Silylation of titanium-containing amorphous silica catalyst: Effect on the alkenes epoxidation with H2O2

Article abstract of DOI:10.1016/j.cattod.2010.07.022

The surface of a Ti/SiO₂ catalyst was silylated using hexamethyldisilazane (HMDS) and tetramethyldisilazane (TMDS) as silylating reagents in vapor phase. The silylation of silanol (Si-OH) on the catalysts was confirmed by diffuse reflectance UV-vis, DRIFT spectroscopy and solid-state ²⁹Si MAS NMR techniques. Titanium-containing mesoporous amorphous silica has been successfully modified by trimethylsilylation enhances its hydrophobicity producing an increase in the hydrogen peroxide efficiency and in selectivity to epoxide in the epoxidation of alkenes with H₂O ₂. These effects are more evident when the hydrogen peroxide concentration is higher. As the results show the silylation degree reached is higher when TMDS is employed, showing the catalysts silylated with TMDS a better hydrogen peroxide efficiency and selectivity to epoxide.

Full text of DOI:10.1016/j.cattod.2010.07.022

Catalysis Today 158 (2010) 103–108
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Silylation of titanium-containing amorphous silica catalyst:
Effect on the alkenes epoxidation with H2O2
M.C. Capel-Sanchez∗, J.M. Campos-Martin, J.L.G. Fierro
Sustainable Energy and Chemistry Group (EQS), Instituto de Catálisis y Petroleoquímica, CSIC, Marie Curie, 2, Cantoblanco, 28049 Madrid, Spain
a r t i c l e i n f o
a b s t r a c t
Keywords:
Silylation
The surface of a Ti/SiO2 catalyst was silylated using hexamethyldisilazane (HMDS) and tetramethyldisi-
lazane (TMDS) as silylating reagents in vapor phase. The silylation of silanol (Si–OH) on the catalysts was
confirmed by diffuse reflectance UV–vis, DRIFT spectroscopy and solid-state 29Si MAS NMR techniques.
Titanium-containing mesoporous amorphous silica has been successfully modified by trimethylsilylation
enhances its hydrophobicity producing an increase in the hydrogen peroxide efficiency and in selectivity
to epoxide in the epoxidation of alkenes with H2O2. These effects are more evident when the hydro-
gen peroxide concentration is higher. As the results show the silylation degree reached is higher when
TMDS is employed, showing the catalysts silylated with TMDS a better hydrogen peroxide efficiency and
selectivity to epoxide.
Ti/SiO2 catalysts
1-Octene
Cyclohexene
Epoxidation
Hydrogen peroxide
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
oxidant that yields an environmentally benign by-product (H2O)
[10]. Therefore, an enhancement of the hydrophobicity is consid-
Despite numerous reports in the literature, the epoxida-
tion of terminal alkenes remains a challenge in petrochemistry.
Many different methods have been developed for the prepara-
tion of epoxides. A large body of work has been generated on
titanium-substituted zeolites in the framework, including TS-1,
Ti-␤, Ti-incorporated ordered mesoporous silica and amorphous
xerogels, which have been extensively investigated [1,2]. An inter-
esting alternative is the use of Ti/SiO2 catalysts [3–7] because
these catalysts are relatively inexpensive, easy to synthesize and
regenerate, and at the same time show high conversion rates.
Notwithstanding, the activity and selectivity of mesoporous Ti-
containing materials reported so far is not as high as expected in
oxidizing relatively less bulky organic substrates compared to that
exhibited by TS-1.
The poor performance of Ti(IV) sites supported on meso-
porous silicas in the epoxidation of alkenes with aqueous hydrogen
peroxide has been attributed to the chemical properties of the
support surface [8,9]. The catalyst surface is hydrophilic, and
the Ti(IV) site is presumably deactivated by the competitive
binding of water molecules, which hinder formation of the key
titanium–hydroperoxide intermediate. There is considerable inter-
est in this problem, since aqueous hydrogen peroxide is a desirable
ered important to improve the activity of Ti/SiO2 and retard the
Ti-leaching in liquid phase oxidation. Moreover, water removal
from the catalyst surface would also reduce the hydrolysis of
the epoxide product and, therefore, would be beneficial for
the selectivity of the reaction. An increased hydrophobicity of
the catalyst surface is also expected to favour the diffusion
of the rather apolar alkene substrates inside the pores of the
catalyst [11]. The hydrophobicity of mesoporous silica should
be increased by removing the unnecessary hydroxyl groups or
replacing them by other groups, which will facilitate the conden-
sation of organic compounds in the pores. A common practice
is to cover the surface with a layer of a silylant agent that is
selectively hydrolyzed by these hydroxyl groups by covalently
bonding organic groups to the inorganic siliceous framework,
i.e. substitution of Si–OH by Si–O–SiR (R = organic chain) [12,13].
Silylation of the catalyst surface is an outstanding procedure
for improving both the activity and selectivity of epoxidation
reactions [14,15]. Several silylating agents have been used and
hexamethyldisilazane has proven to be one of the most efficient
[12,16].
In the present work, we investigate a new silylating agent
1,1,3,3-tetramethyldisilazane (TMDS), to modify Ti/SiO2 catalysts.
The silylation efficiency of this new agent was compared to
1,1,1,3,3,3-hexamethyldisilazane (HMDS). Different levels of sily-
lation were obtained by varying the reaction time. With this aim we
developed a simple method for the silylation of a Ti/SiO2 catalyst.
It was also evaluated the effect that the degree of silylation pro-
∗
Corresponding author. Tel.: +34 91 585 4877; fax: +34 91 585 4760.
E-mail address: mcapel@icp.csic.es (M.C. Capel-Sanchez).
URL: http://www.icp.csic.es/eqs/ (M.C. Capel-Sanchez).
0920-5861/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2010.07.022

104
M.C. Capel-Sanchez et al. / Catalysis Today 158 (2010) 103–108
duces on the activity and the selectivity in the 1-octene epoxidation
reaction with hydrogen peroxide.
2. Experimental methods
Catalysts were prepared by the procedure already reported
in the literature [7]. Titanium (triethanolaminate) isopropoxide
(TYZOR® TE) (2.0 mmol) was dispersed in 2-propanol (25 ml), the
solution was heated to 353 K under stirring and then 5 g of silica
(Grace Davison, XPO 2407) were added and the suspension was
stirred for 2 h. The solid was filtered out and washed twice with
25 ml of 2-propanol, dried at 383 K, and finally calcined at 773 K
for 5 h. Two silylant reagents: 1,1,1,3,3,3-hexamethyldisilazane
(HMDS) and 1,1,3,3-tetramethyldisilazane (TMDS) were used for
the silylation of samples. The procedure was as follows: the
silylant reagent fed continuously by a syringe pump to a con-
tinuous flow of N2 on the sample bed with a temperature of
473 K for 2 or 16 h, then a nitrogen flow was fed for 2 h. The
silylation reagent/catalyst weight ratio was of 0.23 and 1.60,
respectively.
Scheme 1. Silylation scheme with HMDS.
Scheme 2. Silylation scheme with TMDS.
3. Results and discussion
The titanium content of the Ti/SiO2 of the catalysts was deter-
mined using inductively coupled plasma absorption spectrometry,
with a Perkin-Elmer Optima 3300 DV instrument. The amount of
titanium in the Ti/SiO2 catalyst prepared was approximately 1 wt.%.
Elemental analyses (CHN) were performed with a LECO CHNS-932
equipment.
3.1. Catalysts characterization
The reaction between Ti/SiO2 silanol groups and silylating agent
produces NH3 according to Scheme 1 with HMDS and Scheme 2
with TMDS.
Textural properties were determined from the adsorption–
The corresponding specific BET surface areas, pore diameters
and pore volumes of the catalysts are shown in Table 1. The sily-
lation process brought about a slight decrease in the nitrogen
adsorption capacity, reflected in a decrease in the values of specific
areas of all the silylated catalysts. Thus, the value of specific BET
surface area of the reference catalyst – 213 m2 g−1 – decreased to
approximately 193 m2 g−1 in silylated catalysts. A parallel decrease
in pore volume was also observed. The reduction in pore volume
appears to be related to the texture of the reference catalyst, since
coating of the silica particles results in an increase in the size of
the agglomerates or packed particles, eliciting a decrease in the
adsorption capacity of the silylated samples. In addition, pore size
distribution, as determined by applying the BJH model to the des-
orption branch of the nitrogen adsorption–desorption isotherms,
also underwent some changes. The pore sizes of the silylated sam-
ples decreased slightly with respect to the reference catalyst. All
the silylated samples displayed similar decrease in both the specific
surface area and in the pore size. According to the homogeneous
pore size distributions and the pore size of the samples after sily-
lating of the catalysts the ammonia does not damage the porous
structure of the catalyst.
desorption isotherms of nitrogen recorded at 77 K with
a
Micromeretics TriStar 3000. Specific area was calculated by apply-
ing the BET method to the relative pressure (P/P0) range of the
isotherms between 0.03 and 0.3, taking a value of 0.162 nm2 for the
cross-section of adsorbed nitrogen molecule at 77 K. Pore size dis-
tributions were computed by applying the Barrett–Joyner–Halenda
(BJH) model to the desorption branch of the nitrogen adsorption-
desorption isotherms.
Ultraviolet–visible spectra were measured on a Varian Carry
5000 UV–vis spectrophotometer equipped with an integrating
sphere. A BaSO4 disc-sample was used as reference. All spec-
tra were acquired under ambient conditions. Diffuse reflectance
infrared Fourier transform spectra (DRIFTS) were recorded at room
temperature on a Jasco 6300 FTIR spectrophotometer. The instru-
ment incorporated an integration sphere and a Harrick HVC-DRP
environmentally-controlled cell. About 50 mg of the powdered
sample was packed into a sample holder and dried in situ at 423,
523 and 623 K for 1 h while a helium flow (Air Liquid) was passed
through the sample. A DRIFT spectrum of dry KBr was also recorded
as background. For each sample, 100 scans were accumulated at a
spectral resolution of 4 cm−1. Solid-state 29Si CP-MAS NMR with
1H decoupling measurements were recorded on a Bruker AMX 300
spectrometer. The dried powdered samples were loaded into a 4-
mm multinuclear probe BL4 X/Y/1H and spun at 10 kHz according
to the following protocol: ␲/2 pulse, 7 ␮s; CP contact time 2 ms;
high power H-decoupling during detection; a repetition delay of
30 s; and 2000 scans. An internal reference of the spectrometer was
employed to calculate the chemical shifts.
The catalytic activity was performance in a typical run, a suspen-
sion of alkene (0.2 mol), tert-butanol (11 g) and 1 g of catalyst were
heated at 333 K, and then an organic solution of 5 wt.% of H2O2
(in tert-butanol) was added to the reaction vessel during 30 min.
Two hydrogen peroxide concentrations, referred to the whole final
reaction mixture, were tested: 0.6 and 1.2%. The H2O2/substrate
molar ratios were 0.01:0.2 for 0.6% of H2O2 and 0.02:0.2 for 1.2%
H2O2 concentration. The organic compounds were analyzed by GC-
FID (Agilent 6850, equipped with a HP-WAX capillary column).
The hydrogen peroxide was measured by standard iodometric
titration.
The electronic spectra of the samples showed an absorption
associated with the ligand metal-charge transfer (LMCT) from the
oxygen to an empty orbital of the Ti(IV) ion: Ti4+
O
2− → Ti3+O− typi-
cal of this type of catalysts. The wavelength at which this transition
occurs is highly sensitive to the coordination of titanium sites, and
accordingly it has been often considered as a probe to test tita-
nium coordination [17,18]. The UV–vis spectra show absorption
at wavelengths in the region 200–300 nm, which is character-
Table 1
Textural properties of the catalysts.
BET area
Pore diameter
(nm)
Pore volume
(m2 g−1
)
(cm3 g−1
)
Reference
HM 2
HM 16
TM 2
213
198
200
199
193
23.0
22.1
22.3
22.4
22.1
1.37
1.25
1.28
1.26
1.24
TM 16

M.C. Capel-Sanchez et al. / Catalysis Today 158 (2010) 103–108
105
Fig. 1. Diffuse reflectance UV–vis spectra of the samples.
istic of this type of catalyst. Two components can be identified
in this absorption band: the most intense one is placed around
210 nm, typical of isolated Ti(IV) ions in a tetrahedral environ-
ment of oxide ions [17], and a second, broader band located at
somewhat higher wavelengths (ca. 250 nm), originated from tetra-
hedral titanium species coordinated with water molecules [17,18].
It has been shown that the component at ca. 250 nm disappears
upon sample calcination thus avoiding interference of these iso-
lated penta-coordinated Ti centers [7]. Both titanium species are
the active sites of Ti/SiO2 catalysts for oxidation reactions. In addi-
tion, the presence of small Ti–O–Ti oligomers (275–300) [18] or
small TiO2 cluster (370–410 nm) [19] is precluded.
Fig. 2. DIRFT spectra in the OH vibration region of the samples degassed at different
temperatures.
UV–vis spectra of the two silylated samples (TMDS and HMDS)
and the corresponding non-silylated counterpart are included in
Fig. 1. It is emphasized here that the UV–vis spectra were recorded
under ambient conditions, thus the amount of adsorbed water on
the surface depends on the hydrophobic/hydrophilic character of
the surface. Accordingly, the intensity of the band at 250 nm is
related with the hydrophilic character of the surface. These obser-
vations cannot be made if the UV–vis spectra were recorded under
dry atmosphere. At a first glance comparison of these spectra
shows no significant differences although a careful scrutiny of the
250–275 nm energy region, where the adsorption of tetrahedral
species of titanium coordinated with water usually appears [17,18],
reveals some differences. The intensity of the absorption bands of
silylated samples is somewhat lower than the non-silylated one
because its hydrophilic character has been inhibited to some extent
by the chemically grafted surface layer by TMDS and HMDS sily-
lant agents, thus reducing the affinity for water molecules [17,18].
It is also observed that the silylating degree is somewhat higher
in the catalyst silylated with TMDS. This observation is consis-
tent with the sterical hindrance of silyl groups, which makes the
silylation efficiency to be higher for TMDS than for HMDS [20].
Besides, as the samples silylated for 2 and 16 h displayed essentially
the same electronic spectra, it is inferred that silylation is a quick
process.
DRIFT spectra of non-silylated, silylated samples and the bare
silica, used to prepare the samples, degassed at room temperature
and at various pre-treatment temperatures are presented in Fig. 2.
All the spectra at room temperature show a broad band in the OH
stretching vibration energy region (3600–2500 cm−1) associated
with water molecules interacting with surface OH groups [4,16]. An
increase in the degassing temperature leads to the removal of water
molecules adsorbed on the surface (Fig. 2) [16]. Adsorbed water
was fully removed at 623 K. DRIFT spectra of the samples degassed
at 623 K are displayed in Fig. 3. The reference sample shows a
band at about 3700 cm−1, which belongs to the stretching vibration
mode of terminal or geminal silanol group. This band disappears
in silylated samples, which is a clear indication of the chemical
reaction between the organosilane reagent and the terminal and
geminal silanol groups [12]. The broad absorption band between
3600–3400 cm−1 corresponds to H-bonded hydroxyl groups. For
the HM sample, the residual geminal silanols still remaining after
silylation most likely originates from partial silylation of an aggre-
gate of hydrogen-bonded silanol groups. The reaction of HMDS with
these residual silanol groups may be sterically hindered by the pres-
ence of neighbouring trimethylsilyl groups [20]. The DRIFT spectra
of the silylated catalysts show the typical stretching vibration
bands of C–H bonds between 3000 and 2800 cm−1 corresponding to

106
M.C. Capel-Sanchez et al. / Catalysis Today 158 (2010) 103–108
Fig. 3. DIRFT spectra in the of OH vibration region of the samples dried at 623 K.
the –CH3 groups formed during the silylation process [16]. The sam-
ple silylated with TMDS shows an additional peak at 2200 cm−1
due to the presence of Si–H bonds. As already shown by the elec-
tronic spectra, no differences between the DRIFT spectra of the
samples silylated for 2 and 16 h were observed.
Solid-state 29Si MAS NMR spectra (Fig. 4) confirmed the pres-
ence of –SiCH3 groups bound to the surface of the silylated samples.
Distinct resonances can be clearly distinguished: Q4 and Q3 are
responsible for Si bonded to 4 siloxane [(–OSi)4] and three siloxane
groups and one hydroxyl group [(–SiO)3SiOH], respectively [11].
For the HM silylated sample, a new signal was detected at 9.5 ppm,
assigned to (CH3)3Si–O [20], whereas for TM sample this new signal
is placed at −14 ppm and assigned to (CH3)2HSi–O [20]. These spec-
tral features clearly show that the silylant agent reacts with surface
silanol groups, which makes the catalyst surface hydrophobic. The
intensities of Q3 and Q4 components were estimated by calculating
the integral of each peak and then an estimate of the extent of sam-
ple silylation was made by calculating the ratio between the peak
area of Q3 and Q4 components (Table 2). The relative area of Q3 and
Q4 components depend on the silylating treatment. The decrease
of the relative area of Q3 component in silylated samples with
respect the non-silylated one is a consequence of the disappearance
of surface Si–OH groups by chemical reaction with silylant agent
to form (CH3)2HSi–O–Si or (CH3)3Si–O–Si moieties. Such a drop is
more evident for the samples silylated with TMDS. Indeed, these
observations have been confirmed by elemental analysis (Table 2).
After silylation, –CH3 groups are present on the catalyst surface and
Fig. 4. 29Si CP-MAS NMR spectra of samples under ambient conditions.
accordingly the surface density of these groups depends on the sily-
lation degree. The sample silylated with TMDS has a higher carbon
amount than that silylated with HMDS; this difference is even more
pronounced considering that HMDS incorporates three C-atoms
while TMDS incorporates only two C-atoms at each silylation cen-
tre. This result confirms that silylation is more effective by using
TMDS instead HMDS. The silylation efficiency decreased accord-
ing to the sterical hindrance of the silyl groups which is higher for
HMDS than for TMDS [20].
3.2. Catalytic activity
The two silylated catalysts and the non-silylated counterpart
were tested in the liquid phase epoxidation reaction of two alkenes
using hydrogen peroxide as oxidant. For this purpose two olefinic
substrates were selected: 1-octene and cyclohexene. As a primary
alkene, 1-octene is difficult to epoxidize but the epoxide is rela-
tively stable (1-octene), however cyclohexene is relatively easy to
epoxidize but its epoxide is more reactive and consequently it can
be hydrolyzed to the corresponding glycol.
The 1-octene epoxidation was performed using two hydrogen
peroxide concentrations, referred to the whole reaction mixture
(0.6 and 1.2 wt.%) by adding dropwise the oxidant solution during
the first 30 min of the reaction. The slow addition of H2O2 is crucial
Table 2
Chemical analysis of the catalysts.
% C
% H
% N
Q3/Q4
Reference
HM
TM
–
1.85
2.20
–
0.73
0.86
–
0.03
0.07
0.081
0.045
0.039

M.C. Capel-Sanchez et al. / Catalysis Today 158 (2010) 103–108
107
Fig. 6. Hydrogen peroxide efficiency after 1 h of reaction (T = 353 K) in the epoxida-
tion of cyclohexene with two hydrogen peroxide concentrations.
Fig. 5. Hydrogen peroxide efficiency after 1 h of reaction (T = 353 K) in the epoxida-
tion of 1-octene with two hydrogen peroxide concentrations.
Along the epoxidation reaction of cyclohexene two by-products
were observed. Accordingly, the effect of catalyst silylation can be
studied as a function of the amounts of hydrogen peroxide added
to the reactor. In general, the selectivity to epoxide decreased with
increasing concentration of hydrogen peroxide (Fig. 7). This selec-
tivity drop was less pronounced for the silylated sample than for
the non-silylated one.
to avoid, or at least minimize, the useless decomposition of the oxi-
dant and to keep the local water concentration as low as possible
[3–7,21]. The conversion was very high for all three samples and no
other products than 1-octene epoxide were detected. For the low-
est H2O2 concentration (0.6 wt.%) all samples reached very high
hydrogen peroxide efficiency (Fig. 5). At this low H2O2 concentra-
tion, there is not apparent improvement in catalyst performance.
However, when high H2O2 peroxide concentration is employed
(1.2 wt.%), some differences among samples are observed. For all
samples a small decrease in the hydrogen peroxide efficiency is
observed. The efficiency decrease for the non-silylated sample is
very important (90–55%), while the decrease in the silylated sam-
ples is clearly lower. In addition, some differences are observed
when comparing the two silylated samples. In general, the catalyst
silylated with TMDS yields higher efficiency than catalyst sily-
lated with HMDS. This trend is the same than that of the silylation
degree derived from UV–vis, DRIFT and 29Si MAS NMR spectro-
scopic results and also by the hydrogen peroxide efficiency in the
epoxidation of 1-octene. In addition, there is no influence of the
silylation time of the catalysts with TMDS on the performance in
the target reactions, suggesting that the silylation reaction is very
quick. On the contrary, the catalyst silylated with HMDS at differ-
ent times displayed some differences in the performance. These
results suggest that HMDS silylation is slower than with TMDS
as indicated by the improvement of the hydrogen peroxide effi-
ciency achieved with catalysts already silylated for longer periods
of time.
This decrease in the selectivity to the epoxide is related with
the amount of water present in the reaction mixture; as water is
the by-product of hydrogen peroxide reaction, the greater amount
of H2O2 added gives the higher final water concentration. The
water molecules adsorbed in the catalyst surface may react with
epoxide via epoxirane ring-opening reaction, leading to glycols
(diols) [21]. The epoxide ring-opening side reaction is catalyzed
mainly by the acidic titanium sites [22]. Changes in the surface
properties produced by silylation not only affect the acidity of
the surface but also improve the hydrophobicity of the catalyst
surface. This changes in the surface properties yield to a better
performance in the catalytic centre, because the hydrophobicity
of the surface repel water formed far from isolated tetrahedral
Ti(IV) centers. This effect produce an increase in the catalytic activ-
ity by keeping the adsorption site free of water and reducing the
oxirane group hydrolysis because the amount of water molecules
close to the Ti sites is lower. For these reasons, the extent of gly-
col formation is lower in silylated samples, and samples silylated
with TMDS yield higher epoxide selectivity than that silylated
with HMDS. The selectivity of catalyst silylated with HMDS was
found to increase with silylation time, that is, the better selec-
Similarly, the cyclohexene epoxidation was performed using
two hydrogen peroxide concentrations (0.6 and 1.2 wt.%) and the
oxidant solution was also dropwise added during the first 30 min
of the reaction. In this reaction, the desired product is cyclohex-
ene oxide and the potential by-products are 2-cyclohexen-1-ol
and 1,2-cyclohexanediol [21]. For all catalysts and H2O2 concen-
tration studied in this work the oxidant conversion was complete.
For the two hydrogen peroxide concentrations employed the oxi-
dant efficiency was very high (Fig. 6). For all samples an increase
in the amount of hydrogen peroxide added to the reaction pro-
duces an increase in H2O2 efficiency. Silylated samples show a
higher H2O2 efficiency and the one silylated with TMDS yields
higher efficiency than that silylated with HMDS. Similarly to 1-
octene epoxidation, some differences in the performance of the
catalyst silylated with HMDS at different times have been observed.
Again, these results indicate a close relationship between sily-
lation degree revealed by UV–vis, DRIFT and 29Si MAS NMR
techniques and hydrogen peroxide efficiency in the epoxidation of
cyclohexene.
Fig. 7. Selectivity after 1 h of reaction (T = 353 K) to different products in the epoxi-
dation of cyclohexene with hydrogen peroxide.

108
M.C. Capel-Sanchez et al. / Catalysis Today 158 (2010) 103–108
tivity was achieved with the catalysts silylated for the longest
time.
References
[1] D.E. De Vos, B.F. Sels, P.A. Jacobs, Adv. Synth. Catal. 345 (2003) 457.
[2] P. Ratnasamy, D. Srinivas, H. Knözinger, Adv. Catal. 48 (2004) 1.
[3] M.C. Capel-Sanchez, J.M. Campos-Martin, J.L.G. Fierro, M.P. De Frutos, A. Padilla
Polo, Chem. Commun. (2000) 855.
4. Conclusions
[4] M.C. Capel-Sanchez, J.M. Campos-Martin, J.L.G. Fierro, J. Catal. 217 (2003) 195.
[5] A. Campanella, M.A. Baltanas, M.C. Capel-Sanchez, J.M. Campos-Martin, J.L.G.
Fierro, Green Chem. 6 (2004) 330.
[6] M.C. Capel-Sanchez, J.M. Campos-Martin, J.L.G. Fierro, J. Catal. 234 (2005) 488.
[7] M.C. Capel-Sanchez, G. Blanco-Brieva, J.M. Campos-Martin, M.P. De Frutos, W.
Wen, J.A. Rodriguez, J.L.G. Fierro, Langmuir 25 (2009) 7148.
[8] A. Corma, P. Esteve, A. Martinez, J. Catal. 161 (1996) 11.
[9] T. Tatsumi, K.A. Koyano, N. Igarashi, Chem. Commun. (1998) 325.
[10] J.M Campos-Martin, G. Blanco-Brieva, J.L.G. Fierro, Angew. Chem. Int. Ed. 45
(2006) 6962.
[11] K.F. Lin, P.P. Pescarmona, K. Houthoofd, D.D. Liang, G. Van Tendeloo, P.A. Jacobs,
J. Catal. 263 (2009) 75.
[12] N.R.E.N. Impens, P. van der Voort, E.F. Vansant, Micropor. Mesopor. Mater. 28
(1999) 217.
[13] M.R. Prasad, M.S. Hamdy, G. Mul, E. Bouwman, E. Drent, J. Catal. 260 (2008) 288.
[14] R.L. Brutchey, D.A. Ruddy, L.K. Andersen, T.D. Tilley, Langmuir 21 (2005) 9576.
[15] N. Igarashi, S. Kidani, R. Ahemaito, K. Hashimoto, T. Tatsumi, Micropor. Meso-
por. Mater. 81 (2005) 97.
Vapor-phase silylation of Ti-loaded amorphous silica epoxida-
tion catalysts has been successfully achieved with TMDS and HMDS
agents. The UV–vis, DRIFT and 29Si CP-MAS NMR techniques pro-
vide useful information on the silylation degree of the catalysts.
The experimental results indicate that the silylating agent TMDS
improves the catalytic performance of Ti/SiO2 catalyst more sig-
nificantly than the previously reported agent, HMDS, and this is
mainly due to the steric bulk effects of the silylating groups, that is
to say because of the bulky trimethylsilyl groups from HMDS versus
dimethylsilyl groups from TMDS. The silylation degree with TMDS
was found to be higher than with HMDS and no changes in the sily-
lation degree with TMDS was observed for silylation times longer
than 2 h. On the contrary, the slower silylation degree observed
for HMDS implies necessarily long reaction times to achieve high
silylation degrees. The silylation of Ti/SiO2 catalysts led to changes
in their performance for the epoxidation of alkenes (1-octene and
cyclohexene) with dilute hydrogen peroxide solutions. Silylation
treatment enhanced catalyst hydrophobicity producing an increase
in the hydrogen peroxide efficiency and epoxide selectivity and
these effects were more evident by using higher hydrogen perox-
ide concentrations. The H2O2 efficiency and selectivity to epoxide
were higher for the catalysts silylated with TMDS, which in turn
displayed a higher silylation degree.
[16] M.C. Capel-Sanchez, L. Barrio, J.M. Campos-Martin, J.L.G. Fierro, J. Colloid Inter-
face Sci. 277 (2004) 146.
[17] V.A. De La Pe
n˜a O’Shea, M. Capel-Sanchez, G. Blanco-Brieva, J.M. Campos-
Martin, J.L.G. Fierro, Angew. Chem. Int. Ed. 42 (2003) 5851.
[18] M.C. Capel-Sanchez, V.A. De la Pen˜a-O’Shea, L. Barrio, J.M. Campos-Martin, J.L.G.
Fierro, Top. Catal. 41 (2006) 27.
[19] L. Marchese, E. Gianotti, V. Dellarocca, T. Maschmeyer, F. Rey, S. Coluccia, J.M.
Thomas, Phys. Chem. Chem. Phys. 1 (1999) 585.
[20] Y. Liang, R. Anwander, J. Mater. Chem. 17 (2007) 2506.
[21] M. Guidotti, C. Pirovano, N. Ravasio, B. Lazaro, J.M. Fraile, J.A. Mayoral, B. Coq,
A. Galarneau, Green Chem. 11 (2009) 1421.
[22] C.A. Muller, M.S. Schneider, T. Mallat, A. Baiker, J. Catal. 192 (2000) 448.