Liquid and gas-phase Meerwein-Ponndorf-Verley reduction of crotonaldehyde on ZrO2 catalysts modified with Al2O3, Ga 2O3 and In2O3

Article abstract of DOI:10.1016/j.molcata.2011.02.005

A series of catalysts consisting of ZrO₂ modified with Al, Ga and In by coprecipitation with their precursor salts or impregnation of previously synthesized ZrO₂ with the nitrates of the metals was prepared. The solids thus obtained were characterized using a wide range or techniques including TG/DTG; XRD, SEM-EDAX, ICP-MS, FTIR and FT-Raman spectroscopies; and nitrogen adsorption-desorption at 77 K. The catalysts prepared by impregnation were found to contain Al, Ga and In nitrates while those obtained by coprecipitation contained Al₂O₃, Ga ₂O₃ and In₂O₃. The catalysts were analyzed for surface acidity and basicity by thermal programmed desorption of pyridine and CO₂, respectively. The impregnated solids calcined at 300 °C exhibited low basicity and moderate acidity due to Br?nsted and Lewis sites. The coprecipitated solids were slightly more acid and basic than their impregnated counterparts, their acidity being mainly due to the presence of Br?nsted sites. Both types of solids were used as catalysts in the Meerwein-Ponndorf-Verley reduction of crotonaldehyde with 2-propanol as hydrogen donor. The impregnated solids calcined at 300 °C (particularly the Ga/ZrO₂ catalyst) were the most active and selective ones in the liquid phase reaction. The selectivity towards 2-butenol in the reaction in the gas phase was high at temperatures up to 250 °C, above which it dropped in a rapid manner. Again, the impregnated solids calcined at 300 °C were the most active and selective systems in the process (particularly Ga/ZrO ₂).

Full text of DOI:10.1016/j.molcata.2011.02.005

Journal of Molecular Catalysis A: Chemical 338 (2011) 121–129
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Liquid and gas-phase Meerwein–Ponndorf–Verley reduction of crotonaldehyde
on ZrO2 catalysts modified with Al2O3, Ga2O3 and In2O3
Juan F. Min˜ambres, María A. Aramendía, Alberto Marinas, José M. Marinas, Francisco J. Urbano∗
Department of Organic Chemistry, University of Córdoba, Campus de Rabanales, Marie Curie Building (Annex), E-14014 Córdoba, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of catalysts consisting of ZrO2 modified with Al, Ga and In by coprecipitation with their precur-
sor salts or impregnation of previously synthesized ZrO2 with the nitrates of the metals was prepared.
The solids thus obtained were characterized using a wide range or techniques including TG/DTG; XRD,
SEM-EDAX, ICP-MS, FTIR and FT-Raman spectroscopies; and nitrogen adsorption–desorption at 77 K. The
catalysts prepared by impregnation were found to contain Al, Ga and In nitrates while those obtained
by coprecipitation contained Al2O3, Ga2O3 and In2O3. The catalysts were analyzed for surface acidity
and basicity by thermal programmed desorption of pyridine and CO2, respectively. The impregnated
solids calcined at 300 ◦C exhibited low basicity and moderate acidity due to Brønsted and Lewis sites.
The coprecipitated solids were slightly more acid and basic than their impregnated counterparts, their
acidity being mainly due to the presence of Brønsted sites. Both types of solids were used as catalysts in
the Meerwein–Ponndorf–Verley reduction of crotonaldehyde with 2-propanol as hydrogen donor. The
impregnated solids calcined at 300 ◦C (particularly the Ga/ZrO2 catalyst) were the most active and selec-
tive ones in the liquid phase reaction. The selectivity towards 2-butenol in the reaction in the gas phase
was high at temperatures up to 250 ◦C, above which it dropped in a rapid manner. Again, the impregnated
solids calcined at 300 ◦C were the most active and selective systems in the process (particularly Ga/ZrO2).
Received 15 December 2010
Accepted 6 February 2011
Available online 12 February 2011
Keywords:
Zirconia based catalysts
Surface acidity
Surface basicity
Meerwein–Ponndorf–Verley reduction
Crotonaldehyde
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
Traditionally, the MPV reaction has been conducted using a
metal (Al, Zr) alkoxide as catalyst in a homogeneous process. The
The reduction of carbonyl compounds by hydrogen transfer
from an alcohol is known as the “Meerwein–Ponndorf–Verley reac-
reaction mechanism involves the formation of a six-membered
cyclic intermediate where both reactants coordinate to the same
metal site in the alkoxide [4–6]. The past two decades, however,
have seen a rise in research aimed at facilitating conduct of the pro-
cess in a heterogeneous phase on account of the major advantages
of operating in this way in large-scale processes [5]. The reduc-
tion of ␣,␤-unsaturated carbonyl compounds by hydrogen transfer
under heterogeneous catalysis has so far been studied in the pres-
ence of a variety of catalysts including magnesium hydroxides [7],
magnesium oxides [8–10], calcined hydrotalcites [11], hydrous and
calcined zirconia [8,12,13], and a wide range of active components
supported on zeolitic [4,14], mesoporous [15,16] and various other
materials [17].
Although the mechanism behind these heterogeneous hydro-
gen transfer processes is seemingly quite clear, there remains some
uncertainty as to the respective roles of surface acid (Lewis or
Brønsted) and/or basic sites in the catalysts. Rather than a unified
mechanism, researchers have proposed a number of them depen-
dent on the particular catalyst used in the heterogeneous MPV
reaction. Thus, the reduction of carbonyl compounds by hydro-
gen transfer has been hypothesized to occur at Lewis acid sites [4],
tion” or “MPV reaction” in Organic Chemistry. The presence of a C
C
double bond conjugated with the C O group in an ␣,␤-unsaturated
carbonyl compound introduces an additional dimension in the
process: the chemoselective reduction of the C O group in the
presence of the C C bond, which leads to the formation of an
␣,␤-unsaturated alcohol (Scheme 1). This selective synthesis for
primary and secondary alcohols is an important process for the
pharmaceutical, fragrance and food flavouring industries. Their
preparation by catalytic hydrogenation with a metal-supported
catalyst is rather difficult owing to the high reactivity of the C
bond relative to the carbonyl group [1]. However, the MPV reaction
has been successfully used for the selective reduction of the C
C
O
bond in ␣,␤-unsaturated carbonyl compounds to the correspond-
ing unsaturated alcohols [2,3].
∗
Corresponding author. Tel.: +34 957218638; fax: +34 957212066.
E-mail address: FJ.Urbano@uco.es (F.J. Urbano).
1381-1169/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2011.02.005

122
J.F. Mi˜nambres et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 121–129
H
H
R
3
R
1
O
[H]
R
2
R
4
[H]
Saturated Aldehyde (SAL)
R
3
R
H
H
3
H
OH
R
1
O
R
1
R
2
R
4
R
2
R
4
Saturated alcohol (SOL
Unsaturated aldehyde (UAL)
[H]
[H]
R
3
H
R
1
OH
R
2
R
4
Unsaturated alcohol (UOL)
Scheme 1. Reaction network for the reduction of ␣,␤-unsaturated carbonyl compounds.
Brønsted acid sites [12,13], basic sites [18,19] and acid–base pairs
[10,20].
A similar procedure was used to obtain ZrO2 catalysts modi-
fied with a 10 mol.% concentration of Al2O3, Ga2O3 or In2O3 by
coprecipitation. The starting solution contained the correspond-
ing precursors (viz. zirconium oxychloride and Al, Ga or In nitrate),
appropriate amounts of which were dissolved in 500 mL of Milli-Q
water and treated similarly as before. The resulting solids were also
calcined at 175 or 300 ◦C for 6 h.
The reference zirconium oxide above described was split into
four portions three of which were modified by impregnation with
an aqueous solution of aluminium, gallium or indium nitrate (from
Sigma–Aldrich). To this end, 6 g of ZrO2 was suspended in 10 mL of
Milli-Q water and supplied with the amount of nitrate needed to
obtain a 10 mol.% concentration of the corresponding oxide in the
final solid. The mixture was placed in a rotavapor for 2 h to obtain
a homogeneous paste. Residual solvent was then evaporated by
evacuation in a water bath at 60–80 ◦C. Once dry, the solid was
calcined at 175 and 300 ◦C for 6 h, ground and sieved.
The catalysts were designated with the symbol for the
zirconium-modifying element (Al, Ga or In), followed by that
for the synthetic method used (“co” for coprecipitation and “im”
for impregnation) and the calcination temperature used (175 or
300 ◦C). For example, the ZrO2 catalyst containing 10 mol.% Al2O3,
obtained by coprecipitation and calcined at 300 ◦C was designated
Alco300.
Zirconia has proved a highly effective choice among heteroge-
neous catalysts used in the MPV reaction [13]. Zirconium oxide is a
solid with a high thermal stability and corrosion resistance in addi-
tion to a strong amphoteric character [21]. Textural and acid–base
properties of ZrO2 depend largely on its synthetic procedure
and calcination temperature. Adjusting its acid–base properties is
possible by modifying its surface with sulphate ions [22–24], phos-
phate ions [25] and mixtures of other oxides [3,26,27]; this has
proved a highly effective method for tailoring the activity of zirco-
nia towards many organic processes.
Some studies revealed catalytic activity in the MPV reaction to
decrease with increasing calcination temperature (i.e. with increas-
ing loss of surface hydroxyl groups) in hydrous zirconia; this
underlines the significance of proton (Brønsted) sites to the process
[12,13]. Moreover, it has been recently reported that the incorpora-
tion of boron into ZrO2 leads to an increased catalytic activity in the
liquid-phase MPV reduction of cinnamaldehyde consistent with an
increased amount of highly strong proton (Brønsted) acid sites [3].
Modifying the surface chemical properties of ZrO2 through the
incorporation of Al2O3, Ga2O3 or In2O3 could allow to study the
role of surface acid (Brønsted or Lewis) or basic sites on the activ-
ity and selectivity of the MPV reduction of crotonaldehyde with
2-propanol.
2.2. Textural and structural characterization
2. Experimental
Gels were subjected to thermogravimetric and differential ther-
mal analysis on a Setaram Setsys 12 system, using Air at 40 mL/min
as carrier gas, ␣-Al2O3 as reference material and a Pt/Pt–Rh (10%)
thermocouple. The heating rate was 10 ◦C/min and the tempera-
ture range 30–1000 ◦C. The amount of gel used in each test was ca.
20 mg.
The textural properties of the solids were determined from
nitrogen adsorption–desorption isotherms obtained at liquid nitro-
gen temperature on a Micromeritics ASAP-2010 instrument. All
samples were degassed to 0.1 Pa at 110 ◦C prior to measurement.
Surface areas were calculated using the Brunauer–Emmett–Teller
(BET) method [28].
2.1. Synthesis of catalysts
The studied catalysts were prepared by modifying zirconium
oxide with 10% Al2O3, Ga2O3 or In2O3, either by coprecipitation
with their precursor salts or by impregnation of previously synthe-
sized pure ZrO2, which was used as reference.
The reference zirconium oxide was obtained from zirconium
oxychloride octahydrate (from Sigma–Aldrich). To this end, an
appropriate amount of reagent was dissolved in 500 mL of water
and precipitated by dropwise addition of 5 N NH4OH to pH 9.5. The
precipitate was allowed to stand overnight, filtered and washed
with Milli-Q water as many times as required to give a negative
chloride test with silver nitrate. The resulting solid was dried at
110 ◦C for 6 h, ground and sieved prior to calcination at 300 ◦C for
6 h, using a heating rate of 2 ◦C/min.
X-ray diffraction patterns were obtained on a Siemens D5000
diffractometer equipped with a graphite monochromator and using
Co K␣ radiation. The 2ꢀ angle was scanned from 5 to 75◦ with a step
size of 0.05◦.

J.F. Mi˜nambres et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 121–129
123
As far as EDX analyses are concerned, they were performed on a
JEOL JSM-6300 SEM apparatus operating at an accelerating voltage
of 20 keV with a resolution of 65 eV. EDX values corresponded to
the average value of three measurements carried out at different
areas of the solid with amplifications of 600 and 1000×.
the adsorption of the probe molecule, the catalyst (100 mg) was
cleaned by passing an Ar stream (56 mL min−1) up to 300 ◦C (at
10 ◦C min−1) and cooling down in Ar to 50 ◦C. The solids were
then saturated by passing a CO2/Ar stream (56 mL min−1) at 50 ◦C.
Subsequently, a pure Ar stream (56 mL min−1) was passed at the
saturation temperature for 2 h in order to remove any physisorbed
molecules. Once a stable line was obtained, chemisorbed CO2 was
desorbed by heating from saturation temperature up to 300 ◦C in
a programmed fashion, at a rate of 10 ◦C min−1. The selected peaks
were monitored through the whole process. Quantification was
based on the 5% (v/v) CO2/Ar standard.
The elemental analysis of the catalysts was performed on a
Perkin–Elmer ELAN DRC-e ICP-MS instrument following digestion
of the samples in a 1:1:1 mixture of HF, HNO3 and H2O, and dilution
in 3% HNO3. Calibration samples were prepared from appropri-
ate atomic spectroscopy standards (PE Pure Plus, Perkin–Elmer)
in HNO3 (10 ␮g/mL of each metal). Calibration curves were con-
structed over the concentration range 1–100 ppb and included the
results for a blank.
2.4. Meerwein–Ponndorf–Verley reaction
FT-IR spectra were recorded over
a wavenumber range
400–4000 cm−1 on a Bomen MB-100 FT-IR spectrophotometer. The
pellets were prepared by mixing the powdered solid with KBr in a
5:95 (w/w) ratio.
The MPV reaction was conducted in both the liquid phase and
the gas phase. Tests in the liquid phase were performed in a two-
mouthed, round-bottom flask one mouth of which was used to
introduce a 0.5 M solution of crotonaldehyde in isopropyl alcohol
and 0.5 g of catalyst. The flask was fitted with a reflux condenser and
placed in an ethyleneglycol bath that was kept at 130 ◦C through-
out the reaction. The reaction medium was shaken in a continuous
manner for 8 h and 0.2 mL aliquots were withdrawn from it at dif-
ferent times during the process and passed through a nylon filter
of 0.45 ␮m pore size prior to analysis.
FT-Raman spectra were obtained on a Perkin–Elmer 2000 NIR
FT-Raman system equipped with a diode pumped NdYAG laser
(9394.69 cm−1) that was operated at 300 mW laser power and a
resolution of 4 cm−1 throughout the 3500–200 cm−1 range in order
to gather 64 scans.
2.3. Surface acid–base properties
Tests in the gas phase involved using 50 mg of catalyst in
a cylindrical reactor 10 mm in diameter that was placed in a
tubular oven equipped with a 6-segment temperature controller.
Prior to reaction, the solid was heated in a synthetic air stream
at 300 ◦C for 30 min. Then, the catalyst was allowed to cool
down to 200 ◦C (the reaction temperature). The reaction was
started by replacing the synthetic air stream with a nitrogen
stream flowing at 50 mL/min and carrying a 0.5 M solution of
crotonaldehyde in isopropyl alcohol. The solution was injected
at a rate of 1 g/h via a Bronkhorst High-Tech liquid mass flow
controller and evaporated at 130 ◦C in a CEM mixer/evaporator
(Bronkhorst High-Tech).
The above-described, isothermal tests were supplemented with
others in the gas phase which were used to monitor catalyst activ-
ity and selectivity at variable temperatures from 150 to 300 ◦C
obtained by heating at 10 ◦C/min. The highest temperature stud-
ied, 350 ◦C, was held for 30 min and followed by cooling down to
150 ◦C.
Products analysis was carried out on a Fisons Instruments
GC 8000 Series gas chromatograph furnished with a 30 m long,
0.53 mm i.d. Supelcowax-10 semi-capillary column and fitted to
a flame ionization detector (FID). The reaction products obtained
for the crotonaldehyde (UAL) reduction were the crotyl alcohol
(unsaturated alcohol, UOL), 1-butanol (saturated alcohol, SOL) and
butanal (saturated aldehyde, SAL). The initial reduction rate of cro-
tonaldehyde was expressed in terms of catalyst weight (g) and
surface area (m2) and the selectivity towards the unsaturated alco-
hol at variable conversion levels was calculated from the following
expression:
Acid–base properties were assessed by temperature-
programmed desorption-mass spectrometry (TPD-MS) of pyridine
(PY) for total acidity and CO2 for total basicity. TPD-MS experiments
were carried out on a Micromeritics TPD-TPR 2900 instruments
fitted to a VG PROLAB Benchtop QMS (Thermo Scientific).
Pyridine was the probe molecule used to determine the acid
properties of the catalysts. The base peak (m/z = 79) as well as a
secondary one (m/z = 52, 80% abundance) was selected to be mon-
itored in the mass spectrometer. Prior to the adsorption of the
probe molecule, the catalyst (100 mg) was cleaned by passing an
Ar stream (56 mL min−1) up to 300 ◦C (at 10 ◦C min−1) and cooling
down in Ar to 30 ◦C. The solids were then saturated by passing an Ar
stream (56 mL min−1) at room temperature through a pure amine
solution until complete saturation of the catalyst (about 30 min).
Subsequently, a pure Ar stream (56 mL min−1) was passed at the
saturation temperature for 2 h in order to remove any physisorbed
molecules. Once a stable line was obtained, chemisorbed PY was
desorbed by heating from saturation temperature up to 300 ◦C
in a programmed fashion, at a rate of 10 ◦C min−1. The selected
peaks were monitored through the whole process. Calibration was
done by injecting pulses of variable size of a pyridine solution in
cyclohexane. In parallel experiments, the solids saturated with PY,
before being ramped, were analyzed ex situ by FT-Raman spec-
troscopy in order to distinguish between Brønsted (H-bonding or
proton donor) or Lewis acid sites. The most sensitive Raman vibra-
tion of pyridine is its symmetric ring breathing (␯CCN) (vs, v1, A1),
which appears at 991 cm−1 in liquid pyridine. The interaction of
pyridine with acid sites induces a shift in this band to a higher
Raman Shift values. Therefore, the position of the skeletal vibra-
tion band can be used to detect interactions between pyridine and
SUOL=(mol crotyl alcohol/mol crotonaldehyde converted) × 100
protonic weak acid sites through hydrogen bonds (996–1008 cm−1
)
or its chemisorption at strong Brønsted (1007–1015 cm−1) and/or
Lewis acid sites (1018–1028 cm−1) on a solid surface [29–31].
Spectra were collected as stated above and processed with the
software PeakFit v. 4.11 in order to determine the components
for physisorbed and chemisorbed pyridine in their three variants
(hydrogen bonding interactions, Brønsted sites and Lewis sites).
Carbon dioxide (5% CO2 in Argon) was the probe molecule used
to determine the basic properties of the catalysts. The base peak
(m/z = 44) as well as a secondary one (m/z = 12, 10% abundance)
was selected to be monitored in the mass spectrometer. Prior to
3. Results and discussion
3.1. Textural and structural characterization of the catalysts
Table 1 shows the chemical composition and the most relevant
textural, structural and surface chemical properties of the cata-
lysts. Although their syntheses were conducted in such a way as
to obtain a 10 mol.% concentration of modifying oxide, the ICP-MS
results revealed that this theoretical content was never reached; in
fact, the molar proportions of the oxides ranged from 3.4 to 5.6%

124
J.F. Mi˜nambres et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 121–129
Table 1
Chemical composition (SEM-EDAX e ICP-MS) and textural (SBET), structural (TGLOW) and surface chemical (acid–base) properties of the synthesized catalysts.
Catalyst
DTA–TGLOW (◦C)
SBET (m2/g)
Chemical composition
SEM-EDAX (mol.%)
Surface acidity
Surface basicity
ICP-MS (mol.%)
Pyridine (␮mol/g)
Pyridine (␮mol/m2)
CO2 (␮mol/g)
CO2 (␮mol/m2)
Alco300
Gaco300
Inco300
Alim300
Gaim300
Inim300
ZrO2
608
506
445
550
509
459
430
220
163
173
167
175
161
256
4.9
5.4
4.0
9.6
8.2
6.3
–
4.2
5.1
5.6
4.8
5.5
3.4
–
32.3
31.2
22.7
24.3
22.1
20.2
32.3
0.15
0.19
0.13
0.15
0.13
0.13
0.18
329
192
302
49
86
64
1.49
1.18
1.75
0.29
0.49
0.40
0.46
119
Alco175
Gaco175
Inco175
Alim175
Gaim175
Inim175
608
506
445
550
509
459
276
215
267
136
167
151
4.9
5.4
4.0
9.6
8.2
6.3
4.2
5.1
5.6
4.8
5.5
3.4
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
and were essentially similar in the coprecipitated and impregnated
solids. The surface chemical data obtained by SEM-EDAX revealed
more marked surface enrichment with Al, Ga or In in the impreg-
nated solids than in their coprecipitated counterparts. Thus, the
coprecipitated catalysts, consistent with the ICP-MS results, hardly
reached 5 mol.% levels of the added oxides; by contrast, the impreg-
nated catalysts exhibited increased surface concentrations of the
oxides (particularly Alim and Gaim, with 9.6 and 8.2 mol.%, respec-
tively) in greater consistency with the expected outcome for the
ZrO2 impregnation method.
All dry gels obtained immediately before calcination (i.e. all
pre-catalysts) were subjected to thermogravimetry (TG–DTG) and
thermal differential analysis (DTA). The TG–DTG results differed
considerably between coprecipitated and impregnated catalysts.
By way of example, Fig. 1 shows the thermal profiles for the pre-
cursor gels of the catalysts Alco and Alim.
The TG curves for the precursors revealed a similar weight
loss (10–15%) in all solids. The curves exhibited two well-defined
bands, namely: one at ca. 145 ◦C due to the loss of coordination
water and leading to an amorphous zirconium hydroxide and the
other, starting at 250 ◦C and ending at 400 ◦C, due to the gradual
loss of hydroxyl groups to form an amorphous zirconium oxide
[32]. Based on heat flow, all catalysts underwent an atomic rear-
rangement at ca. 400–600 ◦C by which the zirconium oxide gained
structural order via crystallization. This reflected in an exothermic
peak associated to no weight loss and known as “glow exotherm”
the exact position of which is determined by the temperature of
maximum exothermic flow, TGLOW (Table 1) [32,33]. Crystalliza-
tion in our pure ZrO2 occurred at 430 ◦C; however, the presence
of additives in the other solids raised TGLOW (especially in the
aluminium-containing solids). Based on the TGLOW values of Table 1
and the calcination temperatures used (175 and 300 ◦C), the studied
solids were amorphous in nature, as confirmed by the XRD results.
In fact, the XRD patterns for both the coprecipitated catalysts and
their impregnated counterparts contained no diffraction peaks or
bands.
Also, the FT-Raman spectra (Fig. 2) provided virtually no useful
information about the catalysts owing to their amorphous nature,
which was previously confirmed by the XRD results. All spectra
were essentially identical, the sole difference being the presence of
a sharp peak at 1052 cm−1 for the impregnated catalysts that was
assigned to the corresponding Al, Ga and In nitrates (see inset of
Fig. 2). In fact, the nitrates incorporated into the zirconium oxide
duringthe impregnationprocessmust decompose only partly atthe
low–medium calcination temperature used (300 ◦C) and remain to
some extent on the surface of the final catalyst. The absence of this
sharp peak from the spectra for the coprecipitated catalysts can be
ascribed to the supplied nitrates being completely transformed into
hydroxides by precipitation with NH4OH. The subsequent washing
process removed the residual nitrates from the solids.
The infrared (FTIR) spectra for the catalysts were identical with
one another and also with that for the reference zirconium oxide
except for the presence of a strong band at 1385 cm−1 due to the
presence of residual nitrate in the impregnated solids (Fig. 3), in
agreement to the results obtained for FT-Raman. This assignation
was confirmed by obtaining the FTIR spectra for aluminium gallium
and indium nitrates calcined at 300 ◦C for 6 h (not shown). All other
bands in the IR region were assigned to zirconium oxide. Thus, the
broad band at 3400–3500 cm−1 was due to hydroxyl groups in the
hydrated oxide and slightly weaker in the calcined solids, which,
0,8
0
Alco
20
10
0
0,6
0,4
-2
DTG
DTA
TG
-4
-6
0,2
-8
-10
-12
-14
-16
0,0
-10
-20
-0,2
-0,4
0
200
400
600
800
1000
Temperature (ºC)
0,8
0,6
0,4
0,2
0,0
-0,2
Alim
0
20
10
0
DTG
DTA
TG
-2
-4
-6
-8
-10
-12
-14
-16
-10
-20
0
200
400
600
800
1000
Temperature (ºC)
Fig. 1. TG/DTA profiles obtained for the Alco and Alim precursor gels.

J.F. Mi˜nambres et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 121–129
125
consistent with the TG–DTG results, indicates the loss of OH groups
upon calcination at 300 ◦C. The band at ca. 1600 cm−1 was assigned
to atmospheric CO2 adsorbed as bidentate carbonate. Finally, the
strong, broad band at 470 cm−1 corresponded to Zr–O bonds. How-
ever, the typical band at 1000 cm−1 for Zr O bonds was absent
from the spectra, which indicates that the zirconium oxide was in
the form of a polymer with its zirconium atoms linked by oxygen
bonds [32].
Al (NO ) FT-Raman
As regard to textural properties, all isotherms obtained (results
not shown) were of type IV in the IUPAC classification and exhibited
an H2 hysteresis cycle associated to bottleneck pores. Based on pore
3000
2000
1000
ZrO2
-1
Raman Shift (cm
)
˚
size distribution, the average pore size was less than 50 A.
Inim300
Gaim300
The specific surface area of all solids calcined at 300 ◦C was
essentially similar. By contrast, that of the solids calcined at 175 ◦C
differed markedly. In fact, areas ranged from 161 m2/g for Inim300
to 220 m2/g for Alco300 among the former, and from 276 m2/g for
Alco175 to 136 m2/g for Alim175 in the latter. Also, the solids modi-
fied with a given metal (Al, Ga or In) had greater surface areas when
synthesized by coprecipitation than when obtained by impregna-
tion – the opposite was only true for the Ga-modified catalysts
calcined at 300 ◦C. It therefore seems obvious that impregnation of
the reference ZrO2 following calcination at 300 ◦C results in partial
clogging of the pore network by the modifier and, consequently, in
a decreased surface area relative to the support. On the other hand,
the pore network in the coprecipitated solids forms after the mod-
ifier is structurally incorporated and the resulting specific surface
area is greater – and close to that for the reference zirconium oxide
– as a result.
Alim300
Alco300
Gaco300
Inco300
3500 3000 2500 2000 1500 1000 500
Raman Shift (cm-1)
3.2. Surface chemical properties of the catalysts
Surface acidity and basicity in the catalysts were determined by
thermal programmed desorption of two different probes (pyridine
and CO2, respectively) following chemisorption at surface sites.
Table 1 shows the surface acidity and basicity results obtained
by integrating the areas under the pyridine and CO2 desorption
curves, respectively. The catalysts exhibited similar pyridine des-
orption profiles (not shown); pyridine desorption started around
70 ◦C and peaked at 140–150 ◦C. Based on the data of Table 1,
Alco300, Gaco300 and ZrO2 catalysts possessed the highest acidity
per gram of solid. On the other hand, Inim300, Gaim300, Inco300
and Alim300 were the solids with the lowest surface acidity. The
high surface area of Alco300 and, especially, the reference ZrO2
resulted in a low surface acidity (per square meter) relative to
Gaco300. Overall, the coprecipitated catalysts were more acid
than their impregnated counterparts; this was particularly so with
Alco300 and Gaco300. In fact, the impregnated solids were less acid
than the reference ZrO2.
Fig. 2. FT-Raman spectra obtained for the catalyst synthesized in this work. The
inset shows the FT-Raman spectrum obtained for the aluminium nitrate calcined at
300 ◦C for 3 h.
Alco300
Gaco300
Inco300
Alim300
The fact that the pyridine TPD profiles for the catalysts failed
to clearly distinguish the presence of distinct types of acid sites
led us to conduct an ex situ FT-Raman study of the solids fol-
lowing saturation with chemisorbed pyridine at exactly the point
immediately preceding the start of the above-described thermal
programmed desorption tests. Fig. 4 shows the FT-Raman signals
for chemisorbed pyridine in the region of its symmetric ring-
breathing vibration (vs, v1, A1). The interaction of pyridine with
surface acid sites shifts the band to higher Raman wavenumbers.
As a result, the position of the band can help to identify whether the
interaction occurs via hydrogen bonds (996–1008 cm−1) or with
the pyridine chemisorbed at Brønsted (1007–1015 cm−1) or Lewis
acid sites (1018–1028 cm−1) [31]. The most salient conclusion from
this figure is that coprecipitated and impregnated catalysts dif-
fered in their acid sites. Thus, the impregnated solids (particularly
Inim300) exhibited a strong component for Lewis acid sites at ca.
1021 cm−1 in addition to substantial Brønsted acidity (band at ca.
1007 cm−1) and virtually no hydrogen-bonding interaction with
Gaim300
Inim300
4000
3000
2000
1000
Wavenumber (cm-1)
Fig. 3. FTIR spectra corresponding to the catalysts calcined at 300 ◦C.

126
J.F. Mi˜nambres et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 121–129
0,25
Gaco300
0,20
ZrO
Alco300
Alim300
Gaim300
Inim300
0,15
0,10
0,05
Alim300
Gaim300
Inim300
Inco300
ZrO
2
Alco300
Inco300
0,00
0,50
1,00
Basicity(µmol CO2/m2)
1,50
2,00
Gaco300
Fig. 5. Surface acid–base properties map corresponding to the catalysts used in this
work.
1100 1080 1060 1040 1020 1000
980
960
-1
ence ZrO2 as regard to such properties, the sole difference between
the two being a slightly higher acidity in the latter; however, as
shown by the FT-Raman spectra for chemisorbed pyridine, the acid-
ity for impregnated solids was due to Lewis acid sites, which were
absent from the reference ZrO2. Finally, the basicity of the copre-
cipitated solids clearly exceeded that of the impregnated solids and
the reference ZrO2.
Raman Shift (cm
)
Fig. 4. FT-Raman spectra of pyridine chemisorbed over the catalysts calcined at
300 ◦C in the pyridine symmetric ring-breathing vibration region.
pyridine. On the other hand, the coprecipitated catalysts and the
reference zirconium oxide contained mostly Brønsted acid sites
(band at ca. 1010 cm−1) in addition to other, very weak acid sites
that bound pyridine via hydrogen bonds (band at ca. 1000 cm−1). As
shown elsewhere, incorporating Ga into a mesoporous solid such
as MCM-41 increases its acidity (particularly its Lewis acidity) [34].
The surface basic properties of the solids are summarized in
Table 1. The TPD curves for CO2 (not shown) were similar for all
catalysts calcined at 300 ◦C. Desorption of the gas started at 60 ◦C in
all and peaked at temperatures ranging from 100 ◦C for Inco300 to
150 ◦C for the reference ZrO2. As can be seen from Table 1, the copre-
cipitated catalysts contained a significantly increased amount of
basic surface sites per square meter of solid relative to the reference
zirconium dioxide. The difference was greatest for Inco300, which
trebled the basicity of the oxide. By contrast, the impregnated cata-
lysts possessed a surface basicity per square meter of solid similar to
that of the reference ZrO2 except for the Alim300 that is slightly less
basic. Unlike acidity, surface basicity differed markedly between
the catalysts calcined at 300 ◦C (six times between the most basic
solid, Inco300, and the least one, Alim300). The fact that, based on
the FTIR and FT-Raman spectra, the impregnation method provided
solids containing the modifying metal in nitrate form may some-
how have restricted their basicity by effect of the absence of Al2O3,
Ga2O3 and In2O3 from them.
3.3. Catalytic activity
Following synthesis and characterization, the studied solids
were used as catalysts in the Meerwein–Ponndorf–Verley (MPV)
reaction of crotonaldehyde with 2-propanol as hydrogen donor.
The chemoselective reduction to 2-butenol involved was studied
in both the liquid and the gas phase.
The MPV reaction in the liquid phase (Table 2) exhib-
ited moderate–low catalytic activity (conversion after 8 h was
5–26 mol.%) and a medium–high selectivity (44–88%) towards the
unsaturated alcohol. In general, the results provided by the cata-
lysts calcined at 175 ◦C were slightly worse than those obtained
with the solids calcined at 300 ◦C. Also, the impregnated catalysts
(particularly those calcined at 300 ◦C) clearly surpassed their copre-
cipitated counterparts in terms of activity and selectivity. The best
results were those provided by Gaim300 (26% conversion and 86%
selectivity towards 2-butenol), followed by Alim300 (21% conver-
sion and 81% selectivity towards 2-butenol). As can be seen from
Table 2, the average reaction rate after 8 h reaction was significantly
lower than that after 2 h, which suggests that catalytic activity
decreased as the reaction progressed. There was no clear-cut pat-
tern of variation of the selectivity towards 2-butenol with time;
some catalysts, however, exhibited a marked increase in selectiv-
Fig. 5 maps the surface acid–base properties of the catalysts. As
can be seen, the impregnated catalysts were similar to the refer-
Table 2
Liquid-phase MPV reduction of crotonaldehyde over the synthesized catalyst. Molar conversion (%Conv.), selectivity to the unsaturated alcohol (%SUOL) and reaction rate (per
gram and per square meter) at two reaction times.
Catalyst
Reaction time: 2 h
Reaction time: 8 h
%Conv.
%SUOL
rg (␮mol/min g)
rS (␮mol/min m2)
%Conv.
%SUOL
rg (␮mol/min g)
rS (␮mol/min m2)
Alco175
Inco175
Gaco175
Alim175
Inim175
Gaim175
8
3
6
8
4
5
77
77
85
77
76
15.8
5.8
11.5
15.8
7.4
0.057
0.022
0.053
0.116
0.049
0.060
8
5
7
8
9
78
79
51
77
79
87
3.8
2.4
3.3
4.2
4.4
7.7
0.014
0.009
0.015
0.031
0.029
0.046
100
10.0
15
Alco300
Inco300
Gaco300
Alim300
Inim300
Gaim300
4
4
3
10
5
11
4
16
31
68
90
71
7.3
8.4
5.1
20.1
9.6
22.2
0.033
0.049
0.031
0.120
0.060
0.127
5
5
8
21
12
26
44
46
88
81
87
86
2.4
2.4
4.0
10.3
5.8
12.9
0.011
0.014
0.025
0.061
0.036
0.074

J.F. Mi˜nambres et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 121–129
127
100
80
60
40
20
0
100
Conv. (%)
SUOL (%)
80
60
40
20
0
Conv. (%)
SUOL (%)
Gaim300
Gaim300
150
175
200
225
250
275
300
325
350
Temperature (ºC)
0
50
100
150
200
250
300
350
Reaction Time (min)
0,5
0,4
0,3
0,2
0,1
0,0
Gaim300
SAL
UAL
SOL
UOL
0,5
0,4
0,3
0,2
0,1
0,0
Gaim300
SAL
UAL
SOL
UOL
150
175
200
225
250
275
300
325
350
Temperature (ºC)
0
50
100
150
200
250
300
350
Reaction Time (min)
Fig. 6. Temperature-programmed gas-phase MPV reduction of crotonaldehyde over
Gaim300 catalyst. Reaction temperature was ramped from 150 ◦C to 350 ◦C at
10 ◦C/min.
Fig. 7. Isotherm gas-phase MPV reduction of crotonaldehyde over Gaim300 catalyst
at 200 ◦C.
ity with time, particularly those calcined at 300 ◦C. This suggests
that the loss of selectivity to unsaturated alcohol was more marked
in the initial reaction segment by the formation of 1-butanol and
that deactivation was greater at the active sites producing that sat-
urated alcohol. The activity loss was stronger in the solids calcined
at 175 ◦C – Gaim175 excepted – and in the coprecipitated solids
calcined at 300 ◦C. By contrast, the impregnated solids calcined at
300 ◦C exhibited an increased ability to retain their catalytic activ-
ity.
The reaction tests in the gas phase were initially conducted using
a temperature ramp in order to obtain an overview of catalytic per-
formance over a wide temperature range (150–350 ◦C). In any case,
the conclusions thus drawn are subject to the potential influence of
deactivation processes at high temperatures and the fact that the
highest temperature was 50 ◦C higher than that of calcination of
the solids. By way of example, Fig. 6 shows the results for catalyst
Gaim300 and Table 3 summarizes the results derived from the pro-
files for the body of catalysts at four different temperatures. For all
catalysts, the results were essentially similar. Thus, some catalysts
were already active at 150 ◦C – some, such as Alco300, Alim300,
Gaim300 and Inim300, highly active. The reference ZrO2 was also
active at 150 ◦C, but all other catalysts exhibited minimal activity at
the lower end of the temperature range. However, some of the more
active catalysts at 150 ◦C rapidly lost activity by effect of a severe
deactivation process as the temperature was raised. Such was the
case with Alim300, Gaim300 and Inim300, which, interestingly,
were all obtained by impregnation. Solid Alco300, however, pro-
vided a conversion of ca. 50% at 200 ◦C that increased gradually with
increasing temperature; also, it exhibited no signs of severe deac-
tivation throughout the rest of the test (viz. up to 350 ◦C, where it
provided 100% conversion). However, the selectivity of this catalyst
for 2-butenol dropped to levels in the region of 10% at high temper-
atures, where 1-butanol was the dominant product. Butanal levels
with this catalyst were minimal throughout the studied tempera-
ture range; this suggests that the solid allows reduction of the C
C
bond in 2-butenol to give 1-butanol but not that of the C C bond
in crotonaldehyde to give butanal. Previous studies of the dehy-
dration/dehydrogenation of 2-propanol on ZrO2 revealed that the
dehydrogenation of the alcohol to acetone and molecular hydrogen
occurs at high temperatures (above 300 ◦C) [35] and also that the
C
C bond is reduced by H2 produced in the reaction. The reduction
of the C O bond must, however, require hydrogen transfer from the
alcohol to the carbonyl group via a previously reported cyclic inter-
mediate [2,5]. The other catalysts exhibited similar reaction profiles
irrespective of the conversions levels obtained at the different tem-
peratures studied. Thus, the selectivity towards 2-butenol was very
high (>70%) at medium–low temperatures (below 250 ◦C), but fell
by effect of the formation of 1-butanol at higher temperatures. In
any case, butanal was never produced at levels exceeding 4–5%.
Based on the previous results, we chose a temperature of 200 ◦C
for an isothermal study of the MPV reaction in the gas phase, using
all catalysts calcined at 300 ◦C. By way of example, Fig. 7 illustrates
the performance of solid Gaim300 in the process, and Table 4 shows
the molar conversion, selectivity towards the unsaturated alcohol
and reaction rate in the MPV reduction of crotonaldehyde in the gas
phase at 200 ◦C after 80 and 335 min reaction. All catalysts proved
active – albeit with large differences in crotonaldehyde conversion
– at the shorter reaction time (80 min). Thus, the coprecipitated
solids provided conversions from 4% (Inco300) to 16% (Alco300),
whereas their impregnated counterparts resulted in much higher
levels ranging from 42% with Inim300 to 67% with Gaim300. On
Table 3
Temperature-programmed gas-phase MPV reduction of crotonaldehyde over the
catalysts calcined at 300 ◦C. Molar conversion (%Conv.) and selectivity to the unsat-
urated alcohol (%SUOL) at four reaction temperatures.
Catalyst
200 ◦
C
250 ◦
C
300 ◦
C
350 ◦
C
%Conv. %SUOL %Conv. %SUOL %Conv. %SUOL %Conv. %SUOL
Alco300
Gaco300
Inco300
Alim300
Gaim300 70
Inim300
ZrO2
49
15
15
51
79
87
83
87
86
84
83
68
28
21
36
50
23
27
74
78
81
75
80
80
78
91
56
52
42
69
30
60
40
66
66
59
68
64
92
100
79
87
66
73
28
96
10
47
37
38
46
64
34
38
16

128
J.F. Mi˜nambres et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 121–129
Table 4
Isotherm gas-phase MPV reduction of crotonaldehyde (at 200 ◦C) over the catalysts calcined at 300 ◦C. Molar conversion (%Conv.), selectivity to the unsaturated alcohol
(%SUOL) and reaction rate (per gram and per square meter) at two reaction times.
Catalyst
Reaction time: 80 min
Reaction time: 335 min
%Conv.
%SUOL
rg (␮mol/min g)
rS (␮mol/min m2)
%Conv.
%SUOL
rg (␮mol/min g)
rS (␮mol/min m2)
Alco300
Inco300
Gaco300
Alim300
Inim300
Gaim300
16
4
6
63
42
67
83
87
80
88
86
87
34.0
8.4
13.2
133.5
89.5
141.4
0.123
0.032
0.061
0.981
0.591
0.848
0
0
0
14
9
21
0
0
0
88
88
86
0.0
0.0
0.0
30.1
19.1
45.5
0.000
0.000
0.000
0.221
0.126
0.273
the other hand, selectivity was similar with all catalysts and ranged
from 80% with Gaco300 to 88% with Alim300. The longer reaction
time (335 min) resulted in substantial deactivation of the catalysts
(particularly the coprecipitated solids, which lost all activity within
5 h). The impregnated catalysts also exhibited strong deactivation;
however, they retained some of their high initial conversions and
led to final values ranging from 9% for Inim300 to 21% for Gaim300.
Unlike the liquid phase, deactivation of the catalysts with the reac-
tants in the gas phase had no effect on the selectivity towards the
unsaturated alcohol, which remained as high as 86–88%. The best
results in the MPV reduction of crotonaldehyde in the gas phase
at 200 ◦C were those obtained with Gaim300, which exhibited 21%
conversion and 86% selectivity for 2-butenol after 5 h reaction. This
was also the best performing catalyst in the liquid phase, where
it provided 26% conversion and 86% selectivity. Catalyst Gaim300
was followed in performance by the other two impregnated solids
calcined at 300 ◦C (Alim300 and Inim300). This suggests that the
impregnation method provides more active and at least as selective
catalysts as the coprecipitation method.
molecule was adsorbed at Lewis acid sites and that such was
the case with neither the coprecipitated solids nor the reference
ZrO2. The coprecipitated solids are slightly more basic, and also
slightly more acid, than the impregnated solids; the increased
acidity, however, is due to Brønsted sites of medium and low
strength some of them only interacting with pyridine through
hydrogen bonding.
(c) The most active catalysts in the MPV reduction of croton-
aldehyde in both liquid and gas phase are those obtained
by impregnating ZrO2 with Al, Ga or In nitrate. Specifically,
Gaim300 proved the most active solid in the process. All cat-
alysts are very selective (80–88%) towards the unsaturated
alcohol in the gas-phase process conducted at 200 ◦C.
(d) The catalysts are deactivated in the process, both in the liquid
phase and in the gas phase. However, the loss of activity in
the liquid phase is accompanied by an increase in selectivity
towards 2-butenol, whereas that in the gas phase has no such
effect on selectivity.
As noted earlier, the impregnated solids calcined at 300 ◦C had
a low surface basicity and a moderate surface acidity (Table 1).
However, the FT-Raman spectra for pyridine chemisorbed on the
impregnated catalysts revealed the presence of a substantial pro-
portion of pyridine adsorbed at Lewis acid sites, which were absent
from the coprecipitated solids and the reference ZrO2. Also, as
revealed by the FTIR and FT-Raman spectra, the impregnated
catalysts calcined at 300 ◦C contained nitrate and its presence
seemingly led to the formation of Lewis acid sites, which are highly
active in the MPV reaction [36]. The research conducted in this
work could be expanded in the future by studying the mechanisms
by which the studied catalysts are deactivated and whether their
deactivation is related to the presence of any specific type of acid
or basic site.
Acknowledgments
The authors are thankful to MICINN (Projects CTQ2008/01330
and CTQ2010/18126), Junta de Andalucia (P07-FQM-02695, P08-
FQM-3931, and P09-FQM-4781) and FEDER funds for financial
support. The Central Services (SCAI) at the University of Cordoba
are also acknowledged.
References
[1] P. Maki-Arvela, J. Hajek, T. Salmi, D.Y. Murzin, Appl. Catal. A 292 (2005) 1–49.
[2] J.R. Ruiz, C. Jimenez-Sanchidrian, Curr. Org. Chem. A 11 (2007) 1113–1125.
[3] F.J. Urbano, R. Romero, M.A. Aramendia, A. Marinas, J.M. Marinas, J. Catal. 268
(2009) 79–88;
F.J. Urbano, R. Romero, M.A. Aramendia, A. Marinas, J.M. Marinas, J. Catal. 271
(2010) 153.
[4] E.J. Creyghton, R.S. Downing, J. Mol. Catal. A 134 (1998) 47–61.
[5] G.K. Chuah, S. Jaenicke, Y.Z. Zhu, S.H. Liu, Curr. Org. Chem. A10 (2006)
1639–1654.
4. Conclusions
[6] A. Corma, S. Iborra, Adv. Catal. Vol. 49 (2006) 239–302.
[7] M.A. Aramendia, V. Borau, C. Jimenez, J.M. Marinas, J.R. Ruiz, F.J. Urbano, J. Mol.
Catal. A 171 (2001) 153–158.
In this work, we synthesized zirconium oxide catalysts modified
with aluminium, gallium or indium by impregnation or coprecip-
itation. The solids thus prepared were characterized and used as
catalysts in the MPV reduction of crotonaldehyde in the liquid and
gas phase. The results obtained allow us to draw the following
conclusions:
[8] F. Braun, J.I. Di Cosimo, Catal. Today 116 (2006) 206–215.
[9] J.I. Di Cosimo, A. Acosta, C.R. Apesteguia, J. Mol. Catal.
111–120.
A 234 (2005)
[10] J.I. Di Cosimo, A. Acosta, C.R. Apesteguia, J. Mol. Catal. A 222 (2004) 87–96.
[11] M.A. Aramendia, V. Borau, C. Jimenez, J.M. Marinas, J.R. Ruiz, F. Urbano, Appl.
Catal. A 249 (2003) 1–9.
[12] Y. Zhu, S. Liu, S. Jaenicke, G. Chuah, Catal. Today 97 (2004) 249–255.
[13] S.H. Liu, S. Jaenicke, G.K. Chuah, J. Catal. 206 (2002) 321–330.
[14] Y. Zhu, G.K. Chuah, S. Jaenicke, J. Catal. 241 (2006) 25–33.
[15] A. Ramanathan, M.C. Castro Villalobos, C. Wakernaak, S. Telalovic, U. Hanefeld,
Chem. Eur. J. 14 (2008) 961–972.
[16] A. Ramanathan, D. Klomp, J.A. Peters, U. Hanefeld, J. Mol. Catal. A 260 (2006)
62–69.
[17] S. Nishiyama, M. Yamamoto, H. Izumida, S. Tsuyura, J. Chem. Eng. Jpn. 37 (2004)
310–317.
[18] M. Berkani, J.L. Lemberton, M. Marczewski, G. Perot, Catal. Lett. 31 (1995)
405–410.
[19] J. Lopez, J.S. Valente, J.-M. Clacens, F. Figueras, J. Catal. 208 (2002) 30–37.
[20] V. Ivanov, J. Bachelier, F. Audry, J.C. Lavalley, J. Mol. Catal. 91 (1994) 45–59.
(a) The solids produced by the impregnation method contain Al,
Ga or In nitrate that is incompletely decomposed by calcining
at 300 ◦C. This is not the case with the coprecipitated solids,
which consist of ZrO2 and Al2O3, Ga2O3 or In2O3. All solids are
amorphous and possess a high specific surface area.
(b) Based on surface chemical properties, the impregnated solids
calcined at 300 ◦C possess a low basicity and a moderate acid-
ity. However, the FT-Raman spectra for pyridine chemisorbed
on them revealed that a substantial proportion of the probe

J.F. Mi˜nambres et al. / Journal of Molecular Catalysis A: Chemical 338 (2011) 121–129
129
[21] K. Tanabe, T. Yamaguchi, Catal. Today 20 (1994) 185–197.
[22] C.J. Zhang, R. Miranda, B.H. Davis, Catal. Lett. 29 (1994) 349–359.
[23] F. Lonyi, J. Valyon, J. Therm. Anal. 46 (1996) 211–218.
[24] K. Arata, Appl. Catal. A 146 (1996) 3–32.
[30] R. Burch, C. Passingham, G.M. Warnes, D.J. Rawlence, Spectrochim. Acta A 46
(1990) 243–251.
[31] M.A. Aramendia, V. Borau, C. Jimenez, J.M. Marinas, J.R. Ruiz, F.J. Urbano, J.
Colloid Interface Sci. 217 (1999) 186–193.
[25] D. Spielbauer, G.A.H. Mekhemer, T. Riemer, M.I. Zaki, H. Knozinger, J. Phys.
Chem. B 101 (1997) 4681–4688.
[26] K.M. Malshe, P.T. Patil, S.B. Umbarkar, M.K. Dongare, J. Mol. Catal. A 212 (2004)
337–344.
[32] G.Y. Guo, Y.L. Chen, J. Mater. Sci. 39 (2004) 4039–4043.
[33] G.K. Chuah, S. Jaenicke, B.K. Pong, J. Catal. 175 (1998) 80–92.
[34] J.M. Campos, J.P. Lourenc¸ o, A. Fernandes, A.M. Rego, M.R. Ribeiro, J. Mol. Catal.
A 310 (2009) 1–8.
[27] B.Q. Xu, S.B. Cheng, X. Zhang, Q.M. Zhu, Catal. Today 63 (2000) 275–282.
[28] S. Brunauer, P.H. Emmett, E.J. Teller, J. Am. Chem. Soc. 60 (1938) 309–319.
[29] A.A.M. Ali, M.I. Zaki, Colloid Surf. A 139 (1998) 81–89.
[35] M.A. Aramendia, V. Borau, C. Jimenez, J.M. Marinas, A. Porras, F.J. Urbano, J.
Chem. Soc. Faraday Trans. 93 (1997) 1431–1438.
[36] M. Boronat, A. Corma, M. Renz, J. Phys. Chem. B 110 (42) (2006) 21168–21174.