Lewis acid-containing mesoporous molecular sieves as solid efficient catalysts for solvent-free Mukaiyama-type aldol condensation

Article abstract of DOI:10.1016/j.jcat.2005.04.035

The activity of the titanium mesoporous molecular sieve Ti-MCM-41 and amorphous Ti-silicate materials for carbon-carbon bond formation in the Mukaiyama aldol-type reaction between benzaldehyde and methyl trimethylsilyl dimethylketene acetal was studied an

Full text of DOI:10.1016/j.jcat.2005.04.035

Journal of Catalysis 233 (2005) 342–350
www.elsevier.com/locate/jcat
Lewis acid-containing mesoporous molecular sieves as solid efficient
catalysts for solvent-free Mukaiyama-type aldol condensation
Raul Garro, María T. Navarro, Jaime Primo, Avelino Corma ∗
Instituto de Tecnología Química, UPV-CSIC, Universidad Politécnica de Valencia, Avda. de los Naranjos s/n, 46022 Valencia, Spain
Received 14 February 2005; revised 28 April 2005; accepted 29 April 2005
Available online 9 June 2005
Abstract
The activity of the titanium mesoporous molecular sieve Ti-MCM-41 and amorphous Ti-silicate materials for carbon–carbon bond for-
mation in the Mukaiyama aldol-type reaction between benzaldehyde and methyl trimethylsilyl dimethylketene acetal was studied and the
IV
VI
reaction network established. Only tetrahedral Ti is catalytically active, where Ti and highly dispersed TiO are not active or only
2
slightly active for the Mukaiyama reaction and have no effect on selectivity. Whereas in Ti-zeolites only a some of the active sites are acces-
sible to reactants, all of them are accessible in Ti-MCM-41. The structured Ti-MCM-41 offers a clear catalytic advantage over amorphous
Ti-silicalites. Water has a negative effect on catalytic activity and selectivity, and catalyst preactivation and hydrophobicity play an important
role in the final catalytic behavior. The influence of different solvents on conversion was explored.
2005 Elsevier Inc. All rights reserved.
Keywords: Lewis acid catalyst; Mukaiyama reaction; Aldol-type condensation; Heterogeneous catalysis; Mesoporous molecular sieves; Ti-MCM-41
1. Introduction
within stoichiometric or catalytic amounts for the prepara-
tion of crossed aldols by the reaction of silyl enol ethers
or silyl ketene acetals with aldehydes and ketones, known
as the Mukaiyama aldol condensations or Mukaiyama-type
aldol condensation [15–19]. Some heterogeneous catalysts,
such as clays, acid polymers, amorphous silica–alumina, and
metal zeolites (ZSM-5 and Beta), have been reported to be
active for this reaction [20–25]. Especially relevant for us
was the work of Kumar et al. [26–28], who found that TS-1
and Ti-Beta were oxophilic Lewis catalysts able to carry
out the Mukaiyama reaction in the absence of H2O. The
similarity in the activities of TS-1 and Ti-Beta led the au-
thors to speculate that the reaction was mainly occurring
on the external surface of the zeolite, and, consequently, it
may very well be that only a fraction of the potential ac-
tive sites was accessible to reactants. If this is the case, it
seems to us that MCM-41 mesoporous materials could be
better catalysts than metal zeolites for the Mukaiyama reac-
tions, provided the metal Lewis acid site in the mesoporous
metal silicate is tetrahedrally coordinated and isolated on the
walls. Thus, in the present work we show that well-prepared
Developing active and selective solid catalysts for the
formation of new carbon–carbon bonds is a matter of con-
siderable interest. Brønsted solid acids are used commer-
cially to catalyze reactions such as alkylation, dimeriza-
tion, acylations, etc. [1–4]. Meanwhile, solid bases such as
alkaline-substituted zeolites, alkaline-earth oxides [5,6], hy-
drotalcites [7–11], and ALPONs [12,13] have been used
successfully for nucleophillic reactions involving carbanion-
type species for the formation of new carbon–carbon bonds
through aryl-ring side-chain alkylation, Knoevenagel and
aldol condensations, and Michael additions, among oth-
ers. Finally, Lewis acids are also largely used as catalysts
for alkylation, polymerization, Friedel–Crafts, and acyla-
tion, but also for aldol-type condensations and related re-
actions [14]. For instance, Lewis acids derived from Ti,
Sn, Zn, Al, and rare earths, among others, have been used
*
Corresponding author. Fax: +34 96 3877809.
E-mail address: acorma@itq.upv.es (A. Corma).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2005.04.035
2005 Elsevier Inc. All rights reserved.

R. Garro et al. / Journal of Catalysis 233 (2005) 342–350
343
Ti-MCM-41 catalysts [29] are more active than the corre-
sponding ZSM-5 and Beta zeolites for the Mukaiyama re-
action between benzaldehyde (1) and methyl trimethylsilyl
dimethylketene acetal (2), while giving very high selectiv-
ity for the desired condensation product (3a, 3b). We show
the effect of catalyst activation and its hydrophobicity on ac-
tivity. The reaction has been studied with various solvents
and under solvent-free conditions. Finally, the influence of
water and the hydrophobicity of the catalyst on activity and
selectivity is also presented.
2.4.2. UV–vis spectroscopy
All of the analysis was carried out in a Varian Cary 5 with
barium sulfate as an internal standard.
2.4.3. Raman spectroscopy
Raman spectra were collected with a Renishaw inVia
Raman spectrometer equipped with a Leica DMLM micro-
scope and a 514-nm Ar+ ion laser as an excitation source.
A ×50 objective of 0.37-mm optical length was used to fo-
cus the depolarized laser beam on a spot of about 2 µm in
diameter, with a laser power at the sample of 2.0 mW. The
Raman scattering was collected with a CCD array detector,
in the 50–2000 cm−1 spectral region with a resolution of
2 cm−1. Each reported spectrum is the average of 20 scans
with an exposure time of 10 s each.
2. Experimental
2.1. Materials
2.5. Catalytic reaction
Samples of Ti-zeolites and Ti-MCM-41 were prepared
in the following way. TS-1 zeolite was synthesized by a
method based on the wetness impregnation of an amorphous
SiO2–TiO2 solid with a solution of tetrapropylammonium
hydroxide [30], Ti-Beta zeolite was obtained in fluoride me-
dia with a previously reported procedure [31], and Ti-MCM-
41 catalysts were prepared from a gel with the following
molar composition [29]: SiO2:x Ti(OEt)4:0.12 CTABr:0.26
TMAOH:24.3 H2O, where x is varied between 0 and 0.08.
CTABr and TMAOH correspond to cetyltrimethylammo-
nium bromide (from Aldrich) and tetramethylammonium
hydroxide (from Aldrich), respectively. The silica source
was Aerosil-200, from Degussa, and Ti(OEt)4 was supplied
by Alpha Products.
The Mukaiyama reaction between benzaldehyde (1) and
methyl trimethylsilyl dimethylketene acetal (2) was carried
out in a two-necked glass batch reactor under a N2 at-
mosphere. Before the reaction, the calcined catalyst was
activated in vacuum overnight at 250 ◦C. Benzaldehyde
(106 mg; 1 mmol) was placed in another flask, together
with 280 mg of methyl trimethylsilyl dimethylketene ac-
etal (1.5 mmol) and 170 mg of n-dodecane (1 mmol) as an
internal standard, and 2 ml of anhydrous dichloromethane.
The reaction mixture was stirred for a few minutes at am-
bient temperature (20–25 ◦C), and an aliquot was analyzed
and considered the time zero value. The reaction solution
was added to the preactivated catalyst, and the mixture was
stirred magnetically for 24 h at the reflux temperature of
the solvent (40 ◦C). Six aliquots of 25 µl were withdrawn
at different reaction times and subsequently analyzed by gas
chromatography with a 5% phenylsilicone column (ZB-5)
25 m in length.
2.2. Surface silylation
Typically surface silylation was performed as follows
[32]. One gram of calcined Ti-MCM-41 was autogassed at
300 ◦C for 2 h. Then 10 g of solution containing 1.34 g of
hexamethyldisilazane (obtained from ABCR GmbH & Co)
in toluene was added under nitrogen on the dehydrated Ti-
MCM-41. The resulting mixture was refluxed under an inert
atmosphere at 120 ◦C for 2 h. Then the silylated sample was
filtrated and washed with 250 ml of toluene.
3. Results and discussion
3.1. Thermally catalyzed reaction
When silyl ketene acetal (2) is reacted with benzalde-
hyde (1) at the reflux temperature of CH2Cl2 (40 ◦C) in a
1.5:1 mol ratio in the absence of any catalyst, some ther-
mal conversion occurs, and a 6% yield of the corresponding
β-hydroxy ester (3a, 3b) is formed after 24 h (Table 1),
indicating that the homogeneous reaction is slow at this tem-
perature. When the reaction occurs in the presence of well-
crystallized pure silica MCM-41 with a BET surface area of
858 m2 g−1 (Table 2), under the same reaction conditions,
the yield of the condensation product increases to 20% (Ta-
ble 1). The result suggests that the silanol groups, in spite of
their low acidity, are able to catalyze the reaction, and/or the
large adsorption capacity of the MCM-41 sample (concen-
tration effect) produces an increase in the reaction rate of the
2.3. Ti grafting MCM-41 sample
Ti/MCM-41 grafted samples were obtained with a previ-
ously described procedure [33], with the use of titanocene
dichloride as a source of titanium and triethylamine to acti-
vate the surface silanols of the MCM-41 sample.
2.4. Characterization of materials
2.4.1. IR spectroscopy
Experiments were carried out in a Nicolet FT-IR 710
spectrometer, and the samples were heated at 400 ◦C and
10−4 Torr.

344
R. Garro et al. / Journal of Catalysis 233 (2005) 342–350
Table 1
Table 3
Physical properties of some micropore catalysts
a
Silanol reactivity for the Mukaiyama-aldol type reaction
2
−1
)
Catalyst
Conversion
Selectivity (mol%)
Yield of (3a, 3b)
Catalyst
Si/Ti
TiO (wt%)
2
BET area (m g
of (1) (mol%)
(mol%)
(3a, 3b)
Others
Aerosil 200
0
0
36
50
56
0
0
3.2
2.6
2.3
200
500
798
–
Silica gel (Ge0050)
Ti-SAM (ENI)
None
Aerosil
8
9
13
75
89
92
25
11
8
6
8
12
a
TS-1
Silica gel
(GE0050)
MCM-41
a
Ti-Beta
455
a
20
95
5
19
Zeolites synthesized in the ITQ following the procedure described in
the experimental part.
a
Reaction conditions: 1 mmol PhCHO, 1.5 mmol silyl ketene acetal, 2 ml
◦
CH Cl , 0.025 g catalyst activated (250 C, vacuum, overnight), 24 h, at
2
2
reflux temperature of the solvent.
Table 2
a
Physical properties of some mesoporous molecular sieves
Catalyst
TiO
(wt%) (m g
BET area
Total pore vol- Pore size
2
2
−1
−1
)
ume (cc g
)
(Å)
MCM-41
0
858
872
920
950
900
1065
915
956
926
833
912
1020
1.08
1.01
0.99
1.02
0.96
1.01
0.98
0.97
0.98
1.02
1.03
1.06
39.6
39.0
39.6
39.2
39.2
39.5
39.0
39.2
39.3
39.0
40.2
39.5
Ti-MCM-41 (0.4)
Ti-MCM-41 (0.7)
Ti-MCM-41 (1.0)
Ti-MCM-41 (2.3)
Ti-MCM-41 (2.7)
Ti-MCM-41 (4.4)
Ti-MCM-41 (5.0)
Ti-MCM-41 (8.7)
0.4
0.7
1.0
2.3
2.7
4.4
5.0
8.7
Fig. 1. IR spectra of different pure siliceous molecular sieves.
Ti-MCM-41 (11.6) 11.6
b
2Ti/MCM-41
5Ti/MCM-41
a
2.8
5.1
b
All materials synthesized in ITQ.
Titanium incorporated to the mesoporous molecular sieve via grafting
b
following the procedure described in the synthesis of the materials.
bimolecular Mukaiyama condensation reaction. We have an-
alyzed these hypotheses by carrying out the reaction in the
presence of Aerosil-200 (Degussa; BET = 200 m2 g−1; Ta-
ble 3) with external and accessible sylanol groups, and with
amorphous silica GE0050 (Scharlau) with a relatively large
BET surface area (500 m2 g−1). The results obtained (Ta-
ble 1) show a lower activity for the aerosil or the amorphous
silica than for the MCM-41 material, despite the fact that
the number of silanols in the amorphous silicas is not that
different from that of MCM-41 (Fig. 1). This suggests that
the higher activity of MCM-41 should be better related to
the concentration effect achieved with the very high surface
area ordered silica. Our observation agrees with the fact that
pure silica MCM-41 is active in catalyzing the condensation
of silyl ketene acetals with aldehydes [34].
Fig. 2. UV–vis spectra of titanium containing catalysts: (1) TS-1, (2) Ti-
Beta, (3) Ti-MCM-41 (2.7), (4) Ti-SAM (ENI).
Lewis acids can catalyze the Mukaiyama-aldol reaction of
silyl ketene acetals with carbonyl compounds [28,35].
The products obtained during the reaction of silyl ketene
acetal (2) with benzaldehyde (1) in the presence of Ti-MCM-
41 were methyl 2,2-dimethyl-3-phenyl-3-(trimethylsilylo-
xy)propanoic acid methyl ester (3a, 3b), 2-methyl-propanoic
acid methyl ester (4), 2-methyl-propanoic acid trimethylsi-
lyl ester (7), trimethylhydroxysilane (5), and hexamethyldi-
siloxane (6). When the yields of the different products were
plotted versus total conversion (Fig. 3), it was possible to see
that in addition to the very main aldol condensation prod-
uct, minor amounts of 2-methyl-propanoic acid methyl ester
3.2. Lewis acid catalysis: reaction network
The activity of the pure silica MCM-41 strongly in-
creased when a Lewis acid such as Ti was incorporated at the
walls during the synthesis. Indeed, the UV–visible spectrum
of the Ti-MCM-41 catalyst shows a charge transfer band at
∼210 nm (Fig. 2), which is indicative of the presence of
tetrahedrally coordinated Ti. This TiIV can act as a Lewis
acid, and it is known that homogeneous and heterogeneous

R. Garro et al. / Journal of Catalysis 233 (2005) 342–350
345
Scheme 1. Reaction network for Mukaiyama’s reaction between benzaldehyde and methyl trimethylsilyl dimethylketene acetal on Ti-molecular sieves.
Fig. 3. Yield of products vs conversion of benzaldehyde for the Mukaiyama
aldol-type reaction. Reaction conditions: 1 mmol PhCHO, 1.5 mmol silyl
◦
ketene acetal, 2 ml CH Cl , 0.025 g activated catalyst (at 250 C, in vac-
2
2
Fig. 4. Raman spectra of titanium containing catalysts.
uum, overnight) at reflux temperature of the solvent during 24 h. (F) Meth-
yl-2,2-dimethyl-3-phenyl-3-(trimethylsilyloxy)propanoic acid methyl es-
ter (3a, 3b), (2) hexamethyldisiloxane (6), (Q) 2-methyl-propanoic acid
trimethylsilyl ester (7), (") 2-methyl-propanoic acid methyl ester (4).
ples (TS-1 and Ti-Beta) were prepared. UV–visible spec-
troscopy (Fig. 2) shows a well-defined adsorption band at
∼210 nm, which can be associated with tetrahedrally coor-
dinated framework titanium [36]. It should be pointed out
that extraframework TiO2 was not detected by Raman spec-
troscopy in Ti-MCM-41 or in the two Ti-zeolites (Fig. 4).
When the activity of Ti-MCM-41 was compared with that of
TS-1 and Ti-Beta (Table 4 and Fig. 5), the former was much
more active, as judged from the turnover frequency (TOF),
which we calculated by dividing the initial rate by the moles
of Ti in the catalyst, which made it 4 times larger.
The above results support the hypothesis that probably
not all of the Ti sites in the zeolites but only those close to
the external surface are able to react [26,28], whereas in the
case of Ti-MCM-41 the larger pore diameter allows a better
diffusivity of the reactants, which can reach the Ti sites lo-
cated at the pore walls. Nevertheless, the above observations
do not guarantee that all of the Ti sites in Ti-MCM-41 are ac-
cessible through the pores, since some TiIV could be buried
within the walls, making it inaccessible to reactants. Thus,
in order to clarify this point, we prepared a sample of Ti on
MCM-41 (2Ti/MCM-41) by anchoring Ti to the walls of a
(4) and trimethylhydroxysilane (5) were also formed as pri-
mary products, from the hydrolysis of the silyl ketene acetal
by H2O present with the reactants or by H2O remaining ad-
sorbed on the solid, despite the fact that the sample was pre-
activated in vacuum (10−1 Torr) at 250 ◦C. In addition to the
hydrolysis products, 2-methyl-propanoic acid trimethylsilyl
ester (7), and trimethylmethoxysilane (8) were observed as
secondary products, and hexamethyldisiloxane (6) was a pri-
mary plus secondary product, and their formation can occur
through the reaction network presented in Scheme 1.
Nevertheless, it should be noted that in all experiments
carried out with Ti-MCM-41, selectivity for the β-hydro-
xyester (aldol) was always above 95%.
3.3. Accessibility of Ti-sites in MCM-41
It was said before that mesoporous Ti-molecular sieves
could offer diffusional advantages with respect to Ti-zeolites.
In order to check this, two highly crystalline Ti-zeolite sam-

346
R. Garro et al. / Journal of Catalysis 233 (2005) 342–350
Table 4
a
Activity of different titanium silicates for the Mukaiyama-aldol type reaction of benzaldehyde and methyl trimethylsilyl dimethylketene acetal
Catalyst
TiO
(wt%)
Conversion
of (1) (mol%)
Selectivity to the products (mol%)
Yield (3a, 3b)
(mol%)
TON
at 5 h
TOF
2
−1
(h
)
(3a, 3b)
Others
b
TS-1
2.6
2.3
2.3
2.8
5.1
3.2
6
37
77
72
99
64
83
100
100
99
99
98
17
0
0
1
1
5
37
77
71
98
63
4
37
48
61
29
17
1
7
10
12
6
b
Ti-Beta
Ti-MCM-41 (2.3)
2Ti/MCM-41
5Ti/MCM-41
c
c
b
Ti-SAM (ENI)
2
3
a
◦
Reaction conditions: 1 mmol PhCHO, 1.5 mmol silyl ketene acetal, 2 ml CH Cl , 0.025 g of activated catalyst (at 250 C, in vacuum, overnight) at reflux
2
2
temperature of the solvent during 24 h.
b
With 2 h of activation.
With Ti/MCM-41 grafted.
c
(a)
Fig. 5. Yield of the condensation product vs reaction time for Mukaiyama
reaction of PhCHO and silyl ketene acetal over titanium mesoporous mole-
cular sieves. Reaction conditions: 1 mmol PhCHO, 1.5 mmol silyl ketene
◦
acetal, 2 ml of CH Cl , 0.025 g of activated catalyst (at 250 C, in vacuum,
2
2
◦
overnight) at reflux temperature of the solvent (40 C) during 24 h.
pure silica sample of MCM-41 [33]. Note that spectroscopic
characterization shows that in this sample the Ti atoms are
also tetrahedrally coordinated (Fig. 6b). The aldol condensa-
tion activity of 2Ti/MCM-41, per atom of Ti, per hour (i.e.,
TOF), is very close to that of the Ti-MCM-41 (2.3) (see Ta-
ble 4 and Fig. 5), indicating that most, if not all, of Ti sites
in Ti-MCM-41 are accessible through the pores.
Another important point about mesoporous molecular
sieve catalysts is the potential catalytic benefit of the very
regular pores present in these materials, compared with
fully amorphous mesoporous materials. In the case of the
Mukaiyama condensation with solid titanium catalysts, we
have also carried out the reaction with an amorphous meso-
porous Ti-catalyst (SAM) [37]. Characterization of Ti sites
in this sample (Fig. 2) shows that they are tetrahedrally coor-
dinated. Activity results (Table 4) indicate that Ti-MCM-41
is more active than the amorphous material, showing the
benefit of the regular pore dimensions in MCM-41. The rea-
son for this can be related with a better structure of the Ti in
Ti-MCM-41 due to the presence of the surfactant, or to the
presence of bottlenecks at the pores in the case of the amor-
(b)
Fig. 6. UV–vis spectra for Ti-MCM-41 samples with different amount
of TiO . (a) Titanium incorporated to the mesoporous molecular sieve
2
via synthesis: (1) MCM-41, (2) Ti-MCM-41 (2.3), (3) Ti-MCM-41 (5.0),
(4) Ti-MCM-41 (11.6); (b) titanium incorporated via grafting: (1) 2Ti/
MCM-41, (2) 5Ti/MCM-41.
phous material that can limit the accessibility of the reactants
to some Ti sites.
3.4. Influence of Ti site concentration (Ti-MCM-41) on
reactivity
We have attempted to maximize the number of Ti Lewis
active sites by introducing further amounts of Ti during the

R. Garro et al. / Journal of Catalysis 233 (2005) 342–350
347
(a)
Fig. 7. Initial rate (r ) vs wt% of TiO for Mukaiyama reaction of PhCHO
0
2
and silyl ketene acetal over titanium mesoporous molecular sieves. Reaction
conditions: 1 mmol PhCHO, 1.5 mmol silyl ketene acetal, 2 ml of CH Cl ,
2
2
◦
0.025 g of activated catalyst (at 250 C, in vacuum, overnight) at reflux
temperature of the solvent during 24 h. (") Ti-MCM-41 samples with dif-
ferent wt% of titanium (as TiO ); (2) Si-MCM-41 with impregnated tita-
2
nium.
synthesis of the titanosilicate mesoporous molecular sieve. If
all of the Ti atoms were in the form of isolated and tetrahe-
drally coordinated sites, then the activity should be directly
proportional to the Ti content. Thus, if we plot initial reac-
tion rate versus Ti content in Ti-MCM-41 samples, we can
see (Fig. 7) that when the Ti concentration increases above
1–2% not all Ti atoms are in the form of active sites for the
Mukaiyama condensation. A detailed characterization of the
different catalysts by UV–visible and Raman spectroscopy
(Figs. 6 and 8) shows that for Ti-MCM-41 samples with Ti
contents larger than ∼1–2 wt%, in addition to the ∼210-nm
UV band associated with TiIV, absorption bands at ∼270 and
∼320 nm appear, indicating the presence of Ti–O–Ti pairs
and even TiO2. When the Ti content increases further, the
relative intensity of the higher frequency bands increases.
Raman spectroscopy confirms the presence of extraframe-
work TiO2 (bands at 450 and 610 cm−1 assigned to rutile
clusters and four bands at 144, 397, 518, 641 cm−1 assigned
to anatase particles) [38,39] in the samples with larger Ti
content. Our results indicate that octahedrally coordinated
Ti and TiO2, even if highly dispersed (in several samples
TiO2 was not detected by XRD, but only by Raman and UV
spectroscopies), are not catalytically active, or at least they
are much less active than isolated TiIV for the Mukaiyama
condensation between benzaldehyde and silyl ketene acetals.
Results from Table 5 show, however, that the presence of ex-
traframework titanium does not affect the selectivity of the
catalyst. To further confirm this, 2 wt% Ti (in the form of
TiO2) was impregnated on pure silica MCM-41 and, after
calcination at 500 ◦C, was tested under the same reaction
conditions as described above. The activity result is given
in Fig. 7, confirming that extraframework TiO2 has very low
activity.
(b)
Fig. 8. Raman spectra of titanium containing mesoporous molecular sieves.
(a) Region of the spectra where appears characteristic tetrahedral band for
titania: (1) MCM-41, (2) Ti-MCM-41 (2.3), (3) Ti-MCM-41 (5.0), (4) Ti-
MCM-41 (11.6); (b) bands assigned to octahedral titania: (1) MCM-41,
(2) Ti-MCM-41 (2.3), (3) Ti-MCM-41 (5.0), (4) Ti-MCM-41 (11.6).
Table 5
Influence of the titanium incorporation into MCM-41 on the activity and
selectivity for the Mukaiyama reaction
a
TiO (wt%)
2
Conversion of (1) (mol%)
Selectivity (mol%)
2.3
5.0
26
63
70
65
81
82
100
100
100
100
100
100
11.6
b
2.3
b
5.0
b
11.6
a
Reaction conditions: 2 mmol PhCHO, 3 mmol silyl ketene acetal, 2 ml
CH Cl , 0.035 g of Ti-MCM-41 activated (250 C, vacuum, 2 h), 7 h of
reaction at reflux temperature of the solvent.
◦
2
2
b
Reaction made with 0.075 g of catalyst.
In conclusion, higher conversions with very high selec-
tivity for the Mukaiyama reaction can be obtained with
Ti-MCM-41 catalysts containing large amounts (11.6 wt%
TiO2) of Ti. In this case only a part of the Ti, that is, iso-
lated framework TiIV, is active, whereas the extraframework
TiO2 has little if any effect on reactivity. With this Ti-MCM-
41 sample, the activity per weight of titanosilicate is 4 times
higher than for the most active Ti-Beta catalyst at the reflux
temperature of CH2Cl2 (40 ◦C).

348
R. Garro et al. / Journal of Catalysis 233 (2005) 342–350
Table 6
a
Effect of the catalyst activation for the Mukaiyama reaction
Ti-MCM-41
activation
CH Cl
2
(ml)
Time
(h)
Conversion
of (1) (mol%)
Selectivity (mol%)
Hydrolysis product yield (mol%)
2
(3a, 3b)
Others
(6)
(4)
(7)
Yes
No
Yes
No
Yes
No
1
1
0.25
0.25
0.5
0.5
14
23
4
4
5
83
67
84
78
57
49
99
100
99
99
99
1
0
1
1
1
0
3
3
3
0
2
2
8
22
7
1
13
18
4
3
1
1
2
3
5
100
a
Reaction conditions: 0.02 g of Ti-MCM-41 (2.3 wt% TiO ), 1 mmol of PhCHO, 1.5 mmol of silyl ketene acetal, toluene as solvent, reaction made in flask
2
◦
◦
−1
microreactors (2 ml) hermetically closed at 100 C. Activation of the Ti-MCM-41 samples at 250 C, in vacuum (10 mmHg) overnight.
Table 7
a
Effect of the use of 4A molecular sieves in the Mukaiyama reaction
Ti-MCM-41
(g)
Molecular
sieves 4A
Time
(h)
Conversion
of (1) (mol%)
Selectivity (mol%)
Hydrolysis product yield (mol%)
(3a, 3b)
Others
(6)
(4)
(7)
0.035
0.040
0.100
0.100
Yes
No
Yes
No
21.5
21.5
22
69
50
96
60
100
100
98
0
0
2
0
4
5
0
7
7
9
0
1
1
0
0
23
100
11
a
◦
Reaction conditions: Ti-MCM-41 (2.3 wt% TiO ) activated (at 250 C, in vacuum, 2 h), 2 mmol of PhCHO, 3 mmol of silyl ketene acetal, 2 ml of
2
anhydrous CH Cl at reflux temperature of the solvent.
2
2
Table 8
Effect of the silylation of Ti-MCM-41 (2.3 wt% TiO ) on the catalytic activity of the Mukaiyama reaction
a
2
C
Si(Me) /
Conversion
Selectivity (mol%)
Hydrolysis product yield (mol%)
3
(wt%)
SiO
of (1) (mol%)
2
(3a, 3b)
Others
(6)
(4)
(7)
0
7.89
0
0.16
77
90
100
100
0
0
0
1
0
0
0
0
a
◦
Reaction conditions: 0.025 g of Ti-MCM-41 (2.3 wt% TiO ) activated (at 250 C, in vacuum, 2 h), 1 mmol of PhCHO, 1.5 mmol of silyl ketene acetal,
2
2 ml of CH Cl , at reflux temperature of the solvent during 24 h.
2
2
3.5. Effect of catalyst activation and postsynthesis
treatment
We have seen before that the most important competing
reaction was the hydrolysis of the silyl ketene acetals, due to
residual amounts of water remaining adsorbed on the cata-
lyst, and/or present in the benzaldehyde, silyl ketene acetal,
and solvent. It has also been claimed [26–28,40] that wa-
ter poisons Ti Lewis acid sites, decreasing the activity of
Ti-zeolites. We have confirmed this in the case of Ti-MCM-
41, by performing the experiments with a sample exposed to
the open atmosphere, without heating under vacuum before
use, and with a catalyst sample previously activated in situ
at 250 ◦C under a vacuum of 10−1 Torr for 2 h. The results
from Table 6 show that water removal by catalyst activa-
tion has a positive effect on conversion. Further removal of
the water present in reactants and solvent by means of a 4A
molecular sieve further increases the activity of the catalyst
(Table 7).
Fig. 9. Thermogavimetric analysis (TG) of titanium containing mesoporous
molecular sieves with different silylation degree.
given in the Experimental section [32,41]. TG results from
Fig. 9 indicate that silylation of the sample causes the total
amount of adsorbed water and the water desorption tempera-
ture to decrease. Then, when the conversions of silylated and
nonsilylated samples are compared (Table 8), we can see that
Since water has a negative effect on catalyst activity, we
decided to prepare hydrophobic Ti-MCM-41 samples with
a postsynthesis silylation treatment. This was done with
a hexamethyldisilazane treatment, following the procedure

R. Garro et al. / Journal of Catalysis 233 (2005) 342–350
349
Scheme 2. A possible mechanism for Mukaiyama’s reaction between benzaldehyde and methyl trimethylsilyl dimethylketene acetal on Ti-molecular sieves.
Table 9
a
Influence of the solvent in the Mukaiyama reaction
Solvent
Conversion
of (1)
Selectivity
(mol%)
Hydrolysis product
yield (mol%)
(mol%)
(3a, 3b)
Others
(6)
(4)
b
CH Cl
90
85
54
35
7
100
100
100
100
100
100
0
0
0
0
0
0
5
8
4
19
2
3
15
24
13
39
8
2
2
CH CN
3
Toluene
CHCl
3
Dioxane
THF
4
32
a
Reaction conditions: 1 mmol PhCHO, 1.5 mmol silyl ketene acetal, 2 ml
◦
of solvent, 0.02 g of Ti-MCM-41 (2.7 wt% TiO ) activated (at 250 C, in
vacuum, overnight), at 50 C during 22 h of reaction.
b
2
◦
◦
At reflux temperature of the solvent (40 C).
Fig. 10. Yield of the product vs reaction time for Mukaiyama-type reaction
using different solvents. Reaction conditions: 1 mmol PhCHO, 1.5 mmol
silyl ketene acetal, 2 ml of solvent, 0.020 g of activated Ti-MCM-41(2.7) (at
silylation increases the activity of the catalyst by decreasing
the coordination of H2O with Ti sites.
◦
◦
250 C, in vacuum, overnight) at reflux temperature of the solvent (40 C)
during 22 h. (2) with no solvent, (E) with anhydrous CH Cl , (F) re-
2
2
peated experiment with dry CH Cl , (+) with CH CN, (Q) with toluene,
2
2
3
(!) with dry CHCl (4A molecular sieves), (×–) with dry CHCl , (1) with
3.6. Effect of the solvent
3
3
not dry CHCl , (–) with dioxane, (-) with THF.
3
Mukaiyama’s condensations are usually carried out in the
presence of solvents [18,41], and in the case of solid cata-
lysts different organic solvents have been used [25,28,34,42,
43]. In the case of Ti molecular sieves, a plausible mecha-
nism for the Mukaiyama reaction (see Scheme 2) involves a
polarized transition state, and therefore it is not unreasonable
to think that the polarity of the solvent can have an effect on
the stabilization of the transition state. Then a series of sol-
vents with different polarities have been used; the results are
given in Table 9.
Results from Fig. 10 clearly show that when the reaction
is carried out in THF and dioxane, the reaction proceeds
very slowly. This is probably due to a strong adsorption
of the solvent through the oxygen atoms to the Ti sites of
Table 10
a
Solvent free Mukaiyama aldol-type reaction over Ti-MCM-41 (2.3 wt% TiO )
2
Solvent
Temperature
( C)
Ti-MCM-41
(g)
Time
(h)
Conversion of (1)
(mol%)
Selectivity
(mol%)
Hydrolysis product yield (mol%)
◦
(6)
(4)
(7)
Yes
No
Yes
No
Yes
No
100
100
100
100
40
0.100
0.100
0.035
0.035
0.035
0.035
5
5
24
7
21
7
51
79
54
93
50
70
100
100
100
100
100
99
0
0
4
0
5
0
2
0
6
0
9
0
5
1
0
0
0
0
b
40
a
◦
Reaction conditions: Ti-MCM-41 (2.3 wt% TiO ) activated (at 250 C, in vacuum, 2 h), 2 mmol of PhCHO, 3 mmol of silyl ketene acetal, 2 ml of toluene.
2
b
With 2 ml of CH Cl .
2
2

350
R. Garro et al. / Journal of Catalysis 233 (2005) 342–350
Table 11
[2] S. Kobayashi, K. Manabe, S. Limura, Chem. Commun. (2002) 94.
[3] A. Corma, J. Catal. 216 (2003) 298.
[4] S. Kobayashi, Y. Mori, K. Manabe, Tetrahedron 57 (2001) 2537.
[5] A. Hosomi, T. Nakagawa, K. Miura, J. Am. Chem. Soc. 124 (2002)
536.
Solvent free Mukaiyama aldol-type reaction over titanium containing meso-
porous molecular sieves
a
Solvent TiO
Conversion Selectivity
(wt%) of (1)
(mol%)
Hydrolysis product
yield (mol%)
2
(mol%)
[6] M.A. Aramendia, V. Borau, C. Jiménez, J.M. Marinas, J.R. Ruiz, F.J.
Urbano, Appl. Catal. A 255 (2003) 301.
(6)
(4)
(7)
[7] L.M. Parker, N.B. Milestone, R. Newman, Ind. Eng. Chem. Res. 34
(1995) 1196.
[8] M.A. Drezdon, Inorg. Chem. 27 (1998) 4628.
[9] D. Tichit, B. Coq, CatTech 7 (2003) 206.
[10] M.J. Climent, A. Corma, S. Iborra, K. Epping, A. Velty, J. Catal. 225
(2004) 316.
Yes
No
Yes
No
Yes
No
2.3
2.3
5.0
5.0
11.6
11.6
36
100
100
100
100
100
100
0
1
1
1
2
0
0
0
10
0
11
2
0
0
1
0
1
0
78
37
82
53
77
[11] A. Corma, S. Iborra, M. Mifsud, M. Renz, M. Susarte, Adv. Synth.
Catal. 346 (2004) 257.
[12] S. Delsarte, M. Florea, P. Grange, Silicates Industriels 69 (2004) 69.
[13] M.J. Climent, A. Corma, H. García, R. Guil-López, S. Iborra, V. For-
nés, J. Catal. 197 (2001) 385.
a
◦
Reaction conditions: 0.025 g of activated Ti-MCM-41 (at 250 C, in
vacuum, overnight), 1 mmol of PhCHO, 1.5 mmol of silyl ketene acetal,
2 ml of CH Cl , at reflux temperature of the solvent during 5 h.
2
2
[14] A. Corma, H. García, Chem. Rev. 103 (2003) 4307.
[15] T. Mukaiyama, K. Manno, K. Narasaka, J. Am. Chem. Soc. 96 (1974)
7503.
[16] Y. Yamamoto, K. Maruyama, K. Matsumoto, J. Am. Chem. Soc. 105
(1983) 6963.
[17] T. Mukaiyama, Angew. Chem. Int. Ed. 16 (1997) 817.
[18] E.M. Carreira, Comprehensive Asymetric Catálisis I–III 3 (1997) 997,
chap. 29.1.
[19] M. Lemaire, D. Rechavi, Chem. Rev. 102 (2002) 3467.
[20] S. Kobayashi, Eur. J. Org. Chem. 1 (1999) 15.
[21] L. Teck-Peng, L. Xu-Ran, Tetrahedron 55 (1999) 10789.
[22] C.R. Brindaban, S. Manika, B. Sanjay, Tetrahedron Lett. 34 (1993)
1989.
the catalyst. Furthermore, the large difference in activity ob-
served when the reaction is performed in CH2Cl2 and CHCl3
is remarkable. In principle, and from the point of view of
polarity, we did not expect the large differences observed.
Nevertheless, duplication of the experiments confirms the
observation. However, to a first approximation, since the wa-
ter content in CHCl3 is larger than in CH2Cl2 and water has
a negative effect on conversion, this may influence the differ-
ence in activity observed. After the CHCl3 solvent is dried
with the use of a dehydrated A zeolite, an improvement in
conversion occurs. However, there is still a large conversion
difference with respect to CH2Cl2. We have no explanation
for this observation at this point.
[23] M.T. Reetz, D. Giebel, Angew. Chem. Int. Ed. 39 (2000) 2498.
[24] H. Matsuhashi, M. Tanaka, H. Nakamura, K. Arata, Appl. Catal. A 208
(2001) 1.
It should be pointed out that the reaction can be car-
ried out with good activity and selectivity without solvent
(Tables 10 and 11), and we see no benefit to using any of
the above solvents for the Mukaiyama-type condensation re-
action with solid catalysts, where the adsorption properties
have been controlled by direct synthesis or by postsynthesis
treatment, as in the present case.
[25] D.L. Gin, W.J. Zhou, W. Gu, Chem. Mater. 13 (2001) 1949.
[26] M. Sasidharan, R. Kumar, J. Catal. 220 (2003) 326.
[27] R. Kumar, M. Sasidharan, Catal. Lett. 38 (1996) 251.
[28] R. Kumar, P. Vincent, K.V. Srinivasan, V.N. Raju, M. Sasidharan,
Chem. Commun. (1996) 129.
[29] T. Blasco, A. Corma, M.T. Navarro, J. Pérez-Pariente, J. Catal. 156
(1995) 65.
[30] D.P. Serrano, M.A. Uguina, G. Ovejero, R. Van Grieken, M. Camacho,
Micropor. Mater. 4 (1995) 273.
[31] T. Blasco, M.A. Camblor, A. Corma, P. Esteve, J.M. Guil, A. Martinez,
J.A. Perdigón-Melón, S. Valencia, J. Phys. Chem. B 102 (1998) 75.
[32] A. Corma, J.L. Jordá, M.T. Navarro, J. Pérez-Pariente, F. Rey, J. Tsuji,
Stud. Surf. Sci. Catal. 129 (2000) 169.
[33] J.M. Thomas, G. Sankar, F. Rey, T. Maschmeyer, Nature 378 (1995)
159.
[34] M. Iwamoto, H. Ishitani, Tetrahedron Lett. 44 (2003) 299.
[35] R. Mahrwald, Chem. Rev. 99 (1999) 1095.
[36] E. Gianotti, A. Frache, S. Coluccia, J.M. Thomas, T. Maschmeyer,
L. Marchese, J. Mol. Catal. A: Chem. 204–205 (2003) 483.
[37] G. Bellusi, L. Zambelli, A. Corma, M.E. Domine, Eur. Patent Appl.
1403358 (2004), to EniTecnologie.
[38] S. Bordiga, A. Damin, F. Bonino, G. Ricchiardi, C. Lamberti, A. Zec-
china, Angew. Chem. Int. Ed. 41 (2002) 4734.
[39] R. Hoppe, A. Ortlam, J. Rathousky, G. Schulz-Ekloff, A. Zukal, Mi-
cropor. Mater. 8 (1997) 267.
[40] H. Yamamoto, M. Kaneeda, S. Nakamura, K. Ishihara, J. Am. Chem.
Soc. 118 (1996) 12854.
[41] A. Corma, M.E. Domine, J.A. Gaona, J.L. Jorda, M.T. Navarro, F. Rey,
J. Pérez-Pariente, J. Tsuji, B. McCulloch, L.T. Nemeth, Chem. Com-
mun. (1998) 2211.
4. Conclusions
We have demonstrated that well-prepared Ti-MCM-41 is
a better solid catalyst for the Mukaiyama-aldol condensation
than Ti-zeolites, because of the lower diffusional restriction
for reactants and the possibility of introducing isolated TiIV
sites on the walls. The presence of water reduces the cat-
alytic activity and produces the hydrolysis of the silyl ketene
acetal. This problem can be minimized by preactivation of
the catalyst in situ, by drying of the solvent and, particular,
if the catalyst is made more hydrophobic by silylation. A se-
ries of solvents has been studied, and the presence of solvent
does not improve the catalytic performance. Silylated Ti-
MCM-41 can be recycled without any loss in activity or
selectivity.
References
[42] S. Kobayashi, K. Manabe, Y. Mori, Angew. Chem. Int. Ed. 40 (2001)
2816.
[43] S. Kobayashi, S. Nagayama, J. Am. Chem. Soc. 120 (1998) 2985.
[1] H. Yamamoto, A. Hasegawa, K. Ishihara, Angew. Chem. Int. Ed. 40
(2001) 4077.