Microporous and Mesoporous Materials 104 (2007) 269–280

www.elsevier.com/locate/micromeso

Catalytical studies on trimethylsilylated Ti-MCM-41

and Ti-MCM-48 materials
Naoko Igarashi a, Kazuhito Hashimoto a, Takashi Tatsumi b,*

a
Research Center for Advanced Science and Technology, The University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8904, Japan
b
Division of Catalytic Chemistry, Chemical Resources Laboratory, Tokyo Institute of Technology, 4259-R1-9 Nagatsuta, Midori-ku,
Yokohama 226-8503, Japan

Received 5 November 2006; received in revised form 19 February 2007; accepted 19 February 2007
Available online 25 February 2007

Abstract

Titanium-incorporated MCM-41 and MCM-48 materials (Ti-MCM-41 and Ti-MCM-48, respectively) have been trimethylsilylated
by employing various silylating agents (i.e., hexamethyldisilazane (HMDZ) and trimethylchlorosilane (TMCS) in toluene, and a mixture
of TMCS and hexametyldisiloxane (HMDS)). Their catalytic activities in various oxidation reactions as well as their physical properties
are compared. The highly reactive HMDZ has achieved the maximum degree of silylation to introduce 2.35 and 2.06 trimethylsilyl
groups per 1 nm2 on the surface of Ti-MCM-41 and -48, respectively. IR spectra have confirmed that HMDZ can react with hydro-
gen-bonded silanol groups as well as isolated silanol groups, generating a new band at 3700 cm1, while TMCS does not. These silylated
materials show remarkably high catalytic activity in the oxidation of substrates with various molecular sizes (from C6 to C12) with H2O2
and tert-butyl hydroperoxide (TBHP) compared to non-silylated samples. Introducing a small amount of H2O so as to double the H2O
content in a reaction system for the oxidation of cyclododecene using TBHP has approximately doubled the catalytic activity. Such unex-
pected results were only observed with highly hydrophobic samples, Furthermore, by conducting the reaction at 343 K for 6 h over
highly silylated Ti-MCM-48, high conversion of 98% has been attained.
 2007 Elsevier Inc. All rights reserved.

Keywords: Ti-MCM-41; Ti-MCM-48; Hydrophobicity; Trimethylsilylation; Epoxidation catalyst

1. Introduction [7,8] and a wide variety of mesoporous materials synthe-

The discovery of titanium-incorporated molecular sieves
has greatly extended the field of catalytic science and tech-
nology [1]. In particular, microporous titanosilicates such
as TS-1, Ti-beta and Ti-MWW with the MFI, BEA and
MWW structures, respectively, have shown high catalytic
activity in the selective oxidation of various organic sub-
strates with H2O2 [2–6]. However, their applications are
rather limited to the reactants with the molecular sizes
smaller than 7 Å due to their small micropore openings.
On the other hand, the emergence of the M41S family
*
Corresponding author. Tel.: +81 45 339 3943; fax: +81 45 339 3941.
E-mail address: ttatsumi@cat.res.titech.ac.jp (T. Tatsumi).
sized by the surfactant-templating method, having mesop-
ores with uniform sizes ranging from 15 to 100 Å, has
mitigated the size limitation of microporous materials.
Titanium-substituted mesoporous molecular sieves have
actually demonstrated the ability to catalyze the oxidation
of bulky reactant molecules which cannot enter into the
micropores of titanium-incorporated zeolites [9–11]. How-
ever, they show much lower catalytic activity than TS-1 or
Ti-beta in the oxidation of small reactant molecules with
H2O2.
In the preliminary communication we have reported
that the activity of titanium-containing mesoporous molec-
ular sieves in the oxidation of alkenes with H2O2 is remark-
ably enhanced by trimethylsilylation of silanol groups [12].
We assumed that the low activity of titanium-containing

1387-1811/$ - see front matter  2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.micromeso.2007.02.041
270 N. Igarashi et al. / Microporous and Mesoporous Materials 104 (2007) 269–280
mesoporous molecular sieves is caused by the poisoning of
catalytically active Ti sites by H2O molecules adsorbed on
the surface with relatively high hydrophilicity derived from
a large number of silanol groups [10]. This hypothesis has
been verified by the remarkable enhancement of the activ-
ity in the liquid phase oxidation regardless of the method
for introducing organic groups into the titanium-incorpo-
rated mesoporous catalysts; we have also reported that
the oxidation activity has been successfully improved by
direct one-step synthesis of organically modified mesopor-
ous titanium-substituted MCM-41 [13–15].
Catalytic activities can be improved by taking a different
approach to the control of the surface hydrophobicity. It
has been recently reported that the titanium-impregnated
NaY having an amphiphilic external surface obtained by
partly loading alkylsilane is an active phase-boundary het-
erogeneous epoxidation catalyst [16,17]. In this case, the
surface hydrophobicity benefits the catalytic activity in a
macroscopic point of view by allowing the catalyst particles
to lay at the liquid–liquid phase boundary where catalyst is
accessible to both upper organic and lower aqueous phases.
For the trimethylsilylation of silanol groups, various
reagents have been used such as hexamethyldisilazane
(HMDZ), trimethylchlorosilane (TMCS), N-methyl-N-
(trimethylsilyl)trifluoroacetamide (MSTFA), trimethylsily-
limidazole, and hexamethyldisiloxane (HMDS) with/with-
out a solvent [8,18–34], and it has been claimed that the
degree of trimethylsilylation is affected by various factors
such as the dose and reactivity of silylating agent and reac-
tion temperature. Hertl and Hair estimated the activation
energies in the silylation with TMCS and HMDZ at 22.0
and 18.5 kcal/mol, respectively [19,20], showing the higher
reactivity of HMDZ. Actually, TMCS generally requires
higher temperature in silylation than HMDZ. It is also
reported that the conditions for the pre-treatment of silica
surface are important, where the dehydration occurs to
change the environment of silanol groups to be modified
by silylating reagents [23]. A number of researchers sug-
gested that TMCS and HMDS can react with hydrogen-
bonded vicinal silanol groups [18,24,25], and that geminal
silanol groups (@Si(OH)2) are more reactive with silylating
agents than single ones („Si(OH)) [22,30,31].
Thus, a lot of researches have been pursued into the
trimethylsilylation of silica materials. However, as far as
we know, systematic studies on the trimethylsilylation of
titanium-containing mesoporous molecular sieves and its
effect on their catalytic performance have not yet been
reported.
In this study, Ti-MCM-41 and Ti-MCM-48 samples are
silylated under various silylating conditions, and the effec-
tive trimethylsilylating method for obtaining a highly
active catalyst is investigated. Thus silylated materials are
employed as catalysts in the oxidation of cyclic alkenes
and unsaturated alcohols, whose molecular sizes are larger
than micropore sizes of titanium-incorporated zeolites,
using an aqueous solution of H2O2 or TBHP as an
oxidant.
2. Experimental
2.1. Materials
Ti-MCM-41 was synthesized as follows. An aqueous
solution of NaOH and hexadecyltrimethylammonium bro-
mide (C16TMABr) was added to TEOS under vigorous
stirring, and immediately tetrabutyl orthotitanate (TBOT)
was added dropwise. The resulting mixture was stirred
for 1 day at room temperature, followed by the hydrother-
mal treatment at 373 K for 6 days. The molar composition
was TEOS:0.12C16TMABr:0.3NaOH:100H2O:0.02TBOT.
Tetramethylammonium hydroxide (TMAOH) was used
instead of NaOH as a base in the synthesis of Ti-MCM-
41(TMA) for control. Ti-MCM-48 was synthesized in a
similar manner from the mother gel with the molar compo-
sition of TEOS:0.67C16TMABr:0.64NaOH:103H2O:0.02T-
BOT.
The organic templates were removed by the acid treat-
ment using 1 M HCl solution in ethanol at 353 K for
16 h, followed by the calcination at 813 K for 5 h at a heat-
ing rate of 1 K/min. In this treatment, Na cations were
completely removed by ion-exchange, as verified by induc-
tively coupled plasma (ICP) analyses.
2.1.1. Post-synthesis trimethylsilylation
In the post-synthesis trimethylsilylation, TMCS,
HMDZ, or HMDS were used as silylating agents. A typical
surface modification was conducted as follows. 1.0 g of
template-free Ti-MCM-41 was evacuated at 473 K for a
given time and then subsequently dispersed in the mixture
of TMCS and HMDS (total silylating reagent/Ti-MCM-
41 = 0.55 mol/g) at refluxing under N2 atmosphere for
16 h. After filtration of the reaction mixture, the dry
powder obtained was thoroughly washed with dry acetone,
and dried at 393 K.
2.1.2. Analytical procedure
Powder X-ray diffraction (XRD) patterns were collected
on a Mac Science M3X HF-22E instrument equipped with
a CuKa X-ray source. ICP analyses were performed on a
Shimadzu ICPS-8000E spectrometer. The nitrogen and
water adsorption/desorption isotherm measurements were
carried out using Belsorp 28SA and 18 apparatus, respec-
tively. Solid-state 29Si MAS NMR spectra were recorded
on a JEOL JNM-ECA400 spectrometer at a frequency of
79.27 MHz for 29Si. Chemical shifts were referenced to
external tetramethylsilane. A sample spinning rate of
5 kHz, recycle delay time of 15 s, pulse width of 7.0 ls,
and 5000–10,000 scans were taken. Ultraviolet–visible
(UV–VIS) spectroscopy was performed on a Hitachi 340
spectrometer with a diffuse reflectance mode. The in situ
silylated and non-silylated Ti-MCM-41 samples were pre-
pared as self-supporting pellets and placed inside a cell
which allows IR spectrum to be measured after outgassed
at various temperatures. The spectra were recorded on a
Perkin–Elmer 1600 spectrometer.
 N. Igarashi et al. / Microporous and Mesoporous Materials 104 (2007) 269–280 271

In a typical oxidation reaction, 50 mg of Ti-MCM-41

was added as a catalyst to a mixture of 25 mmol of reactant
and 5 mmol of H2O2 (31 wt.% aqueous solution) or TBHP
(70 wt.% aqueous solution). The mixture was allowed to
react under vigorous stirring at 323 K for 3 h.

3. Results and discussion

3.1. Physicochemical properties

The XRD patterns of the Ti-substituted samples before

and after trimethylsilylation exhibit diffraction peaks char-
acteristic of the well-ordered hexagonal or cubic structure
(Fig. 1). UV–VIS spectra show the absorbance peak
around 220 nm both before and after trimethylsilylation,
confirming the incorporation of tetrahedrally coordinated
titanium (Fig. 2). In the case of Ti-MCM-41, the broad
shoulder peak around 250 nm was observed. This peak
can reasonably be assigned to the H2O-coordinated Ti spe-
Fig. 2. UV–VIS spectra of non-silylated and silylated Ti-MCM-41 and -48
cies (5- or 6-coordinated ones) because the removal of samples.
water by trimethylsilylation made this peak disappear.
After the silylation treatments, such broad shoulder peak
decreased probably due to the removal of extra framework
Ti species during the treatment. TMS species in the 29Si MAS NMR spectra, [Q3], [Q4],
The 29Si MAS NMR spectra of the trimethylsilylated and [TMS], respectively, as well as the reaction conditions
samples show three peaks centered at 110, 101, and for post-synthetic trimethylsilylation. These values are
15 ppm assignable to Si(OSi)4 (Q4), HOSi(OSi)3 (Q3), and based on the total Si of the parent samples. The increase
(CH3)3Si(OSi) (designated as TMS), respectively (Fig. 3). in [Q4] by trimethylsilylation, [+DQ4] (= [DQ3]), is also
Table 1 lists the peak area percentages of Q3, Q4 and shown.

Fig. 1. X-ray diffraction patterns for the non-silylated and silylated (a) Ti-MCM-41 and (b) Ti-MCM-48 samples.
272 N. Igarashi et al. / Microporous and Mesoporous Materials 104 (2007) 269–280

Ti-MCM-41 Ti-MCM-41-sil1
Q3 Q4
HO O O
Si Si Si

CH3
H3C CH3
Si

Si

B.G B.G

50 0 -50 -100 -150 50 0 -50 -100 -150

ppm ppm

Ti-MCM-41-sil4

B.G

50 0 -50 -100 -150

ppm
29
Fig. 3. Si MAS NMR spectra of non-silylated and silylated Ti-MCM-41 samples.

Table 1
29
Reaction conditions of trimethylsilylation and Si MAS NMR results
Evacuation temp. (K) Silylating agenta,b Solvent 29
Si MAS NMR peak area ratio (%)c
Q3 Q4 TMS DQ4(DQ3)
Ti-MCM-41 49 51 – –
Ti-MCM-41-sil1 – TMCS + HMDS – 29 71 11 20
Ti-MCM-41-sil2 473 TMCS + HMDS – 31 69 18 18
Ti-MCM-41-sil3 473 TMCS Toluene 36 64 10 13
Ti-MCM-41-sil4 473 HMDZ Toluene 15 85 25 34
Ti-MCM-41(TMA) 27 73 – –
Ti-MCM-41(TMA)-sil2 473 TMCS + HMDS – 13 87 16 14
Ti-MCM-41(TMA)-sil4 473 HMDZ Toluene 7 93 21 20
Ti-MCM-48 36 64 – –
Ti-MCM-48-sil1 – TMCS + HMDS – 22 78 10 14
Ti-MCM-48-sil2 473 TMCS + HMDS – 19 81 18 17
Ti-MCM-48-sil4 473 HMDZ Toluene 12 88 26 24
Post-synthesis reactions are carried out at 383 K for 16 h.
a
TMCS, trimethylchlorosilane; HMDS, hexamethyldisiloxane; HMDZ, hexamethyldisilazane.
b
Total Si of silylating agent/g = 0.55 mol/g.
c
TMS, (CH3)3Si(OSi); Q3, HOSi(OSi)3; Q4, Si(OSi)4; +DQ4, the amount increased by trimethylsilylation (all values are based on the parent non-
silylated samples).

Silylations without pre-evacuation at 473 K have resulting from the enhanced condensation reaction of the
resulted in a larger [+DQ4] (= [DQ3]) than the amount surface silanol groups during silylation at high tempera-
of trimethylsilyl group incorporated [TMS], probably ture. In addition, the final amounts of [Q3], [Q4], [TMS]
 N. Igarashi et al. / Microporous and Mesoporous Materials 104 (2007) 269–280
of the products silylated with HMDZ were fairly similar
between Ti-MCM-41 and Ti-MCM-48 samples, even
though the amount of silanol groups [Q3] of the Ti-
MCM-41 sample before silylation is considerably larger
(49% of total Si) than that of the Ti-MCM-48 sample
(36% of total Si). This result is quite interesting and prob-
ably indicates that the condensation of silanol groups easily
occurs at silylating temperature for the samples with large
amount of silanol groups located close to each other during
the incorporation of TMS groups.
The degree of trimethylsilylation in terms of [TMS] is
significantly increased by the pre-evacuation at 473 K for
both Ti-MCM-41 and -48 samples when they are modified
with a TMCS/HMDS solution. This fact indicates the
great sensitivity of silylation reaction to the adsorbed water
on the parent sample containing a relatively large amount
of silanol groups (40–50 mol% Q3). Water molecules
adsorbed on the solid surface would react with TMCS dur-
ing the silylation, resulting in the decomposition of the sily-
lating agent.
By using HMDZ as a silylating reagent, a higher degree
of trimethylsilylation in the range of 21–26% of [TMS] was
attained for each sample, in agreement with the findings
by Hertl and Hair [20]. Fig. 4 exhibits the in situ IR spectra
of Ti-MCM-41(TMA) silylated with TMCS/HMDS or
HMDZ. After the removal of adsorbed water by the pre-
evacuation at 473 K, the parent Ti-MCM-41(TMA) shows
bands at 3740 and 3500–3700 cm1 due to isolated and
hydrogen-bonded silanol groups, respectively (Fig. 4a).
When this material is allowed to react with TMCS/HMDS
(Fig. 4b), the peak assigned to isolated silanol group disap-
pears, indicating the silylation of isolated silanol groups.
On the other hand, after the reaction with HMDZ, a new
band appears at 3700 cm1 in addition to the band of
hydrogen-bonded silanol groups around 3600 cm1
(Fig. 4c). The former is probably due to the generation
of new isolated silanol groups adjacent to TMS groups pro-
duced from the reaction between hydrogen-bonded silanol
groups and HMDZ. Such a new band of isolated silanol
groups in silylated samples is also observed by other
researchers [27,30,31]. Thus, it can be concluded that
Fig. 4. In situ IR spectra of (a) non-silylated, silylated Ti-MCM-41(TMA) samples (b) Ti-MCM-41(TMA)-sil2, and (c) Ti-MCM-41(TMA)-sil4 evacuated
at various temperatures.
273
HMDZ reacts with hydrogen-bonded silanol groups, while
TMCS does not, which could be the main reason for
HMDZ giving a high degree of silylation.
It is to be noted that the use of TMCS alone in toluene
solvent leads to lower TMS introduction than the use of a
TMCS/HMDS mixture. Kinkel and Unger suggests the
importance of the type of the solvent in the silylation reac-
tion where the solvent molecule interacts with a surface
silanol group resulting in the weakening of Si–O bond
strength due to the pronounced Lewis acid and base char-
acter [35]. Such efficient solvents are dichloromethane, ace-
tonitrile and N,N-dimethylformamide, while less efficient
solvents are benzene, tetrahydrofuran and diethyl ether
[35]. Therefore, although HMDS is a less reactive silylating
reagent, HMDS probably would play a role as an enhanc-
ing solvent.
Various physical data obtained by N2 and H2O adsorp-
tion measurements and elemental analyses are summarized
in Table 2. The unit cell parameter and the pore diameter
of Ti-MCM-41(TMA) are by 3 Å smaller than those of
Ti-MCM-41. The unit cell parameters of Ti-MCM-41,
-41(TMA) and -48 remain unchanged after trimethylsilyla-
tion. In contrast, the pore diameters and pore volumes of
Ti-MCM-41, -41(TMA) and -48 decrease according to
the degree of trimethylsilylation, indicating the successful
substitution of trimethylsilyl groups for silanol protons
on the surface of mesopores. In the presence of Na+ cation,
the titanium incorporation is slightly lower than in the
presence of TMA+. As is well-known in titanosilicate zeo-
lites [3], extraframework Ti species could be formed in the
presence of Na+. However, it can be removed during
the template extraction using HCl-ethanol solution. After
the trimethylsilylation, the Si/Ti ratios are slightly
decreased by the incorporation of trimethylsilyl groups
depending on the degree of trimethylsilylation.
The water adsorption capacity of trimethylsilylated/
non-silylated Ti-MCM-41 and -48 samples is estimated
by the H2O adsorption measurement. The parent samples
show large water adsorption capacities (V H2 O ¼ 91, 48,
and 66 ml/g for Ti-MCM-41, -41(TMA), and -48, respec-
tively), demonstrating rather high affinity to water
274 N. Igarashi et al. / Microporous and Mesoporous Materials 104 (2007) 269–280

Table 2
Change in physical properties upon trimethylsilylation
d100/d211a (Å) Pore diameterb (Å) BET surface areab (m2/g) Pore volumeb (ml/g) Si/Ti ratioc V H2 O (ml/g)d
Ti-MCM-41 44 30 1026 1.2 117 91
Ti-MCM-41-sil1 44 24 823 0.9 130 15
Ti-MCM-41-sil2 44 25 813 0.9 138 9
Ti-MCM-41-sil3 44 25 882 0.9 128 19
Ti-MCM-41-sil4 43 23 580 0.8 146 5
Ti-MCM-41(TMA) 41 27 1029 1.3 77 48
Ti-MCM-41(TMA)-sil2 41 23 631 0.7 90 9
Ti-MCM-41(TMA)-sil4 41 23 675 0.7 93 6
Ti-MCM-48 37 27 1251 1.3 91 66
Ti-MCM-48-sil1 37 23 892 1.0 100 11
Ti-MCM-48-sil2 37 23 891 1.0 108 14
Ti-MCM-48-sil4 37 21 710 0.8 115 6
a
d100 for T-MCM-41 and d211 for Ti-MCM-48.
b
N2 adsorption measurement.
c
ICP analysis.
d
H2O adsorption measurements.

molecules. As shown in Table 2, by the trimethylsilylation
the water adsorption capacity dramatically decreases
according to the degree of silylation, suggesting that the
trimethylsilylated Ti-substituted mesoporous materials are
much more hydrophobic than the non-silylated materials.
3.1.1. Catalysis
The catalytic activities of the trimethylsilylated Ti-
substituted samples in cyclohexene oxidation without sol-
vent (aqueous-organic-solid triphase system) prove to be
remarkably high compared to those of the non-silylated
ones (Table 3). In addition to cyclohexene oxide and cyclo-
hexanediol, allylic oxidation products (2-cyclohexen-1-ol
and 2-cyclohexen-1-one), formed by radical and possibly
singlet molecular oxygen [36], are obtained. Even though
the allylic oxidation is suppressed by trimethylsilylation,
probably following a kinetic model of the suppression by
trimethylsilylation applied on the Ti-silica mixed oxides
proposed by Figueras and Kochkar [37,38], the major
product is cyclohexanediol produced consecutively from
cyclohexene oxide through the acid catalyzed hydrolysis
reaction in the triphase system. Furthermore, non-produc-
tive decomposition of H2O2 has been effectively suppressed
by the trimethylsilylation.
In the cyclohexene oxidation, in a biphase system using
acetonitrile as a solvent, the non-silylated Ti-mesoporous

Table 3
Oxidation of cyclohexene with H2O2 over non-silylated and silylated Ti-MCM-41 and -48 materials in triphase system
Catalyst Conv. (mol%max.) TONa Selectivity (%) H2O2 decomp.b (%)
O HO OH
OH O

Cyclohexene oxidation
Ti-MCM-41c – – – – – – 35
No catalyst 0 – 0 0 0 0 14
MCM-41 0 – 0 0 0 0 2
Ti-MCM-41(117) 1.3 10 16 5 29 50 50
Ti-MCM-41-sil1(130) 21 197 7 6 2 85 22
Ti-MCM-41-sil2(138) 27 291 8 3 1 88 2
Ti-MCM-41-sil3(128) 22 202 7 4 2 87 10
Ti-MCM-41-sil4(146) 34 404 8 2 10 80 9
Ti-MCM-41(TMA)(77) 2.1 10 5 0 14 81 33
Ti-MCM-41(TMA)-sil2(90) 32 216 9 2 3 86 7
Ti-MCM-41(TMA)-sil4(93) 24 178 7 2 7 85 23
Ti-MCM-48(91) 1.5 9 16 9 0 75 59
Ti-MCM-48-sil1(100) 24 168 9 5 1 85 26
Ti-MCM-48-sil2(108) 36 299 8 3 1 87 0
Ti-MCM-48-sil4(115) 35 333 8 3 8 81 10
Catalyst 50 mg, substrate 25 mmol, H2O2 5 mmol, 323 K, 3 h.
Numbers in parentheses denote the Si/Ti ratio.
a
mol(mol-Ti)1.
b
H2O2 decomposition = (H2O2 total  amount used  amount remained)/H2O2 total · 100%.
c
Catalyst 50 mg, H2O2 5 mmol, 323 K, 3 h.
 N. Igarashi et al. / Microporous and Mesoporous Materials 104 (2007) 269–280 275

molecular sieves show much higher catalytic activities than
in a triphase system (Table 4). The catalytic activity is not
improved with increasing reaction temperature by 20 K
from 323 to 343 K. On the other hand, the trimethylsily-
lated samples show high catalytic activity at higher temper-
atures. It is noteworthy that epoxide becomes a major
product when CH3CN is used. Such a high epoxide selec-
tivity is probably due to the following reasons: (i) the basi-
city of CH3CN efficiently neutralizes the weak acidity
which catalyzes the ring opening of epoxide and (ii)
CH3CN dissolves the product adsorbed on the catalyst sur-
face to protect the intermediate epoxide from the further
ring opening reaction. In contrast, when methanol is used
instead of CH3CN the epoxide selectivity as well as the cat-
alytic activity greatly reduced. When the H2O2/cycloxene
ratio is increased from the typical ratio of 1/5 to 5/5, the
catalytic activity in terms of TON is not significantly
improved (Table 4).
The oxidation of bulkier reactants with molecular sizes
beyond the range of micropore opening of TS-1 zeolite is
carried out. In the oxidation of cyclododecene using
H2O2, the catalytic activity of Ti-MCM-48 with the 3-
dimensional pore system is two times as high as that of
Ti-MCM-41 having the 1-dimensional one, probably due
to superior diffusivity in the 3-dimensional pore system
(Table 5). Trimethylsilylation increases the activity in the
oxidation of cyclododecene by from 2 to 4 times. At
323 K, their catalytic activity is enhanced with increasing
degree of silylation for the series of Ti-MCM-41-sil, while
it is decreased for the silylated Ti-MCM-48 (cf. Ti-MCM-
48-sil2 and -sil4). This trend was only observed in the
oxidation of the relatively large molecular size of cyclod-
odecene in contrast to the result of cyclohexene (Tables 3
and 5). The increase in reaction temperature (from 323 to
343 K) enhances the catalytic activity reaching 30–50%
conversions for the silylated Ti-MCM-48-sil catalysts as
clearly shown in Fig. 5. It also shows that the conversion
reaches the maximum within a shorter reaction time (4–
5 h).
The selectivity to diol in the cyclododecene oxidation is
low compared to that in the cyclohexene oxidation proba-
bly due to the structural stability of cyclododecene oxide
compared to cyclohexene oxide. Non-productive decompo-
sition of H2O2 is also retarded by the trimethylsilylation.
The activity improved in the oxidation of cyclododecene
by trimethylsilylation when TBHP is used as an oxidant
(Table 6). This is probably related to high surface hydro-
phobicity that allows trimethylsilylated catalysts to be well
dispersed in the mixture of hydrophobic alkene substrates
and the amphiphilic TBHP. Interestingly, cis-epoxide is
preferentially formed for both the silylated Ti-MCM-41
and -48 (Table 6). The diol product is obtained only over
the non-silylated ones, showing the facile conversion of
epoxide to diol on the hydrophilic Ti-MCM-41 and -48 cat-
alysts having silanol groups that may promote the acid-cat-
alyzed ring-opening reactions.
In the typical oxidation reaction using 5/25 mmol of
TBHP/cyclododecene mixture, 0.2 ml (4 wt.% of the total
reaction system) of water derived from TBHP aqueous
solution is present in the triphase system (Table 6). Surpris-
ingly, when 0.2 ml of water is further added to this system
(thus, making up 8 wt.% H2O of the total system), the

Table 4
Oxidation of cyclohexene with H2O2 over non-silylated and silylated Ti-MCM-41 and -48 materials in biphase system
Catalyst Temperature (K) Conv. (mol% of max) TON (mol(mol-Ti)1) H2O2 decomposition (%) Epoxide selectivity (%)
a
Ti-MCM-41(117) 323 38 29 30 78
333 41 31 35 73
343 39 30 40 80
Ti-MCM-41-sil1(130)a 323 28 26 26 72
333 39 37 27 81
343 41 38 29 49
Ti-MCM-41-sil2(138)a 343 60 65 26 82
Ti-MCM-41-sil4(146)a 343 66 79 16 83
Ti-MCM-41(TMA)-sil2(90)a 343 56 38 15 79
Ti-MCM-48(91)a 323 35 20 23 49
333 46 27 32 75
343 36 21 45 89
Ti-MCM-48-sil1(100)a 333 56 40 16 74
343 51 36 42 87
Ti-MCM-48-sil1(100)b 333 31 22 20 6d
Ti-MCM-48-sil1(100)c 333 14 49 24 52e
Ti-MCM-48-sil2(108)a 343 62 51 14 69
Ti-MCM-48-sil4(115)a 343 69 66 11 83
Numbers in parentheses denote the Si/Ti ratio.
a
Catalyst 0.1 g, substrate 5 mmol, H2O2 1 mmol, CH3CN 10 ml, 323–343 K, 3 h.
b
Catalyst 0.1 g, substrate 5 mmol, H2O2 1 mmol, methanol 10 ml, 333 K, 3 h.
c
Catalyst 0.1 g, substrate 5 mmol, H2O2 5 mmol, CH3CN 10 ml, 333 K, 3 h.
d
Major product is MGE (selectivity 94%).
e
Cyclohexene-1-ol (7%), cyclohexene-1-one (41%).
276 N. Igarashi et al. / Microporous and Mesoporous Materials 104 (2007) 269–280

Table 5
Oxidation of cyclododecene with H2O2 over non-silylated and silylated Ti-MCM-41 and -48 materials in triphase system
Catalyst Temperature (K) Length (h) Conv. (mol%max) TONa Selectivity (%) H2O2 decomposition (%)
OH
O
OH OH
O OH
trans cis cis
trans
b
Cyclododecene oxidation
Ti-MCM-41 323 3 2.5 19(6) 34 52 14 0 39.8
Ti-MCM-48 323 3 4.8 28(9) 22 60 18 0 51.3
Ti-MCM-48 343 16 9.2 54(3) 26 46 28 0 91.0
Ti-MCM-41-sil2 323 3 7.7 83(28) 31 53 16 0 27.4
Ti-MCM-41-sil4 323 3 11 125(42) 22 78 0 0 14.0
Ti-MCM-41-sil4 323 16 17 200(12) 22 66 12 0 27.1
Ti-MCM-41-sil4 343 30 50 597(20) 19 70 11 0 51.1
Ti-MCM-48-sil2 323 3 16 134(45) 21 47 33 0 20.4
Ti-MCM-48-sil2 323 16 20 167(10) 25 57 17 0 60.0
Ti-MCM-48-sil2 343 16 32 268(17) 26 51 23 0 69.9
Ti-MCM-48-sil4 323 3 10 96(32) 20 80 0 0 17.1
Ti-MCM-48-sil4 323 16 17 163(10) 24 54 22 0 46.6
Ti-MCM-48-sil4 343 16 49 470(29) 22 62 16 0 49.4
Ti-MCM-48-sil4 343 30 55 519(17) 22 62 16 0 44.0
Numbers in parentheses denote TOF (mol(mol-Ti h)1).
a
(mol(mol-Ti)1).
b
Cat. 50 mg, substrate 25 mmol, H2O2 5 mmol.

Fig. 5. Time profiles of catalytic activity of non-silylated and silylated Ti-MCM-48 in the oxidation of cyclododecene using H2O2 at 323 and 343 K.

activity doubles reaching 46% conversion for both highly
silylated Ti-MCM-41-sil4 and -48-sil4, while such an
increase is not observed for the non-silylated ones (Table
6 and Fig. 6). The conversion reaches 98% over the highly
silylated Ti-MCM-48 at 343 K for 6 h. Ti-MCM-48 with a
less silyl group content (Ti-MCM-48-sil2) shows only 66%
conversion, demonstrating the strong dependence of the
catalytic performance on the degree of silylation.
This unique catalytic improvement would be strongly
related to the amphiphilic nature of TBHP, having both
hydrophilic and hydrophobic parts, which can solvate
both aqueous and organic phases. Even though three
phases (aqueous–organic–solid) are present in the H2O/
TBHP-cyclododecene/solid reaction system, only a mono-
layer of H2O molecules adsorbs on to the hydrophilic part
of TBHP (based on the 10/5 mmol ratio of TBHP/H2O in
this system), where the three phases would become a emul-
sion having small water molecule droplets covered with
the hydrophilic part of TBHP (Scheme 1). Since hydro-
gen-bondings among H2O molecules and electrostatic
attractions among organic molecules are quite strong in
comparison with van der Waals forces, the introduction
of small amount of H2O is thought to be sufficient to sep-
arate the aqueous phase from the organic phase. By the
complete separation of two phases, TBHP and the hydro-
phobic catalyst become readily available in the oxidation
of cyclododecene. This hypothesis would be fairly reason-
able since the catalytic enhancement of the TBHP/cyclod-
odecene system is observed only over highly hydrophobic
catalysts.
However, the addition of relatively large amount of
water (10 ml) slightly lowers the catalytic activity compared
to the case of addition of 2 ml water. This decrease may be
related to the solubility of TBHP in H2O (15 wt.%); a
large amount of H2O dissolve considerable amount of
TBHP to lower the amount of TBHP in the organic phase.
 N. Igarashi et al. / Microporous and Mesoporous Materials 104 (2007) 269–280 277

Table 6
Oxidation of cyclododecene with TBHP over non-silylated and silylated Ti-MCM-41 and -48 materials in triphase system
Catalyst Temperature (K) Length (h) Conv. (mol%max) TONa Selectivity (%) Total H2O (ml)
OH
O
OH OH
O OH
trans cis cis
trans
b
Cyclododecene oxidation
Ti-MCM-41 323 3 4.9 37(12) 21 43 36 0 0.2
Ti-MCM-41 323 3 2.3 18(6) 32 48 21 0 0.4
Ti-MCM-48 323 3 4.8 28(9) 27 44 29 0 0.2
Ti-MCM-48 323 3 3.2 19(6) 25 50 25 0 0.4
Ti-MCM-41-sil4 323 3 22 256(85) 17 83 0 0 0.2
Ti-MCM-41-sil4 323 3 46 552(184) 19 81 0 0 0.4
Ti-MCM-41-sil4 343 6 95 1130(188) 32 66 2 0 0.4
Ti-MCM-48-sil2 323 3 16 135(45) 19 76 5 0 0.2
Ti-MCM-48-sil2 343 6 66 546(91) 20 78 2 0 10.2
Ti-MCM-48-sil4 323 3 22 210(70) 17 83 0 0 0.2
Ti-MCM-48-sil4 323 3 46 442(147) 19 81 0 0 0.4
Ti-MCM-48-sil4 343 3 86 822(274) 19 80 1 0 0.4
Ti-MCM-48-sil4 343 3 81 769(256) 18 81 1 0 10.2
Ti-MCM-48-sil4 343 6 82 784(131) 18 80 2 0 0.2
Ti-MCM-48-sil4 343 6 98 933(155) 18 80 2 0 0.4
Ti-MCM-48-sil4 343 6 89 849(141) 18 79 3 0 10.2
Ti-MCM-48-sil4c 343 6 39 1465(244) 18 79 3 0 10.2
Numbers in parentheses denote TOF (mol(mol-Ti h)1).
a
(mol(mol-Ti)1).
b
Catalyst 50 mg, substrate 25 mmol, TBHP 5 mmol.
c
Substrate 20 mmol, TBHP 20 mmol.

vents are probably ascribed to the low miscibility of

Fig. 6. Effect of the amount of H2O present in the oxidation of
cyclododecene with TBHP over (a) Ti-MCM-48, 323 K, 3 h; (b) Ti-
MCM-48-sil4, 323 K, 3 h; (c) Ti-MCM-48-sil4, 343 K, 3 h.; (d) Ti-MCM-
48-sil4, 343 K, 6 h.
Catalytic activity for the oxidation of a-terpineol is
examined over the catalysts using H2O2 as oxidants in sol-
vents such as CDCl3 and CH3CN (Table 7). The catalytic
activity in CDCl3 is slightly improved by trimethylsilyla-
tion, even though the conversion is not so high (8–13%
conversion). When TBHP is used as oxidant in CDCl3,
the conversion increases to 24% by trimethylsilylation. It
is to be noted that the product 2 is obtained by the intra-
molecular cyclization of the product 1 probably promoted
by weakly acidic Ti species; concurrently the product 3 is
formed by the acid catalyzed hydrolysis of product 1.
When CH3CN is used in place of CDCl3, the overall cata-
lytic activity increases. Such differences between two sol-
CDCl3 with the aqueous phase containing oxidant H2O2,
resulting in complete separation of aqueous and organic
phases to diminish chances for contact of activated Ti-spe-
cies with organic reactant a-terpineol molecules for non-
silylated catalysts, and to limit chances for Ti-species to
be activated by H2O2 for silylated catalysts. The selectivity
to product 2 is suppressed in the presence of CH3CN com-
pared to that in the presence of CDCl3 probably due to the
nature of solvent as described above.
3.1.2. Stability during liquid phase oxidation
The recovered catalysts were repeatedly used to evaluate
the structural and catalytic stability of the catalysts during
cyclohexene oxidation. The results of the first and the sec-
ond batch reactions over silylated and non-silylated Ti-
MCM-41 samples are summarized in Table 8. The XRD
peak intensities of 100 reflection are practically unchanged
for both non-silylated and silylated samples during the
repeated runs of oxidation, indicating retention of their
structural ordering. In contrast, for the non-silylated sam-
ple the Ti amount is significantly decreased after the first
run of oxidation (from 123 to 576 as the Si/Ti ratio), while
the silylated sample shows a mitigated Ti-leaching (from
139 to 177 as the Si/Ti ratio). Although the catalytic activ-
ity of the silylated sample is reduced by half (from 24.7%
to 12.9% conv.) in the second run, it gives much higher
conversion than the non-silylated sample. This drop of
conversion might be partly due to the adsorption of
278 N. Igarashi et al. / Microporous and Mesoporous Materials 104 (2007) 269–280

emulsion separated phases

hydrophobic catalyst organic phase

=reactant
(cyclododecene)

+H2O

H2O

H2O TBHP
t-Bu -OOH
hydrophobic hydrophilic

Scheme 1.

Table 7
Oxidation of a-terpineol with H2O2 over non-silylated and silylated Ti-MCM-41 and -48 materials in the solvent system
Catalyst Conv. (mol%max) TONa Selectivity (%) H2O2 decomposition (%)
OH
O
O OH OH

OH OH OH
a-Terpineol oxidation 1 2 3
(CDCl3 as solvent)
Ti-MCM-41b 2 17(3) 44 45 10 46
Ti-MCM-41-sil1b 8 82(16) 14 32 54 18
Ti-MCM-48b 5 32(6) 31 54 10 32
Ti-MCM-48-sil1b 10 77(15) 19 27 55 21
Ti-MCM-48c 4 2(0.7) 17 83 0 100
Ti-MCM-48-sil1c 13 9(3) 0 100 0 93
Ti-MCM-48d 8 51(10) 95 5 0 –
Ti-MCM-48-sil1d 24 185(37) 85 15 0 –
(CH3CN as solvent)e
Ti-MCM-41 49 37(12) 57 43 0 37
Ti-MCM-41-sil1 21 20(7) 56 44 0 44
Ti-MCM-48 48 28(9) 46 54 0 40
Ti-MCM-48-sil1 45 32(11) 43 57 0 49
Numbers in parentheses denote TOF (mol(mol-Ti h)1).
a
(mol(mol-Ti)1).
b
Catalyst 50 mg, substrate 18.5 mmol, H2O2 5.5 mmol, CDCl3 5.4 ml, 328 K, 5 h.
c
Catalyst 0.1 g, substrate 5 mmol, H2O2 1 mmol, CDCl3 10 ml, 333 K, 3 h.
d
Catalyst 50 mg, substrate 18.5 mmol, TBHP 5.5 mmol, CDCl3 5.4 ml, 328 K, 5 h.
e
Catalyst 0.1 g, substrate 5 mmol, H2O2 1 mmol, CH3CN 10 ml, 333 K, 3 h.

products on the catalytically active centers, as is often
observed with titanosilicate zeolites systems. Non-produc-
tive decomposition of H2O2 remained suppressed for the
silylated sample.
UV–VIS spectra taken before and after the oxidation
reactions confirm that titanium leaching has been signifi-
cantly suppressed by trimethylsilylation, although the peak
around 220 nm has shifted to 250 nm (Fig. 7). This indi-
cates the change in the coordination of a part of Ti species
from tetrahedral to octahedral probably due to the hydra-
tion of the active sites during the oxidation [37,39]. 29Si
NMR spectra of silylated Ti-MCM-41 before and after oxi-
dation reaction confirm the stability of trimethylsilyl
groups during oxidation (not shown). These facts prove
 N. Igarashi et al. / Microporous and Mesoporous Materials 104 (2007) 269–280 279

Table 8 parameters are calculated according to the following

Structural stabilty of non-silylated and silylated Ti-MCM-41 samples equations:
before and after oxidation of cyclohexene with H2O2a
Catalystb Conv. (mol% H2O2 Peak intensity Si/Ti N TMS ðmolecules nm2 Þ
of max) decomposition (1 0 0) reflection ratio
(%) 6:023  105  ðI TMS Þ
¼ ð1Þ
Cyclohexene oxidation 60  I Q4 þ 69  I Q3 þ 78  I Q2  BETS:A:
Ti-MCM-41- 0.49 36.5 100 123c N OH ðmolecules nm2 Þ
Oxi1
Ti-MCM-41- 0.23 43.8 97.8 576d 6:023  105  ð2  I Q2 þ I Q3 Þ
Oxi2 ¼ ð2Þ
ð60  I Q4 þ 69  I Q3 þ 78  I Q2 Þ  BETS:A:
Ti-MCM-41- 24.7 0 85.3 139c
sil-Oxi1
Ti-MCM-41- 12.9 0 81.9 177d where NTMS and NOH represent the number of trimethylsi-
sil-Oxi2 lyl groups and silanol groups per nm2, respectively, and Ii
a
Silylating conditions for Ti-MCM-41(TMA) sample: HMDS, evacu- represents the relative peak area of the corresponding sili-
atioin at 473 K, [TMS] = 19% of total Si, [Q3] = 7, [Q4] = 93, [+DQ4] con species in the 29Si MAS NMR spectra. It is reported
(= [DQ3]) = 14. that the largest surface coverage of silica surface by the
b
Oxi-1, first run of oxidation; Oxi-2, second run of oxidation.
c trimethylsilyl group based on its molecular size calculation
Si/Ti ratio obtained from ICP analysis before first run.
d
After second run of oxidation. is 2.5 molecule nm2 [30,31]. The maximum numbers of
NTMS obtained for Ti-MCM-41 and -48 are 2.35 and
2.06 molecule nm2, corresponding to 94% and 82% of
the possible coverage, respectively.
The factors influencing the catalytic performance in the
liquid-phase oxidation of organic substrates would be the
following: (i) the amount of trimethylsilyl groups giving
both lipophilicity to encourage the access of organic
substrates and hydrophobicity to lower the H2O2 decom-
position, (ii) the amount of silanol groups giving hydrophi-
licity to encourage the access of oxidant H2O2, and (iii) the
amount and the structure of Ti-species. The hydrophobic-
ity expressed in term of water adsorption capacity should
relate to the amounts of both trimethylsilyl and silanol
Fig. 7. UV–VIS spectra of (a) non-silylated and (b) silylated Ti-MCM-41
samples before and after oxidation of cyclohexene with H2O2.
groups. Therefore, the water adsorption density, the
amount of water adsorption capacity per nm2 (N H2 O ), is
plotted against the difference between NOH and NTMS
that trimethylsilyl groups remain intact during the catalytic (molecule nm2) as shown in Fig. 8a. A clear relation-
oxidation. ship can be seen between the two parameters where
the water adsorption density (N H2 O ) decreases drastically
3.2. Further insights into surface chemistry with decreasing difference in NOH and NTMS, demonstrat-
ing the strong water repelling effect of trimethylsilyl
Both the degree of trimethylsilylation and the amount of groups.
surface silanol groups are important parameters that affect The hydrophobicity should be also important in the
the surface properties of the catalyst samples. These two liquid-phase organic oxidation system, since a supply of

Fig. 8. Comparison of the relation between water adsorption capacity per nm2 (N H2 O ), TON (mol(mol-Ti)1), and H2O2 decomposition, and the difference
between SiOH and TMS per nm2 for the non-silylated and silylated Ti-MCM-41 and -48 samples in the oxidation of cyclohexene in a triphase system.
280 N. Igarashi et al. / Microporous and Mesoporous Materials 104 (2007) 269–280
organic substrates onto the catalytically active sites should
be the rate-determining factor. Fig. 8b shows a clear rela-
tionship between TON (mol(mol-Ti)1) and N H2 O (mole-
cule nm2), demonstrating the strong promoting effect of
hydrophobicity as a result of trimethylsilylation. This
result suggests the importance of lipophilicity/hydropho-
bicity for the high catalytic performance. The H2O2 decom-
position is suppressed for the highly silylated samples with
low water adsorption density (Fig. 8c).
4. Conclusions
Ti-MCM-41 and Ti-MCM-48 materials have been trim-
ethylsilylated by employing various silylating reagents, and
thus silylated materials have proved to show high catalytic
activities in various epoxidation reactions using H2O2 as an
oxidant compared to the non-silylated samples. The highly
reactive HMDZ reacts with hydrogen-bonded silanol
groups in addition to the isolated ones as confirmed by
IR spectra. These highly silylated materials show remark-
ably high catalytic activities in the oxidation of cyclodode-
cene with TBHP especially when a small amount of H2O is
introduced into the reaction system. The effect of steric hin-
drance caused by the high degree of trimethylsilylation was
not observed, while the chances for Ti-species to be acti-
vated by H2O2 was slightly limited. The trimethylsilylated
sample can be reused as a catalyst even though the peak
around 220 nm has shifted to around 250 nm which may
be due to a part of active Ti species has changed their coor-
dination form tetrahedral to octahedral due to the hydra-
tion during oxidation.
Acknowledgement
This work was partly supported by Core Research for
Evolutional Science and Technology (CREST) of JST Cor-
poration to T.T.
References
[1] B. Notari, Adv. Catal. 41 (1996) 253.
[2] T. Tatsumi, M. Nakamura, S. Negishi, H. Tominaga, Chem.
Commun. (1990) 476.
[3] A. Corma, M.A. Camblor, P. Esteve, A. Martı́nez, J. Pérez-Pariente,
J. Catal. 145 (1994) 151.
[4] A. Corma, P. Esteve, A. Martı́nez, S. Valencia, J. Catal. 152 (1995)
18.
[5] T. Tatsumi, N. Jappar, J. Phys. Chem. B 102 (1998) 7126.
[6] P. Wu, T. Tatsumi, T. Komatsu, T. Yashima, J. Phys. Chem. B 105
(2001) 2897.
[7] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, J.S. Beck,
Nature 359 (1992) 710.
[8] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge,
K.D. Schmitt, C.T-U. Chu, D.H. Olson, E.W. Sheppard, S.B.
McCullen, J.B. Higgins, J.L. Schlenker, J. Am. Chem. Soc. 114
(1992) 10834.
[9] P.T. Tanev, M. Chibwe, T.J. Pinnavia, Nature 368 (1994) 321.
[10] T. Blasco, A. Corma, M.T. Navarro, J.P. Pariente, J. Catal. 156
(1995) 65.
[11] K.A. Koyano, T. Tatsumi, Micropor. Mater. 10 (1997) 259.
[12] T. Tatsumi, K.A. Koyano, N. Igarashi, Chem. Commun. (1998) 325.
[13] N. Igarashi, S. Kidani, Rizwan-Ahemaito, T. Tatsumi, Stud. Surf.
Sci. Catal. 129 (2000) 163.
[14] N. Igarashi, S. Kidani, Rizwan-Ahemaito, T. Tatsumi, Zeol. Mater.
1 (2000) 15.
[15] A. Bhaumik, T. Tatsumi, J. Catal. 189 (2000) 31.
[16] H. Nur, S. Ikeda, B. Ohtani, Chem. Commun. (2000) 2235.
[17] H. Nur, S. Ikeda, B. Ohtani, J. Catal. 204 (2001) 402.
[18] L.R. Snyder, J.W. Ward, J. Phys. Chem. 70 (1966) 3941.
[19] M.L. Hair, W. Hertl, J. Phys. Chem. 73 (1969) 2372.
[20] W. Hertl, M.L. Hair, J. Phys. Chem. 75 (1971) 2181.
[21] A.C. Zettlemoyer, H.H. Hsing, J. Coll. Interface Sci. 58 (1977) 263.
[22] E.W. Sindorf, G.E. Maciel, J. Phys. Chem. 86 (1982) 5208.
[23] D.W. Sindorf, G.E. Maciel, J. Phys. Chem. 87 (1983) 5516.
[24] B.A. Morrow, A.J. McFarlan, Langmuir 7 (1991) 1695.
[25] S. Haukka, A. Root, J. Phys. Chem. 98 (1994) 1695.
[26] E.F. Vansant, P. Van Der Voort, K.C. Vroancken, Stud. Surf. Sci.
Cat. 93 (1995) 83.
[27] J. Chen, Q. Li, R. Xu, F. Xiao, Angew. Chem. Int. Ed. Engl. 34
(1995) 2694.
[28] R. Anwander, C. Palm, J. Stelzer, O. Groeger, G. Engelhardt, Stud.
Surf. Sci. Cat. 117 (1998) 135.
[29] A. Corma, M. Domine, J.A. Gaona, J.L. Jordá, M.T. Navarro, F.
Rey, J. Pérez-Pariente, J. Tsuji, B. McCulloch, L.T. Nemeth, Chem.
Commun. (1998) 2211.
[30] X.S. Zhao, G.Q. Lu, A.K. Whittaker, G.J. Millar, H.Y. Zhu, J. Phys.
Chem. B 101 (1997) 6525.
[31] X.S. Zhao, G.Q. Lu, J. Phys. Chem. B 102 (1998) 1556.
[32] J. Bu, H.-K. Rhee, Cat. Lett. 65 (2000) 141.
[33] J. Bu, H.-K. Rhee, Cat. Lett. 66 (2000) 245.
[34] R. Anwander, I. Nagl, M. Widenmeyer, J. Phys. Chem. B 104 (2000)
3532.
[35] J.N. Kinkel, K.K. Unger, J. Chromatogr. 316 (1984) 193.
[36] F. van Laar, D.E. De Vos, D.L. Vanoppen, P.A. Jacobs, 12th
International Zeolite Conference, Stud. Surf. Sci. Catal. (1999) 1213.
[37] H. Kochkar, F. Figueras, J. Catal. 171 (1997) 420.
[38] F. Figueras, H. Kochkar, Catal. Lett. 59 (1999) 79.
[39] D.C.M. Dutoit, M. Schneider, A. Baiker, J. Catal. 153 (1995) 165.