Ethylene Production

CHEMCATCHEM
REVIEWS
DOI: 10.1002/cctc.201200966
Oxidative Dehydrogenation of Ethane: Common Principles

and Mechanistic Aspects
Christian A. Grtner,[a] Andr C. van Veen,[a, b] and Johannes A. Lercher*[a]
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 1 23
These are not the final page numbers!
CHEMCATCHEM
REVIEWS
www.chemcatchem.org
The increasing demand for light olefins and the changing

nature of basic feedstock has stimulated substantial research
activity into the development of new process routes. Steam
cracking remains the most industrially relevant pathway, but
other routes for light-olefin production have emerged. Fluid
catalytic cracking only produces ethene in minor concentrations. Challenged by marked catalyst deactivation, in contrast,
catalytic dehydrogenation ethane up opens a more selective
route to ethene. The oxidative dehydrogenation (ODH) of
ethane, which couples the endothermic dehydration of ethane
with the strongly exothermic oxidation of hydrogen, would potentially be the most attractive alternative route because it
avoids the need for excessive internal heat input, but also consumes valuable hydrogen. In this Review, the current state of
the ODH of ethane is compared with other routes for lightolefin production, with a focus on the catalyst and reactor
system variants. New catalyst systems and reactor designs
have been developed to improve the industrial competitiveness of the ODH reaction of ethane. The current state of our
fundamental understanding of the ODH of light alkanes, in
particular in terms of catalyst and reactor development, is critically reviewed. The proposed mechanisms and the nature of
the active site for the ODH reaction are described and discussed in detail for selected promising catalysts. The reported
catalytic performance and the possible limitations of these
ODH catalysts will be examined and the performance of the
emerging approaches is compared to the currently practiced
methods.
1. Introduction
ers, full-range naphtha (boiling point: 30200 8C) or C6C8 cuts

are typically used as feedstocks. In the Middle East, the feedstock basis has only shifted to ethane in the last decade, thus
leading to attractive production costs.[4] It should be mentioned in passing that the emerging availability of shale gas
will markedly shift the overall picture towards processes that
use light alkanes.
The recurring increase in the price of crude oil and, in particular, the availability of ethane in shale gas, has advanced the
interest in alternative processes for ethene production, including the dehydration of ethanol (thus enabling the utilization of
biomass-derived feeds) and the (oxidative) dehydrogenation of
ethane. This latter process offers conceptual advantages and,
therefore, has been the subject of substantial research activities. However, the lack of suitable catalysts that combine high
activity and selectivity has prevented their industrial realization
so far.
Several excellent reviews have broadly addressed ODH reactions.[610] Because of the special role of the ODH of ethane and
its potential process applications, we have decided to take
a different approach. Therefore, this Review summarizes the
current state of ethene production as a baseline for the potential and challenges that face the development of new processes. This Review also describes insight on the atomic and molecular levels, as well as into the ongoing developments in catalytic processes and reactor engineering.
Ethene is one of the most important building blocks in the

chemical industry. Among organic chemicals, ethene ranks first
with respect to volume, with an annual worldwide production
of 120 106 tons in 2008.[4] Ethene is used to synthesize polymers, styrene, ethene oxide, vinyl chloride and vinyl acetate
monomers, functionalized hydrocarbons (i.e., dichloroethane,
ethylbenzene, acetaldehyde, and ethanol), and many other
basic and intermediate products.
The direct activation and conversion of alkanes into chemicals results in more-complicated process schemes than those
for the analogous processes that employ olefins.[5] Nevertheless, the availability of alkanes, the increasing demand for olefins,[6] and the necessity to minimize the negative environmental impact of a process[7] has stimulated research into new
directions.
Steam cracking, that is, high-temperature pyrolysis in the
presence of diluting steam, is the most established industrial
process for the manufacture of ethene. Typical feedstocks for
this reaction are various grades of naphtha and components of
natural gas. Different steam-cracking processes are known, depending on the length of the carbon chain in the feedstock.
Although naphtha is the main feedstock for ethene production
in Western Europe and Japan, natural-gas-derived feedstocks
are mainly used in the United States and the Middle East. The
lighter feedstocks consist of liquefied petroleum gas (that contains propane and different butanes) or ethane, propane, and
butane, which originate from natural gas. For naphtha crack[a] Dipl.-Ing. MSc. C. A. Grtner, Prof. Dr. A. C. van Veen, Prof. Dr. J. A. Lercher
Department of Chemistry and Catalysis Research Center
Technische Universitt Mnchen
Lichtenbergstr. 4, 85748 Garching (Germany)
Fax: (+ 49) 089-28913544
E-mail: Johannes.lercher@ch.tum.de
[b] Prof. Dr. A. C. van Veen
Current address:
School of Engineering
University of Warwick
Coventry CV4 7AL (UK)
2. Current Practice: Steam Cracking, Ethane

Dehydrogenation and Ethanol Dehydration
2.1. Steam cracking
The steam cracking of hydrocarbons is the most widespread
process for the industrial manufacture of ethene. Steam-diluted alkanes are converted at high temperatures (approximately
800 8C) in reactor tubes, thus leading to homogeneous pyrolysis. A wide variety of feeds with boiling points of up to 600 8C
can be converted through a radical chain mechanism. Having
been practiced for over 50 years, this process is now very
&2&
CHEMCATCHEM
REVIEWS
Christian Grtner, born in Stuttgart
(Germany), studied Chemical Engineering at the University of Stuttgart (Germany) and the University of Wisconsin,
Madison (USA), where he received his
MS in 2009 under the supervision of
Prof. James A. Dumesic. In 2010, he received his diploma from the University
of Stuttgart. Afterwards, he joined the
group of Prof. Johannes A. Lercher at
the Technische Universitt Mnchen
(Germany). His PhD work focuses on
the oxidative dehydrogenation of ethane over supported alkalichloride catalysts.
Andre C. van Veen was born in Issum

(Germany) in 1971 and studied
Chemistry at the Ruhr-University in
Bochum (Germany), where he received
his Diploma in 1997 and his PhD in
2000 with Prof. Martin Muhler. As
a Postdoctoral Researcher, he spent
time at the Institut de Recherches sur
la Catalyse in Villeurbanne (France)
and at the Max-Planck Institut fr Kohlenforschung in Mlheim (Germany). In
2010, he joined the group of Prof. Johannes A. Lercher at the Technische Universitt Mnchen (Germany). His Habilitation in 2011 at the Universit Claude Bernard Lyon
1 (France) focused on structured reactors for heterogeneous catalysis. He became a Lecturer at the TU Mnchen in 2012 and was
later appointed as a Professor at the University of Warwick (UK).
Johannes A. Lercher, born in Vienna

(Austria), studied Chemistry at the
Technische Universitt Wien and, after
a visiting lectureship at Yale University,
he became a Lecturer and Associate
Professor at the same institution. From
1993 to 1998, he was a Professor at
the Department of Chemical Technology, University Twente. In 1998, he
moved to the Department of Chemistry, TU Mnchen, where he assumed
a joint appointment as the Director of
the Institute for Integrated Catalysis at the Pacific Northwest National Laboratory in 2011. He is member of both the Austrian Academy of Sciences and the Academia Europea. His research focuses
on the fundamental aspects of sorption and catalysis on oxides
and molecular sieves, the in situ characterization of catalysts and
catalytic processes, and on catalysis for the sustainable synthesis of
clean-energy carriers.
www.chemcatchem.org
mature. First, the feed is preheated with steam up to the initial
cracking temperature (500680 8C). Subsequently, the mixed
stream is fed into a high-temperature reactor (750875 8C) to
complete the steam pyrolysis, with residence times of 0.1
0.5 s. Cracking furnaces are externally fired and the radiant
tubes (or coils) have different arrangements (split coil or parallel millisecond pyrolysis).[4] The feed is cracked into small olefins and di-olefins. Owing to the high reactivity of the products, the effluent has to be quenched within 0.020.1 s to
avoid product degradation. The products are separated by
a combination of distillation and absorption processes.
Steam cracking can be described by using complex kinetic
models, which account for the progress of the ultrafast reactions in radical chemistry. Parameters that influence the performance and product distribution are residence time, the partial
pressures of the feedstock and steam, and the process temperature (profile) in the reactor. Typical hydrocarbon-conversion
levels reach 70 %,[4] with olefin yields of around 50 % by using
ethane as a feedstock, whereas single-pass conversion yields
are lower in naphtha crackers.
Because the process operates at very high temperatures
with high reactant streams, the requirements on the equipment are demanding. Suitable materials have to be heat-resistant, because temperatures up to 1100 8C are reached in the reaction tubes. Given the importance of heat management, the
sophistication of the engineering of heat exchangers is crucial
and involves great effort.
Although it is the industry standard for ethene production,
steam cracking has disadvantages. It is a highly energy-intensive process; the total energy demand of the produced ethene
is 16 GJ t1 in the case of ethane as a feedstock and 23 GJ t1 if
naphtha is used as a feedstock.[4] A part of the light effluent
from the product gas stream has to be combusted to provide
heat, thereby leading to the formation of CO2 and NOx. Per ton
of ethene, 11.6 t CO2 are produced through this external heating.[4] In addition, notable amounts of coke are formed on the
inside reactor walls, thus requiring periodic reactor shut-down
for maintenance and coke removal[11] by air and steam gasification, because mechanical coke removal is not possible.
2.2. Dehydrogenation of ethane
The catalytic dehydrogenation of alkanes has been industrially
applied since the 1930s. The strongly endothermic and equilibrium-limited process, [Eq. (1)], is performed in fixed-bed
reactors.
C2 H6 C2 H4 H2 Dr H298 136 kJ mol1
Thermodynamic reasons make the use of high temperatures

and low pressures mandatory.[12] For instance, at atmospheric
pressure, reaction temperatures of 550 8C and 700 8C are required to reach ethane equilibrium conversions of 10 % and
40 %, respectively. High pressures shift the equilibrium toward
ethane.
Typical catalysts are based on Cr (i.e., chromiaalumina) or
Pt.[12, 13] The major challenges are the suppression of side reacChemCatChem 0000, 00, 1 23
&3&
CHEMCATCHEM
REVIEWS
www.chemcatchem.org
tions and the formation of a suitable pellet shape of the catalysts, which ensure efficient feed distribution and minimize
pressure drop. Consecutive side reactions lead to the formation of dienes, polymers, and coke.[12] Coke removal from the
catalyst during the frequent regeneration is required to maintain acceptable catalyst lifetimes.
Regular catalyst regeneration occurs by oxidation and the
associated heat of combustion can be recovered in a straightforward manner, especially if several catalyst beds are operated
in parallel. One of the first commercial dehydrogenation processes by using chromiaalumina, Catadiene, was employed
for butadiene production, by using a parallel fixed-bed arrangement of reactors, which are operated in swing mode to
alternate the dehydrogenation and coke removal.
In the late 1980s, the development of new processes broadened the application range to the production of propene from
propane. Catofin employs dehydrogenation and catalyst regeneration in a swing mode that is similar to the Catadiene process. An adiabatic fixed-bed reactor is used and pre-heated
propane is fed into the reactor, with cooling as the reaction
proceeds along the reactor.[14] In the early 1970s, the continuous catalyst regeneration (CCR) concept was introduced by
UOP; the associated process used an adiabatic fluidized-bed
reactor (UOP Oleflex).[10, 12] However, ethane dehydrogenation
by using this route could not compete with steam cracking because of the relatively low activity of the catalysts, thus leading
to low single-pass yields.
2.3. Dehydration of ethanol
3. Oxidative Dehydrogenation of Ethane:

Mechanistic Considerations
In contrast to the above-mentioned processes, the oxidative
dehydrogenation of ethane (ODH) has not been implemented
on a large scale yet. Its conceptual advantage over dehydrogenation, that is, high ethane conversion, potentially lower reaction temperatures (about 400600 8C) as compared to steam
cracking, and the fact that the reaction is exothermic, has spurred growing interest in this process [Eq. (3)].
C2 H6 0:5 O2 ! C2 H4 H2 O Dr H298 105:5 kJ mol1
Short-chain alkanes, in particular ethane that is extracted

from conventional natural gas and shale gas, are cheap and
abundant and, thus, is a suitable feedstock for oxidative dehydrogenation.[19] However, major practical differences are required compared to the established steam cracking. Firstly, O2
is added to the feed, thus imposing additional safety measures
to prevent the thermal runaway of the reaction and consequential explosions. Secondly, the development of suitable catalysts is particularly challenging, because the olefin products
tend to be more reactive than the reacting alkanes. This relatively higher reactivity of the olefins is attributed to enhanced
additional directed bonding to most catalytic surfaces, whereas
alkanes interact almost exclusively through dispersion forces.[1]
Because ethane is converted into ethene through dehydrogenation, some total oxidation, thus forming CO and CO2, as
well as water, may take place. This latter reaction pathway
originates from oxygen insertion into the CH bond during
the primary activation or from the addition of oxygen to
ethene, as shown in Figure 1. Because the latter reaction could
The industrial production of ethanol from biomass, in particular in Brazil, has made it an available commodity for its introduction as a fuel additive or hydrocarbon substitute.[15, 16] Thus,
the dehydration of ethanol, [Eq. (2)], seems feasible as a niche
route to olefins in certain locations.[14] The mildly endothermic
dehydration is catalyzed at around 300 8C by using solid
acids[4] and selectivities of over 99 % are possible.
C2 H5 OH C2 H4 H2 O Dr H298 45:6 kJ mol1
Although the reaction was performed for a long time with

homogeneous catalysts (sulfuric and phosphoric acid), heterogeneous catalysts, such as alumina, silica, and other solid acid
catalysts, have now largely replaced homogeneous
catalysts.[17, 18]
The process layout is straightforward. Ethanol is heated and
fed into the catalytic reactor, followed by a gas/liquid separator, which splits the product stream into a liquid stream that
contains ethanol and water and a gas stream that mainly contains ethene. The ethene stream has to be purified in several
stages. Compared to steam cracking, this process consumes
less energy and the production of carbon and nitrous oxides is
decreased.[4]
Figure 1. ODH reaction of ethane with side reactions.
involve oxygen anions or electrophilic oxygen atoms, it is important to prevent the re-adsorption of ethene and to separate
it from O2. Whereas such oxygenates are intermediates on the
route to total oxidation, many of the targeted uses of, for example, ethene, such as polymerization, require very high purity
of the olefins.[20]
Almost-complete selectivity for a single olefin is only feasible
with ethane as feedstock, because ethene is the only possible
dehydrogenation product and it is sufficiently unreactive towards oxygen addition. The hydrogenolytic CC bond cleavChemCatChem 0000, 00, 1 23
&4&
CHEMCATCHEM
REVIEWS
age of ethane or ethene, which results in the formation of CH4,
is a highly structure-sensitive reaction and, hence, unlikely
under most conditions.[21]
In contrast, this result does not hold true in the case of propene. Nevertheless, the oxidative dehydrogenation of propane
is possible,[22] but the yields of propene are low because the allylic structure that is formed upon further hydrogen abstraction from propene favors further oxidation,[6] thus leading to
oxygenated by-products and, eventually, COx.[23]
The various reactivities of the reactants and the products
can be better understood by considering their CH bond energies. Whereas the weakest CH bond in ethane has a bond
energy of 419.5 kJ mol1, its analogue in propane has a value
of 401.3 kJ mol1. However, the weakest CH bond in ethene
has an energy of 444 kJ mol1 compared to 360.7 kJ mol1 in
propene. A more distinct difference can be seen in terms of
the weakest CC bond, which has a bond energy of
720 kJ mol1 in ethene, because the only CC bond is a double
bond, compared to 413.8 kJ mol1 in propene, in which the
weakest CC bond is a single bond. In general, to achieve high
selectivities, it is mandatory that the weakest CH or CC
bond in the olefin is less than 30 kJ mol1 weaker than the
weakest CH bond in the reacting paraffin.[24] This result is true
for ethane and ethene, but not for propane and propene.
As a consequence, the ODH of ethane is a promising route
with potentially high selectivity.[10, 25] The reported exploratory
approaches range from catalyst development to combining
high activity and selectivity, that is, to tailor the catalyst, in
such a way that it favors hydrogen abstraction, as well as minimizing oxygen insertion,[8, 26] to new reactor concepts.[11, 27, 28]
The emerging paradigm is that the development of suitable
catalysts requires fundamental insight into the reaction mechanism on a molecular level, including the details of the interactions with the catalytic surfaces and the associated gas-phase
reactions.[7]
Because a very diverse range of catalytic materials have
been reported, we will first describe the catalytic chemistry of
the ODH of ethane on these materials, with an emphasis on
the understanding of the catalytically active site. The mechanism and catalytic performance of selected catalysts are discussed with respect to the overall performance, as well as with
respect to the applied reaction conditions. The materials explored include transition-metal, rare-earth-metal, and alkali
metal oxides, as well as supported alkali chlorides. The optimal
reaction conditions for these classes of catalysts differ because
of differences in their catalytic activity, which, for practical reasons, are compensated for by varying the reaction temperature. Because it is rationalized that the first CH activation is
rate determining in most cases, we start by discussing the activation of the first CH bond.
3.1. CH activation and cleavage

Overall, the CH cleavage steps can occur either homolytically
or heterolytically, with homolytic bond cleavage dominant.
www.chemcatchem.org
3.1.1. Single-electron processes
Homolytic CH bond rupture involves unpaired electrons and,
hence, in most cases, higher energy barriers than elementary
steps that involve paired electrons. Moreover, it has to be induced thermally (which can be aided by CH bond polarization) or by radicals.[29] Figure 2 shows a typical one-electron
process, inducing a radical chain mechanism for the oxidative
dehydrogenation of ethane.
Figure 2. Single-electron process for CH bond activation (adapted from

Ref. [1]).
The high energy barriers that are involved in the single-electron process have disadvantages compared to surface catalytic
processes. The involvement of free radicals may induce further
gas-phase reactions, which are complex to control. In ethane
conversion, for example, butanes may form by the recombination of ethyl radicals. Figure 2 shows a good example, because
H2O2 that is formed transiently may decompose into two OHC
radicals, which would further enhance the reaction rate by creating new radical cycles.
The concentration of alkyl radicals and their reactivity
depend on the reaction temperature. Catalytically, one route
for the generation of alkyl radicals involves their formation on
the active surface. Especially at elevated temperatures, the radical is released into the gas phase, in which it reacts further. A
second possibility is that alkyl radicals react close to the catalytic surface, such as in the boundary layer. The third possibility
is that the radical reacts quickly while it is still adsorbed on the
catalyst surface, that is, it undergoes conversion into the corresponding olefin through b-elimination or reaction with carbon
oxides through attack by adsorbed unselective oxygen
species.[7]
Some examples of catalysts that generate radicals on the
surface, with free-radical processes prevailing in the gas phase,
include rare-earth oxides and alkali oxides.[3032] Leveles et al.
exemplified this mechanism for the ODH of propane on MgOsupported Li2O.[33, 34] Their study showed that alkanes were activated on the catalyst surface in the rate-determining step, thus
producing alkyl radicals. However, the overall reaction was
markedly influenced by gas-phase chemistry.
Despite the large number of contributions to this area, the
nature of the active site that abstracts the H atom is still debatChemCatChem 0000, 00, 1 23
&5&
CHEMCATCHEM
REVIEWS
www.chemcatchem.org
ed.[35] A likely model suggests that [Li+O] acts as the active

site to generate the radicals by abstracting a hydrogen atom
and forming an OH group [Eq. (4)].[1]
C2 H6 Li O s ! C2 H5 Li OH s
The alkyl radicals start a chain reaction in the gas phase that
leads to the olefin and further radicals. It should be emphasized that, in the absence of O2, hydrogen and methyl radicals
are the chain propagators with iso- and n-propyl radicals the
main acting species. In the presence of O2, the hydroperoxyl
radical is the main chain propagator and it exists in higher
concentrations than radicals that are formed in the absence of
O2 [Eq. (5)].
C C2 H5 O2 ! C2 H4 C HO2
The hydroperoxyl radical activates hydrocarbons, thus forming H2O2 [Eq. (6)].
C HO2 C2 H6 ! H2 O2 C C2 H5
surface, leads to better control of the reaction and results in

enhanced olefin selectivity.
V- and Ni-containing catalysts also undergo single-electrontransfer processes, but this will be addressed later.
3.1.2. Processes that involve paired electrons

In contrast to single-electron processes for the ODH of ethane,
several catalytic mechanisms involve the simultaneous transfer
of two electrons. In these cases, the first CH bond is cleaved
heterolytically and a redox pair of one or more metal oxides is
involved. In general, the heterolytic cleavage of a CH bond
requires strongly basic sites on the catalyst surface to facilitate
proton abstraction. A nucleophilic lattice oxygen atom of
a basic oxide can serve as such a site.[8] It is disputed whether
the heterolytic splitting can be facilitated by electrophilic
metal cations.[26] Figure 3 shows a typical catalytic cycle, which
involves the transfer of an electron pair, taking a molybdenum
oxide as an example. Vanadium-based catalysts can follow the
same type of paired-electron mechanism.
The latter product decomposes into two hydroxyl radicals,

which act as the main chain propagators[36] and react with hydrocarbons to form water [Eq. (7.1) and (7.2)].
H2 O2 ! 2 C HO
7:1
C HOC2 H6 ! H2 OC C2 H5
7:2
O2 regenerates the surface OH species [Eq. (8.1) and (8.2)].

Li OH s O2 ! Li O s C HO2
8:1
Li OH s C HO2 ! Li O s H2 O2
8:2
Alternatively, regeneration occurs through the ready elimination of water from the surface at elevated reaction temperatures. However, the surface hydroxyl groups are only part of
the cycle and do not have catalytic properties.[33, 37]
Whereas in the prior example, the radical that was formed
on the surface initiated a gas-phase process, homolytic CH
cleavage on the catalyst surface may also be followed by propagation on the catalyst surface.[38] Hence, oxidation reactions in
microporous materials, such as zeolites or redox-active oxides,
may be controlled by surface free-radical reactions. Examples
of catalysts that follow this pathway are vanadium or cobalt
species in zeolites[1] or on oxide surfaces.[38] The catalyst initiates the reaction by the activation of the first CH bond
through the generation of an ethyl radical. This species reacts
quickly on the surface, thereby acting as a chain propagator.
Sauer and co-workers showed that the formation of transient
radicals on an oxide surface is energetically favored over gasphase reactions.[39] However, there are indications that this way
of activating the first CH bond, that is, through a single-electron process, is unique for V-containing catalysts.[40] The absence of radical processes that propagate into the gas phase,
that is, the fact that the reaction is confined to the catalytic
Figure 3. Redox cycle over Mo, which involves a paired electron transfer
(adapted from Ref. [3]).
3.2. Active sites for the ODH of ethane

Having shown that sites with single or paired electrons may be
involved in the oxidative hydrogen-abstraction step, let us turn
to the catalytically active sites. Sites for the ODH reactions of
alkanes can be grouped into those that contain redox-active
metal cations and those that involve anions in the redox cycle.
3.2.1. Redox-active metal oxides
Most of the catalysts that have been reported to be active for
the ODH of ethane fall into this category. The ODH of ethane
on such materials follows a Marsvan Krevelen[41] mechanism.
In a polar route, ethane reacts with a metal-oxide species and
forms an ethoxyhydroxy pair (Figure 3). After the activation
of the b-H atom, ethene is formed and desorbs from the catalyst, thereby leaving a dihydroxy species behind. Water is elimiChemCatChem 0000, 00, 1 23
&6&
CHEMCATCHEM
REVIEWS
www.chemcatchem.org
nated from the dihydroxy species, thereby reducing the metal

oxide.[3] The reduced metal oxide is subsequently re-oxidized
by gas-phase O2[42] .
The cleavage of the first CH bond may occur homolytically
or heterolytically. The most prominent catalysts are based on
vanadium oxide,[9, 38, 43] with g-Al2O3 or SiO2 as typical supports.[44, 45] The oxidic support is not involved in the redox
cycle; however, it does influence the activity and lability of the
oxygen atoms that are associated with V.[46] Three types of
available metaloxide bonds can be distinguished, that is, terminal vanadyloxygen bonds, bridging oxygen atoms between
two V cations, and bridging O atoms between a V cation and
the support. The isolated VO4 species in Figure 4 shows termi-
Figure 5. Mechanism of the ODH reaction over V2O5-based catalysts (S = support, adapted from Ref. [2]).
Figure 4. Different surface vanadium species (adapted from Ref. [1]).
nal V=O bonds and bridging oxygen bonds between a V atom

and the support. The polymeric VO4 species show additional
VO bonds, which bridge between two V atoms. Two questions, that is, which of the oxygen atoms participate in the critical first CH bond activation and what is the structure of the
active site, have been vividly debated, without reaching an unequivocal conclusion.[1]
Ethane can be activated at redox-active metal centers that
involve either one or two centers. The first possibility is that
ethane interacts with the oxygen atoms that are coordinated
to one metal cation, as shown in Figure 3. In this case, two terminal M=O bonds are involved in the ODH of ethane through
the transfer of paired electrons. V2O5-based catalysts with
higher V2O5 loadings, which form polyvanadate species (e.g.,
2 wt. % V2O5 on Al2O3[2]), are believed to operate through
a mechanism in which paired metal centers undergo a singleelectron reduction (in this case from V5+ to V4+).[2] Two neighboring metal sites actively cooperate in the ODH reaction;
however, only one of them performs the first CH cleavage.
This mechanism is shown in Figure 5. The other possibility is
the cooperation between two neighboring redox sites or materials, which involves two or more different anions. However,
in this case, the two redox centers must have different oxidation states. This mechanism is mainly known from selective oxidation of butane,[1, 47] but may also be applied to the ODH of
ethane.
For the ODH of ethane, it is known that neighboring surface
vanadia species lead to enhanced reducibility and, hence, also
to increased reactivity.
Figure 6 shows a mechanism that involves two redox centers

with oxidation numbers +4 and +5. In this case, initial CH activation occurs through heterolytic cleavage.
Figure 6. Cooperation between two redox sites (adapted from Ref. [1]).
3.2.2. Catalysts with non-redox-active metals

The ODH of ethane through homolytic activation of the CH
bonds without a change in the oxidation state of the metal is
also possible. In that case, the anion (in most cases halogens)
assumes the role of the redox-active element. Depending on
the temperature, this process may formally involve single or
paired electrons. These two possibilities cannot be differentiated if the bond-breaking and -forming steps occur without
charge separation.
The first concept in this chapter is an indirect route to oxidative dehydrogenation, which is realized by the halogenation of
CH4 followed by radical formation or acid-catalyzed elimination
and subsequent CC coupling on LaCl3-based catalysts. In particular, HCl and HBr are suitable for such transient halogenations.[1] Of course, ethane can also be directly halogenated and
induced to eliminate HX.[4850] In this case, the formation of
vinyl chloride has been observed.
Mechanistically, O2, in combination with HX, leads to the formation of positively charged halogen atoms on the catalyst
&7&
CHEMCATCHEM
REVIEWS
www.chemcatchem.org
surface. Subsequently, the hydrogen atom in the first CH

bond undergoes electrophilic substitution. To complete the
catalytic cycle, water is eliminated in the next step, by combining one H atom from the alkane and one from HCl, thus replenishing the transferred Cl ion. This process involves unpaired electrons, but not radical gas-phase reactions. However,
the resulting overall chemistry only differs very subtly from
free-radical reactions.
LaCl3-based catalysts are ideally suited to produce chlorinated species because of the suitable strength of the LaCl bond.
If the strength of the metalchloride bond is stronger, similar
catalytic chemistry can be realized, but the chloride transfer is
dramatically decreased and olefins appear as the primary
products.
The second concept comprises a related catalytic route,
which involves halogen species that are based on supported
alkali chlorides. Ethane activation over these catalysts most
likely involves the decomposition of hypochlorite at the surface, thus forming CO and CCl radicals [Eq. (9)].[51]
OCl ! C O C Cl
To complete the list of catalysts that follow this pathway,

rare-earth oxides[30] and LiDyMg mixed oxides[21] also have to
be addressed, even though they do not contain halogen
atoms; they will be explained in detail later on.
A third reaction pathway is realized by catalysts that are
able to form peroxo species,[52, 53] such as vanadium oxides.
Conceptually, V2O5 should allow the formation of such peroxo
species.[53] In such a pathway, the vanadium cation itself does
not have any redox functionality; Figure 7 shows how the oxidation state of the involved oxygen atoms changes instead.
Peroxovanadate species catalyze such a single-site mechanism.
CH bonds on both carbon atoms of ethane are attacked
through homolytic cleavage, followed by the formation of
ethene and two hydroxyl groups on the V atom. Subsequently,
water is eliminated and the V=O bond is formed again.
Subsequently, H is homolytically abstracted from adsorbed

ethane, thus resulting in the formation of a hydroxy group and
an ethyl radical, which recombine to afford ethyl chloride. This
result can also be interpreted as transient halogenation
[Eq. (10)].
Figure 7. ODH reactions with peroxovanadate species.
C2 H6 C O C Cl ! C2 H5 ClOH
10
The next step can proceed through two routes. One possibility is that the ethyl chloride decomposes, thereby forming
the olefin and water, with the release of Cl . In a later step, Cl
can undergo a new oxidation step [Eq. (11)].
C2 H5 ClOH ! C2 H4 H2 OCl
11
The other possibility would be a ClC bond formation on

the surface, thus leading to subsequent homolytic cleavage,
with the release of an ethyl radical. This route would start
a gas-phase radical mechanism[33, 34] [Eq. (12)].
C2 H5 Cl ! C C2 H5 C Cl
12
It is also conceivable that two H atoms are eliminated in

a concerted manner by hypochlorite anions that themselves
are formed upon the concerted reaction of O2 with LiCl.[1] The
hydrogen abstraction leads to the reduction of the hypochlorite anion back into chloride, according to Equations (13.1),
(13.2), and (13.3).
LiCl0:5 O2 ! LiOCl
13:1
LiOClC2 H6 ! LiOH HClC2 H4
13:2
LiOHHCl ! LiClH2 O
13:3
However, V is known to form various metaloxide bonds in

various structural environments, as discussed above.[54] Raman
and UV/Vis studies have shown that peroxo species may not
generally form on certain supports, such as silica, but are conceivable in organometallic compounds.[55] Furthermore, higher
concentrations of Vperoxo moieties can only form at very
low temperatures (90 K),[53] which are not realistic for the ODH
of ethane.
DFT calculations have shown that peroxo species can form
on V2O5 surfaces. However, the substitution of O2 by N2O as an
oxidizing agent (described in more detail in Section 3.3.3) disables the formation of peroxovanadates. It is likely that vanadyl
species are active and selective for the ODH reaction, whereas
peroxo-forming sites are unselective.[56]
3.3. Factors that govern the activity and selectivity of ODH

catalysts
3.3.1. Metaloxygen bond strength
One important parameter, especially for redox-active catalysts,
is the strength of the metaloxygen bond in the active site;
more specifically, its strength and its variation with the degree
of reduction in a catalytic cycle.[57] As the strength of this bond
varies, the activity and selectivity of an oxidation catalyst will
pass through a maximum in terms of the formation rate of the
partial oxidation product, that is, in the specific case discussed
&8&
CHEMCATCHEM
REVIEWS
here, the formation rate of olefins. Such a relation is called
a volcano relationship.
If the metaloxygen bond is weak, the catalyst is active
(e.g., condensed supported vanadium-oxide clusters) but not
very selective, because multiple oxidation steps are facile.[58] As
the bond strength increases (e.g., smaller clusters of supported
vanadia), the reactivity will gradually decrease, but the selectivity will increase.[59] For example, silica supports lead to the
highest selectivity of VO species, because no polymeric vanadia species can be observed.[5]
If one site contains more than one oxygen atom in an optimal material, the release of the second oxygen atom after the
initial reduction step is much more difficult than the first, for
example, in a material with a metal site that is in a high oxidation state combined with a non-reducible oxide site.[60]
Even if the support is not directly involved in most redox
catalysts for the ODH of ethane, its acid/base properties may
play the role of a mediator of the redox-active cluster. A moreacidic support increases the metaloxygen bond strength in
the redox-active cluster, whereas a more-basic oxide decreases
it. This trend can be illustrated with vanadium on a series of
different supports, in which an increasing acidity is correlated
with decreasing activity and increasing olefin selectivity (V2O5/
MgO < V2O5/MgO+Al2O3 < V2O5/Al2O3).[5]
www.chemcatchem.org
3.3.3. Role of O2 and the density of sites
A comprehensive study that compared O2 and N2O as oxidants
was reported by Kondratenko et al.[62] This study on vanadia
that was incorporated into MCM-41 also nicely pointed out the
importance of site isolation and the influence of the density of
active sites. The use of N2O instead of O2 resulted in lower activity, as explained by the fact that N2O had a lower ability to
re-oxidize the VOx sites from the reduced state caused hydrogen abstraction and water elimination. Because the elementary
steps of the reaction are independent of the oxidant and the
same oxygen lattice species are formed regardless whether O2
or N2O is used, the steady-state concentration of lattice
oxygen atoms must account for the difference in reactivity.
Not only is the olefin selectivity markedly higher if N2O is used
as an oxidant, the decrease in olefin selectivity with increasing
activity also becomes less notable with N2O as an oxidant compared to O2. Because N2O tends to re-oxidize the reduced vanadia sites more slowly than O2, the concentration of those
sites is lower, thus lowering the probability that more than
two labile lattice oxygen atoms exist per site. More labile
oxygen atoms per site favor unselective total oxidation reactions and eliminate the possibility that adsorbed oxygen species are in spatial proximity to the vanadia centers (but not incorporated into the lattice),[63] which might enhance total
oxidation.
3.3.4. Desorption and re-adsorption of ethane

3.3.2. Functionality of the active sites and the cooperation between phases
ODH is a complex reaction that requires the active sites to possess multiple qualities. It has to be able to adsorb and chemisorb the alkane, to facilitate the abstraction of two hydrogen
atoms, to be re-oxidized by oxygen from the gas phase, and to
desorb the produced olefin.[57] Typical vanadate or molybdate
anions are capable of performing multiple reaction steps, as
depicted in the examples above. However, highly effective catalysts exist that consist of two phases, which act in a concerted
manner, with the two phases in close spatial proximity to one
another.
The most prominent example of this phenomenon is the
combination of V2O5 and molybdate sites by depositing V2O5
clusters onto a polymolybdate surface that is supported on
alumina.[1, 61] Whereas V2O5 is still considered to be the active
center, Mo seems to decrease the polarity of the metal
oxygen bonds and, hence, increase the acidic character of the
surface without changing the nature of the active V2O5 site. Increasing the V2O5 content up to one monolayer increased the
selectivity for ODH over total oxidation reactions. The ratio between the total oxidation and the ODH reaction was notably
lower than with V2O5 that was supported on Al2O3 without the
molybdates. The presence of the molybdate decreased the interactions between the V cations and the olefin product, thus
leading to accelerated olefin desorption and, hence, higher
selectivities.[1]
Because olefins are more reactive than alkanes, ethene shows

a higher affinity, as well as higher reactivity, with most surfaces
than ethane. Hence, the re-adsorption of the formed olefin on
the catalytic surface limits the selectivity. The stepwise addition
of oxygen is frequently followed by decarbonylation or decarboxylation on acid catalysts. Therefore, the addition of oxygen,
which leads to products such as acetaldehyde or acetic acid,
will, in turn, enhance the selectivity for total oxidation. Because
the re-adsorption of olefins is linked to the presence of accessible metal cations (Lewis acid sites), their presence has to be
minimized. Three options exist to accomplish this goal:
The first option is to use oxides with low specific surface
areas, thereby minimizing the concentration of accessible
Lewis acid sites. Thus, catalysts with high crystallinity show excellent selectivities in the ODH of ethane,[64, 65] as well as in
other oxidation reactions,[66] because they expose a limited
number of unselective sites.
The second concept relies on surfaces that dynamically rearrange to prevent the exposure of cations, such as a supported
molten chloride overlayer. The mobility of the melt prevents
the formation of defect sites (coordinatively unsaturated metal
cations), which are speculated to initiate the reactions that
lead to deep oxidation. In a melt, the surface re-arranges,
thereby quickly limiting the concentration of these exposed
cations by covering the transiently present exposed cations.[51, 67] The fact that predominantly anions that are not able
to attack the double bond of the olefin are present on the
&9&
CHEMCATCHEM
REVIEWS
outer surface further enhances the benefits of a molten
overlayer.
A third concept to improve product desorption is site isolation, which increases the selectivity for partial-oxidation products[68] by restricting the local availability of oxygen. Thus,
labile lattice oxygen or ensembles of atoms that contain
a redox-active site need to be spatially isolated. As pointed out
above, active sites with more than two labile lattice oxygen
atoms facilitate oxygen insertion before desorption, thus eventually leading to COx species.[63] Thus, a good distribution of
labile oxygen atoms and the limitation of the number of
oxygen atoms per site help to prevent total oxidation.[60] On
the other hand, single oxygen atoms on a catalytic surface are
inactive or prone to the generation of allyl radicals, which
react further in the gas phase.[57, 68] Therefore, two labile
oxygen atoms per site appear to be optimal for the highly selective ODH of alkanes.
4. Groups of ODH Catalysts

Having discussed the elementary steps in the ODH process, we
now turn to the description of potential catalysts that are suitable for oxidative dehydrogenation. This discussion should
help to relate the potential elementary steps of reactions to
specific examples of catalysts. Special attention will be paid to
the structures of the active sites on different catalysts for the
ODH of ethane and the common properties and differences
between the various classes of catalysts.
4.1. Transition-metal oxides

4.1.1. Vanadium-oxide-based systems
Supported vanadium-oxide-based catalysts are the most intensively studied of these systems. The V cations may change
their oxidation state between +III, +IV, and +V,[1, 69] thus
making the catalysts redox active, but they can also form
ODH-active peroxo species.[53] In this case, the cation does not
change its oxidation state. It is debated whether or not the activation of the first CH bond occurs through paired- or singleelectron transfer with these catalysts.
In particular, UV/Vis and Raman spectroscopy[70, 71] showed
that vanadium oxide was present in the form of distinct species. VOx sites exist in their isolated or polymeric forms, either
as a 2D layer of interconnected tetrahedral vanadate clusters
(VO43) that are connected through corners[38, 72] or as dimeric
pyrovanadate clusters (V2O74) that are connected through an
oxygen atom, as shown in Figure 4.[73]
To optimize the catalytic performance and to explore the reaction mechanism, many parameters, including the dispersion
of the vanadate centers, the support material (thus potentially
altering the acid/base properties of the catalyst), and the addition of dopants or promoters, were systematically varied. The
most important parameter was the variation of the dispersion
of vanadium oxide on the support, which may markedly influence the reactivity of the catalyst.[74]
www.chemcatchem.org
Because a consensus has not been reached on whether terminal V=O bonds or bridging VOAl bonds form the catalytically active site, let us now introduce the two different leading
paradigms, as exemplified by different experimental approaches and results, followed by their specific interpretations.
Martinez-Huerta et al.[42] studied a vanadia catalyst that was
supported on g-Al2O3 by using temperature-programmed reduction, in situ Raman, IR, UV/Vis diffuse reflectance, and X-ray
photoelectron spectroscopy, combined with kinetic measurements. By synthesizing isolated and polymeric VOx species, as
well as crystalline V2O5 nanoparticles, their combination of
physicochemical and catalytic measurements allowed important conclusions regarding the active site to be drawn.
Whereas at low sub-monolayer vanadia loadings, only isolated surface species with three VO-support bonds and one terminal V=O bond were found, an increase in the vanadia loading resulted in the formation of polymeric vanadia species
with only one VO-support bond, one V=O terminal bond, and
two VO bridging bonds. Only once a surface monolayer of
vanadia was present were 3D V2O5 nanoparticle crystallites
formed on top of the surface monolayer.[75] The catalysts with
higher vanadia loading (mostly polymeric surface vanadia species) showed the lowest reduction temperatures and, thus, the
most reducible. The reduction of isolated surface vanadia species was more difficult for V2O5 nanoparticles. Interestingly, the
V=O bond of the vanadia species becomes stronger with increasing coverage (shift from 1010 to 1023 cm1 for the main
peak in the Raman spectra), even though the reducibility of
the polymeric surface vanadia species was the lowest. Thus,
one may conclude that the rate-determining reduction step
cannot be influenced by the terminal V=O bond, thus suggesting that the terminal V=O bond is not the primary active site
for the ODH of ethane.
Increasing the surface concentration of vanadia species,
which resulted in a higher concentration of bridging VOV
bonds, did not change the catalytic activity, thus suggesting
that those bonds were not catalytically active for ethane activation.[76, 77] Because crystalline V2O5 nanoparticles tend to catalyze total oxidation, the lower selectivity at high V loading can
be explained by the presence of such nanoparticles. However,
it should be kept in mind that the V2O5 nanoparticles might
undergo a solid-state reaction with the support at higher temperatures, thus forming AlVO4 as a new phase. In turn, this reaction would decrease the number of V-free Al2O3 sites, thereby decreasing the trend for total oxidation.[42]
Thus, it was concluded that, in supported surface vanadium
oxide, neither the terminal V=O bond, the bridging VOV
bond, nor V2O5 nanoparticles were active in the ODH reaction.[42] Thus, according to that rationale, the bridging VOAl
bond was the only option left and, therefore, was inferred to
be the active site in this catalytic system. Other authors have
accepted this argument, especially the aspect that the bridging
VOV site cannot be active for the initial CH bond cleavage
in the ODH of ethane.[1]
DFT calculations by using well-defined vanadia clusters with
the metal oxidation state +5 have been found to be an excellent approach for obtaining a more in-depth insight into the
&10&
CHEMCATCHEM
REVIEWS
reactivity.[78, 79] Evaluation of the energy barriers leads to the
conclusion that a hydrogen atom is abstracted through cleavage of the V=O terminal bond, with the formation of transient
radicals and, subsequently, a CV bond.[39, 78, 79] A bridging
oxygen atom could abstract the second hydrogen atom in this
case, thus resulting in ethene formation. One drawback of this
theoretical experiment is that the VO-support bond is not
present in the model vanadia cluster. Thus, further DFT calculations were performed by using a V-substituted silsequioxane
cluster with bridging oxygen atoms between the V atoms and
the support cations.[79] DFT calculations showed that, even in
the presence of VO-support bonds, the terminal V=O (vanadyl) bond was the only active species for the activation and
cleavage of the first CH bond.
To summarize, a consensus exists within the literature that
the bridging VOV bond cannot be responsible for the activation of the first CH bond. However, some discrepancies still
exist between theory and experiment, that is, whether the terminal V=O bond or the VO-support bond is the site for the
activation of the first CH bond.
The concept of site isolation assumes a special role in this
context. Apparently, the spatial proximity of twobut not
morevanadia centers is essential. This result agrees with the
hypothesis that the bridging VOV bond does not boost the
initial CH bond activation in the ODH reaction, but facilitates
the second H abstraction. Thus, an ensemble of two vanadium
centers seems to be the active site, regardless of whether the
V=O terminal bond or the VO-support bond acts as the
active site.
An investigation of the influence of different V loadings
showed that, in accordance with expectations, higher V loading on g-Al2O3 resulted in increased activity and improved
ethene selectivity.[42] This result can be explained by the fact
that V-free Al2O3 cation sites exist on this catalyst, which are reported to be inactive for ODH but to catalyze the total oxidation of the olefin, mainly into CO. Thus, the acid sites of gAl2O3 mainly contribute to the acidity of the supported catalyst
at low V loadings, whereas the vanadate compounds are
mainly responsible for the acidity at higher V loadings. However, in this case, only the vanadia-related acid sites are responsible for product degradation towards carbon oxides.[42] The
overall activity increased with the total concentration of acidic
sites and with the electronegativity of the dopants.[80]
However, there is an upper vanadia-loading limit beyond
which the ODH yield decreases. Whereas the sites on Al2O3
should be covered at such loadings, as-formed V2O5 particles
catalyze the total oxidation of ethene into CO.[42] Neither acid/
base properties nor the chemical composition show pronounced effects on the ODH yield.
The presence of vanadia-free Al2O3 sites leads to the question of whether other supports might be more appropriate.
Indeed, the activity of supported-vanadia-based catalysts
strongly depends on the interactions with the oxide support,
as well as on the structure of the support.[81]
The dispersion and formation of the active sites are influenced by the nature of the support material. A comparison between g-Al2O3 and SiO2 nicely illustrates this fact: g-Al2O3 is
www.chemcatchem.org
a better support than SiO2, owing to the better dispersion of
vanadium oxide and the coordination of the V cations. Its
acidic character plays a less important role in this case.[38] As
discussed above, the active site (two oxygen-bridged tetrahedrally coordinated V cations) are preferentially formed on gAl2O3. These sites can be more easily reduced by ethane than
the tetrahedrally coordinated vanadate monomers, which are
mainly formed on the SiO2 support, and the octahedrally coordinated V species in pure V2O5.[38]
Another proposed support is hexagonal mesoporous silica
(HMS).[82] However, the influence of the V loading on the catalytic performance was explored with respect to the ODH of
propane. Up to 3.2 wt. % V, the ODH activity increased with the
V content. Above 3.2 wt. % V, the reaction selectivity and yield
decreased. This result can be explained by the detection and
characterization of different surface vanadium-oxide species by
temperature-programmed reduction with H2 and UV/Vis spectroscopy, specifically: 1) Monomeric V species in tetrahedral coordination (VO4) that are active and selective in the ODH reaction; 2) oligomeric and distorted tetrahedral vanadium-oxide
species (both up to 3.2 wt. % V) that are highly active but less
selective; and 3) aggregated vanadium-oxide clusters with octahedral coordination (VO6) that have low activity and selectivity. These results confirm that a high V dispersion is crucial for
catalytic performance. Moreover, in this case, the spatial separation of the vanadate centers (site isolation) is a key feature in
determining activity and selectivity.[57, 68]
Another potential catalyst support for vanadate species is
the inexpensive and relatively stable MgO and this catalyst has
been the subject of several studies.[77, 80, 83] Klisinska et al. studied catalysts with a formal concentration of 1.5 monolayers of
vanadium oxide and compared MgO to SiO2 as a support. Dopants, such as K, P, Ni, Cr, Nb, and Mo, influenced the acid/base
and redox properties but did not markedly impact on the
structure of the catalysts. MgO-supported catalysts formed
magnesium orthovanadate (Mg3V2O8)[83] as a sole V-containing
species that was either present in an amorphous bulk phase or
in a thin layer on the support. Only Lewis acid sites were detected on these materials. Their concentration increased with
the electronegativity of the dopant. Although the activities of
the catalysts could not be correlated to their acid/base properties, their selectivity depended on the dopant. The main byproduct in the ODH of ethane was CO2, whereas CO was only
formed in trace amounts.
As MgO is known to have basic sites and vanadate species,
together with the fact that pure V2O5 des not show any basic
sites,[43] the surface acidity and basicity (the acid/base character) of the vanadium species do not seem to have a crucial
impact on the reaction. However, the acid/base properties of
the support may have an impact on the reactivity, which will
be explained below.
Chao et al. investigated MgO as a support by comparing
two groups of mixed-metal oxides, that is, mesostructured Mg/
V oxide and a mixed oxide that was prepared by a solid-state
reaction between MgO and V2O5. For both systems, activity increased with Mg content, but the selectivity for ethene decreased. In comparison, the mesostructured oxides were more
&11&
CHEMCATCHEM
REVIEWS
active and selective. Again, this result can be explained by the
highly dispersed Mg species in the V2O3 phase and the high
surface areas. Mg species are assumed to lower the redox capacities of V-containing species, thereby influencing the activation of ethane.[40]
The influence of the oxidation state of the vanadia species
becomes important if the ODH on vanadia-based systems is initiated by the transfer of paired electrons. For this mechanism,
two vanadia centers have to cooperate as shown in Figure 6.
Whereas only V5+ could be observed in freshly calcined catalysts, V4+ was also detected on used catalysts, especially after
experiments with high conversion levels. Thus, V4+ species
alone seem to favor deep oxidation and, as a consequence,
lead to lower selectivities for the olefin. In passing, a V5+/V4+
ratio of 1:1 yields the maximum selectivity in O-insertion
reactions.[84]
Let us now return to the doping of vanadia-containing catalysts as a means to improve the catalytic performance. Galli
et al. studied the effect of K+ doping into an Al2O3-supported
vanadia catalyst.[85] The incorporation of K+ ions lowers the reducibility of vanadia catalysts. The olefin selectivity in the ODH
of ethane decreased, but, for example, increased in the ODH of
n-butane. The concentration of acidic sites decreased with increasing K+ incorporation, thus suggesting that acidic sites
favor the ODH reactions of short alkanes, whereas basic sites
favor the ODH reactions of longer paraffins.
Other promising catalyst systems include V-containing aluminophosphates (also doped with Co and Mg).[8688] These
active sites, because V and other metals were incorporated
into the microporous aluminophosphate framework, have
been shown, for example, by Concepcin et al. for AlPO-5
frameworks that contained V, Co, and Mg.
Co-APO-5 has been explored in depth, because it should
have the most pronounced redox and acidic properties. This
catalyst was active and selective and CoV-APO-5 was even
much more active for the ODH of ethane. This result was attributed somewhat vaguely to the cooperation between V and
Co sites. A redox couple was assumed to form, but the acidic
character of this pair was assumed to change during the reaction. A similar catalyst, VCoAPO-18, which contained V and Co
in an AEI structure, led to two different redox pairs (V5+/V4+
and Co3+/Co2+). In both cases, the crystal structure and the
presence of the second Co3+/Co2+ redox pair change the redox
properties of the V species. The metal cations also serve as
Lewis acid sites. The catalyst even shows a higher selectivity
than the AFI material. ODH reactions over these catalysts proceed through a Marsvan Krevelen mechanism. The presence
of acidic sites close to the redox center favors a fast desorption
of the olefin intermediates from the catalyst, thus leading to
high selectivity for ethene.[86] Too-high V loadings lead to the
formation of extra-framework VO species, thus favoring total
oxidation. Thus, ideal catalysts have a high degree of incorporation of Co and V into the framework.
The insertion of Mg into the VAPO-5 framework, which is
known to create acidic sites,[89] resulted in a notable increase in
olefin selectivity compared to the Mg-free catalyst (59.7 %
versus 32.1 % at T = 600 8C and a conversion of about 28 %).[88]
www.chemcatchem.org
This example shows again that the spatial proximity of acidic
sites to the redox center promotes the desorption of olefin intermediates, thus leading to a lower total oxidation rate and
higher olefin selectivity at the same conversion.[88] The identical
conversion indicates that the redox sites are active for the paraffin transformation and the improved selectivity in the presence of acidic sites shows that acidic sites are not directly involved in the redox cycle, but rather facilitate the desorption
step.[9095]
This assumption is corroborated by the fact that only low
olefin selectivities are reported in V/Mg-oxide based catalysts
with MgOV pairs, but without the presence of acidic
sites.[88, 96]
Thus, it can be concluded that the spatial proximity of acidic
and redox sites within a molecular-sieve framework is especially positive for ODH reactions. However, the behavior of acidic
sites must not be compared with the influence of acidic sites
in vanadia species that are supported on bulk oxides, in which
acidic centers on the support favor an unselective conversion
of ethane.
Phosphorus is another known dopant. One example is
(VO)2P2O7, both in bulk and supported on TiO2.[97] The supported catalyst contains highly dispersed (VO)2P2O7 and yields a performance that is one order of magnitude higher than the unsupported catalyst. The incorporation of P is another approach
for changing the reducibility of V. The strong interactions with
TiO2 increase the reducibility of V by decreasing the metal
oxygen bond strength (see Section 3.3.1), thereby leading to
very active catalysts.
4.1.2. Mo-based systems
As with V oxides, Mo oxides are also active for the ODH of
ethane.[98101] However, similar to vanadia systems, a consensus
regarding the nature of the active site has not been reached.
On the one hand, the active site has been concluded to be
a molybdenyl group, which activates the first CH bond, inducing a homolytic cleavage through a single-electron process,
because it has been shown that the activity depends on the
concentration of Mo=O bonds.[1, 22] On the other hand, different
authors have studied Al-supported Mo systems, which have
suggested that the anchoring of MoO-support bonds is critical for the ODH reactivity, thus suggesting that oxygen atoms
between the support and the molybdenum species are responsible for the catalytic activity.[102]
Mo can be used on a Si/Ti mixed-oxide support.[103105] The
olefin selectivity is especially increased if Cl ions are
added.[106108] This increase is explained by the formation of
more-complex ligand structures that contain Mo=O species on
SiO2, whereas the formation of these species is hindered on
TiO2. The addition of Cl ions seems to accelerate olefin desorption by weakening the interactions with ethene, thus limiting its further oxidation. Cl ions also lower the oxidation potential of the catalyst by lowering the redox potential of Mo4+
ions,[107] thus leading to a lower mobility of lattice oxygen
atoms. As a result, the addition of Cl ions leads to strongly
bound ethyl species and more weakly bound acetate and forChemCatChem 0000, 00, 1 23
&12&
CHEMCATCHEM
REVIEWS
mate species, which indicate intermediates on the path to
total oxidation.
4.1.3. Mixed-oxide-based systems
Mixed-oxide-based systems appear to be better suited to adjusting the redox potential and acting cooperatively than
single oxides.[64, 109, 110] One of the most popular mixed-oxide
system is based on Mo/V oxides. Mo appears to be particularly
active if employed together with V on a mesoporous Al2O3
support.[98] However, the main active species is still V. Mo and
V do not interact with each other, but the monomeric or polymeric tetrahedral Mo species cover the unselective support
sites, thus increasing the selectivity.
It has been reported that such a catalyst, with an orthorhombic Mo3VOx structure, is very active. The high reactivity
has been associated to the formation of pentagonal Mo6O21
units. The arrangement of these units leads to heptagonal
channels, which constitute micropores. It is speculated that the
catalytic reaction proceeds within these pores.[111]
More-complex systems contain the oxides of V, Mo, Sb, and
Nb, which are mainly prepared by hydrothermal synthesis.[112]
MoVNbOx and MoVSbTe are reported to be very active and selective in the ODH of ethane.[3, 113, 114] The catalysts generally
consist of two crystalline phases, that is, the orthorhombic M1
phase, (AO)22x(A2O)xM20O56 (A = Te, Sb; M = Mo, V, Nb; 0 < x <
1), and the distorted orthorhombic M2 phase, A2OM6O19 (A =
Te, Sb; M = Mo, V, Nb). Further phases have also been found,
that is, TeMo5O16/Sb4Mo10Ox, (V,Nb)xMo5xO14, and bronze
phases.[113] However, the occurrence of a third phase depends
on the synthesis method.[115]
Notably, ethene has a low reactivity, owing to its low affinity
to the M1 phase, which is rich in MoO and VO pairs that
have been reported to be the active sites for the ODH of
ethane. However, for the oxidation of ethene, MoTe sites or
MoNb sites seem to be required, because they form stronger
interactions with olefins.[113]
Botella et al. studied MoVTeNbOx catalysts that were prepared by using different methods.[110, 115] For MoVTeNbOx
mixed-oxide catalysts, cooperation between two crystalline
phases was suggested to explain the structure of an active site
that was assembled from the multifunctional Te2M20O57 orthorhombic phase and the Mo5O14 phase. The orthorhombic
phase contained pentagonal bipyramidic sites and Te-containing hexagonal pores.[110] The catalyst operated at 340 8C with
around 20 % conversion and an ethene selectivity of around
97 %. The calcination temperature was crucial for the formation
of the above-mentioned phases. Similarly, for MoVSbOx catalysts, the (SbO)2M20O56 phase is highly active and selective
(around 65 % conversion and 80 % selectivity). Mo/V-based systems have also been reported to be active if they are promoted with Al, Ga, Bi,[116] and Ce.[117]
4.1.4. Ni- and Co-based systems
NiO-based catalysts are suitable materials for the ODH reaction
that operate at relatively low temperatures (300400 8C). Pure
www.chemcatchem.org
NiO is very reactive towards ethane; however, it exhibits low
selectivity for the ODH of ethane. The selectivity dramatically
improves if supported on an oxide (Al2O3,[118] ZrO2[119]) or if
used as bulk mixed oxide that was doped with promoters,
such as Nb and Sn.[19, 120124]
NiO that is supported on Al2O3 strongly interacts with the
support, thereby forming a non-stoichiometric surface nickelaluminate phase. At higher loadings, capping islands of NiO
particles of increasing size begin to form on top of the nickel/
alumina interface. The interactions with alumina adjust the
electronic properties of NiO such that the decreased oxygen
mobility leads to highly selective catalysts.
The promotion of supported and unsupported NiO-based
systems by the addition of metals cations, including Mo, V, Nb,
Ta, Co, Li, Mg, Al, Ga, and Ti, is possible.[86, 87] The most selective
catalysts are based on Ni/Nb oxides, which achieve ethene selectivities of around 90 % and remain almost constant at conversion levels of up to 20 %. A maximum yield of 46 % has
been reported at 400 8C. Ni cations are speculated to be active
sites for CH activation, with Nb affecting the oxygen
species.[19]
Pure NiO predominantly produces electrophilic O2 and O
species, as well as nucleophilic O2 species, on its surface.[125]
The addition of Nb results in the elimination of the electrophilic oxygen species that are responsible for the total oxidation
of ethane. Thus, the selectivity towards the olefin is improved
by Nb acting as an electron donor. Other dopants increase or
decrease the presence of unselective electrophilic oxygen species.[122] A systematic study of a series of NiO-based mixed
oxides with doping metals that varied from low (+1) to high
valence (+5) showed that the dissolution of cations with
lower/equal valence to nickel (Li+, Mg2+) increased the amount
of non-stoichiometric oxygen species in NiO, whereas the
higher-valence cations (Al3+, Ga3+, Ti4+, Nb5+, Ta5+) acted as
electron donors and decreased the concentration of positive
p+ holes and, consequently, the electrophilic O radicals of the
NiO acceptor.[122]
The ODH of ethane on the NiNbO catalyst follows
a Marsvan Krevelen mechanism. Importantly, the redox cycle
proceeds through a change in the oxidation state of Ni, whereas Nb does not change its oxidation state. In particular, at
large concentrations of Nb (above 15 wt. %), the re-oxidation
of Ni2+ can be retarded by the Nb atoms, thereby decreasing
the catalytic activity of NiNbO compared to pure NiO.[118, 126]
However, the addition of Nb decreases the concentration of
electrophilic oxygen species, thereby increasing the olefin
selectivity.[127, 128]
Notably, an optimum of the reactant partial pressures exists.
Low O2 partial pressures lead to high olefin selectivities, whereas a minimum O2 pressure is needed to keep the catalyst in its
fully oxidized state.[129]
If a catalyst contains more than 15 % Nb, it is deactivated.
This result is related to the formation of a NiNb2O6 phase,
which is thermodynamically stable under the employed ODH
reaction temperatures, but is not active for the ODH
reaction.[128]
&13&
CHEMCATCHEM
REVIEWS
Nickel oxide can be used in combination with ceria as
a mixed-oxide catalyst.[123] Low concentrations of Ce (Ni/Ce > 6)
lead to a marked increase in catalyst activity and selectivity
compared to pure NiO. This increase is related to an increase
in the specific surface area and a decrease in reducibility. The
addition of higher concentrations of ceria (Ni/Ce = 0.20.3)
changes the mechanism. In this case, ceria is responsible for
transportation of the oxygen species from the bulk to the surface, thus forming both NiO and Ce1xNixO2, and leads to
a faster re-oxidation of Ni.
Tungsten is also used to promote NiO.[122] The activity decreases with increasing W content, thus indicating that the
active sites are associated with accessible Ni cations. Again, Wrich and Ni-rich systems behave differently: W-rich systems
mainly form NiWO4 and WO3, whereas Ni-rich systems mainly
form NiO and WOx species. W-rich systems favor ethane decomposition, thus decreasing the selectivity for ethane during
the ODH reaction. However, Ni-rich catalysts show higher
ethene selectivity than pure Ni-based catalysts, but also lower
activity. These catalysts contain small NiO and WOx particles,
with the latter species assumed to block unselective sites of
NiO.
Besides Ni, Co is also an interesting material for the ODH of
ethane. However, Co2+ ions cannot be easily reduced; thus, the
ODH of ethane does not follow the typical Marsvan Krevelen
mechanism.[41] Redox properties are associated with adsorbed
surface-oxygen species, with Co2+ ions acting as a non-redoxactive center.[130] These species may be modified by interactions between Co2+ ions as the active phase and a TiO2 support.[131] Catalysts that contained various Co and P[130] concentrations showed the best performance, with the optimal results
achieved with 7.6 wt. % Co. Whereas the catalyst deactivated
to form a Co/Ti species, a maximum steady-state yield of
13.3 % was achieved at a conversion of around 25 % and a selectivity of around 60 %. Detailed characterization showed that
the active species was Co2+ ions in octahedral coordination.
Co-BaCO3 was explored with CO2 as the oxidant instead of
O2. At 650 8C, ethane conversion reached 48 % with 92.2 %
olefin selectivity.[132] The use of CO2 decreased the flammability
of the reactants and eliminated the total oxidation of ethane
and ethene. Moreover, coke could still be removed from the
catalyst through the reduction of CO2 and the partial oxidation
of C. The active sites are reducible Co4+O species. On the
other hand, BaCO3 is assumed to form defect centers and
trapped electrons, which activate O2 for the ODH reaction. Cooperation between BaCO3 and BaCoO3, as both a redox-active
metal and a source of active oxygen species, is assumed to be
one reason for the good activity and selectivity of this catalyst,
which is a good example of cooperation between sites, as discussed in Section 3.3.2.
4.2. Rare-earth-metal oxides
These catalysts typically consist of La2O3, Sm2O3, CeO2, and
Pr6O11.[30, 133136] Promotion with Na[30] and CaO[137, 138] has also
been investigated. The cleavage of the first CH bond occurs
through a single-electron process and the catalysts belong to
www.chemcatchem.org
the group of non-redox-active materials. The main catalytic
function is the production of ethyl radicals, which are released
into the gas phase, with ethene and CO as the main carboncontaining products.[30] The catalysts were operated between
600 8C and 900 8C and the selectivity was around 50 %; both
the conversion and selectivity for the olefins increased with
temperature. This result is thought to be related to the release
of ethyl radicals, which increasingly takes place at elevated
temperatures. Enhancing the release rates implies that fewer
ethoxy species (intermediates of CO2 production) exist on the
surface at the steady state.[30]
The addition of alkali-earth/alkali metals to rare-earth oxides
improves the catalytic performance.[139] A Sm2O3 catalyst that is
doped with 10 % Na+ cations shows a notably higher selectivity than a pure Sm2O3 catalyst. For both catalysts, olefin selectivity decreases with increasing oxygen concentration; however, this effect is not as pronounced for Na+-doped catalysts
than for pure Sm2O3. Whereas ethene and CO are the main
products with rare-earth oxides, CO2 is observed instead of CO
in the presence of Na+ cations. This result suggests that the
formation of lattice oxygen species plays a role in the activation of the alkane and it points to the operation of a heterogeneous surface-catalyzed mechanism.
This trend is even more clear in the case of La2O3.[30] Adding
+
Sr ions to rare-earth oxides leads to a marked increase in activity and selectivity. At 700 8C, Sr2O3 on neodymium oxide is
the optimal catalyst in terms of activity and selectivity, whereas
Sr+-doped CeO2 and praseodymium oxides show the poorest
performance.[139]
Rare-earth oxides are either employed as bulk materials or
on a MgO support, being per se also a potential catalyst for
the ODH of ethane. By comparing the performance of Sm2O3covered MgO with that of pure MgO, it is interesting that
Sm2O3 does not increase the catalytic activity, although it increases the specific surface area. However, covering MgO in
the same way with La2O3 strongly enhances the catalytic activity. This result cannot be explained by only surface stabilization
at higher specific surface areas; thus, the intrinsic activity of
the catalyst seems to be drastically improved by La3+
species.[30]
CeO2 is an exception because it has two oxidation states in
the catalytic cycle. Ceria-based catalysts are unique because
CO2 can be used as an oxidant.[137] However, higher temperatures are required for the activation of CO2 compared to
oxygen.[140] In this case, a combination of a homogeneous and
heterogeneous mechanisms is assumed. For the heterogeneous catalyst part, Ce4+ species are reduced by ethane into Ce3+
species, thus supplying oxygen, and are re-oxidized through
the reduction of CO2. CeO2 plays a role not only because of its
redox properties, but also because of its oxygen-storage capacity. To further improve the catalytic performance, the addition of Ca2+ ions to Ce4+ ions increases the oxygen-ion mobility
of CeO2, as well as contributes to a higher surface basicity,
thus leading to faster olefin desorption (higher olefin selectivity).[137] Interestingly, olefin desorption is accelerated in the
presence of basic sites on this class of catalysts, in contrast to
redox-active catalysts, in which acid sites that are in spatial
&14&
CHEMCATCHEM
REVIEWS
proximity to redox pairs favor olefin desorption, as mentioned
before.
The improvement of the catalytic performance of rare-earth
oxides can also be achieved by the addition of halides.[141144]
Au and co-workers reported that the addition of 50 mol %
BaCl2 to Ho2O3[141] lowered the reactivity of the oxygen species
and, thus, favored the activation of the weaker CH bond in
ethane (410 kJ mol1) instead of the CH bond in ethene
(452 kJ mol1). Raman spectroscopy showed that the intensity
of the bands of O2 species was higher for the BaCl2-containing
catalyst compared to pure Ho2O3. Thus, the addition of BaCl2
resulted in a higher storage capacity and activation activity of
O2 species. Again, a higher concentration of basic sites, which
was induced by the addition of BaCl2, decreased the adsorption of ethene and, thus, the total oxidation of ethane. It can
be inferred that intrinsically weaker adsorption and site-isolation are the key concepts for avoiding the re-adsorption of
ethane onto these catalysts, thus, maintaining high selectivities. Cation substitution of Ho3+ and Ba2+ is assumed to create
oxygen defects that are active for O2 activation. A similar phenomenon is observed on Sm2O3/LaF3 catalysts by promotion
with BaF3.[142]
4.3. Supported alkali oxides

The simplest and most prominent catalytic system in this class
is MgO-supported Li2O.[145] In general, such oxides belong to
the group of non-redox-active catalysts. However, it is not yet
known whether the first CH bond is activated through
a single-electron process or a paired-electron process.[1]
The active site has been concluded to be [Li+O], thus forming defects on the MgO surface. The fact that [Li+O] sites possess labile O species is of special importance, because the ODH
activity notably decreases upon catalyst deoxygenation. The
O species is assumed to abstract the first hydrogen atom. Evidence was found that Li+ ions formed clusters on the MgO surface, but it was concluded that the Li2O domains on MgO did
not show classical redox properties.[34, 146] It should be emphasized that Li+ ions that are incorporated into a MgO matrix
are also active and selective and can be prepared by a solgel
synthesis.[147] The reaction is assumed to proceed through radical chain reactions, by following a homogeneous/heterogeneous reaction pathway. The catalyst essentially serves as an element to initiate the radical chain mechanism.[34]
In contrast to these conclusions and the assumption that
[Li+O] is the active site, Freund and co-workers concluded
that the paramagnetic [Li+O] species did not exist under the
reaction conditions, because the species were not detected by
EPR spectroscopy.[35, 148, 149] Li+ ions have been found to introduce oxygen vacancies into the MgO matrix. At higher Li+ concentrations, Li/Mg-oxide islands and LixO surface clusters form.
However, Li can desorb into the gas phase at elevated temperatures (> 780 8C), thus resulting in the formation of surface defects. Therefore, despite suggesting that [Li+O] sites may form
under non-equilibrium conditions, Myrach et al. raised doubts
that these are active for the ODH of ethane.[35]
www.chemcatchem.org
The addition of halide anions (Cl , Br) further enhances activity and selectivity.[31, 150, 151] The active sites in the halide-promoted materials have not been unequivocally identified. Wang
et al. suggest that Li2O is the catalytically active phase,[152] with
LiCl forming a thin layer between the support and the catalytically active phase, which modifies the base strength and suppresses the formation of Li2CO3. Fuchs et al. concluded that
the active site consisted of a Li+ cation and nucleophilic
oxygen species.[21] The addition of CO2 to these catalysts markedly decreases the ODH activity, thus suggesting that carbonate formation indeed blocks the active sites. Gaab et al. provided a detailed differentiation between Cl-free and Cl-containing catalysts.[32] They postulated that hypochlorite (OCl)
anions formed in the LiCl melt, thus providing a transient,
highly reactive site with dissolved O2.
Alternatively, a combined surface/gas-phase mechanism has
been suggested, similar to the one for Cl-free catalysts.[6]
However, this mechanism cannot explain the increasing selectivity towards ethene with temperature. A third approach invokes a surface-catalyzed pathway,[32] thus suggesting that the
non-selective sites are covered or dissolved in the molten overlayer.
For Cl-free catalysts, surface ethyl radical formation and
a gas phase radical reaction pathway dominate. The initial activation of ethane to form an alkyl radical in the gas phase is
considered to be rate determining. However, at partial pressures of above 70 mbar, the reaction order decreases, which is
attributed to saturation of the active site.
4.4. Supported alkali chlorides

Whereas supported alkali-oxide catalysts show certain disadvantages, including inhibition by CO2 and limited selectivity,
materials with an overlayer of molten alkali chloride and no addition of alkali oxide have been reported to be catalytically
active for the ODH of ethane.[51, 67] Supported liquid-phase catalysts (SLC) are a new generation of catalysts for the ODH of
ethane, which requires comparatively high reaction temperatures (around 600 8C). However, very high selectivities for
ethane (up to 95 %) can be achieved. The first CH bond is
cleaved homolytically in a single electron-transfer process.
Activity and selectivity have been reported for catalysts with
overlayers of pure alkali chlorides and several eutectica[153] of
alkali chlorides, such as LiCl, KCl, NaCl, LiKCl, LiNaCl, Li
SrCl, and LiBaCl, that are supported on (Dy2O3-promoted)
MgO.[51, 67] Whereas LiCl leads to the most active catalysts,
these catalysts are also the least selective. The selectivity towards ethene inversely depends on the melting point of the
chloride. A linear increase in the selectivity with temperature is
observed below the melting point, whereas it remains essentially constant above the melting point. Metal cations are presumably inactive towards CH activation, but they influence
the catalytic behavior of the molten chloride overlayer. The
oxygen solubility is the highest for LiCl,[51] whereas larger cations (Na+, K+) lead to fewer dissolved oxygen species.[154, 155]
Ethane solubilities in the highly polar melt are negligible.[67]
&15&
CHEMCATCHEM
REVIEWS
www.chemcatchem.org
It has been speculated that dissolved O2 in the molten chloride reacts to form positively charged Cl atoms.[1] In turn, a transient hypochlorite species (OCl) is assumed to catalyze the
ODH reaction. The hypochlorite anion itself has a high redox
potential and is supposed to form at the interface between
the chloride melt and the solid support [Eq. (14)].
2 LiClmoltenO2 dissolved ! 2 LiOCldissolved
14
Hypochlorite is assumed to be the transient species in low

concentrations and diffuses within the melt to the surface, on
which it activates the CH bond. Because those anions are surrounded by alkali cations, spatial separation occurs within the
melt, which causes the active sites to be well-dispersed and
prevents mutual interactions.
However, the exact mechanism is still matter of debate. Cl
radicals are not thought to play a role in alkane activation, neither as initiator nor as radical-chain propagator, and the formation of HCl bonds are unlikely. A key aspect of the adsorption
properties of the melt is the Lewis acidity of the cations. The
addition of, for example, K+ ions, lowers the average Lewis
acidity compared to that of pure Li. Thus, the improved olefin
selectivity in the presence of alkali/alkali-earth chlorides of the
LiClMgDyO catalyst is speculated to be associated with
a decrease in the strength of the Lewis acid.
4.5. Other catalysts
One of the most promising groups of catalysts is based on Ni,
Cu, and Fe (cation)-loaded Y-zeolites.[156] In this case, the first
CH bond is also assumed to be cleaved homolytically
through a single electron transfer.
Acidic catalysts were prepared by ion exchange with metal
sulfates and subsequent reduction with hydrogen. For basic
catalysts, an additional ion exchange with KOH was performed.
Acidic zeolites were significantly better than basic zeolites in
terms of conversion and ethane selectivity. Ni-based materials
showed the best performance, followed by Cu and Fe; the
best catalyst (Ni on acidic Y-zeolite) showed an ethane conversion of 21 % and ethene selectivity of 75 % at 600 8C. Ethene
selectivity increased with increasing temperature. With acidic
zeolites, ethane oxidation (leading to ethene) dominated over
ethane combustion, whereas basic zeolites favored ethane
combustion.
For the ODH reactions of transition-metal-supported zeolites,

two mechanisms are proposed.[156] The first proposal suggests
that two surface oxygen atoms react with ethane, thus forming
one surface metalethoxide group and one surface hydroxy
group. The metal ethoxide can then undergo two subsequent
reactions, that is, a-hydrogen abstraction, thereby forming an
aldehyde plus a metal hydride, or b-hydrogen abstraction,
thereby leading to ethene plus a metalhydroxy group.
The second possibility is the formation of a metalethyl
complex plus a surface hydroxy group after one ethane molecule reacts with one surface oxide group. Subsequently, b-hydrogen abstraction takes place, thus forming ethene plus
a metal hydride. Various techniques have been applied to address the dominating mechanism,[156, 157] although it could not
be clarified whether the reaction took place through an
ethoxyhydroxy mechanism or an ethyl radical. However, it
was shown that ethoxy groups were not stable at elevated
temperatures.[157]
Carbon materials, in particular carbon nanotubes, have also
been reported to be active. However, these catalysts appear to
be challenging to use and ethane conversions of below 5 %
have been reported at 400 8C; the use of higher reaction temperatures is seen as problematic, because of the potential oxidation of the catalyst. This problem can be mediated by surface modification with B2O3 or P2O5, which have the positive
side effect of decreasing the concentration of electrophilic
oxygen species, thereby increasing the selectivity. The nature
of the active sites is not defined and their concentration is too
low to be successfully characterized.[158]
4.6. Conclusions
Table 1 provides a synopsis of a selection of active sites that
have been presented in this Section (only the basic types of
sites are shown). The characteristics of the sites and the specific mechanisms are compiled, together with their typical reaction temperatures and selectivities. A complete list of the referenced catalysts, including their performance, is presented in
the final Section.
The redox-active site is either a cation redox pair (such as
V5+/V4+) or an anion redox pair (such as OCl/Cl). High dilution of the sites and higher acid strength appear to promote
desorption of the ethyl species and favor selectivity. The sites
with metal redox functionality require lower temperatures
Table 1. Different active sites for the ODH of ethane.

Site
Type of site
Activation of first CH bond
Mechanism
Operation range [8C]
Selectivity
Reference
VO
redox
1090
[32, 110]
redox
redox
non-redox
non-redox
non-redox
redox
transient radicals on catalyst surface

or redox cycle
redox cycle
[a]
[a]
radical chain
radical chain
430650
MoO
[NiNbO]
CoO
Sm2O3
[LiMgO]
[OCl]
single-electron process, homolytic

or paired-electron process, heterolytic
[a]
[a]
550580
350400
550
550700
570650
525600
4090
8090
4580
2070
5775
8095
[98, 100]
[16, 87]
[130]
[30]
[100, 111]
[51, 67]
[a] Not reported or still unclear.
&16&
CHEMCATCHEM
REVIEWS
than anion redox pairs and yield high selectivities. In particular,
V- or Ni/Nb-based oxides are among those with the lowest required reaction temperatures. High reaction temperatures also
face the danger of inducing a radical chain mechanism, which
tends to lead to lower selectivities. Thus, temperature is one of
the most critical parameters in determining whether a mechanism is purely heterogeneous (i.e., redox reactions on catalyst
surfaces) or homogeneous (surface-induced gas-phase radical
chain reactions).[7]
5. Advanced Reactor Concepts for the ODH of

Ethane
Clearly, activity and selectivity in the ODH reaction not only
depend on the properties of the catalyst, but are also critically
influenced by the reactor. Because weakly bound oxygen species on the catalyst surfaces are thought to lead to the accelerated formation of COx species, higher O2 partial pressures lead
to lower selectivity, especially with catalysts that contain metal
redox pairs.[9] Thus, various concepts have been investigated to
limit the concentration of this species (such as membranes,
cyclic reactors, and variations in the oxygen/hydrocarbon
feed). In this Section, we focus on two important concepts,
that is, membrane reactors and reactors with a millisecond residence time.
5.1. Membrane reactors

Two main different types of membrane reactors are known
and practically implemented, that is, extractor and distributor
reactors.[27] After shortly introducing both concepts, we want
to rationalize and discuss which type of reactor could be
promising for ODH reactions.
The extractor membrane reactor separates products within
the reactor to prevent consecutive reactions. For equilibriumlimited reactions, it enables higher conversions by selectively
extracting the limiting reaction product, such as H2 from dehydrogenation. However, the H2-extractor membrane does not
provide a solution for coking, which is a major cause of catalyst deactivation in the dehydrogenation of ethane. The removal of ethene, which leads to enhanced selectivity by avoiding consecutive reactions, is not feasible for the ODH reaction,
because membranes with a sufficiently high separation factor
between ethane and ethene are not known, thereby limiting
the potential of the membrane extractor concept.
The distributor membrane reactor disperses a reactant
across the catalyst bed, thereby avoiding high local concentrations, which is of special importance for decreasing the explosion limits with hydrocarbon/oxygen mixtures so that undiluted feed streams can be used.[7] Furthermore, it has the potential to enhance product selectivity, if the distributed reactant
has a higher reaction order in the undesired reaction path
than the desired one, as is the case for O2 in the ODH reaction.[7] Typically, interfacial membrane reactors are used, which
are equipped with catalytic membranes that enable the reaction on the membrane surface.
www.chemcatchem.org
Rodriguez et al. simulated the performance of the ODH reaction of ethane with a NiNbO catalyst that was packed into
a tubular porous inorganic membrane reactor by dosing
oxygen into the shell side across the membrane and ethane
into the center compartment[159] and they compared the results to a model conventional multitubular reactor.[160] Whereas
the conversion of ethane in a tubular porous inorganic membrane reactor remained about the same compared to that in
a fixed-bed reactor, the selectivity for the olefin was higher.
These, higher selectivities were explained by the lower local O2
partial pressures and the smaller temperature gradients that
were achieved with the axial distribution of O2. The operating
conditions have to be finely tuned, in particular by the addition of O2 to the ethane feed, to enhance the reaction rates.
O2-permeation rates, which are higher than O2-consumption
rates, lead to an unfavorable concentration profile with oxygen
accumulation along the catalyst bed, thus inducing lower selectivity. On the other hand, a feed that contains a majority of
oxygen in a mixture with ethane would jeopardize the enhancement by the distributor membrane because it would
leads back to a quasi-conventional reactor with strongly exothermic total oxidation, lower olefin selectivity, and the appearance of hot spots. Concerning the membrane type, a porous
membrane with a weakly temperature-dependent trans-membrane oxygen flux is apparently more amenable than dense
membrane materials with their temperature-related oxygen
permeation.
Iglesia et al. studied the staged feed of oxygen for the ODH
of ethane with a well-known V2O5 catalyst. Similar activities
have been observed compared to a regular fixed-bed reactor,
despite notable differences in the local oxygen concentration.
The kinetics are identical in staged and constant feeds. However, the total oxidation of the produced ethene was lower with
a staged oxygen feed. Reaction studies revealed that the homogeneous part of ethene oxidation was dependent on the
oxygen concentration, whereas the heterogeneous part was
zero-order in oxygen.[161] As the catalyst-to-volume ratio is high
in industrial-scale reactors, homogeneous contributions to the
whole reaction are almost negligible. Thus, the benefits of
a staged oxygen feed become less important in large-scale applications by using a V2O5 catalyst.
Another study compared the performance of a similar catalyst (vanadium oxide supported on gamma-alumina) for the
ODH of ethane in fixed-bed or membrane reactor configurations at high space velocities under excess oxygen.[162] The conversion of ethane was higher in the packed membrane reactor,
although the olefin selectivity decreased. This phenomenon is
particularly important in excess oxygen, because the total oxidation of ethene is higher in the membrane reactor than in
the packed-bed reactor and is related to the local O2 concentrations. In the first part of a membrane reactor, ODH is the
dominant reaction. However, because this part has the longest
contact time, the total oxidation of ethane also occurs. In the
second part, excess oxygen leads to the total oxidation of
ethene, mainly into CO. In the last part, oxygen, which is still
present in excess, leads to the oxidation of ethene and CO.
Given these facts, a catalyst does not always perform equally
&17&
CHEMCATCHEM
REVIEWS
well in fixed-bed and membrane reactors. The kinetics and the
contact time are key factors in influencing the catalyst performance for each reactor type.
To use another well-known ODH catalyst in an advanced reactor concept, MgO that was doped with Li and Sm2O3 was
used as a packed catalyst in a reactor with oxygen dosing
through a porous Al2O3 membrane.[163] The Sm2O3-doped catalyst showed better performance in the membrane reactor. Limited to low ethane/O2 ratios in the feed, splitting the O2 feed
gave marked improvements compared to a standard plug-flow
reactor (PFR). However, this effect disappeared for higher
ethane/O2 ratios,[163] because it is sensitive to contact times. As
the O2 feed rates increased at low feed ratios, the contact
times of oxygen and ethane differed, thus, causing differences
in the behavior of the reactor. However, at higher ethane/O2
feed ratios, the contact times of both reactants became similar
to that of a standard PFR.
This example illustrates that staged O2 admission is beneficial for the ODH of ethane, with yields (under certain conditions) that are about three-times higher than in a fixed-bed reactor. However, the membrane reactor did not perform well at
short contact times. Because long contact times and a low
ethane/O2 ratio are neither productive nor likely, owing to
flammability issues, this improvement is not realistic for largescale applications, owing to a very limited parameter window.
Some membranes catalyze the ODH reaction, even without
additional catalysts. However, coating those membranes with
a catalytically active material enhances the performance of the
membrane reactor.[28] The reactions are run at around 1000 K,
a temperature at which surface-modified dense membranes
allow a constant supply of the catalyst with ionic oxygen (extracted from air) fed into the other side of the membrane. Perovskite membrane (BaSrCoFeO) showed outstandingly
high oxygen permeation; however, the trans-membrane flow
of oxygen decreased with time at lower temperatures (below
about 1000 K).[164]
The membrane surface was modified with Pd nanoclusters
or a V/MgO catalyst on the permeate side-surface. A first effect
of the modification relates to changes in the oxygen permeability, as shown for V-based modification, which doubles the
oxygen flow across the membrane. Ethane conversions increase linearly with temperature up to about 60 % at around
800 8C for a reactor with a bare membrane, whereas the selectivity decreased slightly with increasing temperature. The activation energy was higher for the coated membranes, but the
selectivities decreased even faster. Nonetheless, a highest
ethene yield (about 75 %) was achieved with a catalyst-modified membrane that showed stable performance over at least
23 days. Essential for the operation of those materials was
a reasonably high oxygen flux across the membrane, which
could be controlled with temperature. The membrane transports ionic oxygen, which is converted into the activated species (O2, O , O2) on the catalyst surface without the presence
of O2. Hence, this concept allows the fine-tuning of catalytic
systems, following a MarsVan Krevelen mechanism.[27]
A staged reactor concept has been introduced that combined dehydrogenation with the oxidation of the formed
www.chemcatchem.org
H2.[117] A perovskite-based hollow-fiber membrane, which was
made of mixed oxygen-ion electron-conducting dense BaCo
FeZrO, was used to separate O2 from air. Parts of the membrane were passivated by Au, which created different reaction
zones, that is, the passivated parts were dehydration zones, in
which H2 and ethene were produced, whereas oxygen was
transported into non-passivated parts. Conversions could reach
up to 52 % with selectivities of up to 72 %. The membrane is
claimed to be coke-resistant and to allow long-term operation.
The performance of the reactor concept is supposed to improve upon coating one site of the membrane with a catalyst.
However, the reaction is not a classical ODH reaction, because
of the intermediate formation of H2.
5.2. Short-contact-time partial oxidation of ethane
Schmidt, Huff, and Holmen have utilized monoliths for the
ODH of ethane at very short contact times.[165, 168] By using
a very short contact time of the reactant/product mixtures
(about 104 s) offers various advantages, including a small reactor volume and an autothermal operation mode. Typical catalysts are supported Pt and Rh. The extremely short contact
times help to minimize ethene re-adsorption to enhance olefin
selectivity. The activation of the first CH bond is homolytic
and is thermally induced by a single-electron transfer.
The performance of short-contact-time reactors can be significantly improved by added H2 to the feed stream,[11] thus increasing ethene selectivity up to 85 % with Pt and PtSn catalysts. For very low contact times (around 0.5 ms), the flow rate
did not influence the selectivity. Thus, a homogeneous gasphase mechanism can be ruled out. It is assumed that H2 is
oxidized, which generates heat and consumes O2, thereby lowering the concentration of O2 that is available for total oxidation, because H2 is formed by the dehydrogenation of ethane.
Further studies have shown that the metal coating on the
monolith directs the selectivity. With Pt and PtSn catalysts,
ethene is mainly formed, along with CO, H2, and water as byproducts. However, Rh mainly produces synthesis gas and Pd
leads to severe carbon deposition.[165]
Three different mechanistic scenarios are discussed, that is,
a completely homogeneous gas-phase mechanism, a completely heterogeneous surface mechanism, and a combination of
the two, with H2 oxidation taking place on the surface and
ethane dehydrogenation taking place through radicals in the
gas phase. The results cannot be unequivocally described by
each of these mechanisms.
One prominent way to model the reaction network on Ptcoated monoliths includes the assumption that the catalyst
burns a fraction of ethane, thereby generating heat. This heat
is used to form radicals that start a homogeneous gas-phase
mechanism.[166] The heterogeneous mechanism suggests that
ethane is adsorbed, followed by dissociation into ethene and,
eventually, C and H on the catalyst surface, and the subsequent oxidation of both species. However, this model cannot
fully explain the catalytic chemistry if hydrogen is co-fed into
the reactor. The combined surface gas-phase mechanism[167]
suggests that ethene is formed in the gas phase (in the case of
&18&
CHEMCATCHEM
REVIEWS
Pt and PtSn monoliths), whereas CO, H2, and water
are mainly formed on the surface.
Holmen and co-workers studied the extinction and
ignition behavior of short-contact-time processes.[168]
During ignition and extinction, CO and CO2 were observed to be the main reaction products, together
with a noticeable temperature change during these
processes. These results also suggest that the ODH
reaction takes place in the gas phase, whereas the
by-products, that is, H2, CH4, and COx, are formed on
the Pt surface. Monoliths of rare-earth metals and alkaline-earth-metal oxides are thought to follow the
same reaction pathway.[169, 170]
The microkinetic aspects of short-contact-time
ethane activation over these materials were investigated in greater detail by performing transient kinetic measurements with a focus on the nature of the
active oxide species.[171] It has been found that adsorbed monoatomic and diatomic oxygen species are
active during the ignition of the alkane/oxygen mixture. Whereas the adsorbed monoatomic oxygen
species was active for the ODH reaction, diatomic
oxygen was responsible for heterogeneously catalyzed total oxidation.
Pt/Al2O3 showed a similar behavior, that is, the catalyzed combustion accelerated the ignition of
ethane/oxygen mixtures under autothermal conditions, thereby also showing a homogeneous/heterogeneous mechanism.[172]
This result also holds true for other catalytic materials. Over oxides of non-reducible metals, heterogeneous reaction steps form adsorbed oxide species
that are reactive for the ODH of ethane. Owing to
the exothermic character of this reaction, the temperature of the catalytic bed rises, thus leading to the
thermal dehydrogenation of ethane, which improves
the overall yield of ethene.
An important improvement for short-reactor-time
concepts is heat integration.[173] Reverse-flow operation of a fixed-bed reactor with a coated monolith is
possible (about 15 s per cycle). Whereas selectivities
remain almost unchanged, conversion levels can be
dramatically improved (up to about 90 %).
6. Summary and Outlook on Catalytic

ODH Reactions
www.chemcatchem.org
Table 2. Performance data for selected catalysts.
No.
Catalyst
T [8C]
X(C2H6) [%]
S(C2H4) [%]
Reference
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Ni12Ce
Pt monolith
MoVNb
NiW0.36
NiW0.45
MoV0.39Te0.16Nb0.17O
LiMgCl/Dy2O3
LiZnOCl
V2O5/g-Al (5.2 wt. %)
Co7.6/TiO2
Ni/HY
BaF2/SL1 (Sm2O3LaF3)
BaCl2/Ho2O3
CoBaCO3 (7 wt. %)
SrLaNdO
Li/MgO (3 %)
Mg/Dy/Li/O/Cl
LiMgOCl
LiCl/ZrON
NdLi/SZ
LiMgDyOCl
LiNaMgDyOCl
SrNd2O3
Sm2O3
10CaCe
Vox/Al (20 wt. %)
Ni0.85Nb0.15
MoVSbO
VOP/Ti9
10Val
M10V5
VCo-2
ClMoSiTi
CoVAPO-5
MgVAPO-5
V/Al
SrCl2/Sm2O3
SmOF
275
450
400
400
400
380
570
721
430
550
600
700
640
650
700
625
600
675
650
650
600
650
800
700
750
550
400
400
550
550
580
600
600
600
600
600
640
700
10.4
85
9
54
21
39.8
81.3
82
50
22.2
22
62.9
56.6
48
65.2
53.9
60
11
94.8
93
82
82.4
58
25
21
30
65
40
15
37
33.8
27.8
31
43.9
28.7
60.2
80.3
80.2
59
46
75
38
50
93.9
76.2
78
40
60
74.5
67.7
67.9
92.2
71.2
63.8
83.3
78
71.3
83
77
91
79
60
100
57
70
95
70
57
70.7
74.3
36.2
40.8
59.7
40.2
70.9
91.8
[122]
[168]
[114]
[122]
[122]
[110]
[174]
[146]
[38]
[130]
[156]
[142]
[141]
[132]
[175]
[145]
[176]
[152]
[177]
[177]
[67]
[51]
[139]
[30]
[137]
[44]
[19]
[64]
[97]
[34]
[98]
[86]
[107]
[69b]
[88]
[85]
[143]
[178]
Table 3. Performance data for selected reactor concepts.

No. Catalyst, reactor configuration
39
40
41
42
43
44
45
46
47
48
49
T [8C]
X(C2H6) [%] S(C2H4) [%] Reference
Pt/alumina, reverse-flow reactor

n.r.
90
n.r.
73
Pt/Sn, H2 co-feed
Rh/Pt (10 %) gauze
[a] 34
Pt foam monolith
875
62
600
70
PBR, VOx/g-alumina
n.r.
35
V2O5/Al2O3 (2 wt. %) staged feed
725
100
BaCoxFeyZr1xyO
64
BaCoxFeyZr1xyO, BCFZ membrane 725
V/MgO membrane
777
90
LaSr/CaO
1000
32
[b] 78
Na0.009CaOx
60
83
62
55
30
45
50
67
83
55
72
[173]
[11]
[179]
[165]
[162]
[161]
[180]
[181]
[28]
[182]
[171]
Table 2 and Table 3 show performance data for the

catalysts and the different reactor concepts that have
[a] Surface: 900 8C, gas: 580 8C; [b] ignition: 600 8C, maximum: 927 8C.
been introduced in this Review, respectively; the results are summarized in Figure 8. These data reprecating that a multitude of possible solutions to this problem
sent the best catalytic performance in each of the
exist.
cited references, but they do not consider the space/
High selectivities are limited by the high olefin reactivity and
time yield of the catalytic systems. Interestingly, the concepts
the favorable thermodynamics of paraffin and olefin oxidation.
with the best performance (> 70 % yield of ethene) are not reAs a consequence, all of the catalytic materials show some acstricted to one class of catalyst or reactor concept, thus indi 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&19&
CHEMCATCHEM
REVIEWS
www.chemcatchem.org
als properties, the strategies to synthesize these materials, the
detailed chemistry of the individual reaction steps, and the
fluid dynamics of advanced reactor concepts make it appear
realistic that the ODH of ethane will be competitive to steam
reforming in the near future.
Acknowledgements
C.G. acknowledges support from the TUM Graduate School and
the Faculty Graduate Center Chemistry (FGCH), as well as valuable discussions with Prof. Angeliki Lemonidou.
Figure 8. Performance of catalytic and reactor systems.
tivity for the complete oxidation of paraffin and olefins. Rapid

ethene desorption by adjusting the acid/base properties and
limiting ethene re-adsorption by minimization of the concentrations of Lewis acid sites, as well as coupling surface-catalyzed activation of ethane with gas-phase radical-chain reactions, have been found to be the most promising concepts for
catalyst and process design. Whereas membrane-based processes only show a moderate chance of success, ultrashort reactor concepts appear to be closer to their practical realization.
Suitable catalysts for the ODH of ethane can be divided into
two groups: In the first group, as-ideal-as-possible oxide surfaces minimize the concentration of accessible Lewis acid metal
sites, which are responsible for ethene re-adsorption. In the
second group, molten active components generate dynamically rearranging surfaces, which prevent the re-adsorption of the
olefin by minimizing the concentration of coordinatively unsaturated metal centers.
For conventional reactor systems, catalysts that are based
on complex vanadia mixed oxides and supported molten
chlorides are the best choice. PtSn catalysts have been identified as the best catalysts for reactors with an ultrashort contact
time. Dramatic differences in the materials emphasize how different energies of activation for the multitude of steps in the
complex reaction sequence lead to very different results for
the various optimal catalysts. The complexity of this reaction
also makes it almost impossible to generalize the nature of the
selective and unselective sites. The critical catalytic preferences
will be influenced by the rates of processes in a complex
matrix, which are influenced (by default) by the different energies of activation, pre-exponential factors, and the concentrations of the catalyzing sites. Thus, intermediate oxygen species
could be active and selective for the ODH reaction on one catalyst, but unselective on others.
The examples shown above illustrate that a detailed understanding of the reaction mechanism, the nature of the active
sites, and the reaction kinetics is essential for the design and
optimization of the synthesis and catalytic properties of new
catalysts. The improvement of existing reactor concepts and
the realization of new concepts have to occur in a coordinated
fashion. The unprecedented increase in insight into the materi 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Keywords: CH activation dehydrogenation heterogeneous

catalysis membranes reaction mechanisms
[1] J. A. Lercher, F. N. Naraschewski in Nanostructured CatalystsSelective
Oxidations (Eds.: C. Hess, R. Schlgl), Royal Society of Chemistry, Cambridge, 2011, pp. 5.
[2] K. Chen, A. T. Bell, E. Iglesia, J. Catal. 2002, 209, 35.
[3] E. M. Thorsteinson, T. P. Wilson, F. G. Young, P. H. Kasai, J. Catal. 1978,
52, 116.
[4] H. Zimmermann, R. Walzl, in Ullmanns Encyclopedia of Industrial
Chemistry, Wiley-VCH, Weinheim, 2000.
[5] M. A. Baares, Catal. Today 1999, 51, 319.
[6] F. Cavani, F. Trifiro, Catal. Today 1995, 24, 307.
[7] F. Cavani, F. Trifiro, Catal. Today 1999, 51, 561.
[8] E. A. Mamedov, V. Cortes Corberan, Appl. Catal. A 1995, 127, 1.
[9] F. Cavani, N. Ballarini, A. Cericola, Catal. Today 2007, 127, 113.
[10] H. H. Kung, D. D. Eley, H. Pines, W. O. Haag in Advances in Catalysis,
Vol. 40, Academic Press, 1994, pp. 1.
[11] A. S. Bodke, D. A. Olschki, L. D. Schmidt, E. Ranzi, Science 1999, 285,
712.
[12] M. M. Bhasin, J. H. McCain, B. V. Vora, T. Imai, P. R. Pujado, Appl. Catal. A
2001, 221, 397.
[13] F. M. Ashmawy, J. Appl. Chem. Biochem. 1977, 27, 137.
[14] J. D. Arndt, S. Freyer, R. Geier, O. Machhammer, J. Schwartze, M. Volland, R. Diercks, Chem. Ing. Tech. 2007, 79, 521.
[15] M. Stcker, Angew. Chem. 2008, 120, 9340; Angew. Chem. Int. Ed. 2008,
47, 9200.
[16] G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044.
[17] S. Golay, R. Doepper, A. Renken, Appl. Catal. A 1998, 172, 97.
[18] I. Takahara, M. Saito, M. Inaba, K. Murata, Catal. Lett. 2005, 105, 249.
[19] E. Heracleous, A. A. Lemonidou, J. Catal. 2006, 237, 162.
[20] T. Xie, K. B. McAuley, J. C. C. Hsu, D. W. Bacon, Ind. Eng. Chem. Res.
1994, 33, 449.
[21] S. Fuchs, L. Leveles, K. Seshan, L. Lefferts, A. Lemonidou, J. Lercher,
Top. Catal. 2001, 15, 169.
[22] K. Chen, S. Xie, A. T. Bell, E. Iglesia, J. Catal. 2001, 198, 232.
[23] P. Viparelli, P. Ciambelli, L. Lisi, G. Ruoppolo, G. Russo, J. C. Volta, Appl.
Catal. A 1999, 184, 291.
[24] C. Batiot, B. K. Hodnett, Appl. Catal. A 1996, 137, 179.
[25] S. Albonetti, F. Cavani, F. Trifiro, Catal. Rev. Sci. Eng. 1996, 38, 413.
[26] R. Grabowski, Catal. Rev. Sci. Eng. 2006, 48, 199.
[27] J.-A. Dalmon, A. Cruz-Lopez, D. Farrusseng, N. Guilhaume, E. Iojoiu, J.C. Jalibert, S. Miachon, C. Mirodatos, A. Pantazidis, M. Rebeilleau-Dassonneville, Y. Schuurman, A. C. van Veen, Appl. Catal. A 2007, 325, 198.
[28] M. Rebeilleau-Dassonneville, S. Rosini, A. C. van Veen, D. Farrusseng, C.
Mirodatos, Catal. Today 2005, 104, 131.
[29] F. Khorasheh, M. R. Gray, Ind. Eng. Chem. Res. 1993, 32, 1853.
[30] P. Ciambelli, L. Lisi, R. Pirone, G. Ruoppolo, G. Russo, Catal. Today 2000,
61, 317.
[31] S. Gaab, J. Find, R. K. Grasselli, J. A. Lercher in Studies in Surface Science
and Catalysis, Vol. 147 (Eds.: X. Bao, Y. Xu), Elsevier, 2004, pp. 673.
[32] S. Gaab, J. Find, T. E. Mueller, J. A. Lercher, Top. Catal. 2007, 46, 101.
[33] L. Leveles, K. Seshan, J. A. Lercher, L. Lefferts, J. Catal. 2003, 218, 296.
[34] L. Leveles, K. Seshan, J. A. Lercher, L. Lefferts, J. Catal. 2003, 218, 307.
&20&
CHEMCATCHEM
REVIEWS
[35] P. Myrach, N. Nilius, S. V. Levchenko, A. Gonchar, T. Risse, K.-P. Dinse,
L. A. Boatner, W. Frandsen, R. Horn, H.-J. Freund, R. Schlgl, M. Scheffler, ChemCatChem 2010, 2, 854.
[36] A. Beretta, P. Forzatti, E. Ranzi, J. Catal. 1999, 184, 469.
[37] M. Y. Sinev, V. Y. Bychkov, V. N. Korchak, O. V. Krylov, Catal. Today 1990,
6, 543.
[38] J. Le Bars, A. Auroux, M. Forissier, J. C. Vedrine, J. Catal. 1996, 162, 250.
[39] X. Rozanska, J. Sauer, Int. J. Quantum Chem. 2008, 108, 2223.
[40] Z.-S. Chao, E. Ruckenstein, J. Catal. 2004, 222, 17.
[41] P. Mars, D. W. van Krevelen, Chem. Eng. Sci. 1954, 3, 41.
[42] M. V. Martinez-Huerta, X. Gao, H. Tian, I. E. Wachs, J. L. G. Fierro, M. A.
Banares, Catal. Today 2006, 118, 279.
[43] J. Le Bars, J. C. Vdrine, A. Auroux, S. Trautmann, M. Baerns, Appl.
Catal. A 1994, 119, 341.
[44] T. Blasco, J. M. Lpez Nieto, Appl. Catal. A 1997, 157, 117.
[45] J. M. Lpez Nieto, A. Dejoz, M. I. Vazquez, W. OLeary, J. Cunningham,
Catal. Today 1998, 40, 215.
[46] M. A. Baares, M. V. Martnez-Huerta, X. Gao, J. L. G. Fierro, I. E. Wachs,
Catal. Today 2000, 61, 295.
[47] G. Centi, F. Trifiro, J. R. Ebner, V. M. Franchetti, Chem. Rev. 1988, 88, 55.
[48] A. E. Schweizer, M. E. Jones, D. A. Hickman, WO 09 862 058 ed., USA,
2002.
[49] A. E. Schweizer, M. E. Jones, D. A. Hickman, WO 6 984 63 B2 ed., USA,
2006.
[50] A. E. Schweizer, M. E. Jones, D. A. Hickman, EP1395536 ed., USA, 2007.
[51] C. Kumar, S. Gaab, T. Mller, J. Lercher, Top. Catal. 2008, 50, 156.
[52] M. P. Woods, B. Mirkelamoglu, U. S. Ozkan, J. Phys. Chem. C 2009, 113,
10112.
[53] M.-J. Cheng, K. Chenoweth, J. Oxgaard, A. van Duin, W. A. Goddard, J.
Phys. Chem. C 2007, 111, 5115.
[54] C. Hess in Nanostructured Catalysts - Selective Oxidations (Eds.: C. Hess,
R. Schlgl), Royal Society of Chemistry, Cambridge, 2011, pp. 299.
[55] J. E. Molinari, I. E. Wachs, J. Am. Chem. Soc. 2010, 132, 12559.
[56] X. Rozanska, E. V. Kondratenko, J. Sauer, J. Catal. 2008, 256, 84.
[57] R. K. Grasselli, Top. Catal. 2002, 21, 79.
[58] P. Concepcin, A. Galli, J. M. L. Nieto, A. Dejoz, M. I. Vazquez, Top. Catal.
1996, 3, 451.
[59] O. R. Evans, A. T. Bell, T. D. Tilley, J. Catal. 2004, 226, 292.
[60] J. L. Callahan, R. K. Grasselli, AIChE J. 1963, 9, 755.
[61] H. Dai, A. T. Bell, E. Iglesia, J. Catal. 2004, 221, 491.
[62] E. V. Kondratenko, M. Cherian, M. Baerns, D. Su, R. Schlgl, X. Wang,
I. E. Wachs, J. Catal. 2005, 234, 131.
[63] E. V. Kondratenko, M. Cherian, M. Baerns, Catal. Today 2006, 112, 60.
[64] P. Botella, A. Dejoz, J. M. Lopez Nieto, P. Concepcion, M. I. Vazquez,
Appl. Catal. A 2006, 298, 16.
[65] P. Botella, A. Dejoz, M. C. Abello, M. I. Vazquez, L. Arrua, J. M. Lopez
Nieto, Catal. Today 2009, 142, 272.
[66] A. Trunschke in Nanostructured CatalystsSelective Oxidations (Eds.: C.
Hess, R. Schlgl), Royal Society of Chemistry, Cambridge, 2011, pp. 5.
[67] B. Tope, Y. Zhu, J. A. Lercher, Catal. Today 2007, 123, 113.
[68] R. K. Grasselli, Top. Catal. 2001, 15, 93.
[69] R. Schlgl, C. Hess in Nanostructured CatalystsSelective Oxidations
(Eds.: R. Schlgl, C. Hess), Royal Society of Chemistry, Cambridge,
2011, pp. 355.
[70] M. D. Argyle, K. Chen, A. T. Bell, E. Iglesia, J. Phys. Chem. B 2002, 106,
5421.
[71] I. E. Wachs, Catal. Today 2005, 100, 79.
[72] G. Bergeret, P. Gallezot, K. V. R. Chary, B. R. Rao, V. S. Subrahmanyam,
Appl. Catal. 1988, 40, 191.
[73] J. Haber, A. Kozlowska, R. Kozlowski, J. Catal. 1986, 102, 52.
[74] T. Blasco, A. Galli, J. M. Lopez Nieto, F. Trifiro, J. Catal. 1997, 169, 203.
[75] H. Tian, E. I. Ross, I. E. Wachs, J. Phys. Chem. B 2006, 110, 9593.
[76] X. Gao, M. A. Banares, I. E. Wachs, J. Catal. 1999, 188, 325.
[77] Z. Zhao, X. Gao, I. E. Wachs, J. Phys. Chem. B 2003, 107, 6333.
[78] S. Feyel, D. Schrder, X. Rozanska, J. Sauer, H. Schwarz, Angew. Chem.
2006, 118, 4793; Angew. Chem. Int. Ed. 2006, 45, 4677.
[79] S. Feyel, D. Schrder, H. Schwarz, J. Phys. Chem. A 2006, 110, 2647.
[80] A. Klisinska, S. Loridant, B. Grzybowska, J. Stoch, I. Gressel, Appl. Catal.
A 2006, 309, 17.
[81] I. E. Wachs, B. M. Weckhuysen, Appl. Catal. A 1997, 157, 67.
[82] P. Knotek, L. Capek, R. Bul nek, J. Adam, Top. Catal. 2007, 45, 51.
www.chemcatchem.org
[83] A. Klisinska, K. Samson, I. Gressel, B. Grzybowska, Appl. Catal. A 2006,
309, 10.
[84] F. Cavani, G. Centi, F. Trifiro, R. K. Grasselli, Catal. Today 1988, 3, 185.
[85] A. Galli, J. M. Lopez Nieto, A. Dejoz, M. I. Vazquez, Catal. Lett. 1995, 34,
51.
[86] P. Concepcin, T. Blasco, J. M. Lopez Nieto, A. Vidal-Moya, A. MartinezArias, Microporous Mesoporous Mater. 2004, 67, 215.
[87] P. Concepcin, A. Corma, J. M. L. Nieto, J. Perez-Pariente, Appl. Catal. A
1996, 143, 17.
[88] P. Concepcin, J. M. Lopez Nieto, J. Perez-Pariente, Catal. Lett. 1994,
28, 9.
[89] D. B. Akolekar, J. Catal. 1993, 144, 148.
[90] P. Concepcin, J. M. Lpez Nieto, J. Prez-Pariente, Catal. Lett. 1993,
19, 333.
[91] B. Solsona, T. Blasco, J. M. Lpez Nieto, M. L. Pea, F. Rey, A. VidalMoya, J. Catal. 2001, 203, 443.
[92] P. Concepcin, J. M. Lpez Nieto, A. Mifsud, J. Prez-Pariente, Appl.
Catal. A 1997, 151, 373.
[93] P. Concepcin, A. Corma, J. M. Lpez Nieto, J. Prez-Pariente, Appl.
Catal. A 1996, 143, 17.
[94] T. Blasco, P. Concepcin, J. M. Lpez Nieto, J. Perez-Pariente, J. Catal.
1995, 152, 1.
[95] T. Blasco, L. Fern ndez, A. Martnez-Arias, M. S nchez-S nchez, P. Concepcin, J. M. Lpez Nieto, Microporous Mesoporous Mater. 2000, 39,
219.
[96] P. M. Michalakos, M. C. Kung, I. Jahan, H. Kung, J. Catal. 1993, 140, 226.
[97] P. Ciambelli, P. Galli, L. Lisi, M. A. Massucci, P. Patrono, R. Pirone, G.
Ruoppolo, G. Russo, Appl. Catal. A 2000, 203, 133.
[98] B. Solsona, A. Dejoz, T. Garcia, P. Concepcion, J. M. L. Nieto, M. I. Vazquez, M. T. Navarro, Catal. Today 2006, 117, 228.
[99] M. C. Abello, M. F. Gomez, O. Ferretti, Appl. Catal. A 2001, 207, 421.
[100] E. Heracleous, M. Machli, A. A. Lemonidou, I. A. Vasalos, J. Mol. Catal. AChem. 2005, 232, 29.
[101] E. Heracleous, J. Vakros, A. A. Lemonidou, C. Kordulis, Catal. Today
2004, 91, 289.
[102] A. Christodoulakis, E. Heracleous, A. A. Lemonidou, S. Boghosian, J.
Catal. 2006, 242, 16.
[103] R. B. Watson, U. S. Ozkan, J. Catal. 2000, 191, 12.
[104] R. B. Watson, U. S. Ozkan, J. Catal. 2002, 208, 124.
[105] R. B. Watson, U. S. Ozkan, J. Mol. Catal. A-Chem. 2003, 194, 115.
[106] R. B. Watson, S. L. Lashbrook, U. S. Ozkan, J. Mol. Catal. A-Chem. 2004,
208, 233.
[107] C. Liu, U. S. Ozkan, J. Mol. Catal. A-Chem. 2004, 220, 53.
[108] C. Liu, U. S. Ozkan, J. Phys. Chem. A 2005, 109, 1260.
[109] J. M. Lopez Nieto, P. Botella, M. I. Vazquez, A. Dejoz, Chem. Commun.
2002, 1906.
[110] P. Botella, E. Garcia-Gonzalez, A. Dejoz, J. M. Lopez Nieto, M. I. Vazquez,
J. Gonzalez-Calbet, J. Catal. 2004, 225, 428.
[111] T. Konya, T. Katou, T. Murayama, S. Ishikawa, M. Sadakane, D. Buttrey,
W. Ueda, Catal. Sci. Technol. 2013, 3, 380.
[112] N. Fang Chen, W. Ueda, K. Oshihara, Chem. Commun. 1999, 517.
[113] J. M. Lpez Nieto, B. Solsona, P. Concepcin, F. Ivars, A. Dejoz, M. I.
V zquez, Catal. Today 2010, 157, 291.
[114] A. A. Adesina, N. W. Cant, A. Saberi-Moghaddam, C. H. L. Szeto, D. L.
Trimm, J. Chem. Technol. Biotechnol. 1998, 72, 19.
[115] P. Botella, E. Garca-Gonz lez, J. M. Lpez Nieto, J. M. Gonz lez-Calbet,
Solid State Sci. 2005, 7, 507.
[116] W. Ueda, K. Oshihara, Appl. Catal. A 2000, 200, 135.
[117] J. Guan, S. Wu, H. Wang, S. Jing, G. Wang, K. Zhen, Q. Kan, J. Catal.
2007, 251, 354.
[118] E. Heracleous, A. F. Lee, K. Wilson, A. A. Lemonidou, J. Catal. 2005, 231,
159.
[119] S. Wang, K. Murata, T. Hayakawa, S. Hamakawa, K. Suzuki, J. Chem.
Technol. Biotechnol. 2001, 76, 265.
[122] B. Solsona, J. M. Lpez Nieto, P. Concepcin, A. Dejoz, F. Ivars, M. I.
V zquez, J. Catal. 2011, 280, 28.
[123] B. Solsona, P. Concepcin, S. Hern ndez, B. Demicol, J. M. L. Nieto,
Catal. Today 2012, 180, 51.
&21&
CHEMCATCHEM
REVIEWS
[124] B. Solsona, P. Concepcion, B. Demicol, S. Hernandez, J. J. Delgado, J. J.
Calvino, J. M. Lopez Nieto, J. Catal. 2012, 295, 104.
[125] M. Iwamoto, Y. Yoda, M. Egashira, T. Seiyama, J. Phys. Chem. 1976, 80,
1989.
[126] H. Zhu, S. Ould-Chikh, D. H. Anjum, M. Sun, G. Biausque, J.-M. Basset,
V. Caps, J. Catal. 2012, 285, 292.
[127] Z. Skoufa, E. Heracleous, A. A. Lemonidou, Catal. Today 2012, 192, 169.
[128] B. Savova, S. Loridant, D. Filkova, J. M. M. Millet, Appl. Catal. A 2010,
390, 148.
[129] Z. Skoufa, E. Heracleous, A. A. Lemonidou, Chem. Eng. Sci. 2012, 84, 48.
[130] Y. Brik, M. Kacimi, M. Ziyad, F. Bozon-Verduraz, J. Catal. 2001, 202, 118.
[131] S.-W. Ho, J. M. Cruz, M. Houalla, D. M. Hercules, J. Catal. 1992, 135, 173.
[132] X. Zhang, Q. Ye, B. Xu, D. He, Catal. Lett. 2007, 117, 140.
[133] E. M. Kennedy, N. W. Cant, Appl. Catal. 1991, 75, 321.
[134] E. M. Kennedy, N. W. Cant, Appl. Catal. A 1992, 87, 171.
[135] E. Reverchon, G. D. Porta, D. Sannino, L. Lisi, P. Ciambelli in Studies in
Surface Science and Catalysis, Vol. 118 (Eds.: B. Belmon, P. A. Jacobs, R.
Maggi, J. A. Martens, P. Grange, G. Poncelet), Elsevier, 1998, pp. 349.
[136] L. Ji, J. Liu, X. Chen, M. Li, Catal. Lett. 1996, 39, 247.
[137] R. X. Valenzuela, G. Bueno, V. Cortes Corberan, Y. Xu, C. Chen, Catal.
Today 2000, 61, 43.
[138] R. X. Valenzuela, G. Bueno, A. Solbes, F. Sapina, E. Martinez, V. C. Corberan, Top. Catal. 2001, 15, 181.
[139] V. R. Choudhary, S. A. R. Mulla, V. H. Rane, J. Chem. Technol. Biotechnol.
1998, 71, 167.
[140] V. Cortes Corber n, Catal. Today 2005, 99, 33.
[141] C. T. Au, K. D. Chen, H. X. Dai, Y. W. Liu, C. F. Ng, Appl. Catal. A 1999,
177, 185.
[142] J. Z. Luo, X. P. Zhou, Z. S. Chao, H. L. Wan, Appl. Catal. A 1997, 159, 9.
[143] H. X. Dai, C. F. Ng, C. T. Au, Appl. Catal. A 2000, 202, 1.
[144] S. Sugiyama, K. Sogabe, T. Miyamoto, H. Hayashi, J. B. Moffat, Catal.
Lett. 1996, 42, 127.
[145] E. Morales, J. H. Lunsford, J. Catal. 1989, 118, 255.
[146] D. Wang, M. P. Rosynek, J. H. Lunsford, Chem. Eng. Technol. 1995, 18,
118.
[147] S. Arndt, G. Laugel, S. Levchenko, R. Horn, M. Baerns, M. Scheffler, R.
Schlgl, R. Schomcker, Catal. Rev. Sci. Eng. 2011, 53, 424.
[148] E. Finazzi, C. Di Valentin, G. Pacchioni, M. Chiesa, E. Giamello, H. Gao, J.
Lian, T. Risse, H.-J. Freund, Chem. Eur. J. 2008, 14, 4404.
[149] J. C. Lian, E. Finazzi, C. Di Valentin, T. Risse, H. J. Gao, G. Pacchioni, H. J.
Freund, Chem. Phys. Lett. 2008, 450, 308.
[150] L. Leveles, S. Fuchs, K. Seshan, J. A. Lercher, L. Lefferts, Appl. Catal. A
2002, 227, 287.
[151] M. Machli, C. Boudouris, S. Gaab, J. Find, A. A. Lemonidou, J. A. Lercher,
Catal. Today 2006, 112, 53.
[152] D. J. Wang, M. P. Rosynek, J. H. Lunsford, J. Catal. 1995, 151, 155.
[153] V. L. Cherginets, T. P. Rebrova, Electrochim. Acta 1999, 45, 469.
[154] V. A. Volkovich, T. R. Griffiths, D. J. Fray, M. Fields, J. Chem. Soc. Faraday
Trans. 1997, 93, 3819.
[155] V. A. Volkovich, T. R. Griffiths, D. J. Fray, R. C. Thied, J. Nucl. Mater. 2000,
282, 152.
www.chemcatchem.org
[156] X. Lin, C. A. Hoel, W. M. H. Sachtler, K. R. Poeppelmeier, E. Weitz, J.
Catal. 2009, 265, 54.
[157] G. Busca, E. Finocchio, V. Lorenzelli, G. Ramis, M. Baldi, Catal. Today
1999, 49, 453.
[158] B. Frank, M. Morassutto, R. Schomcker, R. Schlgl, D. S. Su, ChemCatChem 2010, 2, 644.
[159] M. L. Rodriguez, D. E. Ardissone, E. Heracleous, A. A. Lemonidou, E.
Lpez, M. N. Pedernera, D. O. Borio, Catal. Today 2010, 157, 303.
[160] M. L. Rodriguez, D. E. Ardissone, E. Lopez, M. N. Pedernera, D. O. Borio,
Ind. Eng. Chem. Res. 2010, 50, 2690.
[161] T. Waku, M. D. Argyle, A. T. Bell, E. Iglesia, Ind. Eng. Chem. Res. 2003, 42,
5462.
[162] F. Klose, T. Wolff, S. Thomas, A. Seidel-Morgenstern, Catal. Today 2003,
82, 25.
[163] A. L. Y. Tonkovich, J. L. Zilka, D. M. Jimenez, G. L. Roberts, J. L. Cox,
Chem. Eng. Sci. 1996, 51, 789.
[164] A. C. van Veen, M. Rebeilleau, D. Farrusseng, C. Mirodatos, Chem.
Commun. 2003, 32.
[165] M. Huff, L. D. Schmidt, J. Phys. Chem. 1993, 97, 11815.
[166] M. C. Huff, I. P. Androulakis, J. H. Sinfelt, S. C. Reyes, J. Catal. 2000, 191,
46.
[167] F. Dons
, K. A. Williams, L. D. Schmidt, Ind. Eng. Chem. Res. 2005, 44,
3453.
[168] B. Silberova, M. Fathi, A. Holmen, Appl. Catal. A 2004, 276, 17.
[169] O. V. Buyevskaya, M. Baerns, Catal. Today 1998, 42, 315.
[170] S. A. R. Mulla, O. V. Buyevskaya, M. Baerns, J. Catal. 2001, 197, 43.
[171] E. V. Kondratenko, in Nanostructured CatalystsSelective Oxidations
(Eds.: C. Hess, R. Schlgl), Royal Society of Chemistry, Cambridge,
2011, pp. 340.
[172] A. Beretta, E. Ranzi, P. Forzatti, Chem. Eng. Sci. 2001, 56, 779.
[173] T. Liu, V. Gepert, G. Veser, Chem. Eng. Res. Des. 2005, 83, 611.
[174] S. J. Conway, D. J. Wang, J. H. Lunsford, Appl. Catal. A 1991, 79, L1.
[175] S. A. R. Mulla, O. V. Buyevskaya, M. Baerns, Appl. Catal. A 2002, 226, 73.
[176] S. Gaab, M. Machli, J. Find, R. K. Grasselli, J. A. Lercher, Top. Catal. 2003,
23, 151.
[177] S. Wang, K. Murata, T. Hayakawa, K. Suzuki, Chem. Eng. Technol. 2000,
23, 1099.
[178] C. T. Au, X. P. Zhou, H. L. Wan, Catal. Lett. 1996, 40, 101.
[179] D. A. Goetsch, L. D. Schmidt, Science 1996, 271, 1560.
[180] O. Czuprat, S. Werth, S. Schirrmeister, T. Schiestel, J. Caro, Chem. Ing.
Tech. 2009, 81, 1591.
[181] O. Czuprat, S. Werth, S. Schirrmeister, T. Schiestel, J. Caro, ChemCatChem 2009, 1, 401.
[182] L. Olivier, S. Haag, C. Mirodatos, A. C. van Veen, Catal. Today 2009, 142,
34.
Received: December 30, 2012

Published online on && &&, 0000
&22&
REVIEWS
A great suC=Cess: The oxidative dehydrogenation (ODH) of ethane is a process that enables the production of
ethene with high selectivities. Different
catalysts and reactor configurations
have been developed for this process.
This Review focuses on the mechanistic
aspects of heterogeneously catalyzed
ODH reactions and the main features
and common principles of this reaction
are discussed.
C. A. Grtner, A. C. van Veen,

J. A. Lercher*
&& &&
Oxidative Dehydrogenation of Ethane:
Common Principles and Mechanistic
Aspects
&23&

Ethylene Production

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Ethylene Production

Transféré par

Droits d'auteur :

Formats disponibles

CHEMCATCHEM

Oxidative Dehydrogenation of Ethane: Common Principles

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

These are not the final page numbers!

The increasing demand for light olefins and the changing

ers, full-range naphtha (boiling point: 30200 8C) or C6C8 cuts

Ethene is one of the most important building blocks in the

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

2. Current Practice: Steam Cracking, Ethane

These are not the final page numbers!

Andre C. van Veen was born in Issum

Johannes A. Lercher, born in Vienna

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Thermodynamic reasons make the use of high temperatures

These are not the final page numbers!

2.3. Dehydration of ethanol

3. Oxidative Dehydrogenation of Ethane:

Short-chain alkanes, in particular ethane that is extracted

Although the reaction was performed for a long time with

Figure 1. ODH reaction of ethane with side reactions.

These are not the final page numbers!

3.1. CH activation and cleavage

Figure 2. Single-electron process for CH bond activation (adapted from

These are not the final page numbers!

ed.[35] A likely model suggests that [Li+O] acts as the active

surface, leads to better control of the reaction and results in

3.1.2. Processes that involve paired electrons

The latter product decomposes into two hydroxyl radicals,

O2 regenerates the surface OH species [Eq. (8.1) and (8.2)].

3.2. Active sites for the ODH of ethane

These are not the final page numbers!

nated from the dihydroxy species, thereby reducing the metal

Figure 4. Different surface vanadium species (adapted from Ref. [1]).

nal V=O bonds and bridging oxygen bonds between a V atom

Figure 6 shows a mechanism that involves two redox centers

3.2.2. Catalysts with non-redox-active metals

These are not the final page numbers!

surface. Subsequently, the hydrogen atom in the first CH

To complete the list of catalysts that follow this pathway,

Subsequently, H is homolytically abstracted from adsorbed

The other possibility would be a ClC bond formation on

It is also conceivable that two H atoms are eliminated in

LiOClC2 H6 ! LiOH HClC2 H4

2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

However, V is known to form various metaloxide bonds in

3.3. Factors that govern the activity and selectivity of ODH

These are not the final page numbers!

3.3.4. Desorption and re-adsorption of ethane

Because olefins are more reactive than alkanes, ethene shows

These are not the final page numbers!

4. Groups of ODH Catalysts

4.1. Transition-metal oxides

These are not the final page numbers!

These are not the final page numbers!

These are not the final page numbers!

ChemCatChem 0000, 00, 1 23

These are not the final page numbers!

These are not the final page numbers!

4.3. Supported alkali oxides

4.4. Supported alkali chlorides

These are not the final page numbers!

Hypochlorite is assumed to be the transient species in low

For the ODH reactions of transition-metal-supported zeolites,

Table 1. Different active sites for the ODH of ethane.