Water-resistant solid Lewis acid catalysts: Meerwein-Ponndorf-Verley and Oppenauer reactions catalyzed by tin-beta zeolite

Article abstract of DOI:10.1016/S0021-9517(03)00014-9

The catalytic activity of Sn-beta zeolite in the Meerwein-Pondorf-Verley reduction of carbonyl compounds with secondary alcohols as reductants and Oppenauer oxidation of alcohols were performed with quantitative yields to the corresponding product. The catalyst had an excellent activity and selectivity even after four catalytic recycles, and good stereoselectivities to the thermodynamic less favorable cis-alcohol isomer when alkyl-cyclohexanones were used as substrates. A prochiral ketone was reduced within an enantiomeric excess close to 50% when using a chiral alcohol as the reducing reactant. IR studies using cyclohexanone as probe molecule over beta zeolites showed that the more specific Lewis acid sites in the framework of Sn-beta were responsible for its better catalytic activity with respect to Ti- or Al-beta. The order of hydrophobicity of the samples was Ti-beta > Sn-beta > Al-beta. Ti-beta and Sn-beta retained a higher percentage of catalytic activity when water was present in the reaction media. Even with ～ 4 wt % of H₂O (0.2 g) in the media, Sn-beta yielded a higher turnover number than Ti- or Al-beta when working in the absence of water.

Full text of DOI:10.1016/S0021-9517(03)00014-9

Journal of Catalysis 215 (2003) 294–304
www.elsevier.com/locate/jcat
Water-resistant solid Lewis acid catalysts: Meerwein–Ponndorf–Verley
and Oppenauer reactions catalyzed by tin-beta zeolite
∗
Avelino Corma, Marcelo E. Domine, and Susana Valencia
Instituto de Tecnología Química, UPV-CSIC, Universidad Politécnica de Valencia, Avda. de los Naranjos s/n, 46022 Valencia, Spain
Received 27 September 2002; revised 18 December 2002; accepted 18 December 2002
Abstract
The catalytic activity of Sn-beta zeolite in the Meerwein–Pondorf–Verley (MPV) reduction of carbonyl compounds with secondary
alcohols as reductants and Oppenauer oxidation of alcohols were performed with quantitative yields to the corresponding product. This
catalyst exhibits a behavior not observed with Ti- and Al-beta zeolites, with excellent activity and selectivity even after four catalytic recycles,
and good stereoselectivities to the thermodynamic less favorable cis-alcohol isomer when alkyl-cylohexanones were used as substrates.
A prochiral ketone has been reduced within an enantiomeric excess close to 50% when using a chiral alcohol as the reducing reactant.
Other transition metals incorporated in the beta zeolite framework (V or B) led to poor catalytic activities for this reaction. IR studies using
cyclohexanone as probe molecule over beta zeolites demonstrated that the more specific Lewis acid sites in the framework of Sn-beta are
responsibles for its better catalytic activity with respect to Ti- or Al-beta. The influence of the structure of the alcohol and ketone shows
that reaction transition states that by size can be fitted within the pore of beta may have problems forming on the Lewis acid sites due to
the shielding effect of the neighbor framework oxygen atoms. Finally, Sn-beta containing organic groups attached to the structure is able to
perform the MPV in the presence of water, offering a real alternative to conventional aluminum iso-propoxide catalysts.

2003 Elsevier Science (USA). All rights reserved.
Keywords: MPV-O reaction; Tin-beta zeolite; Me-beta catalysts; Solid Lewis acid; Heterogeneous catalysis; Catalysis in aqueous medium
1
. Introduction
reductions [4] whereas metal tert-butoxides are preferred for
oxidations [5]. Other Lewis acid metals such as La or Zr have
been used as homogeneous catalysts for MPVO reactions as
well as magnesium oxides or phosphates in heterogeneous
systems with moderated activity [6–8].
The Meerwein–Ponndorf–Verley (MPV) reduction of
aldehydes and ketones and Oppenauer’s (O) oxidation of
alcohols, together denoted as MPVO reactions, are consid-
ered to be highly selective reactions in which other reducible
groups such as C=C double bonds and C–halogen bonds are
not reacted [1]. In MPV reduction, a secondary alcohol is
the reductant whereas in Oppenauer oxidations a ketone is
the oxidant. It is generally accepted that MPVO reactions
proceed via a transition-state complex in which the carbonyl
group and the alcohol are both coordinated to a Lewis acid
metal center and a hydride transfer from the alcohol to the
carbonyl group occurs (Scheme 1) [2,3].
Usually, aluminum or titanium alkoxides derived from
secondary alcohols are used as homogeneous catalysts in
1
3
*
Scheme 1. MPVO reaction pathways. R = alkyl or aryl; R and R = alkyl
or hydrogen; Me = metal.
Corresponding author.
E-mail address: acorma@itq.upv.es (A. Corma).
0021-9517/03/$ – see front matter  2003 Elsevier Science (USA). All rights reserved.
doi:10.1016/S0021-9517(03)00014-9

A. Corma et al. / Journal of Catalysis 215 (2003) 294–304
295
Recently, Al- and Ti-beta zeolites have been reported
as excellent catalysts for MPVO reactions showing that
Lewis acid sites are the active centers of the catalyst [9–12].
Following this, it appears that the activity and selectivity
of MPVO catalysts can be improved by introducing the
adequate Lewis acidity in the catalyst that can produce the
right polarization of the carbonyl group while coordinating
both the alcohol and the ketone [13]. In this sense, Sn in
beta zeolite seems to be a good candidate since, on the
one hand, it has a higher electronegativity than Ti, and on
the other hand, it can be more adequate for coordinating
both reactants than extraframework Al species. Furthermore,
hydrophobic Sn-beta zeolites could be prepared that should,
in principle, increase the water resistance of the catalyst.
In a very recent short communication [14] we showed that
the Sn-beta is indeed an excellent heterogeneous catalyst
for the reduction of aldehydes and ketones to alcohols and
the oxidations of alcohols to the corresponding carbonyl
group by a MPVO process. The results obtained were
compared with those reported with optimized Al- and Ti-
beta zeolites [9], and the former shows a higher activity and
better selectivity. In this work we have carried out a deep
study of this system and will present that the activity of
zeolite containing different transition metals for the MPVO
reaction correlates well with the Lewis acid strength of
the catalyst. By carrying out the reaction with alcohols
and ketones of different nature and size, and using basic
probe molecules with different degrees of steric hindrance,
we demonstrate that in many cases the catalytic activity
observed is not controlled by steric hindrance due to pore
diameter of the zeolite and the diffusion of reactants or the
size of the transition state into the zeolite cavity, but by the
shielding effect of the framework oxygen atoms bonded to
the Lewis acid. Will also show that high stereoselectivities
and reasonable enantioselectivities can be achieved in some
cases with this catalyst. Finally, we present modifications in
catalyst design and preparation that allow use of Sn-beta in
water-containing reaction media with an excellent activity.
rotation (60 rpm). After the required heating time, the au-
toclaves were cooled and the solids were recovered by fil-
tration and extensively washed with distilled water. Finally,
the solids were dried at 100 C and calcined at 580 C, and
the resultant materials were analyzed by XRD showing high
crystallinities.
◦
◦
2
.2. Synthesis of Al- and Ti-beta zeolites
The Al-beta samples were obtained by hydrothermal
◦
synthesis at 140 C in Teflon-lined stainless-steel auto-
claves with rotation (60 rpm), by hydrolyzing TEOS (98%,
Merck) in an aqueous solution of TEAOH (35%, Aldrich).
Then a solution of metal aluminum (99.95%, Merck) dis-
solved in aqueous TEAOH was added and the mixture was
kept under stirring until the complete evaporation of the
ethanol formed upon hydrolysis of TEOS. Finally, HF (48%,
Aldrich) and, optionally, dealuminated zeolite beta seeds
were added. After the required crystallization time, the au-
toclaves were cooled and the products were filtered and ex-
tensively washed with distilled water. The Si/Al ratios were
5
0 and 15 as determined from chemical analysis. Finally,
◦
◦
the solids were dried at 100 C and calcined at 580 C, and
the resultant materials were analyzed by XRD showing high
crystallinities.
The Al-beta samples containing extraframework alu-
minum were prepared by a careful postsynthetic treatment
with calcinations of the sample in several steps as described
in [9]. In the case of Ti-beta catalyst, a hydrophobic Ti-beta
sample was prepared following the procedure previously de-
scribed in [16].
2
.3. Synthesis of B- and V-beta zeolites
The B-beta sample was prepared from a silica source
(
Aerosil 200, Degussa) plus boric acid (H3BO3) with
TEAOH (35%, Aldrich) as template, by diluting 21.0 g of
TEAOH with 7.9 g of water. Then, 0.42 g of boric acid
was dissolved in this solution, followed by the addition of
6.0 g of silica. This reaction mixture is heated in a Teflon-
2
. Experimental
◦
lined stainless-steel autoclave at 150 C rotated at 60 rpm
2
.1. Synthesis of Sn-beta
for 6 days. The solid product was filtered, washed, dried at
◦ ◦
1
00 C, and then calcined at 580 C. The XRD pattern of this
In order to achieve hydrophobic Sn-beta samples [15],
sample showed the beta structure.
fluoride medium from gels was prepared with the follow-
ing molar compositions: SiO2:xSnO2:0.54TEAOH:7.5H2O:
In the case of V-beta, 37.8 g of aqueous solution of
TEAOH (35%, Aldrich) and 10.0 g silica were mixed.
After 30 min a solution of VOSO4 · H2O (1.35 g) in
water (5.5 g) was added and the mixture was stirred for
30 min. Finally, zeolite beta seeds were added and the
gel was introduced in Teflon-lined stainless-steel autoclaves
0
.54HF where x was 1/120 and 1/240. The gel was pre-
pared by hydrolyzing TEOS (98%, Merck) in an aqueous so-
lution of TEAOH (35%, Aldrich), a solution of SnCl4 ·5H2O
(
98%, Aldrich) in water was added, and then the mixture
◦
was kept under stirring until the complete evaporation of
ethanol formed upon hydrolysis of TEOS. Then HF (48%,
Aldrich) was added and finally dealuminated zeolite beta
and heated at 140 C for 12 days. The final zeolite was
recovered by centrifugation, washed with distilled water,
◦
◦
dried at 100 C overnight, and finally calcined at 580 C for
3 h, resulting in a solid sample with the XRD pattern of beta
structure.
seeds were added. The crystallization was carried out in
◦
6
0-mL Teflon-lined stainless-steel autoclaves at 140 C with

296
A. Corma et al. / Journal of Catalysis 215 (2003) 294–304
2
.4. Synthesis of Sn-MCM-41 material
an IR spectrum was recorded. The same procedure was fol-
lowed for 2,6-dimethylpyridine, 2-ethylpyridine, and pyri-
dine IR adsorption measurements, with the corresponding
desorption performed at 25 and 150 C.
Sn-MCM-41 sample with approximately 10 wt% of
SnO2 was prepared in the following way. An aqueous so-
lution of hexadecyltrimethylammonium hydroxide/bromide
◦
(
C16TAOH/Br) was mixed with a TMAOH solution (25%,
2
.6. Catalytic experiments
Aldrich) and an aqueous solution of SnCl4 · H2O (98%,
Aldrich). After homogenization, the silica (Aerosil, De-
gussa) was added under continuous stirring. The homoge-
neous gel was sealed in Teflon-lined stainless-steel auto-
claves and heated at 135 C under static conditions for 24 h.
The resulting solid was filtered, washed with water, and
dried at 60 C for 24 h. The occluded organic was removed
by heating the solid at 540 C for 1 h in nitrogen flow, fol-
lowed by 6 h in air. The solid obtained presents the XRD of
The experiments for the reduction of ketones with sec-
ondary alcohols (Meerwein–Ponndorf–Verleyreaction) were
performed in a 10-ml round-bottomflask immersed in a ther-
mostated bath and equipped with a condenser, a thermome-
ter, and a magnetic stirrer. Typically, 1 mmol of the substrate
and 60 mmol of the alcohol were homogenized in the flask
under stirring, and the mixture was heated at temperatures
◦
◦
◦
◦
between 85 to 100 C. Then, 75 mg of catalyst was added
2
−1
MCM-41, surface area of 950 m g , and 3.5-nm pore di-
ameter (35 Å) and XRD pattern of MCM-41 structure (spec-
tra not shown).
at once to the reaction mixture (time 0). Small aliquots were
withdrawn from the mixture at different time intervals to fol-
low the kinetics of the reaction. For studying the effect of
catalyst and alcohol concentrations on the reaction rate, the
amount of catalyst was varied between 3 and 75 mg, whereas
the concentrations of alcohol were 48, 20, 9, 6, and 3 mmol.
Silylation of calcined samples was carried out following
17,18].
Phase purity of the catalysts was determined by X-ray
[
diffraction (XRD) in a Philips X’Pert MPD diffractome-
ter equipped with a PW3050 goniometer (Cu-Kα radiation,
graphite monochromator), provided with a variable diver-
gence slit and working in the fixed irradiated area mode.
2
.7. Oppenauer oxidations
The oxidations of sec-alcohols to the corresponding ke-
2
9
119
tones were carried out using 2-butanone as oxidating agent.
The reactions were performed with the same equipment
described above under the following reaction conditions:
Si MAS NMR and Sn MAS NMR were recorded at
spinning rate of 5 kHz on a Varian VXR 400S WB spectrom-
eter. Diffuse reflectance UV–visible (DRUV) spectra were
recorded in a Cary 5 Varian spectrometer equipped with a
1
mmol of the selected alcohol and 80 mmol of 2-butanone
◦
were mixed in the reactor at 80 C. Then the catalyst (75 mg)
was added (time 0).
“
Praying Mantis” cell from Harrick.
Thermogravimetric and differential thermal analysis
(
TGA-DTA) were performed in a Netzsch STA 409 EP ther-
2
.8. Reactions with prochiral substrates
mal analyzer with about 20 mg of sample and a heating rate
of 10 C min in air flow (6 l h ). Chemical composition
was determined by atomic absorption in a Varian SpectrAA-
◦
−1
−1
MPV reduction of different prochiral ketones was made
in the same manner as describes above with (S)-(+)-2-
1
0 Plus and elemental analysis in a Fisons EA1108CHN-S.
butanol as reducing agent. The alcohol/ketone molar ratio
Scanning electron microscopy in a JEOL 6300 microscope
was used in order to know the size and morphology of the
crystals. Nitrogen and argon adsorption experiments at 77
and 85 K, respectively, were carried out in a Micromeritics
ASAP 2000.
◦
was 80, with 75 mg of catalyst at 80–100 C.
The reaction products were analyzed by gas chromatog-
raphy in a Varian 3400 GC equipped with a capillary column
(
5% methylphenylsilicone, 25 m length) and a FID. Product
identification was performed by GC-MS and compared with
available standard compounds. The analysis of chiral prod-
ucts was carried out using a capillary column with chiral sta-
tionary phase (BETA-DEX 120, 30 m length).
2
.5. IR spectroscopic measurements
Infrared spectra were obtained in a Nicolet 710 FTIR
−2
spectrometer using self-supported wafers of 10 mg cm
outgassed overnight at 653 K and 10 Pa. The measure-
ments with adsorption of cyclohexanone (and its 2-methyl
−3
3. Results and discussion
3
.1. Catalyst characterization
◦
and 2,6-dimethyl derivatives) at 25 C and desorption of the
organic molecule in vacuum for 1 h were performed at 50,
The physical properties of the catalysts are summarized in
◦
1
00, and 200 C over selected Me-beta zeolites. First, all the
Table 1. XRD patterns of the Sn-beta (1 and 2 wt% of SnO2),
and Al- (Si/Al = 15 and 50), Ti-, V-, and B-beta samples are
shown in Figs. 1 and 2, respectively. Taking as a reference a
commercial beta zeolite (Zeolyst Intl) it can be said that the
crystallinity of the samples is higher than 90% except for
the V-beta sample in which crystallinity ranges between 80
and 90%.
water must to be removed from the zeolite. Consequently,
◦
the zeolite samples were dehydrated overnight at 400 C in
vacuum, and the first spectrum was recorded. Then the cy-
clohexanone was adsorbed and again an IR spectrum was
taken. Afterward, the organic compound was desorbed suc-
◦
cessively at 50, 100, and 200 C, and after each desorption

A. Corma et al. / Journal of Catalysis 215 (2003) 294–304
297
Table 1
Physical properties of the catalyst samples tested in MPVO reactions
Catalyst Si/Me BET surface Micropore Crystal Desorbed
ratio
area
m g
volume
(cm g
size
(µm)
H O
2
(wt%)
2
−1
3
−1
a
(
)
)
Beta (SiO )
2
–
120
240
65
50
15
12
47
106
450
475
470
454
484
518
–
0.20
0.20
0.20
0.19
0.19
0.20
–
1.0
0.5
0.5
0.5
0.3
0.3
2.0
1.0–5.0
0.2
0.44
2.93
–
1.80
4.87
16.79
16.83
2.63
8.57
Sn-beta
Sn-beta
Ti-beta
Al-beta
Al-beta
b
Al-beta
B-beta
V-beta
460
387
0.19
0.14
a
Calculated by TG analysis.
b
Synthetized the same as van Bekkum and colleagues and calcined with
the same postsynthesis procedure [9].
Fig. 2. XRD pattern of calcined: (a) Al-beta (Si/Al = 15), (b) Al-beta
(Si/Al = 50), (c) Ti-beta, (d) V-beta, and (e) B-beta zeolites samples.
(a)
(b)
◦
Fig. 3. Infrared spectra of pyridine adsorption (at 25 C) and desorption at
◦
Fig. 1. XRD pattern of as-made and calcined Sn-beta: (a) with 1 wt% SnO ,
2
150 C over Al-beta (Si/Al = 15) (a), Al-beta (Si/Al = 50) (b), Sn-beta (c),
(
b) with 2 wt% SnO .
Ti-beta (d), V-beta (e), and B-beta (f).
2
◦
−1
The amount and strength of Lewis acid sites of the
metal-containing zeolites were compared by means of IR
150 C in the 1800–1300 cm region are shown in Fig. 3.
The appearance of the band corresponding to Lewis acid
−1
spectroscopy of the pyridine adsorbed and desorbed at
sites in the IR spectra (1452 cm ) is clearly observed for
the beta samples containing Sn, Ti, and Al. However, the Sn-
◦
1
50 C. The IR spectra obtained after pyridine desorption at

298
A. Corma et al. / Journal of Catalysis 215 (2003) 294–304
beta Lewis acidity exhibits a very weak character similar to
Ti-beta sample [18], while Al-beta presents a large amount
of strong Lewis and Brönsted acidity. When beta zeolites
containing V and B were measured, negligible amounts of
pyridine coordinated to Lewis acid sites were detected after
Table 2
Selective reduction of cyclohexanone with 2-propanol over Sn-beta and
other materials (MPV reaction) (reaction conditions: 85 C, 1 h of reaction,
◦
1
mmol. of substrate, 60 mmol of 2-propanol and 75 mg of catalyst)
Catalyst
Conversion
Selectivity (% mol)
(wt% MeO or Si/Me ratio) (% mol)
◦
2
Cyclohexanol Alquene Others
desorption at 150 C, showing similar absorbance values
than the all-silicon-beta zeolite.
type
1 h
–
0
100
0
0
0
0
0
56.8
91.0
3.9
0
100
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
5.4
3.7
The incorporation of Sn in the framework of beta was
Al(iso-propoxide)
Ti(iso-propoxide)
2% SnO2/SiO2
3
1
19
checked by means of
Sn MAS NMR. Thus, a Sn-beta
3
sample enriched with 1 Sn isotope was synthesized and the
MAS NMR spectrum of the dehydrated Sn-beta gives a peak
at −444 ppm, which is related to tetrahedrally coordinated
Sn in the framework of zeolites [13]. No peak at −690 and
19
a
SnCl · 5H O
4
2
Si-beta
0
0
2% SnO /Si-beta
2
Sn-beta (1 wt% SnO )
Sn-beta (2 wt% SnO )
2
100.0
100.0
100.0
100.0
94.6
97.6
2
−
740 ppm corresponding to octahedrally coordinated Sn in
zeolites was found. In an analogous way the UV–visible
spectra of Ti- and V-beta samples show the incorporation of
these metals by the detection of absorbance bands centered
at 220 and 290 nm for titanium- and vanadium-containing
samples, respectively.
Ti-beta (2 wt% TiO2)
Al-beta (Si/Al = 50)
Al-beta (Si/Al = 15)
15.0
44.6
53.5
b
Al-beta (Si/Al = 12)
0.7
a
With 100 mg of the salt.
Synthetized and calcined the same as van Bekkum and colleagues [9].
b
3
.2. Catalytic activity
Table 3
Selective reduction of cyclohexanone with 2-butanol over Sn-beta and other
materials (MPV reaction) (reaction conditions: 100 C, 1 h of reaction,
The selection of the beta structure for catalyst preparation
◦
was based on the fact that this is a large-pore tridimensional
zeolite that allows the incorporation of transition metals
in framework positions [19], achieving in this way single
isolated Lewis acid sites that can, in principle, be active for
MPVO reactions. Thus, when Sn-, V-, Ti-, B-, and Al-beta
zeolites were tested in the MPV reduction of cyclohexanone
using 2-butanol as reductant, the results given in Fig. 4
show that yields of cyclohexanol obtained with vanadium
and boron beta zeolites were much lower than the analogous
Sn-, Al-, and Ti-beta catalysts. The results obtained for V-
and B-beta materials are not surprising taking into account
the poor Lewis acid character shown by these materials as
measured by pyridine adsorption. This is probably due to
the fact that boron and vanadium have a high tendency to
come out from the framework during the calcinations of the
samples [20,21].
1 mmol of substrate, 60 mmol of 2-butanol, and 75 mg of catalyst)
Catalyst
Conversion
Selectivity (% mol)
(
wt% MeO or Si/Me ratio) (% mol)
2
Cyclohexanol Alquene Others
type
1 h
–
0
100
0
9.9
0
0
100
0
78.5
0
0
0
0
0
3.9
0
0
0
0
17.6
0
0
0
0
0
0
Al(iso-propoxide)
Ti(iso-propoxide)
TiCl
3
3
a
4
2% SnO /SiO
2
2
a
SnCl · 5H O
0
0
0
0
4
2
Si-beta
0
2
% SnO /Si-beta
0
0
0
2
Sn-beta (1% SnO )
64.5
95.4
95.8
91.0
63.2
94.5
6.1
31.5
57.5
23.7
80.3
100.0
100.0
100.0
99.5
98.8
99.2
100.0
95.2
95.5
97.0
98.4
0
0
0
0.1
0.2
0
0
0
2
Sn-beta (2% SnO )
2
119
Sn-beta (2% SnO )
2
0
b
Sn-beta (2% SnO )
0.4
1.0
0.8
0
4.8
3.1
2.3
1.3
2
c
Sn-beta (2% SnO )
2
Sn-beta (2% SnO2)
d
Ti-beta (2% TiO )
2
Al-beta (Si/Al = 50)
Al-beta (Si/Al = 15)
Al-beta (Si/Al = 15)
Al-beta (Si/Al = 12)
1.4
0.7
0.3
b
e
a
With 100 mg of the salt.
b
c
d
e
Alcohol/ketone ratio = 20/1 (mmol).
Alcohol/ketone ratio = 6/1 (mmol).
Zeolite after 4 recycles.
Synthetised and calcined the same as van Bekkum and collegues [9].
Comparative catalytic tests between Sn-beta and other
homogeneous and heterogeneous catalysts were carried out
using cyclohexanone as substrate and 2-propanol and 2-
butanol as reducing agents. The results given in Tables 2
and 3 clearly show that Sn-beta is a more active catalyst
than either Ti-, Al-beta, TiCl4, SnCl4, or SnO2, and almost
3
Fig. 4. Selective reduction of cyclohexanone with 2-butanol as reductant
over different Me-beta catalysts (MPV reaction) (reaction conditions: 1 h
of reaction, 1 mmol of substrate, 60 mmol of the alcohol, 75 mg of catalyst
◦
at 100 C).
as active as Al-(iso-propoxide) . These results indicate not

A. Corma et al. / Journal of Catalysis 215 (2003) 294–304
299
◦
◦
Fig. 5. Infrared spectra of cyclohexanone adsorption (at 25 C) and desorption at 50, 100, and 200 C over Si-beta (a), Sn-beta (b), Ti-beta (c), Al-beta (d),
and V-beta (e).
only the relative good catalytic performance of Sn-beta, but
also that the catalytic active sites in this catalyst are very
probably the Sn atoms in framework positions. If this is
so, these sites should be able to coordinate the ketone and
polarize the carbonyl group. To check this, cyclohexanone
for the pure silicon beta were obtained when SnO2 was
dispersed on the surface of this zeolite sample. For Sn-beta
−1
(Fig. 5b) a shift of 48 cm of the carbonyl band towards
◦
lower wavenumbers after desorption at 100 C occurred,
indicating the formation of an adduct between the carbonyl
oxygen and the tetrahedral Sn in the beta framework. This
◦
has been adsorbed on the Sn-beta in the IR cell at 25 C and
◦
◦
then desorbed at 50, 100, and 200 C for 1 h, while the IR
band persists even after desorption at 200 C. When Sn
spectra were recorded. The results are compared with those
obtained with the other zeolite catalysts in Fig. 5. It can be
seen there that the band associated to the carbonyl group at
is replaced by Ti a smaller shift of the carbonyl band
−1
(32 cm ) was observed (Fig. 5c) and elimination of the
◦
band occurred after desorption at 200 C. Both observations,
−1
1
713 cm is clearly visible on the pure silicon beta zeolite
i.e., the smaller shift and the earlier desorption with Ti-
beta, let us to conclude that the interaction of the carbonyl
oxygen and the titanium was weaker than the observed
for the tin. Incorporation of Al into the zeolite framework
◦
at 25 C. However, in this zeolite almost all cyclohexanone
has been desorbed at 100 C (Fig. 5a), and no carbonyl bands
◦
−1
below 1700 cm can be detected. Identical IR spectra as

300
A. Corma et al. / Journal of Catalysis 215 (2003) 294–304
involves various types of interactions with cyclohexanone
since besides Lewis acidity also Brönsted acidity is present
in this sample [22] (Fig. 5d), and the presence of various
broad carbonyl bands in the IR spectra could be observed.
Table 4
Comparative MPV reduction of cyclohexanone with selected sec-alcohols
over Sn-beta, Sn-MCM-41, and Al(iso-propoxide) (reaction conditions:
3
◦
1
00 C, 6 h of reaction, 1 mmol of substrate, 60 mmol of sec-alcohol and
75 mg of catalyst)
−1
For the V-beta sample (Fig. 5e) a lower shift (< 25 cm
)
Catalyst
wt% SnO )
Alcohol
Conversion Selectivity (% mol)
of the band associated to the carbonyl group occurs and the
a
(
(% mol)
2
Cyclohexanol Others
◦
shift disappears when temperature was increased to 100 C.
Sn-beta (2)
2-Butanol
98.6
95.5
32.0
30.0
100.00
99.08
74.40
73.22
0.00
0.72
25.60
26.78
These observations suggest that the interaction of the
carbonyl group with the Sn center is stronger than with Ti
and V, and more selective than with Al centers in Al-beta
zeolite. These interactions are a measure of the Lewis acidity
of the sites of the catalyst and would indicate that Lewis
acidity in V-beta is not strong enough, and/or the amount
of framework V is too small to carry out efficiently the
reaction. In the case of Al-beta, even if it is catalytically
active, the presence of residual Brönsted acidity catalyzes
undesired reactions and decreases the selectivity to the
desired products.
Cyclopentanol
Me-cyclohexanol
Phenyl-ethanol
Sn-MCM-41 (10)
2-Butanol
54.2
16.0
10.5
21.1
98.64
94.19
85.68
69.23
1.36
5.81
14.32
30.77
Cyclopentanol
Me-cyclohexanol
Phenyl-ethanol
b
Al(iso-propoxide)
2-Butanol
100.0
66.1
39.0
77.4
100.00
99.69
96.99
99.58
0.00
0.31
3.01
0.42
3
Cyclopentanol
Me-cyclohexanol
Phenyl-ethanol
a
(
% SnO ) present in the zeolite.
2
b
With 100 mg of the salt (11.67 wt% of Al).
3
.3. Effects of the molecular structure of the alcohol
A series of sec-alcohols were tested for the MPV reduc-
tion of cyclohexanone using Sn-beta as heterogeneous cata-
lyst. As shown in Fig. 6, all the alcohols studied are reactive,
though the smaller 2-propanol and 2-butanol give the best
activities and selectivities for the MPV reaction. The lower
activity achieved with bulkier alcohols such as cyclopen-
tanol and 2-methylcyclohexanol cannot be due to pore di-
ameter limitations of the zeolite since the pore diameters of
beta are large enough to compile with the size of the transi-
tion state formed between cyclohexanone and the above al-
cohols. Therefore, steric limitations should not occur in the
pore but probably occur near the Sn active site when increas-
ing the diameter of the secondary alcohol molecule or, more
precisely, when bulk substituted are present near the alcohol
function owing to the shielding effect of the oxygens bonded
to the Sn atom. To investigate this hypothesis, the reduction
of cyclohexanone with cyclopentanol was carried out with
Sn-MCM-41, which is also active for this reaction. It is clear
that in the case of mesoporous molecular sieves no limitation
due to pore dimensions should exist for the reaction between
cyclohexanone and cyclopentanol. Thus, since Sn-MCM-41
is active for the reduction of cyclohexanone with 2-butanol
but not with cyclopentanol (Table 4), we could certainly con-
clude that the reaction is not limited by pore dimensions.
Therefore, we propose that the observations made can be
due to geometrical limitations for the formation of the tran-
sition state on the Lewis acid, probably due to the presence
of the framework oxygen atoms bonded to the Lewis acid
site. Note that the activity of the aluminum iso-propoxide
also decreases when the bulkier alcohols are reacted.
Besides the geometrical effect described above it should
be noted that the presence of an aromatic ring in the phenyl-
ethanol reduces cyclohexanol yields. A similar behavior
was found with others secondary aromatics alcohols as
reductants when the alcoholic function is separated by one or
two carbon bonds with respect to the aromatic ring. All these
experimental observations let to us to conclude that presence
of the electron π cloud in the aromatic alcohol competes
with the OH group of the alcohol for the coordination on
the Lewis center resulting in a decrease of the yield of the
desired product. A similar effect is observed when using Al-
(
iso-propoxide)3 as catalyst (Table 4).
3
.4. Effects of the molecular structure of the ketones
By reacting a series of alkyl-cyclohexanones,with methyl
or tert-butyl substituents in different positions of the ring and
using 2-butanol and 2-propanol as the reducing agents, shape
selectivity effects of the Sn-beta catalyst are seen. As it was
found with Ti-beta [12], Sn-beta inhibits the formation of
Fig. 6. Selective reduction of cyclohexanone over Sn-beta (2% SnO ) with
2
different alcohols (MPV reaction) (reaction conditions: 6 h of reaction,
◦
1
mmol of substrate, 60 mmol of the alcohol, 75 mg of catalyst at 100–85 C
for 2-propanol).
4
-alkyl-trans-cyclohexanol isomers but not that of the cis-

A. Corma et al. / Journal of Catalysis 215 (2003) 294–304
301
Table 5
Selective reduction of alkyl-cyclohexanones over Sn-beta (2 wt% SnO )
2
◦
with secondary alcohols (MPVO reaction) (reaction conditions: 100 C,
a
6
h of reaction, 1 mmol of substrate, 60 mmol of alcohol , and 75 mg of
catalyst)
Substrate
Alcohol
Conversion
ketone
Selectivity (% mol)
c
Alcohol
C.P. Others
(% mol)
b
(
cis/trans)
Cyclohexanone 2-Propanol
96.8
98.6
54.2
100.0
100.0
98.6
0.0
0.0
0.0
0.0
0.0
1.4
2-Butanol
-Butanol
d
2
4-Me-cyclo-
hexanone
2-Propanol
95.8
96.5
72.4
100/0
100/0
100/0
0.0
0.0
0.0
0.0
0.0
0.0
2-Butanol
d
2-Butanol
4
-t-Butyl-cyclo- 2-Propanol
97.3
97.2
82.2
99/0.5
96/1.5
60/38.0
0.0
0.6
2.0
0.5
1.9
0.0
hexanone
2-Butanol
d
d
d
d
2
-Butanol
3
2
2
-Me-cyclo-
2-Butanol
2-Butanol
9.3
33.6
100.0
99.1
0.0
0.0
0.0
0.9
hexanone
-Me-cyclo-
hexanone
2-Butanol
2-Butanol
74.0
19.0
50/45.4
72.7/22.0
0.0
0.0
4.6
5.3
-t-Butyl-cyclo- 2-Butanol
0.0
0.0
0.0
0.0
◦
0.0
0.0
0.0
0.0
hexanone
a
2-Butanol
With 2-propanol the reaction temperature was 85 C.
b
Ratio of cis/trans alcohols evaluated by GC and GC-MS vs standard
compounds.
◦
c
Fig. 7. Infrared spectra of aromatic amines adsorption (at 25 C) and
Condensation products.
With Sn-MCM-41 (10 wt% SnO ) mesoporous material as catalysts.
◦
d
desorption at 150 C over Sn-beta zeolite: (a) 2,6-dimethylpyridine,
2
(b) 2-ethylpyridine, and (c) pyridine.
isomers (Table 5). Thus, the less favorable cis-isomers, cis-
-tert-butylcyclohexanol and cis-4-methylcyclohexanol, are
obtained in very high yield with respect to the trans-alcohol
isomers.
the methyl group has been the basis for using 2,6-dimethyl-
pyridine as probe molecule that absorbs on Brönsted but not
on Lewis acid sites [23].
4
More specifically, when cyclohexanone, 2-methylcyclo-
hexanone and 2,6-dimethylcyclohexanone have been ad-
sorbed on Sn-beta in the IR cell, the results presented in
In the case of 2-alkyl-cyclohexanones, the proximity of
the alkyl substituent to the carbonyl group induces steric
hindrance near the tin active site in the transition state
of the reaction. Thus, activity decreases when 2-methyl-
cyclohexanone is reacted instead of cyclohexanone, and it is
totally suppressed when the methyl is changed by the bulkier
t-butyl substituent in the 2 position. For the 3-methyl-
cyclohexanone, the distance between the methyl group and
the carbonyl group is not as determinant as the spatial
configuration (equatorial or axial) that can introduce steric
effects with respect to the walls of the zeolite pore. This
results in poor levels of 3-methylcyclohexanone conversion
compared with 2- and 4-methyl-substitued cyclohexanones.
The importance of hindrance at the framework Lewis acid
sites for alkyl-substituted rings in the vicinity of the electron
donating group of the reactant molecule has been proven
by adsorbing pyridine, 2-ethylpyridine, and 2,6-dimethyl-
pyridine on the Sn-beta sample. The results presented in
Fig. 7 show that while pyridine and 2-ethylpyridine are
◦
Fig. 8 show that after desorption at 100 C, only the two first
reactants remain adsorbed.
From the above discussion and the results given in Table 6
we concluded that transition-state limitation is a key factor
for explaining activities and selectivities found with certain
carbonyl compounds. Consequently, not only the size or
position of the substituent at the ring is important, but
also the type of substituent (alkyl or aryl) can play a very
important role in the coordination of the carbonyl group to
the active site. If we compare all the substrates in Table 6
and their corresponding conversions, we can correlate the
data of the cyclic or bicyclic ketones containing different
substitutes with the level of transition-state limitations that
those can originate. For instance, the fact that the diffusion
of cyclododecanone is restricted within the pores of zeolite
beta can justify the low level of conversion obtained with this
compound. In the case of the norcamphor–camphor couple,
the steric hindrance presented by the latest ketone due to the
methyl substituents absolutely inhibits the MPV-O process.
Nevertheless, the negative resonance or conjugation ef-
fect between the carbonyl band and C=C double bond be-
◦
coordinated to the Lewis acids after desorption at 150 C;
2
,6-dimethylpyridine does not owing to the negative steric
effect of the two methyl groups vicinal to the nitrogen atom
that coordinates to the Lewis center. This hindering effect of

302
A. Corma et al. / Journal of Catalysis 215 (2003) 294–304
◦
◦
Fig. 8. Infrared spectra of alkyl-cyclohexanone derivatives adsorption (at 25 C) and desorption at 50, 100, and 200 C over Sn-beta zeolite: (a) cyclohexanone,
(b) 2-methylcyclohexanone, and (c) 2,6-dimethylcyclohexanone.
comes very clear using cyclohexenone as substrate, taking
into account the low level of alcohol yield achieved with this
substrate compared with those of the above-noted cyclic ke-
tones (Table 6).
3.5. Stereoselectivity and chirality: Oppenauer reaction
As it can be seen in Table 7, when using a 1:1 reac-
tant mixture of cis- and trans-4-alkyl-cyclohexanols,Sn-beta
catalysts show excellent selectivities for reacting the cis-
alcohol, with respect to the trans isomer owing to geometri-
cal constraints for the transition state of the second. This is
in agreement with what occurs when reducing the 4-alkyl-
cyclohexanone and proves that both alcohol and ketone are
coordinated to the zeolite metal site in the transition state.
With respect to the enantioselective reduction of prochi-
ral ketones, it was found [12] that Ti-beta can produce 34%
of enatiomeric excess at 77% conversion after 20 days of re-
action time when phenyl-acetone is used as prochiral ketone
with (S)-2-butanol as reductant. We performed similar ex-
periments with Sn-beta catalysts and we achieved 43.0% e.e.
of the (S)-phenyl-2-propanol, at ∼ 70% conversion in only
In the case of ketones with aryl groups as substituents,
the evident increase in the conversion values in the order:
propiophenone < acetophenone < phenylacetone indicates
that the closer the aromatic ring is to the carbonyl group,
the bigger the steric effect produced. However, besides the
above-noted steric impediments, the activities achieved with
ketones containing aromatic groups can be influenced by
other effects, such as resonance and inductive effects of
the aromatic ring. Thus, the differences in conversion levels
observed for the last three substrates detailed in Table 6
cannot be explained following only the steric hindrance
hypothesis, suggesting that the electronic effects become
more important in this case.
4
h. At complete conversion of the phenyl-acetone (24 h) the
It must be taken into account that the interaction of the
carbonyl group with the metal center leads to polarization
of the carbonyl double bound enhancing the positive charge
on the carbon atom and the negative charge density on the
oxygen. When the aryl substituent group has an electron
donor effect (positive induction effect) stabilization of the
right polarization of the carbonyl double bond occurs,
increasing the effect of the catalyst and increasing the
conversion to the corresponding alcohol product. On the
other hand, a negative inductive effect as in a p-metoxi-
phenyl substituent leads to a poor polarization of the
carbonyl group and then the ketone conversion decreases.
We conclude that for a given metal, zeolite catalyst steric
effects at the Lewis site owing to the size of the transition
state, together with factors derived from the electronic
structure of the reactants, are responsible for the final
conversion observed.
e.e. decreases to 38.9%.
In conclusion, the combination of pore dimensions and
Lewis acidity in Sn-beta zeolites affords a highly active,
stereoselective and moderately enantioselective catalyst for
MPVO reactions.
3
.6. Stability and activity of the catalyst with water
in the reaction media
One of the limitations of using aluminum iso-propoxide
is its fast and irreversible deactivation by water. Moreover,
aluminum alkoxides dispersed on siliceous MCM-41-type
materials were demonstrated to be nonresistant to the pres-
ence of water traces in the MPVO reaction system [24]. In
the case of zeolites, it is known that highly hydrophobic
samples can be prepared. Thus, hydrophobic zeolites con-
taining Lewis acids can be quite suited for the MPVO re-

A. Corma et al. / Journal of Catalysis 215 (2003) 294–304
303
Table 6
Table 8
Selective reduction of different ketones over Sn-beta (2 wt% SnO ) with
Effect of the H O addition to the MPVO activity of Me-zeolites (reaction
conditions: 100 C, 1 h of reaction, 1 mmol of cyclohexanone, 60 mmol of
2-butanol and 75 mg of catalyst)
2
2
◦
◦
2-butanol (MPVO reaction) (reaction conditions: 100 C, 6 h of reaction,
mmol of substrate, 60 mmol of alcohol, and 75 mg of catalyst)
1
−
1
−1
h ) water added
Substrate
Structure
Conversion Selectivity (% mol)
ketone
Catalyst
TON (mol mol
a
Alcohol C.P. Others
0
0.2 g
0.5 g
(% mol)
Ti-beta
Al-beta
Al-beta
1.2
7.0
9.8
109.0
108.0
0.7
0.1
0.5
0.7
0.1
0.2
3.8
48.0
Cyclopentanone
Cyclohexanone
40.9
96.1 0.0
3.9
0.0
a
98.6
1.2
100.0 0.0
Sn-beta
Sn-beta–Sil.
17.8
56.7
b
a
Cyclododecanone
100.0 0.0
97.5 0.6
0.0
1.9
Synthetized the same as van Bekkum and collegues and calcined with
the same postsynthesis procedure [9].
b
Silylated sample.
4-t-Butyl-cyclo-
97.2
hexanone
action when water is present, since its low water adsorption
will have a lower effect on catalyst deactivation. Higher zeo-
lite hydrophobicitiesshould be achieved when no framework
charges or defects (silanols) are present. Taking this into ac-
count, one could expect the order of hydrophobicity of our
samples to be Ti-beta > Sn-beta > Al-beta. The results of
adsorption–desorption of water on these samples (Table 1)
show that indeed this order of hydrophobicity is the one pre-
sented above. This order of hydrophobicitycorresponds well
with the resistance of the catalysts to deactivation by water
(Table 8).
Indeed, Ti-beta and Sn-beta retain a higher percentage of
catalytic activity when water is present in the reaction media.
It is noticeable that even with ∼ 4 wt% of H2O (0.2 g) in the
media Sn-beta gives a higher turnover number than Ti- or Al-
beta when working in absence of water (Fig. 9). Moreover
the sample still retains a reasonable activity when ∼ 10 wt%
of H2O (0.5 g) was present. It appears then that preparing
hydrophobic metal-containing zeolites offers a possibility
for preparing solid Lewis acids that can work in the presence
of water.
Nopinone
12.4
10.1
80.9 0.0 19.1
Norcamphor
100.0 0.0
0.0 0.0
0.0
0.0
Camphor
0.0
25.8
18.9
Cyclohexenone
32.4 37.0 30.6
b
Acetophenone
93.9 6.1
87.4 12.6
0.0
0.0
b
Propiophenone
7.6
b
Phenyl-acetone
65.2
82.9
2.9
93.2 6.8
96.8 3.2
0.0
0.0
b
Phenoxy-acetone
p-Methoxy-phenyl-
0.0 0.0 100.0
b
acetone
a
Condensation products.
At 20 h of reaction.
b
Table 7
Oppenauer oxidation of 1/1 mixtures of cis- and trans-4-alkyl-cyclo-
hexanols over Sn-beta (2 wt% SnO ) with 2-butanone as oxidant (reaction
2
◦
conditions: 80 C, 6 h of reaction, 1 mmol of substrate, 80 mmol of ketone,
and 75 mg of catalyst)
Substrate
Time Conversion
Selectivity (% mol)
b
(h)
Alcohol
Ketone C.P.
Others
a
(cis/trans)
(% mol)
4
-Methylcyclohexanol
1
6
1
6
8.6/1.6
29.1/6.9
25.0/1.0
53.7/1.3
100.0
99.7
99.5
97.3
0.0
0.3
0.5
2.3
0.0
0.0
0.0
1.5
4-t-Butylcyclohexanol
Fig. 9. Loss of activity (as % of TON loss) for different Me-beta catalysts by
addition of water as poisoning agent (MPVO reaction) (reaction conditions:
1 h of reaction, 1 mmol of cyclohexanone, 60 mmol of 2-butanol, 75 mg of
catalyst at 100 C). (a) Synthetized and calcined the same as van Bekkum
and collegues [9].
a
Ratio of cis/trans alcohols evaluated by GC and GC-MS vs standard
◦
compounds.
b
Condensation products.

304
A. Corma et al. / Journal of Catalysis 215 (2003) 294–304
Thus, if catalyst hydrophobicity is an important issue in
References
this case, we can improve such a hydrophobicity by a zeolite
postsynthesis treatment that involves surface silylation [17].
When we did this using hexamethyl-di-silazane as silylating
agent, a Sn-beta-Sil sample was produced that contains
[1] C.F. de Graauw, J.A. Peters, H. van Bekkum, J. Huskens, Synthesis 10
(1994) 1007.
1
.63 wt% of organic and shows a very high catalytic activity
[2] E.J. Creyghton, J. Huskens, J.C. van der Waal, H. van Bekkum, Stud.
Surf. Sci. Catal. 108 (1997) 531.
even when ∼ 10 wt% of H2O was present in the reaction
[3] J.C. van der Waal, P.J. Kunkeler, K. Tan, H. van Bekkum, Stud. Surf.
media (Table 8).
Sci. Catal. 110 (1997) 1015;
J.C. van der Waal, P.J. Kunkeler, K. Tan, H. van Bekkum, J. Catal. 173
4
. Conclusions
(1998) 74.
[
4] T. Markert, Henkel KGAA - DE, WO 1998 EP 05689, 1998.
Sn-beta is an efficient catalyst for the MPV reductions
[5] R. Anwander, C. Palm, G. Gerstberger, O. Groeger, G. Engelhardt,
J. Chem. Soc. Chem. Commun. (1998) 1811.
of carbonyl groups and Oppenauer’s oxidations of alcohols.
The structure as well as pore dimensions added to its
particular Lewis acid properties allows excellent conversion
and selectivity to the corresponding alcohol product. Thus,
the catalyst can be recycled several times and alcohol/ketone
ratios as low as 6 or 3 can be used without important loss
in the activities, this being of clear interest for a potential
industrial application of the catalyst.
[6] R. Anwander, C. Palm, Stud. Surf. Sci. Catal. 117 (1998) 413.
[7] A. Choplin, B. Coutant, C. Dubuisson, P. Leyrit, C. McGill, F. Quig-
nard, R. Teissier, Stud. Surf. Sci. Catal. 110 (1997) 353.
[8] M.A. Aramendía, V. Borau, C. Jiménez, J.M. Marinas, F.J. Romero,
Catal. Lett. 58 (1999) 53.
[9] P.J. Kunkeler, B.J. Zuurdeeg, J.C. van der Waal, J.A. van Bokhoven,
D.C. Koningsberger, H. van Bekkum, J. Catal. 180 (1998) 234.
[
10] J.C. van der Waal, E.J. Creyghton, P.J. Kunkeler, H. van Bekkum, Top.
Catal. 4 (1997) 261.
This MPVO catalyst shows good stereoselectivity an
reasonable enantioselectivity for reducing a prochiral ketone
with a chiral alcohol.
[
11] E.J. Creyghton, S.D. Ganeshie, R.S. Downing, H. van Bekkum, J. Mol.
Catal. A 115 (1997) 457.
By studying the reactivity of alcohols and ketones with
different molecular structures, it has been shown that the
reaction transition states that by size can be fitted within
the pore of Beta zeolite may have problems forming on the
framework Lewis acid sites due to the shielding effect of the
neighbor framework oxygen atoms.
The better activity shown by Sn-beta catalysts compared
with the analogous Ti- and Al-beta zeolites might be
attached to the adequate Lewis acidity of the tin active sites
in the zeolite framework. It is possible to prepare water-
resistant MPVO catalysts, by synthesizing hydrophobic
zeolites. The water resistance can be further improved by
postsynthesis sylilation of the zeolite. These catalysts are
highly active and selective for MPVO process even in the
presence of ∼ 10 wt% of H2O.
[
[
[
12] J.C. van der Waal, K. Tan, H. van Bekkum, Catal. Lett. 41 (1996) 63.
13] A. Corma, L.T. Nemeth, M. Renz, S. Valencia, Nature 412 (2001) 423.
14] A. Corma, M.E. Domine, L.T. Nemeth, S. Valencia, J. Am. Chem.
Soc. 124 (13) (2002) 3194.
[15] A. Corma, S. Valencia (UOP LLC), US Patent 5,968,473, 1999.
[16] T. Blasco, M.A. Camblor, A. Corma, P. Esteve, J.M. Guil, A. Martínez,
J.A. Perdigón-Melón, S. Valencia, J. Phys. Chem. B 102 (1998) 75.
[17] A. Corma, M. Domine, J.A. Gaona, J.L. Jordá, M.T. Navarro, F.
Rey, J. Pérez-Pariente, T. Tsuji, B. McCulloch, L.T. Nemeth, Chem.
Commun. (1998) 2211.
[18] A. Corma, M.E. Domine, J.A. Gaona, M.T. Navarro, F. Rey, S. Valen-
cia, Stud. Surf. Sci. Catal. (2001) 1812.
[19] I.W.C.E. Arends, R.A. Sheldon, M. Wallau, U. Schuchardt, Angew.
Chem. Int. Ed. Engl. 36 (1997) 1144.
[20] K.R. Reddy, A.V. Ramaswamy, P. Ratnasamy, J. Catal. 143 (1993)
275.
[21] R. de Ruiter, K. Pamin, A.P.M. Kentgens, J.C. Jansen, H. van Bekkum,
Zeolites 13 (1993) 611.
Acknowledgments
[
[
[
22] J. Dabrowski, M. Katcha, J. Mol. Struct. 12 (1972) 179.
We gratefully acknowledge financial support by CICyT
23] A. Corma, C. Rodellas, V. Fornes, J. Catal. 88 (1984) 374.
(
Spanish MAT2000-1392) and M.E.D. thanks U.P.V. for
24] J. Wahlen, D.E. De Vos, M. De Bruyn, P.J. Grobet, P.A. Jacobs, Stud.
Surf. Sci. Catal. 135 (2001) 3814.
supporting his doctoral fellowship.