. It has been shown, with both scatterning
dichroism214 and video-microscopy experiments,215 that
magnetic fluids have the capacity of reducing the size of the
structures they are composed of, to decrease their viscous
drag while rotating synchronously with the field.
For an application where low-concentration biomolecules
need to be recovered from a flow by the SPS, it is useful to
have only a small amount of magnetic beads in the bead
plug, as this will improve the purification or concentration
of bound biomolecules per bead. The highest possible
number of biomolecules should bind to the lowest possible
number of beads, to result in the highest biomolecule
concentration per bead. Rida and Gijs205 applied microma-
chined soft magnetic tips to focus the magnetic induction
generated by an electromagnet very locally over a micro-
channel. An oscillating and localized magnetic field was
applied to study the magnetohydrodynamic transport of a
small magnetic SPS consisting of ferromagnetic beads (see
Figure 11b). The particles formed open columnar structures
that had a strong exposure to the fluid flow. The dynamics
of the SPS were analyzed as a function of the flow rate and
the frequency and amplitude of the magnetic field. The
experimental data were well described by a magneto-
hydrodynamic model, where the columns in the SPS were
represented by cylinders. This type of magnetic particle-
driven mixing was also modeled numerically.216
Ferromagnetic beads, which form stable rotating chains
in an alternating magnetic field, easily form bead clusters
and agglomerates after field removal. However, in many
bioanalytical applications, individual beads should be re-
leased from the plug after analyte capture for further
processing. Unfortunately, the dynamic manipulation tech-
nique described in the previous paragraph cannot be readily
applied to superparamagnetic or low-coercivity beads, as the
latter change their magnetic state by Néel relaxation. Rotation
of chains of superparamagnetic beads was achieved using
macroscopic rotating magnetic fields,217-219 but such ap-
proaches are based on complex systems, making microfluidic
integration difficult. Moreover, such a system is not directly
suitable for bead retention in a microchannel, as the magnetic
field gradients are weak. However, dynamic actuation of
superparamagnetic beads in a microchannel could be dem-
onstrated using a quadrupolar magnetic actuation system,21
where cyclic bead motion was achieved in a confined
microchannel volume by superposing a magnetic time-
varying field, as induced by two soft magnetic tips connected
to an electromagnet, and a static field induced by two
permanent magnets.
In other work,206,220 retention of magnetic beads in a
microfluidic channel was achieved, not by focusing the field
at a specific location across the channel, but by periodically
varying the width of the channel. When brought into a
homogeneous magnetic field applied perpendicular to the
channel axis, a solution of magnetic beads flowing through
the microchannel spontaneously self-assembles in periodic
magnetic chains located at the larger cross sections of the
microchannel (see Figure 11c). This configuration benefits
from the lowest total magnetostatic energy, as originating
from magnetic dipole interactions between the individual
magnetic beads. These geometrically trapped nanoparticle
chains are at the basis of a good fluid perfusion through the
magnetic chain structure and a high biomolecule capture
Chemical Reviews, 2010, Vol. 110, No. 3 1533
efficiency. To enhance the interaction of magnetic beads with
molecules present in a flow, magnetic beads were also
chemically fixed to a glass surface, either using poly(ethylene
glycol) (PEG) hydrogels221 or by electrostatic assembly on
micropatterned silane structures.222,223
The dynamic manipulation of magnetic bead aggregates
was also investigated for mixing applications. Active fluid
mixing was demonstrated in microchannels made in a
micromachined microfluidic chip of polymethylmetacrylate
(PMMA). Mixing was based on the manipulation by a local
alternating magnetic field of self-assembled porous structures
of ferromagnetic microbeads that are placed over the section
of the channel.205,224 Using a sinusoidally varying magnetic
field (1 Hz < f < 100 Hz), a rotational motion of the SPS
was induced, thereby strongly enhancing the fluid perfusion
through the SPS that behaved as a dynamic random porous
medium. The mixing is the result of the chaotic splitting of
fluid streams through the dynamic and randomly porous
structure of the SPS and the relative motion of magnetic
entities with respect to the fluid flow. It was quantified by
monitoring the fluorescent intensity of initially parallel
fluorescent and nonfluorescent laminar streams. A 95%
mixing efficiency over a channel length as small as the
channel width (200 µm) and at flow rates of 0.5 cm/s was
obtained. This demonstrates the large lateral mass transfer
induced by the SPS as a consequence of the highly
heterogeneous and dynamic nature of the SPS. On a smaller
length scale, a mixing system consisting of a microfluidic
channel integrated with a number of soft magnetic elements
at the sides of the channel was demonstrated.225,226 The
elements were magnetized by placing the chip in a homo-
geneous external alternating magnetic field. This way,
dynamic plugs of magnetic beads that spanned the micro-
channel were created, and 85% mixing efficiency was
obtained. Similar mixing efficiencies were reported by other
authors using a similar approach.227 Another type of magnetic
force-driven mixer has been fabricated.228 This mixing device
was based on a silicon chip with embedded microconductors
as a magnetic field source, and a microchannel guiding the
fluid streams. By applying a time-dependent control signal
to a row of microconductors, it was found that a two-
dimensional serpentine channel geometry with transverse
electrodes was able to create the stretching and folding of
material lines, resulting in good mixing. In other work,
magnetic beads were chemically linked to create permanent
chains that subsequently were magnetically actuated by a
rotating magnetic field to mix two laminar streams in a
microfluidic channel.217 The mixing via the use of magnetic
beads has also been described in detail using numerical
simulations.228-230
Another technique to realize magnetic bead chains is the
magnetic assembly of microbead chains onto a surface.141,231
It was shown that superparamagnetic particles assemble on
a micropattern of thin ferromagnetic islands, acting as
ferromagnetic traps; magnetic bead trapping could be
controlled by varying the external magnetic field bias. In
addition, a programmable self-assembly method for the
placement of two or more different types of superparamag-
netic colloidal beads onto lithographically defined magnetic
microwell templates has been demonstrated.232 Besides
mixing, the agglomeration of magnetic microbeads into a
SPS also has shown to play a role in magnetic transport.
The large magnetic bead transport velocities (10 mm/s)
obtained by actuation via a planar coil array placed in a
1534 Chemical Reviews, 2010, Vol. 110, No. 3
uniform magnetostatic field only could be explained by the
formation of columnar magnetic objects with a strongly
enhanced magnetic moment and corresponding magnetic
energy.176
4.5. Magnetic Droplets
In contrast to continuous flow microfluidics, where com-
paratively large liquid volumes are pumped through the
microsystem, droplet microfluidics, also called “digital
microfluidics”, handles only small self-enclosed liquid
entities.233,234 The advantage of droplet handling is the strong
reduction of the transported volumes and the possibility of
working simultaneously with a plethora of different samples
in a common system, which is why the manipulation of small
droplets is steadily gaining impact on the development of
lab-on-a-chip systems.234 To perform reactions between the
different droplets, the system however needs to be able to
execute various droplet manipulation steps, such as transport,
merging, mixing, and splitting. Among the solutions pre-
sented in recent years, which range from electrowetting over
dielectrophoresis to acoustic actuation,233 magnetic droplet
manipulation offers the advantage of long-range magnetic
forces, which do not rely on the intrinsic properties of the
manipulated medium and do not interact with most biological
materials.235 This noninteraction, on the other hand, requires
the introduction of magnetically responsive material into the
droplets, where they translate an applied magnetic field
gradient into a force pulling the droplet into the direction of
the gradient.236 Such magnetic particles have the advantage
that they can serve as functionalized components of the
system, such as optical indicators237 or mobile substrates238
in addition to being the force mediators.239 While magnetic
particle actuation in droplets allows a nonlabeling, fast, and
material-independent controlled droplet manipulation, electri-
cal mechanisms are better compatible with smaller liquid
volumes. Electrowetting-on-dielectric (EWOD) is a particu-
larly promising technique, because it uses a simple device
configuration and fabrication, generates large forces on the
microscale, and consumes little energy. EWOD-based ma-
nipulations of aqueous droplets have been demonstrated in
parallel-plate devices filled with oil240 or dry in air.241
A challenge in magnetic droplet manipulation is the
application of the magnetic field gradient. In some cases, an
external magnet was used, which can be moved according
to the requirements of the droplet manipulation.242 In
combination with a suitable three-dimensional structuring of
the surface to create barrier or gating structures, all droplet
manipulation steps, like separation, transport, and fusion, can
be implemented (see Figure 12), but at the cost of increasing
system complexity and decreasing flexibility.243 Droplet
Figure 12. Schematic top view (a) and side view (b) of a magnetic droplet actuation system. Magnetic beads in a first water droplet can
be extracted from the droplet using a permanent magnet and a physical barrier (gate), for subsequent merging with a second water droplet.
Reprinted with permission from ref 243. Copyright 2006 Elsevier.
Gijs et al.
movement, coalescence, and splitting have also been inves-
tigated for droplets moving on an open hydrophobic sur-
face.244 These processes were actuated by magnetic beads
internalized in an oil-coated aqueous droplet using an external
magnet. The results were explained theoretically with a
simple model that balances magnetic, friction, and capillary-
induced drag forces and includes the effects of particle type,
droplet size, surrounding oil layer, surface tension, and
viscosity.
In other work, magnetic particle concentration and separa-
tion was demonstrated in droplets moving in a channel245,246
or near the bottom of an oil-filled reservoir,247 by using both
permanent magnet-induced forces and EWOD droplet ma-
nipulation. Mixing has been demonstrated as well using a
rotating magnetic field acting on magnetic bead chains inside
an aqueous droplet positioned on a superhydrophobic sur-
face.248 A different approach was the use of a matrix of coils
to generate local magnetic field gradients. In this case, the
magnetic particles no longer follow a moving magnet, but
are controlled by the changing topology of the applied
magnetic field.249-251 Because the actuating field gradient
only acts on the particles enclosed inside the droplet, the
magnetic force is transferred onto the droplet via the droplet
wall. The moving particles accumulate at the droplet bound-
ary and push the droplet into the desired direction. As a
consequence, the droplet will move as long as the magnetic
actuation is larger than the friction exerted by the surrounding
liquid and surfaces. In case a droplet should stay in place,
the local friction can be increased, either by introducing
three-dimensional obstacles252 or by varying the wetting
properties of the surface in contact with the droplet.253
5. Magnetic Cell Manipulation
5.1. Cell Types
5.1.1. Cells with Magnetic Character
Few cells with intrinsic magnetic properties exist naturally.
A first type of cell with magnetic character is a deoxygenated
red blood cell (RBC) (see Figure 13a). Hemoglobin is the
iron-containing oxygen-transport metallo-protein in the RBCs
of vertebrates. In mammals, the protein makes up about 97%
of the RBC’s dry content, and around 35% of the total
content (including water). Oxygenated hemoglobin is dia-
magnetic due to the presence of paired electrons on its Fe
atoms. Oxygenated RBCs have therefore a small relative
magnetic susceptibility with respect to that of water: ∆χoxy
≡ χRBC,oxy - χwater ) -0.19 × 10-6,254 with χwater ) -9.0 ×
10-6.255 Deoxyhemoglobin is the form of hemoglobin without
the bound oxygen to the Fe atom, resulting in a RBC relative
Microfluidic Applications of Magnetic Particles Chemical Reviews, 2010, Vol. 110, No. 3 1535

Figure 13. Schematic diagram of (a) a red blood cell, (b) a magnetotactic bacterium, (c) a cell with digested magnetic nanoparticles, and
(d) a cell with magnetic nano- or microparticles attached to the surface via a ligand-receptor interaction.

magnetic susceptibility∆χdeoxy ≡ χRBC,deoxy - χwater ) 3.3 × Table 1. Absolute and Relative Number of Cell Populations and
10-6.256 While this is still a small value, the use of magnetic Subpopulations in Normal Blood (Adapted from Ref 260)
fields in the tesla range allows one to generate sufficient cell type quantity/µL
fraction of
magnetic force to retain or separate deoxygenated RBCs from WBC (%)
a complex matrix.257 Magnetotactic bacteria form a second red blood cells 5 × 106
category of naturally magnetic cells258 (see Figure 13b). reticulocytes (3-7) × 104
platelets (2-5) × 105
These mobile bacteria are present in water-based sediments white blood cells (total) (5-10) × 103 100
and move along the earth’s field lines, a phenomenon called neutrophils (4-8) × 103 40-65
magnetotactism. At the basis of this magnetic interaction is monocytes (2-8) × 102 4-8
a cluster or row of biogenerated magnetic Fe3O4 crystals eosinophils 50-300 1-3
inside the bacteria. basophils 0-100 0-1
lymphocytes (total) 1000-4000 20-40
CD4+ T cells 400-1600 15-20
5.1.2. Nonmagnetic Cells Labeled with Magnetic Beads CD8+ T cells 200-800 7-10
B cells 200-800 7-10
Other cell types are essentially nonmagnetic and can be natural killer cells 100-500 4-6
magnetically “activated” with magnetic nano- or micropar-
ticles via digestion of the particles by the cell (see Figure
13c) or via a (specific) ligand-receptor interaction at the Table 2. Cluster of Differentiation (CD) Markers for Several
Cell Types
cell surface (see Figure 13d). The majority of literature data
on magnetic bead-based cell separation deal with the cell type CD markers
separation of white blood cells (WBCs), cancer cells, or stem cells CD34+, CD31-
bacteria from serum or blood. Blood has two main compo- all white blood cells CD45+
monocytes CD45+, CD14+
nents, plasma and cells, each representing about one-half of T cells CD45+, CD3+
its volume. There are three main types of cells with different B cells CD45+, CD19+, or CD45+, CD20+
functions under different conditions: erythrocytes or RBCs, natural killer cells CD16+, CD56+, CD3-
leukocytes or WBCs, and platelets. RBCs transport oxygen,
WBCs defend against disease, platelets promote clotting,
while plasma proteins perform a variety of functions. Table to a fully differentiated endothelial cell. Table 2 presents
1 gives an overview of the most important type of cells and the CD markers for a few types of cells.
their abundance in normal blood. Depending on the needs
of diagnosis, only some of the blood cells will be of interest, If an assay is seeking to analyze the DNA from WBCs,
and these need to be isolated and purified from other cell a microliter sample volume is sufficient to analyze.
types that are irrelevant or can even compromise good However, rare cells in blood, like bacteria, tumor cells,
analysis of the selected cells. The WBCs are divided into or fetal cells in maternal blood, occur in concentrations
five classes: neutrophils, lymphocytes, monocytes, eosino- of ten to a hundred cells per milliliter. Sampling of 1 µL
phils, and basophils (see Table 1). However, this classifica- volume may not provide even one cell of interest for
tion is too general for many applications. The cluster of subsequent analysis. Clearly, the large initial sample
differentiation (CD) is a protocol used for more precise volume with few numbers of target cells requires a
identification and investigation of cell surface molecules
powerful preconcentration or purification step.261 Here,
present on WBCs and other cell types.259 CD molecules are
proteins that can act in a variety of ways as receptors or magnetic beads can play a prominent role, if they specifically
ligands important to the cell behavior. Usually a combination bind to CD markers on the cell surface. Also, the intracellular
of CD markers is used to classify cells, identify their function, uptake of superparamagnetic iron oxide nanoparticles has
or reveal cellular responses to particular pathological situ- been studied.262 Three different cell types, mesenchymal stem
ations. Cell populations are defined using a plus or a minus cells, cardiac fibroblasts, and cultured hematopoietic stem
symbol to indicate whether a certain cell fraction expresses cells, were exposed to superparamagnetic nanoparticles
or lacks a CD molecule. For example, a CD34+, CD31- treated with a number of different transfection agents. The
cell is one that expresses CD34, but not CD31. This CD study was based on the magnetophoretic analysis of live cells
combination typically corresponds to a stem cell, opposed and supplemented by cytochemical staining.
1536 Chemical Reviews, 2010, Vol. 110, No. 3
5.2. Magnetophoretic Mobility of a Cell
The magnetophoretic mobility of a magnetically tagged
cell is a parameter that distinguishes these cells from
unlabeled cells. Following eq 8, the magnetophoretic mobility
of a simple magnetic particle is the velocity of the particle
in a magnetic energy gradient divided by the magnitude of
the gradient. This velocity is the result of the magnetic force
created on the particle by the interaction of the magnetic
material with the imposed magnetic energy gradient. While
for a magnetic particle, the part of the magnetic force
originating from the particle itself (and not from the external
field) is related to the particle volume and the susceptibility
(V∆χ), the situation for an immuno-magnetically labeled cell
is more complex. Indeed, the magnetic force depends on the
number of magnetic particles on the cell membrane, while
the viscous drag force is due to the complete cell-nanoparticles
complex. Therefore, in case of a two-step binding protocol
on cells (see Figure 14), five parameters will influence the
magnetophoretic mobility of an immunomagnetically labeled
cell:264-266 the antibody binding capacity (ABC) of a cell
population, the secondary antibody binding amplification
factor (ψ), the number of magnetic particles bound to one
antibody n, the particle-magnetic field interaction parameter
of the magnetic particles (V∆χ), and the cell diameter (Dc).
ABC is the number of primary antibodies binding to the cell.
This value includes not only the number of antigen molecules
per cell, but also variables such as valence of antibody
binding, steric hindrance, binding affinities, and nonspecific
binding events. The amplification factor ψ is obtained by
binding a secondary antibody with multiple magnetic particle
interaction sites to the primary antibody, and n is the number
of magnetic particles that conjugate to the (second) antibody.
Summarizing, the magnetophoretic mobility for a cell can
in analogy to eq 8 be written as
ABCψnV∆χ
ξcell ) (14)
3πDcη
The influence of each of these parameters on the magneto-
phoretic mobility has been studied in several papers.263,267
For example,267 ABC values in the range of 50 000-100 000
were reported for the CD34 antigen, which is a marker for
Figure 14. Comparison of immuno-magnetically labeled cells with
different values of antibody binding capacities (ABC), secondary
antibody amplification (ψ), or magnetic-particle field interaction
parameter (V∆χ). An increase in any of these parameters will
increase the amount of magnetic material bound to the cell and
therefore increases the magnetophoretic mobility. Reprinted with
permission from ref 263. Copyright 2003 American Chemical
Society.
Gijs et al.
hematopoietic progenitor cells. These cells are stem cells that
give rise to all of the blood cell types and are found in the
bone marrow of adults or in umbilical cord’s blood, placenta,
and mobilized peripheral blood.
5.3. Cell Separation Methods
Several reviews exist on on-chip separation and analysis
of specific cells from a complex matrix.268-270 Sorting
cells, either for subsequent culture or to purify one cell
type from a complex sample, is practiced using gravita-
tional, chemical, optical, magnetic, or mechanical methods.
The most obvious type of on-chip physical cell separation
is a filter structure that selects cells on the basis of
differences in size and disparity.271 Separation based on
sedimentation is based on gravitation-induced cell migration
across individual fluid lamina with a lateral transport or
migration velocity.272 The driving force of sedimentation is
the difference in cell and fluid densities in the gravity field.
For a mixture of similar cell types, the cell density, shape,
and cell size play a critical role in sedimentation. For cells
with similar properties, such as during separation of cancer-
ous cells from healthy tissue, there may not be sufficient
difference in sedimentation parameters to achieve efficient
separation. However, the main advantages of field-flow
sedimentation are the absence of a label, the ability to
continuously operate the device for high throughput, and the
simplicity of the design. Dielectrophoresis using AC electri-
cal fields is another physical method that has attracted
attention.273 It is most effective when there is a significant
size or dielectric difference between cell types. For disparate
cell separation, such as separating bacteria from blood cells,
there is a strong difference in dielectrophoretic force, and
cells physically settle at different locations on a substrate
surface. In separation by affinity cell capture, the chemical
interaction between cells and a surface is critical for selective
adhesion. Various parameters such as association and dis-
sociation constants, the number of bonds between the cell
and the surface, or the antibody density influence the
adhesion or removal from cells on the surface. In traditional
chemical affinity chromatography, the analyte is finally eluted
from a column by disrupting the target-capture molecular
bond, for example, by changing the pH or ionic strength of
the mobile phase. Such approach is less evident, when
dealing with the separation of cells. However, the specificity
of antibodies to match a desired antigen (CD) on the cell
surface has become a cornerstone in cell separation.
Currently, the majority of commercial cell separation
systems are based on fluorescent flow cytometers or other
complicated instrumentation.274,275 The most commonly used
methods for cell sorting are fluorescence-activated cell sorting
(FACS)274 and magnetic cell sorting (MACS)110 (see Figure
15). FACS instruments operate in the flow-through regime
and use laser-induced fluorescence to count and direct
droplet-based cells stained with fluorophore-conjugated
antibodies against an appropriate CD on the cell surface.
As immunofluorescence is used as the triggering signal
or readout, cells need to be labeled with fluorescent
markers, which adds up to the cost. Also, immunofluo-
rescence imaging requires higher quality optics, excitation
sources, and cameras than normal optical detection. FACS
is expensive, voluminous, and requires dedicated and
trained staff to operate, but is a high-throughput technique,
as up to about a million cells per second can be
characterized and separated. However, FACS is time-
Microfluidic Applications of Magnetic Particles
Figure 15. Schematic diagram of the two most commonly used methods for cell sorting. (left) In fluorescence-activated cell sorting,
separation is done using the electrostatic deviation of specific fluorescently labeled cell-containing droplets. (right) In magnetic cell sorting,
specific magnetic nanoparticle-labeled cells are retained in a high-gradient field magnetic matrix and are eluted from the matrix after
removal of the magnetic field for subsequent analysis. Reprinted with permission from ref 136. Copyright 2001 Wiley-VCH.
consuming due to the large number of RBCs, reticulocytes,
and platelets in a blood sample. FACS analyses have also
been transposed to microfluidic platforms and sorting rates
<100 cells/s have been achieved,276 but the technique still
requires high-quality and expensive detection optics and
electronics. This explains the interest to look for alternative
separation methods. As discussed in section 4.1.2, in MACS,
cells labeled with superparamagnetic nanoparticles are
retained in a high-gradient magnetic field generated by
placing a magnetic column in the field of external magnets.
The column is then removed from the separator, and the
retained cells are eluted (positive selection). Alternatively,
all but the cells of interest are magnetically labeled and are
retained in the column, so that only the unlabeled cells of
interest pass through the column and are recovered (negative
selection). Subsequent cell detection is done using flow
cytometry or microscopy. Selectivity of the separation is
obtained by grafting the magnetic nanoparticles with anti-
bodies against cell proteins on the surface of specific cells.
The MACS method of cell sorting is less expensive and
simpler than FACS, but it is important to note that MACS
represents only a first sample handling step before additional
analysis. Therefore, the idea of performing magnetic separa-
tions on a microfluidic platform is attractive, as this will
facilitate the direct-reading, separation, and counting of the
cells on a single chip. Technical advantages such as the
ability to create large magnetic field gradients using only
modest magnetic sources and the capability for precise
handling of the separated cells make the use of microscale
magnetic sorters a very attractive option for point-of-care
diagnostic devices.
Chemical Reviews, 2010, Vol. 110, No. 3 1537
5.4. Magnetic Cell Separation and Purification
Red Blood Cells
The separation of red blood cells from whole blood was
first demonstrated by Melville et al.257 These authors used a
column packed with magnetic wires, which was placed in a
magnetic induction of 1.75 T with magnetic gradients close
to the wires estimated as 8 × 103 T m-1. After application
of whole blood mixed with a reducing agent, the magnetic
field was switched off and RBCs “eluted” from the column.
Other systems based on a similar retention principle have
been used for the separation of RBCs and WBCs from whole
blood277-279 or for the separation of malaria-infected RBCs
from whole blood.280,281 Using a magnetic dipole system,
RBCs labeled with several types of transition metal ions were
separated by retaining the magnetic beads from a flow at a
surface that was close to a magnetic pole.282 A microfluidic-
based permanent magnet assembly, generating a constant
magnetic force over a small zone of interest in a micro-
channel,124,126 was used to study the magnetophoretic mobili-
ties of oxygenated and deoxygenated RBCs.254 A continuous
separation of RBCs from whole blood or diluted blood was
demonstrated in another type of microfluidic device, by
passing the blood around thick (40-50 µm) magnetized
nickel structures placed in or underneath the microfluidic
channel in the presence of an external magnetic field.153,283,284
The magnetic flux from the external field is attracted toward
the nickel, providing a magnetic field gradient in the
microfluidic channel that acts differently on a WBC and a
RBC. This way, such device proved to separate continuously
up to 93% of RBC and 97% of WBCs from whole blood at
a volumetric flow rate of 5 µL h-1.154 This type of separation
system has also been modeled theoretically.285 In other work,
1538 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 16. Time lapse image showing a single magnetically tagged
fluorescing WBC at different times tracking a magnetic stripe at
an angle of 9.6° to the fluid velocity (110 µm/s) indicated by white
arrows. Red blood cells on the left are from a single image. All
cells entered the chip at the same point approximately 1.5 mm above
the field of view. Reprinted with permission from ref 151. Copyright
2004 American Institute of Physics.
nucleated RBCs have been identified in maternal circula-
tion.286 The detection of these cells provides a potential
source for the monitoring and diagnosis of maternal, fetal,
and neonatal health and disease. Separation of the nucleated
RBCs was done using a two-step enrichment process: (i) a
microfluidic separation based on cell size and the cells’s
ability to deform when moving through an array of posts on
the chip, and (ii) a hemoglobin enrichment module followed
by a batch-type magnetic separation in a MACS column.
White Blood Cells
As shown in Table 1, WBCs are less abundant, giving
high relevance to the separation and purification of this type
of cells from whole blood. Several disease conditions can
be detected or monitored by counting the number of disease-
specific cells. For example, the low concentration of helper
T cells (characterized by a surface expression of CD4, or
CD4+ cells) and a low ratio of the number of those cells to
the number of cytotoxic T cells (CD8+ cells), CD4/CD8, is
a practical measure to quantify the progression of the Human
Immunodeficiency Virus (HIV).287 The separation of WBCs
from whole blood was demonstrated by Inglis et al.151 using
a continuous flow microfluidic device, consisting of magnetic
stripes, magnetized by an externally applied field of 0.08 T
and placed at a small angle (9.6°) to the direction of the
fluid flow, while a narrow stream of cells was carried by the
flow over the stripes. Cells, magnetically labeled using anti-
CD45 magnetic beads, were attracted to the stripes and
tended to follow the stripe direction, while unlabeled cells
did not interact with the stripes and followed the direction
of the fluid flow (see Figure 16). It was pointed out288
that the greatest challenge with this type of system is
preventing the magnetically labeled cells from permanently
sticking to the stripes.
A microfluidic chip was proposed to separate cultured B
cells tagged with 4.5 µm diameter magnetic beads142 from a
flow. The cells were deviated from the fluidic streamlines
or trapped by a bed of packed Ni particles within a second
nearby channel, the role of which was the generation of high
magnetic field gradients in the first microfluidic channel,
when the system was placed in an external magnetic field.
In other work,221 magnetic beads were covalently linked to
a glass surface using a PEG hydrogel. The beads were
completely embedded in the hydrogel and, when placed in
an external magnetic field, could separate B cells, which were
before incubated with anti-CD19 magnetic beads. The same
authors also reported the direct capture of B cells from a
flow when the anti-CD19 magnetic beads were linked to the
Gijs et al.
glass with their active surface exposed to the flow. In another
study,289 enrichment of human T cells from a blood sample
was demonstrated. First, anti-CD3 magnetic beads were
introduced in a chip and captured in a field generated by a
permanent magnet. A blood sample was then introduced, T
cells were captured and rinsed clean, collected at the chip
outlet, and moved off-chip for subsequent analysis. T cell
capture efficiencies up to 37% were demonstrated.
WBCs have also been separated from buffy coats or raw
blood samples using the magnetic quadrupole and bipole flow
sorters developed by The Ohio State University and The
Cleveland Clinic Foundation groups, discussed in section
4.1.1. A buffy coat is the fraction of an anticoagulated blood
sample after density gradient centrifugation that contains
most of the white blood cells and platelets. A typically used
procedure is described in ref 290. Human peripheral leuko-
cytes are collected from healthy donors, and the mononuclear
cell fraction (MNC) is isolated by centrifugation. The typical
composition of the MNC fraction is 90-95% lymphocytes,
with the remainder of the cells identified as monocytes and
a small fraction of erytrhocytes. The lymphocytes are targeted
using mouse antihuman anti-CD45 monoclonal antibody
conjugated to fluorescein isothiocyanate (FITC) and mag-
netized by a secondary rat antimouse polyclonal antibody
conjugated to dextran-coated magnetic beads, as used in a
MACS-type system. To simulate a real cell sample to be
sorted, a small number of magnetically labeled lymphocytes
was used to spike unlabeled lymphocytes, at a final fractional
concentration of a few % of labeled cells. The binding of
the magnetic beads to the target cells increases their
susceptibility and makes them sensitive to an external
magnetic field. At the same time, the use of FITC conjugated
to the primary antibody sensitizes the target cells to UV light
and makes them appear as bright cells in FACS analysis,
while the unlabeled cells are simply part of the cell
autofluorescence background. It was shown that there is a
one-to-one correspondence between fluorescent and magnetic
cell populations,291,292 so that FACS analysis of the separated
cell fractions allows one to investigate the efficiency of the
magnetic separation. An example of cell fluorescence
histograms of the original sample and the two sorted fractions
is shown in Figure 17. These histograms essentially correlate
with the magnetophoretic mobility histograms (of the type
shown in Figure 6c), as determined by cell tracking veloci-
metry in a continuous flow microfluidic magnetic sorter.293,294
A quadrupolar magnetic cell sorter was also used to deplete
peripheral blood leukocytes of T cells.295,296 For an average
initial number of 2 × 108 total cells, a 4 log10 depletion of
CD3+CD45+ T cells was achieved at a sorting speed of
106 cells/s. This technique is of interest for allogeneic CD34+
hematopoietic stem cell transplantation, which is the only
curative option for many patients with hematological ma-
lignancies. A high-level depletion of donor T cells (4-5
log10) from the graft can effectively limit the risk of graft
versus host disease. While a high level of depletion of T
cells on a clinical scale was indeed obtained, the level of
CD34+ stem cells recovered for transplantation was still
suboptimal, as shown by FACS analysis. So, still, large
samples (2 × 1010 total cells) to be sorted, to obtain a
sufficient number of CD34+ cells for a transplant, are
required. A dipole magnetic flow sorter was used by the same
group297 to separate CD4+ and CD8+ T cells labeled with
primary antibody-FITC conjugate and anti-FITC-magnetic
colloid secondary antibody. Postseparation FACS analysis
Microfluidic Applications of Magnetic Particles
Figure 17. Histograms of labeled human CD45 lymphocytes as a function of cell fluorescent intensity. The cells are separated from a
mixture of unlabeled lymphocytes in a magnetic quadrupole flow sorter and subsequently analyzed in a FACS system. Bars and numbers
indicate targeted cell fraction (CD45+ cells). (a′) Original sample with a 3% fraction of positive cells, (a) nonmagnetic cell fraction with
a 1% fraction of positive cells, and (b) magnetic cell fraction with an 85% fraction of positive cells. This shows the significant enrichment
of target cells in the magnetic cell fraction (b) and the depletion of the target cells in the nonmagnetic cell fraction (a). Reprinted with
permission from ref 290. Copyright 1999 Elsevier.
showed that, for CD4+ cells, a 3-fold increase in the total
marker number per cell is observed, when comparing the
highest to the lowest fluorescence fractions. Similarly, a
4-fold increase in total marker number is observed for the
CD8+ cells. Also, CD56+ human natural killer cells were
separated from human peripheral blood using a two-step
antibody sandwich approach and a commercial batch-type
MACS system.298 The sorted cell fractions were subsequently
analyzed by FACS, and histograms of the magnetophoretic
mobilities of the sorted cell fractions, determined using a
microfluidic magnetic cell sorter, were presented.
To reduce the problem of cell loss due to adhesion to the
microchannel walls, a droplet-based magnetically activated
cell separator was presented.299 The system consists of a slide
with a pending droplet, from which magnetically labeled
constituents are retained to the slide surface by a magnet
positioned above the slide. The system was tested using
CD45+ cells that were separated from bone marrow cells
of mice and passed through a 30 µm Nylon mesh prior to
separation. The efficiency of cell separation was comparable
to that of commercial batch-type systems.
Stem Cells
Cell separation can also be used for the isolation of rare
hematopoietic progenitor cells from human umbilical cord
blood and mobilized peripheral blood with the subsequent
use of the cell isolates as a substitute for bone marrow
transplantation in patients having undergone irradiation and
high-dose chemotherapy.300,301 Because these cells are typi-
cally rare (0.1-3%), a commercial batch immunomagnetic
cell retention system was initially used to enrich the cells
prior to cell tracking velocimetry analysis.302 Within the
positive eluents from the MACS column, the range of the
magnetophoretic mobility was 50-fold, representing a plau-
sible 50-fold range in expression of surface CD34 antigen,
a marker for hematopoietic progenitor cells expression. It
was pointed out that monocytes could phagocytose the
magnetic nanobeads and become sufficiently magnetized to
be retained within the magnetic column, reducing the purity
of the positive eluent. A quadrupole magnetic separator was
used to determine a 4 × 104-105 ABC range of the CD34
antigen.267 These values were determined from a number of
clinical apheresis samples, obtained from patients that were
treated with a granulocyte-macrophage colony-stimulating
factor, a cytokine that functions as a white blood growth
factor. CD34+ progenitor cells could be separated to
Chemical Reviews, 2010, Vol. 110, No. 3 1539
64-95% purities and with a throughput of 107 cells/s in a
quadrupole magnetic flow sorter.303
Cancer Cells
The presence of circulating tumor cells (CTCs) in blood
is an indicator of metastatic disease in cancer patients. The
concentration of these cells in the blood has been shown to
correlate with the prognosis following treatment of breast
cancer.304 A count of one CTC per milliliter of whole blood
is related to a lower incidence of relapse following chemo-
therapy. Tissue-specific markers for these cells are often
missing, and the expression patterns of tumor- and epithelial-
specific antigens (Ags) as well as the cell size differ, making
it difficult to find a single reliable method for the isolation
of CTCs.305 As a result, detecting these rare disease-specific
cells from whole blood is an important challenge in cellular
separation technology. Several batch-type commercial sys-
tems have been used for the immunomagnetic separation of
CTCs from whole blood. However, there are concerns about
the limited load capacity and potential irreversible entrapment
of cells inside the magnetic column.298 Recoveries of target
cells in the 10-90% range were reported (see Table 1 of
ref 306 for an overview). Typically, breast cancer tumor cells
spiked in human307 or mouse308 blood were used as target
cells, labeled with both fluorescent markers and magnetic
nanoparticles for fluorescent microscopy or FACS analysis
after MACS separation.
Besides the batch-type magnetic separation systems, also
continuous flow magnetic separation systems have been used
to isolate rare (spiked) cancer cells from a blood sample.
MCF-7 human breast cancer cells were immunomagnetically
labeled by a two-step labeling protocol, similar to what was
discussed before.294,309 Briefly, first a primary mouse antibody
(Ab) against a cell surface Ag was allowed to bind to the
cell. This Ab can be conjugated to a fluorescent label to allow
FACS analysis after magnetic separation. Next, the cells are
labeled with a secondary rat antimouse polyclonal Ab
conjugated to magnetic nanoparticles. Both a magnetic
field294 and gravitation309 have been used to separate the
cancer cells from their matrix, due to differences in magnetic
susceptibility or cellular hydrodynamic diameter of the target
cells, respectively. A quadrupole magnetic flow sorter was
used to separate HCC1954 breast carcinoma cells overex-
pressing the HER-2/neu gene from peripheral blood leuko-
cytes.310 Labeled HCC1954 cells were mixed with unlabeled
(only fluorescent) leukocytes to form the spiked sample that
1540 Chemical Reviews, 2010, Vol. 110, No. 3 Gijs et al.

Figure 18. (A) Schematic representation of the experimental protocol for cell capture and sorting. (a) The channel was initially filled with
a suspension of superparamagnetic beads. (b) The external magnetic field was applied, and the beads were trapped by the Ni micropillars.
(c) Flow of buffer solution, which activates carboxyl groups on the surface of the beads and washes out any untrapped beads simultaneously.
(d) Protein solution was then introduced into the flow stream. (e) Proteins were attached to the beads, and any unbound protein was washed
out of the channel. (f) Cells were introduced into the channel. (g) Cancer cells were captured by specific protein-functionalized beads
anchored to the nickel micropillars. (h) The cancer cells captured by the beads were eluted from the channel, when the external magnetic
field was removed. (B) Covalent attachment of specific proteins to a magnetic bead and its capturing of a cancer cell. Reprinted with
permission from ref 313. Copyright 2007 Wiley-VCH.

was analyzed by cell tracking velocimetry. High-throughput
(3 × 105 cells/s) separations with 89% recovery of the
HCC1954 cells were reported. In a very detailed study,306
the same magnetic separator was used to separate MCF-7
cells from blood using a negative depletion protocol. The
protocol consisted of a red cell lysis step, immunomagneti-
cally staining leukocytes with an anti-CD45-phycoerythrin
(PE) Ab to provide them with a fluorescent label, binding
with the secondary anti-PE-magnetic nanoparticle-conjugated
Ab, immunomagnetic sorting using the flow-through system,
and a final cell analysis step using an automated cell counter.
The process produced an enrichment of the rare cancer cells
of 5 log10 and an average final recovery of 46%. This study
continued with the analysis of the peripheral blood of head
and neck cancer patients.311 After optimization, the process
was able to reduce the number of normal blood cells in a
cancer patient’s blood from 4 × 109 to 8 × 103 cells/mL
and still recover, on average, 2.3 CTC per mL of blood.
Spiking studies of a cancer cell line into normal blood and
subsequent enrichment using the optimized protocol indicated
an average recovery of 83%. The interesting thing of negative
depletion is that the rare cancer cells themselves are not
labeled, so that further molecular analysis of the nonma-
nipulated cell is still possible.
Other types of microfluidic devices have been proposed
to separate rare cancer cells from a blood sample, based on
immunomagnetic labeling. Human cervical cancer cells
(HeLa) were internally labeled with magnetic nanoparticles
following the endocytosis pathway, and fluorescence was
obtained by concomitantly labeling with rhodamine albu-
min.312 First, the magnetic particle uptake by the cells over
different incubation periods was studied. Cell populations
were subsequently sorted according to their acquired mag-
netic moment using a free-flow microfluidic magnetophoresis
device. Another type of integrated microfluidic device
consisting of nickel micropillars, microvalves, and micro-
channels was developed for specific capture and sorting of
cancer cells.313 A regular hexagonal array of micropillars was
integrated on the bottom of a microchannel. The Ni generated
a strong induced magnetic field gradient in the presence of
an external magnetic field to efficiently trap superparamag-
netic beads from a flowing stream. These beads were first
in situ biofunctionalized by covalent attachment of specific
proteins directly to their surface. Because of the microfluidic
approach, only submicroliter volumes of reagents and protein
solutions needed to be used. Figure 18 shows the experi-
mental protocol to in situ functionalize the superparamagnetic
beads in the microchannel and to capture A549 cancer cells
(a human lung carcinoma cell line) from a flow. Based on
the specific interaction between wheat germ agglutinin and
N-acetylglucosamine on the cell membrane, A549 cancer
cells were effectively captured on the magnetic beads. Before,
they were mixed with RBCs at a ratio of 1:10. The cancer
cells were sorted with a capture efficiency of 62-74%. For
the RBCs, no capture occurred under the same experimental
conditions.
HeLa-type and other types of cells were incubated with
ferromagnetic beads that were fluorescently labeled prior to
use.314 Because of the ferromagnetic, rather than superpara-
magnetic, character, magnetic forces on such beads are larger.
The cells showed a high uptake with 2-3 µm-size magnetic
beads. In other work,315 incubation and external labeling of
oral squamous carcinoma cells with 4.5 µm superparamag-
netic beads was reported, followed by microfluidic separation
of the labeled cells. Also, apoptopic cells were selectively
labeled with magnetic beads and subsequently sorted in a
microfluidic device.316 Apoptosis is the process of pro-
grammed cell death that may occur in multicellular organ-
isms. Apoptosis occurs when a cell is damaged beyond
repair, infected with a virus, or undergoing stressful condi-
tions such as starvation. As cells, Jurkat cells (a leukemia-
derived T-cell-line) were treated with apoptosis-inducing
Microfluidic Applications of Magnetic Particles
Figure 19. Separation of E. coli cells mixed with magnetic nanoparticles in a flow (from left to right) of PBS in a 200 µm wide channel.
The composite bright images were generated by overlaying sequential frames of corresponding movies taken at the beginning, middle, and
end (left to right) of the microfluidic channel, in the presence or absence of a neodymium disk magnet placed below the channel (bottom
and top images, respectively). Reprinted with permission from ref 321. Copyright 2006 Springer.
reagents. Apoptopic cells show expression on their surface
of the protein phosphatidylserine. This protein was biotiny-
lated and subsequently bound to magnetic beads grafted with
streptavidin molecules. Magnetic-activated cell sorting has
also been used to separate dead and apoptotic spermatozoa
using colloidal superparamagnetic nanoparticles (∼50 nm
in diameter) to improve sperm quality for assisted reproduc-
tive techniques.317 Jurkat cells have also been magnetically
deposited on the bottom of a microfluidic channel that was
held within an assembly of permanent and soft magnets.318
The Jurkat cells were labeled with both magnetic nanopar-
ticles and fluorescent labels using a two-site sandwich
immunoassay. A good magnetic capture efficiency (>99%)
and the combination with optical (fluorescent) examination
of the cell deposit are the advantageous features318 of this
system.
Bacteria
Magnetic separation and deposition of bacterial cells find
application in rapid detection of microbial contamination in
solutions such as environmental or industrial water, food,
and in clinical microbiology. The previous magnetic deposi-
tion system for Jurkat cells reminds of early work on the
magnetic deposition, quantitation, and identification of
bacteria incubated and reacting with paramagnetic trivalent
Er3+ lanthanide ions (i.e., not using magnetic beads).319,320
The magnetic deposition protocol allowed quantitative detec-
tion of Escherichia coli down to concentrations of 105
colony-forming units (CFU) mL-1. This magnetic deposition
of the bacterial cells on a microscope slide was examined
using dark field microscopy and by light scattering. In more
recent work,321 living E. coli bacteria have been separated
from flowing fluids in a microfluidic system, either alone or
when mixed with RBCs. In this study, 130 nm magnetic
beads were used to label E. coli bacteria (1 × 107 cells/mL)
by incubating the cells with biotinylated anti-E. coli Ab,
mixing them with the magnetic beads grafted with strepta-
vidin, and then injecting them into the source inlet of a
microfluidic circuit equipped with a magnet. Figure 19 shows
the separation of E. coli cells labeled with the magnetic
nanoparticles in a 200 µm wide microfluidic channel. At a
flow rate of 30 µL/h, 89% of the E. coli cells were separated
Chemical Reviews, 2010, Vol. 110, No. 3 1541
from phosphate buffered saline (PBS). In the presence of
RBCs, the collection efficiency was lower due to the
increased viscosity of the RBC containing fluid.
Another microfluidic system, containing a gold interdigi-
tated microelectrode array, was integrated with 145 nm
magnetic nanoparticle-Ab conjugates into an impedance
biosensor to detect pathogenic bacteria in beef samples.322
The magnetic nanoparticle-Ab conjugates were prepared by
conjugating streptavidin-coated magnetic nanoparticles with
biotin-labeled polyclonal goat anti-E. coli Abs and were used
in the magnetic separation and concentration of the target
bacteria near the microelectrode sensor. The impedance
sensor was able to detect in a time frame of 35 min as low
as 160 or 1200 E. coli cells present in a spiked food sample
of pure culture and ground beef, respectively. Another
microfluidic system with on-chip pumps323 had a detection
limit of 2000 CFU/mL and a maximum capture efficiency
from a PBS flow greater than 70% for E. coli cells. Off-
chip polymerase chain reaction (PCR) and capillary elec-
trophoretic analysis were used to determine the capture
efficiencies. Also, a multitarget magnetic activated cell sorter
was reported, which made use of microfluidic technology
to achieve simultaneous spatially addressable sorting of
multiple target cell types in a continuous-flow way.324 Two
types of E. coli cells were labeled via target-specific affinity
reagents with two different magnetic tags with distinct
saturation magnetization and size. The device was engineered
so that the combined effect of the hydrodynamic force (Fd)
produced from the laminar flow and the magnetophoretic
force (Fm) produced from patterned ferromagnetic structures
within the microchannel resulted in the selective purification
of the differentially labeled target cells into multiple inde-
pendent outlets. Figure 20 shows the separation architecture.
The system had the capability to simultaneously sort multiple
magnetic tags with >90% purity and >5000-fold enrichment
and multiple bacterial cell types with >90% purity and >500-
fold enrichment at a throughput of 109 cells/h. Magnetic
deposition microscopy, in comparison with conventional
blood smear tests, has been used to analyze blood samples
of individuals with Plasmodium falciparum malaria symp-
toms.325 Magnetic deposition microscopy increased detection
sensitivity of P. falciparum-infected, hemozoin-containing
1542 Chemical Reviews, 2010, Vol. 110, No. 3 Gijs et al.

Figure 20. Multitarget magnetic activated cell sorter separation architecture. (A) (Step A) The sample contains an excess of nontarget
cells and two different target cells (target 1 and target 2) that are labeled with two different magnetic tags (tag 1 and tag 2) by specific
surface markers. (Step B) The sample is continuously pumped into the device where the two target cell types are sorted into spatially
segregated independent outlets. Separation occurs in two regions of high magnetic field gradient generated by the microfabricated ferromagnetic
strip (MFS) 1 and MFS 2. (Step C) After being sorted, the eluted fractions from each outlet are analyzed via flow cytometry. (B) A
free-body diagram showing the balance of forces at the MFS structures. At MFS 1 (θ1 ) 15°), tag 1-labeled target 1 cells are deflected and
elute through outlet 1 because Fm1 > Fd1 sin(θ1). This is not the case for tag 2-labeled target 2 cells, which are instead deflected at MFS 2
(θ2 ) 5°) because Fm2 > Fd2 sin(θ2), and elute through outlet 2. Nontarget cells are not deflected by either MFS and elute through the waste
outlet. (C) Optical micrographs (magnification 100×) of the tags being separated at the 2 MFS structures at a total flow rate of 47 mL/h
(sample flow ) 5 mL/h, buffer flow ) 42 mL/h). (Left) Tag 1 is deflected by the steep angled MFS 1. (Right) Tag 2 is deflected by MFS
2. Reprinted with permission from ref 324. Copyright 2008 The National Academy of Sciences of the USA.

RBCs from infected humans, while maintaining detection
of ring-stage parasites.
5.5. Cell Biophysics
For applications like magnetic resonance imaging, hyper-
thermia, and separation, specific binding to or uptake of
magnetic beads by a certain cell type is the basis of the
magnetically induced contrast. Superparamagnetic particle
uptake kinetics of living cells and toxicity aspects of such
uptake have been discussed by several authors.326,327 Also,
the incorporation of magnetic beads into cells has provided
a new tool to measure cytoskeleton-associated cell func-
tions.328,329 Superparamagnetic 30 nm beads coated with
monovalent ligands and bound to transmembrane receptors
of mast cells mediated the cellular signal transduction, when
the cell was brought into an external magnetic field.330
Magnetic particle twisting was used to investigate the
mechanical integrity and viscoelastic properties of the
cytoskeleton,331 and, in combination with optical tweezers,
superparamagnetic beads were at the basis of force and torque
measurement systems on single cells.332 Recently, cell-bound
magnetic microparticles, subjected to 0.1 ms magnetic field
pulses, have been used to destruct the targeted cells by
penetration of the beads into the cells or by rupturing the
cells with the beads,333 a nonthermal process that is different
from hyperthermia.
While one could argue that microfluidic aspects in the
studies cited in previous paragraph are not essential, microf-
luidic chips have been used to handle cells for applications
that go beyond the separation and purification of a certain
type of cell from a matrix. For example, a microfluidic
system has been developed for trapping yeast cells.334 Single
cell analysis of budding yeast cells is required to understand
genomic changes that serve as models of human aging and
disease. An automated system to capture individual yeast
cells in a microfluidic device and to analyze each cell as it
buds has been developed. Magnetic capture was chosen to
selectively capture old yeast cells because yeast cells that
are labeled with magnetic beads retain the beads over
multiple divisions and do not pass them onto daughter cells,
which form new cell walls. For labeling, the cells were
incubated with biotin and thereafter incubated with strepta-
vidin-coated 50 nm superparamagnetic beads. A Ni-Co-B
alloy magnetic microstructure was plated into a PDMS chip
and used to capture the magnetic bead-labeled yeast cells
from a flow. In another study,335 no labeling with magnetic
beads was required to trap yeast cells in a locally weakened
magnetic field, obtained by combining an electrical current-
induced field with a uniform magnetic field. Trapping of the
cells was enabled by placing the latter in a biologically
benign ferrofluid matrix with high magnetic susceptibility.
An integrated circuit/microfluidic hybrid system336 was
developed to manipulate and concentrate a small number of
Microfluidic Applications of Magnetic Particles Chemical Reviews, 2010, Vol. 110, No. 3 1543

Figure 21. Magnetic targeting of magnetic nanoparticle-preloaded bovine aortic endothelial cells (BAECs) under flow conditions in vitro
and in vivo. (a) In vitro capture kinetics of magnetically responsive BAECs onto a 304-grade stainless-steel stent in the presence of a
uniform field of 0.1 T and a nonpulsatile flow rate of 30 mL/min. The data were obtained by measuring the fluorescence of the magnetic
nanoparticles. (b and c) Magnetically responsive BAECs captured in vitro onto a 304-grade stainless-steel stent as evidenced by the red
fluorescence of the magnetic nanoparticles (b) and Calcein green staining of live cells (c). (d) Magnetic nanoparticles-loaded BAECs
captured in vivo onto a deployed 304-grade stainless-steel stent in the rat carotid artery. BAECs preloaded with fluorescent magnetic
nanoparticles were transthoracically injected into the left ventricular cavity. Animals were exposed to a magnetic field of 0.1 T for 5 min,
including the period of injection. The animals were killed 5 min after delivery, and the explanted stents were immediately examined by
fluorescence microscopy. (e) Control rats underwent an identical procedure where no magnetic field was used. (Magnification: b-e, ×40.)
Reprinted with permission from ref 338. Copyright 2008 The National Academy of Sciences of the USA.

individual cells with tight spatial control, with a potential to

study cell-cell interactions at the single cell level. Yeast
cells attached to magnetic beads were suspended inside the
microfluidic system, and a microcoil array was used to
produce spatially patterned microscopic magnetic fields that
were reconfigurable in time. Also, bovine capillary endot-
helial cells containing 250 nm diameter magnetic beads have
been manipulated by this hybrid system.337 These cells were
cocultured with peptide-coated magnetic beads, which lead
to the uptake of the beads by the cells through endocytosis.
In a work aiming at creating tissue on stents using magnetic
nanoparticle activation,338 superparamagnetic nanoparticle-
loaded bovine aortic endothelial cells (BAECs) could be
magnetically targeted to steel stent wires in an in vitro study.
The cells were first preloaded with biodegradable polymeric
290 nm superparamagnetic nanoparticles, having a fluores-
cent label, and subsequently brought in the proximity of a
steel stent placed in an external magnetic field. Figure 21
shows some of the experimental results of this study. Also,
in vivo cells transduced with adenoviruses expressing lu-
ciferase were targeted to stents deployed in rat carotid
arteries. A significantly increased luciferase expression was Figure 22. Schematic of the basic steps of an on-chip magnetic
detected in an external magnetic field with respect to the nucleic acid assay using magnetic beads as a substrate for the assay.
expression for a nonmagnetic control experiment.

6. Magnetic Nucleic Acid Assays
An application area where magnetic beads play a standard
role in bioanalytical procedures is that of nucleic acid assays.
In macroscopic lab-bench protocols, magnetic beads are
generally used as mobile substrates for the capture and
extraction of nucleic acids.339,340 With the increasing interest
in the combination of microfluidics and magnetic beads, the
step toward on-chip processing of nucleic acids was logical.
Recent publications demonstrate the feasibility of miniatur-
izing magnetic nucleic acid assays, while maintaining the
procedures and protocols known from batch-type applica-
tions. Figure 22 summarizes the steps followed in an assay,
where magnetic beads can be applied with differing purposes
and at different phases of the assay. The first phase of the
protocol involves the specific capture of the molecules of
interest and is followed by a washing or purification step to
remove the matrix of unbound molecules. Some systems use
the magnetic character of the beads in the detection step,
where their stray field is measured for example using a
magnetoresistive sensor. The processing step, between
purification and detection, requires the miniaturization of the
labeling or PCR amplification steps and is important for
performing a complete on-chip nucleic acid assay. All of
these steps are detailed in the following sections.
6.1. DNA Capture and Purification
An important task in on-chip nucleic acid analysis is the
capture and purification of the molecules of interest. Using
magnetic microbeads in a microfluidic system, the nucleic
acids can be brought into contact with the particle surface
via different means. A popular solution is the incubation of
activated magnetic particles with the sample in a reservoir.341,342
Here, the capture of the nucleic acids is driven by diffusion,
and any subsequent purification step requires the introduction
1544 Chemical Reviews, 2010, Vol. 110, No. 3
of a magnetic field to separate the magnetic beads from the
sample. Other solutions introduce a relative velocity between
the sample and the magnetic particles by either immobilizing
the latter and flushing the sample through or over them.
Among the published systems, the majority is based on the
immobilization of the magnetic particles inside a microchan-
nel before or after the capture step.343-345 The magnetic
particles are hereby held against a flow via a magnetic field
perpendicular to the channel, a principle already illustrated
in Figure 11. The source of the magnetic field is either a
permanent magnet placed underneath the system342,343,346 or
electromagnetic elements integrated into the chip.344 An
inverse approach is the capture of the DNA via the actuation
of magnetic beads inside a small sample reservoir.347 Hereby
the particles can be actuated magnetically via either external
magnets,347 electromagnets,238,348 or passive mixing.345 The
two latter solutions have the advantage of being easily
integrated into a system with subsequent DNA processing
and detection.
Magnetic beads have also been used in a microfluidic chip
for in vitro selection of aptamers from a library.349 Aptamers
are nucleic acid molecules, either RNA or DNA, that bind
to molecular targets with high affinity and specificity. Once
the nucleic acid sequence of the aptamer is identified for a
particular target, it can be produced synthetically, a distinct
advantage over traditional affinity reagents such as antibod-
ies, which require biological processes such as cell culture.
The starting single-stranded DNA (ssDNA) library consisted
of ∼1014 unique sequences, each containing a 60-base
internal randomized region flanked by two 20-base PCR
primer-specific sequences. The target protein, Botulinum
neurotoxin type A, was conjugated to the magnetic beads
through carbodiimide coupling. Target-conjugated beads
were incubated with the heat-treated ssDNA library, and
aptamers that bound to the target protein were separated in
a microfluidic device. The aptamers bound on the target-
coated beads were subsequently amplified via PCR, and
single-stranded products are generated. Finally, the binding
kinetics of the resulting aptamers was measured.
6.2. DNA Processing
After being captured and purified, the nucleic acids can
be subjected to a range of processing steps; they can be
hybridized to add labels for a subsequent detection step or
amplified via a PCR protocol.350 The magnetic beads, which
carry the captured sample molecules, are hereby usually
immobilized in a magnetic field, while primers and labels
are added to the system.351 Generally, such systems combine
Figure 23. (a) Schematic illustration of magnetic tweezers integrated with microelectromagnets, a ring trapper, a fluidic channel, and a
gold-patterned surface. (b) A tethered-DNA magnetic bead is in equilibrium under the action of the magnetic force, DNA elastic force, and
the gravitational force. (c) Stretching of a single DNA molecule linked to thiol-modified beads. Reprinted with permission from ref 356.
Copyright 2006 The Institute of Physics Publishing.
Gijs et al.
the capture, purification, and processing of the DNA on-
chip, because the magnetic actuation of the particles is the
same for the different steps.352 Alternatively, in systems not
employing the particles as a stationary phase, the latter can
be used to concentrate and enrich the purified sample via
local electromagnets before subsequent processing steps, thus
improving the system’s sensitivity.344,353,354 Another interest-
ing approach of using magnetic beads for the hybridization
of DNA was presented by Heer et al.,355 whose system
employs the particles solely for accelerating the binding of
the target DNA to the capture probes attached to the detection
site via magnetic stirring.
In addition to the labeling and enrichment, the captured
DNA molecules can be subjected to mechanical forces via
the magnetic particles. Chiou et al.356,357 presented an
interesting system where single DNA molecules are rotated
and stretched using magnetic tweezers, as demonstrated in
Figure 23. One end of the DNA molecules was hereby
anchored to the chip surface, while the other end was attached
to a magnetic microparticle. Subsequently, any actuation of
the particle was translated into a mechanical force acting on
the DNA molecule. Thus, the mechanical properties of a
single DNA molecule can be easily studied.
Besides being captured at the surface of magnetic micro-
particles, the DNA can also be processed via a loose
interaction with the microparticles. Magnetic SPS within a
microfluidic channel were used as a porous separation
medium for the separation of long DNA molecules. The
structure and porosity of the SPS is a strong function of the
microfluidic channel dimension, particle properties, applied
magnetic field, etc. Gel electrophoresis is the standard method
for separation of DNA by length. However, the efficiency
of gel electrophoresis deteriorates seriously for DNA mol-
ecules longer than about 40 000 base pairs (40 kbp). This
phenomenon was understood in terms of electric field-
induced aggregation of the DNA by the electrical dipole-dipole
interaction.358 Slab gel pulsed-field gel electrophoresis or
pulsed-field capillary gel electrophoresis, using time-varying
drive voltages, can be used to separate the longer chains of
DNA.359 The use of self-assembled magnetic SPS for long
DNA separation in microfluidic channels represents a
convenient solution, because no microlithography is required
to define geometrical constrictions that are simply formed
by the porous magnetic matrix.360 Experimental separations
using the SPS stationary phase have been combined with
theoretical modeling.361 This approach has been presented
in more detail by Minc et al.,362-364 in work where a network
of columns of magnetic microparticles leads to the length-
Microfluidic Applications of Magnetic Particles
dependent separation of DNA in the direction of an elek-
trokinetic flow. Another innovative use of magnetic micro-
particles for the processing of purified DNA samples has
been presented by Park et al.,365 who use hyperthermia for
heating the liquid sample for the different temperature stages
of a PCR procedure. Using small aqueous droplets, these
authors showed that it is possible to heat a discrete liquid
volume to a temperature of 80 °C within 5 min by applying
an oscillating magnetic field. The applicability of the heating
of small droplets for on-chip PCR has already been presented
in the context of magnetic droplet manipulation systems,366,367
where a self-contained liquid sample is transported toward
local heating elements via magnetic particles enclosed in the
droplet. The capture of the DNA molecules can be performed
before or after the PCR steps, depending on the subsequent
detection step.
6.3. DNA Detection
The on-chip detection of nucleic acids offers the
advantage of high sensitivity in combination with a high
selectivity, depending on the chosen protocols and detec-
tion methods. With respect to the latter, two different
principles are generally employed: optical and magnetic
detection.368 Optical detection methods are usually based
on fluorescent tags, which label the molecules of interest
that are captured and concentrated via the magnetic
particles.346,369 Also, a multilayer fluorescent labeling strategy
was proposed to amplify the fluorescent intensity: a DNA
detection limit of 0.25 fmol/mL was achieved.370 The
detection signals could be amplified using multilayers of
biotin-streptavidin conjugated quantum dots based on the
binding with a specific biotinylated linker. In contrast,
magnetic detection directly uses the magnetic beads as
labels354,371,372 or indirectly for the deformation of small
cantilevers.373 A mixed form relies on the optical detection
of the magnetic labels, which can carry fluorescent tags374
or simply influence the opacity of the detection site.375
For the optical detection, the magnetic particles are usually
concentrated at the detection site via external magnets or
electromagnets. Because the strength of the optical signal
indicates the amount of captured molecules, the trapped
magnetic particles need to be washed thoroughly. An optical
detection of the spatial distribution of magnetically labeled
DNA strands, as proposed by Tierno et al.,376 relies on
contrast in the magnetic transport of the labels and would
not require washing. Similarly, the direct magnetic detection
of the particles also profits from the effect that the sensing
element can at the same time act as a trap for the magnetic
particles.354 The magnetic character of the beads allows
hereby their detection via integrated sensors, which are based
Figure 24. Magnetoresistive detection of the magnetic labels. (a) Micrograph sequence of a single 2 µm bead (indicated by the black
arrow) crossing the sensor area. (b) Recorded sensor signal during the passage of a single 2 µm magnetic bead over a spin-valve GMR
sensor with 3 mA sense current. The indicated numbers correspond to the sequence numbers in (a). Reprinted with permission from ref
378. Copyright 2005 Institute of Electrical Engineering.
Chemical Reviews, 2010, Vol. 110, No. 3 1545
either on the GMR effect192 or on the Hall effect.377 Here,
the magnetic particles will be detected, either after being
bound to the sensor surface by a DNA hybridization step or
during the passage over the sensing element. Some publica-
tions demonstrate the possibility of detecting single particles
using the magnetoresistive method;198,202,378 see Figure 24,
as well as using miniaturized Hall sensors.377,379 The advan-
tage of these is the possibility of their direct integration into
the microfluidic system,354 which is an important step toward
the full on-chip DNA capture and detection.
Detection of PCR-amplified DNA on a GMR sensor was
demonstrated using superparamagnetic particles as detection
label.380 The one-step assay was performed on an integrated
and miniaturized detection platform suitable for application
into point-of-care devices. A double-tagged PCR amplifica-
tion product of the LamB gene of the Escherichia coli
bacterium was used to investigate binding kinetics of the
assay. Biological dose-response curves detecting 4-250 pM
amplicon concentrations in a one-step format in total assay
times of less than 3 min were presented. Using various
tag-antibody combinations specific for one of the individual
genes, multianalyte detection was shown for several antibiotic
resistance genes of the food pathogen Salmonella.
6.4. Integrated DNA Analysis Systems Starting
from Cells
Systems have been presented that combine all steps
summarized in Figure 22, starting from crude cell samples.
The majority of such systems rely on the simple trapping of
a plug of magnetic microparticles inside a microchannel via
an external magnet.352,381-383 The sample, washing, and
labeling reagents are subsequently passed though the blocked
array of particles, and the fluorescent signal can be measured.
Alternatively, the magnetic particles can also be separated
from the sample solution after the incubation step, as
presented by Lien et al.384,385 and shown in Figure 25. In
other work by the same authors, a lysed cell solution was
loaded onto a microfluidic chip, which subsequently per-
formed automatic RNA extraction and reverse transcription
processes.386 Total RNA was successfully extracted and
purified from the human β-actin gene extracted from T98
cells, while analysis of the PCR products was done off-chip
via gel electrophoresis.
Most of the integrated nucleic acid assays, summarized
in Table 3, start from samples containing cells or viruses.
These are either collected via the magnetic particles and lysed
before the amplification step,385,387,388 which results in free
nucleic acids in the solution, or lysed before the capture of
the nucleic acids, which are then bound to the magnetic
particles and processed further.83,346,352,382,383,389 Generally,
1546 Chemical Reviews, 2010, Vol. 110, No. 3 Gijs et al.

Figure 25. Integrated reverse transcription (RT)-PCR system. (a) Schematic illustration of the integrated RT-PCR chip. Several components
including a microtemperature module, a bead collection module, and a microfluidic control module are integrated onto a single chip. (b)
Photograph of the RT-PCR chip. The width and length of the chip are 40 and 60 mm, respectively. Note that micropumps, microvalves,
electromagnets, microheaters, and microtemperature sensors were integrated on the same chip. Reprinted with permission from ref 385.
Copyright 2007 Elsevier.

Table 3. Summary of the Main Characteristics of Representative On-Chip Nucleic Acid Assays Employing Magnetic Beadsa
magnetic manipulation step
magnetic assembly for
separation after
detection principle external magnet droplet handling hybridization on
incubation
detection zone
fluorescent simultaneous 10 cells from crude cell high density array, 2 ×
examination of eight sample analyzed in 20 108 beads per cm2 390
DNA samples383 min83
liposome label for
signal amplification, 10
fmol/mL detection
limit346
surface coverage fluidic shear force visualization of
against nonspecific hybridization array by
binding, 1 pmol/mL specific magnetic
detection limit391 particle binding375
electrochemical liposome label fully integrated system,
containing potassium start from crude cell
ferri/ferrohexacyanide samples, pathogenic
and interdigitated bacteria, and SNP
electrodes for detection, detection382
10 µmol/mL detection
limit392
magneto-resistive detection limit of 10 integrated electronics,
µg/mL target DNA368 detection limit10
pmol/mL or 16 pg on
surface378
2.5 mmol/mL target 10 nmol/mL detection
DNA detected using limit, 140 DNA
MgO-based tunnel molecules per sensor394
junction sensor393
PCR amplification of sample integrated chip, slab-gel 500× preconcentration,
for high sensitivity, electrophoresis analysis, detection of 1-50 RNA
detection of SNP352 virus detection limit: copies389
100 PFU/mL385
electrochemical mRNA lab-on-a-CD, Hepatitis
detection for viability of B virus and E. coli, 10
E. coli, down to 100 DNA copies/µL
CFU/mL342 detected388
DNA recovery from 1 34 ng of total mRNA
E. coli cell395 isolated from 10 µg of
total RNA343
a
The systems are categorized on the basis of the magnetic manipulation step and the detection principle. SNP: single nucleotide polymorphism.
PFU: plaque-forming unit. CFU: colony-forming unit.

systems employing PCR as the detection step rely on the systems that use the magnetic particles as labels for the
magnetic capture and purification of cells and viruses with detection perform the capture steps after the cell lysis. Several
a subsequent lysis step to release the nucleic acids, while integrated systems combine the different features; for
Microfluidic Applications of Magnetic Particles Chemical Reviews, 2010, Vol. 110, No. 3 1547

example, PCR amplification can be combined with fluores-

cent detection on the same chip.

7. Magnetic Immunoassays
Magnetic bead-based immunoassays performed in the
microfluidic format have attracted particular interest, because
of the multiple advantages and promising potential applica-
tions.26,27 Basically, an immunoassay consists of using
antibodies (Abs) as chemical reagents to analyze target
molecules, called antigens (Ags). The technique is based on
the recognition of a target antigen (t-Ag) by its Ab. When
brought into contact, Ag and Ab form, due to specific and
strong molecular interactions, a stable immunocomplex
structure (see Figure 26a). Immunoassays are among the most
important techniques used for biological molecule analysis.
They are widely used both in research and in analytical
sciences and have been explored for many applications,
Figure 26. (a) Schematic representation of the reaction between ranging from environmental analysis to clinical diagnosis.
an antibody (Ab) and its specific antigen (Ag) to form an AbAg
immunocomplex. (b) Schematic representation of sandwich immu-
The technique can be divided in two subclasses: that of
nocomplex formation. A surface-linked capture antibody (c-Ab) is competitive and noncompetitive assays, respectively. In a
reacting with its target antigen (t-Ag). A labeled detection antibody competitive immunoassay, the t-Ags to be detected compete
(d-Ab) is then reacting with the t-Ag. (c) Schematic representation with known labeled Ags (Ags*) for immunocomplex forma-
of an immunocomplex formed with a d-Ab labeled with an enzyme. tion. The detection and quantification of the immunocomplex
The conversion of the enzyme substrate to its product by an
enzymatic reaction leads to the formation of a detectable molecule. formed with Ag* permits one to deduce the amount of t-Ag
present in the sample. In a noncompetitive immunoassay,
Table 4. Summary of the Different Application Areas for Microfluidic Magnetic Bead-Based Immunoassays and Overview of
Literature Results, Quoting the Detection Method, the Type of Assay, the t-Ag That Is Analyzed, and the Detection Limit of the Assay
detection technique type of assay target Ag detection limit ref
Proof of Concept
fluorescence direct fluorescein isothiocyanate (FITC) 3.9 µg/mL 207
fluorescence direct biotin-4-fluorescein 1 µg/mL 381
fluorescence direct IgG2a (mouse) 10 µg/mL 381
fluorescence (evanescent field detection) sandwich IgG (rabbit) 120 ng/mL 396, 397
fluorescence sandwich IgG (goat) 5 ng/mL 398
fluorescence direct fluorescently labeled biotin 0.2 µg/mL 399
fluorescence direct antistreptavidin IgG (rabbit) 12.5 pg/mL 400
fluorescence direct antistreptavidin IgG (rabbit) 100 pg/mL 401
fluorescence sandwich mouse IgG 10 pg/mL 402
fluorescence sandwich mouse IgG 250 pg/mL 96
luminescence sandwich human insulin and interleukin-6 250 pg/mL 403
colorimetric Ab transport antimouse IgG (rabbit) 150 ng/mL 404
electrochemical sandwich mouse IgG 50 ng/mL 405
electrochemical sandwich mouse IgG 16 ng/mL 406
surface coverage direct antiovalbumine IgG (rabbit) 10 ng/cm2 407
magnetic force discrimination sandwich mouse IgG 250 pg/mL 408
magnetic force discrimination sandwich IgG (rabbit) 15 ng/mL 408
GMR direct antimouse IgG (goat) n.a. 409
agglutination test direct protein A 1-2 pg/mL 410
magnetic frequency relaxation direct rabbit IgG n.a. 411
Brownian relaxation direct biotinylated T7 phage n.a. 412, 413
surface coverage sandwich streptavidin, IgG 1 fg/mL 414
agglutination test biotinylated bovine serum albumin (BSA) 100 pg/mL 415
fluorescence kinetics of chemical reaction 416
agglutination test kinetics of biomolecular recognition 417
Recombinant Protein Dosing from Cell Culture
fluorescence sandwich IgG (mouse) 1 ng/mL 206
fluorescence of single bead on silicon chip sandwich IgG (mouse) 1 ng/mL 418
Disease Marker Detection
fluorescence sandwich parathyroid hormone 1.4 ng/mL 207
fluorescence sandwich interleukin-5 2 ng/mL 207
fluorescence (evanescent field detection) sandwich interleukin-4 10 ng/mL 397
fluorescence sandwich tumor necrosis factor R 45 pg/mL 419
surface coverage and GMR magneto-sandwich S100ββ protein 27 pg/mL 420
surface coverage magneto-sandwich S100ββ protein 0.2 ng/mL 421
magnetic force discrimination sandwich D. farinae IgE (human) 85 pg/mL 422
magnetic force discrimination sandwich D. pteronyssinus IgE (human) ∼0.04 ng/mL 422
surface coverage sandwich West Nile Virus 700 viral particles 414
surface coverage sandwich Staphylococcus enterotoxin B 1 fg/mL 423
surface coverage via surface plasmon resonance sandwich Staphylococcus enterotoxin B 100 pg/mL 424
fluorescence and Hall effect sandwich dengue virus IgG not reported 425
IgG (mouse) 1 ng/mL
fluorescence sandwich dengue virus 103 PFU/mL 426
electrochemical sandwich various tumor markers 0.5 ng/mL 427
GMR sandwich parathyroid hormone 4 pg/mL 428
inductive sandwich Troponin I (cardiac marker) 0.5 ng/mL 429
optical reflection (evanescent field) sandwich Troponin I <20 pg/mL 430
drugs of abuse <1 ng/mL
1548 Chemical Reviews, 2010, Vol. 110, No. 3
only the t-Ags of the sample are involved in the immuno-
complex formation. The quantitative measure of t-Ag is then
directly obtained by measuring the total amount of immu-
nocomplex formed during the experiment. Both immunoas-
say subclasses have been employed in the microfluidic
format. However, due to efficiency and simplicity, microf-
luidic magnetic bead-based immunoassays have been prin-
cipally developed in the noncompetitive way.
Depending on the experimental design, the number of Abs
involved in the immunocomplex formation reaction is
varying. A t-Ag, like a bacteria toxin, can be directly
immobilized on a reaction substrate and can be quantified
with a labeled detection antibody (d-Ab) in a direct immu-
noassay, involving a single type of Ab. This technique is
however limited, because t-Ags for direct surface im-
mobilization have to be available. A more flexible technique,
called sandwich immunoassay, consists of flanking the t-Ag
to be detected between a capture antibody (c-Ab) linked to
a reaction substrate and a labeled d-Ab (see Figure 26b). In
this technique, the t-Ag is specific for both c-Ab and d-Ab.
It has to be noted that the t-Ag often is an Ab. Also, an
enzyme can be used to label the d-Ab, and, in this case,
flowing the enzyme substrate through the magnetic bead
reaction chamber leads to the formation of a detectable
product (see Figure 26c). An enzyme is a biomolecule able
to convert its specific substrate to a product by catalyzing a
biochemical reaction. The advantage of this technique is the
possibility to generate many detectable molecules with only
a few immunocomplexes, increasing thereby the detection
sensitivity.
The success of magnetic bead-based immunoassays in the
microfluidic format can be explained by their multiple
advantages. A recurrent issue in microfluidic immunoassays
is to control the surface chemistry of the microchannel wall
to allow a reproducible c-Ab surface binding. This drawback
is circumvented when using magnetic beads on-chip, because
particle surface chemistry can be perfectly controlled off-
chip prior to use of the beads. Furthermore, the limited
microfluidic reaction chamber volume (in the nL-µL range)
permits one to efficiently concentrate a small amount of
magnetic beads in a confined space. As a result, the number
of captured t-Ag is concentrated at a small defined location,
thereby enhancing the detection signal intensity. Finally, the
magnetic beads can be easily recovered after analysis for
further applications or analysis. Table 4 summarizes the
principal microfluidic magnetic immunoassay results reported
in literature, categorized by application area and highlighting
the detection method, the t-Ag analyzed, and the detection
limit. The cited papers are discussed in more detail further
in this section.
7.1. Magnetic Beads as Substrate
When magnetic beads are used as a reaction substrate, the
immunocomplexes formed during the assay are linked to the
particle surface. The particle surface chemistry has to be
strictly controlled to allow a reproducible immunocomplex
formation. In a direct immunoassay, the t-Ags are attached
to the particle surface, and specific d-Abs are linked to the
particle via the t-Ags. However, when a sandwich immu-
noassay is executed, c-Abs have to be linked to the particle
surface prior to t-Ag capture. This can be done using a
covalent chemical bond between the activated particle surface
and a chemical group conjugated to the c-Ab. The use of
physicochemical interactions or molecular recognition (for
Gijs et al.
example, forming the biotin/streptavidin complex) are popu-
lar techniques for grafting c-Abs to the magnetic particle
surface. The immunoassay is then executed by flowing
solutions of sample and reagents to the reaction chamber
containing the magnetic beads, followed by signal detection.
The latter can be done using a d-Ab labeled with a detectable
tag (for example, a fluorescent dye). However, when an
additional chemical reaction is needed to generate a detect-
able molecule (for example, in case of an enzymatic
reaction), the signal is detected in the bulk solution.
7.1.1. Direct Fluorescent Detection
Direct signal detection is performed by recording the
fluorescence generated by the labeled d-Abs. This detection
principle benefits from the high sensitivity offered by
fluorescence. In early work, a single rare earth permanent
magnet positioned at the top of a capillary was used to
generate a strong magnetic field gradient within a micro-
capillary. This led to a magnetic force that retained a densely
packed bed of magnetic beads in the microcapillary (see
Figure 27a).207 Flowing appropriate solutions of target sample
and reagents resulted in the formation of detectable immu-
nocomplex at the magnetic bead surface. A direct assay was
demonstrated by direct interaction of FITC with an im-
mobilized anti-FITC conjugate. Also, heterogeneous sand-
wich assays were demonstrated for parathyroid hormone and
interleukin-5 (IL-5) down to concentrations of a few ng/mL
(see Figure 27b). Magnetic bead positioning using external
magnets was also used as a reliable solution to trap magnetic
beads at a glass surface in a method that was called
“magnetically assisted transport evanescent field fluoroim-
munoassay”. For excitation of the fluorescent labels on the
beads, an evanescent wave was generated by total internal
reflection of a laser beam at the optical interface between a
prism and the sample396 or at the surface of a planar
waveguide.397 The waveguide was combined with a multi-
channel microfluidic device for application of sample and
reagents. Rabbit-IgG and interleukin-4 were detected with
detection limits of 120 and 10 ng/mL, respectively.397 The
evanescent technique may reduce or eliminate washing steps,
as interference from fluorescent species in the sample matrix
is strongly minimized. Separation and washing steps could
also be eliminated by configuring the microfluidic device
with an interdigitated electrode microarray for dielectro-
phoretically moving the various Abs and t-Ag to the detection
area.402 A detection limit of 10 pg/mL was reported using
mouse IgG as t-Ag.
Also, pairs of small permanent magnets with opposite
poles were positioned close to a microcapillary, to retain
magnetic beads injected in the capillary. Such system was
used to purify Immunoglobulin E from serum followed by
analysis by capillary electrophoresis.431,432 Also, three dif-
ferent pairs of magnets were used to create three plugs with
differently functionalized magnetic beads in an effort toward
a multiplex immunoassay.381 In other work, retention of
magnetic beads in a microfluidic channel was achieved, not
by focusing the field at a specific location across the channel,
but by periodically varying the width of the channel. Upon
the action of a constant magnetic field, magnetic beads
introduced in the microchannel self-assembled in magnetic
chains (see Figure 11c). The self-assembled magnetic chains
are then positioned over the entire cross-section of the
microchannel, strongly enhancing the liquid-particle interac-
tion, as demonstrated by direct and sandwich immunoassays.
Microfluidic Applications of Magnetic Particles
Figure 27. (a) Micrograph of part of a packed bed of 1-2 µm diameter magnetic particles within a 50 µm diameter fused silica capillary
in the presence of a 0.236 T magnetic field. The total packed bed is ∼1.2 mm in length and has ∼2.4 nL volume. (b) Calibration curve of
interleukin-5 (IL-5) generated from a sandwich immunoassay performed on the magnetic particles. Biotinylated rat monoclonal antimouse
IL-5 was immobilized as c-Ab on the magnetic particles, to bind to a mouse IL-5 t-Ag; FITC-conjugated rat monoclonal antimouse IL-5
was the d-Ab. Reprinted with permission from ref 207. Copyright 2001 American Chemical Society.
A 1 ng/mL detection limit was obtained for mouse IgG. The
assay was executed either integrally on-chip while consuming
nanoliter volume of sample and reagents,206,220 or part of the
protocol was performed off-chip.433 Another promising
solution to retain magnetic beads was to combine solenoids,
for generating a current-activated controlled magnetic field,
with a microfluidic device for the execution of a low-volume
immunoassay.398 When placed close to the microchannel, two
solenoids (obtained by winding copper wires on a ferromag-
netic bar) generated a strong magnetic field gradient that was
sufficient to trap the magnetic beads in a flowing stream and
execute immunoassay experiments. Goat IgG was analyzed
in a sandwich immunoassay with a 5 ng/mL detection limit.
In contrast to their use for magnetic bead retention, perma-
nent magnets can also be used to generate a magnetic force
to displace magnetic particles perpendicularly to multiple
colaminar streams containing sample, washing, and reagent
solutions.399,434 In this way, complete immunocomplex
formation or multistep biochemical processes can be imple-
mented. A magnetic bead-based microflow cytometer with
integrated sample pretreatment module was used for the
purification, concentration, detection, and collection of target
viruses (dengue virus serotype 2).426 Instead of using PCR
techniques, the system proposed to detect the target virus
by using magnetic bead-based flow cytometry. By sandwich
magneto-immunocomplex formation, target viruses in blood
or serum could be captured and separated magnetically,
followed by a sorting step based on the presence of a
fluorescent Ab for identifying the surface t-Ag of the target
viruses. Virus samples with a concentration of 13 plaque-
forming unit/mL (PFU/mL) could be detected.
Also, a sandwich immunoassay using streptavidin-coated
beads as substrate and completely performed on-chip was
presented.96 The beads were electrostatically self-assembled
on aminosilane micropatterns at the bottom of a microfluidic
channel. Mouse IgG diluted in PBS with 1% bovine serum
albumin (BSA) solution was used as t-Ag and detected down
to a concentration of 15 ng/mL in stop-flow mode and 250
pg/mL in continuous flow mode, using 1.3 µL of sample
volume. In other work, a monolithic and fully integrated
complementary metal oxide semiconductor (CMOS) chip
was presented for the manipulation and detection of single
fluorescent magnetic beads in a PDMS microchannel posi-
tioned on top of the chip.418 Magnetic manipulation was done
by current actuation of microcoils on the silicon chip, and
detection was achieved using single photon avalanche diodes
that are located in the center of each microcoil and count
the fluorescent photons originating from a single magnetic
Chemical Reviews, 2010, Vol. 110, No. 3 1549
bead. This approach permitted microscope-less fluorescence
detection with high sensitivity. Detection of murine mono-
clonal Abs with a limit of 1 ng/mL was reported in a
noncompetitive sandwich immunoassay, performed using the
bead surface as assay substrate.
7.1.2. Enzyme Reaction-Based Detection
Enzymatically Generated Fluorescence. Important work
in miniaturized enzymatic reaction-generated fluorescence
detection was reported by Hermann et al.419 These authors
have developed a microfluidic system enabling the execution
of parallel ELISA experiments (see Figure 28). In their
system, the magnetic beads are trapped in a reaction chamber
by an external permanent magnet and capture the t-Ag of
interest. This step is followed by the generation of a
fluorescent signal during an enzymatic reaction in stop flow
conditions. Additionally, the magnet is displaced along
microfluidic chambers, dragging the magnetic beads with it
and allowing an efficient mixing. To reduce the assay noise
due to nonspecific adsorption of Abs to the microchannel
wall, a dual network of channels between which the magnetic
beads are transferred is used. In a first network, the
immunocomplexes are formed by flowing appropriate solu-
tions of sample and reagents, while the enzymatic reaction
is performed in the other channel. The networks are linked
using fluidic bridges controlled by pressure activated valves,
ensuring the complete isolation of the two chambers and
avoiding contamination. After the enzymatic reaction, the
generated fluorescent solutions are driven away from the
magnetic beads (simultaneous stopping all reactions) to
parallel microchannels for detection. An off-chip antistrepta-
vidin immunocomplex was formed by direct linking model
and detection Ab to streptavidin-coated magnetic nanopar-
ticles.400 This direct immuno-assay was quantified on-chip
using enzymatically generated fluorescence down to an
antistreptavidin Ab concentration of 12.5 pg/mL. The same
authors have also demonstrated on-chip the formation of an
immuno-complex consisting of two antibodies401 on magnetic
particles. Antistreptavidin was used as a model Ab and was
detected down to a 100 pg/mL concentration. Detection of
tumor necrosis factor-R (TNF-R) was demonstrated down
to 45 pg/mL in a complete assay time of less than 1 h.419
Other authors have presented a droplet-based assay, where
magnetic bead manipulation was combined with electrowet-
ting, and showed detection of human insulin and interleukin-6
with a detection limit of ∼250 pg/mL using enzymatically
generated chemiluminescent detection.403
1550 Chemical Reviews, 2010, Vol. 110, No. 3 Gijs et al.

Figure 28. Schematic of the immunocomplex formation and detection in a dual channel network microfluidic system. (a) About 106
streptavidin-coated magnetic beads are trapped inside the complexation chamber. The antistreptavidin antibody (t-Ag) binds specifically to
the streptavidin-coated beads during a 5 min incubation period with mixing. Similarly, the alkaline phosphatase (AP)-coupled d-Ab is
added to form the reactive immunocomplex. (b) The valve is transiently opened, and the reactive beads are magnetically transferred into
the reaction chamber. (c) The AP enzyme processes the fluorescein diphosphate (FDP) substrate into the fluorescent molecule FITC. While
the reaction takes place, the solution is homogenized by displacing the beads. (d) The reacted solution is then pushed into the detection
area. (e) Microfluidic ELISA results obtained without pretreatment of the microfluidic channel against nonspecific adsorption: simultaneous
fluorescent detection in all eight channels of the device. (f) Standard curves for the quantification of antistreptavidin antibodies. Each point
represents the average obtained in four separate experiments. Reprinted with permission from ref 401. Copyright 2007 The Royal Society
of Chemistry.

Figure 29. (a) Enzymatic reaction for electrochemical detection: p-aminophenyl phosphate (PAPP) is converted by the enzyme AP, which
is the label chosen for the d-Ab, to the electrochemical product p-aminophenol (PAP). The latter is then converted at the interdigitated
electrodes to 4-quinoneimine in a 2-electron oxidation process. (b) Immunoassay results measured by chronoamperometric detection of the
conversion of PAP. Reprinted with permission from ref 405. Copyright 2002 The Royal Society of Chemistry.

Colorimetric Detection. In colorimetric detection, the
products obtained after enzymatic conversion are chromogen
and have negligible fluorescence. This technique is less
sensitive than fluorescence, but can be developed using a
normal microscope equipped with a camera. Its application
for an immunoassay in the microfluidic format was demon-
strated by Lehmann et al.404 in a water-in-oil droplet-based
system. The system consisted of a two-dimensional array of
coils disposed under an oil bath with a bottom Teflon layer,
surrounded by strong permanent magnets to generate a
permanent magnetic field perpendicular to the Teflon surface.
The latter can be locally hydrophilized by an air plasma to
form several 10 µL reaction zones, in which water droplets
containing sample and reagent solutions were immobilized.
When activating the appropriate coils, the magnetic beads
move within an aqueous droplet along the chip surface and
between the different droplet reservoirs. This unique dis-
placement technique permits one to bring into contact
successively t-Ag (down to a 150 ng/mL concentration) with
enzyme labeled d-Ab, and enzyme substrate, without con-
tamination between each of the steps. Another colorimetric
reaction in a water-in-oil droplet-based system was the
hydrolysis of p-nitrophenylphosphate substrate (down to 500
µg/mL) using alkaline phophatase as the enzyme.435
Electrochemical Detection. When the enzymatic reaction
converts the enzyme substrate to a detectable electrochemical
product, the immunoassay can be detected using electro-
chemistry. The advantage of this technique is the possibility
to convert a biological recognition reaction in an electrical
parameter that can be directly detected by an electronic
sensor and system for further signal analysis. Experimentally,
a potential is applied between electrodes soaked in the sample
solution, and a current due to the oxidation of the electro-
chemical product is detected. As the transport of the
electrochemical product from the bulk solution to the
electrode surface is limited by diffusion, the immunocomplex
has to be located close to the electrochemical sensor for good
detection efficiency.
An efficient solution to localize c-Ab-coated magnetic
particles at the bottom of a microchannel is to use a planar
electromagnet constituted of serpentine coils embedded in
an electroplated permalloy structure.405 When activated, the
coil generates a magnetic field that magnetizes and traps the
magnetic beads. Thereafter, solutions of t-Ag and d-Ab
Microfluidic Applications of Magnetic Particles
labeled with the enzyme p-aminophenyl phosphate (PAPP)
are introduced in the microchannel, and immunocomplexes
form. The t-Ag detection is performed by flowing the enzyme
substrate through the microchannel. The substrate is con-
verted in an electrochemically detectable product, and
electrochemical detection is performed using interdigitated
microelectrodes (see Figure 29) placed on top of the coils.
This way, the detection limit for mouse IgG, which was
chosen as the t-Ag, was <50 ng/mL. Another solution for
magnetic bead trapping in a lab-on-a-chip module was to
use an electroplated permalloy microarray at the bottom of
the microfluidic chip, flanked by two strong permanent
magnets.406 When covered by a microfluidic channel, the
beads are trapped at the location of the magnetic posts. Once
immunocomplexes were formed, an array of interdigitated
microelectrodes disposed at the top of the magnetic pattern
was activated for the electrochemical detection (detection
limit 16 ng/mL for mouse IgG). Another group proposed an
electrochemical immunosensor that pulls t-Ag bound to
magnetic nanoparticles from a laminar flow to a gold
electrode by applying a local magnetic field gradient.427 This
selectively removes the particles from flowing biological
fluids without any washing steps. A set of five disk-shape
gold electrodes coated with c-Ab and an Ag/AgCl reference
electrode was used to detect changes in electrode potential
due to the binding of the t-Ag with the c-Ab. As t-Ag, four
different tumor markers were detected in a simultaneous
multiple assay (detection limit 0.5 ng/mL), by using four
types of magnetic nanoparticles coated with specific Abs.
The complex formation between the t-Ag at the magnetic
bead surface and the d-Ab immobilized on the gold electrode
permits a time-dependent observation of the electrical
potential between the electrodes, the response of which is
directly related to the tumor marker concentration. Note that
no enzyme reaction was necessary in this potentiometric
detection approach.
7.2. Magnetic Beads as Label for Detection
Besides the utilization of magnetic particles as reaction
substrate in an immunoassay, they have also shown promis-
ing results when used as label. The principle of using
magnetic particles as detection label in an immunoassay is
presented in Figure 30. The method consists of first coating
the surface of the biosensor with a c-Ab (using, for example,
techniques developed for microtiter plate-based immunoas-
says) and blocking the remaining sites that are not occupied
Figure 30. Schematic of immunocomplex labeling with a
magnetic nanoparticle. The sandwich immunocomplex with
magnetic nanoparticle label is linked to the reaction substrate.
The immunocomplex formation is detected by detecting the
presence of the particle.
Chemical Reviews, 2010, Vol. 110, No. 3 1551
by c-Ab. Next, the t-Ags are flown through the reaction
chamber, and the immunoreaction between c-Ab and t-Ag
occurs. A solution of magnetic particles functionalized by
the d-Ab that is complementary to the t-Ag is then introduced
in the chip, forming a magneto-sandwich. The formation of
the immunocomplex, and thereby the presence of t-Ag, is
revealed by detecting the presence of the magnetic particles.
As the quantity of immobilized particles will depend on the
number of formed immunocomplexes, a quantitative assay
is possible. Different techniques for detecting the presence
of the magnetic particles, essentially of optical and/or
magnetic nature, have been investigated.
7.2.1. Surface Coverage Measurement
The simplest method to detect surface-linked magnetic
particles is to measure optically the fraction of the substrate
surface that is covered, and then calculate the number of
immuno-magnetosandwich complexes that are formed. Be-
cause the number of particles attached to the surface can
be, in principle, as low as the number of immunocomplexes,
the technique potentially is very sensitive. However, non-
specific surface adsorption of the magnetic particles can pose
a problem and asks for strategies to obtain a minimal
background signal. A perfect control of the surface chemistry
and optimized washing protocols are therefore very impor-
tant. Parameters like the surface zeta potential were optimized
to allow an optimal c-Ab binding, while different types of
magnetic particles, which either bind specifically or non-
specifically to the substrate, were studied.421,436 Techniques
to mechanically remove nonspecifically bonded magnetic
beads have been investigated.407,437 This permits one to
reduce the background noise during measurements and
therefore increases the test sensitivity. Different approaches
using liquid flow shear forces, mechanical pulling using a
magnet, or centrifugation have been proposed.407,437 A
potential bottleneck is that the transport of highly diluted
analytes from the bulk of the solution to the sensor surface
due to normal diffusion is slow, so that only a small fraction
of the analyte can reach the sensor surface in a reasonable
time. This effect strongly limits the sensitivity of the system.
To accelerate the molecular transport beyond what is possible
by diffusion, external forces can be applied, for example,
electrophoretic forces, for analyte concentration near the
surface.414 In other work, 500 nm magnetic beads were
specifically attached to an optically transparent and func-
tionalized sensor surface.430 A specific immunoreaction
resulted in a surface coverage with beads that was quantified
using the internal reflection of a light beam from the back
of the surface. The cardiac marker Troponin I could be
detected down to concentrations of the order of 20 pg/mL
and drugs of abuse down to a concentration of 1 ng/mL,
taking as samples blood plasma and whole saliva. Overall,
the most important advantage of the surface coverage
measurement technique is the absence of the need for a
complex detection system. In principle, a camera mounted
on a microscope and connected to a computer equipped with
image treatment software is sufficient.
Multiple and parallel immunoassays were demonstrated
with good sensitivity.407,437 The protein S100ββ, a marker
for stroke and minor head injury, was detected down to 2
pg/mL.420,421 300 nm magnetic particles in combination with
strong rinsing and a blocking step allowed for the best
specific and sensitive detection of S100ββ in serum over a
wide concentration range. Also, an ultrasensitive, two-step
1552 Chemical Reviews, 2010, Vol. 110, No. 3 Gijs et al.

Figure 31. Dependence of spot brightness on the number of analyte molecules passed through the flow cell and on the magnetic bead
shear flow. The spot brightness is due to the presence of immobilized magnetic particles and is observed by dark-field optical microscopy.
(A) Streptavidin (SA) molecules captured on anti-SA-IgG array from buffer solution (O), from the same buffer containing 1% dialyzed
chicken serum (b), and 10% chicken serum (9) at a flow rate of 20 µL/min. (B) Dependence of the signal on the shear flow upon
electrophoretic capturing of SA for 4 min. Numbers over points denote flow rate in µL/min. (C) Antiovalbumin-IgG captured on an ovalbumin
array at a flow rate of 60 µL/min. (D) Assay of viruses in serum from West Nile virus (WNV)-infected chickens. The serum was dialyzed
against water, diluted 1:200 with buffer, and WNV antigens were captured on an array of anti-WNV-IgG (b). The same procedure was
followed for a sample purified by exclusion chromatography (O). The flow rate of the sample solution was 20 µL/min. The numbers above
the experimental points in panels A, C, and D denote the time of capture in min. Reprinted with permission from ref 414. Copyright 2007
American Chemical Society.

immunoassay method was developed for rapid assay of
proteins and viruses using dark-field optical microscopy
detection of immobilized magnetic beads.414 In the first step,
electrophoresis was used to quickly bring soluble analytes
from a flowing solution to a microarray of probe molecules
immobilized on a semipermeable membrane. In a second
step, the captured analyte was detected by scanning the
microarray with functionalized magnetic beads, which passed
over the array surface by shear flow and were pressed to the
array surface by a magnetic field. This method routinely
detected analytes in concentrations as low as 20 aM (1 fg/
mL) (500-1000 molecules detected) within 2-3 min in the
presence of a 1011-fold excess of other protein molecules in
the sample (see Figure 31), which is a truly impressive result!
In the same line, other authors described a semihomogeneous
implementation of a fluidic force discrimination assay,
detecting in 10 min the toxin RCA (Ricinus communis
agglutinin) and a staphylococcal enterotoxin B with a 35 aM
(1 fg/mL) detection limit.423 In a pure sequential assay,391
the c-Abs, which are immobilized on a surface, capture t-Ag,
a secondary Ab, and Ab-conjugated magnetic bead labels.
In the mentioned semihomogeneous assay, the t-Ag, second-
ary Ab, and the beads are first mixed in solution, and then
applied to the substrate covered with c-Ab. Nonspecifically
bound beads were removed by controlled laminar flow fluidic
forces, and the remaining beads were counted to determine
the t-Ag concentrations.
7.2.2. Magnetic Properties Measurement
Manipulating and positioning functionalized magnetic
beads in microfluidic devices and monitoring the surface
coverage is clearly very interesting for high-sensitivity
immunoassays. An alternative way of detecting the presence
of the beads on a surface, rather than optically observing
them, is to probe the changes in the magnetic environment
due to the presence of the beads. This is an interesting
approach, because magnetic sensors can be integrated in the
chips, and the detection signal can be converted on-chip for
analysis using electronics. Furthermore, the magnetic proper-
ties of the particles are, in most cases, independent from the
immunoassay protocol. A disadvantage for disposable ap-
plications, however, is that this will add cost to the system,
when comparing with the surface coverage method, due to
sensor fabrication (often on a silicon substrate) and integra-
tion of the sensor in a microfluidic package.
Magnetic Force Discrimination. A first approach to use
the magnetic properties of the magnetic beads is to link a
fluorescent polystyrene microparticle (1-5 µm diameter)
coated with c-Ab in an immunoreaction via a t-Ag, to
magnetic nanoparticles (10-50 nm diameter) functionalized
with d-Ab. Applying then a magnetic field gradient into a
microfluidic device induces a displacement of the micropar-
ticle-nanoparticles complex. The trajectory of such complex
can be measured by fluorescent microscopy and directly
correlated to the presence of t-Ag in the sample. Without
t-Ag, no displacement of the fluorescent particle is observed.
This concept was exploited to detect the presence of rabbit
IgG and mouse IgG with a detection limit of 250 pg/mL
and 15 ng/mL, respectively, in a microfluidic device.408 The
latter consisted of a microchannel (connected to a sample
inlet and particles inlet), to which a magnetic field gradient
is applied using a permanent magnet placed at the side of
the microchannel. This measurement principle was further
implemented by integrating a nickel electroplated magnetic
core to enhance the particle deviation. The detection of
human Immunoglobulin E involved in allergen reactions was
Microfluidic Applications of Magnetic Particles
Figure 32. Optical micrographs showing the presence of magnetic
particles on the sensor surface for different concentrations of the
protein S100ββ. The positioning of the magnetic particles before
(at random) and after (aligned) release and alignment with the edges
of the GMR sensors is illustrated. A scale bar is shown in the top
left image. Reprinted with permission from ref 420. Copyright 2007
American Chemical Society.
performed in clinical samples from patients down to a
concentration of 40 pg/mL.422
Detection with a Giant Magnetoresistance (GMR)
Sensor. A sensitive magnetic particle detection technique is
to use the GMR effect.191,409 When a magnetic bead is
positioned over or near a GMR sensor, the resistance of the
magnetoresistor decreases due to the stray field of the
magnetic bead, and the presence of the particle can be
directly correlated to the detected signal.191,372,391,409,420 This
detection technique can also be associated with surface
coverage measurements and can benefit from all advances
made with the surface coverage technique, such as surface
chemistry control, or nonspecific magnetic bead removal. For
example, a GMR chip of the type shown in Figure 10 has
been used to detect 312 pM (10 ng/mL) RCA in serum. De
Palma et al.420 have demonstrated a highly sensitive biosensor
by executing immunoassay experiments in a two-step
protocol. In a first step, magnetic beads are used to label an
optimized magneto-sandwich immunoassay at the surface of
a biosensor. After nonspecific bonded magnetic beads are
removed, the number of remaining beads is proportional to
the t-Ag concentration. Next, in a second step, the remaining
beads are released from the substrate using a NaOH solution
and concentrated using magnetic forces to a GMR sensor
array. This magnetic attraction permits one to relocate the
magnetic beads on the most sensitive parts of the GMR
sensors (see Figure 32). This technique has allowed the
development of an immunoassay for S100ββ proteins with
a two-decade dynamic range and a detection limit of 27 pg/
mL. Other authors proposed a compact GMR biosensor
system that uses actuated magnetic bead labels incorporated
in an immunosandwich.428 For the measurement of parathy-
roid hormone, a detection limit in the 10 pM range (40 pg/
mL) was achieved with a total assay time of 15 min, when
300 nm diameter magnetic beads were used. For 500 nm
particles (with a longer-range magnetic stray field), the
detection sensitivity was 4 pg/mL.
Other Magnetic Bead Detection Principles. An array
of Hall sensors has also been used to measure the presence
of magnetic particles that were substrate-linked via an
immunocomplex. The Hall chip was realized in CMOS
technology and contained 120 sensor elements.425 A sensitiv-
ity of 1 ng/mL was obtained for the detection of mouse IgG,
and also the detection of human antidengue virus IgG in
serum was reported. Another possibility for detecting substrate-
Chemical Reviews, 2010, Vol. 110, No. 3 1553
linked magnetic particles is by placing detection coils
underneath the substrate.429 When the magnetic particles were
present and magnetized, the inductive load of a coil changes,
and the coil inductance or resonance frequency of the coil
circuit (around 7 MHz) provided a measure for the concen-
tration of a t-Ag. As model system, Troponin I, a cardiac
marker, was detected down to 0.5 ng/mL.
In other work, the frequency dependence of the imaginary
part of the magnetic susceptibility of magnetic nanoparticles
suspended in a liquid allowed measurement of the rotational
Brownian relaxation for 40 nm diameter magnetite particles
covered with an avidin layer. This concept was used to
demonstrate experimentally the correlation between the peak
shift of the AC imaginary part of the nanoparticle suscep-
tibility and its hydrodynamic radius, after binding biotinylated
proteins anchored to T7 bacteriophage particles on the avidin
shell.412,413 The nanoparticle hydrodynamic diameter could
also be optically measured by monitoring the frequency-
dependent magnetic relaxation signal.411 When biomolecules
are bonded to the nanoparticle surface, an increase of
hydrodynamic diameter is observed. This diameter change
was observed in an iron oxide ferrofluid exposed to goat
IgG.
A portable system for rapid purification, concentration, and
detection via surface plasmon resonance (SPR) of target
analytes from complex matrixes was presented using antibody-
coated superparamagnetic nanoparticles.424 The SPR signal
due to surface-bound Staphylococcus enterotoxin B was
dramatically increased when superparamagnetic particles
were used as detection amplifiers: a concentration of 100
pg/mL was easily detected in both buffer and stool samples.
7.3. Agglutination Tests
Agglutination tests exploit the fact that dispersed particles
that are grafted with c-Ab form aggregates when they are
mixed with a t-Ag-containing fluid. This liquid medium can
be serum or blood, and agglutination tests are therefore
generally simple, cheap, and do not require sophisticated
equipment, nor highly specialized skills. A number of
agglutination protocols and grafting techniques have been
presented, including the use of magnetic particles.56 In a
magnetic colloid, the particles have a repulsive interaction,
which in principle precludes formation of doublets or larger
clusters. However, when a magnetic field is applied, the
particles can approach each other and form chains. Irrevers-
ible and specific particle clustering or doublet formation is
observed after switching off the magnetic field. Detection
of these formed clusters, when the agglutination test is
performed in a microtiter plate (i.e., a nonmicrofluidic
format), is by turbidimetry.438 For example, a limit of
detection of 1 pmol/L (45 pg/mL) was reported for ovab-
ulmine in phosphate buffer.439 The method also leads to
kinetic information on immunoreactions. A magnetic mi-
croparticle agglutination test in a PDMS microfluidic channel
was reported for the detection of biotinylated Protein A.410
Although dose-response curves were not provided, the first
experiments suggested that analytical sensitivities in the
∼1-2 pg/mL range are achievable. This interesting result
is due to the fact that, by forcing the analyte through a limited
number of magnetic particles using microfluidic flow, sample
capture is much more efficient than when operating the
agglutination assay in the batch format. Other work presents
the dynamic actuation of a confined plug of functionalized
magnetic beads in a microchannel allowing for analyte
1554 Chemical Reviews, 2010, Vol. 110, No. 3
capture with improved efficiency.415 Here, an original dif-
fusion-based detection method of the aggregated particles
was presented. The feasibility of an on-chip agglutination
assay was demonstrated by means of the streptavidin/
biotinylated-BSA model system. A detection limit of 100
pg/mL confirms the high potential of this approach.
Other authors presented a laminar-flow based device for
separation of individual magnetic nanoparticles (5 nm) from
agglutinated clusters.440 The isolated 5 nm magnetic nano-
particles did not exhibit significant magnetophoretic veloci-
ties, but did exhibit high magnetophoretic velocities when
aggregated by the action of a pH-responsive polymer coating.
A simple external magnet is used to magnetophorese the
aggregated particles that have captured a t-Ag (streptavidin)
from a lower pH laminar flowstream (pH 7.3) to a second
higher pH flowstream (pH 8.4) that induces rapid disaggre-
gation of the clusters, after which the individual magnetic
nanoparticles are transported to the output chamber. This
stimuli-responsive reagent system has been shown to transfer
81% of the streptavidin from an input flowstream to a second
flowstream in a continuous flow mode.
8. Catalytic Applications
8.1. Homogenizing Heterogeneous Catalysis
Using Magnetic Particles
A field where magnetic particles have potential to play
an important role is catalysis. Introduced in small quantity
as compared to other reagents in the system, the catalyst is
a substance that enhances the reactivity between reagents,
but is not consumed during the chemical reaction. Catalysis
is generally separated into two subclasses: homogeneous
catalysis, where the catalysts are totally spread (dissolved)
in the reaction media, and heterogeneous catalysis, where
they are attached to a supporting matrix or surface. Homog-
enous catalysis is the most efficient, as all catalytic sites are
easily accessible to reagents. However, homogeneous ca-
talysis suffers from an important drawback: it is difficult to
separate the catalyst from reaction products at the end of
the reaction process.441 Separation methods like distillation
require high temperature, which may cause thermal stress
that degrades the catalyst molecules, while other separation
processes (for example, solvent extraction or chromatogra-
phy) result in catalyst loss. Also, final reaction products are
often contaminated by catalysts. As the last ones are in
general toxic (for example, a heavy metal), contamination
has to be strictly limited.
Heterogeneous catalysis partially overcomes these draw-
backs. Here, the catalysts are immobilized at the surface of
Table 5. Application of Magnetic Nanoparticles as Support for the Catalysis of Typical Chemical Reactions
reaction
atom transfer radical polymerization
epoxidation
Heck
hydroformylation
hydrogenation
hydrogenation
hydrogenation
hydrolysis (biocatalyse)
hydrolysis (biocatalyse)
hydrolysis (biocatalyse)
Sonagashira
Suzuki
PenicilinV degradation
Gijs et al.
a supporting (and when possible soluble) matrix and are
recovered after reaction by separation (for example, precipi-
tation and/or filtration) from the reaction products. However,
a subsequent decrease of catalytic activity and selectivity is
generally observed during reactions. It is due to steric effects
from the support, which limits reagent diffusion to the
surface-anchored catalysts. Also, the bonds between the
catalyst and the ligand are often broken and reformed during
catalytic reactions. If this happens, the catalyst may break
away from the support and become dissolved. This “leach-
ing” problem leads to loss of activity of the catalyst, when
it is used with a continuous liquid flow. Reduced leaching
has been observed when the catalyst was anchored inside
the pores of zeolites or of mesoporous solids.441
Magnetic nanoparticles can be used as the support for
catalysts and can facilitate their separation from the reaction
media.442-446 However, the low surface energy and easy
aggregation of magnetic nanoparticles can hinder their
practical applications. To overcome these drawbacks, mag-
netic nanoparticles have also been introduced in various
mesoporous silica structures.447,448 These types of magnetic
nanoparticle-based materials, which combine the advantages
of both mesoporous silicas and magnetic particles, are
potentially very interesting supports for the immobilization
of catalysts. Table 5 presents typical applications of magnetic
nanoparticles in catalysis. The cited papers are discussed
further in more detail, but here we will not give an exhaustive
overview of the use of magnetic nanoparticles in catalysis,
as most of the work deals with the application of a permanent
magnet to recover the magnetic nanoparticle-supported
catalyst from the reaction mixture. The role of microfluidics
in this field is still minor.449,450
8.2. Transition Metal Catalysts
In many catalysis applications, a transition metal is used
as catalyst to transform or bind reagents. The good catalysis
activity of those elements is due to their ability to have
various oxidation states helping electron transfer or, in the
case of metals, to adsorb other substances onto their surface
and activate them in the process. Palladium (Pd) is employed
to catalyze various chemical reactions, such as hydrogena-
tion,456,457 Heck,445,453,454 Suzuki,445,446,463 or Sonagashira445
cross-coupling reactions. When using magnetic nanoparticles,
different strategies are employed to link Pd atoms to the
magnetic nanoparticle surface. The metal can be either
complexed using ligands, chemically bonded, or directly
immobilized or incorporated to the nanoparticle surface using
a supporting matrix. The functionalized magnetic nanopar-
ticles demonstrate a high conversion yield in the reported
catalyst ref
CuBr 451
Mo 452
Pd 445, 453, 454
Rh 442, 455
Pd 456, 457
Pt 458
Ru 443
Candida rugosa lipase 459, 460
pair of amino acid residues 461
2-pralidoxime 462
Pd 445
Pd 446, 463
β-lactamase 464
Microfluidic Applications of Magnetic Particles
Figure 33. Schematic representation of the immobilization of chiral Ru catalyst on magnetic nanoparticles. Reprinted with permission
from ref 443. Copyright 2005 American Chemical Society.
applications, with a minor loss of catalysis efficiency after
several runs. An interesting approach was demonstrated by
Zheng et al.,446 who have used functionalized 4 nm magnetic
particles to transport and immobilize the catalyst into a
conventional solid-phase resin. After reaction, the catalysts
are recycled from the solid phase reaction support using
magnetic forces, while reaction products are chemically
separated from the resin. Chiral catalysis, using ruthenium
atoms complexed to the magnetic nanoparticle surface, was
demonstrated for the hydrogenation of aromatic ketones.443
Heterogenized asymmetric catalysts were synthesized by
immobilizing preformed Ru catalysts on magnetite nano-
particles via the phosphonate functionality (see scheme of
Figure 33). The immobilized catalysts were easily recycled
by magnetic decantation and reused for up to 14 times
without loss of activity and enantioselectivity.
Rhodium (Rh) is also a well suitable transition metal that
can be used for catalysis. Complexed to the surface of
magnetic nanoparticles using appropriate ligands455 or den-
drons grown on silica-coated nanoparticles,442 Rh was used
to catalyze hydroformylation reactions. The advantage of
using dendrons is the increase of the solubility of the catalysis
body in the reaction medium. Excellent reaction yield and
enantioselectivity were observed for both immobilization
methods. To fully exploit the large surface area of magnetic
nanoparticles, a magnetic cobalt core surrounded by a shell
of Pt metal was used for the hydrogenation of several
compounds.458 The catalytic activity of the particles dem-
onstrated high efficiency, except for reagents with unfavor-
able steric interactions with the magnetic nanoparticle
surface. Under favorable conditions, the catalytic activity still
reached a yield of 100% after seven cycles of reaction and
recycling. Used, for example, in olefin epoxidation, molyb-
denum is another transition metal involved in many catalysis
reactions. Using a mesoporous matrix generated by the
calcination of sol-gel polymerization products, Mo atoms
were attached to a magnetic nanoparticle surface.452 The
catalytic activity of the composite nanoparticles was dem-
onstrated by the epoxidation of several alkenes with high
yield, even after six cycles of recycling. Besides the use of
functionalized magnetic nanoparticles for small molecule
reactions, the polymerization of methyl methacrylate by atom
transfer radical polymerization was investigated using mag-
netite nanoparticles to support a copper(I) bromide (CuBr)
catalyst.451 Under optimized conditions, the catalytic activity
of the functionalized nanoparticles was kept at a high level.
Furthermore, adjusting the experimental conditions permitted
one to synthesize a block copolymer with controlled mo-
lecular weight and low polydispersity.
Chemical Reviews, 2010, Vol. 110, No. 3 1555
8.3. Biocatalysts
Thanks to advances in magnetic nanoparticle functional-
ization with organic molecules or proteins, magnetic particles
have the potential to support biocatalysts and enzymes.
Biocatalysis can be highly selective and permits one to
transform organic compounds. Different biocatalysts are
available to functionalize magnetic nanoparticles. Candida
rugosa lipase has been investigated as biocatalyst for the
hydrolysis of p-nitrophenolbutyrate.459 Results have shown
a lower enzymatic activity for the immobilized enzyme, as
compared to the free one. However, the enzymatic activity
of the functionalized magnetic nanoparticle was kept constant
during several runs (with a slight activity decrease of 2%
due to enzyme denaturation after a period of 14 days) and
found to be higher as compared to lipase immobilized on
standard micrometer-size polymeric microbeads. To exploit
the selectivity of Candida rugosa lipase enzyme, the resolu-
tion of racemic carboxylates was demonstrated on molecules
substituted by different alkyl chains.460 The enantiometric
excess was systematically superior to 99%, as measured by
gas chromatography (comparable to those found for the free
solution enzyme). In other work, the immobilization of
β-lactamase by chemical linkage to a magnetic nanoparticle
surface was used for the degradation of PenicilinV in
phosphate buffered water.464 The enzyme activity was found
to be 54% as compared to the free enzyme in solution, with
a slight decrease in efficiency after catalysis reaction.
An interesting approach for the use of magnetic nanopar-
ticles as biocatalyst support was developed by Zheng et al.461
These authors have used the nanoparticles to support pairs
of amino acid residues for the hydrolysis of pesticides. The
crucial choice of the biomolecular pair permits one to mimic
natural biocatalysts for the cleavage of phosphor ester and
carboxylic ester bonds, without using severe pH or transition
metal-based catalysis. Furthermore, the functionalized par-
ticles kept their catalytic activity after several months of
synthesis. In the same area, the immobilization of common
antidote, 2-pralidoxime, to a magnetic nanoparticle surface
permitted it to degrade by hydrolysis organophosphorus
pesticides responsible for serious diseases.462 With an activity
comparable to that found for transition metal-based catalysts,
this method permits one to develop a safe and nonpolluting
alternative for the destruction of pesticides, without the use
of extreme pH or experimental conditions. Other authors have
electrochemically detected the pesticide carbofuran, by
investigating the inhibition by this pesticide of the enzyme
acetylcholinesterase immobilized on magnetic beads. Car-
bofuran could be detected down to a few ppb thanks to this
on-chip enzymatic inhibition reaction.465
1556 Chemical Reviews, 2010, Vol. 110, No. 3 Gijs et al.

Figure 34. Experimental setup used for the protein digestion experiments. (A) Picture of the PDMS chip with the microchannel (0.25 ×
20 × 1 mm3) and its two inlet-outlet plastic tubings. (B) Expanded view of the magnetic beads immobilized between the two magnets in
the microchannel. (C, left) Details of the magnetic beads arrangement at an early and late stage of microplug formation. The orientation of
the bead arrays parallel to the channel’s axis, providing a low flow resistance and uniform pore size, is visible (white scale bar, 5 µm). (C,
right) View of the whole plug at the beginning and end of the plug’s formation at a lower magnification (black scale bar, 200 µm). Reprinted
with permission from ref 468. Copyright 2008 The Royal Society of Chemistry.

8.4. Magnetic Particle-Based Protein Digestion
A field of activity, where microfluidics has clearly gained
a stronghold, is that of proteolysis or protein digestion.
Proteolysis is the directed degradation (digestion) of proteins
by cellular enzymes, called proteases, or by intramolecular
digestion. For example, trypsin is a serine protease found in
the digestive system, which predominantly cleaves peptide
chains at the carboxyl side of the amino acids lysine and
arginine, except when either is followed by proline. It is used
for numerous biotechnological processes. Trypsin is com-
monly used in biological research to digest proteins into
peptides for mass spectrometry analysis. It can also be used
to dissociate dissected cells (for example, prior to cell fixing
and sorting). A promising field, in which magnetic particles
in combination with microfluidics can play a determinant
role, is the development of microreactors for proteomic
applications.466 Magnetic particles that are grafted with
enzyme very efficiently proteolyze the proteins of interest,
by simply flowing the latter through the porous magnetic
plug. The practical advantage of the microfluidic magnetic
particle-based digestion system is that the magnetic matrix
can be easily replaced by flushing out the beads and loading
new beads into the same channel. Also, grafting on particles
can be performed ex situ in large quantities, allowing for
reduced cost and better reproducibility.
Grafted trypsin magnetic beads were used in a microfluidic
channel for performing protein digestion. Retained as self-
assembled magnetic particle chains in a microchannel,
thereby forming a porous matrix, the particles are used to
digest several types of protein samples from a flow for
subsequent analysis.467 Figure 34 shows the experimental
microfluidic channel with the porous magnetic particle plug.
Kinetics studies of the hydrolysis of a model peptide show
a 100-fold increase in digestion speed obtained by the
microfluidic system when compared to a batch-type system.
High performance and reproducibility for digesting recom-
binant human growth hormone are confirmed by analyzing
the digested products by both capillary electrophoresis and
matrix-assisted laser desorption/ionization time-of-flight mass
spectrometry (MALDI-TOF).467 Anhydrotrypsin, an inactive
form of trypsin, with an affinity for certain peptides was
grafted on magnetic microparticles to digest a model peptide,
human neurotensin, for subsequent mass spectrometry analy-
sis.467 The same system was used for proteinase K-mediated
protein digestion as a step toward the elaboration of a fully
integrated device for the detection of pathological prion
protein.468 Reaction kinetics were typically accelerated 100-
fold as compared to conventional batch reactions. The
proteolytic microsystem was also applied to the digestion
of prion protein from brain tissues, and the difference in
digestion efficiency between the normal and pathological
form was very significant. Other authors reported on the use
of trypsin-functionalized magnetic particles of various size
(50-300 nm diameter) to digest cytochrome c, BSA, and
myoglobin as model proteins, followed by MALDI-TOF
analysis.469-471 In contrast to the conventional digestion
approach of using free trypsin for cytochrome c, with an
incubation time of 12 h, proteins can be digested on-chip in
a matter of minutes.
An allergenic epitope extraction technique, based on
functionalized magnetic microparticles self-organized under
a magnetic field in a microfluidic channel, was applied for
the isolation and identification of prospective allergen
epitopes.472 Ovabulmin, the major protein of egg white and
a typical representative of food allergens was selected as the
model molecule. In a supplementary step after digestion,
capture of allergenic epitopes from the mixture of peptides
was performed by a second magnetic immunoaffinity carrier
with immobilized antiovalbumin IgG molecules. Captured
peptides were eluted subsequently and analyzed with MALDI-
TOF.
9. Conclusions and Outlook
We have reviewed advances in the fabrication, manipula-
tion, detection, and application of magnetic particles in
microfluidic systems. This research area is highly multidis-
ciplinary, requiring scientific knowledge, ranging from
inorganic chemistry involved in the preparation of the
magnetic beads, through biochemistry and medical science
to allow for their functionalization, to the basic physics of
magnetism and magnetic materials. Today many reaction
processes are known for the synthesis of magnetic micro-
and nanoparticles of various size and composition. These
particles can be very reproducibly obtained from a number
of suppliers with a tremendous choice in chemical function-
alization of the particle surface. Nearly all important func-
tions in a bioassay can be realized using magnetic beads:
raw sample purification, providing a solid substrate to the
sample, mixing, labeling, manipulation, transport, and sepa-
Microfluidic Applications of Magnetic Particles
ration. Magnetic labeling is more universal and robust than,
for example, fluorescent labeling, in the sense that the
magnetism of the particles cannot be quenched at normal
working temperatures. Ultrasmall magnetic nanoparticles can
be of the same size as the biomolecules attached to them,
thereby providing minimum disturbance to chemical and
biological processes and benefit from decreased diffusion
times. We have discussed the forces acting on such magnetic
particles and the physics of magnetic actuation. Probably the
most prominent advantage of magnetic beads over other solid
supports is that these particles can be magnetically probed
and manipulated using permanent magnets or electromagnets,
independent of normal chemical or biological processes.
Introducing these particles in miniaturized fluidic or lab-
on-a-chip systems provides further advantages: (i) the
reduced volume of a microfluidic chip with respect to a
batch-type reaction vial strongly limits the consumption of
samples and reagents and results in faster reaction times, as
the latter are less compromised by slow diffusion processes
over long distances, (ii) microfluidic flow patterns can be
engineered by geometrical microchannel design of the chip
to generate very controlled space- and time-dependent
hydrodynamic forces acting on the particles, and (iii)
miniaturized systems that completely integrate various
processes, from raw sample pretreatment until specific
biomolecular detection, are achievable, which qualifies these
systems for point-of-care or in-field testing.
We have identified four important biological application
areas for magnetic particle handling in microfluidic systems:
cell handling and separation, nucleic acid processing and
detection, immunoassays, and catalysis. We showed that
specific magnetic labeling permits one to select or deplete
certain cell types from a complex matrix, thanks to the
combination of magnetic labeling with magnetic separation
principles. When introducing a raw cell sample on a
microfluidic chip or cartridge, it is possible to capture the
DNA, purify and process it, and detect a specific DNA
sequence on-chip. One can also detect, with high sensitivity,
various types of proteins and disease markers, by developing
immunoassays that combine magnetic particles, microfluid-
ics, antigen-antibody interactions, and using a number of
different detection principles. Finally, keeping a magnetically
suspended plug of magnetic particles in a microfluidic flow
is also of interest for catalytic reactions, as demonstrated
for protein digestion.
We foresee a bright future for microfluidic systems that
incorporate magnetic micro- and nanoparticle manipulation.
For cell handling in a clinical environment and cell therapy,
magnetic cell sorting techniques with low complexity have
been developed. Today these are still cross-checked using
more complex fluorescent cell sorting techniques, but, in the
future, will provide stand-alone systems for cell purification.
For example, purification of stem or progenitor cells and
depletion of white blood cells from the sample will be key
in cancer treatment and medical transplantation, if a sufficient
number of cells can be processed (sorted) in a reasonable
time. Microfluidic chips for nucleic acid detection will be
playing an important role for point-of-care diagnosis and in-
field testing. While, in a clinical environment, high-
throughput robotic systems can detect with high sensitivity
and at low cost a large number of samples, due to the use of
a minimum of consumables, diagnosis in remote areas,
quality testing of food, or monitoring of pollution parameters
would exclusively depend on the availability of all-integrated
Chemical Reviews, 2010, Vol. 110, No. 3 1557
microfluidic systems. Immunoassays based on the use of
magnetic particles as reaction substrate or as label have been
very successful and will further increase in importance.
Biomolecules labeled with magnetic nanoparticles will be
magnetically driven to a reaction surface, leading to
antigen-antibody binding reactions that are much faster than
obtainable via normal molecular diffusion. This activated
magnetic transport is especially advantageous for detecting
very low antigen concentrations (∼1 fg/mL or smaller). The
combination with precisely engineered microfluidic flow
patterns on-chip will allow developing finely tuned washing
protocols to remove unspecifically adsorbed particles on the
substrate. Optimized chemical reaction/washing protocols
have been developed for magnetoresistive sensors, where the
magnetic particles need to be immobilized specifically over
the most sensitive part of the sensor area. Yet even more
attractive is the use of these protocols for optical surface
coverage detection of magnetic particles, as there is just a
substrate with capture antibodies needed in the microfluidic
chip. The method requires only an optical microscope for
observation and image processing. As demonstrated, sensi-
tivities obtained with this technique are impressive, and due
to its simplicity, a great future can be expected for this type
of microfluidic force flow discrimination assays. Also, stand-
alone integrated systems, which are comprised of sample
pretreatment, processing, and on-chip detection, will have a
bright future. The use of catalysts deposited on magnetic
nanoparticles for catalyst separation will also be increasingly
developed in microfluidic systems. This approach has found
application for protein digestion as a preparation step for
other analytical techniques, but more work is to be expected
from the positioning of a suspension of catalyst-bearing
magnetic particles in a microfluidic flow, especially for
pharmaceutical or chemical applications. These would strongly
benefit from efficient and reusable catalysts during the
screening of new synthesis routes for the development of
drugs or chemical reagents, with interesting potential for up-
scaling by parallelization of the microfluidic circuits. It is
clear that the convergence of nanotechnologies and bio-
sciences will be one of the leading and most promising areas
of research and development in the 21st century. Magnetic
beads certainly will play an important role in these
developments.
10. Acknowledgments
We gratefully acknowledge Rana Afshar, Jean Baudry,
Jérôme Bibette, Edoardo Charbon, Emile Dupont, Victor
Fernandez, Smail Hadjidj, Heinrich Hofmann, Thomas
Lehnert, Hicham Majd, Yves Moser, Virendra Parashar,
Qasem Ramadan, Amar Rida, Venkataragavalu Sivagnanam,
Cumhur Tekin, and Caroline Vandevyver for helpful dis-
cussions and collaboration. This work is part of the research
programs supported by the Swiss National Science Founda-
tion and the Commission for Technology and Innovation.
We also acknowledge the European Commission funded
project DETECTHIV (# 037118, Sensitive nanoparticle assay
for the detection of HIV).
11. References
(1) Manz, A.; Graber, N.; Widmer, H. M. Sens. Actuators, B: Chem.
1990, 1, 244.
(2) Auroux, P. A.; Iossifidis, D.; Reyes, D. R.; Manz, A. Anal. Chem.
2002, 74, 2637.
(3) Vilkner, T.; Janasek, D.; Manz, A. Anal. Chem. 2004, 76, 3373.
1558 Chemical Reviews, 2010, Vol. 110, No. 3
(4) West, J.; Becker, M.; Tombrink, S.; Manz, A. Anal. Chem. 2008,
80, 4403.
(5) Encyclopedia of Microfluidics and Nanofluidics; Li, D., Ed.; Springer:
New York, 2008; Vols. 1-3.
(6) Song, H.; Chen, D. L.; Ismagilov, R. F. Angew. Chem., Int. Ed. 2006,
45, 7336.
(7) Squires, T. M.; Quake, S. R. ReV. Mod. Phys. 2005, 77, 977.
(8) Whitesides, G. M. Nature 2006, 442, 368.
(9) Madou, M. J. Fundamentals of Microfabrication: The Science of
Miniaturization, 2nd ed.; CRC Press: Boca Raton, FL, 2002.
(10) Reyes, D. R.; Iossifidis, D.; Auroux, P. A.; Manz, A. Anal. Chem.
2002, 74, 2623.
(11) Harrison, D. J.; Manz, A.; Fan, Z. H.; Ludi, H.; Widmer, H. M. Anal.
Chem. 1992, 64, 1926.
(12) Duffy, D. C.; McDonald, J. C.; Schueller, O. J. A.; Whitesides, G. M.
Anal. Chem. 1998, 70, 4974.
(13) McDonald, J. C.; Duffy, D. C.; Anderson, J. R.; Chiu, D. T.; Wu,
H. K.; Schueller, O. J. A.; Whitesides, G. M. Electrophoresis 2000,
21, 27.
(14) Becker, H.; Gartner, C. Electrophoresis 2000, 21, 12.
(15) Kawaguchi, H. Prog. Polym. Sci. 2000, 25, 1171.
(16) Kruis, F. E.; Fissan, H.; Peled, A. J. Aerosol Sci. 1998, 29, 511.
(17) Verpoorte, E. Lab Chip 2003, 3, 60N.
(18) Becker, R. Electromagnetic Fields and Interactions; Dover: New
York, 1982.
(19) Chikazumi, S. Physics of Magnetism; Robert E. Krieger Publishing
Co.: Malabar, FL, 1964.
(20) Häfeli, U.; Schütt, W.; Teller, J.; Zborowski, M. Scientific and Clinical
Applications of Magnetic Carriers; Plenum Press: New York, 1997.
(21) Moser, Y.; Lehnert, T.; Gijs, M. A. M. Appl. Phys. Lett. 2009; 022505.
(22) Shevkoplyas, S. S.; Siegel, A. C.; Westervelt, R. M.; Prentiss, M. G.;
Whitesides, G. M. Lab Chip 2007, 7, 1294.
(23) Sometimes the equation Fm ) µ0-1∇(m · B) is taken as the basic
equation for the magnetic force, instead of eq 1. However, this is
correct only when m has no spatial dependence. Also, we use here
the definition of force, where the moment has units of [tesla m3]. An
alternative dimension in literature for the moment is [A m2], but, in
this case, the factor µ0 in the denominator of eq 1 needs to be replaced
by 1.
(24) Jackson, J. Classical Electrodynamics; Wiley: New York, 1998.
(25) Pankhurst, Q. A.; Connolly, J.; Jones, S. K.; Dobson, J. J. Phys. D:
Appl. Phys. 2003, 36, R167.
(26) Gijs, M. A. M. Microfluid. Nanofluid. 2004, 1, 22.
(27) Pamme, N. Lab Chip 2006, 6, 24.
(28) Lu, A. H.; Salabas, E. L.; Schüth, F. Angew. Chem., Int. Ed. 2007,
46, 1222.
(29) Horák, D.; Babič, M.; Macková, H.; Beneš, M. J. J. Sep. Sci. 2007,
30, 1751.
(30) Landfester, K.; Ramírez, L. P. J. Phys.: Condens. Matter 2003, 15,
S1345.
(31) Bergemann, C.; Müller-Schulte, D.; Oster, J.; à Brassard, L.; Lübbe,
A. S. J. Magn. Magn. Mater. 1999, 194, 45.
(32) Grüttner, C.; Rudershausen, S.; Teller, J. J. Magn. Magn. Mater.
2001, 225, 1.
(33) Grüttner, C.; Teller, J. J. Magn. Magn. Mater. 1999, 194, 8.
(34) Latham, A. H.; Williams, M. E. Acc. Chem. Res. 2008, 41, 411.
(35) Grieve, K.; Mulvaney, P.; Grieser, F. Curr. Opin. Colloid Interface
Sci. 2000, 5, 168.
(36) Trindade, T.; O’Brien, P.; Pickett, N. L. Chem. Mater. 2001, 13,
3843.
(37) Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Annu. ReV. Mater.
Sci. 2000, 30, 545.
(38) Swihart, M. T. Curr. Opin. Colloid Interface Sci. 2003, 8, 127.
(39) Grasset, F.; Labhsetwar, N.; Li, D.; Park, D. C.; Saito, N.; Haneda,
H.; Cador, O.; Roisnel, T.; Mornet, S.; Duguet, E.; Portier, J.;
Etourneau, J. Langmuir 2002, 18, 8209.
(40) Sun, S. H.; Zeng, H. J. Am. Chem. Soc. 2002, 124, 8204.
(41) Park, S. J.; Kim, S.; Lee, S.; Khim, Z. G.; Char, K.; Hyeon, T. J. Am.
Chem. Soc. 2000, 122, 8581.
(42) Puntes, V. F.; Krishnan, K. M.; Alivisatos, A. P. Science 2001, 291,
2115.
(43) Shevchenko, E. V.; Talapin, D. V.; Rogach, A. L.; Kornowski, A.;
Haase, M.; Weller, H. J. Am. Chem. Soc. 2002, 124, 11480.
(44) Sun, S. H.; Murray, C. B.; Weller, D.; Folks, L.; Moser, A. Science
2000, 287, 1989.
(45) Lee, W. B.; Weng, C. H.; Cheng, F. Y.; Yeh, C. S.; Lei, H. Y.; Lee,
G. B. Biomed. MicrodeVices 2009, 11, 161.
(46) Chu, L. Y.; Utada, A. S.; Shah, R. K.; Kim, J. W.; Weitz, D. A.
Angew. Chem., Int. Ed. 2007, 46, 8970.
(47) Hwang, D. K.; Dendukuri, D.; Doyle, P. S. Lab Chip 2008, 8, 1640.
(48) Papell, S. S. U.S. Patent 3215572, 1965.
(49) Reimers, G. W.; Khalfalla, S. E. U.S. Patent 3843540, 1974.
(50) Kelley, J. R. U.S. Patent 4019994, 1977.
Gijs et al.
(51) Sun, S. H.; Zeng, H.; Robinson, D. B.; Raoux, S.; Rice, P. M.; Wang,
S. X.; Li, G. X. J. Am. Chem. Soc. 2004, 126, 273.
(52) Jana, N. R.; Chen, Y. F.; Peng, X. G. Chem. Mater. 2004, 16, 3931.
(53) Park, J.; An, K. J.; Hwang, Y. S.; Park, J. G.; Noh, H. J.; Kim, J. Y.;
Park, J. H.; Hwang, N. M.; Hyeon, T. Nat. Mater. 2004, 3, 891.
(54) Dumestre, F.; Chaudret, B.; Amiens, C.; Renaud, P.; Fejes, P. Science
2004, 303, 821.
(55) Daniel, J.-C.; Schuppiser, J.-L.; Tricot, M. U.S. Patent 4358388, 1982.
(56) Bibette, J.; Charmot, D.; Schorsch, G. U.S. Patent 5242964, 1993.
(57) Bibette, J. J. Magn. Magn. Mater. 1993, 122, 37.
(58) Cornell, R. M.; Schertmann, U. The Iron Oxides: Structure, Proper-
ties, Reactions, Occurrence and Uses; VCH: Weinheim, 1996.
(59) Petri-Fink, A.; Chastellain, M.; Juillerat-Jeanneret, L.; Ferrari, A.;
Hofmann, H. Biomaterials 2005, 26, 2685.
(60) Lawaczeck, R.; Bauer, H.; Frenzel, T.; Hasegawa, M.; Ito, Y.; Kito,
K.; Miwa, N.; Tsutsui, H.; Volger, H.; Weinmann, H. J. Acta Radiol.
1997, 38, 584.
(61) Kim, D. K.; Zhang, Y.; Kehr, J.; Klason, T.; Bjelke, B.; Muhammed,
M. J. Magn. Magn. Mater. 2001, 225, 256.
(62) Liu, Q. X.; Xu, Z. H.; Finch, J. A.; Egerton, R. Chem. Mater. 1998,
10, 3936.
(63) Ban, Z. H.; Barnakov, Y. A.; Golub, V. O.; O’Connor, C. J. J. Mater.
Chem. 2005, 15, 4660.
(64) Nikitenko, S. I.; Koltypin, Y.; Palchik, O.; Felner, I.; Xu, X. N.;
Gedanken, A. Angew. Chem., Int. Ed. 2001, 40, 4447.
(65) Stöber, W.; Fink, A.; Bohn, E. J. Colloid Interface Sci. 1968, 26,
62.
(66) Graf, C.; Vossen, D. L. J.; Imhof, A.; van Blaaderen, A. Langmuir
2003, 19, 6693.
(67) Lu, Y.; Yin, Y. D.; Mayers, B. T.; Xia, Y. N. Nano Lett. 2002, 2,
183.
(68) Ulman, A. Chem. ReV. 1996, 96, 1533.
(69) Molday, R. S. U.S. Patent 4452773, 1984.
(70) Rembaum, A. U.S. Patent 4267234, 1981.
(71) Lee, J. W.; Isobe, T.; Senna, M. Colloids Surf., A 1996, 109, 121.
(72) Suzuki, M.; Shinkai, M.; Kamihira, M.; Kobayashi, T. Biotechnol.
Appl. Biochem. 1995, 21, 335.
(73) Dynal product catalogue (http:/www.invitrogen.com).
(74) Ugelstad, J.; Berge, A.; Ellingsen, T.; Schmid, R.; Nilsen, T. N.;
Mork, P. C.; Stenstad, P.; Hornes, E.; Olsvik, O. Prog. Polym. Sci.
1992, 17, 87.
(75) Ugelstad, J. U.S. Patent 4774265, 1988.
(76) Kleiber, J.; Walter, T.; Harttig, H.; Lesniak, C.; Mennig, M.; Riedling,
M.; Schmidt, H. U.S. Patent 6255477B1, 2001.
(77) Undisclosed inventors. European patent 1154443A1, 2001.
(78) Miltenyi, S.; Muller, W.; Weichel, W.; Radbruch, A. Cytometry 1990,
11, 231.
(79) Häfeli, U. O.; Pauer, G. J. J. Magn. Magn. Mater. 1999, 194, 76.
(80) Häfeli, U. O.; Sweeney, S. M.; Beresford, B. A.; Sim, E. H.; Macklis,
R. M. J. Biomed. Mater. Res. 1994, 28, 901.
(81) Olsvik, O.; Popovic, T.; Skjerve, E.; Cudjoe, K. S.; Hornes, E.;
Ugelstad, J.; Uhlen, M. Clin. Microbiol. ReV. 1994, 7, 43.
(82) Jung, Y. W.; Kang, H. J.; Lee, J. M.; Jung, S. O.; Yun, W. S.; Chung,
S. J.; Chung, B. H. Anal. Biochem. 2008, 374, 99.
(83) Lehmann, U.; Vandevyver, C.; Parashar, V. K.; Gijs, M. A. M.
Angew. Chem., Int. Ed. 2006, 45, 3062.
(84) LesliePelecky, D. L.; Rieke, R. D. Chem. Mater. 1996, 8, 1770.
(85) Hancock, J. P.; Kemshead, J. T. J. Immunol. Methods 1993, 164,
51.
(86) Tarn, M. D.; Hirota, N.; Iles, A.; Pamme, N. Sci. Technol. AdV. Mater.
2009, 10, 014611.
(87) Winkleman, A.; Gudiksen, K. L.; Ryan, D.; Whitesides, G. M.;
Greenfield, D.; Prentiss, M. Appl. Phys. Lett. 2004, 85, 2411.
(88) Gassner, A. L.; Abonnenc, M.; Chen, H. X.; Morandini, J.; Josserand,
J.; Rossier, J. S.; Busnel, J. M.; Girault, H. H. Lab Chip 2009, 9,
2356.
(89) White, F. M. Fluid Mechanics; McGraw-Hill: New York, 1999.
(90) Faxen, H. Ann. Phys. 1922, 68, 89.
(91) Wirix-Speetjens, R.; Fyen, W.; Xu, K. D.; De Boeck, J.; Borghs, G.
IEEE Trans. Magn. 2005, 41, 4128.
(92) Gregory, J. J. Colloid Interface Sci. 1981, 83, 138.
(93) Visser, J. AdV. Colloid Interface Sci. 1981, 15, 157.
(94) Kar, G.; Chander, S.; Mika, T. S. J. Colloid Interface Sci. 1973, 44,
347.
(95) Hunter, R. Zeta Potential in Colloid Science; Academic Press: New
York, 1981.
(96) Sivagnanam, V.; Song, B.; Vandevyver, C.; Gijs, M. A. M. Anal.
Chem. 2009, 81, 6509.
(97) Einstein, A. InVestigations on the Theory of Brownian MoVement;
Dover: New York, 1956.
(98) Schaller, V.; Kraling, U.; Rusu, C.; Petersson, K.; Wipenmyr, J.;
Krozer, A.; Wahnstrom, G.; Sanz-Velasco, A.; Enoksson, P.; Jo-
hansson, C. J. Appl. Phys. 2008, 104, 093918.
Microfluidic Applications of Magnetic Particles
(99) van Ommering, K.; Nieuwenhuis, J. H.; van Ijzendoorn, L. J.;
Koopmans, B.; Prins, M. W. J. Appl. Phys. Lett. 2006, 89, 142511.
(100) Janssen, X. J. A.; Schellekens, A. J.; van Ommering, K.; van
Ijzendoorn, L. J.; Prins, M. W. J. Biosens. Bioelectron. 2009, 24,
1937.
(101) Liu, C. X.; De Palma, R.; Reekmans, G.; Laureyn, W.; Stakenborg,
T.; Lagae, L. Biosens. Bioelectron. 2009, 24, 2294.
(102) Huke, B.; Lucke, M. Rep. Prog. Phys. 2004, 67, 1731.
(103) Derks, R. J. S.; Dietzel, A.; Wimberger-Friedl, R.; Prins, M. W. J.
Microfluid. Nanofluid. 2007, 3, 141.
(104) Derks, R. J. S.; Frijns, A. J. H.; Prins, M. W. J.; Dietzel, A. H. Appl.
Phys. Lett. 2008, 92, 024104.
(105) Dunnill, P.; Lilly, M. D. Biotechnol. Bioeng. 1974, 16, 987.
(106) Mosbach, K.; Andersson, L. Nature 1977, 270, 259.
(107) Kondo, A.; Kamura, H.; Higashitani, K. Appl. Microbiol. Biotechnol.
1994, 41, 99.
(108) Koneracka, M.; Kopcansky, P.; Timko, M.; Ramchand, C. N. J.
Magn. Magn. Mater. 2002, 252, 409.
(109) Kourilov, V.; Steinitz, M. Anal. Biochem. 2002, 311, 166.
(110) Manz, R.; Assenmacher, M.; Pfluger, E.; Miltenyi, S.; Radbruch, A.
Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 1921.
(111) Rheinlander, T.; Kotitz, R.; Weitschies, W.; Semmler, W. J. Magn.
Magn. Mater. 2000, 219, 219.
(112) Ebner, A. D.; Ritter, J. A. AIChE J. 2001, 47, 303.
(113) Pamme, N. Lab Chip 2007, 7, 1644.
(114) Latham, A. H.; Freitas, R. S.; Schiffer, P.; Williams, M. E. Anal.
Chem. 2005, 77, 5055.
(115) Giddings, J. C.; Yang, F. J. F.; Myers, M. N. Science 1976, 193,
1244.
(116) Chalmers, J. J.; Zborowski, M.; Sun, L. P.; Moore, L. Biotechnol.
Prog. 1998, 14, 141.
(117) Hatch, G. P.; Stelter, R. E. J. Magn. Magn. Mater. 2001, 225, 262.
(118) Williams, P. S.; Zborowski, M.; Chalmers, J. Anal. Chem. 1999, 71,
3799.
(119) Carpino, F.; Moore, L. R.; Zborowski, M.; Chalmers, J. J.; Williams,
P. S. J. Magn. Magn. Mater. 2005, 293, 546.
(120) Hoyos, M.; Moore, L. R.; McCloskey, K. E.; Margel, S.; Zuberi,
M.; Chalmers, J. J.; Zborowski, M. J. Chromatogr., A 2000, 903,
99.
(121) Fuh, C. B.; Chen, S. Y. J. Chromatogr., A 1998, 813, 313.
(122) Fuh, C. B.; Chen, S. Y. J. Chromatogr., A 1999, 857, 193.
(123) Williams, P. S.; Carpino, F.; Zborowski, M. J. Magn. Magn. Mater.
2009, 1446.
(124) Nakamura, M.; Zborowski, M.; Lasky, L. C.; Margel, S.; Chalmers,
J. J. Exp. Fluids 2001, 30, 371.
(125) Zhang, H. D.; Moore, L. R.; Zborowski, M.; Williams, P. S.; Margel,
S.; Chalmers, J. J. Analyst 2005, 130, 514.
(126) Moore, L. R.; Zborowski, M.; Nakamura, M.; McCloskey, K.; Gura,
S.; Zuberi, M.; Margel, S.; Chalmers, J. J. J. Biochem. Biophys.
Methods 2000, 44, 115.
(127) Blankenstein, G.; Larsen, U. D. Biosens. Bioelectron. 1998, 13, 427.
(128) Pamme, N.; Manz, A. Anal. Chem. 2004, 76, 7250.
(129) Nandy, K.; Chaudhuri, S.; Ganguly, R.; Puri, I. K. J. Magn. Magn.
Mater. 2008, 320, 1398.
(130) Nedelcu, S.; Watson, J. H. P. Miner. Eng. 2002, 15, 355.
(131) Hoffmann, C.; Franzreb, M. IEEE Trans. Magn. 2004, 40, 462.
(132) Todd, P.; Cooper, R. P.; Doyle, J. F.; Dunn, S.; Vellinger, J.; Deuser,
M. S. J. Magn. Magn. Mater. 2001, 225, 294.
(133) Haik, Y.; Pai, V.; Chen, C. J. J. Magn. Magn. Mater. 1999, 194,
254.
(134) Miltenyi Biotec product catalogue (http://www.miltenyibiotec.com).
(135) Afshar, R.; Moser, Y.; Lehnert, T.; Gijs, M. A. M. Sens. Actuators,
B: Chem. 2009, DOI: 10.1016/j.snb.2009.08.044.
(136) Berger, M.; Castelino, J.; Huang, R.; Shah, M.; Austin, R. H.
Electrophoresis 2001, 22, 3883.
(137) Lyuksyutov, I. F.; Naugle, D. G.; Rathnayaka, K. D. D. Appl. Phys.
Lett. 2004, 85, 1817.
(138) Deng, T.; Prentiss, M.; Whitesides, G. M. Appl. Phys. Lett. 2002,
80, 461.
(139) Guo, S. S.; Deng, Y. L.; Zhao, L. B.; Chan, H. L. W.; Zhao, X. Z.
J. Phys. D: Appl. Phys. 2008, 41, 105008.
(140) Forbes, Z. G.; Yellen, B. B.; Barbee, K. A.; Friedman, G. IEEE Trans.
Magn. 2003, 39, 3372.
(141) Yellen, B.; Friedman, G.; Feinerman, A. J. Appl. Phys. 2003, 93,
7331.
(142) Lin, Y. A.; Wong, T. S.; Bhardwaj, U.; Chen, J. M.; McCabe, E.;
Ho, C. M. J. Micromech. Microeng. 2007, 17, 1299.
(143) Smistrup, K.; Bu, M. Q.; Wolff, A.; Bruus, H.; Hansen, M. F.
Microfluid. Nanofluid. 2008, 4, 565.
(144) Mikkelsen, C.; Hansen, M. F.; Bruns, H. J. Magn. Magn. Mater.
2005, 293, 578.
(145) Furlani, E. P. J. Appl. Phys. 2006, 99, 024912.
(146) Furlani, E. P.; Ng, K. C. Phys. ReV. E 2006, 73, 061919.
Chemical Reviews, 2010, Vol. 110, No. 3 1559
(147) Smistrup, K.; Lund-Olesen, T.; Hansen, M. F.; Tang, P. T. J. Appl.
Phys. 2006, 99, 08P102.
(148) Smistrup, K.; Kjeldsen, B. G.; Reimers, J. L.; Dufva, M.; Petersen,
J.; Hansen, M. F. Lab Chip 2005, 5, 1315.
(149) Lund-Olesen, T.; Bruus, H.; Hansen, M. F. Biomed. MicrodeVices
2007, 9, 195.
(150) Chen, H. T.; Kaminski, M. D.; Rosengart, A. J. Med. Eng. Phys.
2008, 30, 1.
(151) Inglis, D. W.; Riehn, R.; Austin, R. H.; Sturm, J. C. Appl. Phys.
Lett. 2004, 85, 5093.
(152) Han, K. H.; Frazier, A. B. J. Appl. Phys. 2004, 96, 5797.
(153) Han, K. H.; Frazier, A. B. J. Microelectromech. Syst. 2005, 14, 1422.
(154) Han, K. H.; Frazier, A. B. Lab Chip 2006, 6, 265.
(155) Choi, J. W.; Ahn, C. H.; Bhansali, S.; Henderson, H. T. Sens.
Actuators, B: Chem. 2000, 68, 34.
(156) Choi, J. W.; Liakopoulos, T. M.; Ahn, C. H. Biosens. Bioelectron.
2001, 16, 409.
(157) Ahn, C. H.; Allen, M. G.; Trimmer, W.; Jun, Y. N.; Erramilli, S. J.
Microelectromech. Syst. 1996, 5, 151.
(158) Do, J.; Choi, J. W.; Ahn, C. H. IEEE Trans. Magn. 2004, 40, 3009.
(159) Ramadan, Q.; Samper, V.; Poenar, D.; Yu, C. J. Magn. Magn. Mater.
2004, 281, 150.
(160) Smistrup, K.; Hansen, O.; Bruus, H.; Hansen, M. F. J. Magn. Magn.
Mater. 2005, 293, 597.
(161) Smistrup, K.; Tang, P. T.; Hansen, O.; Hansen, M. F. J. Magn. Magn.
Mater. 2006, 300, 418.
(162) Ramadan, Q.; Samper, V.; Poenar, D. P.; Yu, C. Biosens. Bioelectron.
2006, 21, 1693.
(163) Ramadan, Q.; Yu, C.; Samper, V.; Poenar, D. P. Appl. Phys. Lett.
2006, 88, 032501.
(164) Furlani, E. P.; Sahoo, Y.; Ng, K. C.; Wortman, J. C.; Monk, T. E.
Biomed. MicrodeVices 2007, 9, 451.
(165) Mikkelsen, C.; Bruus, H. Lab Chip 2005, 5, 1293.
(166) Barbic, M.; Mock, J. J.; Gray, A. P.; Schultz, S. Appl. Phys. Lett.
2001, 79, 1897.
(167) Song, S. H.; Kwak, B. S.; Park, J. S.; Kim, W.; Jung, H. I. Sens.
Actuators, A: Phys. 2009, 151, 64.
(168) Tondra, M.; Granger, M.; Fuerst, R.; Porter, M.; Nordman, C.; Taylor,
J.; Akou, S. IEEE Trans. Magn. 2001, 37, 2621.
(169) Pekas, N.; Granger, M.; Tondra, M.; Popple, A.; Porter, M. D. J.
Magn. Magn. Mater. 2005, 293, 584.
(170) Deng, T.; Whitesides, G. M.; Radhakrishnan, M.; Zabow, G.; Prentiss,
M. Appl. Phys. Lett. 2001, 78, 1775.
(171) Ostergaard, S.; Blankenstein, G.; Dirac, H.; Leistiko, O. J. Magn.
Magn. Mater. 1999, 194, 156.
(172) Joung, J.; Shen, J.; Grodzinski, P. IEEE Trans. Magn. 2000, 36, 2012.
(173) Yapici, M. K.; Zou, J. Microsyst. Technol.: Micro- and Nanosystems-
Information Storage Process. Syst. 2008, 14, 881.
(174) Lee, C. S.; Lee, H.; Westervelt, R. M. Appl. Phys. Lett. 2001, 79,
3308.
(175) Wirix-Speetjens, R.; Fyen, W.; De Boeck, J.; Borghs, G. J. Appl.
Phys. 2006, 99, 103903.
(176) Rida, A.; Fernandez, V.; Gijs, M. A. M. Appl. Phys. Lett. 2003, 83,
2396.
(177) Ramadan, Q.; Poenar, D.; Yu, C. Microfluid. Nanofluid. 2009, 6,
53.
(178) Yellen, B. B.; Erb, R. M.; Son, H. S.; Hewlin, R.; Shang, H.; Lee,
G. U. Lab Chip 2007, 7, 1681.
(179) Conroy, R. S.; Zabow, G.; Moreland, J.; Koretsky, A. P. Appl. Phys.
Lett. 2008, 93, 203901.
(180) Megens, M.; Prins, M. J. Magn. Magn. Mater. 2005, 293, 702.
(181) Graham, D. L.; Ferreira, H. A.; Freitas, P. P. Trends Biotechnol. 2004,
22, 455.
(182) Tamanaha, C. R.; Mulvaney, S. P.; Rife, J. C.; Whitman, L. J. Biosens.
Bioelectron. 2008, 24, 1.
(183) Baibich, M. N.; Broto, J. M.; Fert, A.; Vandau, F. N.; Petroff, F.;
Eitenne, P.; Creuzet, G.; Friederich, A.; Chazelas, J. Phys. ReV. Lett.
1988, 61, 2472.
(184) Dieny, B.; Speriosu, V. S.; Metin, S.; Parkin, S. S. P.; Gurney, B. A.;
Baumgart, P.; Wilhoit, D. R. J. Appl. Phys. 1991, 69, 4774.
(185) Freitas, P. P.; Silva, F.; Oliveira, N. J.; Melo, L. V.; Costa, L.;
Almeida, N. Sens. Actuators, A: Phys. 2000, 81, 2.
(186) Magnetic Multilayers and Giant Magnetoresistance; Hartmann, U.,
Ed.; Springer Verlag: Berlin-Heidelberg, 1999.
(187) Lagae, L.; Wirix-Speetjens, R.; Das, J.; Graham, D.; Ferreira, H.;
Freitas, P. P. F.; Borghs, G.; De Boeck, J. J. Appl. Phys. 2002, 91,
7445.
(188) Besse, P. A.; Boero, G.; Demierre, M.; Pott, V.; Popovic, R. Appl.
Phys. Lett. 2002, 80, 4199.
(189) Lee, W.; Joo, S.; Kim, S. U.; Rhie, K.; Hong, J.; Shin, K. H.; Kim,
K. H. Appl. Phys. Lett. 2009, 94, 153903.
(190) Ejsing, L.; Hansen, M. F.; Menon, A. K.; Ferreira, H. A.; Graham,
D. L.; Freitas, P. P. Appl. Phys. Lett. 2004, 84, 4729.
1560 Chemical Reviews, 2010, Vol. 110, No. 3
(191) Baselt, D. R.; Lee, G. U.; Natesan, M.; Metzger, S. W.; Sheehan,
P. E.; Colton, R. J. Biosens. Bioelectron. 1998, 13, 731.
(192) Miller, M. M.; Sheehan, P. E.; Edelstein, R. L.; Tamanaha, C. R.;
Zhong, L.; Bounnak, S.; Whitman, L. J.; Colton, R. J. J. Magn. Magn.
Mater. 2001, 225, 138.
(193) Rife, J. C.; Miller, M. M.; Sheehan, P. E.; Tamanaha, C. R.; Tondra,
M.; Whitman, L. J. Sens. Actuators, A: Phys. 2003, 107, 209.
(194) Edelstein, R. L.; Tamanaha, C. R.; Sheehan, P. E.; Miller, M. M.;
Baselt, D. R.; Whitman, L. J.; Colton, R. J. Biosens. Bioelectron.
2000, 14, 805.
(195) Coehoorn, R.; Prins, M. W. J. European patent 1456658, 2004.
(196) Graham, D. L.; Ferreira, H. A.; Freitas, P. P.; Cabral, J. M. S. Biosens.
Bioelectron. 2003, 18, 483.
(197) Tondra, M.; Porter, M.; Lipert, R. J. J. Vac. Sci. Technol., A 2000,
18, 1125.
(198) Li, G. X.; Joshi, V.; White, R. L.; Wang, S. X.; Kemp, J. T.; Webb,
C.; Davis, R. W.; Sun, S. H. J. Appl. Phys. 2003, 93, 7557.
(199) Li, G. X.; Wang, S. X. IEEE Trans. Magn. 2003, 39, 3313.
(200) de Boer, B. M.; Kahlman, J.; Jansen, T.; Duric, H.; Veen, J. Biosens.
Bioelectron. 2006, 2366.
(201) Thilwind, R. E.; Megens, M.; van Zon, J.; Coehoorn, R.; Prins,
M. W. J. J. Magn. Magn. Mater. 2008, 320, 486.
(202) Janssen, X. J. A.; van Ijzendoorn, L. J.; Prins, M. W. Biosens.
Bioelectron. 2008, 23, 833.
(203) Wimberger-Friedl, R.; Nellissen, T.; Weekamp, W.; van Delft, J.;
Ansems, W.; Prins, M.; Megens, M.; Dittmer, W.; de Witz, C.; van
Iersel, B. J. Micromech. Microeng. 2009, 19, 015015.
(204) Hayes, M. A.; Polson, N. A.; Garcia, A. A. Langmuir 2001, 17, 2866.
(205) Rida, A.; Gijs, M. A. M. Appl. Phys. Lett. 2004, 85, 4986.
(206) Lacharme, F.; Vandevyver, C.; Gijs, M. A. M. Anal. Chem. 2008,
80, 2905.
(207) Hayes, M. A.; Polson, N. A.; Phayre, A. N.; Garcia, A. A. Anal.
Chem. 2001, 73, 5896.
(208) Liu, J.; Lawrence, M.; Wu, A.; Ivey, M. L.; Flores, G. A.; Javier,
K.; Bibette, J.; Richard, J. Phys. ReV. Lett. 1995, 74, 2828.
(209) Flores, G. A.; Liu, J.; Mohebi, M.; Jamasbi, N. Int. J. Mod. Phys. B
1999, 13, 2093.
(210) Promislow, J. H. E.; Gast, A. P. Langmuir 1996, 12, 4095.
(211) Promislow, J. H. E.; Gast, A. P. Phys. ReV. E 1997, 56, 642.
(212) Wirtz, D.; Fermigier, M. Langmuir 1995, 11, 398.
(213) Wirtz, D.; Fermigier, M. Phys. ReV. Lett. 1994, 72, 2294.
(214) Melle, S.; Fuller, G. G.; Rubio, M. A. Phys. ReV. E 2000, 61, 4111.
(215) Melle, S.; Calderon, O. G.; Rubio, M. A.; Fuller, G. G. J.
Non-Newton. Fluid Mech. 2002, 102, 135.
(216) Wang, Y.; Zhe, J.; Chung, B. T. F.; Dutta, P. Microfluid. Nanofluid.
2008, 4, 375.
(217) Biswal, S. L.; Gast, A. P. Anal. Chem. 2004, 76, 6448.
(218) Kang, T. G.; Hulsen, M. A.; Anderson, P. D.; den Toonder, J. M. J.;
Meijer, H. E. H. Phys. ReV. E 2007, 76, 066303.
(219) Petousis, I.; Homburg, E.; Derks, R.; Dietzel, A. Lab Chip 2007, 7,
1746.
(220) Lacharme, F.; Vandevyver, C.; Gijs, M. A. M. 21st IEEE International
Conference on Micro Electro Mechanical Systems MEMS 2008,
Tucson, AZ, 2008; p 184.
(221) Pregibon, D. C.; Toner, M.; Doyle, P. S. Langmuir 2006, 22, 5122.
(222) Sivagnanam, V.; Sayah, A.; Gijs, M. A. M. Microelectron. Eng. 2008,
1355.
(223) Sivagnanam, V.; Sayah, A.; Vandevyver, C.; Gijs, M. A. M. Sens.
Actuators, B: Chem. 2008, 132, 361.
(224) Rida, A.; T., L.; Gijs, M. A. M. 7th International Conference on
Miniaturized Chemical and Biochemical Analysis Systems, October
5-9, 2003, Squaw Valley, CA, 2003; p 579.
(225) Lund-Olesen, T.; Buus, B. B.; Howalt, J. G.; Hansen, M. F. J. Appl.
Phys. 2007, 103, 07E902.
(226) Lund-Olesen, T.; Buus, B. B.; Howalt, J. G.; Hansen, M. F. J. Appl.
Phys. 2008, 103, 07E902.
(227) Lee, S. H.; van Noort, D.; Lee, J. Y.; Zhang, B. T.; Park, T. H. Lab
Chip 2009, 9, 479.
(228) Suzuki, H.; Ho, C. M.; Kasagi, N. J. Microelectromech. Syst. 2004,
13, 779.
(229) Zolgharni, M.; Azimi, S. M.; Bahmanyar, M. R.; Balachandran, W.
Microfluid. Nanofluid. 2007, 3, 677.
(230) Calhoun, R.; Yadav, A.; Phelan, P.; Vuppu, A.; Garcia, A.; Hayes,
M. Lab Chip 2006, 6, 247.
(231) Yellen, B.; Friedman, G.; Feinerman, A. J. Appl. Phys. 2002, 91,
8552.
(232) Yellen, B. B.; Friedman, G. Langmuir 2004, 20, 2553.
(233) Mukhopadhyay, R. Anal. Chem. 2006, 78, 1401.
(234) Jensen, K.; Lee, A. Lab Chip 2004, 4, 31N.
(235) Garcia, A. A.; Egatz-Gomez, A.; Lindsay, S. A.; Dominguez-Garcia,
P.; Melle, S.; Marquez, M.; Rubio, M. A.; Picraux, S. T.; Yang, D. Q.;
Aella, P.; Hayes, M. A.; Gust, D.; Loyprasert, S.; Vazquez-Alvarez,
T.; Wang, J. J. Magn. Magn. Mater. 2007, 311, 238.
Gijs et al.
(236) Mukhopadhyay, R. Anal. Chem. 2005, 77, 55A.
(237) Dorvee, J. R.; Derfus, A. M.; Bhatia, S. N.; Sailor, M. J. Nat. Mater.
2004, 3, 896.
(238) Shikida, M.; Nagao, N.; Imai, R.; Honda, H.; Okochi, M.; Ito, H.;
Sato, K. J. Micromech. Microeng. 2008, 18, 035034.
(239) Buriak, J. M. Nat. Mater. 2004, 3, 847.
(240) Pollack, M. G.; Fair, R. B.; Shenderov, A. D. Appl. Phys. Lett. 2000,
77, 1725.
(241) Lee, J.; Moon, H.; Fowler, J.; Schoellhammer, T.; Kim, C. J. 14th
IEEE International Conference on Micro Electro Mechanical Systems
(MEMS 2001), Interlaken, Switzerland, 2001; p 259.
(242) Egatz-Gomez, A.; Melle, S.; Garcia, A. A.; Lindsay, S. A.; Marquez,
M.; Dominguez-Garcia, P.; Rubio, M. A.; Picraux, S. T.; Taraci, J. L.;
Clement, T.; Yang, D.; Hayes, M. A.; Gust, D. Appl. Phys. Lett.
2006, 89, 129902.
(243) Shikida, M.; Koyama, M.; Nagao, N.; Imai, R.; Honda, H.; Okochi,
M.; Tsuchiya, H.; Sato, K. Sens. Actuators, B: Chem. 2009, 137,
774.
(244) Long, Z.; Shetty, A. M.; Solomon, M. J.; Larson, R. G. Lab Chip
2009, DOI: 10.1039/b819818g.
(245) Wang, Y.; Zhao, Y.; Cho, S. K. J. Micromech. Microeng. 2007, 17,
2148.
(246) Shah, G. J.; Kim, C. J. J. Microelectromech. Syst. 2009, 18, 363.
(247) Ohashi, T.; Kuyama, H.; Suzuki, K.; Nakamura, S. Anal. Chim. Acta
2008, 612, 218.
(248) Roy, T.; Sinha, A.; Chakraborty, S.; Ganguly, R.; Puri, I. K. Phys.
Fluids 2009, 21, 027101.
(249) Lehmann, U.; Hadjidj, S.; Parashar, V. K.; Vandevyver, C.; Rida,
A.; Gijs, M. A. M. Sens. Actuators, B: Chem. 2006, 117, 457.
(250) Beyzavi, A.; Nguyen, N. T. J. Phys. D: Appl. Phys. 2009, 42, 015004.
(251) Nguyen, N. T.; Beyzavi, A.; Ng, K. M.; Huang, X. Y. Microfluid.
Nanofluid. 2007, 3, 571.
(252) Shikida, M.; Inouchi, K.; Honda, H.; Sato, K. 17th IEEE International
Conference on Micro Electro Mechanical Systems (MEMS 2004),
Maastricht, The Netherlands, 2004; p 359.
(253) Lehmann, U.; Parashar, V.; Hadjidj, S.; Rida, A.; Gijs, M. A. M.
Digest of Technical Papers of Transducers 2005, Seoul, Korea, 2005;
p 77.
(254) Zborowski, M.; Ostera, G. R.; Moore, L. R.; Milliron, S.; Chalmers,
J. J.; Schechter, A. N. Biophys. J. 2003, 84, 2638.
(255) Spees, W. M.; Yablonskiy, D. A.; Oswood, M. C.; Ackerman, J. J. H.
Magn. Reson. Med. 2001, 45, 533.
(256) The susceptibility values are expressed here in SI units. In the cgs
system, they are a factor of 4π smaller.
(257) Melville, D.; Paul, F.; Roath, S. Nature 1975, 255, 706.
(258) Blakemore, R. P. Annu. ReV. Microbiol. 1982, 36, 217.
(259) Zola, H.; Swart, B.; Banham, A.; Barry, S.; Beare, A.; Bensussan,
A.; Boumsell, L.; Buckley, C. D.; Buhring, H. J.; Clark, G.; Engel,
P.; Fox, D.; Jin, B. Q.; Macardle, P. J.; Malavasi, F.; Mason, D.;
Stockinger, H.; Yang, X. F. J. Immunol. Methods 2007, 319, 1.
(260) Greer, J. P.; Foerster, J.; Lukens, J. N.; Rodgers, G. M.; Paraskevas,
F.; Glader, B. E. Wintrobe’s Clinical Hematology, 11th ed.; Lippincott
Williams & Wilkins: Philadelphia, PA, 2003.
(261) Grodzinski, P.; Yang, J.; Liu, R. H.; Ward, M. D. Biomed.
MicrodeVices 2003, 5, 303.
(262) Jing, Y.; Mal, N.; Williams, P. S.; Mayorga, M.; Penn, M. S.;
Chalmers, J. J.; Zborowski, M. FASEB J. 2008, 22, 4239.
(263) McCloskey, K. E.; Chalmers, J. J.; Zborowski, M. Anal. Chem. 2003,
75, 6868.
(264) McCloskey, K. E.; Chalmers, J. J.; Zborowski, M. Cytometry 2000,
40, 307.
(265) McCloskey, K. E.; Comella, K.; Chalmers, J. J.; Margel, S.;
Zborowski, M. Biotechnol. Bioeng. 2001, 75, 642.
(266) McCloskey, K. E.; Zborowski, M.; Chalmers, J. J. Cytometry 2001,
44, 137.
(267) McCloskey, K. E.; Moore, L. R.; Hoyos, M.; Rodriguez, A.;
Chalmers, J. J.; Zborowski, M. Biotechnol. Prog. 2003, 19, 899.
(268) Pappas, D.; Wang, K. Anal. Chim. Acta 2007, 601, 26.
(269) Yi, C. Q.; Li, C. W.; Ji, S. L.; Yang, M. S. Anal. Chim. Acta 2006,
560, 1.
(270) Toner, M.; Irimia, D. Ann. ReV. Biomed. Eng. 2005, 7, 77.
(271) VanDelinder, V.; Groisman, A. Anal. Chem. 2007, 79, 2023.
(272) Springston, S. R.; Myers, M. N.; Giddings, J. C. Anal. Chem. 1987,
59, 344.
(273) Das, C. M.; Becker, F.; Vernon, S.; Noshari, J.; Joyce, C.; Gascoyne,
P. R. C. Anal. Chem. 2005, 77, 2708.
(274) Bonner, W. A.; Sweet, R. G.; Hulett, H. R.; Herzenbe, L. ReV. Sci.
Instrum. 1972, 43, 404.
(275) Shapiro, H. H. Practical Flow Cytometry; Wiley-Liss: New York,
2003.
(276) Fu, A. Y.; Chou, H. P.; Spence, C.; Arnold, F. H.; Quake, S. R.
Anal. Chem. 2002, 74, 2451.
(277) Graham, M. D. J. Appl. Phys. 1981, 52, 2578.
Microfluidic Applications of Magnetic Particles
(278) Melville, D.; Paul, F.; Roath, S. IEEE Trans. Magn. 1982, 18, 1680.
(279) Takayasu, M.; Kelland, D. R.; Minervini, J. V. 16th International
Conference on Magnet Technology, Ponte Vedra Beach, FL, 1999;
p 927.
(280) Paul, F.; Melville, D.; Roath, S.; Warhurst, D. C. IEEE Trans. Magn.
1981, 17, 2822.
(281) Paul, F.; Roath, S.; Melville, D.; Warhurst, D. C.; Osisanya, J. O. S.
Lancet 1981, 2, 70.
(282) Fuh, C. B.; Su, Y. S.; Tsai, H. Y. J. Chromatogr. 2004, 1027, 289.
(283) Jung, J.; Han, K. H. Appl. Phys. Lett. 2008, 93, 223902.
(284) Qu, B. Y.; Wu, Z. Y.; Fang, F.; Bai, Z. M.; Yang, D. Z.; Xu, S. K.
Anal. Bioanal. Chem. 2008, 392, 1317.
(285) Furlani, E. P. J. Phys. D: Appl. Phys. 2007, 40, 1313.
(286) Huang, R.; Barber, T. A.; Schmidt, M. A.; Tompkins, R. G.; Toner,
M.; Bianchi, D. W.; Kapur, R.; Flejter, W. L. Prenatal Diagn. 2008,
28, 892.
(287) Yu, L. M.; Easterbrook, P. J.; Marshall, T. Int. J. Epidemiol. 1997,
26, 1367.
(288) Inglis, D. W.; Riehn, R.; Sturm, J. C.; Austin, R. H. J. Appl. Phys.
2006, 99, 08K101.
(289) Furdui, V. I.; Harrison, D. J. Lab Chip 2004, 4, 614.
(290) Zborowski, M.; Sun, L. P.; Moore, L. R.; Williams, P. S.; Chalmers,
J. J. J. Magn. Magn. Mater. 1999, 194, 224.
(291) Zborowski, M.; Williams, P. S.; Sun, L.; Moore, L. R.; Chalmers,
J. J. J. Liq. Chromatogr. Relat. Technol. 1997, 20, 2887.
(292) Leigh, D. R.; Steinert, S.; Moore, L. R.; Chalmers, J.; Zborowski,
M. Cytometry, Part A 2005, 66A, 103.
(293) Chalmers, J. J.; Zhao, Y.; Nakamura, M.; Melnik, K.; Lasky, L.;
Moore, L.; Zborowski, M. J. Magn. Magn. Mater. 1999, 194, 231.
(294) Chalmers, J. J.; Haam, S.; Zhao, Y.; McCloskey, K.; Moore, L.;
Zborowski, M.; Williams, P. S. Biotechnol. Bioeng. 1999, 64, 519.
(295) Lara, O.; Tong, X. D.; Zborowski, M.; Farag, S. S.; Chalmers, J. J.
Biotechnol. Bioeng. 2006, 94, 66.
(296) Tong, X. D.; Xiong, Y.; Zborowski, M.; Farag, S. S.; Chalmers, J. J.
Exp. Hematol. 2007, 35, 1613.
(297) Moore, L. R.; Zborowski, M.; Sun, L. P.; Chalmers, J. J. J. Biochem.
Biophys. Methods 1998, 37, 11.
(298) Comella, K.; Nakamura, M.; Melnik, K.; Chosy, J.; Zborowski, M.;
Cooper, M. A.; Fehniger, T. A.; Caligiuri, M. A.; Chalmers, J. J.
Cytometry 2001, 45, 285.
(299) Kim, Y.; Hong, S.; Lee, S. H.; Lee, K.; Yun, S.; Kang, Y.; Paek,
K. K.; Ju, B. K.; Kim, B. ReV. Sci. Instrum. 2007, 78, 074301.
(300) Powles, R.; Mehta, J.; Kulkarni, S.; Treleaven, J.; Millar, B.; Marsden,
J.; Shepherd, V.; Rowland, A.; Sirohi, B.; Tait, D.; Horton, C.; Long,
S.; Singhal, S. Lancet 2000, 355, 1231.
(301) Weissman, I. L. Science 2000, 287, 1442.
(302) Melnik, K.; Nakamura, M.; Comella, K.; Lasky, L. C.; Zborowski,
M.; Chalmers, J. J. Biotechnol. Prog. 2001, 17, 907.
(303) Moore, L. R.; Rodriguez, A. R.; Williams, P. S.; McCloskey, K.;
Bolwell, B. J.; Nakamura, M.; Chalmers, J. J.; Zborowski, M. J.
Magn. Magn. Mater. 2001, 225, 277.
(304) Cristofanilli, M.; Budd, G. T.; Ellis, M. J.; Stopeck, A.; Matera, J.;
Miller, M. C.; Reuben, J. M.; Doyle, G. V.; Allard, W. J.; Terstappen,
L.; Hayes, D. F. New Engl. J. Med. 2004, 351, 781.
(305) Leitzel, K.; Lieu, B.; Curley, E.; Smith, J.; Chinchilli, V.; Rychlik,
W.; Lipton, A. Clin. Cancer Res. 1998, 4, 3037.
(306) Lara, O.; Tong, X. D.; Zborowski, M.; Chalmers, J. J. Exp. Hematol.
2004, 32, 891.
(307) Hager, G.; Tong, D. C. C.; Schiebel, I.; Rezniczek, G. A.; Watrowski,
R.; Speiser, P.; Zeillinger, R. Gynecol. Oncol. 2005, 98, 211.
(308) Allan, A. L.; Vantyghem, S. A.; Tuck, A. B.; Chambers, A. F.; Chin-
Yee, I. H.; Keeney, M. Cytometry, Part A 2005, 65A, 4.
(309) Chalmers, J. J.; Haam, S.; Zhao, Y.; McCloskey, K.; Moore, L.;
Zborowski, M.; Williams, P. S. Biotechnol. Bioeng. 1999, 64, 509.
(310) Nakamura, M.; Decker, K.; Chosy, J.; Comella, K.; Melnik, K.;
Moore, L.; Lasky, L. C.; Zborowski, M.; Chalmers, J. J. Biotechnol.
Prog. 2001, 17, 1145.
(311) Yang, L. Y.; Lang, J. C.; Balasubramanian, P.; Jatana, K. R.; Schuller,
D.; Agrawal, A.; Zborowski, M.; Chalmers, J. J. Biotechnol. Bioeng.
2009, 102, 521.
(312) Pamme, N.; Wilhelm, C. Lab Chip 2006, 6, 974.
(313) Liu, Y. J.; Guo, S. S.; Zhang, Z. L.; Huang, W. H.; Baigl, D.; Xie,
M.; Chen, Y.; Pang, D. W. Electrophoresis 2007, 28, 4713.
(314) Mitrelias, T.; Palfreyman, J.; Jiang, Z.; Llandro, J.; Bland, J. A. C.;
Sanchez-Martin, R. M.; Bradley, M. J. Magn. Magn. Mater. 2007,
310, 2862.
(315) Mauk, M. G.; Ziober, B. L.; Chen, Z. Y.; Thompson, J. A.; Bau,
H. H. Oral-Based Diagnostics 2007, 1098, 467.
(316) Kim, H. S.; Son, O. T.; Kim, K. H.; Kim, S. H.; Maeng, S.; Jung,
H. I. Biotechnol. Lett. 2007, 29, 1659.
(317) Said, T. M.; Agarwal, A.; Zborowski, M.; Grunewald, S.; Glander,
H. J.; Paasch, U. J. Androl. 2008, 29, 134.
Chemical Reviews, 2010, Vol. 110, No. 3 1561
(318) Nath, P.; Strelnik, J.; Vasanji, A.; Moore, L. R.; Williams, P. S.;
Zborowski, M.; Roy, S.; Fleischman, A. Anal. Chem. 2009, 81, 43.
(319) Zborowski, M.; Tada, Y.; Malchesky, P. S.; Hall, G. S. Colloids Surf.,
A 1993, 77, 209.
(320) Zborowski, M.; Tada, Y.; Malchesky, P. S.; Hall, G. S. Appl. EnViron.
Microbiol. 1993, 59, 1187.
(321) Xia, N.; Hunt, T. P.; Mayers, B. T.; Alsberg, E.; Whitesides, G. M.;
Westervelt, R. M.; Ingber, D. E. Biomed. MicrodeVices 2006, 8, 299.
(322) Varshney, M.; Li, Y. B.; Srinivasan, B.; Tung, S. Sens. Actuators,
B: Chem. 2007, 128, 99.
(323) Beyor, N.; Seo, T. S.; Liu, P.; Mathies, R. A. Biomed. MicrodeVices
2008, 10, 909.
(324) Adams, J. D.; Kim, U.; Soh, H. T. Proc. Natl. Acad. Sci. U.S.A.
2008, 105, 18165.
(325) Karl, S.; David, M.; Moore, L.; Grimberg, B. T.; Michon, P.; Mueller,
I.; Zborowski, M.; Zimmerman, P. A. Malar. J. 2008, 7, 66.
(326) Wilhelm, C.; Gazeau, F.; Roger, J.; Pons, J. N.; Bacri, J. C. Langmuir
2002, 18, 8148.
(327) Tiwari, A.; Punshon, G.; Kidane, A.; Hamilton, G.; Seifalian, A. M.
Cell Biol. Toxicol. 2003, 19, 265.
(328) Valberg, P. A.; Butler, J. P. Biophys. J. 1987, 52, 537.
(329) Wang, N.; Butler, J. P.; Ingber, D. E. Science 1993, 260, 1124.
(330) Mannix, R. J.; Kumar, S.; Cassiola, F.; Montoya-Zavala, M.;
Feinstein, E.; Prentiss, M.; Ingber, D. E. Nat. Nanotechnol. 2008, 3,
36.
(331) Moller, W.; Nemoto, I.; Heyder, J. IEEE Trans. Nanobiosci. 2003,
2, 247.
(332) Romano, G.; Sacconi, L.; Capitanio, M.; Pavone, F. S. Opt. Commun.
2003, 215, 323.
(333) Ogiue-Ikeda, M.; Sato, Y.; Ueno, S. IEEE Trans. Nanobiosci. 2003,
2, 262.
(334) Koschwanez, J. H.; Carlson, R. H.; Meldrum, D. R. ReV. Sci. Instrum.
2007, 78, 044301.
(335) Aki, A.; Ito, O.; Morimoto, H.; Nagaoka, Y.; Nakajima, Y.; Mizuki,
T.; Hanajiri, T.; Usami, R.; Maekawa, T. J. Appl. Phys. 2008, 104,
094509.
(336) Lee, H.; Liu, Y.; Westervelt, R. M.; Ham, D. IEEE J. Solid-State
Circuits 2006, 41, 1471.
(337) Lee, H.; Liu, Y.; Ham, D.; Werstervelt, R. M. Lab Chip 2007, 7,
331.
(338) Polyak, B.; Fishbein, I.; Chorny, M.; Alferiev, I.; Williams, D.; Yellen,
B.; Friedman, G.; Levy, R. J. Proc. Natl. Acad. Sci. U.S.A. 2008,
105, 698.
(339) Schlaurman, T.; de Boer, R.; Patty, R.; Kooistra-Smid, M.; van Zwet,
A. J. Microbiol. Methods 2007, 71, 238.
(340) Leb, V.; Stocher, M.; Valentine-Thon, E.; Holzl, G.; Kessler, H.;
Stekal, H.; Berg, J. J. Clin. Microbiol. 2004, 42, 585.
(341) Yeung, S. W.; Lee, T. M. H.; Cai, H.; Hsing, I. M. Nucleic Acids
Res. 2006, 34, e118.
(342) Zhao, W. T.; Yao, S. J.; Hsing, I. M. Biosens. Bioelectron. 2006,
21, 1163.
(343) Jiang, G. F.; Harrison, D. J. Analyst 2000, 125, 2176.
(344) Lien, K. Y.; Lin, J. L.; Liu, C. Y.; Lei, H. Y.; Lee, G. B. Lab Chip
2007, 7, 868.
(345) Lund-Olesen, T.; Dufva, M.; Hansen, M. F. J. Magn. Magn. Mater.
2007, 311, 396.
(346) Kwakye, S.; Baeumner, A. Anal. Bioanal. Chem. 2003, 376, 1062.
(347) Yeung, S. W.; Hsing, I. M. Biosens. Bioelectron. 2006, 21, 989.
(348) Rida, A.; Gijs, M. A. M. Anal. Chem. 2004, 76, 6239.
(349) Lou, X. H.; Qian, J. R.; Xiao, Y.; Viel, L.; Gerdon, A. E.; Lagally,
E. T.; Atzberger, P.; Tarasow, T. M.; Heeger, A. J.; Soh, H. T. Proc.
Natl. Acad. Sci. U.S.A. 2009, 106, 2989.
(350) Mauk, M. G.; Ziober, B. L.; Chen, Z. Y.; Thompson, J. A.; Bau,
H. H. Oral-Based Diagnostics 2007, 1098, 467.
(351) Pioch, D.; Jurgen, B.; Evers, S.; Maurer, K. H.; Hecker, M.;
Schweder, T. Appl. Microbiol. Biotechnol. 2008, 78, 719.
(352) Grover, W. H.; Mathies, R. A. Lab Chip 2005, 5, 1033.
(353) Dubus, S.; Gravel, J. F.; Le Drogoff, B.; Nobert, P.; Veres, T.;
Boudreau, D. Anal. Chem. 2006, 78, 4457.
(354) Ferreira, H. A.; Graham, D. L.; Feliciano, N.; Clarke, L. A.; Amaral,
M. D.; Freitas, P. P. IEEE Trans. Magn. 2005, 41, 4140.
(355) Heer, R.; Eggeling, M.; Schotter, J.; Nohammer, C.; Pichler, R.;
Mansfeld, M.; Bruckl, H. J. Magn. Magn. Mater. 2007, 311, 244.
(356) Chiou, C. H.; Huang, Y. Y.; Chiang, M. H.; Lee, H. H.; Lee, G. B.
Nanotechnology 2006, 17, 1217.
(357) Chiou, C. H.; Lee, G. B. J. Micromech. Microeng. 2005, 15, 109.
(358) Mitnik, L.; Heller, C.; Prost, J.; Viovy, J. L. Science 1995, 267, 219.
(359) Kim, Y. S.; Morris, M. D. Anal. Chem. 1994, 66, 1168.
(360) Doyle, P. S.; Bibette, J.; Bancaud, A.; Viovy, J. L. Science 2002,
295, 2237.
(361) Dorfman, K. D.; Viovy, J. L. Phys. ReV. E 2004, 69, 011901.
(362) Minc, N.; Bokov, P.; Zeldovich, K. B.; Futterer, C.; Viovy, J. L.;
Dorfman, K. D. Electrophoresis 2005, 26, 362.
1562 Chemical Reviews, 2010, Vol. 110, No. 3
(363) Minc, N.; Futterer, C.; Dorfman, K. D.; Bancaud, A.; Gosse, C.;
Goubault, C.; Viovy, J. L. Anal. Chem. 2004, 76, 3770.
(364) Minc, N.; Slovakova, M.; Dorfman, K. D.; Futterer, C.; Bokov, P.;
Bilkova, Z.; Smadga, C.; Taverna, M.; Viovy, J. L. Houille Blanche-
ReVue Internationale De L Eau 2006, 51.
(365) Park, J. H.; Derfus, A. M.; Segal, E.; Vecchio, K. S.; Bhatia, S. N.;
Sailor, M. J. J. Am. Chem. Soc. 2006, 128, 7938.
(366) Pipper, J.; Zhang, Y.; Neuzil, P.; Hsieh, T. M. Angew. Chem., Int.
Ed. 2008, 47, 3900.
(367) Tsuchiya, H.; Okochi, M.; Nagao, N.; Shikida, M.; Honda, H. Sens.
Actuators, B: Chem. 2008, 130, 583.
(368) Schotter, J.; Kamp, P. B.; Becker, A.; Puhler, A.; Reiss, G.; Bruckl,
H. Biosens. Bioelectron. 2004, 19, 1149.
(369) Smistrup, K.; Bruus, H.; Hansen, M. F. J. Magn. Magn. Mater. 2007,
311, 409.
(370) Liu, Y. J.; Yao, D. J.; Chang, H. Y.; Liu, C. M.; Chen, C. Biosens.
Bioelectron. 2008, 24, 558.
(371) Wang, S. X.; Bae, S. Y.; Li, G. X.; Sun, S. H.; White, R. L.; Kemp,
J. T.; Webb, C. D. J. Magn. Magn. Mater. 2005, 293, 731.
(372) Brzeska, M.; Panhorst, M.; Kamp, P. B.; Schotter, J.; Reiss, G.;
Puhler, A.; Becker, A.; Bruckl, H. J. Biotechnol. 2004, 112, 25.
(373) Weizmann, Y.; Patolsky, F.; Lioubashevski, O.; Willner, I. J. Am.
Chem. Soc. 2004, 126, 1073.
(374) Mulvaney, S. P.; Mattoussi, H. M.; Whitman, L. J. BioTechniques
2004, 36, 602.
(375) Stahl, P. L.; Gantelius, J.; Natanaelsson, C.; Ahmadian, A.; Ander-
sson-Svahn, H.; Lundeberg, J. Genomics 2007, 90, 741.
(376) Tierno, P.; Reddy, S. V.; Roper, M. G.; Jofansen, T. H.; Fischer,
T. M. J. Phys. Chem. B 2008, 112, 3833.
(377) Besse, P.-A.; Boero, G. Appl. Phys. Lett. 2002, 80, 4199.
(378) Lagae, L.; Wirix-Speetjens, R.; Liu, C. X.; Laureyn, W.; Borghs,
G.; Harvey, S.; Galvin, P.; Ferreira, H. A.; Graham, D. L.; Freitas,
P. P.; Clarke, L. A.; Amaral, M. D. IEE Proc. Circuits DeVices Syst.
2005, 152, 393.
(379) Ejsing, L.; Hansen, M. F.; Menon, A. K.; Ferreira, H. A.; Graham,
D. L.; Freitas, P. P. J. Magn. Magn. Mater. 2005, 293, 677.
(380) Koets, M.; van der Wijk, T.; van Eemeren, J.; van Amerongen, A.;
Prins, M. W. J. Biosens. Bioelectron. 2009, 24, 1893.
(381) Bronzeau, S.; Pamme, N. Anal. Chim. Acta 2008, 609, 105.
(382) Liu, R. H.; Yang, J. N.; Lenigk, R.; Bonanno, J.; Grodzinski, P. Anal.
Chem. 2004, 76, 1824.
(383) Fan, Z. H.; Mangru, S.; Granzow, R.; Heaney, P.; Ho, W.; Dong,
Q. P.; Kumar, R. Anal. Chem. 1999, 71, 4851.
(384) Lien, K. Y.; Lee, W. C.; Lei, H. Y.; Lee, G. B. 2006 1st Ieee
International Conference on Nano/Micro Engineered and Molecular
Systems, 2006; Vols. 1-3, p 668.
(385) Lien, K. Y.; Lee, W. C.; Lei, H. Y.; Lee, G. B. Biosens. Bioelectron.
2007, 22, 1739.
(386) Liu, C. J.; Lien, K. Y.; Weng, C. Y.; Shin, J. W.; Chang, T. Y.; Lee,
G. B. Biomed. MicrodeVices 2009, 11, 339.
(387) Lien, K. Y.; Liu, C. J.; Lee, G. B. Mems 2008: 21st Ieee International
Conference on Micro Electro Mechanical Systems, Technical Digest,
2008; p 66.
(388) Cho, Y. K.; Lee, J. G.; Park, J. M.; Lee, B. S.; Lee, Y.; Ko, C. Lab
Chip 2007, 7, 565.
(389) Pipper, J.; Inoue, M.; Ng, L. F. P.; Neuzil, P.; Zhang, Y.; Novak, L.
Nat. Med. 2007, 13, 1259.
(390) Barbee, K. D.; Huang, X. H. Anal. Chem. 2008, 80, 2149.
(391) Mulvaney, S. P.; Cole, C. L.; Kniller, M. D.; Malito, M.; Tamanaha,
C. R.; Rife, J. C.; Stanton, M. W.; Whitman, L. J. Biosens.
Bioelectron. 2007, 23, 191.
(392) Kwakye, S.; Goral, V. N.; Baeumner, A. J. Biosens. Bioelectron.
2006, 21, 2217.
(393) Shen, W. F.; Schrag, B. D.; Carter, M. J.; Xie, J.; Xu, C. J.; Sun,
S. H.; Xiao, G. J. Appl. Phys. 2008, 103, 07a306.
(394) Graham, D. L.; Ferreira, H. A.; Feliciano, N.; Freitas, P. P.; Clarke,
L. A.; Amaral, M. D. Sens. Actuators, B: Chem. 2005, 107, 936.
(395) Hong, J. W.; Studer, V.; Hang, G.; Anderson, W. F.; Quake, S. R.
Nat. Biotechnol. 2004, 22, 435.
(396) Wellman, A. D.; Sepaniak, M. Anal. Chem. 2006, 78, 4450.
(397) Wellman, A. D.; Sepaniak, M. Anal. Chem. 2007, 79, 6622.
(398) Liu, Y. J.; Guo, S. S.; Zhang, Z. L.; Huang, W. H.; Baigl, D. M.;
Chen, Y.; Pang, D. W. J. Appl. Phys. 2007, 102, 084911.
(399) Peyman, S. A.; Iles, A.; Pamme, N. Chem. Commun. 2008, 1220.
(400) Herrmann, M.; Veres, T.; Tabrizian, M. Lab Chip 2006, 6, 555.
(401) Herrmann, M.; Roy, E.; Veres, T.; Tabrizian, M. Lab Chip 2007, 7,
1546.
(402) Lee, H.; Yasukawa, T.; Shiku, H.; Matsue, T. Biosens. Bioelectron.
2008, 24, 1000.
(403) Sista, R. S.; Eckhardt, A. E.; Srinivasan, V.; Pollack, M. G.; Palanki,
S.; Pamula, V. K. Lab Chip 2008, 8, 2188.
(404) Lehmann, U.; de Courten, D.; Vandevyver, C.; Parashar, V. K.; Gijs,
M. A. M. Microelectron. Eng. 2007, 84, 1669.
Gijs et al.
(405) Choi, J. W.; Oh, K. W.; Thomas, J. H.; Heineman, W. R.; Halsall,
H. B.; Nevin, J. H.; Helmicki, A. J.; Henderson, H. T.; Ahn, C. H.
Lab Chip 2002, 2, 27.
(406) Do, J.; Ahn, C. H. Lab Chip 2008, 8, 542.
(407) Morozov, V. N.; Morozova, T. Y. Anal. Chim. Acta 2006, 564, 40.
(408) Kim, K. S.; Park, J. K. Lab Chip 2005, 5, 657.
(409) Millen, R. L.; Kawaguchi, T.; Granger, M. C.; Porter, M. D.; Tondra,
M. Anal. Chem. 2005, 77, 6581.
(410) Degre, G.; Brunet, E.; Dodge, A.; Tabeling, P. Lab Chip 2005, 5,
691.
(411) Ku, B. Y.; Chan, M. L.; Ma, Z.; Horsley, D. A. J. Magn. Magn.
Mater. 2008, 320, 2279.
(412) Chung, S. H.; Hoffmann, A.; Bader, S. D.; Liu, C.; Kay, B.;
Makowski, L.; Chen, L. Appl. Phys. Lett. 2004, 85, 2971.
(413) Chung, S. H.; Hoffmann, A.; Guslienko, K.; Bader, S. D.; Liu, C.;
Kay, B.; Makowski, L.; Chen, L. J. Appl. Phys. 2004, 97, 10R101.
(414) Morozov, V. N.; Groves, S.; Turell, M. J.; Baileyt, C. J. Am. Chem.
Soc. 2007, 129, 12628.
(415) Moser, Y.; Lehnert, T.; Gijs, M. A. M. Lab Chip 2009, 9, 3261.
(416) Caulum, M. M.; Henry, C. S. Lab Chip 2008, 8, 865.
(417) Cohen-Tannoudji, L.; Bertrand, E.; Baudry, J.; Robic, C.; Goubault,
C.; Pellissier, M.; Johner, A.; Thalmann, F.; Lee, N. K.; Marques,
C. M.; Bibette, J. Phys. ReV. Lett. 2008, 100, 108301.
(418) Dupont, E.; Labonne, E.; Vandevyver, C.; Lehmann, U.; Charbon,
E.; Gijs, M. A. M. Anal. Chem. 2009, DOI: 10.1021/ac902241j.
(419) Hermann, M.; Veres, T.; Tabrizian, M. Anal. Chem. 2008, 80, 5160.
(420) De Palma, R.; Reekmans, G.; Liu, C. X.; Wirix-Speetjens, R.;
Laureyn, W.; Nilsson, O.; Lagae, L. Anal. Chem. 2007, 79, 8669.
(421) De Palma, R.; Reekmans, G.; Laureyn, W.; Borghs, G.; Maes, G.
Anal. Chem. 2007, 79, 7540.
(422) Hahn, Y. K.; Jin, Z.; Kang, J. H.; Oh, E.; Han, M. K.; Kim, H. S.;
Jang, J. T.; Lee, J. H.; Cheon, J.; Kim, S. H.; Park, H. S.; Park, J. K.
Anal. Chem. 2007, 79, 2214.
(423) Mulvaney, S. P.; Myers, K. M.; Sheehan, P. E.; Whitman, L. J.
Biosens. Bioelectron. 2009, 24, 1109.
(424) Soelberg, S. D.; Stevens, R. C.; Limaye, A. P.; Furlong, C. E. Anal.
Chem. 2009, 81, 2357.
(425) Aytur, T.; Foley, J.; Anwar, M.; Boser, B.; Harris, E.; Beatty, P. R.
J. Immunol. Methods 2006, 314, 21.
(426) Yang, S. Y.; Lien, K. Y.; Huang, K. J.; Lei, H. Y.; Lee, G. B. Biosens.
Bioelectron. 2008, 24, 855.
(427) Tang, D.; Yuan, R.; Chai, Y. Clin. Chem. 2007, 53, 1323.
(428) Dittmer, W. U.; de Kievit, P.; Prins, M. W. J.; Vissers, J. L. M.;
Mersch, M. E. C.; Martens, M. J. Immunol. Methods 2008, 338, 40.
(429) Kiely, J.; Hawkins, P.; Wraith, P.; Luxton, R. IET Sci. Meas. Techn.
2007, 1, 270.
(430) Bruls, D. M.; Evers, T. H.; Kahlman, J. A. H.; van Lankvelt, P. J. W.;
Ovsyanko, M.; Pelssers, E. G. M.; Schleipen, J. J. H. B.; de Theije,
F. K.; Verschuren, C. A.; van der Wijk, T.; van Zon, J. B. A.; Dittmer,
W. U.; Immink, A. H. J.; Nieuwenhuis, J. H.; Prins, M. W. J. Lab
Chip 2009, DOI: 10.1039/B913960E.
(431) Chen, H. X.; Busnel, J. M.; Peltre, G.; Zhang, X. X.; Girault, H. H.
Anal. Chem. 2008, 80, 9583.
(432) Chen, X.; Cui, D. F.; Liu, C. C. Electrophoresis 2008, 29, 1844.
(433) Lacharme, F.; Vandevyver, C.; Gijs, M. A. M. Microfluid. Nanofluid.
2009, 7, 479.
(434) Peyman, S.; Iles, A.; Pamme, N. The 11th International Conference
on Miniaturized Systems for Chemistry and Life Sciences, µTAS
2007 Conference, Paris, France, 2007; p 1270.
(435) Shikida, M.; Takayanagi, K.; Honda, H.; Ito, H.; Sato, K. J.
Micromech. Microeng. 2006, 16, 1875.
(436) De Palma, R.; Liu, C. X.; Barbagini, F.; Reekmans, G.; Bonroy, K.;
Laureyn, W.; Borghs, G.; Maes, G. J. Phys. Chem. C 2007, 111,
12227.
(437) Morozova, T. Y.; Morozov, V. N. Anal. Biochem. 2008, 374, 263.
(438) Baudry, J.; Bertrand, E.; Lequeux, N.; Bibette, J. J. Phys.: Condens.
Matter 2004, 16, R469.
(439) Baudry, J.; Rouzeau, C.; Goubault, C.; Robic, C.; Cohen-Tannoudji,
L.; Koenig, A.; Bertrand, E.; Bibette, J. Proc. Natl. Acad. Sci. U.S.A.
2006, 103, 16076.
(440) Lai, J. J.; Nelson, K. E.; Nash, M. A.; S., H. A.; Yager, P.; Stayton,
P. S. Lab Chip 2009, 9, 1997.
(441) Cole-Hamilton, D. J. Science 2003, 299, 1702.
(442) Abu-Reziq, R.; Alper, H.; Wang, D. S.; Post, M. L. J. Am. Chem.
Soc. 2006, 128, 5279.
(443) Hu, A. G.; Yee, G. T.; Lin, W. B. J. Am. Chem. Soc. 2005, 127,
12486.
(444) Luo, S. Z.; Zheng, X. X.; Xu, H.; Mi, X. L.; Zhang, L.; Cheng, J. P.
AdV. Synth. Catal. 2007, 349, 2431.
(445) Stevens, P. D.; Li, G. F.; Fan, J. D.; Yen, M.; Gao, Y. Chem.
Commun. 2005, 4435.
(446) Zheng, Y.; Stevens, P. D.; Gao, Y. J. Org. Chem. 2006, 71, 537.
Microfluidic Applications of Magnetic Particles
(447) Li, J.; Zhang, Y. M.; Han, D. F.; Gao, Q.; Li, C. J. Mol. Catal. A:
Chem. 2009, 298, 31.
(448) Lee, J.; Lee, D.; Oh, E.; Kim, J.; Kim, Y. P.; Jin, S.; Kim, H. S.;
Hwang, Y.; Kwak, J. H.; Park, J. G.; Shin, C. H.; Hyeon, T. Angew.
Chem., Int. Ed. 2005, 44, 7427.
(449) Kang, J. H.; Park, J. K. Small 2007, 3, 1784.
(450) Kang, J. H.; Park, J. K. Ieee 20th IEEE International Conference on
Micro Electro Mechanical Systems (MEMS 2007), Kobe, Japan,
2007; p 99.
(451) Ding, S. J.; Xing, Y. C.; Radosz, M.; Shen, Y. Q. Macromolecules
2006, 39, 6399.
(452) Shokouhimehr, M.; Piao, Y. Z.; Kim, J.; Jang, Y. J.; Hyeon, T. Angew.
Chem., Int. Ed. 2007, 46, 7039.
(453) Wang, Z. F.; Shen, B.; Zou, A. H.; He, N. Y. Chem. Eng. J. 2005,
113, 27.
(454) Wang, Z. F.; Xiao, P. F.; Shen, B.; He, N. Y. Colloids Surf., A 2006,
276, 116.
(455) Yoon, T. J.; Lee, W.; Oh, Y. S.; Lee, J. K. New J. Chem. 2003, 27,
227.
(456) Teunissen, W.; Bol, A. A.; Geus, J. W. Catal. Today 1999, 48, 329.
(457) Yi, D. K.; Lee, S. S.; Ying, J. Y. Chem. Mater. 2006, 18, 2459.
(458) Jun, C. H.; Park, Y. J.; Yeon, Y. R.; Choi, J. R.; Lee, W. R.; Ko,
S. J.; Cheon, J. Chem. Commun. 2006, 1619.
(459) Dyal, A.; Loos, K.; Noto, M.; Chang, S. W.; Spagnoli, C.; Shafi, K.;
Ulman, A.; Cowman, M.; Gross, R. A. J. Am. Chem. Soc. 2003,
125, 1684.
Chemical Reviews, 2010, Vol. 110, No. 3 1563
(460) Gardimalla, H. M. R.; Mandal, D.; Stevens, P. D.; Yen, M.; Gao, Y.
Chem. Commun. 2005, 4432.
(461) Zheng, Y.; Duanmu, C.; Gao, Y. Org. Lett. 2006, 8, 3215.
(462) Bromberg, L.; Hatton, T. A. Ind. Eng. Chem. Res. 2005, 44, 7991.
(463) Stevens, P. D.; Fan, J. D.; Gardimalla, H. M. R.; Yen, M.; Gao, Y.
Org. Lett. 2005, 7, 2085.
(464) Gao, X.; Yu, K. M. K.; Tam, K. Y.; Tsang, S. C. Chem. Commun.
2003, 2998.
(465) Llopis, X.; Pumera, M.; Alegret, S.; Merkoci, A. Lab Chip 2009, 9,
213.
(466) Krenkova, J.; Foret, F. Electrophoresis 2004, 25, 3550.
(467) Slovakova, M.; Minc, N.; Bilkova, Z.; Smadja, C.; Faigle, W.;
Futterer, C.; Taverna, M.; Viovy, J. L. Lab Chip 2005, 5, 935.
(468) Le Nel, A.; Minc, N.; Smadja, C.; Slovakova, M.; Bilkova, Z.; Peyrin,
J. M.; Viovy, J. L.; Taverna, M. Lab Chip 2008, 8, 294.
(469) Li, Y.; Xu, X. Q.; Yan, B.; Deng, C. H.; Yu, W. J.; Yang, P. Y.;
Zhang, X. M. J. Proteome Res. 2007, 6, 2367.
(470) Li, Y.; Yan, B.; Deng, C. H.; Yu, W. J.; Xu, X. Q.; Yang, P. Y.;
Zhang, X. M. Proteomics 2007, 7, 2330.
(471) Liu, J. Y.; Lin, S.; Qi, D. W.; Deng, C. H.; Yang, P. Y.; Zhang,
X. M. J. Chromatogr. 2007, 1176, 169.
(472) Jankovicova, B.; Rosnerova, S.; Slovakova, M.; Zverinova, Z.;
Hubalek, M.; Hernychova, L.; Rehulka, P.; Viovy, J. L.; Bilkova, Z.
22nd International Symposium on Microscale Bioseparations and
Methods for Systems Biology, Berlin, Germany, 2008; p 64.
CR9001929
1564 Chem. Rev. 2010, 110, 1564–1610

Synthesis of Heterocycles Mediated by Benzotriazole. 1. Monocyclic Systems

Alan R. Katritzky* and Stanislaw Rachwal

Center for Heterocyclic Compounds, Department of Chemistry, University of Florida, Gainesville, Florida 32611-7200

Received May 22, 2009

Contents 9. Five-Membered Rings with Three or Four 1594

Heteroatoms
1. Introduction 1564 9.1. 1,2,3-Triazole 1594
2. Preparation of Simple Derivatives of Benzotriazole 1565 9.2. 1,2,4-Triazole 1594
Used for Heterocyclic Synthesis 9.3. 1,2,4-Oxadiazole 1595
2.1. By Nucleophilic Aliphatic Substitution 1565 9.4. 1,3,4-Oxadiazole 1595
2.2. Addition to Multiple CC Bonds 1566 9.5. Tetrazole 1596
2.3. By Adducts to Carbonyl Groups of Aldehydes 1567 10. Pyridine 1596
and Their Conversions
10.1. Hexahydropyridine (Piperidine) 1596
2.4. By Carbanions Stabilized by a Benzotriazolyl 1568
Substituent 10.1.1. N-Substituent 1596
3. Three-Membered Rings 1568 10.1.2. Ring Formation 1597
3.1. Aziridine and Azirene 1568 10.2. Tetrahydropyridine 1599
3.2. Oxirane 1569 10.3. Transformations of Ring Substituents 1599
4. Four-Membered Rings 1570 10.4. Electrophilic and Nucleophilic Substitutions of 1600
the Aromatic Ring
5. Pyrrole 1570
10.5. Aromatic Ring Formation 1600
5.1. Nonaromatic Rings 1570
10.5.1. Pyridones and Thiopyridones 1600
5.1.1. Pyrrolidines 1570
10.5.2. Pyridines 1602
5.1.2. Pyrrolidinones 1573
11. Pyran and Thiopyran 1603
5.1.3. Pyrrolines 1573
11.1. Pyran 1603
5.2. Aromatic Ring Substituents 1574
11.2. Thiopyran 1604
5.2.1. Substituents at the Nitrogen Atom 1574
12. Six-Membered Rings with Two or More 1605
5.2.2. Electrophilic Substitution at Ring Carbon 1574 Heteroatoms
Atoms
12.1. Pyridazine 1605
5.3. Aromatic Ring Formation 1576
12.2. Pyrimidine 1605
5.3.1. One Substituent at C-2 1576
12.3. Pyrazine 1606
5.3.2. Two Substituents at C-2 and C-3 1577
12.4. Oxazines 1607
5.3.3. Two Substituents at C-2 and C-5 1577
12.5. 1,4-Dithiin 1607
5.3.4. Two Substituents at C-3 and C-4 1577
12.6. 1,3,5-Triazine 1608
5.3.5. Three Substituents at C-2, C-3, and C-4 1578
12.7. 1,2,4,5-Tetrazine 1608
5.3.6. Four Substituents at Carbon Atoms 1579
13. Conclusion 1608
5.3.7. Benzotriazolyl Substituent at the Ring 1579
14. References 1608
6. Furan 1580
6.1. Nonaromatic Rings 1580
1. Introduction
6.1.1. Tetrahydrofurans 1580
6.1.2. 2,5-Dihydrofurans 1580 Throughout the years since our first papers in the 1980s
6.2. Electrophilic Substitution of the Aromatic Ring 1581 on the application of benzotriazole derivatives in organic
6.3. Aromatic Ring Formation 1582 synthesis,1,2 tremendous progress has been achieved in the
7. Thiophene 1583 field of benzotriazole chemistry. Benzotriazole intermediates
7.1. Nonaromatic Rings 1583 are now commonly used for introduction of a variety of
7.2. Electrophilic Substitution of the Aromatic Ring 1583 functional groups into molecules. Five major functions of
benzotriazole in organic transformations are illustrated in
7.3. Aromatic Ring Formation 1583
Figure 1. Many aspects of the application of benzotriazole
8. Five-Membered Rings with Two Heteroatoms 1584 methodology in organic synthesis have been reviewed;
8.1. Pyrazole 1584 however, there is only one outdated review3d of a limited
8.2. Imidazole 1586 scope that is specifically devoted to the synthesis of
8.3. Isoxazole 1589 heterocyclic molecules. With this paper, we fill this existing
8.4. Oxazole 1590 gap and hope to make the work of chemists who are
8.5. Thiazole 1592 struggling with the construction and derivatization of het-
erocyclic systems a bit easier.
10.1021/cr900204u  2010 American Chemical Society
Published on Web 10/02/2009
Synthesis of Heterocycles Mediated by Benzotriazole Chemical Reviews, 2010, Vol. 110, No. 3 1565

Alan R. Katritzky is Kenan Professor of Chemistry and Director for the

Center of Heterocyclic Compounds at the University of Florida. He studied,
researched, and taught in the United Kingdom at the Universities of Oxford,
Cambridge, and East Anglia before crossing the Atlantic to take up his
present post in 1980. He has researched in diverse areas of organic and
physical-organic chemistry but none more deeply than in the chemistry
of benzotriazole, a topic which blossomed in his group after Stan Rachwal
joined him in 1983. Benzotriazole chemistry was first summarized in
Chemical Reviews in 1998; that manuscript has since received 264
citations. It is a particular pleasure for him to have been asked by Stan
to share in the authorship of the present summary, which updates part of
the earlier review.

Figure 1. Reactivity profile of benzotriazole.

Scheme 1

Scheme 2

Stanislaw Rachwal was born in Jodlowka and was raised in Krakow,

Poland. He received a Ph.D. in organic chemistry from Jagiellonian
University in Krakow and was nominated to the position of Adjunct
Professor at that university in 1980. His main research at that time was
focused on the chemistry of ferrocenophanes. During a sabbatical leave
in 1984, he joined Professor Alan R. Katritzky at the University of Florida
to lay a foundation for application of benzotriazole in organic synthesis.
He returned to the University of Florida in 1988, where as a group leader
he effectively contributed to the research on derivatives of benzotriazole.
His collaboration with Professor Katritzky resulted in over 40 scientific
papers on benzotriazole. Since 1993, he has worked in the pharmaceutical
industry specializing in CNS drugs with a primary focus on heterocyclic
compounds.
The aim of this review is to provide practical guidance
for synthetic chemists. Bearing in mind that the major interest
in heterocycles is the synthesis of biologically active
compounds, we arranged the material systematically accord-
ing to the size and shape of the molecules. The nature of
the heteroatoms and their number and positions in the
molecule are used as secondary discriminators. This way,
any chemist searching for bioisosteres of a heterocyclic
scaffold or a heterocyclic substituent will find a whole range
of useful structures.
2. Preparation of Simple Derivatives of
Benzotriazole Used for Heterocyclic Synthesis
2.1. By Nucleophilic Aliphatic Substitution
Alkylation of benzotriazole (1) with alkyl halides or
sulfates in the presence of a base leads to mixtures of
1-alkylbenzotriazoles 2 and 2-alkylbenzotriazoles 3. The ratio
of product 2 to product 3 depends on the bulkiness of the
alkyl group and varies from 78:22 (R ) Et) to 50:50 (R )
C6H11CH2) (Scheme 1).4,5
Direct alkylation of benzotriazole with alcohols in the
presence of triphenylphosphine and NBS provides some
progress (Scheme 2). It is assumed that the first step is the
formation of reactive intermediates 4, which are then attacked
by benzotriazole in the SN2 fashion to give derivatives 5.
The reaction is regioselective and provides exclusively
1-alkyl-, 1-(arylmethyl)-, 1-(2-alken-1-yl)-, and 1-(2-alkyn-
1566 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 3
1-yl)benzotriazoles in 60-84% yields. Secondary alcohols
give the corresponding alkyl derivatives in low yields, while
tertiary alcohols do not alkylate benzotriazole under these
conditions.6
In an aqueous micellar medium with cetyltrimethylam-
monium bromide surfactant, benzotriazole is alkylated re-
gioselectively at N-1 with n-propyl and n-butyl bromides,
but activated alkylating agents (benzyl chloride, allyl bro-
mide, phenacyl chloride, etc.) produce mixtures of benzot-
rizol-1-yl and -2-yl isomers in ratios varying from 55:45 to
80:20.7 Use of ionic liquids as media generally provides
higher regioselectivity; however, the trend is opposite to that
found under micellar conditions: phenacyl bromide and
similar compounds provide exclusively benzotriazol-1-yl
derivatives, and n-alkyl halides give mixtures of benzotriazol-
1-yl and -2-yl derivatives in a ratio of 15:1.8
Microwave irradiation can facilitate the alkylation. Thus,
compound 6 is cleanly prepared in 95% yield upon irradiation
of benzotriazole and the corresponding benzyl bromide in
DMF for 40 s.9 Very often microwave-assisted alkylation
of benzotriazole works best when no solvent is used; for
example, derivative 7 is prepared in this way in 94% yield.10
In another example, benzotriazole with ethyl chloroacetate
and K2CO3 in ethyl acetate catalyzed by polyethylene glycol
(PEG 400) gives a mixture of ethyl benzotriazol-1-ylacetate
(8; 56%) and its benzotriazol-2-yl isomer (15%).11
In the presence of tetrabutylammonium bromide catalyst,
benzotriazole reacts with 1,2-dibromoethane in 40% NaOH
at 60 °C to give a mixture of 2-bromoethyl derivatives 9
and 10. When the reaction mixture is heated at 140 °C, HBr
is eliminated to provide a mixture of 1-vinylbenzotriazole
(11) and 2-vinylbenzotriazole (12). Column chromatography
of the mixture provides 11 in 34% yield and 12 in 10% yield
(Scheme 3).12
2.2. Addition to Multiple CC Bonds
Under strongly acidic conditions (10 molar equiv of
TsOH), benzotriazole adds to unactivated alkenes to afford
a mixture of 1-alkyl- and 2-alkylbenzotriazoles. Because
Katritzky and Rachwal
Scheme 4
Scheme 5
protonation of the double bond with formation of the
corresponding carbocation is the first step, Markovnikov’s
rule is followed, and derivatives with a benzotriazolyl
substituent at the terminal carbon atom are not observed.
Two examples of such additions are depicted in Scheme 4.
Thus, with styrene, benzotriazol-1-yl (13) and -2-yl (14)
isomers are formed in a ratio of 6.5:1 in a total yield of 46%.
Cyclohexene does not react with benzotriazole at 80 °C;
however, at 120 °C, a mixture of derivatives 15 and 16 is
obtained in a ratio of 1.1:1 and a total isolated yield of 56%.13
Mixing benzotriazole with acrolein (17) in equimolar
amounts results in an exothermic reaction leading to adducts
18 and 19 (R1 ) R2 ) R3 ) H) in a ratio of 77:23. Addition
of benzotriazole to crotonaldehyde (R1 ) R2 ) H, R3 ) Me)
requires initiation by heating the reagents to 60-80 °C. A
similar addition to methyl vinyl ketone gives 18 and 19 (R1
) Me, R2 ) R3 ) H) in a 95:5 ratio. More sterically hindered
mesityl oxide generates equimolar amounts of 18 and 19 (R1
) R2 ) R3 ) Me). In most cases, simple recrystallization
of the crude mixtures provides pure benzotriazol-1-yl deriva-
tives 18 (Scheme 5).14
Benzotriazole with ethyl propiolate (20, R1 ) H, R2 )
Et) in refluxing toluene gives a mixture of adducts 21a and
22a in a nearly quantitative yield with a cis:trans (21a:22a)
ratio of 39:61. A similar reaction with methyl 2-butynoate
(20, R1 ) R2 ) Me) requires a CuI catalyst and results in a
1:1 mixture of 21b and 21b (Scheme 6).15
When the electron-deficient alkene contains a good leaving
group X at the double bond, addition of benzotriazole may
be followed by elimination of X (or HX) with restoration of
the double bond. The total effect is a nucleophilic substitution
of group X by a benzotriazolide anion. Two examples of
such reactions are shown in Scheme 7. Thus, adduct 24 is
obtained from a low-temperature reaction of benzotriazole
with methyl 2-(trifluoroacetyl)vinyl sulfone (23). At slightly
elevated temperature, adduct 24 eliminates spontaneously
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 6
Scheme 7
Scheme 8
methanesulfinic acid to give alkenone 25 in 88% yield.16
Addition of benzotriazolide anion to the R-carbon of (E)-
(2-phenylvinyl)phenyliodonium tetrafluoroborate (26) results
in unstable benzyl anion 27, which rapidly eliminates
iodobenzene to afford (E)-1-(2-phenylvinyl)benzotriazole
(28) in 64% yield.17
2.3. By Adducts to Carbonyl Groups of
Aldehydes and Their Conversions
Benzotriazole reacts eagerly with formaldehyde, even in
aqueous solutions, giving well-characterized 1-(hydroxym-
ethyl)benzotriazole (29) in practically quantitative yield.18
Treatment with thionyl chloride converts 29 into 1-(chlo-
romethyl)benzotriazole (30).18 In a more practical one-pot
approach, a mixture of benzotriazole, 37% aqueous form-
aldehyde, and toluene is heated under a Dean-Stark trap to
remove water. Thionyl chloride is then added, and the
mixture is heated under reflux for 2 h to give a solution of
30 in toluene that can be directly used for further transforma-
tions.19
Substitution of the chlorine atom in 30 with nucleophiles
is an easy process giving rise to a variety of synthetically
valuable products 31. The yields are generally good; for the
examples given in Scheme 8, they vary from 44% (Nu )
PhSO2) to 82% (Nu ) N3).20
Condensation of 29 with aromatic amines proceeds rapidly
in refluxing ethanol and provides crystalline 1-(amino-
methyl)benzotriazoles 32 (R1 ) aryl, R2 ) H) in practically
Chemical Reviews, 2010, Vol. 110, No. 3 1567
Scheme 9
Scheme 10
quantitative yields (Scheme 9).21 Dialkylamines and N-
alkylarylamines also react with 29 to give 32 (R1 ) alkyl,
R2 ) alkyl or aryl).22 In solutions, benzotriazol-1-yl deriva-
tives 32 obtained from secondary amines are in equilibria
with their benzotriazol-2-yl isomers 34.23-26 The intermediate
iminium cations 33 can be easily trapped in reactions with
nucleophiles.1,22,27-30
Reactions of primary aliphatic amines with 2 molar equiv
of 29 provide N,N-bis(benzotriazol-1-ylmethyl)amines 35 (R3
) alkyl), in mixtures with their benzotriazol-2-yl isomers.21,22
Heating at reflux a toluene solution of 29 (2 molar equiv)
and an aromatic amine with azeotropic removal of water
allows the preparation of 35 (R3 ) aryl) as a mixture of
three isomers: Bt-1/Bt-1, Bt-1/Bt-2, and Bt-2/Bt-2.31 29 reacts
also with amides to give the corresponding N-benzotriazol-
1-ylmethyl derivatives.28,32 In an example in Scheme 10,
reaction of 29 with formamide gives product 36 in 63% yield.
It is subsequently converted into isocyanide 37 in 66% yield
with phosphorus oxychloride and diisopropylamine.32
The reactions described in Schemes 8-10 are not limited
to 1-(hydroxymethyl)benzotriazole. Higher aliphatic alde-
hydes and aromatic aldehydes with electron-deficient rings
also give solid adducts with benzotriazole, 39, that are stable
enough to be recrystallized and their purity confirmed by
CHN analysis. Condensation with alcohols further stabilizes
the molecule, allowing the preparation of alkoxy compounds
40 derived from a variety of aldehydes in pure form.33
Condensation with primary aromatic amines converts 39 into
stable crystalline aminals 38 (R2 ) aryl, R3 ) H) that are
isolated in 76-98% yields.21,34 Dibenzylamine, pyrrolidine,
1568 Chemical Reviews, 2010, Vol. 110, No. 3 Katritzky and Rachwal

Scheme 11 Scheme 12

piperidine, morpholine, and thiomorpholine also give prod- Scheme 13

ucts 38 stable enough to be characterized.22,35 Thionyl
chloride converts 1-(1-hydroxyalkyl)benzotriazoles 39 into
1-chloroalkyl derivatives 42; in the presence of excess
benzotriazole, 1,1-bisbenzotriazol-1-ylalkanes 41 and their
benzotriazol-2-yl isomers are formed.36 Chlorides 42 react
readily with nucleophiles to give derivatives 43 in high
yields: 74% for R1 ) t-Bu, Nu ) CN; 91% for R1 ) n-Pr,
Nu ) PhCO2; 88% for R1 ) n-Pr, Nu ) PhS; 97% for R1
) i-Pr, Nu ) i-PrO (Scheme 11).19,36

2.4. By Carbanions Stabilized by a Benzotriazolyl

Substituent
Table 1. Examples of Compounds 51 [E(2) ) H] and 52 [E(2) *
Many synthetic applications of benzotriazole derivatives H]
are based on the ability of the benzotriazolyl substituent to
X E(1) E(2) yield (%) ref
stabilize an adjacent carbanion. Even simple 1-(n-alkyl)ben-
zotriazoles 5 can be converted to anions 44 (R1 ) H or alkyl) Me3Si n-C6H13 H 80 39
Me3Si 3-oxocyclohexyl H 70 39
by treatment with n-BuLi. The consecutive treatment with Me3Si n-C6H13 Me 80 39
alkyl halides converts 44 into 1-alkylbenzotriazoles 45 carbazol-9-yl 1-hydroxycyclohexyl H 91 40
bearing secondary alkyl groups, e.g., R1 ) Me, R2 ) Et (72% carbazol-9-yl PhCH2 PhS 95 41
yield), and R1 ) Et, R2 ) n-C10H21 (65% yield). Carbonyl carbazol-9-yl PhCH2 CO2Et 87 41
electrophiles can be used as well to trap 44. Thus, the anion indol-1-yl n-PrCO H 95 42
PhS PhCH2 n-Bu 86 43
derived from 1-methylbenzotriazole (R1 ) H) adds readily MeS n-Bu allyl 89 45
to the carbonyl groups of benzaldehyde to give alcohol 46 EtO PhCtC PhCHOH 73 47
(R1 ) R4 ) H, R3 ) Ph) in 95% yield, acrylaldehyde to
give 46 (R1 ) R4 ) H, R3 ) CH2dCH) in 57% yield, or more complex molecules 52.39 Derivatives 49 with X )
benzophenone to give product 46 (R1 ) H, R3 ) R4 ) Ph) carbazol-9-yl readily form anions 50, which react with
in 70% yield. In a reaction of 44 with ethyl benzoate, ketone electrophiles to generate derivatives 51, in which the
47 (R1 ) H, R5 ) Ph) is obtained in 54% yield, whereas a substituent BtCHX can be considered as a protected carbonyl
reaction of 44 with CO2 gives carboxylic acid 48 (R1 ) H) group of aldehydes.40 Their further lithiation and treatment
in 54% yield.4 Reactions of lithiated 1-(arylmethyl)benzot- with other electrophiles generates 52 as protected forms of
riazoles 5 with electrophiles are even more effective; e.g., complex ketones.41 Other heterocycles, such as pyrrole,
1-benzylbenzotriazole treated with n-BuLi and then n-BuI indole, benzimidazole, imidazole, and 1,2,4-triazole, can also
gives 45 (R1 ) Ph, R2 ) n-Bu) in 82% yield,37 1-(2-methoxy- be effective as substituents X in preparation of 51 and 52.42
3-methylbenzyl)benzotriazole treated with n-BuLi followed Aryl sulfides 49 (X ) ArS) are also suitable for the chemistry
by p-tolualdehyde gives alcohol 46 (R1 ) 2-methoxy-3- outlined in Scheme 13, providing interesting intermediates
methylphenyl, R3 ) p-tolyl, R4 ) H) in 87% yield, and when for further transformations.43-45 Even alkoxy derivatives 51
the same 5 is lithiated and treated with ethyl benzoate, ketone (X ) RO) are acidic enough to be lithiated and converted
47 (R1 ) 2-methoxy-3-methylphenyl, R5 ) Ph) is obtained with electrophiles into 52.46-50 Examples of products syn-
in 76% yield (Scheme 12).38 thesized according to Scheme 13 are given in Table 1.
Anions 50 derived from compounds of the type BtCH2X
(49), where X ) a heteroatom or a group attached by a 3. Three-Membered Rings
heteroatom, allow the generation of a variety of chemical
structures. Thus, compound 49 with X ) Me3Si treated with 3.1. Aziridine and Azirene
n-BuLi followed by electrophiles E(1)+ gives molecules 51,
in which the trimethylsilyl group also can be substituted by Bromination of 12 provides 2-(1,2-dibromoethyl)benzot-
other electrophiles. Repeated lithiation of 51 followed by riazole (53) in 93% yield. Reactions with amines convert
treatment with another electrophile, E(2)+, generates even 53 into 1-alkyl-2-benzotriazol-2-ylaziridines 55. Intermedi-
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 14
Scheme 15
Scheme 16
ates 54 can be observed in the reaction mixtures monitored
by NMR. The yields of aziridines 55 are high (80-96%)
except of that derived from benzylamine (30%) (Scheme
14).51 The analogous reaction sequence of 11 does not give
aziridines but only 1-(1-bromovinyl)benzotriazole.51
Treatment of 30 with lithium bis(trimethylsilyl)amide
results in very unstable anion 56, but when the reaction is
carried out in the presence of Schiff bases, anion 56 can be
effectively trapped, leading to aziridines 57 in high yields
(85-90%) (Scheme 15).51 Due to their reactivity, aziridines
55 and 57 are convenient starting materials for the synthesis
of other heterocycles (see section 5 on pyrrole).
Condensation of 1-(azidomethyl)benzotriazole (31, Nu )
N3) with triphenylphosphine provides [N-(benzotriazol-1-
ylmethyl)imino]triphenylphosphorane (betmip, 58) in a
practically quantitative yield. In a reaction of 58 with
methylmagnesium bromide, the benzotriazolyl moiety is
substituted by a methyl group to give phosphazene interme-
diate 59. Using a one-pot procedure, 59 is treated with
styrene oxide and heated at reflux in THF for 48 h to give
aziridine 60 in 55% yield (Scheme 16).52
Oximes 61 are prepared in 91-93% yields from the
corresponding aryl benzotriazol-1-ylmethyl ketones, hy-
droxylamine hydrochloride, and NaOH in refluxing ethanol/
Chemical Reviews, 2010, Vol. 110, No. 3 1569
Scheme 17
Scheme 18
water. They can be converted into their tosylates 62, but a
large excess ofKOH converts them directly into 2H-azirines
63 (58-66% yields). The benzotriazolyl moiety in azirines
63 is readily substituted by nucleophiles (organomagnesium
reagents, potassium phthalimide, or sodium thiophenoxide)
to give disubstituted azirines 64 in 50-79% yields (Scheme
17).53
3.2. Oxirane
Oxiranes bearing a benzotriazol-1-yl substituent on the ring
can be prepared in practically quantitative yields by epoxi-
dation of benzotriazol-1-ylalkenes with dimethyldioxirane,
e.g., conversion of alkene 28 to oxirane 65 (Scheme 18). At
very low temperatures, substitution of the R-proton in oxirane
65 is possible; just its treatment with LDA at -116 °C
followed by benzyl bromide leads to R-benzyloxirane 67 in
51% yield, via lithiated intermediate 66. Using acyl chlorides
or benzophenone as an electrophile in this reaction provides
the corresponding oxiranes with acyl or diphenylhydroxy-
methyl substituents, respectively. At higher temperatures,
rearrangement of lithiated oxirane 66 to the appropriate
ketone is observed. The stereochemistry is preserved.54
1570 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 19
Scheme 20
1-Alkenylbenzotriazoles 68 are readily prepared by isomer-
ization of the corresponding allyl derivatives catalyzed by
t-BuOK. Reaction of lithiated 68 with electrophiles provides
R-substituted derivatives 69 in 41-80% yields. Epoxidation
of the double bond with m-chloroperbenzoic acid converts
69 into oxiranes 70 in 43% (R ) Me, E ) Ph2COH) to 80%
(R ) H, E ) 1-hydroxycyclohexyl) yields. Oxiranes 70 can
be hydrolyzed to R-hydroxy ketones 71 in good yields
(Scheme 19).55
30 can be converted into its anion by treatment with LDA
at -40 °C. The anion is trapped by ketones to provide a
convenient synthesis of benzotriazol-1-yloxiranes 72 (68-75%
yields). Treated with n-BuLi, oxiranes 72 give lithiated
intermediates 73 that react with various electrophiles to
provide tetrasubstituted oxiranes 74 in 68% (R1 ) Ph, R2 )
Et, E ) PhCO) to 92% (R1 ) R2 ) Ph, E ) PhCHOH)
yields. Upon treatment with perchloric acid, benzoyl deriva-
tives 74 (E ) PhCO) undergo ring-opening with elimination
of benzotriazole to provide 3-hydroxy 1,2-diones, which
themselves are interesting starting materials for other het-
erocycles (Scheme 20).56
R-Benzotriazol-1-ylalkyl ethers 40 with n-BuLi generate
anions 75 that add readily to the carbonyl group of ketones
to give alkoxy anions 76. The presence of zinc bromide
promotes elimination of the benzotriazolide anion (as a zinc
complex) with the formation of phenoxy- or methoxyoxiranes
77 that are isolated in 56% (77a), 49% (77b), and 71% (77c)
yields. Several other ketones and aldehydes provided oxiranes
Katritzky and Rachwal
Scheme 21
Scheme 22
77 that were unstable under the reaction conditions undergo-
ing ring-opening and rearrangement to the corresponding
R-phenoxy or R-alkoxy ketones (Scheme 21).57
4. Four-Membered Rings
N-Substituted 1-(aminomethyl)benzotriazoles 78 treated
with ethyl bromodifluoroacetate, zinc powder, and trimeth-
ylsilyl chloride (Reformatsky-type conditions) give N-
substituted ethyl 3-amino-2,2-difluoropropionates 79a (89%
yield) and 79b (86% yield). Cyclization of 79 promoted by
tert-butylmagnesium chloride furnishes N-protected 3,3-
difluoroazetidin-2-ones 80a (45% yield) and 80b (65% yield)
(Scheme 22). Several other attempts at the synthesis of such
compounds failed.58
Anions generated from 1-acylbenzotriazoles 81 upon their
treatment with LDA add readily to the carbonyl groups of
aldehydes and ketones. Nucleophilic attack of the resultant
alkoxy anion 82 on the acyl carbonyl group followed by
elimination of a benzotriazolide anion results in formation
of oxet-2-one 83. This simple method allows the synthesis
of oxetane derivatives 83 in 42% (R1 ) Et, R2 ) PhCH2CH2,
R3 ) H) to 90% (R1 ) n-C6H13, R2 ) t-Bu, R3 ) H) yields
(Scheme 23).59
5. Pyrrole
5.1. Nonaromatic Rings
5.1.1. Pyrrolidines
Pyrrolidines monosubstituted at C-2 are easily generated
from N-Boc-protected 3-chloropropanamine 84. In the first
step, condensation with formaldehyde and benzotriazole
converts 84 into its N-benzotriazol-1-ylmethyl derivative 85.
In the second step, anion 86 generated from 85 and n-BuLi
undergoes cyclocondensation to pyrrolidine 89.60 In another
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 23
Scheme 24
Scheme 25
approach, pyrrolidinone 87 is reduced by DIBAL-H to
2-hydroxypyrrolidine 88, which upon condensation with
benzotriazole is converted to 89.61 With organomagnesium
reagents catalyzed by zinc chloride, the benzotriazolyl moiety
in 89 is readily substituted by an aryl or a phenylethynyl
group to furnish 2-substituted pyrrolidines 90 in 62-86%
yields.60 Treatment of 89 with (tributylstannyl)lithium affords
2-stannylpyrrolidine 91 in 78% yield (Scheme 24).61
The reaction can be carried out enantioselectively using a
chiral auxiliary instead of Boc. One such possibility is
depicted in Scheme 25. Derivative 92, obtained from
2-pyrrolidinone and 1S,2R-trans-2-cumylcyclohexyl chlo-
roformate, is reduced with DIBAL-H at low temperature and
then treated with benzotriazole to give compound 93. In a
reaction with (tributylstannyl)lithium, the benzotriazolyl
moiety is replaced by a tributylstannyl group to give
predominantly 2-(tributylstannyl)pyrrolidine 94 with the
Chemical Reviews, 2010, Vol. 110, No. 3 1571
Scheme 26
Scheme 27
predominant S configuration (80:20). The diastereomers can
be readily separated by chromatography. With DIBAL-H at
elevated temperature, derivative 94 is converted into (S)-1-
methyl-2-(tributylstannyl)pyrrolidine (95).61
N,N-Bis(benzotriazol-1-ylmethyl)amines 35, with their
benzotriazol-2-yl isomers, are readily prepared by condensa-
tion of 29 with primary amines.21,62,63 In solutions, they exist
in equilibria with ionic forms 96. Such mixtures treated with
samarium diiodide generate radicals 97 that are rapidly
trapped by vinyl groups of styrenes or vinyltrimethylsilane
to provide more stable radicals 98. Consecutive ionization
and reduction with samarium diiodide of the second benzo-
triazolylmethyl substituent provides a diradical that couples
intramolecularly to pyrrolidine 99. This simple process allows
preparation of 1,3-disubstituted pyrrolidines 99 in 49-85%
yields (Scheme 26).64
When one of the substituents on the amine nitrogen atom
is able to trap a radical formed by treatment of N-(R-
aminoalkyl)benzotriazole with SmI2, the intramolecular cy-
clization may occur (Scheme 27). Thus, condensation of the
5-(benzylamino) derivative of trans-2-pentenoic ester (100)
1572 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 28
Scheme 29
with an aldehyde and benzotriazole gives tertiary amine 101.
The R-aminoalkyl radical 102 generated by treatment of 101
with SmI2 is intramolecularly trapped by the double bond,
leading to the more stable radical 103. Final reduction by a
second molecule of SmI2 provides 1,2,3-trisubstituted pyr-
rolidine 104. Among the wide range of starting aldehydes,
the yields of products 104 vary from 29% (R ) PhCH2OCH2)
to 74% (R ) i-Pr). For pyridyl and aliphatic substituents,
cis stereoisomers strongly predominate (e.g., the cis:trans
ratio is 86:14 for R ) n-C10H21), but for most aromatic
substituents, the ratio is reversed (e.g., cis:trans ) 1:9 for R
) Ph).65,66
The same method is effectively used to introduce three
substituents into positions 1, 2, and 4 of pyrrolidines. Thus,
condensation of the 5-(benzylamino) derivative of trans-2-
alkenoic ester 105 with 29 gives tertiary amine 106. The
aminomethyl radical generated from 106 by SmI2 is trapped
by the double bond, leading to pyrrolidine 107. Mixtures of
two stereoisomers in ratios 55:45 (107a), 64:36 (107b), and
65:35 (107c) are produced; however, their identities have
not been established (Scheme 28).67
N-(Benzotriazolylmethyl)-N-[(trimethylsilyl)methy-
l]amines 109 are readily prepared by condensation of
N-[(trimethylsilyl)methyl]amines 108 with formaldehyde and
benzotriazole. Thermally induced desilylation of 109 gener-
ates azomethine ylides 110. Electron-deficient alkenes added
to the reaction mixture trap efficiently 110 to produce
pyrrolidines 111 in 71-90% yields. The cycloaddition
reaction proceeds stereospecifically with retention of the
olefinic dipolarophile configuration (Scheme 29).68
Condensation of succinaldehyde (obtained by hydrolysis
of 2,5-dimethoxytetrahydrofuran) with benzotriazole and (S)-
2-phenylglycinol provides (3S,5R,7aR)-5-benzotriazol-1-yl-
3-phenyloxazolo[2,1-b]pyrrolidine (112). Oxazolopyrrolidine
112 is a convenient synthon for asymmetric syntheses of
Katritzky and Rachwal
Scheme 30
Scheme 31
2-substituted and 2,5-disubstituted pyrrolidines. Thus, treat-
ment with organomagnesium reagents converts 112 into
mixtures of cis (113) and trans (114) pyrrolidines that can
be easily separated by chromatography. For aliphatic groups
R, the yields of cis isomers 113 vary from 53% to 60%,
whereas minor trans isomers 114 are obtained in 17-30%
yields. For R ) Ph, both isomers are equally abundant.
Hydrogenation of the intermediates readily removes the chiral
auxiliary to provide 2,5-disubstituted pyrrolidines 115 and
116; however, with R ) Ph, such a procedure causes opening
of the pyrrolidine ring (Scheme 30).69
In a reaction with allyltrimethylsilane, the benzotriazolyl
moiety in 112 is substituted with an allyl group to provide
derivative 117 in 45% yield. Hydrogenation of 117 cleaves
the chiral auxiliary to give (2R)-2-propylpyrrolidine (118).
Alternatively, a reaction of 112 with triethyl phosphite leads
to chiral phosphonate 120, via intermediate 119, with 68%
overall yield (Scheme 31).69
Alkylation of lithiated compounds 121 with 1-bromo-3-
chloropropane gives chloropropyl derivatives 122 in good
yields (80-89%). Subsequent substitution of the chlorine
atom by an alkylamino group is easily accomplished by
heating 122 and amine R2NH2 in DMF. Intramolecular
substitution of the benzotriazole moiety by the amino group
in amines 123 occurs at room temperature in the presence
of a palladium catalyst to furnish 2-vinylpyrrolidines 124 in
overall 65-85% yields for the last two steps (Scheme 32).70
N-Derivatization of L-proline benzyl ester (125) provides
an example of application of benzotriazole methodology to
the synthesis of biologically important structures. Thus,
condensation of 125 with 29 provides intermediate 126,
which is subsequently subjected to reactions with organo-
Synthesis of Heterocycles Mediated by Benzotriazole Chemical Reviews, 2010, Vol. 110, No. 3 1573

Scheme 32 Scheme 34

Scheme 33

Scheme 35

bismuth reagents generated from RX and BiCl3/Al. Interest-

ingly, the reaction is carried out in aqueous solution to
provide N-substituted proline esters 127 in 18% (R ) Me)
to 70% (R ) allyl) yields (Scheme 33).71 Scheme 36

5.1.2. Pyrrolidinones
Condensation of 2,5-dimethoxy-2,5-dihydrofuran (128)
with benzotriazole and an amine carried out in refluxing
acetic acid produces 5-benzotriazolylpyrrolidin-2-one in good
yield and with strong preference for the benzotriazol-1-yl
isomer 129. Substitution of the benzotriazole moiety in 129
with nucleophiles gives 5-substituted 2-pyrrolidinones
130-132. Since both benzotriazolyl isomers of 129 have
similar reactivity, the isomeric mixtures can be used directly
in this step. Thus, with organozinc reagents, pyrrolidinones
130 are obtained in 49-87% yields. The yields of pyrroli-
dinones 131 are 49-90% and those of 5-oxo-2-pyrrolidi-
nylphosphonates 132 49-85% (Scheme 34).72
Michael addition of benzotriazole to N-methylmethacry-
lamide carried out at 150 °C provides β-benzotriazol-1-
ylpropionamide 133 in 45% yield. The dianion generated
from 133 by treatment with 2 molar equiv of n-BuLi is
trapped by methyl 4-methylbenzoate to give 2-pyrrolidinone white precipitate. Anion 139 derived from imine 138 upon
134 in 63% yield. The stereochemistry is not determined its treatment with n-BuLi adds readily to electron-deficient
(Scheme 35).73 double bonds. In the case of methyl acrylate, pyrrolidine
Acylation of 1 with N-protected L-glutamic acid 135 in anion 140 resulting from such an addition eliminates
the presence of thionyl chloride readily provides derivative spontaneously a benzotriazolide anion to give 1,2-pyrroline
136. Condensation with L-amino acids in aqueous acetonitrile 141 in 96% yield (Scheme 37).75 According to NMR data,
in the presence of triethylamine converts 136 into 2-pyrro- the molecule of 141 has cis geometry. However, trans
lidinones 137. Washing the crude products with 4 N HCl substitution is found in pyrroline 142 (63% yield) derived
gives pure 137 in 58% (a), 88% (b), and 70% (c) yields from a reaction of 139 with acrylonitrile. Reaction of 139
(Scheme 36).74 with dimethyl fumarate leads to pyrroline 143 (95% yield)
with trans,trans orientation of its substituents at C-3, C-4,
and C-5. A similar reaction with diethyl maleate gives a
5.1.3. Pyrrolines
mixture of stereoisomers (Scheme 38).75
Condensation of benzotriazole with benzaldehyde and Alkylation of benzotriazole with 2,3-dichloro-1-propene
ammonia in ethanol provides imine 138 in 92% yield as a and Na2CO3 in DMSO/H2O provides 1-(2-chloropropen-3-
1574 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 37
Scheme 38
yl)benzotriazole (144) in 56% yield and its isomer 2-(2-
chloropropen-3-yl)benzotriazole in 30% yield. Stronger bases
smoothly convert 144 into 1-allenylbenzotriazole (145),
isolated in 90% yield when 20% NaOH is used. LDA is able
to remove an allenyl proton from 145, and the resultant anion
can be trapped by various electrophiles. In a useful approach,
144 is treated with 2 equiv of LDA followed by an
electrophile. Several R-substituted 1-allenylbenzotriazoles are
thus obtained. When N-(4-methoxybenzylidene)aniline is
used as an electrophile, allenyl product 146 rearranges
partially to its propargyl isomer, and both components can
be separated by column chromatography. Heating the allenyl
product in acetic anhydride results in its conversion to 3,4-
pyrroline 147, which is isolated in 30% overall yield (Scheme
39).76
5.2. Aromatic Ring Substituents
5.2.1. Substituents at the Nitrogen Atom
Introduction of a benzotriazol-1-ylmethyl substituent onto
the pyrrole nitrogen atom can be easily achieved on heating
a mixture of pyrrole, 30, and powdered NaOH in DMSO.
Product 148 treated with n-BuLi generates anion 150, which
is subsequently subjected to reactions with electrophiles. In
an example of such reactions, a solution of 148 in THF is
treated with n-BuLi at -78 °C followed by benzyl bromide
and is allowed to warm slowly to room temperature.
Chromatographic purification provides derivative 151 in 47%
yield. The benzotriazolyl moiety can be easily removed from
151 by treatment with LiAlH4 to give 1-phenethylpyrrole
(152) in 96% yield. Reactions of 148 with RMgBr produce
N-substituted pyrroles 149 (42-73% yields) with the groups
Katritzky and Rachwal
Scheme 39
Scheme 40
containing one carbon atom more than in the Grignard
reagent (Scheme 40).42,77
Addition of an anion generated from 148 to the carbonyl
group of p-tolualdehyde results in alcohol 153, which is
separated in 35% yield. An analogous reaction with cyclo-
hexanone provides alcohol 154 in 76% yield. Treatment of
154 with phenylmagnesium bromide allows a convenient
removal of the benzotriazolyl auxiliary, providing alcohol
155 in 50% yield. A similar substitution of the benzotriazolyl
moiety with a phenyl ring in compound 153 gives the
corresponding alcohol in 60% yield (Scheme 41).42,77
In a reaction of lithiated benzotriazolyl derivative 148 with
benzophenone, alcohol 156 is obtained in 72% yield. The
use of carboxylic esters as electrophiles is exemplified by
synthesis of ketone 157 (R ) 4-MeC6H4, 53% yield). To
demonstrate the removal of the benzotriazolyl auxiliary from
such compounds, 157 is treated with activated zinc in acetic
acid to give ketone 160 in 50% yield. In a reaction of lithiated
148 with diphenyl disulfide, disubstituted product 158 is
obtained in 67% yield, whereas a reaction with trimethylsilyl
chloride leads to a mixture of monosubstituted (159) and
disubstituted products in an approximate ratio of 1:3 (Scheme
42).42
5.2.2. Electrophilic Substitution at Ring Carbon Atoms
When the condensation of benzotriazole, formaldehyde
(aqueous solution), and aliphatic amines used in equimolar
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 41
Scheme 42
Scheme 43
amounts is carried out in diethyl ether or dichloromethane,
1-(aminomethyl)benzotriazoles 32 are obtained in high yields
(>90%). In solution, compounds 32 exist in equilibria with
their benzotriazol-2-yl isomers. This equilibration is expected
to proceed through iminium ion 161, and just this cation must
be involved in reactions of 32 with nucleophiles. In reactions
of 32 with pyrrole or 1-methylpyrrole, 2-substituted products
162 are obtained in 62-96% yields. With tertiary amines,
the reactions proceed well in the presence of AlCl3; however,
a milder Lewis acid, ZnCl2, has to be used for secondary
amines (Scheme 43).78
Not only amino, but also other electron-donating groups
at C-R of the benzotriazolyl substituents can facilitate
dissociation of the N-C bond, generating reactive carboca-
tions as well. Thus, complexation of the benzotriazolyl
Chemical Reviews, 2010, Vol. 110, No. 3 1575
Scheme 44
Scheme 45
moiety in furyl and thienyl derivatives 163 with ZnCl2
promotes their dissociation to cations 164 that readily attack
the C-2 atom of N-methylpyrrole to give products 165 in
56% (X ) O) and 52% (X ) S) yields (Scheme 44).79
1-Acylbenzotriazoles 167, readily prepared in reactions of
the corresponding carboxylic acids with 1-(methanesulfo-
nyl)benzotriazole,80 are found to be convenient acylating
agents for pyrroles. Thus, in their reactions with unsubstituted
pyrrole catalyzed by TiCl4, 2-acylpyrroles 166 are obtained
in 21% (R ) 4-Et2NC6H4) to 91% (R ) 2-furyl) yields.
1-Methylpyrrole subjected to similar reactions gives 2-acyl-
1-methylpyrroles 168 in 51% (R ) 4-Et2NC6H4) to 94% (R
) 2-furyl) yields. To direct the acylation reaction to C-3 of
pyrrole, N-protection with a bulky triisopropylsilyl group is
used. 3-Acylpyrroles 169 are isolated in generally good yields
(54-92%). To prove the protection concept, one of the
derivatives 169 (R ) 4-MeC4H4) was deprotected using
tetrabutylammonium fluoride to give 3-acylpyrrole 170 in
98% yield (Scheme 45).81
The benzotriazole approach allows the introduction of
complex groups onto C-2 of pyrroles under mild conditions.
Thus, N-protected 1-(R-aminoacyl)benzotriazoles 171 derived
from chiral R-amino carboxylic acids can be conveniently
prepared by mixing benzotriazole with thionyl chloride
followed by N-protected R-amino carboxylic acid in a 4:1:1
ratio. Friedel-Crafts acylation of pyrrole with 171 in the
presence of AlCl3 provides chiral 2-(aminoacyl)pyrroles 172
in 45-82% yields (Scheme 46).82
1-Cyanobenzotriazole (173), prepared in 92% yield from
benzotriazole, NaH, and cyanogen bromide, is a convenient
reagent for introduction of CN groups onto nucleophilic
carbon atoms. With lithiated 1-methylpyrrole, an electrophilic
attack of 173 occurs at the ring C-2 atom, providing nitrile
174 in 55% yield (Scheme 47).83
1576 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 46
Scheme 47
Scheme 48
5.3. Aromatic Ring Formation
5.3.1. One Substituent at C-2
The anion delivered from 1-propargylbenzotriazole (175)
upon its treatment with n-BuLi readily adds to N-tosy-
larylimines to give sulfonamides 176 in 76-92% yields.
When heated with NaOH in ethanol, compounds 176 are
converted into 2-arylpyrroles 179 in 47-60% yields. The
proposed reaction mechanism involves isomerization of
alkyne 176 to allene 177 followed by intramolecular cy-
cloaddition to form pyrroline 178. Finally, under basic
conditions, the molecule eliminates benzotriazole and the
tosylate protecting group to furnish pyrrole 179 (Scheme
48).84
The reaction of acrolein with benzotriazole and morpholine
provides product 180 in 84% yield as a mixture with its
benzotriazol-2-yl isomers.14 Upon treatment with sodium
hydride, one of the benzotriazolyl substituents is eliminated
to give enamine 181 in 74% yield as a stable crystalline
product.85 Additions of lithiated 181 to Schiff bases presum-
ably give unstable intermediate diamines 182 that by
cyclization and elimination of benzotriazole and morpholine
Katritzky and Rachwal
Scheme 49
Scheme 50
are converted into 1,2-diarylpyrroles 183, isolated in 60-68%
yields (Scheme 49).86
Alkoxy analogues of enamine 181 can be applied in such
reactions as well. Thus, R-ethoxy derivative 184, easily
prepared by condensation of benzotriazole with acrolein
diethyl acetal, undergoes a smooth rearrangement promoted
by ZnBr2 to (γ-ethoxyallyl)benzotriazole (185). After lithia-
tion, the anion is trapped by Schiff bases to give anions 186.
Catalyzed by ZnBr2, intermediates 186 undergo cyclization
with elimination of benzotriazole and ethanol to furnish 1,2-
diarylpyrroles 187 in 55-63% yields (Scheme 50A).87
Lithiated N-allylbenzotriazole (188) readily adds to the
CdN bond of Schiff bases derived from aromatic or
heteroaromatic aldehydes and amines to give amines 189 in
Synthesis of Heterocycles Mediated by Benzotriazole Chemical Reviews, 2010, Vol. 110, No. 3 1577

Scheme 51 Scheme 52

Scheme 53

46-85% yields. In the presence of a palladium catalyst and

cupper(II) oxidizing agent, amines 189 are smoothly con-
verted to pyrroles 190 in 46-79% yields, except 190 with
Ar1) 4-(MeO)C6H4 and Ar2 ) Ph where the yield is only
21% (Scheme 50B).88

5.3.2. Two Substituents at C-2 and C-3

N-(Benzotriazol-1-ylmethyl)thiobenzamide (191) is readily
prepared by condensation of benzotriazole with formaldehyde
and thiobenzamide.89,90 Treatment with n-BuLi followed by
iodomethane converts thiobenzamide 191 into S-methyl
thioimidate 192 in 87% yield. The anion generated from 192
adds to acrylonitrile to give pyrrolidine anion 193a, which
spontaneously eliminates benzotriazole and thiomethoxide
anion to give pyrrole 194a (88% yield). Similar reactions
with 2-vinyl- and 4-vinylpyridines provide 2-phenyl-3-
pyridinylpyrroles 194b (36% yield) and 194c (64% yield). Scheme 54
In the case of substituents X with moderate electron-
withdrawing power, the additions may occur in both modes,
giving rise to mixtures of regioisomers, as demonstrated in
a reaction with methyl vinyl ketone leading to 2,3-disubsti-
tuted pyrrole 195 (34% yield) and 2,4-disubstituted pyrrole
196 (18% yield) (Scheme 51).91
Addition of 58 to PPh3PdCH2 provides phosphonium salt
197, which is easily deprotonated by n-BuLi to give ylide
198. When the reaction mixture is quenched with 1,2-diaryl
1,2-diones, 2,3-disubstituted pyrroles 198a-c are obtained
in 67%, 75%, and 58% yields, respectively (Scheme 52).92,93
riazole moiety is eliminated and the N-protection is removed
5.3.3. Two Substituents at C-2 and C-5 to provide 2,5-disubstituted pyrroles 204a (56% yield) and
204b (43% yield) (Scheme 53).84
The synthetic method for pyrroles monosubstituted at C-2
shown in Scheme 48 can be modified for the preparation of 5.3.4. Two Substituents at C-3 and C-4
2,5-disubstituted pyrroles. Thus, treatment of 175 with 2
molar equiv of n-BuLi generates dianion 200, which is In the presence of a base, isocyanide 37 adds readily to
subsequently alkylated with alkyl iodides to give alkynyl electron-deficient double bonds, and the resultant pyrroline
anions 201. Consecutive addition of N-tosyl-1-naphthylm- intermediates 205 spontaneously eliminate benzotriazole to
ethyleneimine provides adducts 202 in 72-77% yields. give 3,4-disubstituted pyrroles in 92% (206a), 81% (206b),
Heated with an ethanolic solution of NaOH, compounds 202 and 40% (206c) yields (Scheme 54).94
tautomerize to their allene forms and undergo cyclization to Addition of lithiated allylbenzotriazoles 207 to the CdN
pyrrolines 203. Under the reaction conditions, the benzot- bond of isothiocyanates produces anions 208 that are
1578 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 55
Scheme 56
methylated in situ to provide methyl 2,2-disubstituted
3-buteniminothiates 209 in 59-80% yields. Treated with
Lewis acids, compounds 209 undergo cyclization to 2-
(methylthio)pyrroles 210. Yields of 210 determined by gas
chromatography are high (75-94%); however, the yields of
pure materials separated by column chromatography are
much lower (21-65%) due to the low stability of the
electron-rich pyrroles (Scheme 55).95
In the [1 + 2 + 2] annulation process shown in Scheme
56, alkylation of 1-benzylbenzotriazoles 211 with 2-bro-
moacetaldehyde diethyl acetal to give intermediates 212 is
followed by a reaction with N-benzylideneaniline to produce
adducts 213. Subsequent treatment with formic acid causes
cyclization of 213 to 5-ethoxypyrrolidines 214 that then
eliminate ethanol and benzotriazole to give pyrroles 215 in
66-74% yields.96
Due to the high strain energy of the three-membered ring,
aziridines are convenient starting materials for pyrroles. On
heating to 140 °C, the C-N bond in benzotriazol-2-
ylaziridines 55 is cleaved, and the dipolar species 216
undergo [3 + 2] cycloaddition to acetylenedicarboxylate to
form pyrrolines 217 that aromatize to pyrroles 218 by
elimination of benzotriazole (Scheme 57). The reaction is
relatively slow (72 h), providing products 218 in 25% (R )
n-C10H21) to 80% (R ) n-C5H11) yields.51
5.3.5. Three Substituents at C-2, C-3, and C-4
A 1 molar equiv of a base makes possible selective
methylation of the sulfur atom in thioamides 219 to give
thioimidates 220 in high yields. The anions derived from
Katritzky and Rachwal
Scheme 57
Scheme 58
220 on treatment with sodium hydride add readily to electron-
poor double bonds to create unstable anions 221 that
spontaneously eliminate benzotriazole and thiomethoxide to
generate pyrroles 222 (Scheme 58).91 Both reactions, me-
thylation and addition, can be conveniently carried out in
one pot when t-BuOK is used as a base.97 The yields of
pyrroles 222 are generally high (60-99%), except for the
case of R2 ) CF3 (20%).
A similar sequence is employed for the conversion of
N-(benzotriazol-1-ylmethyl)amides 223 into pyrroles 226.
Thus, in reactions with PCl5, amides 223 give chloro imines
224 that on treatment with t-BuOK are converted to nitrile
ylides 225. Benzyl esters of R,β-unsaturated acids acting as
dipolarophiles trap 225 to generate pyrroles 226 in 81-90%
yield (Scheme 59).98
In the presence of t-BuOK, benzotriazol-1-ylmethyl
isocyanide (BetMIC, 37) undergoes alkylation on the
methylene group to give isocyanide 227. The anion
derived from 227, upon treatment with t-BuOK, adds to
the electron-deficient double bonds of R,β-unsaturated
ketones, esters, or nitriles to produce pyrroles 228 in
30-92% yields (Scheme 60).94
Benzotriazol-1-ylaziridines 229 are thermally unstable. On
heating to 100 °C, the C-C bond of the aziridine ring is
cleaved to give azomethine ylides 230 that can be trapped
by diethyl acetylenedicarboxylate. The pyrroline intermedi-
ates that are formed in the first step spontaneously eliminate
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 59
Scheme 60
Scheme 61
benzotriazole to furnish pyrroles 231 in 81-88% yield
(Scheme 61).51
5.3.6. Four Substituents at Carbon Atoms
Similarly to the addition of imine 138 to dimethyl fumarate
(Scheme 38), the anion generated from 138 adds readily to
dimethyl acetylenedicarboxylate to give quantitatively tet-
rasubstituted pyrrole 234. The reaction is believed to proceed
through pyrrolinyl anion 232, which spontaneously expels a
benzotriazolide anion to generate neutral molecule 233,
which eventually rearranges to the more stable aromatic
tautomer 234 (Scheme 62).75
The synthetic method for 2,3,4-trisubstituted pyrroles based
on thioamides (Scheme 58) can be extended to 2,3,4,5-
tetrasubstituted pyrroles. Thus, as given in Scheme 63,
thioamide 235 is prepared in 92% yield by condensation of
thiobenzamide with 2,4-dichlorobenzaldehyde and benzot-
riazole. S-Methylthioimidate 236, obtained by methylation
of 235 with iodomethane in the presence of t-BuOK, is
treated in one pot with cinnamonitrile to give tetrasubstituted
pyrrole 237 in 63% yield.97
Chemical Reviews, 2010, Vol. 110, No. 3 1579
Scheme 62
Scheme 63
Scheme 64
5.3.7. Benzotriazolyl Substituent at the Ring
For most of the synthetic methods described in this
paragraph, pyrrole ring formation involves cycloaddition or
cyclocondensation reactions leading to unstable intermediate
pyrrolines or pyrrolidines that then spontaneously eliminate
benzotriazole and, in the case of pyrrolidines, also another
group to generate an aromatic system. However, under
special circumstances, the benzotriazolyl substituent may be
retained, and other groups leave. Thus, in Scheme 64,
vinamidinium salt 239, obtained in a reaction of benzotriazol-
1-ylacetic acid (238) with DMF and POCl3, reacts with ethyl
esters of glycine and its N-methyl analogue to give 4-ben-
zotriazol-1-ylpyrrole derivatives 241a (89% yield) and 241b
(91% yield). Here, the presumed intermediate pyrrolidines
1580 Chemical Reviews, 2010, Vol. 110, No. 3 Katritzky and Rachwal

Scheme 65 Scheme 67

Scheme 66

Scheme 68

240 preferentially eliminate two molecules of dimethylamine,

leaving the benzotriazolyl substituent unaffected. The only
role for the benzotriazolyl group in this conversion is
activation of R-protons, making the condensation and
elimination easier.99

6. Furan
6.1. Nonaromatic Rings
6.1.1. Tetrahydrofurans
Addition of benzotriazole to 1,2-dihydrofuran (242) pro-
duces a mixture of regioisomers 243 and 244 in a 7:1 ratio,
respectively. The mixture can be separated by chromatog-
raphy. Treatment with organomagnesium reagents smoothly
converts the mixture of 243 and 244 into 2-substituted furans
245 in 59-90% yields (Scheme 65).100
Addition of N-(pyrrolidin-1-ylmethyl)benzotriazole (246)
to 242 results in a mixture of stereoisomers and regioisomers
247 and 248. The mixture can be separated by chromatog-
raphy; however, when treated with phenylmagnesium bro-
mide at elevated temperature, it is converted exclusively into
trans-1,2-disubstituted furan 250 in 75% yield. The stereo-
chemistry of 250 can be explained by formation of tricyclic
ionic intermediate 249 (Scheme 66).101
The method depicted in Scheme 66 can be extended to
relatively complex substituents at C-3. Thus, compound
251a, readily available from a reaction of benzotriazole with
benzaldehyde acetal, adds to 242 in the presence of an acidic
catalyst to produce a complex mixture of regio- and
stereoisomers 252a and 253a. Treatment with methylmag-
nesium iodide allows the substitution of benzotriazole to
provide 2-methyl-3-(1-ethoxybenzyl)tetrahydrofurans
254a-257a. Isomers 254a-257a (each as a pair of enanti-
omers) are separated by column chromatography.101 Addition
of morpholin-4-yl derivative 251b to 242 provides isomers
252b and 253b. The separation of the four stereoisomers of
252b by chromatography is described. Conversions of the
individual stereoisomers 252b with methylmagnesium iodide
allows assignment of stereochemistry to the resultant deriva-
tives 254b-257b (Scheme 67).102
6.1.2. 2,5-Dihydrofurans
The anion generated from 175 readily adds to diaryl
ketones to produce alcohols 258 in high yields. Treatment
with NaOH in ethanol at 60-80 °C allows the rearrangement
of alkynes 258 into allenes 259, which, under the reaction
conditions, undergo cyclization to 2,5-dihydrofurans 260,
isolated in 68-90% yields. On heating of 260 with Grignard
reagents in toluene, the benzotriazolyl moiety is replaced by
an alkyl or aryl group to provide 2,2,5-trisubstituted 2,5-
dihydrofurans 261 in 82-95% yields (Scheme 68).103
The 1,1-disubstituted analogue of allene 259, product 262
(90% yield), forms on treatment of 144 with 2 molar equiv
of LDA followed by p-anisaldehyde. The first equivalent of
LDA is used to generate 1-allenylbenzotriazole, which is
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 69
Scheme 70
deprotonated with the second equivalent of LDA, and the
resulting anion is finally trapped by the aldehyde. Similarly
to 259, allene 262 undergoes cyclization under basic condi-
tions to give 2,3-disubstituted 2,5-dihydrofuran 263 in 60%
yield (Scheme 69).76
The iminium cation generated from ethyl R-(dimethy-
lamino)-R-benzotriazol-1-ylacetate (264) upon treatment with
AlCl3 adds to C-5 of 2-[(trimethylsilyl)oxy]furan with
elimination of the (trimethylsilyl)oxy group to give unstable
2,5-dihydrofuran-2-one 265. Spontaneous elimination of
dimethylamine from 265 in refluxing acetonitrile provides
ethyl (5-oxo-5H-furan-2-ylidene)acetate 266 in 83% yield.
The isomer ratio Z:E is 77:23. The isomers are readily
separated by column chromatography (Scheme 70).104
6.2. Electrophilic Substitution of the Aromatic
Ring
Furan can be lithiated at C-2 using n-BuLi in the presence
of TMEDA. The following treatment with ZnBr2 converts
the lithiated furan into 2-furylzinc bromide, a convenient
reagent for the preparation of furans monosubstituted at C-2.
Thus, condensation of N,N′-diphenylhydrazine with form-
aldehyde and benzotriazole provides trisubstituted hydrazine
267. In solutions, 267 exists in equilibrium with its benzo-
triazol-2-yl analogue, but both isomers are equally reactive.
With 2-furylzinc bromide, the benzotriazolyl moiety in 267
is substituted by a furyl group to give 2-[(N,N′-diphenylhy-
drazino)methyl]tetrahydrofuran (268) in 79% yield (Scheme
71).105
2,5-Disubstituted furans can be readily obtained by C-5
lithiation of 2-substituted furans followed by reactions
with electrophiles. As an example, 2-ethylfuran is first
treated with n-butyllithium and then with 1-(phenylsul-
fonyl)benzotriazole (269) to give derivative 270 in 80%
yield.106 When lithiated 2-ethylfuran is treated with 173,
5-ethyl-2-furonitrile (271) is obtained in 65% yield.83 Acy-
lation of 2-methylfuran with 1-acylbenzotriazoles 167 in the
Chemical Reviews, 2010, Vol. 110, No. 3 1581
Scheme 71
Scheme 72
Scheme 73
presence of a Lewis acid provides 5-acylated derivatives 272
in 54% (R2 ) 2-pyridyl) to 98% (R2 ) 4-Me2NC6H4) yields
(Scheme 72).107
In a Mannich-type reaction of 2-methylfuran with 1-(pi-
peridinomethyl)benzotriazole (273a) catalyzed by ZnCl2, 2,5-
disubstituted furan 274a is obtained in 87% yield. The
reaction can also be used for the introduction of N-
monoalkylated aminomethyl groups to C-5 of 2-methylfuran,
but the yields are somewhat lower: 58% for 274b and 47%
for 274c (Scheme 73).78
Methyl N-(1-benzotriazol-1-ylalkyl)carbamates 275 are
conveniently prepared by condensation of aldehydes with
benzotriazole and methyl carbamate.108 With 2-methylfuran
in the presence of ZnCl2, both the benzotriazolyl and the
carbamoyl groups in 275 are substituted to produce geminal
bis(5-methylfuran-2-yl)alkanes 276 in 88-90% yields (Scheme
74).79
The substitution can be carried out stepwise using two
different heterocyclic systems. Thus, when 2-methylfuran
is subjected to a reaction with 1 molar equiv of carbamate
275a, a mixture of compounds 277 (42%), 278 (17%),
and 276a (15%) is obtained. Treatment of the crude
reaction mixture with 1-methylpyrrole and ZnCl2 provides
2-[1-(1-methylpyrrol-2-yl)-2-methylpropyl]-5-methylfu-
1582 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 74
Scheme 75
Scheme 76
ran (279) in 56% yield. Obviously, both intermediates 277
and 278 have to react with 1-methylpyrrole to give the
same product (Scheme 75).79
6.3. Aromatic Ring Formation
Under reflux in acetic acid, benzotriazole reacts with 128
to give 2-benzotriazol-1-ylfuran (280) in 67% yield (Scheme
76). In the presence of amines, N-substituted 2-benzotriazol-
1-ylpyrroles are formed instead, but the yields are low
(1-10%).109
Alkylation of 185 occurs exclusively at the R-C, producing
derivatives 281, which, in their lithiated forms, add readily
to the carbonyl group of aldehydes. Upon treatment with
ZnBr2, resultant anions 282 are rapidly converted to 2,3-
disubstituted furans 283 that are isolated in 46-52% overall
yields (Scheme 77).88
In the presence of t-BuOK, oxiranes 285, obtained from
lithiated 175 and R-bromo ketones 284, undergo rearrange-
ment to furans 286, which are isolated in 51-61% yields.
When treated with electron-rich heterocycles and ZnCl2, the
benzotriazolyl moiety in 286 is substituted by a heterocyclic
ring to provide derivatives 287 in 57-86% yields. Alterna-
tively, alkylation of the methylene carbon in 286 provides
derivatives 288, which in a reaction with heterocycles and
ZnCl2 are consecutively converted to products 289 in
55-95% yields. In this way, 2,4-disubstituted furans (R1 )
H) or 2,3,5-trisubstituted furans (R1 * H) are prepared
(Scheme 78).110
Deprotonation of N,N-dimethylbenzotriazol-1-ylmethyl-
eneiminium chloride (290), readily available from a reaction
Katritzky and Rachwal
Scheme 77
Scheme 78
of DMF with 1-chlorobenzotriazole or 1-(trimethylsilyl)ben-
zotriazole, by triethylamine provides (dimethylamino)ben-
zotriazol-1-ylcarbene (291). In the presence of trans-1,2-
dibenzoylethylene, 1,4-cycloaddition of the carbene provides
2,3-dihydrofuran derivative 292, which in refluxing benzene
eliminates benzotriazole to give 2,3,5-trisubstituted furan 293,
isolated in 15% yield. A more complex rearrangement
process leads to a side product, furan 294, which is isolated
from the reaction mixture in 6% yield (Scheme 79).111
Benzotriazole with dibenzoylacetylene (295) and triph-
enylphosphine (equimolar amounts) gives a mixture of
compounds 296 (40%) and 297 (45%) (Scheme 80). No
open-chain benzotriazol-2-yl analogue of 296 or benzotriazol-
1-yl analogue of furan derivative 297 is detected.112 Just steric
hindrance in the reaction intermediates may be responsible
for that. To make the cyclization possible, the molecule 296
must first assume an E configuration to bring the carbonyl
groups in close proximity. Isomerization between Z and E
configurations in 296 can be achieved by addition of another
molecule of benzotriazole to the double bond followed by
its elimination, but rotation of the R-benzoyl group imposes
too much strain due to the steric repulsion between its phenyl
and the benzotriazolyl benzenoid ring. The less bulky
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 79
Scheme 80
benzotriazol-2-yl substituent in the benzotriazol-2-yl ana-
logue of 296 allows for such transformations without high
energetic barriers.
7. Thiophene
7.1. Nonaromatic Rings
Condensation of phenylpropargylaldehyde diethyl acetal
with benzotriazole readily provides 1-benzotriazol-1-yl-3-
phenylpropargyl ethyl ether (298). Treated with n-BuLi, 298
forms an anion that adds to a CdS bond in CS2 to give
dithiocarboxylate anion 299. An intramolecular attack of the
sulfur nucleophile on the triple bond in 299 leads to 3,3,5-
trisubstituted 2,3-dihydro-2-thiophenethione 300, isolated in
57% yield. A similar addition of lithiated 298 to the CdS
bond of 1-naphthyl isothiocyanate produces adduct 301,
which reacts further to provide 2,3-dihydro-2-(naphth-1-
ylimino)thiophene 302 in 64% yield (Scheme 81).113
7.2. Electrophilic Substitution of the Aromatic
Ring
Monosubstitution of thiophene at C-2 can be achieved
simply by acylation. 1-Acylbenzotriazoles 167 are efficient
acylating agents under mild conditions (see section 5 on
pyrrole and section 6 on furan). Thus, 167 react smoothly
with thiophene in the presence of a Lewis acid catalyst to
give 2-acylthiophenes 303 in 58% (R ) 4-Et2NC6H4) to 97%
(R ) 1-naphthyl) yields (Scheme 82).107
In another approach to monosubstituted thiophenes,
thiophene itself is lithiated at C-2 with n-BuLi and TMEDA
Chemical Reviews, 2010, Vol. 110, No. 3 1583
Scheme 81
Scheme 82
Scheme 83
and subsequently treated with ZnBr2 to give 2-thienylzinc
bromide. The bromozinc group can be substituted by various
electrophiles. In an example in Scheme 83, 2-thienylzinc
bromide reacts with N-(benzotriazol-1-ylmethyl)-N,N′-diphe-
nylhydrazine (267) to give product 304 in 53% yield.105
29 reacts with 2-methylthiophene in refluxing glacial acetic
acid to give 2,5-disubstituted thiophene 305 in 50% yield.
The anion derived from 305 adds to the carbonyl group of
cyclohexanone to provide alkoxide anion 306. In situ
treatment with ZnBr2 promotes rearrangement of anion 306
with elimination of the benzotriazolide anion and expansion
of the cyclohexanone ring. 2-(5-Methylthien-2-yl)cyclohep-
tanone (307) so obtained is isolated in 67% yield (Scheme
84).114
7.3. Aromatic Ring Formation
R-Substituted 1-allylbenzotriazoles 121 are readily pre-
pared by treatment of 1-allylbenzotriazole with n-BuLi
followed by alkylating agents. Removal of the remaining
R-proton in 121 by n-BuLi and consecutive treatment with
isothiocyanates leads to thioamides 308. Without purification,
thioamides 308 are then subjected to ZnBr2, which causes
cyclization and elimination of benzotriazole to provide
2-aminothiophenes 309 in 25% (R1 ) CH2dCHCH2, R2 )
1584 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 84
Scheme 85
Scheme 86
4-ClC6H4) to 80% (R1 ) 2-ClC6H4CH2, R2 ) Ph) overall
yields for the two steps (Scheme 85).95
By a different route to 2-aminothiophenes, benzotriazol-
1-ylacetone (310) is treated with phenyl isothiocyanate and
KOH to give adduct 311. Without separation, 311 is treated
in situ with phenacyl bromide to provide tetrasubstituted
thiophenes in 82% (312a) and 56% (312b) yields (Scheme
86).115
8. Five-Membered Rings with Two Heteroatoms
8.1. Pyrazole
Condensation of 1,4-diphenylsemicarbazide with formal-
dehyde and benzotriazole carried out in refluxing benzene
with a Dean-Stark trap provides 1,1,4-trisubstituted semi-
carbazide 313. Using ZnBr2 catalyst, the iminium ion
generated from 313 by cleavage of the benzotriazolyl-C
bond is trapped by a vinyl group in CH2dCHOEt, leading
to pyrazolidine 314, separated in 60% yield (Scheme 87).105
The most popular method for the preparation of pyrazoles
unsubstituted at the nitrogen atoms involves condensation
of hydrazine with β-dicarbonyl compounds. However, such
Katritzky and Rachwal
Scheme 87
Scheme 88
compounds are not always readily available. Various deriva-
tives that can be considered to be synthetic equivalents of
β-dicarbonyl compounds may simplify the process. An
example of such reactions is given in Scheme 88. Thus, salt
290, a Vilsmeier-type reagent, reacts with N-propylimines
315 generated from simple ketones to give enamino-iminium
salts 316. In situ treatment with hydrazine converts salts 316
into 3,4-disubstituted pyrazoles 317. The synthesis proceeds
well when imines 315 are derived from symmetrical ketones,
when substituent R1 is unreactive (315b and 315c), or when
there is a significant difference in steric hindrance between
both substituents (315d). In the last case, pyrazole 317d is
contaminated with a minor isomer (5%) formed by conden-
sation of salt 290 with the isobutylmethylene group. For
sterically less differentiated groups, regioselectivity is poor;
e.g., for 315e (R1 ) n-Bu, R2 ) Et), the ratio of regioisomers
in the corresponding pyrazoles 317 is 55:45. Despite these
limitations and moderate yields, 31-83%, the simplicity of
this method makes it an attractive alternative in pyrazole
synthesis.116
Condensation of equimolar amounts of 310 with p-
anisaldehyde in refluxing DMF in the presence of a catalytic
amount of triethylamine gives benzotriazol-1-ylchalcone 318
in 81% yield. Chalcone 318 reacts readily with hydrazine to
provide 3,4,5-trisubstituted pyrazole 319 in 73% yield. This
addition/cyclocondensation process must proceed through a
corresponding pyrazoline that is oxidized under the reaction
conditions to pyrazole 319 without elimination of benzot-
riazole. When an excess of 4-nitrobenzaldehyde is used in
the condensation with 310, both the methylene and the
methyl groups react to give dibenzylideneacetone 320 in 81%
yield. In a reaction with hydrazine, the unsubstituted ben-
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 89
Scheme 90
zylidene site in 320 is favored, providing pyrazole 321 in
61% yield (Scheme 89).117
2-Benzotriazol-1-ylvinamidinium salts 239 are convenient
reagents for the preparation of pyrazoles with a benzotriazolyl
substituent at C-4. In their reactions with arylhydrazines 322
and sodium carbonate in refluxing methanol, salts 239 are
converted into pyrazoles 323 in 75% (R ) OMe) to 91% (R
) Br) yields (Scheme 90).99
In another approach to 4-benzotriazol-1-ylpyrazoles, 310
is treated with N,N-dimethylformamide dimethyl acetal
(DMFDMA) to give enaminone 324 (74% yield). Cyclo-
condensation of 324 with phenylhydrazine provides 1,4,5-
trisubstituted pyrazole 325 in 74% yield. Using the same
reagents but in the reversed order converts 310 into a
different regioisomer, 1,3,4-trisubstituted pyrazole 327 (70%
yield), via phenylhydrazone 326 (85% yield) (Scheme 91).115a
When aldehydes replace ketones, pyrazoles unsubstituted
at C-3 result. Thus, condensation of benzotriazole with
chloroacetaldehyde dimethyl acetal provides benzotriazol-
1-yl derivative 328 as the main product (60% yield) and its
benzotriazol-2-yl analogue as a minor product (21% yield).
Treatment of lithiated acetal 328 with aldehydes results in
alcohols 329 (83-86% yields). Aqueous HCl converts
alcohols 329 to R,β-unsaturated aldehydes 330 in a practi-
cally quantitative manner. Addition/cyclocondensation reac-
tions of aldehydes 330 with methylhydrazine provide pyra-
zolines 332 (57-87% yields) that on treatment with bases
can be converted to 1,5-disubstituted pyrazoles, e.g., 333
Chemical Reviews, 2010, Vol. 110, No. 3 1585
Scheme 91
Scheme 92
(75% yield). Alternatively, alkylation of lithiated pyrazolines
332 converts them to pyrazolines 331 that eliminate benzo-
triazole under basic conditions to furnish 1,4,5-trisubstituted
pyrazoles 334 (52-79% yields). The process is not limited
to aromatic aldehydes nor the benzotriazol-1-yl derivatives
shown in Scheme 92, since similarly good results are also
reported for R1 ) phenethyl and benzotriazol-2-yl analogues
of compounds 328-332.118
R,β-Unsaturated ketones with a benzotriazolyl substituent
at the R-C can be easily prepared by condensation of
2-benzotriazol-1-ylacetophenone (335) with aldehydes in the
presence of a piperidine catalyst. The Z isomers of R,β-
unsaturated ketones 336 are formed exclusively and are
isolated in 55% (R1 ) 3-pyridyl) to 71% (R1 ) Ph) yields.
Treatment with hydrazines converts 336 into cis-substituted
pyrazolines 337 that are isolated in 51% (R1 ) Ph, R2 )
Me) to 80% (R1 ) R2 ) Ph) yields. Elimination of
benzotriazole from pyrazolines 337 in the presence of bases
furnishes pyrazoles 340 in 81-94% yields. Lithiation of
pyrazolines 337 at C-4 followed by treatment with alkylating
agents provides pyrazolines 339 in 73-95% yields. Subse-
quent elimination of benzotriazole initiated by a base converts
1586 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 93
Scheme 94
pyrazolines 339 into 1,3,4,5-tetrasubstituted pyrazoles 338
in 85-95% yields (Scheme 93).119
For preparation of pyrazoles substituted at C-4 and
unsubstituted at C-3 and C-5, the method shown in Scheme
94 is a good option. Thus, addition of N-(benzotriazol-1-
ylmethyl)morpholine (341) to ethyl propiolate catalyzed by
ZnBr2 provides adduct 342 in 80% yield. Apart from the E
isomer 342, minor benzotriazol-2-yl analogues and Z isomers
are detected by NMR in the reaction mixture. Heating with
phenylhydrazine converts 342 into pyrazole 343 in 75%
yield. This step must involve oxidation by atmospheric
oxygen to convert the intermediate pyrazoline, which forms
first by cyclocondensation of 342 with phenylhydrazine with
elimination of the benzotriazole and morpholine, into pyra-
zole 343 (Scheme 94).120
1,3-Dipolar cycloaddition of 1-(3-morpholin-4-ylallyl)ben-
zotriazole85 (181) to nitrilimine 344, generated in situ from
the corresponding phenylhydrazonyl bromide in refluxing
benzene,121 provides pyrazoline 345 in 70% yield. In the
following treatment with hydrochloric acid in ethanol at room
temperature, pyrazoline 345 eliminates morpholine to furnish
pyrazole 347 (90% yield). 1,3-Bisbenzotriazol-1-ylpropene
(346), prepared from epichlorohydrin and benzotriazole
followed by thionyl chloride and finally sodium hydride,
reacts with nitrilimine 344 to give directly pyrazole 347 in
80% yield; in this case, benzotriazole must be eliminated
from an unstable 5-benzotriazol-1-ylpyrazoline intermediate
(Scheme 95).122
In the special case shown in Scheme 96, two of the
benzotriazolyl nitrogens are used to generate the pyrazole
ring. Thus, treated with trifluoroacetic anhydride, sulfoxides
348 undergo conversion to triazapentalenes 349 in high
yields. The process must involve acylation of the sulfoxide
oxygen atom and generation of a carbocation (Pummerer
Katritzky and Rachwal
Scheme 95
Scheme 96
reaction) that attacks the N-2 atom of benzotriazole. Hydro-
genation over Raney nickel cleaves the C-S bond and one
of the N-N bonds to generate 1-(2-aminophenyl)pyrazoles
350.123
8.2. Imidazole
Condensations of N-monosubstituted ethylenediamines 351
with 2 molar equiv of formaldehyde (as a 37% aqueous
solution) and 1 equiv of benzotriazole provide 1,3-disubsti-
tuted imidazolidines 352 in high yields (85-96%). For R1
) Ph, the benzotriazol-1-yl isomer 352 is the only product;
in other cases, the benzotriazol-2-yl isomers are also present
as the minor components in crude product mixtures. Because
there is no significant difference in their reactivity, both
isomers give the same products on treatment with nucleo-
philes. Thus, in reactions with Grignard reagents, the
benzotriazolyl moiety is replaced by an alkyl, vinyl, phe-
nylethynyl, or aryl group to give imidazolidines 354 in
63-96% yields. Similarly, compound 352 reacts with sodium
cyanide to form nitriles 355 (77-97% yields). Additional
reactions of one of the imidazolidines 352 (R1 ) Ph) with
sodium borohydride, benzenethiol and sodium hydride, or
triethyl phosphite and ZnBr2 demonstrate easy substitution
of the benzotriazolyl moiety by a hydrogen atom (353, 96%
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 97
Scheme 98
yield), a phenylthio group (356, 66% yield), or a phosphonate
group (357, 70% yield) (Scheme 97).124
Condensation of ethylenediamine with 3 molar equiv of
formaldehyde and 2 equiv of benzotriazole provides 1,3-
bis(benzotriazol-1-ylmethyl)imidazolidine (358) together with
its minor benzotriazol-1-yl/2-yl isomer. The reaction of crude
358 with KCN in refluxing acetonitrile gives dinitrile 359
in 88% yield. Treatment with Grignard reagents converts 358
into 1,3-dialkylimidazolidines 360 in 68-75% yields (Scheme
98).63
Commercially available N-Boc-protected R-amino acids
can be converted in two easy steps into amides 361 by
coupling with p-toluidine and deprotection of the amino
group. Reduction of amides 361 with LiAlH4 gives optically
active diamines 362. Condensation of these diamines with
formaldehyde and benzotriazole provides single enantiomers
of 3-(benzotriazol-1-ylmethyl)imidazolidines 363 in 85-93%
yields. Subsequent reactions with nucleophiles allow sub-
stitution of the benzotriazole moiety by various functional
groups, resulting in chiral 1,3,4-trisubstituted imidazolidines
364 (Scheme 99).124
Under forcing conditions, refluxing under a Dean-Stark
trap of benzene solutions, amides 365 undergo condensation
with formaldehyde (generated by depolymerization of
paraformaldehyde in the presence of TsOH) and benzotria-
zole to give 1,3,4-trisubstituted imidazolidin-5-ones 366 that
are isolated as mixtures with their minor benzotriazol-2-yl
isomers (isomeric ratio of 3:1 to 5:1) in 91-94% yields
(Scheme 100).125
Chemical Reviews, 2010, Vol. 110, No. 3 1587
Scheme 99
Scheme 100
Scheme 101
N-Benzotriazol-1-ylmethyl derivatives of BOC-protected
amines behave similarly to amides. Thus, treatment of the
anion derived from compound 367 with N-benzylidene-p-
anisidine leads to imidazolidin-2-one 368 (82% yield) with
the substituents at C-4 and C-5 oriented trans. Subsequent
treatment with nucleophiles gives products 369 stereoselec-
tively in 64-68% yields (Scheme 101).126
The aminocarbene generated by treatment of iminium
chloride 290 with triethylamine undergoes [1 + 2 + 2]
cycloaddition with phenyl isocyanate to give imidazolidine-
2,4-dione 370. Quenching the reaction mixture with tetra-
ethylammonium hydrogen sulfide results in reduction of the
C-5 functionality to give 1,3-diphenylimidazolidine-2,4-dione
(1,3-diphenylhydantoin, 371)in 35% yield. On treatment of
370 with sodium borohydride, only the benzotriazolyl moiety
is replaced by a hydride to give 5-(dimethylamino)hydantoin
372 in 51% yield. Similarly, hydantoin 373 (56% yield) is
generated in a reaction of 370 with ethylmagnesium bromide.
Analogous quenching of the reaction mixture with methanol
or morpholine replaces the benzotriazolyl moiety by a
1588 Chemical Reviews, 2010, Vol. 110, No. 3 Katritzky and Rachwal

Scheme 102 Scheme 103

Scheme 104

methoxy (57% yield) or morpholin-4-yl group, respectively,

without affecting the dimethylamino substituent. However,
when propylamine is used as a nucleophile, both the
benzotriazolyl and the dimethylamino groups are eliminated
to give 5-(N-propylimino) derivative 374 (54% yield). A
simple workup of the mixture containing intermediate 370
with water results in imidazolidine-2,4,5-trione 375 in 67%
Scheme 105
yield (Scheme 102).111
1-(Chloroformyl)benzotriazole (377), prepared in a reac-
tion of benzotriazole with phosgene127,128 or more conve-
niently with triphosgene,129 reacts with amino acids 376 to
give derivatives 378. Treatment with thionyl chloride con-
verts acids 378 into acid chlorides 379 that are subsequently
treated with amines to give amides 380. Under mildly basic
conditions, sodium carbonate in acetone, amides 380 undergo
cyclocondensation with elimination of benzotriazole to
furnish hydantoins 381 in 24% (R1 ) Ph, R2 ) Ph2CH) to
88% (R1 ) Me2CHCH2, R2 ) Ph2CH) yields (Scheme
103).130,131
Lithiation of bisbenzotriazol-1-ylmethane (382) followed
by treatment with tosyl azide gives diazide 383 in 22% yield.
In a reaction with triphenylphosphine, diazide 383 is
converted into bis(triphenylphosphoranylidene) derivative simple synthetic method provides 1,5-diarylimidazoles 387
384. Despite the fact that 384 is not stable enough to provide in 23% (R ) Ph) to 85% (R ) H) yields (Scheme 105).94
an analytical sample, thorough NMR studies of the crude
material reveal that the benzotriazol-1-yl substituents have Displacement of benzotriazole in 58 by primary amines
rearranged to the sterically less hindered -2-yl forms. provides derivatives 388 that are treated in situ with benzil
Consecutive treatments of crude 384 with Grignard reagents to give imidazoles 390 (X ) H) in 81-84% yields. With
followed by benzil lead to 2H-imidazoles 385, although the 4-chlorobenzil, the nucleophilic attack occurs primarily on
yields of isolated products are low (15-21%) (Scheme the carbonyl group attached to the 4-chlorophenyl ring,
104).132 leading to intermediates 389 that undergo cyclization with
The anion derived from isocyanide 227, on treatment with elimination of water to give 4-(4-chlorophenyl)imidazoles
t-BuOK, adds readily to Schiff bases to form imidazolines 390 (X ) Cl). Identification of the substituents in a molecule
386. Under stronger basic conditions, imidazolines 386 of imidazole 390 with R ) PhCH2 and X ) Cl is provided
eliminate benzotriazole to produce imidazoles 387. This by X-ray crystallography (Scheme 106).133
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 106
Scheme 107
8.3. Isoxazole
N,N-Bis(benzotriazol-1-ylmethyl)hydroxylamine (391) is
easily prepared by condensation of 1-(hydroxymethyl)ben-
zotriazole with hydroxylamine.22 At elevated temperatures,
in refluxing toluene, 391 eliminates a molecule of benzot-
riazole to give nitrone 392, which can be readily trapped by
1,3-dipolarophiles. Thus, with methyl acrylate in refluxing
toluene for 16 h, 391 gives isoxazolidine 393a quantitatively
and with acrylonitrile isoxazolidine 393b in 88% yield.
Pyridyl derivatives 393c (83% yield) and 393d (81% yield)
are obtained in the same manner from the corresponding
vinylpyridines. The electron-withdrawing ability of the alkene
substituent R seems to be a determining factor, as no such
reaction of 391 is observed with styrene. Use of 1,2-
disubstituted alkenes, e.g., dimethyl fumarate or N-methyl-
maleimide, allows preparation of 1,3,4-trisubstituted isox-
azolidines in this manner (Scheme 107).134
Removal of the benzotriazol-1-ylmethyl substituent from
the isoxazolidine ring in 393 can be achieved by heating
their ethanolic solutions with perchloric acid. Thus, com-
pound 393d undergoes straightforward hydrolysis to 5-(4-
pyridyl)isoxazolidine (394) in 46% yield. Sodium borohy-
dride reduces the benzotriazol-1-ylmethyl group in 393d to
a methyl group, furnishing isoxazolidine 395 in 68% yield.
A similar reaction of 393c gives the corresponding 1-methyl-
isoxazolidine in 60% yield. However, involvement of the
pyridyl nitrogen atom during hydrolysis of 393c complicates
the process, leading to bridged tricyclic system 397, isolated
in 64% yield after 15 min of reflux. To prove its structure,
pyridinium salt 397 was reduced with sodium borohydride
Chemical Reviews, 2010, Vol. 110, No. 3 1589
Scheme 108
Scheme 109
to 12-oxa-1,3-diazatricyclo[7.2.1.03,8]dodec-5-ene (398). Pro-
longed heating of 393c with perchloric acid results in dimeric
product 396 (85% yield) (Scheme 108).134
Treated with NCS, oximes 399 are readily converted
into benzhydroximoyl chlorides 400.135,136 Exposed to mild
bases, chlorides 400 eliminate HCl to give reactive
benzonitrile oxides 401.121 When 188 is added, oxides 401
are trapped in a 1,3-dipolar cycloaddition reaction to give
isoxazolines 402 in 94-100% yields. A similar reaction with
ethoxy derivative 184 provides isoxazoline 403 in 81% yield.
In all these products, the benzotriazol-1-ylmethyl moiety is
located at C-5 of the isoxazoline ring (Scheme 109).137
However, an opposite orientation of the reacting mol-
ecules is required in a 1,3-dipolar cycloaddition of
1-allylbenzotriazoles bearing an electron-donating sub-
stituent at C-3. Thus, 18185 adds to benzonitrile oxide to
provide isoxazoline 404 in 70% yield. Analogous reactions
of benzonitrile oxide with 185 and 346 give isoxazolines
405 (80% yield) and 406 (65% yield), respectively. For
1590 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 110
Scheme 111
isoxazolines 404-406, NMR spectra indicate a trans orienta-
tion of the substituents at C-4 and C-5. On treatment with
hydrochloric acid in ethanol at reflux, isoxazolines 404-406
eliminate the substituent at C-5 to furnish isoxazole 407 in
approximately 90% yield (Scheme 110).122
In a simplified procedure, 175 is heated with benzal-
dehyde oximes, NCS, and wet KHCO3 to give directly in
one pot 3-aryl-5-(benzotriazol-1-ylmethyl)isoxazoles 408
in 52% (X ) 3-NO2) to 100% (X ) H) yields. In this
conversion, the intermediate benzonitrile oxides, generated
by consecutive reactions of the oximes with NCS and
KHCO3, undergo [3 + 2] cycloadditions with the prop-
argyl triple bond to generate the isoxazole ring. To test
possible pathways for the elimination of the benzotriazole
moiety from oxazoles 408, one of them (X ) 4-MeO) is
benzylated at C-R to give derivative 409, which is
subsequently treated with t-BuOK to furnish 4-styrylisox-
azole 410 in 63% yield (Scheme 111).137
Condensation of 335 with aromatic aldehydes readily
provides chalcones 411 bearing a benzotriazol-1-yl sub-
Katritzky and Rachwal
Scheme 112
Scheme 113
stituent at C-R. In a reaction with hydroxylamine, chal-
cones 411 are converted to 4-benzotriazol-1-ylisoxazolines
412 that spontaneously eliminate benzotriazole to provide
3,5-disubstituted isoxazoles 413 in 55-81% yields (Scheme
112).119
Condensation of 310 with 2 molar-equivalents of
4-nitrobenzaldehyde provides 1,4-pentadien-3-one 414 in
81% yield. In a reaction with hydroxylamine, only the
double bond in derivative 414 without a benzotriazolyl
substituent is affected to give isoxazole 415 in 79% yield.
This process must involve an oxidation step (probably by
the atmospheric oxygen) of an isoxazoline intermediate
(Scheme 113).117
Condensation of 310 with N,N-dimethylformamide
dimethyl acetal in refluxing xylene gives enaminone 416
in 74% yield. With an excess of hydroxylamine hydro-
chloride, enaminone 416 is converted into 5-aminoisox-
azole 420 in 68% yield. The process is assumed to go
through dioxime 417, which loses a molecule of water to
give R-cyanooxime 418. An intramolecular addition of
the OH to the CN group in 418 gives 5-iminoisoxazoline
419, which is converted to the more stable tautomeric form
420 (Scheme 114).115
8.4. Oxazole
Similarly to the reaction shown in Scheme 26, the radicals
generated by treatment of N,N-bis(benzotriazolylmethy-
l)amines 35 with SmI2 are trapped by diethyl ketone, leading
to oxazolidines 421; however, the yields are low (30-38%)
(Scheme 115). In this case, generally better results are
obtained when N,N-bis(tosylmethyl)amines are used instead
of derivatives 35.138
Synthesis of Heterocycles Mediated by Benzotriazole Chemical Reviews, 2010, Vol. 110, No. 3 1591

Scheme 114 Scheme 116

Scheme 115

Scheme 117

Cyclocondensation of (S)-2-phenylglycinol with ethyl

glyoxylate, benzotriazole, and formaldehyde provides ox-
azolidine 421 in 80% yield. In reactions with allyltrimeth-
ylsilane and (2-methylallyl)trimethylsilane catalyzed by
ZnBr2, the benzotriazolyl moiety in oxazolidine 421 is readily
substituted to give corresponding oxazolidines 422 in 66%
and 69% yield, respectively. Similar reactions of 421 with Scheme 118
trimethylsilyl enolates provide ketones 423 in 56% (R2 )
Me) and 70% (R ) Ph) yields. Phenylzinc bromide reacts
with 421 to give 3-benzyloxazolidine 424 in 57% yield;
however, aliphatic organozinc reagents fail to give analogous
products. In a similar manner, 421 reacts with triethyl
phosphite to give phosphate 425 in 78% yield and with
1-methoxy-2-methyl-1-[(trimethylsilyl)oxy]propene to give
ester 426 in 80% yield. In all these reactions, the chirality
at C-2 and C-4 is preserved (Scheme 116).139
1-(N-benzylthiocarbamoyl)benzotriazole (427) is prepared
from benzylamine and bisbenzotriazol-1-ylmethanethione in
97% yield.140 The isothiocyanate anion generated from 427
by 2 molar equiv of potassium tert-butoxide adds to carbonyl
groups of ketones to generate oxazolidinethiones 428, which thionyl benzotriazolide with amides. Lithiated 2-methyl-2-
are isolated in 27% (R ) Ph) to 51% (R ) 2-thienyl) yields oxazoline reacts readily with compounds 431 to give
(Scheme 117).141 enamines 433 in 82-96% yields. Imidoyl intermediates 432
Cyclocondensation of acylbenzotriazoles 167 with 2-amino- are not detected by NMR in the product mixtures due to
2-methyl-1-propanol under microwave irradiation produces their rapid tautomerization to forms 433 that are stabilized
2-oxazolines 429. To prevent hydrolysis leading to side by strong intramolecular hydrogen bonding between their
products, SOCl2 is added to the mixture after initial heating nitrogen atoms (Scheme 119).143
at 80 °C for 10 min, and heating is then continued for an The anion derived from isonitrile 37 (BetMIC) on treat-
additional 2 min. In this way, high yields of oxazolines 429 ment with potassium tert-butoxide readily adds to the
(85-97%) are obtained. Use of (S)-2-amino-3-phenyl-1- carbonyl groups of ketones. The benzotriazolyl moiety in
propanol allows the preparation of chiral (S)-4-benzyl-2- the intermediates is substituted by an ethoxy group from the
phenyl-2-oxazoline (430) in 82% yield (Scheme 118).142 added ethanol to give 4-ethoxyoxazolines 434. The reaction
1-Imidoylbenzotriazoles 431 are prepared in a reaction of provides high yields of the products derived from acyclic
1-acylbenzotriazoles 167 with isocyanates or a reaction of ketones (93% for 434a and 92% for 434b) and moderate
1592 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 119
Scheme 120
Scheme 121
yields for the products derived from cycloalkanones (58-65%)
(Scheme 120).144
Addition of the anion derived from isonitrile 37 to the
carbonyl group of aromatic aldehydes results in oxazolines
435 that spontaneously eliminate benzotriazole to furnish
5-aryloxazoles 436. The reaction does not differentiate
between electron-rich and electron-poor aromatic rings, as
the yields of 436 obtained from benzaldehyde (69%),
4-fluorobenzaldehyde (68%), and 4-methoxybenzaldehyde
(63%) are comparable, but heteroaromatic aldehydes give
significantly lower yields: 35% for 2-furyl and 42% for
3-pyridyl (Scheme 121).144
Condensation of benzotriazole with glyoxal in aqueous
acetic acid readily provides 1,2-bisbenzotriazol-1-yl-1,2-
ethanediol (437) almost quantitatively.33 When 437 is heated
with benzamides and a catalytic amount of Amberlyst 15
ion-exchange resin in refluxing toluene with azeotropic
removal of water, diamides 438 are obtained in almost
quantitative yields. Deprotonation of 438 with NaH gives
anions 439 that in a nucleophilic attack on C-β repel the
benzotriazolide anion to provide oxazolines 440. Under the
reaction conditions, rapid elimination of the second benzo-
triazolyl substituent provides oxazoles 441 in 44-62% yields
(Scheme 122).145
Katritzky and Rachwal
Scheme 122
Scheme 123
8.5. Thiazole
As with the analogous reaction of 2-methyl-2-oxazoline
(Scheme 119), lithiated 2-methyl-2-thiazoline reacts with
imines 431 to give substituted 2-(β-enamino)-2-thiazolines
442 in good yields (79-98%). Due to possible hydrolysis
of the heterocyclic rings, thiazolines 442 and the previ-
ously described corresponding oxazolines 433 are con-
sidered as masked β-enamino carboxylic acids (Scheme
123).143
Addition of 335 to phenyl isothiocyanate provides amino
thiol 443, which is trapped by phenacyl bromide to give
2,3-dihydro-2-(1′-benzotriazol-1-yl-1′-benzoylmethylidene)-
3,4-diphenylthiazole (444) in 85% yield (Scheme 124).115
Benzotriazol-1-ylacetonitrile (445) is easily prepared by
condensation of 30 with NaCN in DMSO.20 Hydrolysis of
nitrile 445 with H2O2/K2CO3/H2O gives amide 446 (93%
yield), which is subsequently converted to thioamide 447
(83% yield) using Lawesson’s reagent. Cyclocondensation
of thioamide 447 with phenacyl bromide provides thiazole
448 in 84% yield. Treatment of lithiated 448 with alkyl
halides allows monoalkylation of the methylene group to give
derivatives 449 in 76-83% yields (Scheme 125).146
As is illustrated by transformation of compound 449a, the
lithiation/alkylation process can be repeated to provide
thiazoles 450 (86-93% yields) with the methylene groups
dialkylated with the same or different substituents. In a
Synthesis of Heterocycles Mediated by Benzotriazole Chemical Reviews, 2010, Vol. 110, No. 3 1593

Scheme 124 Scheme 126

Scheme 125

Scheme 127

Scheme 128

reaction with phenylmagnesium bromide in refluxing toluene,

the benzotriazolyl moiety in thiazoles 450 is substituted by
phenyl to furnish 2-(R,R-dialkylbenzyl)-4-phenylthiazoles
452 in 54-75% yields. Thiazole 453 is obtained in 58% yield
in a similar reaction of 450c with methylmagnesium iodide.
Treatment of 450a with 2-methylfuran and ZnBr2 in refluxing
1,2-dichloroethane substitutes the benzotriazolyl moiety by
furyl to provide thiazole 451 in 65% yield. 2-Methyl-
thiophene reacts similarly with 450c to give derivative 454
in 82% yield (Scheme 126).146
The anion obtained from thiazole 448 can also be trapped
by other electrophiles. Thus, lithiated 448 is treated with
phenyl or tert-butyl isocyanate to provide amides 455 in
87-96% yields. Reduction with zinc/acetic acid in refluxing
ethanol removes the benzotriazolyl group to give (4-
phenylthiazol-2-yl)acetamides 456 in 73-86% yields (Scheme
127).146
Using higher R-bromo ketones in the reaction shown in
Scheme 125 allows substitution of C-5 of the ring. Thus,
condensation of thioamide 447 with 2-bromopropiophenone
with benzotriazole in refluxing toluene or dioxane convert
gives thiazole 457 in 59% yield. Subsequent double lithiation/
methylation furnishes thiazole 458 in 56% yield. The diimines 461 into imines 462 in 80-92% yields. Treatment
benzotriazolyl moiety in 458 can be substituted with a phenyl of 462 with n-butyllithium produces anions 463 that add
or 5-methylfuryl ring to give thiazole 459 (59% yield) or readily to isothiocyanates to give intermediate anions 464.
460 (70% yield), respectively (Scheme 128).146 Loss of a benzotriazolide anion results in iminothiazole
Diimines 461 are prepared as precipitates in 75-90% systems 465 that undergo tautomerization to more stable
yields by stirring aromatic aldehydes with a 10-fold excess aminothiazoles 466. This way, 5-aminothiazoles 466 bearing
of an aqueous or methanolic solution of ammonia. Reactions aryl or heteroaryl substituents at C-2 and C-4 are easily
1594 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 129
Scheme 130
prepared in 35% (R1 ) Ph, R2 ) 2-ClC6H4) to 81% (R1 )
R2 ) Ph) yields (Scheme 129).147
9. Five-Membered Rings with Three or Four
Heteroatoms
9.1. 1,2,3-Triazole
1-(Phenylpropiolyl)benzotriazole (467) is readily prepared
by condensation of phenylpropiolic acid with 1-(methylsul-
fonyl)benzotriazole. Its [3 + 2] cycloaddition with benzyl
azide provides smoothly substituted 4-(benzotriazol-1-ylcar-
bonyl)-1,2,3-triazole 468 in 75% yield. The benzotriazolyl
moiety in 468 is easily substituted with amines to give
4-carbamoyl-1,2,3-triazoles 469 in 54-91% yields (Scheme
130).148
9.2. 1,2,4-Triazole
A reaction of benzotriazole with cyanogen bromide carried
out in ethanol in the presence of NaOH provides bisbenzo-
triazolylmethanimine (470) as a mixture of benzotriazol-1-
yl and -2-yl isomers.149,150 To simplify the picture, only the
Katritzky and Rachwal
Scheme 131
benzotriazol-1-yl isomer is shown in Scheme 131. Treatment
with amines converts methanimine 470 under mild conditions
into carboxyimidamides 471 as sole benzotriazol-1-yl iso-
mers. Acyl derivatives 472 are prepared in 62-94% yields
by treatment of imidates 471 with acyl chlorides and
triethylamine in chloroform, and they are stable enough to
be purified by column chromatography. Treatment with
hydrazines substitutes the benzotriazolyl in 472 to give
intermediates 473 that undergo spontaneous cycloconden-
sation to furnish 3-amino-1,2,4-triazoles 475. Yields of 475
are high for R2 ) H or Me (72-95%). However, for less
nucleophilic hydrazines (R2 ) Ph or PhCH2), elimination
of amine can compete with the elimination of benzotriazole,
leading to intermediates 474, and significant amounts of side
products 476 (13-30%) are also obtained.151
When the reactions depicted in Scheme 131 are carried
out on a solid support, no chromatographic purification of
intermediates or final products is needed. Thus, amines 477
bound to StratoSpheres PL-FMP resin ((4-formyl-3-meth-
oxyphenoxy)methyl) by a standard solid-phase protocol
[NaBH(OAc)3, 1% AcOH/DMF] are treated with 470 to give
carboxyimidamides 478 that are washed with THF and
subsequently treated with acid chlorides and DIPEA to
furnish acylated products 479. After being washed with
dichloromethane and 2-propanol, the resins containing 479
are treated with hydrazines and DBU to give 3-amino-1,2,4-
triazole derivatives 480. Thorough washing after this step
removes all side products including benzotriazolyl derivatives
476. Final deprotection with TFA provides 3-amino-1,2,4-
triazoles 481 in 45-65% overall yields and high purity
(Scheme 132).152
In reactions of imines 470 with hydrazines, both hydrazine
nitrogen atoms are involved to give intermediates 482. With
methylhydrazine, a spontaneous cyclocondensation of 482
leads to 5-amino-1,2,4-triazole derivative 483 in 57% yield.
However, for R ) aryl, a reaction of 482 with the second
molecule of hydrazine takes place first to form intermediate
484 before cyclocondensation to 5-hydrazino-1,2,4-triazole
485. Spontaneous oxidation of hydrazines 485 with atmo-
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 132
Scheme 133
spheric oxygen furnishes orange- to red-colored azo deriva-
tives 486, which are isolated in 78-86% yields (Scheme
133).153
9.3. 1,2,4-Oxadiazole
Amidoximes 487 are prepared in 88-92% yields by
refluxing the corresponding nitriles, hydroxylamine hydro-
chloride, and Et3N in ethanol. Cyclocondensation of ami-
doximes 487 with 1-acylbenzotriazoles 167, carried out in
refluxing ethanol in the presence of Et3N, provides 1,2,4-
oxadiazoles 488 in 73-82% yields (Scheme 134).154
1,2,4-Oxadiazoles are valuable bioisosteres of amide
groups in drug design, and benzotriazole methodology allows
efficient conversion of amino acids into their 1,2,4-oxadiazole
analogues. Thus, 1-acylbenzotriazoles 171 are prepared in
high yields in reactions of the corresponding N-protected
amino acids with benzotriazole and thionyl chloride. Their
Chemical Reviews, 2010, Vol. 110, No. 3 1595
Scheme 134
Scheme 135
Scheme 136
reactions with amidoximes 487 are almost instantaneous,
leading to esters 489 that can be separated in stable crystalline
forms. However, heating to reflux in ethanol for a few
minutes in the presence of Et3N converts 489 derived from
p-toluic acid (R2 ) 4-MeC6H4) into 1,2,4-oxadiazoles 490,
which are isolated in high yields (70-94%) by crystalliza-
tion. Amidoxime 487 derived from phenylacetic acid (R2 )
PhCH2) is less reactive toward acyl derivatives 171 requiring
reflux for 90 min, but the yields of products 490 are also
high (83-89%). For 487 derived from 4-pyridinecarboxylic
acid, better results are obtained when the first step is carried
out under acidic conditions (10% HCl, aqueous) before
cyclocondensation of the obtained intermediate 489 in the
presence of Et3N, providing corresponding 1,2,4-oxadiazoles
490 in 90-93% yields. A similar efficient conversion of
several N-protected dipeptides into the corresponding het-
erocyclic derivatives 490 is also described (Scheme 135).154
9.4. 1,3,4-Oxadiazole
290 can readily introduce one carbon atom to a molecular
sequence, and it is a convenient reagent for [4 + 1]
cyclocondensations. Thus, in its reaction with benzoic
hydrazide, 2-phenyl-1,3,4-oxadiazole (491, R ) Ph) is
obtained in 89% yield. A similar reaction of 290 with the
hydrazide derived from 4-pyridinecarboxylic acid gives 1,3,4-
oxadiazole 490 (R ) 4-pyridyl) in 95% yield (Scheme
136).155
Substitution of one of the benzotriazolyl groups in imine
470 with hydrazides derived from benzoic acids results in
intermediates 492 that spontaneously undergo cycloconden-
sation to 2-amino-5-aryl-1,3,4-oxadiazoles 493 with elimina-
tion of the second benzotriazolyl group. Products 493 are
1596 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 137
Scheme 138
obtained in 82-97% yields. 4-Pyridinecarbohydrazide reacts
similarly to give the corresponding oxadiazole in 97% yield,
but acetyl hydrazide gives 1-acetyl-5-amino-3-benzotriazol-
1-yl-1,2,4-triazole (42% yield) instead (Scheme 137).153
8 is easily prepared in a reaction of benzotriazole with
ethyl bromoacetate.156 Treatment with hydrazine converts 8
into hydrazide 494, which is subsequently acylated with
benzoyl chlorides to give N,N′-diacylhydrazines 495 in
52-95% yields.157 Upon refluxing in phosphorus oxychlo-
ride, compounds 495 are converted into disubstituted 1,3,4-
oxadiazoles 496 in 23-75% yields. Electrophilic substitution
of the methylene carbon in 496 can be achieved after
lithiation, as is illustrated by preparation of products 497
(70% yield), 498 (85% yield), and 499 (65% yield); however,
removal of the benzotriazolyl group is difficult due to low
stability of the 1,3,4-oxadiazole ring under the required
conditions (Scheme 138).158
Katritzky and Rachwal
Scheme 139
Scheme 140
9.5. Tetrazole
In an Ugi-type reaction, 37 reacts with 1-pyrrolidin-1-yl-
1-cyclohexene and trimethylsilyl azide in methanol to give
tetrazole derivative 500 in 68% yield. The benzotriazol-1-
ylmethyl group can be easily removed under mild acidic
hydrolysis to provide tetrazole 501 in 98% yield. Examples
of a few other tetrazoles bearing R-aminoalkyl substituents
at their C atom obtained by this method are also reported
(Scheme 139).159
1-Imidoylbenzotriazoles 502 are readily prepared in reac-
tions of the corresponding amides with benzotriazole and
oxalyl or thionyl chloride in yields of 62% (R1 ) R2 )
PhCH2) to 90% (R1 ) PhCH2, R2 ) Ph). They are stable in
water and do not react with sodium azide under neutral
conditions; however, TFA promotes their addition to azide
anion, generating tetrazoles 503, after elimination of ben-
zotriazole. Phase-transfer conditions using tetrabutylammo-
nium bromide (TBAB) accelerate the reaction, allowing its
completion at 20 °C in 30 min to give a variety of
1,5-dusubstituted tetrazoles 503 in high yields (90-95%)
(Scheme 140).160
1-Chlorobenzotriazole (505), a mild transfer reagent of a
“positive” halogen, reacts with 1,3,5-triarylformazans 504
to give 2,3,5-triaryl-2H-tetrazolium chlorides 507 in 70-95%
yields (Scheme 141). The reaction is believed to proceed
through N-chloroformazans 506. Electron-withdrawing sub-
stituents R1 strongly facilitate formation of the product, while
the substituents R2 and R3 have only minor effects on the
reaction rate.161
10. Pyridine
10.1. Hexahydropyridine (Piperidine)
10.1.1. N-Substituent
N-Derivatization of piperidine can be easily achieved by
initial condensation with an aldehyde and benzotriazole and
substitution of the benzotriazole moiety by an alkyl or aryl
group from organometallic reagents in the subsequent step.
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 141
Scheme 142
Scheme 143
In this way more complex substituents may be introduced
that are not readily available by other means. Thus, in an
example given in Scheme 142, piperidine reacts with
benzaldehyde and benzotriazole to give derivative 508, which
is subsequently treated with isopropylmagnesium bromide
to give N-(R-isopropylbenzyl)piperidine (509) in 59% overall
yield.22
With functional groups sensitive to Grignard reagents
present in the molecule, milder acting reagents have to be
used. One such option is shown in Scheme 143. Thus, the
ester group in pipecolinic acid derivative 510 is not affected
during its reactions with organobismuth reagents, producing
derivatives 511 in 33-59% yields.71
10.1.2. Ring Formation
Glutaraldehyde is a common precursor of piperidines. Its
condensation with amines and benzotriazole proceeds con-
veniently in water, giving 2,6-bisbenzotriazolylpiperidines
512 in high yields. In fact, there are six molecular entities
of products 512 present in the reaction mixtures due to cis/
trans and benzotriazol-1-yl/2-yl isomerization, giving rise
to complex NMR spectra. However, their reduction with
Chemical Reviews, 2010, Vol. 110, No. 3 1597
Scheme 144
Scheme 145
sodium borohydride removes the benzotriazolyl substituents
and eliminates the chiral centers to give simple N-substituted
piperidines 513 in 55-80% yields. Treatment of 512 with
Grignard reagents allows substitution of the benzotriazolyl
by alkyl groups. In Scheme 144, benzyl derivative 512 is
converted into 1-benzyl-2,6-dialkylpiperidines 514 that are
isolated in 45-48% yields. According to NMR data, isolated
piperidines 514 have cis orientation of the substituents.162
Similarly to amines, hydrazines and hydrazides react with
glutaraldehyde and benzotriazole to give derivatives 515 as
mixtures of cis/trans and benzotriazol-1-yl/2-yl isomers.
Treatment with sodium borohydride converts mixtures of 515
into 1-amino- or 1-amidopiperidines 516 that are isolated in
79-90% yields. Grignard reactions carried out using two
hydrazide derivatives 515 (R1R2N ) PhCONH and
t-BuO2CNH) give hydrazides 517 as single isomers in
49-72% yields (Scheme 145).162
To prepare piperidines monosubstituted at C-2, a reaction
of N-protected acetal 518 with benzotriazole in the presence
of 10 mol % p-toluenesulfonic acid (p-TSA) gives piperidine
derivative 519 in 88% yield. Treatment with allyltrimethyl-
silane and boron trifluoride etherate allows substitution of
the benzotriazolyl moiety in 519 with an allyl group to
provide N-protected 2-allylpiperidine 520 in 93% yield
(Scheme 146).163
The reaction can be facilitated by using a solid support.
In such an approach, outlined in Scheme 147, suitable amino
acetals are attached to sulfonylethoxycarbonyl-modified
polystyrene. The obtained resin-bound acetals 521 are treated
with a solution of benzotriazole and catalytic amounts of
p-TSA to give derivatives 522 that are subsequently subjected
to a reaction with allyltrimethylsilane. Finally, the products
are cleaved from the resins using sodium methoxide to
furnish 2-allylpiperidines 523 in 71-86% yields. In these
1598 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 146
Scheme 147
Scheme 148
reactions, when R * H, only trans-2,4-disubstituted pip-
eridines are isolated. High diastereoselectivity of this process
is explained by a nucleophilic attack at the less sterically
hindered side of the intermediate iminium ion generated from
522 (Scheme 147).163
Pure enantiomers of C-2-substituted piperidines can be
obtained using chiral auxiliaries in the synthetic process. In
one such approach outlined in Scheme 148, chiral amine 524,
as a salt with tartaric acid, is used as a template for
condensation with glutaraldehyde, giving a pure diastereomer
Katritzky and Rachwal
Scheme 149
of product 525 with R,R configuration of its chiral centers,
as is proved by X-ray crystal structure analysis. Treatment
with sodium cyanoborohydride cleaves selectively the bond
with benzotriazole, and the obtained derivative 526 is
subsequently treated with Grignard reagents to give pure
diastereomers of piperidines 527 in 81-92% yields. Final
deprotection of the piperidine nitrogen by hydrogenolysis
releases chiral piperidines 528 that are isolated in high yields
(93-97%) (Scheme 148).164
In another approach (Scheme 149), condensation of (S)-
2-phenylglycinol with glutaraldehyde (aqueous solution) and
benzotriazole is carried out in methanol at room temperature
to provide bicyclic system 529 in 95% yield as a mixture of
diastereomers and benzotriazol-1-yl/2-yl isomers. Substitu-
tion of the benzotriazolyl moiety in 529 with nucleophiles
(Grignard reagents and lithium diethyl phosphite) provides
derivatives 530 in 57% (Nu ) n-Pr) to 85% (Nu ) Ph)
yields. The reaction seems to be less stereoselective than
that depicted in Scheme 147, as a minor 5-R enantiomer
analogous to derivative 530a (yield 8%) is also isolated from
the reaction mixture. Hydrogenation over a palladium catalyst
cleaves off the asymmetric auxiliary, providing 2-substituted
piperidines 531a (95% yield), 531b (92% yield), and 531c
(68% yield).165 A similar synthesis starting from (R)-2-
phenylglycinol leads to the opposite enantiomers of 530 and
531.166
In the preceding schemes, the piperidine ring is assembled
by making C-N bonds. However, C-C bond formation can
also be utilized; thus, N,N-bis(benzotriazolylmethyl)amines
35 derived from benzyl or phenethylamines undergo cyclo-
condensation with allylsilanes catalyzed by SnCl4 to give
4-chloropiperidines 533 in 58-68% yields. This [3 + 3]
cyclocondensation is assumed to proceed in two steps via
intermediate 532 (Scheme 150).167
Another example of C-C bond formation in construction
of the piperidine ring is depicted in Scheme 151. In this case,
reduction of benzotriazolyl derivatives of 4-penten-1-ylamine
534 with SmI2 generates radicals 535 that are subsequently
rapidly trapped by the alkenyl group and converted to radicals
536. The following reaction with excess SmI2 gives sa-
marium intermediates 537. During aqueous workup, deriva-
tives 537 are hydrolyzed to give 3-methylpiperidines 538
(E ) H) in 50-56% yields. Alternatively, treatment with
electrophiles converts 537 into piperidines 538 with various
substituents at C-3, which are isolated in 30-50% yields.
In general, slight predominance of the cis isomers of
piperidines 538 is observed (Scheme 151).168
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 150
Scheme 151
Scheme 152
10.2. Tetrahydropyridine
Heating of the THF solutions of N-(benzotriazol-1-
ylmethyl)amines 32 (also containing benzotriazol-2-yl iso-
mers), lithium tetrafluoroborate, and dienes produces crys-
talline tetrafluoroborates of 1,2,5,6-tetrahydropyridines 540
that are isolated in 70-90% yields. In reactions with
isoprene, 3-substituted tetrahydropyridines 540a and 540c
are formed regioselectively. To facilitate separation of
crystalline products, 540b and 540d are converted into the
corresponding tetraphenylborates by addition of sodium
tetraphenylborate to the reaction mixtures. The reaction
mechanism must involve generation of iminium tetrafluo-
roborates 539 that subsequently undergo hetero-Diels-Alder
reactions with dienes to give 1,2,5,6-tetrahydropyridinium
salts 540 (Scheme 152).169
Chemical Reviews, 2010, Vol. 110, No. 3 1599
Scheme 153
Scheme 154
Scheme 155
Alkylation of lithiated benzotriazol-1-ylmethyl phenyl
sulfide (541) with allyl bromide gives derivative 542 in 75%
yield, which is subsequently converted into 1-butadien-1-
ylbenzotriazole (543) by treatment with potassium tert-
butoxide. Hetero-Diels-Alder reaction of butadiene 543 with
Eschenmoser’s salt provides N,N-dimethyl-2-benzotriazol-
1-yl-1,2,5,6-tetrahydropyridinium iodide (544) in 77% yield
(Scheme 153).170
10.3. Transformations of Ring Substituents
The rich chemistry of benzotriazole derivatives171 can be
effectively applied to modify functional groups already
attached to the pyridine ring. In an example of such reactions
shown in Scheme 154, the anion generated from thioimidate
545 upon its treatment with NaH adds to the double bond
of 4-vinylpyridine. Obtained pyrrolidine intermediate 546
expels benzotriazole and a thiomethoxide anion to give
pyrrole 547, which is separated in 64% yield. 2-Vinylpyridine
reacts similarly.91
In another advantageous application of benzotriazole
reagents for modification of pyridine substituents (Scheme
155), 2-picoline is lithiated by LDA and treated with
1-acylbenzotriazoles 167 to give pyridin-2-ylmethyl ketones
548 in good yields (60-84%). In comparison with other
synthetic methods for such compounds, this approach simpli-
fies the procedure and provides generally higher yields of
the products. 4-Picoline gives similar products.172
1600 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 156
10.4. Electrophilic and Nucleophilic Substitutions
of the Aromatic Ring
Condensation of 29 with 2-amino-(3, 4, or 6)-methylpy-
ridines 549 in refluxing acetic acid in the presence of catalytic
p-TSA provides corresponding 2-amino-5-(benzotriazol-1-
ylmethyl)pyridines 550 in 60-73% yields. The parent
2-aminopyridine (R1 ) R2 ) R3 ) H) and 4,6-dimethyl-2-
aminopyridine (R1 ) H, R2 ) R3 ) Me) give similar
products in 53% and 61% yields, respectively. Formation
of compounds 550 may involve a direct electrophilic attack
of a cation generated from 29 in the presence of an acid on
C-5 of the pyridine ring or an attack of such a cation on the
amino group followed by Hofmann-Martius rearrangement;
possibly this dual pathway mechanism is operating at the
same time.
In a reaction with n-butylmagnesium bromide carried out
in refluxing THF, the benzotriazolyl moiety in 550 is
smoothly substituted with an n-butyl group to give product
551a in 77% yield. A similar reaction with phenylmagnesium
bromide provides compound 551b in 49% yield. Thioether
derivatives 551c (94% yield) and 551d (48% yield) are
obtained by refluxing solutions of benzotriazolyl derivatives
550 and sodium salts of the corresponding mercaptans in
2-propanol. Ether 551e is obtained in 70% yield from a
reaction of 550e with sodium n-butoxide in refluxing
1-butanol (Scheme 156).173
Due to the relative acidity of the CH2 group attached to
the benzotriazol-1-yl ring, the substituent at C-5 in derivatives
550 can be further modified by lithiation followed by
reactions with electrophiles. In the examples of such reactions
shown in Scheme 157, treatment of 550a with 2 molar equiv
of n-BuLi is followed by bromoethane or benzophenone to
give derivatives 552 and 554 in 60% and 72% yields,
respectively. Substitution of the benzotriazolyl moiety in 552
with a butyl group from n-BuMgBr provides 2-amino-3-
methyl-5-(1-ethylpentyl)pyridine (553) in 65% yield. A
similar reaction of 554 with n-BuMgBr gives alcohol 555
in 70% yield. Use of 3 equiv of n-BuLi followed by 2 equiv
of ethyl bromide allows substitution of both methylene
protons with ethyl groups, resulting in derivative 556, which
is separated in 85% yield. This process can also be run
stepwise using two different electrophiles. Treatment of 556
and similar derivatives with nucleophiles allows substitution
of the benzotriazolyl moiety, forming a complex group at
C-5. In another approach, 556 is treated with NaH in
refluxing THF to eliminate benzotriazole and generate an
alkenyl group in product 557 (58% yield).173
At 155 °C for 30 min, reaction of benzotriazole with
2-chloro-6-methylpyridine (558, R ) Me, X ) Cl) gives
Katritzky and Rachwal
Scheme 157
Scheme 158
benzotriazolyl derivative 559b in 70% yield. The reaction
is accelerated under microwave irradiation, and the yields
are also slightly better: 87% for 559a and 72% for 559b.
2-Bromopyridines 558 (X ) Br) do not perform as well,
giving 559a and 559b in 30% and 47% yields, respectively
(Scheme 158).174
10.5. Aromatic Ring Formation
10.5.1. Pyridones and Thiopyridones
Michael additions of benzotriazol-1-ylmethyl ketones 560
to ethyl acrylate under phase-transfer-catalysis conditions
give the corresponding esters that are in situ hydrolyzed to
δ-ketoacids 561, isolated in 70% (R1 ) Ph) and 85% (R1 )
Me) yields. In refluxing toluene under a Dean-Stark trap,
anilines and benzylamine convert ketones 561 into 3,4-
dihydro-2-pyridones 562 in 56-88% yields. Dehydrogena-
tion of compound 562a by heating with 10% Pd/C at 220
°C results in partial nitrogen loss from the benzotriazole ring,
giving rise to a mixture of 2-pyridones 563a (70%) and 564a
(20%). For 562b, the degree of decomposition is even higher
with yields of 50% for 563b and 40% for 564b, whereas, in
the case 562c, pyridone 564c is the only product, isolated in
85% yield. Interestingly, 3,4-dihydro-2-pyridone 562e is
cleanly dehydrogenated to pyridone 563e (85% yield) without
indication of nitrogen loss from the benzotriazole ring
(Scheme 159).175
The benzotriazol-2-yl analogue of ketone 335, compound
565, adds readily to ethyl acrylate to give ester 566 in 85%
yield. Basic hydrolysis of the ester group in 566 causes
elimination of the benzoyl group, but acidic hydrolysis gives
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 159
Scheme 160
the desired δ-ketocarboxylic acid 567 in 75% yield. Con-
densations of acid 567 with aniline and p-toluidine provide
3,4-dihydro-2-pyridones 568 in 75-78% yields. Contrary to
that of their benzotriazol-1-yl analogues, dehydrogenation
of 568 proceeds smoothly without nitrogen loss to give
2-pyridones 569 in 82% (R ) H) and 75% (R ) Me) yields
(Scheme 160).175
Michael addition of 445 to chalcone, catalyzed by NaOH,
gives intermediate 570. In the following step, nucleophilic
attack of the hydroxy anion on the nitrile promotes cycliza-
tion to tetrahydropyridine 571, which spontaneously elimi-
nates benzotriazole and water to furnish 4,6-diphenyl-2-
pyridone (572) in 47% yield. The side product of this
reaction, 2-ethoxy-4,6-diphenylpyridine (15% yield), results
from involvement of an ethoxy anion in a nucleophilic attack
on intermediate 570. The method does not seem to be
general, as an analogous reaction of nitrile 445 with
4-phenylbut-3-en-2-one fails to give the expected 2-pyridone
(Scheme 161).176
Chemical Reviews, 2010, Vol. 110, No. 3 1601
Scheme 161
Scheme 162
A better synthesis of 4,6-diaryl-2-pyridones involves
Michael addition of readily available benzotriazol-1-ylac-
etamide (446) to chalcones. In this case, formation of the
pyridone backbone occurs immediately after the addition,
which stabilizes the system and prevents side reactions. The
just obtained tetrahydropyridines 573 eliminate benzotriazole
and water to give pyridones 574 in 53-92% yields. Use of
2-benzotriazol-1-ylpropionamide instead of acetamide 446
introduces a methyl group to C-3 of pyridones 574 (Scheme
162).176
Michael addition of acetophenone 335 to benzylidenema-
lononitrile gives intermediate 575, which subsequently
undergoes cyclization to 3,4-dihydro-2-pyridone 576. The
mechanism of this cyclization may be similar to that proposed
for intermediate 570, and/or it may involve other modes;
formally, it is hydrolysis of one of the cyano groups followed
by condensation of the obtained amide with the benzoyl
carbonyl. Unlike the similar system 571, the benzotriazolyl
substituent in a molecule of 576 is not located in a position
suitable for elimination. However, under the reaction condi-
tions, 576 is rapidly oxidized to furnish tetrasubstituted
2-pyridone 577 in 62% yield (Scheme 163).177
Condensation of 310 with aromatic aldehydes gives 4-aryl-
3-benzotriazol-1-yl-3-buten-2-ones in 69% (578a), 81%
(578b), and 71% (578c) yields. Malononitrile in the presence
of catalytic amounts of piperidine converts compounds 578
into 2-pyridones 579 in 75%, 85%, and 92% yields, for a, b
and c, respectively. Again, the reaction is expected to proceed
through intermediates analogous to structures 575 and 576
and involves an oxidation step (Scheme 164).117
1602 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 163
Scheme 164
Enaminone 324 is readily prepared by heating 310 with
N,N-dimethylformamide dimethyl acetal under standard115a
or microwave conditions.115b The outcome of the reaction
of 324 with ethyl cyanoacetate depends strongly on the
reaction conditions. Thus, in the presence of sodium hydride,
the carbonyl group reacts first to give intermediate 580A.
During workup, hydrolysis of the cyano group generates an
amido group that subsequently undergoes addition to the
enamine double bond to form a heterocyclic ring. In the final
stage, elimination of dimethylamine leads to 2-pyridone
580C in 64% yield (Scheme 165).115b An analogous cy-
cloaddition/elimination process converts 2-cyano-5-aryl-5-
(dimethylamino)-2,4-pentadienamides into the corresponding
6-aryl-3-cyano-2-pyridones in 92-95% yields.177a
However, in the presence of acetic acid and ammonium
acetate, the process starts from Michael addition of ethyl
cyanoacetate to the double bond of 324. The following
condensation of the carbonyl group with ammonia generates
imine 580B. In the final stage, addition of the imine to the
cyano group and elimination of dimethylamine results in the
corresponding 2-iminopyridine that tautomerizes to the more
stable form, 2-aminopyridine 580D, isolated in 53% yield
(Scheme 165).115b
The synthesis of 2-thiopyridone 580E (64% yield) shown
in Scheme 165 is similar. Thus, cyclocondensation of 324
with cyanothioacetamide catalyzed by triethylamine consists
of three well-known steps: Michael addition, condensation
Katritzky and Rachwal
Scheme 165
between the thioamido and carbonyl groups, and elimination
of dimethylamine.177b
10.5.2. Pyridines
1-(R-Chloroalkyl)benzotriazoles 581 are easily prepared
in 80-96% yields by stirring solutions of aldehydes with
benzotriazole and SOCl2 in benzene at room temperature.
Substitution of the chloride with azide anions in 581 proceeds
smoothly in DMSO to provide stable azides 582 in 77-97%
yields. Triphenylphosphine converts azides 582 into 1-[R-
(phosphoranylideneamino)alkyl]benzotriazoles 583, which
upon treatment with NaH eliminate benzotriazole to give (N-
vinylimino)phosphoranes 584 as mixtures of E and Z isomers
in 84-94% yields. Prolonged heating of phosphoranes 584
with chalcone in refluxing toluene results in 5-alkyl-2,4-
diphenylpyridines 586 that are isolated in 59-84% yields.178
Two possible routes to intermediate 3,4-dihydropyridines 585
start from (i) Michael-type addition of 584 to the C-3 atom
of chalcone and (ii) Wittig reaction of 584 with the carbonyl
group of chalcone (Scheme 166).
Scheme 167 illustrates application of benzotriazolyl de-
rivatives to the general synthetic method of pyridines based
on 1,5-diketones. The procedure involves prolonged heating
of chalcones with benzotriazol-1-ylacetophenones 587 and
ammonium acetate in acetic acid. The reaction pathway is
believed to start from Michael addition of ketone 587 to
chalcone, giving diketone 588, which subsequently undergoes
condensation with ammonia and elimination of benzotriazole.
The yields of 2,4,6-triarylpyridines 589 obtained by this route
vary from 62% (R1 ) Ph, R2 ) 4-(O2N)C6H4, R3 )
2-napththyl) to 87% (R1 ) 4-MeC6H4, R2 ) 3,4-
(OCH2O)C6H3, R3 ) 2-napththyl).179
The procedure of Scheme 161 can be adapted to synthesize
2-aminopyridines. Thus, use of secondary amines as bases
instead of NaOH provides intermediates 590 that cyclize to
tetrahydropyridines 591. Elimination of water and benzot-
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 166
Scheme 167
Scheme 168
riazole from 591 furnishes 2-aminopyridines 592. As weaker
bases, amines require prolonged heating (48 h), whereas the
reaction with sodium hydroxide proceeds well at room
temperature. The yields of products 592 vary from 35% (R1
) Me, R2R3N ) piperidin-1-yl) to 74% (R1 ) Ph, R2R3N )
pyrrolidin-1-yl) (Scheme 168).176
Nitriles 593, prepared in 65-78% yields by condensation
of β-aminocrotononitrile with ketones in the presence of
Chemical Reviews, 2010, Vol. 110, No. 3 1603
Scheme 169
Scheme 170
TiCl4, react with iminium salt 290 to give pyridines 595 in
57% (R1 ) i-Pr, R2 ) H) to 75% (R1 ) Et, R2 ) Me) yields.
Nitriles 593 derived from cycloalkanones (R1 and R2 are parts
of a ring) react similarly, giving rise to bicyclic systems 595.
The reaction pathway seems to lead through intermediates
594, as such structures can be identified in the reaction
mixtures before their treatment with a base (Scheme 169).180
When the reactions shown in Scheme 164 are run in the
presence of sodium ethoxide as a base, 2-ethoxypyridines
596 are formed instead of 2-pyridones. The proposed
mechanism involves addition of malononitrile to the activated
double bond of ketone 578 followed by addition of an
ethoxide anion to one of the cyano groups to give an imino
ether that undergoes cyclization by a nucleophilic attack of
the imine nitrogen atom on the carbonyl group. The yields
of products 596a and 596b are 76% and 78%, respectively
(Scheme 170).117
11. Pyran and Thiopyran
11.1. Pyran
The benzotriazolide anion (Bt-) attacks electron-
deficient aromatic rings of pyrylium salts 597 in the
position para to the oxygen atom to give 4-benzotriazol-
1-yl-2,6-diaryl-4H-pyrans 598 in 70-95% yields. Crude
benzotriazol-1-yl derivatives 598 are contaminated with
about 10% of their benzotriazol-2-yl isomers, but the
mixtures can be successfully used in further transforma-
tions without separation. The ortho positions in salts 597
are blocked by aryl groups to avoid reactions there.
Derivatives 598 are useful in the synthesis of 4-alkylpyry-
lium salts 600. Thus, treatment of 598 with n-BuLi at -78
°C followed by alkyl or benzyl halides RCH2Y and
quenching of the reaction mixtures with a solution of
ammonium chloride furnishes the corresponding 4-alkyl-
2,6-diarylpyrylium salts, which are isolated as perchlorates
1604 Chemical Reviews, 2010, Vol. 110, No. 3 Katritzky and Rachwal

Scheme 171 Scheme 172

Scheme 173

600 in 28-81% yields. The reactions proceed possibly

via 4-alkyl-4-benzotriazol-1-yl-2,6-diaryl-4H-pyran inter-
mediates 599 that eliminate benzotriazole under acidic
conditions to give pyrylium salts 600 (Scheme 171).181
Addition of lithiated 1-acylbenzotriazoles 601 to R,β-
unsaturated ketones results in intermediate anions 602 that
by a nucleophilic attack of the enolate oxygen on the
carbonyl group and consecutive elimination of a benzotria-
zolide anion provide 3,4-dihydro-2-pyranones 603 and 604.
Due to steric requirements during the initial addition stage,
trans isomers 603 are strongly predominant, with the trans:
cis ratio varying from 4:1 (R1 ) Et, R2 ) R3 ) Ph) to 100:0
(R1 ) n-C6H11, R2 ) R3 ) Ph). The yields are generally
good (53-81%) except in the case of an isopropyl substituent
(R1 ) Me2CH, R2 ) R3 ) Ph), where the yield is only 24%. Scheme 174
Introduction of an additional substituent to C-R in starting
materials 601 inhibits formation of 2-pyranones 603 and 604.
These and other observations lead to the conclusion that the
reaction is kinetically controlled, and the unstable anions
derived from compounds 601 add to R,β-unsaturated ketones
only under preferable steric conditions; otherwise, they
undergo spontaneous disintegration (Scheme 172).182
Condensation of 310 with hippuric acid (605) and N,N-
dimethylformamide dimethyl acetal (DMFDMA) in refluxing
acetic anhydride gives 2-pyranone 609 in 90% yield. The
proposed reaction mechanism involves cyclization of hippuric
acid to oxazoline 606, which subsequently condenses with
DMFDMA to give enamine derivative 607. The consecutive
condensation with 310 gives intermediate 608, which finally
rearranges to 2-pyranone 609 (Scheme 173).115 form 2-alkyl-4,5-dimethyl-2-(trimethylsilyl)-3,6-dihydro-2H-
thiopyrans 612, isolated in 40-64% yields (Scheme 174).183
11.2. Thiopyran In another approach, benzotriazol-1-ylmethyl phenyl ether
Treatment of lithiated monosubstituted R-benzotriazol-1- (613) is lithiated and treated with chlorotrimethylsilane to
ylalkyl thioethers with chlorotrimethylsilane produces R-(tri- give R-trimethylsilyl ether 614. Reaction of 614 with
methylsilyl)alkyl thioethers 610. In reactions with hexame- hexamethyldisilathiane in methanolic HCl generates thio-
thyldisilathiane and cobalt dichloride hexahydrate, thioethers formylsilane 615, which is trapped by 2,3-dimethylbutadiene
610 are converted to thioacylsilanes 611 that can be trapped to give 4,5-dimethyl-2-(trimethylsilyl)-3,6-dihydro-2H-thi-
in a Diels-Alder reaction with 2,3-dimethylbutadiene to opyran (616) in 42% yield (Scheme 175).183
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 175
Scheme 176
12. Six-Membered Rings with Two or More
Heteroatoms
12.1. Pyridazine
Condensation of succinaldehyde with arylhydrazines and
benzotriazole gives 1-aminopyrrolidines 617, isolated in
70-90% yields as complex mixtures of cis/trans and
benzotriazol-1-yl/2-yl isomers, used in further reactions
without separation. Treatment of 617 with Grignard reagents
in refluxing THF results in formation of 1,4,5,6-tetrahydro-
pyridazines 621. For the best results, 5 molar equiv of
organomagnesium reagents are used. The reaction is believed
to proceed via anion 618, which by an intramolecular
nucleophilic attack on C-2 of the pyrrolidine ring and
elimination of a benzotriazolide anion generates aziridine
system 619. Ring enlargement promoted by pulling out
electrons by the remaining benzotriazolyl moiety results in
iminium cation 620, which is converted by an organomag-
nesium reagent into pyridazine 621. Coordination of the
benzotriazolyl N-3 atoms by additional organomagnesium
molecules enhances the leaving ability of the benzotriazolyl
groups. Yields of products 621 vary from 55% to 89% and
are generally higher for phenyl and 4-methylphenyl as R1
than for their 4-chlorophenyl analogues (Scheme 176).184
Hydrazones 622A are smoothly prepared in 76-78%
yields by coupling of benzotriazol-1-ylmethyl methyl ketone
(310) with appropriate aryldiazonium salts in ethanolic
Chemical Reviews, 2010, Vol. 110, No. 3 1605
Scheme 177
NaOH. Prolonged heating of 622A with N,N-dimethylfor-
mamide dimethyl acetal results in formation of 1-aryl-3-
benzotriazol-1-ylpyridazin-4-ones 625A in 79-83% yields.
The reaction pathway involves formation of enaminones
623A that by intramolecular cycloaddition generate 6-(dim-
ethylamino)-1,4,5,6-tetrahydropyridazin-4-ones 624A, which
eliminate dimethylamine to form final products 625A
(Scheme 177A).185a
Cyclocondensation of hydrazone 622B with ethyl cy-
anoacetate under microwave conditions in the presence of
ammonium acetate and acetic acid produces 3-benzotriazol-
1-yl-5-cyano-4-methyl-1-(4-methylphenyl)pyridazin-6-one
(623B) in 91% yield (Scheme 177B).185b
Cyclocondensation of 8 with 2-(phenylhydrazono)-3-oxo-
3-arylpropanals 626 in ethanolic KOH at room temperature
furnishes 3-aroyl-5-benzotriazol-1-yl-1,6-dihydro-1-phe-
nylpyridazin-6-ones 627 in 89-90% yields. 6-Imino ana-
logues of products 627 are obtained in 87-88% yields by
condensation of hydrazones 626 with 445 in THF in the
presence of sodium hydride (Scheme 178).185
12.2. Pyrimidine
Condensation of 1,3-propanediamine with 3 molar equiv
of formaldehyde and 2 molar equiv of benzotriazole provides
1,3-bisbenzotriazol-1-ylhexahydropyrimidine 628 in a mix-
ture with its benzotriazol-2-yl isomer in a ratio of 9:3.
Because there is no difference here in reactivity with
nucleophiles between both benzotriazolyl isomers, crude
1606 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 178
Scheme 179
Scheme 180
product 628 is treated with organomagnesium reagents to
give symmetrically substituted hexahydropyrimidines 629a
(77% yield) and 629b (87% yield). In a reaction with KCN,
both benzotriazolyl groups are replaced to give dinitrile 630
in 87% yield (Scheme 179).63
Use of monosubstituted 1,3-propanediamines as the start-
ing materials in this synthesis allows preparation of unsym-
metrical hexahydropyrimidines. Thus, as is illustrated by the
example in Scheme 180, N-propyl-1,3-propanediamine (631)
reacts with formaldehyde (37% aqueous solution) and
benzotriazole in methanol/water to give hexahydropyrimidine
632 in 94% yield, containing about 17% of the corresponding
benzotriazol-2-yl isomer. Grignard reagents convert 632 (a
mixture of isomers) into hexahydropyrimidines 633 in
90-92% yields.186
Condensation of benzotriazole with aldehydes and urea
or N-methylurea provides derivatives 634 in 80-90% yields.
Treatment of such derivatized ureas with β-ketoesters in the
presence of zinc bromide leads to alkyl 2-oxo-1,2,3,4-
tetrahydropyrimidine-5-carboxylates 635, isolated in 77-97%
yields. Use of thiourea instead of urea in these conversions
gives the corresponding 2-thioxo analogues of 635 (Scheme
181).187
Reaction of bis(triphenylphosphoranylidene) derivative 384
with phenylmagnesium bromide and acetylacetone provides
4,6-dimethyl-2,2-diphenyl-1,2-dihydropyrimidine (636) in
14% yield (Scheme 182).132
Katritzky and Rachwal
Scheme 181
Scheme 182
Scheme 183
2-Benzotriazol-1-ylvinamidinium salts 239 are convenient
for the preparation of pyrimidines with a benzotriazolyl
substituent at C-5. In the reactions with amidines and sodium
ethoxide carried out in refluxing ethanol, salts 239 are
converted into pyrimidines 637 in 60% (R ) H) to 95% (R
) Ph) yields (Scheme 183).99
Condensation of 2-nitro-1,1,3,4,4-tetrachlorobutadiene with
benzotriazole allows replacement of two geminal chlorine
atoms with benzotriazolyl substituents to give bisbenzotria-
zol-1-yl derivative 638. Under mild conditions, 5 °C f room
temperature, in MeOH, one of the benzotriazolyl groups in
638 is readily substituted with amines to give enamines 639
in 57-98% yields. Treatment of 639 in DMSO with amidines
(as HCl salts) and NaH results in formation of tetrasubstituted
pyrimidines 643 that are isolated in 40-85% yields. The
reaction pathway is believed to lead through intermediates
640-642 (Scheme 184).188
12.3. Pyrazine
Condensation of ethyl esters of N-monosubstituted glycines
644 with formaldehyde and benzotriazole carried out in water
at room temperature provides benzotriazolyl derivatives 645
in 90-98% yields as mixtures with their benzotriazol-2-yl
isomers. Upon treatment with sodium hydride in refluxing
THF, compounds 645 are converted to esters of trans-
piperazine-2,3-dicarboxylic acids 646, isolated in 20% (R
) Me) to 80% (R ) PhCH2) yields. Molecular structures of
646 are unequivocally assigned on the basis of their NMR
spectra and X-ray crystallographic data; however, the mech-
anism of their formation is not quite clear (Scheme 185).189
Synthesis of Heterocycles Mediated by Benzotriazole
Scheme 184
Scheme 185
Scheme 186
12.4. Oxazines
Hetero-Diels-Alder reaction of 543 with nitrosobenzene
provides 6-benzotriazol-1-yl-2-phenyl-3,6-dihydro-2H-1,2-
oxazine (647) in 90% yield. X-ray crystallographic data of
the product confirm its molecular structure and exclude an
alternative structure with the benzotriazolyl substituent in
the 3-position (Scheme 186).170
Condensation of 1-benzoylbenzotriazoles 648 with 3-ami-
nopropanol under microwave irradiation provides 2-aryl-5,6-
dihydro-4H-1,3-oxazines 649 in 96% (X ) Cl) and 84% (X
) NO2) yields. The best procedure consists of two steps:
(a) a solution of the reagents in chloroform is irradiated at
80 °C for 10 min to give the corresponding N-(3-hydrox-
ypropyl)benzamides; (b) thionyl chloride is added, and
irradiation is continued for an additional 2 min to close the
oxazine ring (Scheme 187).142
Chemical Reviews, 2010, Vol. 110, No. 3 1607
Scheme 187
Scheme 188
Scheme 189
Benzotriazole derivatives 650 are simply prepared by
refluxing benzamide, benzotriazole, and aldehyde
XC6H4CHO in toluene with azeotropic removal of water. In
the presence of aluminum chloride as a Lewis acid, com-
pounds 650 undergo cyclocondensation with alkynes to give
4H-1,3-oxazines 653 in high yields (76-94%). The proposed
reaction mechanism is based on ionization of derivative 650
promoted by aluminum chloride to acyliminium cation 651,
which in an electrophilic attack on the triple bond of alkyne
produces intermediate 652. In the final stage, ring closure
occurs by electron shifts and elimination of a proton to
furnish oxazine 653 (Scheme 188).190
12.5. 1,4-Dithiin
In a reaction with thionyl chloride, the adduct of benzo-
triazole to glyoxal, 437, is converted into dichloro derivative
654 in 90% yield. Compound 654 is a convenient building
block delivering a two-carbon unit to a heterocyclic ring. In
an example of such reactions in Scheme 189, dichloride 654
reacts with 1,2-ethanedithiol disodium salt to provide tet-
rahydro-2,3-bisbenzotriazol-1-yl-1,4-dithiin (655) in 80%
yield. The stereochemistry of 655 is not defined.191
1608 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 190
Scheme 191
12.6. 1,3,5-Triazine
Cyclocondensation of N-acyl-benzotriazol-1-ylcarboximi-
dates 472 with urea or N-methylurea in the presence of
potassium tert-butoxide provides 1,3,5-triazin-2-ones 656 in
76-91% yields. A similar reaction of carboximidamides 472
with thioureas leads to 1,3,5-triazine-2-thiones 657 in
44-88% yields (Scheme 190).192
In the case of N-phenylurea, with the N-phenyl nitrogen
atom being less reactive, only the first step, substitution of
the benzotriazolyl group, occurs under standard conditions,
giving derivatives 658 in 61-89% yields. Using HMDS as
a dehydrating agent, compounds 658 are easily cyclized into
triazinones 659 in 51% (R2 ) t-Bu) to 95% (R2 ) Ph) yields
(Scheme 191).192
12.7. 1,2,4,5-Tetrazine
Aminomethylation of formazan 660 with 1-(dialkylami-
no)benzotriazoles 32 provides unstable intermediates 661 that
under the reaction conditions (barium hydroxide base)
undergo cyclization to 1,2,3,4-tetrahydro-1,2,4,5-tetrazines
662. Rapid oxidation by atmospheric oxygen converts 662
into stable 3-aminoverdazyl radicals 663 that are isolated as
dark-colored crystalline substances in 37-69% yields (Scheme
192).193
Katritzky and Rachwal
Scheme 192
13. Conclusion
Rapid progress in the chemistry of benzotriazole deriva-
tives in the past two decades has led to a wide range of
valuable synthetic methods for all major classes of organic
compounds. Application of the benzotriazole methodology
to heterocyclic chemistry improves the synthesis of many
heterocyclic systems and in some cases allows construction
of the molecules with combination of their substituents
difficult to achieve by other methods. With heterocycles
being the major building blocks in designing biologically
active molecules, we hope that this review will provide a
useful aid to medicinal, crop protection, and other chemists
dealing with heterocyclic systems on a daily basis.
14. References
(1) (a) Katritzky, A. R.; Rachwal, S.; Rachwal, B. J. Chem. Soc., Perkin
Trans. 1 1987, 805. (b) Katritzky, A. R.; Kuzmierkiewicz, W.
J. Chem. Soc., Perkin Trans. 1 1987, 819.
(2) Katritzky, A. R.; Drewniak, M. J. Chem. Soc., Perkin Trans. 1 1988,
2339.
(3) (a) Katritzky, A. R.; Lan, X.; Yang, J. Z.; Denisko, O. V. Chem.
ReV. 1998, 98, 409. (b) Katritzky, A. R.; Belyakov, S. A. Aldrichimica
Acta 1998, 31, 33. (c) Katritzky, A. R.; Qi, M. Collect. Czech. Chem.
Commun. 1998, 63, 599. (d) Katritzky, A. R.; Henderson, S. A.; Yang,
B. J. Heterocycl. Chem. 1998, 35, 1123. (e) Katritzky, A. R.; Qi, M.
Tetrahedron 1998, 54, 2647. (f) Katritzky, A. R. J. Heterocycl. Chem.
1999, 36, 1501. (g) Katritzky, A. R.; Li, J.; Xie, L. Tetrahedron 1999,
55, 8263. (h) Katritzky, A. R.; Denisko, O. V. Pure Appl. Chem.
2000, 72, 1597. (i) Katritzky, A. R.; Toader, D. Synlett 2001, 458.
(j) Katritzky, A. R.; Rogovoy, B. V. Chem.sEur. J. 2003, 9, 4587.
(k) 1656 Katritzky, A. R.; Suzuki, K.; Wang, Z. Synlett 2005, 1656.
(l) Katritzky, A. R.; Manju, K.; Singh, S. K.; Meher, N. K.
Tetrahedron 2005, 61, 2555. (m) Katritzky, A. R.; Kirichenko, K.
ArkiVoc 2006, iV, 119.
(4) Katritzky, A. R.; Wu, J.; Kuzmierkiewicz, W.; Rachwal, S. Liebigs
Ann. Chem. 1994, 1.
(5) Krollpfeiffer, F.; Rosenberg, A.; Muhlhausen, C. Ann. Chem. 1935,
515, 113.
(6) Katritzky, A. R.; Oniciu, D. C.; Ghiviriga, I. Synth. Commun. 1997,
27, 1613.
(7) Mashraqui, S. H.; Kumar, S.; Mudaliar, C. D. Bull. Chem. Soc. Jpn.
2001, 74, 2133.
(8) Le, Z. G.; Chen, Z. C.; Hu, Y.; Zheng, Q. G. Heterocycles 2004, 63,
1077.
(9) Nanjunda Swamy, S.; Basappa Sarala, G.; Priya, B. S.; Gaonkar,
S. L.; Shashidhara Prasad, J.; Rangappa, K. S. Bioorg. Med. Chem.
Lett. 2006, 16, 999.
(10) Perez, E. R.; Loupy, A.; Liagre, M.; de Guzzi Plepis, A. M.; Cordeiro,
P. J. Tetrahedron 2003, 59, 865.
Synthesis of Heterocycles Mediated by Benzotriazole
(11) Liu, Z.; Yu, Y. Synth. React. Inorg. Met.-Org. Chem. 2002, 32, 265.
(12) Katritzky, A. R.; Li, J.; Malhotra, N. Liebigs Ann. Chem. 1992, 843.
(13) Katritzky, A. R.; Puschmann, I. B.; Stevens, C. V.; Wells, A. P.
J. Chem. Soc., Perkin Trans. 2 1995, 1645.
(14) Katritzky, A. R.; Rachwal, S.; Hughes, C. V.; Wang, Z. Pol. J. Chem.
1992, 66, 1633.
(15) Wender, P. A.; Cooper, C. B. Tetrahedron 1986, 42, 2985.
(16) Krasovsky, A. L.; Pissarev, S. A.; Nenajdenko, V. G.; Balenkova,
E. S. Russ. Chem. Bull. 2003, 52, 1791.
(17) Zhang, P. F.; Chen, Z. C. J. Chem. Res., Synop. 2002, 338.
(18) Burckhalter, J. H.; Stephens, V. C.; Hall, L. A. R. J. Am. Chem.
Soc. 1952, 74, 3868.
(19) Katritzky, A. R.; Rachwal, S.; Rachwal, B. J. Org. Chem. 1989, 54,
6022.
(20) Katritzky, A. R.; Rachwal, S.; Caster, K. C.; Mahni, F.; Law, K. W.;
Rubio, O. J. Chem. Soc., Perkin Trans. 1 1987, 781.
(21) Katritzky, A. R.; Rachwal, S.; Rachwal, B. J. Chem. Soc., Perkin
Trans. 1 1987, 799.
(22) Katritzky, A. R.; Yannakopoulou, K.; Lue, P.; Rasala, D.; Urogdi,
L. J. Chem. Soc., Perkin Trans. 1 1989, 225.
(23) Lindsay Smith, J. R.; Sadd, J. S. J. Chem. Soc., Perkin Trans. 1
1975, 1181.
(24) Katritzky, A. R.; Yannakopoulou, K.; Kuzmierkiewicz, W.; Aur-
recoechea, J. M.; Palenik, G. J.; Koziol, A. E.; Szczesniak, M.
J. Chem. Soc., Perkin Trans. 1 1987, 2673.
(25) Katritzky, A. R.; Yannakopoulou, K. Heterocycles 1989, 1121.
(26) Katritzky, A. R.; Rachwal, S.; Rachwal, B. Int. J. Chem. Kinet. 1995,
27, 351.
(27) Katritzky, A. R.; Belyakov, S. A.; Sorochinsky, A. E.; Steel, P.;
Schall, O. F.; Gokel, G. W. J. Org. Chem. 1996, 61, 7585.
(28) Katritzky, A. R.; Strah, S.; Belyakov, S. A. Tetrahedron 1998, 54,
7167.
(29) Ahn, J. H.; Joung, M. J.; Yoon, N. M.; Oniciu, D. C.; Katritzky,
A. R. J. Org. Chem. 1999, 64, 488.
(30) Katritzky, A. R.; Nair, S. K.; Qiu, G. Synthesis 2002, 199.
(31) Katritzky, A. R.; Rachwal, S.; Wu, J. Can. J. Chem. 1990, 68, 446.
(32) Katritzky, A. R.; Sutharchanadevi, M.; Urogdi, L. J. Chem. Soc.,
Perkin Trans. 1 1990, 1847.
(33) Katritzky, A. R.; Rachwal, S.; Rachwal, B. J. Chem. Soc., Perkin
Trans. 1 1987, 791.
(34) Katritzky, A. R.; Pilarski, B.; Urogdi, L. Org. Prep. Proced. Int.
1989, 135.
(35) Idzik, K. R; Cabaj, J.; Soloducho, J.; Abdel-Fattah, A. A. A. HelV.
Chim. Acta 2007, 90, 1672.
(36) Katritzky, A. R.; Kuzmierkiewicz, W.; Rachwal, B.; Rachwal, S.;
Thomson, J. J. Chem. Soc., Perkin Trans. 1 1987, 811.
(37) Katritzky, A. R.; Gordeev, M. F.; Greenhill, J. V.; Steel, P. J. J. Chem.
Soc., Perkin Trans. 1 1992, 1111.
(38) Katritzky, A. R.; Lan, X.; Lam, J. N. Chem. Ber. 1991, 123, 1819.
(39) Katritzky, A. R.; Lam, J. N. Heteroat. Chem. 1990, 1, 21.
(40) Katritzky, A. R.; Yang, Z.; Lam, J. N. J. Org. Chem. 1991, 56, 2143.
(41) Katritzky, A. R.; Yang, Z.; Lam, J. N. J. Org. Chem. 1991, 56, 6917.
(42) Katritzky, A. R.; Drewniak-Deyrup, M.; Lan, X.; Brunner, F.
J. Heterocycl. Chem. 1989, 829.
(43) Katritzky, A. R.; Afridi, A. S.; Kuzmierkiewicz, W. HelV. Chim. Acta
1991, 74, 1931.
(44) Katritzky, A. R.; Ghiviriga, I.; Oniciu, D. C.; Soti, F. J. Heterocycl.
Chem. 1996, 33, 1927.
(45) Katritzky, A. R.; Oniciu, D. C.; Ghiviriga, I.; Soti, F. J. Org. Chem.
1998, 63, 2110.
(46) Katritzky, A. R.; Zhang, G.; Jiang, J. J. Org. Chem. 1995, 60, 7589.
(47) Katritzky, A. R.; Lang, H. J. Org. Chem. 1995, 60, 7612.
(48) Katritzky, A. R.; Lang, H.; Wang, Z.; Zhang, Z.; Song, H. J. Org.
Chem. 1995, 60, 7619.
(49) Katritzky, A. R.; Wang, Z.; Lang, H. Organometallics 1996, 15, 486.
(50) Katritzky, A. R.; Toader, D.; Xie, L. Synthesis 1996, 1425.
(51) Katritzky, A. R.; Yao, J.; Bao, W.; Qi, M.; Steel, P. J. J. Org. Chem.
1999, 64, 346.
(52) Katritzky, A. R.; Jiang, J.; Urogdi, L. Synthesis 1990, 565.
(53) Katritzky, A. R.; Wang, M.; Wilkerson, C. R.; Yang, H. J. Org. Chem.
2003, 68, 9105.
(54) Katritzky, A. R.; Maimait, R.; Denisenko, A.; Denisenko, S. N.
ArkiVoc 2001, V, 68.
(55) Katritzky, A. R.; Heck, K. A.; Li, J.; Wells, A.; Garot, C. Synth.
Commun. 1996, 26, 2657.
(56) Katritzky, A. R.; Manju, K.; Steel, P. J. J. Org. Chem. 2003, 68,
407.
(57) Katritzky, A. R.; Xie, L.; Serdyuk, L. J. Org. Chem. 1996, 61, 7564.
(58) Lacroix, S.; Cheguillaume, A.; Gerard, S.; Marchand-Brynaett, J.
Synthesis 2003, 2483.
(59) Wedler, C.; Kleiner, K.; Kunath, A.; Schick, H. Liebigs Ann. 1996,
881.
(60) Katritzky, A. R.; Luo, Z.; Fang, Y. Tetrahedron Lett. 2000, 41, 9691.
Chemical Reviews, 2010, Vol. 110, No. 3 1609
(61) Gawley, R. E.; Barolli, G.; Madan, S.; Saverin, M.; O’Connor, S.
Tetrahedron Lett. 2004, 45, 1759.
(62) Katritzky, A. R.; Rachwal, S.; Wu, J. Can. J. Chem. 1989, 68, 446.
(63) Katritzky, A. R.; Pilarski, B.; Urogdi, L. J. Chem. Soc., Perkin Trans.
1 1990, 541.
(64) Katritzky, A. R.; Fang, Y.; Qi, M.; Feng, D. Heterocycles 1998, 48,
2535.
(65) Aurrecoechea, J. M.; Fernandez-Acebes, A. Synlett 1996, 39.
(66) Aurrecoechea, J. M.; Fernandez, A.; Gorgojo, J. M.; Saornil, C.
Tetrahedron 1999, 55, 7345.
(67) Bustos, F.; Gorgojo, J. M.; Suero, R.; Aurrecoechea, J. M. Tetra-
hedron 2002, 58, 6837.
(68) Katritzky, A. R.; Koditz, J.; Lang, H. Tetrahedron 1994, 50, 12571.
(69) Katritzky, A. R.; Cui, X. L.; Yang, B; Steel, P. J. J. Org. Chem.
1999, 64, 1979.
(70) Katritzky, A. R.; Yao, J.; Yang, B. J. Org. Chem. 1999, 64, 6066.
(71) Katritzky, A. R.; Allin, S. M. Synth. Commun. 1995, 24, 7612.
(72) Katritzky, A. R.; Mehta, S.; He, H. Y.; Cui, X. J. Org. Chem. 2000,
65, 4364.
(73) Katritzky, A. R.; Szajda, M.; Lam, J. N. Heterocycl. Chem. 1993,
30, 1261.
(74) Katritzky, A. R.; Angrish, P.; Todadze, E.; Ghiviriga, I. Bioorg. Med.
Chem. Lett. 2007, 17, 6000.
(75) Katritzky, A. R.; Hitchings, G. J.; Zhao, X. Synthesis 1991, 863.
(76) Katritzky, A. R.; Verin, S. V. J. Heterocycl. Chem. 1995, 32, 323.
(77) Katritzky, A. R.; Lang, H.; Lan, X. Tetrahedron 1993, 49, 2829.
(78) Katritzky, A. R.; Yang, Z.; Lam, J. N. Tetrahedron 1992, 48, 4971.
(79) Katritzky, A. R.; Xie, L.; Fan, W. Q. J. Org. Chem. 1993, 58, 4376.
(80) Katritzky, A. R.; Pastor, A. J. Org. Chem. 2000, 65, 3679.
(81) Katritzky, A. R.; Suzuki, K.; Singh, S. K.; He, H. Y. J. Org. Chem.
2003, 68, 5720.
(82) Katritzky, A. R.; Jiang, R.; Suzuki, K. J. Org. Chem. 2005, 70, 4993.
(83) Katritzky, A. R.; Akue-Gedu, R.; Vakulenko, A. V. ArkiVoc 2007,
iii, 5.
(84) Katritzky, A. R.; Li, J.; Gordeev, M. F. Synthesis 1994, 93.
(85) Katritzky, A. R.; Rachwal, S. Pol. J. Chem. 1992, 66, 1653.
(86) Katritzky, A. R.; Chang, H. X.; Verin, S. V. Tetrahedron Lett. 1995,
36, 343.
(87) Katritzky, A. R.; Wu, H.; Xie, L.; Rachwal, S.; Rachwal, B.; Jiang,
J.; Zhang, G.; Lang, H. Synthesis 1995, 1315.
(88) Katritzky, A. R.; Zhang, L.; Yao, J.; Denisko, O. V. J. Org. Chem.
2000, 65, 8074.
(89) Katritzky, A. R.; Drewniak, M.; Lue, P. J. Org. Chem. 1988, 53,
5854.
(90) Katritzky, A. R.; Drewniak, M. Tetrahedron Lett. 1988, 29, 1755.
(91) Katritzky, A. R.; Zhu, L.; Lang, H.; Denisko, O.; Wang, Z.
Tetrahedron 1995, 51, 13271.
(92) Katritzky, A. R.; Jiang, J.; Greenhill, J. V. J. Org. Chem. 1993, 58,
1987.
(93) Katritzky, A. R.; Jiang, J.; Steel, P. J. J. Org. Chem. 1994, 59, 2850.
(94) Katritzky, A. R.; Cheng, D.; Musgrave, R. P. Heterocycles 1997,
44, 67.
(95) Katritzky, A. R.; Wang, X.; Denisenko, A. J. Org. Chem. 2001, 66,
2850.
(96) Katritzky, A. R.; Wang, Z.; Li, J.; Levell, J. R. J. Heterocycl. Chem.
1997, 34, 1379.
(97) Katritzky, A. R.; Huang, T. B.; Voronkov, M. V.; Wang, M.; Kolb,
H. J. Org. Chem. 2000, 65, 8819.
(98) Katritzky, A. R.; Zhang, S.; Wang, M.; Kolb, H.; Steel, P. J.
J. Heterocycl. Chem. 2002, 39, 759.
(99) Gupton, J. T.; Hicks, F. A.; Smith, S. Q.; Main, A. D.; Petrich, S. A.;
Wilkinson, D. R.; Sikorski, J. A.; Katritzky, A. R. Tetrahedron 1993,
49, 10205.
(100) Katritzky, A. R.; Rachwal, S.; Rachwal, B. J. Chem. Soc., Perkin
Trans. 1 1990, 1717.
(101) Katritzky, A. R.; Rachwal, S.; Rachwal, B.; Steel, P. J. J. Org. Chem.
1992, 57, 4925.
(102) Katritzky, A. R.; Rachwal, S.; Rachwal, B.; Steel, P. J. J. Org. Chem.
1992, 57, 4932.
(103) Katritzky, A. R.; Li, J.; Gordeev, M. F. J. Org. Chem. 1993, 58,
3038.
(104) Piper, S.; Risch, N. ArkiVoc 2003, i, 86.
(105) Katritzky, A. R.; Qiu, G.; Yang, B. J. Org. Chem. 1997, 62, 8210.
(106) Katritzky, A. R.; Abdel-Fattah, A. A. A.; Vakulenko, A. V.; Tao, H.
J. Org. Chem. 2005, 70, 9191.
(107) Katritzky, A. R.; Suzuki, K.; Singh, S. K. Croat. Chem. Acta 2004,
77, 175.
(108) Katritzky, A. R.; Urogdi, L.; Mayence, A. J. Org. Chem. 1990, 55,
2206.
(109) Katritzky, A. R.; Cui, X.; Long, Q.; Mehta, S.; Steel, P. J. ArkiVoc
2000, iV, 471.
(110) Katritzky, A. R.; Li, J. J. Org. Chem. 1995, 60, 638.
1610 Chemical Reviews, 2010, Vol. 110, No. 3
(111) Katritzky, A. R.; Cheng, D.; Leeming, P.; Ghiviriga, I.; Hartshorn,
C. M.; Steel, P. J. J. Heterocycl. Chem. 1996, 33, 1935.
(112) Yavari, I.; Alizadeh, A.; Anary-Abbasinejad, M. Tetrahedron Lett.
2002, 43, 9449.
(113) Katritzky, A. R.; Lang, H. J. Org. Chem. 1995, 60, 7812.
(114) Katritzky, A. R.; Xie, L.; Toader, D.; Serdyuk, L. J. Am. Chem. Soc.
1995, 117, 12015.
(115) (a) Al-Saleh, B.; Abdel-Khalik, M. M.; Darwich, E.; Abdel-Motaleb
Salah, O.; Elnagdi, M. H. Heteroat. Chem. 2002, 13, 141. (b) Al-
Saleh, B.; Behbehani, H.; El-Apasery, M. A.; Elnagdi, M. H. J. Chem.
Res. 2004, 575. (c) Al-Saleh, B.; El-Apasery, M. A.; Elnagdi, M. H.
J. Heterocycl. Chem. 2005, 42, 483.
(116) Katritzky, A. R.; Denisenko, A.; Arend, M.; Denisenko, S. N.
J. Heterocycl. Chem. 2000, 37, 1309.
(117) Al-Omran, F.; El-Khair, A. A. J. Heterocycl. Chem. 2004, 41, 327.
(118) Katritzky, A. R.; Vakulenko, A. V.; Akue-Gedu, R.; Gromova, A. V.;
Witek, R.; Rogers, J. W. ArkiVoc 2007, i, 9.
(119) Katritzky, A. R.; Wang, M.; Zhang, S.; Voronkov, M. V.; Steel, P. J.
J. Org. Chem. 2001, 66, 6787.
(120) Katritzky, A. R.; Fali, C. N.; Oniciu, D. C. Tetrahedron 1995, 51,
1069.
(121) Kuehne, M. E.; Weaver, S. J.; Franz, P. J. Org. Chem. 1964, 29,
1582.
(122) Katritzky, A. R.; Musgrave, R. P.; Breytenbach, J. C. J. Heterocycl.
Chem. 1996, 33, 1637.
(123) Kim, T.; Kim, K.; Park, Y. J. Eur. J. Org. Chem. 2002, 493.
(124) Katritzky, A. R.; Suzuki, K.; He, H. Y. J. Org. Chem. 2002, 67,
3109.
(125) Katritzky, A. R.; Suzuki, K.; He, H. Y. J. Org. Chem. 2002, 67,
8224.
(126) Katritzky, A. R.; Luo, Z.; Fang, Y. J. Org. Chem. 2001, 66, 2858.
(127) Katritzky, A. R.; Fali, C. N.; Li, J.; Ager, D. J.; Prakash, I. Synth.
Commun. 1997, 27, 1623.
(128) Butula, I.; Jadrijevic-Mladar Takac, M. Croat. Chem. Acta 2000, 73,
569.
(129) Kalcic, I.; Zovko, M.; Jadrijevic-Mladar Takac, M.; Zorc, B.; Butula,
I. Croat. Chem. Acta 2003, 76, 217.
(130) Rajic, Z.; Zorc, B.; Raic-Malic, S.; Ester, K.; Kralj, M.; Pavelic, K.;
Balzarini, J.; De Clercq, E.; Mintas, M. Molecules 2006, 11, 837.
(131) Opacic, N.; Barbaric, M.; Zorc, B.; Cetina, M.; Nagl, A.; Frkovic,
D.; Kralj, M.; Pavelic, K.; Balzarini, J.; Andrei, G.; Snoeck, R.; De
Clercq, E.; Raic-Malic, S.; Mintas, M. J. Med. Chem. 2005, 48, 475.
(132) Katritzky, A. R.; Wrobel, L.; Savage, G. P. J. Chem. Res., Synop.
1990, 330.
(133) Katritzky, A. R.; Jiang, J.; Harris, P. A.; Steel, J. P. Heterocycles
1990, 31, 2187.
(134) Katritzky, A. R.; Hitchings, G. J.; Zhao, X. J. Chem. Soc., Perkin
Trans. 1 1990, 2371.
(135) Liu, K. C.; Shelton, B. R.; Howe, R. K. J. Org. Chem. 1980, 45,
3916.
(136) Larsen, K. E.; Torssell, K. B. G. Tetrahedron 1984, 40, 2985.
(137) Katritzky, A. R.; Button, M. A. C.; Denisenko, S. N. J. Heterocycl.
Chem. 2000, 37, 1505.
(138) Katritzky, A. R.; Feng, D.; Qi, M. Tetrahedron Lett. 1998, 39, 6835.
(139) Katritzky, A. R.; Cobo-Domingo, J.; Yang, B.; Steel, P. J. J. Chem.
Res., Synop. 1999, 162.
(140) Katritzky, A. R.; Ledoux, S.; Witek, R. M.; Nair, S. K. J. Org. Chem.
2004, 69, 2976.
(141) Katritzky, A. R.; Khashab, N. M.; Haase, D. N.; Yoshioka, M.;
Ghiviriga, I.; Steel, P. J. J. Org. Chem. 2007, 72, 6742.
(142) Katritzky, A. R.; Cai, C.; Suzuki, K.; Singh, S. K. J. Org. Chem.
2004, 69, 811.
(143) Katritzky, A. R.; Donkor, A.; Fang, Y. J. Org. Chem. 2001, 66, 4041.
(144) Katritzky, A. R.; Chen, Y. X.; Yannakopoulou, K.; Lue, P. Tetra-
hedron Lett. 1989, 30, 6657.
(145) Katritzky, A. R.; Wu, H.; Xie, L. J. Heterocycl. Chem. 1995, 32,
1651.
(146) Katritzky, A. R.; Chen, J.; Yang, Z. J. Org. Chem. 1995, 60, 5638.
(147) Katritzky, A. R.; Wang, X.; Maimait, R. J. Org. Chem. 2000, 65,
8077.
(148) Katritzky, A. R.; Zhang, Y.; Singh, S. K.; Steel, P. J. ArkiVoc 2003,
xV, 47.
(149) Thambidurai, S.; Jeyasubramanian, K.; Ramalingam, S. R. Polyhedron
1996, 15, 4011.
(150) Katritzky, A. R.; Rogovoy, B. V.; Chassaing, C.; Vvedensky, V. J.
Org. Chem. 2000, 65, 8080.
(151) Katritzky, A. R.; Rogovoy, B. V.; Vvedensky, V. Y.; Kovalenko,
K.; Steel, P. J.; Markov, V. I.; Forood, B. Synthesis 2001, 897.
(152) Makara, G. M.; Ma, Y.; Margarida, L. Org. Lett. 2002, 4, 1751.
(153) Katritzky, A. R.; Vvedensky, V.; Cai, X.; Rogovoy, B.; Steel, P. J.
ArkiVoc 2002, Vi, 82.
Katritzky and Rachwal
(154) Katritzky, A. R.; Shestopalov, A. A.; Suzuki, K. ArkiVoc 2005, Vii,
36.
(155) Katritzky, A. R.; Huang, T. B.; Voronkov, M. V. J. Org. Chem. 2000,
65, 2246.
(156) Schellhammer, C. W.; Schroeder, J.; Joop, N. Tetrahedron 1970, 26,
497.
(157) El-Khawass, S. M.; Habib, N. S. J. Heterocycl. Chem. 1989, 26,
177.
(158) Katritzky, A. R.; Tymoshenko, D. O.; Chen, K.; Fattah, A. A. ArkiVoc
2001, ii, 101.
(159) Domling, A.; Beck, B.; Magnin-Lachaux, M. Tetrahedron Lett. 2006,
47, 4289.
(160) Katritzky, A. R.; Cai, C.; Meher, N. K. Synlett 2007, 1204.
(161) Katritzky, A. R.; Belyakov, S. A.; Lam, J. N.; Durst, H. D.; Karpenko,
D. V. Heterocycles 1994, 39, 73.
(162) Katritzky, A. R.; Fan, W. Q. J. Org. Chem. 1990, 55, 3205.
(163) Veerman, J. J. N.; Klein, J.; Aben, R. W. M.; Scheeren, H. W.; Kruse,
C. G.; van Maarseveen, J. H.; Rutjes, F. P. J. T.; Hiemstra, H. Eur.
J. Org. Chem. 2002, 3133.
(164) (a) Xu, X.; Lu, J.; Li, R.; Ge, Z.; Dong, Y.; Hu, Y. Synlett 2004,
122. (b) Wang, X.; Dong, Y.; Sun, J.; Xu, X.; Li, R.; Hu, Y. J. Org.
Chem. 2005, 70, 1897.
(165) Katritzky, A. R.; Yang, B.; Qiu, G.; Steel, P. J. J. Org. Chem. 1998,
63, 6699.
(166) Zheng, J. F.; Jin, L. R.; Huang, P. Q. Org. Lett. 2004, 6, 1139.
(167) Katritzky, A. R.; Luo, Z.; Cui, X. L. J. Org. Chem. 1999, 64, 3328.
(168) Katritzky, A. R.; Luo, Z.; Fang, Y.; Feng, D.; Ghiviriga, I. J. Chem.
Soc., Perkin Trans. 2 2000, 1375.
(169) Katritzky, A. R.; Gordeev, M. F. J. Org. Chem. 1993, 58, 4049.
(170) Katritzky, A. R.; Gupta, V.; Gordeev, M. J. Heterocycl. Chem. 1993,
30, 1073.
(171) Rachwal, S.; Katritzky, A. R. ComprehensiVe Heterocyclic Chemistry
III; Elsevier Ltd.: Oxford, U.K., 2008; Vol. 5, Chapter 01.
(172) Katritzky, A. R.; Abdel-Fattah, A. A. A.; Akhmedova, R. G. ArkiVoc
2005, Vi, 329.
(173) Katritzky, A. R.; El-Zemity, S.; Lang, H. J. Chem. Soc., Perkin Trans.
1 1995, 3129.
(174) Vera-Luque, P.; Alajarin, R.; Alvarez-Builla, J.; Vaquero, J. J. Org.
Lett. 2006, 8, 415.
(175) Katritzky, A. R.; Shcherbakova, I. V. J. Heterocycl. Chem. 1996,
33, 2031.
(176) Katritzky, A. R.; Belyakov, S. A.; Sorochinsky, A. F.; Henderson,
S. A.; Chen, J. J. Org. Chem. 1997, 62, 6210.
(177) (a) Alnajjar, A.; Abdelkhalik, M. M.; Al-Enezi, A.; Elnagdi, M. H.
Molecules 2009, 14, 68. (b) Al-Saleh, B.; Abdelkhalik, M. M.; El-
Apasery, M. A.; Elnagdi, M. H. J. Heterocycl. Chem. 2003, 40, 171.
(178) Katritzky, A. R.; Mazurkiewicz, R.; Stevens, C. V.; Gordeev, M. F.
J. Org. Chem. 1994, 59, 2740.
(179) Katritzky, A. R.; Abdel-Fattah, A. A. A.; Tymoshenko, D. O.;
Essawy, S. A. Synthesis 1999, 2114.
(180) Katritzky, A. R.; Denisenko, A.; Arend, M. J. Org. Chem. 1999, 64,
6076.
(181) Katritzky, A. R.; Du, W.; Denisenko, S. N.; Czerney, P.; Steel, P. J.
J. Prakt. Chem. (Weinheim, Ger.) 1999, 341, 152.
(182) Katritzky, A. R.; Denisko, O. V. J. Org. Chem. 2002, 67, 3104.
(183) Degl’Innocenti, A.; Capperucci, A.; Oniciu, D. C.; Katritzky, A. R.
J. Org. Chem. 2000, 65, 9206.
(184) Katritzky, A. R.; Pleynet, D. P. M.; Yang, B.; Yao, J. Synthesis 1998,
1627.
(185) (a) Al-Omran, F.; Al-Awadl, N.; Yousef, O.; Elnagdi, M. H.
J. Heterocycl. Chem. 2000, 37, 167. (b) Al-Mousawi, S.; Elassar,
A. Z.; El-Apasery, M. A. Phosphorus, Sulfur Silicon Relat. Elem.
2006, 181, 1755.
(186) Katritzky, A. R.; Singh, S. K.; He, H. J. J. Org. Chem. 2002, 67,
3115.
(187) Abdel-Fattah, A. A. A. Synthesis 2003, 2358.
(188) Zapol’skii, V. A.; Namyslo, J. C.; Altug, C.; Mimoza, G.; Kaufmann,
D. E. Synthesis 2008, 304.
(189) Katritzky, A. R.; Wang, Z.; Lang, H. Heterocycl. Commun. 1996, 2,
109.
(190) Katritzky, A. R.; Pernak, J.; Fan, W. Q. J. Prakt. Chem./Chem.-Ztg.
1992, 334, 114.
(191) Katritzky, A. R.; Fan, W. Q. J. Heterocycl. Chem. 1990, 27, 1543.
(192) Katritzky, A. R.; Rogovoy, B. V.; Vvedensky, V. Y.; Hebert, N.;
Forood, B. J. Org. Chem. 2001, 66, 6797.
(193) Katritzky, A. R.; Belyakov, S. A.; Durst, H. D.; Xu, R.; Dalal, N. S.
Can. J. Chem. 1994, 72, 1849.
CR900204U
 Chem. Rev. 2010, 110, 1611–1641 1611

Hydrogen Autotransfer in the N-Alkylation of Amines and Related Compounds

using Alcohols and Amines as Electrophiles

Gabriela Guillena,* Diego J. Ramón,* and Miguel Yus*

Instituto de Sı́ntesis Orgánica and Departamento de Quı́mica Orgánica, Universidad de Alicante, Facultad de Ciencias,
Apartado 99, E-03080 Alicante, Spain

Received June 8, 2009

Contents
1. Introduction and Definitions 1611
2. Alcohols as Source of Electrophiles 1613
2.1. Heterogeneous Catalysts 1613
2.1.1. Derived from Silicon and Aluminum 1613
2.1.2. Derived from Nickel 1614
2.1.3. Derived from Copper 1616
2.1.4. Derived from Platinum 1619
2.1.5. Others 1619
2.2. Homogeneous Catalysts 1620
2.2.1. Without Transition-Metal Catalysts 1620
2.2.2. Derived from Ruthenium 1621
2.2.3. Derived from Iridium 1628
2.2.4. Others 1632 Gabriela Guillena was born in Alicante (Spain) in 1970 and received her
B.Sc. degree (1993) from the University of Alicante. After spending one
3. Amines as Source of Electrophiles 1633 year as postgraduate student in the group of Prof. D. Seebach (ETH,
3.1. Heterogeneous Catalysts 1633 Zürich), she returned to the University of Alicante and received her M.Sc.
3.1.1. Derived from Nickel 1633 (1995) and Ph.D. (2000) degrees. After two years as a postdoctoral fellow
3.1.2. Derived from Copper 1634 in the research group of Prof. G. van Koten (University of Utrecht,
Netherlands), she returned to the University of Alicante where she became
3.1.3. Derived from Palladium 1634 assistant professor (2003) and associate professor (2007) through the
3.1.4. Derived from Platinum 1635 National Habilitation process. Her current research interest is focused on
3.1.5. Others 1635 new organic methodologies and asymmetric organocatalysis.
3.2. Homogeneous Catalysts 1636
3.2.1. Derived from Ruthenium 1636 The unquestionable interaction between nitrogen-contain-
3.2.2. Others 1637 ing compounds with living organisms has impulsed the
4. Conclusions and Outlook 1638 pharmaceutical and agrochemical industries into the develop-
ment of many different molecules containing this atom.4 In
5. Acknowledgments 1638
fact, the vast majority of compounds used in pharmaceutical
6. References 1638 companies contain at least one nitrogen atom. Moreover,
nitrogen-containing compounds are of great importance in
1. Introduction and Definitions other aspects of the chemical industry, such as in the cases
The chemistry of amines, amides, and other nitrogen- of preparation of dyes, detergents, surfactants, fabric soften-
containing compounds plays a central role in the organic ers, emulsion and pigment stabilizers, epoxy hardeners,
vulcanizing agents, and additives in the petroleum industry.5
synthesis. In fact, the first synthesized organic compound
was urea.1 The great importance of this type of nitrogen- Among nitrogen-containing compounds, amines are the
main and most important ones, and therefore, there are a
containing compound takes its roots from several factors.
plethora of different methods and variants to prepare them,6
The first one is that Nature has chosen different nitrogen including electrophilic alkylation processes,7 reductive alkyl-
derivatives as their typical building blocks in the construction ation processes,8 and amination of aryl halides.9
of Life, such as amino acids and nucleotides.2 But not only All of the above protocols have reached excellent levels
the main constituents of Nature are included in this category; of yield. However, since the beginning of the 1990s,10 the
other important minor compounds such as neurotransmissors, attention of chemists has been drawn to the necessity to
natural toxics, alkaloids, and other active biomolecules should develop methods with increasing environmental awareness
be included in this group.3 and integrated in sustainable processes.11 A re-examination
of the established strategy of synthesis, utilization of hazard-
ous chemicals, and utility of solvents, as well as their
* Phone: +34 965903986. Fax: +34 965903549. E-mail: gabriela.guillena@
ua.es; djramon@ua.es; yus@ua.es. URLs: http://iso.ua.es/ and http:// environmental impact, has to be taken into account. Thus,
www.ua.es/dept.quimorg/index.html. following these environmental aspects, it is necessary not
10.1021/cr9002159  2010 American Chemical Society
Published on Web 11/23/2009
1612 Chemical Reviews, 2010, Vol. 110, No. 3
Diego J. Ramón was born in Alicante (Spain) in 1965 and received his
B.Sc. (1988), M.Sc. (1989), and Ph.D. (1993) degrees from the University
of Alicante. After spending two years as a postdoctoral fellow at the
Eidgenössische Technische Hochschule in Zürich (ETH-Zentrum), he
returned to the University of Alicante and, after a short stay at Miguel
Hernández University, became associate professor (2000) in the former
university. Dr. Ramón has been visiting professor at Debye Institute
(University of Utrecht, Netherlands, 2001). In 1994, he was awarded the
National Prize for Young Scientists of the Spanish Royal Society of
Chemistry. His current research interest is focused on organometallic
chemistry and asymmetric synthesis.
Miguel Yus (left) was born in Zaragoza (Spain) in 1947 and received his
B.Sc. (1969), M.Sc. (1971), and Ph.D. (1973) degrees from the University
of Zaragoza. After spending two years as a postdoctoral fellow at the
Max Planck Institut für Kohlenforschung in Mülheim a.d. Ruhr, he returned
to Spain to the University of Oviedo where he became associate professor
in 1977, being promoted to full professor in 1987 at the same university.
In 1988 he moved to a chair in Organic Chemistry at the University of
Alicante where he is currently the head of the Organic Synthesis Institute.
Professor Yus has been visiting professor at different institutions and
universities such as ETH-Zentrum, Oxford, Harvard, Uppsala, Marseille,
Tucson, Okayama, Paris, and Strasbourg. He is co-author of more than
450 papers (among them 7 review articles in this journal) mainly in the
field of the development of new methodologies involving organometallic
intermediates. His current research interest is focused on the preparation
of very reactive functionalized organometallic compounds and their use
in synthetic organic chemistry, arene-catalyzed activation of different
metals, and preparation of new metal-based catalysts for homogeneous
and heterogeneous selective reactions. Among others, he has received
the Spanish-French Prize (1999), twice the Japan Society for the
Promotion of Science Prize (2000, 2007), and the Stiefvater Memorial
Lectureship Award (2001). Professor Yus has belonged, among others,
to the advisory board of the journals Tetrahedron, Tetrahedron Letters,
European Journal of Organic Chemistry, Chemistry Letters, Current Organic
Chemistry, and Trends in Organic Chemistry. Professor Yus, Dr. Ramón,
and other members of the ISO founded the new chemical company
MEDALCHEMY S.L. to commercialize fine chemicals.
only to maximize the chemical yields but also to minimize
the waste production and ideally to eliminate it.
Guillena et al.
Scheme 1. General Scheme for a Hydrogen Autotransfer
Process
There are different parameters and ratios to evaluate the
environmental impact of a reaction.12 Perhaps the most
simple and easy-to-calculate parameter is the atom efficiency,
which gives a rapid idea of the percentage of the material
incorporated to the product, as well as its waste (see eq a).
According to this equation, the aforementioned methodology
for the preparation of amines has a very low incorporation
of reagents into the final product and a very high amount of
waste. The presence of leaving groups with high molecular
weights in the reagents, as well as the use of stoichiometric
amount of reagents that only transfer small portions of them
to the final product, decreases enormously the atom efficiency
of the process.
atom efficiency (%) ) yield (%) ×
Mw of final product
(a)
∑
i)reagents
(equiv × Mw)i + ∑
j)catalysts and additives
(equiv × Mw)j
There are only few general and attractive methodologies
for the synthesis of amines that take into account the atom
efficiency factor. One of them is the hydroamination of
olefins and alkynes,13 which is limited due to the electrophilic
character of the nitrogen reagent, as well as the existence of
the corresponding olefin. Using this approach, it is impossible
to perform the methylation, benzylation, and related pro-
cesses. Moreover, the regioselectivity of the addition is an
important aspect that never should be underestimated.
Another interesting protocol is the hydrogen autotransfer
process,14 known also as either borrowing hydrogen15 or self-
supplying system for active hydrogen.16 This type of reaction
is included in the general hydrogen transfer reaction field17
and always starts with the abstraction of a hydrogen atom
from the starting reagent (R1-H) by the corresponding
catalyst to generate a new reagent (R1). The abstracted
hydrogen is further returned and incorporated to the final
product, hence the reaction name (see Scheme 1). In the case
of carbon-carbon bond-formation reactions, two subtypes
of pathways have been found, depending on the sequence of
reaction of the three formed reagents (R1, R2, and Cat.-H).
One of them is the so-called activation of the intermediate,18
in which the second step is the condensation reaction between
both reagents R1 and R2, and finishes with the reduction of
the formed double bond by a hydride catalyst. Another one
is the so-called activation of reagent,19 in which the second
step is the hydrometalation of the initial reagent R2 by the
hydride catalyst, and finishes with the addition reaction of
the above organometallic to the in situ generated reagent R1.
Among all different protocols for the syntheses of amines,
the hydroamination and hydrogen autotransfer processes are
clearly superior and their superiority is illustrated, for
instance, by the fact that industries synthesizing amine
intermediates for foam booster in household detergents and
other applications only use these two strategies. So, whereas
P&GsChemicals, Albermarle, and Huntsman use the hy-
Hydrogen Autotransfer in the N-Alkylation of Amines Chemical Reviews, 2010, Vol. 110, No. 3 1613

Scheme 2. General Scheme for the N-Alkylation of Amines Table 1. N-Alkylation of Aniline with Primary Alcohols using
through a Hydrogen Autotransfer Process Silicon or Aluminum Oxides

entry catalyst R T (°C) no. yield (%) no. yield (%)

1 SiO2 H 362 3a 20 4a 25
2 SiO2 Me 385 3b 47 4b 13
droamination processes, Lonza (U.S.A.), Clariant (Germany), 3 SiO2 Et 385 3c 37 4c 12
Kao (Japan), and Feixiang Chem. (China) use the hydrogen 4 SiO2 Pr 400 3d 27 4d 9
autotransfer protocols. 5 SiO2 Me 400 3b 28 4b 5
The alkylation of amines through a hydrogen autotransfer 6 V2O5-SiO2 Me 400 3b 66 4b 15
7 Al2O3a H 400 3a 32 4a 50
process should be classified into the intermediate activation 8 γ-Al2O3 H 400 3a 15 4a 2
set (Scheme 2), in which either an alcohol or an amine is 9 γ-Al2O3 H 375 3a 24 4a 18
the source of the electrophilic agent, changing the usual 10 γ-Al2O3 H 200 3a 41 4a 5
typical nucleophilic behavior of these compounds. This a
Prepared from Al(OPri)3.
strategy has obvious advantages toward hydroamination
processes such as the simplicity of protocol, the availability One of the oldest examples of N-alkylation by a hydrogen
of starting reagents, and the important number of catalysts autotransfer process was performed by using silica gel as
able to perform this reaction. catalyst.26 In this early work, simple aniline (1a) was
Although some previous reviews have covered partially alkylated with different aliphatic primary alcohols (2,
the use of hydrogen autotransfer process for the preparation 100-200 mol %) at very high temperatures to give a mixture
of amines,20 there is not so far a general article showing the of substituted anilines 3 and 4 with, in general, moderated
great potential of this strategy for these thrifty times. The yields (Table 1, entries 1-4). The detection of the corre-
review is assorted by the type of general source of the used sponding aldehydes and hydrogen as byproduct showed that
electrophile, with the initial morphology of the reaction the reaction pathway was the hydrogen autotransfer process,
media, and taking into consideration the type of metal used. with the silica gel being a modest catalyst to promote this
reaction.
2. Alcohols as Source of Electrophiles The aforementioned efficiency could be enhanced by
mixing another metal oxide on the silica gel support.27 Thus,
The alkylation of amines by alcohols with loss of water when vanadia (V2O5), prepared by calcinations of NH4VO3,
is a thermodynamically favored process where the loss of a was used as the catalyst for the ethylation of aniline under
carbon-oxygen bond for forming a carbon-nitrogen bond flow reaction conditions, only 21% of compound 3b was
is compensated by the gain of an oxygen-hydrogen bond obtained, together with 6% of 4b. However, when vanadia
from a nitrogen-hydrogen bond.21 In fact, the suitability of was supported on silica gel (2-20 wt %) by wet impregna-
this reaction was already proven in 1901, when the alkylation tion of silica with the appropriate amount of NH4VO3 and
of aniline was successfully performed using different sodium oxalic acid, the results were better. So, the conversion
alkoxides.22 reached up to 93%, showing that the support enhanced the
results (Table 1, entry 6) compared with the use of either
2.1. Heterogeneous Catalysts the vanadia or the silica alone (Table 1, entry 5).
Along this section, different examples of catalysts which Conversely to typical acid metal oxides, basic metal oxides
were initially in a different phase (solid, liquid, gas) from clearly catalyzed the N-alkylation of amines. Thus, different
the reagents will appear. The heterogeneous catalysts23 have types of alumina (Al2O3) have been used as catalysts in this
some advantages over the homogeneous one, with the process. When alumina, prepared from aluminum isopro-
possible easy recovery being the most important one.24 poxide, was used as heterogeneous catalyst in a conventional
flow reactor and using a vaporized mixture of aniline (1a)
2.1.1. Derived from Silicon and Aluminum and methanol (2a, 15 000 mol %) at 400 °C, the main product
was the double alkylated compound 4a. However, when the
Silicon and aluminum oxides are collected in the same catalyst used was commercially available γ-Al2O3 under the
section because of their similar behavior. In fact, different same reaction conditions, the main product was compound
metal oxides can drive the N-alkylation of amines through 3a with lower yield (compare entries 7 and 8 in Table 1).
an acid-catalyzed mechanism. However, there are several These facts highlighted the crucial role of the preparation
facts that could show us that, in the below reported cases, of the catalyst.28a A systematic study using a fixed bed of
the reaction pathway goes through a hydrogen autotransfer γ-Al2O3 was carried with an integral-flow reactor,28b showing
process. One of them, favoring the hydrogen autotransfer that the yield increased with the reaction temperature, with
process, is the presence of the corresponding aldehyde or compound 4a being the main one. Also, the increase of
alkane in the reaction media, which came from the oxidation methanol/aniline molar ratio enhanced the selectivity toward
of the alcohol and a decarbonylation process. Other argu- the amine 4a. The maximum amount of N-methylaniline (3a)
ments are the absence of reaction with tertiary alcohols, as was achieved at 375 °C and with an alcohol/aniline ratio of
well as the total regioselectivity for secondary alcohols, or 2 (Table 1, entry 10). The kinetic measurements showed that
the no-relationship between accessible acid sites and kinet- the first methylation was a pseudofirst-order reaction with
ics.25 All these facts may indicate that the process goes respect to the aniline concentration, with the apparent
through a hydrogen autotransfer mechanism and not through activation energies for the first and second methylation being
an acid-catalyzed process. 62.7 and 48.3 kJ mol-1, respectively.
1614 Chemical Reviews, 2010, Vol. 110, No. 3
Better selectivity was obtained using gas-phase conditions
and γ-Al2O3 at atmospheric pressure and lower temperature
(Table 1, entry 10). Under similar conditions, other different
alcohols such as 1- and 2-propanol (30 000 mol %) could
be used as the initial source of the electrophile. Whereas for
the primary alcohol yields up to 72% for 3c and 8% for 4c
were reached, only 10% of the monomethylated aniline
derived from the secondary one was obtained, probably due
to steric effects.29 Other amines such as octylamine, ben-
zylamine, and chiral R-methylbenzylamine could be alkylated
under similar reaction conditions using methanol, 1-propanol,
or benzylic alcohol, with high conversions and selectivities
for the monoalkylated products. Even ammonia (5) could
be alkylated using benzylic alcohol, although in this case
the reaction was performed at 170 °C under ammonia
atmosphere, giving only 48% of benzylamine as well as 23%
of di- and trialkylated amines.30
The selectivity of this process could also be changed to
the dialkylated product just by changing the reaction condi-
tions. Thus, the alkylation of 1-butylamine with methanol
(alcohol/amine ratio ) 3) at 320 °C and 4 atm gave N,N-
dimethyl-1-butylamine as the main product, with this com-
pound being an important component for the preparation of
different polymers.31
2.1.2. Derived from Nickel
Different nickel catalysts have been historically used in
the N-alkylation of amines using alcohols as electrophilic
source. Thus, particles of nickel obtained by reduction of
NiO at 300 °C were effective catalysts in the alkylation of
different aromatic amines, such as aniline (1a) or para-
toluidine, with aliphatic alcohols, such as methanol, ethanol,
or cyclohexanol. For instance, the corresponding N-ethyla-
niline (3b) was obtained in 30% yield, by using 10% weight
of nickel catalyst, 2/1 ratio of ethanol/aniline, at 180 °C in
an autoclave for 12 h. In a similar way, it was possible to
obtain cyclohexylamine or dicyclohexylamine from cyclo-
hexanol and ammonia (5) by just varying the reaction
temperature from 150 °C for the primary amine to 190 °C
for the secondary one.32
Better results were achieved by using nickel particles (100
mesh) as catalyst and potassium as base in the alkylation of
aniline (1a) with benzyl alcohol (6a, 200 mol %), as well as
the corresponding 4-substituted derivatives, in xylene at 150
°C (see Table 2).33 The reaction was performed in a
Dean-Stark apparatus until water evolution was completed.
Interestingly, whereas the presence of electron-withdrawing
Table 2. N-Alkylation of Aromatic Amines with Benzylic
Alcohols Using Nickel Particles
entry R1 R2 no. yield (%)
1 H H 7a 89
2 H Cl 7b 78
3 H Me 7c 89
4 H MeO 7d 84
5 Cl H 7e 87
6 Me H 7f 93
7 MeO H 7g 88
Guillena et al.
groups in the benzyl alcohol derivative 6 decreased the
reaction rate, the opposite effect was observed for anilines
1. Under these conditions, the reaction could also be
performed using simple aliphatic primary alcohols 2, such
as hexanol or decanol.
Other catalysts such as nickel supported in SiO2 or
mixtures of SiO2/Al2O3 were able to catalyze the reaction
of 2-aminoethanol with ammonia (5) to give ethylenedi-
amine,34 with the second supported catalyst being more
effective than the first one. However, yields never surpassed
25% of the product.
Without doubt, the so-called Raney-nickel or nickel sponge
catalyst has been the most used catalyst in this context. These
catalysts are produced when a block of nickel-aluminum
alloy is treated with concentrated sodium hydroxide. This
treatment, called activation, dissolves most of the aluminum
out of the alloy. The resulting porous structure has a large
surface area, which gives high catalytic activity, obtaining
different types of Raney-nickel depending on the basic
treatment, in all cases with a high amount of nickel (>85%),
with the remaining aluminum helping to preserve the porous
structure of the catalyst.35 Thus, N-alkyl substituted anilines
of type 3 were prepared using Raney-nickel (ca. 60% weight)
in a refluxing mixture of aniline and an excess of alcohol
(500 mol %) for 16 h. For straight-chain primary alcohols
different from methanol, yields ranked 78-83%, while those
derived from branched primary alcohols gave the corre-
sponding N-alkyl anilines in only 41-49% yields.36 Shorter
reaction times (ca. 2 h) were sufficient if a large amount of
catalyst was employed in similar reaction conditions. Fol-
lowing a similar protocol, not only aniline (1) but also 2,5-
dimethoxyaniline and 2-naphthlyamine could be alkylated
selectively using primary and secondary alcohols, such as
ethanol, 2-propanol, 1-butanol, cyclohexanol, and benzyl
alcohol, with yields ranging from 22 to 82%.37 When 1 equiv
of aluminum tert-butoxide was added into the reaction
mixture, only 30% of Raney-nickel was necessary, with the
amount of alcohol being decreased to 3-1 equiv with respect
to the aniline. Under these conditions, the corresponding
alkylated aniline derivatives could be obtained with excellent
results (96-98%). This procedure could also be applied to
other amines, such as cyclohexylamine, with similar results.38
A kinetic study of the process was carried out, finding a
reaction-rate law that depended on first order for aniline and
aluminum tert-butoxide concentrations.39 Also the influence
of the substitution of the aniline in the reaction rate was
studied. The data showed that the presence of electron-
donating groups in the aromatic ring increased the reaction
rate, whereas electron-withdrawing groups decreased it, with
the exception of 4-methoxyaniline (1d) and the related
3-methoxyaniline, which reacted slower than aniline (1a).
A plausible explanation for this behavior came from the
possible complexation of the methoxy group and the
aluminum atom, exerting the possible electron-donating
effect of the methoxy group. According to the observed rate
law, a mechanism scheme was postulated, in which, after
the oxidation of alcohol to the corresponding carbonyl
compound catalyzed by the Raney-nickel, the obtained
carbonyl compound coordinates to the aluminum alkoxide,
giving an activated carbonyl species that, in turn, reacts with
aniline to yield the corresponding imine, which is reduced
finally by the alcohol in a Meerwein-Ponndorf-Verley type
process.
Hydrogen Autotransfer in the N-Alkylation of Amines
Table 3. N-Alkylation of Indole with Secondary Alcohols using
Raney-Nickel
entry R1 R2 no. yield (%)
1 Me Me 10a 90
2 Me Et 10b 94
3 (CH2)5 10c 70
Scheme 3. Double N-Alkylation of a Primary Amine by an
Alcohol and Another Amine Using Raney-Nickel
Following the former protocol, indole (8a) could be
successfully alkylated using an excess of a secondary alcohol
(9, 4 000 mol %) in the presence of an excess of aluminum
tert-butoxide, and catalyzed by Raney-nickel W-2, after three
days at toluene reflux (Table 3).40
Raney-nickel was effective not only in the monoalkylation
of amines but also in their double alkylation, as is depicted
in Scheme 3.41 In this case, the alkylation of 1,6-hexanedi-
amine (11) using equimolecular amounts of 1,4-butanediol
(12) and a Dean-Stark apparatus surprisingly gave the
azepine derivative 13, with a hydroxy and an amine group,
being the source of the successive electrophiles (for the use
of amines as the source of electrophiles, see section 3). The
considerable amount of ammonia evolved from the reaction
media showed that the hydrogen autotransfer pathway was
involved in both alkylation processes. The first step might
be the oxidation of an amine group to the corresponding
imine, followed by condensation of an amine group with
the in situ formed imine, to give the expected cyclic imine
and ammonia, and the final reduction gives the corresponding
hexahydroazepine. Finally, the second alkylation takes place
with the alcohol 12.
Several secondary and tertiary amines were prepared in
34-80% yield using nickel catalyst (12 mol %) under a
hydrogen atmosphere (100 atm pressure) at 200 °C in a few
hours, using different aliphatic and cyclic amines as nucleo-
philes, and ethanol, 1-butanol, and cyclohexanol (200 mol
%) as electrophiles.42 The involved pathway is similar to
the aforementioned one, with the nickel catalyst favoring the
oxidation of the alcohol to the corresponding aldehyde or
ketone and generating nickel hydride intermediates. Then,
after formation of the corresponding imine, by condensation
of the amine and the in situ formed carbonyl compound, the
final reduction took place by the nickel hydride intermediate.
Although the last reduction could be performed partially by
the hydrogen present in the reaction media, the hydrogen
autotransfer process was also acting. In fact further studies
on similar processes showed that the presence or absence of
hydrogen in different hydrogen autotransfer processes neither
altered the kinetics nor the results of the processes, only
affecting the deactivation of the nickel catalyst. In fact, the
presence of hydrogen prevented the formation of an inactive
nitride layer on the metal surface, as well as the formation
of different nickel-carbonyl derivatives, permitting a de-
crease of the amount of the catalyst.43
Chemical Reviews, 2010, Vol. 110, No. 3 1615
Scheme 4. Double N-Alkylation of 1-Aminopropan-2-ol
Using Raney-Nickel
Different nitrogen-containing heterocyclic compounds
were prepared using amino alcohol derivatives as starting
materials and Raney-nickel as catalyst under a hydrogen
atmosphere. For instance, 1-aminopropan-2-ol (14) under-
went a double hydrogen autotransfer process to form a
mixture of cis/trans-piperazine 15 and the related aromatic
pyrazine 16.44 The careful study of the reaction conditions
permitted the authors to obtain the aliphatic heterocycle 15
as the main product (Scheme 4).
The same reaction using Raney-nickel (ca. 40 mol %) with
different 4-aminobutan-1-ol derivatives under a hydrogen
atmosphere (80-90 atm) at 195 °C gave the corresponding
pyrrolidines with, in general, moderate yields (50-92%).45
Raney-nickel is a typical catalyst for the hydrogenation
of different functionalities containing nitrogen atoms. Inter-
estingly, N-alkylated amines could be formed also during
the hydrogenation of nitriles. Thus, in the hydrogenation of
different arylalkylnitriles using nickel at 115-125 °C under
hydrogen atmosphere and using cyclohexanol as solvent, two
compounds were isolated; one of them was the expected
arylalkylamine, and the other one was the corresponding
N-cyclohexyl arylalkylamine with yields ranging from 5 to
45% for the last compound.46 Although the authors did not
give any clear explanation for the formation of the second
product, it looks probable that, after the standard hydrogena-
tion to the primary amine, the alcoholic solvent suffered an
oxidation process to give a nickel hydride intermediate. Then,
the condensation of the formed primary amine with cyclo-
hexanone gave the corresponding imine, which was finally
reduced by the formed nickel hydride to the secondary amine
byproduct. Although the last reduction could be performed
partially by the hydrogen present in the reaction media, the
hydrogen autotransfer process was also acting, as was
mentioned previously. The use of a primary alcohol as
solvent gave similar results.
Using a similar strategy for the alkylation of the in situ
formed amines, different thioamides could be transformed
into the corresponding N-ethyl amines derivatives by using
a large excess of Raney-nickel as catalyst under a hydrogen
atmosphere at reflux of ethanol after a long period of time.47
In a similar way, hydrazobenzene and azoxybenzene could
be converted into the corresponding N-ethylaniline (3b) by
using Raney-nickel (500% in weight with respect to the
starting reagent) under hydrogen atmosphere in refluxing
ethanol with 43% and 36% yield, respectively.48
With this idea in mind, recently different N-sulfinylamides
17 have been transformed into the corresponding N-alkylated
amines 18 using a large excess of Raney-nickel, through a
tandem process involving initially a desulfinylation step to
liberate the primary amine followed by the hydrogen
autotransfer process with the alcohols, which are at the same
time the solvent and the source of electrophile (Table 4).
Yields were in general good, although the reaction times for
primary alcohols were in the range of hours whereas those
for the secondary ones were days. The reaction gave similar
results for aromatic or aliphatic derivatives, with the presence
1616 Chemical Reviews, 2010, Vol. 110, No. 3 Guillena et al.

Table 4. Desulfinylation-Alkylation of N-Sulfinylamides with Table 5. N-Alkylation of Aliphatic Amines with Secondary
Alcohols Using Raney-Nickel Alcohols Using Copper-Chromite Catalyst

entry R1 R2 R3 no. yield (%) entry R1 R2 R3 R4 no. yield (%)

1 Ph Me H 18a 81 1 CH3(CH2)4 H Me H 21a 39
2 4-MeOC4H4 Me H 18b 83 2 CH3(CH2)4 H Me Me 21b 43
3 PhCH2 Me H 18c 0a 3 CH3(CH2)4 H Et H 21c 15
4 Ph(CH2)2 Me H 18d 89 4 (CH2)5 Me H 21d 31
5 Ph(CH2)2 Me Me 18e 76 5 (CH2)5 Me Me 21e 46
6 PhCHMe Me H 18f 78 6 (CH2)5 (CH2)5 21f 59
7 PhCHMe Me Me 18g 80 7 PhCHMe Me Me H 21g 67
a
Decomposition products. 8 BuiCHMe H Et H 21h 61

Scheme 5. N-Alkylation Process by an Indirect aza-Wittig Scheme 6. Transalkylation of Triethylamine with Primary
Reaction of N-(Triphenylphosphoranylidine)aniline with Alcohols Using Copper-Chromite Catalyst
Primary Alcohols Using Nickel Nanoparticles

BaCr2O4 as catalyst, several aliphatic amines 20 have been

of substituents on the aromatic ones not seeming to have a
detrimental effect.49 The same protocol has been carried out
with the related simple primary amines, obtaining similar
results. The main drawback of the reaction was the large
amount of catalyst required, although it could be recovered
by filtration without losing its efficiency, at least after four
cycles, but with longer reaction times being needed.
The nanoparticles of nickel obtained in the reduction of
nickel(II) chloride with a large excess of lithium powder and
a substoichiometric amount of 4,4′-di-tert-butylbiphenyl were
used as catalyst for the N-alkylation process through an
indirect aza-Wittig reaction between N-(triphenylphosphor-
anylidine)aniline (19) and different primary alcohols, produc-
ing moderate yields (Scheme 5). The obtained results seemed
to be constant and independent of the primary alcohol used.
So, benzyl alcohol, linear alkyl alcohols, and the related
branched ones reached practically the same chemical yield.50
2.1.3. Derived from Copper
Simple copper or its oxide have been used as catalysts in
the preparation of pyrazine from 2-aminoethanol, similar to
that depicted in Scheme 4, by a hydrogen autotransfer process
in a continuous flow reactor. However, the achieved yield
for the aromatic compound was very low (6%), even at
temperatures as high as 300 °C. After a few hours, the copper
catalyst lost its activity, attributable, probably, to its reduc-
tion.51
In order to stabilize copper catalysts, other complexes have
been tested. Thus, one of them is the so-called copper-chromite
catalyst (CuCr2O4-BaCr2O4), in which the association of
copper and chromium increases the metal surface area, the
hydrogen adsorption and storage, and the acidity sites. This
catalyst can be easily prepared by mixing barium nitrate,
copper nitrate, and ammonium chromate, which gives a
complex mixture of different chromates. The ignition of this
mixture at 350-450 °C yields the copper-chromite catalyst,
in which the initial amount of reagents could be modified in
order to obtain the desired aspect.52 So, using CuCr2O4–
converted to the corresponding secondary or tertiary amines
21 with modest yields using an equimolecular amount of a
primary or secondary alcohol 9 by heating under a hydrogen
atmosphere (Table 5).53
The aforementioned catalyst has been used in the tran-
salkylation of triethylamine (22a) with a low amount of a
primary alcohol 2, such as octanol and dodecanol (Scheme
6). The reaction was performed under a high hydrogen
pressure and only took place at 250 °C. At lower tempera-
tures, the reaction did not take place. In all cases tested, the
reaction gave the amine 23 as the main product with a small
amount of the double transalkylated amine 24.54 Although
the whole mechanism is not clear, the reaction might start
with the dehydrogenation of starting amine 22a to give the
corresponding iminium derivative, which liberates diethy-
lamine (and acetaldehyde); the same process will be dis-
cussed in section 3. Then, the reaction pathway goes probably
through the known hydrogen autotransfer process, yielding
the amine 23. A second similar sequence, but starting from
the tertiary amine 23, might explain the presence of the
byproduct 24. This protocol has been used in the transalky-
lation of triethylamine (22a) with mixtures of commercially
available long-chain surfactant alcohols at 250-300 °C and
under 19 atm of hydrogen, obtaining the corresponding
amines 23 with good yields (74-89%).55 It should be pointed
out that these amines could be used as efficient surfactants.
The CuCr2O4-BaCr2O4 catalyst has also been successfully
used in the preparation of cyclohexylamine or aniline by
reaction of a 9:1 mixture of cyclohexanol/cyclohexanone with
ammonia (5) in the presence of hydrogen.56 Cyclohexylamine
was obtained in 88% yield when the reaction temperature
was below 240 °C. However, the main product was aniline
(1a, 85%) at 250 °C.
Supported copper and chromium catalysts have also been
used in a tandem process57 involving first the reduction of
dodecanonitrile in methanol and second the double methy-
lation of dodecylamine to give N,N-dimethyldodecylamine.58
The reaction was performed in a stainless steel tubular flow
reactor at 250 °C and 50 atm of hydrogen. Whereas the
reaction using copper supported on alumina gave only the
Hydrogen Autotransfer in the N-Alkylation of Amines
Table 6. N-Alkylation of Dimethylamine with Primary Alcohols
Using Supported Copper-Chromium Catalyst
entry catalyst R feed 25/2/H2 no. yield (%)
1 Cu-Cr on SiO2 CH3(CH2)6 8.3:6:6 26a 97
2 Cu-Cr on SiO2 CH3(CH2)10 5.6:6:6 26b 97
3 Cu-Cr on SiO2 CH3(CH2)14 4.5:6:6 26c 96
4 CuO on γ-Al2O3 Ph 43:26:12 26d 86
5 CuO on γ-Al2O3 PhCH2 38:26:12 26e 87
6 CuO on γ-Al2O3 Ph(CH2)2 33:25:12 26f 90
initial process of reduction (78% of dodecylamine), and using
only chromium on alumina gave a complicated mixture (the
main product was tridodecylamine in 32%), the reaction
using a mixture of copper-chromium gave the desired N,N-
dimethyldodecylamine in good yield (77%). Significant
variations on the activity and selectivity were observed, and
they were attributed to the relative amount of copper used.
Although the mixed catalyst was more stable than the
monometallic one, the modification of the catalyst during
the reaction affected the final results. Thus, the formation
of water seemed to inhibit the catalyst. Therefore, the
presence of hydrogen was necessary in the hydrogen au-
totransfer step since it reduced the strong adsorption of water
and avoided the formation of other reaction products. The
presence of ammonia and amines enhanced the modification
of the catalyst surface by the formation of Cu3N, the
irreversible adsorption of amines, and the deposition of
carbonaceous or nitrogenous compounds, which restricted
the accessibility of reagents to the active centers. The
exceptional good behavior of the 4:5 mixture of copper/
chromium was attributed to the increase of both, the
hydrogen activation, and the acidity resulting from the
selective adsorption of water on Cr(III) species.
Supported catalysts bearing copper and chromium have
been found to be effective and highly selective for the
synthesis of long-chain aliphatic amines, which are widely
used in many fields, such as corrosion inhibitors, epoxy
hardeners, textile additives, etc. Thus, in a continuous fixed-
bed reactor, a heterogeneous catalyst formed with CuO
(25%), Cr2O3 (1%), Na2O (0.1%), SiO2 (70%), and water
(4%) was able to catalyze the alkylation of dimethylamine
(25) with dodecanol with practically quantitative yields
(Table 6, entry 1).59 A further careful study of this process
showed that the optimal parameters depended on the ap-
paratus used, so the temperature was 300 °C and the optimal
ratio of amine 25 and dodecanol ranged from 1.0 to about
1.5 for a stirred autoclave.60 Meanwhile, for a fixed-bed
reactor, the range of temperature was lower (225-235 °C)
and the feed ratio was higher (25/2 ) 5.5). One of the most
important factors studied was the effect of the hydrogen
pressure. Whereas in a fixed-bed reactor a considerable
decrease in the activity and selectivity was observed in a
short period of time in the absence of hydrogen, in the batch
process the presence of hydrogen had a very small influence,
since the hydrogen needed for the hydrogenation of the imine
intermediate was furnished by the dehydrogenation of the
alcohol (hydrogen autotransfer). Although initial studies
seemed to conclude that the main reason for the catalyst
deactivation was the irreversible adsorption of byproduct
coming from the dimerization of alcohol 2, as well as
formation of large aggregates of copper by thermal diffusive
fusion,61 more conscientious kinetic studies showed that the
Chemical Reviews, 2010, Vol. 110, No. 3 1617
Scheme 7. N-Alkylation of Dimethylamine with Diols Using
Supported Copper Catalyst
catalyst deactivation was due to the formation of a bulk or
surface copper nitride.43,62 This nitride could be formed by
reaction of the ammonia, originated from a disproportion
reaction of initial amine (see section 3), with copper at
temperatures as low as 197 °C. The in situ formed aldehyde
could also inhibit the reaction by adsorption. These studies
also showed that the dehydrogenation of the alcohol 2 was
the only rate-determining step, with copper catalyzing this
step as well as the hydrogenation of imine intermediate.
Therefore, hydrogen did not have any influence on the overall
reaction rate and only prevented the catalyst deactivation. It
should be pointed out that, although the disproportion of
dimethylamine (25) is thermodynamically favored, it only
occurred in a small extent due to the inhibition by the alcohol.
Also CuO supported in γ-Al2O3 (60 wt %) could be used
as catalyst to prepare different N,N-dimethylphenylalky-
lamines 26 with high yields following a similar procedure
to that previously described for long-chain amines (Table 6,
entries 4-6).63
A similar catalyst derived from CuO on γ-Al2O3 (63 wt
%) has been used in the preparation of 2-methylaminoal-
kanols 28 by reaction of dimethylamine (25) with the
corresponding aliphatic diol 27 (Scheme 7). The reaction of
ethyleneglycol (27a: n ) 1) with the amine 25 in 1:1 ratio
in a fixed-bed reactor exhibited a maximum at 230 °C.64 The
hydrogen partial pressure did not have any influence on the
conversion, but it was necessary to prevent the catalyst
deactivation. The same reaction was performed with 1,6-
hexanodiol (27b: n ) 5) using a lower loaded catalyst of
CuO on γ-Al2O3 (54 wt %), giving the expected amine 28
with 90% yield at 180 °C.65 As in the previous case, the
conversion of the amine 25 showed no dependence on the
hydrogen partial pressure nor on the partial pressure ratio
of reagent, which indicated that amine 25 is not involved in
the rate-determining step.
Several cyclic amines, most of which are very important
intermediates for some pharmaceuticals, detergents, and
additives, have been synthesized starting from the corre-
sponding aminoalcohol 29 by using copper supported either
on γ-Al2O3 or on MgO in a continuous-flow, fixed-bed
reactor (Table 7).66 Although the same results were obtained
when the reaction was performed under hydrogen or nitrogen,
a deactivation of the catalyst was observed in the absence
of hydrogen. In all cases and using CuO on γ-Al2O3 (63 wt
%), cyclic amines 30 were obtained with high yields and
selectivities (Table 7, entries 1, 4, and 7). Yields were slightly
lower when methanol was used as solvent (Table 7, entries
2, 3, 5, 6, and 8), with the corresponding N-methyl cyclic
amine being obtained as byproduct. This byproduct came
from the N-methylation of cyclic amines 30 through a new
hydrogen autotransfer process with methanol, similar to the
process shown in Scheme 3. The amount of these byproducts
was increased when the reaction was performed on MgO as
support. Whereas this protocol could not be applied to the
synthesis of four-membered ring amines, the reaction using
2-(2-aminobenzyl)ethanol gave the corresponding indoline
with an excellent 94% yield.
1618 Chemical Reviews, 2010, Vol. 110, No. 3 Guillena et al.

Table 7. Cyclization of Aminoalcohols Using Supported Copper purpose.16 The initial copper reagent showed to be very
Catalyst important. For instance, copper stearate showed a higher
activity than copper acetylacetonate or even CuO-CuCr2O4
in the N-alkylation of dimethylamine (25) with dodecylal-
cohol to give the corresponding amine 26b with 40, 0, and
7% yield, respectively.70 However, the equimolecular mixture
entry metal oxide (solvent) n T (°C) no. yield (%)a of Cu(C17H35CO2)2 and Ni(C17H35CO2)2 gave even higher
1 γ-Al2O3 1 200 30a 90 activity than the copper stearate alone (26b, 72%), with
2 γ-Al2O3(MeOH) 1 225 30a 45 (13) nickel stearate not being active. The Cu/Ni ratio had a great
3 MgO (MeOH) 1 225 30a 62 (38) impact on the results, with the best results being obtained
4 γ-Al2O3 2 210 30b 75 with ratios from 5:1 to 8:1. It should be pointed out that,
5 γ-Al2O3(MeOH) 2 225 30b 89 (10)
6 MgO (MeOH) 2 225 30b 62 (37)
under these conditions, small metal particles were observed
7 γ-Al2O3 3 225 30c 95 during the reaction. The role of the stearic acid moiety was
8 MgO (MeOH) 3 225 30c 84 (13) carefully studied, finding that it was an indispensable catalyst
a component, since its function seemed to be to prevent the
In parentheses are the isolated yields of the corresponding N-methyl
cyclic amine. copper sinterization or the coagulation of the Cu-Ni two-
component nanoparticles. A maximum catalytic activity was
Table 8. N-Alkylation of Ethylenediamine with Secondary obtained at 37 mol %. The stearic acid had to be added as
Alcohols Using Supported Copper Catalyst a salt of an alkaline or an alkaline-earth metal ion, since
they were not reduced under the reducing amination condi-
tions (standard electrode potential of about -2.8 V). Beside
the role of the fatty acid residue, the second catalyst
component was optimized, finding that nickel stearate
entry R1 R2 no yield (%) showed higher activity than other derivatives, from metals
1 H H 32a 21 such as Mn, Fe, Co, or Zn. The addition of alkaline or
2 Me H 32b 50 alkaline-earth metal stearates stabilized the catalyst, with
3 Ph H 32c 45 Ba(C17H35CO2)2 giving the best activity and selectivity.
4 (CH2)5 32d 73 Amine 26b could be obtained with an excellent 96% yield
when the stearic salt of Cu/Ni/Ba was used in a 5:1:1 ratio.
It should be pointed out that copper supported on γ-Al2O3 The X-ray diffraction analysis of the reduced catalyst
was known to be effective in the synthesis of nitriles from indicated that Cu metal, CuO, and Ni metal were colloidally
alcohols by reaction with ammonia at 325 °C and atmo- dispersed in the system. A further addition of calcium stearate
spheric pressure with yields ranging from 87 to 96%.67 to the initial catalytic mixture increased the yield up to 99%
Commercially available CuO-ZrO-γ-Al2O3 has been for the amine 26b (5:1:1:1 ratio of Cu/Ni/Ba/Ca).71 The
used as catalyst for the methylation of n-butylamine (amine/ incorporation of calcium in the Cu-Ni core resulted in the
methanol ) 10/39) in a hydrogen atmosphere at 185 °C, possible formation of Ca-Ni-based alloys such as CaNi5 and
rendering the corresponding N-methyl-1-butylamine with a their hydrides, always in a colloidal state, which increased
54% yield.68 This study evidenced the correlation between the hydrogenolysis activity and prevented the transalkylation
the catalytic activity of the reduced catalyst and its ionic of dimethylamine, which occurred with the simple Cu-Ni-
copper content. The obtained results confirmed that the rate- based colloidal catalyst. The transmission electron micro-
determining step of the alkylation of primary amines with scope (TEM) images of the catalysts showed the presence
alcohols was the dehydrogenation of the alcohol on ionic of Cu-Ni nanoparticules with a diameter of 125 nm.72 The
copper active sites of the catalysts. excellent results of this catalyst were demonstrated in a
commercial plant using 5 763 kg (1 equiv) of dodecyl
A catalyst prepared by coprecipitation of Na2CO3,
alcohol, 1.36 equiv of amine 25, 0.29 equiv of hydrogen,
Cu(NO3)2, Zn(NO3)2, and Al(NO3)2 and subsequent calcina-
and only 1 000 ppm of the catalyst (5:1:1 ratio of copper,
tion has been used in the alkylation of ethylenediamine (31)
nickel, and barium stearates) at 210 °C during 4 h, obtaining
with different alcohols 9 (Table 8). The reaction was
the tertiary amine 26b in 90%.
performed in a stainless-steel autoclave with an alcohol/amine
ratio of 5. The results in the selective monoalkylation of the The generation of small amounts of carbon monoxide by
amine 31 were accountably lower for methanol than for other the decarbonylation of the in situ generated dodecanal was
alcohols, suggesting that the methanol dehydrogenation step a side reaction in this process, which deactivated the Cu-Ni
is the rate-determining step. The use of ZnO-Al2O3 as catalyst.73 The generation of CO was a serious problem when
catalyst did not render the alkylated amine under these the amine supply was insufficient during the catalytic
reaction conditions (for the use of alumina as catalyst, see activation. Therefore, a closed system using a CO absorber,
section 2.1.1), showing that the activity was due to copper which was composed of CuCl2, CuCl, ethyleneglycol, a 50%
species. Nevertheless, the synergy effect between different aqueous solution of amine 25, and liquid amine 25, was
copper and zinc species was observed since the activity of designed to prevent the aforementioned catalyst deactivation.
CuO-ZnO-Al2O3 catalyst was much higher than the related However, the use of the CO absorber is not economical, and
CuO-Al2O3 one.69 therefore, a CO-resistant catalyst was prepared by the simple
As was pointed out previously, tertiary long-chain addition of P(OPh)3 (0.16 equiv) to the standard Cu-Ni-Ba
aliphatic amines are very important industrial intermedi- stearate catalyst.
ates, and for that reason, different catalysts have been Although the catalyst is a colloidal system, it could not
tested in their preparation. So, a colloidal catalyst bearing be separated by simple filtration after the amination reaction
copper, nickel, and barium was initially developed for this because of its small particle size (1.5 nm), but its recovery
Hydrogen Autotransfer in the N-Alkylation of Amines
could be performed by distillation.74 The distillation residue
could be reused four times with only a slight decrease on
both the yield and the initial reaction rate after the third use.
The activity of the recovered catalyst decreased drastically
when the used catalyst did not have barium stearate, and
this fact was attributed to the coagulation of the Cu-Ni based
colloid.
When the catalytic activity of colloidal Cu-Ni-Ba
catalyst was compared with others such as CuO-NiO-SiO2
and Raney-nickel, the colloidal system showed seven times
higher catalytic activity than the solid one and more
selectivity than Raney-nickel.75 This effect was even higher
when it was compared to the related Cu-Ni-Ba-Ca
catalyst.
The aforementioned three-component colloidal catalyst has
also been used in the alkylation of dimethylamine (25) with
different diols 27. For instance, the reaction with 1,6-
hexanodiol gave the corresponding N,N,N′,N′-tetramethyl-
hexanediamine with 85% yield. Other different alcohols such
as mixtures of commercial long-chain oxoalcohols were also
used as the source of the electrophile, giving the correspond-
ing N,N-dimethylamines with good results (78-89%).76
2.1.4. Derived from Platinum
Platinum derived catalysts have been shown to be effective
to promote the N-alkylation of amines by a hydrogen
autotransfer process. Thus, a vapor phase containing cyclo-
hexanol and ammonia (5) over platinum supported on silica
(ca. 5% wt loading) was transformed into cyclohexylamine
and aniline in a continuous-flow reactor at atmospheric
pressure with yields of 30 and 58% at 267 °C, respectively
(compare these results with those presented in section 2.1.3
using a copper-chromite catalyst, which is more selective).77
Other platinum group metals also catalyzed the reaction, with
the difference between platinum and rhodium being that the
activity of the least active metal rhodium is within 1 order
of magnitude lower. The yields of the reaction were
dependent on the temperature as well as the contact time,
increasing as those parameters increased. However, the
selectivity changed for aniline (1a) at higher temperatures.
It was found in this process that the catalytic activity per
surface metal atom and product selectivity did not change
so much with the platinum size. A deactivation of the catalyst
was detected as the yield of each product decreased with
the time, and it seems to occur at the surface, so it was found
that the weight of the catalyst was increased by 10% after
performing the reaction. It seems to be clear that the process
started with the dehydrogenation of cyclohexanol to give
cyclohexanone, which is the rate-determining step, although
alternative mechanisms have also been suggested.78
Benzylic amine derivatives have been alkylated by elec-
trolysis in alcoholic solutions using a platinum black-coated
anode and a coiled platinum anode.79 In this process, ethanol
or methanol, which contained LiClO4 or LiNO3 as electrolyte,
were used as solvents. After adding the corresponding amine,
the galvanostatically or potentiostatically electrolysis at room
temperature gave the expected alkylated amines with yields
ranging from 27 to 91%. The electrolysis of alcohols in the
absence of amines gave directly the corresponding aldehydes
and hydrogen via an anodic oxidation of the alcohols and
cathodic reduction of protons. In the presence of an amine,
the condensation proceeded feasibly to give the correspond-
ing imine,80 which is hydrogenated in the platinum black
surface of the anode under an open circuit.
Chemical Reviews, 2010, Vol. 110, No. 3 1619
Scheme 8. Triple N-Alkylation of Ammonia with Primary
Alcohols Using Platinum Supported Catalyst
Ammonia (5) could also be converted into the correspond-
ing tertiary amines 22 at room temperature by means of a
photoirradiation procedure by a 400 W high-pressure mercury
lamp, under catalysis with Pt-TiO2, prepared in turn by
simply mixing Pt black with anatase powder. The reaction
of ammonia with an extraordinary high excess of a primary
alcohol (2, 41 000 mol %) such as methanol, ethanol, or
1-butanol gave the corresponding amines 22 with moderate
yields (Scheme 8). The photoirradiation in the absence of
ammonia rendered the corresponding alcohol, which indi-
cated that the whole process seemed to go through a
hydrogen autotransfer process. When this reaction was
performed in the presence of a small amount of water (4-8
mol %), the yield of the tertiary amine 22 decreased at the
same extent as the increase of the yield of the secondary
amine.81 This protocol could also be extended to the use of
simple primary or secondary amines instead of ammonia (5),
giving in these cases the corresponding nonsymmetrical
tertiary amines with slightly better results (28-93%).82
2.1.5. Others
Besides the already presented catalysts, other transition-
metal derivatives have been used in the N-alkylation of
amines and related compounds. For instance very recently,
unmodified commercial magnetite has been shown as an
excellent catalyst for the N-alkylation of different aromatic
amines using benzylic alcohols (6) as the source of the
electrophile.83 The surface of Fe3O4 (111) is terminated by
a hexagonal oxygen layer covered by one-quarter monolayer
of iron atoms, with these metallic centers being responsible
for the catalysis action.84 After optimizing the reaction
conditions for aniline (1a) with benzyl alcohol (6a, 4 equiv),
the expected N-benzylaniline (7a) was obtained with a good
88% yield by using Fe3O4 (20 mol %) in the presence of
potassium tert-butoxide (2 equiv) at 90 °C in dioxane after
seven days. The results were independent of the electron-
withdrawing or -donor character of the substituents at the
four position of alcohols 6, with results ranging from 80 to
83%. However, when different substituted anilines were used
as nucleophiles, the results were dependent on the substitu-
tion, with the lowest nucleophilic 3-chloroaniline giving a
surprisingly better result (99% yield) than the highest
nucleophilic 4-methoxyaniline (42%). Therefore, this pro-
tocol was applied to different electron-poor heteroaromatic
amine derivatives 33 with unbeatable results (Table 9). As
was expected, the use of an electron-rich heteroaromatic
amine such as 2-methylthiazol-2-ylamine gave a modest 33%
yield in its reaction with benzyl alcohol. The selectivity was
notable since the competitive reactions between a high
nucleophilic aliphatic amine and a low nucleophilic amine
such as aniline with benzyl alcohol only rendered N-
benzylaniline (7a), with the selectivity between aliphatic and
benzylic alcohols being exclusively favored to the benzylic
derivatives.
The magnetite catalyst could be simply and easily recov-
ered by using an external magnet and reused for at least eight
cycles of reaction without a detrimental effect on the initial
1620 Chemical Reviews, 2010, Vol. 110, No. 3
Table 9. N-Alkylation of Heteroamoatic Amines with Benzylic
Alcohols Using Magnetite
entry Z Y X R no. yield (%)
1 CH CH N H 34a 96
2 CH CH N Cl 34b >99
3 CH CH N Me 34c >99
4 CH CH N MeO 34d >99
5 N CH CH H 34e >99
6 CH N N H 34f >99
Scheme 9. N-Alkylation of Sulfonamides with Benzylic
Alcohols Using Ruthenium-Impregnated Magnetite As
Catalyst
results. In order to study the possible degradation of the
catalysts, the Brunauer-Emmett-Teller (BET) surface and
TEM images have been determined before and after eight
reaction cycles, with the results being almost similar and
showing that there was not a significant sinterization process.
Magnetite impregnated85 with ruthenium hydroxide
[Ru(OH)x-Fe3O4] has been used in the alkylation of sul-
fonamides 35. The reaction using an excess of benzylic
alcohol (6, 400 mol %) and substoichiometric amounts of
Ru(OH)x-Fe3O4 (0.4 mol %) and K2CO3 (2 mol %) at 150
°C without solvent gave the expected derivatives 36 in
general with excellent results (Scheme 9). From the obtained
results, it could be concluded that the reaction worked well
not only for aromatic and heteroaromatic sulfonamides but
also for aliphatic sulfonamide derivatives. Other benzylic
derivatives different from 4-substituted ones could be suc-
cessfully used, as well as the related heteroarylmethanol
derivatives. As in the previous case, the catalyst could be
separated from the reaction media by using an external
magnetic bar and directly reused five times.86 The catalyst
was characterized by different techniques including X-ray,
X-ray photoelectron spectroscopy (XPS), BET area, and
TEM images. The ruthenium particles were of a small size
(1.5-5 nm), whereas the magnetite support showed particles
of around 100 nm. The BET area measures showed no
changes before and after the reaction cycles. The XPS studies
showed equal binding energies of 461.7 and 461.9 eV for
the Ru 3p3/2 peaks in the new and reused catalyst, respec-
tively. All these results suggested that the catalyst contained
Ru(0) nanoparticles immobilized on the surface of the
crystalline magnetite, which remain stable and highly
dispersed during the hydrogen autotransfer process. The
labeling experiments, together with other competitive ones,
showed a primary kinetic isotope effect for the dehydroge-
nation step, indicating that the benzylic C-H bond breaking
was the rate-determining step.
The use of palladium metal (22 mol %) was less effective
for the reaction of 2-phenylethylamine with benzyl alcohol
Guillena et al.
Scheme 10. N-Alkylation of Amines with Secondary
Alcohols Using Palladium Catalyst
(6a; alcohol/amine ratio ) 1.2) to give the expected alkylated
amine (25% yield).87 However, a similar reaction but using
a small excess of amine 20 (1.1 equiv) and 1-phenylethanol
(37) gave the expected amines 38 with good results (Scheme
10).88 When the reaction was performed using other alcohols
such as benzyl or allyl alcohol, the results were accountably
lower (12-34% yields).
Other metal oxides, such as tungsten oxide89 or thorium
oxide,90 have also been used as catalysts for the alkylation
of ammonia (5) with benzylic, aliphatic, or cyclic alcohols
at 330 °C to give the corresponding primary amines, although
the corresponding secondary and tertiary amines could be
detected as byproducts. A similar protocol, but using aliphatic
primary amines, gave the expected unsymmetrical secondary
amines. Finally, it should be pointed out that the reaction of
ammonia with alcohols at higher temperatures gave the
corresponding symmetrical secondary amine as the main
product.
2.2. Homogeneous Catalysts
An argument often made in favor of the homogeneous
catalysts is that these types of catalysts frequently allow
reactions to occur at a lower temperature, with higher
selectivity, and more easily than heterogeneous catalysts. In
fact, as was previously mentioned, the first example of
N-alkylation of aniline by alcohols was made in a homoge-
neous solution in the absence of transition-metal catalyst in
1901.22 However, the homogeneous version using transition-
metal catalysts was established in 1981, with the simulta-
neous introduction of different rhenium91 and ruthenium92a
catalysts, as will be presented in the corresponding sections.
2.2.1. Without Transition-Metal Catalysts
Reactions that are always performed under catalytic
conditions can often be also performed in the absence of a
catalyst, by just varying the reaction conditions, usually using
a higher temperature, pressure, amount of reagents, reaction
times, or using solvent-free conditions. This is also the case
of the N-alkylation of amines using alcohols as the source
of the electrophiles that could be made in the absence of a
catalyst using just a strong base at higher temperature. For
instance, when aniline (1a, 1.2 equiv) was heated at 250-300
°C with sodium 1-pentoxide for several hours, it gave the
expected N-pentylaniline (3e, 40%).22 Instead of sodium
alkoxides, the reaction could also be performed using
aluminum alkoxides giving the corresponding alkylated anilines
3 with good results.93 For instance, the reaction with aniline
(1a) at 275 °C using stoichiometric amounts of aluminum
triethoxide gave after 4 h the expected amine 3b with a 94%
yield.
In a similar way, several aniline derivatives 1 could be
alkylated with benzyl alcohol (6a, 140 mol %) to give the
corresponding derivatives 7 with high yields (82-99%),
using only potassium hydroxide (40 mol %) as base and
heating the mixture at 250 °C during several hours (14-110
h), with continuous distilling of in situ formed water.94 For
instance, this protocol has been extended to the synthesis of
Hydrogen Autotransfer in the N-Alkylation of Amines Chemical Reviews, 2010, Vol. 110, No. 3 1621

Table 10. N-Alkylation of Aromatic Amines with 2.2.2. Derived from Ruthenium
Pyridylmethanol Derivatives Using Potassium Hydroxyde as
Catalyst Ruthenium complexes have been extensively used in
organic synthesis as the prototype for the transition-metal
catalysis.102 Therefore, several ruthenium derived species
have been reported as excellent catalysts in the homogeneous
hydrogen autotransfer process for the N-alkylation of amines,
with RuCl2(PPh3)3 being probably the most widely used. This
ruthenium complex was initially used in the alkylation of
entry R Y Z no. yield (%) aniline derivatives 1 with primary alcohols 2 (Table 11) in
1 H N CH 40a 63
a 1:5-12 amine/alcohol ratio, as well as with secondary ones,
2 Me N CH 40b 50 to yield the corresponding N,N-dialkylated amines as the
3 MeO N CH 40c 71 main products in the absence of solvent.92 Whereas very good
4 EtO N CH 40d 68 yields were obtained for the dialkylated products 44, the
5 H2N N CH 40e 73 related reaction with secondary alcohols failed, giving only
6 EtO CH N 40f 60
the corresponding monoalkylated amine with low yields
(25-28%).
Scheme 11. N-Alkylation of Benzamide with Primary The introduction of electron-donating groups at the four
Alcohols without Catalyst
position of the aromatic aniline derivative 1 enhanced both
the reaction rate and the yields. Meanwhile, when the
substitution was located at the two position, the monoalky-
lated product was the main isolated compound. Importantly,
the control on the ratio of aniline/alcohol regulated the
proportion of the obtained compounds 43 and 44. So, when
2-benzylamino pyridine (34a) with an excellent 99% yield.95 the above ratio was reduced to one, the monoalkylated
It should be pointed out that the reaction could be accelerated compound 43 was predominantly formed in good yields (see,
by the addition of a small amount of benzaldehyde or by for instance, entries 6 and 7 in Table 11). A zero-order
the use of a higher base concentration. Other amino dependence on the aniline concentration was obtained from
heterocyclic compounds, such as 2-aminobenzimidazol de- kinetic studies, as well as a first-order dependence on the
rivatives, could be successfully benzylated using this protocol alcohol and a first-order dependence on the catalyst. The
with nearly quantitative yields.96 Alternatively, the reaction activation energy from the Arrhenius plot was 73.6 kJ mol-1,
of compounds 1a and 6a could be performed in the presence with the enthalpy and entropy being 70.2 kJ mol-1 and -123
of potassium tert-butoxide at only 90 °C but increasing the J mol-1 K-1, respectively.
reaction time to 10 days (7a, 67%).83 The use of the same catalyst was further extended to the
alkylation of more nucleophilic aliphatic primary amines with
Pyridylmethanol compounds 39 could also be used as the methanol (2a). Thus, the reaction of primary amines 45 with
source of the electrophile in the alkylation of different aniline a large excess of methanol (2a, 3 000 mol %) at 180 °C
derivatives 1 (Table 10). The reaction of equimolecular catalyzed by 3 mol % of RuCl2(PPh3)3 gave the correspond-
amounts of these compounds in the presence of substoichio- ing dimethylamine derivative 26 with good yields (26-94%).
metric amounts of a base gave the expected amines 40 with However, when the amount of methanol was decreased to
moderated yields.97 A further modified protocol using 200 500 mol %, the main product was the monomethyl tertiary
mol % of amine, nitrobenzene (100 mol %) as the initial amine 46 (Scheme 12), with a small amount of the corre-
oxidant, and xylene as solvent did not change the previous sponding amine 26 as byproduct (1-19%).103 The formation
results.98 of compounds 46 could be easily explained by taking into
The selective monoalkylation of aniline derivatives 1 could account that the initial amine 45 could be the source of the
also be carried out by using aliphatic alcohols,99 although, electrophile in a transalkylation amine process. The oxidation
in this case, the amount of alcohol had to be increased up to
200 mol % and metallic sodium (68 mol %) was used to Table 11. Dialkylation of Aniline with Primary Alcohols Using
generate the corresponding base. The reaction performed in RuCl2(PPh3)3
an autoclave at 270-300 °C during 4-6 h gave the expected
compounds 3 generally with modest results (20-72%).
More recently, supercritical methanol conditions (239.4
°C, 79.9 atm) have been used in the methylation of aniline
1a to give after 2 h a mixture of compounds 3a (49%) and
4a (1%).100 About 70% of hydrogen bonds among methanol
molecules are broken down under these conditions, producing entry R1 R2 no. yield (%) no. yield (%)
dimers and monomers and, therefore, increasing the reactivity
1 H Me 43a 13 44a 74
of the alcohol. 2 H Et 43b 10 44b 88
Finally, it should be pointed out that the reaction of 3 H Prn 43c 15 44c 79
benzamide (41a) with primary alcohols 2 in a sealed tube 4 MeO Prn 43d 7 44d 91
5 Me Prn 43e 15 44e 85
gave the corresponding N-alkylated benzamides 42 in general 6a H Prn 43c 79 44c 6
with moderate yields (Scheme 11), together with other 7a MeO Prn 43d 99 44d
byproducts such as benzoic acid, and the corresponding alkyl a
1:1 ratio of compounds 1/2.
benzoate.101
1622 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 12. Autoalkylation and N-Methylation of Aliphatic
Amines Using RuCl2(PPh3)3 Catalyst
Scheme 13. N-Alkylation of Symmetrical Secondary Amines
with Long-Chain Alcohols Using RuCl2(PPh3)3 Catalyst
Table 12. N-Dialkylation of Aniline with Primary Alcohols
Using RuCl2(PPh3)
entry R no. yield (%)
1 Bun 51a 63
2 But 51b 58
3 HO 51c 56
of the amine 45 catalyzed by ruthenium species gave the
corresponding imine, which suffered the addition of the initial
amine 45 to form a new N-alkyl imine derivative, with
liberation of ammonia, and the final hydrogenation of this
N-alkyl imine derivative by the ruthenium hydride species
gave the corresponding symmetrical dialkylamine (see sec-
tion 3.2.1). The final step was the expected methylation
through a hydrogen autotransfer process.
Symmetrical secondary amines 47 could also be alkylated
with aliphatic long-chain primary alcohols 48 (n ) 9, 13,
15, 17) to give the corresponding amines 49 with moderate
to good results in the absence of solvent (Scheme 13). It
should be pointed out that, whereas aromatic amines gave
modest results, aliphatic amines gave poor results (yields
lower than 28%), with dimethylamine reaching the highest
yields.104 This catalytic reaction showed an induction period
of about 30 min before significant amounts of amine 49 were
formed. This induction period could be reduced by the
addition of 4 (mol of PPh3)/(mol of RuCl2(PPh3)3). This
acceleration effect was attributed to the displacement of the
formed aldehydes by the phosphine from the ruthenium
catalytic species, thus facilitating the condensation with the
amine and accelerating the hydrogenation of the in situ
formed iminium derivative to the tertiary amine 49. When
the same reaction procedure was applied to secondary
alcohols 9, the reaction failed.
The double alkylation of primary amines can be also
performed by using the corresponding diol. For instance, 1,1-
ferrocenedimethanol (50) was reacted with different anilines
1 using N-methylpyrrolidin-2-one as solvent to give the
corresponding ferrocenylamine derivatives 51 (Table 12).
The results seemed to be independent of the aniline substitu-
tion, keeping the same level of yields when aliphatic or
benzylic amines were used instead of compounds 1.105
Guillena et al.
Scheme 14. N-Alkylation of Amines with Ethanediol Using
RuCl2(PPh3)3 Catalyst
Scheme 15. Cyclization of Aromatic Amines with Diols
Using RuCl2(PPh3)3 Catalyst
Scheme 16. Cyclization of Diols with Anilines Using
RuCl2(PPh3)3 Catalyst: Synthesis of N-Substituted
Perhydroazepines
On the contrary, the reaction of secondary amines 20 with
an excess of ethanediol 27a (740 mol %) in the absence of
solvent gave selectively the aminoethanol derivatives 52 with
good yields (Scheme 14). However, when the catalyst was
changed by RuCl3 · nH2O, the main product was the corre-
sponding ethylenediamine. Surprisingly, the same reaction
catalyzed by RuCl2(PPh3)3 using primary amines (R2 ) H
in Scheme 2) gave a mixture of three products including the
aminoethanol 52, the related ethylenediamine, and N,N′-
dialkylpiperazine, with the ratio being strongly influenced
by the alkyl moiety and the presence of an extra amount of
PPh3.106
When 1,5-diol derivatives 54 (150 mol %) were reacted
with aromatic amines 53 in dioxane, the corresponding
N-substituted piperidines 55 were obtained in moderate to
good yields (Scheme 15), with the results being independent
of the electron character of the substituents on the aromatic
ring. It should be pointed out that the yields were lower when
aliphatic amines were used instead of aromatic ones.
However, in this case, the use of RuCl3 · nH2O combined with
tributylphosphine increased the results to the initial level.107
This strategy has been applied to the synthesis of biologically
active substances such as 4-[2-(benzhydryloxy)ethyl]mor-
pholine (from diethyleneglycol X ) O in Scheme 15, 47%),
which is an antihistaminic agent, or N-methyl-N′-benzhy-
drylpiperazine (from diethanolamine X ) NH in Scheme
15, 19%), which is a drug preventing travel sickness.
Perhydroazepines 56 could be prepared by reaction
between aniline derivatives 1 and 1,6-hexanodiol (27b, 150
mol %) in dioxane at 180 °C (Scheme 16). The RuCl2(PPh3)3
complex was used as catalyst, although it could be prepared
in situ by adding RuCl3 · nH2O and PPh3. Yields were
generally good, although for 2-substituted aniline derivatives,
benzyl amine or lauryamine, the corresponding heterocyclic
compound was obtained with lower results (12-29%).108
Some of the aforementioned heterocyclic compounds could
also be obtained by using the corresponding N-substituted
amino alcohols 57 as the appropriate starting material.109
Hydrogen Autotransfer in the N-Alkylation of Amines Chemical Reviews, 2010, Vol. 110, No. 3 1623

Table 13. Cyclization of Amino Alcohols Using RuCl2(PPh3)3 Table 15. Cyclization of Aminophenylethanol Derivatives
Catalyzed by RuCl2(PPh3)3: Synthesis of Indoles

entry R n no. yield (%)

1 Pri 2 58a 70
2 Ph 2 58b 88 entry no. R yield (%)
3 Bun 3 58c 71 1 8a H >99
4 Ph 3 58d 90 2 8b 2-Ph 99
5 Bun 4 58e 71 3 8c 3-Me 79
4 8d 4-Cl 92
Scheme 17. Cyclization of Alkynildiols with Primary Amines 5 8e 5-MeO 94
Using RuCl2(PPh3)3 Catalyst: Synthesis of N-Substituted 6 8f 6-Cl 82
7 8g 6-Me 80
Pyrroles

Table 16. N-Alkylation of Carboxamides with Primary Alcohols

Using RuCl2(PPh3)3

Table 14. Cyclization of 1,2-Diols with Aniline Derivatives

Catalyzed by RuCl2(PPh3)3: Synthesis of N-Substituted Indoles
entry R1 R2 no. yield (%)
1 Me Me 64a 43
2 Me CH3(CH2)6 64b 24
3 Ph CH3(CH2)6 64c 74
4 PhCH2 Me 64d 15
5 Prn CH3(CH2)6 64e 28
entry R1 R2 R3 R4 no. yield (%)
1 Me H H H 10d 51 derivative by a double autotransfer hydrogen alkylation
2 Et H H H 10e 34
3 Me H Me Me 10f 58 process, (b) monodehydrogenation to give the corresponding
4 H H Me Me 10g 46 imine (iminium) derivative, (c) ortho-metalation of amine
5 H 2-Cl Me Me 10h 72 moiety, (d) intramolecular addition to the imine (iminium)
6 H 4-Cl Me Me 10i 89 group, and (e) β-elimination of starting amine 61. When the
7 H 2-Me Me Me 10j 50 reaction was performed using 1,3-propanodiol derivatives,
8 H 4-Me Me Me 10k 80
9 Me H H Me 10l 50a the corresponding substituted quinolines were obtained, with
10 H H Ph H 10m 43 this approach being an alternative to the classical Friedländer
11 H 4-Cl (CH2)4 10n 65 synthesis.112
a
1:1 mixture of possible regioisomers. Substituted indoles 8 were easily prepared by reaction of
the corresponding substituted aminophenylethanol derivatives
63 with excellent results (Table 15). Yields seemed to be
Their reactions in dioxane gave the corresponding azacy- independent of the nature of substitution, the electronic
cloalkanes 58 with in general good yields (Table 13). character of the group, or the position in the indol.113 The
Heteroaromatic compounds could also be prepared fol- whole transformation implied first an intramolecular alky-
lowing the former strategy using RuCl2(PPh3)3 as catalyst. lation process through a hydrogen autotransfer process to
Thus, the reaction of aliphatic primary amines 45 with give the corresponding indoline, which suffered a further final
4-butyne-1,4-diol (59, 150 mol %) in dioxane at 150 °C gave dehydrogenation process to yield the aromatic compound.
the corresponding pyrroles 60 with moderate yields (Scheme Alternatively, the related nitroaromatic derivative could also
17). The catalytic activity was slightly affected by the be used as starting material, with an initial hydrogenation
addition of a phosphorus ligand, with the addition of step being necessary in these cases.
bis(diphenylphosphinoethane) as a bidentate ligand or tri- The RuCl2(PPh3)3 complex has also been used in the
cyclohexylphosphine as a bulky ligand suppressing the preparation of other heterocyclic compounds, such as ben-
catalytic activity. However, better results were obtained when zoxazoles and benzimidazoles, using alcohols as electrophiles
butane-1,4-diol (12) was used, obtaining in this case the in the reaction,114 as well as quinoxalines.115 Alternatively,
corresponding pyrrolidine derivative (45-91%).110 RuH2(PPh3)3(CO) and xantphos have been recently proposed
The reaction of substituted anilines 61 (250 mol %) with for the same transformation.116 Imidazo[1,2-a]pyridines were
1,2-diols 62 catalyzed by the RuCl2(PPh3)3 complex gave obtained in the reaction of 2-aminopyridines and 1,2-
the indoles 10 with moderate to good yields (Table 14). The ethanediols.117 However, it should be pointed out that these
results seemed not to be related with the position of reactions are not standard hydrogen autotransfer processes,
substitution as well as its electronic character. Only the although they are closely related.
presence of a substituent on position two of aniline 61 had Finally, it should be highlighted that RuCl2(PPh3)3 has been
a small detrimental impact on the yield. This protocol could used as catalyst in the N-alkylation of simple carboxamides
also be applied to cyclic diols with similar results.111 A (Table 16). The reaction of amides 41 with different primary
mechanistic study showed that the reaction goes through (a) alcohols (320-850 mol %) gave the corresponding amides
the formation of the corresponding N,N′-diamino ethane 64, in general with modest yields. The obtained results were
1624 Chemical Reviews, 2010, Vol. 110, No. 3 Guillena et al.

Scheme 18. N-Alkylation of Aliphatic Primary Amines with Table 18. N-Alkylation of Heteroaromatic Amines with Primary
Primary Alcohols Using RuH2(PPh3)4 Catalyst Alcohols Using Ru(COD)(COT) Catalyst

entry Z Y X R no. yield (%)

Table 17. Cyclization of Triisopropanolammonium Chloride
with Aniline Derivatives Catalyzed by RuH2(PPh3)4: Synthesis of 1 CH CH CH Me 68a 3
Indoles 2 CH CH CH Me 68a 27a
3 CH N CH Me 68b 79
4 N CH CH Me 68c 70
5 CH N N Me 68d 37
6 N CH CH H 68e 67
7 N CH CH Et 68f 82
8 N CH CH Ph 68g 45
a
entry no. R yield (%) Reaction performed in the presence of 12.5 mol % of pyridine.
1 8a H 85
2 8h 5-Cl 31 Scheme 19. Methylation of Symmetrical Secondary Amines
3 8i 5-Me 68 Using RuClCp(PPh3)2
4 8e 5-MeO 80
5 8j 4,7-(MeO)2 21

slightly better when benzamide derivatives were used as

nucleophiles, where even benzyl alcohol could be used as
electrophile and yields were about 50% in all cases. When
different lactams or 3-pyridinecarboxamide were employed
in the aforementioned protocol, the reactions failed. The
protocol could not be extended to R,β-unsaturated carboxa-
mides, since, besides the formation of the alkylation product,
a hydrogenation of the double carbon-carbon bond also took
place.118
Other well-defined ruthenium complexes have also been
used in different hydrogen autotransfer reactions. For in-
stance, RuH2(PPh3)4 is a highly selective catalyst able to
perform the reaction between high nucleophilic aliphatic
primary amines 45 and primary alcohols 2 to render the
corresponding secondary amines 65 with moderate to high
yields (Scheme 18). The reaction was carried out using
equimolecular amounts of reagents in a stainless steel
autoclave and in the absence of solvent.119 The intramolecular
version of this reaction provides an excellent method for the
preparation of cyclic amines. The process was performed
starting from amino alcohols 29, which were transformed
into the cyclic compounds 30 (n ) 0-2) with yields ranging
from 41 to 79%. This protocol has also been used in the
preparation of N-substituted cyclic amines of type 58 by
reaction of either amino alcohols with primary alcohols or
primary amines with diols, with the last approach being the
most convenient since the yields were better. This approach
has been successfully used in the preparation of tetrahy-
droisoquinolines starting from 2-(2-hydroxymethylphenyl)-
ethanol.
The RuH2(PPh3)4 catalyst has proven to be efficient for
the preparation of amides by condensation of amines with
nitriles120 and esters by oxidative condensation with primary
alcohols.121 Very recently, RuH2(PPh3)3(CO) and xantphos
have been proposed as an alternative for the last tansforma-
tion.122
RuH2(PPh3)4 has been used, together with SnCl2 · 2H2O,
in the heteroannulation between an excess of aniline deriva-
tives 66 (1 000 mol %) and triethanolammonium chloride
(67) to give the corresponding 6-substituted indoles 8 (Table
17). The results were affected by the electron character of
the substituent, with electron-withdrawing groups giving
lower yields. Moreover, related 2- and 3-substituted aniline
derivatives provided similar or better results than the
corresponding 4-substituted ones.123 When the reaction was
performed using substituted triisopropanolammonium chlo-
ride as the initial source of nucleophile (secondary alcohol),
3-methylindoles were selectively formed. Other catalysts,
such as RuCl3 · nH2O with either PPh3 or bis(diphenylphos-
phino)methane, gave lower yields. Although the reaction
mechanism is not fully understood, the reaction seemed to
proceed through (a) a N-alkylation of aniline by a hydrogen
autotransfer process, (b) oxidation of the trialkylamine moiety
to the corresponding iminium, (c) ortho-metalation of the
aniline ring, (d) intramolecular addition to the iminium group,
and (e) β-elimination of diethanolamine (similar mechanism
to that mentioned in Table 14).
The RuH2(PPh3)4 complex has been used in the transfor-
mation of amino alcohols into the corresponding lactams with
moderated yields.124
The well-defined catalyst (η4-1,5-cyclooctadiene)(η6-1,3,5-
cyclooctatriene)ruthenium [Ru(COD)(COT)] was able to
catalyze the N-alkylation of poor electrophilic heteroaromatic
amines with a large excess of primary alcohols 2 (2 100 mol
%).125 This catalyst displayed a high activity in the selective
alkylation of aminopyridine derivatives (Table 18), although
it was ineffective in the alkylation of simple aniline, and
the ethylation took place with a modest yield and only when
pyridine was added as coligand. This fact suggested that the
coordination of the nitrogen atom of the pyridine ring with
the ruthenium atom is essential for the formation of the true
catalytic species, with the possible chelation being intra- or
intermolecular when using aminopyridine derivatives. Dif-
ferent primary alcohols were used with similar results, even
when benzyl alcohol was used as electrophile.
The half-sandwich ruthenium complex RuCl(η5-C5H5)(PPh)2
has been used as catalyst in kinetic studies of the reaction
of high nucleophilic aliphatic secondary amines 47 (as well
as primary ones) with a large excess of methanol (2a, 10 000
mol %) to give the corresponding N-methyl (or N,N-
dimethyl) tertiary amines (Scheme 19). The reaction rates
depended strongly on the nature of the amine substituents,
with the reaction rate increasing as the basicity (nucleophi-
Hydrogen Autotransfer in the N-Alkylation of Amines
Table 19. N-Monoalkylation of Aniline with Primary Alcohols
Using [RuCl2(PPh3)2(MeCN)]BPh4 Catalyst
entry R1 R2 no. yield (%)
1 H H 43f 60
2 H Me 43a 40
3 H Et 43b 28
4 H Prn 43c 19
5 Me Prn 43e 20
licity) of the starting amine increased. For instance, the
reaction rate decreased along the series piperidine > N-
methylpiperazine > morpholine. For primary amines, the
highest basic amine, benzyl amine (with a pKb of 4.67), gave
the highest reaction rate. Furthermore, poor nucleophilic
amines such as aniline (1a) did not react under these reaction
conditions.126 The mechanistic studies indicated apparently
that the reaction proceeded stepwise, with the starting amine
being transformed into the N,N-dialkyl formyl imine deriva-
tive whose concentration remained very low (ca. 5%) and
almost constant throughout the reaction. The further hydro-
genation of this imine gave the product 69. The process was
zero order for the amine, with the presence of free PPh3 or
chloride anions inhibiting the reaction. Measurements of
31
P{1H}-RMN showed that the dissolution of the catalyst in
CD3OD occurred with chloride dissociation and formation
of the corresponding cationic solvent complex. Free PPh3
appeared into the solution when this complex was heated at
80 °C together with the corresponding imine. The Ru–H–
formaldehyde complex intermediate, which is probable the
key species in the catalytic cycle, might be formed from
the initial cationic complex via PPh3 dissociation, which
provides a vacant coordination site in the ruthenium metal
center. This hypothesis was in agreement with the inhibi-
tion observed by the addition of PPh3 or chloride anions
to the reaction media.
Different aniline derivatives 1 were monoalkylated with
a large excess of refluxing primary alcohols (2, > 24 500
mol %) in the presence of a base and substoichiometric
amounts of the cationic ruthenium complex [RuCl2(PPh3)2-
(MeCN)]BPh4 (Table 19). Yields were dependent on the chain
length of the alkyl group of the alcohol, with the methanol
(2a) giving the best results.127 When benzyl alcohol was used
as electrophile, a mixture of mono- and dialkylated products
were obtained, along with the corresponding imine, with
similar yields.
Ruthenium pincer type complexes have been used as
promoter for the reaction reviewed here. For instance, the
N,N′,N-ruthenium complex depicted in Scheme 20 was able
to catalyze the selective reaction of aniline (1a) with different
terminal diols [12 (1,4), 54a (1,5), 27b (1,6), and 70 (1,10)]
to give, in all cases, the corresponding amino alcohols with
moderate to good yields (42-70%).128 The analysis of
reaction mixture with the time showed that the possible imine
intermediates were not produced during the reaction. There-
fore, it was concluded that the first condensation between
the in situ formed aldehydes and the amine to give the
expected imine must be the rate-limiting step, taking place
in the coordination sphere of the ruthenium cationic center.
Chemical Reviews, 2010, Vol. 110, No. 3 1625
Scheme 20. N-Alkylation of Aniline with 1,10-Decanodiol
Catalyzed by a Pincer Ruthenium Complex
Table 20. N-Alkylation of Ammonia with Primary Alcohols
Catalyzed by a Ruthenium Pincer Complex: Synthesis of
Primary Amines
entry no. R yield (%)
1 45a Ph 87
2 45b 4-FC6H4 91
3 45c 4-MeC6H4 83
4 45d 4-MeOC6H4 78
5 45e Bun 61
6 45f MeOCH2 95
7 45g (CH2)5CH 82
The no-formation of the cyclic amines as major products
was explained as a function of the difficulty for the iminium
intermediate formation in the cationic ruthenium species, with
the formation of these byproducts proceeding most probably
by a enamine intermediate on the coordination sphere of the
complex. Ruthenium carbonyl complexes, formed by de-
composition of aldehyde-ruthenium intermediates, were
detected in some reactions, with these carbonyl complexes
being inactive for the hydrogen autotransfer reactions.
A ruthenium acridine-based pincer complex has shown its
efficiency in the selective monoalkylation of ammonia (5)
with primary alcohols (2) at toluene reflux to give the
corresponding primary amines (45) in high yields (Table 20).
Benzyl alcohol derivatives bearing electron-donating groups
reacted faster than those containing electron-withdrawing
groups. Heteroaromatic and aliphatic alcohols could also be
used as initial source of electrophiles under these conditions.
Furthermore, the reaction also proceeded on water, with its
presence enhancing the selectivity toward the formation of
the primary amine. Probably, the hydrogenation of the imine
intermediate is favored in the presence of water. Surprisingly,
the reaction gave good results even with nonsoluble primary
alcohols; meanwhile water-soluble alcohols, such as pyridine-
2-methanol, gave a complex mixture of products in water in
sharp contrast to the excellent result obtained in toluene
(96%).129
Other ruthenium pincer-type complexes have been suc-
cessfully used as catalysts for the formation of either amides
from the reaction of alcohols with amines130 or acetals by
trimerization of alcohols.131
Besides the use of well-defined ruthenium-complex cata-
lysts, other catalysts generated in situ could be employed
for the hydrogen autotransfer processes. The first ones were
1626 Chemical Reviews, 2010, Vol. 110, No. 3 Guillena et al.

Table 21. N-Methylation of Anilines Catalyzed by a Ruthenium Scheme 21. N-Alkylation of Aniline with 1,7-Heptanediol
and Tributhylphosphime Using RuCl3 · nH2O Catalyst

entry R no. yield (%)

1 H 72a 80
2 Cl 72b 86 Scheme 22. Cyclization of 1,3-Propanodiol with Aniline
3 Me 72c 76 Derivatives Using RuCl3 · nH2O Catalyst: Synthesis of
4 MeO 72d 71 Quinolines

Table 22. N-Alkylation of Cyclic Amines with Diols Using

RuCl3 · nH2O

Table 23. N-Alkylation of Azoles with Primary Alcohols Using

RuCl3 · nH2O

entry m n no. yield (%)

1 1 1 74a 79
2 2 1 74b 79
3 2 2 74c 59
4 2 3 74d 81a entry Z Y R1 R2 R3 no. yield (%)
5 2 4 74e 77a,b
1 CH N H H H 79a 97
a
Reaction performed using 6 mol % of PBun3. b At 100 °C. 2 N CH H H H 79b 67
3 N CH H Me H 79c 85
4 N CMe Me H H 79d 94
obtained by combination of RuCl3 · nH2O with phosphines. 5 N CMe Me H Me 79e 32
Hence, RuCl3 · nH2O and PBun3 have shown their efficiency 6 N CMe Me H Ph 79f >99
7 N N H H H 79g 47
as catalyst in the N-methylation of anilines 1 using methanol
(2a, 400 mol %) as source of the electrophile in dioxane
(Table 21). Although the yields were very high, the obtained process. The oxidation of primary alcohol, followed by an
results were lower when the protocol was extended to the intramolecular Friedel-Crafts and final dehydrogenation,
methylation of N-alkylanilines of type 43.132 gave quinolines 77.136
Piperidine (30b, m ) 2) could be alkylated with primary Indoles 8 were synthesized in a similar way to that
alcohols (2, 300 mol %), such as methanol, ethanol, and presented in Table 17 by reaction of the corresponding
1-butanol, in good yields (70-92%) using RuCl3 · nH2O (1 anilines 66 with triethanolamine (instead of the ammonium
mol %) at 220 °C.133 The reaction with the secondary alcohol chloride used in Table 17) using RuCl3 · nH2O (7 mol %),
2-butanol gave, however, a lower yield (8%). When a similar PPh3 (2 mol %), and SnCl2 · 2H2O (100 mol %) in dioxane
reaction was performed using cyclic aliphatic amines (30, at 180 °C for 20 h.137 Yields depended on the electronic
300 mol %) and different diols (Table 22) in dioxane at 180 character of the substituent, with electron-withdrawing groups
°C, the expected tertiary diamines 74 were obtained with giving poor results (9-21%), one electron-donating group
good results.133 The use of tributylphosphine was compulsory giving moderate results (33-66%), and two electron-
to obtain good yields in the case of using longer-chain diols donating groups giving good results (86-99%). When
(12 and 54a).134 However, if the reaction was carried out tri(propan-2-ol)amine was used as the source of the electro-
using Ru3(CO)12 as the source of ruthenium species, triph- phile, the two possible regioisomers were isolated, with the
enylphosphine was the ligand of choice, and the temperature
corresponding 2-methylindole being the major one.
had to be raised up to 220 °C in THF.133
The ruthenium source RuCl3 · nH2O has been used in the Azoles 78 could be alkylated with primary alcohols (400
formation of perhydroazocines 76 by reaction of the corre- mol %) to give the corresponding heterocycles 79 with good
sponding aniline derivative 1 with 1,7-heptanediol (75, 200 results (Table 23) independently of the substitution of the
mol %) in the presence of triphenylphosphine (Scheme 21). starting azole. It should be pointed out that the reaction with
Yields were very modest, with the substitution on the ethanol gave moderate yields.138
aromatic ring not affecting the results.135 Different carboxamides 41 could be alkylated in a similar
Slightly better results were obtained in the preparation of way to that presented in Table 16 by using an excess of a
quinolines 77 by reaction of anilines 66 with 1,3-propanediol primary alcohol 2 (300 mol %) and catalyzed by the mixture
(73, 150 mol %) in the presence of a hydrogen scavenger of RuCl3 · nH2O (0.6 mol %) and tributylphosphine (8 mol
(nitroarene) in dioxane at 180 °C (Scheme 22). The mech- %) in the absence of solvent at 210 °C for 10 h. Yields were
anism of the reaction showed that the pathway started with good for the acetamide alkylation (88-99% for 64) but
the N-alkylation of the amine 66 to give the corresponding moderate for both the specific case of ethylation and the use
3-arylaminopropan-1-ol through a hydrogen autotransfer of other aliphatic amides (55-71%).139 This protocol could
Hydrogen Autotransfer in the N-Alkylation of Amines
Table 24. N-Alkylation of Primary Aliphatic Amines with Secondary Alcohols Using Ru3(CO)12 and Phosphines
a
At 120 °C.
be successfully expanded to different lactams, and Ru3(CO)12
could alternatively be used with slightly lower results.
Another source of ruthenium used in the former processes
of hydrogen autotransfer was Ru3(CO)12, which was an
effective catalyst for the reaction of different amines 20 with
ethanediol (27a, 300 mol %) to give, after 24 h, the
corresponding 2-aminoethanol derivatives 52, as was de-
picted in Scheme 14, with moderate yields (9-42%).106b
The obtained results using other more hindered phosphines
were much more successful. The reaction of primary amines
80 with different alcohols 9 (500 mol %), using Ru3(CO)12
and the phosphines depicted in Table 24, gave the expected
secondary amines 18 in general with good results in the
absence of solvent.140 It should be pointed out that ortho-
tolylphosphine140a gave always worse results than the related
N-phenyl-2-(dicyclohexylphosphanyl)pyrrole.140b These dif-
ferences were the highest when R-substituted primary amines
were used as nucleophiles (compare entries 7 and 8 in Table
24). The reaction rendered similar yields for either aliphatic
or benzylic primary and secondary alcohols, and even acid-
sensitive alcohols (such as furanyl and thienylmethanol)
could be used with similar results. The presence of a
substituent at the four position of benzylamine did not have
any appreciable effect on the final result. However, the
Chemical Reviews, 2010, Vol. 110, No. 3 1627
Scheme 23. Cyclization of 1,2-Ethanodiamine with 1,2-Diols
Using Ru3(CO)12: Synthesis of Substituted Piperazines
reaction failed for poor nucleophilic anilines under these
conditions, even using electron-donating substituted ones.
The aforementioned protocol using the N-phenyl-2-(dicy-
clohexylphosphanyl)pyrrole ligand has been expanded to
secondary amines 20, using in this case tert-amyl alcohol as
solvent. The obtained yields for amines 21 were generally
very good, with the best results being obtained with cyclic
aliphatic derivatives (47-97%).141
Piperazines 81 were prepared by the reaction of equimo-
lecular amounts of the diamine 31 with 1,2-diols 62 in THF
(Scheme 23).142 Other phosphanes, sources of ruthenium
species, and solvents were tested, giving in all cases lower
results.
The combination of the dimeric (para-cymene)ruthenium
dichloride [RuCl2(p-cymene)]2 with different phosphanes
seemed to be a more versatile protocol, having been used
with very different nitrogen-containing compounds. Initially,
1628 Chemical Reviews, 2010, Vol. 110, No. 3 Guillena et al.

Scheme 24. N-Alkylation of 2-(Piperazin-1-yl)pyrimidine Table 25. N-Alkylation of Sulfonamides with Primary Alcohols
with Benzo[d][1,3]dioxol-5-ylmethanol Using Using [RuCl2(p-cymene)]2
[RuCl2(p-cymene)]2: Synthesis of Piribedil

entry R1 R2 no. yield (%)

[RuCl2(p-cymene)]2 was used in combination with 1,1′-
bis(diphenylphosphanoferrocene) (dppf) as the catalytic
system for the alkylation of equimolecular amounts of
secondary amines 20 with primary alcohols 2 in the presence
of molecular sieves, with good yields (62-97%).143 For
illustrating the possibilities of this protocol, it was used in
the synthesis of piribedil (Scheme 24), which is a piperazine
dopamine agonist used in the treatment of Parkinson’s
disease. Even ammonium acetate could be used as the source
of the nucleophile to give, in this case, tribenzylamine by
reaction with 5 equiv of benzyl alcohol, in the absence of
solvent.
The above protocol was also used with either very crowded
primary aliphatic amines, such as tert-butylamine, or poor
nucleophilic aniline (1a) and 2-aminopyridine (33a) to give
the corresponding secondary amine derivatives with good
yields (69-99%).144 It should be pointed out that, in these
cases, the amount of the catalyst has to be double, and the
presence of K2CO3 (10 mol %) was necessary in order to
achieve good yields.
A further modification of the protocol, essentially by
changing the ferrocene diphosphane derivative for bis(2-
diphenylphosphanophenyl)ether, improved the previous re-
sults using [RuCl2(p-cymene)]2 and permitted the synthesis
of anti-inflammatory agents such as antergan, tripelennamine,
pheniramine, and chlorpheniramine. The new protocol
permitted the use of secondary alcohols 9 as the source of
the electrophile in the alkylation of secondary amines 20 to
give the expected tertiary amines 21 with good results
(36-99%). This combination was also able to catalyze the
reaction of anilines 66 with different diols (12, 54a, and 27b)
to give the corresponding N-aryl substituted pyrrolidines,
piperidines, and perhydroazepines, respectively, in the pres-
ence of triethylamine (10 mol %) as base. Mechanistic studies
showed that the in situ formed aldehydes could dissociate
from the ruthenium coordination sphere, and therefore, the
imine formation did not necessarily take place while it was
coordinated. Moreover, it should be highlighted that this
combination of ruthenium source and phosphane was able
to alkylate simple sulfonamides 85 with primary alcohols 2
(100 mol %) to give the expected compounds 86 with good
yields for the first time (Table 25). The results seemed to be
independent of the nature of the sulfonamide or alcohol
used.145 A similar catalytic combination has been successfully
used in the conversion of 1,4-alkynediols into pyrroles.146
Finally, it should be pointed out that a modification of
the above protocol using triphenylphosphane has been used
for the alkylation of trimethylsilylethanesulfonamide to give
the corresponding N-alkyl sulfonamide, which by treatment
with cesium fluoride gave directly the related amine 45 with
yields from 53 to 74% for both steps. This strategy is an
1 Me Ph 86a 91
2 Ph Ph 86b 92
3 4-O2NC6H4 Ph 86c 72
4 4-ClC6H4 Ph 86d 89
5 4-MeC6H4 Ph 86e 95
6 4-MeOC6H4 Ph 86f 99
7 4-MeC6H4 (CH2)2CH 86g >99
8 4-MeC6H4 (CH2)5CH 86h 91
Scheme 25. N-Benzylation of Diphenylphosphinic Amide
Using [RuCl2(p-cymene)]2
indirect alternative to the N-alkylation of ammonia (5). In
this context, the first alkylation of a phosphinic amide was
described with a good yield (Scheme 25).147 This catalytic
system has also been used in the preparation of amides from
amines and alcohols.148
2.2.3. Derived from Iridium
Iridium complexes have catalyzed different reactions such
as carbon-carbon forming reactions, as well as isomerization
and hydrogen autotransfer reactions.149 One of these iridium
complexes, which has shown its efficiency in the reaction
reviewed here, was the dimeric η5-pentamethylcyclopenta-
dienyliridium(III) dichloride [(IrCl2Cp*)2]. This complex has
been used as catalyst in the N-alkylation of different amine
and nitrogen-containing derivatives. Hence, this catalyst was
able to promote the reaction between aniline derivatives 1
with primary and secondary alcohols (9, 100 mol %) in the
presence of substoichiometric amounts of K2CO3 to give
compounds 89 with excellent yields (Table 26). The elec-
tronic nature of the substituent in aniline did not have any
influence on the obtained results. Lower yields were obtained
when either other primary aliphatic amines (such as benzyl-
amine, pentylamine, or octylamine) or secondary alcohols
were used (61-88%).150 Under similar reaction conditions,
N-methylaniline and N-methylbenzylamine could be benzy-
lated with 75 and 93% yield, respectively.
Slightly better results were obtained when NaHCO3 (1-3
mol %) was used as base, achieving products 89 with yields
ranging from 71 to 98%. Other substituents such as methyl,
bromo, nitro, cyano, and ester moieties in the aniline
derivative 1 were well-tolerated.151 This improved protocol
was applied to the alkylation of secondary amines 20 with
secondary alcohols 9 to give the expected tertiary amines
21 with, in general, good results (Table 27). The reaction
required a slightly higher amount of catalyst loading with
less basic amines such as N-methylaniline. The reaction of
an imine in the presence of a hydrogen donor was performed
Hydrogen Autotransfer in the N-Alkylation of Amines Chemical Reviews, 2010, Vol. 110, No. 3 1629

Table 26. N-Alkylation of Anilines with Secondary Alcohols Scheme 26. Double N-Alkylation of Ammonium
Using (IrCl2Cp*)2 Tetrafluoroborate with Secondary Alcohols Using
(IrCl2Cp*)2

Table 28. Cyclization of Primary Amines with Diols Using

entry R1 R2 R3 no. yield (%) (IrCl2Cp*)2
1 H H Ph 89a 88
2 H H 4-ClC6H4 89b 83
3 H H 4-MeOC6H4 89c 95
4 H H 2-MeOC6H4 89d 95
5 H H CH3(CH2)6 89e 79
6 Cl H Ph 89f 93
7 MeO H Ph 89g 95 entry R1 R2 n no. yield (%)
8 H Me (CH2)5 89h 69
1 Ph H 3 93a 70
2 4-MeOC6H4 H 3 93b 90
Table 27. Alkylation of Secondary Amines with Secondary 3 CH3(CH2)7 H 3 93c 81
Alcohols Using (IrCl2Cp*)2 4 PhCH2 H 3 93d 72
5 PhCH2 H 4 93e 91
6 PhCH2 H 5 93f 73
7 PhCH2 Ph 4 93g 78
8 (R)-PhCHMe Ph 4 93h 76a
a
Reaction performed using 6 mol % potassium acetate as base, with
entry R1 R2 R3 R4 no. yield (%) the (R,S)-diasteroisomer being the major product (92% diastereomeric
1 Me Ph H Ph 21i 91 excess (de)).
2 Me Ph (CH2)5 21j 83
3 (CH2)4 H Ph 21k 92
4 (CH2)4 (CH2)5 21l 88 corresponding cyclic tertiary amines 93 with good yields
5 PhCH2 PhCH2 H Ph 21m 75 (Table 28), with the presence of electron-donating groups
6 PhCH2 Me (CH2)5 21n 44 on the aromatic ring of aniline derivatives enhancing the
results.154 This heterocyclization reaction has been scaled up
under catalytic conditions to understand the possible reaction in the case of compound 93e (0.1 mol) with practically the
mechanism, finding only traces of the expected amine. This same yield.154b The diastereoisomeric version of the reaction
result seemed to indicate that the imine formation must occur could be performed by using the corresponding chiral (R)-
in the coordination sphere of the iridium complex. Finally, phenethylamine. In this case, the reaction was carried out
it should be pointed out that this protocol has been success- using sodium acetate as base to give the expected N-
fully used for the sequential double alkylation of benzylamine phenethylpiperidine 93h. Although the diastereoisomeric
first with a benzylic alcohol and second with an aliphatic excess was good, the enantiomeric excesses of the isolated
one, obtaining the corresponding tertiary amine with good diastereoisomers were 86 and 93% for the major (R,S)-93h
yields (42-88%). and minor (R,R)-93h, respectively. This racemization could
The above protocol has been recently used in the be explained by isomerization of the initial imine or the
preparation of N,N′-disubstituted piperazines from the iminium intermediate.
corresponding N-substituted 2-aminoethanol derivatives This strategy has been used in the diastereoselective
of type 57 (n ) 0).152 synthesis of noranabasamine (94), which is an amphibian
Tertiary amines 22 could also be obtained by reaction of
alkaloid isolated from the Columbian poison dart frog
ammonium acetate with a slight excess of a great variety of
Phyllobates terribilis.155 For this purpose, diol 92e was
primary alcohols 2 (360 mol %) using the former protocol
reacted with chiral (R)-phenethylamine (80h) to give the
but increasing the amount of base (NaHCO3 5-30 mol %)
expected piperidine 93i with good yield (Scheme 27). The
and the temperature (130 °C).153 Although yields were very
final debenzylation of the chiral auxiliary, followed by
high for benzylic alcohols 6 (71-92%), in the case of
chlorination of the pyridine ring and a final Suzuki reaction,
aliphatic alcohols, a higher amount of catalyst should be
loaded [(IrCl2Cp*)2 5 mol %; NaHCO3 30 mol %]. Surpris- gave the expected alkaloid 94. The enantiomeric compound
ingly, when ammonium tetrafluoroborate (90) was used was also prepared with similar results just by changing the
instead of acetate, a high selectivity toward the formation absolute configuration of the initial amine.
of the corresponding symmetrical secondary amine 91 was Other heterocyclic compounds have been prepared by
observed using only a little excess of the corresponding means of a N-alkylation process catalyzed by (IrCl2Cp*)2.
alcohol in the absence of solvent (Scheme 26; 220 mol % For instance, indoles 8 were synthesized by reacting amino
for primary alcohols R1 ) H, and 300 mol % for secondary alcohols 63 with 5 mol % of (IrCl2Cp*)2 and 10 mol % of
ones). The reaction procedure could be extended to the K2CO3 in toluene at 110 °C for 20 h, with good yields
formation of heterocyclic amines such as 2-phenylpyrrolidine (68-99%).156 2-Nitrophenylethanol could be used, instead
and 2-phenylpiperidine with 62 and 85% yield, respectively, of the corresponding amino derivative, as the starting
just by starting from the corresponding diol 92. material, rendering the expected indole 8a in 68%. In a
The reaction of diols 92 with different amines 80 (150 similar process, other heterocyclic compounds 96 could be
mol %), instead of ammonium trifluoroborate, gave the obtained from the aminophenyl alcohols 95 (Scheme 28).
1630 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 27. Cyclization of 1-(6-Methoxypyridin-3-yl)butane-
1,4-diol with (R)-1-Phenylethylamine: Synthesis of
Noranabasamine
Scheme 28. Cyclization of 2-Aminophenylalkyl Alcohol
Derivatives Using (IrCl2Cp*)2
Scheme 29. Cyclization of 2-Aminophenylaminoethanol
Derivatives Using (IrCl2Cp*)2
Scheme 30. Cyclization of 1,2-Diamine Derivatives with
1,2-Diols Catalyzed by (IrCl2Cp*)2: Synthesis of Substituted
Piperazines
The presence of a nitrogen atom in the aliphatic chain did
not change the previous results. Thus, different 2-aminoet-
hanol derivatives 97 could be converted into the correspond-
ing 1,2,3,4-tetrahydroquinooxalines 98 (Scheme 29). The
reaction needed a higher catalyst loading, as well as longer
reaction times, and the yields were slightly lower than
previously presented in Scheme 28. Moreover, the presence
of the N-methyl moiety in the tertiary amine 97 was of great
importance, since the reaction with the corresponding
secondary amine failed, probably due to the coordination of
iridium atom at this position.157
The cyclization of diols 62 with different 1,2-diamines 99
catalyzed by (IrCl2Cp*)2 was accomplished in water as
solvent to give the corresponding piperazines 100 (Scheme
30). It should be pointed out that the diastereoselectivities
on the newly formed sterogenic centers were from good [in
the case of the formation of two new sterogenic centers (75%
de)] to excellent [in the case of the formation of only one
stereogenic center (>95% de, R4 ) H)].158
Guillena et al.
Table 29. N-Alkylation of Carbamates with Primary Alcohols
Using (IrCl2Cp*)2
entry R1 R2 no. yield (%)
1 Me Ph 102a 46
2 Bun Ph 102b 87
3 Bun 4-ClC6H4 102c 55
4 Bun 4-MeC6H4 102d 92
5 Bun CH3(CH2)2 102e 65
Mechanistic studies on the homogeneous iridium-catalyzed
alkylation of amines with alcohols have been performed
using density functional theory (DFT) calculations.159 This
study showed that the found reaction energetic barriers were
consistent with the experimental requirements of elevated
temperatures. The reaction was composed of three multistep
processes: the first one was the iridium-catalyzed oxidation
of the alcohol, the second one was the nucleophilic addition
of the amine to the formed aldehydes, and the third one was
the iridium-catalyzed reduction of the imine to the final
amine, requiring the coordination of the alcohol and/or the
imine to the metal coordination sphere. Because the amine
dissociation from the metal center was a high energetic step,
although it is required to regenerate the catalytic species,
high temperature must be used to overcome this energetic
barrier. The presence of a carbonate as ancillary ligand has
been found to decrease the energetic barrier in many different
steps. It participates in the dehydrogenation of the alcohol
by hosting the proton, while the hydride coordinates to the
metal, and in the final step it has the reverse effect as proton
reservoir for the reduction of the imine. The dehydrogenation
of the alcohol via proton transfer, followed by β-elimination
of a hydride, has a lower energetic barrier than the related
one with the amine. This is connected with the iridium-oxygen
bond in the alkoxy intermediate compared with the
iridium-nitrogen bond in the possible amido intermediate.
These facts contribute to the production of the required
aldehyde by oxidation of the initial alcohol rather than the
generation of the imine intermediate by oxidation of the
initial amine. Moreover, the imine formed by condensation
of the initial amine and the aldehyde is more easily
hydrogenated that the aldehyde, favoring the overall reaction
pathway.
The former catalyst has been used not only in the
alkylation of amines but also in the alkylation of amides.
The reaction of different carbamates 101 with primary
alcohols 2 (400 mol %) at 130 °C without solvent gave the
expected carbamic derivatives 102 (Table 29). The presence
of a base accelerated the reaction, with sodium acetate being
the most effective one. The temperature was also an
important aspect, since temperatures higher or lower than
130 °C gave lower yields. The best results were obtained
with the benzyl alcohol derivative bearing electron-donating
groups at the four position, with the yields using aliphatic
alcohols being moderated. However, the reaction failed when
secondary alcohols were used as electrophilic partners.160
Owing to the facile deprotection of carbamates under basic
methanol/water conditions, these carbamates could be re-
garded as an ammonia equivalent. In fact, the reaction of
compound 102b with NaOH in MeOH/H2O gave the
corresponding benzyl amine in 92% yield. The same protocol
Hydrogen Autotransfer in the N-Alkylation of Amines Chemical Reviews, 2010, Vol. 110, No. 3 1631

Scheme 31. N-Alkylation of Primary Amines with Scheme 32. Double N-Alkylation of 2,6-Diaminopyridine
2-(1H-indol-3-yl)ethanol Using [IrCl(COD)]2: Synthesis of with Primary Alcohols Using [IrCl(COD)]2
N-Alkylated Tryptamine Derivatives

Table 30. N-Alkylation of Heteroaromatic Amines with Primary

Alcohols Using [IrCl(COD)]2

Table 31. N-Alkylation Process through an Indirect aza-Wittig

Process of N-(Triphenylphosphoranylidine)aniline with Primary
Alcohols Using [IrCl(COD)]2

entry R1 Z Y R2 no. yield (%)

1 H CH CH Ph 106a 92 entry R no. yield (%)
2 4-CF3 CH CH Ph 106b 67 1 Ph 3e 91
3 4-MeO CH CH Ph 106c 92 2 4-NO2C6H4 3f 38
4 H CH N Ph 106d 93 3 4-MeOC6H4 3g 81
5 H N CH Ph 106e 97 4 2-furyl 3h 71
6 H CH N 4-ClC6H4 106f 82 5 PhCH2 3i 85
7 H CH N 4-MeOC6H4 106g 97

A further optimization of the reaction conditions showed

could also be applied to simple amides 41 to give the
corresponding N-alkylated amides 64 with good yields
(59-91%).
Another iridium complex widely used in this type of
reaction was the dimer of (η4-1,5-cyclooctadiene)iridium
chloride {[IrCl(COD)]2}. Initially, it was used in the prepara-
tion of N-alkylated tryptamine 104 (Scheme 31). Tryptamine
is a substance presented in various naturally occurring
products, as well as in synthetic ones, that exhibit an
important pharmacological activity. Its derivatives were
prepared by reaction of tryptophol (103) with different
primary amines 45 in the presence of 1,1-bis(diphenylphos-
phanoferrocene) (dppf) with excellent yields.161 The same
structures 104 could be obtained by reaction of primary
tryptamine with the corresponding primary alcohols 2,
although with lower yields (64-86%). This protocol has been
applied to the reaction of tryptamine with diols, such as 1,4-
butanediol (12), 1,5-pentanediol (54a), and 1,6-hexanediol
(27b), rendering the corresponding pyrrolidine, piperidine,
and azepine derivative, respectively, with moderate to good
yields.
The combination of the aforementioned iridium source
with ligands of type N-phosphanopyridin-2-amine has shown
its potential in the alkylation of different substituted (het-
ero)aromatic amino derivatives 105 with primary alcohols
2 (110 mol %). The reaction gave good yields independently
of the substitution position on the amine derivative 105.
However, lower yields were obtained when electron-
withdrawing group were placed in the four position of both
the amine derivative 105 or the benzylic alcohol (Table 30).
It should be pointed out that the presence of strong bases
diminished the tolerance of the process toward the use of
basic sensitive compounds. Surprisingly, the reaction using
high nucleophilic aliphatic amines failed.162
that the process could be performed at 70 °C, with a lower
iridium loading (0.1-0.6 mol %) but longer reaction times
(24 h), affording the expected products 106 with similar
yields. Under these new conditions, the double alkylation
of 2,6-diaminopyridine (107) was attempted using primary
alcohols 2 (210 mol %), yielding the corresponding amine
108 after 2 days with excellent results (Scheme 32). In the
case of using benzylic alcohols, the presence of electron-
donating groups was well-tolerated, but in the case of
electron-withdrawing groups, yields were slightly lower.163
The related unsymmetrical alkylated diaminopyridine deri-
vates could be obtained in a two-step procedure with similar
results to those for symmetrical amine ones.
The first alkylation of amines through an indirect aza-
Wittig process was performed using the former iridium
source (Table 31). The reaction using a slight excess of the
starting compound 19 (110 mol %) gave the expected
compound 3 in general with good yields.164
A similar iridium complex has been successfully used in
the oxidative dimerization of alcohols to give the corre-
sponding esters.165
N-Heterocyclic carbenes have attracted much attention
because of their behavior as stable ligands for homogeneous
catalysis. The first example of their application to the
alkylation of aniline (1a) with primary alcohols 2 (100 mol
%) is depicted in Scheme 33. The reaction gave better results
for benzyl alcohol than for 1-butanol, with the presence of
a strong base being mandatory to get good results.166
A similar complex has been used in the alkylation of
primary amines with primary and secondary alcohols 9
(500-1000 mol %) to give a mixture of products, with the
selectivity being highly dependent on the combination of
amines and alcohol used without any apparent trend (Table
32). The presence of silver trifluoromethanosulfonate (Ag-
1632 Chemical Reviews, 2010, Vol. 110, No. 3 Guillena et al.

Scheme 33. N-Alkylation of Aniline with Primary Alcohols Table 33. Cyclization of Naphthylamine Derivatives with
Using a Simple Carbine-Iridium Complex 1,2-Diols Catalyzed by IrCl3: Synthesis of Substituted
1H-Benzo[g]indoles

entry R1 R2 R3 no. yield (%)

1 H H H 112a 47
Table 32. N-Alkylation of Primary Amines with Secondary 2 H Me Me 112b 96
Alcohols Using a Simple Carbene-Iridium Complex 3 H (CH2)3 112c 90
4 H (CH2)3 113d 98a
5 4-Me H H 112e 90
6 4-Me Me Me 112f 88
7 5-MeO H H 112g 72
8 7-MeO H H 112h 72
9 5-Me2N H H 112i 59
a
The same result was obtained using either cis- or trans-diol.

the corresponding alkylated amines, in general, with moder-

ate results.169
entry R1 R2 R3 no. yield (%) no. yield (%) An in situ generated iridium complex was also active
1 Ph Ph H 18l 90 109a 0 catalyst in the hydrogen autotransfer process. Thus, the
2 PhCH2 Ph H 18m 0 109b 90 combination of nearly equimolecular amounts of iridium
3 PhCH2 Prn H 18n 33 109c 62 trichloride and 2,2′-bis(diphenylphosphano)-1,1′-binaphthyl
4 CH3(CH2)5 Ph Me 18h 67 109d 33 (BINAP) was an efficient catalyst for the N-heterocyclization
5 PhCH2 Ph Me 18j 90 109e 0
of different substituted naphthylamines 111 with 1,2-diols
(62) to give the corresponding 1H-benzo[g]indoles 112
Scheme 34. N-Alkylation of Primary Amines with Primary (Table 33). The reaction was faster in the presence of air
Alcohols Using a Carbene-Iridium Complex and did not show any effect due to the presence of
substitution on the starting amine 111.170 The protocol was
further expanded to the use of 1,3-propanediol 73, rendering
the corresponding benzoquinolines with good to excellent
results (66-95%). The mechanism seemed to go through
the alkylation of 2 equiv of the starting amine with the diol
to give the corresponding diamine, which suffers an intramo-
lecular Friedel-Crafts reaction, liberating the starting amine
111. The final dehydrogenation by the oxygen in the air gave
the corresponding heterocyclic compound.

OTf) is compulsory in order to perform the reaction 2.2.4. Others

successfully. The reaction of highly nucleophilic tert-
butylamine with benzyl alcohol (6a) gave the corresponding
benzylated amine in 95% yield. Following the same proce-
dure, a secondary amine, such as N-methylaniline, could also
be alkylated with benzyl alcohol to give the corresponding
tertiary amine in a excellent 95% yield. However, the reaction
failed with other secondary amines.167
Very recently, the aforementioned iridium catalyst has
been used in the conversion of waste glycerol into the aniline
derivative 3c in a poor yield (10%).168 Glycerol was
converted into 1,3-propanediol (73) by fermentation using
Clostridium butyricum, and this aqueous culture supernatant
was directly transformed into N-propylaniline by reaction
with an excess of aniline (1a, 200 mol %) in a two-phase
media using an ionic liquid.
An iridium complex bearing a chelating pyrimidine-
functionalized N-heterocyclic carbene has been used as the
catalyst in the N-alkylation of aliphatic and aromatic primary
amines 80 with an equimolecular amount of primary alcohols
2 (Scheme 34). The reaction was active in the presence of
weak bases, such as NaHCO3, and molecular sieves, yielding
Other different complexes were active in the hydrogen
autotransfer process. Hence, simple copper(II) acetate has
been very recently proposed as a catalytic alternative in the
alkylation of sulfonamides 85 with different alcohols of type
2.171
The RhH(PPh3)4 complex was effective in the alkylation
of primary amines 113 with a large excess of primary
alcohols 2 (such as methanol, ethanol, or benzyl alcohol)
used both as source of the electrophile and solvent at the
same time, to give the expected N-alkylated amines 114
(Scheme 35). The results were in general excellent, as the
yields and time were concerned. Only when the secondary
amine pyrrolidine was used as nucleophilic partner was the
yield decreased to 74%.91
Scheme 35. N-Alkylation of Primary Amines with Primary
Alcohols Using RhH(PPh3)4 Catalyst
Hydrogen Autotransfer in the N-Alkylation of Amines
Scheme 36. Double N-Alkylation of Primary Amines with
Primary Alcohols Using PtCl2(PhCN) Complex
The last example of this section is depicted in Scheme
36. The reaction of different primary amines 80 using an
excess of a primary alcohol 2 (1200 mol %) catalyzed by
PtCl2(PhCN)2 in the presence of tin(II) chloride, gave the
double alkylation process, rendering the corresponding
amines 115. Allylic alcohols, instead of aliphatic ones, could
also be used, yielding in these cases the corresponding
monoalkylated products.172
3. Amines as Source of Electrophiles
Surprisingly, the alkylation of amines using other amines
as initial source of the electrophile has been performed using
many different heterogeneous and homogeneous catalysts,
although it should be highlighted that this process is less
developed than the one using alcohols as electrophiles. The
hydrogen autotransfer process of amine alkylation does not
have to be confused with an alkylation using in situ formed
ammonium derivatives through a typical SN2.173 The process
reviewed here always begins with the oxidation174 of an
alkylamine to the corresponding imine (iminium), followed
by addition of a second nucleophilic amine to the in situ
formed imine, generating an aminal intermediate, elimination
of the initial dealkylated amine and formation of a new imine
(iminium), with a final hydrogenation of this imine. The
thermodynamics of the whole process is in some cases highly
favorable, above all when ammonia is one of the products
of the reaction.175
3.1. Heterogeneous Catalysts
Initially, heterogeneous catalysts were used in a different
reduction process, with the N-alkylation of amines by other
amines being an unwanted side reaction. However, later on
these processes have gained the attention of the scientific
community.
3.1.1. Derived from Nickel
Historically, nickel was the first metal catalyst used to
perform the N-alkylation of amines with other amines being
the source of the electrophile. This catalytic ability was found
when Raney-nickel was used as catalyst to carry out the
reduction of nitrile derivatives. When this process was
performed at 20 atm of hydrogen and 110-130 °C, in tetralin
or decalin, symmetrical secondary amines were formed in
low yields (2-47%), together with the expected primary
amine.46 The presence of the main side product was explained
by reaction of the intermediate imine formed, either in the
first step of the reduction of the starting nitrile or by
dehydrogenation of final primary amine product, with the
final primary amine to give an aminal intermediate, which
in its turn liberated ammonia and gave a N-substituted imine.
The final hydrogenation of this imine led to the formation
of the secondary amine. The yield of this side product could
be minimized by carrying out the reaction in the presence
of ammonia.53 Surprisingly, moderate yields of the corre-
sponding secondary amines (45 and 66%, respectively) were
Chemical Reviews, 2010, Vol. 110, No. 3 1633
Scheme 37. Auto N-Alkylation of Amylamine Using
Raney-Nickel
Table 34. N-Alkylation of Cyclic Amines with Primary Amines
by Raney-Nickel
entry Y R no. yield (%)
1 CH2 CH3(CH2)4 119a 40
2 CHMe CH3(CH2)4 119b 72
3 O CH3(CH2)4 119c 52
4 CH2 PhCH2 119d 58
5 CHMe Ph(CH2)3 119e 72
Scheme 38. Auto N-Alkylation of Primary Amines Using
Raney-Nickel
obtained when the reduction of octanenitrile or nonanenitrile
with hydrogen (75 atm) was performed at lower temperatures
(65-80 °C) in ethanol using Raney-nickel as catalyst (15
mol %).176
The former alkylation process was also observed in the
reduction of other nitrogen-containing derivatives, such as
amides and oximes. Thus, in the reduction of benzaldoxime,
benzamide, and other compounds, in the absence of solvent,
and using Raney-nickel (8-15 mol %) at 200 °C under 100
atm of hydrogen, the corresponding symmetrical secondary
amine was obtained in low yields (6-23%). Better results
were obtained in the direct alkylation of cyclohexylamine
and 3-methylbutan-1-amine (116) under similar reaction
conditions (Scheme 37), affording the corresponding dialky-
lamine with 36 and 99% yield, respectively, and liberating
ammonia.42
The shown auto N-alkylation process of amines (Scheme
37) was also observed as a side reaction when thioamides
were reduced in the presence of Raney-nickel (1 150-1 600
mol %) at reflux of ethanol or dioxane, yielding the
corresponding symmetrical secondary amines in moderate
yields (40-45%).47
The auto N-alkylation of primary amines, under benzene,
toluene, or xylene reflux to give the corresponding sym-
metrical secondary amines, has been accomplished by the
use of Raney-nickel in the absence of hydrogen with good
yields (70-86%). This protocol was also extended to the
alkylation of cyclic secondary amines 118 (400 mol %) using
primary alkyl amines 45 as the source of the electrophile
(Table 34), to give the corresponding tertiary amines 119 in
moderate yields, with concomitant liberation of ammonia.177
The large excess of Raney-nickel required above could
be reduced in the auto N-alkylation of primary amines 45 to
give the corresponding symmetrical secondary ones 120
when the reaction was performed under xylene reflux
conditions, retaining the good yields (Scheme 38).178
1634 Chemical Reviews, 2010, Vol. 110, No. 3 Guillena et al.

3.1.2. Derived from Copper Table 35. N-Alkylation of Secondary Amines with Secondary
Amines by Palladium Black
Several copper species have been used as catalysts either
in the transalkylation or in the autoalkylation of amines. For
instance, diethylamine could be obtained from ethylamine
by using porous copper in a closed recycling system at 256
°C after 2.5 h under a hydrogen atmosphere (1 atm), with
entry R1 R2 R3 R4 no. yield (%)
only 14% yield.179 The kinetic data obtained from this
transformation showed a zero-order equation rate, with the 1 H CH3(CH2)5 H Ph 121a 5
2 H CH3(CH2)4 Prn Me 121b 34
desorption of ammonia and absorption of ethylamine being 3 H CH3(CH2)4 Ph Me 121c 20
the possible controlling mechanism pathway. Both, the 4 H Ph Ph Me 121d 19
amount of ammonia and hydrogen presented on and over 5 (CH2)4 Ph Me 121e 56
the catalysts surface influenced strongly the reactivity and 6 (CH2)5 Prn Me 121f 83
the product distribution of the reaction. 7 (CH2)5 Ph Me 121g 87

Alumina-supported copper (56 wt %) has been used in a

continuous bead reactor at an atmospheric hydrogen pressure
and 227-257 °C to perform the autoalkylation and tran-
salkylation of benzylamine (45a) and benzylmethylamine,
respectively. When benzylamine was the starting material,
the main reaction product was dibenzylamine (42-50%
yield), with the corresponding imine (N-benzylidene-1-
phenylmethanamine) being also detected. However, when
benzylmethylamine was the initial reagent, the main product
was dibenzylmethylamine and methylamine, albeit in lower
yields (14-17%).180 In these processes catalyzed by copper
and other metals, the presence of hydrogen was necessary
in order to prevent the formation of bulk copper nitride,
which deactivated the catalyst. Temperature-programmed
desorption (TPD) and surface reaction (TPSR) studies
indicated that three different deactivation processes had to
be considered, namely, the formation of metal nitride, the
formation of metal carbide, and carbonaceous species,181
independently of the metal used (copper, nickel, or cobalt).
While only copper nitride formation was observed with
copper, all of the other three deactivation processes occurred
with nickel and cobalt.
The copper-chromite catalyst (CuCr2O4-BaCr2O4, 93
mol %) has been used in the synthesis of the symmetrical
secondary amine dilaurylamine by hydrogenation of laura-
mide at 220 °C and under 340-365 atm of hydrogen
pressure. The corresponding didodecylamine was obtained
with a moderate 52% yield.54 Also, the reduction of dode-
cylnitrile in a stainless steel tubular flow reactor at 150 °C
and under 30 atm of hydrogen pressure, using in this case a
copper-chromium catalyst supported in alumina, gave the
aforementioned didodecylamine in a modest 34% yield.58
3.1.3. Derived from Palladium
Heterogeneous palladium catalysts have been extensively
used in the N-alkylation of amines using another (or the
same) amine as the source of the electrophile. At the
beginning of the previous century, the optimal conditions to
obtain secondary amines from primary ones by using
palladium as catalyst were discovered. So, when palladium
on Ba2SO4 (2.5 wt %, 2.5 mol %) was used as catalyst in
the presence of hydrogen, benzylamine (45a) was trans-
formed into the corresponding dibenzylamine with 90% yield
under xylene reflux conditions and after 2 h.182
When the aforementioned reaction was performed using
palladium black (18 mol %) in ethanol at reflux, dibenzy-
lamine was also quantitatively obtained after 3 h. However,
when the same protocol was applied to 2-phenylethylamine,
only a modest 10% yield of the corresponding symmetrical
secondary amine was obtained.87
Table 36. Auto N-Alkylation of Primary Amines using
Palladium on Carbon
entry R1 R2 no. yield (%)
1 H Prn 91a 47
2 H Pri 91b 50
3 Me Et 91c 41
Following the above ideas, different evaporated metal films
(Pt, Pd, Ni, W, and Co) were incorporated in carbon to
prepare the corresponding catalysts. These catalysts were
tested in the reaction of methylamine and dimethylamine (25)
at 227 °C under a hydrogen atmosphere: only palladium on
carbon showed activity but rendered mixtures of dimethy-
lamine and trimethylamine as main products.183
After these preliminary studies on the activity of palladium,
the studies were revisited. Hence, when benzylamine (45a)
was treated with palladium black (5 mol %) at 80 °C, 41%
of dibenzylamine was formed after a few hours, together
with 41% of the corresponding imine N-benzylideneaniline.
However, allylamine was converted into N-propylideneal-
lylamine (95%). By heating the reaction at higher temper-
atures, secondary amines were converted into the corre-
sponding tertiary ones, albeit in general with low yields
(3-60%). The interesting results obtained when amines 20
were alkylated by other secondary amines 110 to give the
corresponding tertiary amines 121 are depicted in Table 35.184
Following a similar protocol, pyrrolidine (30: m ) 1) was
transformed into 1,4-di(pyrrolidin-1-yl)butane (74f: m ) 1;
n ) 3) in 75% yield, after being heated at 200 °C during
5 h.185
The aforementioned protocol using palladium black was
extended to the transalkylation of two different tertiary
amines 22, affording a mixture of all statistically possible
tertiary amines. For instance, when equimolecular amounts
of tributylamine and trihexylamine were reacted, a mixture
of tributylamine (16% yield), trihexylamine (16% yield),
dibutylhexylamine (31% yield), and butyldihexylamine (31%
yield) was obtained.186
The above alkylation process was used in the selective
synthesis of different 1,1- or 1,2-diamino compounds, which
are interesting molecules because of their potential ability
as metal cation and anion receptor molecules, as well as
biologically active substances. Thus, the treatment of an
excess of triamines 122 (333 mol %) with azetidine 123 gave
the polyamines 124 in moderate to good yields (Scheme
39).187 Studies on the mechanism of this exchange reaction
Hydrogen Autotransfer in the N-Alkylation of Amines
Scheme 39. Selective N-Alkylation of Primary Amines with
Azetidine Using Palladium Black
Table 37. Auto N-Alkylation of Primary Amines using Platinum
on Carbon
entry R no. yield (%)
1 Bui 47a 97
2 Bus 47b 37
3 CH3(CH2)4 47c 73
4 (CH3)3CCH2 47d 75
5 (CH3CH2)2CH 47e 21
6 (CH2)4CH 47f 37
7 CH3(CH2)5 47g 51
8 (CH2)5CH 47h 26
showed that the first step was a rapid equilibrium between
the starting electrophilic amine and its planar imine inter-
mediate. The formed imine underwent a rapid equilibrium
toward the enamine-palladium complex, which suffered the
addition of the nucleophilic amine to give an aminal, which
after a reductive cleavage gave the final amine. In addition,
palladium black (40 mol %) was also able to catalyze the
hydrolysis of tertiary amines, such as tributylamine, to afford
the corresponding secondary ones in moderate yields
(22-60%) at 200 °C in water and HCl (35 mol %).188
The auto alkylation of amines 113 could be performed
either by continuous microwave irradiation (2.45 GHz, 50
W) or by heating at 170 °C in water in the presence of
palladium on carbon (Table 36). It should be pointed out
that, besides the amine 91, the corresponding imine and its
carbonyl compound derivative were also observed as byprod-
ucts.189
Recently, palladium on carbon has been successfully used
in the alkylation of different substituted anilines 66 using
nitriles under hydrogen atmosphere.190 Similar compounds
could be obtained when nitroarenes were used instead of
anilines.
3.1.4. Derived from Platinum
The use of platinum catalysts has been reported more
recently for the N-alkylation of amines using other amines
as source of the electrophile. Thus, platinum black supported
in titanium dioxide (5 wt %, 22 mol %) was used as catalyst
to convert primary alkylamines 45 into the corresponding
symmetrical secondary ones 120 by means of a photoirra-
diation (300 nm with a 500 W high-pressure mercury arc)
at room temperature in low yields (24-33%). When the same
protocol was followed for terminal diamines, the correspond-
ing cyclic aliphatic amines 30 were obtained (20-67%
yield).191
Primary amines 80 suffered an autoalkylation process upon
microwave irradiation using platinum on carbon to give the
corresponding secondary ones 47, in general, with moderate
results (Table 37). It should be pointed out that microwave
heating gave better results than conventional heating at 156
Chemical Reviews, 2010, Vol. 110, No. 3 1635
Scheme 40. N-Alkylation of r-Substituted Primary Amines
with Arylmethylamine Derivatives Using Pt Electrodes
°C and that the yields for R-substituted primary amines were
quite lower, with the corresponding ketone being the main
isolated compound.192
Secondary amines could also be obtained electrolytically
starting from primary ones. However, the presence of highly
unstable cationic radicals, which are rapidly deprotonated
and attached to the electrode surface, makes the overall
process very difficult. A biomimetic pathway has been used
in order to overcome this difficulty. Thus, the alkylation of
R-substituted primary amines 113 was carried out using
arylmethyl amines 125 (150 mol %) as initial source of the
electrophile, giving the expected amines 126 with moderate
results (Scheme 40). From a mechanistic point of view, the
process started with the condensation of 4-acetyl-5-hydroxy-
6-iminocyclohexa-2,4-dienone with the arylmethylamine to
give the corresponding imine and liberating ammonia. Then,
isomerization of the above imine gave the corresponding
N-arylidene imine derivative, which reacted with the nu-
cleophilic amine 113, rendering a new iminic arylidene
alkylamine and 3-amino-2,4-dihydroxyacetophenone. The
ketone was oxidized on the anode to the starting 4-acetyl-
5-hydroxy-6-iminocyclohexa-2,4-dienone, while the arylidene
alkylamine was reduced on the cathode to the corresponding
arylmethyl alkylamine 126. The presence of different func-
tional groups on the nucleophilic amine 113, as well as on
the aromatic ring of the alkylating amine 125 (such as
hydroxy, alkoxy, or fluoro moieties), was well-tolerated.193
3.1.5. Others
The auto N-alkylation reaction of amines has also been
carried out using ion-exchange montmorillonites as catalyst
(20 wt %) at 220 °C. Primary amines 80 gave the corre-
sponding secondary ones 47, liberating ammonia. Cyclic
amines, such as pyrrolidine (30: m ) 1), were transformed
into a mixture of products including 1,4-di(pyrrolidin-1-
yl)butane (74f: m ) 1; n ) 3) in 5% yield and 4-(pyrrolidin-
1-yl)butaneamine in 21% yield after 74 h.194
Other pure metals, such as rhodium and ruthenium, were
used as catalysts in some alkylation processes. The best
results were obtained in the hydrogenation of aniline (1a)
with rhodium (0.05 mol %) at 145 °C under 54-68 atm of
hydrogen. This reaction gave not only the hydrogenation of
aniline but also its alkylation to form dicyclohexylamine
(47h, 25% yield).195
Rhodium on carbon (5 mol %) has also been used in the
alkylation of different aliphatic primary amines 45, using
nitriles under hydrogen atmosphere, in good yields
(71-96%).190 Different amines have been used as source of
electrophiles in the alkylation of lithium areneselenolates
catalyzed by ruthenium, with good yields for the alkylare-
neselenides formed.196
1636 Chemical Reviews, 2010, Vol. 110, No. 3 Guillena et al.

3.2. Homogeneous Catalysts Scheme 41. Cyclization of r-Substituted Primary Amines

Using RuCl2(PPh3)3
In the N-alkylation of amines using other amines as
electrophiles, the use of homogeneous catalysts has been
more successful than the heterogeneous ones, with better
results and a wider scope being achieved.

3.2.1. Derived from Ruthenium

As in the previous case of N-alkylation of amines using Table 40. Cyclization of Aniline Derivatives with
Diallylammonium Derivatives Catalyzed by RuCl2(PPh3)3:
alcohols as electrophiles, ruthenium complexes have been Synthesis of Quinolines
the most successful and widely used catalysts to perform a
similar transformation but using amines as electrophiles.
RuCl2(PPh3)3 has been used in a substoichiometric amount
to promote the auto N-alkylation process of primary amines
80 to give the corresponding symmetrical secondary amines
with excellent yields in the absence of solvent (Table 38).
When R-substituted amines were used, the results were
slightly worse but they could be improved either by increas- entry no. R yield (%)
ing the amount of catalyst or by using tetrahydrofuran (THF) 1 129a H 57
as solvent.197 2 129b 6-Cl 34
Following a similar procedure, the aforementioned catalyst 3 129c 6-MeCO 63
was able to convert N,N-dimethylalkylamines 26 into the 4 129d 6-Me 64
5 129e 6-MeO 58
corresponding N-methyldialkylkamines 46 in good yields 6 129f 7-MeO 50
(Table 39). The process formally involved the dehydroge- 7 129g 8-Me 45
nation of the initial amine, on one hand to the methylene
derivative, followed by hydrolysis to give the corresponding
N-methylalkylamine (nucleophilic partner of the reaction), The results were good independent of the ring size, although
and on the other hand to the alkylene derivative (electrophilic in these cases the use of diphenyl ether as solvent showed a
partner of the reaction). The nucleophilic addition to give beneficial effect on the yields.199 It should be pointed out
that the same reaction could also be performed by using
the corresponding aminal, followed by methylamine elimina-
RuCl3 · nH2O and an adequate amount of PPh3 with similar
tion, formation of iminium intermediate, and final hydroge-
results. Moreover, this in situ preparation of the ruthenium
nation, rendered the amine 46. In fact, the same products
catalyst was used in the auto alkylation of primary amines
were obtained when the corresponding N-methylalkylamines
45 to give the corresponding tertiary amines 22 with good
were used as initial reagents.198
yields (51-98%).
Different terminal primary diamines 127 have been cy-
Mechanistic studies performed in the reaction of auto
clized using the former approach as depicted in Scheme 41.
N-alkyaltion of primary amines 45 to give the corresponding
Table 38. Auto N-Alkylation of Primary Amines using RuCl2(PPh3)3 symmetrical secondary derivatives 120 catalyzed by
RuCl2(PPh3)3 revealed a first-order dependence on both the
catalyst and the amine concentration, with the rate-determin-
ing step being the amine dehydrogenation. During the
reaction, the free secondary amine 120 was observed,
indicating that the dissociation of this nitrogen-containing
entry R no. yield (%) ligand from the ruthenium intermediate complex was pos-
1 Pri 47i 72a sible. The imine initially formed from the starting amine must
2 Bun 47j 96 react quickly either with the starting amine or with the
3 CH3(CH2)5 47g 98 liberated ammonia, since it was not observed in the reaction
4 CH3(CH2)11 47k 99 media.200
5 PhCH2 47l 99 As in previous examples, different quinolines could be
6 (CH2)5CH 47h 90
prepared by reaction of the diallylammonium derivative 128
a
THF used as solvent. with different anilines 66 (Table 40). The reaction gave, in
all cases, moderate yields, independent of either the electronic
Table 39. Auto N-Alkylation of N,N-Dimethylalkylamines using character of the substituent or its position. The reaction seems
RuCl2(PPh3)3 to proceed via the initial dehydrogenation of the amine
included in 128 to render the corresponding iminium
derivative, which suffers a nucleophilic condensation with
the aniline 66 to give the corresponding imine. This aniline
derivative is hydrogenated to the corresponding N-allyla-
entry R no. yield (%)
niline, closing the hydrogen autotransfer cycle. The final steps
involved the isomerization of this allylamine to the corre-
1 Prn 46a 68a
2 CH3(CH2)4 46b 80a
sponding enamine-imine, auto Mannich-type reaction, fol-
3 CH3(CH2)7 46c 78 lowed by an intramolecular Friedel-Crafts cyclation; the
4 Ph 46d 90 final dehydrogenation and dehydroamination processes gave
a the corresponding 2-ethyl-3-methylquinoline 129.201 The same
Reaction performed using THF as solvent.
quinolines 129 could be prepared starting from the corre-
Hydrogen Autotransfer in the N-Alkylation of Amines
Table 41. N-Alkylation of Aromatic Amines with Primary
Amines using [(η4-Ph4C4CO)Ru(CO)3]2 Complex
entry Ar R no. yield (%)
1 Ph CH3(CH2)7 130a 99
2 4-O2NC6H4 CH3(CH2)5 130b 20
3 4-FC6H4 CH3(CH2)5 130c 99
4 2-MeC6H4 CH3(CH2)5 130d 93
5 4-MeOC6H4 CH3(CH2)5 130e 98
6 2,4,6-Me3C6H2 CH3(CH2)5 130f 34
7 2-pyridyl CH3(CH2)5 130g 96
8 Ph (CH2)5CH 130h 99
sponding nitrobenzene derivatives and tripropylamine, in-
stead of anilines derivatives 66 and the ammonium salt 128,
respectively. In these cases, yields were similar (22-85%),
with the two- and four-position substituted nitroarenes giving
the best results.202
In situ generated ruthenium complexes, obtained by
reaction of RuCl3 · nH2O (5 mol %) and PPh3 (15 mol %) in
the presence of SnCl2 · 2H2O (100 mol %), promoted the
reaction of aninile derivatives 66 with triallylamine in
dioxane at 180 °C after 20 h to give the corresponding
quinolines 129 with moderate yields (24-61%), with the
worst results being obtained with anilines bearing electron-
withdrawing groups.203 When bis(diphenylphosphane)methane
was used as a phosphane, in a similar protocol, the obtained
results for the synthesis of quinolines 129 were slightly better
(21-86%), with the presence of hex-1-ene, acting as an
hydrogen scavenger, being mandatory.204
Other simple catalysts showed to be very effective in the
N-alkylation of amines using different amines as the source
electrophile. For instance, Ru3(CO)12 (0.42 mol %) catalyzed
the transalkyaltion reaction between triethylamine (22a) and
tripropylamine at 100 °C under a nitrogen atmosphere (7
atm) to give diethylpropylamine and ethyldipropylamine in
25 and 27% yield, respectively.205 Mechanistic and kinetic
studies for this reaction showed a very complex rate
dependence.206 This catalyst was also able to catalyze the
hydrogen-deuterium exchange in tertiary amines from
deuterium oxide, showing a very high selectivity toward R
versus β position.207
Perhaps, the bisruthenium complex depicted in Table 41
is the most useful catalyst for the reaction described along
this section. Different aromatic amines 53 (200 mol %) could
be alkylated with primary amines 80 to afford the corre-
sponding secondary amines 130 with excellent yields. The
electronic nature of the substituent and its position at the
aromatic ring did not have any impact on the results. Only
in the case of nitroaniline derivatives was the yield lower,
and this fact could be attributed to the possible reduction of
the nitro group to give different compounds. A very broad
scope of alkyl amines 80, including aliphatic primary and
secondary amines, benzylic amine derivatives, and the
corresponding heterocyclic methylamines, could be used as
source of the electrophile, affording in all cases the corre-
sponding amines 130 with excellent yields.208
The above-cited protocol was further extended to the
reaction of aniline (1a, 200 mol %) with different sym-
Chemical Reviews, 2010, Vol. 110, No. 3 1637
Scheme 42. Double N-Alkylation of Aniline Derivatives with
Cyclic Amines Using a Bisruthenium Complex
Table 42. N-Alkylation of tert-Octylamine Amines with Primary
Amines using [(η4-Ph4CCO)Ru(CO)3]2 Complex
entry R no. yield (%)
1 Ph 133a 58
2 4-MeOC6H4 133b 89
3 PhCH2 133c 75
4 CH3(CH2)6 133d 90
metrical secondary amines 120 to give the expected N-
alkylated anilines 3 with excellent yields. Not only secondary
amines but also tertiary amines 22 could be used successfully.
Moreover, the use of aliphatic R-substituted primary amines
113 did not have any detrimental effect on the previous
excellent results.209
The use of cyclic amines of type 30 as source of the
electrophile permitted to perform the double alkylation of
aniline derivatives 66 to give the corresponding N-phenyl
substituted cyclic amines 131, but it was less successful
(Scheme 42), with the lowest yields (around 25%) being
obtained for anilines bearing an electron-withdrawing group
at the four-position, such as a halogen atom. Notwithstanding
these yields, the use of cyclic amines as dialkylating agents
was proved.210
Finally, it should be pointed out that the aforementioned
protocol has also been extended to the alkylation of aliphatic
amines by other amines. So, tert-butyl amine derivatives such
as the amine 132 (300 mol %), which could not suffer a
dehydrogenation process to form imines, was alkylated with
different primary amines 45 to give selectively the corre-
sponding secondary amines 133 (Table 42). Instead of
primary amines 45, symmetrical secondary amines 120 (even
the related tertiary ones 22 as well as alkoxy functionalized
ones) could be used as electrophiles with similar results.
Another nucleophilic amine such as adamantylamine, could
be used, with its alkylation being achieved with slightly better
chemical yields.211
3.2.2. Others
Other transition-metal complexes were able to catalyze the
alkylation of amines with other amines. Thus, Os3(CO)12
(0.42 mol %) was effective in the standard transalkylation
reaction between triethylamine (22a) and tripropylamine at
150 °C under 7 atm of nitrogen pressure during 3 h, affording
1638 Chemical Reviews, 2010, Vol. 110, No. 3
Table 43. N-Alkylation of Anilines with Secondary Amines
Using a Carbene-Iridium Complex
entry R1 R2 R3 no. yield (%)
1 H H Ph 89a 94
2 H H CH3(CH2)5 89i >95
3 H H CH3(CH2)10 89j >95
4 H (CH2)5 89k >95
5 Me H CH3(CH2)5 89l 63
6 Cl H CH3(CH2)5 89m >95
7 MeO H CH3(CH2)5 89n 25
a mixture of diethylpropylamine and ethyldipropylamine in
27 and 28% yield, respectively.212 This transformation was
also catalyzed by Rh6(CO)16 but rendering the same mixture
with lower yields,213 with this catalyst being able to perform
the hydrogen-deuterium exchange in amines from deuterium
oxide.214
Very recently, the carbene-iridium complex shown in
Table 43 has been able to catalyze very successfully the
alkylation of anilines 1 (200 mol %) with different R-sub-
stituted primary amines 113, in general, with excellent
results.167 Only in the cases of using anilines having an
electron-donating group were the yields accountably lower.
The PtCl2(PPh3)2 complex (0.5 mol %) in the presence of
SnCl2 · 2H2O (25 mol %) promoted the auto alkylation
process of primary amines 80 in benzene at 180 °C after 5 h
to give the corresponding secondary ones 47 with moderate
to good yields (27-81%), independently of the substitution
on the starting amine. The protocol has also been used in
the cyclization of 1,4-butanediamine 127a (n ) 3) to give
the expected pyrrolidine in low yields.215
4. Conclusions and Outlook
With this review, we have tried to show the versatility
and usefulness of heterogeneous and homogeneous catalysts
in the alkylation of amines through the so-called hydrogen
autotransfer process. The use of alcohols and amines, which
are normally considered as nucleophiles in organic textbooks,
acting in these transformations as electrophiles is highly
convenient owing to the easy handling, stability, wide
availability, and low cost of such compounds. Furthermore,
the use of this strategy of alkylation substitutes the classical
ones employing hazardous and expensive alkyl halides,
sulfonates, or sulfates and allows a greener process, since
the only generated waste is water or ammonia. Moreover,
the low molecular weight of wastes makes this synthetic
strategy unbeatable as far as the atom efficiency numbers
are concerned.
These facts, together with the plethora of new and useful
homogeneous and heterogeneous catalysts, make it possible
to perform these reactions with excellent selectivities and
also make this synthetic strategy very attractive not only for
academia but also for industrial purposes. In the near future,
these processes might be applied to a wide range of nitrogen-
Guillena et al.
containing compounds, which are interesting because of their
application in several chemical industries (such as pharma-
ceutical, agrochemical, detergent, dye, or textile industries),
with the whole synthetic protocol being cleaner, greener,
safer, and more economical and meeting the sustainability,
all goals highly demanded by our present day society.
In the near future, the development of recoverable, more
effective or chiral catalysts, as well as the application of this
type of transformation to other new substrates will improve
the scope of this protocol, with the information review here
serving to envisage the new possibilities and the new niches
unoccupied nowadays.
5. Acknowledgments
We are grateful to the Spanish Ministerio de Ciencia e
Innovación (Consolider Ingenio 2010 CSD2007-00006,
CTQ2007-65218/BQU) and Generalitat Valenciana (Proyecto
Prometeo 2009/039) for the continuous financial support. We
gratefully acknowledge the polishing of our English by O. C.
Townley.
6. References
(1) (a) Wöhler, F. Ann. Phys. Chem. 1828, 12, 253. (b) Kauffman, G. B.;
Chooljian, S. H. Chem. Educ. 2001, 6, 121.
(2) (a) Chemical Biology; Waldmann, H., Janning, P., Eds.; Wiley-VCH:
Weinheim, Germany, 2004. (b) Chemical Biology: From Small
Molecules to Systems Biology and Drug Design; Schreiber, S. L.,
Kapoor, T., Wess, G., Eds.; Wiley-VCH: Weinheim, Germany, 2007;
Vols. 1-3.
(3) (a) Hesse, M. Alkaloids: Nature’s Curse or Blessing?; VHCA: Zürich,
2002. (b) Bhat, S. V.; Nagasampagi, B. A.; Sivakumar, M. Chemistry
of Natural Products; Springer: Berlin, 2005.
(4) (a) Pesticide Chemistry and Bioscience; Brooks, G. T., Roberts, T. R.,
Eds.; Royal Society of Chemistry: Cambridge, U.K., 1999. (b)
Pharmaceuticals: Classes, Therapeutic Agents, Areas of Application;
McGuire, J. L., Ed.; Wiley-VCH: Weinheim, Germany, 2000; Vols.
1-4.
(5) (a) Egan, R. R. J. Am. Oil Chem. Soc. 1968, 45, 581. (b) Arif, S. In
Surfactant Science Series: Detergency of Specialty Surfactants;
Friedli, F., Ed.; Marcel Dekker: New York, 2001; Vol. 98; pp 71-
115. (c) Hayes, K. S. Appl. Catal., A 2001, 221, 187. (d) Wittcoff,
H. A.; Reuben, B. G.; Plotkin, J. S. Industrial Organic Chemicals;
Wiley-Interscience: Hoboken, NJ, 2004.
(6) (a) Malpass, J. R. In ComprehensiVe Organic Chemistry; Barton,
D., Ollis, W. D., Eds.; Pergamon Press: Oxford, U.K., 1979; Vol. 2,
pp 3-59. (b) Gladych, J. M. Z.; Hartley, D. In ComprehensiVe
Organic Chemistry; Barton, D., Ollis, W. D., Eds.; Pergamon Press:
Oxford, U.K., 1979; Vol. 2, pp 61-130. (c) Lindsay, R. J. In
ComprehensiVe Organic Chemistry; Barton, D., Ollis, W. D., Eds.;
Pergamon Press: Oxford, U.K., 1979; Vol. 2, pp 131-184. (d)
Mitsunobu, O. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Fleming, I., Eds.; Pergamon Press: Oxford, U.K., 1991; Vol. 6, pp
65-101. (e) Gilchrist, T. L. In ComprehensiVe Organic Synthesis;
Trost, B. M., Fleming, I., Eds.; Pergamon Press: Oxford, U.K., 1991;
Vol. 8, pp 381-402. (f) Salvatore, R. N.; Yoon, C. H.; Jung, K. W.
Tetrahedron 2001, 57, 7785.
(7) Smith, M. B.; March, J. In March’s AdVanced Organic Chemistry;
Wiley-Interscience: New York, 2001; pp 499-501 and literature
quoted therein.
(8) (a) Emerson, W. S. In Organic Reactions; Adams, E., Ed.; Wiley:
New York, 1948; Vol. 4, pp 175-255. (b) Gribble, G. W. Chem.
Soc. ReV. 1998, 27, 395. (c) Olsen, C. A.; Franzyk, H.; Jaroszewski,
J. W. Synthesis 2005, 2631.
(9) For recent reviews, see: (a) Schlummer, B.; Scholz, U. AdV. Synth.
Catal. 2004, 346, 1599. (b) Hartwig, J. F. H. Acc. Chem. Res. 2008,
41, 1534. (c) Surry, D. A.; Buchwald, S. L. Angew. Chem., Int. Ed.
2008, 47, 6338.
(10) (a) Trost, B. M. Science 1991, 254, 1471. (b) Green Chemistry:
Designing Chemistry for the EnVironment; Anastas, P. T., William-
son, T. C., Eds.; ACS: Washington, DC, 1996. (c) Sheldon, R. A.
Pure Appl. Chem. 2000, 72, 1233. (d) Trost, B. M. Acc. Chem. Res.
2002, 35, 695. (e) Sheldon, R. A.; Arends, I.; Hanefeld, U. Green
Chemistry and Catalysis; Wiley-VCH: Weinheim, Germany, 2007.
(11) (a) Metzger, J. O.; Eissen, M. C. R. Chim. 2004, 7, 569. (b) Jenck,
J. F.; Agterberg, F.; Droescher, M. J. Green Chem. 2004, 6, 544. (c)
Hydrogen Autotransfer in the N-Alkylation of Amines
Thomas, J. M.; Raja, R. Annu. ReV. Mater. Res. 2005, 35, 315. (d)
Parthasarathy, G.; Hart, R.; Jamro, E.; Miner, L. Clean Technol.
EnViron. Policy 2005, 7, 219. (e) Noyori, R. Chem. Commun. 2005,
1807. (f) Ranke, J.; Stolte, S.; Störmann, R.; Arning, J.; Jastorff, B.
Chem. ReV. 2007, 107, 2183. (g) Polshettiwar, V.; Varma, R. S. Acc.
Chem. Res. 2008, 41, 629. (h) Murzin, D. Y.; Leino, R. Chem. Eng.
Res. Des. 2008, 86, 1002. (i) Clark, J. H.; Deswarte, F. E. I.; Farmer,
T. J. Biofuels, Bioprod. Biorefin. 2009, 3, 72.
(12) (a) Sheldon, R. A. Green Chem. 2007, 9, 1273. (b) Sheldon, R. A.
Chem. Commun. 2008, 3352.
(13) For recent reviews, see: (a) Beller, M.; Tillack, A.; Seayad, J. In
Transition Metals for Organic Synthesis: Building Blocks and Fine
Chemicals; Beller, M., Bolm, C., Eds.; Viley-VCH: Weinheim,
Germany, 2004; Vol. 2, pp 403-414. (b) Hong, S.; Marks, T. J.
Acc. Chem. Res. 2004, 37, 673. (c) Severin, R.; Doye, S. Chem. Soc.
ReV. 2007, 36, 1407. (d) Aillaud, I.; Collin, J.; Hannedouche, J.;
Schulz, E. Dalton Trans. 2007, 5105. (e) Müller, T. E.; Hultzsch,
K. C.; Yus, M.; Foubelo, F.; Tada, M. Chem. ReV. 2008, 108, 3795.
(14) (a) Martı́nez, R.; Brand, G. J.; Ramón, D. J.; Yus, M. Tetrahedron
Lett. 2005, 46, 3683. (b) Martı́nez, R.; Ramón, D. J.; Yus, M.
Tetrahedron 2006, 62, 8982. (c) Martı́nez, R.; Ramón, D. J.; Yus,
M. Tetrahedron 2006, 62, 8988.
(15) Edwards, M. G.; Williams, J. M. J. Angew. Chem., Int. Ed. 2002,
41, 4740.
(16) Hoshino, F.; Kimura, H.; Matsutani, K. Chem. Abstr. 1980, 92,
41325c Ger. Offen. Patent DE2907869, 1979.
(17) (a) Zassinovich, G.; Mestroni, G.; Gladiali, S. Chem. ReV. 1992, 92,
1051. (b) Gladiali, S.; Alberico, E. Chem. Soc. ReV. 2006, 35, 226.
(c) Samec, J. S. M.; Bäckvall, J.-E.; Andersson, P. G.; Brandt, P.
Chem. Soc. ReV. 2006, 35, 237. (d) Hydrogen-Transfer Reactions;
Hynes, J. T., Klinman, J. P., Limbach, H.-H., Schowen, R. L., Eds.;
Wiley-VCH: Weinheim, Germany, 2007; Vols. 1-4.
(18) (a) Guillena, G.; Ramón, D. J.; Yus, M. Angew. Chem., Int. Ed. 2007,
46, 2358. (b) Hamid, M. H. S. A.; Slatford, P. A.; Williams, J. M. J.
AdV. Synth. Catal. 2007, 349, 1555. (c) Nixon, T. D.; Whittlesey,
M. K.; Williams, J. M. J. Dalton Trans. 2009, 753.
(19) (a) Shibahara, F.; Krische, M. J. Chem. Lett. 2008, 37, 1102. (b)
Bower, J. F.; Kim, I. S.; Patman, R. L.; Krische, M. J. Angew. Chem.,
Int. Ed. 2009, 48, 34.
(20) (a) Klyuev, M. V.; Khidekel′, M. L. Russ. Chem. ReV. Engl. Transl.
1980, 49, 14. (b) Baiker, A.; Kijenski, J. Catal. ReV. Sci. Eng. 1985,
27, 653. (c) Roundhill, D. M. Chem. ReV. 1992, 92, 1.
(21) Pašek, J.; Kondelik, P.; Richter, P. Ind. Eng. Chem., Prod. Res. DeV.
1972, 11, 333.
(22) Nef, J. U. Liebigs Ann. Chem. 1901, 318, 137.
(23) For recent reviews, see: (a) Gladysz, J. A. Chem. ReV. 2002, 102,
3215. thematic issue on recoverable catalysts and reagents. (b) Corma,
A.; Garcı́a, H. Chem. ReV. 2003, 103, 4307. (c) Dai, L.-X. Angew.
Chem., Int. Ed. 2004, 43, 5726. (d) Itsuno, S.; Haraguchi, N.;
Arakawa, Y. Recent Res. DeV. Org. Chem. 2005, 9, 27. (e) Ding,
K.; Wang, Z.; Shi, L. Pure Appl. Chem. 2007, 79, 1531. (f)
Matsumoto, T.; Ueno, M.; Wang, N.; Kobayashi, S. Chem. Asian J.
2008, 3, 196. (g) Felpin, F.-X.; Fouquet, E. ChemSusChem 2008, 1,
718.
(24) For recent examples from our laboratory, see: (a) Forrat, V. J.; Ramón,
D. J.; Yus, M. Tetrahedron: Asymmetry 2006, 17, 2054. (b) Forrat,
V. J.; Ramón, D. J.; Yus, M. Tetrahedron: Asymmetry 2009, 20, 65.
(25) Takita, Y.; Nishida, Y.; Seiyama, T. Bull. Chem. Soc. Jpn. 1976,
49, 3699.
(26) Brown, A. B.; Reid, E. E. J. Am. Chem. Soc. 1924, 46, 1836.
(27) Narayanan, S.; Prasad, B. P. J. Chem. Soc., Chem. Commun. 1992,
1204.
(28) (a) Matshuhashi, H.; Arata, K. Bull. Chem. Soc. Jpn. 1991, 64, 2605.
(b) Ko, A.-N.; Yang, C.-L.; Zhu, W.-d.; Lin, H.-e. Appl. Catal., A
1996, 134, 53.
(29) Valot, F.; Fache, F.; Jacquot, R.; Spagnol, M.; Lemaire, M.
Tetrahedron Lett. 1999, 40, 3689.
(30) Rosenmund, K. W.; Joithe, A. Ber. Dtsch. Chem. Ges. 1925, 58B,
2054.
(31) Li, K.-T.; Peng, Y.-C. Appl. Catal., A 1994, 109, 225.
(32) Guyot, A.; Fournier, M. Bull. Soc. Chim. Fr. 1930, 47, 203.
(33) Pratt, E. F.; Frazza, E. J. J. Am. Chem. Soc. 1954, 76, 6174.
(34) Barnes, C. M.; Rase, H. F. Ind. Eng. Chem., Prod. Res. DeV. 1981,
20, 399.
(35) (a) Billica, H. R.; Adkins, H. In Organic Synthesis, CollectiVe Volume
3; Horning, E. C., Ed.; John Wiley & Sons: New York, 1955; Vol.
3, pp 176-180. (b) Mozingo, R. In Organic Synthesis, CollectiVe
Volume 3; Horning, E. C., Ed.; John Wiley & Sons: New York, 1955;
Vol. 3, pp 181-183 and literature quoted therein.
(36) Rice, R. G.; Kohn, E. J. J. Am. Chem. Soc. 1955, 77, 4052.
(37) Ainsworth, C. J. Am. Chem. Soc. 1956, 78, 1635.
(38) Botta, M.; De Angelis, F.; Nicoletti, R. Synthesis 1977, 722.
Chemical Reviews, 2010, Vol. 110, No. 3 1639
(39) De Angelis, F.; Ferretti, G.; Botta, M.; Grgurina, I.; Nicoletti, R.
Gazz. Chim. Ital. 1982, 112, 267.
(40) De Angelis, F.; Crasso, M.; Nicoletti, R. Synthesis 1977, 335.
(41) Rice, R. G.; Kohn, E. J.; Daasch, L. W. J. Org. Chem. 1958, 23,
1352.
(42) Winans, C. F.; Adkins, H. J. Am. Chem. Soc. 1932, 54, 306.
(43) Baiker, A. Ind. Eng. Chem., Prod. Res. DeV. 1981, 20, 615.
(44) Langdon, W. K.; Levis, W. W.; Jackson, D. R. Ind. Eng. Chem.,
Process. Des. DeV. 1962, 1, 153.
(45) (a) Venot, A.; Glacet, C. C. R. Chim. 1971, 273, 718. (b) Venot, A.;
Glacet, C. C. R. Chim. 1971, 273, 1266. (c) Venot, A. Bull. Soc.
Chim. Fr. 1972, 4732. (d) Venot, A. Bull. Soc. Chim. Fr. 1972, 4736.
(e) Venot, A. Bull. Soc. Chim. Fr. 1972, 4744.
(46) von Braun, J.; Blessing, G.; Zobel, F. Ber. Dtsch. Chem. Ges. 1923,
56B, 1988.
(47) Kornfeld, E. C. J. Org. Chem. 1951, 16, 131.
(48) Mozingo, R.; Spencer, C.; Folkers, K. J. Am. Chem. Soc. 1944, 66,
1859.
(49) Garcı́a Ruano, J. L.; Parra, A.; Alemán, J.; Yuste, F.; Mastranzo,
V. M. Chem. Commun. 2009, 404.
(50) Alonso, F.; Riente, P.; Yus, M. Eur. J. Org. Chem. 2008, 4908.
(51) Aston, J. G.; Peterson, T. E.; Holowchak, J. J. Am. Chem. Soc. 1934,
56, 153.
(52) Laizer, W. A.; Arnold, H. R. In Organic Synthesis, CollectiVe Volume
2; Horning, E. C., Ed.; John Wiley & Sons: New York, 1943; Vol.
2, pp 142-145.
(53) Schwoegler, E. J.; Adkins, H. J. Am. Chem. Soc. 1939, 61, 3499.
(54) Schneider, H.; Adkins, H.; McElvain, S. M. J. Am. Chem. Soc. 1952,
74, 4287.
(55) Takehara, K.; Okajima, S.; Agawa, T.; Komori, S. J. Am. Oil Chem.
Soc. 1971, 48, 748.
(56) Becker, J.; Niederer, J. P. M.; Keller, M.; Hölderich, W. F. Appl.
Catal., A 2000, 197, 229.
(57) For reviews, see: (a) Ho, T.-L. Tandem Organic Reaction; John Wiley
& Sons: New York, 1992. (b) Tietze, L. F.; Beifuss, U. Angew.
Chem., Int. Ed. Engl. 1993, 32, 131. (c) Tietze, L. F. Chem. ReV.
1996, 96, 115. (d) Parsons, P. J.; Penkett, C. S.; Shell, A. J. Chem.
ReV. 1996, 96, 195. (e) Multicomponent Reactions; Zhu, J., Bienaymé,
H., Eds.; Wiley-VCH: Weinheim, Germany, 2005. (f) Wessjohann,
L. A.; Ruijter, E. Mol. DiVersity 2005, 9, 159. (g) Ramón, D. J.;
Yus, M. Angew. Chem., Int. Ed. 2005, 44, 1602. (h) Dömling, A.
Chem. ReV. 2006, 106, 17. (i) Dondoni, A.; Massi, A. Acc. Chem.
Res. 2006, 39, 451. (j) Lieby-Muller, F.; Simon, C.; Constantieux,
T.; Rodriguez, J. QSAR Comb. Sci. 2006, 25, 432. (k) Guillena, G.;
Ramón, D. J.; Yus, M. Tetrahedron: Asymmetry 2007, 18, 693. (l)
Groenendaal, B.; Ruijter, E.; Orru, R. V. A. Chem. Commun. 2008,
5474. (m) Ismabery, N.; Lavilla, R. Chem.sEur. J. 2008, 14, 8444.
(58) Barrault, J.; Essayem, N.; Guimon, C. Appl. Catal., A 1993, 102,
151.
(59) Baiker, A.; Richarz, W. Tetrahedron Lett. 1977, 1937.
(60) Baiker, A.; Richarz, W. Ind. Eng. Chem., Prod. Res. DeV. 1977, 16,
261.
(61) Baiker, A.; Richarz, W. HelV. Chim. Acta 1978, 61, 1169.
(62) (a) Baiker, A.; Richarz, W. Stud. Surf. Sci. Catal. 1981, 7B, 1428.
(b) Baiker, A.; Caprez, W.; Holstein, W. L. Ind. Eng. Chem., Prod.
Res. DeV. 1983, 22, 217.
(63) Baiker, A.; Richarz, W. Synth. Commun. 1978, 8, 27.
(64) Runeberg, J.; Baiker, A.; Kijenski, J. Appl. Catal. 1985, 17, 309.
(65) Vultier, R.; Baiker, A.; Wokaun, A. Appl. Catal. 1987, 30, 167.
(66) (a) Hammerschmidt, W.; Baiker, A.; Wokaun, A.; Fluhr, W. Appl.
Catal. 1986, 20, 305. (b) Kijeński, J.; Niedzielski, P. J.; Baiker, A.
Appl. Catal. 1989, 53, 107.
(67) Card, R. J.; Schmitt, J. L. J. Org. Chem. 1981, 46, 754.
(68) Göbölös, S.; Hegedüs, M.; Kolosova, I.; Maciejewski, M.; Margitfalvi,
J. L. Appl. Catal., A 1988, 169, 201.
(69) Yamakawa, T.; Tsuchiya, I.; Mitsuzuka, D.; Ogawa, T. Catal.
Commun. 2004, 5, 291.
(70) Abe, H.; Yokota, Y.; Okabe, K. Appl. Catal. 1989, 52, 171.
(71) Kimura, H.; Taniguchi, H. Appl. Catal., A 2005, 287, 191.
(72) Kimura, H.; Matsutani, K.; Tsutsumi, S.-i.; Nomura, S.; Ishikawa,
K.; Hattori, Y.; Itahashi, M.; Hoshino, H. Catal. Lett. 2005, 99, 119.
(73) Kimura, H.; Ishikawa, K.; Nishino, K.; Nomura, S. Appl. Catal., A
2005, 286, 120.
(74) Kimura, H.; Tsutsumi, S.-i.; Tsukada, K. Appl. Catal., A 2005, 292,
281.
(75) Kimura, H.; Yokota, Y.; Sawamoto, Y. Catal. Lett. 2005, 99, 133.
(76) Kimura, H.; Taniguchi, H. Catal. Lett. 1996, 40, 123.
(77) Hamada, H.; Kuwahara, Y.; Sato, T.; Wakabayashi, K. Bull. Chem.
Soc. Jpn. 1987, 60, 55.
(78) Richardson, J. T.; Lu, W.-C. J. Catal. 1976, 42, 275.
(79) Ohtani, B.; Nakagawa, K.; Nishimoto, S.-i.; Kagiya, T. Chem. Lett.
1986, 1917.
1640 Chemical Reviews, 2010, Vol. 110, No. 3
(80) Martı́nez, R.; Ramón, D. J.; Yus, M. Tetrahedron Lett. 2005, 46,
8471.
(81) Ohtani, B.; Osaki, H.; Nishimoto, S.-i.; Kagiya, T. Tetrahedron Lett.
1986, 27, 2019.
(82) Ohtani, B.; Osaki, H.; Nishimoto, S.-i.; Kagiya, T. J. Am. Chem.
Soc. 1986, 108, 308.
(83) Martı́nez, R.; Ramón, D. J.; Yus, M. Org. Biomol. Chem. 2009, 7,
2176.
(84) For the recent use of commercial magnetite as soft Lewis acid catalyst,
see: (a) Martı́nez, R.; Ramón, D. J.; Yus, M. AdV. Synth. Catal. 2008,
350, 1235. (b) Mojtahedi, M. M.; Abaee, M. S.; Alishiri, T.
Tetrahedron Lett. 2009, 50, 2322.
(85) For other examples of metal-impregnated magnetite derivatives; see:
(a) Basińska, A.; Jóźwiak, W. K.; Gǒralski, J.; Domka, F. Appl.
Catal., A 2000, 190, 107. (b) Basińska, A.; Klimkiewicz, R.; Domka,
F. Appl. Catal., A 2001, 207, 287. (c) Wang, Z.; Shen, B.; Aihua,
Z.; he, N. Chem. Eng. J. 2005, 113, 27. (d) Uheida, A.; Iglesias, M.;
Fontàs, C.; Hidalgo, M.; Salvadó, V.; Zhang, Y.; Muhammed, M. J.
Colloid Interface Sci. 2006, 301, 402. (e) Kotani, M.; Koike, T.;
Yamaguchi, K.; Mizuno, N. Green Chem. 2006, 8, 735.
(86) Shi, F.; Tse, M. K.; Zhou, S.; Pohl, M.-M.; Radnik, J.; Hübner, S.;
Jähnisch, K.; Brückner, A.; Beller, M. J. Am. Chem. Soc. 2009, 131,
1775.
(87) Kindler, K. Liebigs Ann. Chem. 1931, 485, 113.
(88) Murahashi, S.-I.; Shimamura, T.; Moritani, I. J. Chem. Soc., Chem.
Commun. 1974, 931.
(89) Sabatier, P.; Mailhe, A. C. R. Chim. 1909, 148, 898.
(90) (a) Sabatier, P.; Mailhe, A. C. R. Chim. 1911, 153, 160. (b) Sabatier,
P.; Mailhe, A. C. R. Chim. 1912, 153, 1204.
(91) Grigg, R.; Mitchell, T. R. B.; Sutthivaiyakit, S.; Tongpenyai, N.
J. Chem. Soc., Chem. Commun. 1981, 611.
(92) (a) Watanabe, Y.; Tsuji, Y.; Ohsugi, Y. Tetrahedron Lett. 1981, 22,
2667. (b) Watanabe, Y.; Tsuji, Y.; Ige, H.; Ohsugi, Y.; Ohta, T. J.
Org. Chem. 1984, 49, 3359.
(93) Lazier, W. A.; Adkins, H. J. Am. Chem. Soc. 1924, 46, 741.
(94) Sprinzak, Y. J. Am. Chem. Soc. 1956, 78, 3207.
(95) Sprinzak, Y. In Organic Synthesis, CollectiVe Volume 4; Rabjohn,
N., Ed.; John Wiley & Sons: New York, 1963; Vol. 4, pp 91-92.
(96) Zvezdina, E. A.; Pozharskii, A. F.; Sokolov, V. I. Chem. Heterocycl.
Compd. 1970, 6, 389.
(97) (a) Miyano, S. Chem. Pharm. Bull. 1965, 13, 1135. (b) Miyano, S.;
Uno, A.; Abe, N. Chem. Pharm. Bull. 1967, 15, 515.
(98) Miyano, S.; Abe, N.; Abe, A. Chem. Pharm. Bull. 1967, 15, 511.
(99) Miyano, S.; Nakao, M. Chem. Pharm. Bull. 1972, 20, 1328.
(100) Horikawa, Y.; Uchino, Y.; Sako, T. Chem. Lett. 2003, 32, 232.
(101) Reid, E. E. Am. Chem. J. 1911, 45, 38.
(102) For reviews on ruthenium in organic chemistry, see: (a) Naota, T.;
Takaya, H.; Murahashi, S.-I. Chem. ReV. 1998, 98, 2599. (b) Trost,
B. M.; Toste, F. D.; Pinkerton, A. B. Chem. ReV. 2001, 101, 2067.
(c) Ritleng, V.; Sirlin, C.; Pfeffer, M. Chem. ReV. 2002, 102, 1731.
(d) Grubbs, R. H. Chem. Eng. News 2003, 81 (36), 112. (e) Ruthenium
in Organic Synthesis; Murahashi, S.-I., Ed.; Wiley-VCH: Weinheim,
Germany, 2004. (f) Ruthenium Catalysts and Fine Chemistry in
Topics in Organometallic Chemistry; Bruneau, C., Dixneuf, P. H.,
Eds.; Springer: Berlin, 2004; Vol. 11. (g) Murahashi, S.-I.; Zhang,
D. Chem. Soc. ReV. 2008, 37, 1490.
(103) Arcelli, A.; Khai, B.-T.; Porzi, G. J. Organomet. Chem. 1982, 235,
93.
(104) (a) Ganguly, S.; Joslin, F. L.; Roundhill, D. M. Inorg. Chem. 1989,
28, 4562. (b) Ganguly, S.; Roundhill, D. M. Polyhedron 1990, 9,
2517.
(105) Yamaguchi, I.; Sakano, T.; Ishii, H.; Osakada, K.; Yamamoto, T. J.
Organomet. Chem. 1999, 584, 213.
(106) (a) Marsella, J. A. J. Org. Chem. 1987, 52, 467. (b) Marsella, J. A.
J. Organomet. Chem. 1991, 407, 97.
(107) Tsuji, Y.; Huh, K.-T.; Ohsugi, Y.; Watanabe, Y. J. Org. Chem. 1985,
50, 1365.
(108) Shim, S. G.; Doh, C. H.; Woo, B. W.; Kim, H. S.; Huh, K. T.; Park,
W. H.; Lee, H. J. Mol. Catal. 1990, 62, L11.
(109) Felföldi, K.; Klyavlin, M. S.; Bartók, M. J. Organomet. Chem. 1989,
362, 193.
(110) Tsuji, Y.; Yokoyama, Y.; Huh, K.-T.; Watanabe, Y. Bull. Chem. Soc.
Jpn. 1987, 60, 3456.
(111) (a) Tsuji, Y.; Huh, K.-T.; Watanabe, Y. Tetrahedron Lett. 1986, 27,
377. (b) Tsuji, Y.; Huh, K.-T.; Watanabe, Y. J. Org. Chem. 1987,
52, 1673.
(112) For examples of the indirect Friedländer synthesis of quinolines using
an usual hydrogen transfer process, see: (a) Cho, C. S.; Kim, B. T.;
Kim, T.-J.; Shim, S. C. Chem. Commun. 2001, 2576. (b) Cho, C. S.;
Kim, B. T.; Choi, H.-J.; Kim, T.-J.; Shim, S. C. Tetrahedron 2003,
59, 7997. (c) Motokura, K.; Mizugaki, T.; Ebitani, K.; Kaneda, K.
Tetrahedron Lett. 2004, 45, 6029. (d) Taguchi, K.; Sakaguchi, S.;
Ishii, Y. Tetrahedron Lett. 2005, 46, 4539. (e) Cho, C. S.; Ren, W. X.;
Guillena et al.
Shim, S. C. Bull. Korean Chem. Soc. 2005, 26, 1286. (f) Cho, C. S.;
Ren, W. X.; Shim, S. C. Bull. Korean Chem. Soc. 2005, 26, 2038.
(g) Cho, C. S.; Seok, H. J.; Shim, S. C. J. Heterocycl. Chem. 2005,
42, 1219. (h) Cho, C. S.; Ren, W. X.; Shim, S. C. Tetrahedron Lett.
2006, 47, 6781. (i) Martı́nez, R.; Ramón, D. J.; Yus, M. Eur. J. Org.
Chem. 2007, 1599. (j) Vander Mierde, H.; Ledoux, N.; Allaert, B.;
Van Der Voort, P.; Drozdzak, R.; De Vos, D.; Verpoort, F. New.
J. Chem. 2007, 31, 1572. (k) Cho, C. S.; Ren, W. X. J. Organomet.
Chem. 2007, 629, 4182. (l) Vander Mierde, H.; Van Der Voort, P.;
De Vos, D.; Verpoort, F. Eur. J. Org. Chem. 2008, 1625. (m)
Martı́nez, R.; Ramón, D. J.; Yus, M. J. Org. Chem. 2008, 73, 9778.
(n) Vander Mierde, H.; Van Der Voort, P.; De Vos, D.; Verpoort, F.
Tetrahedron Lett. 2008, 49, 6893. (o) Vander Mierde, H.; Van Der
Voort, P.; De Vos, D.; Verpoort, F. Tetrahedron Lett. 2009, 50, 201.
For a recent review on general Friedländer synthesis, see: (p) Marco-
Contelles, J.; Pérez-Mayoral, E.; Samadi, A.; Carreiras, M. C.;
Soriano, E. Chem. ReV. 2009, 109, 2653.
(113) (a) Tsuji, Y.; Huh, K.-T.; Yokoyama, Y.; Watanabe, Y. J. Chem.
Soc., Chem. Commun. 1986, 1575. (b) Tsuji, Y.; Kotachi, S.; Huh,
K.-T.; Yokoyama, Y.; Watanabe, Y. J. Org. Chem. 1990, 55, 580.
(114) Kondo, T.; Yang, S.; Huh, K.-T.; Kobayashi, M.; Kotachi, S.;
Watanabe, Y. Chem. Lett. 1991, 1275.
(115) Cho, C. S.; Oh, S. G. Tetrahedron Lett. 2006, 47, 5633.
(116) Blacker, A. J.; Farah, M. M.; Hall, M. I.; Marsden, S. P.; Saidi, O.;
Williams, J. M. J. Org. Lett. 2009, 11, 2039.
(117) Kondo, T.; Kotachi, S.; Ogino, S.-i.; Watanabe, Y. Chem. Lett. 1993,
1317.
(118) Watanabe, Y.; Ohta, T.; Tsuji, Y. Bull. Chem. Soc. Jpn. 1983, 56,
2647.
(119) Murahashi, S.-I.; Kondo, K.; Hakata, T. Tetrahedron Lett. 1982, 23,
229.
(120) Murahashi, S.-I.; Naota, T.; Saito, E. J. Am. Chem. Soc. 1986, 108,
7846.
(121) Murahashi, S.-I.; Naota, T.; Ito, K.; Maeda, Y.; Taki, H. J. Org. Chem.
1987, 52, 4319.
(122) Owston, N. A.; Nixon, T. D.; Parker, A. J.; Whittlesey, M. K.;
Williams, J. M. J. Synthesis 2009, 1578.
(123) Cho, C. S.; Kim, J. H.; Kim, T.-J.; Shim, S. C. Tetrahedron 2001,
57, 3321.
(124) Naota, T.; Murahashi, S.-i. Synlett 1991, 693.
(125) Watanabe, Y.; Morisaki, Y.; Kondo, T.; Mitsudo, T.-a. J. Org. Chem.
1996, 61, 4214.
(126) Del Zotto, A.; Baratta, W.; Sandri, M.; Verardo, G.; Rigo, P. Eur.
J. Inorg. Chem. 2004, 524.
(127) Naskar, S.; Bhattacharjee, M. Tetrahedron Lett. 2007, 48, 3367.
(128) Abbenhuis, R. A. T. M.; Boersma, J.; van Koten, G. J. Org. Chem.
1998, 63, 4282.
(129) Gunanathan, C.; Milstein, D. Angew. Chem., Int. Ed. 2008, 47, 8661.
(130) (a) Gunanathan, C.; Ben-David, Y.; Milstein, D. Science 2007, 317,
790. (b) Ulrik, L.; Vogt, H.; Madsen, R. J. Am. Chem. Soc. 2008,
130, 17672. (c) Zweifel, T.; Naubron, J.-V.; Grützmacher, H. Angew.
Chem., Int. Ed. 2009, 48, 559.
(131) Gunanathan, C.; Shimon, L. J. W.; Milstein, D. J. Am. Chem. Soc.
2009, 131, 3146.
(132) Huh, K.-T.; Tsuji, Y.; Kobayashi, M.; Okuda, F.; Watanabe, Y. Chem.
Lett. 1988, 449.
(133) Bitsi, G.; Schleiffer, E.; Antoni, F.; Jenner, G. J. Organomet. Chem.
1989, 373, 343.
(134) Huh, K.-T.; Shim, S. C.; Doh, C. H. Bull. Korean Chem. Soc. 1990,
11, 45.
(135) Shim, S. C.; Doh, C. H.; Yoon, J. H.; Lee, D. Y.; Youn, Y. Z. Bull.
Korean Chem. Soc. 1991, 12, 649.
(136) Tsuji, T.; Nishimura, H.; Huh, K.-T.; Watanabe, Y. J. Organomet.
Chem. 1985, 286, C44.
(137) Cho, C. S.; Lim, H. Y.; Shim, S. C.; Kim, T. J.; Choi, H.-J. Chem.
Commun. 1998, 995.
(138) Tanaka, N.; Hatanaka, M.; Watanabe, Y. Chem. Lett. 1992, 575.
(139) Jenner, G. J. Mol. Catal. 1989, 55, 241.
(140) (a) Tillack, A.; Hollman, D.; Michalik, D.; Beller, M. Tetrahedron
Lett. 2006, 47, 8881. (b) Hollman, D.; Tillack, A.; Michalik, D.;
Jackstell, R.; Beller, M. Chem. Asian J. 2007, 2, 403.
(141) Tillack, A.; Hollman, D.; Mevius, K.; Michalik, D.; Bähn, S.; Beller,
M. Eur. J. Org. Chem. 2008, 4745.
(142) Jenner, G.; Bitsi, G. J. Mol. Catal. 1988, 45, 165.
(143) Hamid, M. H. S. A.; Williams, J. M. J. Tetrahedron Lett. 2007, 48,
8263.
(144) Hamid, M. H. S. A.; Williams, J. M. J. Chem. Commun. 2007, 725.
(145) Hamid, M. H. S. A.; Allen, C. L.; Lamb, G. W.; Maxwell, A. C.;
Maytum, H. C.; Watson, A. J. A.; Williams, J. M. J. J. Am. Chem.
Soc. 2009, 131, 1766.
(146) Pridmore, S. J.; Slatford, P. A.; Daniel, A.; Whittlesey, M. K.;
Williams, J. M. J. Tetrahedron Lett. 2007, 48, 5115.
Hydrogen Autotransfer in the N-Alkylation of Amines
(147) Lamb, G. W.; Watson, A. J. A.; Jolley, K. E.; Maxwell, A. C.;
Williams, J. M. J. Tetrahedron Lett. 2009, 50, 3374.
(148) Watson, A. J. A.; Maxwell, A. C.; Williams, J. M. J. Org. Lett. 2009,
11, 2667.
(149) For reviews on iridium complexes in different hydrogen autotransfer
reactions, see: (a) Ishii, Y.; Sakaguchi, S. Bull. Chem. Soc. Jpn. 2004,
77, 909. (b) Fujita, K.-i.; Yamaguchi, R. Synlett 2005, 560.
(150) Fujita, K.-i.; Li, Z.; Ozeki, N.; Yamaguchi, R. Tetrahedron Lett. 2003,
44, 2687.
(151) Fujita, K.-i.; Enoki, Y.; Yamaguchi, R. Tetrahedron 2008, 64, 1943.
(152) Fujita, K.-i.; Kida, Y.; Yamaguchi, R. Heterocycles 2009, 77, 1371.
(153) Yamaguchi, R.; Kawagoe, S.; Asai, C.; Fujita, K.-i. Org. Lett. 2008,
10, 181.
(154) (a) Fujita, K.-i.; Fuji, T.; Yamaguchi, R. Org. Lett. 2004, 6, 3525.
(b) Fujita, K.-i.; Enoki, Y.; Yamaguchi, R. In Organic Synthesis;
Curran, D. P., Ed.; John Wiley & Sons: Hoboken, NJ, 2006; Vol.
83, pp 217-221.
(155) Miao, L.; DiMaggio, S. C.; Shu, H.; Trudell, M. L. Org. Lett. 2009,
11, 1579.
(156) Fujita, K.-i.; Yamamoto, K.; Yamaguchi, R. Org. Lett. 2002, 4, 2691.
(157) Eary, C. T.; Clausen, D. Tetrahedron Lett. 2006, 47, 6899.
(158) Nordstrøm, L. U.; Madsen, R. Chem. Commun. 2007, 5034.
(159) Balcells, D.; Nova, A.; Clot, E.; Gnanamgari, D.; Crabtree, R. H.;
Eisenstein, O. Organometallics 2008, 27, 2529.
(160) Fujita, K.-i.; Komatsubara, A.; Yamaguchi, R. Tetrahedron 2009,
65, 3624.
(161) Cami-Kobei, G.; Slatford, P. A.; Whittlesey, M. K.; Williams, J. M. J.
Bioorg. Med. Chem. Lett. 2005, 15, 535.
(162) Blank, B.; Madalska, M.; Kempe, R. AdV. Synth. Catal. 2008, 350,
749.
(163) Blank, B.; Michlik, S.; Kempe, R. Chem.sEur. J. 2009, 15, 3790.
(164) Cami-Kobeci, G.; Williams, J. M. J. Chem. Commun. 2004, 1072.
(165) Izumi, A.; Obara, Y.; Sakaguchi, S.; Ishii, Y. Tetrahedron Lett. 2006,
47, 9199.
(166) Pontes da Costa, A.; Viciano, M.; Sanaú, M.; Merino, S.; Tejeda, J.;
Peris, E.; Royo, B. Organometallics 2008, 27, 1305.
(167) Prades, A.; Corbeŕan, R.; Poyatos, M.; Peris, E. Chem.sEur. J. 2008,
14, 11474.
(168) Liu, S.; Rebros, M.; Stephens, G.; Marr, A. C. Chem. Commun. 2009,
2308.
(169) Gnanamgari, D.; Sauer, E. L.; Schley, N. D.; Butler, C.; Incarvito,
C. D.; Crabtree, R. H. Organometallics 2009, 28, 321.
(170) Aramoto, H.; Obara, Y.; Ishii, Y. J. Org. Chem. 2009, 74, 628.
(171) Shi, F.; Tse, M. K.; Cui, X.; Gördes, D.; Michalik, D.; Thurow, K.;
Deng, Y.; Beller, M. Angew. Chem., Int. Ed. 2009, 48, 5912.
(172) Tsuji, Y.; Takeuchi, R.; Ogawa, H.; Watanabe, Y. Chem. Lett. 1986,
293.
(173) Snyder, H. R.; Carnahan, R. E.; Lovejoy, E. R. J. Am. Chem. Soc.
1954, 76, 1301.
(174) For a review on oxidation of amines, see: Murahashi, S.-I. Angew.
Chem., Int. Ed. Engl. 1995, 34, 2443.
(175) Issoire, J.; van Long, C. Bull. Soc. Chim. Fr. 1960, 2004.
(176) Borrows, E. T.; Hargreaves, B. M. C.; Page, J. E.; Resuggan, J. C. L.;
Robinson, F. A. J. Chem. Soc. 1947, 197.
(177) Klindler, K.; Melamed, G.; Matthies, D. Liebigs Ann. Chem. 1961,
644, 23.
(178) De Angelis, F.; Grgurina, I.; Nicoletti, R. Synthesis 1979, 70.
(179) Pommersheim, J. M.; Coull, J. AIChE J. 1971, 17, 1075.
(180) Kijeński, J.; Burger, J.; Baiker, A. Appl. Catal. 1984, 11, 295.
(181) Baiker, A.; Monti, D.; Fan, Y. S. J. Catal. 1984, 88, 81.
Chemical Reviews, 2010, Vol. 110, No. 3 1641
(182) Rosenmund, K. W.; Jordan, G. Ber. Dtsch. Chem. Ges. 1925, 58B,
51.
(183) Anderson, J. R.; Clark, N. J. J. Catal. 1966, 5, 250.
(184) Yoshimura, N.; Moritani, I.; Shimamura, T.; Murahashi, S.-I. J. Am.
Chem. Soc. 1973, 95, 3038.
(185) Yoshimura, N.; Moritani, I.; Shimamura, T.; Murahashi, S.-I. J. Chem.
Soc., Chem. Commun. 1973, 307.
(186) Murahashi, S.-I.; Hirano, T.; Yano, T. J. Am. Chem. Soc. 1978, 100,
348.
(187) Murahashi, S.-I.; Yoshimura, N.; Tsumiyama, T.; Kojima, T. J. Am.
Chem. Soc. 1983, 105, 5002.
(188) Murahashi, S.-I.; Watanabe, T. J. Am. Chem. Soc. 1979, 101, 7429.
(189) Miyazawa, A.; Tanaka, K.; Sakakura, T.; Tashiro, M.; Tashiro, H.;
Prakash, G. K. S.; Olah, G. A. Chem. Commun. 2005, 2104.
(190) Sajiki, H.; Ikawa, T.; Hirota, K. Org. Lett. 2004, 6, 4977.
(191) Nishimoto, S.-i.; Ohtani, B.; Yoshikawa, T.; Kagiya, T. J. Am. Chem.
Soc. 1983, 105, 7180.
(192) Miyazawa, A.; Saitou, K.; Tanaka, K.; Gädda, T. M.; Tashiro, M.;
Prakash, G. K. S.; Olah, G. A. Tetrahedron Lett. 2006, 47, 1437.
(193) Largeron, M.; Fleury, M.-B. Org. Lett. 2009, 11, 883.
(194) Ballantine, J. A.; Purnell, H.; Rayanakorn, M.; Thomas, J. M.;
Williams, K. J. J. Chem. Soc., Chem. Commun. 1981, 9.
(195) Greenfield, H. J. Org. Chem. 1964, 29, 3082.
(196) Murahashi, S.-I.; Yano, T. J. Am. Chem. Soc. 1980, 102, 2456.
(197) Khai, B.-T.; Concilio, C.; Porzi, G. J. Organomet. Chem. 1981, 208,
249.
(198) Arcelli, A.; Khai, B.-T.; Porzi, G. J. Organomet. Chem. 1982, 231,
C31.
(199) Khai, B.-T.; Concilio, C.; Porzi, G. J. Org. Chem. 1981, 46, 1759.
(200) Jung, C. W.; Fellmann, J. D.; Garrou, P. E. Organometallics 1983,
2, 1042.
(201) Cho, C. S.; Kim, J. S.; Oh, B. H.; Kim, T.-J.; Shim, S. C.; Yoon,
N. S. Tetrahedron 2000, 56, 7747.
(202) Cho, C. S.; Kim, T. K.; Kim, B. T.; Kim, T.-J.; Shim, S. C. J.
Organomet. Chem. 2002, 650, 65.
(203) Cho, C. S.; Oh, B. H.; Shim, S. C. Tetrahedron Lett. 1999, 40, 1499.
(204) Cho, C. S.; Oh, B. H.; Kim, J. S.; Kim, T.-J.; Shim, S. C. Chem.
Commun. 2000, 1885.
(205) Laine, R. M. J. Mol. Catal. 1983, 21, 119.
(206) Wilson, R. B.; Laine, R. M. J. Am. Chem. Soc. 1985, 107, 361.
(207) Shvo, Y.; Thomas, D. W.; Laine, R. M. J. Am. Chem. Soc. 1981,
103, 2461.
(208) Hollmann, D.; Bähn, S.; Tillack, A.; Beller, M. Angew. Chem., Int.
Ed. 2007, 46, 8291.
(209) Hollmann, D.; Bähn, S.; Tillack, A.; Beller, M. Chem. Commun. 2008,
3199.
(210) Hollmann, D.; Bähn, S.; Tillack, A.; Parton, R.; Altink, R.; Beller,
M. Tetrahedron Lett. 2008, 49, 5742.
(211) Bähn, S.; Hollmann, D.; Tillack, A.; Beller, M. AdV. Synth. Catal.
2008, 350, 2099.
(212) Shvo, Y.; Laine, R. M. J. Chem. Soc., Chem. Commun. 1980, 753.
(213) Laine, R. M.; Thomas, D. W.; Cary, L. W. J. Am. Chem. Soc. 1982,
104, 1763.
(214) Laine, R. M.; Thomas, D. W.; Cary, L. W.; Buttrill, S. E. J. Am.
Chem. Soc. 1978, 100, 6527.
(215) Tsuji, Y.; Shida, J.; Takeuchi, R.; Watanabe, Y. Chem. Lett. 1984,
889.
CR9002159
1642 Chem. Rev. 2010, 110, 1642–1662

Mechanisms for DNA Charge Transport

Joseph C. Genereux and Jacqueline K. Barton*

Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91125

Received June 25, 2009

Contents
1. Introduction
2. Properties of Long-Range Charge Transport in
DNA
2.1. Coupling to the DNA
2.2. Global Structural Integrity
2.3. Local Structural Integrity
2.4. Conformational Gating
2.5. Back Electron Transfer
2.6. Injection and Migration Effects
2.7. Energetics
2.8. Proton-Coupled Electron Transfer
2.9. Characteristics of DNA CT
3. DNA-Mediated CT Mechanisms
3.1. Transport through Water, Ions, and
Phosphates
3.2. Superexchange
3.2.1. Coupling Constants in DNA
3.2.2. Reorganization Energy
3.2.3. Superexchange in DNA
3.3. Localized Hopping
3.3.1. Nearest-Neighbor Models
3.3.2. Thermally Induced Hopping
3.3.3. Variable-Range Models
3.4. Delocalized Mechanisms
3.4.1. Polaron Hopping and Gating Mechanisms
3.4.2. Domain Delocalization
4. Summary
5. Unanswered Questions
6. Acknowledgments
7. References
1. Introduction
DNA charge transport (CT) chemistry has received
considerable attention by scientific researchers over the past
15 years since our first provocative publication on long-range
CT in a DNA assembly.1,2 This interest, shared by physicists,
chemists, and biologists, reflects the potential of DNA CT
to provide a sensitive route for signaling, whether in the
construction of nanoscale biosensors or as an enzymatic tool
to detect damage in the genome. Research into DNA CT
chemistry began as a quest to determine whether the DNA
double helix, a macromolecular assembly in solution with
π-stacked base pairs, might share conductive characteristics
with π-stacked solids. Physicists carried out sophisticated
experiments to measure the conductivity of DNA samples,
* To whom correspondence should be addressed. E-mail: jkbarton@
caltech.edu.
10.1021/cr900228f  2010 American Chemical Society
Published on Web 11/23/2009
1642
1643
1643
1644
1645
1646
1647
1648
1648
1650
1651
1651 Joseph C. Genereux was born in Burbank, CA, in 1981. He received his
1651 B.A. with high honors in chemistry from Swarthmore College in 2001,
where he worked in the laboratory of Prof. Robert F. Pasternack. He
received a B.S. in physics from the University of CaliforniasIrvine in 2002.
1652 He is currently a graduate student in the research group of Prof. Jacqueline
1652 K. Barton at Caltech, studying DNA-mediated charge transport.
1653 but the means to connect discrete DNA assemblies into the
1653 devices to gauge conductivity varied, as did the conditions
1653 under which conductivities were determined. Chemists
1653 constructed DNA assemblies to measure hole and electron
1654 transport in solution using a variety of hole and electron
1655 donors. Here, too, DNA CT was seen to depend upon the
1656 connections, or coupling, between donors and the DNA base
1656 pair stack. Importantly, these experiments have resolved the
debate over whether DNA CT is possible. Moreover, these
1657
studies have shown that DNA CT, irrespective of the oxidant
1658 or reductant used to initiate the chemistry, can occur over
1658 long molecular distances but can be exquisitely sensitive to
1659 perturbations in the base pair stack.
1659 Here we review some of the critical characteristics of DNA
charge transport chemistry, taking examples from a range
of systems, and consider these characteristics in the context
of their mechanistic implications. This review is not intended
to be exhaustive but instead to be illustrative. For instance,
we describe studies involving measurements in solution using
pendant photooxidants to inject holes, conductivity studies
with covalently modified assemblies, and electrochemical
studies on DNA-modified electrodes. We do not focus in
detail on the differences among these constructs but instead
on their similarities. It is the similarity among these various
systems that allows us to consider different mechanisms to
describe DNA CT. Thus, we review also the various
mechanisms for DNA CT that have been put forth and
attempt to reconcile these mechanistic proposals with the
many disparate measurements of DNA CT. Certainly the
debate among researchers has shifted from “Is DNA CT
possible?” to “How does it work?”. In this review we explore
this latter question in detail.
Mechanisms for DNA Charge Transport Chemical Reviews, 2010, Vol. 110, No. 3 1643

Dr. Jacqueline K. Barton is the Arthur and Marian Hanisch Memorial

Professor of Chemistry and Chair of the Division of Chemistry and
Chemical Engineering at the California Institute of Technology. Barton
was awarded the A.B. summa cum laude at Barnard College in 1974
and a Ph.D. at Columbia University in 1978. After a postdoctoral fellowship
at Bell Laboratories and Yale University, she became an assistant professor
at Hunter College, City University of New York. In 1983, she returned to
Columbia University, becoming an associate professor of chemistry and
biological sciences in 1985 and a professor in 1986. In the fall of 1989,
she joined the faculty at Caltech. In 2009, she began her term as Chair
of the Division. Prof. Barton has pioneered the application of transition-
metal complexes to probe recognition and reactions of double-helical DNA,
in particular studies to elucidate electron-transfer chemistry mediated by
the DNA double helix. Barton has received numerous awards. These
include the NSF Waterman Award, the ACS Award in Pure Chemistry,
the ACS Breslow Award in Biomimetic Chemistry, and a MacArthur
Foundation Fellowship. She is a member of the American Academy of
Arts and Sciences, the American Philosophical Society, and the National
Academy of Sciences.

Figure 2. DNA-mediated CT requires electronic coupling to the

base pair stack. (A) Electrochemical reduction of an electronically
well-coupled anthraquinone (AQ) is facile, while that of a poorly
coupled AQ is suppressed.18 (B) MutY competently reduces an
oxidized nitroxide spin label that is well-coupled to thymidine, but
not the nitroxide conjugated through the partially unsaturated
linker.160 (C) For a series of polypyridyl RuIII(bpy)2L ground-state
oxidants, the yield of oxidative damage to DNA scales with the
size and planarity of the intercalating ligand.15

Figure 1. Transverse and longitudinal perspectives of DNA. The
sugar phosphate backbone envelops the hydrophobic base pairs.
The planar base pairs form a one-dimensional π-stack down the
center of the DNA, insulated by the backbone.
2. Properties of Long-Range Charge Transport in
DNA
Among the most interesting characteristics of charge
transport in DNA is the long distance over which it occurs
(Figure 1).3-6 Nevertheless, there are some DNA systems
that do not mediate charge over long distances. How DNA
CT occurs depends upon coupling and the structure and
dynamics of the DNA assembly. The chemistry and photo-
physics of the photoexcited acridine (Acr+*) containing
systems, which mediate CT over only a few base pairs, have
been particularly well-characterized in this regard.7,8 It is
important to note that the same physical laws apply to all
CT processes.9 The essential distinctions are with respect to
the relative roles which different mechanisms play, and it is
in this respect that long-range CT, with effective coupling
to the base pair stack, differs from short-range CT, with poor
coupling. Here we focus on long-range CT, where transport
is through the base pair stack, and discuss short-range
systems only to the necessary extent to clarify the distinctions
between the two regimes.
2.1. Coupling to the DNA
It is notable that initial measurements of DNA-mediated
charge transport for both photooxidation experiments10 and
device experiments11 found rates and conductivities spanning
several orders of magnitude over comparable distances,
depending on the experimental conditions. This foreshad-
owed the same observation in scanning tunneling microscopy
(STM) measurements of conductivity through other molec-
ular bridges12 and, ultimately, was for the same reason. For
short molecular bridges, it has been established both experi-
mentally12 and theoretically13 that the coupling between the
bridge and the donor (or acceptor) can dominate the observed
conductivity. Similarly, when DNA is the bridge, the
coupling can have a dramatic effect on both charge-transport
rates and yields (Figure 2).14-18 Characteristically, conductiv-
ity measurements that have not provided covalent contact
1644 Chemical Reviews, 2010, Vol. 110, No. 3
between the DNA and the device yield a spectrum of
behavior: from insulating to superconductive.11,19-23
In the case of DNA, the essential coupling is into the
π-stack of the bases. This is a marked challenge, as DNA is
essentially “insulated”, with sugars and phosphates flanking
the periphery of the bases.24 This insulation, in part, explains
why early experiments on dry DNA found insulating
behavior, in contrast to that observed with conducting organic
polymers. A series of well-conjugated charge donors and
acceptors are now employed by various groups,25,26 including
metallointercalators, organic intercalators, organic end-cap-
pers, and modified bases. In several cases, direct comparison
has been made between similar photooxidant pairs that differ
primarily in their ability to couple well with the base stack.
These examples include the adenine analogues ethenoadenine
and 2-aminopurine (Ap),14 two different coupling strategies
for ethidium bromide,27 and, most notably, a series of
intercalating ruthenium analogues with decreasing planarity
in the intercalating ligand.15 As an extreme case, for two
ruthenium complexes that are unable to intercalate, and that
are attached on opposite ends of a short DNA duplex via
terminal phosphate modification, the CT rate was found to
be ∼10-6 s-1;28 this is what would be expected for the rate
were the metal complexes connected solely through their σ
tethers. Similarly, electrochemiluminescence studies find the
same rates for DNA-mediated CT between a DNA-modified
gold electrode and tethered Ru(bpy)32+ as are observed
through solely the tether itself.29 In electrochemical studies
of methylene blue covalently attached to a DNA duplex,
effective transport is found only when the methylene blue
is stacked in the helix, not under high salt conditions, where
the dye, although still linked by a σ-bonded tether, is
unstacked.30 Indeed, electrochemical measurements on DNA
films generally have been shown to be rate-limited by tether
linking the DNA to the electrode surface.31 In each case, it
is clear that the coupling between the donor/acceptor pair
and the bridge is dominating the measurement and that the
bridge is the π-stack of DNA.
2.2. Global Structural Integrity
The structure of DNA is central to its extraordinary
effectiveness as the genetic template for the cell. This
relationship between structure and function is underscored
by the extent of the biological function that was first predicted
in the landmark papers that reported the proper three-
dimensional structure.32,33 Hence, it is not surprising that
DNA-mediated CT is also substantially affected by the global
structure of a DNA sample. This is clear when considering
the results of conductivity measurements on single or a few
DNA strands that have been performed in recent years.
Various measurements from 1996 to the present have found
DNA conductivities covering several orders of magnitude.
Furthermore, conductivity has been found to be dependent
on the sequence, hydration, length, temperature, and hybrid-
ization in some experiments, while independent of each of
those in others. Ultimately, the vast differences in observa-
tions can be largely reconciled by comparing the sample
preparation methodologies of the individual studies.11 Condi-
tions that cause global DNA conformational changes or
damage can both increase and decrease the observed
conductivity. In one extreme case, it was found that imaging
conditions commonly used prior to conductance measure-
ments led to a morphological change in the structure of the
DNA, which is itself correlated with increased conductivity.20
Genereux and Barton
Among experiments that examine undamaged DNA, a
profound difference is always observed between single-
stranded and double-stranded DNA: single-stranded DNA
does not mediate CT over long distances. This has been
observed by direct conductivity studies,34 photooxidation,35
transient absorption,36 electrochemical AFM,37,38 STM,39
electrogenerated chemiluminescence,29 and electrochemical
experiments in DNA monolayers.40 The caveat in interpreting
studies on single-stranded DNA is, however, that its structure
and, importantly, stacking are heterogeneous and extremely
dependent on the sequence.
DNA, stabilized by a variety of hydrophobic and hydro-
philic interactions and evolved for an aqueous environment,
undergoes gross structural changes as a result of moving from
a hydrated to a dehydrated environment.41 Critically, these
changes are to the equilibrium conformation of DNA; the
effects of dehydration on DNA dynamics are not well-
understood. Highly bound waters play a major role in the
dynamics that gate molecular recognition and other bio-
chemical interactions between macromolecules.42
Regrettably, the first and many recent measurements of
DNA conductivity were performed under a vacuum. A
vacuum is ideal for conductivity measurements due to the
suppression of voltage leak and the associated background
current. Even experiments performed in the presence of water
frequently deposit the DNA under vacuum conditions.
Similar to the previous case of poorly coupled versus well-
coupled systems, there is wide disparity between the
conductivities observed under conditions of low humidity.
Recently, progress has been made in understanding the role
of humidity in many of the poorly coupled systems.43 Even
without strong coupling into the DNA base stack, water
adsorbed on the DNA and in DNA bundles can mediate ionic
conduction. The amount of adsorbed water will depend
strongly on the humidity and also on the adsorption environ-
ment of the DNA. This helps explain why many systems in
which coupling to DNA was poor were still observed to
conduct.19,23,44
Not surprisingly, experiments that have preserved the DNA
in its native conformation, with leads covalently coupled to
the bridge, have shown remarkably similar (and substantial)
conductivities (Figure 3).34,37,44-46 The conductivity measured
by Xu et al. across a dodecanucleotide with terminal
propanethiol-Au contacts (>40 Å) is comparable (6 × 10-4
G0, where G0 is the quantum unit of conductance) to that
found across the much smaller benzenedimethanethiol (∼10
Å) under the same experimental conditions,47 though this
comparison is complicated by the possibility that DNA
accommodates internal stretching during the measurement
rather than extruding gold from the molecular junction, as
is postulated for benzenedimethanethiol.
As is the case with water, the ionic strength can dictate
the conformation of DNA. A high ionic strength drives the
transition from the B-form to the more extended Z-form of
DNA. Poor base stacking, associated with this condensed
structure, leads to less efficient DNA-mediated CT.48 Con-
versely, a sufficiently low ionic strength leads to strand
dehybridization, which also suppresses CT.
Beyond issues of ionic strength, there is conflicting
evidence as to whether the identity of the counterion affects
CT. Some calculations have shown that counterion identity
does not affect the electric field inside the DNA,49 while
others have found that movement of a single sodium has
profound effects on base energies.50 Similarly contrasting
Mechanisms for DNA Charge Transport
Figure 3. Devices for measurement of single-molecule DNA
conductivity through molecular contacts. In each case, currents
between 10 and 100 nA are obtained for modest source-drain and
gating voltages. (A) A gold nanoparticle allows strong coupling
between the EC-AFM tip and an individual 26-mer DNA molecule
on a gold electrode.37 (B) The gold STM tip is slowly brought into
contact with thiol-modified DNA (8-mer), allowing a histogram of
conductance over many different orientations.45 (C) A single 15-
mer DNA is covalently attached across a gap between single-walled
carbon nanotubes.34
results have been observed in experimental work.51-54 For
solvent-exposed donors and acceptors, an ion pair can form
between the dye and counterion that itself profoundly affects
the CT rate and yield.55
More chemically controlled experiments elucidate the
structural basis of environmental effects. RNA/DNA hybrids
and double-stranded RNAs adopt A-form structures, while
alternating purine-pyrimidine sequences under certain con-
ditions adopt Z-form structures. Both conformations support
DNA charge transport, though the Z-form is an inferior
bridge relative to the A-form and B-form for electrochemi-
cal,48 but not photooxidation,56 assays. Not surprisingly, the
competence for mediating CT has been shown to follow the
extent of base stacking, both in solution studies with Ap as
the photooxidant57 and in electrochemical experiments
monitoring the efficiency of reducing an intercalated redox
probe.48 Again, different couplings of the redox probes into
these different conformations mean that they cannot be
quantitatively compared.
Perhaps most interesting is the comparison of rates of
intrastrand versus interstrand base-base CT in DNA as-
semblies modified with Ap.14,35,57 Here for the B-conforma-
tion, intrastrand CT is found to be 3 orders of magnitude
faster than interstrand CT, consistent with the fact that
stacking in the B-conformation is exclusively intrastrand; CT
across strands requires CT across a hydrogen bond. However,
in the A-form there is a mix of interstrand and intrastrand
stacking down the helix, and here we observe that the rates
of intrastrand and interstrand CT are comparable.
2.3. Local Structural Integrity
A variety of studies have found similar effects of disrupting
the base stack locally. The assays include electrochemical
experiments in both films30,58,59 and devices34 and solution
experiments using time-resolved fluorescence,60 irreversible
trapping of the chemical product,61 and transient absorption
measurements.62 The presence of mismatches lowers both
the rate and yield of DNA-mediated CT, and the extent of
this attenuation scales with the base pair lifetime.63 Abasic
Chemical Reviews, 2010, Vol. 110, No. 3 1645
sites64 and destabilizing lesions65 also interfere with CT
through DNA films.
Regarding those experiments that utilize product trapping,
however, it is important to note that the results are convoluted
with two effective clocks. The first is the rate of back electron
transfer (BET), if it occurs.66 The second is the rate of product
trapping.66-69 A disruption of the π-stack will only be
observable in product trapping experiments if it is sufficient
to disrupt equilibration of the radical cation on the time scale
of BET and product trapping.70 Toward this end, guanine
damage assays have recently been replaced by assays for
fast decomposition of a radical trap. N2-Cyclopropylgua-
nosine (CPG),71,72 N6-cyclopropyladenosine (CPA),17,73,74 and
N4-cyclopropylcytidine (CPC)75,76 are synthetically accessible,
cause minimal perturbation to the DNA duplex, and undergo
irreversible picosecond ring-opening upon oxidation or
reduction.
Not all modifications of DNA suppress long-range CT. A
dephosphorylation nick to the backbone, despite causing a
substantial change to the local ion density, does not have a
measurable effect.77,78 Furthermore, some modifications can
enhance CT. Most notably, DNA with adenines replaced by
the lower potential base deazaadenine or the better stacking
benzodeazaadenine improves the rate and yield of DNA-
mediated CT.79-81
In addition to global changes in structural integrity, subtle
modulations to structure can also have profound effects on
the rates and yields of DNA-mediated CT. DNA-mediated
CT is attenuated by the presence of mismatches, even though
mismatched base pairs cause only minor distortions to the
structure of DNA.82,83 Nevertheless, mismatch discrimination
has been observed in charge transport through DNA
films,58,59,84 single-molecule devices,34,85 and photooxidation
systems.35,61 This attenuation is identical for oxidative and
reductive CT,86 implying mechanistic similarity. Importantly,
the extent of mismatch discrimination corresponds to the base
pair lifetime associated with the specific mismatch.63 In the
extreme case, an abasic site completely suppresses CT.64,87,88
Subtle lesions, such as the oxidative guanine products O6-
methylguanine and 8-oxoguanine, attenuate CT.65 It should
be noted that although 8-oxoguanine terminates DNA-
mediated CT as a thermodynamic and kinetic trap at high
driving force,89,90 this and other damaged base products also
attenuate CT even under driving forces incompetent for direct
oxidation. This property of DNA-mediated CT has led to
the development of a new class of DNA-detection devices91,92
and might be relevant to damage detection in the cell.93
Local changes to structure can also be induced. Inside the
cell, proteins can bend, twist, and dehybridize DNA, and
some can extrude bases as well. Not surprisingly, many of
these binding events have severe effects on DNA-mediated
CT. Monitoring DNA-mediated electrochemistry to SoxR,
a transcription factor that initiates the oxidative stress
response in Escherichia coli, is consistent with the prediction
that the oxidized form twists DNA to initiate transcription,94
as has since been validated by a recent crystal structure.95
CT is also inhibited by the sequence-specific binding of
TATA-binding protein (TBP),58,96 which bends DNA. One
particularly informative experiment involved the methylase
M.HhaI, which extrudes a cytidine within its recognition
sequence, replacing it with glutamine. DNA with the HhaI
recognition site shows attenuated CT in the presence of the
protein. The Q237W mutant, in which the intercalating
glutamate is replaced with the aromatic ligand tryptophan,
1646 Chemical Reviews, 2010, Vol. 110, No. 3
however, barely affects CT compared to the absence of
protein.58,97 Importantly, proteins that do not distort the DNA
π-stack, such as antennapedia homeodomain protein or
unactivated R.PVuII, do not attenuate DNA-mediated CT.97
Indeed, the rigidification associated with R.PVuII binding
increased CT through the dynamically flexible TATA binding
site. An interesting exception is the case of R.BamHI, a
restriction endonuclease which does not bend the DNA
π-stack, but contains an asparagine guanidinium that hydro-
gen bonds to a guanine in its cognate site. It has been shown
that R.BamHI attenuates DNA CT to guanines both within
and beyond its binding site, presumably due to electrostatic
modulation of the intervening DNA via the hydrogen-
bonding interactions.98
Protein binding can also affect the fate of radicals in
DNA.99-101 Guanine radical has been shown to cross-link to
histones and short peptides. In one case,102 excitation of an
anthraquinone (AQ)-DNA conjugate bound to a reconsti-
tuted nucleosome particle led to DNA-protein cross-links;
experiments with a different photooxidant with facile back
electron transfer found no such effect.103 Since it is clear
that migration can occur over long distances in both isolated
nuclei104 and mitochondria,105 this might not be a fast
pathway for radical quenching; the time scale of protein-DNA
cross-linking in the presence of guanine radical is similar to
that for guanine radical decomposition to 8-oxoguanine.106
These cross-links are reversible under the processing condi-
tions used to convert base damage to strand cleavage and
hence are likely hidden in typical gel analysis experiments.
Furthermore, protein cross-links can be formed as a second-
ary product after radical degradation in DNA to 8-oxogua-
nine.107 Recently, it has been shown that a protein that
disrupts the π-stack, Hbb of Borrelia burgdorferi, affects
the conversion of radical damage to interstrand cross-links.108
DNA CT is sensitive to even more subtle deviations in
stacking integrity. The strongest stacking interactions occur
between consecutive purines, as has been shown both experi-
mentally and computationally. Extended purine-pyrimidine
runs correspond to the minimal extent of base-stacking, while
purine-purine runs, particularly adenine tracts, correspond to
the maximal extent for DNA containing natural nucleotides.
This relationship is borne out in the sequence dependence of
CT.51,109-111 Note that in photoactivated studies of electron
transport, runs of pyrimidines, which are more easily reduced,
are the preferred sequences.112 Likely here too, the decreased
flexibility of homopurine-homopyrimidine sequences plays
a role.
2.4. Conformational Gating
The rate and efficiency of charge transfer is centrally
related to the structure of the individual pathway(s) that
mediates CT between the donor and acceptor. Over long
distances, however, it is inevitable that fluctuations will be
induced at nonzero temperature, such that the equilibrium
structure only reflects an average over the ensemble. If these
fluctuations are sufficiently large and slower than CT for the
equilibrium conformation, then CT no longer is properly
described by a unitary rate constant, but will instead proceed
with a complicated time dependence that convolutes the
conformational dynamics with the CT rates of the different
conformations.113 To avoid this, the best performing molec-
ular wires are chosen for, among other properties, rigid
homogeneous conformations.114 For example, the conduction
of certain oligo(phenyleneethynylene)s can be substantially
Genereux and Barton
Figure 4. Sequence and temperature dependence of single-step
oxidation of guanine by photoexcited 2-aminopurine (Ap) in
DNA.35,54,117 The CT yield in well-matched ApA4G increases with
temperature (+∆), up to duplex melting. Two perturbations that
disturb CT due to poor stacking dynamics, an AA mismatch and
the sequence ATAT, attenuate CT at room temperature but are
comparable to the A4 sequence at higher temperature, while CT
through a perturbation that disturbs CT due to an electronic barrier,
AAIA, is only partially recovered at high temperature. This argues
that the CT activation is related to the flexibility of the bridge. At
low temperature (77 K), an intervening adenine eliminates CT from
Ap to G, implying that the equilibrium conformation is not the
CT-active conformation.
enhanced by the presence of bulky side groups that limit
the conjugation-breaking rotation around the acetylene
bonds.115,116
It is not surprising that DNA, which even as a relatively
short 15-mer contains several hundred atoms, exhibits
substantial conformational flexibility. Interestingly, the effect
of conformational gating in DNA is generally to increase
CT rather than to decrease it. Duplexes frozen in glass show
no attenuation in CT between the photooxidant Ap and
neighboring guanine hole acceptor (Figure 4).117 The inser-
tion of a single adenine, however, between the donor and
acceptor completely suppresses CT. The temperature depen-
dence of CT between Ap* and G for varying bridge length
has been studied by both femtosecond transient absorption
spectroscopy35 and steady-state fluorescence quenching mea-
surements.54 The temperature dependence has also been
measured for CT between tethered, photoexcited
[Rh(phi)2(bpy′)]3+ and CPG.68 For all bridges, CT efficiency
increases with temperature, and the temperature dependence
is greater with increasing bridge length.
Although thermal activation on its own does not imply a
mechanism, the sequence dependence establishes a strong
relationship between the bridge structure and the activated
process. The temperature dependence of the electronic
contribution to the rate should be the same irrespective of
the increasing bridge length. The temperature dependence
of the nuclear contribution, due to bridge length dependences
on the driving force and reorganization energy (see below),
can naturally be bridge-length-dependent, but should disap-
pear for distances above 10-15 Å,118 which are exceeded
here. Conformational gating to reach a CT-active state,
however, is expected to lead to increased CT with temper-
Mechanisms for DNA Charge Transport Chemical Reviews, 2010, Vol. 110, No. 3 1647

ature, and it is also expected that this increase will be greater

when more bridge units are required to align, i.e., for longer
bridges.
Interestingly, this increase in the rate with respect to
increasing temperature is even more profound when the
bridge is poorly stacked ATAT54 or contains an AA
mismatch,35 consistent with the model whereby these se-
quences disrupt DNA due to poor dynamic stacking (Figure
4). When the bridge is AAIA, which stacks well but
attenuates CT from Ap* due to the inosine potential barrier,
the increase in CT with respect to temperature is only modest.
These experiments strongly suggest that the equilibrium
conformation of DNA is not the active conformation for
long-distance CT, but that conformational gating allows the
formation of CT-active states. This is in direct contrast to
the usual role dynamic disorder plays in molecular wires,
where distortion from the equilibrium conformation decreases
coupling and transport.119
There are two different implications of conformational
gating. In one sense, reorientation of the photooxidant with
respect to the DNA to form a CT-active conformation can Figure 5. (A) The rate of coherent deazaguanine (ZG) oxidation
be rate-limiting. Alternatively, formation of a CT-active state by ethidium bromide is the same over short distances for both the
flexible linkage (Et1, black, both CT rates shown) and the rigid
in the DNA itself can be the rate-limiting step for CT. Both linkage (Et2, gray).27 For two intervening nucleotides, a sharp drop
senses are represented by the case of ethidium bromide. in rate is observed for the rigid Et+, but the rate is unaffected for
Ethidium bromide (Et+) is competent for DNA-mediated the flexible Et+. (B) This steep drop in rate over short distances is
oxidation and reduction of deazaguanine (ZG) and consistent with that observed for oxidation of guanine by a
[Rh(phi)2(bpy′)]3+, respectively.27,60,120,121 Femtosecond tran- photoinduced sugar radical (circles)207 and CT between hairpin
sient absorption and fluorescence up-conversion spectroscopy capping stilbenes (squares, photooxidation of stilbene 4,4′-diether
(Sd) by stilbene-4,4′-dicarboxamide (Sa))125 and has been attributed
of Et+ site-selectively intercalated in DNA found that the to a crossover between coherent superexchange and incoherent
Et+ oxidized ZG with two rate constants.120 One (5 ps) hopping. In the latter case, comparison of injection and hole arrival
corresponds to Et+ that is already present in a CT-active rates supports superexchange for one or two intervening base pairs
conformation. The other (75 ps) corresponds to the gating and hopping for three or more base pairs between the stilbenes.
of Et+ to reach a CT-active conformation with respect to
the DNA. This is the first sense of conformational gating. dramatically decreases the fluorescence intensity, via CT
As the distance between Et+ and ZG increases, the rates of quenching.122,123 The robust fluorescence in the absence of
each component are unaffected, but the amplitudes mono- guanine was taken as evidence for a lack of CT between
tonically decrease, suggesting that the increase in distance stilbene and adenine, despite the presence of several low-
lowers the yields by virtue of changing the population of yield picosecond decay components. Eventually, these
CT-active states, rather than by affecting the inherent rates components were assigned to CT between the excited stilbene
of CT through the DNA. This represents the second sense and adenine.124,125 Either the charge-separated state can
of conformational gating. undergo recombination or, in the presence of distal guanine
or a lower energy stilbene 4,4′-diether (Sd), the hole migrates
To further test this model, a new method of conjugating
to the acceptor, leading to CT quenching.126 It was argued
Et+ to DNA as a rigid base pair surrogate was developed
that the recombination recovers the excited state and that
(Figure 5).27 For the Et+ separated from the hole acceptor,
Z hence charge injection in the Sa-AT constructs only
G, by a single base pair, the rate is similar to that for CT
minimally quenches fluorescence.124 This is distinct from
from the intercalated Et+. This rate drops 4 orders of
exciplex emission, as the emission spectrum is similar to
magnitude if another adenine is inserted between the Et+
the unconjugated fluorophore. The stilbene radical anion is
and ZG. Beyond a distance of two base pairs, the rate is
solvent exposed, and the motions of DNA, associated
constant. The authors interpreted this system as one that held
counterions, and bound water allow rapid relaxation of
the Et+ rigidly in a CT-inactive state. The rate-limiting step
dyes,127 so it is unlikely that recombination-induced emission
is injection, which for a poorly coupled donor exhibits a steep
would be of the same energy as radiation from the initial
distance dependence. At sufficient donor-acceptor separa-
excited state. An alternate explanation is that this fast
tion, reorientation of the donor is competitive with the slow
injection is limited to a small population that is in a CT-
CT from poorly coupled Et+. Now the apparent distance
active conformation. The remaining population fluoresces
dependence is flat, for the same reason as for the intercalating
normally in the absence of nearby guanine. In that context,
Et+.
the slow, strongly distance-dependent direct charge transfer
A similar explanation might serve for the slow rate of from excited stilbene to guanine can be interpreted in the
charge injection into stilbene-capped hairpins where several context of a donor that is not in a CT-active state with respect
AT base pairs separate the photoexcited stilbene hole donor to the π-stack.122
from a guanine hole acceptor. Stilbene-4,4′-dicarboxamide
incorporated as a bridging and capping element in short AT- 2.5. Back Electron Transfer
containing hairpins (Sa-AT) has an extended nanosecond
lifetime and hence higher fluorescence quantum yield versus Inevitably, photoinduced charge separation events are
the free dye, but a single GC pair close to the stilbene followed by charge recombination, also termed back electron
1648 Chemical Reviews, 2010, Vol. 110, No. 3
Table 1. Experimental and Calculated Oxidation Potentials of Nucleotides
nucleotide oxidation potential (V vs NHE)
method solvent dG
pulse radiolysis aqueous 1.29
electrochemistry MeCN 1.49
electrochemistry DMSO 1.44
CHCl3 1.62
CHCl3 + dC 1.28
time-resolved quenching aqueous 0.97
electrochemistry DMF 1.52
DFT organic 1.88
organic + dC 1.48
organic + dT
pulse radiolysis aqueous, DNA 1.22
transfer. After all, if charge recombination is thermodynami-
cally unfavorable, then charge separation is thermodynami-
cally favorable and will not require photoexcitation. The
effect of BET varies by the nature of the assay. Assays for
the presence of the charge-separated state, such as the slow
oxidation of guanine cation radical, will generate yields that
are convoluted with BET processes. In two extreme cases,
thionine128 and Ap,73 which are competent for efficient charge
separation, are not competent for the formation of permanent
guanine oxidation products.
The case of the two excited electronic states of AQ offers
a nice comparison of photooxidation with fast and slow BET.
Irradiation of DNA-conjugated AQ at 350 nm promotes it
to the singlet excited state, which relaxes to the triplet state.
Both states are competent for direct oxidation of all four
bases in DNA, but only the triplet radical anion reduces
oxygen to superoxide. The singlet radical anion undergoes
rapid BET, regenerating the initial state, while the charge
injected by the triplet radical anion is persistent and can
equilibrate along the DNA on a longer time scale.129,130 This
scheme explains the incompetence of AQ to oxidatively
repair cyclobutane thymine dimer;131,132 repair can only
proceed from the singlet state.
Experiments that rely on slow product trapping at guanine
need BET to provide the fundamental clock that allows
discrimination of CT attenuation.66 Hence, although the
results will be qualitatively diagnostic, the quantitative
accuracy will only hold relative to the BET rate for that
system. For example, a single negatively charged phosphate
group near the intercalation site of a tethered rhodium
photooxidant changes the observed ratio of damage between
a distal and proximal GG site by an order of magnitude,
indicating that the distal/proximal damage ratio is not solely
determined by the intervening bridge.52
Short-range CT is particularly subject to BET, as the
recombination has a steeper distance dependence than
separation, most likely due to greater separation in donor-
bridge-acceptor energies, as discussed in section 3.2.122 This
was first exploited in guanine damage systems using AQ as
the donor, but has since been systematically studied in
Acr+-phenothiazine (Ptz) and napthalimide (NI)-Ptz
systems.133-136 In the former, suppressing BET allowed the
extension of a canonical short-range CT system into one that
exhibited persistent CT separation over a long range!136
2.6. Injection and Migration Effects
Even among well-coupled donors and acceptors, substan-
tial variations in CT yields and rates have been observed.
The fastest observed rates (subnanosecond) over long
distances are for the RuII*/III/RhIII/II pair1,137 and for the
Genereux and Barton
dA dC dT ref
1.42 1.6 1.7 272
1.96 2.14 2.11 144
145
145
145
1.2 174
146
2.01 2.18 2.25 273
273
1.81 273
149
oxidation of ZG by Et+* 120 and the reduction of RhIII by
Et+*.121 Oxidation of guanine by Ap*,138 excited stilbene,122
or even guanine radical after initial oxidation by photoexcited
stilbene139,140 is substantially slower. To first order, this trend
reflects the relative stacking of donors and acceptors with
the DNA duplex.
In addition, some of these results may be reconciled in
the context of considering the effect of electrostatics on hole
migration.113 This effect was directly demonstrated by a study
with RhIII* as the photooxidant wherein the position of a
terminal phosphate was varied.52 In this experiment, com-
parative damage at GG sites proximal and distal to the Rh
intercalation site was determined. Since the decomposition
of the guanine cation radical is slow and the DNA between
the GG sites was short and undamaged, the final yield reflects
both the relative extent of BET and the potentials at the GG
sites. When a phosphate anion is added to the end opposite
to the rhodium, there is a small increase in damage
distribution toward the distal site. An extra phosphate anion
on the same end as the rhodium, however, leads to a several-
fold change in relative damage. For one sequence, relative
damage at the proximal site increases from 16% to 56%.
This argues that local charge can have a strong effect, both
on the rate of BET and on the rate of migration. For AQ,
which irreversibly injects a hole due to rapid oxygen
quenching of the triplet AQ radical anion, this effect is not
present,141 consistent with a lack of BET,66 although given
the low amount of distal damage in these constructs, it is
not clear whether a subtle change would be detectable.142
Importantly, in biological systems the initial oxidation
product is generally a guanine cation radical, without an
anion radical also being localized on the DNA. Hence,
Coulombic attraction will not inhibit transport away from
the injection site, as is the case for Ap* and Ap(-H)•,
stilbene, AQ, and other neutral photosensitizers.
2.7. Energetics
The natural nucleosides of DNA are resistant to mild
oxidation and reduction, and the radical cations and anions
undergo secondary chemical reactions on the microsecond
time scale. Hence, the reversible potentials are not trivial to
acquire.143 Approaches for determining the nucleoside po-
tentials fall into four categories: computational, electro-
chemical, pulse radiolysis, and photooxidation studies (Table
1). A common conclusion from all of these studies is that
the oxidation potentials increase in the order G < A < C ≈
T.
Electrochemical measurements of base potentials are
limited to organic solvents, generally acetonitrile, DMSO,
or DMF, due to the relative facile oxidation of water versus
Mechanisms for DNA Charge Transport
the four bases. Considering the hydrophobic interior of DNA,
potentials determined under these conditions may be more
relevant to DNA than potentials determined in aqueous
solution. To date, reversible electrochemistry has not been
achieved. Irreversible oxidation potentials of all four nucleo-
sides have been measured, which should be close to the
standard potential.144 The values for dG are similar in
acetonitrile and DMSO,145 but substantially more positive
and more negative values have been found in chloroform145
and DMF,146 respectively. In chloroform, it was found that
the presence of dC, which allows the possibility of proton-
coupled electron transfer (PCET), lowers the oxidation peak
potential of guanine by 340 mV.145
The closest that electrochemical experiments have come
to measuring the oxidation of guanine in its natural context
in DNA are the electrocatalytic experiments using mediators
on indium tin oxide.147 That [Ru(bpy)3]3+ (E1/2 ≈ 1.3 V; all
potentials herein are vs NHE) is a facile mediator for
electrocatalytic oxidation of DNA indicates a comparable
potential for guanine.148 Analysis of the oxidation rate using
a variety of metal complex mediators of different potentials
supports this value. Notably, a sufficiently high ionic strength
was used in these experiments to deconvolute the potential
from the affinity of the metal complex. This result was later
validated by pulse radiolysis experiments in DNA,149 where
a potential of 1.22 ( 0.02 eV was found for guanine in
multiple sequence contexts. Although the absolute potentials
from electrocatalysis are approximate, they provide strong
evidence for 5′-GG-3′ being about 0.15 V lower in potential
than G with a 5′-pyrimidine.150
The latter result, that the 5′-G of GG doublets is lower in
oxidation potential versus G, has been extensively exploited
in studies of the migration of charge in DNA, where GG
sites are used as shallow hole traps. Calculation and oxidative
yield experiments indicate that GGG acts as an even deeper
well, although smaller differences in potential between these
sequences are found experimentally than by calculations,151
probably due to solvent interactions.152 Both calculation and
experimental work support preferential hole density on the
5′-G.99,153-156 More generally, there is a strong correlation
between the calculated ab initio ionization potential of
guanine in the different stacking environments and the
relative oxidative damage found between different guanines
under conditions that allow full equilibration of the injected
charge.157 Stacking affects the energies of all the bases, with
the strongest perturbation due to the 5′-base.150,158
Vitally, all of these experiments probe the equilibrium
potentials of the bases. Random sequences of DNA have
rugged potential landscapes, corresponding to extensive static
disorder. There are many conformational modes in DNA,
and its hydration layer with time scales from picoseconds
to seconds.127 As discussed in more detail below, the
energetics of the bases are coupled to these modes, introduc-
ing both static and dynamic disorder to the system.
Given the challenge in properly coupling an individual
photooxidant to DNA, it is not surprising that few studies
have attempted to determine rationally the driving force
dependence of CT. It has been established from several
studies that CT does not occur from an excited photooxidant
hole donor to a higher energy hole acceptor. For example,
the metallointercalator [Rh(phi)2(bpy′)]3+, tethered to DNA,
can oxidize A, G, or C from its excited state (E3+*/E2+ )
2.0 eV), but the metallointercalator [Ru(phen)(dppz)(bpy′)]3+
(E3+/E2+ ) 1.6 eV) oxidizes G, but not C.76 Similarly,
Chemical Reviews, 2010, Vol. 110, No. 3 1649
Figure 6. Oxidation and repair of a thymine dimer (∼1.8 eV) by
tethered photoexcited [Rh(phi)2(bpy′)]3+ (2.0 V) is unaffected by
the intervening double guanine site (1.2 eV). Oxidation of double
guanine sites by [Ru(phen)(bpy′)(dppz)]3+ (1.5 eV) is unaffected
by the presence of the thymine dimer, which this oxidant lacks
sufficient driving force to repair. The latter result implies that the
guanine radical is not competent to repair thymine dimers, in
accordance with the known potentials. Hence, either the guanine
radical oxidized by [Rh(phi)2(bpy′)]3+ does not relax prior to
migration to the thymine dimer or the guanine radical is not an
intermediate in DNA-mediated oxidation of the thymine dimer by
[Rh(phi)2(bpy′)]3+.
[Ru(phen)(dppz)(bpy′)]3+ and ethidium bromide (E+*/E0 )
1.2 V)121 are not competent to repair thymine dimers, but
photoexcited [Rh(phi)2(bpy′)]3+ performs this repair, as does
napthaldiimide (NDI) (E*/E- > 1.9 V).131,159 These studies
include measurements with the hole donor tethered far from
the acceptor, with intervening low-potential double guanine
sites, indicating that, even after charge is injected into DNA,
there is some memory of the energy of its initial state (Figure
6).
In support of this interpretation, CT can still occur far
below the potential of the DNA. In one example,160 an
oxidized nitroxide (NO+/NO• ≈ 1 V) was incorporated into
a duplex by covalent attachment to thymine. The [4Fe-4S]
protein MutY (E3+/E2+ ) 0.1 V) was added, and the
generation of the reduced nitroxide spin radical was observed
by EPR (yield ∼50%). This chemistry was demonstrated to
be DNA-mediated by two controls. A noncleavable substrate
analogue for MutY incorporated into DNA far from the label
increased the reduction yield, and partially saturating the
electronic conjugation between the label and the DNA
substantially decreased the reduction yield.
A similar conundrum is involved in DNA-mediated CT
in self-assembled monolayers on gold (Figure 7).161 In these
systems, a DNA monolayer is covalently self-assembled on
an electrode, and CT through the DNA is measured with an
electrochemical probe. At an applied potential greater than
0.4 V, the DNA adopts an upright conformation. Although
elastic motions can bring the DNA into contact with the
surface,162 these conditions require high salt and fairly
1650 Chemical Reviews, 2010, Vol. 110, No. 3 Genereux and Barton

Figure 7. Scale diagram describing the relevant potentials for DNA-mediated CT through DNA self-assembled monolayers on gold. The
potentials of the individual nucleotides are not accessible within the window of electrochemistry of DNA monolayers on Au. Nevertheless,
facile DNA-mediated electrochemistry is observed for redox probes over DNA bridges. For all probes and sequences of well-matched
duplexes, the tunneling through the alkane linker is rate-limiting (∼30 s-1). Shown, in order from the top, are daunomycin, methylene blue,
Redmond Red, and a [4Fe-4S] cluster similar to those in the redox-active repair proteins EndoIII and MutY.

positive potentials. DNA mediation has been established for

our system by the methods discussed above, i.e., mismatch
and binding event discrimination, probe conjugation, and
linker length dependence, and by the differential redox
potential between direct contact and DNA-mediated reduc-
tion.31,163-165 The window for these experiments on Au is
limited by the potential of Au-S reduction, about -0.5 V.
DNA-mediated CT has been observed for a dozen different
probes spanning a full volt below this value.91,163-166
Importantly, the reductive limit is 600 mV below the
reduction potential of cytosine.
Photooxidant-bridged DNA hairpins were employed to
measure systematically the dependence of the CT rate on
the driving force.167 In these experiments, five photooxidants,
Sa (E*/- ) 1.68 V), naphthalene-2,6-dicarboxamide (E*/-
) 1.74 V), diphenylacetylene-4,4′-dicarboxamide (PA; E*/-
) 2.02 V), NDI (E*/- ) 2.93 V), and phenanthrene-2,7-
dicarboxamide (E*/- ) 1.43 V), were employed as hole
donors. Guanosine, inosine, deazaguanosine, and 8-oxogua-
nosine were used as the hole acceptors, either immediately
adjacent to the bridging dye or separated by a bridge of two
AT base pairs. The measured charge separation and recom-
bination rates fit well to the Marcus-Levich-Jortner equa-
tion, with reorganization energies of 1.2 and 1.3 eV,
respectively (Figure 8). The agreement is improved if a
molecule-based model for the solvent is used and the Figure 8. The driving force dependence for CT in photooxidant-
Q-model is employed for the energy surfaces; in this case bridged DNA hairpins is determined from time-resolved transient
higher reorganization energies are found.168 Further experi- absorption studies of a series of five stilbene-derived photooxidants
ments varied the bridge length for the Sa and PA duplexes and four hole acceptor bases, following both charge separation
(filled symbols) and charge recombination (empty symbols): cases
and found that the distance dependence increased for greater I and II (circles), the donor and acceptor are in contact; cases III
donor-bridge energy separation.169 and IV (triangles), the donor and acceptor are separated by two
TA base pairs. Cases I and III are fit only to charge separation
2.8. Proton-Coupled Electron Transfer rates (dotted lines), while cases II and IV are fit to both charge
separation and charge recombination rates (solid lines). Similar
If a participant in electron transfer is also an acid or base, reorganization energies, about 1 eV for the nuclear reorganization
the reduced and oxidized species will often have different energy and ∼0.2 V for the solvent reorganization energy, are found
pKA values as well. In this case, it is likely that CT will be for both case II and case IV. Adapted from ref 167. Copyright 2000
coupled to proton transfer.170 PCET may be more the rule American Chemical Society.
than the exception in biological systems. Most redox-active whether the electron transfer is proton-coupled, but whether
groups in biology are also subject to protonation or depro- the coupling is significant so that proton transfer becomes
tonation, with the pKA dependent on the redox state. Since rate-limiting. In the case of class I E. coli ribonucleotide
complete proton transfer is unnecessary for substantial effects reductase, PCET has been found to occur over multiple
on the coupled electron transfer rate, the question is not steps.171-173
Mechanisms for DNA Charge Transport
Each nucleotide in the double strand participates in stable
hydrogen bonding with its complement in a base pair. Hence,
it is not surprising that CT between nucleotides should be
proton-coupled, although likely in a way that cannot be
probed by the usual assay of pH dependence, given that the
base pair protons are excluded from solvent. It has been
shown that oxidation of the aqueous, isolated nucleosides
by Ap* is not proton-coupled,174 but that does not exclude
the possibility of PCET in the context of protons in the base
pair. Theoretical work predicts that double proton transfer
between guanine and cytosine lowers the guanine potential.175
Indeed, the cytosine radical has been directly observed by
transient absorption spectroscopy, after oxidation of DNA
by SeO3•- and SO4•- ions generated by pulse radiolysis.176
This might not be general, though, as the mechanism of GC
oxidation in pulse radiolysis is strongly dependent on the
chemical interaction with the oxidizing radical.149,177 Similar
evidence for PCET reduction of thymidine base-paired to
adenine has also been found.178
Furthermore, the pKA of the guanine cation radical has
been measured to be 4, near that of cytosine (4.5),179 and
the neutral guanine radical is observed by nanosecond
transient absorption and EPR after oxidation by intercalated
∆-[Ru(phen)2(dppz)]3+,180,181 supporting deprotonation of the
guanine cation radical, presumably to the paired cytosine,
on a faster time scale. For DNA ionized by γ-irradiation at
77 K, ESR measurements find that the equilibrium strongly
favors the neutral guanine radical over the cation radical;
this holds for guanine stacked between cytidine and for each
guanine of the GGG triplet.155 Although this evidence
strongly supports proton transfer, it does not establish
coherent PCET, as proton transfer consequent to oxidation
has been shown to be favorable.182
Isotope experiments, which would establish whether proton
transfer is rate-limiting, have not been straightforward.
Certainly, for oxidation by SO4•- ions, the charge injection
yield is decreased in D2O.177 In one experiment, deuterium
replacement of acid protons led to a 3-fold decrease in the
relative yield of damage of distal GGG to proximal G sites.87
In another, CT between Ap(-H)• and G was found to exhibit
a small differential between D2O and H2O, consistent with
PCET.183 A similar small differential was observed in some
sequences, but not in others, for CT between photoexcited
NI and Ptz.184 However, in the fluorescence experiments, the
substitution of D2O for H2O also affects excited-state
lifetimes.
At first, it might seem that the facile oxidation of CPC in
competition with CPG supports a PCET model.73 However,
CP
C oxidation is increased by base-pairing with inosine, a
high-potential guanine analogue, indicating that PCET is not
the mechanism of CPC oxidation. Instead, this mechanism is
enticingly similar to the proposed mechanism for excited-
state relaxation in GC base pairs,185,186 which involves proton-
coupled exciplex formation and has also received experi-
mental support,187,188 although it appears that guanine-guanine
stacking might prevent this relaxation.189 On the basis of this
accumulated evidence, it seems likely that PCET is involved
in at least some charge injections to guanine and that neutral
guanine radical is the persistent form of injected radical.
2.9. Characteristics of DNA CT
It is apparent that DNA mediates CT over long distances
and that the rate and yield are sensitive to both the donor
and acceptor identities and the integrity of the intervening
Chemical Reviews, 2010, Vol. 110, No. 3 1651
π-stack. Structural distortion of the DNA, or poor coupling
of the donor or acceptor to the DNA, sharply attenuates long-
range CT. Furthermore, rapid CT is conformationally gated,
and the equilibrium conformation is not necessarily the CT-
active conformation.
3. DNA-Mediated CT Mechanisms
Tautologically, all mechanisms of charge transport incor-
porate an electron moving from a donor orbital to an acceptor
orbital. The variation consists in the identification of orbitals
that mediate this transition and the pathways that are coupled
to it. In a large biomolecule, such as DNA, complexity arises
from the sheer number of atoms involved. In this section,
we will evaluate postulated mechanisms of long-range CT
in DNA in the context of the properties discussed above.
3.1. Transport through Water, Ions, and
Phosphates
An obvious source of conductivity in DNA is the highly
charged phosphate backbone. Indeed, one of the earliest
models of CT through DNA involved transport through the
phosphates.190 A recent measurement of delocalization of a
hole produced on a single phosphate lends some credence
to this model,191 although it is unclear whether this delocal-
ization can be transduced into conduction, and comparable
measurements have not observed this delocalization.192 In
the phosphate conduction model, phosphates on the edge of
the DNA are directly ionized, and the hole rapidly hops
through isoenergetic phosphates. For this to occur, coupling
between the phosphates must be substantial. Even more
importantly, oxidative damage must preferentially occur at
phosphates versus the base stack. Some calculations found
that this was the case,193 but later work demonstrated that
this was due to neglecting the presence of water and
counterions that can shield the phosphate group’s negative
charge.194 Theoretical and experimental work suggests that
photoionization is initiated at bases and not at phosphates195
and that the energies of the ions, phosphates, and sugar states
are far from the Fermi energy.196
Alternatively, the motions of water and ions can lead to
apparent conduction. DNA adsorbs several layers of high
dielectric water,42 a primary condensation layer of cations,
and a secondary layer of condensed anions. Even under
relatively dry conditions, water and cations are still adsorbed.
This layer plays a major role in the conformational dynamics
of DNA and mediates molecular recognition events with
other biomolecules. In particular, it seems certain that early
conduction measurements were measuring ionic conduction
along the DNA, rather than properties of the DNA molecule
itself.43
Ultimately, however, it is difficult to rationalize these
models with the marked sensitivity of long-range DNA-
mediated CT to the integrity of stacking, as described in
section 2. In contrast, changing the pH, ionic strength, or
identity of the salt has at best a minor effect on CT, as long
as well-coupled donors and acceptors are employed. Even
the removal of a phosphate along the bridge does not cause
a measurable difference in CT yield.77,78 Adding extra
intervening phosphates, via the construction of triplex DNA,
actually lowers the competence for CT.197 Hence, it is
apparent that DNA CT must proceed through the base pairs,
in the interior of the duplex.
1652 Chemical Reviews, 2010, Vol. 110, No. 3
3.2. Superexchange
Any medium is a superior pathway for charge transport
in comparison to a vacuum. Superexchange is coherent
orbital-mediated tunneling, where, for electron (hole) trans-
port, the high-energy LUMOs (HOMOs) on the pathway are
virtually occupied, allowing a probability and corresponding
rate of transmission from the donor to the acceptor. Follow-
ing the Born-Oppenheimer approximation, the rate of
superexchange can be separated into a nuclear factor, νn, and
an electronic factor, νe:
kCT ) νnνe
where
νn ) exp ( (∆G + λ)2
4πλRT ) distance dependences to incoherent processes; at least it
and
2π
νe ) 0 2
|HDA | exp(-βd)
p√4πλRT
∆G is the driving force, H0DA is the donor-acceptor coupling
extrapolated to zero bridge length, β is a decay parameter
characteristic of the bridge, and d is the bridge length. The
nuclear factor is a function solely of the identities and the
environment of the donor and acceptor. The electronic factor
represents the electronic coupling between the donor and
acceptor, mediated by the bridge states. In the adiabatic limit,
electronic coupling is sufficiently strong that the nuclear
motion will determine the rate of charge transfer. In the
nonadiabatic limit, the electronic coupling is sufficiently
weak that the electronic transition probability is less than
unity at the transition state.9 Hence, long-range (greater than
1 nm) CT systems are generally treated as nonadiabatic,
though it has increasingly been recognized that changing the
structure of even long bridges can have a nontrivial effect
on both ∆G and λ.7,168,185,198 Ignoring this effect, the only
dependence of the rate on the donor-acceptor distance is
the exponential decay of the donor-acceptor coupling with
d, the length of the bridge, characterized by the parameter
β. It is important to note that β is generally not what is
directly measured in experimental systems. For systems that
measure the yield of irreversible chemical products, compet-
ing processes such as BET or equilibration will inevitably
convolute with the inherent rate of charge separation. Even
for very fast charge traps, or spectroscopy based measure-
ments that can directly measure kCT, the exponential drop-
off will not necessarily correspond to β if the nuclear factor
is itself distance-dependent.199 This restriction can be miti-
gated for long-range CT, where the iterative changes in the
bridge length are unlikely to affect ∆G and λ, but are
significant for short-range CT.7
Furthermore, it is important to note that calculation of the
CT rate requires precise knowledge of the intervening
electronic structure, which in turn is dependent on the
molecular structure. If the mechanism or pathway changes
with an increase in bridge length, then the distance depen-
dence will not be well-represented by the electronic factor
β. Also conformational dynamics can lead to a time-
dependent rate. If the equilibrium structure is the best coupled
structure, the dynamics will decrease the apparent CT rate.
Genereux and Barton
Another important consideration is that β is not indepen-
dent of the bridge and donor energies. For increasing
difference between the donor and bridge energies, β increases
according to
[  ]
2 ∆ε ∆ε2
β) ln + 1+
a 2V 4V2
where a is the intersite separation, V is the intersite coupling,
and ∆ε is the donor-bridge energy separation. As this
separation decreases to below V, direct injection will
successfully compete with tunneling. Hence, if tunneling is
occurring, β is limited to about 0.3 Å-1 (numerical calcula-
tions that properly treat the bridge as finite find about 0.2
Å-1).200 This supports the assignment of extremely shallow
excludes superexchange mediated by orbitals on the indi-
vidual bases of DNA. This model was supported by
experiments in photooxidant-capped adenine tract hairpins.169
The oxidations of guanine by photoexcited Sa and of ZG by
photoexcited phenanthrene-2,7-dicarboxamide are of similar
driving force, but the latter pair is 0.25 eV lower in potential
than the former pair. For each pair, the rate constants were
measured for varying lengths of an intervening adenine tract,
and the distance dependence was greater for the PA-ZG pair.
Superexchange has been most thoroughly characterized
as a mechanism for CT within and between redox-active
proteins; charge-transfer reactions among proteins are es-
sential to all organisms. To a rough approximation, proteins
can be treated as a homogeneous medium with a single
characteristic β of 0.9 Å-1.201 The scatter for individual
proteins, however, spans several orders of magnitude,
indicating that the electronic structure and pathways vary
strongly with the identity of the protein and the location of
the donor and acceptor.202 For some pathways in proteins,
conformational dynamics have been shown to play an
important role in dictating which pathways are available.203
It is clear that proteins optimize charge transport not only
by controlling the donor-acceptor distance and driving force,
but by allowing a specific pathway, or combination of
pathways, to be available for superexchange. An essential
lesson from superexchange in proteins is that the most facile
pathways determine the overall rate and yield. Although
DNA might appear to be a simple one-dimensional system,
owing to the extensive π-stacking, experiments suggest a
more complicated system.
3.2.1. Coupling Constants in DNA
No model of superexchange can be properly constructed
without first considering appropriate values for the coupling
constants along the bridge.204 Given the structural complexity,
nontrivial assumptions are necessary to allow tractable
calculations, each of which have certain disadvantages.
Furthermore, the stacking interaction of bases is a particularly
challenging one to computationally describe.205 It is important
to consider these couplings when developing a theoretical
model. For example, a two-stranded model206 for coherent
DNA-mediated CT was recently published to fit the results
of work207 by Giese and co-workers. The model was only
able to fit the data by taking intrastrand AA coupling to be
0.52 eV, nearly an order of magnitude greater than the time-
averaged value found in typical calculations.208,209 The
average couplings appear to be about 80 meV for intrastrand
Mechanisms for DNA Charge Transport
GG, with somewhat lower instrastrand coupling for AA and
smaller values for interstrand purine-purine couplings.
Increasingly, it has been clear that couplings are highly
dependent upon the geometry of the stacked bases. An early
demonstration of this concept was the calculation of coupling
constants between base pairs for coordinates drawn for a
large family of crystallized duplexes.210 Even though this
measurement was only for coordinates from a set of crystals,
each of which presumably corresponds to an equilibrium
conformation, variations in couplings were on the order of
half of the values. In addition to this static disorder, DNA is
subject to extensive dynamic disorder on a full spectrum of
time scales. A later study considered fluctuations from
equilibrium conformations, using MD simulations to access
the transient structures (Figure 9).211 This study found even
larger variations in coupling and found that HDA for
GAAAAG varied by more than an order of magnitude over
the course of the 40 ps simulation. Interestingly, they also
found that transverse base motions, which affect stacking,
are more significant than longitudinal motions; this is
consistent with recent work that found that shear, twisting,
and stretching within base pairs also affects coupling
constants.212 They also found that peaks in coupling over
the bridge were more significant for GAAAAG than GTTTTG
and nearly absent in GATATG, in accordance with measured
CT yields. Since then, similarly large fluctuation-dependent
variations have been found using a variety of computational
approaches,49,208,213,214 with fluctuations being most significant
on the picosecond time scale.213 These studies have dem-
onstrated that the conformation also has a profound effect
on the transient nucleobase energies, as does solvent polar-
ization.215 Interestingly, calculations indicated that base
energies tend to be correlated in the duplex,182 although the
relative ordering of base energies is preserved.208 These
results offer a natural explanation for the conformational
gating that has been observed in long-range systems.
3.2.2. Reorganization Energy
Given the excellent correlation of theory and experiment
for the driving force dependence of CT in stilbene-capped
hairpins,167 the reorganization energy for those systems is
certainly close to 1 eV. For systems where the donor and
acceptor are internal to the π-stack, as with intercalators, it
is less clear, as these sites are far less solvent exposed than
the end-capped agents, such as Sa, Ptz, AQ, and NI.
Even for transfers between sites in DNA, the reorganiza-
tion energy can vary substantially. The sequence context,
which affects couplings and site energies, is expected to
affect the reorganization energy as well. Delocalization
among multiple bases, which decreases the effective amount
of charge that must be transferred, lowers the reorganization
energy.216 One study found, by sampling many molecular
dynamics configurations of oligopurine•oligopyrimidine DNA,
the reorganization energy for nearest-neighbor hops to be
about 1.1 eV for both adenine to adenine and for guanine to
guanine.49 Another study estimated 0.5 eV for adenine to
adenine on the basis of the spectral density of intercalated
ethidium bromide.217
It has been found that for short-range CT in DNA, changes
in bridge length can induce substantial changes in the
reorganization energies.7,118 This finding is consistent with
Marcus’s classical description of the solvent reorganization
energy:
Chemical Reviews, 2010, Vol. 110, No. 3 1653
(
∆q2 1
)( )
1 2 1 1
λs ) + - - st
2 aD aA RDA εop ε
where ∆q is the change in charge, aD and aA are the donor
and acceptor radii, respectively, and εop and εst are the optical
and static dielectric constants. λ explicitly depends on the
donor-acceptor separation.9 An indirect length dependence
of λ will also be incorporated through the effect of the
molecular structure on the dielectric.
3.2.3. Superexchange in DNA
Long-range charge transport over more than 50 Å seems
incompatible with superexchange, given its inherently strong
distance dependence. Even a β of 0.1 Å-1 implies a loss of
over 8 orders of magnitude in rate over 200 Å. However,
most long-range measurements either neglect yield122,159 or
measure products formed on long time scales.3 In the former
case, long-range transport can reflect a small yield, while,
in the latter case, products might be formed only on the time
scale of milliseconds to seconds. Fluorescence quenching
of photoexcited 2-aminopurine by guanine 35 Å away54 has
shown that long-range CT can occur on a time scale that is
defined by the nanosecond lifetime of the 2-aminopurine
excited state. Furthermore, the distance-independent decom-
position of CPA by photoexcited [Rh(phi)2(bpy′)]3+ over 40
Å17 (Figure 9) demonstrates that CT occurs at high yield at
least as fast as BET between oxidized adenine and
[Rh(phi)2(bpy′)]2+; the energetically similar reduction of
[Rh(phi)2(phen′)]3+ by [Ru(phen′)2dppz]2+ over 41 Å is faster
than 3 ns.1 Hence, it is clear that DNA CT can occur over
long distances on relatively short time scales, and any model
must account for this. For these reasons, superexchange
models are not satisfactory for DNA-mediated CT over long
distances.
3.3. Localized Hopping
The apparent contrast between theory and experiment led
to extraordinary efforts to challenge the validity of the
measured CT rates and yields. Hopping models offer an
alternative that does not require exceptional coupling between
bridge sites. Hopping, a type of diffusive, incoherent
transport, is the concatenation of multiple superexchange
steps, or “hops”, in which charge occupies the bridge between
each hop. Hopping has been proven as a mechanism in both
natural171 and synthetic218 protein models. The distance
dependence of hopping is geometric and hence shallower
than the distance dependence of superexchange.202,219 The
reason for the shallower distance dependence is intuitive;
long, slow, superexchange steps are avoided.
3.3.1. Nearest-Neighbor Models
Although formal ballistic models do not distinguish
superexchange from hopping, it is most straightforward to
treat hopping as a multistep process. In this case, an injected
charge resides on the lowest potential base, guanine. This
charge can diffusively migrate along the DNA, mediated by
short single-step superexchange with neighboring guanines.219
The rate of a hop will depend on the distance and sequence
context. The fastest hop will be to neighboring guanine; hops
through other nucleotides (i.e., GAG, GCG, or GTG) will
be slower. For the case of GCG, hopping is also allowed to
1654 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 9. CT from photoexcited [Rh(phi)2(bpy′)]3+ to N6-cyclo-
propyladenosine (CPA) across an adenine tract is distance-
independent over 14 adenines.17 The rate of CT across the adenine
tract, then, must be much faster than BET from the first adenine to
the reduced rhodium. The driving force for recombination is only
about 1.7 V, implying that BET should not be in the inverted region,
consistent with evidence that BET from adenine to this rhodium
complex is facile.68 The lack of distance dependence, in a system
with a rapid competing process in BET and a charge trap that
samples pre-equilibrium CT dynamics, implies extensive delocal-
ization across the bridge.
Figure 10. Time-dependent couplings between guanines separated
by three different four-base sequence contexts, based on conforma-
tions generated through molecular dynamics. It is clear that the
average value of coupling can be several orders of magnitude lower
than the maximum coupling. For the poorly stacked, flexible ATAT
sequence, strong coupling between the guanines is not achieved
over the time scale of the simulation. Reprinted from ref 211.
Copyright 2002 American Chemical Society.
the guanine on the complementary strand (G+CG/CGC f
GCG/CG+C f GCG+/CGC).
The most impressive evidence in support of this model is
a series of photooxidation experiments by the Majima group
and a series of STM conduction experiments performed by
Genereux and Barton
the Tao group.45 Using an elegant experimental setup,220 they
form unambiguously covalent contacts to single DNA
molecules under aqueous conditions. They found a geometric
dependence of the conductance on the length for (GC)n
sequences, but insertion of an (AT)m sequence into the (GC)n
sequence led to an exponential decrease of the conductance
with the length of the (AT)m sequence, as predicted by the
hopping model. However, they did not investigate sequences
that allowed purine-purine stacking and were unable to find
evidence for thermal activation,221 although this was possibly
due to the limited window of temperatures and potentials
that allowed device stability. Calculations on averaged
structures confirm that the alternating purine-pyrimidine
sequence attenuates delocalization for this system and
reproduce the data well.222 Furthermore, more recent experi-
ments using the same system223 and a similar approach using
a mechanical break junction224 found a much smaller effect
on conduction from increasing AT content. This was ascribed
to the latter experiments being performed with DNA that
was more likely to form the B-conformation versus the (GC)n
sequences. For DNA covalently bridging a carbon nanotube
gap, there appears to be no sequence dependence when the
GC content of random sequence DNA is varied.225
Majima and his colleagues have used transient absorption
to study the oxidation of Ptz by photoexcited NI, linked by
varying sequences of DNA.4 The sequences near the hole
acceptors and donors were kept constant, and a central region
varied with (GA)n or (GT)n; complementary studies separated
the donor and acceptor by only adenines.226,227 Each system
fits well to a geometric dependence of the rate on the bridge
length. They found the rate of hopping to be 2 × 1010 s-1
for adenine to adenine hopping, 7.6 × 107 s-1 for G+-A-G
f G-A-G+ hopping, and about 2 × 105 s-1 for G+-T-G
f G-T-G+ hopping. Although this is strong evidence for
multistep hopping in these systems, they have noted that it
does not necessarily require that the intermediate states be
completely localized on individual nucleotides.227 More
generally, the researchers were able to distinguish between
hole injection and hole arrival, showing that the two are not
coincident over long distances. Similar results have been
observed for CT between DNA-capping stilbenes, with the
transition between single-step and multistep CT at about two
intervening AT base pairs.125,126
3.3.2. Thermally Induced Hopping
Although hopping between guanine sites can explain many
features of the propagation of holes in mixed sequences, it
is not sufficient to explain facile charge transport through
adenine tracts.228 Occupation of adenines during CT has been
demonstrated both by a direct chemical probe73 and by the
observation of facile and weakly distance-dependent transport
of holes across long adenine bridges,17,31,207,229-231 even when
BET competes with equilibration.31 This can be explained
by a reasonable modification of the hopping model, whereby
a hole on guanine can be thermally excited to occupy an
adenine tract.228,232,233 Although this will be disfavored in
the case of mixed-sequence DNA in preference to guanine
hopping, it can be much more favorable than the slow hop
through a long AT sequence. For isoenergetic adenines, this
model does not sufficiently explain the distance indepen-
dence,234 but if the adenines on the edge of the tract are
higher in potential than the interior adenines, then the
apparent yield of CT will be distance-independent.235 This
explanation, although reasonable for An bridges on the strand
Mechanisms for DNA Charge Transport
Figure 11. The variable-range hopping model predicts a shallow distance dependence for the rate of CT between G and GGG across an
adenine tract on the opposing strand. Delocalized states, even in the absence of disorder (dashed lines), yield larger and shallower CT rates
due to the smaller reorganization energy and a shorter effective bridge length. In the presence of static disorder (solid lines), localized
hopping is substantially attenuated due to the rugged energy landscape. Delocalized hopping, however, is relatively unaffected by static
disorder, as the coupling is strong enough to allow tunneling through local barriers. Reprinted from ref 217. Copyright 2003 American
Chemical Society.
complementary to the guanine sites, does not support the
shallow distance dependence observed when the An bridge
is in the same strand,229,236 for which the edge adenines
should be of lower energy versus internal adenines. It would
be interesting to determine what effect incorporating the
stacking-dependent adenine energetics would have on the
theoretical predictions of the thermally induced hopping
model.228
The most compelling evidence for thermal activation
comes from a biochemical trapping assay of G+/An/GGG,
where the yield of GGG versus G damage was quantified
after hole injection from a sugar radical near the single G
site.207 A steep distance dependence for n e 3 was followed
by a flat distance dependence for n g 3. This is consistent
with two mechanisms at play, where the steep distance
dependence corresponds to CT through superexchange across
the AT bridge, and the flat regime is where superexchange
is sufficiently slow for thermally induced hopping across the
adenine tract to become the dominant mechanism. It is
important to note, however, that this dependence looks
identical to that found for the rigid Et+ base pair surrogate.27
In the latter case, the dependence was caused not by a
fundamental shift in mechanism, but rather by the rate-
limiting injection from the hole donor. Stilbene-capped
hairpin systems125 and AQ-capped duplexes236 show a
similar, though much more gradual, positive second-order
change in the slope with distance. As shall be discussed in
greater detail below, delocalized mechanisms can also explain
facile transport through adenine tracts. Ultimately, a change
in slope on its own is not sufficient to justify a crossover in
mechanism.237
3.3.3. Variable-Range Models
All the mechanisms listed above can be considered
together as components of a variable-range hopping model.
Here, a hole is allowed to migrate by superexchange to any
other site, rather than being limited to nearest neighbors.238
The most probable sites will be the closest low-potential sites,
i.e., guanines. Hence, the hole will hop from guanine to
guanine through the DNA, preferring intrastrand transfer, but
able to exploit interstrand transfer or thermally induced
hopping onto adenine tracts where the sequence does not
Chemical Reviews, 2010, Vol. 110, No. 3 1655
allow more favorable pathways. Even unfavorable pathways
are possible, although slow. Theoretical treatments using this
model have been successful in modeling some biochemical
experiments,239 although it was demonstrated that introducing
static disorder substantially degrades the success of variable-
range hopping models that rely on localized states (Figure
11).217 In turn, dynamic disorder, analogous to the confor-
mational gating discussed above, can assist hopping in a
rugged landscape.113
One challenge that is common to all localized hopping
models is the explanation of the mismatch discrimination
that has been observed in nearly every system studied. One
proposal was that mismatches allow water access to prefer-
entially quench CT at guanine through proton abstraction or
other chemical reactions.232 This model was supported by
the observation that GT mismatches affected the distal yield
more than AA mismatches and that methylation of the most
acidic residue of a guanine opposite an abasic site restores
CT.87 This model has not stood up to more extensive
measurements, however, as AC mismatches attenuate CT
more than GT mismatches. It has also been shown that GT
mismatches lower the yield of CT by lowering the rate,4 and
GT mismatches attenuate CT even under applied potentials
insufficient for guanine oxidation.30 Alternatively, mis-
matched base pairs might have lower couplings to the
neighboring bases than matched pairs.62 Particularly, they
are less stacked and sample more unstacked configurations.
A similar argument can then be made to explain how DNA-
binding proteins that bend the π-stack also attenuate CT.
A more profound problem with localized hopping models
is the apparent “memory” that a charge has of the energy of
the state from which is was injected. The intermediate in a
localized hopping model is the cation or neutral radical on
guanine or adenine. Oxidation of cytidine or thymidine by
these species is taken to be highly unfavorable. Hence, the
energy of injected charge should not affect the nature of the
intermediate over long distances. This is not consistent with
the evidence from thymine dimer159 or CPC75,76 oxidation,
where oxidants competent for guanine oxidation, but not
pyrimidine oxidation, were unable to decompose these
species over a long distance. Oxidants that are competent
for thymine dimer repair or CPC decomposition, however,
1656 Chemical Reviews, 2010, Vol. 110, No. 3
remain competent to decompose these oxidation reporters
even with intervening low-energy guanines. For an extended
A20 bridge separating Ptz and photoexcited NI, the central
double guanine does attenuate the CT yield by about half,
indicating that over a very long piece of DNA relaxation of
the cation does occur.227
Localized hopping is also inconsistent with electrochemical
measurements, where the Fermi level is maintained up to a
volt below the potentials of the bases.58,88,164,240,241 Consider
as an example the case where the electrode is at the potential
of a [4Fe-4S]-containing protein (0.1 V) and injection is
into cytidine (certainly <-0.9 V), the most readily reduced
base, for an unfavorable driving force of at least 1.0 V.
Although the coupling between cytidine and the metal is
mediated by a long saturated linker, and hence will be small,
let the coupling correspond to a generous value for two
stacked bases, HDA ≈ 0.2 eV, and take the reorganization
energy as being 0.5 V. According to a simple nonadiabatic
Marcus-derived expression,116,242 the injection rate is no
greaterthan0.002s-1;forrealisticvaluesforthemolecule-metal
coupling this injection rate would necessarily be far lower.
This is slower than the linker-limited rate found through
DNA of about 30 s-1.18,240 Effectively, this discrepancy
reflects the inherent unfavorability of thermal activation far
from the bridge potential.
3.4. Delocalized Mechanisms
The models discussed so far each assume localization of
a hole on a single base. Although the couplings between
bases might be expected to allow delocalization, disorder in
the bath should rapidly localize charges onto a single site as
long as the reorganization energy is greater than interbase
coupling. However, there is some experimental evidence for
delocalization of charges, such as the effect of stacking
interactions on the pKA of the adenine cation radical243 or
the competition of CPC with CPG for oxidation.76 It has also
been demonstrated that static disorder attenuates rapid
hopping by creating low-potential bottlenecks.238 This can
be alleviated by allowing delocalization of the charge; in
this case, static disorder is partially averaged.217,244 In
conjuction with the known role of conformational gating,
the obvious candidate for the delocalized state is the
polaron.245,246
3.4.1. Polaron Hopping and Gating Mechanisms
Whenever charge is injected into a molecule, the environ-
ment will polarize in response, effectively partially delocal-
izing the charge and lowering the energy of the system.247
Since the energy of the polaron is different from that of the
purely localized charge, the presence of a polaron will affect
the CT behavior of the system, in a way largely dependent
on the polaron size. Much like PCET is inevitable for any
charge-transfer participant with acidic protons, polaron
formation is inevitable whenever CT proceeds with bridge
occupation. The essential questions are whether the polariza-
tion occurs on a time scale that can impact the CT process,
which relaxation modes will be coupled to the polaron
formation, and how much the polaron is stabilized relative
to the completely localized state.
At first order, polarization of the environment in response
to charge injection does not violate the tight-binding as-
sumption. In this case, although the DNA conformation, ion
distribution, and water orientation all restructure as a result
Genereux and Barton
of charge migration, the effect on the CT efficiency and rate
will be via a change in the site energies on the bases and
will be gated by the time scale of environmental polarization.
It is important to note that small polaron formation slows
charge migration, as the site energy is lowered, and hence,
the activation energy of each hop is increased; this leads to
dynamic disorder, distinct from the static disorder discussed
above. The exception is if the polaron can move by drift,
where the orbitals of the donor and acceptor states overlap,
so that CT occurs in the adiabatic limit. This results in
transport that is faster than hopping, especially as it can be
activationless.247 However, drift is most rapid between
isoenergetic sites, so it is not a likely mechanism in the
presence of static disorder, unless gated by conformational
fluctuation of site energies.
Any description of polarons must take into account the
structural rearrangement that provides the polarization.
Lattice motion has been well-treated in terms of deformations
along the hydrogen bonds between the base pairs.248-250 This
treatment is particularly instructive regarding the effect of
increased coupling between the lattice motion and the charge;
high coupling implies a higher activation energy for indi-
vidual hops and a higher probability for trapping. Hence,
thermal activation is taken as evidence for small polaron
trapping. There have been contradictory results on the
temperature dependence of CT in conductivity measure-
ments,23,221 though photooxidation studies have unambigu-
ously shown an increase in the long-range CT rate184 and
yield54,68 with temperature. Whether this temperature depen-
dence is due to conformational gating, small polaron activa-
tion, or activation of localized hopping is not immediately
obvious. Ultimately, the distinction between these cases is
not sharp. Conformational dynamics influence the bridge
energy and hence the activation energy for polaron drift.
Calculations suggest that ion fluctuations, in particular, could
sufficiently modulate the potential of a bridging sequence
of DNA to permit polaron equilibration between two sites.50
Given the ambiguity in experiments where the counterion
identity and concentration have been varied,51-54 water
reorientation is more likely than ion motion to gate polaron
formation.
Sufficient polarization of the environment will lead to
formation of a large polaron. In this case, the polarization
distortion extends far beyond the lattice site, i.e., the
individual base (Figure 12). Polarization over a large range
must involve the medium, water. Calculation has supported
that these polarons form by water reorientation delocalized
over 2-5 base pairs, depending on the sequence.251,252 Large
polaron formation can have both positive and negative
consequences for CT: self-trapping can decrease the rate of
individual hopping steps, as is the case for small polarons,
but delocalization decreases the distance between individual
steps. Furthermore, for periodic sequences, drift can sub-
stantially increase the rate of individual hops, by lowering
the activation energy for incoherent transport.
Critically, polaron drift can explain important features of
DNA-mediated CT as discussed in the previous sections. As
discussed above, the observed dependences of the CT rates
and yield on the distance across adenine tracts is too shallow
to be readily reconciled with thermal activation and localized
hopping.17,125,207,236 Rapid polaron drift across adenine tracts,
in concert with inhibition of BET from adenine to guanine
due to polaron self-trapping, has been predicted to provide
a shallow distance dependence.253 Furthermore, since the
Mechanisms for DNA Charge Transport
Figure 12. Formation of a large polaron. Upon charge injection,
a hole is initially localized on a single base (red). Reorientation of
the environment, including neighboring bases and the hydration
layer, lowers the energy of the hole. Delocalization occurs to the
extent that the coupling between the bases balances the unfavorable
decrease in the reorganization energy.
calculated polaron size is ∼4 adenines, the steep distance
dependence that has been found for tracts shorter than this
length naturally corresponds to these sequences not support-
ing polaron formation.
This mechanism is not limited to adenine tracts, as polaron
formation is predicted over mixed polypurine sites, with
significant population of high-energy pyrimidines.252,254 This
is consistent with the oxidative damage observed at the fast
trap CPC despite the presence of guanine75 and the preferential
damage at thymine in constructs containing only thymine
or adenine.69 As described above, the resulting delocalization
serves as a mechanism for dynamic motions, which allow
polaron formation, to alleviate the barrier to CT generated
by the static disorder of site energies in DNA and is
consistent with the observed long-range migration of CT
through mixed DNA sequences.3,103 In this model, the
effective hopping rates observed by the Majima group could
correspond to hops between delocalized polaron sites.4,226,227
Physical identification of the polarization medium allows
calculation of the polaron properties, particularly the speed
limit on polaron migration imposed by the rate of repolar-
ization. For drift along an adenine tract, water reorientation
limits the polaron mobility to about 3 × 10-3 cm2/(V s).255
This mobility can be related to the conductivity of a single
DNA between two carbon nanotubes,34 where a mixed,
aperiodic sequence should decrease the mobility of a polaron.
Here, a resistance of about 3 MΩ was observed in a 15 base
pair duplex. Although the number of charge carriers was
unknown, it certainly cannot be less than unity or greater
than 15, the number of base pairs. Within that range, the
mobility is constrained to between 3 × 10-2 and 5 × 10-1
cm2/(V s). It will be interesting if theoretical evaluation is
Chemical Reviews, 2010, Vol. 110, No. 3 1657
able to rationalize these values, as they appear inconsistent
with polaron drift.
3.4.2. Domain Delocalization
Evidence for delocalization in DNA has come from recent
insights into long-lived excited states and exciplexes. It is
well-known that the individual nucleotides of DNA rapidly
relax upon excitation, with a low fluorescence quantum
yield.256 This property is essential for the molecule of life,
as long-lived excited states would render the genetic material
prone to photodamage. For mixed-sequence DNA as well,
relaxation is rapid (subpicosecond to picosecond), but in a
manner highly dependent on the specific sequence context.257
Recently, however, femtosecond studies of purine repeats
in both oligonucleotides and duplexes have found much
longer lifetimes for ground-state recovery,189 possibly due
to exciton or exciplex states. Critically, for B-form DNA, it
is base-stacking rather than base-pairing interactions that are
most critical in achieving long-lived states.258 The extent of
delocalization is still under debate. Although some states
might extend over 4-8 base pairs,259,260 others are delocalized
over only 2 base pairs,261 particularly in single strands. There
is some computational support for this delocalization as well.
DFT calculations find that eximers can delocalize over
several adenines,262 and it has been found that fluctuations
in the on-site energies of neighboring bases are highly
correlated.214 Recent calculations that incorporate static and
dynamic disorder and solvent effects have shown that such
transient delocalization can occur over several base pairs.263
Despite experimental and computational support for de-
localization, as described above, there are profound theoreti-
cal arguments against delocalization models in DNA. A
variety of computational studies have found that solvent and
ion motions will strongly localize injected charge to a single
nucleotide or only a few nucleotides.152,264,265 Importantly,
these studies have mostly been limited to considerations of
equilibrated or averaged structures. It will be important to
determine whether solvent-induced localization is maintained
in the presence of dynamic disorder.
A recent computational study266 on the dynamics of electric
fields produced by DNA, water, and ions was able to
reproduce time-resolved Stokes shift data that directly
measure those dynamics.127 This study found no evidence
of subpopulations where the DNA had particular electric
fields beyond the Gaussian distribution. However, the Gauss-
ian tails could in principle be the very states that mediate
CT. All the data suggest that delocalization must be highly
transient, and over very long distances, a substantial amount
of incoherent hopping will also occur.
Perhaps the most distinctive feature of DNA-mediated CT
that has been observed in recent years is the periodic length
dependence of the CT yield across adenine tracts for some
systems (Figure 13). This dependence was clear, with the
same period of 3-4 base pairs, for coherent transport from
Ap* to guanine54 and for total CT from [Rh(phi)2(bpy′)]3+*
to CPG.68 This periodicity was shown to be with respect to
the adenine tract length, rather than with respect to the
donor-acceptor distance; by measuring the decomposition
of CPA moved serially along an adenine tract of constant
length, no periodicity is observed.17 For CT from photoex-
cited AQ to guanine, the periodicity is less apparent, but
this is likely due to the quenching of the radical anion of
AQ by oxygen, which allows charge equilibration. Interest-
ingly, Ap* oxidation of CPG across adenine tracts is smoothly
1658 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 13. Equivalent periodicities with the same period and temperature dependence are observed for (A) the total oxidation of CPG by
photoexcited [Rh(phi)2(bpy′)]3+ and (B) the single-step oxidation of guanine by Ap*. The temperature increases from purple to red. Errors
are given in (A) as 90% SEM.54,68 Reprinted from refs 54 and 68. Copyright 2004 and 2008, repectively, American Chemical Society.
monotonic, but separating the Ap* from the adenine tract
with three inosines restores the non-monotonicity. Clearly,
the rapid BET associated with Ap* allows duplexes that are
well-suited to forward transport also to better mediate BET,
suppressing the periodicity. It should be noted that the inosine
tract is a high-potential barrier to oxidation by Ap.120 It
lowers both forward CT and BET, but since the former
competes with the nanosecond Ap* fluorescence lifetime,
and the latter competes with picosecond ring-opening, BET
should be comparatively more attenuated. With BET sup-
pressed, periodicity is again apparent.
A periodic A-tract dependence indicates that some adenine
tract lengths mediate CT superior to others. On the basis of
our experiments, this length is about three or four base pairs.
In light of the extensive evidence for delocalization cited
above, we characterize this CT-active tract as a delocalized
domain. The role of conformational gating, then, is to
generate this CT-active state. An adenine tract length that
allows an integer number of these states allows facile CT;
transport across other tracts requires dephasing processes,
such as drift or hopping. Because these domains are, by their
nature, transient, these effects will only be seen in experi-
ments where the donor and acceptor are well-coupled to the
bridge, and where injection and arrival can be observed on
a fast time scale, decoupled from other pathways, such as
BET. Critically, domain delocalization readily explains the
facile competition between CPC and CPG74 and the ability of
DNA to mediate CT far below the base potentials.161,164
4. Summary
It is clear that DNA, when adequately coupled between
the donor and acceptor, can competently mediate CT over
long distances. This property is dependent on, and hence
diagnostic of, the integrity of base stacking. Furthermore,
long-range DNA-mediated CT is thermally activated in a
manner dependent on the dynamical stacking of the bridge,
indicating that conformational gating is convoluted with the
CT rate. Theoretically, CT over long molecular distances
cannot be assigned to superexchange. Incoherent transport
must play a role, although evidence does support coherent
transport over at least 30 Å in some systems. Assigning the
intermediates as guanine cation radicals in the context of a
variable-range hopping model is sufficient to explain some
gross features of DNA-mediated CT, but this model cannot
explain long-range coherence. Transient delocalization plays
an important role, at least with some sequences. Identifying
the extent to which delocalization occurs, including via
Genereux and Barton
polaron formation, will be particularly important for under-
standing DNA CT mediated at potentials below those of the
individual nucleotides.
Any model for DNA CT must consider the effects of static
and dynamic disorder. For most models, static disorder
attenuates long-range CT. Since DNA has many sources of
static disorder in the site energies, intersite couplings, and
reorganization energies, it is unlikely that calculations
performed on uniform ideal structures with a single repeating
base pair will be relevant to understanding experimental
results. On the other hand, dynamic disorder has the potential
to alleviate the challenge posed by static disorder, by
allowing transient structures to form with less rugged
energetic landscapes. As long as the equilibrium conforma-
tion is not the most CT-active conformation, this condition
will hold for most pathways, whether incoherent or coherent.
Computational studies have begun to appear that consider
what CT-active states look like;267 it will be a challenge to
experimentalists to evaluate these exciting predictions.
CT between a donor and acceptor will always proceed
through the fastest pathways available. In a dynamic,
structurally complex molecule such as DNA, multiple time
scales describe the energetic and coupling landscapes, and
hence, there will be a time-dependent ensemble of pathways.
This ensemble is even larger when delocalized states are
allowed, whether they are transiently formed prior to,
concurrently with, or after charge injection. For conditions
that deplete available pathways, whether through rigidifying
the duplex, disrupting donor and acceptor coupling to the
bridge, or introducing structural distortion, slower CT and
conduction will inevitably result.
In this context, correlating the distance dependence to the
β value of the electronic factor of the CT rate equation
requires a high level of experimental support. It is unlikely
that any of the measured distance dependences correspond
to the distance dependence of the purely electronic compo-
nent of CT through DNA. Nevertheless, the effective distance
dependence over long distances compares favorably with
common molecular wires such as oligo(phenylenevinylene)
and oligo(phenyleneethynylene), indicating a promising role
for DNA in molecular electronics.
5. Unanswered Questions
It should be clear from this review that DNA-mediated
CT does not pose a challenge to the fundamental theories of
electron and hole transport. Ultimately, charge-transfer events
only occur with the rates predicted by Marcus’s theory. For
Mechanisms for DNA Charge Transport
a molecule as large and complicated as DNA, however, the
parameters for the Marcus equation are not trivial to
determine. Each conformation of a given DNA offers many
pathways, and the extent of dynamical disorder leads to the
failure of the Condon approximation. Furthermore, in the
context of hopping and drift, the nature of the states that
mediate charge transport vary with the sequence and
sequence-dependent dynamics. What these states are, local-
ized radical cations, localized neutral radicals, large polarons,
delocalized domains, or a combination, will be different on
the basis of the properties of the specific donor, DNA bridge,
and acceptor. Understanding what conditions lead to what
mechanism of transport is important, as the physical nature
of charge injection and migration in DNA undoubtedly
influences CT between DNA and redox-active DNA-binding
proteins5,17,93,94 and the cellular defense against oxidizing
radicals.105,268-271
Particular experiments that require more attention by
theorists are the electrochemical experiments in DNA films.
In these experiments, the Fermi level is held to potentials
far from those of the bridge states, and yet many of the same
properties are observed here as are observed in solution and
device experiments that are at profoundly different energies.
Insight into this process will undoubtedly also help elucidate
DNA-mediated CT in general.
Ultimately, single-molecule conductivity experiments have
the most potential for determining the details of DNA-
mediated CT, due to the strong control of the driving force
and online measurement of the current. The main challenges
for these experiments are maintaining the DNA in its native
structure and establishing that the observed current is, in fact,
due to the DNA. These can be easily determined by the
proper choice of controls.
If the past 15 years of DNA-mediated CT are any
indication, the synergy between the applications of DNA in
devices and biology, and theoretical and experimental efforts
to elucidate the mechanism, will continue to advance both
areas of study. Certainly, bringing these different perspectives
together offers both a challenge and an opportunity.
6. Acknowledgments
We thank the NIH (Grant GM49216) for their financial
support of this work. We also thank the many researchers
in the field of DNA CT chemistry that have contributed their
many and varied perspectives.
7. References
(1) Murphy, C. J.; Arkin, M. R.; Jenkins, Y.; Ghatlia, N. D.; Bossmann,
S. H.; Turro, N. J.; Barton, J. K. Science 1993, 262, 1025.
(2) Murphy, C. J.; Arkin, M. R.; Ghatlia, N. D.; Bossman, S. H.; Turro,
N. J.; Barton, J. K. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 5315.
(3) Nuñez, M. E.; Hall, D. B.; Barton, J. K. Chem. Biol. 1999, 6, 85.
(4) Takada, T.; Kawai, K.; Fujitsuka, M.; Majima, T. Proc. Natl. Acad.
Sci. U.S.A. 2004, 101, 14002.
(5) Lee, P. E.; Demple, B.; Barton, J. K. Proc. Natl. Acad. Sci. U.S.A.
2009, 106, 13164.
(6) Osakada, Y.; Kawai, K.; Fujitsuka, M.; Majima, T. Proc. Natl. Acad.
Sci. U.S.A. 2006, 103, 18072.
(7) Davis, W. B.; Hess, S.; Naydenova, I.; Haselsberger, R.; Ogrodnik,
A.; Newton, M. D.; Michel-Beyerle, M.-E. J. Am. Chem. Soc. 2002,
124, 2422.
(8) von Feilitszch, T.; Tuma, J.; Neubauer, H.; Verdier, L.; Haselsberger,
R.; Feick, R.; Gurzadyan, G.; Voityuk, A. A.; Griesinger, C.; Michel-
Beyerle, M. E. J. Phys. Chem. B 2008, 112, 973.
(9) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265.
(10) Turro, N. J.; Barton, J. K. J. Biol. Inorg. Chem. 1998, 3, 201.
(11) Endres, R. G.; Cox, D. L.; Singh, R. R. P. ReV. Mod. Phys. 2004,
76, 195.
Chemical Reviews, 2010, Vol. 110, No. 3 1659
(12) Chen, F.; Hihath, J.; Huang, Z.; Li, X.; Tao, N. J. Annu. ReV. Phys.
Chem. 2007, 58, 535.
(13) Müller, K.-H. Phys. ReV. B 2006, 73, 045403.
(14) Kelley, S. O.; Barton, J. K. Science 1999, 283, 375.
(15) Delaney, S.; Pascaly, M.; Bhattacharya, P. K.; Han, K.; Barton, J. K.
Inorg. Chem. 2002, 41, 1966.
(16) Tada, T.; Kondo, M.; Yoshizawa, K. ChemPhysChem 2003, 4, 1256.
(17) Augustyn, K. E.; Genereux, J. C.; Barton, J. K. Angew. Chem., Int.
Ed. 2007, 46, 5731.
(18) Gorodetsky, A. A.; Green, O.; Yavin, E.; Barton, J. K. Bioconjugate
Chem. 2007, 18, 1434.
(19) Fink, H. W.; Schoenberger, C. Nature 1999, 398, 407.
(20) de Pablo, P. J.; Moreno-Herrero, F.; Colchero, J.; Gómez Herrero,
J.; Herrero, P.; Baró, A. M.; Ordejón, P.; Soler, J. M.; Artacho, E.
Phys. ReV. Lett. 2000, 85, 4992.
(21) Kasumov, A.Yu.; Kociak, M.; Guéron, S.; Reulet, B.; Volkov, V. T.;
Klinov, D. V.; Bouchiat, H. Science 2001, 291, 280.
(22) Storm, A. J.; van Noort, J.; de Vries, S.; Dekker, C. Appl. Phys.
Lett. 2001, 79, 3881.
(23) Yoo, K.-H.; Ha, D. H.; Lee, J.-O.; Park, J. W.; Kim, J.; Kim, J. J.;
Lee, H.-Y.; Kawai, T.; Choi, H. Y. Phys. ReV. Lett. 2001, 87, 198102.
(24) Odom, D. T.; Dill, E. A.; Barton, J. K. Chem. Biol. 2000, 7, 475.
(25) Schuster, G. B., Ed. Long-Range Charge Transfer in DNA, I and II;
Springer: New York, 2004; Vols. 236 and 237.
(26) Wagenknecht, H. A., Ed. Charge Transfer in DNA; Wiley-VCH:
Weinheim, Germany, 2005.
(27) Valis, L.; Wang, Q.; Raytchev, M.; Buchvarov, I.; Wagenknecht,
H.-A.; Fiebig, T. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 10192.
(28) Meade, T. J.; Kayyem, J. F. Angew. Chem., Int. Ed. 1995, 34, 352.
(29) Lyndon Pittman, T.; Miao, W. J. Phys. Chem. C 2008, 112, 16999.
(30) Boon, E. M.; Jackson, N. M.; Wrightman, M. D.; Kelley, S. O.; Hill,
M. G.; Barton, J. K. J. Phys. Chem. B 2003, 107, 11805.
(31) Drummond, T. G.; Hill, M. G.; Barton, J. K. J. Am. Chem. Soc. 2004,
126, 15010.
(32) Watson, J. D.; Crick, F. H. C. Nature 1953, 171, 737.
(33) Watson, J. D.; Crick, F. H. C. Nature 1953, 171, 964.
(34) Guo, X.; Gorodetsky, A. A.; Hone, J.; Barton, J. K.; Nuckolls, C.
Nat. Nanotechnol. 2008, 3, 163.
(35) O’Neill, M. A.; Becker, H. C.; Wan, C.; Barton, J. K.; Zewail, A. H.
Angew. Chem., Int. Ed. 2003, 42, 5896.
(36) Melvin, T.; Botchway, S.; Parker, A. W.; O’Neill, P. J. Chem. Soc.,
Chem. Commun. 1995, 6, 653.
(37) Cohen, H.; Nogues, C.; Naaman, R.; Porath, D. Proc. Natl. Acad.
Sci. U.S.A. 2005, 102, 11589.
(38) Cohen, H.; Nogues, C.; Ullien, D.; Daube, S.; Naaman, R.; Porath,
D. Faraday Discuss. 2006, 131, 367.
(39) Ceres, D. M.; Barton, J. K. J. Am. Chem. Soc. 2003, 125, 14964.
(40) Boon, E. M.; Ceres, D. M.; Drummond, T. G.; Hill, M. G.; Barton,
J. K. Nat. Biotechnol. 2000, 18, 1096.
(41) Calladine, C. R.; Drew, H. R.; Luisi, B. F.; Travers, A. A.
Understanding DNA: The Molecule and How It Works; Elsevier
Academic Press: San Diego, 2004; p 304.
(42) Pal, S. K.; Zewail, A. H. Chem. ReV. 2004, 104, 2099.
(43) Yamahata, C.; Collard, D.; Takekawa, T.; Kumemura, M.; Hash-
iguchi, G.; Fujita, H. Biophys. J. 2008, 94, 63.
(44) Han Ha, D.; Nham, H.; Yoo, K.-H.; So, H.-m.; Lee, H.-Y.; Kawai,
T. Chem. Phys. Lett. 2002, 355, 405.
(45) Xu, B.; Zhang, P.; Li, X.; Tao, N. Nano Lett. 2004, 4, 1105.
(46) Nogues, C.; Cohen, S. R.; Daube, S.; Apter, N.; Naaman, R. J. Phys.
Chem. B 2006, 110, 8910.
(47) Xiao, X.; Xu, B.; Tao, N. J. Nano Lett. 2004, 4, 267.
(48) Boon, E. M.; Barton, J. K. Bioconjugate Chem. 2003, 14, 1140.
(49) Steinbrecher, T.; Koslowski, T.; Case, D. A. J. Phys. Chem. B 2008,
112, 16935.
(50) Barnett, R. N.; Cleveland, C. L.; Joy, A.; Landman, U.; Schuster,
G. B. Science 2001, 294, 567.
(51) Williams, T. T.; Odom, D. T.; Barton, J. K. J. Am. Chem. Soc. 2000,
122, 9048.
(52) Williams, T. T.; Barton, J. K. J. Am. Chem. Soc. 2002, 124, 1841.
(53) Douki, T.; Angelov, D.; Cadet, J. J. Am. Chem. Soc. 2001, 123, 11360.
(54) O’Neill, M. A.; Barton, J. K. J. Am. Chem. Soc. 2004, 126, 11471.
(55) Blaustein, G. S.; Demas, B.; Lewis, F. D.; Burin, A. L. J. Am. Chem.
Soc. 2009, 131, 400.
(56) Abdou, I. M.; Sartor, V.; Cao, H.; Schuster, G. B. J. Am. Chem.
Soc. 2001, 123, 6696.
(57) O’Neill, M. A.; Barton, J. K. J. Am. Chem. Soc. 2002, 124, 13053.
(58) Boon, E. M.; Salas, J. E.; Barton, J. K. Nat. Biotechnol. 2002, 20,
282.
(59) Okamoto, A.; Kamei, T.; Saito, I. J. Am. Chem. Soc. 2006, 128, 658.
(60) Kelley, S. O.; Barton, J. K. Chem. Biol. 1998, 5, 413.
(61) Bhattacharya, P. K.; Barton, J. K. J. Am. Chem. Soc. 2001, 123, 8649.
(62) Osakada, Y.; Kawai, K.; Fujitsuka, M.; Majima, T. Nucleic Acids
Res. 2008, 36, 5562.
1660 Chemical Reviews, 2010, Vol. 110, No. 3
(63) Bhattacharya, P. K.; Cha, J.; Barton, J. K. Nucleic Acids Res. 2002,
30, 4740.
(64) Buzzeo, M. C.; Barton, J. K. Bioconjugate Chem. 2008, 19, 2110.
(65) Boal, A. K.; Barton, J. K. Bioconjugate Chem. 2005, 16, 312.
(66) Williams, T. T.; Dohno, C.; Stemp, E. D. A.; Barton, J. K. J. Am.
Chem. Soc. 2004, 126, 8148.
(67) Furrer, E.; Giese, B. HelV. Chim. Acta 2003, 86, 3623.
(68) Genereux, J. C.; Augustyn, K. E.; Davis, M. L.; Shao, F.; Barton,
J. K. J. Am. Chem. Soc. 2008, 130, 15150.
(69) Joy, A.; Ghosh, A. K.; Schuster, G. B. J. Am. Chem. Soc. 2006,
128, 5346.
(70) Delaney, S.; Yoo, J.; Stemp, E. D. A.; Barton, J. K. Proc. Natl. Acad.
Sci. U.S.A. 2004, 101, 10511.
(71) Nakatani, K.; Dohno, C.; Saito, I. J. Am. Chem. Soc. 2001, 123, 9681.
(72) O’Neill, M. A.; Dohno, C.; Barton, J. K. J. Am. Chem. Soc. 2004,
126, 1316.
(73) Dohno, C.; Ogawa, A.; Nakatani, K.; Saito, I. J. Am. Chem. Soc.
2003, 125, 10154.
(74) Elias, B.; Genereux, J. C.; Barton, J. K. Angew. Chem., Int. Ed. 2008,
47, 9067.
(75) Shao, F.; O’Neill, M. A.; Barton, J. K. Proc. Natl. Acad. Sci. U.S.A.
2004, 101, 17914.
(76) Shao, F.; Augustyn, K. E.; Barton, J. K. J. Am. Chem. Soc. 2005,
127, 17445.
(77) Liu, T.; Barton, J. K. J. Am. Chem. Soc. 2005, 127, 10160.
(78) Osakada, Y.; Kawai, K.; Fujitsuka, M.; Majima, T. Chem. Commun.
2008, 23, 2656.
(79) Nakatani, K.; Dohno, C.; Saito, I. J. Am. Chem. Soc. 2000, 122, 5893.
(80) Kodera, H.; Osakada, Y.; Kawai, K.; Majima, T. Nat. Chem. 2009,
1, 156.
(81) Okamoto, A.; Nakatani, K.; Saito, I. J. Am. Chem. Soc. 2003, 125,
5066.
(82) Hunter, W. N.; Brown, T.; Kennard, O. Nucleic Acids Res. 1987,
15, 6589.
(83) Hunter, W. N.; Brown, T.; Kneale, G.; Anand, N. N.; Rabinovich,
D.; Kennard, O. J. Biol. Chem. 1987, 262, 9962.
(84) Kumamoto, S.; Watanabe, M.; Kawakami, N.; Nakamura, M.;
Yamanato, K. Bioconjugate Chem. 2008, 19, 65.
(85) Takada, T.; Fujitsuka, M.; Majima, T. Proc. Natl. Acad. Sci. U.S.A.
2007, 104, 11179.
(86) Elias, B.; Shao, F.; Barton, J. K. J. Am. Chem. Soc. 2008, 130, 1152.
(87) Giese, B.; Wessely, S. Chem. Commun. 2001, 20, 2108.
(88) Boal, A. K.; Yavin, E.; Lukianova, O. A.; O’Shea, V. L.; David,
S. S.; Barton, J. K. Biochemistry 2005, 44, 8397.
(89) Gasper, S. M.; Schuster, G. B. J. Am. Chem. Soc. 1997, 119, 12762.
(90) Hickerson, R. P.; Prat, F.; Muller, J. G.; Foote, C. S.; Burrows, C. J.
J. Am. Chem. Soc. 1999, 121, 9423.
(91) Drummond, T. G.; Hill, M. G.; Barton, J. K. Nat. Biotechnol. 2003,
21, 1192.
(92) Inouye, M.; Ikeda, R.; Takase, M.; Tsuri, T.; Chiba, J. Proc. Natl.
Acad. Sci. U.S.A. 2005, 102, 11606.
(93) Boal, A. K.; Yavin, E.; Barton, J. K. J. Inorg. Biochem. 2007, 101,
1913.
(94) Gorodetsky, A. A.; Dietrich, L. E. P.; Lee, P. E.; Demple, B.;
Newman, D. K.; Barton, J. K. Proc. Natl. Acad. Sci. U.S.A. 2008,
105, 3684.
(95) Watanabe, S.; Kita, A.; Kobayashi, K.; Miki, K. Proc. Natl. Acad.
Sci. U.S.A. 2008, 105, 4121.
(96) Gorodetsky, A. A.; Ebrahim, A.; Barton, J. K. J. Am. Chem. Soc.
2008, 130, 2924.
(97) Rajski, S. R.; Barton, J. K. Biochemistry 2001, 40, 5556.
(98) Nakatani, K.; Dohno, C.; Ogawa, A.; Saito, I. Chem. Biol. 2002, 9,
361.
(99) Kurbanyan, K.; Nguyen, K. L.; To, P.; Rivas, E. V.; Lueras, A. M. K.;
Kosinski, C.; Steryo, M.; González, A.; Mah, D. A.; Stemp, E. D. A.
Biochemistry 2003, 42, 10269.
(100) Perrier, S.; Hau, J.; Gasparutto, D.; Cadet, J.; Favier, A.; Ravanat,
J. L. J. Am. Chem. Soc. 2006, 128, 5703.
(101) Cadet, J.; Douki, T.; Ravanat, J.-L. Acc. Chem. Res. 2008, 41, 1075.
(102) Bjorklund, C. C.; Davis, W. B. Biochemistry 2007, 46, 10745.
(103) Nuñez, M. E.; Noyes, K. T.; Barton, J. K. Chem. Biol. 2002, 9, 403.
(104) Nuñez, M. E.; Holmquist, G. P.; Barton, J. K. Biochemistry 2001,
40, 12465.
(105) Merino, E. J.; Barton, J. K. Biochemistry 2008, 47, 1511.
(106) Nguyen, K. L.; Steryo, M.; Kurbanyan, K.; Nowitzki, K. M.;
Butterfield, S. M.; Ward, S. R.; Stemp, E. D. A. J. Am. Chem. Soc.
2000, 122, 3585.
(107) Johansen, M. E.; Muller, J. G.; Xu, X. Y.; Burrows, C. J. Biochemistry
2005, 44, 5660.
(108) Peng, X.; Pigli, Y. Z.; Rice, P. A.; Greenberg, M. M. J. Am. Chem.
Soc. 2008, 130, 12890.
(109) Liu, C.-S.; Schuster, G. B. J. Am. Chem. Soc. 2003, 125, 6098.
Genereux and Barton
(110) Lewis, F. D.; Daublain, P.; Cohen, B.; Vura-Weiss, J.; Shafirovich,
V.; Wasielewski, M. R. J. Am. Chem. Soc. 2007, 129, 15130.
(111) O’Neill, M. A.; Barton, J. K. Proc. Natl. Acad. Sci. U.S.A. 2002, 99,
16543.
(112) Shao, F.; Barton, J. K. J. Am. Chem. Soc. 2007, 129, 14733.
(113) Grozema, F. C.; Tonzani, S.; Berlin, Y. A.; Schatz, G. C.; Siebbeles,
L. D. A.; Ratner, M. A. J. Am. Chem. Soc. 2008, 130, 5157.
(114) Davis, W. B.; Ratner, M. A.; Wasielewski, M. R. J. Am. Chem. Soc.
2001, 123, 7877.
(115) Smalley, J. F.; Sachs, S. B.; Chidsey, C. E. D.; Dudek, S. P.; Sikes,
H. D.; Creager, S. E.; Yu, C. J.; Feldberg, S. W.; Newton, M. D.
J. Am. Chem. Soc. 2004, 126, 14620.
(116) Newton, M. D.; Smalley, J. F. Phys. Chem. Chem. Phys. 2007, 9,
555.
(117) O’Neill, M. A.; Barton, J. K. J. Am. Chem. Soc. 2004, 126, 13234.
(118) Siriwong, K.; Voityuk, A. A.; Newton, M. D.; Rösch, N. J. Phys.
Chem. B 2003, 107, 2595.
(119) Kocherzhenko, A. A.; Patwardhan, S.; Grozema, F. C.; Anderson,
H. L.; Siebbeles, L. D. A. J. Am. Chem. Soc. 2009, 131, 5522.
(120) Wan, C. Z.; Fiebig, T.; Kelley, S. O.; Treadway, C. R.; Barton, J. K.;
Zewail, A. H. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 6014.
(121) Kelley, S. O.; Holmlin, E. R.; Stemp, E. D. A.; Barton, J. K. J. Am.
Chem. Soc. 1997, 199, 9861.
(122) Lewis, F. D.; Wu, T.; Xhang, Y.; Letsinger, R. L.; Greenfield, M. R.;
Wasielewski, M. R. Science 1997, 277, 673.
(123) Lewis, F. D.; Wu, T.; Liu, X.; Letsinger, R. L.; Greenfield, S. R.;
Miller, S. E.; Wasielewski, M. R. J. Am. Chem. Soc. 2000, 122, 2889.
(124) Lewis, F. D.; Zhu, H.; Daublain, P.; Fiebig, T.; Raytchev, M.; Wang,
Q.; Shafirovich, V. J. Am. Chem. Soc. 2006, 128, 791.
(125) Lewis, F. D.; Zhu, H.; Daublain, P.; Cohen, B.; Wasielewski, M. R.
Angew. Chem., Int. Ed. 2006, 45, 7982.
(126) Lewis, F. D.; Zhu, H.; Daublain, P.; Sigmund, K.; Fiebig, T.;
Raytchev, M.; Wang, Q.; Shafirovich, V. Photochem. Photobiol. Sci.
2008, 7, 534.
(127) Andreatta, D.; Pérez Lustres, J. L.; Kovalenko, S. A.; Ernsting, N. P.;
Murphy, C. J.; Coleman, R. S.; Berg, M. A. J. Am. Chem. Soc. 2005,
127, 7270.
(128) Dohno, C.; Stemp, E. D. A.; Barton, J. K. J. Am. Chem. Soc. 2003,
125, 9586.
(129) Armitage, B.; Yu, C.; Devadoss, C.; Schuster, G. B. J. Am. Chem.
Soc. 1994, 116, 9847.
(130) Sanii, L.; Schuster, G. B. J. Am. Chem. Soc. 2000, 122, 11545.
(131) Vicic, D. A.; Odom, D. T.; Núñez, M. E.; Gianolio, D. A.;
McLaughlin, L. W.; Barton, J. K. J. Am. Chem. Soc. 2000, 122, 8603.
(132) Dotse, A. K.; Boone, E. K.; Schuster, G. B. J. Am. Chem. Soc. 2000,
122, 6825.
(133) Kawai, K.; Osakada, Y.; Takada, T.; Fujitsuka, M.; Majima, T. J. Am.
Chem. Soc. 2004, 126, 12843.
(134) Kawai, K.; Osakada, Y.; Fujitsuka, M.; Majima, T. Chem.sEur. J.
2008, 14, 3721.
(135) Kawai, K.; Takada, T.; Nagai, T.; Cai, X.; Sugimoto, A.; Fujitsuka,
M.; Majima, T. J. Am. Chem. Soc. 2003, 125, 16198.
(136) Kawai, K.; Osakada, Y.; Fujitsuka, M.; Majima, T. J. Phys. Chem.
B 2008, 112, 2144.
(137) Arkin, M. R.; Stemp, E. D. A.; Holmlin, R. E.; Barton, J. K.;
Hörmann, A.; Olson, E. J. C.; Barbara, P. F. Science 1996, 273, 475.
(138) Wan, C. Z.; Fiebig, T.; Schiemann, O.; Barton, J. K.; Zewail, A. H.
Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 14052.
(139) Lewis, F. D.; Liu, J.; Zuo, X.; Hayes, R. T.; Wasielewski, M. R.
J. Am. Chem. Soc. 2003, 125, 4850.
(140) Senthilkumar, K.; Grozema, F. C.; Guerra, C. F.; Bilkelhaupt, F. M.;
Lewis, F. D.; Berlin, Y. A.; Ratner, M. A.; Siebbeles, L. D. A. J. Am.
Chem. Soc. 2005, 127, 14894.
(141) Santhosh, U.; Schuster, G. B. J. Am. Chem. Soc. 2002, 124, 10986.
(142) O’Neill, M. A.; Barton, J. K. In Topics in Current Chemistry;
Schuster, G. B., Ed.; Springer: Berlin, 2004; Vol. 237, p 67.
(143) Boussicault, F.; Robert, M. Chem. ReV. 2008, 108, 2622.
(144) Seidel, C. A. M.; Shulz, A.; Sauer, M. H. M. J. Phys. Chem. 1996,
100, 5541.
(145) Caruso, T.; Carotenuto, M.; Vasca, E.; Peluso, A. J. Am. Chem. Soc.
2005, 127, 15040.
(146) Sheu, C.; Foote, C. S. J. Am. Chem. Soc. 1995, 117, 6439.
(147) Thorp, H. H. In Topics in Current Chemistry; Schuster, G. B., Ed.;
Springer: Berlin, 2004; Vol. 237, p 159.
(148) Johnston, D. H.; Glasgow, K. C.; Thorp, H. H. J. Am. Chem. Soc.
1995, 117, 8933.
(149) Shinde, S. S.; Maroz, A.; Hay, M. P.; Anderson, R. F. J. Am. Chem.
Soc. 2009, 131, 5203.
(150) Sistare, M. F.; Codden, S. J.; Heimlich, G.; Thorp, H. H. J. Am.
Chem. Soc. 2000, 122, 4742.
(151) Lewis, F. D.; Liu, X.; Hayes, R. T.; Wasielewski, M. R. J. Am. Chem.
Soc. 2000, 122, 12037.
Mechanisms for DNA Charge Transport
(152) Kurnikov, I. V.; Tong, G. S. M.; Madrid, M.; Beratan, D. N. J. Phys.
Chem. B 2002, 106, 7.
(153) Saito, I.; Takayama, M.; Sugiyama, H.; Nakatani, K. J. Am. Chem.
Soc. 1995, 117, 6406.
(154) Sugiyama, H.; Saito, I. J. Am. Chem. Soc. 1996, 118, 7063.
(155) Adhikary, A.; Khanduri, D.; Sevilla, M. D. J. Am. Chem. Soc. 2009,
131, 8614.
(156) Hall, D. B.; Holmlin, R. E.; Barton, J. K. Nature 1996, 382, 731.
(157) Saito, I.; Nakamura, T.; Nakatani, K.; Yoshioka, Y.; Yamaguchi,
K.; Sugiyama, H. J. Am. Chem. Soc. 1998, 120, 12686.
(158) Voityuk, A. A.; Jortner, J.; Bixon, M.; Rösch, N. Chem. Phys. Lett.
2000, 324, 430.
(159) Dandliker, P. J.; Nuñez, M. E.; Barton, J. K. Biochemistry 1998, 37,
6491.
(160) Yavin, E.; Stemp, E. D. A.; O’Shea, V. L.; David, S. S.; Barton,
J. K. Proc. Natl. Acad. Sci. U.S.A. 2006, 3610.
(161) Treadway, C. R.; Hill, M. G.; Barton, J. K. Chem. Phys. 2002, 281,
409.
(162) Anne, A.; Demaille, C. J. Am. Chem. Soc. 2008, 130, 9822.
(163) Wong, E. L. S.; Gooding, J. J. Anal. Chem. 2006, 78, 2138.
(164) Gorodetsky, A. A.; Buzzeo, M. C.; Barton, J. K. Bioconjugate Chem.
2008, 19, 2285.
(165) Ikeda, R.; Akaishi, A.; Chiba, J.; Inouye, M. ChemBioChem 2007,
8, 2219.
(166) Hartwich, G.; Caruana, D. J.; de Lumley-Woodyear, T.; Wu, Y.;
Campbell, C. N.; Heller, A. J. Am. Chem. Soc. 1999, 121, 10803.
(167) Lewis, F. D.; Kalgutkar, R. S.; Wu, Y.; Liu, X.; Liu, J.; Hayes, R. T.;
Miller, S. E.; Wasielewski, M. R. J. Am. Chem. Soc. 2000, 122,
12346.
(168) LeBard, D. N.; Lilichenko, M.; Matyushov, D. V.; Berlin, Y. A.;
Ratner, M. A. J. Phys. Chem. B 2003, 107, 14509.
(169) Lewis, F. D.; Liu, J.; Weigel, W.; Rettig, W.; Kurnikov, I. V.; Beratan,
D. N. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 12536.
(170) Huynh, M. H. V.; Meyer, T. J. Chem. ReV. 2007, 107, 5004.
(171) Stubbe, J.; Nocera, D. G.; Yee, C. S.; Chang, M. C. Y. Chem. ReV.
2003, 103, 2167.
(172) Reece, S. Y.; Hodgkiss, J. M.; Stubbe, J.; Nocera, D. G. Philos. Trans.
R. Soc. London, B 2006, 361, 1351.
(173) Seyedsayamdost, M. R.; Yee, C. S.; Reece, S. Y.; Nocera, D. N.;
Stubbe, J. J. Am. Chem. Soc. 2006, 128, 1562.
(174) Fiebig, T.; Wan, C.; Zewail, A. H. ChemPhysChem 2002, 3, 781.
(175) Gervasio, F. L.; Boero, M.; Parrinello, M. Angew. Chem., Int. Ed.
2006, 45, 5606.
(176) Anderson, R. F.; Shinde, S. S.; Maroz, A. J. Am. Chem. Soc. 2006,
128, 15966.
(177) Kobayashi, K.; Yamagami, R.; Tagawa, S. J. Phys. Chem. B 2008,
112, 10752.
(178) Yamagami, R.; Kobayashi, K.; Tagawa, S. J. Am. Chem. Soc. 2008,
130, 14772.
(179) Huber, R.; Fiebig, T.; Wagenknecht, H.-A. Chem. Commun. 2003,
15, 1878.
(180) Stemp, E. D. A.; Arkin, M.; Barton, J. K. J. Am. Chem. Soc. 1997,
119, 2921.
(181) Schiemann, O.; Turro, N. J.; Barton, J. K. J. Phys. Chem. B 2000,
104, 7214.
(182) Kumar, A.; Sevilla, M. D. J. Phys. Chem. B 2009, 113, 11359.
(183) Shafirovich, V.; Dourandin, A.; Geacintov, N. E. J. Phys. Chem. B
2001, 105, 8431.
(184) Takada, T.; Kawai, K.; Fujitsuka, M.; Majima, T. Chem.sEur. J.
2005, 11, 3835.
(185) Sobolewski, A. L.; Domcke, W. Phys. Chem. Chem. Phys. 2004, 6,
2763.
(186) Sobolewski, A. L.; Domcke, W.; Hattig, C. Proc. Natl. Acad. Sci.
U.S.A. 2005, 102, 17903.
(187) Abu-Riziq, A.; Grace, L.; Nir, E.; Kabelac, M.; Hobza, P.; de Vries,
M. S. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 20.
(188) Schwalb, N. K.; Temps, F. J. Am. Chem. Soc. 2007, 129, 9272.
(189) Crespo-Hernández, C. E.; de La Harpe, K.; Kohler, B. J. Am. Chem.
Soc. 2008, 130, 10844.
(190) Brillouin, L. In Horizons in Biochemistry; Kasha, M., Pullman, B.,
Eds.; Academic Press: New York, 1962; p 295.
(191) Sekiguchi, I.-S.; Sekiguchi, T. Phys. ReV. Lett. 2007, 99, 228102.
(192) Baba, Y.; Sekiguchi, T.; Shimoyama, I.; Hirao, N.; Nath, K. G. Phys.
ReV. B 2006, 74, 205433.
(193) Zakjevskii, V. V.; King, S. J.; Dolgounitcheva, O.; Zakrzewski, V. G.;
Ortiz, J. V. J. Am. Chem. Soc. 2006, 128, 13350.
(194) Close, D. M.; Øhman, K. T. J. Phys. Chem. A 2008, 112, 11207.
(195) Gabelica, V.; Rosu, F.; Tabarin, T.; Kinet, C.; Rodolphe, A.; Broyer,
M.; De Pauw, E.; Dugourd, P. J. Am. Chem. Soc. 2007, 129, 4706.
(196) Hübsch, A.; Enders, R. G.; Cox, D. L.; Singh, R. R. P. Phys. ReV.
Lett. 2005, 94, 178102.
(197) Haruna, K.-I.; Iida, H.; Tanabe, K.; Nishimoto, S.-I. Org. Biomol.
Chem. 2008, 6, 1613–1617.
Chemical Reviews, 2010, Vol. 110, No. 3 1661
(198) Vladimirov, E.; Ivanova, A.; Rösch, N. J. Phys. Chem. B 2009, 113,
4425.
(199) Yonemoto, E. H.; Saupe, G. B.; Schmehl, R. H.; Hubig, S. M.; Riley,
R. L.; Iverson, B. L.; Mallouk, T. E. J. Am. Chem. Soc. 1994, 116,
4786.
(200) Berlin, Y. A.; Grozema, F. C.; Siebbeles, L. D. A.; Ratner, M. A. J.
Phys. Chem. C 2008, 112, 10988.
(201) Page, C. C.; Moser, C. C.; Chen, X.; Dutton, P. L. Nature 1999,
402, 47.
(202) Gray, H. B.; Winkler, J. R. Q. ReV. Biophys. 2003, 36, 341.
(203) Balabin, I. A.; Beratan, D. N.; Skourtis, S. S. Phys. ReV. Lett. 2008,
101, 158102.
(204) Rösch, N.; Voityuk, A. A. In Topics in Current Chemistry; Schuster,
G. B., Ed.; Springer: Berlin, 2004; Vol. 237, p 37.
(205) Sponer, J.; Riley, K. E.; Hobza, P. Phys. Chem. Chem. Phys. 2008,
10, 2595.
(206) Wang, X. F.; Chakraborty, T. Phys. ReV. Lett. 2006, 97, 106602.
(207) Giese, B.; Amaudrut; Köhler, A.-K.; Sporman, M.; Wessely, S. Nature
2001, 412, 318.
(208) Voityuk, A. A. J. Chem. Phys. 2008, 128, 115101.
(209) Migliore, A.; Corni, S.; Varsano, D.; Klein, M. L.; Di Felice, R. J.
Phys. Chem. B 2009, 113, 9402.
(210) Troisi, A.; Orlandi, G. Chem. Phys. Lett. 2001, 344, 509.
(211) Troisi, A.; Orlandi, G. J. Phys. Chem. B 2002, 106, 2093.
(212) Mallajosyula, S. S.; Gupta, A.; Pati, S. K. J. Phys. Chem. A 2009,
113, 3955.
(213) Reha, D.; Barford, W.; Harris, S. Phys. Chem. Chem. Phys. 2008,
10, 5437.
(214) Kubař, T.; Elstner, M. J. Phys. Chem. B 2008, 112, 8788.
(215) Bongiorno, A. J. Phys. Chem. B 2008, 112, 13945.
(216) Kubař, T.; Elstner, M. J. Phys. Chem. B 2009, 113, 5653.
(217) Renger, T.; Marcus, R. A. J. Phys. Chem. A 2003, 107, 8404.
(218) Shih, C.; Museth, A. K.; Abrahamsson, M.; Blanco-Rodriguez, A. M.;
Di Bilio, A. J.; Sudhamsu, J.; Crane, B. R.; Ronayne, K. L.; Towrie,
M.; Vicek, A.; Richards, J. H.; renger and marcusWinkler, J. R.;
Gray, H. B. Science 2008, 320, 1760.
(219) Jorner, J.; Bixon, M.; Langenbacher, T.; Michel-Beyerle, M. E. Proc.
Natl. Acad. Sci U.S.A. 1998, 95, 12759.
(220) Xu, B.; Tao, N. Science 2003, 301, 1221.
(221) Hihath, J.; Chen, F.; Zhang, P.; Tao, N. J. Phys.: Condens. Matter
2007, 19, 215202.
(222) Mallajosyula, S. S.; Lin, J. C.; Cox, D. L.; Pati, S. K.; Singh, R. R. P.
Phys. ReV. Lett. 2008, 101, 176805.
(223) Hihath, J.; Xu, B.; Zhang, P.; Tao, N. Proc. Natl. Acad. Sci. U.S.A.
2005, 102, 16979.
(224) Dulic, D.; Tuukkanen, S.; Chung, C.-L.; Isambert, A.; Lavie, P.;
Filoramo, A. Nanotechnology 2009, 20, 115502.
(225) Barton, J. K.; Nuckols, C. Unpublished results.
(226) Takada, T.; Kawai, K.; Cai, X.; Sugimoto, A.; Fujitsuka, M.; Majima,
T. J. Am. Chem. Soc. 2004, 126, 1125.
(227) Takada, T.; Kawai, K.; Fujitsuka, M.; Majima, T. J. Am. Chem. Soc.
2006, 128, 11012.
(228) Jortner, J.; Bixon, M.; Voityuk, A. A.; Rösch, N. J. Phys. Chem. A
2002, 106, 7599.
(229) Sartor, V.; Boone, E.; Schuster, G. B. J. Phys. Chem. B 2001, 105,
11057.
(230) Yoo, J.; Delaney, S.; Stemp, E. D. A.; Barton, J. K. J. Am. Chem.
Soc. 2003, 125, 6640.
(231) Kendrick, T.; Giese, B. Chem. Commun. 2002, 18, 2016.
(232) Giese, B.; Wessely, S. Angew. Chem., Int. Ed. 2000, 39, 3490.
(233) Bixon, M.; Jortner, J. J. Am. Chem. Soc. 2001, 123, 12556.
(234) Bixon, M.; Jortner, J. Chem. Phys. 2002, 271, 393.
(235) Bixon, M.; Jortner, J. Chem. Phys. 2006, 326, 252.
(236) Liu, C.-S.; Hernandez, R.; Schuster, G. B. J. Am. Chem. Soc. 2004,
126, 2877.
(237) Goldsmith, R. H.; DeLeon, O.; Wilson, T. M.; Finkelstein-Shapiro,
D.; Ratner, M. A.; Wasielewski, M. R. J. Phys. Chem. A 2008, 112,
4410.
(238) Yu, Z. G.; Song, X. Phys. ReV. Lett. 2001, 86, 6018.
(239) Berlin, Y. A.; Burin, A. L.; Ratner, M. A. J. Am. Chem. Soc. 2001,
123, 260.
(240) Kelley, S. O.; Jackson, N. M.; Hill, M. G.; Barton, J. K. Angew.
Chem., Int. Ed. 1999, 38, 941.
(241) Hartwich, G.; Caruana, D. J.; de Lumley-Woodyear, T.; Wu, Y.;
Campbell, C. N.; Heller, A. J. Am. Chem. Soc. 1999, 121, 10803.
(242) Traub, M. C.; Brunschwig, B. S.; Lewis, N. S. J. Phys. Chem. B
2007, 111, 6676.
(243) Adhikary, A.; Kumar, A.; Khanduri, D.; Sevilla, M. D. J. Am. Chem.
Soc. 2008, 130, 10282.
(244) Jakobsson, M.; Stafström, S. J. Chem. Phys. 2008, 129, 125102.
(245) Schuster, G. B. Acc. Chem. Res. 2000, 33, 253.
(246) Conwell, E. M.; Rakhmanova, S. V. Proc. Natl. Acad. Sci. U.S.A.
2000, 97, 4556.
1662 Chemical Reviews, 2010, Vol. 110, No. 3
(247) Kuznetsov, A. M. Charge Transfer in Physics, Chemistry and
Biology: Physical Mechanisms of Elementary Processes and an
Introduction to the Theory; Gordon and Breach Science Publishers:
Amsterdam, 1995; p 2.
(248) Maniadis, P.; Kalosakas, G; Rasmussen, K. Ø.; Bishop, A. R. Phys.
ReV. B 2003, 68, 174304.
(249) Komineas, S.; Kalosakas, G.; Bishop, A. R. Phys. ReV. E 2002, 65,
061905.
(250) Chang, C.-M.; Castro Neto, A. H.; Bishop, A. R. Chem. Phys. 2004,
303, 189.
(251) Conwell, E. M. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 8795.
(252) Conwell, E. M.; Bloch, S. M. J. Phys. Chem. B 2006, 110, 5801.
(253) Conwell, E. M.; Park, J.-H.; Choi, H.-Y. J. Phys. Chem. B 2005,
109, 9760.
(254) Conwell, E. M.; Bloch, S. M.; McLaughlin, P. M.; Basko, D. M.
J. Am. Chem. Soc. 2007, 129, 9175.
(255) Conwell, E. M.; Basko, D. M. J. Phys. Chem. B 2006, 110, 23603.
(256) Middleton, C. T.; de La Harpe, K.; Su, C.; Law, Y. K.; Crespo-
Hernández, C. E.; Kohler, B. Annu. ReV. Phys. Chem. 2009, 60, 217.
(257) Schwalb, N. K.; Temps, F. Science 2008, 322, 243.
(258) Crespo-Hernández, C. E.; Cohen, B.; Kohler, B. Nature 2005, 436,
1141.
Genereux and Barton
(259) Buchvarov, I.; Wang, Q.; Raytchev, M.; Trifonov, I.; Fiebig, T. Proc.
Natl. Acad. Sci. U.S.A. 2007, 104, 4794.
(260) Kadhane, U.; Holm, A. I. S.; Hoffmann, S. V.; Nielsen, S. Phys.
ReV. E 2008, 77, 021901.
(261) Takaya, T; Su, C.; de La Harpe, K.; Crespo-Hernández, C. E.; Kohler,
B. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 10285.
(262) Lange, A. W.; Herbert, J. M. J. Am. Chem. Soc. 2009, 131, 3913.
(263) Hatcher, E.; Balaeff, A.; Keinan, S.; Venkatramani, R.; Beratan, D. N.
J. Am. Chem. Soc. 2008, 130, 11752.
(264) Uskov, D. B.; Burin, A. L. Phys. ReV. B 2008, 78, 073106.
(265) Voityuk, A. A. J. Chem. Phys. 2005, 122, 204904.
(266) Sen, S.; Andreatta, D.; Ponomarev, S. Y.; Beveridge, D. L.; Berg,
M. A. J. Am. Chem. Soc. 2009, 131, 1724.
(267) Voityuk, A. A. J. Chem. Phys. 2008, 128, 045104.
(268) Friedman, K. A.; Heller, A. J. Phys. Chem. B 2001, 105, 11859.
(269) Heller, A. Faraday Discuss. 2000, 116, 1.
(270) Merino, E. J.; Barton, J. K. Biochemistry 2007, 46, 2805.
(271) Merino, E. J.; Barton, J. K. Biochemistry 2009, 48, 660.
(272) Steenken, S.; Jovanovic, S. V. J. Am. Chem. Soc. 1997, 119, 617.
(273) Crespo-Hernández, C. E.; Close, D. M.; Gorb, L.; Leszczynski, J. J.
Phys. Chem. B 2007, 111, 5386.
CR900228F
 Chem. Rev. 2010, 110, 1663–1705 1663

Unexpected Inversions in Asymmetric Reactions: Reactions with Chiral

Metal Complexes, Chiral Organocatalysts, and Heterogeneous Chiral
Catalysts

Mihály Bartók
Department of Organic Chemistry, University of Szeged, Stereochemistry Research Group of the Hungarian Academy of Sciences,
Dóm tér 8, H-6720 Szeged, Hungary

Received July 1, 2009

Contents 4.5. Asymmetric Cyclopropanation 1698

4.6. Asymmetric Aziridination 1699
1. Introduction 1663 4.7. Summary 1699
2. Unexpected Inversions in Metal Complex 1664 5. Conclusion 1701
Catalyzed Asymmetric Reactions
6. List of Abbreviations 1701
2.1. Hydrogenation and Transfer Hydrogenation 1664
7. Acknowledgments 1702
2.2. Asymmetric Hydroformylation 1667
8. References 1702
2.3. Asymmetric Oxidations 1667
2.3.1. Sharpless Asymmetric Epoxidation 1667
1. Introduction
2.3.2. Asymmetric Oxidation of Sulfides 1667
2.4. Asymmetric Alkylations 1668 The recognition of the optical activity of organic com-
2.5. Asymmetric Aldol and Related Additions 1670 pounds, followed by the detailed analysis of optical isomer-
2.6. Other Asymmetric Carbon-Carbon Coupling 1671 ism,1 laid the foundation for the development of asymmetric
Reactions syntheses, a group of reactions of fundamental significance.
2.6.1. Asymmetric Henry Reaction 1671 Asymmetric syntheses, available in great varieties, are the
2.6.2. Asymmetric Mannich Reaction 1671 indispensable source of a large number of chiral compounds
that have become essential for human society. Some of the
2.6.3. Asymmetric Heck Reaction 1672
most important procedures that belong to this group are
2.7. Homogeneous Catalyzed Asymmetric 1673
Cycloaddition Reactions asymmetric organocatalytic, asymmetric homogeneous and
heterogeneous metal catalytic reactions, and enzyme-
2.7.1. Diels-Alder Addition 1673
catalyzed procedures. One of the main tasks of pertinent
2.7.2. 1,3-Dipolar Cycloaddition Reactions 1675 research is the development of chiral catalysts that enable
2.7.3. [3 + 2]-Cycloaddition Reaction 1676 the preparation of the required product in the highest possible
2.7.4. Other Asymmetric Reactions 1676 enantiomeric excess (ee). The value of ee depends on a
2.8. Summary 1676 number of known and yet unknown factors not only for a
3. Unexpected Inversions in the Organocatalytic 1678 given reaction but also for each individual compound within
Asymmetric Reactions a reaction type.
3.1. Asymmetric Alkylations 1678 Experimental observations occasionally reveal unknown
3.2. Asymmetric Aldol Additions 1679 phenomena, toosone of these is the so-called unexpected
3.3. Asymmetric Michael Reactions 1680 sense of enantioselectivity, the subject of the present review.
3.4. Asymmetric Baylis-Hillman Reaction 1682 This phenomenon has been mentioned in the literature under
3.5. Asymmetric β-Lactone Synthesis 1683 a variety of names including “unexpected inversion of
3.6. Asymmetric β-Lactams Synthesis 1683 enantioselectivity”, “change/dramatic change in the sense/
3.7. Asymmetric Aza-Henry Reaction 1684 direction of enantioselectivity”, “chirality inversion”, “switch
3.8. Summary 1684 of the expected chiral sense”, and “unexpected reversal of
4. Unexpected Inversions in Heterogeneous Catalytic 1685 the enantioselectivity/enantioselection/enantiodifferentiation/
Asymmetric Reactions chiral induction/stereochemistry/configuration”. Authors des-
4.1. Asymmetric Hydrogenation 1685 ignate unexpected experimental observations in their manu-
4.1.1. Hydrogenation of Ketones over Ni 1686 scripts as “dramatic”, “remarkable”, “surprising”, and
Catalysts “interesting”. Of course, these expressions do not facilitate
4.1.2. Hydrogenation of Activated Ketones over 1686 orientation in the literature.
Pt Catalysts Experimental results exhibiting a stereochemistry dif-
4.1.3. Hydrogenation over Pd Catalyst 1694 ferent from the relationships generally accepted for the
4.2. Asymmetric Aldol Addition 1696 sense of enantiodifferentiation (ED) will be referred to
4.3. Asymmetric Diels-Alder Addition 1697 as “unexpected inversion” or simply “inversion” in this
4.4. Asymmetric Hetero Diels-Alder Additions 1698 review. The detailed possible definition is as follows: the
absolute configuration of a product in a reaction is
10.1021/cr9002352  2010 American Chemical Society
Published on Web 10/29/2009
1664 Chemical Reviews, 2010, Vol. 110, No. 3
Mihály Bartók was born and raised in Szeged, Hungary. He received his
M.Sc. degree in organic chemistry from University of Szeged under the
direction of Professor Gabor Fodor. In 1966 he received his candidate of
chemistry degree in the Moscow State University under the supervision
of Professors N. I. Shuikin and A. F. Plate in the field of heterogeneous
catalytic organic reactions. He was dean of the Faculty of Science and
director of Department of Organic Chemistry in University of Szeged. He
is a member of the Hungarian Academy of Sciences (1987). Professor
Bartók became professor emeritus at University of Szeged in 2003. His
research interests are stereochemistry and mechanism of heterogeneous
catalytic organic reactions and heterogeneous enantioselective catalysis.
unexpected if, in earlier reactions using a chiral catalyst
of identical absolute configuration and substrates with not
significantly different structures, the absolute configuration
of the product was the opposite. There may be, however,
further criteria depending on the reaction type, especially
as more and more knowledge is accumulating. For
example, in the case of chiral metal complex catalysts,
metal and counterions may also play determinant roles in
addition to the chiral ligand (L*). When the roles of the
individual components of the chiral system have been
identified, however, the phenomenon gradually ceases to
be unexpected and a new unexpected result may appear.
In addition to the main components of a reaction, other
factors that may play a role are the experimental param-
eters, namely, the temperature, the solvents, the various
achiral additives, the L*/substrate or the chiral catalyst/
substrate ratio, or the concentration of the catalyst, to
mention just the most common ones.
When faced with the interpretation of a new phenom-
enon, a scientist would first examine any antecedents in
closely related fields. Such antecedents may be expected
to be found among homogeneous catalytic asymmetric
procedures employing the thoroughly investigated metal
complex catalysts as well as among organocatalytic
processes and heterogeneous catalytic enantioselective
reactions; the latter two fields have been developed
enormously in the past few years.
The objective of this manuscript is to draw, from the
collected experimental observations, some generalized
conclusions that may set further tasks. The results obtained
in the course of studying unexpected inversions are
important contributions to the understanding of the mech-
anisms of the reactions studied and, more specifically, to
the identification of the origin of chiral induction.
Naturally, the rapid development of and the volume of
data available on various reactions prevented us from a
full discussion of all fields. On the other hand, in most
cases, there is no opportunitysmainly due to consider-
ations of sizesto also include details published in easily
Bartók
accessible journals. For the same reason, the review of
other important reactions (e.g., addition of organometallic
reagents, polymerizations, asymmetric autocatalysis, etc.)
was also omitted from the program; these could be treated
at length on another occasion. The manuscript calls
attention to the antecedents of the relevant reaction by
citing a few characteristic publications and reviews.
Comprehensive reviews have been published mainly with
the aim of highlighting the possibility of novel syntheses
that enable the preparation of both enantiomers with the
help of the same chiral catalyst.2–5 Although unexpected
inversion has already been discussed in two short subsec-
tions in the reviews on heterogeneous catalytic enanti-
oselective hydrogenations,6 to our best knowledge similar
chapters/subchapters are absent from the monographs and
reviews addressing asymmetric organocatalytic and hetero-
geneous catalytic reactions.
This review gives insight into the homogeneous catalytic
asymmetric reactions, the organocatalytic reactions, and the
heterogeneous catalytic asymmetric reactions. The last part
attempts to give a full account of unexpected inversions
observed in the course of studies on heterogeneous metal
catalytic enantioselective hydrogenations. Finally, after a
critical analysis of the numerous experimental observations
described, the review ends with the compilation of generaliz-
able conclusions and the designation of further tasks.
Findings and interpretation of unexpected inversion have
provided important data with respect to the stereochemistry.
It is my hope that the present review catalyze similar
research.
2. Unexpected Inversions in Metal Complex
Catalyzed Asymmetric Reactions
In asymmetric reactions/syntheses of this type, chiral
induction is supplied by chiral molecules of natural origin
(amino acids, carbohydrates, alkaloids) and their synthetic
derivatives, or other synthetic chiral compounds. These chiral
molecules participate in asymmetric syntheses as the ligands
of metal complexes. The results obtained in the field of metal
complex catalyzed reactions have been the subject of
numerous monographs and reviews, which also give continu-
ous account of the state of these highly significant syntheses
(in addition to the well-known early works,7–13 see, e.g., the
reviews published in the past few years14–16).
As regards the stereochemical course of metal complex
catalyzed reactions, experimental observations have already
allowed the formulation of empirical rules for certain widely
used chiral ligands. However, similar relationships have not
been established for a large number of chiral ligands,
since the principal aim in these instances is the optimization
of the preparation of a given enantiomer in >90% ee. Perhaps
the most important parameter of optimization is the input of
the appropriate chiral ligand. Continuous completions and
corrections of the rules of the stereochemistry are, therefore,
not surprising.
2.1. Hydrogenation and Transfer Hydrogenation
Most of these methods were discovered by Kagan,
Knowles, and Noyori.17–19 Studies on the stereochemistry of
homogeneous asymmetric reactions have taken their models
from the hydrogenation and transfer hydrogenation of
prochiral compounds with CdC bonds and prochiral ketones
in the presence of Rh and Ru complexes. The large number
Unexpected Inversions in Asymmetric Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1665

Scheme 1. Hydrogenation of E and Z Isomers over Scheme 3. Two Competitive Pathways in the Hydrogenation
DuPhos-Rh Catalyst of Methyl 2-(Acetylamino)cinnamate Using TRAP-Rh
Catalyst

of studies allowed the establishment of certain empirical

rules.11–14 According to ref 14: (i) Remarkably, the DuPhos-
Rh system provides excellent enantioselectivities for both Z
and E isomeric substrates, and the hydrogenation products
are formed with the same configuration (Scheme 1). (ii) An
(R,R)-(S,S)-i-Bu-TRAP-Rh catalyst provides 96% ee for
hydrogenation of a tetrahydropyrazine carboxamide deriva-
tive; interestingly, a related (R,R)-(S,S)-Me-TRAP-Rh catalyst
provides the hydrogenation product with a different config-
uration (Scheme 2).20b
In the asymmetric catalytic hydrogenation of R-(aceta-
mido) acrylates using chiral TRAP-Rh catalysts, remark-
able effects of ligand P-substituent and hydrogen pressure
were observed. Only a small change in the ligand
P-substituent influenced the catalytic activity and enan-
tioselectivity dramatically, indicating that a unique and suppresses path B, because low hydrogen pressure is
efficient asymmetric environment is created around the unfavorable for the oxidative addition of hydrogen (Scheme
metal center of TRAP-Rh complex. The most surprising 3).21a
unusual experimental observation made in studies on the In the same reaction in the presence of triethylamine, the
hydrogenation of compounds containing prochiral CdC (S)-BINAP-Ru complex promoted the formation of the (S)-
bonds20,21 is the effect of hydrogen pressure on the sense product, whereas (S)-BINAP-Rh complex catalyzed that of
of ED (Table 1).21a the (R)-product in higher ee (84%).21b In contrast, the (R)-
The dramatic effect of hydrogen pressure may be explained BINAP-Rh and (R)-BINAP-Ru complexes give the same
on the basis of the assumption that the hydrogenation result with the opposite product configuration (see, e.g., in
proceeds through two competitive pathways. Path A involves ref 12, p 38). Today these results no more sound unusual,
the coordination of olefin to complex followed by the but in 1985 it could well be unexpected to obtain products
oxidative addition of hydrogen, giving (R)-product prefer- of opposite configurations using the chiral ligand with the
entially. Path B involves the oxidative addition of hydrogen same configuration.
prior to the coordination of the CdC bond, favoring the Burgess et al. synthesized novel iridium complexes and
formation of (S)-product. Decrease in the hydrogen pressure studied their effects in the asymmetric hydrogenation of

Scheme 2. Hydrogenation of a Tetrahydropyrazine Carboxamide Derivative

Table 1. Inversion in Hydrogenation of Dehydroamino Acid Esters

entry ee (%) conv. (%) solvent temp. (°C) conv. (%) ee (%)
1 36 100 DCM 15 100 77* a
2 64 100 DCM 30 100 69*
3 18 100 MeOH 30 100 79*

a
Unexpected inversion, marked with * in the manuscript.
1666 Chemical Reviews, 2010, Vol. 110, No. 3 Bartók

Scheme 4. Inversion in Enantioselective Hydrogenation of 2-(4′-Methoxyphenyl)-but-1-ene

Table 2. Inversion in Transfer Hydrogenation of Acetophenone

entry ee (%) conv. (%) L* additive, Add/L complex L* additive, Add/L conv. (%) ee (%) ref
1 53 88 L1a i-PrOK, 1 Ir(III)compl. L1a i-PrOK, 10 84 46* 22
2 65 92 L1b i-PrOK, 1 Ir(III)compl. L1b i-PrOK, 10 83 48* 22
3 61 95 L1c i-PrOK, 1 Ir(III)compl. L1c i-PrOK, 10 79 48* 22
4 53 93 L1d i-PrOK, 1 Ir(III)compl. L1d i-PrOK, 10 86 49* 22
5 55 23 L1e i-PrOK, 1 Ir(III)compl. L1e i-PrOK, 10 93 47* 22
6 95 78 L2 i-PrOK, 1 Ir(III)compl. 22
7 90 59 L2 i-PrOK, 10 Ir(III)compl. 22
8 20* 35 Lb i-PrONa, 5 [Ru(pcy)Cl2]2 La i-PrONa, 5 83 92 23a
9 81* 49 L2b i-PrONa, 5 [Ru(pcy)Cl2]2 L2a i-PrONa, 5 5 0 23a
10 92* 57 L2b i-PrONa, 10 LiCl, 10 [Ru(pcy)Cl2]2 23a
11 95* 88 L2b i-PrONa, 5 LiCl, 5 [RhCl2 cp] 23a

arylalkenes. The most striking21c of the numerous experi-
mental results obtained under high-pressure/low-temperature
or high-temperature/low-pressure conditions are shown in
Scheme 4.
The authors interpret their results according to ref 21a,
emphasizing that this unexpected observation implies that it
is of a steric, rather than an electronic, origin. Deuterium-
labeling experiments provide evidence for other types of
competing mechanisms that lead to D incorporation at
positions that do not correspond to direct addition to the
double bond.21c
In the field of transfer hydrogenation, experimental
observations of unexpected inversion are also found in
the literature, two examples of which are presented in
Table 2.
Monosulfonated diamines, with an axially chiral biaryl
backbone in combination with different Ir(III) complexes,
were investigated in the catalytic transfer hydrogenation
of acetophenone under i-PrOH/i-PrOK conditions (entries
1-5). The resulting catalysts showed an unexpected base-
dependent enantioselectivity. Less base than chiral catalyst
resulted in a (R)-secondary alcohol; excess of base gave
the (S)-enantiomer.22 By altering the amount of base, the
authors were able to influence not only the activity but also
the enantioselectivity of the reaction. With a small increase
from 0.6 to 1.4 equiv, the ee changed from 82% (R) to 49%
(S). This very unusual finding was observed with all ligands
1a-e, but not with ligand 2 (entries 6 and 7).22 The authors
assumed that the availability of the binding sites at different
pH levels may play a crucial role in this change of
mechanism. However, the structures of the two catalysts at
low and high base concentrations have not been published
since then.
The other example for unexpected inversion (entries
8-11) is a consequence of a subtle change in the structure
of the chiral ligand.23a A change in the ligand structure,
namely, replacement of the amide oxygen in Boc-protected
amino acid amides by sulfur La f Lb, L2a f L2b, and
modification of the catalytic system with a lithium salt, lead
to a novel and most efficient class of Ru and Rh catalysts
for the asymmetric transfer hydrogenation of aromatic
ketones in propan-2-ol. In addition, the replacement of the
amide functionality for the corresponding thioamide resulted
in a dramatic switch of the product enantioselectivity. Under
optimized conditions, the secondary alcohol products were
obtained in high yield and enantioselectivity (up to 97% ee)
using only 0.25 mol % catalyst loading. The authors also
confirmed the phenomenon in substituted acetophenones.23
The switch of the sign of enantioselectivity on going
from amides to the appropriate thioamides may arise from
Unexpected Inversions in Asymmetric Reactions
a different mode of coordination due to significant
differences in the acidity of the NH functions in amides
and carbamates relative to those in thioamides. In their
recent reports published in this field, however, they supply
no evidence supporting the interpretation of the phenom-
enon discovered.23b,c
2.2. Asymmetric Hydroformylation
Although evidence on the stereochemistry of hydroformy-
lation was already available at the beginning of the 1980s,24
the formulation of relationships depending on the absolute
configuration of the chiral ligand requires further experi-
ments.25 Consiglio and Pino performed asymmetric hydro-
formylation of butene isomers under identical experimental
conditions and observed ED sense inversion depending on
the structure of the butenes involved.24 Inversion was not
unexpected, since the reactions of but-1-ene, (E)-but-2-ene,
and (Z)-but-2-ene exhibit different stereochemistries due to
the excessive differences between the steric structures of
these compounds. Table 3, however, shows unexpected
inversion depending on the temperature in the presence of
the same Pt complexes.26,27
Kollár et al. and later Hanson and co-workers reported a
very interesting Pt-catalyzed hydroformylation of styrene in
which a change from (S)- to (R)-enantioselectivity was seen
as a function of temperature: catalysis by [(2S,4S)-BDPP]-
Pt(SnCl3)Cl gave the branched aldehyde with 63% ee (S) at
40 °C but 17% ee (R) at 100 °C.26
Kollár et al. proposed that the reversal of enantioselectivity
might be due to a temperature-dependent change in the
conformation of the catalyst’s six-membered chelate ring. It
is more likely, however, that the enantioselectivity-determin-
ing step changed with temperature.27 Casey et al. have
reported deuterioformylation studies and the CO and H2
pressure dependence of ee, which show that, at low tem-
perature, enantioselectivity is set by largely irreversible
platinum hydride addition to styrene. At high temperature,
in contrast, platinum hydride addition is reversible and
enantioselectivity is set by a combination of partially
reversible alkyl migration to CO and hydrogenolysis of the
Table 3. Inversion in Hydroformylation of Styrene
yield temp. CO H2 yield
entry ee (%) (%) (°C) (psi) (psi) n/ia (%) ee (%) ref
1 50 580 580 20 63
2 17* 18 100 580 580
3 40 500 500 2.2 19 60
4 10* 3 100 500 500 2.4
a
n/i ) normal aldehyde/isoaldehyde.
Scheme 5. Stereochemistry of the Sharpless Asymmetric Epoxidation
Chemical Reviews, 2010, Vol. 110, No. 3 1667
platinum acyl intermediate. The explanation as to why the
selectivity-determining step changes as a function of tem-
perature is based on a detailed analysis of reaction kinetic
data obtained at various temperatures.27
2.3. Asymmetric Oxidations
2.3.1. Sharpless Asymmetric Epoxidation
Scheme 5 represents the experimentally verified basic
scheme of the stereochemical course of enantioselective
oxidation in the presence of (R,R)-DMT and (S,S)-DMT, in
the case of the widely investigated model compound trans-
hex-2-en-1-ol.
In 1980 Sharpless and Katsuki achieved the enantioselec-
tive epoxidation of primary allylic alcohols.28 The details of
the oxidation are also summarized in some of the pertinent
reviews.12,29 According to Scheme 5 in the case of (R,R)-
DMT as ligand, the product formed in excess is the (2S,3S)-
epoxide, whereas in the case of (S,S)-DMT as ligand, it is
the compound with the opposite absolute configuration,
(2R,3R)-epoxide.30a
According to experiments reported in 2002, in the presence
of soluble polymer-supported tartrate ester ligands, the
reaction surprisingly exhibited a stereochemistry of the
opposite direction, depending on the molecular weight of
the polyethylene glycol monomethyl ether (MPEG), under
otherwise identical experimental conditions (Table 4, entries
2 and 3).30b Janda et al. investigated this fascinating effect
in more detail (Table 4, entries 4-14).31
The results have clarified that the enantioselectivity of this
reaction can be reproducibly reversed solely as a function
of the molecular weight of the appended PEG.31 By preparing
a range of tartrate ligands with varying PEG chains lengths,
the reversal was found to occur within a molecular weight
change of only 800. As the PEG chain did not affect the
inherent chirality of the ligand, the enantioreversal was
proposed to occur as a result of two Ti-ligand complexes
that differ in their molecularity of ligand, one monomeric in
ligand and the other dimeric. These investigations into the
nature of Sharpless asymmetric epoxidations catalyzed by
MPEG tartrate esters have revealed several interesting details
regarding the mechanism of catalysis.31
2.3.2. Asymmetric Oxidation of Sulfides
Unexpected inversion was observed in the course of the
oxidation of sulfides over Ti-complex catalysts containing
chiral (R,R)-p,p′-disubstituted 1,2-diphenylethane-1,2-diol
ligands (Table 5).32a The use of the p-CF3 substituted L*
26 dramatically decreased the ee (entry 4) and, unexpecedly,
26 gave the p-tolyl methyl sulfoxide with opposite stereochem-
27 istry (R) with respect to those obtained with other diols
27 (entries 1 and 3). According to the authors, L* containing
coordinating moieties (OMe and CF3) can lead to the
formation of different Ti complexes with different structures
and, consequently, with new reactivities and opposite senses
of ED.
The formation of (R)-sulfoxide may be regarded as
unexpected, since the configurations of L* are identical.
1668 Chemical Reviews, 2010, Vol. 110, No. 3 Bartók

Table 4. Inversion in the Sharpless Asymmetric Epoxidation

entry ee (%) yield (%) ligand yield (%) ee (%) ref

1 85 94 (R,R)-DMT 30b
2 70 81 (R,R)-TA-MPEG750 30b
3 (R,R)-TA-MPEG2000 85 93* 30b
4 (R,R)-TA-MPEG2000 54 75* 31
5 (R,R)-TA-MPEG750 66 67* 31
6 (R,R)-TA-MPEG550 85 8* 31
7 75 95 (R,R)-TA-MPEG350 31
8 80 93 (R,R)-TA-MPEG207 31
9 89 83 (R,R)-TA-MPEG163 31
10 85 72 (R,R)-TA-MPEG119 31
11 93 75 (R,R)-TA-MPEG75 31
12 78 63 (S,S)-TA-MPEG2000 31
13 96 96 (R,R)-DIPT 31
14 96 87 (R,R)-DIPT + MPEG2000 31

Table 5. Inversion in Oxidation of Sulfides used that involved either the generation of the nucleophile
in situ, using dimethyl malonate and BSA with catalytic
potassium acetate in DCM, or the use of preformed nucleo-
phile, sodium dimethyl malonate in tetrahydrofuran (THF).
The palladium catalyst was formed by mixing the allyl
entry ee (%) yield (%) Ar R X yield (%) ee (%) palladium chloride dimer with 2 molar equivalents of the
1 p-tolyl Me H 62 80 chiral ligand.
2 Ph Bn H 73 99
3 26* 70 p-tolyl Me CF3 According to the authors, it is striking that, despite each
4 Ph Bn CF3 80 18 ligand possessing identical backbone chirality, ligand L1
gives (S)-enantiomer in excess while ligands L2 and L3 both
give (R)-enantiomer in excess under each of the reaction
conditions. To offer a hypothesis for the dramatic reversal
of enantioselection, the authors examined the possible
conformation of the allyl intermediates for this reaction,
which could lead to the observed enantiomers of the product
(Figure 1).
Pd-catalyzed allylic alkylations of the rac-1,3-diphenyl-
Similar observations have been reported by other laboratories 2-propenyl esters with the dimethyl malonate nucleophile
in the oxidation of sulfides in the presence of L*32b of similar using carbohydrate bidentate phosphinites35b and various
structures and (R)-BINOL.32c chiral mono- and bis(oxazoline) ligands also were studied.36,37
Hoarau et al. have discovered an example of asymmetric
2.4. Asymmetric Alkylations synthesis leading to the formation of (S) or (R) isomers, both
This subsection enumerates examples for allylation and in high ee (92% and 90%, respectively) by using an
Friedel-Crafts alkylation. The Pd(0)-catalyzed allylation enantiogenic catalyst based on ligand L4 or L5 characterized
developed by Trost and Tsuji is useful for creating organic by the same chiral backbone and configuration (Table 6, entry
frameworks that have a variety of polar functional groups.33 3). It was demonstrated that this shift in the control of the
The reaction is formally viewed as a combination of an allylic ED was due to the presence of a hydroxy group on the side
cation and a carbanion. Further results are summarized in chain. Since the configurations of the chiral C atoms in the
reviews.8–11,34 Two examples for unexpected inversion are ligands L4 and L5 are identical, formation of the (R)-product
presented below. of the opposite configuration by the effect of L5 can be
The effectiveness of the chiral chelate nitrogen-phosphorus regarded as unexpected inversion. The direction of the
ligands derived from (S)-valine was investigated by Anderson nucleophilic attack and, consequently, the absolute config-
et al.35a using the standard palladium catalyzed substitution uration of the product formed are determined by the structure
of 1,3-diphenyl-2-propenyl acetate with dimethylmalonate of the ICs, which in turn depends on L* (Scheme 6).36b The
(Table 6, entries 1 and 2). Two common procedures were stuctures of ICs were determined by XRD.
Unexpected Inversions in Asymmetric Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1669

Table 6. Inversion in the Pd-Catalyzed Allylations

entry ee (%) yield (%) L* catalyst temp. (°C) L* yield (%) ee (%) ref
1 80 96 L2 Pd(II) r.t. L1 94 83* 35a
2 60 93 L3 Pd(II) r.t. L1 94 83* 35a
3 90* 98 L5 Pd(II) 26 L4 98 92 36a

Other studies also confirmed allylations exhibiting unex-
pected stereochemistries: (i) O-allylations of phenols38a using
chiral P,N-heterodonor ligands; (ii) C-allylation38b using
chiral P,S-heterodonor ligands with a binaphthalene frame-
work; and (iii) N-allylation38c with novel metallocene-based
planar chiral diphosphine ligands.
Tang and co-workers observed dramatic solvent effects
in the highly enantioselective alkylation of indoles with
alkylidene malonates using novel trisoxazoline-Cu(II) com-
plexes as catalysts (Scheme 7).39
The use of alcohols as the solvents not only accelerates
the reaction dramatically but also improves the ee. Strongly

Figure 1. Possible intermediate π-allyl Pd complexes.

Scheme 6. Inversion in the Nucleophilic Attack Caused by the OH-Nu Hydrogen Bonding

Scheme 7. Inversion in the Friedel-Crafts Alkylation

1670 Chemical Reviews, 2010, Vol. 110, No. 3 Bartók

coordinating solvents (e.g., alcohols) gave the product with

(S) configuration, while weakly coordinating solvents af-
forded the product with the opposite configuration. It was
postulated that the reversal of enantioselectivity resulted from
the change of the coordination geometry of copper center in
different solvents.39 Although the attempt to develop single
crystals of trisoxazoline-Cu(II) was unsuccessful, the authors
could support the hypothesis by the preparation of (R)- and
(S)-products and their further investigations by 1H NMR and
in situ spectroscopic methods (see Schemes 6-10 and Figure
1 in ref 39).

2.5. Asymmetric Aldol and Related Additions

The aldol addition is one of the most important asymmetric
syntheses of carbon-carbon bonds with new chiral centers.7,11–13 Figure 2. Assumed intermediate complexes in aldol reaction.
This subsection describes three examples of aldol additions
that can be classified as unexpected inversions. According Table 8. Inversion in Aldol Addition of Methyl Ynones
to Table 7, in the presence of L1 and L2 1,2-diamines as
ligands, the Mukaiyama-type aldol reaction yields products
with opposite absolute configurations in high ee.40
Both diamines L1 and L2 were prepared from L-proline,
and the absolute configuration of the C2 is S in both cases. entry time (h) solvent temp. (°C) yield (%) ee (%)
The difference is the fusion point of the benzene ring 1 4 toluene 0 63 44 (+)
connected to the piperidine moiety. It was surprising that 2 4 toluene -25 27 72* (-)
the slight difference in the structure of the chiral sources 3 7 THF 0 61 83 (+)
4 2 THF -25 nd 69* (-)
completely reversed the enantiofacial selectivity. In addition
to the unique selectivities, the reaction provides convenient
methods for the preparation of both enantiomers of syn-2,3-
dihydroxy thioesters. As shown in Table 7, adducts with the
2S,3R configuration were obtained by chiral diamine L1,
while adducts with the 2R,3S configuration were produced
by chiral diamine L2. In every case, the selectivities were
very high; almost complete syn selectivities and more than
98% ee of the syn adducts were obtained.40
In order to clarify the origin of the selectivities, the
structures of the tin(II) triflate-chiral diamine complexes
were examined. A possible explanation for the stereoselec-
the asymmetric aldol addition of methyl ynones to pyruval-
tivity is the formation of bicyclic tin(II) intermediates of
dehyde ketals in the course of the optimization studies (Table
different conformations (Figure 2). In chiral diamine L1, the
L1-Sn(II) conformation is favored; on the other hand, in 8).41
chiral diamine L2, the L2-Sn(II) conformation is preferred, According to the authors’ opinion, the experimental data
which was supported by nuclear Overhauser effect (NOE) clearly suggest that dinuclear Zn catalytic species are formed
experiments. as the reaction progresses (Table 8). The recognition of
In another selected example, Trost et al. reported dramatic enantioselective autoinduction by the authors may prove to
reversal of ED due to a temperature-dependence effect in have important implications on related systems.41
Table 7. Inversion in the Mukaiyama Aldol Addition

entry ee (%) yield (%) R yield (%) ee (%)

1 98 86 Ph 82 98*
2 98 61 Et 63 98*
3 98 74 MeCHdCH 81 98*
4 99 71 PhCHdCH 80 98*
5 97 86 2-furyl 86 99*
6 96 86 MeCHdCHCHdCH 78 99*
Unexpected Inversions in Asymmetric Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1671

Scheme 8. Inversion in Aldol Reaction with Trimethylsilyl Thioenolates

Table 9. Inversion in the Henry Reaction

Unexpected inversions were also observed by Evans et
al.42a and by Kunieda and co-workers42b in the aldol reaction
with trimethylsilyl thioenolate as a competent aldol reagent
catalyzed by chiral bis(oxazoline)-Cu(II) complexes (Scheme
8). The inversion was observed using PyBOX ligands, too. ee yield yield ee
In order to characterize the structure of ICs and to interpret entry (%) (%) R L* solvent (%) (%)
the experimental observations, Evans et al. applied various 1 50 90 Me L1a THF 80 77*
methods to study the effect of BOX-Cu complexes,42a,c often 2 39 54 Me L1c THF 73 47*
tested previously in other reactions. In these studies, unex- 3 47 76 Me L2a THF 56 38*
4 63 30 Me L2c THF 25 6*
pected inversion occurred in the presence of catalysts 5 Me L1d hexane 97 84*
consisting of ligands of identical configuration and structure 6 70 55 Me L1e
and of identical metal ions; the only difference is in 7 Bu L1d hexane 92 82
8 Ph(CH2)2 L1d hexane 95 71
counterion composition (SbF6-, TfO-) (see section 2.7 for
details).
Kunieda et al.42b have described a new class of sterically
congested and conformationally rigid chiral bis(oxazoline)
ligands with methylene and ethylene spacers between the
oxazoline rings, from which the derived Cu(OTf)2 complexes
serve as catalysts in asymmetric aldol reactions, resulting in
the reversal of ee, depending on six- and seven-membered
chelate sizes. The difference between the structures of the
two ligands lies in the methylene and ethylene spacers; still
the inversion can be regarded as unexpected, because the
configurations of the chiral carbon atoms of the ligands are
identical. The experimental antecedents, however, pointed
to the key role of stereochemistry, namely, the length of the
spacer between the oxazoline rings.

2.6. Other Asymmetric Carbon-Carbon Coupling

Reactions
2.6.1. Asymmetric Henry Reaction
The Henry reaction is a C-C coupling reaction between
a nitroalkane and a carbonyl compound.13,43 β-Nitro alcohols
can be produced by this reaction. Unexpected inversion of
ED in asymmetric Henry reaction was achieved with the
same chiral ligand by changing the Lewis acid center from
Cu(II) to Zn(II)44 (Table 9). In the course of their earlier
experiments, the authors demonstrated that, in the presence
of L1-Cu(II) and L2-Cu(II) complexes, the (S)-products
are formed in higher ee.44a In the presence of the same L1-
and L2-BOX ligands, under identical experimental condi-
tions and using the same reactant, but under the effect of
Et2Zn, the (R)-products are found to be formed in higher
ee.44b In both cases, ee values are up to 70-80%.
The various studies and data of literature show that the
NH groups in ligands play a very significant role in
controlling both the yields and ED of the Henry products.
The authors assumed intermediate B for the L*-Cu(II)
Figure 3. Assumed intermediate Cu(II)- and Zn(II)-complexes in
asymmetric Henry reaction.
complex catalyzed reaction and presumed intermediate C in
Et2Zn catalytic system (Figure 3).44b According to Figure 3,
in the case of Cu(II)-catalyst, the R-keto ester is activated
by Cu(II) and the nitronate is oriented by hydrogen bond (B
IC); therefore, the reaction takes place from the Si-face. In
the case of Et2Zn, a dinuclear Zn catalyst forms. Nitronate
is oriented by two Zn atoms (C IC): the reaction proceeds
from the Re-face.
2.6.2. Asymmetric Mannich Reaction
The Mannich-type reactions are important and useful
synthetic methods for the construction of nitrogen-containing
molecules.45 The Mannich reaction was also utilized in
asymmetric homogeneous catalyzed reactions between gly-
cine derivatives and imines. Unexpected inversion observed
in a reaction of this type is presented in Table 10.46
1672 Chemical Reviews, 2010, Vol. 110, No. 3 Bartók

Table 10. Inversion in the Mannich-type Reaction

entry ee (%) yield (%) anti/syn L* L* syn/anti yield (%) ee (%)

1 97 97 85/15 L1 L3 91/9 89 99*
2 99 96 88/12 L2 L4 91/9 96 96*
3 99 93 92/8 L1a L3
4 99 92 96/4 L2a L4a 95/5 97 99*

a
-78 °C.

2.6.3. Asymmetric Heck Reaction

Figure 4. Calculated structures of intermediate complexes in an
enantioselective Mannich reaction.
The change in the sense of diastereoselectivity elicited by
the ligand of identical configuration under nearly identical
conditions can be regarded as unexpected. The inversion can
be attributed to a change in the electronic factor of the
ligands. L1 and, furthermore, the L2 ligands containing an
electron-donating methoxy group mainly catalyze the forma-
tion of anti-product, whereas ligands L3 and L4 having
electron-withdrawing CF3 and F groups promote the forma-
tion of syn-product.46
The stereochemistry can be understood based on the
fact that the Re-face is blocked by the i-Pr group of the
ligand (Figure 4B). A working model is proposed by the
authors to correlate the observed stereochemistry. Imine
approaches the Ca anion center in a staggered conformation
with the N atom pointing to Cu. The Ts group occupies
the valley formed by the two arene groups if the two rings
are electron-deficient in ligand L4, giving a (2S,3R)
product (Si-face for imine). The imine attacks the Ca with
its Re-face when the arene rings are electron-rich in ligand
L2.
Table 11. Inversion in Intermolecular Heck Reaction
The antecedents of the asymmetric Heck reaction have
been reviewed.47 In 2008, two studies reporting on dramatic
stereo- and enantiodivergency in the intramolecular asym-
metric Heck reaction were published. The results reported
in ref 48a are shown in Table 11.
According to experimental data in Table 11, a dramatic
switch in ED is realized using (S)-1 or (S)-2 ligands (with
H or with methyl as substituents). In the case of (S)-1
ligands, (R)-products were formed, while using (S)-2
ligands, (S)-products were formed. X-ray analysis of (S)-
1a-PdCl2 and (S)-2a-PdCl2 complexes as well as density
functional calculations on the stereochemistry and mech-
anism provide a rational explanation for these interesting
observations. Ligands S1 and S2 coordinate with Pd in
syn and anti seven-membered-ring conformations, respec-
tively; due to the stronger trans effect of P over N, the
Pd-bound phenyl group prefers trans to P in the transition
state so that it is more activated with a lower barrier for
addition to Pd-coordinated 2,3-dihydrofuran substrate;
because of the unsymmetrical environment caused by the
two phenyl groups on P, the trans-syn-Si and the trans-
anti-Re transition states are more favorable than others
(Figure 5).48a
In the other article,48b Rubina et al. described the effect
of the novel PHOX ligands with rigid chiral cyclopropyl
backbones outlined in Figure 6 on the stereochemistry of
the Heck reaction. They obtained unexpected inversion
similar to that shown in Table 11. The new observations of

entry ee (%) yield (%) L* Ar L* yield (%) ee (%)

1 78 62 (S)-1a Ph (S)-2a 86 81*
2 93 79 (S)-1b Ph (S)-2b 78 10*
3 87 66 (S)-1b p-F-C6H4 (S)-2a 83 87*
Unexpected Inversions in Asymmetric Reactions
Figure 5. syn-Conformation (S)-1a-Pd and the anti-conformation
(S)-2a-Pd intermediate complexes in the Heck reaction.
Figure 6. Novel PHOX ligands with a chiral cyclopropyl backbone
in the Heck reaction.
each of these reports48a,b significantly enriched our knowledge
of the asymmetric Heck reaction.49
In addition to the reactions described in section 2.6,
unexpected inversions have also been observed in other
carbon-carbon coupling reactions (Wittig reaction,50 cy-
anosilylations51).
2.7. Homogeneous Catalyzed Asymmetric
Cycloaddition Reactions
In this chapter, examples of unexpected inversion observed
in Diels-Alder (DA) reactions, 1,3-dipolar cycloadditions,
and one [3 + 2]-cycloaddition are reviewed.
Table 12. Inversion in the Diels-Alder Reaction
a
IL: 1-ethyl-3-methylimidazolium bis[(trifluoromethanesulfonyl)imide]. b SILC: silica supported ionic liquid phase catalyst. c Configuration refers
to the stereogenic center indicated by the *. L*: see Figure 8; mainly endo additions occur.
Chemical Reviews, 2010, Vol. 110, No. 3 1673
2.7.1. Diels-Alder Addition
Among homogeneous catalyzed asymmetric reactions, the
DA reaction, a test reaction widely studied (especially with
BOX chiral ligands), is the one suitable for an attempt of
drawing conclusions based on the analysis/evaluation of the
unexpected experimental results recorded.14,52 A characteristic
set of unexpected experimental observations on C2-sym-
metrical bis(oxazoline)-Lewis acid complexes (BOX-
complexes) are summarized in Table 12. Among various
chiral Lewis acid catalysts, those containing the chiral BOX
ligands have proven useful in many applications since the
pioneering work of Corey and Evans.53a,b
The recently published review on the diverse stereochem-
istry of the DA reaction52g offers, among others, the following
highlights: (i) The chiral C2-symmetric bis(oxazoline)-Lewis
acid complexes coordinate dienophiles, which react easily
and enantioselectively with dienes. (ii) It appears that the
stereochemistry of the reaction is affected by the substituents
of BOX and their configurations, the coordination geometry
of metal cations, counterions, additives, solvents, immobi-
lizations, and experimental conditions. (iii) The 3D structure
of the catalyst complex is formed by Lewis acids and BOX
ligands. (iv) ICs containing readily dissociating (noncoor-
dinating, like SbF6- or ClO4-) anions usually have tetrahedral
structure, whereas the structure of those containing coordi-
nating TfO- anions is usually octahedral. (v) For example,
in the case of (R)-Me-BOX-Ph-Mg(II) complexes if ICs with
tetrahedral structure favor the formation of (S)-DA products,
then ICs with octahedral structure will favor the formation
of (R)-DA products. (vi) When several chiral centers were
present in a Ph-BOX ligand, their configuration had a
dramatic influence on enantioselectivity.
The first observation of unexpected inversion in the DA
reaction was reported by Kobayashi et al.54a,b The chiral
catalyst obtained from the Lewis acid Yb(OTf)3, (R)-BINOL,
1674 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 7. Kobayashi BINOL Lewis acid catalyst in the Diels-Alder
reaction.
and a tertiary amine catalyzes the formation of the (R) and
(S) enantiomers in high enantioselectivity by the effect of
and depending on various achiral additives, under otherwise
identical conditions54a (Table 12, entry 1). The inversion is
strongly dependent on the specific coordination number of
Yb(OTf)3. The structure of the chiral catalyst proposed by
the authors is shown in Figure 7. Two years later, Ghosh et
al. confirmed the same result using chiral BOX-Lewis acid
catalysts.54c
BOX-magnesium perchlorate and magnesium triflate
chiral catalysts (Table 12) have also been successfully applied
in the DA reaction.55 Catalysts containing magnesium
perchlorate used in conjunction with any of the three (R)-
MeBOX-Ph ligands studied (Figure 8) yielded the (S)-product
in different ee values in the absence of H2O and ROH. When
2 equiv of OH-ligands were added, the ED is reversed and
(R)-products were obtained (Table 12, entries 2-4). In the
case of magnesium triflate, always (R)-products were formed
Figure 8. BOX ligands in enantioselective Diels-Alder reactions.
Figure 9. Stereochemistry of a Diels-Alder reaction catalyzed by tetrahedral or octahedral intermediate BOX-Mg complexes.
Bartók
(entries 5-7). The structures of the different Mg(II) com-
plexes were verified by NMR measurements. The magnesium
perchlorate intermediates proved to be tetrahedral complexes
in the absence of water and alcohols; water and alcohols,
however, can expand the coordination number of magnesium
perchlorate, and octahedral complexes are formed (Figure
9). These experimental results are in agreement with the
above conclusions. In the reactions described, DA addition
products of (R) and (S) configuration were obtained in ee
over 90%.55 It has to be noted that DA addition studies have
also been performed with PyBOX-Lewis acid complexes,
and the formation of (S)- and (R)-DA products in higher ee
was observed. Because of the significantly different structure
of these catalysts, however, these experimental results cannot
be classified as conversions associated with unexpected
inversion.52d,53c
Ionic liquid (IL) effect was observed on the reversal of
configuration for the (S)-BOX-Mg(II) and (S)-BOX-Cu(II)
catalyzed reaction in both homogeneous and heterogeneous
media (entries 9-13).56 Compared with reaction performed
in dichloromethane or diethyl ether, an enhancement in ee
is observed with a large increase in reaction rate. In addition,
for nonsterically hindered bis(oxazoline) ligands, that is,
phenyl functionalized ligands, a reversal in configuration is
found in the IL, compared with molecular solvents. Unex-
pected inversion was observed using heterogenized com-
Unexpected Inversions in Asymmetric Reactions
plexes in DA and hetero-DA reactions56–58 (entries 10, 14,
and 15; in detail, section 4).
Surprisingly, the reports on unexpected inversions in the DA
reaction do not address the effect of the configuration of the
BOX ligand on enantioselectivity.55–58 The reason for this may
be that no results have been obtained using BOX ligands and
Lewis acids that differ only in the configuration of the chiral
carbon atoms but are otherwise identical, under identical
experimental conditions. Thus, it is not possible to propose a
general relationship. Unfortunately, the review recently pub-
lished does not adopt a definite standpoint on the effect of (S)-
and (R)-BOX ligands on the sense of ED in the DA reaction.52g
In some cases, inaccurate information can also be found (e.g.,
(R)-1 BOX data among (S)-1 BOX data). On the basis of
experimental data adequate for this type of comparison
(Table 12, entries 2, 10, and 12), both the (R)- and (S)-Me-
BOX-Ph-Mg(ClO4)2 and the (S)-Me-BOX-Ph-Mg(OTf)2
catalysts promote the formation of the (S)-DA product. In
contrast, the (R)-Me-BOX-Ph-Mg(ClO4)2/Mg(OTf)2 catalysts
give the (R)-DA product at different temperatures. It cannot
be determined whether the reversal ee is due to the change
in BOX configuration, the presence of a different Lewis acid,
or the change in experimental conditions.
Although a large number of experimental data points on
the DA reaction could be collected, these do not yet allow
the formulation of relationships that would explain the
majority of the data. In agreement with the opinionsand
hopessexpressed by Desimoni et al. in 2006, “In enanti-
oselective catalysis with BOX complexes there are still
unanswered questions, but given the spectacular development
of the field, the authors are confident that these will be solved
in the near future.”52g
2.7.2. 1,3-Dipolar Cycloaddition Reactions
The second most important enantioselective pericyclic
reaction after the Diels-Alder reaction is 1,3-dipolar
Scheme 9. Inversion in Asymmetric [1,3]-Cycloaddition of Nitrones with Ketene Acetals
Table 13. Inversion in 1,3-Dipolar Cycloadditions of Nitrones
entry ee (%) yield (%) R L*a
1 82* 100 H (R)-Me-BOX-Ph
2 79* 72 Me (R)-Me-BOX-Ph
3 Me (S)-Me-BOX-Ph
4 70* 100 H (R)-Me-BOX-Ph
a
L* see Figure 8.
Chemical Reviews, 2010, Vol. 110, No. 3 1675
cycloaddition.52d,f,60 Numerous experiments have been per-
formed that used Lewis acids and complexes containing
BINOL, BOX, and PyBOX chiral ligands. The new results
were reviewed by Desimoni et al. in 2003 and 2006.52d,g
Unexpected inversion was first observed by Schreeren et al.,
who studied 1,3-dipolar cycloaddition between nitrones and
ketene acetals over chiral oxazaborolidine Lewis acid
catalysts in 199561 (Scheme 9).
As demonstrated in Scheme 9, a dramatic solvent effectsas
the authors put itson enantioselectivity was observed.61b
When the catalyst was prepared in BH3 · THF (cosolvent is
THF), the value of ee was 62% (-), whereas if it was
prepared in BH3 · SMe2 in the presence of diphenyl ether
cosolvent, remarkable reversal of the ee (58 (+)) of the
reaction occurred (dimethyl sulfide probably has a role in
this reaction). Some characteristic data on the formation of
enantiomers of different configurations, obtained using chiral
BOX ligands, are listed in Table 13.
Mostly endo-addition takes place in the reaction. Enantiose-
lectivity is found to be dependent on the presence of MS 4 Å.
The stereochemistry of the process can be interpreted similarly
to that of the DA reaction, i.e., it is determined by the
coordination geometry of metal cations; ICs are octahedral in
the presence of water, whereas in its absence (i.e., in the
presence of MS), they have tetrahedral structure. The former
favor the formation of the (3R,4S)-product, whereas the latter
promote that of the product of the opposite absolute configuration.
In addition to the above, Kawamura and Kobayashi found
relatively high ee (85-96%) by the effect of the chiral Lewis
acid catalyst Yb(OTf)3 + (S)-BINOL + N-methyl-bis[(R)-
(1-naphthyl)ethyl]amine in the presence of MS 4 Å and
preferential formation of the product with the opposite
configuration in its absence (76-88%).63 They conclude their
report with stating that “Further studies to clarify the role of
MS 4 Å from a mechanistic point of view are now in
progress.” However, new information has not been published
since 1999.
Lewis acid R1 yield (%) ee (%) ref
MgI2 Ph 100 48 62a
MgI2 Ph 73 46 62a,b
MgI2 Bn 82 rac. 62b
Mg(ClO4)2 Ph 100 48 62c,d
1676 Chemical Reviews, 2010, Vol. 110, No. 3 Bartók

2.7.3. [3 + 2]-Cycloaddition Reaction therefore, will attack from the bottom face of C2-1a; hence,
the enantioselectivity is reversed.64
Inversion was recently observed in the case of [3 +
2]-cycloaddition, using ligands of identical configuration.64
Some experimental data are presented in Table 14. 2.7.4. Other Asymmetric Reactions
As regards the stereochemistry of the process, on first sight In 1998 Sibi et al. reported their experiments in which
an unexpected inversion took place, since inversion occurred both enantiomers were obtained in the conjugate addition
by the effect of L* of identical configuration (although they of R,β-unsaturated pyrazole amides in the presence of
contain either a primary amino group or a dimethylamino O-benzylhydroxylamine, BOX-complexes, and Lewis acids
group), in the presence of iminoesters of diverse composi- (depending on the latter), under otherwise identical
tions under identical experimental conditions.64 As soon as conditions.70a Scheme 10, however, represents a temperature-
the main properties of the reaction mechanism became known dependent reversal of stereochemistry in asymmetric con-
to the authors, however, the inversion ceased to be surprising. jugate amine addition of R,β-unsaturated oxazolidinone
The reason for this is that, in the case of an L* containing amide in the presence of the same Lewis acid.70b
a primary amino group, there is a hydrogen bonding Specifically, at room temperature the (R)-product and at
interaction between L*, Ag+, and the substrate. Although -60 °C the (S)-product is formed in higher ee. The authors
the use of hydrogen bonding to accelerate or catalyze certain varied several experimental parameters (temperature, Lewis
reactions has been well-documented,65 reversal of enanti- acids, BOX ligands, substrate structure), mainly with the aim
oselectivity directed by hydrogen bonding has been rarely of elucidating the general character of the reaction and of
reported.36a,66 Interpretation of the stereochemistry of the maximizing ee. In the authors’ opinion, expounded in detail
reaction was based on experiments using iminoesters and in the manuscript, the probable cause of the phenomenon is
L*s of different structures, density-functional theory studies that more than one complex may be present at a given
on the structure of ICs, and, last but not least, the results of temperature. In spite of the fact that the manuscript presents
1
H NMR titration measurements of hydrogen-bonding com- several reactions in which the only difference between the
plexes. These studies confirmed the existence of the hypo- conditions of the formation of the two enantiomers is in
thetic ICs of dimethylamino (C2-1a) and primary amino temperature, the inversion is still unexpected, because it is
(C2-1b) type (Figure 10).64 not characteristic of the majority of the reactions.
Figure 10 indicates that it is favorable for C2-1b to be
attacked from the top face, while in C2-1a, the dimethy-
lamino group cannot form hydrogen bonds, and the methyl
2.8. Summary
group will cause steric repulsion. The dimethyl maleate, Reports on unexpected inversion may be classified on the
Table 14. Inversion in [3 + 2]-Cycloaddition of Iminoesters basis of several different criteria. In Table 15, the collected
data are listed following the order of discussion according
to reactions, indicating a few parameters. The data in the
individual columns of Table 15 reflect the diversity of
unexpected inversion: (i) Inversion may occur in a variety
of reactions. (ii) Experimental data are available for cases
entry R L* temp. (°C) yield (%) ee (%) with L* of a wide range of types. (iii) In addition to high ee
1 p-ClC6H4 1a 0 95 -76 values (entries 10, 12, 15, 18, 19, 27, and 28), medium and
2 p-ClC6H4 1b 0 91 83* low ee values also occur, since in many cases optimization
3 p-ClC6H4 1c 0 95 -84
4 p-ClC6H4 1d 0 94 84*
has not been undertaken. (iv) The publications call attention
5 p-ClC6H4 1c -25 95 -92 to the diversity of effects causing inversion. (v) In many
6 p-ClC6H4 1d -25 90 92* cases, especially in more recent reports, the suggested
7 2-naphthyl 1c -25 91 -87 interpretations are also supported by experimental results.
8 2-naphthyl 1d -25 98 91*
The reviewer could next set out to identify any possible
relationships on the basis of expediently chosen parameters.
Theoretically, such parameters could be reaction types, chiral
ligands, Lewis acids (cations, counterions), other experi-
mental conditions, or effects (supposed to bring about the
unexpected inversion). In accordance with very recently
published reviews, I have to note that, at present, no
experimental data suitable for the formulation of generalized
conclusions are available. Looking at the data in Table 15,
it appears that, in the absence of suitable experimental data,
it is as yet impossible to attempt even a qualitative
comparison of the structural differences among chiral ligands
and of other effects causing unexpected inversion. As regards
the effects of the structural differences of ligands on
inversion, according to the pertinent definition these can only
be minor differences, since the absolute configurations of
the L* pairs participating in the reaction must be identical
Figure 10. Optimized structures of C2-1b and C2-1a inter-
(otherwise the inversion would not be unexpected). In nearly
mediate complexes of asymmetric [3 + 2]-cycloaddition (the one-half of the examples listed, BOX-type ligands partici-
hydrogen atoms that are not involved in the reactions are omitted pated in the reactions, whereas in the rest, ligands of diverse
for clarity). structures were present.
Unexpected Inversions in Asymmetric Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1677

Scheme 10. Inversion in Enantioselective Conjugate Amine Addition

Table 15. Inversions in Metal Complex-Catalyzed Reactions

interpret- Table/Scheme in
entry reactions chiral metal complexes Rmax/Smax effects ationa manuscript ref
1 hydrogenation (R,R)-(S,S)-EtTRAP-Rh(III) 18/79 H2 pressure A Table 1 21a
2 hydrogenation oxazoline-imidazoline-Ir(III) -64/+89 temperature H2 pressure A, C Scheme 4 21c
3 transfer hydrogenation Ax. biaryl-Ir(III) 60/50 [Base] A, C Table 2 22
4 transfer hydrogenation amide or sulfamide-Ru(III) 20/92 L* structure A, B Table 2 23
5 hydroformylation (S,S)-BDPP-PtCl(SnCl3) 17/68 temperature A Table 3 26
6 hydroformylation (S,S)-BDPP-PtCl(SnCl3) 10/60 temperature D Table 3 27
7 Sharpless epoxidation PEG-TA-Ti(IV) 75/67 additive: MPEG A, B Table 4 31
8 oxidation of sulfides Ar-diols-Ti(IV) 26/80 CF3 groups on Ph-s A, B Table 5 32a
9 allylation chiral N,P-Pd(0) 80/83 subst. on N A, B Table 6 35a
10 allylation BOX-diol/Bz-Pd(II) 90/92 H bonding D Table 6 36a,b
11 alkylation trisoxazoline-Cu(II) 65/90 solvent D Scheme 7 39
12 aldolization L-proline amines-Sn(II) 98/98 subst. on N D Table 7 40
13 aldolization bisdiphenyl prolinol-Zn2 -69/+83 temperature A, C Table 8 41
14 aldolization (S)-BOX-Cu(II) 91/64 TfO-, SbF6- A, C Scheme 8 42a
15 aldolization (S,R)-BOX-Cu(II) 92/80 spacer-length A, B Scheme 8 42b
16 Henry reaction BOX-Cu(II) and BOX-Zn(II) 84/76 N-H effect on Cu A, C Table 9 44b
17 Mannich reaction MOX-ferrocenyl-Cu(II) syn/anti electronic factor of L* D Table 10 46
18 Heck reaction Pfaltz-Helmchen P,N-Pd(II) 87/87 IC conform-ation D Table 11 48b
19 DA reaction (R)-BINOL-Ln(III) 81/93 achiral additives A, B Table 12 entry 1 54a
20 DA reaction (R)-BOX-Mg(II) 76/94 coordination geometry D Table 12 entries 2-8 55
21 DA reaction (S)-BOX-Mg(II) (S)-BOX-Cu(II) 79/62 ionic liquid A Table 12 entries 9-13 56
22 DA reaction (S)-BOX-Cu(II) 33/59 heterogenization C Table 12 entry 14 57
Table 39 entries 8-10
23 DA reaction (S)-BOX-Mg(II) 30/60 heterogenization D Table 12 entry 14 58
Table 39 entries 13-15
24 DA reaction (S)-BOX-Cu(II) (S)-BOX-Zn(II) 64/98 Cu(II) or Zn(II) D Table 12 entries 16 and 17 59
25 1,3-dipol. ca. oxazaborolidine -62/+79 solvent A, B Scheme 9 61
26 1,3-dipol. ca. (R)-BOX-Mg(II) 48/82 additive: 4 Å MS D Table 13 62
27 1,3-dipolar cycloadd. (R)-BINOL-Yb(III) -96/+88 additive: 4 Å MS A, B 63
28 [3 + 2]-ca. PPFA-AgOAc -92/+92 H bonding D Table 14 64
29 cyclopropanation (S)-BOX-Cu(II) 34/42 heterogenization, solv. D Table 41 68
30 aziridination (S)-BOX-Cu(II) 43/82 heterogeniz. A Table 42 69a
31 conj. addition (S,R)-BOX-MgBr2 50/43 temperature A, B Scheme 10 70b
a
A ) assumed, B ) planned, C ) there are some experiments, and D ) verified.

In the case of BOX-type L* pairs, the following can be
established regarding the effect of minor structural changes
in chiral ligands on ED: (i) In the case of allylation in the
presence of BOX ligands containing OH groups, ED was
controlled by the H bonding structure of the IC (entry 10).
(ii) In aldol addition, the sense of ED may depend on a
different counteranion (entry 14) or chelate ring size (entry
15). (iii) In the Henry reaction, in the case of Cu-BOX, ED
is controlled by H bonding, whereas in the case of Zn-BOX,
this is not possible (entry 16). (iv) In DA reaction, in the
case of t-Bu-BOX and Ph-BOX complexes, inversion of the
sense of ED is due to the different conformations of the IC
(entry 24). (v) In DA reaction, ED is controlled by the
coordination geometry of the central metal cation (tetrahedral
or octahedral geometries) (entry 20).
In the case of L* pairs of diverse structures, the following
can be stated about the effects of minor structural changes
in chiral ligands: (i) In transfer hydrogenation on Ru catalyst
containing amide/thioamide L* pairs, due to the acidity of
the NH functions, inversion may arise from a different mode
of coordination (entry 4). (ii) In the sulfide f sulfoxide
reaction on p,p′-disubstituted-1,2-diphenylethane 1,2-diols,
the inversion may arise from formation of new Ti complexes
with CF3 substituents on Ph groups (entry 8). (iii) The
inversions have been interpreted on the basis of different
conformations of IC complexes: in the allylation using
valphos-N,P-Pd(0) catalysts (entry 9), in the aldolization
using L-proline diamine-Sn(II) catalysts (entry 12), in the
Heck reaction using phenyloxazolinephosphine-Pd(II) cata-
lyst (entry 18). (iv) In Mannich reaction using oxazoline-
based ferrocenyl phosphine-CuClO4 catalyst, the origin of
the inversion has electronic character (entry 17). (v) In [3 +
2]-cycloaddition, inversion may be interpreted by the forma-
tion of H bonding, in the case of L* containing a primary
amino group, whereas in the case of a tertiary amine such
an interaction is not possible (entry 28).
As regards the role of experimental conditions (tempera-
ture, solvents, additives, heterogenized catalysts, and, in the
case of hydrogenations, hydrogen pressure), on the basis of
the suggestions/conclusions of the reports compiled in this
manuscript, it is impossible to derive any additional experi-
mentally verified statements of a more specific nature.
1678 Chemical Reviews, 2010, Vol. 110, No. 3
In the course of studying the pertinent literature, I found
catalysts with BOX ligands to be the most intensively studied
and, therefore, the best elucidated subject. The coordination
geometry of metal cations has a determinant role in develop-
ing the structure of the IC responsible for ED. It has also
been shown that the geometry of the metal cation is
profoundly influenced by counteranions and various addi-
tives. It is important to stress that it was systematic studies
on the unexpected stereochemical changes observed in the
case of BOX complexes that made possible the successful
formulation of generalizable relationships for BOX-type
catalysts. This and the knowledge of the properties of metal
ions and counterions have permitted us to design certain
chemical processes utilizing this type of chiral catalyst.
3. Unexpected Inversions in the Organocatalytic
Asymmetric Reactions
Although organocatalysts have been used in organic
chemistry for decades, their utilization has not skyrocketed
until the years after 2000, as illustrated by the list in a figure
in ref 71. The rapid development of organocatalysis is also
confirmed by the monographs published in the past few
years.71–75 Our group started to study the Michael-type
addition catalyzed by cinchona alkaloids at the end of the
1990s. Although studies on the stereochemistry of the
reaction yielded some unexpected results,76 our research
capacity was fully engaged in enantioselective heterogeneous
catalytic hydrogenations.6a
The present state of research in the field of the stereo-
chemistry of enantioselective organocatalytic reactions al-
ready allows the formulation of generalized relationships in
the case of certain reactions and catalysts. However, it has
not been possible to define similar relationships in the large
number of reaction types involved and for most of the
organocatalysts applied, due to a lack of required experi-
mental materials, because the main objective in these studies
is the production of the given enantiomers in >90% ee. That
may be the reason why the stereochemical relationships
directing certain reactions have not yet been elucidated, and
no review calling attention to the significance of the
observations made to date on unexpected inversion has been
published.
Table 16. Inversion in the Cinchona-Catalyzed Alkylation
entry ee (%) yield (%) base
1 KOH
2 NaOH
3 KOH
4 40* 96 NaOH
5 KOH
6 40* 93 NaOH
7 KOH
8 38* 80 NaOH
Bartók
In accordance with the objectives listed in the Introduction,
this section will describe a number of organocatalytic
asymmetric reactions involving unexpected inversion, pub-
lished in the literature.
3.1. Asymmetric Alkylations
Among the procedures for creating C-C bonds, asym-
metric organocatalytic alkylations are easily performable
reactions.71–74,77 The salts of chiral organic bases are espe-
cially often utilized as phase-transfer catalysts.77
According to the experimental data of Table 16, new
cinchonidinium salts bearing a 3,5-dialkoxybenzyl group
show an alkaline metal base-dependent reversal of enanti-
oselectivity when used as phase-transfer catalysts in the
asymmetric alkylation of N-(diphenylmethylene)glycine iso-
propyl ester with benzyl bromide.78a The use of potassium
hydroxyde as base in this alkylation reaction afforded the
(S)-enantiomer, whereas using sodium hydroxide under the
same conditions afforded the corresponding (R)-enantiomer.
The nature of the solvent and the temperature seems to
play important roles in the switching of stereoselectivity
when changing the base. The concentration of the base also
seems to be crucial to the change in stereoselectivity. The
authors have assumed that the presence of the C9 alkoxy
groups of the catalyst is a key factor in the observed inversion
of enantioselectivity. A similar observation is described in
ref 78b. Studies showing that the configuration of the product
is determined by that of the chiral atoms of the cinchona
alkaloid catalyst, irrespective of the bulkiness of the sub-
stituents of C9-OH, are also interesting,78c–f because dif-
ferent experimental observations also exist, especially in
heterogeneous metal-catalyzed enantioselective hydrogena-
tions (see below).
Denmark and Fu79a call attention to an unexpected chiral
formamide catalyzed allylation in the “Enantioselective
Catalysis” thematic issue14 of Chemical ReViews (Table 17).
The chiral catalyst in stoichiometric amount promotes the
allylation of the aldehyde to give the (R)-adduct in 68%
ee.79b,c Interestingly, when a catalytic amount of catalyst is
used, the (S)-product is obtained in low yield and selectivity.
A change in the sense of enantioselectivity is clearly
indicative of the operation of dual catalytic pathways for
Catalyst* temp. (°C) yield (%) ee (%)
1 -20 80 24
1 -20 70 24
2 -20 80 58
2 -20
2 -40 91 66
2 -40
3 -20 98 44
3 -20
Unexpected Inversions in Asymmetric Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1679

Table 17. Inversion in Allylation with Chiral Formamide Table 18. Inversion in the L-Proline- and L-Prolinol-Catalyzed
Catalyst Direct Aldol Reaction

entry Catalyst* conversion (%) selectivity (%) eeanti (%) eesyn (%)
1 L-proline 60 97 82 54
ee yield Catalyst* Catalyst* yield ee 2 D-proline 68 97 -81a -56a
(%) (%) (mol %) HMPA time (mol %) (%) (%) ref 3 L-prolinol 40 86 -22a * -26a *
32* 12 10 0 7h 50 45 44 79b a
30* 25 25 0 7h 100 81 68 79b The excess enantiomers had opposite configuration in comparison
100 7h 25 33 94 79c with those obtained in the reactions catalyzed by L-proline.
100 7h 100 89 96 79c
100 14 ha 20 80 98 79b
Sauer-Wiechert reaction82b and found that proline is an
effective organocatalyst for intermolecular direct asymmetric
aldol reactions.83a The manuscript describes several examples
for unexpected inversion observed in L-proline-catalyzed
asymmetric aldol addition.81b,83
According to Table 18, the asymmetric organocatalytic
aldol reaction has been extended to ketone + R-fluoro-β-
a
Solvent: EtCN. keto ester aldol addition.76b It has been shown that this
unprecedented reaction can be carried out with readily
formamides as well. The HMPA as an additive enhances available chiral amine catalysts, obtaining good enantiose-
the yield and ee and accelerates the catalytic cycle. The lectivities in the reaction of the R-monofluorinated com-
reason why HMPA increases the yield and ee remains unclear pounds. Surprisingly, when L-prolinol was used as catalyst,
in the author’s opinion. In the interpretation of the stereo- the sense of the ED was opposite to that obtained with
chemistry of the reaction, a cyclic chairlike transition L-proline. The direct aldol reaction catalyzed by L-prolinol
structure was assumed (Table 17). shown in Table 18 may be considered unexpected, since the
Another example for unexpected inversion depending on configuration of the product formed is identical with that
a slight modification of substrate structure is shown for the obtained in the case of D-proline rather than L-proline.
case of asymmetric alkylation in Scheme 11: alkylation of Since many proline derivatives have become popular,84
2-oxygenated diphenylmethane derivatives using s-BuLi and the authors synthesized new L-prolinamide derivatives with
(-)-spartein gave ee’s up to 60% with allylbromide. When rigid structures and axial and central chirality for the purpose
compounds with a free hydroxy in the 2-position were of the experiments outlined in Table 19 and tested them in
alkylated, the selectivity was reversed. Alkylations with the aldol reaction indicated.85a In the experiments, molecular
methyl electrophiles were poorly selective.80a sieves were used as water scavengers. Several unexpected
According to the authors, this reversal of selectivity opens events were observed, the interpretation of which, for the
up the possibility of obtaining either enantiomer of a given time being, was not addressed by the authors. Most con-
derivative, as desired, by using a protected or unprotected spicuously, out of the 14 experiments reported in this study,
starting material. This greatly increases the possibilities for only in one case was a product with (2R,1′R) configuration
this chemistry since only one enantiomer of sparteine is formed. As the authors say, “It was noteworthy that trans-
readily available. The reason for these results is not clear. 4-hydroxy-L-proline-derived organocatalysts 3c and 3d gave
The change in the sense of stereoselection from reactions of the opposite sense of asymmetric induction.” On the basis
ether derivative to those of hydroxy compound is difficult of the experimental results available, it is impossible to
to explain. The authors assume these reactions involve a formulate an unambiguous, definitive standpoint on the role
thermodynamic/kinetic resolution of an equilibrating pair of of the chirality of the organocatalysts studied in determining
complexes.80a Excellent results also have been obtained by the stereochemistry of the process. It appears as if the
several researchers using sparteine in asymmetric alkylation determinant factors of the stereochemistry were other than
of benzylic methylene group.80b,c these absolute configurations. The authors propose the key
role of MSs. They found that the presence of water had a
remarkable effect on catalytic activity and stereoselectivity.
3.2. Asymmetric Aldol Additions Perhaps the most characteristic experimental result is that
Reviews published recently verify the importance of aldol catalyst 3d with MS 4 Å forms a (2R,1′R) adduct (entry 3),
additions.71,72,74,81 List et al.82a reinvestigated the Hajos-Eder- whereas with MS 3 Å it forms a (2R,1′S) adduct (entry 4)
Scheme 11. Inversion in Organocatalytic Allylation of Diphenylmethane Derivatives
1680 Chemical Reviews, 2010, Vol. 110, No. 3 Bartók

Table 19. Inversion in L-Prolinamide-Catalyzed Aldol Reaction

a
0 °C, 24 h. b 3 Å MS.

Table 20. Inversion in the Chiral Phosphoramide-Catalyzed Aldol Reaction

entry ee (%) yield (%) solvent Catalyst* (L*) yield (%) ee (%)
1 92 92 DCM (S,S)-L
2 DCM (R,R)-L 87 92
3 EtCN (S,S)-L 88 90*

with reversed ee under identical experimental conditions.85a 3.3. Asymmetric Michael Reactions
Further research along these lines can be expected to yield
important information. Beside aldol addition, Michael additions are the most
It is mentioned that the steric and stereoelectronic effects commonly applied C-C bond-forming organocatalyzed
that control the enantioselectivity in the cross-aldol asymmetric syntheses.72,74,77a,81b,c Four examples of Michael-
addition of acetone to isatin catalyzed by L-proline have type asymmetric reactions will be presented below, with three
been studied by means of density functional theory (DFT) using cinchona and one using proline as catalyst.
and atoms in molecule (AIM) calculations.85b This reaction Table 21 summarizes the experimental data on unexpected
results in a reversal of enantioselectivity compared with the inversion in the reaction between 2-acetylbutyrolactone and
corresponding cross-aldol addition to 4,6-dibromoisatin and methyl vinyl ketone.76a Interestingly, QN gave higher ee
aldehydes. Because of the relatively large difference between values than CD, while QD gave slightly lower ee values than
the two reactants compared, the result obtained cannot be CN. The senses of the ee induced by the four basic cinchona
regarded as unexpected; therefore, it is not described in detail alkaloids were surprising. The levorotatory enantiomer was
in this manuscript.
Table 21. Inversion in the Cinchona-Catalyzed Michael
Significant solvent effect and rate enhancement were also Reaction of 2-Acetylbutyrolactone to Methyl Vinyl Ketone
observed in the chiral phosphoramide catalyzed aldol reac-
tions of aldehydes with trichlorosilyl enolates as competent
aldol reagents (Table 20).86
In DCM as solvent, (S)-aldol was produced in high optical
yield in the presence of (S,S)-L catalyst (entry 1), whereas
(R)-aldol was formed in the presence of (R,R)-L catalyst
entry Catalyst* yield (%) a
[R]D25 ee (%)
(entry 2). In propionitrile, however, in the presence of (S,S)-L
catalyst, the product had the configuration opposite to the 1 CD 95 -6.2 8
one expected, i.e., the (R)-product was produced in high 2 QN 95 23 33*
3 CN 82 17.4 25
optical yield. This unexpected inversion also aroused the 4 QD 84 -15.1 22*
attention of the authors of the recently published mono-
graph.72 The reports do not propose an explanation for the
a
The specific rotations of the optically pure products [R]D25: +69.5
and -69.5 (ethyl alcohol, c 10).
unexpected inversion.
Unexpected Inversions in Asymmetric Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1681

Table 22. Inversion in the Cinchona-Catalyzed Michael Reaction of Dimethyl Malonate to Chalcon

entry ee (%) yield (%) solvent yield (%) ee (%)

1 61 96 DMSO
2 56 92 toluene
3 46 94 DCM
4 [bmim]PF6 99 50*
5 [bmim]BF4 97 44*
6 42 97 [bpy]BF4

Table 23. Inversion in the Cinchonine-Catalyzed Michael Reaction of Benzyl Phosphonate to p-Chloronitrostyrene

entry ee (%) yield (%) additive entry additive yield (%) ee (%)
1 100 81 none
2 100 15 Me4-ethylenediamine 7 N-Me-morpholine 25 80*
3 100 17 OdP(NMe2)3 8 4-dimethylaminopyridine 51 60*
4 100 56 ethylenediamine 9 Et2O 60 60*
5 100 55 ethylene glycol 10 Et2Oa 55 100*
6 100 44 12-Crown-4
a
Used as solvent without any THF.

obtained in excess with CD and QD and the dextrorotatory
by the use of CN and QN as catalysts. The result obtained
is doubly surprising. On the one hand, the members of the
catalyst pair CD-QN containing C atoms of identical
configuration (8S, 9R) and those of the CN-QD pair (8R,
9S) catalyzed the formation of products with opposite
configurations in higher ee. On the other hand, in the first
pair, it was QN containing an OMe group, whereas in the
second pair, it was CN containing no OMe group that
catalyzed the formation of the dextrorotatory product in
higher ee. Interpretation of the phenomenon calls for further
research because, among other reasons, studies on other
cyclic β-ketoesters under identical conditions revealed
surprises of a different character.76a
Making use of the experiences of previous research87a,d
(ILs, cinchona derivatives), Salunkhe et al. reported that the
enantioselective Michael addition of dimethyl malonate to
1,3-diphenylprop-2-en-1-one (chalcon) promoted by a qua-
ternary ammonium salt derived from QN as a PTC in
different ILs, 1-butyl-3-methylimidazolium hexafluorophos-
phate, [bmim]PF6, 1-butyl-3-methylpyridinium tetrafluo-
roborate, [bpy]BF4, and 1-butyl-3-methylimidazolium tet-
rafluoroborate [bmim]BF4, as in conventional organic solvents,
was studied87c (Table 22). The reactions in ionic liquids
afforded excellent yields of the product in relatively short
periods of time, but interestingly and surprisingly, the ED
was reversed in the reactions in [bmim]BF4 and [bmim]PF6,
whereas it remained the same in [bpy]BF4, as was the case
for the conventional organic solvents under investigation.
In order to ascertain the factor responsible for the reversal
of ED, the results indicated that the reversal of ED was not
due to the PTC but can be attributed to the cation associated
with the anion of the IL. This unexpected phenomenon was
also noted by Hashimoto and Maruoka.77a The third Michael-
type addition catalyzed by a cinchona alkaloid is shown in
Table 23.
In the absence of any additives, high-yield diastereo- and
enantioselectivities were obtained as reported earlier (Table
23, entry 1).88a Similar stereoselectivities, but low yields,
were obtained when achiral additives were used (Table 23,
entries 2-6).88b Surprisingly, the addition of other achiral
additives to the precatalyst, that is, CN-Li complex,
provided low yields and caused a reversal of enantioselec-
tivity (entries 7-10). More importantly, data in Table 23
reveal that the two enantiomers (R,R) and (S,S) could be
synthesized by performing the reaction in two different
solvents, THF and ether, respectively. Although certain
preliminary experiments have been performed, the authors
indicate that detailed investigations to determine the exact
origin of the unexpected inversion are currently underway
in their laboratory. This unusual result suggested that the
stereochemistry at C8 and C9 in CN has no influence on
selectivity. In other words, the stereoinducing region of the
chiral ligand is away from the reaction site in this case. This
1682 Chemical Reviews, 2010, Vol. 110, No. 3 Bartók

report presents a complex system. The IC responsible for
enantioselection is a CN-phosphonate-Li complex. This
catalyst system appears to represent a transition between
organocatalysts and metal complex catalysts. Since the
publication reports several unexpected phenomena (solvent
effect, inversion, achiral additives, configuration of cinchona
alkaloids) in a Michael-type reaction, further results are
eagerly awaited.
Knowing and fully applying the antecedents from a report
published in 200889 disclosed a mild and efficient procedure
for Michael additions of cyclohexanone to chalcones (Table
24).90a In the presence of L-proline-achiral IL organocatalyst,
cyclohexanone reacted with various chalcones to afford
Michael adducts in high yields (80-99%) and moderate to
good ee (16-94%), accompanied by an unexpected solvent-
dependent inversion of the ED. The authors repeated the
literature data to determine the configurations of the Michael
adducts.90b
The authors assumed that the substrate is probably
anchored to the catalyst through a strong hydrogen bond
between catalyst and the amine group. Although attempts
to detail the reaction mechanism have not yet been under-
taken, the authors now propose a plausible transition state,
representing the stereoselective and solvent dependence of
Michael addition reactions of cyclohexanone with
chalcones.90a
3.4. Asymmetric Baylis-Hillman Reaction
The Baylis-Hillman (BH) reaction allows the direct pre-
paration of R-methylene-β-hydroxy carbonyl compounds
from the corresponding R,β-unsaturated ketones and aldehy-
des.91a,b In the special thematic issue of Chemical ReViews
on Organocatalysis,71 recently published reports on asym-
metric BH-type reactions are summarized in as many as two
subsections.81c,83a Table 25 reports on an asymmetric in-
tramolecular BH reaction, in which unexpected inversion
took place.92
The BH reaction of hept-2-enedial with L-proline was
examined in various solvents. Most reactions gave rise to
the corresponding (S)-6-hydroxy-cyclohex-1-enecarbal-
dehyde, albeit with variable yields and enantioselectivities.
Reactions in DMF and MeCN gave the highest yields,
enantioselectivities, and reaction rates.92 Surprisingly, the
authors found that, in the presence of imidazole, the
enantioselectivity of the reaction was completely reversed
(entries 1-4). This selectivity is highly sensitive to the nature
of the solvent and the temperature. In the case of D-proline,
the phenomenon is similar, but the reaction proceeds with
an ee of opposite direction (entries 5-8). This unprecedented
and striking inversion of selectivity is most likely due to the
formation of a new reactive intermediate, which also includes
imidazole.92 A short version of the reaction mechanism
proposed by the authors is shown in Figure 11.

Table 24. Inversion in the L-Proline Ionic Liquid Catalyzed Michael Reaction

entry ee (%) yield (%) Ar1 Ar2 yield (%) ee (%)

1 78* 95 Ph Ph 98 86
2 72* 98 4-MeC6H4 Ph 90 37
3 65* 94 4-ClC6H4 Ph 98 60
4 72* 88 Ph 4-ClC6H4 97 23
5 39* 90 4-BrC6H4 Ph 80 44
6 91* 85 Ph 4-NH2C6H4 87 94
7 16* 99 2-ClC6H4 4-MeC6H4 99 29

Table 25. Inversion in Intramolecular Baylis-Hillman Reaction

entry ee (%) conv. (%) L-pro/imid (equiv) solvent temp. (°C) L-proline (equiv) conv. (%) ee (%)
1 59* 71 0.1/0.1 MeCN r.t. 0.1 67 15
2 24* 79 0.1/0.1 DMF r.t. 0.1 73 45
3 80* 73 0.1/0.1 MeCN 0
4 93* 72 1/1 MeCN 0

entry ee (%) conv. (%) D-proline (equiv) solvent temp. (°C) D-pro/imid (equiv) conv. (%) ee (%)
5 41 75 0.1 DMF r.t.
6 MeCN 15 0.1/0.1 76 77*
7 MeCN 0 0.1/0.1 77 96*
8 MeCN 0 1/1 74 98*
Unexpected Inversions in Asymmetric Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1683

Figure 11. Proposed reaction pathways for the L-proline and the L-proline-imidazole cocatalyzed intramolecular Baylis-Hillman reactions.

3.5. Asymmetric β-Lactone Synthesis structure of the β-lactone formed in the presence of β-ICP
is (1S, 2R). This corresponded to the product obtained using
A new procedure for the synthesis of β-lactones has been AcOQN of configuration (8S,9R) rather than to that obtained
developed by Wynberg et al.93a,b This process made use of in the presence of the AcOQD having identical configuration
the nucleophilic properties of O-acetyl quinine and O-acetyl (8R,9S) with β-ICP. The authors conclude the description
quinidine to promote a [2 + 2]-cycloaddition between of this indeed surprising result by admitting that “we are
aldehydes and ketenes.93c,d Romo and co-workers disclosed not able to offer a satisfying explanation at this time”.94b To
the first examples of nucleophilic catalyzed aldol-lactoniza- our best knowledge, they have not published new results in
tion reactions with nonactivated aldehydes utilizing C9- this field ever since. It appears plausible that, in unexpected
acylated cinchona alkaloids as nucleophilic catalysts.94 inversion, the OH group of β-ICP may play a role in the
A further case of unexpected inversion was recognized in formation of the IC responsible for ED.
the course of studies on the stereochemistry of the reaction
(Table 26).94b Namely, the presence of β-ICP (β-isocuprei-
dine), a catalyst of rigid structure, was found to bring about
3.6. Asymmetric β-Lactams Synthesis
complete reversal in the sense of ED. The result was β-Lactams, compounds with structures similar to that of
surprising since, according to earlier concepts, the chirality β-lactones, can be prepared by a synthesis similar to that
of a reaction product is determined by the configurations of developed by Staudinger. Owing to the significance of the
atoms C8 and C9 of the cinchona alkaloid present. The 3D type of compound involved, the method has been widely
used.93c,d,95 The planar chiral catalyst (PPY derivative Cata-
Table 26. Inversion in Intramolecular Asymmetric lyst*) applied in the interesting reversal in diastereoselectivity
Aldol-Lactonization recognized in the Staudinger reaction is not purely an
organocatalyst according to the accepted definition, because
it also contains an inorganic atom.96 Nonetheless, inversion
is discussed in this section for two reasons. First, it is closely
related to the previous subsection, and second, this reaction
ee conv. conv. ee is probably discussed in ref 71 for a similar reason. The
entry (%) (%) Catalyst* solvent Catalyst* (%) (%) reaction exhibiting unexpected stereochemistry is shown in
1 92 54 AcOQD MeCN AcOQN 51 86 Scheme 12.96b
2 92 54 AcOQD MeCN β-ICP 42 90*
3 92 21 AcOQD DCM β-ICP 18 90*
It is well-known that catalytic asymmetric syntheses of
β-lactams are generally cis-selective.96a,97 According to
Scheme 12, cis-selectivity is reversed merely by the exchange
of the protecting group for the imine from Ts to Tf. In the
case of trans compounds, ee values as high as 89% were
achieved, depending on the substituents. The probable
mechanism of the process is described.93d,96b The key step
of the reaction mechanism depends on whether the Catalyst*
reacts with the ketone or with the imine first. Since the
Catalyst* reacts quantitatively with an N-triflyl imine, in
1684 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 12. Inversion in the Staudinger Reaction to Form
β-Lactam Enantiomers
this case an “imine-first” pathway is operative, whereas in
the case of N-tosyl imine, the reaction takes the “ketone-
first” pathway (Figure 12).
3.7. Asymmetric Aza-Henry Reaction
A series of chiral guanidines and bisguanidines were
synthesized, and their effects on the catalyzed aza-Henry
reaction were studied (Scheme 13). Lovick and Michael
observed an unexpected inversion elicited by organocatalysts
in the course of the aza-Henry reaction.98
Results with respect to the use of the two (R)-(S)
enantiomer pairs achieving the highest ee values of opposite
senses are outlined. In the case of monomeric catalysts (M1,
M2), the (R)-products are formed in higher ee, whereas in
the presence of dimeric catalysts (D1, D2), (S)-products are
formed in higher ee. The mention of very recently published,
unexpected preliminary data in this review is considered
justifiable. A plausible model for the interpretation of
unexpected inversion by authors is presented in Figure 13.
Figure 12. “Ketene-first” and an “imine-first” pathways for
nucleophile-catalyzed Staudinger reactions.
Bartók
Scheme 13. Inversion in Enantioselective Aza-Henry
Reaction
Figure 13. Assumed intermediate complex in the aza-Henry
reaction catalyzed by monoguanidine-type organocatalyst.
3.8. Summary
Research on organocatalytic reactions, as compared to that
of metal complex catalyzed processes, is still in the stage of
data collection, in my opinion not only in the field discussed
here. This is well-demonstrated by the number of examples
listed in Table 27.
In Table 27, the examples collected in the literature are
grouped according to reactions, with a few parameters
indicated similarly to Table 15. The largest number of
reactions exhibiting unexpected inversion of the sense of ED
were found among aldol additions and Michael reactions.
One example was identified in each of the Baylis-Hillman
reaction, syntheses of β-lactone and β-lactam, and the aza-
Henry reaction. The chiral organocatalysts used in these
reactions can be classified into 3 groups: (i) base catalysts:
alkaloids (mainly cinchona alkaloids), L-proline, and its
derivatives; (ii) the salts of the former used as phase-transfer
catalysts; (iii) other chiral organocatalysts (formamide- and
phosphoramide-substituted, ferrocenyl-type (PPY derivative),
guanidines). The largest number of examples for unexpected
inversion was found among organocatalytic reactions utiliz-
ing L-proline and its derivatives, as well as cinchona
alkaloids.
As regards the ee values attained, it is remarkable that it
is possible to obtain both enantiomers in outstandingly high
ee on the chiral catalyst of identical configuration after
modifying some other parameter (Table 27, entries 6, 7, 9,
and 12). The effects presumed to be responsible for the
inversions are listed in column 5 of Table 27. The majority
of these effects are minor differences in organocatalyst
structure, achiral additives, and solvents used. The data in
column 6 refer to the interpretation of unexpected inversions.
This column, however, also illustrates the current state of
organocatalytic research, because in this field only one
experimentally verified interpretation could be found, a fact
calling attention to further tasks.
The proposals of the authors regarding the interpretation
of the reactions are included with the description of the
experimental data. The authors usually plan to experimentally
verify the observed unexpected phenomena in future experi-
ments. On the basis of the outstanding results achieved in
the field of organocatalysis in the last 10 years, it feels
Unexpected Inversions in Asymmetric Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1685

Table 27. Inversions in Organocatalytic Reactions

entry reactions Catalysts* Rmax/Smax effects interpret-ationa Table/Scheme in manuscript ref
1 alkylation BnCD+Br- 40/66 KOH, NaOH, temperature A, B Table 16 78a
2 allylation formamide 68/30 [Catalyst*] HMPA A, B Table 17 79b,c
3 alkylation (-) sparteine 60/46 substrate A, B Scheme 11 80
4 aldolization L-proline 68/21 heterogeniz. A Table 38 83c
5 aldolization L-proline, L-prolinol -26/+54 catalyst* structure Table 18 76b
6 aldolization L-proline amides 99/91 additive: 4 Å MS A, B Table 19 85a
7 aldolization phosphoramides 90/92 solvent Table 20 86
8 Michael r. BnQN+ Br- 50/61 solvent A, B Table 22 87c
9 Michael r. CN 100/100 achiral addit. Solvent A, B Table 23 88a,b
10 Michael r. L-proline- IL+ -78/+86 solvent A, B Table 24 90a
11 BH r. L-proline 93/45 additive: imidazoline A, B Table 25 92
12 β-lactone synthesis AcOQD, β-ICP 92/90 C8*, C9* configuration B Table 26 94b
13 β-lactam synthesis ferrocenyl-PPY 63/98 protecting Ts or Tf D Scheme 12 96b
14 aza-Henry r. guanidines 30/77 monomers, dimers A, B Scheme 13 98
a
A ) assumed, B ) planned, and D ) verified.

justified to expect the discovery of many more unexpected
phenomena and significant progress in their research.
4. Unexpected Inversions in Heterogeneous
Catalytic Asymmetric Reactions
The attention of researchers aware of the advantages of
heterogeneous catalysis was aroused by the research of
heterogeneous catalyzed asymmetric syntheses. The results
achieved have been continuously summarized and evaluated
by reviews and monographs.99,6b In the “Handbuch of
Asymmetric Heterogeneous Catalysis”, published very re-
cently, the results of studies on different variations of
asymmetric syntheses using heterogenized/immobilized chiral
catalysts of a great variety of types are reviewed in about
200 subsections.99j
The new monograph reveals that the present objective of
research on asymmetric heterogeneous catalysis is the
development of active and reusable catalysts with stable
structures. Simple test reactions are used, and studies on the
reactions mechanism are not in the forefront of interest.
Consequently, unexpected inversions are mostly observed
in hydrogenations and only a few examples are found in the
literature of other asymmetric syntheses, which naturally sets
the direction of future tasks.
4.1. Asymmetric Hydrogenation
It is not surprisingsfollowing the discovery made by
Sabatier and Senderens100asthat research addressing the
preparation and investigation of the hydrogenation over
heterogeneous chiral catalysts was started as early as the first
Scheme 14. Enantioselective Hydrogenation of β-Ketoesters over TA-MNi Catalyst
one-third of the 20th century.100b,c It is well-known that the
discovery of hydrogenations on metals with large specific
surface areas significantly boosted the development of
organic chemistry.
Chiral modifiers (M*) used for the preparation of hetero-
geneous chiral catalysts were natural materials available such
as hydroxycarboxylic acids, amino acids, chiral bases, and
their easily synthesizable derivatives. Research led to the
recognition of “modified catalysts”.101 Further multifaceted
efforts resulted in the development of two modified chiral
hydrogenation catalytic systems, namely, Ni catalysts modi-
fied by (R,R)-tartaric acid (TA-MNi)102 and Pt catalysts
modified by cinchona alkaloids (Pt-cinchona alkaloid; Orito
reaction).103 Similar catalyst systems employing other metals
were also developed at the later stages of research.
It was found in the course of the experiments that, similarly
to enzyme-catalyzed reactions, the two catalyst systems
cannot be applied to enantioselective hydrogenations in
general; in other words, they enable the attainment of high
ee only in the hydrogenation of certain types of compounds.
High ee values were initially achieved mainly in the
hydrogenation of β-ketoesters using the TA-MNi catalyst
system in the presence of NaBr (Scheme 14, Figure 14) and
in that of R-ketoesters using the Pt-cinchona catalyst system
(Scheme 15, Figure 15). The utilization of these two catalyst
systems in various hydrogenations are the most intensively
studied enantioselective heterogeneous catalytic reactions,
which are also exploited on an industrial scale.99b,104,105
The main objective of recent studies on these two reactions
was to expand their field of utilization, to elucidate the
reaction mechanism, and to interpret the origin of ED. The

Figure 14. Chiral modifiers in preparation of modified Ni catalysts.

1686 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 15. Stereochemistry of the Orito Reaction
significance of these reactions is underlined not only by the
high enantioselectivities (above 90%) observed (in the case
of TA-MNi106,107 and in the case of the Orito reaction108–110)
but also by numerous reviews discussing and evaluating the
steady flow of novel results in the heterogeneous catalytic
enantioselective hydrogenations.6,99f–i,111–115
As regards the stereochemistry of the processes outlined
in Schemes 14 and 15, it was recognized already at the time
of the discovery of the reactions that the catalyst modified
by (R,R)-TA promotes the formation of an excess of the (R)-
product, whereas the one modified by (S,S)-TA promotes
the formation of the (S)-product in excess (Scheme 14). As
shown in Scheme 15, the presence of C8(S),C9(R) cinchonas
(CD,QN) promotes the formation of (R)-R-hydroxy carboxy-
lic acid esters in excess, whereas C8(R),C9(S) cinchonas
(CN,QD) induce the formation of (S)-R-hydroxy carboxylic
acid esters. The term “inversion of enantioselectivity” in the
title of the manuscript implies the formation of products with
opposite configurations from Schemes 14 and 15.
Unexpected inversion of enantioselectivity in the hetero-
geneous catalyzed enantioselective hydrogenations reaction
had been reported,106a,111b,116–121 butsdue to the low ee values
involvedsthese results aroused little attention. Since the
publication of refs 122and 123, however, the inversion of
ED has become a preferred research objective, because it
yields important new information regarding the reaction
mechanism.
4.1.1. Hydrogenation of Ketones over Ni Catalysts
Scheme 14 demonstrates the experimentally verified basic
scheme of the stereochemical course of enantioselective
hydrogenation on Ni catalyst modified by TA on the most
often studied model substrate, methyl acetoacetate. According
to Scheme 14, hydrogenation of methyl acetoacetate produces
an excess of (R)-methyl hydroxybutyrate on (R,R)-TA-Ni
catalyst and an excess of (S)-methyl hydroxybutyrate on
(S,S)-TA-Ni catalyst.111b,113d
It was soon recognized that a careful observance of the
conditions of catalyst preparation is essential for achieving
high ee values. For example, according to observations made
at the initial stages of these studies, the enantioselectivity of
Figure 15. Proposed enantiodifferentiation models in hydrogena-
tion of ketones over (R,R)-TA-MNi catalyst.
Scheme 16. Inversion in Enantioselective Hydrogenation of Methyl Acetoacetate
Bartók
Scheme 17. Inversion in Enantioselective Hydrogenation of
Ketones
Ni catalysts modified by various amino acids was profoundly
influenced by the temperature and the pH during the
preparation of the chiral catalyst111b (Scheme 16).
During the decades elapsed since the discovery of the
reaction, research on the reaction mechanism has mostly been
focused on catalyst modified by (R,R)-TA. Whether or not
hydrogenation governed by a stereochemistry different from
that shown in Scheme 14, i.e., one that yields an excess of
the (S)-product on a catalyst modified by (R,R)-TA, was
observed depended on the structure of the ketone to be
hydrogenated (Scheme 17).
After the principles governing the direction of ee formation
(namely, that on Ni catalysts modified by (R,R)-TA hydro-
genation of R-, β-, and γ-ketocarboxylic acid esters yields
(R)-hydroxycarboxylic acids in excess, whereas hydrogena-
tion of δ- and ε-ketocarboxylic acids esters, as well as of
alkanones, produces the corresponding (S)-compounds in
excess) became widely known,113d,f,115a to our best knowledge
unexpected inversion has not been mentioned in the literature.
From the multiple variants of models interpreting the
stereochemistry of these processes, two models, namely, the
Two Hydrogen Bonds stereochemical model (2P model) and
the One Hydrogen Bond and a Steric Repulsion model (1P
model) are schematically represented in Figure 15.
On the basis of reflection absorption infrared spectroscopy
(RAIRS) and scanning tunneling microscopy (STM) studies,
Baddeley et al. proposed that the altered ee can be attributed
to the presence of the diketo and enol tautomers of methyl
acetoacetate.124
4.1.2. Hydrogenation of Activated Ketones over Pt
Catalysts
The chiral molecules utilized for the modification of chiral
Pt catalysts are summarized in Figure 16.
Hydrogenation of Pyruvates. EtPy is the most commonly
used model compound of studies on the Orito reaction; it is,
therefore, not surprising that the discovery of unexpected
inversion is also associated with EtPy. In 1993, Augustine
et al. were the first to report unexpected inversion under the
Unexpected Inversions in Asymmetric Reactions
Figure 16. Cinchona alkaloids and derivatives used in this section.
conditions of the Orito reaction118 (Table 28, entry 2),
namely, the formation of (S)-EtLt was observed when very
low concentrations of the alkaloid modifier were used,
whereas at higher modifier concentrations, (R)-EtLt was
produced. The formation of (R)-EtLt was accompanied by
an increase in hydrogenation rate. The formation of (S)-EtLt
is attributed mainly to the corner atoms, whereas the adatoms
are considered to be responsible for (R)-EtLt formation. It
is unfortunate that the measurements were not performed in
AcOH at systematically varied CD concentrations, where
high ee can be achieved. Thus, even though the surface active
sites were adequately characterized, the optimal reaction
conditions required for high ee were not provided. That is
why the role of active sites of different types in ED could
not be unequivocally verified.
In 2002, a remarkable observation made on the hydroge-
nation of EtPy, the most commonly studied model substrate
of the Orito reaction, was reported in ref 123, methyl
acetoacetate. According to ref 123, hydrogenation carried
out in toluene, in the presence of β-ICN, an ether derivative
with C8C9 configuration identical with CN, yielded (R)-EtLt
in 48% ee, even though according to the relationships
accepted at the time the product formed in excess should
have been (S)-EtLt (entry 6). A significant solvent effect was
observed in the course of the studies on the chiral catalyst
β-ICN-Pt: in AcOH, unlike in toluene, the expected (S)-EtLt
was formed in excess (Figure 17). This was the first
significant experimental observation indicating that, in enan-
tioselective hydrogenation initiated by cinchona alkaloids,
it is not solely the C8 chiral center of the alkaloid that
controls the sense of chiral induction.
On the basis of the significant solvent effect, the inversion
was explained by a change in the reaction mechanism.123,126
It is therefore expedient, for the interpretation of inversion,
to outline the existing views on the mechanism of the
Chemical Reviews, 2010, Vol. 110, No. 3 1687
enantioselective hydrogenation of activated ketones, char-
acterized by the structure of the IC responsible for ED
(Figure 18).6,114
The quinuclidine nitrogen of β-ICN acts either as a
nucleophile (C or D type in Figure 18) or as an electrophile
(upon protonation) (A, B, or F type) to interact with the
R-carbonyl group of the EtPy. Consequently, the structure
of the IC responsible for chiral induction depends on the
solvent applied (AcOH, toluene). The proposed structures
of 1:1 β-ICN-EtPy IC-s in AcOH and in toluene are shown
in Figure 19. In AcOH, β-ICN participates in the formation
of the 1:1 complex as a protonated electrophile (Figure 19A),
whereas in toluene it binds EtPy as a nucleophile (Figure
19B). To interpret the enantioselective hydrogenation in the
Pt/β-ICN chiral catalyst, the role of other organometallic type
surface complex (E) may not be ruled out either.134a,b
The ability of β-ICN to cause inversion was also demon-
strated over Rh/alumina catalyst in the hydrogenation of not
only EtPy but also ethyl 3-methyl-2-oxobutyrate127 (entries
7 and 19-22). According to the authors: (i) the formation
of the opposite enantiomer in small excess in protic solvents
is attributed to the formation of solvent-substrate and
solvent-modifier complexes that disturb the enantioselection
on cinchona-modified Rh; (ii) the adsorption modes of β-ICN
and CN during enantioselective hydrogenation on Rh are
considerably different. The latter can be agreed with, because
with adsorption being one of the steps of the mechanism, a
change in adsorption mode may entail a change in reaction
mechanism.
As shown in Table 28, the majority of these studies
revealed unexpected inversion upon varying the concentration
or structure of cinchonas (C9-OR cinchonas)127–132 (entries
8-18). The authors of ref 128 emphasize the change in
adsorption mode of the chiral modifier in their interpretation
of unexpected inversion; in our opinion, such a change may
1688 Chemical Reviews, 2010, Vol. 110, No. 3
Table 28. Inversions in Hydrogenation of Pyruvates
Figure 17. Hydrogenation of EtPy to (R)- and (S)-EtLt on β-ICN
modified Pt/alumina in toluene and AcOH mixtures (room tem-
perature (r.t.), 1 bar H2).
be associated with changes in certain steps of the reaction
mechanism. Dependence of inversion on the concentration
of the chiral modifier is reported in ref 129. This was the
first example where enantio-inversion was induced solely as
a function of the chiral modifier concentration (entries 9 and
10).
In the study of ether derivatives of CD,121,128,129,131,132 the
crucial role of steric bulkiness of the ether groups in the
inversion of the sense of enantioselection was shown as
Bartók
Figure 18. Assumed intermediate complexes in the enantioselec-
tive hydrogenation of activated ketones (X ) activating group, Ql
) quinolinyl).
Figure 19. Proposed structures of intermediate complexes in
electrophilic (A) and nucleophilic-type (B) interactions betveen EtPy
and β-ICN.
the bulky ether group occupies the chiral space available for
the adsorption of the substrate in case of the CD. Demonstra-
tive results of studies on CD, PhOCD, and 2-PyOCD under
identical experimental conditions are presented in Figure
20.132 The comparison of PhOCD and 2-PyOCD modifiers
provides an even more intriguing example as their van der
Waals volumes and adsorption modes on Pt/Al2O3 are very
Unexpected Inversions in Asymmetric Reactions
Figure 20. Hydrogenation of EtPy in THF over CD-Pt, 2-PyOCD-
Pt, and PhOCD-Pt chiral catalysts.
similar, but an additional substrate-modifier interaction
possible only for 2-PyOCD inverts the ee.
Inversion has also been observed under conditions other
than those of the Orito reaction (entries 11 and 12).130 It has
been shown that enantioselective hydrogenation of pyruvates
can be carried out using gas-phase reactants,130 thereby
avoiding the complicating factor of solvent effects. At low
ClBzHQD modifier concentration, the (S)-lactate is the
preferred product, and at higher concentrations, (R)-lactate
is favored. The inversion effect is discussed by the authors
in terms of the interaction of the substrate and modifier with
the catalyst surface. The modifier interactions with different
sites on the heterogeneous catalyst lead to different senses
of enantioselection.
In the continuous fixed-bed reactor (CFBR), unexpected
inversion was observed in the hydrogenation of EtPy over
CN-Pt catalyst after removal of the soluble fraction of the
modifier in 50 °C, that is, in the presence of chemisorbed
CN (entry 18).133 On the authors opinion these modifications
can be summarized as (i) alteration of platinum crystallites
inducing some ee retention and (ii) some sort of support
effect inducing a reversal in ee.133
Table 29. Inversions in Hydrogenation of Esters of Phenylglyoxylic Acids
Chemical Reviews, 2010, Vol. 110, No. 3 1689
Figure 21. Assumed structures of intermediate complexes in
electrophilic (A) and nucleophilic-type (B) interactions betveen
phenylglyoxylic acid esters and modifiers.
Hydrogenation of Phenylglyoxylic Acid Esters. The
enantioselective hydrogenation of phenylglyoxylic acid esters
was significantly slower than that of EtPy. The ee values in
unexpected inversion in hydrogenations in toluene over
β-ICN-Pt,134c but especially over PhOCD-Pt chiral catalysts,
are remarkably high (Table 29). The highest ee so far
reported in unexpected inversion is 78% (S) (entry 12).
The experimental data of conversions and ee suggest that
both electronic and steric factors may play a role in
determining the rate of enantioselective hydrogenation and
ee. In the case of compounds containing methyl and
cycloalkyl groups (entries 1-4), steric factors predominate,
whereas in the case of aromatic groups, mainly electronic
factors that may naturally affect hydrogenation rate and ee
predominate (entries 5-7).
On the basis of investigations in the hydrogenation of the
bulkier R-ketoesters, steric effects seem to play a more
important role in the inversion of ED than in hydrogenation
without inversion. For a CD-Pt chiral catalyst in AcOH, the
IC is generated through the interaction of the protonated CD
(Figure 21A). The inversion of bulky phenylglyoxylic acid
esters on β-ICN-Pt chiral catalyst, like that of EtPy, was
interpreted on the basis of the nucleophilic mechanism. The
surface complex responsible for the enantioselection is
probably formed by the interaction between the nucleophilic
N atom of the quinuclidine skeleton and the electrophilic C
atom of the keto group of the substrate134c (Figure 21B). In
the case of PhOCD, the inversion is probably due to the steric
bulkiness of the phenoxy group relative to that of the OH
function and also to a change in the adsorption geometry on
the alkaloid, resulting in a shift in the position of the
interacting function, the quinuclidine N atom.136
Hydrogenation of Ketopantolactone. The first experi-
mental observation was made by Baiker’s group in 2003.137
1690 Chemical Reviews, 2010, Vol. 110, No. 3 Bartók

The experimental data revealed that, in the presence of the
chiral modifiers CD, MeOCD, and EtOCD, the product with
the expected configuration, i.e., (R)-PL, was formed in higher
ee, whereas in the presence of bulky C9-ethers (PhOCD,
Me3SiOCD) as chiral modifiers, the formation of (S)-PL was
induced in excess (Table 30, entries 1, 2, 9, and 10).
In both cases, the configurations of the C8,C9 carbon
atoms of CD and its derivatives, responsible for chiral
induction, were identical, which means that the inversion of
enantioselectivity was brought about by the bulky C9-ethers.
As no experimental evidence was available at the time for
the interpretation of the phenomenon, it was concluded that
introduction of the bulky trimethylsilyl or phenyl substituents
changes dramatically the chiral pocket available for the
adsorption of KPL over the Pt surface and leads to the
favored adsorption of KPL on the opposite enantioface.
It was assumed that Me3SiOCD and PhOCD do not adsorb
via the quinoline ring, being approximately parallel to the
Pt surface (π-bonding) but rather in a tilted position (N-lone
pair bonding).137 This change in the adsorption geometry
should result in a considerably weaker adsorption of these
modifiers compared to the adsorption of CD. In this tilted
position, the modifier adsorbs less strongly via the quinoline
N, and also the position of the interacting function, the
quinuclidine N, is shifted. This shift results in a different
shape and size of the “chiral pocket” available for adsorption
of the activated ketone substrate.128,141
The performance of a new modifier, 2-PyOCD, is com-
pared to that of PhOCD and CD.132 In the hydrogenation of
KPL, the bulky O-phenyl group favors the (S)-enantiomer,
whereas in the case of the 2-pyridyl group, the (R)-alcohol
is the major product (entry 11). Various catalytic studies,
ATR-IR spectroscopy using conditions of Orito reaction, and
theoretical calculations of the modifier-substrate interactions
suggest that formation of two N-H-O-type H bondss
involving the quinuclidine and pyridine N atoms, and the
two keto-carbonyls in the substratescontrols the adsorption
of the substrate during hydrogen uptake.132
We have also studied the enantioselective hydrogenation
of KPL in toluene on β-ICN-Pt chiral catalyst.139 Enanti-
oselective hydrogenation yielded an excess of (R)-PL, i.e.,
inversion of enantioselection took place, since the (R)-
configuration is opposite to what is expected from the
absolute configuration of the CN backbone (entries 4-6).
Scheme 18. Stereochemistry of Enantioselective
Hydrogenation of Ketopantolactone
Scheme 19. Nucleophilic Mechanism in Hydrogenation of
Ketopantolactone over β-ICN-Pt Chiral Catalyst
Both the reactant and the chiral modifier used in these studies
were rigid molecules of well-known structures. Thus, these
studies have yielded new information, contributing to a
deeper understanding of the enantioselective hydrogenation
of activated ketones, because structural rigidity prevents
conformational movements. The stereochemistry of the
enantioselective hydrogenation of KPL and its dependence
on the solvent as well as the presence of chiral modifiers
CD, CN, R-ICN, or β-ICN are summarized in Scheme 18.
In this case, we proposed again the nucleophilic mecha-
nism for the interpretation of inversion (Scheme 19), in view
of the fact that in toluene there was inversion, whereas in
AcOH there was no inversion (entry 4).
On the basis of our experimental data and on the verified
open-3 conformation of β-ICN142,143 as well as on the widely
accepted adsorption model,6,113b,c,114,115c,144 the proposed
structure of the ICs responsible for the enantioselectivity is
outlined in Figure 22. The formation of such nucleophilic
complexes has been verified by NMR measurements in the
liquid phase.145a There is a correlation between the solution-
state concentration of the nucleophilic 1:1 modifier-substrate
complex and the ee on enantioselective hydrogenation of
KPL using β-ICN-Pt chiral catalyst.145b These results confirm
the earlier suggestion regarding the direction of ED:145c the
sense of ED is controlled by the conformation of the adsorbed

Table 30. Inversion in Hydrogenation of Ketopantolactone

entry ee (%) M* solvent H2 (bar) M* solvent ee (%) ref

1 40-80 CD, MeOCD toluene 40 PhOCD toluene 55* 137
EtOCD THF Me3SiOCD THF 35*
2 40-73 CD, MeOCD PhCF3 40 ArOCD PhCF3 37*-53* 128
EtOCD Ar ) Ph, Xyl., HFXyl., β-naphthyl
a
3 25-40 CD aprotic solvents 10 PhOCD aprotic solvents 15*-25* 138
4 46* β-ICN toluene 1 β-ICN AcOH 5 139
5 46* β-ICN toluene 1 CN toluene, AcOH 52 139
6 1 CN 55
7a 54* β-ICN toluene 10 CN toluene, AcOH 25 127
8a 11* β-ICN AcOH 10 CN AcOH 1(R) 127
9 16 HFXylOCD THF 1 PhOCD THF 21* 140
10 1 XylOCD THF 36* 140
11 50, 31 CD, 2-PyOCD THF 10 PhOCD THF 21* 132
a
Rh/alumina.
Unexpected Inversions in Asymmetric Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1691

Table 32. Inversions in Hydrogenation of β-Diketones

ee ee
entry (%) M* solvent R M* solvent (%) ref
1 31* HFXylOCD toluene Me MeOCD toluene 68 128
2 13* MeOCD DMF t-Bu MeOCD AcOH 15 147a
3 8* CD DMF t-Bu CD THF 36 147a
4 22* MeOCD AcOH Ph MeOCD toluene 2 147a

Scheme 20. 1-Phenyl-1,2-propanedione Hydrogenation

Figure 22. Proposed structure of pro (R)-intermediate complex

between KPL and β-ICN (black spheres ) N atoms, pink spheres
) O atoms, yellow spheres ) C atoms, white spheres ) H
atoms).

reactant-chiral modifier (1:1) complex. It was repeatedly

verified that solvents may have a profound effect on reaction
mechanism.
The ability of β-ICN to cause inversion was also confirmed
in the hydrogenation of KPL on Rh/alumina catalyst (entries among those outlined in Scheme 20 (k1, k2) are shown in
7 and 8).127 Comparison of the experimental data of KPL Table 31.146 The data in Table 31 indicate the determinant
hydrogenation in toluene and AcOH, over chiral catalysts role of modifiers in inversion. In this respect, the effect of
CN-Pt, β-ICN-Pt, CN-Rh, and β-ICN-Rh, reveals that, C9-OR modifiers on the sense of ED is especially remark-
although the experimental conditions are somewhat different, able. The most extensive inversion was brought about by
studying the origin of chiral induction still demands intensive cinchonine ethers during hydrogenation in toluene (entries
experimentation in this field. 8-11). A characteristic example is the case of MeOCN and
Hydrogenation of r- and β-Diketones. The most often CN: under identical experimental conditions, unexpected (R)-
studied model compound of the Orito reaction of R-diketones product was formed in 32% ee in the hydrogenation of the
is 1-phenylpropane-1,2-dione (PPD), whereas those of C1 ketone group of PPD, whereas the expected (S)-product
β-diketones are 1,1,1-trifluoro-2,4-diketones. Experimental was generated in 18% ee on CN-Pt chiral catalyst (entry
results published on inversion observed in the enantioselec- 8).146c This is an important observation concerning the nature
tive hydrogenation of R-diketones and β-diketones are of the chiral site on the Pt surface. The observed inversion
summarized in Tables 31 and 32, respectively. Because of of ED induced by the cinchona alkaloid ether modifiers in
the presence of the two carbonyl groups, the hydrogenation toluene indicates that the ED steps over the Pt-cinchona
process is the result of several competitive and consecutive alkaloid and Pt-cinchona alkaloid ether chiral systems, that
reactions, which are illustrated for PPD hydrogenation in is, different types of substrate-modifier complexes are
Scheme 20. The experimental data reported in ref 146 shed involved.146c
light on numerous relationships governing the individual To the best of our knowledge, the activated β-diketones
reactions of the hydrogenation of R-diketones. that have been subjected to enantioselective studies are the
The experimental data of the unexpected inversion of the 3 compounds shown in Table 32.128,147 In cases where
primary reactions proceeding with the highest selectivity identical modifiers were used, the experimental data suggest
Table 31. Inversion in Hydrogenation of
a solvent effect (entries 2-4); in hydrogenations carried out
1-Phenylpropane-1,2-dione under identical experimental conditions, the data confirm the
determinant role of chiral modifiers.
In the interpretation of inversion, the authors emphasize
the change in the adsorption mode of the chiral modifier,
which has been discussed in detail for KPL. This assumption
entry ee (%) M* M* ee (%) ref is in agreement with the results of studies on the hydrogena-
1 57 CD MeOCD 2* 146a,b tion mechanism of trifluoromethyl-β-diketones, which re-
2a 7 CD MeOCD 2* 146a vealed by a combination of catalytic, NMR, and FTIR
3 16* MeOCN CN 27 146b
4 35 QN MeOQN 3* 146b spectroscopic and theoretical methods that the two phenom-
5 60 CD MeOCD 10* 146c ena are coupled, offering the unique possibility for under-
6 60 CD PhOCD 27* 146c standing the substrate-modifier-metal interactions. The high
7a 11 CD PhOCD 5* 146c chemo- and enantioselectivities are attributed to the formation
8 32* MeOCN CN 18 146c
9 30* PhOCN CN 18 146c
of an ion pair involving the protonated amine function of
10 28* Me3SiOCN CN 18 146c the chiral modifier and the enolate form of the substrate.147b
11 28* t-BuMe3SiOCD CN 18 146c Hydrogenation of r-Hydroxyketones. Baiker et al.
performed multifaceted investigations on the enantioselective
a
Solvent ) AcOH.
hydrogenation of R-hydroxyketones and one of their ether
1692 Chemical Reviews, 2010, Vol. 110, No. 3
derivatives (2-methoxyacetophenone).132,148 The experimental
data on inversion that have allowed many important conclu-
sions to be drawn are summarized in Table 33. As a result
of the very first experiments, it was established148a that CD
showed by far the best catalytic performance affording ee’s,
between 57 and 82% depending on the substrate. MeOCD,
in turn, showed poor ED. PhOCD favored the opposite
enantiomer compared to CD (entries 1 and 2). Among
solvents, t-butyl methyl ether proved to be the most suitable.
All four reactants show the same general trend with respect
to dependence on solvent, hydrogen pressure, and modifier.
The phenyl ring or a fixed system is essential for ED.
Furthermore, the oxygen in R-position to the ketone plays a
crucial role for achieving high ee.
Changing the modifier from CD to PhOCD resulted in a
switch of the major enantiomer of the product on Rh/alumina
catalyst as well. Hydrogenation of 2-hydroxyacetophenone
showed a switch from 73% ee in favor of the (R)-product to
68% ee for the (S)-product when the modifier was changed
from CD to PhOCD (entry 3).148b
The inversion of the ee is interpreted in terms of repulsive
modifier-reactant interactions, which become more pro-
nounced as the steric demand of the C9-OR group of the
modifier increases. The obvious importance of an oxygen-
containing group (ketone, hydroxyl, methoxy) in R-position
to the ketone that hydrogenated is rather assigned to a
lowering of the transition state energy for hydrogenation due
Table 33. Inversions in Hydrogenation of r-Hydroxyketones
b
Rh/alumina, 15 °C.
Bartók
to hydrogen bonding, as previous calculations suggested.148a
One of the latest reports on the inversion of the sense of ED
presents new experimental data on the enantioselective
hydrogenation of 2-methoxyacetophenone (entries 9-13) and
on the effect of the new chiral catalyst 2-PyOCD-Pt.132 The
main conclusions of these studies have been discussed in
the subsections on EtPy and KPL.
Hydrogenation of r-Trifluoromethylketones. This sub-
section summarizes inversions observed in the course of the
enantioselective hydrogenation of trifluoromethylalkyl, cy-
cloalkyl, and aryl ketones (Table 34). Ethyl trifluoroacetoac-
etate is hydrogenated over MeOCD-Pt catalyst, in AcOH in
90 (S) % ee (entry 1).149 The authors later observedsunder
different experimental conditionssa so far unknown inver-
sion at later stages of conversion (entry 2).122 Figure 23 shows
that the ee value decreased gradually after an initial constant
period (87% (S)) and dropped to 27% (S) at full conversion.
The sense of ED was inverted, and the (R)-product became
dominant. According to the data obtained by NMR, inversion
is not due to hydrogen addition at the carbonyl group but to
hydrogenolysis of the C-O bond in the geminal diol
(Scheme 21).
Inversion of ED sense has also been observed with the
application of CD-OAr chiral modifiers (entries 3-5).128
Inversion was interpreted on the basis of a change in the
adsorption mode of the chiral modifier, similarly to the case
of EtPy and KPL hydrogenation.128 Hydrogenation of trif-
Unexpected Inversions in Asymmetric Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1693

luoromethylalkyl and trifluoromethylcycloalkyl ketones over

CD-Pt chiral catalyst proceeds in moderate ee in weakly polar
solvents with and without TFA. In alcohols (EtOH, i-PrOH),
however, inversion takes place (entries 6-8).150,151a The
probable reason for this inversion is additional H bonding
with the solvent.
In the course of the enantioselective hydrogenation of
2,2,2-trifluoroacetophenone (TFAP) and its p-trifluoromethyl
analogue on the chiral catalyst MeOCD-Pt, a low extent of
inversion was observed in 2001: instead of the expected (R)-
product, (S)-alcohols were formed (entries 9 and 10)151b
(Figure 23).
According to our new experimental data, the hydrogena-
tion of TFAP in the presence of the chiral modifiers CN,
QN, QD, and their C9-ether derivativessunlike CDsproduces
low ee values, accompanied, in the majority of cases, by
unexpected inversion152 (entries 11-14). The inversion
cannot be interpreted on the basis of the IC containing the
Figure 23. ee versus yield during enantioselective hydrogenation
protonated quinidine skeleton, because there is no inversion of ethyl trifluoroacetoacetate on a CD-Pt chiral catalyst.
in the presence of TFA. In our opinion, however, the role of
Table 34. Inversions in Hydrogenation of r-Trifluoromethylketones

a
∆ee ) (ee1Y1 - ee2Y2)/(Y2 - Y1), Y ) yield, for CN 0 °C, 10 bar.
1694 Chemical Reviews, 2010, Vol. 110, No. 3 Bartók

Scheme 21. Hydrogenation of Ethyl 4,4,4-Trifluoroacetoacetate in the Presence of Water

the hydrogen-bridged IC binding to the quinuclidine
skeletonswhich represents an interaction weaker than the
one with the protonated onesand also that of the nucleophilic
complex proposed several times before (Figure 24) cannot
be excluded either.145
Hydrogenation of Methyl Aryl Ketones. In view of the
fact that, in the studies to date, the Orito reaction has been
shown to be applicable only in the hydrogenation of activated
ketones, it appears that only the hydrogenation of methyl
aryl ketones with electron-withdrawing groups attached to
the phenyl group could bring the expected result (Table 35).
The influence of the type of solvent, pressure, temperature,
and modifier/substrate/Pt molar ratios was investigated in the
hydrogenation of fluoro- and trifluoromethyl-substituted
acetophenones (entries 2-10). Modification of a catalyst by
CD afforded the corresponding (S)-1-phenylethanol. Working
in strongly polar solvents, addition of TFA (entry 5) in a
Figure 24. Proposed interaction between TFAP and CD on Pt in
aprotic solvents.
Table 35. Inversions in Hydrogenation of Methyl Aryl Ketones
weakly polar solvent, and replacing CD by its ether deriva-
tives (entries 7-10) resulted in the inversion of ED.
According to the authors’ opinion,153a inversion in the
presence of strongly polar and acidic solvents is attributed
to special interactions with the OH function of CD, and to
the formation of a CD-acid ion pair, respectively. A possible
explanation for the moderate ee’s in the hydrogenation of
ring-substituted acetophenones is that a reaction pathway
without involvement of the OH function of CD is also
feasible. This competing pathway is even faster and provides
low ee to the opposite enantiomer. The inversion of ee is
usually attributed to changes in the reaction mechanism.
Unexpected inversion of ED was also observed after
replacement of CD by ether derivatives on Rh/alumina
catalyst (entries 8-10).153b Interestingly PhOCD is a more
effective chiral modifier for the reaction than CD.
4.1.3. Hydrogenation over Pd Catalyst
Hydrogenation of Pyruvates. The first observation of
inversion was reported in 1988 in the enantioselective
hydrogenation of activated ketones:116 in hydrogenation of
methyl pyruvate on Pd/C catalyst (i.e., on a catalyst other
than Pt, the regular catalyst of the Orito reaction), in the
presence of CD the formation of (S)-methyl lactate was
observed in a very slow reaction (Table 36, entry 1).
According to their most important conclusion (in 1988!), the
results strongly suggest that the enhanced adsorption of the
pyruvate ester is due to a stereochemically favorable interac-

entry ee (%) M* solvent X Y M* solvent ee (%) ref

1 3* MeOCD toluene H H CD toluene 16 153a
2 4* MeOCD toluene H F CD toluene 30 153a
3 6* MeOCD toluene H CF3 CD toluene 43 153a
4 4* MeOCD toluene CF3 CF3 CD toluene 45 153a
5 10* CD toluene + TFA CF3 CF3 CD toluene 45 153a
6 15* CD DMF CF3 CF3 CD toluene 45 153a
7 5* PhOCD toluene CF3 CF3 CD toluene 45 153a
8a 3* MeOCD toluene CF3 CF3 CD toluene 27 153b
9a 7* EtOCD toluene CF3 CF3 CD toluene 27 153b
10a 33* PhOCD toluene CF3 CF3 CD toluene 28 153b
a
Rh/alumina.

Table 36. Inversion in Hydrogenation of Methyl Pyruvate

entry ee (%) M* solvent catalyst M* solvent ee (%) ref

1 Pd/C CD MeOH 4* 116
2 13* CN EtOH Pd/Fe2O3 CD EtOH 5* 117
3 12* CN THF Pd/alumina CD THF 10* 120
4 11* CN EtOH Pd/alumina CD EtOH 5* 120
5 Pd/alumina CD MEK 7* 120
6a 15 CD THF Pd/alumina or Pd/C CD EtOH 14* 120
a
EtPy, 10 bar H2.
Unexpected Inversions in Asymmetric Reactions
Table 37. Inversions in Hydrogenation of Prochiral Alkenes on Pd Catalyst
tion with CD on the metal surface. In experiments carried
out 8 years later117 (1996; entry 2), they established that the
hydrogenation of methyl pyruvate over Pd differed from the
corresponding reaction over Pt in every important particular.
The ee was low (high over Pt) and in the reverse sense (e.g.,
CD modification provided an S-excess in the product over
Pd but an R-excess over Pt). The crucial role of the solvent
in determining the stereochemistry of the Orito reaction is
also verified by further experiments using cinchonas (entries
3 and 4). These experiments called attention to the solvent-
dependent hydrogenation of chiral modifiers, to be taken into
account in the interpretation of unexpected inversion. On the
basis of studies in deuterium, it was also concluded that
methyl pyruvate hydrogenation over Pd is a kinetically fast
hydrogenation of adsorbed enol formed via dissociative
adsorption of the R-ketoester. (On Pt, the ketone group of
the substrate is directly hydrogenated.)
Hydrogenation of Prochiral Alkenes. From the catalysts
most commonly used for the hydrogenation of alkenes and
their various substituted derivatives, Pd-based catalysts were
chosen for utilization in heterogeneous catalytic enantiose-
lective hydrogenations.6b,113a,114a,c,e,154 The first reproducible,
although low-ee CdC hydrogenations were reported at the
end of the 1980s.155,156 The number of published results of
heterogeneous catalytic enantioselective alkene hydrogena-
tions falls behind that of ketones. Probably that is why only
a handful of publications have reported on the stereochem-
istry of hydrogenations and, within this field, on inversion
(Table 37).
The first unexpected inversion was recognized in 1988 by
Nitta and Shibata in the hydrogenation of (E)-R-phenylcin-
Chemical Reviews, 2010, Vol. 110, No. 3 1695
namic acid over MeOCD-Pd chiral catalyst.119 In this case,
the (R)-product was formed in higher ee as compared to the
expected (S)-product on CD-Pt catalyst (entries 1 and 2).
The authors suggested that the interaction of the hydroxyl
group at C9 of CD with the carbonyl group in the substrate
is crucial for the induction of high ee.119 They proposed a
two-point interaction model for the interpretation of the
formation of the (S)-product.
Baiker et al. recognized in 2000 that 2-pyrones can be
hydrogenated in high ee over Pd modified with cinchona
alkaloids to produce the corresponding dihydropyrones, also
known as unsaturated δ-lactones.160 In the enantioselective
hydrogenation of 4-hydroxy-6-methyl-2-pyrone, MeOCD
was an inefficient modifier because the H-bonding interaction
of the OH group of CD with the carbonyl group of substrate
is hindered (entries 3-6). The ee was close to zero in
acetonitrile, but in other solvents, such as i-PrOH, AcOH,
and 3-pentanone, the opposite enantiomer (R)-product formed
with 10, 12, and 14% ee, respectively.157 The small but
significant ee to the opposite enantiomer is an indication of
some changes in the mechanism in the latter solvents.
According to the authors, a feasible model based on
2-pyrone-CD interactions is given in Figure 25. These pro
(S) and pro (R) ICs are strongly supported by the catalytic
and spectroscopic studies presented in ref 157. The bidentate
interaction in pro (S) model affords up to 85% ee to the (S)-
enantiomer. When this interaction is disfavored by a basic
or protic solvent or prevented by blocking the OH group of
CD (in MeOCD), the interaction shifts to the monodentate
model (Figure 25 pro (R)). The single attractive interaction
1696 Chemical Reviews, 2010, Vol. 110, No. 3 Bartók

observation is in accordance with our suggestion on the role

of the β-substituents in determining the sense of the chiral
induction in the hydrogenation of R,β-unsaturated acids over
Cinchona-modified Pd.161

4.2. Asymmetric Aldol Addition

Figure 25. Proposed intermediate complexes for enantioselective
hydrogenation of 2-pyrones.
is poorly effective: it afforded at best only 14% ee to the
(R)-enantiomer.157
Cinchona-modified Pd/alumina catalysts can be effective
for the enantioselective hydrogenation of N-acetyl dehydro-
phenylalanine methyl ester (NADPME) to the N-acetyl
phenylalanine methyl ester (NAPME).158 At low alkaloid/
NADPME molar ratios, CN gave (S)-NAPME and CD gave
(R)-NAPME. However, at higher alkaloid concentrations, the
sense of enantioselectivity inverted for CD, representing one
of the first examples of this type of behavior (entries 7-10).
The authors consider that the effect may be due to an
interaction of the modifier with specific Pd sites at low
modifier concentrations.158
The hydrogenation of 2-acetamidocinnamic acid resulted
in (R)-N-acetyl phenylalanine of 36% optical purity over CD
modified Pd/TiO2 under low H2 pressure. Increasing the
pressure led to an interesting inversion in the sense of the
ED, which was more pronounced if benzylamine was used
as additive (entries 11 and 12).159 In the hydrogenations of
2-acetamidoacrylic acid, no inversion in the ED occurred as
an effect of changes in the H2 pressure. This phenomenon
seems to be a special case, characteristic of R-acetamido-
β-unsaturated carboxylic acids and esters. Obviously the
inversion is related with the presence of the R-acetamido
group complemented with the steric effect of the β-phenyl
substituent. Interestingly, the configurations of the products
obtained in excess in the hydrogenation of 2-acetamidocin-
namic acid and 2-acetamidoacrylic acid using the same
modifier under low H2 pressures were opposite. This
Table 38. Inversion in the L-Proline-Catalyzed Direct Aldol Reaction
A few studies to immobilize L-proline on silica, MCM-
41, SILC, and dendrimers for direct asymmetric aldol
reactions have been reported;162 however, an inversion
phenomenon was not observed. The test reaction employed
in the relevant research is shown in Table 38. Table 38
describes two examples for unexpected inversion observed
in heterogenized/immobilized L-proline-catalyzed asymmetric
aldol addition. In one of these, inversion was brought about
by heterogenization of the L-proline organocatalyst, whereas
in the other one, inversion is due to the solvent effect
occurring in the course of the application of the immobilized
catalyst. To our best knowledge, this was the first unexpected
reversal inversion in heterogeneous asymmetric catalysis
observed in a reaction other than hydrogenation.
The L-proline-catalyzed homogeneous aldol reaction gave
the aldol product with (R)-configuration (entry 1).83c When
γ-Al2O3 was added to the aldol reaction mixture, the ee value
was reduced (entry 2). When the concentration of γ-Al2O3
was decreased, aldol with an (S)-configuration was unexpect-
edly obtained and the ee values increased from 4 to 21%
(entries 3 and 4). The results suggest that a unique catalyst
is formed on the adsorption of L-proline on γ-Al2O3. This
inversion phenomenon is found to be general for different
types of amino acids adsorbed on γ-Al2O3 (entries 6-12).83c
To explain the inversion, the adsorption of L-proline on
γ-Al2O3 and UV-Raman spectra of L-proline in solid form
and on γ-Al2O3 were obtained. These experiments indicate
that L-proline strongly interacts with γ-Al2O3 and that the
carboxylate groups are involved in the interaction. The
L-proline/γ-Al2O3 preadsorbed with pyridine resulted in a
dramatically reduced ee value, clearly showing that the acidic

entry ee (%) conv. (%) Catalyst* Catalyst* conv. (%) ee (%) ref
1 68 80 L-proline 83c
2 22 61 7.2b 83c
3 5b 80 4* 83c
4 3.3b 78 21* 83c
5 64 78 3.3b (γ-Al2O3 silylated) 83c
6 47 L-leucine + γ-Al2O3
L-leucine 15* 83c
7 48 L-alanine + γ-Al2O3
L-leucine 5* 83c
8 46 L-tryptophan + γ-Al2O3
L-leucine 8* 83c
9 23 L-phenylalanine + γ-Al2O3
L-leucine 4* 83c
10 20 L-threonine + γ-Al2O3
L-leucine 5* 83c
11 20 L-glutamine + γ-Al2O3
L-leucine 2* 83c
12 13 L-lysine + γ-Al2O3
L-leucine 6* 83c
13 77 68 PEG-Proa in DMF PEG-Proa in acetone 23 21 163

a
PEG-Pro ) poly(ethyleneglycols) supported (2S,4R)-4-hydroxyproline. b
L-Proline/Al2O3 (molecules/nm2).
Unexpected Inversions in Asymmetric Reactions Chemical Reviews, 2010, Vol. 110, No. 3 1697

Table 39. Inversion in the Diels-Alder Reaction

entry ee (S %) conv. (%) Lewis acid R L*a or Catalyst 1-3 Lewis acid conv. (%) ee (R %) ref
1 Me Catalyst 1 100 38 164
2 Me Catalyst 2 99 3 164
3 18* 100 Me Catalyst 3 164
4 3 100 Mg(ClO4)2c,d H (S)-Me-BOX-Ph Mg(ClO4)2d 100 66* 56
5 22 100 Mg(OTf)2c,d H (S)-Me-BOX-Ph Mg(OTf)2cd 100 67* 56
6 32 100 Cu(OTf)2c,d H (S)-Me-BOX-t-Bu 56
7 4 13 Cu(OTf)2c,d H (S)-Me-BOX-Ph 56
8 20 100 Cu(OTf)2 H (S)-H-BOX-Ph Cu(OTf)2 100 33* 57b
9 59 50 Cu(OTf)2 H (S)-Me-BOX-t-Bu 57b
10 57 43 Cu(OTf)2b H (S)-Me-BOX-t-Bu 57b
11 90 97 Cu(OTf)2 Me (S)-Me-BOX-t-Bu 58
12 85 43 Cu(OTf)2b Me (S)-Me-BOX-t-Bu 58
13 15 100 Cu(OTf)2 Me (S)-Me-BOX-Ph Cu(OTf)2 100 31* 58
14 60 95 Mg(OTf)2 Me (S)-Me-BOX-Ph Mg(OTf)2 46 30* 58
15 22 97 Zn(OTf)2 Me (S)-Me-BOX-Ph Zn(OTf)2 95 24* 58

a
L*: see Figure 8; mainly endo additions occur. b Silica supported. c SILC: silica-supported ionic liquid phase catalyst. d Et2O.

sites on γ-Al2O3 play an important role in the aldol reaction,
because the preadsorbed pyridine can occupy these sites.
Silylated γ-Al2O3 gave results similar to those obtained with
the free L-proline (entry 5). These results suggest that the
surface hydroxyl groups on γ-Al2O3 are essential to induce
the inversion of enantioselectivity.
The authors have assumed that coupling of L-proline with
the γ-Al2O3 surface gives an organo-inorganic bifunctional
catalyst for direct asymmetric aldol reactions. The amine
group of the adsorbed proline activates acetone, and the
hydroxyl group on γ-Al2O3 activates the carbonyl of p-
nitrobenzaldehyde through hydrogen bonding (see IC in
Table 38). The Si face on the aldehyde may be more easily
attacted by the enamine on γ-Al2O3, resulting in formation
of the product with an S configuration. As regards the
experimental observation described in entry 13, Table 38,
the authors of ref 163 reported no more than stating that
remarkably, and quite surprisingly, the use of acetone as the
reaction solvent gave an aldol of the opposite absolute
configuration. Ref 163 provides no more information regard-
ing inversion. In addition to the test reaction, the report
describes the results of the aldol condensation of other
compounds and of the reuse of immobilized catalysts. The
extensive studies on this chiral inversion on a solid surface
help clarify the adsorption mode of modifier and its interac-
tion with reactant and, hence, are helpful in explaining the
origin of the ED.
4.3. Asymmetric Diels-Alder Addition
Important antecedents regarding the asymmetric DA
reaction are summarized in subsection 2.7.1, whereas the
recently published monograph reviews our present knowl-
edge on the use of chiral heterogeneous catalysts.99j In the
course of experiments on heterogenized catalysts of a great
variety of structures, the first phenomenon of unexpected
inversion was observed on a polymer with Taddol-type chiral
surface sites164 (entry 3, Table 39). In the presence of
heterogenized Catalyst 3, the major product had an opposite
configuration (2S) as compared to the product obtained in
homogeneous phase (2R) that is in the presence of Catalyst
1 and Catalyst 2. In the case of Taddol derivatives containing
2-naphthyl groups instead of 3,5-Me2C6H3 groups, no inver-
sion was observed, as the major product had (2S) configu-
ration in all cases. Occurrences of inversion were later
encountered in the course of studies on chiral BOX
complexes.56–58
Supported IL catalysts (SILC) have been developed using
surface-modified silica, which show good reactivity and
reversal of enantioselectivity for the case of the magnesium-
1698 Chemical Reviews, 2010, Vol. 110, No. 3
Table 40. Inversion in the Hetero Diels-Alder Reaction
ee ee
entry 2R,4S(%) 2R,4R(%) conv. (%) Lewis acid (La) conditions
1 21 94 Cu(OTf)2 hom.
2 het.
3 56 100 Cu(OTf)2 hom.
4 24 100 Cu(OTf)2 het.
5 hom.
6 het.
based BOX complexes.56 Irrespective of the anion of the
Lewis acid, in the case of the application of the SILC variety
of (S)-BOX-Ph-Mg complexes, inversion takes place in the
DA reaction as compared to the reaction in IL (entries 4
and 5). In the case of the corresponding Cu complexes, no
inversion happens (entries 6 and 7).
A reversal in ee when changing from a homogeneous to
a heterogeneous system has been previously reported57,58 and
was thought to be due to the dissociation of the catalyst anion
on the support, resulting in a change in catalyst geometry.
For the SILC-mediated reactions where the catalysis takes
place in a solid-supported thin ionic liquid film, it is more
likely that the change in catalyst geometry is induced by the
large concentration of anions in an analogous manner to that
found under homogeneous reaction conditions.
Table 39 presents examples for unexpected inversion, in
the majority of which the sense of ED observed in homo-
geneous reaction was altered by the immobilization/hetero-
genization of the chiral complexes.57a,58 Two chiral BOX-Cu(II)
complexes have been immobilized on silica via H-bonding
interactions.57a The immobolized catalysts were tested in a
standard DA reaction and gave surprising results. Where the
immobilized (S)-H-BOX-Ph-Cu(OTf)2 catalyst was used, the
predominant enantiomer formed was the opposite of that
produced using the same catalyst in a homogeneous reaction
(entries 8-10). The behavior of the immobilized t-Bu catalyst
((S)-Me-BOX-t-Bu-Cu) is very different from that of its
phenyl analogue. The phenyl-substituted catalyst maintained
its activity and enantioselectivity and was highly recyclable,
whereas the t-Bu catalyst was obviously less stable and less
effective when immobilized by this technique. No further
information is supplied on experimental observations relevant
to the interpretation of unexpected inversion in either ref 57a
or more recent publications from the same laboratory.
Chiral BOX-complexes of Cu(II)-, Mg(II)-, and Zn(II)-
triflates have been immobilized on silica support via hydrogen-
bonding interactions (entries 11-15).58 Moderate or, occa-
sionally, good ee values could be attained in the outlined
DA reaction. Similar relative behavior was observed57a in
the case of t-Bu- and Ph-BOX complexes.58 A surprising
observation from these studies is that the configuration of
the product changed on going from the homogeneous to the
heterogeneous system in the case of BOX-Ph complexes.
This is of both theoretical and practical importance, as it
indicates that immobilization alters the active catalytic
species. The establishment of hydrogen-bonding interactions
between the silanol groups of the support and the triflate
anions was verified by IR studies.58 Hydrogen bonding can
provide a simple way for the immobilization of homogeneous
catalysts, which requires neither modification of the catalysts
nor functionalization of the surface.165
Bartók
ee ee
L* Lewis acid (La) conv. (%) 2S,4S (%) 2S,4R (%) ref
(S)-Me-BOX-Ph 165a
(S)-Me-BOX-Ph CuHY 16 41* 165a
(S)-Me-BOX-t-Bu 57a
(S)-Me-BOX-t-Bu 57a
(S)-H-BOX-Ph Cu(OTf)2 100 27 57a
(S)-H-BOX-Ph Cu(OTf)2 100 18 57a
4.4. Asymmetric Hetero Diels-Alder Additions
Earlier results showing that Cu-BOX complexes were
identified as one of the best catalyst type of DA, and hetero
Diels-Alder additions (HAD) are summarized in reviews.52g,166
Table 40 includes some experimental data from two reports
describing remarkable and interesting results in HAD addi-
tion in the presence of heterogenized chiral Cu-BOX
catalysts.
An unexpected effect was observed for the heterogeneous
catalytic HAD reaction, which was not apparent in the
homogeneously catalyzed process, namely, the reversal in
enantioselectivity of the dihydropyran product.167 Initially,
the 2R,4S product is observed, and subsequently, this
switches to the 2S,4R product. This reversal of enantiose-
lectivity could be attributed to the confinement effect of the
porous supports.
The peculiarity of the experimental data published in the other
report is that, under identical experimental conditions, there is
no difference between the senses of the EDs of the homoge-
neous and heterogeneous HAD reactions.57a However, inver-
sion is observed in both the homogeneous and the hetero-
geneous reaction in the presence of BOX catalysts bearing
the t-Bu and the Ph substituents but otherwise of identical
configuration. This is another characteristic example for the
reaction of heterogenized complexes: the sense of ED is
significantly affected not only by the configuration of the
chiral catalyst but also by the substituents attached to the
chiral center.
In my opinion, the experimental facts showing that, in the
case of DA and HAD reactions carried out on the same chiral
catalyst under identical experimental conditions, unexpected
inversion takes place in one (DA) and is absent in the other
(HAD) also have special significance57a (see data in Tables
39 and 40).
I have some doubts regarding the conclusions, because the
configurations of the main products obtained under the condi-
tions of homogeneous catalysis are not identical ((2R,4S) and
(2R,4R), respectively). Consequently, neither are the configura-
tions of the compounds with opposite configurations identical
in the two reports.57a,167a It has earlier been suggested in studies
on the mechanism of the HAD reaction that the mechanism
might be stepwise rather than concerted.167b
4.5. Asymmetric Cyclopropanation
Reviews on asymmetric cyclopropanation have been pub-
lished continuously, giving repeatedly updated informa-
tion.52d,e,g,67a These reviews have given account of surprising
experimental data classifiable as unexpected inversion. Below
follows the description of cyclopropanation, in which the sense
Unexpected Inversions in Asymmetric Reactions
Table 41. Inversion in Cyclopropanation
entry (1R,2S) (%) trans/cis L*
1 8 71/29 (S)-Me-BOX-Ph
2 8 68/32 (S)-Me-BOX-Ph
3 (S)-Me-BOX-t-Bu
4 (S)-Me-BOX-t-Bu
of ED observed in homogeneous catalysis was reversed by
immobilization of the chiral complex. In the cyclopropanation
of styrene with ethyl diazoacetate, inversion was observed in
the presence of (S)-Me-BOX-Ph-Cu catalyst immobilized on
clay (laponite) by electrostatic interaction.68
Mayoral et al. studied cyclopropanation and observed an
interesting change in stereoselectivity: both cis/trans selectivity
and enantioselectivity were changed when laponite-immobilized
catalyst was used. In DCM, in both the homogeneous-phase
and the heterogeneous-phase reaction, the trans/cis selectivity
was changed (Table 41, entry 1).68a,d
An unexpected observation was made in styrene: in
homogeneous phase, the trend is the same as in DCM,
whereas on the immobilized catalyst both the trans/cis ratio
and the sense of enantioselection was reversed in each of
the apolar solvents studied. Unexpected inversion brought
about by heterogenization has also been confirmed in DCM
by recent studies.68d As shown in Table 41, heterogenization
shifted the product ratio toward the formation of the cis
enantiomers. The phenomenon has also been observed on
supported ionic liquid films.68e Although there is some
difference between the effects of t-Bu- and Ph-BOX com-
plexes, it is less characteristic than those observed in, e.g.,
the DA reactions.
The authors explain that this solvent effect is due to the
difference in the distance between the BOX complex and
laponite. In the case of apolar solvents, this distance is shorter,
which leads to considerable changes in both conformational
relations and steric interactions. To explain the formation of
the (1S,2R)-product in higher ee as a result of unexpected
Figure 26. Assumed intermediate complex on the clay surface in
cyclopropanation.
Table 42. Inversion in Aziridination
entry ee (S%) yield (%) L*a
1 28* 80 (S)-H-BOX-t-Bu
2 24* 58 (S)-Me-BOX-t-Bu
3 77* 72 (S)-H-BOX-t-Bu
4 82* 68 (S)-Me-BOX-t-Bu
a
L*: see Figure 8; HY zeolite.
Chemical Reviews, 2010, Vol. 110, No. 3 1699
solvent trans/cis (1S,2R) (%)
DCM 28/72 32*
styrene 9/91 41*
DCM 53/47 10*
styrene 15/85 48*
inversion, the authors refer to the IC68b outlined in Figure 26,
which is also supported by DFT calculations.68c
In their latest publications, they emphasize the role of
surface-confinement effects. According to the authors, ex-
perimentally verifying the mechanistic proposal is not
possible. Only indirect evidence and molecular modeling
studies are able to shed light on these interesting but
extremely complicated catalytic systems.
4.6. Asymmetric Aziridination
The first enantioselective aziridination catalyst was found
only in 1998.168 The achieved results have been review-
ed.52d,e,g,67b Rechavi and Lemaire52e summarized the experi-
mental results of Hutchings et al.69a according to Table 42.
Namely, the (S)-BOX-Cu(II) complexes gave the (R)-
aziridines, whereas in the heterogeneous phase the (S)-
aziridines were obtained.69a This reversal of induction
indicates that, in this case, the zeolite HY pores significantly
influence the substrate-catalyst interaction. Similarly to
cyclopropanation, the surface significantly affects confor-
mational interactions.
4.7. Summary
In the previous chapters, examples for unexpected
inversion in a variety of reactions, using a large selection
of catalysts, were enumerated. In consideration of the
shortage of published experimental observations of un-
expected inversion in the field of heterogeneous catalytic
asymmetric reactions (with the exception of hydrogena-
tion) (see the introduction of section 4), this summary
reviews the large number of experimental data points
obtained in hydrogenations. Observations made in other
heterogeneous catalytic asymmetric reactions were men-
tioned in subsections 2.8 and 3.8. Namely, these unex-
pected inversions take place in the heterogenized variants
of homogeneous catalytic reactions. The conclusion that
can be drawn from the few inversion reactions observed
to date is that, in the heterogeneous asymmetric reactions,
X yield (%) ee (R%)
Me 85 8
Me 78 35
NO2 42 31
NO2 79 43
1700 Chemical Reviews, 2010, Vol. 110, No. 3 Bartók

Table 43. Inversions in Heterogeneous Catalytic Enantioselective Hydrogenations

a
For abbreviations, see Figures 14 and 16.

the reversal of enantioselectivity could be attributed to
the confinement effect of the porous supports.
Section 4 discusses only two widely studied examples
that have stereochemistries not fitting in with empirical
rules and that have been observed in the presence of a
great variety of substrates and chiral modifiers, in
hydrogenations also utilized in industrial applications.
Table 43 shows a selection of these experimental data,
namely, the results showing the largest values of unex-
pected ee. On the one hand, these can be compared to the
Unexpected Inversions in Asymmetric Reactions
methyl acetoacetate (MAA) reference compound on TA-
MNi; on the other hand, the data obtained on Pt modified
with cinchona alkaloid derivatives can be compared to
those obtained on Pt modified with the parent cinchonas.
At the time of its discovery, the TA-MNi catalyst was
eminently suitable for the hydrogenation of β-oxoesters.
Hydrogenation of methyl acetoacetate on TA-MNi yielded
(R)-product (Scheme 14, Table 43, entry 1). In the case of
a γ-oxoester (entry 2), ee was found to decrease, and in the
case of the other oxoesters and simple ketones, it was already
the (S)-products that were formed in higher ee (entries 3-5).
This finding was at that time unexpected.
Studies using Pt catalysts modified with cinchona
alkaloids gave unexpected results when derivatives of the
parent alkaloids were used (Table 43). The objective of
these studies was the elucidation of the reaction mecha-
nism because, after adequate optimization, the four
inexpensive parent cinchonas (CD, CN, QN, QD) allowed
the attainment of high ee values (over 90%). According
to the generally accepted empirical rule, ED is determined
by the configuration of the C8 and C9 atoms of cinchonas.
CD and QN (both 8(S),9(R) configurations) yield (R)-
products, whereas CN and QD (both 8(R),9(S) configura-
tions) give (S)-products. Deviations from this rule were
unexpected. The unexpected inversion was the first
significant experimental observation indicating that, in
enantioselective hydrogenation over cinchona alkaloids
modified catalysts, it is not the C8 chiral center of the
alkaloid that controls the sense of the chiral induction.123
Maximal unexpected inversion in ee was caused by the
modifiers β-ICN and PhOCD in the hydrogenation of each
substrate studied (Table 43, accented in bold).
It was revealed in the course of research that those of
the real effects, which do not alter the configuration of
stereogenic centers C8 and C9, can be attributed to a
modified conformation of the substrate-modifier surface
complex responsible for ED. According to some research
groups, this means a change in adsorption mode, whereas
others assume that a change in the reaction mechanism
or one of its steps is responsible. These details have been
discussed above.
As regards the experimental verification of the inter-
pretation of these inversions, in contrast to those discussed
in the previous sections, there exists only indirect evidence
with respect to the existence of the short-lived surface
ICs of fast surface reactions. The following indirect
confirmations of substrate-modifier interactions have been
published: (i) ATR-IR spectroscopy;141,169 (ii) NMR
spectroscopy145a,b under conditions similar to hydrogena-
tion in solution; (iii) RAIRS and STM studies on Pt(III)
in ultrahigh vacuum conditions;170 and (iv) theoretical (DFT
calculations).171
5. Conclusion
Asymmetric reactions proceeding on three different types
of catalyst systems (chiral metal complexes, chiral organo-
catalysts, and chiral heterogeneous catalysts), in which the
formation of enantiomers was accompanied by unexpected
inversion, were collected. The experimental data compiled
are summarized in Tables 15, 27, and 43. The results of the
reviewed research can be summarized in the following
(without going into the details of subsections 2.8, 3.8, and
4.7): (i) Unexpected inversion is not a unique phenomenon
among enantioselective catalytic reactions of various types.
Chemical Reviews, 2010, Vol. 110, No. 3 1701
(ii) Since the necessary experimental data are not available,
it is not yet possible to formulate general relationships
between the chiralities of the individual catalyst types and
the sense of ED. (iii) Whereas chiral catalysts are naturally
responsible for the origin of chiral induction, determination
of the sense of ED can be attributed not only to the absolute
configuration of chiral catalysts but also to a large number
of other factors (Tables 15, 27, 43). (iv) A sensitive
equilibrium of attractive and repulsive interactions depending
on the structures of the chiral catalyst, L*, M*, and the
substrates controls the conformation of the IC responsible
for ED, which favors either Re-face or Si-face selectivity.
(v) The theoretical as well as the practical importance of
detailed investigations on the unexpected phenomena ac-
companying enantioselective reactions will increase in the
future. (vi) It is of theoretical importance to explore the origin
of chiral induction as thoroughly as possible, and it is of
practical importance to create the means for the synthesis
of enantiomer pairs in high ee using the same chiral raw
material.
On the basis of the summary tables cited above, a large
number of questions can be posed to which it is as yet
impossible to give answers supported by concrete experi-
mental evidence. In order to find the answers to these
questions and, most importantly, to attain the goals listed
above, many tasks await realization, including the fol-
lowing: (i) Due to the great significance of asymmetric
syntheses, detailed analysis of unexpected phenomena
revealed in the course of research is not only justified but
also necessary. (ii) In spite of the availability of a large
number of experimental data points, studies varying only
one parameter, e.g., the absolute configuration of the chiral
atoms of catalysts and keeping the experimental condi-
tions, the techniques applied, and the origin of the starting
materials constant are regrettably rare. (iii) The develop-
ment of the heterogenized versions of enantioselective
reactions requires more extensive research. (iv) Reactions
with unusual ED are expected to be discovered in
increasing numbers in futures studies, and their interpreta-
tion requires novel research techniques. (v) The solution
to these problems may become part of designable synthetic
procedures.
Because of volume restrictions, many important topics
in enantioselective reactions are not even mentioned in
this review that focuses on a relatively narrow field. The
author asks for the reader’s forgiveness for any possible
deficiencies.
6. List of Abbreviations
AcOQD O-acetylquinidine
AcOQN O-acetylquinine
ATR-IR attenuated total reflection spectroscopy
BARF tetrakis[3,5-bis(trifluoromethyl)phenyl]borate
BDPP 2,4-bis(diphenylphosphino)pentane
BH Baylis-Hillman
BINAP 2,2′-bis(diphenylphosphino)-1,1′-binaphthyl
BINOL 2,2′-binaphthol
[bmim]BF4 1-butyl-3-methylimidazolium tetrafluoroborate
[bmim]PF6 1-butyl-3-methylimidazolium hexafluorophos-
phate
[bpy]BF4 1-butyl-3-methylpyridinium tetrafluoroborate
Boc t-butoxycarbonyl
BOX bis(oxazoline)
BSA N,O-bis(trimethylsilyl)acetamide
CD cinchonidine
1702 Chemical Reviews, 2010, Vol. 110, No. 3 Bartók

CFBR continuous fixed-bed reactor (3) Sibi, M. P.; Liu, M. Curr. Org. Chem. 2001, 5, 719.
CN cinchonine (4) Zanoni, G.; Castronovo, F.; Franzini, M.; Vidari, G.; Giannini, E.
DA Diels-Alder Chem. Soc. ReV. 2003, 32, 115.
(5) Tanaka, T.; Hayashi, M. Synthesis 2008, 3361.
dba dibenzylideneacetone (6) (a) Bartók, M. Curr. Org. Chem. 2006, 10, 1533. (b) Mallat, T.;
DCM dichloromethane Orglmeister, E.; Baiker, A. Chem. ReV. 2007, 107, 4863.
DMF dimethylformamide (7) Morrison, J. D., Ed. Asymmetric Synthesis; Academic Press: New
DMSO dimethyl sulfoxide York, 1985.
DuPhos substituted 1,2-bis(phosphonalo)benzene (8) Bosnich, B. Asymmetric Catalysis; Martinus Nijhoff: Dordrecht, The
ED enantiodifferentiation Netherlands, 1986.
ee enantiomeric excess (9) Trost, B. M., Fleming, I., Eds. ComprehensiVe Organic Synthesis;
Pergamon Press: Oxford, 1991.
ee* unexpected ee (in tables) (10) Brunner, H.; Zettlemeyer, W. Handbook of EnantioselectiVe Catalysis
EtLt ethyl lactate with Transition Metal Compounds; Wiley-VCH: Weinheim, Ger-
EtPy ethyl pyruvate many, 1993.
HMPA hexamethylphosphoramide (11) Ojima, I., Ed. Catalytic Asymmetric Synthesis; Wiley-VCH: Wein-
IC intermediate complex heim, Germany, 1993.
ICN isocinchonine (12) Noyori, R. Asymmetric Catalysis in Organic Synthesis; Wiley-VCH:
ICP isocupreidine New York, 1994.
(13) Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds. ComprehensiVe
IL ionic liquid Asymmetric Catalysis; Springer: Berlin, 1999; Vols. 1-3.
KPL ketopantolactone (14) Bolm, C.; Gladysz, J. A. Chem. ReV. (EnantioselectiVe Catalysis
L* chiral ligand (Special Issue)) 2003, 103, 2761.
laponite laponite RD (clay from Laponite Inc.) (15) Blaser, H. U., Schmidt, E., Eds. Asymmetric Catalysis on Industrial
LDA i-Pr2NH + BuLi Scale. Challenges, Approaches and Solutions; Wiley-VCH: New
M* chiral modifier York, 2004.
MPEG poly(ethylene glycol) monomethyl ether (16) de Vries, J. G., Cornelis, J., Eds. Handbook of Homogeneous
Hydrogenation; Wiley-VCH: Amsterdam, The Netherlands, 2007.
MR Mukaiyama’s reagent (17) (a) Knowles, W. S.; Sabacky, M. J. Chem. Commun. 1968, 1445.
MRNi modified Raney-Ni (b) Knowles, W. S.; Sabacky, M. J.; Vineyard, B. D. Chem. Commun.
MS molecular sieves 1972, 10. (c) Knowles, W. S. Acc. Chem. Res. 1983, 16, 106.
Ns p-nitrophenylsulphonyl (18) (a) Kagan, H. B.; Dang, T. P. Chem. Commun. 1971, 481. (b) Kagan,
Nu- nucleophile H. B.; Dang, T. P. J. Am. Chem. Soc. 1972, 94, 6429.
PEG poly(ethylene glycol) (19) (a) Noyori, R. Chem. Soc. ReV. 1989, 18, 187. (b) Noyori, R. Science
1990, 248, 1194.
PhOCD O-phenylcinchonidine
(20) (a) Tang, W.; Zhang, X. Chem. ReV. 2003, 103, 3029. (b) Kuwano,
PHOX phosphanyloxazoline R.; Ito, Y. J. Org. Chem. 1999, 64, 1232. (c) Berens, U.; Selke, R.
PPD 1-phenylpropane-1,2-dione Tetrahedron: Asymmetry 1996, 7, 2055. (d) Reetz, M. T.; Sell, T.;
PTC phase-transfer catalyst Meiswinkel, A.; Mehler, G. Angew. Chem., Int. Ed. 2003, 42, 790.
p-TSA p-toluenesulfonic acid (e) Reetz, M. T.; Mehler, G. Tetrahedron Lett. 2003, 44, 4593.
PyBOX pyridine-2,6-bis(oxazoline) (21) (a) Kuwano, R.; Sawamura, M.; Ito, Y. Bull. Chem. Soc. Jpn. 2000,
QD quinidine 73, 2571. (b) Ikariya, T.; Ishhi, Y.; Kawano, H.; Arai, T.; Saburi,
M.; Yoshikawa, S.; Akutagawa, S. J. Chem. Soc., Chem. Commun.
QN quinine 1985, 922. (c) Perry, M. C.; Cui, X.; Powell, M. T.; Hou, D.-R.;
RAIRS reflection absorption infrared spectroscopy Reibenspies, J. H.; Burgess, K. J. Am. Chem. Soc. 2003, 125, 113.
SILC supported ionic liquid catalysts (22) Furegati, M.; Rippert, A. J. Tetrahedron: Asymmetry 2005, 16, 3947.
STM scanning tunneling microscopy (23) (a) Zaitsev, A. B.; Adolfsson, H. Org. Lett. 2006, 8, 5129. (b) Ahlford,
TA tartaric acid K.; Zaitsev, A. B.; Ekström, J.; Adolfsson, H. Synlett 2007, 2541.
TA-MNi (R,R)-tartaric acid modified Ni catalyst (c) Wettergren, J.; Zaitsev, A. B.; Adolfsson, H. AdV. Synth. Catal.
2007, 349, 2556.
TBS t-butyldimethylsilyl (24) Consiglio, G.; Pino, P. Top. Curr. Chem. 1982, 105, 77.
Tf triflyl (25) (a) Breit, B.; Seiche, W. Synthesis 2001, 1. (b) Bohnen, H.; Cornils,
TFA trifluoroacetic acid B. AdV. Catal. 2002, 47, 1.
TFAP 2,2,2-trifluoroacetophenone (26) (a) Kollár, L.; Bakos, J.; Tóth, I.; Heil, B. J. Organomet. Chem. 1988,
TMS trimethylsilyl 350, 277. (b) Tóth, I.; Guo, I.; Hanson, B. E. Organometallics 1993,
TRAP bis[(dialkylphosphino)ethyl]biferrocenes 12, 848.
Ts p-toluenesulfonyl (27) Casey, C. P.; Martins, S. C.; Fagan, M. A. J. Am. Chem. Soc. 2004,
126, 5585.
(28) Katsuki, T.; Sharpless, K. B. J. Am. Chem. Soc. 1980, 102, 5974.
7. Acknowledgments (29) Asymmetric Methods of Epoxidation. In ComprehensiVe Organic
Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon Press: Oxford,
The author wishes to express his gratitude to Professors 1991; Vol. 7, Chapter 3, p 389.
(30) (a) Gao, Y.; Hanson, R. M.; Klunder, J. M.; Ko, S. Y.; Masamune,
A. Molnár, F. Notheisz, Drs. Gy. Szöllősi, K. Balázsik, K. H.; Sharpless, K. B. J. Am. Chem. Soc. 1987, 109, 5765. (b) Guo,
Felföldi, I. Bucsi and K. Szőri, and Sz. Cserényi, who made H.; Shi, X.; Qiao, Z.; Hou, S.; Wang, M. Chem. Commun. 2002,
significant contributions to the development of the hetero- 118.
geneous catalyzed reactions. In particular, I am indebted to (31) Reed, N. N.; Dickerson, T. J.; Boldt, G. E.; Janda, K. D. J. Org.
Chem. 2005, 70, 1728.
Prof. Árpád Molnár for making himself promptly available (32) (a) Superchi, S.; Donnoli, M. I.; Rosini, C. Tetrahedron Lett. 1998,
for extremely helpful discussions, for a detailed scrutiny of 39, 8541. (b) Yamamoto, K.; Ando, H.; Shuetake, T.; Chikamatsu,
the manuscript, and for his critical comments. I would like H. J. Chem. Soc., Chem. Commun. 1989, 754. (c) Reetz, M. T.; Merk,
C.; Naberfeld, G.; Rudolph, J.; Griebenow, N.; Goddard, R. Tetra-
to warmly thank my wife for her indispensable help in the hedron Lett. 1997, 38, 5273.
preparation of this review. Financial assistance by the (33) (a) Trost, B. M.; Dietsche, T. J. J. Am. Chem. Soc. 1973, 95, 8200.
Hungaryen National Science Foundation (OTKA K72065) (b) Trost, B. M.; Strege, P. E. J. Am. Chem. Soc. 1977, 99, 1649. (c)
is also gratefully acknowledged. Trost, B. M. Acc. Chem. Res. 1980, 13, 385. (d) Tsuji, J. J.
Organomet. Chem. 1986, 300, 281.
(34) (a) Trost, B. M.; Van Vranken, D. L. Chem. ReV. 1996, 96, 395. (b)
8. References Trost, B. M.; Crawley, M. L. Chem. ReV. 2003, 103, 2921.
(35) (a) Anderson, J. C.; Cubbon, R. J.; Harling, J. D. Tetrahedron:
(1) Pasteur, L. Ann. Chem. 1948, 442. Asymmetry 1999, 10, 2829. (b) Nomura, N.; MermetBouvier, Y. C.;
(2) Kim, Y. H. Acc. Chem. Res. 2001, 34, 955. RajanBabu, T. V. Synlett 1996, 745.
Unexpected Inversions in Asymmetric Reactions
(36) (a) Hoarau, O.; Aı̈t-Haddou, H.; Daran, J.-C.; Cramailère, D.;
Balavoine, G. G. A. Organometallics 1999, 18, 4718. (b) Aı̈t-Haddou,
H.; Hoarau, O.; Cramailère, D.; Pezet, F.; Daran, J.-C.; Balavoine,
G. G. A. Chem.sEur. J. 2004, 10, 699.
(37) (a) Chelucci, G.; Pinna, G. A.; Saba, A.; Valentini, R. Tetrahedron:
Asymmetry 2000, 11, 4027. (b) Li, X. G.; Cheng, X.; Ma, J. A.; Zhou,
Q. L. J. Organomet. Chem. 2001, 640, 65. (c) Danjo, H.; Higuchi,
M.; Yada, M.; Imamoto, T. Tetrahedron Lett. 2003, 45, 603.
(38) (a) Trost, B. M.; Toste, F. D. J. Am. Chem. Soc. 1999, 121, 4545.
(b) Zhang, W.; Shi, M. Tetrahedron: Asymmetry 2004, 15, 3467. (c)
Xie, F.; Liu, D. L.; Zhang, W. B. Tetrahedron Lett. 2008, 49, 1012.
(39) Zhou, J.; Ye, M.-C.; Huang, Z.-Z.; Tang, Y. J. Org. Chem. 2004,
69, 1309.
(40) Kobayashi, S.; Horibe, M. J. Am. Chem. Soc. 1994, 116, 9805.
(41) Trost, B. M.; Fettes, A.; Shireman, B. T. J. Am. Chem. Soc. 2004,
126, 2660.
(42) (a) Evans, D. A.; Kozlowski, M. C.; Murry, J. A.; Burgey, C. S.;
Campos, K. R.; Connell, B. T.; Staples, R. J. J. Am. Chem. Soc.
1999, 121, 669. (b) Matsunaga, H.; Yamada, Y.; Ide, T.; Ishizuka,
T.; Kunieda, T. Tetrahedron: Asymmetry 1999, 10, 3095. (c) Evans,
D. A.; Burgey, C. S.; Kozlowski, M. C.; Tregay, S. W. J. Am. Chem.
Soc. 1999, 121, 686.
(43) (a) Ghosh, A. K.; Mathivanan, P.; Cappiello, J. Tetrahedron:
Asymmetry 1998, 9, 1. (b) Helmchen, G.; Pfaltz, A. Acc. Chem. Res.
2000, 33, 336. (c) Shibasaki, M.; Sasai, H.; Arai, T. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 1236. (d) Luzzio, F. A. Tetrahedron 2001,
57, 915. (e) Shibasaki, M.; Yoshikawa, N. Chem. ReV. 2002, 102,
2187.
(44) (a) Lu, S. F.; Du, D. M.; Zhang, S. W.; Xu, J. Tetrahedron:
Asymmetry 2004, 15, 3433. (b) Du, D. M.; Lu, S. F.; Fang, T.; Xu,
J. X. J. Org. Chem. 2005, 70, 3712.
(45) (a) Nishiwaki, N.; Knudsen, K. R.; Gothelf, K. V.; Jørgensen, K. A.
Angew. Chem., Int. Ed. 2001, 40, 2992. (b) Salter, M. M.; Kobayashi,
J.; Shimizu, Y.; Kobayashi, S. Org. Lett. 2006, 8, 3533. (c) Cutting,
G. A.; Stainforth, N. E.; John, M. P.; Kociok-Köhn, G.; Willis, M. C.
J. Am. Chem. Soc. 2007, 129, 10632.
(46) Yan, X. X.; Peng, Q.; Li, Q.; Zhang, K.; Yao, J.; Hou, X. L.; Wu,
Y. D. J. Am. Chem. Soc. 2008, 130, 14362.
(47) (a) Dounay, A. B.; Overman, L. E. Chem. ReV. 2003, 103, 2945. (b)
Mellah, M.; Voituriez, A.; Schulz, E. Chem. ReV. 2007, 107, 5133.
(48) (a) Wu, W. Q.; Peng, Q.; Dong, D. X.; Hou, X. L.; Wu, Y. D. J. Am.
Chem. Soc. 2008, 130, 9717. (b) Rubina, M.; Sherrill, W. M.; Rubin,
M. Organometallics 2008, 27, 6393.
(49) (a) Tietze, L. T.; Ila, H.; Bell, H. P. Chem. ReV. 2004, 104, 3453.
(b) Helmchen, G.; Kudos, S.; Sennhenn, P.; Steinhagen, H. Pure
Appl. Chem. 1997, 69, 513.
(50) Dai, W. M.; Wu, A.; Wu, H. Tetrahedron: Asymmetry 2002, 13,
2187.
(51) (a) Kunz, H.; Sager, W.; Pfrengle, W.; Schanzenbach, D. Tetrahedron
Lett. 1988, 29, 4397. (b) Yabu, K.; Masumoto, S.; Yamasaki, S.;
Hamashima, Y.; Kanai, M.; Du, W.; Curran, D. P.; Shibasaki, M.
J. Am. Chem. Soc. 2001, 123, 9908. (c) Kato, N.; Mita, T.; Kanai,
M.; Therrien, B.; Kawano, M.; Yamaguchi, K.; Danjo, H.; Sei, Y.;
Sato, A.; Furusho, S.; Shibasaki, M. J. Am. Chem. Soc. 2006, 128,
6768.
(52) (a) Kagan, H. B.; Riant, O. Chem. ReV. 1992, 92, 1007. (b) Pindur,
U.; Lutz, G.; Otto, C. Chem. ReV. 1993, 93, 741. (c) Deloux, L.;
Srebnik, M. Chem. ReV. 1993, 93, 763. (d) Desimoni, G.; Faita, G.;
Quadrelli, P. Chem. ReV. 2003, 103, 3119. (e) Rechavi, D.; Lemaire,
M. Chem. ReV. 2002, 102, 3467. (f) Chen, Y.; Yekta, S.; Yudin,
A. K. Chem. ReV. 2003, 103, 3155. (g) Desimoni, G.; Faita, G.;
Jøergensen, K. A. Chem. ReV. 2006, 106, 3561. (h) Kobayashi, S.,
Jøergensen, K. A., Eds. Cycloaddition Reactions in Organic Syn-
thesis; Wiley-VCH: Weinheim, Germany, 2001.
(53) (a) Corey, E. J.; Imai, N.; Zhang, H.-Y. J. Am. Chem. Soc. 1991,
113, 728. (b) Evans, D. A.; Woerpel, K. A.; Hinman, M. M.; Faul,
M. M. J. Am. Chem. Soc. 1991, 113, 726. (c) Desimoni, G.; Faita,
G.; Guala, M.; Laurenti, A.; Mella, M. Chem.sEur. J. 2005, 11,
3816. (d) Johnson, J. S.; Evans, D. A. Acc. Chem. Res. 2000, 33,
325.
(54) (a) Kobayashi, S.; Ishitani, H. J. Am. Chem. Soc. 1994, 116, 4083.
(b) Kobayashi, S.; Ishitani, H.; Hachiya, I.; Araki, M. Tetrahedron
1994, 50, 11623. (c) Ghosh, A. K.; Mathivanan, P.; Cappiello, J.
Tetrahedron Lett. 1996, 37, 3815.
(55) (a) Desimoni, G.; Faita, G.; Invernizzi, A. G.; Righetti, P. Tetrahedron
1997, 53, 7671. (b) Carbone, P.; Desimoni, G.; Faita, G.; Filippone,
S.; Righetti, P. P. Tetrahedron 1998, 54, 6099. (c) Desimoni, G.;
Faita, G.; Righetti, P. P. Tetrahedron Lett. 1996, 37, 3027.
(56) Goodrich, P.; Hardacle, C.; Paun, C.; Pârvulescu, V. I.; Podolean, I.
AdV. Synth. Catal. 2008, 350, 2473.
(57) (a) O’Leary, P.; Krosveld, N. P.; De Jong, K. P.; v. Koten, G.;
Gebbink, R. J. M. K. Tetrahedron Lett. 2004, 45, 3177. (b) Sibi,
M. P.; Matsunaga, H. Tetrahedron Lett. 2004, 45, 5925.
Chemical Reviews, 2010, Vol. 110, No. 3 1703
(58) Wang, H.; Liu, X.; Xia, H.; Liu, P.; Gao, J.; Ying, P.; Xiao, J.; Li,
C. Tetrahedron 2006, 62, 1025.
(59) (a) Evans, D. A.; Miller, S. J.; Lectka, T.; von Matt, P. J. Am. Chem.
Soc. 1999, 121, 7559. (b) Evans, D. A.; Miller, S. J.; Lectka, T. J. Am.
Chem. Soc. 1993, 115, 6460. (c) Evans, D. A.; Murry, J. A.; von
Matt, P.; Norcross, R. D.; Miller, S. J. Angew. Chem., Int. Ed. 1995,
34, 798.
(60) Gothelf, K. V.; Jøergensen, K. A. Chem. ReV. 1998, 98, 863.
(61) (a) Seerden, J. P. G.; Schotte op Reimer, A. W. A.; Scheeren, H. W.
Tetrahedron Lett. 1994, 35, 4419. (b) Seerden, J. P. G.; Kuypers,
M. M. M.; Scheeren, H. W. Tetrahedron: Asymmetry 1995, 6, 1441.
(c) Seerden, J. P. G.; Boeren, M. M. M.; Scheeren, H. W. Tetrahedron
1997, 53, 11843.
(62) (a) Gothelf, K. V.; Hazell, R. G.; Jøergensen, K. A. J. Org. Chem.
1998, 63, 5483. (b) Gothelf, K. V.; Hazell, R. G.; Jøergensen, K. A.
J. Org. Chem. 1996, 61, 346. (c) Crosignani, S.; Desimoni, G.; Faita,
G.; Filippone, S.; Mortoni, A.; Righetti, P. P.; Zema, M. Tetrahedron
Lett. 1999, 40, 7007. (d) Desimoni, G.; Faita, G.; Mella, M.; Boiocchi,
M. Eur. J. Org. Chem. 2005, 1020.
(63) Kawamura, M.; Kobayashi, S. Tetrahedron Lett. 1999, 40, 3213.
(64) Zeng, W.; Chen, G. Y.; Zhou, Y. G.; Li, Y. X. J. Am. Chem. Soc.
2007, 129, 750.
(65) (a) Noyori, R.; Hashiguchi, S. Acc. Chem. Res. 1997, 30, 97. (b)
Dalko, P. I.; Moisan, L. Angew. Chem., Int. Ed. 2004, 43, 5138. (c)
Taylor, M. S.; Jacobsen, E. N. Angew. Chem., Int. Ed. 2006, 45,
1520.
(66) Beliczey, J.; Giffels, G.; Kragl, U.; Wandrey, C. Tetrahedron:
Asymmetry 1997, 8, 1529.
(67) (a) Doyle, M. P.; Forbes, D. C. Chem. ReV. 1998, 98, 911. (b) Müller,
P.; Fruit, C. Chem. ReV. 2003, 103, 2905.
(68) (a) Fernandez, A. I.; Fraile, J. M.; Garcı́a, J. I.; Herrerı́as, C. I.;
Mayoral, J. A.; Salvatella, L. Catal. Commun. 2001, 2, 165. (b) Fraile,
J. M.; Garcı́a, J. I.; Martı́nez-Merino, V.; Mayoral, J. A.; Salvatella,
L. J. Am. Chem. Soc. 2001, 123, 7616. (c) Cornejo, A.; Fraile, J. M.;
Garcı́a, J. I.; Gil, M. J.; Herrerı́as, C. I.; Legarreta, G.; Martı́nez-
Merino, V.; Mayoral, J. A. J. Mol. Catal. A: Chem. 2003, 196, 101.
(d) Garcı́a, J. I.; López-Sánchez, B.; Mayoral, J. A.; Pires, E.; Villalba,
I. J. Catal. 2008, 258, 378. (e) Castillo, M. R.; Fousse, L.; Fraile,
J. M.; Garcı́a, J. I.; Mayoral, J. A. Chem.sEur. J. 2007, 13, 287.
(69) (a) Taylor, S.; Gullick, J.; McMorn, P.; Bethell, D.; Page, P. C. B.;
Hancock, F. E.; King, F.; Hutchings, G. J. J. Chem. Soc., Perkin
Trans. 2001, 2, 1714. (b) Taylor, S.; Gullick, J.; McMorn, P.; Bethell,
D.; Page, P. C. B.; Hancock, F. E.; King, F.; Hutchings, G. J. J. Chem.
Soc., Perkin Trans. 2001, 2, 1724. (c) Gullick, J.; Taylor, S.; Kerton,
O.; McMorn, P.; King, F.; Hancock, F. E.; Bethell, D.; Page, P. C. B.;
Hutchings, G. J. Catal. Lett. 2001, 75, 151.
(70) (a) Sibi, M. P.; Shay, J. J.; Liu, M.; Jasperse, C. P. J. Am. Chem.
Soc. 1998, 120, 6615. (b) Sibi, M. P.; Gorikunti, U.; Liu, M.
Tetrahedron 2002, 58, 8357.
(71) List, B. Chem. ReV. (Organocatalysis (Special Issue)) 2007, 107,
5413.
(72) Berkessel, A.; Gröger, H. Asymmetric Organocatalysis; Wiley-VCH:
Weinheim, Germany, 2005.
(73) New Frontiers in Asymmetric Catalysis; Mikami, K., Lautens, M.,
Eds.; Wiley: New York, 2007.
(74) EnantioselectiVe Organocatalysis; Dalko, P. I., Ed.; Wiley-VCH: New
York, 2007.
(75) (a) Asymmetric SynthesissThe Essentials; Christmann, M., Bräse,
S., Eds.; Wiley-VCH: New York, 2007; p 388. (b) Supramolecular
Catalysis; van Leeuwen, P. W. N. M., Ed.; Wiley-VCH: New York,
2008.
(76) (a) Szöllősi, Gy.; Bartók, M. Chirality 2001, 13, 614; (b) London,
G.; Szöllősi, Gy.; Bartók, M. J. Mol. Catal. A: Chem. 2007, 267,
98. (c) Szöllősi, Gy.; London, G.; Baláspiri, L.; Somlai, Cs.; Bartók,
M. Chirality 2003, 15, 90.
(77) (a) Hashimoto, T.; Maruoka, K. Chem. ReV. 2007, 107, 5656. (b)
Maruoka, K.; Ooi, T. Chem. ReV. 2003, 103, 3013. (c) O’Donnell,
M. J. Acc. Chem. Res. 2004, 37, 506. (d) Lygo, B.; Andrews, B. I.
Acc. Chem. Res. 2004, 37, 518. (e) Ooi, T.; Takeuchi, M.; Kameda,
M.; Maruoka, K. J. Am. Chem. Soc. 2000, 122, 5228.
(78) (a) Mazón, P.; Chinchilla, R.; Nájera, C.; Guillena, G.; Kreiter, R.;
Gebbink, R. J. M. K.; v. Koten, G. Tetrahedron: Asymmetry 2002,
13, 2181. (b) Chinchilla, R.; Mazón, P.; Nájera, C.; Ortega, F. J.
Tetrahedron: Asymmetry 2004, 15, 2603. (c) Chinchilla, R.; Mazón,
P.; Nájera, C. AdV. Synth. Catal. 2004, 346, 1186. (d) Thierry, B.;
Plaquevent, J. C.; Cahard, D. Tetrahedron: Asymmetry 2001, 12, 983.
(e) Thierry, B.; Plaquevent, J. C.; Cahard, D. Tetrahedron: Asymmetry
2003, 14, 1671. (f) Danelli, T.; Annunziata, R.; Benaglia, M.;
Cinquini, M.; Cozzi, F.; Tocco, G. Tetrahedron: Asymmetry 2003,
14, 461.
(79) (a) Denmark, S. E.; Fu, J. Chem. ReV 2003, 103, 2763. (b) Iseki, K.;
Mizuno, S.; Kuroki, Y.; Kobayashi, Y. Tetrahedron Lett. 1998, 39,
1704 Chemical Reviews, 2010, Vol. 110, No. 3
2767. (c) Iseki, K.; Mizuno, S.; Kuroki, Y.; Kobayashi, Y. Tetra-
hedron 1999, 55, 977.
(80) (a) Wilkinson, J. A.; Rossington, S. B.; Leonard, J.; Hussein, N.
Tetrahedron Lett. 2004, 45, 1191. (b) Beak, P.; Basu, A.; Gallagher,
D. J.; Park, Y. S.; Thayumanavan, S. Acc. Chem. Res. 1996, 29,
552. (c) Nakamura, S.; Nakawaga, R.; Watanabe, Y.; Toru, T. J. Am.
Chem. Soc. 2000, 122, 11340.
(81) (a) Modern Aldol Reactions; Mahrwald, R. E., Ed.; Wiley-VCH
VerlagGmbH & Co. KgaA: Weinheim, Germany, 2004; Vols. 1 and
2. (b) Mukherjee, S.; Yang, J. W.; Hoffmann, S.; List, B. Chem.
ReV. 2007, 107, 5471. (c) Davie, E. A. C.; Mennen, S. M.; Xu, Y.;
Miller, S. J. Chem. ReV. 2007, 107, 5759.
(82) (a) List, B.; Lerner, R. A.; Barbas, C. F. J. Am. Chem. Soc. 2000,
122, 2395. (b) Eder, U.; Sauer, G.; Wiechert, R. Angew. Chem., Int.
Ed. 1971, 10, 496. (c) Hajos, Z. G.; Parrish, D. R. J. Org. Chem.
1974, 39, 1615.
(83) (a) Doyle, A. G.; Jacobsen, E. N. Chem. ReV. 2007, 107, 5713. (b)
Denmark, S. E.; Stavenger, R. A. Acc. Chem. Res. 2000, 33, 432.
(c) Zhong, L.; Xiao, J.; Li, C. J. Catal. 2006, 243, 442.
(84) (a) Tang, Z.; Jiang, F.; Yu, L. T.; Cui, X.; Gong, L. Z.; Mi, A. Q.;
Jiang, Y. Z.; Wu, Y. D. J. Am. Chem. Soc. 2005, 127, 9285. (b)
Rodriguez, B.; Rantanen, T.; Bolm, C. Angew. Chem., Int. Ed. 2006,
45, 6924. (c) Revell, J. D.; Wennemers, H. Tetrahedron 2007, 63,
8420.
(85) (a) Li, X. J.; Zhang, G. W.; Wang, L.; Hua, M. Q.; Ma, J. A. Synlett
2008, 1255. (b) Corrêa, R. J.; Garden, S. J.; Angelici, G.; Toasini,
C. Eur. J. Org. Chem. 2008, 736.
(86) (a) Denmark, S. E.; Stavenger, R. A. J. Am. Chem. Soc. 2000, 122,
8837. (b) Denmark, S. E.; Su, X.; Nishigaichi, Y. J. Am. Chem. Soc.
1998, 120, 12990.
(87) (a) Corey, E. J.; Xu, F.; Noe, M. C. J. Am. Chem. Soc. 1997, 119,
12414. (b) Corey, E. J.; Zhang, F. Y. Angew. Chem., Int. Ed. 1999,
38, 1931. (c) Dere, R. T.; Pal, R. R.; Patil, P. S.; Salunkhe, M. M.
Tetrahedron Lett. 2003, 44, 5351.
(88) (a) Rai, V.; Mobin, S. M.; Namboothiri, I. N. N. Tetrahedron:
Asymmetry 2007, 18, 2719. (b) Rai, V.; Namboothiri, I. N. N.
Tetrahedron: Asymmetry 2008, 19, 767.
(89) (a) Luo, S.; Mi, X.; Zhang, L.; Liu, S.; Xu, H.; Cheng, J. P. Angew.
Chem., Int. Ed. 2006, 45, 3093. (b) Ni, B.; Zhang, Q.; Headley, A. D.
J. Org. Chem. 2006, 71, 9857. (c) Fukumoto, K.; Yoshizawa, M.;
Ohno, H. J. Am. Chem. Soc. 2005, 127, 2398.
(90) (a) Qian, Y.; Xiao, S.; Liu, L.; Wang, Y. Tetrahedron: Asymmetry
2008, 19, 1515. (b) Wang, J.; Li, H.; Zu, L.; Wang, W. AdV. Synth.
Catal. 2006, 348, 425.
(91) (a) Basavaiah, D.; Rao, A. J.; Satyanarayana, T. Chem. ReV. 2003,
103, 811. (b) Langer, P. Angew. Chem., Int. Ed. 2000, 39, 3049.
(92) Chen, S. H.; Hong, B. C.; Su, C. F.; Sarshar, S. Tetrahedron Lett.
2005, 46, 8899.
(93) (a) Wynberg, H.; Staring, E. G. J. J. Am. Chem. Soc. 1982, 104,
166. (b) Wynberg, H.; Staring, E. G. J. J. Chem. Soc., Chem.
Commun. 1984, 1181. (c) Wurz, R. P. Chem. ReV. 2007, 107, 5570.
(d) Gaunt, M. J.; Johansson, C. C. C. Chem. ReV. 2007, 107, 5596.
(94) (a) Cortez, G. S.; Tennyson, R. L.; Romo, D. J. Am. Chem. Soc.
2001, 123, 7945. (b) Cortez, G. S.; Oh, S. H.; Romo, D. Synthesis
2001, 1731.
(95) Palomo, C.; Aizpurua, J. M.; Ganboa, I.; Oiarbide, M. Curr. Med.
Chem. 2004, 11, 1837.
(96) (a) Hodous, B. L.; Fu, G. C. J. Am. Chem. Soc. 2002, 124, 1578. (b)
Lee, E. C.; Hodous, B. L.; Bergin, E.; Shih, C.; Fu, G. C. J. Am.
Chem. Soc. 2005, 127, 11586.
(97) (a) Taggi, A. E.; Hafez, A. M.; Wack, H.; Young, B.; Drury, W. J.,
III; Lectka, T. J. Am. Chem. Soc. 2000, 122, 7831. (b) France, S.;
Shah, M. H.; Weatherwax, A.; Wack, H.; Roth, J. P.; Lectka, T.
J. Am. Chem. Soc. 2005, 127, 1206.
(98) Lovick, H. M.; Michael, F. E. Tetrahedron Lett. 2009, 50, 1016.
(99) (a) Blaser, H. U.; Müller, M. Stud. Surf. Sci. Catal. 1991, 59, 73
(Heterogeneous Catalysis and Fine Chemicals II). (b) Blaser, H. U.
Tetrahedron: Asymmetry 1991, 2, 843. (c) Jannes, G., Dubois, V.,
Eds. Chiral Reactions in Heterogeneous Catalysis; Plenum Press:
New York, 1995; (d) De Vos, D. E., Vankelecom, I. F. J., Jacobs,
P. A., Eds. Chiral Catalyst Immobilization and Recycling; Wiley-
VCH: Weinheim, Germany, 2000; (e) Gladysz, J. A. Chem. ReV.
(RecoVerable Catalyst and Reagents (Special Issue)) 2002, 102, 3215.
(f) Murzin, D. Y.; Maki-Arvela, P.; Toukoniitty, E.; Salmi, T. Catal.
ReV.sSci. Eng. 2005, 47, 175. (g) Baiker, A. Catal. Today 2005,
100, 159. (h) Hutchings, G. J. Annu. ReV. Mater. Res. 2005, 35, 143.
(i) Heitbaum, M.; Glorius, F.; Escher, I. Angew. Chem., Int. Ed. 2006,
45, 4732. (j) Ding, K., UozumiY., Eds. Handbook of Asymmetric
Heterogeneous Catalysis; Wiley-VCH: Weinheim, Germany, 2008.
(100) (a) Sabatier, P.; Senderens, J. R. Action du nickel sur l’éthène. C. r.
1897, 124, 1358. (b) Schwab, G. M.; Rudolph, L. Naturwissen-
schaften 1932, 20, 363. (c) Lipkin, D.; Stewart, T. D. J. Am. Chem.
Soc. 1939, 61, 3295.
Bartók
(101) (a) Ogston, A. G. Nature 1948, 162, 963. (b) Terent’ev, A. P.;
Klabunovskii, E. I.; Patrikeev, V. V. Dokl. Akad. Nauk SSSR 1950,
74, 947. (c) Akabori, G. M.; Sakurai, S.; Izumi, Y.; Fujii, Y. Nature
1956, 178, 323.
(102) (a) Izumi, Y.; Imaida, M.; Fukawa, H.; Akabori, S. Bull. Chem. Soc.
Jpn. 1963, 36, 155. (b) Tatsumi, S.; Imaida, M.; Fukuda, Y.; Izumi,
Y.; Akabori, S. Bull. Chem. Soc. Jpn. 1964, 37, 846. (c) Izumi, Y.
Angew. Chem., Int. Ed. Engl. 1971, 10, 871.
(103) (a) Orito, Y.; Imai, S.; Niwa, S. J. Chem. Soc. Jpn. 1979, 1118. (b)
Orito, Y.; Imai, S.; Niwa, S.; Hung, N. G. J. Synth. Org. Chem. 1979,
37, 173.
(104) Tai, A. Chem. Ind. Catalysis of Organic Reactions; Dekker: New
York, 2003; Vol. 89, pp 191-223.
(105) Blaser, H. U., Schmidt, E., Eds. In Asymmetric Catalysis on Industrial
Scale. Challenges, Approaches and Solutions; Wiley-VCH: New
York, 2004.
(106) (a) Sugimura, S.; Osawa, T.; Nakagawa, S.; Harada, T.; Tai, A. Stud.
Surf. Sci. Catal. 1996, 101, 231. (b) Nakagawa, S.; Sugimura, T.;
Tai, A. Chem. Lett. 1997, 859.
(107) Blaser, H. U.; Spindler, F.; Studer, M. Appl. Catal., A 2001, 221,
119.
(108) (a) Studer, M.; Burkhardt, S.; Blaser, H. U. Chem. Commun. 1999,
1727. (b) Schürch, M.; Künzle, N.; Mallat, T.; Baiker, A. J. Catal.
1998, 176, 569.
(109) (a) von Arx, M.; Mallat, T.; Baiker, A. Tetrahedron: Asymmetry 2001,
12, 3089. (b) von Arx, M.; Bürgi, T.; Mallat, T.; Baiker, A.
Chem.sEur. J. 2002, 1430. (c) Török, B.; Balázsik, K.; Török, M.;
Szöllösi, Gy.; Bartók, M. Ultrason. Sonochem. 2000, 7, 151. (d)
Török, B.; Felföldi, K.; Szakonyi, G.; Bartók, M. Ultrason. Sonochem.
1997, 4, 301.
(110) (a) Szöri, K.; Sutyinszki, M.; Felföldi, K.; Bartók, M. Appl. Catal.,
A 2002, 237, 275. (b) Sutyinszki, M.; Szöri, K.; Felföldi, K.; Bartók,
M. Catal. Commun. 2002, 3, 125. (c) Török, B.; Felföldi, K.;
Balázsik, K.; Bartók, M. Chem. Commun. 1999, 1725.
(111) (a) Fish, M. J.; Ollis, D. F. Catal. ReV. Sci. Eng. 1978, 18, 259. (b)
Izumi, Y. AdV. Catal. 1983, 32, 215. (c) Bartók, M.; Wittmann, Gy.
In Stereochemistry of Heterogeneous Metal Catalysis; BartókM., Ed.;
Wiley: Chichester, U.K., 1985; Chapter 11. (d) Tai, A.; Harada, T.
In Tailored Metal Catalysts; Iwasawa, Y., Ed.; Reidel: Dordrecht,
The Netherlands, 1986; pp 265-324. (e) Klabunovskii, E. I. Russ.
Chem. ReV. 1991, 60, 980.
(112) (a) Webb, G.; Wells, P. B. Catal. Today 1992, 12, 319. (b) Baiker,
A. J. Mol. Catal. A: Chem. 1997, 115, 473. (c) Blaser, H. U.; Jalett,
H. P.; Müller, M.; Studer, M. Catal. Today 1997, 37, 441. (d) Osawa,
T.; Harada, T.; Tai, A. Catal. Today 1997, 37, 465. (e) Baiker, A.;
Blaser, H. U. In Handbook of Heterogeneous Catalysis; Ertl, G.,
Knözinger, H., Weitkamp, J., Eds.; Wiley-VCH: Weinheim, Ger-
many, 1997; Vol. 5, p 2422.
(113) (a) Smith, G. V.; Notheisz, F. Heterogeneous Catalysis in Organic
Chemistry; Academic Press: New York, 1999. (b) Wells, P. B.; Wells,
R. P. K. In Chiral Catalyst Immobilization and Recycling; De Vos,
D. E., Vankelecom, I. F. J., Jacobs, P. A., Eds.; Wiley-VCH:
Weinheim, Germany, 2000; p 123. (c) Baiker, A. In Chiral Catalyst
Immobilization and Recycling; De Vos, D. E., Vankelecom, I. F. J.,
Jacobs, P. A., Eds.; Wiley-VCH: Weinheim, Germany, 2000; p 155.
(d) Tai, A.; Sugimura, T. In Chiral Catalyst Immobilization and
Recycling; De Vos, D. E., Vankelecom, I. F. J., Jacobs, P. A., Eds.;
Wiley-VCH: Weinheim, Germany, 2000; p 173. (e) Baiker, A. J.
Mol. Catal. A: Chem. 2000, 163, 205. (f) Osawa, T.; Harada, T.;
Takayasu, O. Top. Catal. 2000, 13, 155.
(114) (a) Studer, M.; Blaser, H. U.; Exner, C. AdV. Synth. Catal. 2003,
345, 45. (b) Burgi, T.; Baiker, A. Acc. Chem. Res. 2004, 37, 909.
(115) (a) Osawa, T.; Harada, T.; Takayasu, O. Curr. Org. Chem. 2006,
10, 1513. (b) Klabunovskii, E.; Smith, G. V.; Zsigmond, Á.
Heterogeneous EnantioselectiVe Hydrogenation; Springer: New York,
2006. (c) Blaser, H. U.; Studer, M. Acc. Chem. Res. 2007, 40, 1348.
(d) Zaera, F. J. Phys. Chem. C 2008, 112, 16196.
(116) Blaser, H. U.; Jalett, H. P.; Monti, D. M.; Reber, J. F.; Wehrly, J. T.
Stud. Surf. Sci. Catal. 1988, 41, 153.
(117) Hal, l T. J.; Johnston, P.; Vermeer, W. A. H.; Watson, S. R.; Wells,
P. B. Stud. Surf. Sci. Catal. 1996, 101, 221.
(118) Augustine, R. L.; Tanielyan, S. K.; Doyle, L. K. Tetrahedron:
Asymmetry 1993, 4, 1803.
(119) Nitta, Y.; Shibata, A. Chem. Lett. 1998, 161.
(120) Collier, P. J.; Hall, T. J.; Iggo, J. A.; Johnston, P.; Slipszenko, J. A.;
Wells, P. B.; Whyman, R. Chem. Commun. 1998, 1451.
(121) Blaser, H. U.; Jalett, H. P.; Lottenbach, W.; Studer, M. J. Am. Chem.
Soc. 2000, 122, 12675.
(122) von Arx, M.; Mallat, T.; Baiker, A. Angew. Chem., Int. Ed. 2001,
40, 2302.
(123) Bartók, M.; Sutyinszki, M.; Felföldi, K.; Szöllősi, Gy. Chem.
Commun. 2002, 1130.
Unexpected Inversions in Asymmetric Reactions
(124) (a) Jones, T. E.; Urquhart, M. E.; Baddeley, C. J. Surf. Sci. 2005,
587, 69. (b) Jones, T. E.; Baddeley, C. J. Langmuir 2006, 22, 148.
(125) Szöllősi, Gy.; Somlai, Cs.; Szabó, P. T.; Bartók, M. J. Mol. Catal.
A: Chem. 2001, 170, 165.
(126) (a) Bartók, M.; Sutyinszki, M.; Bucsi, I.; Felföldi, K.; Szöllősi, Gy.;
Bartha, F.; Bartók, T. J. Catal. 2005, 231, 33. (b) Bartók, M.;
Sutyinszki, M.; Balázsik, K.; Szöllősi, Gy. Catal. Lett. 2005, 100,
161. (c) Bartók, M.; Sutyinszki, M.; Felföldi, K. J. Catal. 2003, 220,
207.
(127) Hoxha, F.; Mallat, T.; Baiker, A. J. Catal. 2007, 248, 11.
(128) Diezi, S.; Mallat, T.; Szabo, A.; Baiker, A. J. Catal. 2004, 228, 162.
(129) Jenkins, R. L.; Dummer, N.; Li, X. B.; Bawaked, S. M.; McMorn,
P.; Wells, R. P. K.; Burrows, A.; Kiely, C. J.; Hutchings, G. J. Catal.
Lett. 2006, 110, 135.
(130) Dummer, N. F.; Jenkins, R.; Li, X. B.; Bawaked, S. M.; McMorn,
P.; Burrows, A.; Kiely, C. J.; Wells, R. P. K.; Willock, D. J.;
Hutchings, G. J. J. Catal. 2006, 243, 165.
(131) Cserényi, Sz.; Felföldi, K.; Balázsik, K.; Szöllősi, Gy.; Bucsi, I.;
Bartók, M. J. Mol. Catal. A: Chem. 2006, 247, 108.
(132) Hoxha, F.; Konigsmann, L.; Vargas, A.; Ferri, D.; Mallat, T.; Baiker,
A. J. Am. Chem. Soc. 2007, 129, 10582.
(133) Gao, F.; Chen, L.; Garland, M. J. Catal. 2006, 238, 402.
(134) (a) Bartók, M.; Balázsik, K.; Notheisz, F. React. Kinet. Catal. Lett.
2002, 77, 363. (b) Bartók, M.; Balázsik, K.; Bartók, T.; Kele, Z.
Catal. Lett. 2003, 87, 235. (c) Szőri, K.; Balázsik, K.; Felföldi, K.;
Bartók, M. J. Catal. 2006, 241, 149.
(135) (a) Exner, C.; Pfaltz, A.; Studer, M.; Blaser, H. U. AdV. Synth. Catal.
2003, 345, 1253. (b) Balázsik, K.; Bucsi, I.; Cserényi, Sz.; Szöllősi,
Gy.; Bartók, M. J. Mol. Catal. A: Chem. 2008, 285, 84.
(136) Diezi, S.; Reimann, S.; Bonalumi, N.; Mallat, T.; Baiker, A. J. Catal.
2006, 239, 255.
(137) Diezi, S.; Szabó, A.; Mallat, T.; Baiker, A. Tetrahedron: Asymmetry
2003, 14, 2573.
(138) Maris, M.; Mallat, T.; Baiker, A. J. Mol. Catal. A: Chem. 2005, 242,
151.
(139) Bartók, M.; Balázsik, K.; Bucsi, I.; Szöllősi, Gy. J. Catal. 2006, 239,
74.
(140) Bonalumi, N.; Vargas, A.; Ferri, D.; Baiker, A. J. Phys. Chem. C
2007, 111, 9349.
(141) Bonalumi, N.; Vargas, A.; Ferri, D.; Burgi, T.; Mallat, T.; Baiker,
A. J. Am. Chem. Soc. 2005, 127, 8467.
(142) Braje, W.; Frackenpohl, J.; Langer, P.; Hoffmann, H. M. R.
Tetrahedron 1998, 54, 3495.
(143) Thiel, J.; Katrusiak, A.; Fiedorow, P. J. Mol. Struct. 2001, 561, 131.
(144) (a) Bartók, M.; Felföldi, K.; Török, B.; Bartók, T. Chem. Commun.
1998, 2605. (b) Bartók, M.; Felföldi, K.; Szöllösi, Gy.; Bartók, T.
Catal. Lett. 1999, 61, 1. (c) Balázsik, K.; Bartók, M. J. Catal. 2004,
224, 463.
(145) (a) Martinek, T. A.; Varga, T.; Fülöp, F.; Bartók, M. J. Catal. 2007,
246, 266. (b) Martinek, T. A.; Varga, T.; Balázsik, K.; Szöllősi, Gy.;
Fülöp, F.; Bartók, M. J. Catal. 2008, 255, 296. (c) Bartók, M.;
Felföldi, K.; Szöllösi, Gy.; Bartók, T. React. Kinet. Catal. Lett. 1999,
68, 371.
(146) (a) Toukoniitty, E.; Busygin, I.; Leino, R.; Murzin, D. Yu. J. Catal.
2004, 227, 210. (b) Busygin, I.; Toukoniitty, E.; Leino, R.; Murzin,
D. Y. J. Mol. Catal. A: Chem. 2005, 236, 227. (c) Busygin, I.; Wärnå,
J.; Toukoniitty, E.; Murzin, D. Y.; Leino, R. J. Catal. 2008, 254,
339.
(147) (a) Hess, R.; Diezi, S.; Mallat, T.; Baiker, A. Tetrahedron: Asymmetry
2004, 15, 251. (b) Diezi, S.; Ferri, D.; Vargas, A.; Mallat, T.; Baiker,
A. J. Am. Chem. Soc. 2006, 128, 4048.
(148) (a) Sonderegger, O. J.; Ho, G. M. W.; Burgi, T.; Baiker, A. J. Mol.
Catal. A: Chem. 2005, 229, 19. (b) Sonderegger, O. J.; Ho, G. M. W.;
Burgi, T.; Baiker, A. J. Catal. 2005, 230, 499.
(149) von Arx, M.; Mallat, T.; Baiker, A. J. Catal. 2000, 193, 161.
(150) (a) Felföldi, K.; Varga, T.; Forgó, P.; Bartók, M. Catal. Lett. 2004,
97, 65. (b) Varga, T.; Felföldi, K.; Forgó, P.; Bartók, M. J. Mol.
Catal. A: Chem. 2004, 216, 181.
Chemical Reviews, 2010, Vol. 110, No. 3 1705
(151) (a) Vargas, A.; Hoxha, F.; Bonalumi, N.; Mallat, T.; Baiker, A. J.
Catal. 2006, 240, 203. (b) von Arx, M.; Mallat, T.; Baiker, A.
Tetrahedron: Asymmetry 2001, 12, 3089.
(152) Szőri, K.; Balázsik, K.; Cserényi, Sz.; Szöllősi, Gy.; Bartók, M. Appl.
Catal., A 2009, 362, 178.
(153) (a) Hess, R.; Vargas, A.; Mallat, T.; Bürgi, T.; Baiker, A. J. Catal.
2004, 222, 117. (b) Hess, R.; Krumeich, F.; Mallat, T.; Baiker, A. J.
Mol. Catal. A: Chem. 2004, 212, 205.
(154) Tungler, A.; Sı́pos, É.; Háda, V. Curr. Org. Chem 2006, 10, 1569.
(155) (a) Bartók, M.; Wittmann, Gy.; Göndös, Gy.; Smith, G. V. J. Org.
Chem. 1987, 52, 1139. (b) Wittmann, Gy.; Bartók, G. B.; Bartók,
M.; Smith, G. V. J. Mol. Catal. 1990, 60, 1.
(156) Tungler, A.; Kajtár, J.; Máthé, T.; Tóth, T.; Fogassy, E.; Petró, J.
Catal. Today 1989, 5, 159.
(157) Huck, W. R.; Burgi, T.; Mallat, T.; Baiker, A. J. Catal. 2003, 219,
41.
(158) Colston, N. J.; Wells, R. P. K.; Wells, P. B.; Hutchings, G. J. Catal.
Lett. 2005, 103, 117.
(159) Szöllősi, Gy.; Szabó, E.; Bartók, M. AdV. Synth. Catal. 2007, 349,
405.
(160) Huck, W. R.; Mallat, T.; Baiker, A. J. Catal. 2000, 193, 1.
(161) (a) Szöllősi, Gy.; Niwa, S. I.; Hanaoka, T. A.; Mizukami, F. J. Mol.
Catal. A: Chem. 2005, 230, 91. (b) Szöllősi, Gy.; Balázsik, K.; Bartók,
M. Appl. Catal., A 2007, 319, 193. (c) Szöllősi, Gy.; Szöri, K.; Bartǒk,
M. J. Catal. 2008, 256, 349.
(162) (a) Sakthivel, K.; Notz, W.; Bui, T.; Barbas, C. F. J. Am. Chem.
Soc. 2001, 123, 5260. (b) Calderón, F.; Fernandez, R.; Sánchez, F.;
Mayoralas, A. F. AdV. Synth. Catal. 2005, 347, 1395. (c) Gruttadauria,
M.; Riela, S.; Meo, P. L.; D’Anna, F.; Noto, R. Tetrahedron Lett.
2004, 45, 6113. (d) Bellis, E.; Kokotos, G. J. Mol. Catal. A: Chem.
2005, 241, 166.
(163) Benaglia, M.; Cinquini, M.; Cozzi, F.; Puglisi, A.; Celentano, G.
AdV. Synth. Catal. 2002, 344, 533.
(164) Altava, B.; Burguete, M. I.; Fraile, J. M.; Garcia, J. I.; Santiago,
V. I.; Mayoral, J. A.; Vicent, M. J. Angew. Chem., Int. Ed. 2000, 39,
1503.
(165) (a) Bianchin, C.; Burnaby, D. G.; Evans, J.; Frediani, P.; Meli, A.;
Oberhauser, W.; Psaro, R.; Sordelli, L.; Vizza, F. J. Am. Chem. Soc.
1999, 121, 5961. (b) de Rege, F. M.; Morita, D. K.; Ott, K. C.; Tumas,
W.; Broene, R. D. Chem. Commun. 2000, 1797. (c) Raja, R.; Thomas,
J. M.; Jones, M. D.; Johnson, B. F. G.; Vaughan, D. E. W. J. Am.
Chem. Soc. 2003, 125, 14982.
(166) (a) Jørgensen, K. A.; Johannsen, M.; Yao, S.; Audrain, H.; Thorhauge,
J. Acc. Chem. Res. 1999, 32, 605. (b) Jørgensen, K. A. Angew. Chem.,
Int. Ed. 2000, 39, 3558.
(167) (a) Wan, Y.; McMorn, P.; Hancock, F. F.; Hutchings, G. J. Catal.
Lett. 2003, 91, 145. (b) Ghosh, A. K.; Mathivanan, P.; Cappiello,
P.; Krishnan, J. Tetrahedron: Asymmetry 1996, 7, 2165.
(168) Langham, C.; Piaggio, P.; Bethell, D.; Lee, D. F.; McMorn, P.;
Bulman Page, P. C.; Willock, D. J.; Sly, C.; Hancock, F. E.; King,
F.; Hutchings, G. J. Chem. Commun. 1998, 1601.
(169) Bonalumi, N.; Bürgi, T.; Baiker, A. J. Am. Chem. Soc. 2003, 125,
13342.
(170) (a) Lavoie, S.; Laliberte, M. A.; McBreen, P. H. J. Am. Chem. Soc.
2003, 125, 15756. (b) Lavoie, S.; Laliberte, M. A.; McBreen, P. H.
Catal. Lett. 2004, 97, 111. (c) Bonello, J. M.; Williams, F. J.;
Lambert, R. M. J. Am. Chem. Soc. 2003, 125, 2723.
(171) (a) Nieminen; Taskinen, A.; Toukonitty, E.; Hotokka, M.; Murzin,
D. Y. J. Catal. 2006, 237, 131. (b) Meier, D. M.; Urakawa, A.; Turrà,
N.; Rüegger, H.; Baiker, A. J. Phys. Chem. A 2008, 112, 6150. (c)
Vargas, A.; Santarossa, G.; Iannuzzi, M.; Baiker, A. J. Phys. Chem.
C 2008, 112, 10200.
CR9002352
1706 Chem. Rev. 2010, 110, 1706–1745

Novel Syntheses of Bridge-Containing Organic Compounds

Wenyi Zhao
Shasun Pharma Solutions, Incorporated, 10 Knightsbridge Road, Pistcataway, New Jersey 08854

Received July 4, 2009

Contents 8. Important Classes of Compounds 1732

8.1. Reactions in Synthesis of FR66979 and 1732
1. Introduction 1706 FR900482
2. [4 + 2] Cycloaddition Reactions 1707 8.2. Reactions in Synthesis of Platensimycin and 1733
2.1. Intermolecular [4 + 2] Cycloadditions 1707 Carbaplatensimycin
2.1.1. Catalyst-Promoted Asymmetric Diels-Alder 1707 8.2.1. Synthesis of the Cyclohexane Ring System 1733
Reactions 8.2.2. Synthesis of the Cyclopentane Ring 1734
2.1.2. Hetero [4 + 2] Cycloadditions 1709 System
2.1.3. Other [4 + 2] Cycloaddition Reactions 1710 8.2.3. Synthesis of Tetrahydrofuran and 1736
2.2. Intramolecular [4 + 2] Cycloadditions 1710 Tetrahydropyran Ring Systems
2.2.1. Lewis-Acid-Catalyzed Diels-Alder 1710 8.3. Reactions in Synthesis of Cortistatin A 1738
Reactions 8.3.1. Synthesis of the C Ring System (Route I) 1739
2.2.2. Hetero [4 + 2] Cycloadditions 1711 8.3.2. Synthesis of the ABD Ring System (Route 1740
2.2.3. Other [4 + 2] Cycloaddition Reactions 1712 II)
3. [n + m] Cycloaddition Reactions 1713 8.3.3. Synthesis of the BC Ring System (Route 1741
3.1. [3 + 2] Dipolar Cycloadditions 1713 III)
3.1.1. 1,3-Dipolar Cycloaddition of Nitrones 1713 8.3.4. Synthesis of the BC Ring System (Route 1741
IV)
3.1.2. Metal-Catalyzed [3 + 2] Cycloadditions 1714
8.4. Reactions in the Synthesis of Welwistatin and 1741
3.2. [5 + 2] Cycloadditions 1714 Analogues
3.3. [8 + 2] Cycloadditions 1715 9. Concluding Remarks 1741
3.4. [4 + 3] Cycloadditions 1716 10. References 1741
3.5. [6 + 3] Cycloadditions 1717
3.6. [6 + 4] Cycloadditions 1717 1. Introduction
4. Molecular Rearrangements 1717
4.1. Ring Rearrangements 1717 Bridge-containing segments exist in numerous natural
products and active pharmaceutical ingredients (API) and
4.2. Schmidt Reactions 1719
intermediates, creating more structural complexity. In addi-
4.3. Pinacol Rearrangements 1720 tion, bridge-containing segments make the molecules struc-
4.4. Claisen/Diels-Alder Cascade Process 1720 turally less flexible compared to their homologous counter-
4.5. Polonovski Rearrangements 1720 parts. Accordingly, installation of the bridge segment(s) onto
4.6. Prins-Type Reactions 1720 the existing ring system makes it more challenging and
5. Free-Radical Reactions 1722 interesting for synthetic organic chemists, and much effort
5.1. Intramolecular Radical Cycloadditions 1722 has been made to develop new cyclization technologies.
5.2. Other Radical Reactions 1723 Among the variety of methods documented in the litera-
6. Transition-Metal-Catalyzed Cyclization Reactions 1724 ture, the Diels-Alder (DA) reaction1 is the most commonly
6.1. Gold-Catalyzed Reactions 1724 used method for the construction of carbo- and heterocyclic
6.2. Cu(II)-Catalyzed Cyclizations 1724 compounds. A cyclic diene is required if the intermolecular
6.3. Rhodium-Catalyzed Reactions 1724 DA reaction is involved in the synthesis of bridge-containing
organic compounds (Scheme 1, eq 1). The intramolecular
6.3.1. Rh(II)-Catalyzed Reactions 1724
DA (IMDA) reaction has been classified into two categories
6.3.2. Rh(I)-Catalyzed Reactions 1725 by Shea, type 1-IMDA (Scheme 1, eq 2) and type 2-IMDA
6.4. Palladium-Catalyzed Reactions 1725 (Scheme 1, eq 3).1f,2 The dienophile in the type 2-IMDA
7. Other Reactions 1726 reaction is joined at the 2 position of the diene. This type
7.1. Iodine-Promoted Cyclizations 1726 2-IMDA reaction is distinguished by the fact that a strained
7.2. Pictet-Spengler Cyclizations 1727 bridgehead double bond is formed in the cycloaddition step.
7.3. Nucleophilic Substitutions 1727 It is well known that DA reactions occur usually at elevated
7.4. Nucleophilic Additions 1729 temperature. However, in the presence of Lewis-acid catalyst
7.5. Aldol-Type Condensation Reactions 1730 some reactions can be carried out at relatively low temper-
7.6. Friedel-Crafts Cyclizations 1731 ature.3
7.7. Ring-Closing Metathesis 1731 The [n + m] dipolar cycloaddition reactions have been
found in many useful organic syntheses to access bridge-
10.1021/cr9002402  2010 American Chemical Society
Published on Web 10/15/2009
Bridge-Containing Organic Compounds
Wenyi Zhao was born in Shenyang, China, and received his B.S. and
M.S. degrees from Lanzhou University, and Ph.D. degree from Nanjing
University, China. He spent two years at the Institute of Chemistry, Chinese
Academy of Sciences, as Postdoctoral Associate. In 1995, he joined Dr.
Shine’s group at Texas Tech University, Lubbock, TX, as Senior Research
Associate. He started his industrial career at Roche Carolina in 2001 as
a process chemist. His research interests are focused on organic synthesis,
reaction mechanism, and development of chemical processes for large-
scale production of pharmaceutical intermediates and APIs. He has
developed a Mechanism-guided Process Development (MPD) Strategy
and applied it toward his process research and development.
containing organic compounds. The polarized unsaturated
bond(s) is allowed to create a 1,3-dipolar reaction center in
the ring-closure step. 1,3-Dipolar [3 + 2] cycloadditions can
be catalyzed by metal cations. Nitrones as 1,3-dipoles in [3
+ 2] cycloaddition reactions are capable of reacting with a
great variety of dipolarophiles giving rise to a vast array of
products. Besides [3 + 2] cycloadditions, [5 + 2], [8 + 2],
[4 + 3], [6 + 3], and [6 + 4] cycloadditions are also widely
employed to construct bridge-containing organic compounds.
Molecular rearrangement plays an important role in the
formation of bridged compounds. Intramolecular ring-open-
ing/closure process and ring fragmentation are commonly
used approaches to synthesize bridged cyclic organic com-
pounds. Utilizations of intramolecular Schmidt reaction,
Polonovski and Prins-type cyclization, and Pinacol rear-
rangement are proved to be versatile synthetic tools in the
preparation of bridge-containing organic molecules. Another
powerful method is the combination of Claisen rearrangement
with Diels-Alder reaction.
Free-radical reactions as one of the synthetic tools have
been frequently applied toward the total synthesis of natural
products containing bridge segments.
Transition-metal-catalyzed4 annulation reaction is of in-
creasing importance in synthetic organic chemistry for the
construction of complex organic compounds including
bridged-ring systems. Compared to the traditional organic
transformations, most of the transition-metal-catalyzed trans-
formations go through sequential processes involving more
Scheme 1
Chemical Reviews, 2010, Vol. 110, No. 3 1707
Scheme 2
than one reactive species. This makes it difficult in both
elucidation of the reaction mechanisms and design of new
reactions.
This review will address the applications of the most
recently developed asymmetric D-A reactions (section 2),
[n + m] cycloaddition reactions (section 3), molecular
rearrangements (section 4), free-radical reactions (section 5),
and transition-metal-catalyzed cyclization reactions (section
6) toward the syntheses of APIs and natural products with
bridge-containing ring systems. Furthermore, this review will
also cover other newly developed methods such as halogen-
promoted cyclizations, nucleophilic additions and substitu-
tions, condensation reactions, etc. (section 7). The literature
from 2000 to 2009 will be included in this review.
2. [4 + 2] Cycloaddition Reactions
The Diels-Alder cycloaddition reaction is one of the most
fundamental synthetic transformations and has been the
cornerstone in countless total syntheses.5 In particular, the
asymmetric catalytic DA reaction has recently received
unprecedented attention, presumably due to the ability to
rapidly provide complex enantiopure carbocycles from
simple substrates.6 The bridged bicyclic compounds can be
prepared via either intermolecular or intramolecular DA
reactions. In the intermolecular cycloaddition, cyclic diene
is required to build a bridged bicyclic product wherein the
endo cycloadduct usually predominates. When a molecule
contains both diene and dienophile functionalities, an in-
tramolecular Diels-Alder reaction can take place given that
the tethered chain is in the right length to afford bridge-
containing organic compounds.
2.1. Intermolecular [4 + 2] Cycloadditions
2.1.1. Catalyst-Promoted Asymmetric Diels-Alder
Reactions
Lewis-acid-catalyzed asymmetric DA reactions have at-
tracted great attention due to their ability to construct
complex carbocyclic frameworks in an enantiomerically
enriched form.7 Studies on the unsaturated aldehydes,8
bidentate alkenoyl-oxazolidinones,9 and R,β-unsaturated
amide bearing chiral auxiliaries10 as dienophiles have been
reported. Corey’s research group has pioneered enantiose-
lective DA reactions catalyzed by chiral Lewis acids and
demonstrated their applications to the synthesis of complex
molecules. A great review written by Corey addressed these
Lewis-acid-catalyzed asymmetric [4 + 2] cycloaddition
reactions.11
Activation of R,β-unsaturated ketone (A) by Lewis acid
or by forming iminium ion (B) is illustrated in Scheme 2.
Due to the fact that the equilibrium dynamics and π-orbital
electronics of iminium (B) are inherent to the complex of
ketone-Lewis acid (A), MacMillan revealed an attractive
prospect that chiral amines might function as enantioselective
catalysts for a range of transformations that traditionally
utilize metal salts.
As a result, an enantioselective DA reaction of R,β-
unsaturated aldehydes with cyclopenta-1,3-diene was devel-
1708 Chemical Reviews, 2010, Vol. 110, No. 3 Zhao

Scheme 3 Scheme 5

Scheme 4

Scheme 6

oped.12 The use of 5 mol % of chiral imidazolidin-4-one

hydrochloride salt afforded an exo/endo mixture (1:1) of
bicyclo[2.2.1]hept-5-ene-2-carbaldehyde derivatives 3 and 4
with good enantioselectivity (exo > 84% ee, endo > 90%
ee) in high yields (75-99%) (Scheme 3).12 The R group can
vary from alkyl, such as methyl, propyl, and isopropyl, to
aryl, such as phenyl and furyl.
This methodology was further extended to R,β-unsaturated
ketones 5, and the first efficient asymmetric Diels-Alder
reaction using chiral amine catalyst was reported by Mac-
Millan (Scheme 4).13 All ketones except isopropyl-substituted
ketone (Scheme 4, entry 5) led to high yields with good
enantioselectivity.
Analogous to chiral imidazolin-4-one derivatives as DA
cycloaddition catalyst, Ishihara demonstrated that chiral
primary amines 9 could activate R,β-unsaturated aldehydes
7 for highly enantioselective DA reactions (Scheme 5).14 In
contrast to the results reported by MacMillan,13 the asym-
metric DA cycloaddition catalyzed by 9a or 9b afforded exo-
bicyclo[2.2.1] adducts 8 as major products. Moreover, when 11 (Scheme 6, eq 1). Cleavage of the lactone bridge 12
the reaction was conducted in the presence of 10 mol % of provides a useful asymmetric route to a wide range of cyclic
water, both the yield and the enantioselectivity were in- chiral building block, such as chiral ketone 15 (Scheme 6,
creased. Water might promote the hydrolysis of the aldimine eq 2), amenable for further synthetic elaborations.
intermediate to increase the catalytic turnover rate. A catalytic enantioselective [4 + 2] cycloaddition of cyclic
Following the work done by MacMillan13 and Ishihara,14 electron-rich dienes 2 or 16 with electron-deficient propi-
Deng15 and co-workers developed an enantioselective DA olamide derivatives 17 was achieved by Ishihara (Scheme
reaction by applying cinchona alkaloid-derived chiral cata- 7).16 As expected, a temperature effect was noticed for a
lysts toward the [4 + 2] cycloaddition. By screening various reaction of cyclopentadiene 2 with 17a and a bridged
cinchona alkaloids, Deng’s group uncovered that among six cycloadduct 18 was obtained with 93% ee in 50% yield at
cinchona alkaloids 9-amino QD-14 gave the best results for -78 °C and with 88% ee in 91% yield at -40 °C,
the DA reaction of 2-pyrone 10 with R,β-unsaturated ketones respectively.
Bridge-Containing Organic Compounds
Scheme 7
Scheme 8
2.1.2. Hetero [4 + 2] Cycloadditions
The equilibrium of valence isomerism between cyclohep-
tatriene I and norcaradiene II is influenced by the substituents
(R) at C7. An electron-withdrawing R group tends to shift
the equilibrium toward the norcaradiene II, whereas a
π-electron-donating group favors cycloheptatriene (Scheme
8, eq 1). Although the equilibrium has been studied on
cycloheptatrienes having various electronically differentiated
substituents,17 very little is known about the effect of silyl-
substituted cycloheptatrienes (I, R ) SiR3) and their silyl-
methyl homologues (I, R ) CH2SiR3).18 Nevertheless, a [4
+ 2] cycloaddition of norcaradiene II in the equilibrium of
19 with nitroso dienophiles generated in situ by oxidation19
of hydroxylamine with tetrabutylammonium periodate (n-
Chemical Reviews, 2010, Vol. 110, No. 3 1709
Scheme 9
Scheme 10
Bu4NIO4) led to bridged products 20 in good yields (Scheme
8, eq 2).20 The resulting bridged compounds 20 could be
converted into valuable synthetic intermediates 21 in good
yields (Scheme 8, eq 3). Formation of the vinyl group
probably occurs through the ring opening of the cyclopropane
moiety of 20a,b, with concomitant desilylation.
The structure-inherent instability of the heterobicyclo
[2.2.n] ring systems (n ) 1 or 2) is utilized by Vassiliko-
giannakis’s group21 to cultivate a versatile and general
method for the synthesis of bis-spiroketal. Bis-spiroketal
functionality exists in a wide range of interesting natural
products, such as spirolides,22 pinnatoxins,23 pteriatoxins,24
and the azaspiracids.148,25 As depicted in Scheme 9, a tandem
sequence began with a [4 + 2] cycloaddition reaction, leading
to a transient bridged intermediate 23.26 The subsequent
reactions including SN2-type ring opening, peroxide 24
reduction, and cyclization afforded the desired bis-spiroketal
26.
Further utilization of the attractive singlet-oxygen-mediated
tandem process to synthesize tetrahydrofuran-3-one 32 was
demonstrated by the same group in Scheme 10.27 Obviously,
an intramolecular nucleophilic ring opening of cycloadduct
ozonide 28 is unlikely because of an unfavorable 4-exo-
cyclization compared to the 6-exo-cyclization of 23 (Scheme
9). However, an intermolecular nucleophilic ring opening
was effected when the reaction was conducted in methanol,
resulting in a mixture of products 29 and 30. Without
1710 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 11
Scheme 12
isolation, exposure of the mixture of 29 and 30 to dimethyl
sulfide (DMS) followed by treatment with a catalytic amount
of p-TsOH furnished 32 in 53% yield.
In the absence of nucleophiles, a bicyclo[2.2.2] ring
system, such as 34 (Scheme 11, eq 1),28 could be isolated in
85% yield as colorless crystals. Surprisingly, the resulting
bridged bicyclic endoperoxide 34 is fairly stable with a
melting point of 123-126 °C. When a [4 + 2] reaction of
tridachiahydropyrone 35 was carried out in the presence of
0.2 equiv of methylene blue, complete conversion into the
desired product, oxytridichiahydro-pyrone 36 (Scheme 11,
eq 2),29 was realized in quantitative yield.
The isolated endoperoxide 34 (Scheme 11, eq 1) was in
turn selectively reduced with thiourea under very mild
conditions followed by acetylation in pyridine, affording the
diacetate 37 in 87% yield (Scheme 12).28 In the event,
stereoselective synthesis of bishomo-allo-inositol and bishomo-
chiro-inositol was realized in four and three additional steps,
respectively.
2.1.3. Other [4 + 2] Cycloaddition Reactions
2,5-Disubstituted cyclopentadiene 38 is highly competent
in Diels-Alder cycloadditions with a wide variety of
dienophiles under very mild reaction conditions. Gleanson
and co-workers employed a room-temperature Diels-Alder
cycloaddition of the substituted cyclopentadiene 38 toward
the synthesis of the bridged [4 + 2] cycloadduct 42, a
potential intermediate for construction of the E ring in
palau’amine-marine alkaloid (Scheme 13).30
A recent report by Buszek and co-workers describes a
useful intermolecular DA cycloaddition of indole arynes. It
was found that while 4,5- and 5,6-indolynes show virtually
no regioselectivity in Diels-Alder cycloadditions, the 6,7-
indolyne by contrast is highly regioselective, presumably due
to the polarization influence of the pyrrole ring. This
regioselective DA cycloaddition of 6,7-indolyne, being
Zhao
Scheme 13
Scheme 14
formed in situ from dibromide 45 in the presence of
n-butyllithium, furnished a mixture of ether-bridged com-
pounds 46 and 47 (Scheme 14).31
2.2. Intramolecular [4 + 2] Cycloadditions
2.2.1. Lewis-Acid-Catalyzed Diels-Alder Reactions
Lewis-acid-catalyzed intramolecular Diels-Alder (IMDA)
reactions usually proceed under mild conditions with high
reaction rates, resulting in good selectivity owing to the
reduction in the degrees of freedom of the transition. Type
2-IMDA of 48 wherein the diene and dienophile are
separated by a cyclohexene moiety occurred to afford taxane
skeleton 49 (Scheme 15).32
A combination of intramolecular Diels-Alder cycload-
dition with intermolecular Diels-Alder cycloaddition, de-
veloped by Baran’s group,33 constitutes a powerful tool to
access a key tricyclic carbon skeleton 56 (Scheme 17) in
the total synthesis of vinigrol.34 A Lewis-acid-catalyzed
intermolecular Diels-Alder cycloaddition of cyclohexadiene
50 with R,β-unsaturated ester 51 afforded bicyclic ketone
52. Without isolation, a DA reaction of alkoxide 53 at 105
Bridge-Containing Organic Compounds
Scheme 15
Scheme 16
Scheme 17
°C gave 54, whose reaction with TBAF furnished the
tetracycle 55 in 75% overall yield (Scheme 16).
Grob fragmentations are generally performed under nu-
cleophilic or basic conditions, albeit some Grob fragmenta-
tions are conducted under certain acid conditions.35 The
fragmentation process may proceed through a concerted or
stepwise pathway, depending on stereochemical and stereo-
electronic factors.36 Notably, tetracyclic intermediate 55 is
unreactive toward Grob fragmentation, presumably due to
the lack of correct antiperiplanar atomic arrangement.
Eventually, a Grob fragmentation of its diastereoisomer 58,
being inverted from 55, was realized to furnish the desired
bridged 56, Scheme 17.
2.2.2. Hetero [4 + 2] Cycloadditions
The Diels-Alder reaction is a powerful method for
creation of complex polycyclic structures containing a
Chemical Reviews, 2010, Vol. 110, No. 3 1711
Scheme 18
number of stereogenic centers in a single step.5c The bridged
adduct 61 (or 62) features vicinal quaternary stereocenters
representing the correct relationship for the vicinal quaternary
stereogenic centers in the core of daphnilactone B (Scheme
18). A nitro-substituted olefin 60 with a pendant diene
functional group underwent heterointramolecular [4 + 2]
cycloaddition reactions in the presence of Lewis acid SnCl4,
leading to cycloadducts 61 and 62 (Scheme 18).37
The type-2 IMDA cycloaddition of heteroatom variants
provides a general route toward heterocycles with alkene
bridgeheads.1f,2,38 Under oxidative conditions, hydroxy-
lamines or hydrazides 63 were converted in situ into
dienophiles, N-acylnitroso-64 and N-acylazo-66, whose in-
tramolecular DA cycloadditions provide an efficient method
for the preparation of bridged bicyclic heterocycles (Scheme
19).39 A condition of using tetraalkylammonium (meta)pe-
riodate (Et4NIO4 or n-Bu4NIO4) as oxidant was adopted as
a general method to convert hydroxylamines or hydrazides
63 into the corresponding N-acylnitroso-64 or N-acylazo-66
dienophiles (Scheme 19).
Employing a protocol described by Evans and co-work-
ers,40 N-acylazo 66b was isolated in 96% yield (Scheme 20,
eq 1), which was used to examine the subsequent IMDA
under both thermal and Lewis-acid-catalyzed conditions.
N-Acylazo dienophile 66b was found to be unstable at
temperatures >40 °C in benzene, and cycloadduct 67b was
not observed. The best results were obtained at 40 °C,
producing cycloadduct 67b in only 58% yield. Lewis-acid
catalyst allows a DA reaction to occur at mild conditions,
which circumvents the decomposition of labile organic
compounds. Thus, in the presence of 10 mol % of ZnCl2, a
DA cycloaddition of 66b led to cycloadduct 67b in 78%
yield (Scheme 20, eq 2).39b
A complex mixture was obtained when application of the
oxidative cycloaddition condition (Et4NIO4, CHCl3, 0 °C)
to 63e due to other competing processes. This obstacle was
overcome by a thermal generation of the N-acylnitroso
species from a 9,10-dimethylanthracene adduct 69. In the
event, 1,3-adduct 70 along with its isomeric 1,4-adduct 71
was obtained in 60% and in a 1:1 ratio (Scheme 21, eq 1).39a
R-Substituted hydroxylamines 72 underwent regioselective
IMDA reaction, and 1,3-adducts 73a and 73b were obtained
1712 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 19
Scheme 20
Scheme 21
as single diastereomers in 83% and 70% yields, respectively
(Scheme 21, eq 2).39a
Furthermore, stereoselective IMDA reactions were ob-
served for R-substituted N-acylazo dienophiles 74 and 77
under either thermal (for 75) or Lewis-acid-catalyzed (for
78) conditions, affording a single diastereomer 76 or 79, in
yields of 91% and 71%, respectively (Scheme 22, eqs 1 and
2).39b Hydrogenation and successive reductive N-N bond
Zhao
Scheme 22
cleavage of the two bridged cycloadducts 76 and 79 afforded
cis-3,6-disubstituted caprolactam 81 and cis-3,7-disubstituted
enantholactam 83, respectively (Scheme 22, eqs 3 and 4).39b
This sequence of transformations offers a convenient route
for the diastereoselective synthesis of medium-ring nitrogen
heterocycles.
Recently, Nicolaou’s group revealed an intramolecular [4
+ 2] cycloaddition reaction in the synthesis of the sporolide
ring framework 85 (Scheme 23).41 This IMDA reaction of
o-quinone 84 occurred in toluene at 120 °C, affording the
desired cycloadduct 85 in 50% yield.
2.2.3. Other [4 + 2] Cycloaddition Reactions
A report from Johansson’s group uncovered a construction
of a transtaganolide carbon skeleton through an intramo-
lecular DA cycloaddition under very mild condition (Scheme
24).42 It was found that under aqueous basic conditions at
50 °C a bridged lactone 87 was obtained in 61% yield with
a dr ratio of 2:1, while heating the carboxylic acid 86 in
Bridge-Containing Organic Compounds
Scheme 23
Scheme 24
Scheme 25
toluene at 120 °C or in benzene at 80 °C resulted in
decarboxylation.
11-Acetoxy-4-deoxyasbestinin D43 and asbestinin-1244
have attracted much attention by synthetic chemists due to
their intriguing biological properties and unique chemical
structures. Both molecules have a fully substituted tetrahy-
drofuran moiety with 9 contiguous stereocenters in 11-
acetoxy-4-deoxyasbestinin D and 10 in asbestinin-12, re-
spectively. The enantioselective total synthesis of 11-acetoxy-
4-deoxyasbestinin D and asbestinin-12 involves a common
ketone intermediate 91 wherein the C2 and C9 are linked
together by an oxygen bridge. The synthesis of the ketone
intermediate 91 is outlined in Scheme 25. Wittig reaction of
aldehyde 88 led to 89, which triggered a rapid in-situ IMDA
cycloaddition, furnishing tricyclic compound 90. The conver-
sion of 90 to the desired ketone 91 was effected in three
steps (Scheme 25).45
Chemical Reviews, 2010, Vol. 110, No. 3 1713
Scheme 26
3. [n + m] Cycloaddition Reactions
3.1. [3 + 2] Dipolar Cycloadditions
The applications of dipolar cycloaddition reactions have
been found in many useful organic syntheses, particularly
with respect to the preparation of complex molecules
including natural products with chiral centers.
3.1.1. 1,3-Dipolar Cycloaddition of Nitrones
Nitrones have proved to be very useful intermediates in
the construction of structurally complex molecules.46 Nitrones
as 1,3-dipoles in cycloaddition reactions are capable of
reacting with a great variety of dipolarophiles, giving rise
to a vast array of products. Owing to their high reactivity,
nitrones are usually generated in situ.
1,3-Dipolar cycloaddition of nitrone to olefin has been used
in the synthesis of five- and six-membered nitrogen-contain-
ing heterocycles.47 Palma and et al.48 described an intramo-
lecular 1,3-dipolar cycloaddition reaction of nitrones 93 to
access bridged compounds 94 (Scheme 26). Nitrones 93 were
generated in situ from oxidation of ortho-allyl-N-benzyla-
nilines 92 according to a methodology described by Mura-
hashi.49
Despite the oxidative method, nitrone 98 was generated
from intermolecular nucleophilic addition of oxime 95 and
used in situ toward intramolecular [3 + 2] dipolar cycload-
dition to access an oxygen-bridged cycloadduct 97, a key
intermediate in the total synthesis of cylindricine C (Scheme
27).50
Nitrones can also be generated thermally and used in situ
as dipoles. Scheme 2851 demonstrates a [3 + 2] cycloaddition
of nitrone intermediate 101 in the synthesis of a bridged
tricyclic nitrile 100, an intermediate in the total synthesis of
(-)-HTX 285A. The nitrone 101 was generated by a
microwave (MW) thermally induced 1,3-dipolar cyclorever-
sion of 99 with accompanying extrusion of styrene (Scheme
28).52
1,3-Dipolar cycloaddition of azomethine ylides with alk-
enes or alkynes is a very effective method for the construc-
tion of N-containing heterocycles.53 Due to their high
reactivity, azomethine ylides are usually generated in situ.
Recently reported by Pandey and co-workers, azomethine
ylide was employed in the [3 + 2] cycloaddition reaction to
access challenging vicinal quaternary and tertiary stereo-
centers of the 5,10b-ethanophenanthridine skeleton 103 of
maritidine. This nonstabilized azomethine ylide 104 was
generated in situ by removing a TMS group (Scheme 29).54
1714 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 27
Scheme 28
Scheme 29
3.1.2. Metal-Catalyzed [3 + 2] Cycloadditions
An approach for generation of R-carbonyl carbenoid 107
and subsequent formation of azomethine ylide equivalent 108
Zhao
Scheme 30
was reported by Shin et al. (Scheme 30, eq 1).55 Azabicyclo
[3.2.1]octanes 110 were synthesized by a gold(III)-catalyzed
cycloaddition of N-alkylnitrones 109 followed by [3 + 2]
dipolar cycloaddition (Scheme 30, eq 2).55 Various substit-
uents on the enyne skeleton were well tolerated.
This gold(III)-catalyzed reaction is proposed to involve a
gold-carbenoid intermediate 112,56 generated by Au(III)-
promoted 6-exo-dig cycloaddition of 109 (R1 ) Bn, R2 )
R3 ) R4 ) H) followed by retro-electrocyclization to cleave
the N-O bond (Scheme 31). An intramolecular addition of
imine to this carbenoid leads to 113 (or 114). Ultimately, a
dipolar cycloaddition leads to the observed bicyclo[3.2.1]octane
product 110.
Analogously, Pt-mediated [3 + 2] cycloaddition of enynal
115 (or 117) was developed to access ether-bridged multiring
system 116 (or 118) (Scheme 32, eqs 1 and 2).57
A possible reaction mechanism was proposed and is
illustrated in Scheme 33.57 A Huisgen-type [3 + 2] cycload-
dition of Pt(II)-activated enynal 115 (or 117) occurs to yield
a tetracyclic platinum-carbene complex 120 via platinum-
pyrylium intermediates 119. The platinum-carbene 120 is
trapped by benzylic C-H bond insertion58 to furnish the
product 116 (or 118).
In addition to intramolecular [3 + 2] cycloaddition,
application of intermolecular [3 + 2] cycloaddition toward
a regioselective pyrrole synthesis was developed by Park and
co-workers (Scheme 34).59
Scheme 35 outlines the proposed reaction pathway which
involves a heterocyclic bridged intermediate 127, formed via
[3 + 2] cycloaddition, whose spontaneous decarboxylation
furnishes the product 124.
3.2. [5 + 2] Cycloadditions
Wender and et al. developed an intramolecular [5 + 2]
cycloaddition approach to synthesize a BC ring system of
1R-alkyldaphnanes 131 (Scheme 36).60
Harrowven designed a total synthesis of colombiasin A
and elisapterosin B which involves a common intermediate
132 (Scheme 37).61 Colombiasin A was obtained by in-
tramolecular DA cycloaddition followed by deprotection with
2 equivalents of diethyl ether-trifluoroborane. In contrast,
direct treatment of 132 with 2 equivalents of BF3 · OEt2
afforded the [5 + 2] cycloaddition product, elisapterosin B.
Bridge-Containing Organic Compounds Chemical Reviews, 2010, Vol. 110, No. 3 1715

Scheme 31

Scheme 32 Scheme 34

Scheme 35

Scheme 33

3.3. [8 + 2] Cycloadditions
The [8 + 2]-cycloaddition reaction is a potentially power-
ful method for the preparation of 10-membered ring systems,
although it is limited to geometrically constrained tetraenes
wherein the terminal atoms at positions 1 and 8 are rigidly
held in close proximity. Yu and et al. proposed two reaction
pathways for formation of the bridged 10-membered ring
systems:62 a stepwise pathway via formation of either a
diradical or a zwitterionic intermediate followed by ring Notably, the selectivity for the [8 + 2] cycloadduct 136
closure and a pathway which starts from a concerted, (except entry 4) is attributed to the planar orientation in 134
asynchronous [4 + 2] reaction and a diradical [1,5]-vinyl between the enol ether and isobenzofuran (Scheme 38).63
shift. Thus, these two pathways can compete with one The steric effect of TMS-substituted enol ether 134 (entry
another to furnish both [8 + 2] and [4 + 2] cycloadducts. 4), in contrast, features an orthogonal orientation between
1716 Chemical Reviews, 2010, Vol. 110, No. 3 Zhao

Scheme 36 Scheme 38

Scheme 39

Scheme 37

Scheme 40
the enol ether and isobenzofuran which would prevent [8 +
2] cycloaddition.

3.4. [4 + 3] Cycloadditions
The intramolecular [4 + 3] cycloaddition of allyl cations
with cyclic dienes such as furan and cyclopentadiene
provides stereochemically defined oxabicyclic and carbobi-
cyclic scaffolds where subsequent stereocontrolled ring-
opening reactions afford useful intermediates for the syn-
theses of natural products. The use of epoxy enol triethylsilanes
as oxyallyl cation precursors in the [4 + 3] cycloaddition
was developed by Chiu and co-workers64 to produce hy-
droxylated cycloheptanoids. Scheme 39 demonstrates the
application of this intramolecular [4 + 3] cycloaddition
reaction toward the synthesis of a bridged tricyclic ring
system with high diastereoselectivity (entries 1-6). Only in well, leading to an ether-bridged seven-membered ring
the case of the hindered substrate (entry 7) was the reaction system 143 (Scheme 40, eq 1).64 The sterically demanding
high yielding but unselective. 2,5-dimethylfuran 144 was reluctant toward undergoing
Furthermore, intermolecular [4 + 3] cycloaddition of the cycloaddition, giving diminished yields of 143 (R ) H)
epoxy enol silanes 141 with furan 142 performed equally (Scheme 40, eq 2).64
Bridge-Containing Organic Compounds
Scheme 41
Takeda and co-workers reported that a stereoselective
construction of 2-oxabicyclo[3.3.2]decenone derivatives 147
was effected by Brook rearrangement-mediated [4 + 3]
cycloaddition (Scheme 41).65 As the enolate of cyclohep-
tenone 146 is unstable at -80 °C, replacement of THF with
cyclopentyl methyl ether (CPME, mp < -140 °C) was able
to perform this cycloaddition at -98 °C. As a result, the
yields of cycloadducts 147 were improved.
Recently, a diastereoselective Brook rearrangement-medi-
ated [4 + 3] annulation was developed by the same group.
Applications of this [4 + 3] annulation toward a formal
synthesis of (+)-laurallene are illustrated in eqs 1-3 of
Scheme 42. A bridged ketone 150 was prepared by using [4
+ 3] annulation of enantiomerically enriched (-)-trans-149
(90% ee) with potassium enolate 148 of 2-cycloheptenone
(Scheme 42, eq 1).66 [4 + 3] annulation of lithium enolate
of 151, being formed in situ, with acryloylsilane 152 gave
bridged ketone 153 in 76% yield as a single isomer (Scheme
42, eq 2).67 When Davis’ oxaziridine 155 was added to a
THF solution of sodium enolate of 154 with an extra amount
of NaHMDS and 18-crown-6, R-hydroxy ketone 156 was
obtained in 64% yield (Scheme 42, eq 3).67
Harmata and co-workers employed intramolecular [4 +
3] cycloaddition to access a bridged cycloadduct 158 bearing
a bromo substituent at a bridgehead (Scheme 43).68
A [4 + 3] cycloaddition/quasi-Favorskii rearrangement
approach was introduced in the synthesis of (()-sterpurene
by the same group in which a mixture of trifluoroethanol
(TFE) and benzene was used as reaction solvents (Scheme
44).69 A tricyclic bridged ketone 161 was obtained via an
intermolecular [4 + 3] cycloaddition. A subsequent quasi-
Favorskii rearrangement followed by reduction with LAH
afforded tricyclic alcohol 162.
Furthermore, Harmata’s [4 + 3] cycloaddition/quasi-
Favorskii rearrangement approach was applied toward the
syntheses of the spatane ring system 167 (Scheme 45)70 and
spatol ring system 172 (Scheme 46).71
3.5. [6 + 3] Cycloadditions
A phosphine-catalyzed [6 + 3] annulation is a simple and
expedient method for constructing bridged carbocycles
(Scheme 47).72
The reaction mechanism is rationalized in Scheme 48.72
The formation of [6 + 3] cycloaddition product 179 is
presumably through two intermediates 177 and 178 being
generated via 1,6-conjugate additions of ylide 176 onto 174.
With the tert-butyl carbonate derivative 173 (X ) OBoc),
the reaction can be carried out in the absence of base because
of the in-situ-generated tert-butoxide anion.
Chemical Reviews, 2010, Vol. 110, No. 3 1717
An enantioselective palladium-catalyzed [6 + 3] cycload-
dition reaction of trimethylenemethane (Pd-TMM) with
cycloheptatrienone (tropones) was developed by Trost and
co-workers, which allowed direct access to asymmetric
substituted bicyclo[4.3.1] decadienes 182 (Scheme 49).73
3.6. [6 + 4] Cycloadditions
Ingenol offers several synthetic challenges including a
bicyclo[4.4.1]undecanone unit with the BC ring substructure
as well as a highly strained inside, outside intrabridgehead
stereochemical relationship at the BC ring juncture. A facile
entry into the ingenane core 184 using a Lewis-acid-catalyzed
intramolecular [6 + 4] cycloaddition of 183 has been
developed (Scheme 50, eq 1).74 The out,out cycloadduct 184
was obtained in 80% yield in low enantioselectivity (40%
ee). Furthermore, an expeditious construction of a highly
functionalized ABC core of ingenol via an intramolecular,
metal-free [6 + 4] cycloaddition of cycloheptatrienone and
furan was reported by the same group (Scheme 50, eq 2).75
Isomerization of the out,out stereoisomer 187 into the
highly strained inside-outside stereoisomer 189 was effected
by Pd-promoted rearrangement of an allylic epoxide followed
by 1,5-H migration (Scheme 51).76
Despite the popular intramolecular [6 + 4] cycloadditions,
an intermolecular [6 + 4] cycloaddition was observed when
heating a mixture of 8,8-dicyanoheptafulvene 190 with
electron-rich dienes 191, leading to 192 in high yields
(Scheme 52).77
4. Molecular Rearrangements
4.1. Ring Rearrangements
Development of a methodology for the synthesis of the
bicyclo[5.3.1]undecane ring system, a structurally important
feature of the taxane, is an attractive area for organic
chemists.78 The ring strain of a fused small-sized (3- and
4-membered) ring system can, under certain conditions, cause
ring fragmentation, resulting in a bridged bicyclic ring
system. The intramolecular photocycloaddition/fragmentation
reaction has been applied to the synthesis of inside-outside
trans-bicyclo[n.3.1]alkanones,79 perhydrohistronicotoxin,80
and saudin.81 Acid-catalyzed fragmentation of [2 + 2]
photoadduct 194 led to a trans-bicyclo[5.3.1]undecan-11-
one 195 in 80% overall yield (Scheme 53, eq 1).82 The
inside-outside stereoisomer structure is unambiguously
proven by X-ray. In addition, the use of base in this photo
[2 + 2] cycloaddition/4-membered-ring fragmentation af-
forded a bridged tricyclic ketone 198a, an intermediate for
the total synthesis of (()-ingenol (Scheme 53, eq 2).83
A 4-membered ring in tricyclic compound 199 could be
cleaved by Wagner-Meerwein rearrangement (Scheme 54,
eq 1).84 In addition, fragmentation of cyclopropane derivative
201 was effected by samarium(II) catalyst (Scheme 54, eq
2).85
The intramolecular ring-opening/closure process is a
commonly used approach to synthesize bridged cyclic
organic compounds. In order to ensure the ring-opening/
closure tandem process can be easily achieved, a reaction
design typically involves a small-sized ring, such as cyclo-
propane and epoxide. Owing to the dipolar nature of the
cyclopropane moiety in 204, a cascade process involving
dioxasilinane ring cleavage/intramolecular nucleophilic cy-
clopropane opening and simultaneously 1,3-dioxolane forma-
1718 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 42
Scheme 43
Scheme 44
Scheme 45
tion led to an ether-bridged tricyclic anhydrosugar 205 in
80% yield (Scheme 55).86
A Dieckmann cyclization and epoxide ring-opening se-
quence was employed by the Maier group in the synthesis
of oxabicyclooctane ring system 208, a fully functionalized
core structure of abyssomicin C (Scheme 56, eq 1).87
Zhao
Scheme 46
Scheme 47
Analogously, Nicolaou and co-workers used this strategy to
synthesize oxabicyclic intermediate 210 (Scheme 56, eq 2).88
A similar approach was applied to prepare bridged 6-oxa-
bicyclo[3.2.1]-octane derivative 212 by opening the epoxide
ring in 211 (Scheme 56, eq 3).89
A natural product, (+)-pinnatoxin A, is comprised of a
27-membered carbocycle incorporating a unique A,G-spiro-
imine and E,F-6,8-dioxa-bicyclo[3.2.1]octane segment. This
E/F ring system is bridged by an ether bond and assembled
by a cascade process (Scheme 57).90 The cascade process
involves deprotection, epoxide ring formation, and simulta-
neous cleavage of the 1,3-dioxane ring with extrusion of
benzaldehyde to furnish alkoxide 215. An intramolecular
nucleophilic attack on the carbonyl group furnishes 7-mem-
bered oxepane 216 whose nucleophilic epoxide ring opening
affords the oxygen-bridged E/F ring system 214.
Bridge-Containing Organic Compounds Chemical Reviews, 2010, Vol. 110, No. 3 1719

Scheme 48 Scheme 50

Scheme 49

Scheme 51

A cascade reaction was designed by Nicolaou and co-

workers to construct the “left domain” of haplophytine Scheme 52
(Scheme 58).91 Upon epoxidation of tetrasubstituted olefin
217, the epoxide intermediate 218 underwent regioselective
epoxide ring opening, followed by skeleton rearrangement,
furnishing a desired bridged cyclization product 220.

4.2. Schmidt Reactions

Synthesis of the bridged bicyclic lactams with the nitrogen
at a bridgehead position is challenging because the “twisted
amide” tends to undergo rapid hydrolysis. Utilization of
intramolecular Schmidt reaction allows accessing bridged
bicyclic lactams. A Lewis-acid-catalyzed Schmidt reaction
of ketones 221 led to two lactams, bridged bicyclic 222 and
fused bicyclic 223 (Scheme 59).92 The nature of the sub-
stituent groups in the keto azide 221 affects the distribution
of the two products, which is largely controlled by the steric
effect and 1,3-cation-π interaction (R2 ) Ar).92 This Lewis-
acid-catalyzed Schmidt reaction may go through any of the
intermediates (224, 225, or 226). In the event, reaction of
221c or 221d with R2 ) aromatic groups (especially with
an electron-releasing group) proceeds for the most part
through intermediate 226, leading predominantly to bridged
lactam 222c or 222d by cleaving the C(a)-C(c) bond. In
contrast, reaction of 221a occurrs through intermediate 224,
providing 223a as the only product in 96% yield.
The intramolecular Schmidt reaction of 227 led to tricyclic
diketone 228 in 82% (Scheme 60, eq 1).93 The tricyclic
diketone 228 is the key intermediate in the total synthesis
of (+)-aspidospermidine.94 Surprisingly, reaction of the
monoprotected ketone 229 occurred through a process
involving presumably two intermediates, 231 (formed via
ring opening of ketal) and 232 (via carbon migration),
affording only 230 (Scheme 60, eq 2).93
1720 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 53
Scheme 54
Scheme 55
4.3. Pinacol Rearrangements
Many natural products bearing the β-araneosene ring
system95,96 are biologically active.97 Pinacol rearrangement
allows 1,2-diol in the presence of acid to generate ketone
by migrating the alkyl group. It was found, by Corey’s
group98 during the total synthesis of β-araneosene, that the
conformational rigidity of trans-12-membered ring 233
resulted in undesired Pinacol rearrangement product 234 via
cleavage of the C(a)-C(b) bond (Scheme 61, eq 1). This
importance of conformation is also evidenced by a very facile
rearrangement of hydroxy ketone 235 to 236 by exposure
to silica gel at 23 °C (Scheme 61, eq 2). Under identical
conditions as of eq 1, however, the cis-diol isomer 237
provided the desired bicyclo-ketone 238 through cleavage
of C(a)-C(c) bond with migration of tertiary carbon (Scheme
61, eq 3).
4.4. Claisen/Diels-Alder Cascade Process
A chemical transformation of xanthone derivative 239 into
242 and 243 was developed by Nicolaou and co-workers
Zhao
Scheme 56
(Scheme 62).99 It was observed that this Claisen/Diels-Alder
cascade sequence was accelerated by water. The concurrent
acceleration of the Diels-Alder component is due to the
hydrophobic effect100 and the unique internal pressure of
water.101
4.5. Polonovski Rearrangements
Application of a modified Polonovski reaction for serra-
tinine 244 resulted in generation of serratezomine A 245
(Scheme 63).102
The bridged lactone 245 could be converted into lactone
246 by treatment with p-TsOH or conducting the reaction
at 20 °C. This modified Polonovski reaction showed a strong
solvent effect, and THF or toluene as solvent shut down the
reaction completely (Table 1).102
A possible mechanism of this one-pot reaction was
proposed and is outlined in Scheme 64.102 Obviously, the
trifluoroacetylation of the N-oxide 247 serves as the driving
force for formation of hemiketal 249 and iminium 250.
4.6. Prins-Type Reactions
Reiβig and co-workers developed a synthetic approach by
utilization of Lewis-acid-catalyzed Prins-type cyclization
reactions to access bridged bicyclic compounds,103 which
presents an efficient and stereodivergent synthesis of C2-
branched 4-amino sugars.
Coordination of the Lewis acid to the more easily
accessible oxygen of the acetonide in 252 leads to ring-
opening intermediate 253 (Scheme 65). This electrophilic
species attacks the enol ether moiety similar to Prins-type
reaction to form intermediate 254. With the R1 group being
either 2-(trimethylsilyl)ethyl (TMSE) or 4-(p-methoxybenzyl)
(PMB), a bridged bicyclic ketone 255 is produced through
cleavage of the R1 group.
Table 2 demonstrates the scope of the Prins-type cycliza-
tion reactions. Although the rearrangement could be ac-
Bridge-Containing Organic Compounds
Scheme 57
Scheme 58
complished with different Lewis acids, dibutylboron triflate,
trimethylsilyl triflate, and tin tetrachloride proved to be the
best promoters. To introduce the more stable tert-butyldim-
ethylsilyl (TBS) protecting group, syn-252 was treated with
tBuMe2SiOTf (3 equiv) and then with triethylamine to
generate bicyclic products 255 (entries 3-5).
In a similar manner, rearrangement of diastereomeric 1,2-
oxazine anti-252 led to protected bicyclic 1,2-oxazine 256
(Scheme 66).103c
Chemical Reviews, 2010, Vol. 110, No. 3 1721
Mascareñas and co-workers developed a Lewis-acid-
promoted Prins-type cyclization to assemble oxabicy-
clo[n.3.1] ring systems 258 (Scheme 67, eq 1).104,105 10-
Oxabicyclo[4.3.1]decane 258a was obtained in 82% yield
as a mixture of isomers (8:2) at the tertiary center from
acetals 257a with a tethered terminal alkene group (n ) 1)
(Scheme 67, eq 2).104 Remarkably, the same reaction with
trimethylsilylalkene 257b produced only one stereoisomer
258b in 86% yield. The use of Prins-like cyclization allowed
1722 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 59
Scheme 60
access to the oxabicyclo[3.3.1] ring system 258c (Scheme
67, eq 3).105
5. Free-Radical Reactions
5.1. Intramolecular Radical Cycloadditions
Free-radical reaction as a synthetic tools has been fre-
quently applied to the total synthesis of natural products.
Myers’ group presented an excellent application of free-
radical reaction toward the synthesis of amide-bridged
tetracyclic compound 260, a key intermediate in the total
synthesis of alkaloid stephacidin B (Scheme 68).106 This
challenging transformation involves formation of an ami-
noacyl radical intermediate107 whose attack on the more
substituted carbon of the enamide C-C double bond and
expulsion of thiophenyl radical leads to the desired product
260.
Pattenden’s group108 utilized a radical domino reaction to
construct the taxane skeleton. Exposure of iodo compound
Zhao
Scheme 61
262 to radical reaction conditions (Bu3SnH/AIBN) led to
tricyclic bridged diketone 263 as a mixture of two diaster-
eomers (3:1) albeit in only 25% yield (Scheme 69).
A radical cycloaddition of bromide 264 occurred at
elevated temperature, affording a tricyclic core of alkaloid
(-)-aphanorphine analogs 265 (Scheme 70).109
A radical process was developed to extend the one-carbon
bridge into a two-carbon bridge (Scheme 71).110 Two
products were obtained with the desired rearranged product
267 in 77% yield along with about 15% of 268. The
bicyclo[2.2.2] diene 267 is a key intermediate in the synthesis
of platencin. Two competitive radical additions are described
in Scheme 71: addition of radical 269 onto the exo double
bond to lead to the desired product 267 via radical intermedi-
ate 270 and addition onto the tethered double bond to produce
the tricyclic 268.
Taking advantage of labile diazo compound 272, a radical
reaction was employed to build a methylene bridge 274 via
a biradical intermediate 273 (Scheme 72).111
It was found that a thermally more stable radical initiator
276 allows for cleaner, continuous generation of radicals
required for the radical cycloaddition of alkoxycarbonyl
selenide 275 during the total synthesis of (-)-pseudolaric
acid B (Scheme 73).112 In the event, two radical cycloaddition
products 277 and 278 were obtained from secondary alkoxy-
carbonyl selenide 275. A double-bond isomer 278 was
converted cleanly to a desired bridged lactone 277 by
addition of DBU to the reaction mixture after completion of
the cycloaddition. A unique feature of this radical cycload-
dition is that the radical generation is controlled by the low
rate of decomposition of radical initiator 276; consequently,
an operationally simple procedure could be adopted by
mixing all of the reagents together at the beginning.
Bridge-Containing Organic Compounds Chemical Reviews, 2010, Vol. 110, No. 3 1723

Scheme 62

Scheme 63 Scheme 65

Table 1. Solvent and Temperature Effect102

yield, %
solvent T (°C) 244 (recovered) 245 246
CH2Cl2 -50 30 6
CH2Cl2 -20 48 27
CH2Cl2 0 17 38 oxidative coupling.114 Mechanistically, it is believed that
CH2Cl2 20 65 conversion of indole derivative 279 into its corresponding
CHCl3 -20 27 13 bridged compound 280 by forming a key C6-C22 bond in
CH3CN -20 17 6 the enantioselective total synthesis of stephacidin A involves
THF -20 66 0 0
toluene -20 54 0 0 radical species (Scheme 74).
In addition to radical initiator, radical species can be
Scheme 64
generated by means of an electron-transfer process as well.
Formation of bridged lactones 283 was hypothesized to
involve a radical process initiated by an electron-transfer
reaction between dihydropyridine 285115 and triflic anhydride
(Schemes 75 and 76).116 A radical combination of trifluoro-
methyl radical ( · CF3), generated by dissociation of SO2 from
unstable trifluoro-methanesulfonyl radical (CF3SO2 · ),117 with
dihydropyridine cation radical provides 287. The final
trifluoro-methanesulfonic acid anion-promoted intramolecular
cycloaddition gives the desired product 283.
Notably,atandemalkoxyradicalfragmentation-etherification
sequence was designed to access a tricyclic compound 289,
a key intermediate in the synthesis of the AB subunit of
angelmicin B (Scheme 77).118 Formation of ether-bridged
289 is presumably via β-iodine intermediate 291.
The use of an intramolecular hydrogen abstraction reaction
of alkoxy radicals was applied toward the synthesis of the
2,7-dioxabicyclo[2.2.1]heptane ring system by Suárez and
co-workers (Scheme 78).119 The resulting C radical is
5.2. Other Radical Reactions oxidized to give an oxycarbenium ion that is then internally
Recent findings in Baran’s laboratory113 demonstrate that trapped by the nucleophilic alcohol to furnish 2,7-
C-C bond formation can occur through a metal-mediated dioxabicyclo[2.2.1]heptanes 293.
1724 Chemical Reviews, 2010, Vol. 110, No. 3 Zhao

Table 2. Reaction Conditions for the Conversion of syn-252 to 255

entry R1 R2 R3 LA solvent T (°C) R4 yield, % refs
1 TMSE CH3 CH3 SnCl4 (3 equiv) CH3CN -30 H quant. 103c
2 TMSE CH3 CH3 Me3SiOTf (0.05 equiv) CH2Cl2 -30 TMS 79 103c
3 TMSE CH3 CH3 tBuMe2SiOTf (3 equiv) CH2Cl2 0 TBS quant. 103c
4 TMSE -(CH2)5- tBuMe2SiOTf (3 equiv) CH2Cl2 0 TBS quant. 103c
5 TMSE Ph H tBuMe2SiOTf (3 equiv) CH2Cl2 0 TBS quant. 103c
6 PMB CH3 CH3 SnCl4 CH3CN -30 H 58 103b
7 TMSE CH3 CH3 BF3 OEt2 CH2Cl2 rt H 63 103b
8 TMSE H SPh Me3SiOTf (2 equiv) CH2Cl2 -30 H 75 103a

Scheme 66 tandem process (Scheme 79, eq 1).121 The reaction tolerates

different substituents on the substrate alkynes 294, and in
addition, a number of alcohols proved to be effective
nucleophiles. A highly efficient cycloisomerization of bis-
homopropargylic diols 296 catalyzed by Au(I) or Au(III)
under very mild conditions furnished functionalized strained
bicyclic ketals 297 (Scheme 79, eq 2).122
Mechanistically, these reactions involve additions of
Scheme 67
hydroxyl group onto the metal-activated triple bond, leading
to vinyl gold intermediates 298 or 300. The resulting
intermediates may then be protonolyzed, providing enol
ethers 299, which then undergo another intramolecular
addition123 of the remaining hydroxyl group, leading to the
cyclic ketals 297 (Scheme 80, eq 1). In the absence of internal
hydroxyl group, the enol ethers 301 may go through a Prins-
type cyclization, affording the ether-bridged 295 (Scheme
80, eq 2).
A new and highly convenient Ag-catalyzed intramolecular
oxycyclization of alkynes proved to be expedient in the
synthesis of epoxy-bridged tetrahydropyran skeletons, which
are present in a wide range of natural products, including
(+)-exo-brevicomin, (-)-frontalin, (+)-dedemniserinolipid
B, (+)-xanthane epoxide, and phyllaemblic acid.124 Oxycy-
clization of bis-homopropargylic diols 302 in the presence
of 5 mol % of AgOTf led to bridged tetrahydropyrans 303
in good yields (Scheme 81, eq 1).125 Extension of this Ag-
catalyzed cyclization to other diols is demonstrated in eqs 2
and 3 of Scheme 81. Interestingly, cyclization of carboxylic
acid 308 was also successful by heating, giving bridged
lactone 309 in 70% yield (Scheme 81, eq 4).125

6. Transition-Metal-Catalyzed Cyclization 6.2. Cu(II)-Catalyzed Cyclizations

Reactions
6.1. Gold-Catalyzed Reactions
Electrophilic transition-metal salts and complexes are
uniquely selective Lewis acids that have a high affinity for
π bonds. Owing to the electron configuration
([Xe]4f145d106s1), gold catalysts mainly exist in the +1 and
+3 oxidation states. As a soft Lewis acid, gold catalysts can
activate unsaturated functionalities such as alkynes, alkenes,
and allenes to create carbon-carbon and carbon-heteroatom
bonds under extremely mild conditions.120 Furthermore, gold
catalysts can catalyze the formation of carbon-carbon and
carbon-heteroatom (oxygen and nitrogen) bonds by activat-
ing the sp, sp2, and sp3 carbon-hydrogen bonds.120 Most
Au(I)-catalyzed reactions are well tolerated by oxygen, water,
and alcohols, which is in sharp contrast to most air- and
moisture-sensitive Lewis-acid- or transition-metal-catalyzed
reactions.
A diastereoselective synthesis of eight-membered car-
bocycles 295 involves gold- or platinum-catalyzed 6-exo-
cycloisomerization followed by a Prins-type cyclization
Diazo compounds are used as precursors to carbenes,
which are generated by light and heat or in the presence of
transition-metal catalyst. These carbenes in turn can undergo
intramolecular cyclizations to form bridged carbocycles. The
use of Cu(II) catalyzed intramolecular cyclization of alkynes
bearing a diazo functionality allowed access to ether-bridged
ketones 311 (Scheme 82).126
In another case, this Cu(II)-catalyzed cyclization was
brilliantly combined with ring expansion of oxonium ylide
314, affording ether-bridged macrocyclic ketone (Z)-315 in
excellent yield (Scheme 83).127
6.3. Rhodium-Catalyzed Reactions
6.3.1. Rh(II)-Catalyzed Reactions
The transition-metal-catalyzed carbenoid insertion reaction
of R-diazocarbonyl compounds is an important tool in
synthetic organic chemistry.128 Rh(II) has important applica-
tions in the catalytic generation of carbenoid intermediates129
from diazo compounds. Intramolecular 1,3-dipolar cycload-
dition of 317, being generated from Rh(II)-carbenoid,
Bridge-Containing Organic Compounds
Scheme 68
Scheme 69
Scheme 70
furnished ether-bridged polycyclic system 318 in 90% yield
(Scheme 84).130
Similarly, an ether-bridged cycloadduct 321131 was formed
from dipole intermediates 320, generated via trapping the
rhodium carbenoid intermediates by the adjacent amido
carbonyl group (Scheme 85).132 This cycloaddition process
was applied toward the generation of a pentacyclic skeleton
in the synthesis of indole alkaloids,133 such as (()-3H-
epivincamine and (()-tacamonine.132
6.3.2. Rh(I)-Catalyzed Reactions
Analogous to Rh(II), Rh(I) reveals great diversity in C-H
bond activation, which results in an alkenylation/electro-
cyclization process (Scheme 86).134 A 9-membered ring
326 with an exocyclic double bond is presumably formed
via intermediate 324 and 10-membered Rh complex 325.
Electrocyclization of the resulting azatriene 326 gives the
bicyclic enamine 323 with bridgehead unsaturation (Scheme
86).134
Chemical Reviews, 2010, Vol. 110, No. 3 1725
6.4. Palladium-Catalyzed Reactions
Murrayazoline, isolated from the genus Murraya,135 dis-
plays potent antiplatelet aggregation activity.136 During the
synthesis of murrayazoline, a Pd-based intramolecular
Buchwald-Hartwig C-O coupling reaction137 was utilized
to assembly a [3.3.1] heterocyclic ring system by Chida’s
group in Japan. Construction of the methylene bridge in this
hexa-heterocyclic structure is the final step of the total
synthesis of murrayazoline. This Pd-mediated C-O coupling
reaction of 327 furnished the desired murrayazoline in 80%
yield (Scheme 87).138 However, this C-O coupling reaction
would need some optimization prior to any practical ap-
plication due to the prerequisite of stoichiometric amounts
of Pd(OAc)2.
Paterson and co-workers139 presented another application
of palladium(II)-catalyzed cyclization toward the preparation
of bicyclo[3.2.1] acetal 330 in the total synthesis of the
orthinine decarboxylase inhibitors saliniketals A and B.
Saliniketals A and B were isolated in 2007 by Fenical and
co-workers from the marine actinomycete Salinispora areni-
cola.140 The use of catalytic PdCl2 and CuCl2 in THF under
an atmosphere of oxygen at 0 °C was found to give a good
yield (88%) of the acetal 330. The desired [3.2.1]-dioxabi-
cycle 330 likely arises from the expected 5-exo attack of
the C13 hydroxyl group in diol 328 onto the palladium-
activated terminal double bond followed by β-H elimination
to lead to intermediate 329. The subsequent acid-catalyzed
acetal formation furnishes the desired bicyclo[3.2.1]acetal
330 (Scheme 88).139
Similar to gold(I) or gold(III), Pd(II)-catalyzed cycloi-
somerization of alkyne 331 was applied toward construct-
ing the bridged 6,8-dioxabicyclo[3,2,1]octane core 332
in the total synthesis of didemniserinolipid B (Scheme
89).141
A novel and efficient palladium-catalyzed intramolecular
allylic alkylation was performed by Wipf’s research group,
affording bridged adducts 334a and 334b in excellent yields
(Scheme 90).142
1726 Chemical Reviews, 2010, Vol. 110, No. 3 Zhao

Scheme 71

Scheme 72 Scheme 75

Scheme 73
Scheme 76

7. Other Reactions
7.1. Iodine-Promoted Cyclizations
The triterpene garsubellin A, first isolated by Fukuyama
and colleagues from the wood of Garcinia subelliptica, is a
Scheme 74
molecule with biological activity and structural complexity.
Not surprisingly, the total synthesis of garsubellin has caught
the interest of many research groups.143 Iodine-promoted
cyclization allows the formation of C-C or C-heteroatom
bonds under basic but mild conditions, which has been used
in total synthesis.
Using Nicolaou’s condition (I2, KI, KHCO3, THF-H2O),143a,b
Danishefsky and co-workers were able to synthesize diiodide
336, an intermediate of garsubellin A, in one step (Scheme
91, eq 1)144 A new and short enantioselective pathway for
the synthesis of the anti-influenza neuramidase inhibitor
oseltamivir phosphate (Tamiflu) was developed by Corey’s
group.145,146 As one of the intermediates, a bridged bicyclic
lactam 338 was synthesized by iodolactamization of 337
using the Knapp protocol147 in 84% yield (Scheme 91, eq
2).
Bridge-Containing Organic Compounds Chemical Reviews, 2010, Vol. 110, No. 3 1727

Scheme 77 Scheme 79

Scheme 80
Scheme 78

Nicolaou’s group employed an intramolecular iodoetheri-

fication to construct the G ring system during the total
synthesis of azaspiracid-1, a neurotoxin isolated from mus-
sels, and iodoether 340 was obtained in 62% yield (Scheme
92).148

7.2. Pictet-Spengler Cyclizations

Multicomponent reactions (MCRs) involving domino
processes149 have emerged as powerful tools used in modern
organic chemistry. A multicomponent domino reaction
involving Pictet-Spengler cyclization was developed to
synthesize polyfunctionalized 2,6-diazabicyclo[2.2.2]octane
derivative 344 (Scheme 93).150
Scheme 94150 underscores this three-component reaction
pathways. Molecular sieves (4 Å) act as both dehydrating
agent and heterogeneous catalyst for the initial Michael
addition.151 The key steps of the present one-pot process are
the successive formation of three iminium intermediates, 347,
348, and 349 via Pictet-Splengler reaction.
Despite the total synthesis of ecteinascidin 743,152 an
extremely potent antitumor agent,153 which has been achieved
by Corey and co-workers154 and others,155 investigations into
new synthetic approaches are still of great interest for
synthetic chemists. A pentacyclic product 351 containing the
ABCD ring system156 was constructed by an intramolecular
Pictet-Spengler cyclization of aldehyde 350 triggered by
cleavage of the Boc protection group (Scheme 95).157
7.3. Nucleophilic Substitutions
A process involving intramolecular conjugate SN2-type
displacement was devised by Clive and co-workers for
constructing a broad range of carbocycles.158 This metal-
free process offers different opportunities from palladium-
catalyzed cyclization159 of allylic acetates and should tolerate
the presence of palladium-sensitive groups. A bridged
bicyclic compound 353 was synthesized by an intramolecular
SN2-type displacement reaction (Scheme 96).158 In the
1728 Chemical Reviews, 2010, Vol. 110, No. 3 Zhao

Scheme 81 Scheme 84

Scheme 85

Scheme 82

Scheme 86

Scheme 83

Scheme 87

presence of DBU and 50 min of reaction time, a mixture of

bicycles 353 and 354 was isolated. However, with the
inorganic base (Cs2CO3) only bridged bicyclic 353 was
obtained in 99% yield. Interestingly, 354 could be converted
into 353 by exposure to DBU.
A hypothesis for the formation of 353 and 354 in the
presence of DBU is described in Scheme 97.158 A DBU-
facilitated cycloaddition occurred through iminium 355,
leading to bicycle 354. Formation of 353 from 354 may go
through DBU-promoted ring rearrangement with two possible
intermediates 356 and 357 or through Claisen rearrangement.
Bridge-Containing Organic Compounds Chemical Reviews, 2010, Vol. 110, No. 3 1729

Scheme 88 Scheme 90

Scheme 91

Scheme 89

SN2-type cycloaddition/elimination reactions were utilized

to form the [2.2.2] bridged bicyclic ring system in asym-
metric total syntheses of the fungal metabolites (-)-stepha-
cidin A, (+)-stephacidin B, and (+)-notoamide B. Conver-
sion of allyl chloride 358 to the desired bridged bicycle 359
was completed via a SN2-type cycloaddition/elimination
sequence in 60% yield as a single diastereoisomer (Scheme
98, eq 1).160 Another example of a SN2-type displacement
was found in the total synthesis of (()-welwitinone A
isonitrile, which gave not the expected dehydration of the
C20 alcohol 362 but an undesired bridged product 361
(Scheme 98, eq 2).161
Mesylate is considered an excellent leaving group in
nucleophilic substitution reactions. The use of mesylates in
the SN2-type cyclizations was found in the syntheses of (+)-BE-43472B, an antitumor agent, is composed of two
cytosine, 11-norcytisine, and (+)-serratezomine A (Scheme different anthraquinone-type molecules joined together by a
99). The SN2 displacement of mesylates by nitrogen atoms sterically hindered carbon-carbon bond and an oxygen
was demonstrated in the formation of bridged cytisine bridge in an assembly containing no less than eight rings.
skeletons 364 and 366 (Scheme 99, eqs 1 and 2).162 In Nicolaou and co-workers reported the first total synthesis of
addition, in-situ-generated carboxylate also participated in this antibiotic. Scheme 102166 illustrates the synthesis of a
an analogous SN2 displacement in the formation of bridged key 8-membered ring intermediate 377 through a cascade
lactone in (+)-serratezomine A (Scheme 99, eq 3).163 sequence consisting of Diels-Alder reaction, hemiketal
An unexpected rearrangement of hemeanthamine-type formation, and SNAr ipso substitution. The DA adduct
alkaloids in the presence of halogenating agents has been intermediate 375 undergoes an intramolecular nucleophilic
observed.164 When hemeanthamine 368 was treated with a ipso substitution expelling a molecule of MeOH to afford
large excess of thionyl chloride (20 equiv), 370 was formed the desired octacycle 377. This novel cyclization is appar-
in 71% yield via intermediate 369, which underwent ring ently facilitated by the presence of the adjacent phenolic
rearrangement containing a SN2-type cyclization/displace- quinone moiety.
ment sequence (Scheme 100).164
In addition to SN2-type cyclization, SN1-type cyclization
appears to be equally important in the organic synthesis.
7.4. Nucleophilic Additions
An N-acyliminium cyclization was promoted with thio- The Michael addition is an important method for diastereo-
phile, AgBF4, giving 372 as a single diastereomer (Scheme and enantioselective C-C bond formation. A stereoselective
101).165 intramolecular Michael addition of ketone 378 in the
Furthermore, the use of SNAr ipso substitution in the total presence of t-BuOK yielded a bridged tricyclic ketone 379
of synthesis was described. Bisanthraquinone natural product (only the endo isomer) (Scheme 103, eq 1).167 Hopeahainol
1730 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 92
Scheme 93
A168 and hopeanol169 are two recently disclosed molecules,
each thought to be biosynthetically derived from two
molecules of resveratrol. Nicolaou and co-workers explored
the total synthesis of hopeahainol A and hopeanol and found
that the transformation of hopeahainol A to hopeanol
occurred upon exposure of hopeahainol A to 1 equiv of
NaOMe in MeOH at 25 °C, giving hopeanol in 80% yield
(Scheme 103, eq 2).170 It was hypothesized that the base-
promoted γ-lactone ring opening allows C1b-C7a bond
formation to be effected through an intramolecular conjugate
nucleophilic addition of intermediate 380.
In addition to the base-promoted nucleophilic cycloaddi-
tion, an acid-catalyzed reaction was also observed, delivering
a bridged bicyclo[3.2.1]octene 382 (Scheme 104, eq 1).171
Interestingly, no reaction occurred for the diastereomer (β-
OMe) under identical conditions, suggesting that rearrange-
ment of 381 may be controlled by the relative configuration
of the tetrasubstituted SP3 carbon. An intermediate 385
containing a 6,9-trioxatricyclo[3.3.2.03,7]decane ring system
in a total synthesis of norhalichondrin B was effected
involving TBS group removal and an intramolecular hetero-
Michael addition (Scheme 104, eq 2).172 Subjecting the
mixture of diastereoisomers 384 to nonaqueous workup
conditions (CaCO3, DOWEX 50WX8-400, MeOH) resulted
in the desired product 385 in 64% yield.
Utilization of intramolecular nucleophilic addition of
R,β-unsaturated ketones 386 toward synthesis of bridged
compounds 387 was reported by Goeke and co-workers
(Scheme 105).173 This Lewis-acid-catalyzed cyclization
Zhao
involves a 1,2-H shift followed by 1,2-alkyl group shift
cascade process.
The versatility and practicality of this Michael-type
intramolecular cycloaddition was nicely demonstrated by
Patir’s group in the synthesis of a framework of uleine-type
alkaloids. A bridged tetracyclic compound 392 was obtained
from sequential multistep reactions involving a Michael-type
intramolecular cycloaddition of iminium intermediate 390
(Scheme 106).174
Furthermore, N-tosyl removal provides free indole, which
serves as a nucleophile, leading to a bicyclo[3.3.1] ring
system 395 (Scheme 107).175
Intramolecular carbolithiation (anionic cyclization) offers
a valuable alternative to the related radical cyclization and
provides enhanced or complementary stereoselectivity.
Tin-lithium exchange and anionic cyclization have been
used to construct the azabicyclo[2.2.1]heptane ring system.176
On treatment of the cis isomer 396 with n-butyllithium,
the 7-azabicyclo[2.2.1]heptane 397 was formed in good
yield and as a single diastereomer (Scheme 108, eq 1).176 In
the presence of electrophiles, 2-substituted 7-azabicyclo-
[2.2.1]heptane derivatives 398 were obtained (Scheme 108,
eq 2).176
Besides the intramolecular nucleophilic addition, a com-
bination of intermolecular nucleophilic addition with in-
tramolecular cycloaddition has been employed in the con-
struction of bridged compounds 401 during the synthesis of
tropinones (Scheme 109, eq 1).177 Formation of 3-oxo-8-
aza-bicyclo[3.2.1]octane 401 involves two iminium triflate
intermediates 402 and 403 (Scheme 109, eq 2).
7.5. Aldol-Type Condensation Reactions
Aldol condensation is one of the important organic
transformations in organic synthesis. Weinreb and co-workers
reported an intramolecular condensation reaction to construct
bridged molecules such as tetracyclic lactone 406 via
Bridge-Containing Organic Compounds
Scheme 94
Scheme 95
Scheme 96
transesterification of intermediate 405 (Scheme 110, eq 1).178
Reductive lactone ring opening of 407 occurred with DIBAL-
H, leading to transitory aldehyde 408 followed by intramo-
lecular aldol cycloaddition to bridged bicyclic alcohol 409
(Scheme 110, eq 2).179
7.6. Friedel-Crafts Cyclizations
An important use of the Friedel-Crafts alkylation reaction
is to effect ring closure. The most common method is to
heat the aromatic compound with aluminum chloride.
Friedel-Crafts alkylation promoted by SnCl4 allowed
Chemical Reviews, 2010, Vol. 110, No. 3 1731
cyclization of acetals 411 at cryogenic temperature and an
optically pure 412, which were obtained in excellent yields
(Scheme 111).105
7.7. Ring-Closing Metathesis
Ring-closing metathesis (RCM) is a powerful tool to
assemble macrocycles. The RCM method was applied toward
the construction of bicyclo[4.2.1]nonan-9-one 414, a bridged
intermediate in the synthesis of cyclooctanoids (Scheme 112,
eq 1).180 Although ring-closing metathesis has been widely
used in organic synthesis, its application in constructing
medium-sized bridged carbocycles is less common.181 RCM
was utilized to generate the strained tricylic skeleton of
platencin. Under reflux in methylene chloride in the presence
of Grubbs’ second-generation catalyst (7 mol %), triene 415
was converted to 416 in 90% isolated yield (Scheme 112,
eq 2).182 A nice ring-opening metathesis (ROM)-RCM
strategy was developed by Ghosh to synthesize the tricycles
418 (Scheme 112, eq 3).183
1732 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 97
Scheme 98
8. Important Classes of Compounds
8.1. Reactions in Synthesis of FR66979 and
FR900482
Anticancer agents FR66979 and FR900482 are structurally
related to the mitomycins and are believed to act by a similar
mechanism but with less toxicity. Both FR900482 and
FR66979 possess unique structural features, including hy-
droxylamine hemiacetal and aziridine functional groups,
Zhao
Scheme 99
Scheme 100
making it an attractive target for synthetic organic chemists.
Although there were a number of reports184 regarding the
synthesis of FR900482 (R ) CHO),185 six individual research
groups186-191 recently revealed new synthetic approaches
toward the synthesis of FR900482 and FR66979.
Bridge-Containing Organic Compounds
Scheme 101
Williams and co-workers187 reported a one-pot process for
the construction of a bridged bicyclo[3.3.1] intermediate 421
in the synthesis of FR66979 (R ) CH2OH) and FR900482.
This one-pot protocol involves cleavage of the N-p-meth-
oxybenzyl residue and oxidization of the amine to the
corresponding hydroxylamine, thus forming the desired
hydroxylamine hemiketal. Reactions of 419 furnished 421
as the only isolated product in 30-50% yield, along with
recovered starting material (40-50%) (Scheme 113).
An analogous approach was developed by Fukuyama’s
group187 toward an enantioselective total synthesis of (+)-
FR900482 (Scheme 114). Hydroxymethylation was best
effected by treatment of ketone 422 with formalin in the
presence of a catalytic amount of LiOH to furnish the desired
423 with high diastereoselectivity (94:6). Subsequent acidi-
fication afforded hemiacetal 424, whose acetonide formation
gave the bridged pentacyclic compound 425 in 56% yield
from ketone 422.
Two research groups, Ciufolini188 and Paleo,189 demon-
strated independently another synthetic method to access the
core bridged bicycle 427 (Scheme 115). Exposure of the
acetyl-protected hydroxylamine 426 to hydrazine gave
directly the transannular cyclization product 427 in quantita-
tive yield.
Scheme 102
Chemical Reviews, 2010, Vol. 110, No. 3 1733
Oxidative ring expansion can be realized in the presence
of oxidant by heterolytic cleavage of the C-N bond of a
nitrogen-containing heterocyclic compound. Jimenez and co-
workers190 reported that the oxygen-bridged bicyclo[3.3.1]
intermediate 434 was assembled by oxidative ring expansion
of bicyclo[3.3.0] 428 and successive intramolecular cycliza-
tion (Scheme 116) via intermediates such as iminium salt
430 and ring expansion 433.
Trost and co-workers modified this oxidative ring expan-
sion with mCPBA as the oxidant, which improved the yield
of 436 (Scheme 117).191 This conversion of 435 to 436 went
through two intermediates, hydroxylamine 437192 and 8-mem-
bered ketone 438.
8.2. Reactions in Synthesis of Platensimycin and
Carbaplatensimycin
Platensimycin has potent activity against Gram-positive
bacteria including multiresistant strains of staphylococci and
enterococci. Platensimycin has an intriguing structure (Scheme
118), which features a hydrophilic aromatic unit and a
lipophilic tetracyclic unit linked together by an amide bond.
8.2.1. Synthesis of the Cyclohexane Ring System
The total synthesis of platensimycin was pioneered by
Nicolaou’s group, and the first total synthesis of the racemate
was reported in 2006.193 Scheme 119 (eq 1) describes a
synthesis of cyclohexane ring through a SmI2-mediated
cyclization reaction, affording 440 and its diastereoisomer
in 46% combined yield. Analogously, the same cyclization
conditions were applied toward asymmetric total syntheses
of platensimycin, and a desired secondary alcohol 442 was
obtained in moderate yield (39%) as a single diastereoisomer
(Scheme 119, eq 2).194
Further development by Nicolaou and co-workers resulted
in another approach toward the synthesis of a 6-membered
ring system of carbaplatensimycin (Scheme 120).195 Treat-
1734 Chemical Reviews, 2010, Vol. 110, No. 3 Zhao

Scheme 103

Scheme 104

ment of 1-ethoxyethyl (EE) ether 443 with KHMDS induced 8.2.2. Synthesis of the Cyclopentane Ring System
intramolecular conjugate addition of the transient anion so
generated onto the bisenone subunit to afford tricycle 444 A platensimycin analog 446 was synthesized by exposure
in 70% yield as a single epimer at C10. of diol 445 to a catalytic amount of AuCl3 (Scheme 121).196
Bridge-Containing Organic Compounds
Scheme 105
Scheme 106
Scheme 107
Mechanistically, the reaction involves addition of one
hydroxyl group onto the metal-activated triple bond, leading
to vinyl gold intermediate 447. The resulting intermediate
may then be protonolyzed, providing enol ethers 448, which
then undergoes another intramolecular addition123 of the
Chemical Reviews, 2010, Vol. 110, No. 3 1735
Scheme 108
Scheme 109
remaining hydroxyl group, leading to the cyclic ketal 446
(Scheme 122).
Analogously, a Au(I)-catalyzed cyclization of acetylenic
diene 449 in toluene-methanol (10:1) at 25 °C, reported by
Nicolaou’s group, afforded a bicyclo[3.2.1] ketone 450 in
94% yield (Scheme 123).110
Although a radical approach toward the target molecule
was nicely documented in a recent review,193b conjugate
radical cycloaddition of bromoalcohol 451 was performed
by Ghosh and co-workers197 to access bicyclo[3.2.1] alcohols
452 and 453, whose subsequent reactions gave bicyclic
lactone 454 (Scheme 124).198
Another conjugate radical cycloaddition of 455 yielded
the tricyclic caged framework 456 (Scheme 125).199 This
radical-mediated cyclization was proved to depend critically
on the reaction temperature. An initial attempt at 100 °C
gave 456 as a mixture of epimers in a 2.5:1 ratio with 75%
yield, while lowering the reaction temperature to ∼65 °C
improved both the yield and selectivity (81%, 4.5:1).
Addition to the radical cyclization, Corey and co-worker
released a one-step desilylation of 457 and subsequent SN2-
type cyclization to produce 458 (Scheme 126).200
Analogous bicyclo[3.2.1] ring systems 460 and 461 were
synthesized by Nicolaou’s group via a radical cyclization
process shown in Scheme 127.201 This radical cycloaddition
1736 Chemical Reviews, 2010, Vol. 110, No. 3 Zhao

Scheme 110 Scheme 113

Scheme 114

Scheme 111

Scheme 112

8.2.3. Synthesis of Tetrahydrofuran and Tetrahydropyran

of 459 occurred by application of the sequential oxymercu-
ration/reductive alkylation methodology pioneered by
Giese.202
Ring Systems
Treatment of an inseparable diastereoisomer mixture of
secondary alcohol 440 with TFA led to smooth etherification,
affording a cage-like structure 462 in 87% yield (Scheme
128, eq 1).193 Analogously, the tetrahydrofuran/tetrahydro-
pyran ring system 462 was realized from cyclization of
endocyclic olefin 442 under identical conditions in the same
yield (Scheme 128, eq 2).194
On application of Nicolaou’s ether linkage condition,193
an inseparable 1:1 mixture of diol 463 was treated with TFA
in CH2Cl2 to effect formation of the desired ring system 464
(Scheme 129).203 This two-step sequence afforded intermedi-
ate 464 with axial alcohol in 39% yield and the equatorial
alcohol in 42% yield.
Furthermore, Nicolaou and co-workers developed another
novel approach toward the tetrahydrofuran system (Scheme
130).201 Stereoselective reduction and cycloaddition of ketone
465 resulted, upon acidic workup (1 N aq HCl), in formation
of the desired cage hydroxy compound 466 in 80% yield.
Corey’s group employed a bromine-promoted cyclization
reaction toward the synthesis of a tetrahydrofuran ring system
Bridge-Containing Organic Compounds
Scheme 115
Scheme 116
Scheme 117
Scheme 118
of Platensimycin (Scheme 131).200 Conversion of phenol 467
to bromoether 468 was effected in two steps with high
diastereoselectivity (>10:1 ratio) in 84% overall yield.
A radical cyclization was adopted to construct the cyclo-
pentane ring system in carbaplatensimycin (Scheme 132).195
Chemical Reviews, 2010, Vol. 110, No. 3 1737
Scheme 119
Scheme 120
Scheme 121
Scheme 122
Scheme 123
Xanthate 469 undertook a Barton-McCombie deoxygen-
ation204 furnishing hydrocarbon 470 in 65% yield.
An enantioselective route to the oxatetracyclic core of (-)-
platensimycin has been investigated by using an intramo-
lecular Diels-Alder (IMDA) reaction as the key step.205 An
initial attempt with enol ether 471, in the presence of a
catalytic amount of 2,6-di-tert-butyl-4-methylphenol (BHT)
as a radical inhibitor, provided the expected oxatetracyclic
core 472 in 39% yield (Scheme 133, eq 1). Incorporating
the desired methyl group on the dienophile in 473, the desired
DA cycloadduct 474 was obtained in 44% yield at 270 °C
(Scheme 133, eq 2). The low yield could be due to side
reactions at elevated temperature.
The use of rhodium(II)-catalyzed carbonyl ylide from
diazoketone 475 was able to construct the desired ketone
477 via ylide 476 and its subsequent [3 + 2] cycloaddition
reaction (Scheme 134).206
1738 Chemical Reviews, 2010, Vol. 110, No. 3 Zhao

Scheme 124 Scheme 129

Scheme 130

Scheme 131

Scheme 125

Scheme 132

Scheme 126

Scheme 133

Scheme 127

Scheme 128

Scheme 134

Robinson annulation was accomplished in one pot using

L-proline as a chiral control element to mediate an initial
intramolecular Michael addition followed by sodium hy-
droxide treatment to finish the aldol dehydration. The
tetracyclic core structure 462 was obtained along with its
C-9 epimer in a dr ratio of 5:1 (Scheme 135).207
8.3. Reactions in Synthesis of Cortistatin A
Cortistatin A has received a great deal of attention from
the synthetic community due to its promising biological
Bridge-Containing Organic Compounds
Scheme 135
Scheme 136
Scheme 137
Scheme 138
activity and unusual structural features. Since its isolation
as a trace component from a marine sponge, Corticulum
simplex, by Kobayashi in 2006,208 numerous research groups
have embarked on the synthesis of cortistatin A.212,209 The
Chemical Reviews, 2010, Vol. 110, No. 3 1739
Scheme 139
Scheme 140
Scheme 141
Scheme 142
first synthesis of cortistatin A was recently achieved by Baran
and co-workers.210 Scheme 136 outlines four different
synthetic pathways developed in recent years toward the
construction of an ether-bridged BC ring system.
8.3.1. Synthesis of the C Ring System (Route I)
The substitute X () OMs) in route I (Scheme 136) as a
leaving group was replaced during the intramolecular alky-
lation, leading the desired ether-bridged tetracyclic products
480 and 482 (Scheme 137, eqs 1 and 2).209d Phenoxide
intermediates, presumably generated by treatment of com-
pounds 479 or 481 with TBAF at room temperature,
underwent the desired alkylative dearomatization upon
heating at 130 °C.
With X ) Br as a leaving group, compound 483 cyclized,
readily affording the ether-bridged intermediate 484 with
moderate yield (Scheme 138).209e
1740 Chemical Reviews, 2010, Vol. 110, No. 3 Zhao

Scheme 143 Scheme 146

Scheme 147

Scheme 144

Scheme 145 Scheme 148

Recent studies have shown that cortistatins may protect

against loss of vision.211 Yamashita et al. employed a Scheme 149
conjugated radical cycloaddition of iodo compound 485 to
synthesize the framework of cortistatin A (Scheme 139).212
This one-step reaction was conducted in the presence of Et3B
and (TMS)3SiH in THF, furnishing the C ring of pentacyclic
compound 486.213

8.3.2. Synthesis of the ABD Ring System (Route II)

Synthesis of the ABD ring system 488 was effected by
treatment of macrocycle 487 with Pd(OAc)2 in the presence
of LiBr in 37% yield (Scheme 140).209h
This palladium-mediated [4 + 3] cycloaddition occurs
through a mechanism depicted in Scheme 141.209h The initial
(allene)palladium complex 489 produces a σ-allylpalladium
intermediate, which rapidly equilibrates to the corresponding
(π-allyl)palladium intermediate 490. Intramolecular nucleo-
philic attack by the furan ring gives the [4 + 3] cycloaddition
Bridge-Containing Organic Compounds
product 491, which loses a proton in the presence of
potassium carbonate to lead to the desired ABCD ring system
488.
8.3.3. Synthesis of the BC Ring System (Route III)
Nicolaou and co-workers designed a 1,4-hydroxy addition/
aldol/dehydration cascade process providing the desired BC
ring system 493 upon heating hydroxy enone-enal 492 at
reflux in dioxane in the presence of K2CO3 in 52% yield
(Scheme 142).209a The ether-bridged product 493 was
produced through two intermediates, 494 and 495, generated
presumably from conjugate addition and Aldol condensation,
respectively.
8.3.4. Synthesis of the BC Ring System (Route IV)
The BC ring system in cortistatin A can be assembled
through route IV as described in Scheme 136. A mild SN1-
type cyclization mediated by Lewis acid to close the final
ring of the cortistatin skeleton was accomplished by
Baran’s group (Scheme 143, eq 1).210 A mild alkylative
dearomatization was realized under oxidative conditions
(Scheme 143, eq 2).209b In the event, a key pentacyclic core
499 was synthesized via a tandem oxidative dearomatization/
cyclization in the presence of hypervalent iodine PhI(OAc)2.
An enantioselective synthesis of the ABC ring system of
cortistatin A has been achieved by Shair and co-workers
using a highly diastereoselective aza-Prins cyclization coupled
with transannular etherification (Scheme 143, eq 3).209g
Scheme 144 describes the plausible reaction mechanism
involving an intermediate 503, generated by aza-Prins
cyclization coupled with transannular etherification. Oxonum
ion release from 503 affords pentacyclic 501. Experimental
evidence supports this sequence of events rather than MEM
deprotection preceding aza-Prins cyclization.209g
8.4. Reactions in the Synthesis of Welwistatin
and Analogues
The synthesis of welvistatin poses a significant synthetic
challenge because of its unique four-ring compact chemical
structure.214 An approach to the total synthesis of the
antimicrotubule agent welwistatin is described by Funk’s
group.215 A key transformation of dioxin 504 into bicyclo[4.3.1]
decanone 506 involves an intramolecular conjugate addition
reaction with no indication of competing epimerization at
C(15) (Scheme 145).
In an attempt to synthesize the N-methylwelwitindoli-
none skeleton, an efficient and convergent synthesis of
the core bicyclo[4.3.1]decane ring system of welwitin-
dolinones was developed by Rawal and co-workers.216 A
key step in the synthesis includes an intramolecular pal-
ladium-catalyzed enolate arylation of bromoindole derivative
507 to create the desired bicyclic skeleton 508 (Scheme 146).
Interestingly, the O-arylated vinyl ether 509 was not observed.
An aldol-type intramolecular condensation in the presence
of triethylsilane allows the transformation of R,β-unsaturated
aldehyde 510 to the bridged indole 511, an intermediate in
the synthesis of welwitinodonlinone alkaloid skeleton (Scheme
147, eq 1).217 In contrast, this aldol-type ring-closure reaction
of 510 in the absence of triethylsilane generated a mixture
of cyclohexanone-bridged indoles 511 and 512 (Scheme 147,
eq 2).217
Chemical Reviews, 2010, Vol. 110, No. 3 1741
This acid-mediated transformation of 510 to the bridged
indole 511 can be illustrated in Scheme 148.217 Dehydration
of the alcoholic intermediate 513 leads to iminium species
514, whose subsequent interaction with 513 by hydride
transfer gives product 511 along with R,β-unsaturated ketone
512 in a 1:1 ratio. In the presence of external hydride source
(Et3SiH) the iminium 514 is reduced to 511 in high yield.
In addition, a key intermediate 516 that contains the
complete carbon framework of welwitindolinone was syn-
thesized by means of RCM with Grubbs’ catalyst (first
generation) (Scheme 149).218
9. Concluding Remarks
The recent syntheses of bridge-containing organic
compounds have great accomplishments as a result of the
most modern synthetic methods derived from total syn-
theses of natural products and pharmaceutical intermedi-
ates or APIs. Subsequent transformations of bridged ring
systems, in turn, have a huge impact on the organic
syntheses, especially on those of asymmetric synthetic
tasks. The design of general methods for obtaining
enantiopure compounds and, in particular, extension of
the most efficient methods described to asymmetric
synthesis, with special attention on transition-metal-
catalyzed reactions, continues to be of great interest for
synthetic chemists. As expected, exploration of expedient
and environmentally benign synthetic methods remains
as the main challenge for the synthetic community.
10. References
(1) (a) Smith, A. B., III; Liverton, N. J.; Hrib, N. J.; Sivaramakrishman,
H.; Winzenberg, K. J. Am. Chem. Soc. 1986, 108, 3040. (b) Grieco,
P. A.; Lis, R.; Zelle, R. E.; Finn, J. J. Am. Chem. Soc. 1986, 108,
5908. (c) Sainte, F.; Serckx-Poncin, B.; Hesbain-Frisque, A.-M.;
Ghosez, L. J. Am. Chem. Soc. 1982, 104, 1428. (d) Fitzsimmons,
B. J.; Leblanc, Y.; Rokach, J. J. Am. Chem. Soc. 1987, 109, 285.
(e) Shea, K. J.; Wise, S.; Burke, L. D.; Davis, P. D.; Gilman,
J. W.; Greeley, L. J. Am. Chem. Soc. 1982, 104, 5708. (f) Shea,
K. J.; Wada, E. J. Am. Chem. Soc. 1982, 104, 5715. (g) Shea,
K. J.; Fruscella, W. M.; Carr, R. C.; Burke, L. D.; Cooper, D. K.
J. Am. Chem. Soc. 1987, 109, 447. (h) Schmidt, R. R. Acc. Chem.
Res. 1986, 19, 250. (i) Hwang, Y. C.; Fowler, F. W. J. Org. Chem.
1985, 50, 2719. (j) Corey, E. J.; Danheiser, R. L. Tetrahedron
Lett. 1973, 14, 4477. (k) Maier, M. E.; Perez, C. Synlett 1998,
159. (l) Gharagozloo, P.; Miyauchi, M.; Birdshall, B.; Birdshall,
N. J. M. J. Org. Chem. 1998, 63, 1974. (m) Magnuson, S. R.;
Sepp-Lorenzino, L.; Rosen, N.; Danishefsky, S. J. J. Am. Chem.
Soc. 1998, 120, 1615. (n) Kozmin, S. A.; Rawal, V. H. J. Am.
Chem. Soc. 1997, 119, 7165.
(2) Bear, B. R.; Sparks, S. M.; Shea, K. J. Angew. Chem., Int. Ed. 2001,
40, 820.
(3) (a) Danishefsky, S. J.; Pearson, W. H.; Harvey, D. F.; Maring, C. J.;
Springer, J. P. J. Am. Chem. Soc. 1985, 107, 1265. (b) Danishefsky,
S. J.; Harvey, D. F. J. Am. Chem. Soc. 1985, 107, 6647. (c) Roush,
W. R.; Myers, A. G. J. Org. Chem. 1981, 46, 1509.
(4) For reviews, see:(a) Nakamura, I.; Yamamoto, Y. Chem. ReV. 2004,
104, 2127. (b) Zeni, G.; Larock, R. C. Chem. ReV. 2004, 104, 2258.
(5) (a) Nicolaou, K. C.; Sorenen, E. J. Classics in Total Synthesis I;
Wiley-VCH: New York, 1996. (b) Nicolaou, K. C.; Snyder, S. A.
Classics in Total Synthesis II; Wiley-VCH: New York, 2003. (c)
Nicolaou, K. C.; Snyder, S. A.; Montagnon, T.; Vassilikogiannakis,
G. Angew. Chem., Int. Ed. 2002, 41, 1668.
(6) Northrup, A. B.; MacMillan, D. W. C. J. Am. Chem. Soc. 2002, 124,
2458.
(7) For recent reviews of enantioselective Diels-Alder reactions, see:
(a) Maruoka, K. In Catalytic Asymmetric Synthesis, 2nd ed.; Ojima,
I., Ed.; Wiley-VCH: New York, 2000; p 467. (b) Evans, D. A.;
Johnson, J. S. In ComprehensiVe Asymmetric Catalysis; Jacobsen,
E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer: New York, 1999;
Vol. 3, p 1177. (c) Oppolzer, W. In ComprehensiVe Organic
Synthesis; Trost, B. M., Ed.; Pergamon Press: New York, 1991; Vol.
5. (d) Kagan, H. B.; Riant, O. Chem. ReV. 1992, 92, 1007. (e) Dias,
L. C. J. Braz. Chem. Soc. 1997, 8, 289.
1742 Chemical Reviews, 2010, Vol. 110, No. 3
(8) Ishihara, K.; Gao, Q.; Yamamoto, H. J. Am. Chem. Soc. 1993, 115,
10412.
(9) (a) Evans, D. A.; Barnes, D. M.; Johnson, J. S.; Lectka, T.; Matt, P.;
Miller, S. J.; Murry, J. A.; Norcross, R. D.; Shaughnessy, S. A.;
Campos, K. R. J. Am. Chem. Soc. 1999, 121, 7582. (b) Singh, R. S.;
Adachi, S.; Tanaka, F.; Yamauchi, T.; Inui, C.; Harada, T. J. Org.
Chem. 2008, 73, 212.
(10) Liu, X.; Snyder, J. K. J. Org. Chem. 2008, 73, 2935.
(11) Corey, E. J. Angew. Chem., Int. Ed. 2002, 41, 1650.
(12) Ahrendt, K. A.; C. Borths, C. J.; MacMillan, D. W. C. J. Am. Chem.
Soc. 2000, 122, 4243.
(13) (a) Northrup, A. B.; MacMillan, D. W. C. J. Am. Chem. Soc. 2002,
124, 2458. (b) MacMillan, D. W. C. Nature 2008, 455, 304.
(14) Sakakura, A.; Suzuki, K.; Ishihara, K. AdV. Synth. Catal. 2006, 348,
2457.
(15) Singh, R. P.; Bartelson, K.; Wang, Y.; Su, H.; Lu, S.; Deng, L. J. Am.
Chem. Soc. 2008, 130, 2422.
(16) Ishihara, K.; Fushimi, M. J. Am. Chem. Soc. 2008, 130, 7532.
(17) Adam, W.; Balci, M.; Pietrzak, B. J. Am. Chem. Soc. 1979, 101,
6285.
(18) Ashe, A. J., III J. Org. Chem. 1972, 37, 2053.
(19) (a) Vogt, P. F.; Miller, M. J. Tetrahedron 1998, 54, 1317. (b) Streith,
J.; Defoin, A. Synthesis 1994, 1107. (c) Kibayashi, C.; Aoyagi, S.
Synlett 1995, 873. (d) Bodnar, B. S.; Miller, M. J. J. Org. Chem.
2007, 72, 3929. (e) Krchnak, V.; Moellmann, U.; Dahse, H.-M.;
Miller, M. J. J. Comb. Chem. 2008, 10, 94.
(20) Beniazza, R.; Desvergnes, V.; Landais, Y. Org. Lett. 2008, 10, 4195.
(21) Montagnon, T.; Tofi, M.; Vassilikogiannakis, G. Acc. Chem. Res.
2008, 41, 1001.
(22) (a) Hu, T.; Curtis, J. M.; Oshima, Y.; Quilliam, M. A. S.; Walter,
J. A.; Watson-Wright, W. M.; Wright, J. L. C. J. Chem. Soc., Chem.
Commun. 1995, 2159. (b) Hu, T.; Burton, I. W.; Cembella, A. D.;
Curtis, J. M.; Quilliam, M. A. S.; Walter, J. A.; Wright, J. L. C. J.
Nat. Prod. 2001, 64, 308.
(23) (a) Uemura, D.; Chou, T.; Haino, T.; Nagatsu, A.; Fukuzawa, S.;
Zeng, S.-Z.; Chen, H.-S. J. Am. Chem. Soc. 1995, 117, 1155. (b)
Chou, T.; Kamo, O.; Uemura, D. Tetrahedron Lett. 1996, 37, 4023.
(c) Chou, T.; Haino, T.; Kuramoto, M.; Uemura, D. Tetrahedron
Lett. 1996, 37, 4027. (d) Takada, N.; Umemura, N.; Suenaga, K.;
Chou, T.; Nagatsu, A.; Haino, T.; Yamada, K.; Uemura, D.
Tetrahedron Lett. 2001, 42, 3491.
(24) Takada, N.; Umemura, N.; Suenaga, K.; Uemura, D. Tetrahedron
Lett. 2001, 42, 3495.
(25) Satake, M.; Ofuji, K.; Naoki, H.; James, K. J.; Fruey, A.; McMahon,
T.; Silke, J.; Yasumoto, T. J. Am. Chem. Soc. 1998, 120, 9967.
(26) Georgiou, T.; Tofi, M.; Montagnon, T.; Vassilikogiannakis, G. Org.
Lett. 2006, 8, 1945.
(27) Tofi, M.; Koltsida, K.; Vassilikogiannakis, G. Org. Lett. 2009, 11,
313.
(28) Baran, A.; Balci, M. J. Org. Chem. 2009, 74, 88.
(29) Sharma, P.; Lygo, B.; Lewis, W.; Moses, J. E. J. Am. Chem. Soc.
2009, 131, 5966.
(30) Hudon, J.; Cernak, T. A.; Ashenhurst, J. A.; Gleason, J. L. Angew.
Chem., Int. Ed. 2008, 47, 8885.
(31) Brown, N.; Luo, D.; Velde, D. V.; Yang, S.; Brassfield, A.; Buszek,
K. R. Tetrahedron Lett. 2009, 50, 63.
(32) (a) Petit, M.; Chouraqui, G.; Phansavath, P.; Aubert, C.; Malacria,
M. Org. Lett. 2002, 4, 1027. (b) Chouraqui, G.; Petit, M.; Phansavath,
P.; Aubert, C.; Malacria, M. Eur. J. Org. Chem. 2006, 1413. (c)
Winkler, J. D.; Kim, H. S.; Kim, S. Tetrahedron Lett. 1995, 36, 687.
(33) Maimone, T. J.; Voica, A.-F.; Baran, P. S. Angew. Chem., Int. Ed.
2008, 47, 3054.
(34) Studies toward vinigrol:(a) Paquette, L. A.; Guevel, R.; Sakamoto,
S.; Kim, I. H.; Crawford, J. J. Org. Chem. 2003, 68, 6096. (b)
Paquette, L. A.; Efremov, I.; Liu, Z. S. J. Org. Chem. 2005, 70,
505. (c) Paquette, L. A.; Efremov, I. J. Org. Chem. 2005, 70, 510.
(d) Paquette, L. A.; Liu, Z. S.; Efremov, I. J. Org. Chem. 2005, 70,
514. (e) Grise, C. M.; Tessier, G.; Barriault, L. Org. Lett. 2007, 9,
1545. (f) Morency, L.; Barriault, L. J. Org. Chem. 2005, 70, 8841.
(g) Morency, L.; Barriault, L. Tetrahedron Lett. 2004, 45, 6105. (h)
Devaux, J. G.; Hanna, I.; Lallemand, J. Y. J. Org. Chem. 1997, 62,
5062. (i) Devaux, J. F.; Hanna, I.; Lallemand, J. Y. J. Org. Chem.
1993, 58, 2349. (j) Gentric, L.; Hanna, I.; Ricard, L. Org. Lett. 2003,
5, 1139. (k) Mehta, G.; Reddy, K. S. Synlett 1996, 625. (l) Kito, M.;
Sakai, T.; Haruta, N.; Shirahama, H.; Matsuda, F. Synlett 1996, 1057.
(m) Kito, M.; Sakai, T.; Shirahama, H.; Miyashita, M.; Matsuda, F.
Synlett 1997, 219. (n) Matsuda, F.; Kito, M.; Sakai, T.; Okada, N.;
Miyashita, M.; Shirahama, H. Tetrahedron 1999, 55, 14369. (o)
Souweha, M. S.; Enright, G. D.; Fallis, A. G. Org. Lett. 2007, 9,
5163. (p) Tessier, G.; Barriault, L. Org. Prep. Proced. Int. 2007, 39,
311. (q) Devaux, J.-F.; Hanna, I.; Lallemand, J.-Y.; Prange, T.
J. Chem. Res. Syn. 1996, 32.
Zhao
(35) (a) Bucsi, I.; Molnr, A.; Bartk, M.; Olah, G. A. Tetrahedron 1995,
51, 3319. (b) Kato, M.; Tooyama, Y.; Yoshikoshi, A. Bull. Chem.
Soc. Jpn. 1991, 64, 50.
(36) (a) Wharton, P. S.; Hiegel, G. A. J. Org. Chem. 1965, 30, 3254. (b)
Grob, C. A.; Bolleter, M.; Kunz, W. Angew. Chem., Int. Ed. Engl.
1980, 19, 708. (c) Ochiai, M.; Ukita, T.; Iwaki, S.; Nagao, Y.; Fujita,
E. J. Org. Chem. 1989, 54, 4832. (d) Deslongchamps, P. Stereo-
electronic Effects in Organic Chemistry; Pergamon Press: Oxford,
1983; p 259.
(37) Denmark, S. E.; Baiazitov, R. Y. Org. Lett. 2005, 7, 5617.
(38) (a) Lease, T. G.; Shea, K. J. J. Am. Chem. Soc. 1993, 115, 2248. (b)
Shea, K. J.; Lease, T. G.; Ziller, J. W. J. Am. Chem. Soc. 1990, 112,
8627.
(39) (a) Sparks, S. M.; Chow, C. P.; Zhu, L.; Shea, K. J. J. Org. Chem.
2004, 69, 3025. (b) Molina, C. L.; Chow, C. P.; Shea, K. J. J. Org.
Chem. 2007, 72, 6816.
(40) Evans, D. A.; Johnson, D. S. Org. Lett. 1999, 1, 595.
(41) Nicolaou, K. C.; Wang, J.; Tang, Y. Angew. Chem., Int. Ed. 2008,
47, 1432.
(42) Larsson, R.; Sterner, O.; Johansson, M. Org. Lett. 2009, 11, 657.
(43) Morales, J. J.; Lorenzo, D.; Rodrigues, A. D. J. Nat. Prod. 1991,
54, 1368.
(44) Rodrı́guez, A. D.; Co′bar, O. M.; Martı́nez, N. J. Nat. Prod. 1994,
57, 1638.
(45) Crimmins, M. T.; Ellis, J. M. J. Org. Chem. 2008, 73, 1649.
(46) (a) Stanley, L. M.; Sibi, M. P. Chem. ReV. 2008, 108, 2887. (b) Nair,
V.; Suja, T. D. Tetrahedron 2007, 63, 12247.
(47) (a) Hamer, J.; Macaluso, A. Chem. ReV. 1964, 64, 473. (b) Kametani,
T.; Nagahara, T.; Honda, T. J. Org. Chem. 1985, 50, 2327. (c)
Mzengeza, S.; Yang, C. M.; Whitney, R. A. J. Am. Chem. Soc. 1987,
109, 276. (d) Mzengeza, S.; Whitney, R. A. J. Org. Chem. 1988, 53,
4074. (e) Varlamov, A.; Kouznetsov, V.; Zubkov, F.; Chernyshev,
A.; Shurupova, O.; Vargas, L.; Palma, A.; Rivero, J.; Rosas, A.
Synthesis 2002, 771.
(48) Ayala, S. L. G.; Stashenko, E.; Palma, A.; Bahsas, A.; Amero-Luis,
J. M. Synlett 2006, 14, 2275.
(49) Murahashi, S.-I.; Mitsui, H.; Shiota, T.; Tsuda, T.; Watanabe, S. J.
Org. Chem. 1990, 55, 1736.
(50) Flick, A. C.; Caballero, M. J. A.; Padwa, A. Org. Lett. 2008, 10,
1871.
(51) Macdonald, J. M.; Horsley, H. T.; Ryan, J. H.; Saubern, S.; Holmes,
A. B. Org. Lett. 2008, 10, 4227.
(52) Davison, E. C.; Fox, M. E.; Holmes, A. B.; Roughley, S. D.; Smith,
C. J.; Williams, G. M.; Davies, J. E.; Raithby, P. R.; Adams, J. P.;
Forbes, I. T.; Press, N. J.; Thompson, M. J. J. Chem. Soc., Perkin
Trans. 1 2002, 1494.
(53) For reviews on azomethine ylides, see:(a) Huisgen, R. In 1,3-Dipolar
Cycloaddition Chemistry; Padwa, A., Ed.; Wiley: New York, 1984;
Vol. 1, pp 1-177. (b) Tufariello, J. J. In 1,3-Dipolar Cycloaddition
Chemistry; Padwa, A., Ed.; Wiley: Chichester, U.K. 1984; Vol. 2,
pp 83168. (c) Gothelf, K. V.; Jørgensen, K. A. Chem. ReV. 1998,
98, 863. (d) Harwood, L. M.; Vickers, R. J. In The Chemistry of
Heterocyclic Compounds: Synthetic Applications of 1,3-Dipolar
Cycloaddition Chemistry toward Heterocycles and Natural Products;
Padwa, A., Pearson, W. H., Eds.; Wiley and Sons: New York, 2002;
Vol. 59, pp 169-252. (e) Nájera, C.; Sansano, J. M. Curr. Org. Chem.
2003, 7, 1105. (f) Coldham, I.; Hufton, R. Chem. ReV. 2005, 105,
2765. (g) Bonin, M.; Chauveau, A.; Micouin, L. Synlett. 2006, 2349.
(h) Pandey, G.; Banerjee, P.; Gadre, S. R. Chem. ReV. 2006, 106,
4484.
(54) Pandey, G.; Gupta, N. R.; Pimpalpalle, T. M. Org. Lett. 2009, 11,
2547.
(55) Yeom, H.-S.; Lee, J.-E.; Shin, S. Angew. Chem., Int. Ed. 2008, 47,
7040.
(56) Shapiro, N. D.; Toste, F. D. J. Am. Chem. Soc. 2007, 129, 4160.
(57) Oh, C. H.; Lee, J. H.; Lee, S. J.; Kim, J. I.; Hong, C. S. Angew.
Chem., Int. Ed. 2008, 47, 7505.
(58) (a) Seyferth, D.; Mai, V. A.; Gordon, M. E. J. Org. Chem. 1970, 35,
1993. (b) Adams, J.; Poupart, M. A.; Grenier, L.; Schaller, C.;
Ouimet, N.; Frenette, R. Tetrahedron Lett. 1989, 30, 1749.
(59) Kim, Y.; Kim, J.; Park, S. B. Org. Lett. 2009, 11, 17.
(60) Wender, P. A.; D’Angelo, N.; Elitzin, V. I.; Ernst, M.; Jackson-
Ugueto, E. E.; Kowalski, J. A.; McKendry, S.; Rehfeuter, M.; Sun,
R.; Voigtlaender, D. Org. Lett. 2007, 9, 1829.
(61) Harrowven, D. C.; Pascoe, D. D.; Demurtas, D.; Bourne, H. O.
Angew. Chem., Int. Ed. 2005, 44, 1221.
(62) Chen, Y.; Ye, S.; Jiao, L.; Liang, Y.; Sinha-Mahapatra, D. K.;
Herndon, J. W.; Yu, Z.-X. J. Am. Chem. Soc. 2007, 129, 10773.
(63) Luo, Y.; Herndon, J. W.; Cervantes-Lee, F. J. Am. Chem. Soc. 2003,
125, 12720.
(64) Chung, W. K.; Lam, S. K.; Lo, B.; Liu, L. L.; Wong, W.-T.; Chiu,
P. J. Am. Chem. Soc. 2009, 131, 4556.
(65) Sawada, Y.; Sasaki, M.; Takeda, K. Org. Lett. 2004, 6, 2277.
Bridge-Containing Organic Compounds
(66) Nakai, Y.; Kawahata, M.; Yamaguchi, K.; Takeda, K. J. Org. Chem.
2007, 72, 1379.
(67) Sasaki, M.; Hashimoto, A.; Tanaka, K.; Kawahata, M.; Yamaguchi,
K.; Takeda, K. Org. Lett. 2008, 10, 1803.
(68) Harmata, M.; Bohnert, G.; Kürti, L.; Barnes, C. L. Tetrahedron Lett.
2002, 43, 2347.
(69) Harmata, M.; Bohnert, G. J. Org. Lett. 2003, 5, 59.
(70) Harmata, M.; Rashatasakhon, P. Tetrahedron Lett. 2001, 42, 5593.
(71) Harmata, M.; Rashatasakhon, P. Org. Lett. 2001, 3, 2533.
(72) Du, Y.; Feng, J.; Lu, X. Org. Lett. 2005, 7, 1987.
(73) Trost, B. M.; McDougall, P. J.; Hartmann, O.; Wathen, P. T. J. Am.
Chem. Soc. 2008, 130, 14960.
(74) Rigby, J. H.; Fleming, M. Tetrahedron Lett. 2002, 43, 8643.
(75) Rigby, J. H.; Chouraqui, G. Synlett 2005, 16, 2501.
(76) Rigby, J. H.; Bazin, B.; Meyer, J. H.; Mohammadi, F. Org. Lett.
2002, 4, 799.
(77) Nair, V.; Abhilash, G.; Biju, A. T.; Suresh, E. Synthesis 2007, 12,
1833.
(78) Rigby, J. H.; Niyaz, N. M.; Bazin, B. Tetrahedron 2002, 58, 4879.
(79) Winkler, J. D.; Hong, B.-C.; Hey, J. P.; Williard, P. G. J. Am. Chem.
Soc. 1991, 113, 8839.
(80) Winkler, J.; Hershberger, P. J. Am. Chem. Soc. 1989, 111, 4852.
(81) Winkler, J.; Doherty, E. J. Am. Chem. Soc. 1999, 121, 7425.
(82) Winkler, J. D.; Hey, J. P. J. Am. Chem. Soc. 1986, 108, 6425.
(83) Winkler, J. D.; Rouse, M. B.; Greaney, M. F.; Harrison, S. J.; Jeon,
Y. T. J. Am. Chem. Soc. 2002, 124, 9726.
(84) Grota, J.; Domke, I.; Stoll, I.; Schröder, T.; Mattay, J.; Schmidtmann,
M.; Bögge, H. Synthesis 2005, 14, 2321.
(85) Sheikh, S. E.; Kausch, N.; Lex, J.; Neudörfl, J.-M.; Schmalz, H.-G.
Synlett 2006, 10, 1527.
(86) Yu, M.; Pagenkopf, B. L. J. Am. Chem. Soc. 2003, 125, 8122.
(87) Rath, J.-P.; Kinast, S.; Maier, M. E. Org. Lett. 2005, 7, 3089.
(88) Nicolaou, K. C.; Harrison, S. T. Angew. Chem., Int. Ed. 2006, 45,
3256.
(89) Wang, Q.-G.; Deng, X.-M.; Zhu, B.-H.; Ye, L.-W.; Sun, X.-L.; Li,
C.-Y.; Zhu, C.-Y.; Shen, Q.; Tang, Y. J. Am. Chem. Soc. 2008, 130,
5408.
(90) Stivala, C. G.; Zakarian, A. J. Am. Chem. Soc. 2008, 130, 3774.
(91) Nicolaou, K. C.; Majumder, U.; Roche, S. P.; Chen, D. Y.-K. Angew.
Chem., Int. Ed. 2007, 46, 4715.
(92) Yao, L.; Aube, J. J. Am. Chem. Soc. 2007, 2766.
(93) Iyengar, R.; Schildknegt, K.; Aube, J. Org. Lett. 2000, 1625.
(94) Iyengar, R.; Schildknegt, K.; Morton, M.; Aube, J. J. Org. Chem.
2005, 70, 10645.
(95) Borschberg, H.-J. Ph.D. Thesis, Eidenøssischen Technischen Hoch-
schule (ETH) Zûrich, No. 5578, 1975.
(96) Jenny, L.; Borschberg, H.-J. HelV. Chim. Acta 1995, 78, 715.
(97) For a review, see:Rodrı́guez, A. D.; González, E.; Ramı́rez, C.
Tetrahedron 1998, 54, 11683.
(98) Kingsbury, J. S.; Corey, E. J. J. Am. Chem. Soc. 2005, 127, 13813.
(99) Nicolaou, K. C.; Xu, H.; Wartmann, M. Angew. Chem., Int. Ed. 2005,
44, 756.
(100) For a review of hydrophobic effects, see:(a) Ben-Naim, A. Hydro-
phobic Interactions; Plenum: New York, 1980; p 303. (b) Tanford,
C. The Hydrophobic Effect, 2nd ed.; Wiley: New York, 1980; Chapter
1.
(101) For Diels-Alder reactions accelerated in water, see:(a) Breslow, R.
Acc. Chem. Res. 2004, 37, 471. (b) Rideout, D.; Breslow, R. J. Am.
Chem. Soc. 1980, 102, 7816. (c) Lindstroem, U. Chem. ReV. 2002,
102, 2751. (d) Harano, Y.; Sato, H.; Hirata, F. J. Am. Chem. Soc.
2000, 122, 2289. (e) Grieco, P. A.; Yoshida, K.; Garner, P. J. Org.
Chem. 1983, 48, 3137. (f) Grieco, P. A.; Kaufman, M. D. J. Org.
Chem. 1999, 64, 6041. (g) Otto, S.; Engberts, J. B. F. N.; Kwak,
J. C. T. J. Am. Chem. Soc. 1998, 120, 9517. (h) Blake, J. F.;
Jorgensen, W. L. J. Am. Chem. Soc. 1991, 113, 7430.
(102) Morita, H.; Kobayashi, J. J. Org. Chem. 2002, 5378.
(103) (a) Pfrengle, F.; Lentz, D.; Reiβig, H.-U. Angew. Chem., Int. Ed.
2009, 48, 3165. (b) Pfrengle, F.; Al-Harrasi, A.; Brüdgam, I.; Reiβig,
H.-U. Eur. J. Org. Chem. 2009, 282. (c) Al-Harrasi, A.; Reiβig, H.-
U. Angew. Chem., Int. Ed. 2005, 44, 6227. (d) Pulz, R.; Al-Harrasi,
A.; Reiβig, H.-U. Synlett 2002, 5, 817.
(104) López, F.; Castedo, L.; Mascareñas, J. L. J. Am. Chem. Soc. 2002,
124, 4218.
(105) (a) López, F.; Castedo, L.; Mascareñas, J. L. Org. Lett. 2005, 7, 287.
(b) Lüpez, F.; Mascareñas, J. L. Chem.sEur. J. 2007, 13, 2172.
(106) Herzon, S. B.; Myers, A. G. J. Am. Chem. Soc. 2005, 127, 5342.
(107) (a) Jackson, L. V.; Walton, J. C. Chem. Commun. 2000, 2327. (b)
Bella, A. F.; Jackson, L. V.; Walton, J. C. Org. Biomol. Chem. 2004,
2, 421.
(108) Hitchcock, S. A.; Pattenden, G. Tetrahedron Lett. 1992, 33, 4843.
(109) Donets, P. A.; Goeman, J. L.; Van der Eycken, J.; Robeyns, K.; Van
Meervelt, L.; Van der Eycken, E. V. Eur. J. Org. Chem. 2009, 793.
Chemical Reviews, 2010, Vol. 110, No. 3 1743
(110) Nicolaou, K. C.; Tria, G. S.; Edmonds, D. J. Angew. Chem., Int. Ed.
2008, 47, 1780.
(111) Maiti, A.; Gerken, J. B.; Masjedizadeh, M. R.; Mimieux, Y. S.; Little,
R. D. J. Org. Chem. 2004, 69, 8574.
(112) Trost, B. M.; Waser, J.; Meyer, A. J. Am. Chem. Soc. 2008, 130,
16424.
(113) (a) Baran, P. S.; Richter, J. M. J. Am. Chem. Soc. 2004, 126, 7450.
(b) Baran, P. S.; Richter, J. M.; Lin, D. W. Angew. Chem., Int. Ed.
2005, 44, 609.
(114) Baran, P. S.; Guerrero, C. A.; Ambhaiker, N. B.; Hafensteiner, B. D.
Angew. Chem., Int. Ed. 2005, 44, 606.
(115) Katritzky, A. R.; Zhang, S.; Kurz, T.; Wang, M. Org. Lett. 2001, 3,
2807.
(116) Rudler, H.; Parlier, A.; Sandoval-Chaves, C.; Herson, P.; Daran, J.-
C. Angew. Chem., Int. Ed. 2008, 47, 6843.
(117) Billard, T.; Langlois, B. R.; Large, S.; Anker, D.; Roidot, N.; Roure,
P. J. Org. Chem. 1996, 61, 7545.
(118) Li, J.; Todaro, L. J.; Mootoo, D. R. Org. Lett. 2008, 10, 1337.
(119) (a) Francisco, C. G.; Herrera, A. J.; Suárez, E. J. Org. Chem. 2002,
67, 7439. (b) Francisco, C. G.; Herrera, A. J.; Suárez, E. J. Org.
Chem. 2003, 68, 1012.
(120) Hashmi, A. S. K. Chem. ReV. 2006, 107, 3180.
(121) Barluenga, J.; Diéguez, A.; Fernández, A.; Rodrı́guez, F.; Faòanás,
F. J. Angew. Chem., Int. Ed. 2006, 2091.
(122) Antoniotti, S.; Genin, E.; Michelet, V.; Genêt, J.-P. J. Am. Chem.
Soc. 2005, 127, 9976.
(123) This step may also be catalyzed by gold, see:Hoffman-Roder, A.;
Krause, N. Org. Lett. 2001, 3, 2537.
(124) For reviews, see:(a) Jun, J.-G. Synlett 2003, 1759. (b) Kotsuki, H.
Synlett 1992, 97. (c) Zhang, Y.-J.; Nagao, T.; Tanaka, T.; Yang, C.-
R.; Okabe, H.; Kouno, I. Biol. Pharm. Bull. 2004, 27, 251.
(125) Oh, C. H.; Yi, H. J.; Lee, J. H. New J. Chem. 2007, 31, 835.
(126) Murphy, G. K.; West, F. G. Org. Lett. 2005, 7, 1801.
(127) Clark, J. S.; Hayes, S. T.; Wilson, C.; Gobbi, L. Angew. Chem., Int.
Ed. 2007, 46, 437.
(128) Wee, A. G. H. Curr. Org. Synth. 2006, 3, 499.
(129) Karche, N. P.; Jachak, S. M.; Dhavale, D. D. J. Org. Chem. 2001,
66, 6323.
(130) Muthusamy, S.; Srinivsan, P. Tetrahedron 2009, 65, 1567.
(131) For a theoretical investigation of this reaction, see:Weingarten, M. D.;
Prein, M.; Price, A. T.; Snyder, J. P.; Padwa, A. J. Org. Chem. 1997,
62, 2001.
(132) England, D. B.; Padwa, A. J. Org. Chem. 2008, 73, 2792.
(133) (a) Gmeiner, P.; Feldman, P. L.; Chu-Moyer, M. Y.; Rapoport, H. J.
Org. Chem. 1990, 55, 3068. (b) Takano, S.; Yonaga, M.; Morimoto,
M.; Ogasawara, K. J. Chem. Soc., Perkin Trans.1 1985, 305. (c)
Schultz, A. G.; Malachowski, W. P.; Pan, Y. J. Org. Chem. 1997,
62, 1223. (d) Lounasmaa, M.; Jokela, R. Heterocycles 1986, 24, 1663.
(e) Kuehne, M. E. J. Am. Chem. Soc. 1964, 86, 2949.
(134) Yotphan, S.; Bergman, R. G.; Ellman, J. A. J. Am. Chem. Soc. 2008,
130, 2452.
(135) Isolation of racemic murrayazoline:(a) Dutta, N. L.; Quasim, C.;
Wadia, M. S. Indian J. Chem. 1969, 7, 1061. (b) Kureel, S. P.; Dapil,
R. S.; Popli, S. P. Tetrahedron Lett. 1969, 3857. (c) Wu, T.-S.; Wang,
M.-L.; Wu, P.-L. Phytochemistry 1996, 43, 785. Isolation of (+)-
murrayazoline, see: (d) Furukawa, H.; Wu, T.-S.; Ohta, T.; Kuoh,
C.-S. Chem. Pharm. Bull 1985, 33, 4132.
(136) Wu, T.-S.; Chan, Y.-Y.; Liou, M.-J.; Lin, F.- W.; Shi, L.-S.; Chen,
K.-T. Phytother. Res. 1998, 12, S80.
(137) (a) Vorogushin, A. V.; Huang, X.; Buchwald, S. L. J. Am. Chem.
Soc. 2005, 127, 8146. (b) Shelby, Q.; Kataoka, N.; Mann, G.;
Hartwig, J. F. J. Am. Chem. Soc. 2000, 122, 10718.
(138) Ueno, A.; Kitawaki, T.; Chida, N. Org. Lett. 2008, 10, 1999.
(139) Paterson, I.; Razzak, M.; Anderson, E. A. Org. Lett. 2008, 10, 3295.
(140) Williams, P. G.; Asolkar, R. N.; Kondratyuk, T.; Pezzuto, J. M.;
Jensen, P. R.; Fenical, W. J. Nat. Prod. 2007, 70, 83.
(141) Ramana, C. V.; Induvadana, B. Tetrahedron Lett. 2009, 50, 271.
(142) Woo, G. H. C.; Kim, S.-H.; Wipf, P. Tetrahedron 2006, 62, 10507.
(143) (a) Nicolaou, K. C.; Pfefferkorn, J. A.; Kim, S.; Wei, H. X. J. Am.
Chem. Soc. 1999, 121, 4724. (b) Nicolaou, K. C.; Pfefferkorn, J. A.;
Cao, G.-Q.; Kim, S.; Kessabi, J. Org. Lett. 1999, 1, 807. (c) Usuda,
H.; Kanai, M.; Shibasaki, M. Org. Lett. 2002, 4, 859. (d) Spessard,
S. J.; Stoltz, B. M. Org. Lett. 2002, 4, 1943. (e) Young, D. G. J.;
Zeng, D. X. J. Org. Chem. 2002, 67, 3134. (f) Usuda, H.; Kanai,
M.; Shibasaki, M. Tetrahedron Lett. 2002, 43, 3621. (g) Kraus, G. A.;
Nguyen, T. H.; Jeon, I. Tetrahedron Lett. 2003, 44, 659. (h) Ciochina,
R.; Grossman, R. B. Org. Lett. 2003, 5, 4619. (i) Klein, A.; Miesch,
M. Tetrahedron Lett. 2003, 44, 4483. (j) Mehta, G.; Bera, M. K.
Tetrahedron Lett. 2004, 45, 1113. (k) Usuda, H.; Kuramochi, A.;
Kanai, M.; Shibasaki, M. Org. Lett. 2004, 6, 4387. (l) Nicolaou, K. C.;
Carenzi, G. E. A.; Jeso, V. Angew. Chem., Int. Ed. 2005, 44, 3895.
(m) Kuramochi, A.; Usuda, H.; Yamatsugu, K.; Kanai, M.; Shibasaki,
M. J. Am. Chem. Soc. 2005, 127, 14200.
1744 Chemical Reviews, 2010, Vol. 110, No. 3
(144) Siegel, D.; Danishefsky, S. J. J. Am. Chem. Soc. 2006, 128, 1048.
(145) (a) Yeung, Y.-Y.; Hong, S.; Corey, E. J. J. Am. Chem. Soc. 2006,
128, 6310. (b) Nie, L.-D.; Shi, X.-X.; Ko, K. H.; Lu, W.-D. J. Org.
Chem. 2009, 74, 3970.
(146) (a) Kim, C. U.; Lew, W.; Williams, M. A.; Liu, H.; Zhang, L.;
Swaminathan, S.; Bischofberger, N.; Chen, M. S.; Mendel, D. B.;
Tai, C. Y.; Laver, G.; Stevens, R. C. J. Am. Chem. Soc. 1997, 119,
681. (b) Rohloff, J. C.; Kent, K. M.; Postich, M. J.; Becker, M. W.;
Chapman, H. H.; Kelly, D. E.; Lew, W.; Louie, M. S.; McGee, L. R.;
Prisbe, E. J.; Schultze, L. M.; Yu, R. H.; Zhang, L. J. Org. Chem.
1998, 63, 4545. (c) Albrecht, S.; Harrington, P.; Iding, H.; Karpf,
M.; Trussardi, R.; Wirz, B.; Zutter, U. Chimia 2004, 58, 621.
(147) (a) Knapp, S.; Gibson, F. S. Organic Syntheses; Wiley & Sons: New
York, 1998; Collect. Vol. IX, p 516. (b) Knapp, S.; Levorse, A. T.
J. Org. Chem. 1988, 53, 4006.
(148) Nicolaou, K. C.; Koftis, T. V.; Vyskocil, S.; Petrovic, G.; Tang, W.;
Frederich, M. O.; Chen, D. Y.-K.; Li, Y.; Ling, T.; Yamada, Y. M. A.
J. Am. Chem. Soc. 2006, 128, 2859.
(149) For representative reviews, see:(a) Tietze, L. F. Chem. ReV. 1996,
96, 115. (b) Tietze, L. F.; Lieb, M. E. Curr. Opin. Chem. Biol. 1998,
2, 363. (c) Rodrı́guez, J. Synlett 1999, 505. (d) Zhu, J.; Bienaymé,
H. Multicomponent Reactions; Wiley-VCH: New York, 2005.
(150) Liéby-Muller, F.; Constantieux, T.; Rodriguez, J. J. Am. Chem. Soc.
2005, 127, 17176.
(151) Simon, C.; Peyronel, J. F.; Rodriguez, J. Org. Lett. 2001, 3, 2145.
(152) (a) Rinehart, K. L.; Holt, T. G.; Fregeau, N. L.; Stroh, J. G.; Keifer,
P. A.; Sun, F.; Li, L. H.; Martin, D. G. J. Org. Chem. 1990, 55,
4512. (b) Sakai, R.; Rinehart, K. L.; Guan, Y.; Wang, A. H.-J. Proc.
Natl. Acad. Sci. U.S.A. 1992, 89, 11456.
(153) (a) Zhou, B.; Guo, J.; Danishefsky, S. J. Org. Lett. 2002, 4, 43. (b)
Zhou, B.; Edmondson, S.; Padron, J.; Danishefsky, S. J. Tetrahedron
Lett. 2000, 41, 2039. (c) Zhou, B.; Guo, J.; Danishefsky, S. J.
Tetrahedron Lett. 2000, 41, 2043.
(154) Corey, E. J.; Gin, D. Y.; Kania, R. S. J. Am. Chem. Soc. 1996, 118,
9202.
(155) (a) Endo, A.; Yanagisawa, A.; Abe, M.; Tohma, S.; Kan, T.;
Fukuyama, T. J. Am. Chem. Soc. 2002, 124, 6552. (b) Chen, J.; Chen,
X.; Bois-Choussy, M.; Zhu, J. J. Am. Chem. Soc. 2006, 128, 87.
(156) For synthesis of analogous ring systems, see:(a) Martinez, E. J.;
Corey, E. J. Org. Lett. 2000, 2, 993. (b) Chan, C.; Heid, R.; Zheng,
S.; Guo, J.; Zhou, B.; Furuuchi, T.; Danishefsky, S. J. J. Am. Chem.
Soc. 2005, 127, 4596. (c) Chan, C.; Zheng, S.; Zhou, B.; Guo, J.;
Heid, R. M.; Wright, B. J. D.; Danishefsky, S. J. Angew. Chem., Int.
Ed. 2006, 45, 1749.
(157) Zheng, S.; Chan, C.; Furuuchi, T.; Wright, B. J. D.; Zhou, B.; Guo,
J.; Danishefsky, S. J. Angew. Chem., Int. Ed. 2006, 45, 1754.
(158) Prabhudas, B.; Clive, D. L. J. Angew. Chem., Int. Ed. 2007, 46, 9295.
(159) (a) Trost, B. M. Angew. Chem., Int. Ed. 1989, 28, 1173. (b) Heumann,
A.; Réglier, M. Tetrahedron 1995, 51, 975. (c) Trost, B. M.; Oslob,
J. D. J. Am. Chem. Soc. 1999, 121, 3057. (d) Trost, B. M.; Brickner,
S. J. J. Am. Chem. Soc. 1983, 105, 568.
(160) Artman, G. D., III; Grubbs, A. W.; Williams, R. M. J. Am. Chem.
Soc. 2007, 129, 6336.
(161) Reisman, S. E.; Ready, J. M.; Weiss, M. M.; Hasuoka, A.; Hirata,
M.; Tamaki, K.; Ovaska, T. V.; Smith, C. J.; Wood, J. L. J. Am.
Chem. Soc. 2008, 130, 2087.
(162) (a) O’Neill, B. T.; Yohannes, D.; Bundesmann, M. W.; Arnold, E. P.
Org. Lett. 2000, 2, 4201. (b) Yohannes, D.; Hansen, C. P.; Akireddy,
S. R.; Hauser, T. A.; Kiser, M. N.; Gurnon, N. J.; Day, C. S.; Bhatti,
B.; Caldwell, W. S. Org. Lett. 2008, 10, 5353.
(163) Chandra, A.; Pigza, J. A.; Han, J.-K.; Mutnick, D.; Johnston, J. N.
J. Am. Chem. Soc. 2009, 131, 3470.
(164) Cedrón, J. C.; Estévez-Braun, A.; Ravelo, A. G.; Gutiérrez, D.; Flores,
N.; Bucio, M. A.; Pérez-Hernández, N.; Joseph-Nathan, P. Org. Lett.
2009, 11, 1491.
(165) (a) Magnus, P.; Mathews, K. S. J. Am. Chem. Soc. 2005, 127, 12476.
(b) Wu, Y.-C.; Bernadat, G.; Masson, G.; Couturier, C.; Schlama,
T.; Zhu, J. J. Org. Chem. 2009, 74, 2046.
(166) Nicolaou, K. C.; Lim, Y. H.; Becker, J. Angew. Chem., Int. Ed. 2009,
48, 3444.
(167) Ramesh, N. G.; Hassner, A. Synlett 2004, 6, 975.
(168) Ge, H. M.; Zhu, C. H.; Shi, D. H.; Zhang, L. D.; Xie, D. Q.; Yang,
J.; Ng, S. W.; Tan, R. X. Chem.sEur. J. 2008, 14, 376.
(169) Ge, H. M.; Xu, C.; Wang, X. T.; Huang, B.; Tan, R. X. Eur. J. Org.
Chem. 2006, 5551.
(170) Nicolaou, K. C.; Wu, T. R.; Kang, Q.; Chen, D. Y.-K. Angew. Chem.,
Int. Ed. 2009, 48, 3440.
(171) Wang, M.; Wu, A.; Pan, X.; Yang, H. J. Org. Chem. 2002, 67, 5405.
(172) Jackson, K. L.; Henderson, J. A.; Motoyoshi, H.; Phillips, A. J.
Angew. Chem., Int. Ed. 2009, 48, 2346.
(173) Goeke, A.; Mertl, D.; Brunner, G. Angew. Chem., Int. Ed. 2005, 44,
99.
(174) Patir, S.; Uludag, N. Tetrahedron 2009, 65, 115.
Zhao
(175) Sapeta, K.; Kerr, M. A. Org. Lett. 2009, 11, 2081.
(176) Coldham, I.; Fernàndez, J.-C.; Price, K. N.; Snowden, D. J. J. Org.
Chem. 2000, 65, 3788.
(177) Albrecht, U.; Armbrust, H.; Langer, P. Synlett 2004, 1, 143.
(178) Liu, P.; Hong, S.; Weinreb, S. M. J. Am. Chem. Soc. 2008, 130,
7562.
(179) Maity, P.; Lepore, S. D. J. Am. Chem. Soc. 2009, 131, 4196.
(180) (a) Michaut, A.; Miranda-Garcı́a, S.; Menéndez, J. C.; Coquerel, Y.;
Rodriguez, J. Eur. J. Org. Chem. 2008, 4988. (b) Michaut, A.; Garcia,
S. M.; Menendez, J. C.; Rodriguez, J. Org. Lett. 2004, 6, 3075.
(181) Morehead, A., Jr.; Grubbs, R. Chem. Commun. 1998, 275.
(182) Tiefenbacher, K.; Mulzer, J. J. Org. Chem. 2009, 74, 2937.
(183) Malik, C. K.; Ghosh, S. Org. Lett. 2007, 9, 2537.
(184) (a) Fukuyama, T.; Xu, L.; Goto, S. J. Am. Chem. Soc. 1992, 114,
383. (b) Schkeryantz, J. M.; Danishefsky, S. J. J. Am. Chem. Soc.
1995, 117, 4722. (c) Katoh, T.; Itoh, E.; Yoshino, T.; Terashima, S.
Tetrahedron 1997, 53, 10229. (d) Yoshino, T.; Nagata, Y.; Itoh, E.;
Hashimoto, M.; Katoh, T.; Terashima, S. Tetrahedron 1997, 53,
10239. (e) Katoh, T.; Nagata, Y.; Yoshino, T.; Nakatani, S.;
Terashima, S. Tetrahedron 1997, 53, 10253.
(185) Uchida, I.; Takase, S.; Kayakiri, H.; Kiyoto, S.; Hashimoto, M. J. Am.
Chem. Soc. 1987, 109, 4108.
(186) (a) Judd, T. C.; Williams, R. M. Angew. Chem., Int. Ed. 2002, 41,
4683. (b) Judd, T. C.; Williams, R. M. J. Org. Chem. 2004, 69, 2825.
(187) (a) Suzuki, M.; Kambe, M.; Tokuyama, H.; Fukuyama, T. Angew.
Chem., Int. Ed. 2002, 41, 4686. (b) Suzuki, M.; Kambe, M.;
Tokuyama, H.; Fukuyama, T. J. Org. Chem. 2004, 69, 2831.
(188) Ducray, R.; Ciufolini, M. A. Angew. Chem., Int. Ed. 2002, 41, 4688.
(189) Paleo, M. R.; Aurrecoechea, N.; Jung, K.-Y.; Rapoport, H. J. Org.
Chem. 2003, 68, 130.
(190) Jimenez, L. S.; Colandrea, V. J.; Rajaraman, S. Org. Lett. 2003, 5,
785.
(191) Trost, B. M.; O’Boyle, B. M. Org. Lett. 2008, 10, 1369.
(192) Mithani, S.; Drew, D. M.; Rydberg, E. H.; Taylor, N. J.; Mooibrock,
S.; Dmitrienko, G. I. J. Am. Chem. Soc. 1997, 119, 1159.
(193) (a) Nicolaou, K. C.; Li, A.; Edmonds, D. J. Angew. Chem., Int. Ed.
2006, 45, 7086. (b) For review on the synthesis of platensimycin,
see: Mulzer, J.; Tiefenbacher, K. Angew. Chem., Int. Ed. 2008, 47,
2548.
(194) Nicolaou, K. C.; Edmonds, D. J.; Li, A.; Tria, G. S. Angew. Chem.,
Int. Ed. 2007, 46, 3942.
(195) Nicolaou, K. C.; Tang, Y.; Wang, J.; Stepan, A. F.; Li, A.; Montero,
A. J. Am. Chem. Soc. 2007, 129, 14850.
(196) Yeung, Y.-Y.; Corey, E. J. Org. Lett. 2008, 10, 3877.
(197) Ghosh, A. K.; Xi, K. J. Org. Chem. 2009, 74, 1163.
(198) For review on synthesis of bicycle[3.2.1]octanes, see:Filippini, M.-
H.; Rodriguez, J. Chem. ReV. 1999, 99, 27.
(199) Lee, D.; Zheng, J.-C.; Yun, S. Y. J. Am. Chem. Soc. 2009, 131, 8413.
(200) Lalic, G.; Corey, E. J. Org. Lett. 2007, 9, 4921.
(201) Nicolaou, K. C.; Pappo, D.; Tsang, K. Y.; Gibe, R.; Chen, D. Y.-K.
Angew. Chem., Int. Ed. 2008, 47, 944.
(202) (a) Giese, B. Radicals in Organic Synthesis: Formation of Carbon-
Carbon Bonds; Pergamon Press: Oxford, 1986. (b) Weinges, K.;
Reichert, H. Synlett 1991, 785.
(203) Zou, Y.; Chen, C.-H.; Taylor, C. D.; Foxman, B. M.; Snider, B. B.
Org. Lett. 2007, 9, 1825.
(204) Barton, D. H. R.; McCombie, S. W. J. Chem. Soc. Perkin Trans. 1
1975, 1571.
(205) (a) Ghosh, A. K.; Xi, K. Org. Lett. 2007, 9, 4013. (b) Ghosh, A. K.;
Xi, K. J. Org. Chem. 2009, 74, 1163.
(206) Kim, C. H.; Jang, K. P.; Choi, S. Y.; Chung, Y. K.; Lee, E. Angew.
Chem., Int. Ed. 2008, 47, 4009.
(207) Li, P.; Payette, J. N.; Yamamoto, H. J. Am. Chem. Soc. 2007, 129,
9534.
(208) Aoki, S.; Watanabe, Y.; Sanagawa, M.; Setiawan, A.; Kotoku, N.;
Kobayashi, M. J. Am. Chem. Soc. 2006, 128, 3148.
(209) (a) Nicolaou, K. C.; Sun, Y.-P.; Peng, X.-S.; Polet, D.; Chen, D. Y.-
K. Angew. Chem., Int. Ed. 2008, 47, 7310. (b) Simmons, E. M.;
Hardin, A. R.; Guo, X.; Sarpong, R. Angew. Chem., Int. Ed. 2008,
47, 6650. (c) Dai, M.; Danishefsky, S. J. Heterocycles 2009, 77, 157.
(d) Dai, M.; Danishefsky, S. J. Tetrahedron Lett. 2008, 49, 6610.
(e) Dai, M.; Wang, Z.; Danishefsky, S. J. Tetrahedron Lett. 2008,
49, 6613. (f) Kürti, L.; Czakó, B.; Corey, E. J. Org. Lett. 2008, 10,
5247. (g) Lee, H. M.; Nieto-Oberhuber, C.; Shair, M. D. J. Am. Chem.
Soc. 2008, 130, 16864. (h) Craft, D. T.; Gung, B. W. Tetrahedron
Lett. 2008, 49, 5931.
(210) Shenvi, R. A.; Guerrero, C. A.; Shi, J.; Li, C.-C.; Baran, P. S. J. Am.
Chem. Soc. 2008, 130, 7241.
(211) Czakó, B.; Kürti, L.; Mammoto, A.; Ingber, D. E.; Corey, E. J. J. Am.
Chem. Soc. 2009, 131, 9014.
Bridge-Containing Organic Compounds
(212) Yamashita, S.; Iso, K.; Hirama, M. Org. Lett. 2008, 10, 3413.
(213) For selected examples, see:(a) Middleton, D. S.; Simpkins, N. S.
Tetrahedron 1990, 46, 545. (b) Lee, E.; Tae, J. S.; Park, C. M.
Tetrahedron Lett. 1993, 34, 4831.
(214) Avendaño, C.; Menéndez, J. C. Curr. Org. Synth. 2004, 1, 65.
(215) Greshock, T. J.; Funk, R. L. Org. Lett. 2006, 8, 2643.
(216) MacKay, J. A.; Bishop, R L.; Rawal, V. H. Org. Lett. 2005, 7, 3421.
Chemical Reviews, 2010, Vol. 110, No. 3 1745
(217) Baudoux, J.; Blake, A. J.; Simpkins, N. S. Org. Lett. 2005, 7, 4087.
(218) Wood, J. L.; Holubec, A. A.; Stoltz, B. M.; Weiss, M. M.; Dixon,
J. A.; Doan, B. D.; Shamji, M. F.; Chen, J. M.; Heffron, T. P. J. Am.
Chem. Soc. 1999, 121, 6326.
CR9002402
1746 Chem. Rev. 2010, 110, 1746–1787

Ruthenium-Based Heterocyclic Carbene-Coordinated Olefin Metathesis

Catalysts†
Georgios C. Vougioukalakis‡ and Robert H. Grubbs*,§
Institute of Physical Chemistry, National Centre of Scientific Research “Demokritos”, 15310 Agia Paraskevi, Greece, and Division of Chemistry and
Chemical Engineering, California Institute of Technology, Pasadena, California 91125

Received July 6, 2009

Contents 10. Removal of Ruthenium Impurities from Metathesis 1778

Products and Ruthenium Recycling Strategies
1. Brief Introduction to Olefin Metathesis 1746 11. Decomposition Studies 1780
2. Introduction to the Use of Heterocyclic Carbenes 1748 12. Conclusions and Perspectives 1783
as Ancillary Ligands in Ruthenium-Based 13. Acknowledgments 1783
Metathesis Catalysts
14. References 1783
3. Heterocyclic Carbene Frameworks Used in 1749
Ruthenium-Based Metathesis Catalysts
3.1. Symmetrical Imidazol- and 1749 1. Brief Introduction to Olefin Metathesis
Imidazolin-2-ylidenes The fascinating story of olefin (or alkene) metathesis (eq
3.2. Unsymmetrical Imidazol- and 1754 1) began almost five decades ago, when Anderson and
Imidazolin-2-ylidenes Merckling reported the first carbon-carbon double-bond
3.3. Chiral Monodentate N-Heterocyclic Carbenes 1756 rearrangement reaction in the titanium-catalyzed polymeri-
3.4. Chiral Bidentate N-Heterocyclic Carbenes 1757 zation of norbornene.1 Nine years later, Banks and Bailey
3.5. Four- and Six-Membered Ring N-Heterocyclic 1758 reported “a new disproportionation reaction . . . in which
Carbenes olefins are converted to homologues of shorter and longer
3.6. 1,2,4-Triazol-5-ylidenes, Cyclic (Alkyl)(amino) 1759 carbon chains . . .”.2 In 1967, Calderon and co-workers
Carbenes, Thiazol-2-ylidenes, and Other named this metal-catalyzed redistribution of carbon-carbon
Heterocyclic Carbene Ligands double bonds olefin metathesis, from the Greek word
3.7. Carbene Biscoordination 1760 “µετŔθεση”, which means change of position.3 These
4. Phosphine-free Heterocyclic Carbene-Coordinated 1761 contributions have since served as the foundation for an
Ruthenium Catalysts amazing research field, and olefin metathesis currently
4.1. Chelating Alkoxybenzylidene Ligands 1761 represents a powerful transformation in chemical synthesis,
4.2. Chelating Thioether and Chelating Sulfoxide 1764 attracting a vast amount of interest both in industry and
Benzylidene Ligands academia.4-6
4.3. Mono- and Bis(pyridine)-Coordinated Catalysts 1765
4.4. Chelating Quinolin- and Quinoxalin-ylidenes 1767
4.5. Bidentate Alkylidenes Chelated through Imine 1768
Donors
4.6. 14-Electron Phosphonium Alkylidenes 1768
5. Ruthenium Alkylidene Variation: Fischer-Type 1769
Carbenes, Indenylidenes, Vinylidenes, Cyclic
Ruthenium Alkylidenes, and Other Alkylidene The generally accepted mechanism for olefin metathesis
Ligands was originally proposed by Chauvin and Hérisson in 1970.7
6. Variation of the Phosphine Ligand 1771 According to this mechanism, olefin metathesis proceeds
7. Anionic Ligand(s) Variation 1772 through metallacyclobutane intermediates, generated by the
7.1. Halides 1772 coordination of the olefin(s) to a metal alkylidene, via a series
7.2. Monodentate and Bidentate Aryloxides 1772 of alternating [2 + 2]-cycloadditions and cycloreversions
7.3. N,O-, P,O-, and O,O-Bidentate Ligands 1772 (Scheme 1).8 Because of the reversibility of all individual
steps in the catalytic cycle, an equilibrium mixture of olefins
7.4. Carboxylates and (Alkyl)sulfonates 1773
is obtained. For the metathesis to be productive and useful,
7.5. Nitrile- and Isonitrile-Coordinated 1774 it is necessary to shift the equilibrium in one direction. These
Alkylidene-Free Ruthenium Catalysts
inceptive mechanistic explorations, followed by highly
8. N-Heterocyclic Carbene-Coordinated 1775 sophisticated attempts to synthesize metal alkylidenes and
(η6-Arene)ruthenium Metathesis Catalysts
metallacyclobutanes, eventually led to the synthesis of the
9. N-Heterocyclic Carbene-Coordinated Ruthenium 1777 first well-defined olefin metathesis catalysts. However, as will
Catalysts Designed for Homogeneous Metathesis
in Water and Protic Solvents be also discussed in the following sections, there are many
details of the olefin metathesis mechanism that still remain
* Corresponding author. unclear.
†
This article is part of the Carbenes special issue.
‡
National Centre of Scientific Research “Demokritos”. The most important olefin metathesis subtypes are pre-
§
California Institute of Technology. sented in Scheme 2. The ring-opening metathesis polymer-
10.1021/cr9002424  2010 American Chemical Society
Published on Web 12/14/2009

Georgios C. Vougioukalakis was born on Crete, Greece, in 1976. He
received his B.Sc. from the University of Crete and his Ph.D. from the
same university under the direction of Michael Orfanopoulos. He then
moved to the California Institute of Technology where he worked for two
years with Robert H. Grubbs. After another postdoctoral year with Nikos
Hadjichristidis at the University of Athens, he joined the National Centre
of Scientific Research ″Demokritos″. His research interests include organic
and organometallic chemistry and photochemistry, homogeneous catalysis
and photocatalysis, and the development of polymeric structures and
nanostructures mostly related to energy issues and nanotechnology.
Robert H. Grubbs was born near Possum Trot, KY, in 1942. He received
his B.A. and M.Sc. degrees from the University of Florida working with
Merle Battiste and his Ph.D. from Columbia University for work with Ron
Breslow. After a postdoctoral year with Jim Collman at Stanford University,
he joined the faculty at Michigan State University. In 1978, he moved to
the California Institute of Technology, where he is now the Victor and
Elizabeth Atkins Professor of Chemistry. His research interests include
polymer chemistry, organometallic catalysis, and development of new
synthetic organic methodology.
ization (ROMP) of monomers containing strained, unsatur-
ated rings was one of the earliest commercial applications
of olefin metathesis.4,5,9 The driving force for ROMP is the
ring-strain release,10 upon going to the polymerized products.
The ring-strain release also determines the irreversible nature
of ROMP, as the pathway back to the cyclic compound(s)
has to overcome a significant thermodynamic barrier. Ring-
closing metathesis (RCM) is widely used in organic synthe-
sis. The driving force for RCM is primarily entropic, because
one substrate molecule affords two molecules of product;
furthermore, since the small molecules released from this
reaction are volatile (if not gaseous), RCM is practically
irreversible and can proceed to completion. On the other
hand, cross metathesis (CM) is more challenging than both
RCM and ROMP, as it lacks the entropic driving force of
RCM and the ring-strain release of ROMP, which can lead
to relatively low yields of the desired cross-product.5 For
Chemical Reviews, 2010, Vol. 110, No. 3 1747
Scheme 1. Catalytic Cycle of Olefin Metathesis
these reasons, CM has been an underutilized metathesis
transformation. Other types of olefin metathesis reactions
include acyclic diene metathesis polymerization (ADMET),
ring-opening cross-metathesis (ROCM), ring-rearrangement
metathesis (RRM), and ethenolysis (ethenolysis is the cross-
metathesis of ethylene with an internal olefin).4,5
Scheme 2. Most Common Types of Olefin Metathesis
Reactions
Early metathesis catalysts were multicomponent systems
formed in situ from transition-metal halides and main-group
metal alkyl cocatalysts. Some representative examples
include WCl6/EtAlCl2, WCl6/BuSn4, and MoO3/SiO2.4,5 Oc-
casionally, a third component had to be added to the catalytic
system as an activator, e.g., the Calderon catalyst (WCl6/
EtAlCl2/EtOH).3a However, these catalytic systems were of
limited use in organic synthetic applications, mainly because
of the harsh reaction conditions they require and their
prolonged initiation periods. Additionally, the propagating
species were neither quantitatively nor uniformly formed,
resulting in lack of reaction control. The first single-
component metathesis catalysts were based upon titanium,11
tantalum,12 and tungsten,13 with the synthesis of the first
members of these catalytic families reported in the late 1970s.
Later on, well-defined molybdenum-based catalysts were also
synthesized.14 Unfortunately though, despite the high catalytic
activity of these early transition-metal catalysts, their some-
what limited functional group tolerance and high sensitivity
toward oxygen and moisture render them difficult to use in
many cases.15,16 In addition to the necessity for careful
handling, time-consuming protecting-group strategies have
to be utilized when substrates bear alcohols or aldehydes.
Many of the functional group tolerance and oxophilicity
problems in these early transition-metal systems were ad-
dressed by the development of well-defined ruthenium-based
metathesis catalysts.6c,16,17 Although the first reports regarding
1748 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 1. First well-defined metathesis-active ruthenium alkylidene
complex.
Figure 2. First- and second-generation ruthenium-based metathesis
catalysts.
ill-defined ruthenium-catalyzed ROMP were published as
early as the 1960s, using RuCl3(H2O)n,18 this late transition
metal had to wait 20 more years until it came back into the
metathesis game in the late 1980s.19 Unlike its early
transition-metal counterparts, ruthenium was remarkably
tolerant toward oxygen, water, and functional groups, at least
in these early ill-defined systems. Furthermore, one of the
most important findings in these studies was the suggestion
that the active species was a ruthenium alkylidene.20 On this
basis, the synthesis of the first metathesis-active ruthenium
alkylidene complex (1, Figure 1) was accomplished in
1992;21 nevertheless, catalyst 1 showed relatively low
reactivity and was only effective in the ROMP of highly
strained olefins.
Although the basic structure of the currently used ruthenium-
based catalysts still resembles that of the original complex,
composed of a ruthenium alkylidene along with two anionic
and two neutral ligands, contemporary catalysts (2-5, Figure
2) are much more robust and functional group tolerant. For
example, first-generation catalyst 2 has much better functional
group compatibility than all of the early transition-metal olefin
metathesis initiators.22 Substitution of one of the tricyclo-
hexylphosphine ligands with the bulky N-heterocyclic car-
bene (NHC) ligand H2IMes produced ruthenium complex
3, which displays improved catalytic activity, maintaining
the high functional group tolerance and thermal stability of
2.23 This improvement has been attributed to the increased
affinity of the NHC-substituted ruthenium center for π-acidic
olefins relative to σ-donating phosphines (vide infra).24
Furthermore, substitution of the second phosphine ligand for
a bidentate alkylidene (complexes 4 and 5) led to ruthenium-
based catalysts with even higher thermal stability.25-27 More
recent studies have led to the development of ruthenium
catalysts that, among others,28 initiate asymmetric olefin
metathesis reactions,29-36 with applications in aqueous and
protic solvent systems,37-43 or even carry out the challenging
formation of tetrasubstituted carbon-carbon double bonds.44,45
All the above-mentioned developments in the field of
heterocyclic carbene-coordinated ruthenium-based metathesis
catalysts will be extensively discussed in the following
sections of this article.
Vougioukalakis and Grubbs
2. Introduction to the Use of Heterocyclic
Carbenes as Ancillary Ligands in
Ruthenium-Based Metathesis Catalysts
The earliest reports on the synthesis of carbene-
coordinated organometallic complexes by Öfele and
Wanzlick (complexes 6 and 7, Figure 3) date back to the
late 1960s.46 Lappert and co-workers followed this work with
a series of studies on the chemistry of late transition-metal
carbene complexes in the early 1970s.47 However, carbenes
did not draw much of the chemical community’s attention
until 1991, when Arduengo reported the isolation of the first
stable NHC (8, Figure 3).48 Ever since, the use of heterocyclic
carbenes has had a great impact both on organometallic
chemistry,49 where they are extensively utilized as ligands,
and on organocatalysis.50 Carbenes are strong Lewis bases,
acting as excellent σ-donors and poor π-acceptors, and afford
metal-carbon bonds that are usually less labile than the
related metal-phosphine bonds.49a,51 This decreased lability
of carbenes, compared to the phosphine ligands, is believed
to be one of the reasons for the improved thermal and
oxidative stability of the corresponding organometallic
complexes. Moreover, the electronic properties and the steric
environment of heterocyclic carbenes can be easily and, in
some cases, rationally modified, by changing the substituents
on the carbene framework, thereby fine-tuning the catalytic
properties of the resulting organometallic complexes.
Typically, heterocyclic carbene precursors are easily
accessible through well-established synthetic routes start-
ing from commercially available reagents.52 Many of the
existing carbenes are stable enough to be isolated; however,
their in situ generation and subsequent reaction with the
desired metal source is more straightforward and, conse-
quently, more popular. The most common methods for the
generation of free heterocyclic carbene ligands are illustrated
in Scheme 3. Thus, deprotonation of imidazolium or imi-
dazolinium salts with a strong base, such as potassium
hexamethyldisilazane (KHMDS) or potassium tert-butoxide
(KOt-Bu), afford the corresponding free NHCs (eq 5).23,26,52-54
Alternatively, the desired carbene species can be thermolyti-
cally generated from the related 2-trichloromethyl, 2-pen-
tafluorophenyl, 2-carboxylated, or 2-dithiocarboxylated “pro-
tected” NHC adducts (eq 6).53,55 The as-generated carbene
can be then either isolated or in situ reacted with the
appropriate ruthenium precursor, to afford the targeted
organometallic complex. This is usually achieved by displac-
ing one (or more) phosphine or pyridine ligand(s). In another
approach, especially useful when the in situ generated
carbenes tend to dimerize, Ag2O can be used to initially form
the corresponding heterocyclic carbene-Ag complexes (eq
7);56 these silver salts, which usually are very efficient
carbene-transfer agents, subsequently afford the desired
heterocyclic carbene-coordinated ruthenium species by trans-
metalation.
The heterocyclic carbenic frameworks used thus far in
ruthenium-based metathesis catalysts are illustrated in
Figure 3. First reported NHC-coordinated organometallic com-
plexes and isolable carbene.
 Chemical Reviews, 2010, Vol. 110, No. 3 1749
Scheme 3. Most Common Methods for the Generation of
Free Heterocyclic Carbene Ligands

Figure 4. Ruthenium-based heterocyclic carbene-coordinated

metathesis catalysts reported to date.

Figure 4.57 The most successful and well-studied ruthenium

catalysts bear either symmetrical or unsymmetrical imidazol-
or imidazolin-2-ylidenes. Triazol-5-ylidenes, tetrahydropy-
rimidin-2-ylidenes, and a four-membered ring diaminocar-
bene have also been utilized, to afford the corresponding
complexes. More recently, a series of ruthenium complexes Figure 5. First reported ruthenium-based carbene-coordinated
metathesis catalysts.
coordinated with cyclic (alkyl)(amino)carbenes and thiazol-
2-ylidenes have been synthesized. The synthesis, structure,
and the catalytic activity of all (pre)catalysts families in
Figure 4 will be thoroughly discussed in the following
sections.

3. Heterocyclic Carbene Frameworks Used in

Ruthenium-Based Metathesis Catalysts
3.1. Symmetrical Imidazol- and
Imidazolin-2-ylidenes
In 1998, Herrmann and co-workers reported the synthesis
of the first heterocyclic carbene-containing ruthenium-based
metathesis catalysts (9-13, Figure 5) in which both phos-
phine ligands were replaced by NHCs.58 Despite their high
stability, these complexes did not show a significant im-
provement in metathesis activity, mostly due to their slow
initiation rates, attributed, in these carbene biscoordinated
complexes, to one of the NHCs (less labile than phosphines),53,59
which has to dissociate from the metal center for the catalyst
to be initiated (vide infra).
Soon thereafter, the synthesis of heteroleptic ruthenium
complexes 14a and 14b (Figure 6), by combining a nonlabile
(1,3-dimesityl-imidazol-2-ylidene) with a labile (tricyclo-
hexylphosphine or triphenylphosphine) ligand, was publi-
shed.60-62 Species 14a and 14b exhibited not only higher
RCM activity affording even tetrasubstituted cycloolefins,
at that time out of reach of ruthenium catalysts,60 but also
improved thermal stability compared to the parent bis(tri-
Figure 6. Ruthenium-based metathesis catalysts 14-18.
cyclohexylphosphine) complex. Almost simultaneously, het-
eroleptic ruthenium-based metathesis catalysts 15-18 (Figure
6) were also reported.59,63,64
These first reports on heterocyclic carbene-coordinated
ruthenium catalysts paved the way for one of the major
breakthroughs in the field of olefin metathesis, which
occurred with the synthesis of second-generation complex
3 (Figure 7).23 The synthesis of carbene-coordinated ruthenium-
based complexes 19 and 20 (Figure 7) was also published
in that same work. The motivation for the preparation of
complexes 3, 19, and 20 originated from the expected
increased basicity of the corresponding saturated NHCs,65
when compared to their unsaturated counterparts in Figure
6.23 The increased basicity of these saturated NHCs was, in
turn, anticipated to translate into an increased activity of the
resulting ruthenium complexes. Complexes 3, 19, and 20
were synthesized in high yields, starting from the first-
1750 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 7. Ruthenium-based metathesis catalysts 3, 19, and 20.
generation catalyst (2), by substituting one of the phosphine
ligands with the in situ generated free carbenes.23 Complex
3 expanded the scope of ruthenium metathesis catalysts
significantly, as it was proved to be not only air-, water-,
and functional group-tolerant66,67 but also highly efficient in
the RCM of sterically demanding dienes,23 in the ROMP of
low-strain substrates,68 and in the realization of challenging
CM reactions.66,69 This kind of reactivity was previously
possible only with the more active, though highly air- and
water-sensitive, early transition-metal catalysts. What is
more, 3 remains efficient at catalyst loadings as low as 0.05
mol % for RCM and 0.0001 mol % for ROMP reactions.
An alternative, one-pot synthesis for complex 3 was pub-
lished a couple of years later.70 According to this report,
potassium t-amylate can be used for the deprotonation of
the carbene precursor, an imidazolinium salt, whereas
isolation of 3 can be achieved by simple filtration. Another
route toward the preparation of 3 was reported in 2003.53 In
this work, the NHC-alcohol or -chloroform adducts were
used as carbene precursors. These easy-to-synthesize and
easy-to-use “protected” forms of NHCs are air-stable and
easier to handle than their free carbene analogues, while the
desired carbenes can be readily released in solution, providing
direct access to the metal-NHC complexes.
The reasons for the increased activity of the second-
generation phosphine-containing catalysts, compared to their
first-generation analogues, have been investigated both
experimentally and theoretically. Initially, it was assumed
that the higher activity of the second-generation catalysts
originates from the higher electronic trans-influence of the
NHCs, compared to the phosphine ligands, that was in turn
expected to lead to a lower barrier to phosphine dissociation
(faster initiation). However, a series of mechanistic studies,24,71
later on confirmed by gas-phase experiments,72 showed that
this is not the case. Thus, whereas both the first- (2) and
second-generation (3) tricyclohexylphosphine-containing cata-
lysts were found to initiate via a dissociative mechanism
(Scheme 4), the initiation rate (k1) of 2 was measured to be
2 orders of magnitude higher than that of 3.73 Nevertheless,
the overall catalytic activity of 3 was found to be about 2
orders of magnitude higher than that of 2. To account for
these contradictory observations, it was proposed that the
partitioning (k2/k-1) between the coordination of the alkene
substrate (k2) and the phosphine ligand (k-1, return to the
resting state of the catalyst), by the corresponding 14-electron
ruthenium intermediates (I, Scheme 4), is about 4 orders of
magnitude greater for 3 relative to 2.24,71 Therefore, the
increased activity of the second-generation catalyst(s) was
rationalized on the basis of an increased affinity of the NHC-
Vougioukalakis and Grubbs
Scheme 4. Proposed Catalytic Mechanism of
Phosphine-Containing Ruthenium-Based Catalysts
substituted ruthenium center for π-acidic olefins relative to
σ-donating phosphines. It should be noted that improvement
of initiation alone does not necessarily lead to a better
catalyst, since catalyst efficiency depends on initiation,
phosphine rebinding, reaction of the 14-electron ruthenium
intermediate with olefin, and rate of catalyst decomposition.
As mentioned above, there are also a number of theoretical
studies, sometimes contradictory to each other, dealing with
the increased activity of the second-generation catalysts. For
instance, differences in the barriers to the rotation of the
ancillary neutral ligands (phosphine or NHC), in the 14-
electron ruthenacyclobutane intermediates, were initially
proposed as the source of the observed differences in activity
between the first- and second-generation catalysts.74 Later
on, it was suggested that these differences in reactivity
originate from the difference in the energy of the 14-electron
benzylidenes and the corresponding 14-electron ruthenacy-
clobutane intermediates.75 Other studies showed that both
steric and electronic effects dictate the differences in
reactivity.76,77 Finally, in a more recent work, it was proposed
that the difference in the initiation rates for phosphine- and
NHC-coordinated catalysts is determined by attractive non-
covalent interactions.78
In another significant contribution to the field of ruthenium-
based metathesis, the Hoveyda group reported the synthesis
of isopropoxystyrene-coordinated catalyst 5 in 2000,26 one
year after the report of its first-generation analogue 4 (Figure
8).25 The Blechert group published the synthesis of the same
phosphine-free catalyst almost simultaneously.27 Compared
to its phosphine-containing analogue 3, catalyst 5 shows
improved thermal stability, oxygen- and moisture-tolerance.
On the other hand, the decreased initiation rate of 5 quite
often comprises a major disadvantage. A variety of steric
and electronic modifications of the chelating benzylidene
Figure 8. Isopropoxystyrene-coordinated ruthenium catalysts 4
and 5.
 Chemical Reviews, 2010, Vol. 110, No. 3 1751
Scheme 5. Proposed Catalytic Mechanism of
Isopropoxystyrene-Coordinated Ruthenium Catalysts

ether ligand were aimed at resolving this problem. These

efforts will be discussed in section 4.1, which is exclusively
dedicated to the chelating benzylidene ether ligands.
The proposed catalytic mechanism of the chelating ben-
zylidene ether ruthenium complexes is slightly different from
that discussed above for the phosphine-containing complexes.
Initially, 14-electron intermediate II is formed through the Figure 10. NHC-coordinated ruthenium-based catalysts 24-27.
dissociation of the benzylidene ether chelating group (Scheme
5). Coordination of the alkene substrate, followed by
metathesis, leads to the formation of the catalytically active
species III and a molecule of isopropoxystyrene (or a related
derivative).25,26,79 Therefore, the initial catalyst, or precatalyst,
exists in equilibrium with the catalytically active species,
and when the olefin is completely consumed, the catalyst
may return to its resting state by rebinding the isopropoxy-
styrene that was eliminated at the first step (release/return
mechanism).25,26 Also note that both phosphine-containing
(Scheme 4) and phosphine-free catalysts (Scheme 5) provide
the same propagating species (III, Scheme 5) after a single
turnover.
In 2000, Nolan and co-workers reported the synthesis of
complex 21 (Figure 9), bearing two bulky aryl groups (2,6-
diisopropylphenyl), in an effort to study the influence of the Figure 11. Fluorinated ruthenium-based catalysts 28 and 29.
bulkiness of the NHCs on the catalytic activity of the
corresponding complexes.80 Likewise, the groups of Fürst- and improved ADMET polymerization efficiency compared
ner,81 Mol,82 and Wagener83 reported the synthesis and the to complex 22.83
catalytic activity evaluation of complexes 22 and 23 (Figure The Fürstner group has also published the preparation of
9), the saturated phosphine-containing and phosphine-free complex 24 (Figure 10).81 Remarkably, substitution of the
analogues of 21. At ambient temperature, catalyst 22 shows backbone of this unsaturated NHC with two chlorine atoms
effective turnover numbers 6 times higher than those of 3,82 has little effect on the reactivity of the resulting complex,
along with very high initiation rates (turnover number is despite the obviously altered electronics of the ligand. The
defined as the number of substrate molecules that are synthesis of complexes 2585 and 26,86 presented in Figure
converted into product per catalyst molecule); nevertheless, 10, was published in 2005; 25, bearing an internal double
its decomposition rate is also increased,83 especially when bond, was prepared in order to be used as a starting point
utilized in challenging transformations.84 In sharp contrast, for the preparation of further functionalized NHC-coordinated
phosphine-free catalyst 23 displays increased thermal stability catalysts.85 The unsaturated backbone in 25 remains intact
even at elevated temperatures, most probably due to the low
metathesis reactivity of unstrained cycloolefins such as
cyclohexenes. The reactivity of 25 in the RCM of N,N-diallyl
tosylamine was found to be slightly lower than that of the
second-generation isopropoxystyrene-coordinated ruthenium
catalyst 5 (Figure 8). Triphenylphosphine-containing complex
26 was reportedly purified by simple hexane washings (i.e.,
there was no need for column chromatography).86 Its catalytic
efficiency was evaluated in the self-CM of acrylonitrile,
where it showed activity similar to the second-generation
catalyst 3 (Figure 7). Complex 27 (Figure 10), bearing the
deuterated analogue of the NHC in 26, was prepared to
investigate the mechanism of olefin isomerization in me-
tathesis reactions carried out by H2IMes-containing ruthe-
nium catalysts.87
Figure 9. Ruthenium catalysts 21-23 bearing sterically demanding Complexes 28 and 29 (Figure 11), bearing o-fluorinated
NHCs. aryl groups on the NHC ligand, were reported in 2006.88
1752 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 12. Ruthenium-based catalysts 30-34 without o-substit-
uents on the N-bound aryl rings.
Phosphine-containing 28 catalyzes the RCM of diethyl
diallylmalonate with a significantly increased reaction rate
compared to both parent complexes 3 (Figure 7) and 5
(Figure 8), whereas, in the same benchmark reaction,
phosphine-free 29 shows a slower rate than both 3 and 5.
This contradictory behavior of 28 and 29 was ascribed to an
unprecedented fluorine-ruthenium interaction (Ru-F dis-
tance ) 3.2 Å) observed in the solid-state structure of 29
(IV, Figure 11) via X-ray diffraction.
Despite the development of the highly active and func-
tional group-tolerant catalysts described above, RCM to form
tetrasubstituted olefins remained one of the weaknesses of
ruthenium-catalyzed metathesis until 2007, when a series of
new catalysts with increased activity for this transformation
was published.44,45 On the basis of an earlier observation,
according to which catalysts with reduced bulk at the ortho
position of the two N-bound aryl groups of the NHCs exert
increased efficiency in the formation of sterically demanding
substrates,34,35 catalysts 30 and 31 (Figure 12) were designed
and synthesized.44 Indeed, these two catalysts, along with
complex 29 in Figure 11, performed significantly better than
all the ruthenium catalysts available at the time in the RCM
formation of many tetrasubstituted olefins. Further improve-
ments led to the syntheses of complexes 32-34 (Figure 12),
bearing the unsaturated analogue of the NHC in 31.44 Overall,
34 was the most efficient catalyst in that work regarding
RCM to afford tetrasubstituted olefins. It is also worth
mentioning that the biphenyl group on the isopropoxyben-
zylidene moiety of complex 34 is known to afford rapidly
initiating phosphine-free catalysts (refer to section 4.1).89
Unfortunately, however, the difficult preparation of catalyst
34 rendered its large-scale production uneconomical and
imposed significant drawbacks regarding its commercializa-
tion. Research in the direction of more easily prepared
catalysts eventually led to the evolution of 35-40 (Figure
13), the syntheses of which can be easily performed on a
large scale.45 Catalysts 35-40 proved to be very efficient in
Vougioukalakis and Grubbs
Figure 13. Ruthenium catalysts 35-40 with increased efficiency
in the formation of tetrasubstituted olefins via RCM.
Figure 14. Ruthenium-based catalysts 41-45 bearing bulky NHC
ligands.
the RCM of dimethallylmalonates, a family of sterically
demanding benchmark substrates, while N-tolyl complexes
35 and 38 were the most successful catalysts in that work.
Later on, 38 was also found to exert increased efficiency
for the formation of sterically challenging disubstituted
olefins by CM.90 On the basis of solution- and solid-state
structural data, and in conjunction with a series of theoretical
calculations, the outstanding catalytic activity of 35 and 38
toward sterically demanding substrates was proposed to result
from a significantly more open steric environment around
the ruthenium center.91 This was suggested to originate from
the accessibility of conformations in which the N-tolyl rings
are rotated away from approaching and coordinating olefins.
In two other recent reports, the syntheses of ruthenium-
based catalysts 41,92 42,92 and 43-4593 were described
(Figure 14). Complexes 41 and 42 were initially targeted
with the aim of increasing the diastereoselectivity of ring-
rearrangement metathesis reactions. Although both 41 and
42 were found to be of limited stability in solution, even in
the absence of olefin substrates, 41 indeed showed some
promising results in diastereoselective ring-rearrangement
metathesis reactions, affording improved E/Z selectivities.92
In addition, 42 led to the isolation of a ruthenium complex
relevant to the deactivation of NHC-coordinated ruthenium-
based metathesis catalysts (refer to section 11). On the other
hand, 43-45 were proven to be efficient RCM catalysts in
benchmark reactions: 43 and 44 showed reactivity similar
to that of the existing second-generation phosphine-
coordinated catalyst (3), whereas 45 outperformed 3 by about
an order of magnitude, in terms of turnover frequencies for
complete conversion.

Figure 15. NHC-backbone-substituted ruthenium catalysts 46-53.
A series of decomposition studies concerning NHC-
substituted ruthenium complexes, extensively discussed in
section 11 of the present article, has shown that N-aryl-
substituted NHC complexes without ortho-substituents on
the N-aryl groups are more prone to degradation compared
to complexes bearing ortho-substituted N-aryl NHCs.92,94,95
This lack of stability has been attributed to an easier rotation
of the N-aryl groups in the former, which brings the ortho-
aryl C-H bonds closer to the ruthenium center, thereby
facilitating degradation through C-H bond activation. For
this reason, it was anticipated that, by placing bulky
substituents on the backbone of the NHC, the rotation of
the N-aryl groups (about the N-C bond) should be restricted,
thereby rendering this decomposition pathway unfavorable.
In this context, complexes 46 and 47 (Figure 15), bearing a
tetramethyl-substituted NHC ligand, were synthesized, and
47 was the first stable ruthenium metathesis catalyst coor-
dinated with an N,N-diphenyl-substituted NHC with a
saturated backbone.96 Species 47 was proven to be an
efficient olefin metathesis catalyst, carrying out a series of
model RCM, CM, and ROMP reactions. 47 is also one of
the most efficient catalysts in the RCM of the sterically
demanding diethyl dimethallylmalonate.
A subsequent more-detailed study, concerning the effects
of NHC-backbone substitution on the efficiency of ruthenium
metathesis catalysts, led to the synthesis of complexes 48-53
(Figure 15).97 In that work, the catalytic activities of 48-53
were evaluated by the use of a highly sensitive Symyx robotic
system. Both backbone and aryl substitution were found to
significantly impact catalyst stability and activity. Thus, low
N-aryl bulk on the NHC ligand led to increased activity and
decreased stability, while increased backbone substitution
increased catalyst lifetimes and decreased reaction rates.
Furthermore, the relative importance of catalyst stability and
activity on efficiency was found to depend on the steric
encumbrance of the specific RCM reaction. Whereas for
substrates with low steric demands catalyst stability is
important for success at low catalyst loadings, for sterically
hindered substrates catalyst activity becomes more important
than catalyst stability.
Chemical Reviews, 2010, Vol. 110, No. 3 1753
Figure 16. Catalysts 54-73 bearing saturated and unsaturated
symmetrical NHC ligands.
In other work, based mostly on electrochemical and NMR
studies, imidazolylidene- and imidazolinylidene-coordinated
complexes 54-71 (Figure 16) were synthesized.98 The major
goal of this study was to investigate the existence of
intramolecular π-π interactions and whether such interac-
tions influence the electronic density at the ruthenium center
as well as the catalytic ability of the corresponding com-
plexes. The reactivity of some representative complexes in
selected RCM and CM reactions was found to systematically
depend on the electronic properties of substituents R (Figure
16). Complex 54, bearing the electron-donating NEt2 group,
was the most catalytically active complex. It was furthermore
suggested by the authors that the differences in reactivity
between saturated and unsaturated NHCs do not originate
from different electron densities at the ruthenium center, and
that the electron-donating abilities of the saturated and the
unsaturated NHCs (bearing the same substituents R) are
similar.
Complexes 72 and 73 (Figure 16), bearing a pH-responsive
NHC ligand, have also been prepared by Schanz and co-
workers.99 In organic solvents, and in the absence of acid,
72 and 73 show reactivity similar to that of the parent H2IMes
complexes 3 and 5 (Figures 7 and 8, respectively) in
representative RCM and ROMP transformations. Upon
addition of HCl, the NMe2 groups in 72 and 73 get
protonated, affording the corresponding dicationic complexes,
which show increased decomposition rates. A protocol was
developed to remove the residual ruthenium from RCM
reaction mixtures by acidification and subsequent filtration
of protonated 73 (refer also to section 10 dedicated to this
issue).
Phosphine-free complexes 74 and 75 (Figure 17) were the
first reported isolable ruthenium-based catalysts bearing
aliphatic side groups on both nitrogen atoms of their saturated
NHC ring.100 Despite their higher catalytic activity in the
ROMP of 1,5-cyclooctadiene, compared to the parent
bis(mesityl)-substituted catalyst 5 in Figure 8, both 74 and
75 show decreased activity in model RCM and CM reactions.
The authors suggested that this low efficiency originated from
1754 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 17. Catalysts 74 and 75 bearing saturated NHC ligands
with two aliphatic side groups.
the increased steric bulk of the alkyls relative to the usual
aromatic groups.
3.2. Unsymmetrical Imidazol- and
Imidazolin-2-ylidenes
The first report on ruthenium complexes coordinated with
unsymmetrical NHCs came from the Fürstner group in
2001.81 Specifically, 76-78 (Figure 18) were targeted as
complexes able to metathesize their own heterocyclic carbene
ligands, affording the corresponding chelates, with the goal
of regenerating themselves after the quantitative consumption
of the substrate. In that same work, complexes 79 and 80
(Figure 18), incorporating a silylether and a perfluoroalkyl
chain, respectively, were also reported. Complexes 76-80
are efficient in the RCM of N,N-dimethallyl-N-tosylamide
to form the corresponding tetrasubstituted carbon-carbon
double bond. The catalytic activity of 76-78 also showed a
systematic dependence on the tether length between the
alkene group and the ruthenium center. This effect was
postulated to depend on the different capacities of 76-78
to form chelate complexes in situ.
To further enhance the steric bulk and the electron-
donating ability of the NHC N-substituents in the second-
generation ruthenium catalysts, Mol and co-workers at-
tempted the synthesis of [RuCl2(dCHPh)(H2IAd)(PCy3)]
(H2IAd ) 1,3-di(1-adamantyl)-4,5-dihydroimidazol-2-yli-
dine).101 Although the synthesis of this symmetrical NHC
complex was proven impossible, supposedly due to the
increased three-dimensional bulkiness of the H2IAd ligand,
its unsymmetrical counterpart 81 (Figure 19), bearing the
mixed 1-adamantyl/mesityl ligand, was isolated as a green
solid in good yield. However, 81 was proved to be a very
poor olefin metathesis catalyst and only catalyzed the ROMP
of norbornene, one of the least challenging cyclic olefins to
polymerize. The explanation offered for the exceptionally
low reactivity of 81 was based on the extreme steric
hindrance imposed by its NHC ligand.
Unsymmetrical complexes 82a-82d and 83a-83c (Figure
19), bearing protected or unprotected hydroxyl groups on
Figure 18. Complexes 76-80 coordinated with unsymmetrical
NHC ligands.
Vougioukalakis and Grubbs
Figure 19. Ruthenium-based complexes 81-84 bearing unsym-
metrical NHC ligands.
Figure 20. Complexes 85-88 coordinated with unsymmetrical
NHCs.
the side-chains of their NHC ligands, were synthesized in
an attempt to prepare catalysts capable of undergoing
immobilization on various supports (for a discussion on
ruthenium metathesis catalysts tagged with insoluble materi-
als or soluble functionalities, refer to section 10).102 An
unanticipated molecular rearrangement was observed during
the deprotection of 82d under acidic or mildly basic
conditions. Thus, instead of the expected 83d, rearranged
84c with its neutral ligands in a cis-configuration was isolated
in high yield (Figure 19). The same phenomenon was
observed during immobilization attempts of 83b and 83c on
silica gel, when rearranged complexes 84b and 84a were
respectively isolated. It was speculated that this reorganiza-
tion process is promoted by the terminal hydroxyl groups.
The ability of the hydroxyl function to effect this transforma-
tion becomes increasingly facile as this group is brought
closer to the ruthenium center. cis-Configured 84a-84c are
active metathesis catalysts at elevated temperatures, where,
as suggested by 31P NMR data, they reconvert into their
trans-isomers 83b-83d.
The design of complexes 85-88 (Figure 20) was based
on the anticipation that the unsymmetrical nature of their
NHC ligands might alter the environment of key intermedi-
ates in the metathesis pathway, leading to improved E/Z
selectivity in CM reactions and diastereoselectivity in RCM
reactions.103 Moreover, the enhanced electron-donating ability
of alkyl substituents was anticipated to lead to increased
catalyst activity. Complexes 85-88 were synthesized from
commercially available reagents in good to high yields. NOE-
difference NMR experiments (NOE ) nuclear overhauser
effect) suggest that only one rotational isomer exists for both
85 and 86 in solution, with the benzylidene moiety located
directly under the mesityl ring of the NHC. Similarly, in the
solid state, a single isomer was isolated for 87 and 88, with
the mesityl group situated directly above the benzylidene
proton. In a model RCM reaction, complexes 85-88 showed

Figure 21. Unsymmetrical NHC-coordinated complexes 89-98.
activities similar to those of their parent bis(mesityl)-
substituted complexes 3 and 5 (Figures 7 and 8, respectively).
Moreover, 85 and 87 showed significantly different E/Z ratios
in selected CM transformations and improved selectivities
in a diastereoselective RCM reaction compared to the parent
complexes 3 and 5.
Ledoux, Verpoort, and co-workers synthesized and evalu-
ated another series of ruthenium catalysts coordinated with
N-alkyl-N-aryl-substituted NHCs (89-98, Figure 21).100,104,105
Phosphine-containing 89, 91, and 92 were demonstrated to
surpass the parent second-generation catalyst (3) in the
ROMP of 1,5-cyclooctadiene,104 whereas 93 showed only
fair metathesis activity.105 The decreased catalytic activity
of 90 was attributed to its increased steric bulk in close
proximity to the metal center.104 Phosphine-free complexes
94, 96, and 97 display reduced catalytic activity compared
to the parent bis(aryl) N-substituted symmetrical catalysts
(5 and 23, Figures 8 and 9, respectively).100
A family of ruthenium-based complexes bearing unsym-
metrical NHCs with fluorinated N-aryl groups (99-106,
Figure 22) has been also synthesized.106,107 These complexes
are readily accessible in one or two steps from commercially
available first-generation catalyst 2. Among others, 99-106
promote the RCM of diethyl diallylmalonate and diethyl
allylmethallylmalonate, the ROMP of 1,5-cyclooctadiene, and
the CM of allyl benzene with cis-1,4-diacetoxy-2-butene, in
some cases surpassing the existing second-generation cata-
lysts 3 and 5 in efficiency. Especially in the CM of allyl
benzene with cis-1,4-diacetoxy-2-butene, complexes 99-106
demonstrate similar or higher activity than the second-
generation ruthenium catalysts and, more importantly, afford
improved E/Z ratios of the desired cross-product at conver-
sion above 60%. This was quite an important finding, since,
as mentioned in the Introduction, compared to RCM and
ROMP, CM is an underutilized olefin metathesis transforma-
tion, not only because it lacks the entropic driving force of
RCM and the ring-strain release of ROMP, but also because
Chemical Reviews, 2010, Vol. 110, No. 3 1755
Figure 22. Fluorinated unsymmetrical NHC-coordinated com-
plexes 99-106.
Figure 23. Complexes 107-109 coordinated with 1-mesityl-3-
phenyl-substituted unsymmetrical NHC ligands.
it often leads to relatively low statistical yields of the desired
cross-product, as well as poor E/Z cross-product selectivity.108
The E/Z selectivity in CM reactions at high conversion is
usually governed by thermodynamic factors; that is, second-
ary metathesis promotes isomerization of the product to the
favored E isomer, with the E/Z selectivity being controlled
by the stability of the olefin isomers rather than the selectivity
of the catalyst.
The influence of the unsymmetrical NHC ligands in
99-106 on the initiation rate of the irreversible reaction of
these ruthenium complexes with butyl vinyl ether was also
studied.106,107 The measured rate constants and activation
parameters for all phosphine-containing catalysts in Figure
22 suggest rate-determining phosphine dissociation. On the
contrary, rate constants and activation parameters for the phos-
phine-free catalysts 5, 100, 102, 104, and 106 are indicative
of an associative mechanism.106,107,109 Finally, the synthesis
of the related Rh(CO)2Cl(NHC) complexes allowed for the
study of the electronic properties of all unsymmetrical NHC
ligands in 99-106, by measuring the corresponding carbonyl
stretching frequencies.107
Complexes 107-109, coordinated with the 1-mesityl-3-
phenyl-substituted NHC ligands presented in Figure 23, have
been also synthesized. 107 and 108 were prepared as model
complexes during a series of catalyst decomposition studies
1756 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 24. Unsymmetrical NHC-coordinated complexes 110-112.
(see section 11) and, being highly unstable, their metathesis
catalytic activity has not been evaluated.92 On the other hand,
109 was targeted as a stabilized analogue of 108 and
efficiently catalyzes the RCM of diethyl diallyl, diethyl
allylmethallyl, and diethyl dimethallylmalonate, as well as
the CM of allyl benzene with cis-1,4-diacetoxy-2-butene.96
In 2008, phosphine-containing complexes 110-112 (Fig-
ure 24), coordinated with unsymmetrical bis(N-aryl)-
substituted NHCs, were reported.98 As had been found earlier
for other similar unsymmetrical systems,106,107 complexes
110-112 exist as pairs of atropisomers.98 During a series of
electrochemical studies, it was also observed that the
orientation of the aryl rings relative to the benzylidene moiety
dictates the ruthenium redox potential. The metathetic
catalytic activity of 110-112 has not been evaluated.
3.3. Chiral Monodentate N-Heterocyclic Carbenes
Although the first ruthenium-based catalysts bearing chiral
monodentate NHCs110 (19 and 20, Figure 25) were reported
in 1999,23 the first asymmetric metathesis reaction catalyzed
by these kinds of complexes was published two years later.29
Besides 19 and 20, chiral complexes 113-116 (Figure 25)
Figure 25. Ruthenium complexes 19, 20, and 113-116 coordi-
nated with chiral monodentate NHCs.
Vougioukalakis and Grubbs
Figure 26. Possible geometries of the intermediate olefin complex.
were synthesized and evaluated. In these asymmetric com-
plexes, the chirality is transferred from the 4- and 5-positions
of the NHC imidazolyl ring to the N-bonded aromatic groups,
forcing the ortho-substituents of the N-aryl rings to reside
on the NHC-face opposite to the bulky groups on the
backbone (a so-called “gearing” effect). Complexes 113,
114a, 115a, and 116a are air-stable solids easily purified on
the bench by column chromatography, whereas 114b-116b
and 114c-116c can be generated in situ by the addition of
excess LiBr or NaI (vide infra). Crystallographic evidence
of the conformation of these chiral NHCs was obtained by
conversion of 114a to the corresponding bis(pyridine)
complex (see section 4.3).
Complexes 19, 20, and 113-116 were evaluated in the
enantioselective desymmetrization of achiral trienes (3-
allyloxy-2,4-dimethylpenta-1,4-dienes), also known as asym-
metric ring-closing metathesis (ARCM).29 It was found that
catalysts encompassing the (1R,2R)-1,2-diaminocyclohexane
group (19, 113, and 115) exhibit lower enantioselectivities
than those having the (1R,2R)-1,2-diphenylethylenediamine
moiety (20, 114, and 116). Replacement of the mesityl
substituents (in 19 and 20) with o-methylaryl (in 113 and
114) or o-isopropylaryl groups (in 115 and 116) also
increases the enantioselectivity. Finally, changing the halide
ligands from chlorides to iodides improves the enantiose-
lectivity; however, the conversion to the metathesized
products is simultaneously reduced, supposedly due to the
lower stability of the diiodide ruthenium intermediates. None
of these catalysts showed a significant temperature- or
solvent-dependent change in its enantioselectivity. Catalyst
116c afforded the highest enantiomeric excess measured in
that study.
On the basis of the ligand effects described above and the
stereochemical outcome of the studied reactions, a bottom-
face (trans to the NHC) olefin binding pathway (intermediate
V in Figure 26) was excluded. Among the two side-on (cis
to the NHC) olefin binding pathways in Figure 26, interme-
diate VII was favored, although VI was not excluded.29 On
the contrary, theoretical work published in 2004 suggested
bottom-face olefin binding (V, Figure 26), given that the
other two intermediates (VI and VII) were calculated to be
of remarkably higher energy.111 For a more detailed discus-
sion on olefin coordination and the geometry of ruthenacy-
clobutane intermediates, the reader may refer to section 4.6.
Subsequent studies, aimed at the enhancement of the
enantioselectivity displayed by 116, together with the goal
of expanding the substrate scope of ARCM, led to the
synthesis of chiral complexes 117-120 (Figure 27).34,35
While 117 and 118 showed enantioselectivities similar to
those of catalyst 116, 119 displayed increased enantioselec-
tivity to the extent that, in a number of substrates, even its
dichloride version (119a) could be used at very low catalyst
loadings (<1 mol %) to afford high enantiomeric excesses
and conversions.34 Thus, 116c and 119a were utilized to ring-
close alkenyl ether- and silyl ether-prochiral trienes, affording
five- to seven-membered rings; conversions up to >98% and
enantiomeric excesses up to 92% were obtained. The

Figure 27. Ruthenium complexes 117-120 bearing chiral mono-
dentate NHC ligands.
influence of solvent and temperature on conversion and
enantiomeric excess was also studied. Finally, according to
the olefin-binding pathways proposed, if the incoming olefin
binds cis to the NHC, the stereodefining interaction is the
face of the ruthenium to which the olefin binds. If, on
the other hand, the incoming olefin binds trans to the NHC,
the stereodefining interaction is the position of the alkylidene
under the N-bound aryl ring. Either way, the position of the
pendent olefin in the forming ring also plays an important
role in the transition state.
116-120 (Figures 25 and 27) were later proven to also
be highly active in asymmetric ring-opening cross-metathesis
(AROCM) reactions.35 118a was found to be the most
selective of dichloride catalysts 116-120, whereas the use
of diiodide catalyst 118b slightly improved the enantiomeric
excess values (up to 82% in the AROCM of norbornene
derivatives with styrene). Nevertheless, since diiodide cata-
lysts are generally less reactive than their dichloride coun-
terparts, the loading of 118b had to be increased to 3 mol %
to achieve activity similar to that observed with 1 mol %
loading of 118a. A ruthenium benzylidene, rather than a
ruthenium methylidene as the propagating species, along with
a trans coordination pathway were proposed to be operative
in these AROCM reactions. Moreover, in the same work,
the first examples of asymmetric cross-metathesis (ACM)
reactions, the most challenging of the asymmetric metathesis
transformations, were reported.35 The enantiomeric excess
values obtained in the ACM reactions of protected 1,4-
pentadien-3-ols, hexa-1,5-diene-3,4-diol, or hepta-1,6-diene-
3,5-diol with cis-1,4-diacetoxy-2-butene ranged from 37%
to 52%.
Collins and co-workers prepared another family of ruthe-
nium catalysts coordinated with chiral monodentate NHC
ligands (121-124, Figure 28).112,113 Instead of the 1,2-
diphenyl enantioinducting backbone of the NHCs in 20, 114,
and 116-120 (Figures 25 and 27), complexes 121-124
encompass a bulky 1,2-di-tert-butyl unit. Furthermore, the
NHCs in 121-124 are unsymmetrical, having one N-aryl
and one N-methyl substituent adjacent to the carbenic center.
121-124 showed high reactivities and high enantiomeric
excess values in representative ARCM reactions, without the
use of halide additives. It was also proposed that the barriers
to the rotation of the NHC ligands in 121-124 play a
significant role in determining the reactivity of the corre-
sponding catalysts.
Chemical Reviews, 2010, Vol. 110, No. 3 1757
Figure 28. Ruthenium complexes 121-124 bearing chiral mono-
dentate NHC ligands.
Figure 29. Ruthenium complexes 125-128 coordinated with chiral
monodentate NHCs.
In 2008, the synthesis of complexes 125-128 (Figure 29)
was reported by the groups of Buchmeiser, Blechert, and
Grisi.114,115 The catalytic activity of 125 was not investigated
extensively, since it was essentially prepared as the precursor
of the corresponding phosphine-free, pyridine-coordinated
complex, utilized in alternating copolymerizations (refer to
section 4.3).114 126-128, bearing saturated NHCs with two
nonaromatic N-substituents, were found to promote RCM,
ARCM, CM, and ROMP reactions.115 In these benchmark
ROMP and CM reactions, 126-128 showed activities
between that of the first- and second-generation phosphine-
containing catalysts (Figure 2, catalysts 2 and 3, respectively),
with 127 being the most catalytically active complex in the
series. In the CM of allyl benzene with cis-1,4-diacetoxy-
2-butene, 126-128 afforded improved E/Z ratios toward Z
double bond formation. It should be noted that 126, which
does not have chirality in the backbone of the NHC, was
completely unable to give enantioinduction in ARCM,
whereas, in the same transformation, 128 afforded modest
enantioselectivities (33%). This difference in reactivity was
a key observation and highlights the significance of chiral
substitution on the backbone and the minor role of the chiral
N-substituents in chirality induction with this catalyst type.
3.4. Chiral Bidentate N-Heterocyclic Carbenes
Hoveyda and co-workers have developed a series of
ruthenium complexes coordinated with bidentate NHC
1758 Chemical Reviews, 2010, Vol. 110, No. 3 Vougioukalakis and Grubbs

the low initiation rate of 129, were also reported in that same
work (for a detailed discussion on this issue, see section 4.1).
Complexes 130-135 showed enhanced catalytic activity, in
general requiring lower catalyst loadings than 129, and in
some cases promoting asymmetric reactions that cannot be
effected by 129.31,116 More recently, complex 136 (Figure
31), the iodide-containing analogue of complex 133, was also
Figure 30. Ruthenium complex 129 bearing a chiral bidentate reported.32 Both 133 and 136 were found to be efficient and
NHC. highly enantioselective, affording up to 98% enantiomeric
excesses in the AROCM of low-strain oxabicyclic olefins,
allowing access to a variety of 2,6-disubstituted pyrans.32
136 was shown to catalyze these AROCM reactions with
significantly higher asymmetric induction than 133.
The most significant drawback to the synthesis and,
therefore, the extensive use of the above binaphthyl-based
catalysts is their lengthy, chiral auxiliary directed synthesis.
To overcome these difficulties, Hoveyda and co-workers
synthesized biphenolate, NHC-coordinated complexes 137
and 138 (Figure 31).33 The synthetic route to the precursor
of the asymmetric carbene contained in 137 and 138 is
considerably shorter and, more significantly, does not require
the use of optically pure, axially chiral amino alcohols.
Although 137 is not stable to chromatography, it can be
prepared and in situ catalyze a series of AROCM reactions.33
On the contrary, iodide-containing 138, while less active than
137, can be chromatographically purified and promotes a
variety of AROCM reactions, in many cases affording higher
enantioselectivities than its binaphthyl-based analogues, 133
and 136.33,36,117

3.5. Four- and Six-Membered Ring N-Heterocyclic

Carbenes
With the purpose of investigating the role of the NHC
ring size on the activity of ruthenium-based catalysts,
additional structural modifications of the diaminocarbene
Figure 31. Ruthenium complexes 130-138 coordinated with chiral family led to the synthesis of complexes 139-141 (Figure
bidentate NHCs. 32). 139, coordinated with a six-membered NHC, was
synthesized in moderate yield via a four-step synthetic
ligands, bearing biphenolate or binaphtholate moieties that route.118 Unfortunately, however, 139 displayed lower cata-
displace one of the chlorides in the coordination sphere of lytic activity than its counterparts with five-membered NHCs,
ruthenium. More specifically, in 2002 they reported the most probably due to the increased steric bulk of its six-
synthesis of chiral complex 129 (Figure 30), bearing a membered NHC in close proximity to the ruthenium center.
bidentate binaphthol NHC moiety.30 This was the first report This pronounced steric influence, clearly observed in the
regarding a ruthenium-based metathesis catalyst in which the X-ray structure of 139, was proposed to disfavor olefin
chiral information of the NHC ligand is transferred directly
to the ruthenium center. Complex 129, isolated in >98%
diastereo- and enantiomeric purity without resolution, is air-
and moisture-stable and can be recycled at the end of the
reaction, by column chromatography, with up to 96% catalyst
recovery. Because of the less electronegative nature and the
increased steric bulk of the naphthoxide ligand in 129,
compared to the corresponding chloride ligand in parent
complex 5, 129 is less active, requiring longer reaction times
and elevated temperatures for the same transformations.
Nevertheless, 129 was proven to efficiently catalyze a series
of AROCM reactions to afford high enantioselectivities.30
Aiming at the enhancement of the activity of 129, by
increasing the electronegativity of the naphtholate moiety,
the Hoveyda group developed the trifluoromethyl-substituted
chiral NHC ligand incorporated in complexes 134 and 135
(Figure 31).31 Indeed, 134 and 135 showed reactivities more
than 2 orders of magnitude higher than parent complex 129.
A number of modifications of the chelating isopropoxyben- Figure 32. Complexes 139-141 bearing four- and six-membered
zylidene ligand (in complexes 130-133 and 135), addressing NHC ligands.

binding and/or ruthenacyclobutane formation, resulting in
reduced activity. Soon thereafter, complex 140, featuring a
six-membered NHC similar to that in 139 along with a
chelating benzylidene ether ligand, was also reported.119 140
is highly active in benchmark RCM and ROCM reactions
and moderately active in enyne metathesis. The significantly
different activity of 139 and 140 is most probably related to
the existence of the chelating benzylidene ether ligand in
140, versus the tricyclohexylphosphine in 139, rather than
to the difference in the backbone structure of the corre-
sponding NHCs.
In order to study the impact of employing more strained
diaminocarbene frameworks than in the ordinary five-
membered NHCs, complex 141 was synthesized.120 X-ray
crystallographic analysis of 141 showed that the geometry
around the nitrogen atoms is not strictly planar, indicating a
reduced pπ overlap between the carbenic carbon atom and
the adjacent nitrogen atoms. Moreover, by measuring the
carbonyl stretching frequencies in the corresponding
Rh(CO)2Cl(NHC) complex, it was found that the four-
membered NHC ligand in 141 is a slightly less effective
σ-donor than its dihydroimidazol-2-ylidene analogue. This
difference was attributed to the more bent carbene angle in
the former. Complex 141 was shown to catalyze selected
RCM, CM, and ROMP transformations, albeit in a slower
rate than the parent, phosphine-free second-generation
catalyst (5).
3.6. 1,2,4-Triazol-5-ylidenes, Cyclic (Alkyl)(amino)
Carbenes, Thiazol-2-ylidenes, and Other
Heterocyclic Carbene Ligands
The first report of the employment of a carbene structure,
other than the typical diaminocarbene framework in a
ruthenium-based metathesis catalyst, was published in 2001
by Fürstner et al. (complex 142, Figure 33).81 Despite the
straightforward synthesis of 142, given that the 1,2,4-triazol-
5-ylidene is commercially available, this complex is not an
efficient metathesis catalyst, due to its high instability in
solution.53,81 The rapid decomposition of 142 was proposed
to originate from the facile dissociation of the triazolylidene
from the metal center.53 Nevertheless, 142 was shown to exert
a high initial activity in the formation of tetrasubstituted
cycloalkenes,81 providing one of the first hints regarding the
Figure 33. 1,2,4-Triazol-5-ylidene-coordinated complex 142 and
cyclic (alkyl)(amino) carbene-coordinated complexes 143-145.
Chemical Reviews, 2010, Vol. 110, No. 3 1759
Figure 34. Ruthenium-based complexes 146-152 bearing thiazol-
2-ylidene ligands.
creation of a “more open” steric environment around the
ruthenium center in order to accommodate more sterically
demanding substrates; this effect is discussed above for
complexes 30-40 (Figures 12 and 13) and complexes 46,
47, and 51-53 (Figure 15).
In a more recent work, a series of ruthenium catalysts
bearing cyclic (alkyl)(amino) carbenes was synthesized and
characterized (143-145, Figure 33).121 Cyclic (alkyl)(amino)
carbenes are more σ-electron-donating than their conven-
tional NHC counterparts and, at the same time, introduce a
unique Cs- or C1-symmetric steric environment. These
unusual steric properties of cyclic (alkyl)(amino) carbenes
may have implications for the microscopic reversibility of
the olefin binding and cycloreversion steps along the me-
tathesis catalytic cycle. Complexes 143-145 are air- and
moisture-stable compounds that can be isolated and purified
by column chromatography in low (145) to high (144) yields.
In the solid state, 143-145 position the N-aryl rings of their
cyclic (alkyl)(amino) carbenes above the benzylidene moiety
and the quaternary carbon adjacent to the carbenic center
over the empty coordination site. 1H NMR spectroscopy data
suggest that the solid-state conformation of 143-145 is
maintained in solution. The catalytic activity of 143-145
was evaluated in RCM,121 CM, and ethenolysis reactions.122
The RCM efficiency of 143-145, in the formation of
representative di- and trisubstituted cycloalkenes, was found
to be comparable to that of the second-generation complexes
3 and 5 (Figures 7 and 8, respectively).121 In the CM of allyl
benzene with cis-1,4-diacetoxy-2-butene, 143-145 exhibit
lower E/Z ratios relative to most NHC-substituted com-
plexes.122 Furthermore, in the ethenolysis of methyl oleate,
143-145 afford good selectivity for the formation of terminal
olefins versus internal olefins (originating from undesired
self-metathesis and secondary metathesis), with complex 145
being the most efficient ethenolysis catalyst examined to date,
achieving 35 000 turnover numbers.
Another family of ruthenium-based metathesis catalysts,
coordinated with a series of thiazol-2-ylidene ligands
(146-152, Figure 34), was reported in 2008.123 The steric
1760 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 35. Homodinuclear ruthenium catalysts 153 and 154.
environment of the thiazol-2-ylidenes encompassed in
146-152 is unique, in the sense that they bear only one
exocyclic substituent adjacent to the carbenic center. The
synthesis of the thiazol-2-ylidene precursors, namely, the
corresponding 3-aryl-4,5-dimethylthiazolium chlorides, is a
two-step, straightforward procedure, while 146-152, which
are stable to chromatography, can be prepared in one step
from commercially available 2 or 4 (Figure 2). In the solid
state, the N-aryl substituents of the thiazol-2-ylidene ligands
are located above the empty coordination site of the
ruthenium center. This is a rather interesting find, since all
phosphine-free ruthenium complexes bearing unsymmetrical
carbenes with only one exocyclic aryl substituent adjacent
to the carbenic center, reported thus far, are isolated with
this aryl group located directly above the benzylidene proton.
Despite the decreased steric protection of their ligands,
complexes 146-152 were demonstrated to efficiently pro-
mote a series of benchmark RCM, macrocyclic RCM,
ROMP, and CM reactions, showing stability and activity
comparable to the conventional NHC-containing ruthenium
catalysts.123 The phosphine-free catalysts of this family were
found to be more stable than their phosphine-containing
counterparts. Upon removing the steric bulk from the ortho-
positions of the N-aryl group of the thiazol-2-ylidenes, the
phosphine-free catalysts lose stability, but when the substit-
uents become too bulky, the resulting catalysts show
prolonged induction periods. Among the five thiazol-2-
ylidene ligands examined, 3-(2,4,6-trimethylphenyl)- and
3-(2,6-diethylphenyl)-4,5-dimethylthiazol-2-ylidene afforded
the most efficient and stable catalysts (148 and 149). Unlike
all previously evaluated catalysts, the steric bulk of these
thiazol-2-ylidene-containing complexes is correlated to the
observed E/Z ratio of the cross-product in the cross-
metathesis reaction of allyl benzene with cis-1,4-diacetoxy-
2-butene. Thus, decreasing the steric demand of the ortho
substituents on the N-aryl groups from i-Pr to H results in
an increased kinetic E-selectivity from ∼4 to ∼6.5. In
addition, in the macrocyclic ring-closing of a 14-membered
lactone, the E/Z profile of catalysts 146-152 is completely
different than that of H2IMes catalysts 3 and 5 (Figures 7
and 8, respectively) and more similar to the stereoselectivity
displayed by the first-generation catalyst 2 (Figure 2).
Homodinuclear ruthenium complexes 153 and 154 (Figure
35) were designed as “double-centered” metathesis catalysts
that should, upon reaction with the appropriate R,ω-dienes,
promote cyclodimerization rather than cyclization or oligo-
merization.124 Indeed, although 153 is unstable and decom-
poses after some hours in solution, phosphine-free 154 was
shown to competently promote dimer ring-closing metathesis
Vougioukalakis and Grubbs
of dienes with the appropriate length, at the suitable effective
molarity. A trapping experiment suggested that both ruthe-
nium centers in homobimetallic 154 are simultaneously
metathetically active.
3.7. Carbene Biscoordination
Ruthenium-based complexes bearing two heterocyclic
carbene ligands are prepared in one or two steps via the
substitution of two or more labile ligands by the correspond-
ing carbenes. The ease of this carbene biscoordination has
been suggested to depend both on the steric and the electronic
properties of the utilized carbene(s).53,123 Moreover, according
to the established mechanistic model, one of the carbenes
has to dissociate from the metal center for a biscarbene
ruthenium complex to initiate.24-26,71,79 In this regard,
although it is generally accepted that carbene ligands bind
strongly to the metal centers,49,51 carbene dissociation/transfer
have been repeatedly proven feasible.53,100,105,125,126
As mentioned earlier, the first heterocyclic carbene
biscoordinated ruthenium complexes were reported in 1998
(9-13, Figure 5).58 These complexes were applied in the
ROMP of functionalized norbornenes,127 as well as in
benchmark RCM reactions,128 albeit displaying slow initiation
rates due to the relatively low lability of the corresponding
NHCs. Later on, carbene biscoordinated ruthenium com-
plexes 155 and 156 (Figure 36) were also prepared.53 The
complexes 155 and 156 could not be obtained directly from
first-generation catalyst 2 (Figure 2), even when a large
excess of the NHC was used. This was attributed to the
significant decrease of the phosphine exchange rate when
one of the tricyclohexylphosphine ligands was replaced by
a NHC.53 Nevertheless, bis-substitution was achieved using
complex 157 (Figure 36) bearing two labile pyridine ligands
(see section 4.3).129 Both 155 and 156 are highly stable and
can be purified by column chromatography on silica gel.53
In the solid state, both of the Ru-NHC distances in 156 are
longer than those in either of the corresponding monocarbene
complexes (complexes 14 and 3 in Figures 6 and 7,
respectively), which was suggested to originate from the
greater steric congestion in 156 and possibly also from a
more electron-rich ruthenium center. 155 shows low RCM
and ROMP activities below 40 °C, but this picture signifi-
cantly improves at 80 °C. Therefore, elevated temperatures
are required for an efficient initiation of 155 to occur. When
155 is heated in the presence of PCy3 or the first-generation
Figure 36. Carbene biscoordinated ruthenium complexes 155 and
156 and bis(pyridine) complex 157.

Figure 37. Heterocyclic carbene biscoordinated complexes
158-160.
Figure 38. Complexes 161-163 coordinated with chelating NHC
ligands.
catalyst 2 (Figure 2), the heteroleptic, second-generation
catalyst 3 (Figure 7) is formed, confirming NHC dissociation.
It should also be noted that a bimolecular NHC transfer
mechanism has been similarly proposed by Herrmann and
co-workers for the formation of complex 16 (Figure 6).63
Arylphosphines are generally weaker donor ligands and
bind more weakly to the metal centers than their alkylphos-
phine analogues (for a related discussion, refer to section
6). On this basis, complexes 158-160 (Figure 37) were
prepared in one step by simply reacting their corresponding
bis(triphenylphosphine) precursors with the appropriate NHC
ligand.126,130 158 was isolated as a crystalline, air-stable solid
and was fully characterized, whereas the isolation of 159
was proven impossible and was only observed in situ,
because of its high solubility.130 The metathetical catalytic
activities of 158 and 159 have not been evaluated. Moreover,
160 shows a rather low RCM activity at 40 °C, ascribed to
its slow initiation, but at 80 °C it efficiently ring-closes
diethyl diallylmalonate and diallyl malononitrile.
Carbene biscoordinated complexes 161-163 (Figure 38)
have been also prepared.131,132 161 and 162, bearing two
9-membered chiral bidentate NHCs, were isolated as air-
stable solids, but their metathesis activity has not yet been
evaluated.131 On the other hand, the catalytic activity of
“pincer” pyridine-dicarbene complex 163 was examined in
benchmark RCM and ROMP reactions.132 Both the first- and
Chemical Reviews, 2010, Vol. 110, No. 3 1761
Figure 39. Heterocyclic carbene biscoordinated complexes
164-166.
the second-generation phosphine-containing catalysts (com-
plexes 2 and 3, respectively, Figure 2) outperformed 163,
as the same metathesis transformations required higher
catalyst loadings and longer reaction time with 163.
More recently, the synthesis of complexes 164 and 165
(Figure 39) was accomplished by substituting both tricyclo-
hexylphosphine ligands in the first-generation catalyst 2
(Figure 2) in one step.105 The exclusive formation of the NHC
biscoordinated complexes 164 and 165, when even an
equimolar amount of the corresponding heterocyclic carbene
was utilized, was assigned to a higher phosphine exchange
rate in the heteroleptic (phosphine-NHC) intermediate
complex compared to its precursor 2. Interestingly, it was
also found that both NHCs on biscoordinated complexes 164
and 165 can be exchanged with an excess of tricyclohexy-
lphosphine, affording the initial bis(phosphine) complex (2).
164 and 165 were found to be catalytically active in the
ROMP of 1,5-cyclooctadiene at elevated temperatures, and
initiation was proposed to occur via NHC dissociation.
An analogous situation was observed during a series of
attempts to isolate the tricyclohexylphosphine-containing
ruthenium complex bearing a 3-(2,6-diisopropylphenyl)-4,5-
dimethylthiazol-2-ylidene, when only the formation of the
carbene biscoordinated complex 166 (Figure 39) was ob-
served by 1H NMR and high-resolution mass spectroscopy.123
In that case, it was proposed that the coordination of the
first bulky thiazol-2-ylidene ligand at the ruthenium center,
followed by the dissociation of the remaining tricyclohexy-
lphosphine and coordination of a second thiazol-2-ylidene,
is highly energetically favorable due to the formation of an
“empty pocket” in the coordination sphere of the intermediate
complex. This empty pocket was suggested to better accom-
modate the second unsymmetrical carbene ligand compared
to the C3-symmetric tricyclohexylphosphine. Complex 166
is not stable enough to be purified by column chromatography.
4. Phosphine-free Heterocyclic
Carbene-Coordinated Ruthenium Catalysts
4.1. Chelating Alkoxybenzylidene Ligands
As mentioned in section 3.1, the synthesis of the first
isopropoxystyrene-containing heterocyclic carbene-coordi-
nated catalyst (5, Figure 8) was independently and almost
1762 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 40. Ruthenium-based complexes 167-176 bearing chelat-
ing isopropoxybenzylidene ligands.
simultaneously published by the Hoveyda and the Blechert
groups in 2000.26,27 Compared to its phosphine-containing
counterpart (3, Figure 7), catalyst 5 displays enhanced
oxygen- and moisture-tolerance; however, its decreased
initiation rate presents a major disadvantage. To increase
initiation rate, either the bulk around the ether moiety is
increased or a para-electron-withdrawing substituent that
decreases the basicity of the ether group is introduced. In
all these cases, the catalyst activity is controlled by initiation.
After the first turnover, the properties of the catalytically
active species are the same for all. The modifications of the
chelating benzylidene ether moiety’s sterics and electronics
that have been attempted thus far are comprehensively
discussed below.
The first and rather adventitious advancement in this
direction came from the group of Blechert in 2002.133 Highly
active and air-stable 167 (Figure 40), bearing a binol-based
ruthenium alkylidene, was prepared as a potent ARCM
catalyst. Despite the absence of asymmetric induction in
ARCM, 167 showed a large improvement in catalytic
activity. This high reactivity was suggested to originate from
an improved leaving group ability of the binol-substituted
isopropoxystyrene, because of its increased steric bulk in
comparison with the isopropoxystyrene in the parent complex
(5, Figure 8). Soon thereafter, the synthesis of catalyst 168,
showing a markedly greater catalytic efficiency than either
5 or 167, and without any loss of stability in air, was
published.89 Again, the enhanced reactivity of 168 was
attributed to the faster dissociation of the phenyl-substituted
isopropoxystyrene ligand, owing to a weakening of the
chelation bond as a result of steric crowding. This assumption
was later confirmed via a series of ligand-exchange experi-
ments.134 A more efficient and practical synthetic route to
168 was also reported in the same work.
In order to obtain more detailed information regarding the
effect of the steric and the electronic environment of the
chelating isopropoxybenzylidene ligands on the rate of
Vougioukalakis and Grubbs
Figure 41. Ruthenium complexes 177-180 coordinated with
chelating isopropoxybenzylidene ligands.
metathesis, Blechert and co-workers prepared complexes
169-176 (Figure 40).135 By comparing the RCM activity of
5 and 168-176, and by taking into account the σm+ and σp+
values of the corresponding substituents, they concluded that
an increase in the electrophilicity of either substituent (R1
through R3), namely, decreasing the electron density at both
the RudC and the Ru-O bonds, leads to a faster initiation
rate, with the electronic character of the RudC bond being
the dominant factor. As expected, increased steric hindrance
ortho to the isopropoxy group was also found to enhance
initiation rates. In another publication, addressing the lower
catalyst loading limit that can effect a RCM transformation,
complexes 177-180 (Figure 41) were reported.136 177-179
were found to deliver similar turnover numbers to each other,
whereas the sterically modified 180 delivered lower turnover
numbers, perhaps due to its faster initiation rate, which was
proposed to result in faster decomposition.
A variety of electronically and sterically modified catalysts,
bearing chelating alkoxybenzylidene ligands, have been also
synthesized by Grela and co-workers. Thus, 181 (Figure 42)
was found to be significantly more reactive than its parent
complex (5, Figure 8), without being less air- or moisture-
stable.137 The drastically increased reactivity of 181 was
rationalized in terms of a reduced chelating ability of the
isopropoxy fragment, due to a decrease in the electron density
of the corresponding oxygen atom. Moreover, complex 182
(Figure 42), encompassing an inexpensive and easily acces-
sible chelating methoxystyrene fragment, was synthesized
as a cheap alternative of the parent catalyst (5, Figure 8).138
182 was successfully tested in RCM, CM, and enyne
metathesis reactions, showing slightly improved reactivities
in comparison with 5. A more detailed study regarding the
activity of catalysts bearing nitro-substituted chelating alkox-
ybenzylidenes (181 and 183-187, Figure 42) in RCM, CM,
and enyne metathesis reactions was subsequently pub-
lished.139 It was found that catalysts coordinated with meta-
and para-nitro-substituted isopropoxybenzylidenes (181 and
183) are significantly more active than the parent catalyst 5,
although not as reactive as 168 (Figure 40). On the other
hand, attempts to combine electronic and steric activation
in the same isopropoxybenzylidene induced a drastic decrease
in the stability of the resulting complexes (184-186);
furthermore, 187 proved to be a less efficient catalyst than
its isopropyl counterpart (181). Subsequently, an improved
synthetic route to various 3- and 5-substituted alkoxystyrenes,
and a practical, large-scale preparation of the corresponding
NHC-coordinated alkoxybenzylidene complexes, was also
published.140,141
Further research led to the synthesis of phosphine-free
complexes 188 and 189 (Figure 43), the initiation efficiency

Figure 42. Ruthenium-based complexes 181-187 bearing chelat-
ing alkoxybenzylidene ligands.
Figure 43. Ruthenium complexes 188-192 coordinated with
chelating isopropoxybenzylidene ligands.
of which can be controlled on demand.142 Thus, in the
absence of acid, 188 shows no activity in the RCM of diethyl
allylmethallylmalonate, whereas its in situ formed salts (VIII)
via the addition of 1 equiv of organic acids are highly active,
in one case outperforming even the parent complex 5 (Figure
Chemical Reviews, 2010, Vol. 110, No. 3 1763
Figure 44. Complexes 193-197 bearing chelating alkoxyben-
zylidene ligands.
8). Similarly, the reactivity of 189 can be dramatically
enhanced using Ph2SnCl2, due to the formation of carbocation
IX (Figure 43). Complexes 190-192 have been also
prepared.139,143 190 and 191 are slightly less efficient than
the parent catalyst 5, while the metathetical catalytic behavior
of 192 has not been reported.
In complexes 193-195 (Figure 44), the aliphatic end group
of the styrenyl ether has been functionalized by the attach-
ment of an ester or an acid moiety.144,145 “Scorpio catalysts”
193 and 194 were shown to be highly efficient in model
RCM and CM reactions, sometimes outperforming even the
very active 168 (Figure 40).144 Interestingly, 194, which
combines a coordinating ester function with an electron-
withdrawing NO2 group, requires 10 times lower catalyst
loadings than the parent NO2-bearing complex 181 (Figure
42) in order to afford the same result under identical
conditions in a model CM reaction. Moreover, during an
attempt to prepare the free carboxylic acid analogue of 193,
ruthenium carboxylate 195 was isolated in 84% yield.145
Although catalytically dormant, 195 can be chemically (via
the addition of 1 equiv of acid) and thermally activated in
situ, promoting a series of RCM and CM reactions. 196 and
197 (Figure 44) were also synthesized and evaluated in
benchmark RCM reactions, mainly because their isopro-
poxybenzylidene moiety can be simply prepared in three
steps from inexpensive starting materials.146 Both 196 and
its dimeric analogue 197 show catalytic activity higher than
that of the parent complex 5.
Additional studies led to the synthesis of complexes
198-205 (Figure 45). 198 was prepared with the purpose
of carrying out metathesis reactions both in organic and
aqueous solvents, as well as in ionic media (for a survey on
water-soluble and ionic liquid tagged catalysts, see sections
9 and 10, respectively).147 However, although highly active
in model RCM reactions, the recyclability levels of 198 in
ionic solvents were very low. Complexes 199, 201, and 202
were also shown to be significantly more active than the
1764 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 45. Ruthenium-based complexes 198-205 coordinated
with chelating alkoxybenzylidene ligands.
Figure 46. Ruthenium complexes 206-210 coordinated with
chelating isopropoxyarylidene ligands.
parent complex 5 (Figure 8);148 nevertheless, 201 and 202
are also less thermally- and air-stable and, therefore, less
efficient than both 199 and 5. Finally, phosphine-free
catalysts 203-205 were targeted in view of the tunable steric
and electronic properties of their aminocarbonyl group
(amide or carbamate).149 203-205 initiate faster than 5, while
204 is the best-performing and 203 is the worst-performing
catalyst in the series. These differences in the initiation rates
were rationalized on the basis of the switchable electronic
properties of the aminocarbonyl function.
In 2008, Barbasiewicz et al. reported the preparation of
complexes 206-210 (Figure 46), bearing five different
isopropoxyarylidene chelates.150 207, which was found to
be the most efficient RCM catalyst in the series, showed
metathesis activity similar to 5 (Figure 8). On the contrary,
systems 206 and 208 were found to be latent, initiating only
Vougioukalakis and Grubbs
Figure 47. Complexes 211-215 coordinated with chelating
benzylidene ligands.
at temperatures above 110 °C. These pronounced differences
in the catalytic activity of 206-210, also taking into
consideration a series of structural and spectroscopic studies,
were attributed to the partially aromatic character of the
ruthenafurane ring, present in all ruthenium alkoxyben-
zylidenes, which inhibits initiation and thereby decreases the
catalytic activity of the complexes.
An analogous class of NHC-coordinated complexes bear-
ing carbonyl- or carboxyl-substituted chelating benzylidene
ligands (211-215) is presented in Figure 47. 211 is
significantly less metathesis active than both first-generation
catalyst 2 (Figure 2) and its heteroleptic (phosphine-NHC)
analogues 14a and 14b (Figure 6).151 On the contrary,
complexes 212-215, which have an unusual cis-dichloro
arrangement, most commonly observed in complexes bearing
strong chelating ligands such as pyridyl or quinoline (vide
infra), were found to be thermally switchable in the ROMP
of functionalized norbornenes.152 For example, aldehyde
derivative 212, which showed the lowest initiation efficiency
in this study, is barely active at room temperature but is an
efficient polymerization catalyst at temperatures higher than
45 °C. In the ester series (213-215), initiation efficiencies
were found to increase with the decreasing steric bulk of
the alkoxy substituent (R). It should be noted that the
development of thermally switchable, chemically switchable,
or photoswitchable catalysts is highly important in polymer
chemistry, as quite often the mixing of the monomer with
the catalyst and the polymerization reaction have to be carried
out at different times (and/or reactors) and, therefore,
accurately controlled. This is vital in processes such as
spraying or inkjet printing, where a constant and low
viscosity is required, or when the monomer/catalyst mixture
has to be shaped and profiled prior to polymerization
(“curing”).
4.2. Chelating Thioether and Chelating Sulfoxide
Benzylidene Ligands
In 2008, Lemcoff and co-workers published the synthesis
of ruthenium-based complexes 216a and 217-220 (Figure
48) bearing a series of chelating thioether benzylidene
ligands.153,154 As demonstrated by NMR experiments and
single-crystal X-ray analysis data, these complexes display
a cis-dichloro arrangement similar to that of 212-215 (Figure
47). 216a and 217-220 are highly stable toward oxygen and
moisture and, equally importantly, thermally switchable. For
example, catalyst 216a could be repeatedly switched on and
off in the RCM of diethyl diallylmalonate by heating to 80
°C and cooling to 25 °C, respectively.153 It was also found
that the initiation efficiency of 216a and 217-220 systemati-
cally depends on the steric bulk of substituent R, with 218
being the most reactive catalyst in the series.

Figure 48. Complexes 216-227 coordinated with chelating
thioether and chelating sulfoxide benzylidene ligands.
More recently, Grela and co-workers reported the synthesis
and catalytic activity evaluation of complexes 221-227 and
216b, an isomer of 216a having the usual trans-dichloro
arrangement (Figure 48).155 221-227, which are coordinated
with chelating sulfoxide benzylidene ligands, were proven
to be metathetically inactive at room temperature but showed
good diene and enyne RCM catalytic activity at elevated
temperatures. 227 was found to be the most efficient
metathesis catalyst in this study, combining thermal and air
stability with good catalytic activity.
4.3. Mono- and Bis(pyridine)-Coordinated
Catalysts
Further modifications of the ligand environment in het-
erocyclic carbene-coordinated ruthenium complexes led to
the synthesis of catalysts 157 and 228-230 (Figure 49),
bearing one or two pyridine ligands. As mentioned earlier,
because of the lability of these pyridine ligands, mono- and
bis(pyridine)-coordinated complexes can be used as versatile
starting materials for the synthesis of other NHC-coordinated
ruthenium complexes.53,129,156 Complexes 157, 229, and 230
can be easily prepared and purified on a multigram scale,
Figure 49. Bis(pyridine)-coordinated H2IMes complexes 157 and
228-230.
Chemical Reviews, 2010, Vol. 110, No. 3 1765
Figure 50. Ruthenium-based NHC-coordinated complexes
231-233.
requiring little or no solvent, by simply adding an excess of
the appropriate pyridine to 3 (Figure 7).129,157 As shown in
Figure 49, pyridine ligands bind in a cis geometry, occupying
the coordination sites trans to the benzylidene and the NHC
ligands. In the solid-state structure of 157, the Ru-N bond
of the pyridine located trans to the benzylidene is more than
0.15 Å longer than the Ru-N bond of the pyridine positioned
trans to the NHC, indicating that the benzylidene exerts a
significantly larger trans influence than the NHC.129 Besides
being the precursor of bis(NHC)-coordinated complexes 155
and 156 (Figure 36),53 157 reacts instantaneously with PCy3
to regenerate the parent complex 3, with NaI to afford
mono(pyridine) complex 231, with potassium tris(pyra-
zolyl)borate to give 232, or with potassium t-butoxide to
afford complex 233 (Figure 50).129 Mono(pyridine) complex
228 can be prepared from 157 upon loss of one pyridine
ligand under vacuum.
Probably even more importantly, bis(pyridine) ruthenium
benzylidenes were proven to be efficient in the challenging
CM of acrylonitrile, and they are among the fastest-initiating
ruthenium systems studied thus far.157 Specifically, the
initiation rate in the irreversible reaction of complex 229 with
ethyl vinyl ether is at least 6 orders of magnitude higher
than the corresponding initiation rate of second-generation
catalyst 3. This very fast initiation of catalyst 229 has proven
to be extremely useful in the production of polymers with
very narrow polydispersity and for the synthesis of block
copolymers.158
The procedure for the preparation of complexes 234-240
(Figure 51) is analogous to that of 157 and 229.86,159 The
isolation of mono(pyridine)-coordinated 234, instead of the
expected bis(pyridine) complex, was rationalized on the basis
of the increased steric bulk and donor ability of the ancillary
six-membered NHC ligand, compared to the five-membered
NHC in 157 (Figure 49).159 234 was used in the ROMP of
two enantiomerically pure norbornene derivatives. Bis(py-
ridine)-coordinated 235 was shown to be highly efficient in
the CM of acrylonitrile with various functionalized alkenes.86a,b
The activity of 235 decreases in coordinating solvents,
whereas utilization of Lewis acids, which prevent the
coordination of the cyano functionality with the ruthenium
center, improves both the reaction rate and the yield of the
CM reactions. In another more detailed study, mono(pyri-
dine) complexes 236-239 were prepared from the corre-
sponding triphenylphosphine-containing NHC-coordinated 26
(Figure 10).86c Complex 236 can also be prepared from
1766 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 51. NHC-coordinated ruthenium complexes 234-240.
Scheme 6. Latent Ruthenium-Based Metathesis Catalysts
241-243
bis(pyridine)-coordinated 235 under vacuum, and conversely,
236 can be transformed into 235 in the presence of an excess
of pyridine. The metathetic catalytic activity of complexes
235 and 236 was shown to be quite similar, while 238 was
the most catalytically active complex studied, namely, among
complexes 235-240, in the RCM of diallyl malononitrile
and the CM of acrylonitrile with terminal olefins.
On a different note, pyridine-containing tethered catalysts
241-243 (Scheme 6) initiate slowly, but since the same
propagating species is provided after one turnover, they
maintain the high activity of parent catalyst 3 (Figure 7).160
As mentioned earlier, metathesis catalysts exerting attenuated
initiation rates are very important in a number of ROMP
applications, because they allow for longer fabrication times
of the monomer/catalyst mixture and, therefore, for a more
uniform polymeric material to be synthesized.161 Complexes
241a and 241b, of Cs and C1 symmetry, respectively, are
isomers in equilibrium, with 241b being the thermodynami-
cally favored species. Also, 241a and 241b are both latent
initiators relative to complexes 3, 5 (Figures 7 and 8,
Vougioukalakis and Grubbs
Figure 52. Pyridine-containing ruthenium-based complexes
244-248.
respectively), and 157 (Figure 49), with 241b initiating much
slower than 241a in RCM and ROMP transformations. The
difference in initiation rate between 241a and 241b was
attributed to the fact that the tethered pyridine ligand in 241a
is trans to the strongly σ-donating NHC ligand and, as a
result, dissociates to give the active 14-electron species much
more quickly than that in 241b. Substitution on the pyridine
ring was found to have a much less significant effect on
catalytic activity, and 243 was found to be a faster initiator
than 241 and 242, presumably due to the steric crowding of
the o-methyl group on the tethered pyridine.
A series of other heterocyclic carbene-coordinated ruthe-
nium complexes bearing pyridine or pyridine-based ligands
are presented in Figure 52. Cationic complex 244 was
isolated in an attempt to enforce intramolecular displacement
of one of the chloride atoms, by the hydroxyl group in the
side chain, to form a chelate structure.102 244 is metathesis
inactive in standard RCM reactions, supposedly due to the
pyridine ligands that are tightly bound to the cationic metal
center and cannot be released even upon addition of
p-toluenesulfonic acid. On the other hand, the tridentate
carbene in 245 was designed as a more labile and less rigid
alternative of the pincer NHC ligand in 163 (Figure 38) and
was anticipated to lead to an improved metathesis catalyst
compared to 163.132 Nevertheless, 245 was isolated in a very
low yield, preventing evaluation of its catalytic activity.
Pyridine adducts 246 and 247 (Figure 52) were also isolated,
during attempts to prepare the corresponding (not observed)
tricyclohexylphosphine complexes.121 Complexes 246 and
247, coordinated with cyclic (alkyl)(amino)carbenes, showed
relatively low RCM efficiency, which was suggested to result
from increased catalyst decomposition. Finally, 248 was
targeted as a catalyst with improved initiation efficiency
compared to its tricyclohexylphosphine-containing counter-
part 125 (Figure 29).114 Indeed, 248 was shown to be an
efficient ROMP initiator in sequence-selective copolymeriza-
 Chemical Reviews, 2010, Vol. 110, No. 3 1767

Figure 54. Pyridine-coordinated H2IMes ruthenium complexes 254

and 255.

Figure 53. Pyridine-containing ruthenium complexes 249-253.

tions. This alternating ROMP selectivity by 248 was at-

tributed to the steric interaction of the 2-phenethyl substituent
of the NHC with the growing polymer chain.
The concept of using labile pyridine ligands to prepare
faster-initiating and highly active ruthenium metathesis
catalysts that do not suffer from incomplete initiation has
been also employed in complexes 249-253 (Figure 53),
Figure 55. Water-soluble, pyridine-coordinated complex 256.
which bear alkylidene ligands other than benzylidene (note
that section 5 is dedicated to the systematic variation of the 4.4. Chelating Quinolin- and Quinoxalin-ylidenes
alkylidene ligand in heterocyclic carbene-coordinated cata-
lysts). Thus, Wagener and co-workers developed catalysts Quinoline- and quinoxaline-containing tethered complexes
249 and 250, bearing a ruthenium ethylidene and a ruthenium 257 and 258 (Scheme 7), bearing five-membered chelate
dimethylvinylidene, respectively, and successfully used them rings, have also been prepared.169 Similarly to their pyridine-
in ADMET and ROMP transformations.162,163 In other work, containing counterparts 241-243 (Scheme 6),160 these
251 was prepared by treatment of the corresponding tricy- complexes are initially isolated in the trans-dichloro geom-
clohexylphosphine-coordinated indenylidene complex with etry (257a and 258a); nevertheless, upon prolonged storage
an excess of pyridine.164 Displacing the phosphine ligand in solution they isomerize to the thermodynamically favored
by pyridine in the indenylidene precursor of 251 is signifi- Scheme 7. Ruthenium-Based Catalysts 257 and 258 Bearing
cantly slower than the same substitution reaction in ben- Chelating Quinolin- and Quinoxalin-ylidenes
zylidene-bearing complex 3 (Figure 7). This was attributed
to the stronger electron-donating ability and, therefore, trans
influence, as well as to the increased steric bulk of the
indenylidene, as compared to the benzylidene ligand. Nev-
ertheless, 251 is also more thermally stable and, furthermore,
more active in both RCM and ROMP compared to its
benzylidene counterpart 228 (Figure 49).164,165 Quite simi-
larly, both pyridine-coordinated ruthenium indenylidenes 252
and 253 were shown to be highly active in the ROMP of
1,5-cyclooctadiene.166
Finally, polyvinyl-, poly(ethylene glycol)-, and phospho-
rylcholine-substituted pyridine ligands, in complexes 254,
255, and 256, respectively (Figures 54 and 55), have been
successfully utilized to prepare immobilized (254)167 (see
section 10) or water-soluble metathesis catalysts (255 and
256)168 (see section 9).
1768 Chemical Reviews, 2010, Vol. 110, No. 3
cis-dichloro isomers (257b and 258b, respectively). Both 257
and 258 are air-stable in solution, showing RCM, enyne
metathesis, and thermally triggered ROMP activity.169,170
Moreover, trans-dichloro complexes 257a and 258a initiate
faster than their cis-dichloro isomers 257b and 258b, while
the quinoxaline-containing 258a is faster than its quinoline
analogue (257a) in model RCM and enyne metathesis
reactions.169
4.5. Bidentate Alkylidenes Chelated through
Imine Donors
The first NHC-coordinated ruthenium alkylidenes contain-
ing an imine donor tethered to the alkylidene (259 and 260,
Figure 56) were reported in 2005 by Slugovc et al.171 These
air- and moisture-stable complexes exert thermally switchable
ROMP behavior, showing high efficiency at temperatures
around 110 °C and very low initiation rates at room
temperature. 260 has a higher switching temperature and a
lower polymerization rate than 259. This difference in the
initiation rates of 259 and 260 was ascribed to the varying
chelate ring sizes (five- versus six-membered, respec-
tively).171 However, this hypothesis was challenged one year
later, by the suggestion that the placement of the imine bond
(exocyclic in 259 versus endocyclic in 260) is the factor that
primarily determines the initiation behavior.172 In that study,
complexes 261-270 (Figure 56) were prepared and evaluated
in RCM and ROMP transformations. Exocyclic imine
catalysts 261 and 262 were found to be highly active, and
certainly not latent, in the RCM of diethyl diallylmalonate,
and 261 initiated somewhat faster than 262. On the contrary,
endocyclic imine complexes 263-267 are thermally triggered
latent catalysts that show an almost on/off polymerization
behavior in the ROMP of dicyclopentadiene. This different
initiation behavior among the exocyclic and the endocyclic
imine frameworks was attributed to unfavorable steric
interactions in the exocyclic case with the rest of the catalyst
framework. Thus, a weaker Ru-N bond results in a more
Figure 56. Ruthenium-based metathesis catalysts 259-270 bearing
chelating carbenes with imine functionalities.
Vougioukalakis and Grubbs
efficient initiator and, ultimately, leads to a higher activity.
Moreover, endocyclic imine complexes 263-267 efficiently
ring-close diethyl diallylmalonate at elevated temperatures,
with an order of activity 263 > 264 > 265 > 266 > 267. This
result was rationalized on the basis of the relative donating
ability and steric demand of the imine substituents. Namely,
the electron-poor phenyl substituent affords the fastest-
initiating catalyst and the small methyl group affords the
slowest-initiating catalyst. The three-point chelating alkyl-
idenes in 269 and 270 were designed as potentially even
slower metathesis initiators since two successive ligand
dissociation events must take place before a catalytically
active fragment is generated. However, 268 and 269 show
essentially identical RCM activities, indicating that the
oxygen atom does not bind tightly enough to measurably
impact the catalysis. Complex 270, on the other hand, shows
a lower initiation rate than both 268 and 269, which suggests
that incorporating an appropriate third point of attachment
may indeed have a major impact on catalysis.
4.6. 14-Electron Phosphonium Alkylidenes
In 2004, Piers and co-workers published the synthesis of
NHC-coordinated ruthenium complex 271 (Figure 57) bear-
ing a 14-electron phosphonium alkylidene.173 Surprisingly,
the four-coordinate complex 271, which models the presumed
active species formed upon dissociation of the labile ligand
in NHC-coordinated catalysts, is air- and moisture-stable and,
furthermore, highly active in model RCM reactions. More
importantly, this system provides rapid metathesis initiation,
outperforming even bis(3-bromopyridine) complex 229
(Figure 49). Initiation in these 14-electron phosphonium
alkylidenes is more energetically favorable than in the classic
five- or six-coordinated systems, as it consists of a low-barrier
olefin-binding event without the need for a ligand to
dissociate. Soon after the synthesis of 271, the analogous
14-electron complexes 272 and 273 (Figure 57) were also
prepared.174,175
The ability of these phosphonium alkylidenes to initiate
at very low temperatures has additionally proven useful in a
series of low-temperature mechanistic studies that resulted
in the direct observation of ruthenacyclobutane intermediates
relevant to olefin metathesis.174-177 Given that ruthenacy-
clobutanes are known to play a key role in the determination
of the regio- and stereochemical outcome of metathesis, a
better understanding of their geometry is essential to the
rational design of diastereo- and enantioselective catalysts.
These studies are suggestive of bottom-face olefin coordina-
Figure 57. 14-electron NHC-coordinated ruthenium catalysts
271-273 bearing phosphonium alkylidenes.

Scheme 8. Initial Steps Proposed for the Bottom-Face
Pathway in the Ruthenium-Catalyzed Olefin Metathesis
Mechanism
Scheme 9. Side-on and Bottom-Face Bound Ruthenium
Olefinic Complexes
tion and metallacycle generation, that is, trans to the NHC
ligand. Spectroscopic data for all reported ruthenacyclobu-
tanes are consistent with a symmetric structure with a flat,
kite-shaped four-membered ring (species XII, Scheme
8).174-178 Ruthenacyclobutanes show dynamic structure,
proceeding through a series of nonproductive metallacycle
formations/cycloreversions prior to olefin exchange. How-
ever, most systems studied to date are relatively simple, and
therefore, one should be very careful in generalizing these
observations to more complicated ruthenacyclobutane spe-
cies. Also note that there has been a long-standing debate
regarding the site of olefin coordination to the ruthenium
catalyst that leads to ruthenacyclobutane formation (side- or
bottom-bound). Thus, in 1997 Snapper and co-workers
reported the isolation of complex 274 (Scheme 9) in which
a chelating olefin coordinates trans to the PCy3 ligand
(bottom face).179 Complexes 275a and 275b (Scheme 9) were
also subsequently isolated, suggesting a side-bound olefin
intermediate.180 In a similar vein, the reaction between 1,2-
divinylbenzene and a series of NHC-coordinated ruthenium
complexes led to the formation of two types of side-bound
olefin adducts (XIII and XIV, Scheme 9) that undergo
dynamic interconversion.181,91
5. Ruthenium Alkylidene Variation: Fischer-Type
Carbenes, Indenylidenes, Vinylidenes, Cyclic
Ruthenium Alkylidenes, and Other Alkylidene
Ligands
The first NHC-coordinated ruthenium indenylidenes
(276-279, Figure 58) were reported in 1999 by Nolan and
co-workers.182,183 These complexes were demonstrated to be
highly thermally stable and efficient in the RCM of bench-
mark R,ω-dienes, forming five-, six-, and seven-membered
di-, tri-, and tetrasubstituted cycloalkenes, in the ROMP of
1,5-cyclooctadiene, as well as in a series of CM transforma-
tions.166,182-186 The preparation of indenylidene complexes
280, 281, and 282 (Figure 58), bearing NHC ligands with
Chemical Reviews, 2010, Vol. 110, No. 3 1769
Figure 58. NHC-coordinated ruthenium indenylidene complexes
276-282.
Figure 59. Ruthenium complexes 283-289 bearing dimethylvi-
nylidene and alkylidene ligands.
saturated backbones, was published in 2001, 2008, and 2009,
respectively.164,187,188 Complex 280 was found to initiate the
ROMP of 1,5-cyclooctadiene faster than both 281 and 3
(Figure 7), due to the more labile nature of the PPh3 ligand
as compared to the PCy3 (also refer to section 6). Moreover,
281 shows an increased induction period compared to its
benzylidene analogue 3 (Figure 7) and, therefore, lower
activity in both RCM and ROMP transformations. It should
also be noted that (pre)catalysts 3, 280, and 281, as well as
all H2IMes-coordinated ruthenium complexes, provide the
same propagating species (X, Scheme 8) after a single
turnover. That is, as long as ligand “L” and the two anionic
ligands in X are the same in two (pre)catalysts, the catalytic
behavior of these species will only differ in the initiation
step. Finally, complex 282 was found to be more competent
than both 276 and 281 in the RCM of unhindered and
moderately hindered dienes and enynes.
Ruthenium vinylalkylidene complexes 283-285 (Figure
59) have also been isolated.66a,130,162 283 was successfully
utilized in the RCM and CM of a variety of electron-deficient
olefins,66a whereas 285 and ruthenium ethylidene 286 were
shown to be efficient ADMET and ROMP catalysts.162,163
Ruthenium alkylidenes 286-288 (Figure 59) initiate faster
than the parent benzylidene complex 3 (Figure 7),189 while,
in contrast, methylidene 289 is a very poor metathesis
catalyst, in part as a result of its extremely low phosphine
1770 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 60. Ruthenium complexes 290-296 bearing Fischer-type
carbenes.
Figure 61. Ruthenium-based complexes 297-300.
dissociation rate.24 289 represents an important intermediate
of low stability in metathesis reactions of terminal olefins
initiated by H2IMes catalyst 3, but can be isolated and
purified by column chromatography.24 All related catalyst
decomposition studies are discussed in section 11.
Ozawa and co-workers were the first to report the
preparation of an NHC-coordinated ruthenium complex
bearing a Fischer-type carbene (290, Figure 60),190 shown
to be efficient in the ROCM of endo-5,6-disubstituted
norbornenes using phenyl vinyl selenide as an acyclic olefin.
Soon thereafter, Fischer-type ruthenium complexes 291-295
(Figure 60) were also synthesized and were demonstrated
to be structurally similar, but inherently more stable than
their carbon analogues.191 In solution, complex 294 exists
in a temperature-dependent equilibrium with the chelate
complex 296 and free PCy3. Complexes 291-295 efficiently
catalyze a series of model RCM and ROMP reactions,
although with significantly lower rates than their correspond-
ing alkylidene counterparts. The rate of the RCM reaction
of diethyl diallylmalonate was found to be highly dependent
on the R-heteroatom; complex 293 was proven to be the most
active catalyst in the series with the relative activities
following the trend E ) C > N > S > O. Along these lines,
care should be taken when ethyl vinyl ether is used as the
quenching agent in ROMP reactions, since complex 295 is
metathesis active, at least under some conditions.
Furthermore, ruthenium allenylidene 297 (Figure 61),
reported in 1999, shows high thermal stability but very low
RCM activity, supposedly due to the relatively high bonding
energy of the allenylidene moiety.192 However, vinylidenes
298-300 (Figure 61) are efficient RCM and ROMP catalysts,
Vougioukalakis and Grubbs
Figure 62. Ruthenium fluoromethylidene complexes 301-303.
Figure 63. Terminal carbide 304 and phosphoniomethylidene 305.
although not as reactive as the corresponding benzylidene
complex 3 (Figure 7).193
Despite the great progress that has been made in the field
of ruthenium-catalyzed metathesis, ruthenium catalysts do
not efficiently carry out the metathesis of directly halogenated
alkenes (i.e., vinyl halides and related substrates), because
of the electron-withdrawing nature of the pendent halogens.
This is especially true in CM, since there are some examples
of RCM transformations involving R-chloro- and R-fluoro-
R,ω-dienes.194 The first example of a successful metathesis
reaction between an NHC-coordinated ruthenium complex
and a vinyl halide was reported in 2001, where ruthenium
difluoromethylidene 301 (Figure 62) was prepared by treating
the parent benzylidene complex (3, Figure 7) with an
atmosphere of 1,1-difluoroethylene.195 Although the
F2CdCH2 double bond was cleaved metathetically, this
reaction was proven not catalytic. Nevertheless, 301 effects
the ROMP of 1,5-cyclooctadiene, albeit significantly less
efficiently than 3. The poor catalytic activity of 301 was
attributed to its insufficient initiation and, therefore, could
be slightly improved by additives that promote phosphine
dissociation.
On the basis of another study, Johnson and co-workers
published a procedure toward the synthesis of monofluo-
romethylidene complexes 302 and 303 (Figure 62).196 302
and 303 are significantly more reactive than 301, but slow
compared to 3, in the RCM and CM of model alkenes.
Bis(pyridine) complex 303 initiates more quickly than 302,
as anticipated; however, it also suffers from a higher
decomposition rate. Isolation of the monochloromethylidene
analogue of 302 was not possible, even though its transient
formation could be observed at -70 °C by NMR.197 Instead,
terminal carbide 304 and phosphoniomethylidene 305 (Figure
63) were formed upon reaction of ruthenium benzylidene 3
with vinyl chloride.
In 2001, Fürstner et al. reported the synthesis of complexes
306 and 307 (Figure 64), featuring a chelating N-to-Ru tether,
as catalysts that could be regenerated upon consumption of
monomer; nevertheless, the catalytic activities of 306 and
 Chemical Reviews, 2010, Vol. 110, No. 3 1771

Figure 64. Cyclic ruthenium alkylidene complexes 306-312.

307 were not evaluated in that early work.81 Later on,

however, catalyst 309 (Figure 64) was found to mediate the
synthesis of cyclic polymers via ROMP of strained cyclic
monomers such as cis-cyclooctene,198 1,5-cyclooctadiene, and Figure 65. Ruthenium-based complexes 313-317.
1,5,9-trans-cis-trans-cyclododecatriene.199 In brief, this ring-
expansion metathesis polymerization (REMP) was suggested
to proceed via a ring-expansion initiation event, from a cyclic
ruthenium alkylidene catalyst, and propagate as cyclic
monomers are incorporated into the growing cyclic polymer
(XV, Figure 64) that remains attached to the metal center
throughout the entire polymerization process. This approach
circumvents the problems involved in other more typical
routes to cyclic polymers, which require the intramolecular
macrocyclization of linear precursors at very low concentra-
tions. Further investigations, with the aim of systematically
studying the impact of the tether length and the electronic
properties of the NHC, via backbone saturation, on different
aspects of the polymerization mechanism, led to the synthesis
of cyclic catalysts 307-312, presented in Figure 64.200
Whereas increasing the N-to-Ru tether length was found to
result in enhanced rates of polymerization, shorter tethers
were more efficient for catalyst release from the polymer.
Utilizing a saturated NHC backbone (311 and 312) was
shown to boost polymerization rates to a greater extent than
increasing the length of the tether.
Attempts to prepare active chelated catalysts led to the
preparation of an unusual imidazolium-substituted ruthenium
alkylidene (314, Figure 65).201 Complexes 313 and 314 were Figure 66. Phosphine-coordinated ruthenium catalysts 3 and
318-330.
isolated during attempts to synthesize a ruthenium alkylidene
complex monocoordinated with the corresponding bidentate
aryloxy-NHC ligand. 313 and 314 were purified by column are different only in their phosphine ligand provide the same
chromatography and fully characterized, although in very propagating species (X, Scheme 8) upon phosphine dissocia-
low isolated yields (6 and 12%, respectively). As expected, tion. Consequently, by varying the phosphine, one can
314 proved to be a poor olefin metathesis catalyst. manipulate initiation and phosphine rebinding without chang-
The isolation of alkynyl-substituted alkylidene ruthenium ing the metathesis ability of the catalyst. Accordingly,
complexes 315-317 (Figure 65) was reported in 2009.202 phosphine-containing complexes 318-327 (Figure 66) were
In that work, Lee and co-workers showed that substituents prepared for a systematic study of the effect of different
on alkynyl ruthenium alkylidenes can efficiently adjust their phosphine ligands.24,129,156,203 The data obtained for complexes
reactivity and metallotropic [1,3]-shift behavior. 315, char- 322-327 (bearing phosphine ligands with the same cone
acterized via single-crystal X-ray analysis, was proved to angle)204 by either magnetization transfer experiments or by
be moderately active in the RCM of a model enyne substrate. the stoichiometric initiation with ethyl vinyl ether24,156
revealed the existence of a linear free energy relationship
6. Variation of the Phosphine Ligand between the phosphine dissociation rate constant and the
Hammett constant σp (that is, phosphine σ-donor strength),
With regard to phosphine-containing NHC ruthenium with the more electron-rich phosphines dissociating at slower
catalysts, it should be re-emphasized that complexes which rates than electron-poor ones.156 Moreover, arylphosphine
1772 Chemical Reviews, 2010, Vol. 110, No. 3
dissociation was generally found to be faster than alkylphos-
phine dissociation. Thus, initiation in H2IMes phosphine-
containing catalysts can be easily adjusted by tuning
phosphine electronics. On the other hand, phosphine reas-
sociation showed no direct correlation with phosphine
electronics. In addition to electronics, the steric properties
of phosphine ligands have a major impact on phosphine
dissociation, and thus complex 318 was essentially metathesis
inactive at room temperature.203
Recently, complexes 328-330 (Figure 66), coordinated
with a series of chelating phosphine-carboxylate ligands,
were prepared.205 These chelated complexes exhibit slow
initiation rates at temperatures up to 40 °C; nevertheless, at
elevated temperatures, 328 and 330 efficiently ring-close
diethyl diallylmalonate and diallyl malononitrile, outperform-
ing 329.
7. Anionic Ligand(s) Variation
7.1. Halides
It has been already mentioned that, in some cases,
changing the halide ligands from chlorides to iodides
improves the enantioselectivity of chiral ruthenium metath-
esis catalysts (sections 3.3 and 3.4). In the case of chiral
monodentate NHCs, the diiodide complexes are prepared in
situ, by dissolving the dichloride catalyst in THF in the
presence of NaI,29,34,35 whereas chiral bidentate iodide
complexes are stable enough to be isolated and chromato-
graphically purified.32,33 Additionally, halide ligands have
been shown to have a significant impact on the initiation
rates of second-generation catalysts.24,206 For example, di-
bromide 331 and diiodide 332 (Figure 67) initiate 3 and 250
times faster than the dichloride parent complex 3 (Figure
66), respectively.24 This initiation rate enhancement was
predominantly attributed to the increased steric bulk of
bromide or iodide ligands, since cis electronic effects (i.e.,
between the halide(s) and the phosphine ligand) are generally
relatively small in dissociative ligand-substitution reactions.
However, despite the increased initiation efficiency of 332,
its olefin metathesis activity is comparable to, or even lower
than, that of the parent dichloride complex 3 due to slower
turnover rates.
7.2. Monodentate and Bidentate Aryloxides
Motivated by the easily tunable steric and electronic
properties of aryloxides, and in order to inhibit the formation
of chloride-bridged ruthenium species that lead to catalyst
decomposition, Fogg and co-workers developed “pseudoha-
lide” ruthenium catalysts 333-337 (Figure 68).207-210 “Halide-
free” 333, the first NHC-coordinated complex of this type
to be reported, proved to be a highly active metathesis
catalyst, efficiently carrying out the ring-closing of model
R,ω-dienes, even at very low catalyst loadings.207 Equally
Figure 67. Dibromide and diiodide ruthenium catalysts 331 and
332.
Vougioukalakis and Grubbs
Figure 68. Ruthenium-based catalysts 333-337 bearing aryloxide
ligands.
efficient metathesis catalysts 334 and 335 followed soon
thereafter.208 Remarkably, 333-335 also showed very high
affinity for silica gel, facilitating their efficient removal from
the metathesized compound(s). For example, in the RCM
of diethyl diallylmalonate using 5 mol % catalyst loading
of 333-335, followed by flash chromatography, ruthenium
levels lower than 100 ppm in the product were achieved.
This is quite important since difficulties in removing residual
ruthenium impose major drawbacks in multistep synthetic
procedures that employ metathesis (for a more extended
discussion on this issue, refer to section 10). Finally,
o-sulfonato and catecholato aryloxide complexes 336 and
337 were also synthesized and evaluated.209 As can be seen
in Figure 68, o-sulfonato complex 336 is isolated in the form
of two isomers (336a and 336b); while catecholato derivative
337 was shown to be a highly active RCM catalyst, the
pyridine ligand in 336 is nonlabile, and consequently, 336
shows modest RCM activity.
7.3. N,O-, P,O-, and O,O-Bidentate Ligands
The first ruthenium catalysts featuring this class of ligands
to be reported are complexes 338-341 (Figure 69), prepared
by exchanging tricyclohexylphosphine with the correspond-
ing chelating pyridinyl alcoholate ligand, which display low
metathesis activity at room temperature.211 This effect was
attributed to the chelate stabilization induced by the dangling
pyridine ligand. Nevertheless, at 60 °C, 338-341 effect the
ROMP of both norbornene and cyclooctene, showing reac-
tivities similar to the tricyclohexylphosphine-containing
parent complexes. In a more recent study, Jordaan and
Vosloo prepared the structurally similar complex 342 (Figure
69) and evaluated its catalytic performance in the self-
metathesis of 1-octene in the absence of a solvent.212
Complex 342 displays a lower initiation rate than phosphine-
containing 3 (Figure 66); however, at 60 °C, 342 has a higher
activity and stability compared to 3.
Schiff base N,O-bidentate ligands were introduced for the
first time in NHC-coordinated ruthenium complexes by
Verpoort and co-workers.213-215 Utilizing this ligand frame-
work is quite appealing, not only because of the fine-tuning

Figure 69. Ruthenium catalysts 338-342 coordinated with chelat-
ing pyridinyl-alcoholato ligands.
Figure 70. NHC-coordinated ruthenium catalysts 343-350 bearing
bidentate Schiff base ligands.
possibility of both the sterics and electronics at the ruthenium
center but also because of the high-yielding and usually
single-step procedures by which they are accessible. Thus,
complexes 343-349 (Figure 70) were synthesized and shown
to efficiently catalyze the RCM and ROMP of a series of
benchmark substrates.213,214,216 The catalytic activity of
343-348 was proved to depend strongly and systematically
on the steric and electronic environment of the Schiff base,
with an order of activity 343 > 344 > 345 > 346 > 347 >
348. In addition, complexes 346-349 initiate extremely
slowly compared to their phosphine-containing analogue 3
(Figure 66), showing very low metathesis activity at room
temperature. Nevertheless, at 90 °C, 346-349 are very
efficient ROMP initiators, on account of their very high
thermal stability.216 It was also found that the initiation
efficiency of latent 346 can be chemically controlled on
demand.217 Specifically, upon addition of Brönsted or Lewis
acids, such as HSiCl3, HSiMeCl2, BF3, or AlCl3, at room
temperature, 346 can be transformed into a very reactive
catalyst, affording high turnover numbers in the ROMP of
Chemical Reviews, 2010, Vol. 110, No. 3 1773
Figure 71. Ruthenium catalysts 351-354 bearing chelating
carboxylate ligands.
both 1,5-cyclooctadiene and dicyclopentadiene, as well as
in the RCM of diethyl diallylmalonate. The triggering
mechanism was proposed to involve the reversible formation
of an adduct between the acid and the electron pair on the
nitrogen of the Schiff base. Ruthenium indenylidene 350
(Figure 70), with both ROMP and controlled radical polym-
erization reactivity, has been also prepared.218
In 2005, Hahn et al. reported the synthesis of halide-free
351 (Figure 71), coordinated with two bidentate 2-pyridine-
carboxylato ligands.219 Although complex 351 is metathesis
inactive, upon addition of 2 equiv of HCl it generates a
catalytically active species by protonation of at least one of
the 2-pyridine-carboxylato ligands. While less active than
second-generation catalyst 3 (Figure 66), this in situ gener-
ated species effects the RCM of model R,ω-dienes in CH2Cl2
and MeOH.
In another approach, N,O-, P,O-, and O,O-bidentate
complexes 352-354 (Figure 71) were developed.220 Quite
similar to many of the above-reported chelated catalysts,
352-354 exert low metathesis activity at room temperature,
while O,O-chelate 354 also suffers from a high decomposi-
tion rate. Interestingly, however, the initiation of both 352
and 353 is significantly enhanced upon addition of CuCl.
Whereas CuCl-activated 352 shows reduced stability, CuCl-
activated 353 is a quick-initiating and highly efficient
catalytic system in the RCM of diethyl diallyl and diethyl
allylmethallylmalonate. Additionally, intermediate trapping
experiments suggest that this CuCl-assisted initiation mech-
anism involves reversible coordination of the prolinate ligand
to CuCl, thereby facilitating an open coordination site on
ruthenium.
7.4. Carboxylates and (Alkyl)sulfonates
Substitution of the anionic chloride ligand(s) by perfluo-
rosulfonates and, more often, perfluorocarboxylates, in
ruthenium metathesis catalysts is typically connected with
the preparation of immobilized catalysts on solid supports
(see section 10). However, homogeneous catalytic applica-
1774 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 72. Ruthenium-based catalysts 355-369 coordinated with
carboxylate and (alkyl)sulfonate ligands.
tions of this family of complexes turned out to be quite
successful as well. Thus, catalysts 355-357 (Figure 72) were
prepared and found to be efficient in the RCM of diethyl
diallylmalonate, 1,7-octadiene, diallyldiphenylsilane, N,N-
diallyltrifluoroacetamide, and other related substrates, as well
as in enyne metathesis and ROCM reactions.221 In other
work, 357-362 were tested in a series of benchmark RCM
transformations.222 These complexes proved to exert similar
or lower reactivity compared to the parent chlorine-contain-
ing systems (catalysts 3 and 5, Figures 66 and 8, respec-
tively). Quite similarly, 363-368 (Figure 72) were prepared
by substituting one or two of the chloride ligands in complex
5 with 1 or 2 equiv of the corresponding silver carboxylates,
respectively.223,224 Note that this rather typical substitution
reaction is usually complete within minutes, driven by the
precipitation of silver(I) chloride. In terms of catalytic
efficiency, monosubstituted catalysts (363, 365, and 367)
outperform those in which both chloride ligands are ex-
changed (364, 366, and 368), with the most efficient, 363,
showing activity similar to the parent dichloride catalyst 5.
Complex 369 was shown to be a highly active metathesis
catalyst in both RCM and ROCM transformations.119
In 2006, Braddock and co-workers published a related
study on Cl-, Br-, CF3CO2-, and C2F5CO2-substituted ruthe-
Vougioukalakis and Grubbs
nium isopropoxybenzylidenes, the results of which reveal
significant implications for all kinds of anionic substituents
in these types of ruthenium complexes.225 In solution, all
examined complexes were shown to constantly undergo
anionic ligand exchange, under mild conditions typical for
olefin metathesis reactions. The mechanism that was pro-
posed to account for this ligand exchange involves the
intermediacy of halide-bridged dimers, which are more easily
accessible in the case of carboxylate-containing complexes
due to steric reasons. This effect should be taken into
consideration in the design of both homogeneous and
heterogeneous ruthenium catalysts in the case of immobiliza-
tion through the anionic ligand(s). One year later, the same
group reported the vacuum-driven anionic ligand exchange
of free perfluorocarboxylic acids with ruthenate-bound per-
fluorocarboxylates.226
7.5. Nitrile- and Isonitrile-Coordinated
Alkylidene-Free Ruthenium Catalysts
This class of compounds was targeted by Buchmeiser and
co-workers with the aim of developing phototriggered ROMP
catalysts. Whereas both 370 and 371 (Figure 73) were shown
to initiate the ROMP of norbornene at room temperature in
the absence of irradiation, therefore being unsuitable pho-
totriggered initiators for this monomer, polymerization of
norborn-5-ene-2-ylmethanol required UV irradiation (172
nm) and concurrent heating at 40 °C.227 Thus, 370 and 371
are in principle suitable for the UV-initiated ROMP of this
monomer. Initiation of 370 and 371 was postulated to involve
the phototriggered dissociation of at least two of the three
phenyl isonitrile groups. Although structurally similar to 370
and 371, carboxylate-containing catalysts 372 and 373
(Figure 73) decompose upon heating in the presence of a
series of functionalized norbornenes, supposedly due to an
imine metathesis-type reaction of dissociated phenyl isonitrile
with any in situ generated ruthenium alkylidene complex.228
Nevertheless, the turning point in this family of catalysts
occurred with the preparation of nitrile-coordinated cationic
complexes 374 and 375 (Figure 73).229 These two complexes
proved to be the first thermally stable, UV-initiated ROMP
catalysts. Both 374 and 375 can be handled in air and, in
the absence of light, are completely unreactive toward
cyclooctene, dicyclopentadiene, and a number of norbornene
derivatives up to 45 °C. Exposing mixtures of either 374 or
375 with these same monomers to 308 or 254 nm irradiation,
at room temperature, led to the formation of the correspond-
Figure 73. Nitrile- and isonitrile-coordinated ruthenium-based
complexes 370-375.

Scheme 10. Proposed Mechanism for the Photoinitiated
ROMP Activity of Complex 374
ing polymers, with 254 nm excitation being considerably
more efficient. On the basis of NMR data, laser flash and
steady-state photolysis experiments, and a series of theoretical
calculations, the mechanism shown in Scheme 10 was
proposed to account for the phototriggered ROMP activity
of 374. According to their proposal, photolysis of the
precatalyst (374) initially leads to the formation of species
XVI, which then binds one monomer molecule to form
intermediate XVII. In the key step for the alkylidene
formation, a 1,2-hydrogen atom shift on the carbon-carbon
double bond of the alkene π-complex (XVII) affords the
active ruthenium(IV) species (XVIII) that initiates the ROMP
cascade.
8. N-Heterocyclic Carbene-Coordinated
(η6-Arene)ruthenium Metathesis Catalysts
Complexes 376-378 depicted in Figure 74 were the first
(η6-arene)ruthenium species230 bearing NHC ancillary ligands
to be reported.231 These half-sandwich complexes, which are
isolated in high yields and display high thermal stability,
outperform their phosphine-containing counterparts in the
RCM of diethyl diallylmalonate. It should be noted that
the typical alkylidene moiety that serves as the precursor
for the generation of the well-established ruthenacyclobutane
intermediates (via a [2 + 2]-cycloaddition step) is not
preinstalled in 376 and 377 and, therefore, an in situ
generated alkylidene species is likely operating in this case.
Figure 74. Half-sandwich NHC-coordinated ruthenium metathesis
catalysts 376-381.
Chemical Reviews, 2010, Vol. 110, No. 3 1775
Figure 75. NHC-coordinated half-sandwich ruthenium complexes
382-391.
The metathetic activity of 376-378 in the RCM of diethyl
diallylmalonate was proven not to be photoinduced, since
exposure to UV or neon light, known to accelerate or in some
cases to be necessary for the initiation of the phosphine-
containing analogues, had no effect on these systems.231 More
recently, 376 was shown to efficiently promote the CM of
functionalized styrenes to afford the corresponding sym-
metrical and unsymmetrical stilbenes, as well as the ring-
closing of dimethallyl tosylamide.232 Homobimetallic com-
plex 379 (Figure 74), which bears structural resemblance to
376-378, was also isolated and shown to be efficient in
RCM and ROMP transformations.127,128 In a more recent
study, the isolation of homobimetallic ruthenium species 380
and 381 (Figure 74) was reported as well.233 Upon reacting
with R,ω-dienes, 380 and 381 were found to unselectively
catalyze both ring-closing and cycloisomerization transfor-
mations simultaneously.
Noels and co-workers conducted a series of more detailed
investigations on this kind of NHC-coordinated (η6-arene)ru-
thenium metathesis catalysts. Specifically, they initially
synthesized complexes 382-389 (Figure 75) and tested their
metathesis activity in the ROMP of cyclooctene, including
complexes 376 and 377 (Figure 74) for comparison.234 While
all alkyl-substituted imidazol-2-ylidene-containing complexes
(382-386) showed extremely low catalytic activity, their
aryl-substituted imidazol- and imidazolin-2-ylidene-coordi-
nated counterparts (376 and 377, as well as 387-389)
efficiently polymerize cyclooctene, either by visible light
activation or by trimethylsilyldiazomethane initiation. Tri-
methylsilyldiazomethane was proposed to act through the
formation of an [Ru]dCHSiMe3 intermediate species, as had
been previously observed with other analogous systems.
Alternatively, visible light irradiation was observed to lead
to p-cymene decoordination, generating highly reactive,
coordinatively unsaturated ruthenium species that were
suggested to trigger metathesis. In the absence of trimeth-
ylsilyldiazomethane, the ROMP of cyclooctene by 376 and
377 was found to depend on the presence of light, in contrast
to the RCM of diethyl diallylmalonate discussed above.
Photoinitiation only required an ordinary 40 W “cold white”
fluorescent tube or a 250 W incandescent light bulb fixed at
10 cm from the Pyrex reaction flasks. The most efficient
ROMP catalyst in this study proved to be 376, slightly
outperforming the second most efficient, 377. In related more
recent work, Ledoux et al. reported that preparation of 389
is extremely problematic due to its high decomposition
rate.235 Instead, they prepared chelated complexes 390 and
391 (Figure 75), the phenolate ligand of which dissociates
1776 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 76. Ruthenium complex 392, imidazolium chloride 393,
and imidazolinium chlorides 394a-394h.
Figure 77. Imidazolium chlorides 395a-395i and imidazolinium
chlorides 396a-396e.
upon treatment with HCl to afford the corresponding
monodentate analogues of 389. These species were highly
unstable as well.
Interestingly, it was also found that preformed 376 is
almost as catalytically efficient as when generated in situ
from the homobimetallic complex 392 (Figure 76) and the
corresponding NHC, which is also obtained in situ from
imidazolium chloride 393 (Figure 76) by deprotonation with
KOt-Bu.234,236 This three-component catalytic system (i.e.,
[RuCl2(p-cymene)]2/NHC-precursor salt/base) is simpler and
more straightforward to utilize, as it requires only stable and
commercially available reagents to generate the active
species. Subsequently, the same approach was utilized with
imidazolium and imidazolinium chlorides 394-396 (Figures
76 and 77) as NHC ligand precursors.237-239 It was found
that the presence of a C4-C5 double bond in the imidazole
ring of the NHC ligand is not a prerequisite for high catalytic
activities in the photoinduced ROMP of cyclooctene and
norbornene.237,238 However, blocking all the ortho positions
on the N-aryl substituents of the NHCs is necessary to afford
efficient photoinitiating ROMP polymerization catalysts. This
effect was proposed to originate from the tendency of
ruthenium complexes lacking ortho substituents to undergo
Vougioukalakis and Grubbs
Figure 78. NHC-coordinated half-sandwich ruthenium complexes
397 and 398.
ortho-metalation240 of the N-aryl moiety (also refer to section
11). Moreover, changing the sterics and the electronics of
the remote aryl groups of the biphenyl units in imidazolium
chlorides 395a-395h and imidazolinium chlorides
396a-396e (Figure 77) had only limited influence on the
polymerization activity.239 Mesityl-based in situ generated
376 and 389 were the most efficient catalysts in these studies.
Finally, Buchmeiser and co-workers prepared complexes 397
and 398 (Figure 78), which are the trifluoroacetate-
coordinated analogues of 376 and 389, and examined their
thermally initiated and photoinitiated ROMP activity.227,228
While 397 and 398 were indeed shown to be suitable for
the UV-triggered polymerization of norborn-5-ene-2-yl-
methanol, norbornene was uncontrollably polymerized by
both 397 and 398 in the absence of irradiation at room
temperature.
Dixneuf and co-workers utilized a quite similar, though
thermally initiated rather than photoinitiated, three-compo-
nent catalytic system (i.e., [RuCl2(p-cymene)]2/NHC-precur-
sor salt/Cs2CO3) to carry out enyne metathesis and RCM
reactions.241-244 The presence of a terminal alkyne as an
activator was necessary in the latter set of transformations.
Furthermore, the same in situ prepared three-component
system was found to promote the ROMP of cyclooctene.245
Thus, heating of [RuCl2(p-cymene)]2 (392), NHC-precursors
394g or 394h, and Cs2CO3 at 80 °C in chlorobenzene led to
high- and moderate-yielding cyclooctene polymerization with
394g and 394h, respectively. Two different experimental
procedures were used, with addition of cyclooctene to the
catalytic system either before or after its activation process.
Cs2CO3 was proposed not only to deprotonate the NHC
precursor but also to modify the catalyst, possibly by
substituting the chloride ligand(s) and/or favoring the dis-
sociation of p-cymene.
Electron-rich bis(imidazolinylidene) olefins, containing at
least one pendant N-arylmethyl group on each imidazoli-
nylidene moiety, have also been used as carbene precursors
to afford chelated NHC-coordinated (η6-arene)ruthenium
complexes 399-405 (Scheme 11).246-248 Complexes
399-405 are not metathesis-active; however, chloride ab-
straction with AgOTf and reaction with 1,1-diphenyl prop-
argyl alcohol (HCtCCPh2OH) afford the corresponding
unstable ruthenium allenylidene intermediates XIX (Scheme
11) that, depending on the nature of the substrate, catalyze
the RCM and/or the cycloisomerization of R,ω-dienes.246,247
Furthermore, following the same activation procedure, 399,
400, and 402-405 effect the ROMP of norbornene, with
402 and 403 being the most efficient catalysts in this work.248
The metathetically active ruthenium species in these studies
were proposed to be indenylidene intermediates XX (Scheme
11), in situ generated by rearrangement of the initially formed
allenylidenes XIX.249

6
Scheme 11. NHC-Coordinated (η -Arene)ruthenium
Complexes 399-405
9. N-Heterocyclic Carbene-Coordinated
Ruthenium Catalysts Designed for Homogeneous
Metathesis in Water and Protic Solvents
In addition to the potential environmental and economic
benefits of aqueous olefin metathesis, successful materializa-
tion of such a process would also be important for numerous
biological applications. In this context, water- and protic
solvent-soluble NHC-coordinated ruthenium catalysts were
targeted in an attempt to overcome the relatively low stability
and activity of the early bis(phosphine) water-soluble
catalysts.37-39,250 In fact, the first report of olefin metathesis
utilizing NHC-coordinated complexes in protic media in-
volved the use of conventional 3 and 5 (Figures 66 and 8,
respectively), which were shown to effect the RCM and, to
a lesser extent, CM of model substrates in MeOH, as well
as in MeOH-water and DMF-water mixtures.251
Two kinds of functionalities have been employed thus far
to solubilize the desired NHC-bearing (pre)catalysts in water:
(i) poly(ethylene glycol) (PEG) chains (406 and 407, Figure
79);40,41 and (ii) quaternary ammonium groups (408-411,
Figure 79).43,99,168,252,253 As can be seen in Figure 79, these
solubilizing moieties have been attached: (i) to the NHC
ligand, as in 406-407; (ii) through the benzylidene, as in
408-410; or (iii) via the anionic ligand, as in 411. In
particular, 406 efficiently initiates the ROMP of strained
cyclic olefins in both water and methanol.40b In the former
case, the presence of 1 equiv of HCl, relative to 406, is
necessary in order to protonate the dissociated tricyclohexy-
lphosphine, thereby inhibiting its reassociation to the ruthe-
nium center and preventing catalyst decomposition by base.
Phosphine dissociation in water was proposed to be disfa-
vored due to the energetic cost of solvating two neutral
molecules. Catalyst 406, which remains in solution through-
out the entire metathesis reaction in water or MeOH, was
also found to catalyze the RCM of benchmark dienes in
MeOH. With the intention of avoiding the incorporation of
the PEG-carbamoyl-benzyl moiety, which was suggested
to reduce the stability of 406, the PEG group (number
average molecular weight ≈ 2600) has been alternatively
appended on the backbone of the NHC ligand (407).41
Indeed, water-soluble complex 407, which is also soluble in
common organic solvents such as dichloromethane and
toluene, exerts improved stability and activity in water,
compared to both 406 and all previously reported water-
Chemical Reviews, 2010, Vol. 110, No. 3 1777
Figure 79. Ruthenium metathesis catalysts 406-411 for use in
water and protic solvents.
soluble bis(phosphine) catalysts. Thus, 407 efficiently carried
out the ROMP of norbornene derivatives, the unprecedented
RCM of a series of water-soluble R,ω-dienes, and the self-
CM of cis-2-butene-1,4-diol. In an analogous fashion,
complexes 255 and 256 (Figures 54 and 55, respectively),
bearing PEG- and phosphorylcholine-substituted pyridine
ligands, were more recently shown to initiate the ROMP of
a PEG-containing oxanorbornene monomer under a variety
of conditions.168
Further studies have furnished small-molecule catalysts
408 and 409,252 as well as 410253 and 41143 (Figure 79). In
brief, 408 and 409 efficiently mediate a series of ROMP and
RCM transformations in water,252 whereas 410 performs
RCM and CM reactions in water (only for X ) I), alcohols,
and homogeneous alcohol-water mixtures, even in the
presence of air.253 In micellar solutions, 410 acts both as an
initiator and a surfactant promoting RCM and CM under
heterogeneous aqueous conditions. Complex 411 also proved
to be an efficient RCM catalyst in alcohols and homogeneous
alcohol-water mixtures in air.43 Complex 73 (Figure 16)
can be transformed into its moderately water-soluble bispro-
tonated analogue by the addition of 2 equiv of HCl.99
Unfortunately though, this bisprotonated complex suffers
from a high decomposition rate, owing to the hydrolysis of
the NHC-ruthenium bond.
Olefin metathesis in water can also be carried out by
occluding existing homogeneous ruthenium catalysts in a
hydrophobic matrix of polydimethylsiloxane and then using
the resulting polydimethylsiloxane slabs in heterogeneous
reactions.42,254 For a more detailed discussion of heteroge-
neous olefin metathesis, refer to section 10. Finally, note that
Raines and co-workers successfully utilized conventional
1778 Chemical Reviews, 2010, Vol. 110, No. 3 Vougioukalakis and Grubbs

NHC-coordinated catalysts 3 and 5 (Figures 66 and 8,

respectively) in RCM and CM reactions in homogeneous
water/organic mixtures, achieving high conversions for a
variety of substrates.255

10. Removal of Ruthenium Impurities from

Metathesis Products and Ruthenium Recycling
Strategies
Despite the widespread use of ruthenium-catalyzed me-
tathesis, removal of ruthenium byproducts at the end of the Figure 80. Variable positions for NHC-coordinated ruthenium
reaction is still rather challenging. In addition to pharma- catalysts tagging. The dashed line between Ar and L indicates that
the coordinating ligand L may or may not be connected to the
ceutical chemistry applications, where the acceptable ruthe- alkylidene.
nium content is <10 ppm in the final compound, efficient
purification of olefin metathesis products is also highly
important in the case of polymeric materials, especially when
these are to be used in electronics and other technologically
advanced applications. Of equal importance is the unsuc-
cessful exclusion of ruthenium impurities during the produc-
tion of fine chemicals, which invokes the danger of undesired
side-reactions in subsequent steps. Another closely related
and very significant issue, from both an environmental and
an economic point of view, is catalyst recycling and
regeneration.
The most common strategies that have been thus far
employed to address these problems are based on ruthenium
catalyst tagging with (i) inorganic materials (e.g., silica gel);
(ii) insoluble polymers; (iii) ionic liquid functionalities; (iv)
perfluorinated hydrocarbons; or (v) soluble polymers or
small-molecule functionalities. Ideally, these modified cata-
lysts can be easily recovered from reaction mixtures by
filtration (supports i and ii above) or by extracting the catalyst
into the ionic liquid or fluorous phase (functionalities iii and
iv, respectively), resulting in reduced ruthenium impurities.
Another procedure involves precipitation controlled on
demand, or purification via a chromatographic procedure
(approach v). As illustrated in Figure 80, attachment of NHC-
coordinated ruthenium catalysts with the above-mentioned
functionalities can be achieved through the NHC nonlabile
ligand (positions R1 or R2), anionic ligand(s) X1 and/or X2,
the benzylidene moiety (Ar), or phosphine- or pyridine-based
labile ligand(s) L. Besides affecting recycling along with
easier and more efficient purification, immobilizing the
catalytic complex onto a solid support has also been proposed
to improve catalyst stability and prevent the undesirable
bimolecular decomposition pathways, by inhibiting inter-
molecular catalyst-catalyst interactions via a phenomenon
known as site isolation.256 In view of the fact that this
research field has been very well described in a series of
recent review articles,257 herein we only briefly discuss some Figure 81. Representative NHC-coordinated heterogeneous ru-
representative examples.258 thenium catalysts 412-417.
Catalysts of the types shown in Figure 81 (412-417)102,259-262
are immobilized onto solid insoluble supports and are utilized efficiently promotes both ROMP and RCM transformations,
in heterogeneous catalysis. Thus, 412 was prepared by while ruthenium contamination of the products was found
immobilizing the precursor alcohol adduct on commercial to be as low as 70 ppm.260 Silica-supported catalysts 416
silica gel, which was pretreated with MeSiCl3 or PhSiCl3 in and 417 were shown to competently catalyze a variety of
order to install the necessary chlorosilane anchoring func- model RCM and CM reactions.261 416 and 417 can be
tionalities on its surface. 412 effects the RCM of a series of efficiently recycled multiple times and, most importantly, do
R,ω-dienes and could be reused up to three times; however, not leach ruthenium, as revealed by inductively coupled
it was proved to exert lower catalytic activity compared to plasma-mass spectrometry (ICP-MS) analysis of filtered
its homogeneous analogue.102 413, immobilized onto non- reaction solutions (ruthenium contamination of filtrate < 5
porous silica with a 0.5 wt % ruthenium loading, also shows ppb). 415 also could be recycled up to five times with no
modest RCM activity in slurry-type reactions.259 On the other significant loss of activity in the CM of a series of highly
hand, catalyst 414, prepared with a loading of 1.4 wt %, electron-deficient alkenes;262 other ruthenium isopropoxy-
 Chemical Reviews, 2010, Vol. 110, No. 3 1779

Figure 83. Ruthenium-based catalyst 423 bearing two ferrocenyl

moieties.

ro(methylcyclohexane). Fluorous-tagged complex 422 (Fig-

ure 82) was also shown to be highly active in RCM and
CM transformations of terminal olefins.270 Removal of
ruthenium residues from the metathesized products was
achieved either by fluorous-phase extraction or by filtration
through fluorous-phase silica gel, resulting in ruthenium
contamination levels as low as 500 ppm.
Figure 82. Examples of ionic-tagged and fluorous-tagged ruthe- An alternative, on-demand purification and recycling
nium catalysts 418-422. strategy, employing the redox-switchable ferrocenyl moieties
in complex 423 (Figure 83), was published by Süβner and
benzylidenes, supported on monolithic silica discs, were Plenio in 2005.271 In particular, after utilizing soluble catalyst
reportedly reused in 20 cycles.263 Note that in the case of 423 to catalyze the RCM of N,N-diallyl tosylamine, they
“boomerang-type” catalysts resembling 415 (i.e., anchored were able to in situ oxidize its two ferrocenyl moieties,
through the benzylidene ligand), the active species are causing its precipitation and separation from the reaction
homogeneous. Nevertheless, it is proposed that a large products; precipitated and washed 423 could then be easily
fraction of these catalytically active ruthenium species are redissolved by reduction. By repeating the same protocol,
recaptured during metathesis, forming the more stable 423 could efficiently perform up to three consecutive
chelating isopropoxybenzylidenes, whereas complexes that metathesis-redox cycles.
decompose remain in solution. Ruthenium contamination levels as low as 41 ppm were
The use of ionic liquids as alternative solvents provides achieved by simply extracting the RCM reaction mixtures
many potential advantages over their conventional counter- carried out by 407 (Figure 79) with water.272 Furthermore,
parts. These advantages include their high chemical and treatment of the ring-closed products with activated carbon
thermal stability, extremely low vapor pressure, insolubility after aqueous extraction led to ruthenium levels below 0.04
or immiscibility with either aqueous or organic reaction ppm (ICP-MS). Other published strategies, attempting to
media, and good ability to solvate both polar and nonpolar address the difficulties in removing ruthenium residues
species. In this context, specially designed catalysts incor-
porating an ionic moiety into their structure, such as 198,147a
410,253 or 418-420264-268 in Figures 45, 79, and 82,
respectively, have been exploited in ruthenium-catalyzed
biphasic (i.e., organic solvent/ionic liquid) olefin metathesis
reactions, aiming at the recovery and reusability of the
catalyst. In brief, 198147a and 410253 efficiently promote olefin
metathesis in organic solvents, aqueous media, and ionic
liquids, leading to levels of ruthenium contamination in the
products as low as 25 and 12 ppm, respectively, after a
simple filtration through silica gel; however, they both
display poor recyclability, and their activity is significantly
reduced in the second cycle. On the contrary, complexes
419265,266 and 420267b are very efficient RCM catalysts and
display relatively high recyclability (i.e., they can be reused
up to 8 and 17 times, respectively, with no significant loss
of activity). Furthermore, 419 affords low ruthenium con-
tamination levels in the ring-closed products (1-22 ppm).
As noted above, attempts to eliminate ruthenium contami-
nation have been carried out by utilizing ruthenium com-
plexes bearing fluorous tags, either via filtration through a
short pad of fluorous-phase silica gel or by fluorous-phase
extraction. For example, complex 421 (Figure 82) was found
to be efficient in the RCM of benchmark R,ω-dienes under
both monophasic (CH2Cl2) and biphasic (CH2Cl2/fluorous
solvent mixtures) conditions; the rate acceleration observed
in the latter case was proposed to arise from phase transfer
of the dissociated fluorous phosphine.269 Moreover, 421 could
be recycled up to three times with no significant loss of Figure 84. Ruthenium-based metathesis catalysts degradation
activity, by extracting the reaction mixtures with perfluo- adducts 424-429.
1780 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 12. Proposed Decomposition Mechanism for Ruthenium Methylidene 289
coming from homogeneous metathesis catalysts, include (i)
purification of the products on silica gel along with treatment
with activated carbon (ruthenium contamination levels as low
as 60 ppm);273 (ii) use of ruthenium scavengers, such as
dimethyl sulfoxide,274 Ph3PdO,274 or lead tetraacetate,275 in
combination with column chromatography (residual ruthe-
nium levels as low as 240 ppm); (iii) treatment of the
metathesized products with amine-modified silica (ruthenium
contamination less than 2000 ppm);276 and (iv) treatment of
the metathesis product(s) mixture with isocyanide
CNCH2CO2K (residual ruthenium as low as 120 ppm)277 or
tris(hydroxymethyl)phosphine.278
11. Decomposition Studies
Understanding the decomposition pathways of existing
ruthenium-based metathesis catalysts is crucial for the
development of new, more efficient catalysts, by rationally
designing and utilizing adjusted ligand environments that
reduce reactions that result in alkylidene loss. Along these
lines, hydridocarbonyl chlorides 424-426 and phenylcar-
bonyl chlorides 427 and 428 (Figure 84), formed in basic
alcoholic solutions upon prolonged heating of the corre-
sponding benzylidenes, comprise the first reported family
of heterocyclic carbene-coordinated ruthenium catalysts
degradation adducts.53,279,280 Note that 424 is also formed
upon prolonged heating of parent complex 3 (Figure 66) in
the presence of oxygen-containing substrates such as ethyl
vinyl ether,53 and 427 can also be produced by the reaction
Vougioukalakis and Grubbs
of solid 3 with oxygen in 29% isolable yield.279 Complexes
424-428 are derived through alcohol decarbonylation,
although the exact mechanism of this process is still
unknown. Moreover, while many of these complexes are
highly efficient hydrogenation- and olefin isomerization87
catalysts, they usually do not impose significant problems
on olefin metathesis reactions carried out in alcoholic
solvents, due to the high temperatures and prolonged reaction
times needed for their production.279,280 Structurally similar
species 429 (Figure 84), encompassing an H2IMes ligand
that has undergone C-H bond activation on one of its ortho-
methyl groups, is formed when 3 is prepared under a
moderately rigorous inert atmosphere.53 However, it should
be noted that none of the above decomposition adducts
(424-429) are formed from typical metathesis conditions
employing aprotic solvents (e.g., dichloromethane, benzene,
or toluene), and consequently, their generation cannot be
considered universal.
On the contrary, by taking into consideration that ruthe-
nium methylidenes such as 289 (Scheme 12) are common
intermediates in most metathesis reactions, studying their
decomposition was expected to shed some light on ruthenium
catalyst degradation in general. Initial investigations revealed
that 289 decomposes rapidly (t1/2 ) 5 h 40 min) compared
to the parent benzylidene complex 3 (Figure 66), via a
unimolecular pathway, despite exhibiting very low initiation
rates.24,281 While decomposition of 3 was found to be
inhibited by adding free phosphines, this was certainly not
 Chemical Reviews, 2010, Vol. 110, No. 3 1781
the case with 289. Subsequent studies led to the isolation of Scheme 13. Decomposition of Ruthenium Complexes 289
the first well-characterized decomposition products of 289, and 322
namely, 430 and 431 (Scheme 12).282 As shown by X-ray
crystallographic analysis, dinuclear ruthenium hydride 431
bears a bridging carbide between the two ruthenium centers
(Ru1 and Ru2), whereas the complete loss of phosphine
ligands is accompanied by η6-binding of Ru2 to one of the
mesityl rings in the NHC on Ru1. The proposed mechanism
for the formation of methyltricyclohexylphosphonium chlo-
ride (430) and binuclear complex 431 is illustrated in Scheme
12. Decomposition of 289 commences by nucleophilic attack
of dissociated tricyclohexylphosphine on the methylidene
moiety of XXI. Next, the 12-electron species XXII, formed
upon elimination of phosphonium ylide CH2dPCy3, binds
one of the mesityl rings of XXI to afford XXIII. Terminal
alkylidyne species XXIV, along with 430, are then generated
through HCl abstraction by CH2dPCy3. In the final step,
insertion into the alkylidyne C-H bond in XXIV with
concomitant migration of the two chlorides leads to the
formation of 431, isolated as an orange-yellow crystalline
solid in 46% yield. It is important to emphasize that complex
431 was found to catalyze alkene isomerization under
metathesis conditions, suggesting that the above-described
decomposition route of methylidene 289, and accordingly
(pre)catalyst 3, could be responsible for competing unwanted
alkene isomerization reactions during olefin metathesis
transformations carried out by 3.
Expanding this decomposition study, to include other
heteroleptic (phosphine-NHC) model ruthenium meth- Scheme 14. Proposed Mechanistic Pathway for the
ylidenes, confirmed the assumption of phosphine attack on Decomposition of 434
the methylidene carbon along the major decomposition
pathway.95 This was also found to be the case in decomposi-
tion experiments performed in the presence of ethylene as a
model olefin substrate. Thus, after five days at room
temperature, in a toluene solution under an atmosphere of
ethylene, complex 322 (Scheme 13) was found to quantita-
tively afford methyltricyclohexylphosphonium chloride 430
along with binuclear complex 432 (in about 70% yield). With
the exception of the necessary ortho-methyl C-H bond
activation step of the NHC ligands, the proposed mechanistic
pathway for the generation of 432 was essentially the same
as for complex 431. Finally, tris(pyridine) decomposition
adduct 433 (Scheme 13) was isolated in 29% yield during
attempts to prepare the corresponding bis(pyridine) ruthenium
methylidene.
In related studies, N-phenyl-substituted NHC-coordinated
ruthenium complexes were shown to also be prone to C-H
bond activation.94 In particular, when complex 434 (Scheme
14) was heated in benzene at 60 °C for 3 days, decomposition
adduct 435 precipitated in 58% yield, together with traces
(<2%) of 436 (Scheme 14). When 434 was heated in
dichloromethane at 40 °C, the isolated yields of 435 and
436 after 12 h were 24% and 38%, respectively. The
structures of both 435 and 436 were elucidated by X-ray
crystallographic analysis, and the mechanism proposed to
rationalize their generation is illustrated in Scheme 14.
Intermediate XXVI, formed by the oxidative addition of an
ortho C-H bond of one of the N-phenyl NHC substituents 436 is finally generated via a second C-H insertion and
to the ruthenium center, undergoes hydride insertion at the PCy3-mediated elimination of HCl.
R-carbon atom of the benzylidene to afford XXVII. This is NHC-coordinated alkoxybenzylidene complexes lacking
followed by reductive elimination between the metalated ortho substituents on the N-aryl groups of the NHCs show
phenyl carbon atom of the NHC and the R-carbon atom of a high decomposition tendency via ortho C-H bond activa-
benzylidene to yield complex 435. Decomposition adduct tion. Hence, in 2007, Blechert and co-workers reported the
1782 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 15. Decomposition of Ruthenium Complexes 42 and
108
isolation of oxidative degradation products 437 and 438,
derived from solutions of complexes 42 and 108, respec-
tively, in the presence of oxygen (Scheme 15).92 As expected,
437 and 438 were found to be completely metathesis inactive.
The proposed mechanistic pathway for the formation of 438
(Scheme 15) begins with a pericyclic cyclization reaction
of valence structure XXVIII leading to XXIX. Reaction with
oxygen (XXX) followed by elimination and rearomatization
(XXXI) affords the final insertion product 438. As also
mentioned in section 3.1, this deactivation route, involving
ortho C-H bond activation of N-phenyl groups in NHC-
coordinated ruthenium complexes, can be shut off by placing
bulky substituents on the backbone of the NHC, thereby
restricting the intramolecular rotation of the N-aryl groups
that brings the ortho-aryl C-H bonds closer to the ruthenium
center.96
In 2008, Piers and co-workers published a detailed work
on the thermal decomposition of 14-electron phosphonium
alkylidene species 271 and 272 (Scheme 16).283 During their
studies, in which they utilized 1,1-dichloroethylene as a
trapping agent, the formation of cationic trichloride-bridged
dimer 439 (Scheme 16) was observed in 40-45% NMR
yield, along with methylphosphonium chloride [H3CPR3]+-
[Cl]-. As can be seen in Scheme 16, 439 contains a
dichloromethylidene and a vinyl-modified NHC ligand at
each ruthenium center. The mechanism proposed to account
for all findings (characterization of decomposition adducts
as well as kinetic isotope effects and deuterium-labeling
studies) includes the C-H bond activation of an ortho-methyl
group of one of the N-mesityl substituents (XXXII), followed
by elimination of the methylphosphonium species (isolated)
to yield the cyclometalated ruthenium benzylidene XXXIII.
Intermediate species XXXIV, formed after the CM reaction
of highly reactive XXXIII with 1,1-dichloroethylene, even-
tually undergoes loss of a chloride anion and dimerization
to afford the final degradation product 439.
Vougioukalakis and Grubbs
Scheme 16. Proposed Decomposition Mechanism for
Complexes 271 and 272
H2IMes-coordinated ruthenium complexes bearing phos-
phine ligands have been also found to undergo carbon
monoxide- and aryl isocyanide-promoted alkylidene insertion
into the aryl substituent of their H2IMes ligand.277,284,285 In
fact, catalyst degradation adducts of this kind (440-455,
Scheme 17) were initially observed during attempts to
develop a rapidly quenching procedure for metathesis reac-
tions by blocking any available coordination sites with carbon
monoxide. Aryl isocyanides promote the same insertion
reaction for isopropoxybenzylidene-coordinated H2IMes
ruthenium complexes, but only after initial displacement of
the coordinated ether by a phosphine (456, 457, Scheme
17).285 Ruthenium complexes 440-457 were reported to form
through carbon monoxide or aryl isocyanide coordination-
triggered carbene cyclopropanation of the closest “double
bond” of the mesityl ring, followed by electrocyclic ring-
opening of the resulting cyclopropane derivative to afford
the final cycloheptatriene.284,285 As discussed in section 10,
this isocyanide-promoted degradation route has also been
utilized as a “cleanup” procedure for metathesis transforma-
tions.277 Finally, it should be noted that a number of
theoretical calculations have dealt with the decomposition
of ruthenium olefin metathesis catalysts.286 For example, on
the basis of a series of DFT calculations, van Rensburg and
co-workers have suggested a substrate-induced decomposi-
tion mechanism involving a β-hydride transfer from a
ruthenacyclobutane intermediate.286a,b
To summarize, the most important decomposition modes
of NHC-coordinated catalyst precursors and intermediates
include (i) degradation with primary alcohols, producing

Scheme 17. Carbon Monoxide- or Aryl Isocyanide-Promoted
Transformation of H2IMes-Substituted Ruthenium
Alkylidenes
ruthenium hydrido carbonyls; (ii) insertion of ruthenium into
the C-H bond of the methyl groups of the H2IMes ligand;
(iii) carbon monoxide- or aryl isocyanide-promoted ben-
zylidene or methylidene insertion into one of the aryl
substituents of their NHC ligand; (iv) nucleophilic attack of
a dissociated phosphine molecule on the methylidene carbon
of ruthenium methylidenes, forming the corresponding
carbide-bridged dimers; and (v) ortho C-H bond activation
of one of the N-aryl NHC substituents in ruthenium
complexes lacking ortho substituents on the N-aryl groups
of the NHCs.
12. Conclusions and Perspectives
As discussed above, nearly 400 ruthenium heterocyclic
carbene-coordinated olefin metathesis catalysts have been
prepared. They offer a wide array of structures and activities
that will benefit specific applications such as aqueous and
asymmetric reactions. In spite of all these structures, it is
pleasing to recognize that, for most applications, a few
structures will provide excellent results. The N-mesityl and
N-tolyl NHC-coordinated complexes bearing chelating alkox-
ybenzylidene ligands provide an excellent starting point for
most applications. It also appears as though the mechanisms
of all the complexes involve the formation of a 14-electron
species that adds an olefin to initiate the reaction. The general
reactivity can be understood in terms of the effect of a ligand
on the initiation formation of the 14-electron species and
the turnover of the olefin complex. Given the increasing rate
at which new catalysts are now appearing, we look forward
to further surprises and control mechanisms.
13. Acknowledgments
The authors are grateful for financial support provided by
the National Science Foundation, the National Institutes of
Chemical Reviews, 2010, Vol. 110, No. 3 1783
Health, and the 6th European Community Framework
Programme.
14. References
(1) (a) Anderson, A. W.; Merckling, N. G. (Du Pont de Nemours &
Co.) U.S. Patent 2,721,189, 1955. Also see: (b) Truett, W. L.;
Johnson, D. R.; Robinson, I. M.; Montague, B. A. J. Am. Chem.
Soc. 1960, 82, 2337.
(2) The term “disproportionation”, used by Banks and Bailey, actually
refers to the overall process of olefin metathesis and olefin
isomerization: Banks, R. L.; Bailey, G. C. Ind. Eng. Chem., Prod.
Res. DeV. 1964, 3, 170.
(3) (a) Calderon, N.; Chem, H. Y.; Scott, K. W. Tetrahedron Lett. 1967,
34, 3327. (b) Calderon, N. Acc. Chem. Res. 1972, 5, 127.
(4) Ivin, K. J.; Mol, J. C. Olefin Metathesis and Metathesis Polymeri-
zation; Academic Press: San Diego, CA, 1997.
(5) Grubbs, R. H. Handbook of Metathesis; Wiley-VCH: Weinheim,
Germany, 2003.
(6) For the 2005 Nobel Prize in Chemistry Lectures, see: (a) Chauvin,
Y. Angew. Chem., Int. Ed. 2006, 45, 3740. (b) Schrock, R. R. Angew.
Chem., Int. Ed. 2006, 45, 3748. (c) Grubbs, R. H. Angew. Chem.,
Int. Ed. 2006, 45, 3760.
(7) (a) Hérisson, J.-L.; Chauvin, Y. Makromol. Chem. 1971, 141, 161.
(8) For a series of mechanistic studies that followed the hypothesis of
the metallacyclobutane mechanism, see: (a) Katz, T. J.; McGinnis,
J. J. Am. Chem. Soc. 1975, 97, 1592. (b) Grubbs, R. H.; Burk, P. L.;
Carr, D. D. J. Am. Chem. Soc. 1975, 97, 3265. (c) Katz, T. J.;
Rothchild, R. J. Am. Chem. Soc. 1976, 98, 2519. (d) Grubbs, R. H.;
Carr, D. D.; Hoppin, C.; Burk, P. L. J. Am. Chem. Soc. 1976, 98,
3478. (e) Katz, T. J.; McGinnis, J. J. Am. Chem. Soc. 1977, 99, 1903.
Also see ref 5 and (f) Leconte, M.; Basset, J.-M.; Quignard, F.;
Larroche, C. Mechanistic Aspects of the Olefin Metathesis Reaction.
In Reactions of Coordinated Ligands; Braterman, P. S., Ed.; Plenum:
New York, 1986; Vol. 1, pp 371-420.
(9) Slugovc, C. Macromol. Rapid Commun. 2004, 25, 1283.
(10) Wiberg, K. B. Angew. Chem., Int. Ed. Engl. 1986, 25, 312.
(11) Tebbe, F. N.; Parshall, G. W.; Reddy, G. S. J. Am. Chem. Soc. 1978,
100, 3611.
(12) (a) Schrock, R. R. J. Am. Chem. Soc. 1975, 97, 6577. (b) Wood,
C. D.; McLain, S. J.; Schrock, R. R. J. Am. Chem. Soc. 1979, 101,
3210. (c) Rocklage, S. M.; Fellmann, J. D.; Rupprecht, G. A.;
Messerle, L. W.; Schrock, R. R. J. Am. Chem. Soc. 1981, 103, 1440.
(d) Wallace, K. C.; Liu, A. H.; Dewan, J. C.; Schrock, R. R. J. Am.
Chem. Soc. 1988, 110, 4964.
(13) (a) Katz, D. J.; Lee, S. J.; Acton, S. Tetrahedron Lett. 1976, 47,
4247. (b) Kress, J.; Wesolek, M.; Osborn, J. A. J. Chem. Soc., Chem.
Commun. 1982, 514. (c) Katz, T. J.; Sivavec, T. M. J. Am. Chem.
Soc. 1985, 107, 737. (d) Quignard, F.; Leconte, M.; Basset, J.-M.
J. Chem. Soc., Chem. Commun. 1985, 1816. (e) Kress, J.; Aguero,
A.; Osborn, J. A. J. Mol. Catal. 1986, 36, 1. (f) Schaverien, C. J.;
Dewan, J. C.; Schrock, R. R. J. Am. Chem. Soc. 1986, 108, 2771.
(g) Kress, J.; Osborn, J. A.; Greene, R. M. E.; Ivin, K. J.; Rooney,
J. J. J. Am. Chem. Soc. 1987, 109, 899. (h) Schrock, R. R.; DePue,
R. T.; Feldman, J.; Schaverien, C. J.; Dewan, J. C.; Liu, A. H. J. Am.
Chem. Soc. 1988, 110, 1423.
(14) (a) Schrock, R. R.; Murdzek, J. S.; Bazan, G. C.; Robbins, J.; DiMare,
M.; O’Regan, M. B. J. Am. Chem. Soc. 1990, 112, 3875. (b) Bazan,
G. C.; Khosravi, E.; Schrock, R. R.; Feast, W. J.; Gibson, V. C.;
O’Regan, M. B.; Thomas, J. K.; Davis, W. M. J. Am. Chem. Soc.
1990, 112, 8378. (c) Bazan, G. C.; Oskam, J. H.; Cho, H.-N.; Park,
L. Y.; Schrock, R. R. J. Am. Chem. Soc. 1991, 113, 6899. (d) Aeilts,
S. L.; Cefalo, D. R.; Bonitatebus, P. J., Jr.; Houser, J. H.; Hoveyda,
A. H.; Schrock, R. R. Angew. Chem., Int. Ed. 2001, 40, 1452.
(15) Armstrong, S. K. J. Chem. Soc., Perkin Trans. 1 1998, 371.
(16) Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18.
(17) Grubbs, R. H. Tetrahedron 2004, 60, 7117.
(18) (a) Michelotti, F. W.; Keaveney, W. P. J. Polym. Sci. 1965, A3, 895.
(b) Rinehart, R. E.; Smith, H. P. Polym. Lett. 1965, 3, 1049.
(19) (a) Novak, B. M.; Grubbs, R. H. J. Am. Chem. Soc. 1988, 110, 960.
(b) Novak, B. M.; Grubbs, R. H. J. Am. Chem. Soc. 1988, 110, 7542.
(20) (a) France, M. B.; Paciello, R. A.; Grubbs, R. H. Macromolecules
1993, 26, 4739. (b) France, M. B.; Grubbs, R. H.; McGrath, D. V.;
Paciello, R. A. Macromolecules 1993, 26, 4742.
(21) Nguyen, S. T.; Johnson, L. K.; Grubbs, R. H.; Ziller, J. W. J. Am.
Chem. Soc. 1992, 114, 3974.
(22) (a) Schwab, P.; France, M. B.; Ziller, J. W.; Grubbs, R. H. Angew.
Chem., Int. Ed. Engl. 1995, 34, 2039. (b) Schwab, P.; Grubbs, R. H.;
Ziller, J. W. J. Am. Chem. Soc. 1996, 118, 100. Ruthenium-based
metathesis catalysts (L2X2RudCHR) are usually divided into two
families: the first- and the second-generation ones. In the first-
generation catalysts, both neutral ligands (L) are phosphines, while
1784 Chemical Reviews, 2010, Vol. 110, No. 3
in the second-generation ones, one of the neutral ligands is a
heterocyclic carbene.
(23) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett. 1999, 1,
953.
(24) Sanford, M. S.; Love, J. A.; Grubbs, R. H. J. Am. Chem. Soc. 2001,
123, 6543.
(25) Kingsbury, J. S.; Harrity, J. P. A.; Bonitatebus, P. J., Jr.; Hoveyda,
A. H. J. Am. Chem. Soc. 1999, 121, 791.
(26) Garber, S. B.; Kingsbury, J. S.; Gray, B. L.; Hoveyda, A. H. J. Am.
Chem. Soc. 2000, 122, 8168.
(27) Gessler, S.; Randl, S.; Blechert, S. Tetrahedron Lett. 2000, 41, 9973.
(28) Review articles on the applications of olefin metathesis include the
following: (a) Grubbs, R. H.; Miller, S. J.; Fu, G. C. Acc. Chem.
Res. 1995, 28, 446. (b) Schuster, M.; Blechert, S. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 2037. (c) Grubbs, R. H.; Chang, S.
Tetrahedron 1998, 54, 4413. (d) Armstrong, S. K. J. Chem. Soc.,
Perkin Trans. 1 1998, 371. (e) Ivin, K. J. J. Mol. Catal. A: Chem.
1998, 133, 1. (f) Buchmeiser, M. R. Chem. ReV. 2000, 100, 1565.
(g) Fürstner, A. Angew. Chem., Int. Ed. 2000, 39, 3012. (h) Connon,
S. J.; Blechert, S. Angew. Chem., Int. Ed. 2003, 42, 1900. (i) Schrock,
R. R.; Hoveyda, A. H. Angew. Chem., Int. Ed. 2003, 42, 4592. (j)
Donohoe, T. J.; Orr, A. J.; Bingham, M. Angew. Chem., Int. Ed.
2006, 45, 2664. (k) Binder, J. B.; Raines, R. T. Curr. Opin. Struct.
Biol. 2008, 12, 767.
(29) Seiders, T. J.; Ward, D. W.; Grubbs, R. H. Org. Lett. 2001, 3, 3225.
(30) van Veldhuizen, J. J.; Garber, S. B.; Kingsbury, J. S.; Hoveyda, A. H.
J. Am. Chem. Soc. 2002, 124, 4954.
(31) van Veldhuizen, J. J.; Gillingham, D. G.; Garber, S. B.; Kataoka,
O.; Hoveyda, A. H. J. Am. Chem. Soc. 2003, 125, 12502.
(32) Gillingham, D. G.; Kataoka, O.; Garber, S. B.; Hoveyda, A. H. J. Am.
Chem. Soc. 2004, 126, 12288.
(33) van Veldhuizen, J. J.; Campbell, J. E.; Giudici, R. E.; Hoveyda, A. H.
J. Am. Chem. Soc. 2005, 127, 6877.
(34) Funk, T. W.; Berlin, J. M.; Grubbs, R. H. J. Am. Chem. Soc. 2006,
128, 1840.
(35) Berlin, J. M.; Goldberg, J. M.; Grubbs, R. H. Angew. Chem., Int.
Ed. 2006, 45, 7591.
(36) Giudici, R. E.; Hoveyda, A. H. J. Am. Chem. Soc. 2007, 129, 3824.
(37) Lynn, D. M.; Mohr, B.; Grubbs, R. H.; Henling, L. M.; Day, M. W.
J. Am. Chem. Soc. 2000, 122, 6601.
(38) Lynn, D. M.; Grubbs, R. H. J. Am. Chem. Soc. 2001, 123, 3187.
(39) Rölle, T.; Grubbs, R. H. Chem. Commun. 2002, 1070.
(40) (a) Varray, S.; Lazaro, R.; Martinez, J.; Lamaty, F. Organometallics
2003, 22, 2426. (b) Gallivan, J. P.; Jordan, J. P.; Grubbs, R. H.
Tetrahedron Lett. 2005, 46, 2577.
(41) Hong, S. H.; Grubbs, R. H. J. Am. Chem. Soc. 2006, 128, 3508.
(42) Mwangi, M. T.; Runge, M. B.; Bowden, N. B. J. Am. Chem. Soc.
2006, 128, 14434.
(43) Binder, J. B.; Guzei, I. A.; Raines, R. T. AdV. Synth. Catal. 2007,
349, 395.
(44) Berlin, J. M.; Campbell, K.; Ritter, T.; Funk, T. W.; Chlenov, A.;
Grubbs, R. H. Org. Lett. 2007, 9, 1339.
(45) Stewart, I. C.; Ung, T.; Pletnev, A. A.; Berlin, J. M.; Grubbs, R. H.;
Schrodi, Y. Org. Lett. 2007, 9, 1589.
(46) (a) Öfele, K J. Organomet. Chem. 1968, 12, 42. (b) Wanzlick, H. W.;
Schönherr, H. J. Angew. Chem., Int. Ed. Engl. 1968, 7, 141.
(47) (a) Cardin, D. J.; Cetinkaya, B.; Lappert, M. F. Chem. ReV. 1972,
72, 545. (b) Cardin, D. J.; Doyle, M. J.; Lappert, M. F. J. Chem.
Soc., Chem. Commun. 1972, 927.
(48) (a) Arduengo, A. J., III; Harlow, R. J.; Kline, M. J. Am. Chem. Soc.
1991, 113, 361. (b) Arduengo, A. J., III; Kline, M.; Calabrese, J. C.;
Davidson, F. J. Am. Chem. Soc. 1991, 113, 9704. Note that the first
stable carbenes were isolated by Bertrand and co-workers. (c) Igau,
A.; Grutzmacher, H.; Baceiredo, A.; Bertrand, G. J. Am. Chem. Soc.
1988, 110, 6463. (d) Igau, A.; Baceiredo, A.; Trinquier, G.; Bertrand,
G. Angew. Chem., Int. Ed. Engl. 1989, 28, 621. For two excellent
review articles on stable carbenes, see (e) Bourissou, D.; Guerret,
O.; Gabbai, F. P.; Bertrand, G. Chem. ReV. 2000, 100, 39. (f) Canac,
Y.; Soleilhavoup, M.; Conejero, F.; Bertrand, G. J. Organomet. Chem.
2004, 689, 3857.
(49) Representative review articles: (a) Herrmann, W. A.; Köcher, C.
Angew. Chem., Int. Ed. Engl. 1997, 36, 2162. (b) Herrmann, W. A.
Angew. Chem., Int. Ed. 2002, 41, 1290. (c) Perry, M. C.; Burgess,
K. Tetrahedron: Asymmetry 2003, 14, 951. (d) Hahn, F. E. Angew.
Chem., Int. Ed. 2006, 45, 1348. (e) Kantchev, E. A. B.; O’Brien,
C. J.; Organ, M. G. Angew. Chem., Int. Ed. 2007, 46, 2768. In
addition: Coordination Chemistry ReViews has devoted a special issue
on the organometallic chemistry of NHCs: 2007, 251 (5-6), 595-
896. Chemical ReViews has also devoted a special issue on carbenes:
2009, 109 (8), 3209-3884. European Journal of Inorganic Chemistry
has devoted a special issue on N-heterocyclic carbene complexes:
2009, (13), 1663-2007.
Vougioukalakis and Grubbs
(50) Representative review articles: (a) Enders, D.; Balensiefer, T. Acc.
Chem. Res. 2004, 37, 534. (b) Nair, V.; Bindu, S.; Sreekumar, V.
Angew. Chem., Int. Ed. 2004, 43, 5130. (c) Marion, N.; Dı́ez-
González, S.; Nolan, S. P. Angew. Chem., Int. Ed. 2007, 46, 2988.
(51) (a) Cardin, D. J.; Cetinkaya, B.; Lappert, M. F. Chem. ReV. 1972,
72, 545. (b) Hu, X.; Castro-Rodriguez, I.; Olsen, K.; Meyer, K.
Organometallics 2004, 23, 755. (c) Süβner, M.; Plenio, H. Chem.
Commun. 2005, 5417. (d) Cavallo, L.; Correa, A.; Costabile, C.;
Jacobsen, H. J. Organomet. Chem. 2005, 690, 5407. (e) Jacobsen,
H.; Correa, A.; Costabile, C.; Cavallo, L. J. Organomet. Chem. 2006,
691, 4350.
(52) (a) Herrmann, W. A. Synthetic Methods of Organometallic and
Inorganic Chemistry; Georg Thieme Verlag: Stuttgart, New York,
1995. (b) Herrmann, W. A.; Kocher, C.; Goossen, L. J.; Artus, G. R.
Chem.sEur. J. 1996, 2, 1627. (c) Arduengo, A. J., III; Krafczyk,
R.; Schmutzler, R.; Craig, H. A.; Goerlich, J. R.; Marshall, W. J.;
Unverzagt, M. Tetrahedron 1999, 55, 14523. (d) Weskamp, T.;
Böhm, V. P. W.; Herrmann, W. A. J. Organomet. Chem. 2000, 600,
12. (e) Enders, D.; Gielen, H. J. Organomet. Chem. 2001, 617-
618, 70. (f) Altenhoff, G.; Goddart, R.; Lehmann, C.; Glorius, F.
J. Am. Chem. Soc. 2004, 126, 15195.
(53) Trnka, T. M.; Morgan, J. P.; Sanford, M. S.; Wilhelm, T. E.; Scholl,
M.; Choi, T. L.; Ding, S. D.; Day, M. W.; Grubbs, R. H. J. Am.
Chem. Soc. 2003, 125, 2546.
(54) Upon prolonged heating ruthenium-chloride complexes in the
presence of KOt-Bu, substitution of the chlorides by tert-butoxide
occurs: Sanford, M. S.; Henling, L. M.; Day, M. W.; Grubbs, R. H.
Angew. Chem., Int. Ed. 2000, 39, 3451. This side-product can be
avoided by using hexafluoro-tert-butoxide instead (ref 23)
(55) (a) Nyse, G. W.; Csihony, S.; Waymouth, R. M.; Hedrick, J. L.
Chem.sEur. J. 2004, 10, 4073. (b) Voutchkova, A. M.; Appelhans,
L. N.; Chianese, A. R.; Crabtree, R. H. J. Am. Chem. Soc. 2005,
127, 17624. (c) Tudose, A.; Delaude, L.; Andre, B.; Demonceau, A.
Tetrahedron Lett. 2006, 47, 8529. (d) Voutchkova, A. M.; Feliz, M.;
Clot, E.; Eisenstein, O.; Crabtree, R. H. J. Am. Chem. Soc. 2007,
129, 12834.
(56) (a) Wang, H. M. J.; Lin, I. J. B. Organometallics 1998, 17, 972. (b)
Lin, I. J. B.; Vasam, C. S. Comments Inorg. Chem. 2004, 25, 75. (c)
de Frémont, P.; Scott, N. M.; Stevens, E. D.; Ramnial, T.; Lightbody,
O. C.; Macdonald, C. L. B.; Clyburne, J. A. C.; Abernethy, C. D.;
Nolan, S. P. Organometallics 2005, 24, 6301. (d) Garrison, J. C.;
Youngs, W. J. Chem. ReV. 2005, 105, 3978. (e) Yu, X. Y.; Patrick,
P. O.; James, B. R. Organometallics 2006, 25, 2359.
(57) Review articles on NHC-coordinated ruthenium complexes: (a)
Jafarpour, L.; Nolan, S. P. J. Organomet. Chem. 2001, 617-618,
17. (b) Dragutan, I.; Dragutan, V.; Delaude, L.; Demonceau, A.
ArciVoc 2005, x, 206. (c) Despagnet-Ayoub, E.; Ritter, T. Top.
Organomet. Chem. 2007, 21, 193. (d) Schrodi, Y.; Pederson, R. L.
Aldrichim. Acta 2007, 40, 45. (e) Colacino, E.; Martinez, J.; Lamatay,
F. Coord. Chem. ReV. 2007, 251, 726. (f) Dragutan, V.; Dragutan,
I.; Delaude, L.; Demonceau, A. Coord. Chem. ReV. 2007, 251, 765.
(g) Samojlowicz, C.; Bieniek, M.; Grela, K. Chem. ReV. 2009, 109,
3708. Also see (h) Deshmukh, P. H.; Blechert, S. Dalton Trans. 2007,
2479.
(58) Weskamp, T.; Schattenmann, W. C.; Spiegler, M.; Herrmann, W. A.
Angew. Chem., Int. Ed. 1998, 37, 2490.
(59) Weskamp, T.; Kohl, F. J.; Hieringer, W.; Gleich, D.; Herrmann, W. A.
Angew. Chem., Int. Ed. 1999, 38, 2416.
(60) Scholl, M.; Trnka, T. M.; Morgan, J. P.; Grubbs, R. H. Tetrahedron
Lett. 1999, 40, 2247.
(61) Huang, J.; Stevens, E. D.; Nolan, S. P.; Petersen, J. L. J. Am. Chem.
Soc. 1999, 121, 2674.
(62) For an improved preparation protocol of complex 14a, see: Jafarpour,
L.; Nolan, S. P. Organometallics 2000, 19, 2055.
(63) Weskamp, T.; Kohl, F. J.; Herrmann, W. A. J. Organomet. Chem.
1999, 582, 362.
(64) Huang, J.; Schanz, H. J.; Stevens, E. D.; Nolan, S. P. Organometallics
1999, 18, 5375.
(65) More recent studies of the carbonyl stretching frequencies in NHC-
containing transition-metal complexes imply that saturated NHCs are
marginally less electron-donating than their unsaturated analogues.
Moreover, calorimetric studies suggest relatively small differences
in the donor capacities between saturated and unsaturated NHCs.
For a recent review article on the stereochemical and electronic
properties of NHCs, see: Diez-Gonzalez, S.; Nolan, S. P. Coord.
Chem. ReV. 2007, 251, 874.
(66) (a) Chatterjee, A. K.; Morgan, J. P.; Scholl, M.; Grubbs, R. H. J. Am.
Chem. Soc. 2000, 122, 3783. The performance of 3 in the CM of
acrylate esters can be enhanced by the addition of p-cresol: (b)
Forman, G. S.; Tooze, R. P. J. Organomet. Chem. 2005, 690, 5863.
(67) For a study regarding the dependence of the RCM performance of 3
on solvent, see: Adjiman, C. S.; Clarke, A. J.; Cooper, G.; Taylor,
P. C. Chem. Commun. 2008, 2806.

(68) Bielawski, C. W.; Grubbs, R. H. Angew. Chem., Int. Ed. 2000, 39,
2903.
(69) Chatterjee, A. K.; Grubbs, R. H. Org. Lett. 1999, 1, 1751.
(70) Jafarpour, L.; Hillier, A. C.; Nolan, S. P. Organometallics 2002, 21,
442.
(71) Sanford, M. S.; Ulman, M.; Grubbs, R. H. J. Am. Chem. Soc. 2001,
123, 749.
(72) Adlhart, C.; Chen, P. HelV. Chim. Acta 2003, 86, 941.
(73) On the basis of element-specific X-ray spectroscopies and theoretical
calculations, the increased initiation rate of the phosphine-coordinated
catalysts was rationalized on the basis of a higher electron density
on the metal center: (a) Getty, K.; Delgado-Jaime, M. U.; Kennepohl,
P. J. Am. Chem. Soc. 2007, 129, 15774. Also see (b) Antonova, N. S.;
Carbó, J. J.; Poblet, J. M. Organometallics 2009, 28, 4283.
(74) Adlhart, C.; Chen, P. Angew. Chem., Int. Ed. 2002, 41, 4484.
(75) Adlhart, C.; Chen, P. J. Am. Chem. Soc. 2004, 126, 3496.
(76) Straub, B. F. Angew. Chem., Int. Ed. 2005, 44, 5974.
(77) Tsipis, A. C.; Orpen, A. G.; Harvey, J. N. Dalton Trans. 2005, 2849.
(78) Zhao, Y.; Truhlar, D. G. Org. Lett. 2007, 9, 1967.
(79) Kingsbury, J. S.; Hoveyda, A. H. J. Am. Chem. Soc. 2005, 127, 4510.
(80) Jafarpour, L.; Stevens, E. D.; Nolan, S. P. J. Organomet. Chem. 2000,
606, 49.
(81) Fürstner, A.; Ackermann, L.; Gabor, B.; Goddard, R.; Lehmann,
C. W.; Mynott, R.; Stelzer, F.; Thiel, O. R. Chem.sEur. J. 2001, 7,
3236.
(82) Dinger, M. B.; Mol, J. C. AdV. Synth. Catal. 2002, 344, 671.
(83) Courchay, F. C.; Sworen, J. C.; Wagener, K. B. Macromolecules
2003, 36, 8231.
(84) Ritter, T.; Hejl, A.; Wenzel, A. G.; Funk, T. W.; Grubbs, R. H.
Organometallics 2006, 25, 5740.
(85) Weigl, K.; Köhler, K.; Dechert, S.; Meyer, F. Organometallics 2005,
24, 4049.
(86) (a) Bai, C. X.; Lu, X. B.; He, R.; Zhang, W. Z.; Feng, X. J. Org.
Biomol. Chem. 2005, 3, 4139. (b) Bai, C. X.; Zhang, W. Z.; He, R.;
Lu, X. B.; Zhang, Z. Q. Tetrahedron Lett. 2005, 46, 7225. (c) Zhang,
W. Z.; He, R.; Zhang, R. Eur. J. Inorg. Chem. 2007, 5345.
(87) Courchay, F. C.; Sworen, J. C.; Ghiviriga, I.; Abboud, K. A.;
Wagener, K. B. Organometallics 2006, 25, 6074.
(88) Ritter, T.; Day, M. W.; Grubbs, R. H. J. Am. Chem. Soc. 2006, 128,
11768.
(89) Wakamatsu, H.; Blechert, S. Angew. Chem., Int. Ed. 2002, 41, 2403.
(90) Stewart, I. C.; Douglas, C. J.; Grubbs, R. H. Org. Lett. 2008, 10,
441.
(91) Stewart, I. C.; Benitez, D.; O’Leary, D. J.; Tkatchouk, E.; Day, M. W.;
Goddard, W. A., III; Grubbs, R. H. J. Am. Chem. Soc. 2009, 131,
1931.
(92) Vehlow, K.; Gessler, S.; Blechert, S. Angew. Chem., Int. Ed. 2007,
46, 8082.
(93) Luan, X.; Mariz, R.; Gatti, M.; Costabile, C.; Poater, A.; Cavallo,
L.; Linden, A.; Dorta, R. J. Am. Chem. Soc. 2008, 130, 6848.
(94) Hong, S. H.; Chlenov, A.; Day, M. W.; Grubbs, R. H. Angew. Chem.,
Int. Ed. 2007, 46, 5148.
(95) Hong, S. H.; Wenzel, A. G.; Salguero, T. T.; Day, M. W.; Grubbs,
R. H. J. Am. Chem. Soc. 2007, 129, 7961.
(96) Chung, C. K.; Grubbs, R. H. Org. Lett. 2008, 10, 2693.
(97) Kuhn, K. M.; Bourg, J.-B.; Chung, C. K.; Virgil, S. C.; Grubbs, R. H.
J. Am. Chem. Soc. 2009, 131, 5313.
(98) Leuthäuβer, S.; Schmidts, V.; Thiele, C. M.; Plenio, H. Chem.sEur.
J. 2008, 14, 5465.
(99) (a) Balof, S. L.; P’Pool, S. J.; Berger, N. J.; Valente, E. J.; Shiller,
A. M.; Schanz, H. J. Dalton Trans. 2008, 5791. (b) Balof, S. L.; Yu,
B.; Lowe, A. B.; Ling, Y.; Zhang, Y.; Schanz, H. J. Eur. J. Inorg.
Chem. 2009, 1717.
(100) Ledoux, N.; Linden, A.; Allaert, B.; Mierde, H. V.; Verpoort, F. AdV.
Synth. Catal. 2007, 349, 1692.
(101) Dinger, M. B.; Nieczypor, P.; Mol, J. C. Organometallics 2003, 22,
5291.
(102) Prühs, S.; Lehmann, C. W.; Fürstner, A. Organometallics 2004, 23,
280.
(103) Vehlow, K.; Maechling, S.; Blechert, S. Organometallics 2006, 25,
25.
(104) Ledoux, N.; Allaert, B.; Pattyn, S.; Mierde, H. V.; Vercaemst, C.;
Verpoort, F. Chem.sEur. J. 2006, 12, 4654.
(105) Ledoux, N.; Allaert, B.; Linden, A.; van der Voort, P.; Verpoort, F.
Organometallics 2007, 26, 1052.
(106) Vougioukalakis, G. C.; Grubbs, R. H. Organometallics 2007, 26,
2469.
(107) Vougioukalakis, G. C.; Grubbs, R. H. Chem.sEur. J. 2008, 14, 7545.
(108) Chatterjee, A. K.; Choi, T. L.; Sanders, D. P.; Grubbs, R. H. J. Am.
Chem. Soc. 2003, 125, 11360.
(109) Hejl, A. H. Ph.D. Thesis, California Institute of Technology,
Pasadena, CA, 2007.
Chemical Reviews, 2010, Vol. 110, No. 3 1785
(110) For a review article on the use of chiral NHCs in transition-metal
complexes, see: Snead, D. R.; Seo, H.; Hong, S. Curr. Org. Chem.
2008, 12, 1370.
(111) Costabile, C.; Cavallo, L. J. Am. Chem. Soc. 2004, 126, 9592.
(112) Fournier, P. A.; Collins, S. K. Organometallics 2007, 26, 2945.
(113) (a) Fournier, P. A.; Savoie, J.; Stenne, B.; Bédard, M.; Grandbois,
A.; Collins, S. K. Chem.sEur. J. 2008, 14, 8690. (b) Grandbois, A.;
Collins, S. K. Chem.sEur. J. 2008, 14, 9323.
(114) Vehlow, K.; Wang, D.; Buchmeiser, M. R.; Blechert, S. Angew.
Chem., Int. Ed. 2008, 47, 2615.
(115) Grisi, F.; Costabile, C.; Gallo, E.; Mariconda, A.; Tedesco, C.; Longo,
P. Organometallics 2008, 27, 4649.
(116) Hoveyda, A. H.; Gillingham, D. G.; van Veldhuizen, J. J.; Kataoka,
O.; Garber, S. B.; Kingsbury, J. S.; Harrity, J. P. A. Org. Biomol.
Chem. 2004, 2, 8.
(117) For a comparative study of chiral molybdenum-and ruthenium-based
metathesis catalysts, see: Cortez, G. A.; Baxter, C. A.; Schrock, R. R.;
Hoveyda, A. H. Org. Lett. 2007, 9, 2871.
(118) Yun, J.; Marinez, E. R.; Grubbs, R. H. Organometallics 2004, 23,
4172.
(119) Yang, L.; Mayr, M.; Wurst, K.; Buchmeiser, M. R. Chem.sEur. J.
2004, 10, 5761.
(120) Despagnet-Ayoub, E.; Grubbs, R. H. Organometallics 2005, 24, 338.
(121) (a) Anderson, D. R.; Lavallo, V.; O’Leary, D.; Bertrand, G.; Grubbs,
R. H. Angew. Chem., Int. Ed. 2007, 46, 7262. Synthesis of cyclic
(alkyl)(amino) carbenes. (b) Lavallo, J.; Canac, Y.; Praesang, C.;
Donnadieu, B.; Bertrand, G. Angew. Chem., Int. Ed. 2005, 44, 5705.
(c) Jazzar, R.; Dewhurst, R. D.; Bourg, J.-B.; Donnadieu, B.; Canac,
Y.; Bertrand, G. Angew. Chem., Int. Ed. 2007, 46, 2899. (d) Jazzar,
R.; Bourg, J.-B.; Dewhurst, R. D.; Donnadieu, B.; Bertrand, G. J.
Org. Chem. 2007, 72, 3492.
(122) Anderson, D. R.; Ung, T.; Mkrtumyan, G.; Bertrand, G.; Grubbs,
R. H.; Schrodi, Y. Organometallics 2008, 27, 563.
(123) Vougioukalakis, G. C.; Grubbs, R. H. J. Am. Chem. Soc. 2008, 130,
2234.
(124) Tzur, E.; Ben-Asuly, A.; Diesendruck, C. E.; Goldberg, I.; Lemcoff,
N. G. Angew. Chem., Int. Ed. 2008, 47, 6422.
(125) (a) Liu, S. T.; Reddy, K. R. Chem. Soc. ReV. 1999, 28, 315. (b)
Simms, R. W.; Drewitt, M. J.; Baird, M. C. Organometallics 2002,
21, 2958. (c) Titcomb, L. R.; Caddick, S.; Cloke, F. G. N.; Wilson,
D. J.; McKerrecher, D. Chem. Commun. 2001, 1388.
(126) Zhang, W.; Bai, C.; Lu, X.; He, R. J. Organomet. Chem. 2007, 692,
3563.
(127) Frenzel, U.; Weskamp, T.; Kohl, F. J.; Schattenmann, W. C.; Nuyken,
O.; Herrmann, W. A. J. Organomet. Chem. 1999, 586, 263.
(128) Ackermann, L.; Fürstner, A.; Weskamp, T.; Kohl, F. J.; Herrmann,
W. A. Tetrahedron Lett. 1999, 40, 4787.
(129) Sanford, M. S.; Love, J. A.; Grubbs, R. H. Organometallics 2001,
20, 5314.
(130) Conrad, J. C.; Yap, G. P. A.; Fogg, D. E. Organometallics 2003, 22,
1986.
(131) Marshall, C.; Ward, M. F.; Harrison, W. T. A. J. Organomet. Chem.
2005, 690, 3970.
(132) Wright, J. A.; Danopoulos, A. A.; Motherwell, W. B.; Carroll, R. J.;
Ellwood, S. J. Organomet. Chem. 2006, 691, 5204.
(133) Wakamatsu, H.; Blechert, S. Angew. Chem., Int. Ed. 2002, 41, 794.
(134) Dunne, A. M.; Mix, S.; Blechert, S. Tetrahedron Lett. 2003, 44, 2733.
(135) Zaja, M.; Connon, S. J.; Dunne, A. M.; Rivard, M.; Buschmann, N.;
Jiricek, J.; Blechert, S. Tetrahedron 2003, 59, 6545.
(136) Maechling, S.; Zaja, M.; Blechert, S. AdV. Synth. Catal. 2005, 347,
1413.
(137) Grela, K.; Harutyunyan, S.; Michrowska, A. Angew. Chem., Int. Ed.
2002, 41, 4038.
(138) Grela, K.; Kim, M. Eur. J. Org. Chem. 2003, 963.
(139) Michrowska, A.; Bujok, R.; Harutyunyan, S.; Sashuk, V.; Dolgonos,
G.; Grela, K. J. Am. Chem. Soc. 2004, 126, 9318.
(140) Bujok, R.; Bieniek, M.; Masnyk, M.; Michrowska, A.; Sarosiek, A.;
Stȩpowska, H.; Arlt, D.; Grela, K. J. Org. Chem. 2004, 69, 6894.
(141) Bieniek, M.; Michrowska, A.; Gułajski, Ł.; Grela, K. Organometallics
2007, 26, 1096.
(142) Gułajski, Ł.; Michrowska, A.; Bujok, R.; Grela, K. J. Mol. Catal. A:
Chem. 2006, 254, 118.
(143) Ettari, R.; Micale, N. J. Organomet. Chem. 2007, 692, 3574.
(144) Bieniek, M.; Bujok, R.; Cabaj, M.; Lugan, N.; Lavigne, G.; Arlt,
D.; Grela, K. J. Am. Chem. Soc. 2006, 128, 13652.
(145) Gawin, R.; Makal, A.; Woźniak, K.; Mauduit, M.; Grela, K. Angew.
Chem., Int. Ed. 2007, 46, 7206.
(146) Bieniek, M.; Bujok, R.; Stȩpowska, H.; Jacobi, A.; Hagenkötter, R.;
Arlt, D.; Jarzembska, K.; Makal, A.; Woźniak, K.; Grela, K. J.
Organomet. Chem. 2006, 691, 5289.
(147) (a) Rix, D.; Clavier, H.; Coutard, Y.; Gułajski, Ł.; Grela, K.; Mauduit,
M. J. Organomet. Chem. 2006, 691, 5397. For some representative
articles regarding the utilization of neutral NHC-coordinated ruthe-
1786 Chemical Reviews, 2010, Vol. 110, No. 3 Vougioukalakis and Grubbs

nium complexes in ionic liquid solvents, see (b) Buijsman, R. C.; (187) Randl, S.; Gessler, S.; Wakamatsu, H.; Blechert, S. Synlett 2001,
van Vuuren, E.; Sterrenburg, J. G. Org. Lett. 2001, 3, 3785. (c) Mayo, 430.
K. G.; Nearhoof, E. H.; Kiddle, J. J. Org. Lett. 2002, 4, 1567. (d) (188) Clavier, H.; Urbina-Blanco, C. A.; Nolan, S. P. Organometallics 2009,
Ding, X.; Lv, X.; Hui, B.; Chen, Z.; Xiao, M.; Guo, B.; Tang, W. 28, 2848.
Tetrahedron Lett. 2006, 47, 2921. (e) Williams, D. B. G.; Ajam, (189) Williams, J. E.; Harner, M. J.; Sponsler, M. B. Organometallics 2005,
M.; Ranwell, A. Organometallics 2006, 25, 3088. 24, 2013. Ruthenium alkylidenes 272-274 were also prepared
(148) Barbasiewicz, M.; Bieniek, M.; Michrowska, A.; Szadkowska, A.; starting from their bis(3-bromopyridine) analogues.
Makal, A.; Woźniak, K.; Grela, K. AdV. Synth. Catal. 2007, 349, (190) Katayama, H.; Urushima, H.; Nishioka, T.; Wada, C.; Nagao, M.;
193. Ozawa, F. Angew. Chem., Int. Ed. 2000, 39, 4513.
(149) Rix, D.; Caijo, F.; Laurent, I.; Boeda, F.; Clavier, H.; Nolan, S. P.; (191) Louie, J.; Grubbs, R. H. Organometallics 2002, 21, 2153.
Mauduit, M. J. Org. Chem. 2008, 73, 4225. (192) Schanz, H. J.; Jafarpour, L.; Stevens, E. D.; Nolan, S. P. Organo-
(150) Barbasiewicz, M.; Szadkowska, A.; Makal, A.; Jarzembska, K.; metallics 1999, 18, 5187.
Woźniak, K.; Grela, K. Chem.sEur. J. 2008, 14, 9330. (193) Opstal, T.; Verpoort, F. J. Mol. Catal. A: Chem. 2003, 200, 49.
(151) Fürstner, A.; Thiel, O. R.; Lehmann, C. Organometallics 2002, 21, (194) For some representative examples, see: (a) Chao, W. C.; Weinreb,
331. S. M. Org. Lett. 2003, 5, 2505. (b) Salim, S. S.; Bellingham, R. K.;
(152) Slugovc, C.; Perner, B.; Stelzer, F.; Mereiter, K. Organometallics Satcharoen, V.; Brown, R. C. D. Org. Lett. 2003, 5, 3403. (c) de
2004, 23, 3622. Matteis, V.; van Delft, F. L.; de Gelder, R.; Tiebes, J.; Rutjes, F.
(153) Ben-Asuly, A.; Tzur, E.; Diesendruck, C. E.; Sigalov, M.; Goldberg, Tetrahedron Lett. 2004, 45, 959.
I.; Lemcoff, N. G. Organometallics 2008, 27, 811. (195) Trnka, T. M.; Day, M. W.; Grubbs, R. H. Angew. Chem., Int. Ed.
(154) Kost, T.; Sigalov, M.; Goldberg, I.; Ben-Asuly, A.; Lemcoff, N. G. 2001, 40, 3441.
J. Organomet. Chem. 2008, 693, 2200. (196) Macnaughtan, M. L.; Johnson, M. J. A.; Kampf, J. W. Organome-
(155) Szadkowska, A.; Makal, A.; Woźniak, K.; Kadyrov, R.; Grela, K. tallics 2007, 26, 780.
Organometallics 2009, 28, 2693. (197) Macnaughtan, M. L.; Johnson, M. J. A.; Kampf, J. W. J. Am. Chem.
(156) Love, J. A.; Sanford, M. S.; Day, M. W.; Grubbs, R. H. J. Am. Chem. Soc. 2007, 129, 7708.
Soc. 2003, 125, 10103. (198) Bielawski, C. W.; Benitez, D.; Grubbs, R. H. Science 2002, 297,
(157) Love, J. A.; Morgan, J. P.; Trnka, T. M.; Grubbs, R. H. Angew. 2041.
Chem., Int. Ed. 2002, 41, 4035. (199) Bielawski, C. W.; Benitez, D.; Grubbs, R. H. J. Am. Chem. Soc.
(158) Choi, T. L.; Grubbs, R. H. Angew. Chem., Int. Ed. 2003, 42, 1743. 2003, 125, 8424.
(159) Wang, D.; Yang, L.; Decker, U.; Findeisen, M.; Buchmeiser, M. R. (200) Boydston, A. J.; Xia, Y.; Kornfield, J. A.; Gorodetskaya, I. A.;
Macromol. Rapid Commun. 2005, 26, 1757. Grubbs, R. H. J. Am. Chem. Soc. 2008, 130, 12775.
(160) Ung, T.; Hejl, A.; Grubbs, R. H.; Schrodi, Y. Organometallics 2004,
(201) Occhipinti, G.; Bjørsvik, H. R.; Törnroos, K. W.; Fürstner, A.; Jensen,
23, 5399.
V. R. Organometallics 2007, 26, 4383.
(161) For a recent review article on latent olefin metathesis catalysts bearing
(202) Yun, S. Y.; Kim, M.; Lee, D.; Wink, D. J. J. Am. Chem. Soc. 2009,
chelating alkylidene ligands, see: Szadkowska, A.; Grela, K. Curr.
131, 24.
Org. Chem. 2008, 12, 1631.
(162) Lehman, S. E.; Wagener, K. B.; Akvan, S. J. Polym. Sci., Part A: (203) Bolton, S. L.; Williams, J. E.; Sponsler, M. B. Organometallics 2007,
Polym. Chem. 2005, 43, 6134. 26, 2485. The ruthenium propylidene analogue of 301 was also
prepared.
(163) Lehman, S. E.; Wagener, K. B. Organometallics 2005, 24, 1477.
(164) (a) Monsaert, S.; Drozdzak, R.; Dragutan, V.; Dragutan, I.; Verpoort, (204) Review articles on the steric properties of phosphines: (a) Tolman,
F. Eur. J. Inorg. Chem. 2008, 432. (b) Monsaert, S.; de Canck, E.; C. A. Chem. ReV. 1977, 77, 313. (b) Brown, T. L.; Lee, K. J. Coord.
Drozdzak, R.; van der Voort, P.; Verpoort, F.; Martins, J. C.; Chem. ReV. 1993, 128, 89. For a recent review article on the
Hendrickx, P. M. S. Eur. J. Org. Chem. 2009, 655. computational description of phosphorus ligands and the metal-
(165) Burtscher, D.; Lexer, C.; Mereiter, K.; Winde, R.; Karch, R.; Slugovc, phosphorus bond, see. (c) Fey, N.; Orpen, A. G.; Harvey, J. N. Coord.
C. J. Polym. Sci., Part A: Polym. Chem. 2008, 46, 4630. Chem. ReV. 2009, 253, 704.
(166) de Frémont, P.; Clavier, H.; Montembault, V.; Fontaine, L.; Nolan, (205) Zhang, W.; Liu, P.; Jin, K.; He, R. J. Mol. Catal. A: Chem. 2007,
S. P. J. Mol. Catal. A: Chem. 2008, 283, 108. 275, 194.
(167) Mennecke, K.; Grela, K.; Kunz, U.; Kirschning, A. Synlett 2005, (206) For a recent, related theoretical study, see: Straub, B. F. AdV. Synth.
19, 2948. Catal. 2007, 349, 204.
(168) Samanta, D.; Kratz, K.; Zhang, X.; Emrick, T. Macromolecules 2008, (207) Conrad, J. C.; Amoroso, D.; Czechura, P.; Yap, G. P. A.; Fogg, D. E.
41, 530. Organometallics 2003, 22, 3634.
(169) Barbasiewicz, M.; Szadkowska, A.; Bujok, R.; Grela, K. Organo- (208) Conrad, J. C.; Parnas, H. H; Snelgrove, J. L.; Fogg, D. E. J. Am.
metallics 2006, 25, 3599. Chem. Soc. 2005, 127, 11882.
(170) Gstrein, X.; Burtscher, D.; Szadkowska, A.; Barbasiewicz, M.; Stelzer, (209) Monfete, S.; Fogg, D. E. Organometallics 2006, 25, 1940.
F.; Grela, K.; Slugovc, C. J. Polym. Sci., Part A: Polym. Chem. 2007, (210) (a) Conrad, J. C.; Snelgrove, J. L.; Eeelman, M. D.; Hall, S.; Fogg,
45, 3494. D. E. J. Mol. Catal. A: Chem. 2006, 254, 105. (b) Conrad, J. C.;
(171) Slugovc, C.; Burtscher, D.; Stelzer, F.; Mereiter, K. Organometallics Fogg, D. E. Curr. Org. Chem. 2006, 10, 185.
2005, 24, 2255. (211) Denk, K.; Fridgen, J.; Herrmann, W. A. AdV. Synth. Catal. 2002,
(172) Hejl, A.; Day, M. W.; Grubbs, R. H. Organometallics 2006, 25, 6149. 344, 666.
(173) Romero, P. E.; Piers, W. E.; McDonald, R. Angew. Chem., Int. Ed. (212) Jordaan, M.; Vosloo, H. C. M. AdV. Synth. Catal. 2007, 349, 184.
2004, 43, 6161. (213) de Clercq, B.; Verpoort, F. Tetrahedron Lett. 2002, 43, 9101.
(174) Wenzel, A.; Grubbs, R. H. J. Am. Chem. Soc. 2006, 128, 16048. (214) de Clercq, B.; Verpoort, F. J. Organomet. Chem. 2003, 672, 11.
(175) van der Eide, E. F.; Romero, P. E.; Piers, W. E. J. Am. Chem. Soc. (215) Review articles on ruthenium-based complexes bearing bidentate
2008, 130, 4485. Schiff base ligands: (a) Drozdzak, R.; Allaert, B.; Ledoux, N.;
(176) Romero, P. E.; Piers, W. E. J. Am. Chem. Soc. 2005, 127, 5032. Dragutan, I.; Dragutan, V.; Verpoort, F. AdV. Synth. Catal. 2005,
(177) Romero, P. E.; Piers, W. E. J. Am. Chem. Soc. 2007, 129, 1698. 347, 1721. (b) Drozdzak, R.; Ledoux, N.; Allaert, B.; Dragutan, I.;
(178) The same structure has been also computationally predicted. For Dragutan, V.; Verpoort, F. Cent. Eur. J. Chem. 2005, 3, 404. (c)
example, see ref 75. Drozdzak, R.; Allaert, B.; Ledoux, N.; Dragutan, I.; Dragutan, V.;
(179) Tallarico, J. A.; Bonitatebus, P. J., Jr.; Snapper, M. L. J. Am. Chem. Verpoort, F. Coord. Chem. ReV. 2005, 249, 3055. (d) Ding, F.; Sun,
Soc. 1997, 119, 7157. Y. G.; Monsaert, S.; Drozdzak, R.; Dragutan, I.; Dragutan, V.;
(180) Trnka, T. M.; Day, M. W.; Grubbs, R. H. Organometallics 2001, Verpoort, F. Curr. Org. Synth. 2008, 5, 291.
20, 3845. (216) Allaert, B.; Dieltiens, N.; Ledoux, N.; Vercaemst, C.; van der Voort,
(181) (a) Anderson, D. R.; Hickstein, D. D.; O’Leary, D.; Grubbs, R. H. P.; Stevens, C. V.; Linden, A.; Verpoort, F. J. Mol. Catal. A: Chem.
J. Am. Chem. Soc. 2006, 128, 8386. (b) Anderson, D. R.; O’Leary, 2006, 260, 221.
D.; Grubbs, R. H. Chem.sEur. J. 2008, 14, 7536. (217) (a) Ledoux, N.; Allaert, B.; Schaubroeck, D.; Monsaert, S.; Drozdzak,
(182) Jafarpour, L.; Schanz, H. J.; Stevens, E. D.; Nolan, S. P. Organo- R.; van der Voort, P.; Verpoort, F. J. Organomet. Chem. 2006, 691,
metallics 1999, 18, 5416. 5482. (b) Monsaert, S.; Drozdzak, R.; Verpoort, F. Chem. Today
(183) For a recent review article on ruthenium indenylidene complexes, 2008, 26, 93.
see: Boeda, F.; Clavier, H.; Nolan, S. P. Chem. Commun. 2008, 2726. (218) Opstal, T.; Verpoort, F. Angew. Chem., Int. Ed. 2003, 42, 2876.
(184) Fürstner, A.; Thiel, O. R.; Ackermann, L.; Schanz, H. J.; Nolan, S. P. (219) Hahn, F. E.; Paas, M.; Fröhlich, R. J. Organomet. Chem. 2005, 690,
J. Org. Chem. 2000, 65, 2204. 5816.
(185) Clavier, H.; Nolan, S. P. Chem.sEur. J. 2007, 13, 8029. (220) Samec, J. S. M.; Grubbs, R. H. Chem.sEur. J. 2008, 14, 2686.
(186) Boeda, F.; Bantreil, X.; Clavier, H.; Nolan, S. P. AdV. Synth. Catal. (221) Krause, J. O.; Nuyken, O.; Wurst, K.; Buchmeiser, M. R. Chem.sEur.
2008, 350, 2959. J. 2004, 10, 777.

(222) Halbach, T. S.; Mix, S.; Fisher, D.; Maechling, S.; Krause, J. O.;
Sievers, C.; Blechert, S.; Nuyken, O.; Buchmeiser, M. R. J. Org.
Chem. 2005, 70, 4687.
(223) Vehlow, K.; Maechling, S.; Köhler, K.; Blechert, S. Tetrahedron Lett.
2006, 47, 8617.
(224) Vehlow, K.; Maechling, S.; Köhler, K.; Blechert, S. J. Organomet.
Chem. 2006, 691, 5267.
(225) Tanaka, K.; Böhm, V. P. W.; Chadwick, D.; Roeper, M.; Braddock,
D. C. Organometallics 2006, 25, 5696.
(226) Braddock, D. C.; Tanaka, K.; Chadwick, D.; Böhm, V. P. W.; Roeper,
M. Tetrahedron Lett. 2007, 48, 5301.
(227) Zhang, Y.; Wang, D.; Lönnecke, P.; Scherzer, T.; Buchmeiser, M. R.
Macromol. Symp. 2006, 236, 30.
(228) Buchmeiser, M. R.; Wang, D.; Zhang, Y.; Naumov, S.; Wurst, K.
Eur. J. Inorg. Chem. 2007, 3988.
(229) Wang, D.; Wurst, K.; Knolle, W.; Decker, U.; Prager, L.; Naumov,
S.; Buchmeiser, M. R. Angew. Chem., Int. Ed. 2008, 47, 3267.
(230) For a recent review article on η6-arene ruthenium complexes, see:
Therrien, B. Coord. Chem. ReV. 2009, 253, 493.
(231) Jafarpour, L.; Huang, J.; Stevens, E. D.; Nolan, S. P. Organometallics
1999, 18, 3760.
(232) Lo, C.; Cariou, R.; Fischmeister, C.; Dixneuf, P. H. AdV. Synth. Catal.
2007, 349, 546.
(233) Sauvage, X.; Borguet, Y.; Noels, A. F.; Delaude, L.; Demonceau,
A. AdV. Synth. Catal. 2007, 349, 255.
(234) Delaude, L.; Demonceau, A.; Noels, A. F. Chem. Commun. 2001,
986.
(235) Ledoux, N.; Allaert, B.; Verpoort, F. Eur. J. Inorg. Chem. 2007,
5578.
(236) An alternative method for the in situ generation of some of these
NHC ligands has been also reported:Tudose, A.; Demonceau, A.;
Delaude, L. J. Organomet. Chem. 2006, 691, 5356.
(237) Delaude, L.; Spyza, M.; Demonceau, A.; Noels, A. F. AdV. Synth.
Catal. 2002, 344, 749.
(238) Delaude, L.; Demonceau, A.; Noels, A. F. Curr. Org. Chem. 2006,
10, 203.
(239) Maj, A. M.; Delaude, L.; Demonceau, A.; Noels, A. F. J. Organomet.
Chem. 2007, 692, 3048.
(240) (a) Hitchcock, P. B.; Lappert, M. F.; Pye, P. L.; Thomas, S. J. Chem.
Soc., Dalton Trans. 1979, 1929. (b) Simal, F.; Jan, D.; Delaude, L.;
Demonceau, A.; Spirlet, M. R.; Noels, A. F. Can. J. Chem. 2001,
79, 529.
(241) Sémeril, D.; Cléran, M.; Bruneau, C.; Dixneuf, P. H. AdV. Synth.
Catal. 2001, 343, 184.
(242) Ackermann, L.; Bruneau, C.; Dixneuf, P. H. Synlett 2001, 3, 397.
(243) Sémeril, D.; Bruneau, C.; Dixneuf, P. H. HelV. Chim. Acta 2001,
84, 3335.
(244) Sémeril, D.; Bruneau, C.; Dixneuf, P. H. AdV. Synth. Catal. 2002,
344, 585.
(245) Castarlenas, R.; Alaoui-Abdallaoui, I.; Sémeril, D.; Mernari, B.;
Guesmi, S.; Dixneuf, P. H. New J. Chem. 2003, 27, 6.
(246) Cetinkaya, B.; Demir, S.; Özdemir, I.; Toupet, L.; Sémeril, D.;
Bruneau, C.; Dixneuf, P. H. New J. Chem. 2001, 25, 519.
(247) Cetinkaya, B.; Demir, S.; Özdemir, I.; Toupet, L.; Sémeril, D.;
Bruneau, C.; Dixneuf, P. H. Chem.sEur. J. 2003, 9, 2323.
(248) Özdemir, I.; Demir, S.; Cetinkaya, B.; Toupet, L.; Castarlenas, R.;
Fischmeister, C.; Dixneuf, P. H. Eur. J. Inorg. Chem. 2007, 2862.
(249) Castarlenas, R.; Vovard, C.; Fischmeister, C.; Dixneuf, P. H. J. Am.
Chem. Soc. 2006, 128, 4079.
(250) For two recent review articles on aqueous olefin metathesis, see: (a)
Burtscher, D.; Grela, K. Angew. Chem., Int. Ed. 2009, 48, 442. (b)
Zaman, S.; Curnow, O. J.; Abell, A. D. Aust. J. Chem. 2009, 62, 91.
(251) Connon, S. J.; Rivard, M.; Zaja, M.; Blechert, S. AdV. Synth. Catal.
2003, 345, 572.
(252) Jordan, J. P.; Grubbs, R. H. Angew. Chem., Int. Ed. 2007, 46, 5152.
(253) (a) Michrowska, A.; Gułajski, Ł.; Kaczmarska, Z.; Mennecke, K.;
Kirschning, A.; Grela, K. Green Chem. 2006, 8, 685. (b) Michrowska,
A.; Gułajski, Ł.; Grela, K. Chem. Today 2006, 24, 19. (c) Gułajski,
Ł.; Michrowska, A.; Naroznik, J.; Kaczmarska, Z.; Rupnicki, L.;
Grela, K. ChemSusChem 2008, 1, 103.
(254) Runge, M. B.; Mwangi, M. T.; Bowden, N. B. J. Organomet. Chem.
2006, 691, 5278.
(255) Binder, J. B.; Blank, J. J.; Raines, R. T. Org. Lett. 2007, 9, 4885.
(256) (a) Collman, J. P.; Belmont, J. A.; Brauman, J. J. J. Am. Chem. Soc.
1983, 105, 7288. (b) Drago, R. S.; Pribich, D. C. Inorg. Chem. 1985,
24, 1983. (c) Tollner, K.; Popovitv-Biro, R.; Lahav, M.; Milstein,
D. Science 1997, 278, 2100. (d) Annis, D. A.; Jacobsen, E. N. J. Am.
Chem. Soc. 1999, 121, 4147.
(257) (a) Buchmeiser, M. R. New J. Chem. 2004, 28, 549. (b) Sommer,
W. J.; Weck, M. Coord. Chem. ReV. 2007, 251, 860. (c) Clavier, H.;
Grela, K.; Kirschning, A.; Mauduit, M.; Nolan, S. P. Angew. Chem.,
Chemical Reviews, 2010, Vol. 110, No. 3 1787
Int. Ed. 2007, 46, 6786. (d) Copéret, C.; Basset, J.-M. AdV. Synth.
Catal. 2007, 349, 78. (e) Šebesta, R.; Kmentová, I.; Toma, S. Green
Chem. 2008, 10, 484. (f) Śledź, P.; Mauduit, M.; Grela, K. Chem.
Soc. ReV. 2008, 37, 2433. (g) Buchmeiser, M. R. Chem. ReV. 2009,
109, 303.
(258) For some recent representative studies, see: (a) Rix, D.; Caijo, F.;
Laurent, I.; Gulajski, L.; Grela, K.; Mauduit, M. Chem. Commun.
2007, 3771. (b) van Berlo, B.; Houthoofd, K.; Sels, B. F.; Jacobs,
P. A. AdV. Synth. Catal. 2008, 350, 1949. (c) Chen, S.-W.; Kim,
J. H.; Shin, H.; Lee, S.-G. Org. Biomol. Chem. 2008, 6, 2676. (d)
Lim, J.; Lee, S. S.; Ying, J. Y. Chem. Commun. 2008, 4312. (e)
Elias, X.; Pleixats, R.; Man, M. W. C. Tetrahedron 2008, 64, 6770.
(f) Clavier, H.; Nolan, S. P.; Mauduit, M. Organometallics 2008,
27, 2287. (g) Wakamatsu, H.; Saito, Y.; Masubuchi, M.; Fujita, R.
Synlett 2008, 12, 1805. (h) Consorti, C. S.; Aydos, G. L. P.; Ebeling,
G.; Dupont, J. Org. Lett. 2008, 10, 237. (i) Keraani, A.; Renouard,
D.; Fischmeister, C.; Bruneau, C.; Rabiller-Baudry, M. ChemSus-
Chem 2008, 1, 927. (j) Schoeps, D.; Buhr, K.; Dijkstra, M.; Ebert,
K.; Plenio, H. Chem.sEur. J. 2009, 15, 2960. (k) Kim, J. H.; Park,
B. Y.; Chen, S.-W.; Li, S.-G. Eur. J. Org. Chem. 2009, 2239. (l)
Chen, S.-W.; Kim, J. H.; Ryu, K. Y.; Lee, W.-W.; Hong, J.; Lee,
S.-G. Tetrahedron 2009, 65, 3397.
(259) Mayr, M.; Buchmeiser, M. R.; Wurst, K. AdV. Synth. Catal. 2002,
344, 712.
(260) Mayr, M.; Mayr, B.; Buchmeiser, M. R. Angew. Chem., Int. Ed. 2001,
40, 3839.
(261) Allen, D. P.; van Wingerden, M. M.; Grubbs, R. H. Org. Lett. 2009,
11, 1261.
(262) Randl, S.; Buschmann, N.; Connon, S. J.; Blechert, S. Synlett 2001,
10, 1547.
(263) Kingsbury, J. S.; Garber, S. B.; Giftos, J. M.; Gray, B. L.; Okamoto,
M. M.; Farrer, R. A.; Fourkas, J. T.; Hoveyda, A. H. Angew. Chem.,
Int. Ed. 2001, 40, 4251.
(264) Audic, N.; Clavier, H.; Mauduit, M.; Guillemin, J. C. J. Am. Chem.
Soc. 2003, 125, 9248.
(265) Clavier, H.; Audic, N.; Mauduit, M.; Guillemin, J. C. Chem. Commun.
2004, 2282.
(266) Clavier, H.; Audic, N.; Mauduit, M.; Guillemin, J. C. J. Organomet.
Chem. 2005, 690, 3585.
(267) (a) Yao, Q.; Zhang, Y. Angew. Chem., Int. Ed. 2003, 42, 3395. (b)
Yao, Q.; Sheets, M. J. Organomet. Chem. 2005, 690, 3577.
(268) For another type of ionic liquid functionality attached to the
benzylidene ligand, see: Thurier, C.; Fischmeister, C.; Bruneau, C.;
Olivier-Bourbigou, H.; Dixneuf, P. H. J. Mol. Catal. A: Chem. 2007,
268, 126.
(269) (a) Corrêa da Costa, R.; Gladysz, J. A. Chem. Commun. 2006, 2619.
(b) Corrêa da Costa, R.; Gladysz, J. A. AdV. Synth. Catal. 2007, 349,
243.
(270) Matsugi, M.; Curran, D. P. J. Org. Chem. 2005, 70, 1636.
(271) Süβner, M.; Plenio, H. Angew. Chem., Int. Ed. 2005, 44, 6885.
(272) Hong, S. H.; Grubbs, R. H. Org. Lett. 2007, 9, 1955.
(273) Cho, J. H.; Kim, B. M. Org. Lett. 2003, 5, 531.
(274) Ahn, Y. M.; Yang, K. L.; Georg, G. I. Org. Lett. 2001, 3, 1411.
(275) Mendez-Andino, J.; Paquette, L. A. Org. Lett. 2000, 2, 1263.
(276) McEleney, K.; Allen, D. P.; Holliday, A. E.; Crudden, C. M. Org.
Lett. 2006, 8, 2663.
(277) Galan, B. R.; Kalbarczyk, K. P.; Szczepankiewicz, S.; Keister, J. B.;
Diver, S. T. Org. Lett. 2007, 9, 1203.
(278) Ferguson, M. L.; O’Leary, D. J.; Grubbs, R. H. Org. Synth. 2003,
80, 85.
(279) Dinger, M. B.; Mol, J. C. Eur. J. Inorg. Chem. 2003, 2827.
(280) Banti, D.; Mol, J. C. J. Organomet. Chem. 2004, 689, 3113.
(281) Ulman, M.; Grubbs, R. H. J. Org. Chem. 1999, 64, 7202.
(282) Hong, S. H.; Day, M. W.; Grubbs, R. H. J. Am. Chem. Soc. 2004,
126, 7414.
(283) Leitao, E. M.; Dubberley, S. R.; Piers, W. E.; Wu, Q.; McDonald,
R. Chem.sEur. J. 2008, 14, 11565.
(284) Galan, B. R.; Gembicky, M.; Dominiac, P. M.; Keister, J. B.; Diver,
S. T. J. Am. Chem. Soc. 2005, 127, 15702.
(285) Galan, B. R.; Pitak, M.; Gembicky, M.; Keister, J. B.; Diver, S. T.
J. Am. Chem. Soc. 2009, 131, 6822.
(286) For some representative examples, see:(a) van Rensburg, W. J.;
Steynberg, P. J.; Meyer, W. H.; Kirk, M. M.; Forman, G. S. J. Am.
Chem. Soc. 2004, 126, 14332. (b) van Rensburg, W. J.; Steynberg,
P. J.; Kirk, M. M.; Meyer, W. H.; Forman, G. S. J. Organomet. Chem.
2006, 691, 5312. (c) Mathew, J.; Koga, N.; Suresh, C. H. Organo-
metallics 2008, 27, 4666.
CR9002424
1788 Chem. Rev. 2010, 110, 1788–1806

Catalysis and Surface Organometallic Chemistry: A View from Theory and

Simulations

Philippe Sautet* and Françoise Delbecq

Université de Lyon, Laboratoire de Chimie, Institut de Chimie de Lyon, École Normale Supérieure de Lyon and CNRS, 46, allée d’Italie,
69364 Lyon Cedex 07, France

Received September 1, 2009

Contents
1. Introduction
2. A Brief Review of Methods and Tools
3. Mechanisms of Interaction between a Metallic
Complex and a Solid Surface
4. Nature and Structure of the Surface
Organometallic Species. Identification of the
Surface Catalytic Active Site
4.1. Clusters as Surface Models
4.2. Molecular Models as Supports
4.3. Periodic Systems as Models of Supports
5. Catalytic Reactivity of the Surface Organometallic
Complexes
5.1. Alkane Hydrogenolysis and Metathesis on
Transition Metal Hydrides
5.2. Olefin Metathesis
5.3. Polymerization
6. Conclusion
7. References
1. Introduction
The development of new catalysts is a key objective for a
cleaner and sustainable chemistry. Two important families
of catalysts coexist nowadays. Metal complexes with ligands
offer a flexible framework for molecular catalytic species,
reaching high activity and selectivity for a large class of
chemical reactions, finely tunable by changing the ligands.
The structure of the catalytic complex itself can be studied
by a variety of spectroscopic and crystallographic techniques,
and although the exact nature of the active site in catalytic
conditions is sometimes difficult to determine, a single
molecular species is clearly responsible for the activity. The
knowledge of reaction mechanisms and pathways is generally
well advanced. The reaction is usually conducted in a
homogeneous phase with the reactant and product, which
leads to the well-known difficulty of separating the product
from the catalyst. The necessity for a solvent is another
constraint. On the other end of the spectra, heterogeneous
catalysis uses a variety of solid surfaces (metals, oxides,
sulfides) to accelerate a wide range of reactions. The large
number of industrial processes using a solid catalyst is a clear
demonstration of the efficiency of those catalysts. The surface
of the solid presents a variety of sites, including defects such
as steps with variable coordination for the active element.
Moreover, the nature of the surface can change upon the
* To whom correspondence should be addressed. Fax: (+33) 472728860.
E-mail: philippe.sautet@ens-lyon.fr.
10.1021/cr900295b  2010 American Chemical Society
Published on Web 10/29/2009
1788
1789
1790
1793
1793
1795
1795
1797
1797
Philippe Sautet has studied at “Ecole Polytechnique” in Paris and defended
his doctorate in Theoretical Chemistry at Orsay University (Paris XI) in
1800 1989. He then entered CNRS at the Institute of Research on Catalysis in
1802 Lyon, where he developed and led a group devoted to the applications
1805 of theoretical chemistry to heterogeneous catalysis. He is now director of
the “Laboratory of Chemistry” at the Ecole Normale Supérieure of Lyon
1805 and at CNRS, and director of the “Institut de Chimie de Lyon”. He has
published over 190 scientific papers. He received the Descartes-Huygens
prize in 1998, the silver medal of CNRS in 2007, and the Paul Pascal
Prize of the French Academy of Science in 2008. His research interests
are in the theory of the electronic structure at the solid-gas interface
and the modeling of elementary steps of heterogeneous catalysis. Recent
studies focus on the modeling of catalytic surfaces in realistic conditions
and on the surface structure of alumina and its interaction with active
sites, such as organometallic complexes. They also aim at exploring
reaction pathways for molecules on these catalytic sites, with a special
goal at understanding selectivity in heterogeneous catalysis. He col-
laborates with several experimental groups in the fields of surface chemistry
and catalysis.
reaction conditions, and its structure is more difficult to
characterize from microscopy and spectroscopy than it is for
a molecular species. The determination of the active site is
hence a challenge, and most probably several possible active
sites compete on the solid surface. As a result, although the
activity can be high and the products easily separable from
the catalysts, heterogeneous catalysts generally suffer from
a lack of selectivity. The fabrication of efficient enantiose-
lective catalyst for fine chemical and pharmacological
purposes is, for example, much less developed in heteroge-
neous catalysis than in the homogeneous field.
Surface organometallic chemistry attends to bridge the gap
between these two fields. The organometallic chemist would
say that it uses a solid surface as a ligand for the complex,
hence attaching it to a particle. The solid-state chemist would
say that the surface is modified by grafting a metal complex,
hence creating a new catalytic species. The approach aims
at creating a direct interaction between the metal and the
Catalysis and Surface Organometallic Chemistry
Françoise Delbecq graduated from the University of Lyon (France) in 1971
and obtained a Master degree in organic chemistry in 1972. At the same
date she obtained a permanent position as a researcher at CNRS, Centre
National de la Recherche Scientifique. She defended her Ph.D. in 1976
in the field of Organic Chemistry at the University of Lyon in the laboratory
of Prof. Goré. She then moved to the laboratory of Prof. Salem at the
University of Paris XI (Orsay) where she oriented her research toward
Quantum Chemistry. She worked there on the mechanism of radical
organic reactions. In 1985, she moved to the Institut de Recherches sur
la Catalyse at Villeurbanne mainly devoted to the study of catalytic
reactions. There she oriented her research toward the investigation of
the mechanism of catalytic reactions by means of density functional theory
(DFT) calculations with a particular focus on the selective and enanti-
oselective reactions. Since 2003, she belongs to the Laboratory of
Chemistry at Ecole Normale Supérieure de Lyon where she continues to
work in the field of catalysis.
solid surface, and it is hence different from supported
homogeneous catalysts where one of the ligands is attached
to a solid surface at a position far from the metal. It has
already led to several catalytic successes,1-3 but opens a wide
range of fundamental questions. What is the mechanism of
the chemical reaction leading to the attachment of the
complex on the surface? What are the possible interactions
at its origin? Which species is formed? Is it unique, or do
several forms coexist? What is the influence of the solid
surface itself? How are the electronic properties of the metal
modified? Are the functions of the complex kept, or new
properties generated? Can unprecedented catalytic properties
be obtained?
Those hybrid molecular-solid species, at the border
between two different chemical fields, are difficult to
characterize.4,5 Computational methods6 can hence be a very
fruitful tool to better understand those fundamental questions.
The purpose of this Article is to review the contributions of
theoretical chemistry in this field of catalysis with surface
organometallic species, in a broad definition, to summarize
the insights that theory can bring to the above-listed
questions. All types of surfaces and modes of interaction
with the organometallic complex are considered, and struc-
tural, spectroscopic, and catalytic reactivity aspects are
included.
2. A Brief Review of Methods and Tools
There are two main classes of models to simulate a
metallic complex grafted on a solid support, also issued from
the two fields that we discussed previously: molecular and
solid-state chemistry. The molecular-based approach consid-
ers a fragment of the support surface and includes it simply
as a ligand of the complex (Figure 1). In the past, the
fragment of the surface was usually very small, which was
a severe limitation for the representation of the support.
Chemical Reviews, 2010, Vol. 110, No. 3 1789
Figure 1. Example of cluster model for a surface organometallic
system. A Zr complex is grafted on an alumina support, modeled
by a cluster saturated by hydrogen atoms (Zr, blue; C, gray; Al,
pink; O, red; H, white).
Nowadays, however, due to the growing computer power,
large fragments can be considered. The termination of the
fragment remains an important region, and, in the case of
oxide materials, dangling bonds are often saturated by
hydrogen or pseudohydrogen atoms. In some cases, the
influence of the “missing” part of the solid can be described
by a proper embedding technique, such as an array of point
charges for an ionic solid, for example. The limitations of
this approach are generally linked with the choice of a too
small cluster and with the insufficient description of the
environment of the chosen surface site.
This molecular approach (also called cluster model) for
the support and grafted complex allows the use of the full
machinery of molecular quantum chemistry. A compromise
has to be achieved between the size of the model and the
accuracy of the quantum mechanical description. Small
models (10 atoms) can be tackled by highly accurate
explicitly correlated techniques7 (CASSCF, CI, ...), while
more realistic models (from 10 to a few hundreds atoms)
are generally considered with density functional theory
(DFT),8 which proposes an averaged and efficient way to
treat the electron-electron interactions. Today, the various
DFT approximations have become very popular, because
they offer a good balance between accuracy and cost (hence,
size of systems accessible). For large systems (large ligands
remaining on the complex, for example), hybrid methods
(QM/MM),9 which combine a quantum chemical treatment
for a core part, and a semiempirical approach for the
surrounding can also be used.
The second class of models consists of describing the
surface by a periodic slab, from a unit cell that is repeated
in two directions10 (Figure 2). In the direction perpendicular
to the surface, the slab is composed of a finite number of
atomic layers. A metal complex can be grafted on each of
these unit cells.
When the grafted complex bares large ligands, the size of
the unit cell must be big enough to avoid lateral interactions
between neighboring complexes on the surface. Such periodic
systems can be tackled with methods based on Bloch’s
theorem and coming from the solid-state community. These
methods describe the electronic structure with the density
functional theory,11,12 with the Hartree-Fock method,13 or
recently for small systems with perturbation correlation
techniques14 (MP2). The surface is modeled as a two-
dimensional extended system, with a finite number of layers
in the perpendicular direction. This number of layers can,
however, be varied until convergence of the desired property
is obtained. Drawbacks arise from the fully ordered nature
of the model surface, the possible lateral interactions between
1790 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 2. Example of a 2-D periodic slab model for a surface
organometallic system, a Zr complex grafted on a periodic model
of γ-alumina. Two boxes of the periodic system are shown.
the complex and its periodic images, and the more limited
number of quantum methods available.
Both from the cluster and from the slab models, total
energy calculations allow one to optimize the geometry of
various structures for the grafted complexes, and to compare
their stability. These structures can be compared to experi-
mental results, for example to EXAFS data. Density func-
tional theory generally gives molecular and solid structures
that are in very good agreement with crystallographic results.
The accuracy in the formation or adsorption energies is
usually good but with a sizable error bar of 10-30 kJ mol-1.
Indeed, if DFT formalism is exact in principle, one has to
rely on approximate exchange-correlation functionals,6-8 and
this is the main source of error. Other approximations relate
to numerical aspects, such as completeness of the basis set,
number of k points, or use of effective core potential to
describe atomic cores, but these points are usually well
controlled in modern codes and lead to smaller errors in the
formation energy. DFT is usually not very good at describing
systems with weak interactions, such as dispersive van der
Waals interactions. Explicitly correlated techniques are more
accurate, but can only address small systems.
The comparison with experimental data can be enriched
by the simulation of spectroscopic properties. Vibrational
spectra can be simulated at various levels of approximations:
harmonic approximation, anharmonic corrections, and cal-
culations of the intensities. Methods to simulate NMR spectra
have also been developed, both for molecular15 and for solid-
state16 systems.
The exploration of catalytic mechanisms requires specific
methods to follow reaction pathways from the reactants to
the products, and to understand the influence of the grafted
complex on this molecular reactivity. The “nudged elastic
band” method17 is widely used, together with ab initio
molecular dynamics. The description of dynamic trajectories
allows one to go beyond the static description of structure
and reactivity and to incorporate fluxional behavior and
temperature effects.18
Generally speaking, there has been a great evolution in
the calculations, because of the development of new algo-
Sautet and Delbecq
rithms and the growth of the computer power. Very large
catalytic systems can now be simulated, in realistic conditions.
3. Mechanisms of Interaction between a Metallic
Complex and a Solid Surface
Before considering the structure and properties of the
surface organometallic complex, let us first describe the
chemical mechanism leading to its formation, from a
complex and specific reactive functions on the solid surface.
Properties of the final grafted complex will be described in
the following sections.
A large number of chemical mechanisms can be involved
in the grafting of organometallic complex on a solid surface.
We will only describe those that have been explored by a
theoretical approach.
The simplest mechanism is obtained when one of the
ligands of the complex possesses another chemical function,
not implied for the interaction with the metal atom and free
for a coordination with the surface. This process is, however,
borderline of surface organometallic chemistry because the
surface is not directly interacting with the metal atom. This
is the case, for example, for Ru(4,4′-dicarboxylate-2,2′-
bipyridine)(CO)2I2. The bipyridine part of the ligand is
involved in the complexation with Ru, and hence the
carboxylate function is free. It was shown to interact with
Ti Lewis sites on anatase TiO2(101).19 Bidentate binding
from the two carboxylate groups was found to be preferred
despite the mismatch between the O-O distance between
them (6.3 Å) and Ti-Ti distance on the support (Figure 3a).
The flexibility of the ligand is here important for the bidentate
coordination.
Another mechanism described by quantum chemical
calculations is related to the nature of the Ti(IV) active sites
obtained when grafting Cp2TiCl2 (Cp ) cyclopentadienyl)
complexes on mesoporous silica samples.20 The grafting sites
on the oxide support are the surface silanols. The calculations
show that the first stable surface species result from a triple
anchoring, (tSiO)3TiCp, consuming three surface silanols
and replacing the two Cl and one Cp ligand around Ti. After
calcination, (tSiO)3TiOH and (tSiO)2Ti(OH)2 are formed,
and they are more stable than the terminal oxo (tSiO)2TidO
system (Figure 3b). These structures predicted after calcina-
tions are in good agreement with EXAFS data, although
EXAFS cannot easily distinguish between the mono- and
dihydroxyl centers. Significant deviations between calculated
and EXAFS structures are, however, observed for the
cyclopentadienyl intermediate, especially in the Ti-C dis-
tance (2.4 Å in the calculation versus 2.0 Å from EXAFS).
For the quantum calculations, the surface silanols are
represented by minimal cluster models H3SiOH, with some
constraints in their positions. This is, however, a minimal
approach to model the influence of the solid support. This
is completed by a force field calculation for a large
amorphous silica glass model of the silica substrate.
The global anchoring reaction is modeled, but the detailed
elementary steps of the process are not described. The
formation of the hydroxyl function from the (tSiO)3TiCp
surface complex is considered as a hydrolysis reaction:
(tSiO)3TiCp + H2O f (tSiO)3TiOH + CpH
and subsequent hydration and hydrolysis of these surfaces
models have also been modeled.
Catalysis and Surface Organometallic Chemistry Chemical Reviews, 2010, Vol. 110, No. 3 1791

Figure 3. (a) Grafting of Ru(4,4′-dicarboxylate-2,2′-bipyridine)(CO)2I2 on TiO2 anatase described with a cluster model. Reprinted with
permission from ref 19. Copyright 2002 Elsevier. (b) Proposed structures and relative energies for the Ti(IV) active sites obtained after
calcinations when grafting Cp2TiCl2 on silica. Surface silanols are represented by minimal H3SiOH models. Reprinted with permission
from ref 20. Copyright 1997 American Chemical Society.

Adsorption of dirhenium decacarbonyl Re2(CO)10 on

γ-alumina follows from a different mechanism and implies
different surface sites.21 Oxygen atoms bridging two octa-
hedral Al atoms, or bridging one octahedral and one
tetrahedral Al, have been considered as grafting sites of the
metal-carbonyl complexes, using cluster models for the
γ-alumina surface. The grafting proceeds via a decarbony-
lation reaction, and the interaction energy of the unsaturated
subcarbonyl complexes Re(CO)3, Re(CO)4, and Re2(CO)9
with the surface has been calculated. Binding energies range
between 300 and 600 kJ mol-1. The coordination of Al was
found to have a considerable influence on the interaction
energy, the mixed tetrahedral-octahedral site yielding a
significantly stronger interaction. A decarbonylation scheme
has been proposed (Figure 4). A first removal of a CO group
on Re2(CO)10 enables one to attach Re2(CO)9 on the bridging
oxygen. Heating then results in the breaking of the
metal-metal bond and the initial formation of two grafted
Re(CO)4 centers. Further decarbonylation yields the forma-
tion of the stable Re(CO)3/Al2O3. The reaction profile is
globally significantly exothermic, from the strong bond
formed by the unsaturated Re atom and the surface oxygen.
Figure 4. Proposed reaction mechanism for the adsorption and
Another complex used to generate Re surface sites, active decarbonylation of Re2(CO)10 on a γ-alumina surface. Reprinted
in olefin metathesis reaction, is CH3ReO3. Understanding the with permission from ref 21. Copyright 2001 Elsevier.
mechanism leading to the grafting of such complex on a
silica, silica-alumina, or alumina surface is of key impor- involved on the surface. An additional bond is created
tance to determine the structure and chemical properties of between an adjacent bridging oxygen atom (AlOSi) and the
the potential active sites. Two studies combining spectro- Re atom. Structural comparison between EXAFS and DFT
scopic characterization and theory have appeared recently. leads to a very good agreement, with the RedO bond
The first one, combining NMR, IR, EXAFS, and DFT involved in the grafting elongated by 0.1 Å. The authors
calculations, considers the grafting as a Lewis acid-base hence conclude that the grafting creates a well-defined site
interaction, the organometallic species being kept intact.22 with a two-point attachment, which is electronically different
The fully dehydrated silica-alumina support is described by from the molecular precursor. The second study considers a
a siloxane-capped aluminosilsesquioxane monosilanol cube, partially dehydroxylated γ-alumina support with the same
where Al has a tetrahedral coordination. In the most combination of techniques.23 The γ-alumina support is
energetically favorable situation, one of the oxo ligands of modeled by a periodic slab.24-26 The calculations show that
CH3ReO3 interacts with the Al site. Comparison between indeed the main adduct results from the interaction between
EXAFS and DFT suggests that a five-coordinate Al is the oxo ligand and the Al Lewis acid sites (Figure 5a).
1792 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 5. The two proposed grafting modes of CH3ReO3 on
γ-alumina: (a) interaction of the oxygen lone pair with the surface
Al lewis site; (b) dissociation of a C-H bond on a Al-O surface
site. The second species is proposed to be the active site in olefin
metathesis. Re, yellow; C, gray; Al, pink; O, red; H, white.
However, this majority species is not active for the olefin
metathesis reaction. The active site is proposed to be a
minority species resulting from the C-H activation of the
methyl ligand of CH3ReO3 at reactive Al-O dehydroxylated
sites of alumina (Figure 5b). This creates a surface hydroxyl
group and a grafted surface Re methylene complex
AlCH2ReO3, which is the initiating center for the carbene
species propagating olefin metathesis.
From NMR, the surface density of these active Re
methylene complexes was determined to be 0.15 per nm2,
as compared to 1.05 per nm2 for the majority CH3ReO3
surface units. This completely agrees with the result of
catalytic tests with labeled ethene, which shows that the
catalyst contains about 14-15% of active sites, and that the
methylene groups of those sites are involved in the formation
of the propagating carbene species. Hence, the major surface
species, which is seen by EXAFS spectroscopy, is not
responsible for the catalytic event.
The grafting mechanism on γ-alumina was also detailed
in the case of the chemisorptive interaction of two electron-
poor complexes: [Zr(CH2tBu)4] and [W(tCtBu)(CH2tBu)3].27
These are d0 complexes, where the metal is at its highest
oxidation state. This reduces the number of potential mech-
anisms for the grafting to alumina. The chemical active sites
on the γ-alumina surface are the hydroxyl groups resulting
from a partial hydration of the surface in usual operating
conditions. The basic mechanism for the grafting process is
a σ-bond metathesis scheme between Zr-C on the complex
and O-H on the surface, to form the new Zr-O bond and
a C-H bond with the evolution of an alkane molecule,
consuming one of the alkyl ligands of the complex. This
reaction can proceed several times for the same complex,
leading to multipodal interactions between the complex and
the surface.
The calculated energy pathway for the successive grafting
steps of the model Zr(CH3)4 complex on γ-alumina is given
in Figure 6. The initial grafting step proceeds as indicated
with a σ-bond metathesis mechanism between one of the
Zr-C bonds and the surface OH bond. The Zr-C and the
OH bond are broken, while two new bonds are formed in a
concerted manner: a Zr-O bond, covalently attaching the
complex to the alumina surface, and a CH bond liberating a
methane molecule in the gas phase. The transition state is a
four-member ring with a triangular shape, the H atom being
located along the O-C edge. The barrier for this elementary
grafting step is small (37 kJ mol-1), and the reaction is largely
exothermic (200 kJ mol-1). The system has then two
possibilities to evolve further. A second Zr-C bond can
interact with a neighboring OH group, forming a second
covalent Zr-O bond, or, alternatively, one methyl ligand
Sautet and Delbecq
Figure 6. Calculated energy pathway for the successive grafting
steps of the model Zr(CH3)4 complex on γ-alumina and structure
of the proposed final surface complex (Zr, blue; C, gray; Al, pink;
O, red; H, white).
Scheme 1. Two Possible Pathways for the Initial Grafting
Step of the Model W(CH3)3(tCCH3) Compound on
γ-Alumina
remaining on the Zr can shift to a neighboring Al Lewis
acid center, hence creating an ion pair: Zrδ+, Al(CH3)δ-. Both
processes are easy and exothermic, and, in fact, they happen
together successively in an indifferent order to yield the most
probable grafted species, which is shown in Figure 6. The
same process is involved when the real neopentyl ligands
are considered. In that case, the neopentyl group is not
bridged between the Zr and Al, but totally displaced on Al.
Two covalent Zr-O-Al bonds are formed, liberating two
neopentane molecules, in agreement with the mass balance
analysis, and consuming surface OH bonds, as clearly seen
in the evolution of IR spectra. A third neopentyl is shifted
toward a surface Al, hence creating the ion pair, with a
cationic Zr center bearing a single remaining alkyl ligand.
The surface Zr complex is further stabilized by a dative bond
from a surface aluminoxane Al-O-Al bridge toward the
metal. A third elimination, although possible in principle, is
excluded from a kinetic argument.
In the case of the W complex, for which alkyl and carbyne
ligands are present, two reaction pathways are possible and
have been compared (see Scheme 1). The first one goes
through a σ-bond metathesis between W-CH3 and O-H to
form the O-W and H3C-H bonds. The second one results
from the addition of the hydroxyl group onto the carbyne to
give a carbene, which undergoes an R-H-abstraction, thus
releasing an alkane molecule in the gas phase and restoring
Catalysis and Surface Organometallic Chemistry
the WtC bond. While the addition of an AlsOH onto the
carbyne requires one to go through a rather high activation
barrier, that is, ∆rEq ) 126 kJ mol-1, the direct electrophilic
cleavage of the W-C bond by AlsOH gives a much lower
activation energy (81 kJ mol-1), which shows that grafting
probably occurs via the latter pathway, implying the alkyl
ligand as it was the case for Zr. Elimination of a second
alkyl ligand to form a bipodal complex involves a high
barrier, especially if the complete ligands are taken into
account (100 kJ mol-1). It is hence proposed that the grafting
process here stops at the monopodal complex. It can be
underlined that activation barriers for the grafting are higher
in the case of the W complex than for the Zr complex. This
was explained by the rigidity in the complex due to the
presence of the carbyne ligand, rending the deformation in
the pentacoordinated structure of the transition state more
difficult in the case of W. This is well in line with the slower
grafting kinetics in the case of W.
A very natural approach to study the mechanism for
grafting a complex on a surface is ab initio molecular
dynamics, if this process involves energy barriers that are
low enough to be passed after a rather short time for the
trajectory (typically a few picoseconds). This approach
was used to study the mechanism of deposition of
Cu-hexafluoroacetylacetonato-trimethylvinylsilane on
tantalum surfaces.28 This relates to applications in micro-
electronics, but the fundamental aspects are very similar to
surface organometallic chemistry. The dynamic simulations
show that the Ta surfaces are very active and that the organic
ligand undergoes spontaneous decomposition, which results
in a very contaminated and disordered interface. Passivation
of the surface by N2, and formation of a surface nitride,
strongly reduces their activity, and the ligand remains intact
and can be liberated to the gas phase, with the formation of
a clean interface with Cu.
4. Nature and Structure of the Surface
Organometallic Species. Identification of the
Surface Catalytic Active Site
A large number of theoretical studies are devoted to the
determination of the structure and electronic properties of
grafted organometallic complexes. Such approaches are
generally conducted in parallel with experimental charac-
terizations such as EXAFS, NMR, or vibrational spectros-
copy, which would be very difficult to interpret alone. There
are very good illustrations of the potential synergy between
such experimental data and simulations, toward a better
understanding of these complex surface structures.
Indeed, it is generally not easy to determine which
complex-surface structure represents the observed surface
complex as indicated in the methods section. Three types of
models are used for these calculations.
4.1. Clusters as Surface Models
Initially, the oxide surface, in many cases silica, was
modeled by a small cluster with one, two, or three Si atoms
(or equivalent). These small models have been very helpful
in understanding the main characteristics of the bonding
between the surface site and the metal complex. They bear,
however, a number of limitations, inherent to the small size
of the model and to the incomplete description of the
environment of the grafting site. A large class of systems
has been described with such a cluster approach.
Chemical Reviews, 2010, Vol. 110, No. 3 1793
Figure 7. (a) Proposed Ti-peroxo species. All distances are shown
in angstroms, and experimental values are given in parentheses (i,
η2; ii, η1). Reprinted with permission from ref 29. Copyright 2002
PCCP Owner Societies. (b) Structure of dinuclear Cr(II) models
on silica, in a constraint (Dn) or in an open (Dw) geometry. The
elements are coded on a gray scale according to increasing atom
number, H (white) < O < F < Si < Cr (dark gray). Reprinted with
permission from ref 32. Copyright 2002 Elsevier.
A combined DFT-EXAFS study has been proposed in that
spirit to investigate the structure and coordination of the
oxygen donor for the oxidation catalysts formed by Ti in
MCM-41 exposed to tert-butyl hydroperoxide.29 Several
Ti(η2-peroxo) and Ti(η1-peroxo) species have been consid-
ered, both with the model hydrogen peroxide and with the
completely substituted oxygen donor. In contrast with
previous estimates,30 the calculations show that Ti(η1-OOtBu)
is significantly more stable than the η2 structure.
A 6-coordinate Ti(η1-OOR) species (R ) H or R ) tBu)
is then proposed for the peroxide/titanosilicate complex.
EXAFS confirms the 6-coordinate species of Figure 7a but
is unable to distinguish between η1 and η2, both structures
giving equally good fits. Other species such as η2-O2, η2-
OO-, and η1-OO- are ruled out. A previously unidentified
complex Ti(η1-O2H2) is found to be stable with respect to
reactant from the calculations, but does not fit well the
experimental data.
Such Ti-based oxidation catalysts can be modified by
ligands with Lewis base properties, to change the Lewis
acidity on the Ti center and to control the activity and
selectivity of the catalyst.31 Various O-O, O-N, and N-N
bidentate ligands were compared. The Lewis acidity of the
surface complex was tested by the coordination of a NH3
molecule. The energy level of the LUMO of the surface
complex, an indicator of the Lewis acidity, was shown to
give a very good correlation with the activity and the
selectivity of the catalyst for cyclohexene epoxidation (the
lower is the LUMO energy, the higher are the activity and
selectivity). Theory hence appears here as a very useful tool
to design new active and selective catalysts, within this class
of compounds and structures.
In the context of Phillips catalysis for the polymerization
of ethene, mononuclear Cr(II) and Cr(III) sites, as well as
dinuclear Cr(II) sites, have been studied on silica models
(cf., Figure 7b) using density functional theory.32 For a
1794 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 8. Two structures for aluminum trimethyl grafted on an OH group of silica. One methane molecule is evolved in the gas phase.
Optimized structures are obtained from MP2 calculations. Reprinted with permission from ref 39. Copyright 2000 Elsevier.
detailed characterization, calculated harmonic frequencies and
infrared intensities for CO exposed catalysts are compared
to experimental spectra. The mononuclear Cr(II) site, with
a pseudotetrahedral structure, adsorbs two molecules of CO,
and, together with monocarbonyl species, they explain the
triplet band found in room-temperature spectra. The evolution
of this spectra with increasing CO pressure is explained first
by dicarbonyl and tricarbonyl species at dinuclear divalent
chromium sites, and then by the formation of tricarbonyl
complexes at mononuclear Cr(III) sites. An important
consequence of this reassignment is that dichromium species
exist in significant amount on the catalyst.
Ni(II) sites on silica have been tackled with a similar
approach.33 One type of monocarbonyl species, 3-fold
coordinated on the silica, is formed on Ni(II) ions. Only the
model cluster bearing a -1 charge allows reproducing the
Ni-O distances obtained from EXAFS and yields a CO
frequency in agreement with experimental data. It is also
shown that the charge of the cluster strongly influences the
CO vibration, while the size of the cluster has a much smaller
effect. This suggests a rather local nature of the metal-substrate
interaction.
A similar approach was carried out for amide-copper-
silica systems,34 as models for C-Cl bond metathesis
catalysts, and for silica-supported Mo-allyl,35,36 or uranyl
complexes.37 In the latter study, two types of bridged
structure for uranyl silicates have been proposed, which are
consistent with the EXAFS data. A series of metal aquo-
complexes has also been studied, interacting with disiloxane
sites (metal ions ) Mg, Ca, Sr, Ba, Zn, Cd, and Hg).38 One
water molecule in the aquo-complex is simply replaced by
an oxygen atom of a Si-O-Si bridge on the silica surface.
The formation energy of the metal aquo-ion/disiloxane
system and the inverse of the metal-oxygen bond length in
the metal aquo-ions are shown to be pertinent descriptors of
the stability constant for surface complexation.
The coordination of methyl aluminum, zinc, and boron
derivatives on silica was also studied with small clusters in
the context of chemical vapor deposition for electronic
materials.39 Indeed, metal alkyls are promising precursors
for gas-phase epitaxy, thin film growth, and novel surface
design. The metal alkyl molecule M-(CH3)n reacts with the
surface OH group, with a σ-bond metathesis type reaction
(cf., section 3), to form the grafted metaloxy derivative
Si-O-M-(CH3)n-1, one methane molecule being liberated
in the gas phase. A specificity of aluminum trialkyl is to
Sautet and Delbecq
yield a cyclic structure arising from an additional Lewis
acid-base interaction between Al and an oxygen neighbor
of the Si grafting site (Figure 8). Calculation of the free
energy shows that the cyclic structure is slightly less stable
than the structure without this extra Lewis acid-Lewis base
interaction (by 0-4 kJ mol-1 depending on the method).
Not only oxide supports can be addressed, and, for
example, Ziegler-Natta polymerization catalysts can be
modeled by the interaction between a TiCl4 unit and a MgCl2
support, here described as small clusters (with 1 or 2 Mg
ions).40 The structure of the supported complex resembles
the chain-like structure of β-TiCl3, with a vacant site on the
octahedral environment of Ti that allows the further coor-
dination of the alkene, the first step of the polymerization
process (see also section 5.3).
The interaction between transition metal carbonyl com-
plexes and oxide supports has also been the subject of
numerous studies, using such a cluster approach. The
decomposition of Re carbonyls on MgO powder results in
Re(CO)3n+ fragments in interaction with the oxide, which
have been characterized by DFT calculations, in relation with
EXAFS and IR spectra.41 Both dehydroxylated and hydroxy-
lated MgO surface cluster models were considered.
The surface acts as a polydentate ligand, creating new
adsorbate-surface (Re-O) bonds that are as strong as
metal-ligand bonds in usual transition metal complexes.
Hydroxylation of the oxide support can strongly reduce the
adhesion between the organometallic fragment and the
surface. The calculated structural and vibrational parameters
are in good agreement with experimental data, including the
strong red shift for ν(C-O) frequency observed on hydroxy-
lated versus dehydroxylated sites. This approach was further
extended to other examples such as cationic Rh dicarbonyl
complex embedded in the cavities of dealuminated Y
zeolite.42,43 In this case, EXAFS cannot directly determine
the location of the complex in the cavities, because three
different models (with different number of O atoms of the
zeolite interacting with the Rh) are compatible with the data.
This indetermination has been resolved by the DFT calcula-
tions and their correlation with the EXAFS and IR results.
The proposed planar ML4 structure, shown in Figure 9, gives
very good agreement between calculations and experiments
for the bond distances and the CO frequency shifts. This
study clearly illustrates the synergy between theory and
spectroscopy to determine the structure of a metal complex
anchored to a structurally well-defined oxide support. Using
Catalysis and Surface Organometallic Chemistry
Figure 9. Proposed structure for the cationic Rh dicarbonyl
complex in the cavity of dealuminated Y zeolite from a combination
of EXAFS and DFT calculations. Rh, blue sphere; C, gray sphere;
O, red sphere and stick; Si, blue stick; Al, green stick. Reprinted
with permission from ref 42. Copyright 2004 Elsevier.
Figure 10. Proposed structure of niobium-silsesquioxane com-
plexes coordinating three methyl ligands (anion). Reprinted with
permission from ref 47. Copyright 2002 Elsevier.
the same computational approach, the structure of Rh(I)
monocarbonyl species in zeolite was determined and their
IR frequency assigned.44 It was shown that the weak band
at 2093 cm-1 is not associated with RhCO+, as assumed
before, but corresponds to species with additional hydrogen
or dihydride ligands, such as Rh(H2)CO+ or RhH2CO+.
To complete this paragraph on carbonyl complex, we
should also mention a study of Co carbonyls in interaction
with silica, where different types of bonding are compared:
(CO)xCo-Si, (CO)xCo-O-Si, and (CO)x-1Co-CO-Si.
Bonding via the carbonyl ligand was found to be the weakest
interaction, while the strongest one is obtained by bonding
with surface oxygen.45
Figure 11. Optimized structures, labeling, and energies of the three regioisomers of [R-PW11O39{Ru(η6-C6H6)(H2O)}]5-. Energies (in kcal
mol-1) are given relative to the most stable isomer (values in parentheses include the effect of the water solvent treated as a continuum).
For reasons of clarity, the hydrogen atoms on the benzene ligands are not represented. Reprinted with permission from ref 48. Copyright
2006 American Chemical Society.
Chemical Reviews, 2010, Vol. 110, No. 3 1795
4.2. Molecular Models as Supports
In recent calculations, larger clusters tend to be used. An
elegant bridge between simulation model and reality is to
use molecular models of oxide particles, such as silsesqui-
oxane or polyoxometalates. The silsesquioxane ligand (for-
mula R7Si8O12O) has a cubic shape with the 8 Si atoms in
the corners and 12 O atoms in the middle of the edges. The
electronic properties of the model ligand L ) H7Si8O12O
were characterized on a L2TiH2 complex by DFT calcula-
tions.46 It was found to be less electron donating than the
Ph3SiO ligand, in good agreement with the lower deproto-
nation energy and pKa value of H7Si8O12(OH) as compared
to Ph3SiOH. The transition metal atom can also be incor-
porated in the cube framework. This is the case for
niobium-silsesquioxane complexes (see Figure 10), which
have been studied as models for silica-supported transition-
metal catalysts and compared to niobium coordination
complexes.47
The structure around Nb can be described as trigonal
antiprismatic. The Nb(CH3)3 fragment is easily accom-
modated by the silsesquioxane, due to the flexible nature of
the Si-O-Si framework (especially the Nb-O-Si angle).
The electronic structure is simple, with a closed-shell singlet
ground state.
Polyoxometalates are also interesting molecular supports
to graft transition metal complexes. RuII and OsII derivatives
have been incorporated in the monolacunary Keggin anion
[R-PW11O39]7-, and they show a bidentate coordination on
two nonequivalent oxygen atoms of the lacuna. The obtained
complexes have been studied by DFT calculations.48
These calculations show that the observed nonsymmetrical
coordination mode is not governed by charge control but by
orbital interactions. The shape of the HOMO of the lacunary
anion in the geometry from X-ray structures explains the
bidentate and regiospecific linkage. The three possible
regioisomers resulting from the bidentate grafting of the
complex have been studied (Figure 11). The nonsymmetrical
one implying two nonequivalent O atoms (noted R(O1O4))
is indeed the most stable one, but the two others regioisomers
are only slightly less stable.
4.3. Periodic Systems as Models of Supports
A third class of models, which is getting more and more
popular, is to describe the oxide support with a periodic slab,
studied with periodic calculations. In these cases, the
extended nature of the support is well described. The periodic
1796 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 12. Optimized structures for [(tSiO)Re(tCtBu)-
(dCHtBu)(CH2tBu)] using cristobalite as a surface model. Re, blue;
C, gray; Al, yellow; O, red; H, small white. Reprinted with
permission from ref 49. Copyright 2006 The Royal Society of
Chemistry (RSC) for the Centre National de la Recherche Scien-
tifique (CNRS) and the RSC.
and hence perfect “crystalline” nature of the model is,
however, a limitation, because in most cases the support is
a high surface area powder.
The silica-supported olefin metathesis catalyst
[(tSiO)Re(tCR)(dCHR)(CH2R)] (R ) Me and R ) tBu)
has been modeled in such a way.49 This case is especially
interesting because the supported system is highly active for
olefin metathesis, while molecular analogues [(X3SiO)-
Re(tCR)(dCHR)(CH2R)] (X3SiO is triphenylsiloxy or
polymeric silsesquioxane) are inactive. The cristobalite and
edingtonite ideal surfaces are used in the calculations. The
calculated structure agrees with EXAFS data and also yields
very good comparisons for the alkylidene ligand JC-H
coupling constants and νC-H stretching frequency. This
alkylidene fragment shows two conformations (syn and anti,
see Figure 12), and the associated isomers are close in
energy, the syn with the R ligand pointing away from the
surface being always more stable. The striking result,
however, is that the geometry and electronic structure of the
Re species are essentially the same when grafted on the silica
surface, or when attached to the molecular analogues. The
calculations suggest that the siloxy group of the first
coordination sphere of Re determines the metal properties.
The silica surface hence appears just as a bulky chelating
ligand, electronically equivalent to other simpler siloxy
groups. Therefore, the specific catalytic reactivity of the
supported organometallic complex is not related in this case
to an electronic effect induced by the support. The more
likely influence of the solid support is to separate the highly
reactive intermediates and to prevent their deactivation via
dimerization, which is difficult to avoid in liquid phase. A
similar approach has been used in the case of
[(tSiO)W(dNAr)(dCHtBu)(CH2tBu)].50
Beside silica, the alumina support has also been described
by using periodic models. We have already mentioned in
section 3 the study of the grafting mechanisms of Zr and W
complexes on γ-alumina.27 In the case of the Zr tetra-
neopentyl complex, a cationic bipodal structure was proposed
(see Figure 6) with one of the two remaining alkyl ligands
shifted toward an Al atom on the support. This surface
species has been further validated from the simulation of
the 13C NMR spectra, and it has been shown that the
neopentyl ligand interacts with a tricoordinated Al (AlIII,
defect). A similar system has been recently considered by
Marks et al.51 with Cp2Zr(CH3)2 grafted on fully dehydroxy-
Sautet and Delbecq
Figure 13. Most stable adsorption mode of cationic Cp2Zr(CH3)+
on dehydrated alumina. Zr, green; C, gray; H, blue; Al, pink; O,
red. Adapted and reprinted with permission from ref 51. Copyright
2008 American Chemical Society.
Figure 14. Optimized structure of [(tAlsO)W(tCtBu)(CH2tBu)2].
The distances from the EXAFS analysis are indicated in parentheses.
W, blue; C, gray; Al, pink; O, red; H, white. Reprinted with
permission from ref 27. Copyright 2006 American Chemical
Society.
lated γ-alumina (110) (Figure 13). A periodic approach is
used, the electronic wave function being expanded in atom-
centered basis functions. Because OH groups are not present
in that case, the above-described elimination of CH4 from a
CH3 ligand cannot happen, and grafting proceeds by the
formation of an ion pair (Cp2Zr(CH3)+CH3-). The cation
interacts with the surface O atoms (dioxo-bridged or oxo-
bridged structure at both the µ2-O and the µ3-O sites), while
CH3- interacts with a 3-fold surface Al, as evidenced by
NMR.
In the case of the W complex [W(tCtBu)(CH2tBu)3], the
structure of the grafted complex on a surface hydroxyl group
of γ-alumina, involving elimination of one neopentyl ligand
(see section 3), has been validated by a combination of DFT
calculations, EXAFS spectroscopy, and IR spectroscopy.27,52
The optimized structure of the grafted complex is shown in
Figure 14, together with the distances obtained from the
EXAFS analysis. The agreement is very good.
A second way to evaluate this model is to consider how
the stretching frequencies of the remaining surface OH
groups on alumina are affected by the grafting of the
complex.27 Experimentally, after the grafting process, some
of the OH vibrations are shifted and a broad signal at 3650
cm-1 appears. This was interpreted as resulting from an
interaction between the surface OH groups and ligands
remaining on the complex. Several azimuthal orientations
of the complex can occur, leading to different interactions.
In the structure of Figure 14, one surface OH weakly interacts
with the π system of the carbyne ligand (dashed line), and
its stretching frequency is decreased from 3700 to 3630 cm-1.
Other conformers give a smaller decrease of the frequency.
There is hence a clear indication that complex-surface
interaction affects the properties of the remaining surface
OH. This effect is intrinsic to cases where a large number
Catalysis and Surface Organometallic Chemistry Chemical Reviews, 2010, Vol. 110, No. 3 1797

of ligands remain on the complex and was not seen, for Scheme 2. Potential Mechanisms for the σ-Bond Metathesis
example, in the previous case of zirconium. of Ethane on Silica-Supported Ta Monohydridesa
Finally, this structure allows one to reproduce the six major
peaks in the 13C NMR spectrum of the W complex. Again,
the interaction between surface and ligands is important, now
affecting the NMR chemical shifts of the C atoms on the
ligands. Two conformers exist, one showing an interaction
of a vicinal OH group with the carbyne and the other with
its neopentyl ligand, and they allow a full interpretation of
the spectrum.
The interaction between Co2+ cations and γ-alumina has
also been tackled from a slab model.53 The calculations show
that two Co2+ cations can be accommodated in adjacent
positions on the surface with a pseudo tetrahedral coordina-
tion. Only this structure adequately reproduces all experi-
mental data, with a short Co-Co distance of 3.2 Å in
agreement with EXAFS.

5. Catalytic Reactivity of the Surface

Organometallic Complexes
Whereas the reactivity of organometallic complexes in
homogeneous catalysis is well documented from a theoretical
point of view, only a few groups have tackled the study of
the reactivity of supported complexes. This is due to the
difficulty in combining the representation of the support
surface and the calculation of reaction pathways. However,
from the development of new methods and of the growing
power of computers, such studies have become feasible.
The considered grafted organometallic complexes can be
divided into two classes active in two important reactions:
the metal hydrides and alkyls involved in hydrogenolysis
and alkane metathesis and the metal carbenes involved in
alkene metathesis. We will review below these two classes
of catalysts and end this part centered on reactivity by a
section dealing with alkene polymerization.
5.1. Alkane Hydrogenolysis and Metathesis on
Transition Metal Hydrides
Two electron-poor metal hydrides have essentially been
considered in the theoretical studies of alkane activation: Ta
and Zr. In one case, hydrides of vanadium and niobium have
been compared. Finally, the reactivity of an alkyl-tungsten
complex has also been reported.
A first series of papers tackles the activation of alkanes
on silica-supported group VB metal hydrides and more
particularly tantalum hydrides. The silica is represented by
a fragment H4Si2O5(OH)2 extracted from the SiO2 β-cristo-
balite structure. Calculations are based on DFT with the
B3LYP functional. The first paper deals with σ-bond
metathesis of ethane.54 The key difficulty is to determine
the nature and oxidation state of the active surface species.
Starting from a bigrafted tantalum(III) monohydride, the
following reactions can be envisaged (Scheme 2): the first
one involves the cleavage of a C-H bond, then a second
ethane molecule reacts with the ethyl-Ta complex to give
a methyl-Ta complex and the formation of propane. Finally,
the ethyl-Ta complex is regenerated with elimination of
methane.
These reactions can occur in two steps, oxidative addition
and reductive elimination with formation of a Ta ethyl
dihydride intermediate complex, or in a one-step concerted
mechanism with a four-center transition state (σ-bond
metathesis). None is, however, here satisfactory.
a
Reprinted with permission from ref 54. Copyright 2003 Springer
Science+Business Media.
Figure 15. Mechanism for ethane hydrogenolysis involving
π-complexes and carbene complexes. Reprinted with permission
from ref 55. Copyright 2003 Springer Science+Business Media.
Indeed, on one hand, the transition state for the concerted
mechanism could not be detected. On the other hand, the
two-step mechanism proceeds by the formation of a
dihydride-ethyl complex, more stable than the reaction
products by 18 kcal mol-1. The second step and the formation
of the products require passing a high barrier, and hence the
two-step mechanism is difficult, the reaction being blocked
at the intermediate. The mechanism hence remains unre-
solved at this point.
In the second paper,55 hydrogenolysis of ethane is explored
but this time with a different hydride model, a TaV trihydride,
because the calculated IR spectrum is in better agreement
with the experimental spectrum than the one of the TaIII
monohydride (Figure 15). Besides DFT/B3LYP results, MP2
calculations are tested for comparison. The calculation of
the reaction pathways demonstrates that ethane hydrogenoly-
sis may proceed by a mechanism involving π-complexes and
carbene complexes. The first step, metathesis of a C-H bond
giving a dihydride-ethyl Ta complex, is now allowed
because the starting point is a more stable Ta trihydride.
Hence, the catalytic cycles based on a silica-supported
tantalum trihydride can explain the hydrogenolysis activity.
The authors propose that this Ta trihydride is also the active
site for alkane metathesis, following the two-step mechanism
of Scheme 2.
1798 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 3
Scheme 4a
a
Reprinted with permission from ref 59. Copyright 2003 Wiley-VCH
Verlag GmbH & Co. KGaA.
In a third paper, the same group compares the activity of
vanadium, niobium, and tantalum hydrides toward metathesis
and hydrogenolysis of alkanes.56 The calculations are con-
ducted at the CASSCF level because the pathways are studied
both in triplet and in singlet states. A correct description of
the vanadium complexes requires a multiconfigurational
approach. A possible role of the triplet states in the reactions
catalyzed by niobium and vanadium complexes is pointed
out. Indeed, a decrease of the activation energy is observed
with these metals due to a crossing between the triplet and
singlet paths, which leads to a new low-energy high-spin
pathway.
The second type of complexes widely used in alkane
activation is constituted by zirconium hydrides. A review
article summarizes several works dealing with methane
activation and alkane hydrogenolysis.57 The catalysts are
mono- and dihydrides of zirconium supported on silica
(Scheme 3). The support is modeled by the small Si3O6F3
cluster. F atoms are used instead of OH groups to terminate
the cluster. The DFT calculations use the PBE functional
and are spin-unrestricted if necessary.
Butane hydrogenolysis is compared on mono- and dihy-
dride catalysts. On the monohydride, the first step is a C-H
activation of butane on Zr-H, with elimination of H2 (σ-
bond metathesis). The further step involves a β-ethyl transfer,
which is rate determining with a relatively high barrier (31
kcal mol-1). Moreover, the reaction is globally endothermic,
which suggests that on the monohydride the reaction cannot
occur in mild conditions. On dihydrides, the mechanism is
different due to the presence of the second hydrogen atom:
from a Zr butyl hydride complex, the β-ethyl transfer still
has a high barrier (34 kcal mol-1), but it leads directly to a
stable dialkyl complex without passing through the formation
of the nonfavorable olefin complex, as it is the case of
monohydride. Hence, the hydrogenolysis of butane on
dihydrides is feasible at the elevated experimental temper-
atures (423 K).
A similar conclusion on the better reactivity of the
dihydride (tSiO)2ZrH2 as compared to the monohydride
(dSiO)3ZrH has been reached when modeling silica by a
polyhedral silsesquioxane terminated also by fluorine atoms.58
Sautet and Delbecq
Figure 16. Transition-state geometries for H exchange reactions.
Reprinted with permission from ref 59. Copyright 2003 Wiley-VCH
Verlag GmbH & Co. KGaA.
The H/D exchange reaction of alkanes has also been
studied on grafted ZrH by means of DFT cluster calcula-
tions.59 The silica was represented by the same small cluster
as in Scheme 3 but terminated by H instead of F. The PW91
GGA functional was used. Four different pathways were
considered (Scheme 4). In each case, the transition state was
located and characterized.
All transition states correspond to a metal center sur-
rounded by three ligands in a coplanar arrangement (Figure
16). Their energy depends strongly on the nature and the
position of the ligands (H versus CH3) and varies as follows:
H-H-H , H-H-CH3 < CH3-H-CH3 , H-CH3-H
Thus, having a hydrogen as the central atom is an important
factor for a low-energy transition state.
The main conclusion is that the mechanism for H/D
exchange in a CH4/H2 mixture does not involve a direct
exchange (eq 1) or a methyl exchange (eq 4) but rather a
two-step process, C-H activation (eq 3) followed by
hydrogenolysis (eq 3). These pathways always allow the use
of the transition states with the lowest energy.
The last three studies dealing with the reactivity of
supported Zr-H complexes are different in that the support
is not represented by small clusters but by surfaces treated
with periodic calculation methods. The topic of the first paper
is the depolymerization of polyethylene (here modeled by
propane) by hydrogenolysis on silica-supported zirconium
monohydrides.60 The (100) and (111) surfaces of β-cristo-
balite are used as two possible models of a silica surface.
The authors consider (2 × 2) and c(4 × 2) unit cells for the
(100) and (111) surfaces, respectively. The calculations are
based on the Car-Parrinello Molecular Dynamics (CPMD)
method with a mixture of localized and plane wave func-
tions.6 After a study of the surface structures, the grafting
of the ZrH4 complex has been thermodynamically considered.
The formation of SiOZrH3 releases 50 kcal mol-1, and the
formation of (dSiO)3ZrH is also exothermic. The depolym-
erization reaction proceeds in several steps (Figure 17): first,
a σ-bond metathesis to attach propane to Zr with liberation
of H2, then a β-alkyl transfer releasing ethylene, model of a
polymer one C shorter with a terminal double bond, and
finally an insertion of this double bond into a Zr-H bond at
another zirconium site. Methane and ethane, the products of
the hydrogenolysis reaction, are then formed by σ-bond
metathesis with H2.
An important result for this study is that the formed alkene
desorbs and does not remain π-bonded, which avoids a
repolymerization. This contrasts with the previous static
studies on model clusters where the alkene was remaining
adsorbed. A common point, however, is that the C-C
cleavage step (β-alkyl transfer) is found to be the rate-limiting
one. The nature of the surface is important because the
Catalysis and Surface Organometallic Chemistry
Figure 17. Energy diagram for the hydrogenolysis (depolymeri-
zation) of propane on a zirconium monohydride site grafted on a
(100) surface (top) and on a (111) surface (bottom) of cristobalite.
Reprinted with permission from ref 60. Copyright 2000 American
Chemical Society.
barriers for this β-alkyl transfer are 47 and 35 kcal mol-1
on the (111) and (100) surfaces, respectively. The barrier
on the (100) surface is similar to that previously pre-
sented on a cluster model of the Zr monohydride. An
important factor is also the coverage in ZrH sites on the
surface. If the density of ZrH sites is low, the site on which
the β-alkyl transfer occurs must be regenerated by hydro-
genolysis before the alkene can react, which requires a
supplementary energy barrier. If the ZrH sites are numerous,
Figure 18. Reaction path for the formation of [(AlO)2Zr(H)(µ-H)Al]. Energies in kJ mol-1. Reprinted with permission from ref 61. Copyright
2007 American Chemical Society.
Chemical Reviews, 2010, Vol. 110, No. 3 1799
a second site can be used for the next steps of the reaction.
To summarize, this study underlines the important role of
the surface nature and of the density of active sites. It also
underlines the importance of a dynamical treatment in the
case of weakly bound intermediates.
The two last studies for this hydrogenolysis section
differ by the nature of the support, γ-alumina instead of
silica. Studies on this support are less common. As
described in sections 3 and 4, the interaction of
Zr(CH2tBu)4 and W(tCtBu)(CH2tBu)3 with γ-alumina
leads to well-characterized complexes that can be trans-
formed by hydrogen treatment into hydrides or react with
alkanes. The mechanism of formation of alumina-sup-
ported zirconium hydrides by treatment of the grafted
complex [(AlO)2Zr(CH2tBu)+-(CH2tBu)Al-] (see -
Figure 6) with H2 has been investigated by a combined
experimental and theoretical study.61,62 The calculations
were performed within the framework of DFT using a
periodic description of the system, with the PW91
functional and a plane wave basis set. The surface was
modeled by a four-layer slab, the two bottom layers being
maintained fixed. A model complex was chosen where
the neopentyl ligands were replaced by methyls. The
formation of a hydride with elimination of methane occurs
from the hydrogenolysis of a Zr-C bond through σ-bond
metathesis (Figure 18). In the starting complex, two kinds
of CH3 exist, one terminal and one bridging between Zr
and Al. The calculations showed that the most favorable
pathway starts with the hydrogenolysis of the terminal
methyl (barrier of 60 kJ mol-1) and continues with the
bridging methyl (barrier of 108 kJ mol-1). The structure
of the obtained dihydride, with one terminal and one
bridging H, has been evidenced by the match between the
experimental and theoretical vibrational spectra.
Experimentally, the formation of the Zr-hydrides is
accompanied by the concomitant formation of methane and
ethane from hydrogenolysis of the formed neopentane. As a
model, hydrogenolysis of butane has been studied with this
alumina-supported zirconium hydride. The first step is the
C-H activation of butane by σ-bond metathesis. Several
pathways were compared depending on the position of the
activated C-H bond, at a primary or secondary carbon. The
reaction then proceeds through a β-alkyl transfer that requires
a high barrier of 135 kJ mol-1, similar to the one found on
1800 Chemical Reviews, 2010, Vol. 110, No. 3
Scheme 5. First Hypothesis for Propane Metathesis on
(AlO)W(tCtBu)(CH2tBu)2a
a
Reprinted with permission from ref 63. Copyright 2007 Elsevier.
silica.57,60 Hence, the β-alkyl transfer is also the rate-
determining step on alumina support as it is on silica.
The monoaluminoxy complex (tAlO)W(tCtBu)
(CH2tBu)2, formed by the interaction of [W(tCtBu)(CH2tBu)3]
on partially hydroxylated γ-alumina, is active in propane
metathesis. Olefin metathesis has been proposed as the key
process in alkane metathesis, because olefins have been identi-
fied as primary products. The mechanism for the initial alkane
dehydrogenation step to form olefins has been studied by means
of DFT periodic calculations.63 The method was the same as
in the previous study; in the model complex, the neopentyl
ligands were replaced by methyls and (tCtBu) by (tCMe).
The activation of propane by the grafted complex can
proceed in three steps (see Scheme 5): the first one is a C-H/
W-C σ-bond metathesis and is highly activated (140 kJ
mol-1). The second step is a β-H abstraction by the carbyne
ligand and leads to a carbene and a coordinated olefin on
WIV. The formation of a metallacyclobutane, central inter-
mediate for olefin metathesis, can then occur, yielding a d2
complex with a high barrier (130 kJ mol-1).
The activation barriers of this process are too high,
particularly if the entropic effects are introduced, to account
for the experimental results. However, it has been shown
by a combined theoretical and experimental study that the
hydroxylated surface of γ-Al2O3-500 presents defects (as
dehydroxylated AlIII atoms) that can activate H2 and CH4.64
Such activation can be applied to propane: one C-H bond
of propane interacts with a tricoordinated Al atom with a
low barrier (25 kJ mol-1) to give an alkyl-aluminum.
Propene is then obtained by β-H elimination with a high
barrier (140 kJ mol-1) that can be reduced if entropy is taken
into account. The main conclusion of this study is that the
support can have a direct chemical role and take part in the
reaction.
The next step of the whole process, the metathesis of the
formed propene, will be discussed in the next part devoted
to olefin metathesis.
5.2. Olefin Metathesis
While many theoretical investigations of the olefin me-
tathesis reaction have been reported at the molecular level,
modeling homogeneous catalysis, only a few theoretical
studies have been undertaken for reactions proceeding in
heterogeneous catalysis, with complexes supported on a
surface. In a way at the frontier between the two approaches,
a first set of studies considers molecular complexes but with
a siloxy OSiH3 ligand to represent a silica surface. The
Sautet and Delbecq
Scheme 6. Active Site and Molybdacyclobutane
Intermediates for Olefin Metathesis on an
Oxo-Carbene-Molybdene Complex Grafted on Aluminaa
a
Reprinted with permission from ref 69. Copyright 2002 Elsevier.
chosen reaction is ethylene metathesis with d0 Schrock-type
catalysts M(tER1)(dCHR2)(X)(Y). In the first paper,65
the catalyst is the Re alkylidyne complex Re(tCtBu)-
(dCHtBu)(X)(Y). A key step of the mechanism is the
structural preparation of the catalyst for the coordination of
ethylene. The associated barrier depends on the nature of X
and Y. It is the smallest when X is a good σ-donor and Y a
poor one. This explains the high efficiency of the silica-
supported system Re(tCtBu)(dCHtBu)(CH2tBu)(OSit) as
compared to its homogeneous equivalent Re(tCMe)-
(dCHMe)(OR)2. In the second paper,66 several metals are
compared (Re, Mo, W). In all cases, the catalysts are more
efficient when they are unsymmetrical (X * Y). However,
in the case of W or Mo, the OSiH3 group does not give the
same barrier decrease as in the case of Re. This means that
grafting the W or Mo catalyst on silica would not improve
the reaction rate. In the last paper of the series,67 the authors
aim at explaining the deactivation of the surface complex
Re(tCtBu)(dCHtBu)(CH2tBu)(OSit) observed experimen-
tally. This deactivation starts by a β-H transfer trans to the
weak σ-donor ligand (siloxy) at the intermediate metallacy-
clobutane. The most accessible pathway then proceeds with
the insertion of ethylene in the Re-H bond followed by R-H
abstraction yielding byproducts or β-H abstraction leading
to degrafting. The calculated barriers are in the same range
as that of the metathesis reaction, which explains the
formation of the byproducts as primary products and the rapid
deactivation of the catalyst.
Let us now switch to more realistic models of the solid
support. The series of papers by Handzlik and co-workers
deals with olefin metathesis catalyzed by molybdenum
complexes supported on alumina or silica. The first articles
are based on cluster calculations and the last ones on periodic
calculations. In the first three studies, very small Al2(OH)6
model clusters were used for representing the alumina
support, and the active center was a Mo methylidene. The
DFT calculations were carried out with the B3LYP exchange
correlation functional. The comparison of MoVI and MoIV
methylidene in alkene metathesis leads to the conclusion that
the MoIV centers are not active because of a high energy
barrier for the formation of the intermediate metallacyclobu-
tane (singlet case) or a too high stability of this intermediate
(triplet case).68 Hence, only MoVI centers are considered in
the further studies. The addition of ethene to the molybde-
namethylidene leads to a molybdacyclobutane with a trigonal
bipyramid (TBP) geometry, which can rearrange, by a Berry
pseudorotation, to the more stable complex with a square
pyramidal (SP) geometry69 (see Scheme 6).
No transition state could be localized directly connecting
the reactants and the SP cyclobutane. Hence, the metathesis
Catalysis and Surface Organometallic Chemistry Chemical Reviews, 2010, Vol. 110, No. 3 1801

Scheme 7. Two Possible Grafted Sites for Molybdene-Oxo-Carbene Fragment on the (100) Surface of Alumina, Modeled by a
Clustera

a
Reprinted with permission from ref 72. Copyright 2005 Elsevier.

Scheme 8. Degenerate Metathesis

reaction must involve the TBP complex as intermediate. The
free energy barrier for the TBP-SP rearrangement and the
one for the TBP cyclobutane decomposition into ethylene
and molybdenamethylidene, the second step of the metathesis
reaction, are calculated at similar values (∆Gq ≈ 36 kJ
mol-1). Hence, the two reactions are in competition, and the
exothermic rearrangement TBP-SP decreases the overall rate
of ethene metathesis. The same conclusion has been reached
for propene metathesis70 that leads to ethylene and 2-butene.
In this case, the catalytic cycle must be continued by propene where Mo is grafted to a bridged oxygen has a better
addition to the formed Mo-ethylidene species, for which metathesis activity.
the barrier is a little lower than for propene addition to It is known that productive metathesis is always ac-
Mo-methylidene. The cross-metathesis of ethene and 2-butene companied by so-called degenerate metathesis that gives back
has also been studied. The activation energies depend on the starting reactant. It has been observed experimentally that
the length of the alkene and of the nature of the Mo-alkylidene degenerate metathesis on heterogeneous molybdenum sys-
center. The conclusion of this series of works is that, insofar tems proceeds faster than the productive one (Scheme 8).
as the small cluster models are reliable, the rate of alkene This problem has been addressed by comparing the activity
metathesis is slowed by the TBP-SP transformation of the of [Mo]dCH2 and [Mo]dCHCH3 on small Al2(OH)6 mod-
intermediate metallacyclobutane. els.73
The conclusion is that the degenerate metathesis is indeed
It is known experimentally that the activity of heteroge-
favored over the productive one if Mo-ethylidene centers
neous Mo catalysts strongly depends on their structure.
are considered. However, this is not true for the Mo-me-
Hence, the metathesis activity toward propene of monomeric
thylidene sites.
Mo-methylidene centers has been compared on two different
Until now, monomeric Mo centers have been studied. In
surfaces of γ-alumina: (100) and (110).71 The same method
the next paper, dimeric Mo centers are considered and
as before was used, but the clusters are larger and hence
compared to monomeric ones, always with the same cluster
they represent better the surface structure. Two differences
models.74 The dimeric Mo centers are found more active in
have been found between the two surfaces. The first one
alkene metathesis because the cyclo-reversal step leading to
concerns the formation of the molybdacyclobutane: on the
the products of metathesis is preferred over the isomerization
(100) surface it is formed directly, whereas on the (110)
of the TBP metallacyclobutane into the SP form.
surface a propene π-complex is obtained because of the
Finally, the same mechanism has also been studied for
reduced electron density on molybdenum. The second
Mo-alkylidene centers grafted on silica, modeled by
difference is that the TBP conformation of the molybdacy-
clusters.75 The main feature is that the transformation of the
clobutane is more stable than the SP one on the (110) surface.
TBP molybdacyclobutane into the SP conformation is
It results that the (110) surface has a better catalytic activity
kinetically unfavorable, whatever the considered site, which
in alkene metathesis. This study shows the influence on the
explains the high activity in metathesis of the Mo-alkylidene
reactivity of the support nature and of the stability of the
complexes grafted on silica, by comparison with alumina.
intermediates.
The conclusion of this series of studies about alkene
Another important factor is also the nature of the grafting metathesis with molybdenum complexes is that the efficiency
site.72 The Mo-methylidene moiety can replace two terminal of the catalyst depends on two main factors: the ability of
hydroxyl groups on octahedral aluminum atoms (Scheme 7, the complex to make a π-adduct with the alkene and the
left), as it is the case in the papers commented above, or feasibility of the transformation between the TBP and SP
replace only one terminal hydroxyl and be coordinated forms of the intermediate molybdacyclobutane, which com-
directly to a bridged oxygen of alumina (Scheme 7, right). petes with the direct metathesis pathway.
As in the previous paper, a π intermediate complex is The last study on the same catalytic system is based on
formed in the case of the latter grafted complex, because of a different approach because the calculations are periodic.76,77
a reduced electron density on molybdenum, which results The use of an alumina surface instead of clusters allows
in a pathway with moderate energy barriers. Hence, the site for the exploration of more numerous grafting sites and
1802 Chemical Reviews, 2010, Vol. 110, No. 3 Sautet and Delbecq

Figure 19. Structure of the TBP and SP metallacyclobutane intermediate with Mo-methylidene species grafted on a periodic model of
γ-alumina. Reprinted with permission from ref 76. Copyright 2008 Elsevier.

Figure 20. Proposed mechanism for butane metathesis on a W alkyl carbyne complex grafted on alumina, including dehydrogenation/
hydrogenation on the support and olefin metathesis on the supported complex. Reprinted with permission from ref 63. Copyright 2007
Elsevier.
imposes constraints to the structure of the species involved
in the mechanism. The DFT calculations are performed
with the PW91 functional and a plane wave basis set. Two
surfaces of γ-Al2O3 are compared, (100) and (110). A first
result is that the Mo-methylidene species are more stable
on the (110) than on the (100) surface. The same
intermediates are investigated as for the cluster calcula-
tions, the π-complex of ethylene and the metallacyclobu-
tane (see Figure 19).
The calculation of the Gibbs free energy leads to the
conclusion that the relative stability of the different
Mo-methylidene species varies with temperature and water
vapor pressure. Their activity in alkene metathesis strongly
depends on their geometry and location on the γ-Al2O3
surface. As it is the case with the cluster models, the TBP
and SP molybdacyclobutane intermediates are found decisive
for the efficiency of the catalyst. In many cases, these species
are more stable than the reactants, and hence their opening
to restore carbene and olefin is the rate-limiting step of olefin
metathesis. The large majority of sites are blocked at the
molybdacyclobutane step, and their catalytic activity is small.
The species active at low temperature has a reduced stability
because of constraint imposed by the surface that deforms
the Mo environment as compared to a coordination complex.
Therefore, the key result of this study is that the various
described Mo species show a very different activity, as a
function of the constraints imposed by the grafting to the
solid support. Hence, only a small fraction of the Mo sites
are active. To produce efficient catalysts, the grafting process
has therefore to be chemically controlled.
Olefin metathesis also been has been proposed recently
as the central step for the carbon-carbon bond formation
process in alkane metathesis, the transformation of a given
alkane into its higher and lower homologue. In that context,
the mechanism of ethylene metathesis has recently been
studied on a Ta-hydride-carbene complex grafted on a
small molecular model of silica.78 A low-energy pathway is
found, hence underlining the possibility to follow olefin
metathesis pathway within the alkane metathesis mechanism.
One originality of this system is that the standard direct
metathesis pathway with [2 + 2] addition and cycloreversion
steps is forbidden, because it would require passing through
a high energy intermediate with the carbene trans to the
hydride. The pathway involves several low-energy steps of
rearrangement of the metallacyclobutane, to avoid that
configuration.
Similarly, propene metathesis has been proposed as the
central step for the propane metathesis mechanism on
(tAlO)W(tCtBu)(CH2tBu)2 (see section 5.1).63 The whole
catalytic cycle is given in Figure 20.
To obtain the carbene required for metathesis, the carbyne
ligand isomerizes to a bis-carbene with a small barrier (56
kJ mol-1). The rest of the mechanism then proceeds with
low barriers, in contrast to the direct propane metathesis
(Scheme 5). The reason is that, in propane metathesis, the
metallacycle is a d2 complex of WIV, whereas, in propene
metathesis, the metallacycle is a more stable d0 complex of
WVI. Hence, these results confirm the hypothesis that olefin
metathesis is the central process in alkane metathesis.
5.3. Polymerization
One first example in this domain is the mechanistic study
of ethylene polymerization on silica-supported zirconium
Catalysis and Surface Organometallic Chemistry
Figure 21. Reaction energy, free energy profiles (at 298 K), and structure of intermediates for ethylene polymerization on grafted Zr-H
species. Reprinted with permission from ref 54. Copyright 2003 Springer Science+Business Media.
hydrides (tSiO)3ZrH and (tSiO)2ZrH2 (Figure 21).54 The
support is represented by a very small cluster Si3O6F3, and
the calculations are performed with the PBE functional. The
first step of the reaction involves insertion of ethylene into
the Zr-H bond, and then the chain grows by successive
insertions into the newly formed Zr-C bonds. The rate of
ethylene polymerization is found to be very similar if the
monohydride or the dihydride is considered as the active site.
In both cases, the reaction is easy with barrier in free energy
not exceeding 20 kcal mol-1.
Ethylene polymerization reactivity has been also explored
by Marks et al.51 on Cp2ZrCH3+ adsorbed on dehydrated
alumina. Two adsorption sites of the cation are compared
(on a µ2 or on a µ3 O atom on alumina), and the polymer-
ization activity is compared to that of the homogeneous
catalyst. The activity is correlated with the energy of the
ethylene π-complex formation. The two surface cationic
complexes give a different reactivity, only the least stable
one being highly active (more than the homogeneous
analogue). This is in good agreement with experimental data,
which indicates that only a fraction of the Zr sites on alumina
are active.
The next two studies do not involve organometallic
complexes. They can nevertheless be related to surface
organometallic catalysis because the initiation step is the
formation of a metal-alkyl bond giving the reactive complex.
The first study deals with ethylene polymerization cata-
lyzed by a pseudotetrahedral mononuclear Cr(II)/SiO2 site.79
The calculations are based on DFT, and the catalyst is
modeled by small silicachromate clusters. The initiation step,
shown in Figure 22, is supposed to be the formation of either
an ethenylhydridochromium(IV), an ethylidenechromium, or
a chromacyclopentane from the interaction of one or two
ethylene molecules.
From these species, several starting points for polymeri-
zation are derived and compared for the propagation phase.
The first conclusion is that chromacyclopentane, obtained
by fixation of two ethylene molecules, is the dominating
initial species. Chain growth proceeds through direct ethylene
insertion into a Cr-C bond with a large barrier of 119 kJ
mol-1 (Figure 23). Experimentally, the formation of 1-alkene
in the early stage of the reaction is reported. From the formed
chromacycloalkane, a β-H transfer to the R carbon at the
other end of the carbon chain leads to a coordinated alkene.
Chemical Reviews, 2010, Vol. 110, No. 3 1803
Figure 22. Initiation step for ethylene polymerization on model
Cr(II)/SiO2. Energies in kJ mol-1 relative to free complex and
ethylene. Reprinted with permission from ref 79. Copyright 2000
Elsevier.
This reaction is in competition with the growth of the chain
by further insertion of ethylene into the chromacycloalkane
and shows a slightly smaller barrier (98 kJ mol-1), hence
explaining the formation of 1-alkene. The barrier for ethylene
insertion into this cyclopentane (119 kJ mol-1) is too high
to explain the polymerization. A key factor to control the
barrier seems to be the OCrO angle between the chromate
and the siloxanes, which means here again that the strain
induced by the support plays an important role.
In a second paper by the same authors,80 two other
mononuclear Cr(II) sites are considered: a highly strained
pseudo-octahedral and a pseudotetrahedral with a supple-
mentary coordinating silanol group. With a pseudo-octahedral
site, the barrier for ethylene insertion is largely reduced but
not enough to be competitive with the release of small
alkenes. In contrast, the pseudotetrahedral site assisted by a
vicinal silanol seems to be a good precursor for polymeri-
zation: a hydrogen transfer occurs from the silanol to form
a four-coordinated ethylchromium(IV) species on which
subsequent insertion of ethylene proceeds with a modest
barrier. Nevertheless, the best model cluster found for
interpreting the polymerization of ethylene by the Cr/SiO2
system is a Cr dinuclear site,81 due to a low insertion barrier.
The comparison of these three articles shows the impor-
tance of the environment of the active site: ligand field of
the metal and participation of the surface to the reaction.
Nevertheless, the small size of the clusters simulating the
surface could introduce spurious effects and allows geom-
1804 Chemical Reviews, 2010, Vol. 110, No. 3
Figure 23. Stationary points along the reaction path of 1-hexene formation on model Cr(II)/SiO2. Activation energies in kJ mol-1. Reprinted
with permission from ref 79. Copyright 2000 Elsevier.
Figure 24. Most important steps in the insertion reaction for the
5-fold site in the case of Ziegler-Natta polymerization of ethylene
with titanium active sites supported on the (110) surface of MgCl2:
the ethylene approach (a), the π-complex formation (b), the
transition state (c), and the formed chain (d). Reprinted with
permission from ref 82. Copyright 1998 American Chemical
Society.
etries for approaching ethylene that are not realistic on a
real surface.
The second topic concerns the Ziegler-Natta polymeri-
zation of ethylene with titanium active sites supported on
the (110) surface of MgCl2.82 The method used is totally
different from the previous one because the study is based
on DFT periodic molecular dynamics using the Car-Parrinello
scheme (CPMD). Nevertheless, as previously, the initiation
step requires the pre-existence of a metal-carbon bond and
of an under-coordinated site (Ti(IV)). Two different con-
figurations are compared for titanium. In both, Ti is sur-
rounded by four Cl and one CH3 ligands but in different
geometries. The barriers for ethylene insertion are calculated
to be low (13-15 and 7 kcal mol-1, respectively), which
can be explained by the formation of an agostic interaction
between Ti and one H of the methyl group that reduces the
steric hindrance with the approaching ethylene. The best site
is a 5-fold site in which one Cl atom bound to Ti has no
bond with the substrate and can easily be displaced, allowing
the alkene to approach (see Figure 24). Hence, the reactivity
depends on the local geometry.
The further step of the mechanism, the insertion of a
second ethylene, has also been investigated, starting from
(d) in Figure 24. The ethylene coordination requires a large
displacement of the chain, which orientates parallel to the
surface. The reaction then proceeds with the same low-energy
Sautet and Delbecq
barrier as for the first ethylene molecule and is also agostic-
assisted. This work therefore depicts a realistic pathway for
explaining the polymerization of alkenes by the Ziegler-Natta
catalysts and provides support for the agostic-assisted inser-
tion of alkenes in the Ti-C bond. The results compare well
with those of calculations done on nonsupported Ti cata-
lysts.83
In a second paper by the same authors, the problem of
isotacticity is addressed with the same method as previ-
ously.84 The polymerization of propene is studied with the
same Ti catalyst on MgCl2 (110). On the 5-fold catalytic
center, propene can approach only with CH3 far from the
substrate. For the approach of a second propene molecule,
several geometries are tested. One of them, where propene
inserts into the Ti-CH3 bond in the same position as for
the first insertion, leads to stereoselectivity. In this case, the
barrier is the lowest among all possible situations (10.5 kcal
mol-1), and the insertion is R-agostic assisted. It is clear that
the resulting stereoselectivity is a consequence of the local
geometry of the active site.
Finally, a comparison between various surfaces of MgCl2,
(100), (110), and (104), and between mononuclear, as studied
before, and binuclear complexes is made in a third study.85
After having studied the structure of the three surfaces, the
authors investigate the grafting of TiCl4 and Ti2Cl6 precursors
of the alkyl-Ti complexes needed for polymerization. Stable
TiCl4 adducts are allowed only on the (110) surface where
Mg atoms have a low coordination. In contrast, Ti2Cl6 binds
to the (100) surface with five bonds. To test the ability of
this binuclear complexes to achieve polymerization, the
alkylated center is simulated by substituting one Cl with an
isobutyl group. During the complexation of propene on one
of the Ti atoms, a disruption of the dinuclear species is
observed with the other TiCl4 moiety being expelled. The
active Ti loses also some bonds with the substrate and
becomes Ti(II). The same conclusion is reached if the
calculations are done on the triplet states. Hence, dinuclear
species do not remain stable during polymerization.
The conclusion of these studies on Ziegler-Natta catalysts
is that the only active site is a mononuclear complex grafted
on the (110) surface with a 5-fold configuration allowing
flexibility. These results indicate that the consideration of
Catalysis and Surface Organometallic Chemistry
the surface in the calculations is essential and show that the
use of small cluster models can be insufficient to take into
account the complexity of the Ziegler-Natta catalysts.
6. Conclusion
The various theoretical studies exposed in this Review
bring several types of insights on surface organometallic
complexes: mechanism of formation, nature of bonding with
the surface, structure of the surface organometallic complex,
interpretation of spectroscopic data, nature of reaction
intermediates, and finally catalytic reactivity. Their contribu-
tion is especially important when combined with adequate
experimental data. The association with EXAFS gives a
direct input on the structure. Other spectroscopic techniques
such as IR or NMR play also a central role, because they
establish a link between the reality of the experimental
system (in conditions as close as possible to the catalysis
ones) and the virtual world of models. The simulations of
the spectra are nowadays developing quickly, which gives a
direct (and nonforgiving) test of the theoretical models
against the experimental spectra. For the reactivity, theory
has the major advantage to propose, after computer intensive
simulations, a microscopic view of the mechanism. A
simulation of the reaction kinetics based on this energy
profile from first principles allows again the confrontation
with the experiment.
Concerning surface organometallic chemistry, one of the
key questions is whether the grafted complex is electronically
similar or different as compared to the related molecular
complex. Can unprecedented catalytic properties be obtained?
Contrasting answers are proposed. In some cases, it is
claimed that the oxide support (as silica) is just a ligand
similar to any alkoxy-ligand, nothing special. This is the case
if monopodal complexes are considered, as obtained, for
example, for the silica-supported olefin metathesis catalyst
[(tSiO)Re(tCR)(dCHR)(CH2R)] (R ) Me and R ) tBu).
In contrast, for dipodal coordinations and higher, the rigidity
of the surface “ligand” is an important aspect, imposing
constraints on the grafted complex especially on the angles
around the metal. These constraints will destabilize the
grafted complex, and hence make it intrinsically more
reactive than the homogeneous counterpart. Surface orga-
nometallic chemistry might hence produce a well-controlled
species, which is electronically different from the molecular
precursor and that can exist as isolated molecular species.
In contrast, a simple grafting by a ligand, with no direct
metal-surface interactions, cannot produce this effect. Such
constraints can also be obtained, however, by large poly-
dentate ligands, but their rigidity is reduced as compared to
the surface.
The stability of the active site is also an important aspect
for catalysis. In solution, the electron-deficient complexes
can meet, dimerize, and hence deactivate. When grafted on
the support, such dimerization is impossible, which is an
important indirect aspect for the reactivity of the grafted
system. If not intrinsically more active, the surface organo-
metallic complex can become more stable and have a longer
lifetime.
Another recurrent question is whether or not single sites
are produced. Cases of true single sites are indeed rare.
Comparison of various grafting modes, for example, for
CH3ReO3 or molybdena on alumina, underlines cases where
a mixture of sites, and not a single site, is obtained with
only one being highly active. In such systems, the ultimate
Chemical Reviews, 2010, Vol. 110, No. 3 1805
goal of surface organometallic chemistry is not attained
because only a small fraction of the metal atoms are active.
In such cases, the majority surface species, which is the one
seen by spectroscopy, is not responsible for the catalytic
event. There is a strong added value of modeling in such
cases, because various species, including slightly less stable
ones, can be explored and compared, but the link with
experimental characterization is more difficult.
The next frequently asked question is about the influence
of the support. Will a different structure and reactivity be
obtained if silica, alumina, or zirconia is used? Theory gives
key answers on the role and differences between supports.
Silica and alumina are often used, and sometimes can give
markedly different results. Both supports present, for ex-
ample, OH groups enabling the grafting of alkyl complex.
Alumina, however, presents a richer chemistry. The theoreti-
cal studies have shown that it can play a dual role, as a
support and as a Lewis acid cocatalyst. Indeed, it can form,
in addition to the covalent link, an ion pair with a cationic
metal center showing an enhanced electrophilic character,
as in the case of Zr. This explains why the Zr-alkyl complex
supported on alumina is active for olefin polymerization,
while it is inactive on the silica support, except if Lewis
acid cocatalysts are used. In other cases, a direct chemical
role of the support is involved because the catalyst activation
step occurs on the support. Finally, just changing the structure
of the support ((100) vs (110) surface for β-cristobalite as a
model of silica) can change the reactivity of the grafted
complex.
On more fundamentals aspects, theory brings important
and new insights on the elementary mechanisms, explaining
the formation of the surface complex. In some cases, various
possible pathways (as σ bond metathesis and addition) can
be discriminated, and this is important for the design of new
grafting modes. Theory can also define reactivity indexes,
which allow the comparison of various systems for their
potential reactivity, without undertaking the tedious task to
explore all reaction mechanisms. It can also probe the
importance of the spacing between the grafted complexes,
hence explaining the influence of loading. Two neighboring
sites can be involved in the reaction at high loading, which
is a specific aspect of surface organometallic chemistry.
Simulation of structure and reactivity related to surface
organometallic chemistry is a young research field. Large
perspectives are opened. The pertinence of the approach lies
in the combination of methods: theory, spectroscopy, struc-
ture, reactivity. A strong coupling with kinetics and with in
situ spectroscopic techniques is essential to understand such
catalysts in conditions close to the catalytic ones. Experiment
and simulations could be extended to other spectroscopic
techniques such as Raman, able to probe the bonds between
the complex and the surface, or NEXAFS. In situ NMR is
also a technique of choice. For the simulation techniques,
hybrid methods will certainly develop due to their ability to
describe large systems from a combination of two levels of
theory, a precise one for the metallic complex and a less
accurate one for the environment and the rest of the support.
This will allow the study of catalysis in realistic conditions.
7. References
(1) Copéret, C.; Chabanas, M.; Petroff Saint-Arroman, R.; Basset, J.-M.
Angew. Chem., Int. Ed. 2003, 42, 156.
(2) Cornils, B.; Herrmann, W. A. Applied Homogeneous Catalysis with
Organometallic Compounds, 2nd ed.; Wiley-VCH: New York, 2002;
Vol. 1.
1806 Chemical Reviews, 2010, Vol. 110, No. 3
(3) Cornils, B.; Herrmann, W. A. Applied Homogeneous Catalysis with
Organometallic Compounds, 2nd ed.; Wiley-VCH: New York, 2002;
Vol. 2.
(4) Guzman, J.; Gates, B. C. Dalton Trans. 2003, 3303.
(5) Fierro-Gonzalez, J. C.; Kuba, S.; Hao, Y.; Gates, B. C. J. Phys. Chem.
B 2006, 110, 13326.
(6) van Santen, R. A., Sautet, P., Eds. Computational Methods in Catalysis
and Materials Science; Wiley-VCH: Weinheim, 2009.
(7) Helgaker, T.; Jorgensen, P.; Olsen, J. Molecular Electronic-Structure
Theory; Wiley: Chichester, 2004.
(8) Parr, R. G.; Yang, W. Density-Functional Theory of Atoms and
Molecules; Oxford University Press: New York, 1989.
(9) Morokuma, K.; Maseras, F. J. Comput. Chem. 1995, 16, 1170.
(10) Hoffmann, R. ReV. Mod. Phys. 1988, 60, 601.
(11) Kresse, G.; Furthmueller, J. Phys. ReV. B 1996, 54, 11169.
(12) Segall, M. D.; Lindan, P. J. D.; Probert, M. J.; Pickard, C. J.; Hasnip,
P. J.; Clark, S. J.; Payne, M. C. J. Phys.: Condens. Matter 2002, 14,
2717.
(13) Pisani, C.; Dovesi, R.; Roetti, C. Hartree-Fock Ab-initio of Crystalline
Systems, Lecture Notes in Chemistry; Springer Verlag: Heidelberg,
1988; Vol. 48.
(14) Pisani, C.; Maschio, L.; Casassa, S.; Halo, M.; Schütz, M.; Usvyat,
D. J. Comput. Chem. 2008, 29, 2113.
(15) Ditchfield, R. J. Chem. Phys. 1972, 56, 5688.
(16) Pickard, C. J.; Mauri, F. Phys. ReV. B 2001, 63, 245101.
(17) Jónsson, H.; Mills, G.; Jacobsen, K. W. In Classical and Quantum
Dynamics in Condensed Phase Simulations; Berne, B. J., Ciccotti,
G., Coker, D. F., Eds.; World Scientific: Singapore, 1998; p 385.
(18) Boero, M.; Parrinello, M.; Terakura, K. J. Am. Chem. Soc. 1998, 120,
2746, and references herein.
(19) Haukka, M.; Hirva, P. Surf. Sci. 2002, 511, 373.
(20) Sinclair, P. E.; Sankar, G.; Catlow, C. R.; Thomas, J. M.; Maschmeyer,
T. J. Phys. Chem. B 1997, 101, 4232.
(21) Räty, J.; Suvanto, M.; Hirva, P.; Pakkanen, T. A. Surf. Sci. 2001, 492,
243.
(22) Moses, A. W.; Ramsahye, N. A.; Raab, C.; Leifeste, H. D.; Chatto-
padhyay, S.; Chmelka, B. F.; Eckert, J.; Scott, S. L. Organometallics
2006, 25, 2157.
(23) Salameh, A.; Joubert, J.; Baudouin, A.; Lukens, W.; Delbecq, F.;
Sautet, P.; Basset, J.-M.; Copéret, C. Angew. Chem., Int. Ed. 2007,
46, 3870.
(24) Digne, M.; Sautet, P.; Raybaud, P.; Euzen, P.; Toulhoat, H. J. Catal.
2002, 211, 1.
(25) Digne, M.; Sautet, P.; Raybaud, P.; Euzen, P.; Toulhoat, H. J. Catal.
2004, 226, 54.
(26) Joubert, J.; Fleurat-Lessard, P.; Delbecq, F.; Sautet, P. J. Phys. Chem.
B 2006, 110, 7392.
(27) Joubert, J.; Delbecq, F.; Sautet, P.; Le Roux, E.; Taoufik, M.;
Thieuleux, C.; Blanc, F.; Copéret, C.; Thivolle-Cazat, J.; Basset, J. M.
J. Am. Chem. Soc. 2006, 128, 9157.
(28) Machado, E.; Kaczmarski, M.; Ordejón, P.; Garg, D.; Norman, J.;
Cheng, H. Langmuir 2005, 21, 7608.
(29) Barker, C. M.; Gleeson, D.; Kaltsoyannis, N.; Catlow, C. R. A.; Sankar,
G.; Thomas, J. M. Phys. Chem. Chem. Phys. 2002, 4, 1228.
(30) Sinclair, P. E.; Catlow, C. R. A. J. Phys. Chem. B 1999, 103, 1084.
(31) Fraile, J. M.; Garcia, J. I.; Mayoral, J. A.; Salvatella, L.; Vispe, E.;
Brown, D. R.; Fuller, G. J. Phys. Chem. B 2003, 107, 519.
(32) Espelid, Ø.; Børve, K. J. J. Catal. 2002, 205, 177.
(33) Costa, D.; Martra, G.; Che, M.; Manceron, L.; Kermanec, M. J. Am.
Chem. Soc. 2002, 124, 7210.
(34) Smirnov, V. V.; Golubeva, E. N. J. Mol. Catal. A 2000, 158, 487.
(35) Griffe, B.; Sierraalta, A.; Ruette, F.; Brito, J. J. Mol. Catal. A: Chem.
2001, 168, 265.
(36) Griffe, B.; Sierraalta, A.; Ruette, F.; Brito, J. J. Mol. Struct.
(THEOCHEM) 2003, 625, 59.
(37) Wheaton, V.; Majumdar, D.; Balasubramanian, K.; Chauffe, L.; Allen,
P. G. Chem. Phys. Lett. 2003, 371, 349.
(38) Chang, C. M.; Jalnout, A. F.; Lin, C. J. Mol. Struct. (THEOCHEM)
2003, 664-665, 27.
(39) Bagatur’yants, A. A.; Ignatov, S. K.; Razuvaev, A. G.; Gropen, O.
Mater. Sci. Semicond. Process. 2000, 3, 71.
(40) Puhakka, E.; Pakkanen, T. T.; Pakkanen, T. A. Surf. Sci. 1995, 334,
289.
(41) Hu, A.; Neyman, K. M.; Staufer, M.; Belling, T.; Gates, B. C.; Rösch,
N. J. Am. Chem. Soc. 1999, 121, 4522.
(42) Neyman, K. M.; Vayssilov, G. N.; Rösch, N. J. Organomet. Chem.
2004, 689, 4384.
Sautet and Delbecq
(43) Goellner, J. F.; Gates, B. C.; Vayssilov, G. N.; Rösch, N. J. Am. Chem.
Soc. 2000, 122, 8056.
(44) Vayssilov, G. N.; Rösch, N. J. Am. Chem. Soc. 2002, 124, 3783.
(45) Suvanto, S.; Hirva, P.; Pakkanen, T. A. Surf. Sci. 2000, 465, 277.
(46) Duchateau, R.; Cremer, U.; Harmsen, R. J.; Mohamud, S. I.;
Abbenhuis, H. C. L.; van Santen, R. A.; Meetsma, A.; Thiele, S. K. H.;
van Tol, M. F. H.; Kranenburg, M. Organometallics 1999, 18, 5447.
(47) Belanzoni, P.; Rosi, M.; Sgamellotti, A. J. Mol. Struct. (THEOCHEM)
2002, 579, 181.
(48) Laurencin, D.; Villanneau, R.; Gérard, H.; Proust, A. J. Phys. Chem.
A 2006, 110, 6345.
(49) Solans-Montfort, X.; Filhol, J.-S.; Coperet, C.; Eisenstein, O. New
J. Chem. 2006, 30, 842.
(50) Rhers, B.; Salameh, A.; Baudoin, A.; Quadrelli, E. A.; Taoufik, M.;
Copéret, C.; Lefebvre, F.; Basset, J. M.; Solans-Monfort, X.; Eisen-
stein, O.; Lukens, W. W.; Lopez, L. P. H.; Sinha, A.; Schrock, R. R.
Organometallics 2006, 25, 3554.
(51) Motta, A.; Fragalà, I. L.; Marks, T. J. J. Am. Chem. Soc. 2008, 130,
16533.
(52) LeRoux, E.; Taoufik, M.; Copéret, C.; de Mallmann, A.; Thivolle-
Cazat, J.; Basset, J. M.; Maunders, B. M.; Sunley, G. J. Angew. Chem.,
Int. Ed. 2005, 44, 6755.
(53) Taniike, T.; Tada, M.; Morikawa, Y.; Sasaki, T.; Iwasawa, Y. J. Phys.
Chem. B 2006, 110, 4929.
(54) Mikhailov, M. N.; Bagatur’yants, A. A.; Kustov, L. M. Russ. Chem.
Bull., Int. Ed. 2003, 52, 30.
(55) Mikhailov, M. N.; Bagatur’yants, A. A.; Kustov, L. M. Russ. Chem.
Bull., Int. Ed. 2003, 52, 1928.
(56) Mikhailov, M. N.; Kustov, L. M. Russ. Chem. Bull., Int. Ed. 2005,
54, 300.
(57) Besedin, D. V.; Ustynyuk, L. Yu.; Ustynyuk, Y. A.; Lunin, V. V.
Top. Catal. 2005, 32, 47.
(58) Thieuleux, C.; Quadrelli, E. A.; Basset, J.-M.; Döbler, J.; Sauer, J.
Chem. Commun. 2004, 1729.
(59) Copéret, C.; Grouillet, A.; Basset, J.-M.; Chermette, H. ChemPhy-
sChem 2003, 4, 608.
(60) Mortensen, J. J.; Parrinello, M. J. Phys. Chem. B 2000, 104, 2901.
(61) Joubert, J.; Delbecq, F.; Thieuleux, C.; Taoufik, M.; Blanc, F.; Copéret,
C.; Thivolle-Cazat, J.; Basset, J. M.; Sautet, P. Organometallics 2007,
26, 3329.
(62) Joubert, J.; Delbecq, F.; Copéret, C.; Basset, J. M.; Sautet, P. Top.
Catal. 2008, 48, 114.
(63) Joubert, J.; Delbecq, F.; Sautet, P. J. Catal. 2007, 251, 507.
(64) Joubert, J.; Salameh, A.; Krakoviack, V.; Delbecq, F.; Sautet, P.;
Copéret, C.; Basset, J.-M. J. Phys. Chem. B 2006, 110, 23944.
(65) Solans-Monfort, X.; Clot, E.; Copéret, C.; Eisenstein, O. J. Am. Chem.
Soc. 2005, 127, 14015.
(66) Poater, A.; Solans-Monfort, X.; Clot, E.; Copéret, C.; Eisenstein, O.
J. Am. Chem. Soc. 2007, 129, 8207.
(67) Leduc, A.-M.; Salameh, A.; Soulivong, D.; Chabanas, M.; Basset, J.-
M.; Copéret, C.; Solans-Monfort, X.; Clot, E.; Eisenstein, O.; Böhm,
V. P. W.; Röper, M. J. Am. Chem. Soc. 2008, 130, 6288.
(68) Handzlik, J.; Ogonowski, J. J. Mol. Catal. A: Chem. 2001, 175, 215.
(69) Handzlik, J.; Ogonowski, J. J. Mol. Catal. A: Chem. 2002, 184, 371.
(70) Handzlik, J. J. Catal. 2003, 220, 23.
(71) Handzlik, J. Surf. Sci. 2004, 562, 101.
(72) Handzlik, J.; Ogonowski, J.; Tokarz-Sobieraj, R. Catal. Today 2005,
101, 163.
(73) Handzlik, J. J. Mol. Catal. A: Chem. 2004, 218, 91.
(74) Handzlik, J. Surf. Sci. 2007, 601, 2054.
(75) Handzlik, J. J. Phys. Chem. B 2005, 109, 20794.
(76) Handzlik, J.; Sautet, P. J. Catal. 2008, 256, 1.
(77) Handzlik, J.; Sautet, P. J. Phys. Chem. C 2008, 112, 14456.
(78) Schinzel, S.; Chermette, H.; Copéret, C.; Basset, J. M. J. Am. Chem.
Soc. 2008, 130, 7984.
(79) Espelid, Ø.; Børve, J. K. J. Catal. 2000, 195, 125.
(80) Espelid, Ø.; Børve, J. K. J. Catal. 2002, 205, 366.
(81) Espelid, Ø.; Børve, J. K. J. Catal. 2002, 206, 331.
(82) Boero, M.; Parrinello, M.; Terakura, K. J. Am. Chem. Soc. 1998, 120,
2746.
(83) Kawamura-Kuribayashi, H.; Koga, N.; Morokuma, K. J. Am. Chem.
Soc. 1992, 114, 2359.
(84) Boero, M.; Parrinello, M.; Hüffer, S.; Weiss, H. J. Am. Chem. Soc.
2000, 122, 501.
(85) Boero, M.; Parrinello, M.; Weiss, H.; Hüffer, S. J. Phys. Chem. A
2001, 105, 5096.
CR900295B
Additions and Corrections Chemical Reviews, 2010, Vol. 110, No. 3 1807

ADDITIONS AND CORRECTIONS

Yuguo Zheng,* Xiaolong Chen, and Yinchu Shen.

Commodity Chemicals Derived from Glycerol, an Important
Biorefinery Feedstock (Chem. Rev. 2008, 108, 5253.)
This paper was withdrawn at the request of the Editor-
in-Chief due to violations of the Ethical Guidelines to
Publication of Chemical Research of the American
Chemical Society.

CR100058U
10.1021/cr100058u
Published on Web 02/25/2010