Support effect in Meerwein-Ponndorf-Verley reduction of benzaldehyde over supported zirconia catalysts

Article abstract of DOI:10.1016/S1872-2067(11)60370-7

A series of zirconia catalysts supported on Si-MCM-41 and Al-doped MCM-41 (Si/Al = 50) mesoporous molecular sieves, silica, γAl₂O ₃, and MgO were prepared by the wet impregnation method. The catalytic activities of these materials in

Full text of DOI:10.1016/S1872-2067(11)60370-7

CHINESE JOURNAL OF CATALYSIS
Volume 33, Issue 6, 2012
Online English edition of the Chinese language journal
Cite this article as: Chin. J. Catal., 2012, 33: 914–922.
ARTICLE
Support Effect in Meerwein-Ponndorf-Verley Reduction of
Benzaldehyde over Supported Zirconia Catalysts
ZHANG Bo*, TANG Minhui, YUAN Jian, WU Lei
Research Institute of Industrial Catalysis, Zhejiang University of Technology, Hangzhou 310014, Zhejiang, China
Abstract: A series of zirconia catalysts supported on Si-MCM-41 and Al-doped MCM-41 (Si/Al = 50) mesoporous molecular sieves, silica,
ꢀ-Al2O3, and MgO were prepared by the wet impregnation method. The catalytic activities of these materials in the Meer-
wein-Ponndorf-Verley reduction (MPV) of benzaldehyde with 2-propanol as reducing agent were investigated, and compared to that of
hydrous zirconia. The materials were characterized by X-ray diffraction, nitrogen adsorption-desorption, X-ray photoelectron spectroscopy,
UV-Vis diffuse reflectance spectroscopy, and Fourier transform infrared and thermal desorption of pyridine. Loading zirconia on
Si-MCM-41, Al-MCM-41, and SiO2 gave improved catalytic activity. This is attributed to a strong interaction of zirconia with the support to
form Si–O–Zr bonds, which gave a significant increase in the amount of exposed Zr–OH groups and stronger Lewis acidity as well as an
appearance of Brönsted acid sites. The activity of 5%ZrO2/Si-MCM-41 was the highest, followed by those of 5%ZrO2/Al-MCM-41 and
5%ZrO2/SiO2. However, 5%ZrO2/Al2O3 and 5%ZrO2/MgO gave very low catalytic activities. This is ascribed to that the acidities of
5%ZrO2/Al2O3 and the ꢀ-Al2O3 support were similar due to the weak interaction of zirconia with ꢀ-Al2O3, and 5%ZrO2/MgO had no acidity
because of the strong interaction between zirconia and MgO.
Key words: support effect; Meerwein-Ponndorf-Verley reduction; zirconia; MCM-41 mesoporous molecular sieve; benzaldehyde
The selective hydrogenation of unsaturated carbonyl com-
pounds is a very important step in the preparation of fine
chemicals. The hydrogen-transfer approach, Meerwein-
Ponndorf-Verley (MPV) reduction, can provide a better way to
enhance the chemoselectivity to allylic alcohol than the tradi-
tional use of molecular hydrogen. Furthermore, unlike con-
ventional hydrogenation, MPV reactions do not require an
elaborate experimental setup or high pressure reactors. A
typical reducing reagent is 2-propanol. Traditionally, the reac-
tion is catalyzed homogeneously by metal alkoxides such as
aluminum or titanium alkoxides [1]. It is generally accepted
that the process proceeds via a six-membered cyclic transi-
tion-state complex in which the carbonyl group and the alcohol
are both coordinated to a Lewis acid metal center, and a hy-
dride transfer from the alcohol to the carbonyl group occurs.
The homogenously catalyzed MPV reaction has some un-
avoidable drawbacks in that the catalyst is required in
stoichiometric amounts and the separation of the catalyst is a
tedious process. Hence, heterogeneous catalysts have gained
increasing attention in recent years. These include magnesium
oxides [2], hydrous zirconia [3,4], layered double hydroxides
[5], Zr-ꢁ [6] and Sn-ꢁ zeolites [7], and metal alkoxides immo-
bilized on mesoporous materials [8]. More than one MPV
mechanisms over heterogeneous catalysts have been proposed
depending on the particular catalyst used. The reduction of
carbonyl compounds by hydrogen transfer has been hypothe-
sized to occur at Lewis acid and/or Brönsted acid sites, basic
sites, and acid-base pairs.
Zirconium-containing heterogeneous catalysts are reported
to exhibit good resistance to the presence of water in addition
to having good catalytic activity. Among these, the catalytic
activity of hydrous zirconia is the lowest. However, hydrous
zirconia has the advantages over other zirconium-containing
catalysts of low cost and easy preparation. In order to improve
the catalytic efficiency of metal oxides, the presence of a
support can provide new means to tailor catalytic performance
by altering the exposure of the active sites and modifying their
nature by interaction with the support. In our previous study [9],
the catalytic activities of pure MgO and MgO supported on
activated carbon (AC), mesoporous molecular sieve MCM-41,
Received 14 December 2011. Accepted 16 January 2012.
*Corresponding author. Tel: +86-571-88320417; E-mail: zb10006093@zjut.edu.cn
Copyright © 2012, Dalian Institute of Chemical Physics, Chinese Academy of Sciences. Published by Elsevier BV. All rights reserved.
DOI: 10.1016/S1872-2067(11)60370-7

ZHANG Bo et al. / Chinese Journal of Catalysis, 2012, 33: 914–922
silica, and ꢀ-alumina in the MPV reduction of acetophenone
selected as silica source and aluminium source, respectively.
Cetyltrimethylammonium bromide (CTAB, C16H33(CH3)3NBr)
was used as the structure-directing agent. The molar composi-
tion of the resulting gel was 1.0SiO2:xAl2O3:0.2CTAB:120H2O,
where x was 0 for Si-MCM-41 and 0.01 for Al-MCM-41,
respectively. The resulting gels were crystallized in a Tef-
lon-lined autoclave at 393 K for 3 d under autogenous pressure.
The materials were recovered by filtration, washed with de-
ionized water, dried at 383 K overnight, and then calcined in air
at 823 K for 6 h in order to remove the template. Solid
Mg(OH)2 was obtained by precipitating 1 mol/L magnesium
nitrate solution with 1 mol/L sodium hydroxide solution, and
then calcined in the air at 773 K for 6 h to yield the support of
magnesium oxide. Silica and ꢀ-alumina were purchased from
Shanghai Henye Chemical Plant (Shanghai, China).
with 2-propanol as reducing agent were investigated. The
higher activity of MgO/AC to those of pure MgO and the other
supported MgO catalysts was ascribed to the high surface area
of AC and that there was no strong interaction of MgO with AC
support, which favored the high dispersion of MgO on the
surface of AC to form very small crystallites and increased the
amount of basic sites on the surface of catalyst. On the other
hand, the concentration of acetophenone near the active sites in
the catalyst was increased because of the ꢂ-electron interaction
between the benzene ring in acetophenone and the graphite
layer of the AC support, which also enhances the activity.
However, the support effect on the catalytic performance of
supported zirconia in the MPV reduction of carbonyl com-
pounds has not been sufficiently addressed.
Due their textural and surface chemical features, silica [10],
alumina [11], and magnesium oxide [12] are widely employed
as catalyst supports. MCM-41 mesoporous molecular sieve is
the most studied member of the M41S meso-structured family.
It has some striking properties such as very high surface area,
regular nano-sized pore structure, and adjustable heteroatom
compositions. This material has been shown to be an excellent
support for preparing supported catalysts with activity and
selectivity superior to those of amorphous silica, alumina, and
even zeolites [13]. In the present contribution, a series of zir-
conia supported on Si-MCM-41 and Al-doped MCM-41 (Si/Al
= 50) mesoporous molecular sieves, silica, ꢀ-alumina, and
magnesium oxide were prepared by the wet impregnation
method using zirconium nitrate as the precursor. The catalytic
activities of these supported zirconia samples in the MPV
reduction of benzaldehyde were investigated and compared to
that of pure hydrous zirconia. The catalysts were characterized
by X-ray diffraction (XRD), nitrogen adsorption-desorption,
X-ray photoelectron spectroscopy (XPS), UV-Vis diffuse re-
flectance spectroscopy, and Fourier transform infrared (FT-IR)
and thermal desorption of pyridine. The activities of the cata-
lysts were correlated with the characterization results to gain a
better insight into the support effect in the MPV reduction of
benzaldehyde over supported zirconia catalysts.
Si-MCM-41, Al-MCM-41, silica, ꢀ-alumina, and magne-
sium oxide supported zirconia catalysts were prepared by the
wet impregnation method. A calculated amount of zirconium
nitrate (Zr(NO3)4·5H2O ) was dissolved in deionized water,
then the finely powered support was added to this solution.
Excess water was evaporated on a water bath with continuous
stirring. The resulting solid was dried in air at 383 K for 12 h,
and further calcined in air at 573 K for 8 h. The zirconia content
in the catalyst was maintained at 5%. The samples were labeled
as 5%ZrO2/Si-MCM-41, 5%ZrO2/Al-MCM-41, 5%ZrO2/SiO2,
5%ZrO2/Al2O3, and 5%ZrO2/MgO. Hydrous zirconia, labeled
as ZrO2, was prepared by the precipitation of a 10% zirconium
nitrate solution with excess 5 mol/L ammonium hydroxide.
The precipitate was washed with deionised water and dried
overnight at 373 K, then calcined at 573 K in air for 8 h. 573 K
was reported to be the optimum calcination temperature of
hydrous zirconia used in the MPV reduction of cinnamalde-
hyde and citral [3,4], so the catalysts in this work were calcined
at this temperature.
1.2 Characterization of catalysts
Powder XRD measurements were carried on a Thermo ARL
X’TRa diffractometer with Cu Kꢀ radiation at a voltage of 40
kV and current of 40 mA. Samples were scanned in the 2ꢁ
range 1.0oꢃ70.0o at a rate of 0.06o/s. The textural properties of
samples were determined from nitrogen adsorption-desorption
isotherms obtained at 77 K on a Micromeritics ASAP 2010 M
instrument. Prior to the measurement, the sample was degassed
at 523 K for 5 h under high vacuum. The specific surface area
was calculated from the adsorption isotherm by the Brun-
auer-Emmett-Teller (BET) method, while the mean pore di-
ameter and cumulative pore volume were estimated by the
Barrett-Joyner-Halenda (BJH) model using the desorption
branch. XPS were performed with a Kratos ULTRA X photo-
electron spectrometer equipped with a monochromatic Al Kꢀ
X-ray source (1486.6 eV). The C 1s peak of contaminant car-
bon at 284.8 eV was used as an internal standard for the cor-
1 Experimental
1.1 Preparation of catalysts
All the chemicals employed in the preparation of catalysts
and reactions were analytical grade supplied by China Medi-
cine (Group) Huadong Chemical Reagent Co. (Hangzhou,
China). They were used as received without further purifica-
tion.
Si-MCM-41 and Al-doped MCM-41 (Al-MCM-41, Si/Al =
50) mesoporous molecular sieves were synthesized by the
hydrothermal method using methods that have been described
elsewhere [13]. Sodium silicate and aluminium nitrate were

ZHANG Bo et al. / Chinese Journal of Catalysis, 2012, 33: 914–922
rection of binding energies. The uncertainty in the determina-
tion of the binding energy values was estimated to be 0.2 eV.
UV-Vis spectra were collected in the 200ꢃ800 nm range on a
Shimadzu 2550 spectrometer equipped a diffuse reflectance
accessory at ambient temperature. BaSO4 was the reference.
Acid properties of samples were analyzed by FT-IR and ther-
mal desorption of pyridine. Pyridine adsorption was performed
in a conventional flow adsorption system using N2 as carrier for
30 min. Physisorbed pyridine was removed by evacuation
under vacuum for 30 min. Then, the thermal desorption was
performed from room temperature to 573 K. The samples were
analyzed in the range of 2000ꢃ1300 cmꢃ1 using a Nicolet
Nexus FTIR spectrometer. Band intensities were normalized
by sample weight.
(a)
(1)
(2)
(3)
(4)
(b)
MgO
ꢀ-Al2O3
(5)
(6)
(7)
10
1.3 Catalytic experiments
0
20
30
40
50
60
70
2ꢁ/(o )
The MPV reduction of benzaldehyde with 2-propanol was
performed in a 25 ml three-necked round-bottom flask im-
mersed in a thermostated bath and equipped with a condenser, a
thermometer, and a magnetic stirrer. A reaction mixture of 3
mmol of the benzaldehyde and 60 mmol of 2-propanol was
placed in the flask and refluxed with stirring at 355 K. The
reaction was started by feeding 300 mg of freshly calcined
catalyst into the mixture. After the reaction of 8 h, the catalyst
was removed by filtration and reaction products were analyzed
using a gas chromatograph (Agilent 6890N) equipped with a
HP Innowax capillary column (30 m × 0.25 mm i.d.) and a
flame ionization detector (FID). The various products were
identified by GC/MS using a similar column and by compari-
son with calibration samples.
Fig. 1. XRD patterns of different samples. (1) Si-MCM-41; (2)
5%ZrO2/Si-MCM-41; (3) Al-MCM-41; (4) 5%ZrO2/Al-MCM-41; (5)
5%ZrO2/MgO; (6) 5%ZrO2/Al2O3; (7) 5%ZrO2/SiO2.
visible in XRD pattern. This was due to the occurrence of
disorder in the hexagonal structure of MCM-41. During the
impregnation, the higher acidity of the zirconium nitrate solu-
tion (pH = 2ꢃ3) could have caused a transformation of some
framework aluminum species of Al-MCM-41 into ex-
tra-framework octahedral aluminum species.
The XRD patterns of 5%ZrO2/SiO2, 5%ZrO2/Al2O3, and
5%ZrO2/MgO are shown in Fig. 1(b). 5%ZrO2/SiO2 exhibited a
broad hump in the range of 2ꢁ = 5oꢃ40o, which was ascribed to
the amorphous structure of the silica support [10].
5%ZrO2/Al2O3 and 5%ZrO2/MgO presented weak diffraction
peaks of ꢀ-Al2O3 [11] and the characteristic peaks of MgO [2],
respectively. No patterns of any bulk zirconia crystal phase [14]
were observed in the high angle range for all supported ZrO2
samples, implying that supported zirconia had an amorphous
nature.
2 Results and discussion
2.1 XRD
The powder XRD patterns of 5%ZrO2/Si-MCM-41 and
5%ZrO2/Al-MCM-41 and the corresponding supports are
displayed in Fig. 1(a). The small angle XRD patterns of
Si-MCM-41 and 5%ZrO2/Si-MCM-41 showed one major (100)
reflection and two minor reflections corresponding to the (110)
and (220) planes, indicating a highly ordered hexagonal pore
structure [13]. The similarity in the intensity of the peaks of
Si-MCM-41 and 5%ZrO2/Si-MCM-41 revealed that the
mesoporous structure was intact after the loading of zirconia.
5%ZrO2/Al-MCM-41 and Al-MCM-41 also gave a typical
XRD pattern for a 2D hexagonal mesophase. The decrease in
the intensity of the diffraction peaks of Al-MCM-41 compared
to Si-MCM-41 can be ascribed to that the substitution of Si4+
by Al3+ ions resulted in an alteration of the TꢃOꢃT bond angle
and a slight decrease in the long range order of the
mesostructure [13]. Upon loading zirconia into Al-MCM-41, a
significant decrease in the intensity of the [100] peak was
2.2 N2 adsorption-desorption
The textural parameters of hydrous zirconia, support mate-
rials, and supported zirconia catalysts are summarized in Table
1. The N2 adsorption-desorption isotherms (not shown) of
Si-MCM-41, 5%ZrO2/MCM-41, Al-MCM-41, and 5%ZrO2/
Al-MCM-41 samples exhibited Type IV isotherms with a
completely reversible nature characteristic of solids with uni-
formly sized mesopores [13]. This indicated that the ordering
of the hexagonal arrays of the mesopores in Si-MCM-41 and
Al-MCM41 was not affected upon loading zirconia. The sur-
face area and total pore volume decrease from 1034 to 776
m2/g and 0.92 to 0.65 cm3/g or 998 to 737 m2/g and 0.95 to 0.65
cm3/g, respectively, after addition of 5% of zirconia into

ZHANG Bo et al. / Chinese Journal of Catalysis, 2012, 33: 914–922
Table 1 Textural properties of different samples
BET surface
Mean pore
Pore volumeb
(cm3/g)
0.14
Zr 3d3/2
Zr 3d5/2
Sample
area (m2/g) diametera (nm)
ZrO2
154
1034
776
998
737
176
97
208
205
86
2.6
3.5
3.5
3.8
3.3
7.7
9.2
10.6
10.0
18.2
6.6
Si-MCM-41
5%ZrO2/Si-MCM-41
Al-MCM-41
5%ZrO2/Al-MCM-41
SiO2
5%ZrO2/SiO2
ꢀ-Al2O3
5%ZrO2/Al2O3
MgO
0.92
0.65
0.95
0.65
0.34
0.22
0.55
0.51
(1)
(2)
(3)
(4)
(5)
(6)
0.32
0.18
5%ZrO2/MgO
112
aCalculated by the BJH method from the desorption isotherm.
190 188 186 184 182 180 178 176
Binding energy (eV)
bTotal pore volume at p/p0 = 0.99.
Fig. 2. XPS spectra of different samples. (1) 5%ZrO2/Si-MCM-41; (2)
5%ZrO2/Al-MCM-41; (3) 5%ZrO2/SiO2; (4) 5%ZrO2/MgO; (5) 5%ZrO2/
Al2O3; (6) ZrO2.
Si-MCM-41 or Al-MCM-41. This indicated the added zirconia
was mainly located inside the Si-MCM-41 or Al-MCM-41
mesoporous channels. No change in the mean pore diameter
between Si-MCM-41 and 5%ZrO2/MCM-41 was observed.
This result revealed that loaded zirconia was highly dispersed
inside the mesopores of the Si-MCM-41 support. However, a
slight decrease in the mean pore diameter from 3.8 to 3.3 nm
was found when loading zirconia into Al-MCM-41, implying a
lower dispersion of zirconia in 5%ZrO2/Al-MCM-41 than in
5%ZrO2/Si-MCM-41. This may be due to the probable forma-
tion of extra-framework aluminum species during the prepara-
tion of 5%ZrO2/Al-MCM-41.
Compared with above samples, the silica, ꢀ-Al2O3, MgO
supports, and the corresponding supported zirconia samples
had lower BET surface areas. Remarkable decreases in the
BET surface area and total pore volume, as well as an increase
in the mean pore diameter after the addition of zirconia into the
silica support were observed. The surface area, pore volume,
and mean pore diameter of 5%ZrO2/Al2O3 were the same as the
pure ꢀ-Al2O3 support. The incorporation of zirconia into the
MgO support caused an effective increase in the surface area
and drastic decreases in the mean pore size and the pore vol-
ume. These results suggested the lower dispersion of zirconia
in the silica, ꢀ-Al2O3, and MgO supports than in the
Si-MCM-41 and Al-MCM-41 supports.
was an increase in the binding energies of the Zr 3d5/2 and Zr
3d3/2 core electron levels in the 5%ZrO2/Si-MCM-41, 5%ZrO2/
Al-MCM-41, and 5%ZrO2/SiO2 samples. This shift towards
higher values can be interpreted in terms of a strong interaction
between zirconia and the support oxides, i.e., the formation of
SiꢃOꢃZr bonds [15]. There is enough evidence in the literature
for an interaction between zirconia and other oxide supports
[16,17]. The higher binding energy of Zr 3d indicated a higher
positive charge on the Zr in the SiꢃOꢃZr linkages, due to the
smaller electronegativity of Zr than that of Si (Pauling values,
Zr 1.4, Si 1.8) [16]. In addition, in comparison to hydrous
zirconia, a trough between the Zr 3d5/2 peak and Zr 3d3/2 peak
and a broadening of the Zr 3d peaks were seen for
5%ZrO2/Al-MCM-41 and 5%ZrO2/SiO2. This can be attributed
to the presence of more than one type of surface Zr4+ species
with different chemical environments [17]. The extent of the Zr
3d peak broadening was less in the case of 5%ZrO2/
Si-MCM-41. This suggested that Zr4+ species had a more uni-
form chemical environment in the Si-MCM-41 support than in
the Al-MCM-41 and silica supports. The results from the N2
adsorption-desorption analysis indicated that the dispersion
degree of zirconia in 5%ZrO2/Si-MCM-41 was the highest,
2.3 XPS
Table 2 XPS data of hydrous zirconia and various supported zirconia
catalysts
All the supported zirconia samples and pure hydrous zirco-
nia were investigated by XPS. The photoelectron peaks for Zr
3d for hydrous zirconia and various supported zirconia samples
are shown in Fig. 2. Table 2 lists the binding energies of the
metal elements in the samples. Figure 2 shows the binding
energies of the Zr 3d peaks of hydrous zirconia at ~182.2 and
~184.5 eV for Zr 3d5/2 and Zr 3d3/2 lines, respectively, which
are in agreement with the value for Zr4+ in pure zirconia in the
literature [14]. Compared to those of hydrous zirconia, there
Binding energy (eV)
Sample
Zr 3d5/2
182.2
183.6
183.5
183.0
182.2
181.1
Si 2p
—
103.9
103.8
103.6
—
Al 2p
—
—
74.7
—
74.6
—
Mg 2p
—
—
—
—
ZrO2
5%ZrO2/Si-MCM-41
5%ZrO2/Al-MCM-41
5%ZrO2/SiO2
5%ZrO2/Al2O3
5%ZrO2/MgO
—
49.0
—

ZHANG Bo et al. / Chinese Journal of Catalysis, 2012, 33: 914–922
followed by those of 5%ZrO2/Al-MCM-41 and 5%ZrO2/SiO2.
visible region (above 400 nm) in the electronic spectrum of all
zirconia-containing samples [19], suggesting the d0 configura-
tion of the Zr4+ ions.
It is a known fact that more Zr atoms can interact with the
support to form ZrꢃOꢃSi bonds when the dispersion of zirco-
nia is higher. Hence, it can be concluded that the amount of
SiꢃOꢃZr bonds in the three samples decreased in the sequence
An intensive absorption band near 210 nm and a weak and
broad shoulder extending from 270 to 390 nm were observed in
the spectrum of hydrous zirconia. These adsorption bands can
be assigned to O2ꢃ (2p) toꢄZr4+ (4d) charge transfer transitions.
The absorption band at higher energy was due to the normal
energy gap transition in the bulk zirconia phase. The low en-
ergy shoulder was attributed to the O2ꢃ ꢄ Zr4+ transitions from
O2ꢃ in low coordination sites at the surface of small particles
[4,12,19]. Compared with hydrous zirconia, with the
5%ZrO2/Si-MCM-41 sample, the intensity of the adsorption
band below 230 nm was significantly decreased and the edge
position of the low energy step was shifted to lower wavelength
(high energy), while the intensity of the absorption band in the
region of 230ꢃ270 nm was distinctly increased with the
maximum adsorption at 250 nm. According to previous reports
[19ꢃ21], in the UV-Vis spectra of Zr4+ species, eight-coordi-
nated Zr4+ species (like those in cubic and tetragonal zirconia)
are responsible for the adsorption in the 200ꢃ210 nm range,
seven-coordinated Zr4+ species (like those in monoclinic zir-
conia) are associated to the absorption near 240 nm, and
six-coordinated (octahedral) Zr4+ species give rise to the ab-
sorption located at 250ꢃ350 nm, i.e., the absorption by Zr4+
species shifts to higher wavelengths (lower energy) with the
decrease of the coordination number of the Zr atoms. Hence,
loading zirconia into the Si-MCM-41 support resulted in a
clear decrease in the coordination number of the Zr species.
This can be attributed to the interaction of high dispersed zir-
conia with the surface of the Si-MCM-41 support to form
ZrꢃOꢃSi bonds during the calcination. It was reported through
characterization by extended X-ray absorption fine structure
(EXAFS) [22] that monolayer-like dispersion of zirconia on
the silica MCM-41 support calcined at 673 K showed a de-
creased ZrꢃO coordination as compared to that of the bulk
zirconia phase. This was ascribed to the partial dehydration of
zirconium hydroxide to anchor it to the silica surface [22].
Compared with 5%ZrO2/Si-MCM-41, 5%ZrO2/Al-MCM-41
showed a more intense absorption edge below 240 nm and a
weaker absorption band at 250 nm. This indicated less decrease
in the coordination number of the Zr species in
5%ZrO2/Al-MCM-41 than in 5%ZrO2/Si-MCM-41. This could
be due to the lower dispersion of zirconia in the former. In
contrast to 5%ZrO2/Si-MCM-41 and 5%ZrO2/Al-MCM-41,
the UV-Vis spectra profiles of 5%ZrO2/SiO2, 5%ZrO2/Al2O3,
and 5%ZrO2/MgO resembled that of hydrous zirconia, but the
edge position of the low energy absorption was shifted to a
lower wavelength (higher energy) from 270 to 250 nm. In
addition, the intensity of the low energy shoulder of
5%ZrO2/SiO2 was larger than that of 5%ZrO2/Al2O3,
5%ZrO2/MgO, and hydrous zirconia. These results indicated
that the coordination of the Zr4+ species in 5%ZrO2/Al2O3 and
of 5%ZrO2/Si-MCM-41
5%ZrO2/SiO2.
>
5%ZrO2/Al-MCM-41
>
In the case of 5%ZrO2/Al2O3, well resolved Zr 3d lines with
high intensity were observed and the binding energies of Zr 3d
were in agreement with those of pure hydrous zirconia. This
indicated a weak interaction between zirconia and the ꢀ-Al2O3
support.
From Fig. 2 and Table 2, a clear decrease in the binding en-
ergy of Zr 3d and an extensive broadening of the Zr 3d lines
can be noted for the 5%ZrO2/MgO sample compared to hy-
drous zirconia. This indicated that there was a stronger inter-
action between zirconia and the MgO support, perhaps with the
formation of MgꢃOꢃZr bonds. The lower binding energy of Zr
3d indicated a lower positive charge density on the Zr atoms,
which was attributed to the higher electronegativity of Zr than
Mg (Pauling values, Zr 1.4, Mg 1.3) [16,18]. The observed
extensive broading of the Zr 3d lines implied that some unde-
fined ZrOx species were located on the surface of
5%ZrO2/MgO.
For these supported zirconia samples, the binding energies
of Si 2p (~103.7 eV), Al 2p (~74.6 eV), and Mg 2p (~49.0 eV)
were essentially in agreement with the values of pure SiO2,
ꢀ-Al2O3, and MgO reported in the literature. This was due to
the significantly high content of the support (95%) [15ꢃ18].
2.4 UV-Vis
Figure 3 shows the UV-Vis spectra of supported zirconia
samples with the hydrous zirconia spectrum as a reference. No
feature characteristics of dꢃd transitions were displayed in the
(1)
(2)
(3)
(4)
(5)
(6)
200
300
400
500
600
700
800
Wavelength (nm)
Fig. 3. UV-Vis spectra of different samples. (1) ZrO2; (2) 5%ZrO2/
Si-MCM-41; (3) 5%ZrO2/Al-MCM-41; (4) 5%ZrO2/SiO2; (5) 5%ZrO2/
Al2O3; (6) 5%ZrO2/MgO.

ZHANG Bo et al. / Chinese Journal of Catalysis, 2012, 33: 914–922
5%ZrO2/MgO was basically like that in pure hydrous zirconia,
while 5%ZrO2/SiO2 has more Zr4+ species with a lower ZrꢃO
coordination number. Hence, the amount of Zr4+ species with a
lower ZrꢃO coordination number in 5%ZrO2/Si-MCM-41 was
the highest, followed by those of 5%ZrO2/Al-MCM-41 and
5%ZrO2/SiO2. This was in agreement with the order of amount
of SiꢃOꢃZr bonds in these samples. The amounts of Zr4+ spe-
cies with a lower ZrꢃO coordination number in 5%ZrO2/Al2O3
and 5%ZrO2/MgO were the lowest.
(a)
0.1
(1)
(2)
(3)
(4)
(5)
(6)
(b)
0.01
2.5 FT-IR and thermal desorption of pyridine
(1)
The acid properties of the samples were analyzed by FT-IR
and thermal desorption of pyridine. Only three weak infrared
bands at 1597, 1446, and 1577 cmꢃ1 were observed for the
Si-MCM-41 and silica supports (not shown). These were as-
cribed to hydrogen bonded pyridine, where the ꢃOH groups
were not acidic enough to transfer their proton to pyridine (very
weak interaction) [23]. This suggested a relatively weak sur-
face acidity for the Si-MCM-41 and silica supports. Besides
these three bands, the Al-MCM-41 support presented another
very weak peak at 1490 cmꢃ1 (not shown here), assigned to
chemisorbed pyridine on either Lewis or Brönsted acid sites
[24]. This indicated that the substitution of a small amount of
Si4+ by Al3+ ions (Si/Al = 50) in Si-MCM-41 caused a slight
increase in the acidity.
Figure 4 shows the FT-IR spectra of pyridine adsorption on
hydrous zirconia and supported zirconia samples evacuated at
room temperature, 373 K, and 423 K. In Fig. 4(a), only two
very weak adsorption peaks were observed on hydrous zirconia
at 1443 and 1603 cmꢃ1, which were attributed to coordinated
pyridine species on Lewis acid sites [22,23]. This showed that
pure hydrous zirconia possessed few Lewis acid sites. For
5%ZrO2/Si-MCM-41, 5%ZrO2/ Al-MCM-41, and 5%ZrO2/
SiO2 as compared to the corresponding supports and pure
zirconia, the intensities of the adsorption bands at 1597, 1577,
and 1446 cmꢃ1 were significantly increased. This suggested an
effective increase in the amount of ꢃOH groups on the surface
interacting with pyridine by hydrogen bonding. This was at-
tributed to the appearance of plentiful ZrꢃOH groups due to the
dispersion of zirconia on the surface of the supports and the
higher coordination number of Zr with respect to Si [25]. The
intensities of the above desorption bands decrease in the order
of 5%ZrO2/Si-MCM-41 > 5%ZrO2/Al-MCM-41 > 5%ZrO2/
SiO2, suggesting that the amount of ZrꢃOH groups in the three
samples decreases in the same order. In addition, bands at 1638,
1628, 1545, and 1490 cmꢃ1 were observed for 5%ZrO2/
Si-MCM-41and 5%ZrO2/ Al-MCM-41, and a weak band at
1490 cmꢃ1 was found for 5%ZrO2/SiO2. The comparison with
literature data suggested that the bands at 1638 and 1545 cmꢃ1
were due to pyridinium ion on Brönsted acid sites and the band
at 1628 cmꢃ1 was attributed to coordinated pyridine species on
Lewis acid sites [13,23ꢃ26]. It is clear that both Brönsted acid
(2)
(3)
(4)
(5)
(6)
(c)
0.01
(1)
(2)
(3)
(4)
(5)
(6)
2000 1900 1800 1700 1600 1500 1400 1300
Wavenumber (cmꢂ1)
Fig. 4. Pyridine-FT-IR spectra of samples after desorption at room
temperature (a), 373 K (b), and 423 K (c). (1) 5%ZrO2/Si-MCM-41; (2)
5%ZrO2/Al-MCM-41; (3) 5%ZrO2/SiO2; (4) 5%ZrO2/Al2O3; (5)
5%ZrO2/MgO; (6) ZrO2.
and Lewis acid sites interacting strongly with pyridine were
present on these supported samples. This result means that the
acidities of these samples were stronger than those of pure
zirconia and the corresponding supports. This was probably
associated with the interaction of zirconia with the supports
and the coordination number of Zr4+ species. It had been re-
ported [27,28] that on monoclinic zirconia, in which the Zr4+
cation is heptacoordinated, there exists both Brönsted acid and
Lewis acid sites, while only Lewis acid sites were detected by
IR measurement after pyridine adsorption on tetragonal zirco-
nia in which the Zr4+ cation is octacoordinated. Furthermore,
the surface acid site density is higher on monoclinic zirconia
than on tetragonal zirconia. On the other hand, for these sup-
ported zirconia samples, the interface between zirconia and
support can be considered a silica-zirconia mixed oxide due to
the formation of SiꢃOꢃZr bonds. According to Tanabe’s hy-
pothesis [29,30], acidity generation in a binary oxide is caused
by an excess of negative or positive charge. For the sil-
ica-zirconia mixed oxide, the Brönsted acidity was related the
presence of an excess of negative charge due to that the coor-
dination numbers of Si (4) and Zr (8, 7, or 6) elements are
different but the oxygen anions adopt the coordination number

ZHANG Bo et al. / Chinese Journal of Catalysis, 2012, 33: 914–922
of the host oxide (in silica-rich region). The higher positive
Lewis acid sites and surface ꢃOH groups of 5%ZrO2/Al2O3
were mainly due to the ꢀ-Al2O3 support, due to the small loaded
amount (5%) and low dispersion of zirconia in the catalyst.
In comparison with the other supported zirconia samples and
hydrous zirconia, almost no adsorption band can be found in
the IR spectrum of pyridine adsorption for 5%ZrO2/MgO. It
implied negligible acidity for 5%ZrO2/MgO. This can be ex-
plained by the strong interaction between zirconia and the
MgO support as confirmed by XPS results, which lead to the
decrease of the positive charge of Zr when in the MgꢃOꢃZr
linkages [23,33].
charge on Zr in the SiꢃOꢃZr linkage, indicated by the XPS
analysis, implied more Lewis acid sites on zirconium in the
silica-zirconia mixed oxides.
Comparing Fig. 4(a)ꢃ(c), for 5%ZrO2/Si-MCM-41,
5%ZrO2/Al-MCM-41, and 5%ZrO2/SiO2, upon outgassing at
increasing temperatures, the intensities of all the bands de-
creased. This was more pronounced for the bands at 1597 and
1446 cmꢃ1, confirming that hydrogen bonded pyridine was
very weakly held on the supports. At the same time, an ap-
pearance of a band at 1609 cmꢃ1 and a shift of the band from
1446 to 1448 cmꢃ1 were found. The two new bands were as-
signed to coordinated pyridine on Lewis acid sites [26], over-
lapped by the bands at 1597 and 1446 cmꢃ1 at room tempera-
ture. It was observed that the bands corresponding to Lewis
acid sites were stronger and more abundant than those of
Brönsted acid sites, indicating the amount of Lewis acid sites
was much higher than that of Brönsted acid sites. The bands at
1609, 1490, and 1448 cmꢃ1 still remained while the bands
assigned to Brönsted acid sites had almost diminished at 423 K,
implying that the strength of the Lewis acid sites was stronger
than that of the Brönsted acid sites. On the other hand, the
higher wavenumbers of the Lewis acid sites for these samples
compared to those observed for pure zirconia meant that the
strength of the Lewis acid sites was stronger.
2.6 Catalytic studies
During the MPV reduction of benzaldehyde using
2-propanol as reducing agent, benzaldehyde is reduced to
benzyl alcohol and 2-propanol is oxidized to acetone. Table 3
shows the conversion of benzaldehyde and selectivity for
benzyl alcohol over hydrous zirconia and the various supported
zirconia catalysts. All the supports were inactive (conversion of
benzaldehyde < 2%, not listed). Hydrous zirconia showed a
low conversion of benzaldehyde (33.7%) and quite high se-
lectivity for benzyl alcohol (99.2%) after a reaction of 8 h. The
conversion of benzaldehyde was significantly enhanced upon
loading zirconia into Si-MCM-41, Al-MCM-41, and SiO2, and
at the same time, the selectivity for benzyl alcohol was main-
tained. The activities of the three catalysts increased in the
order 5%ZrO2/SiO2 (51.5%) < 5%ZrO2/Al-MCM-41 (56.1%)
< 5%ZrO2/Si-MCM-41 (88.5%). In contrast, loading zirconia
into ꢀ-Al2O3 or MgO resulted in a dramatic fall of benzalde-
hyde conversion (< 5%) and a slight decrease of benzyl alcohol
selectivity. The activities of 5%ZrO2/Al2O3 and 5%ZrO2/MgO
were closed to those of the ꢀ-Al2O3 and MgO supports.
The intensity of the bands attributed to Brönsted and Lewis
acid sites in the spectra of 5%ZrO2/Si-MCM-41 was the high-
est, followed by those of 5%ZrO2/Al-MCM-41 and
5%ZrO2/SiO2, suggesting that the amount of Brönsted and
Lewis acid sites followed the same order, which agreed well
with the order of amounts of SiꢃOꢃZr bonds and Zr species
with a lower coordination number.
From Fig. 4(a)ꢃ(c), weak bands at 1614, 1595, 1490, and
1445 cmꢃ1 were observed for 5%ZrO2/Al2O3 at room tem-
perature. With the increase of outgassing temperature, the band
at 1595 cmꢃ1 rapidly disappeared while the band at 1444 cmꢃ1
shifted to 1448 cmꢃ1. The bands at 1595 and 1445 cmꢃ1 were
attributed to hydrogen bonded pyridine. The intensities of these
were significantly lower in comparison with those of
5%ZrO2/Si-MCM-41, 5%ZrO2/Al-MCM-41, and 5%ZrO2/
SiO2. The bands at 1614, 1490, and 1448 cmꢃ1 were assigned to
pyridine adsorbed on Lewis acid sites [24ꢃ30]. The
wavenumbers of these bands were slightly higher than those
observed for pure zirconia but close to those reported for the
Lewis acid sites on ꢀ-Al2O3 [31,32]. No Brönsted acid sites
were observed, as revealed by the absence of the pyridinium
bands at about 1630 and 1540 cmꢃ1. In the literature [31,32], it
was reported that ꢀ-Al2O3 only possesses Lewis acidity. This
indicated that the acidity of 5%ZrO2/Al2O3 is basically similar
to that of the ꢀ-Al2O3 support. XPS results confirmed there was
a quite weak interaction between zirconia and the ꢀ-Al2O3
support. Hence, the acidity of 5%ZrO2/Al2O3 would resemble
that of a mechanical mixture of zirconia and ꢀ-Al2O3. The
Hydrous zirconia is an active catalyst for the MPV reduction
of cinnamaldehyde and citral using 2-propanol as the hydrogen
donor [5,6]. The pretreatment temperature of hydrous zirconia
has a big effect on the activity. The oxide calcined at 573 K was
the most active, and above it, the catalytic activity decreased
with calcination temperature. Chuah et al. [3,4] suggested that,
besides surface area and porosity, the exposed hydroxyl groups
Table 3 MPV reduction of benzaldehyde over hydrous zirconia and
supported zirconia
Conversion
(%)
Selectivity for
benzyl alcohol (%)
Sample
ZrO2
33.7
88.5
56.1
51.5
3.9
3.3
99.2
99.1
99.5
99.8
96.2
97.8
5%ZrO2/Si-MCM-41
5%ZrO2/Al-MCM-41
5%ZrO2/SiO2
5%ZrO2/Al2O3
5%ZrO2/ MgO
Reaction conditions: benzaldehyde 3 mmol, 2-propanol 60 mmol, cata-
lyst 300 mg , 355 K, 8 h.

ZHANG Bo et al. / Chinese Journal of Catalysis, 2012, 33: 914–922
of hydrous zirconia have a big influence on activity. They
port has an important role in the surface acidity of the sup-
ported zirconia catalysts, which correlated strongly with the
catalytic activity. For an interaction between zirconia and the
support that favors a significant increase in the amount of
surface ZrꢃOH groups and the strengthening of Lewis acidity,
and appearance of Brönsted acid sites, there is enhancement of
catalytic activity.
proposed that hydroxyl groups act as sites for ligand exchange
with 2-propanol to form 2-propoxide on the catalyst. The car-
bonyl compound coordinates to the zirconium metal center.
This activates the carbonyl group and initiates a hydride
transfer from 2-propoxide to the carbonyl through a cyclic
six-membered transition state, as has been proposed for beta
zeolite. Acetone is formed and subsequent alcoholysis leads to
the product and regeneration of the active catalyst. According
to above mechanism, the increase in the number of exposed
ZrꢃOH groups and strength of Lewis acidity (Zr4+ center) favor
the improvement of catalytic activity. Urbano et al. [34] in-
vestigated the catalytic activity of zirconia modified with boron
and alkaline-earth metal calcined at 673 K in the MPV reduc-
tion of cinnamaldehyde with 2-propanol. They found that an
increase of surface hydroxyl groups of the catalyst led to an
enhancement of its activity, and proton (Brönsted) acid sites of
medium-high strength were the most catalytically active sites,
while a potential contribution of Lewis acid sites cannot be
ruled out. They also reported that the solids with an increased
proportion of weak acid sites interacting with pyridine by
hydrogen bonding or a high density of basic sites were much
less active.
The comparison of the acidity characterization and MPV
reduction of benzaldehyde test results showed that the catalytic
activity is closely related to surface acidity. The higher cata-
lytic activities of 5%ZrO2/Si-MCM-41, 5%ZrO2/Al-MCM-41,
and 5%ZrO2/SiO2 than that of hydrous zirconia can be attrib-
uted to a significant increase in the amount of surface ZrꢃOH
groups and the strengthening of Lewis acidity, as well as the
appearance of Brönsted acid sites. For these three catalysts, the
activity increased with the amount of surface ZrꢃOH groups,
Brönsted and Lewis acid sites. The very low activity of
5%ZrO2/Al2O3 can be attributed to that its surface acidity
resembled that of the ꢀ-Al2O3 support. The very low activity of
5%ZrO2/MgO can be ascribed to the absence of acidity of its
surface. This is different from the results of Urbano et al. [34]
that a significant increase in the amount of weak acid sites, i.e.
ZrꢃOH groups interacting with pyridine by hydrogen bonding,
favors the enhancement of catalytic activity. This could be due
to the differences in the preparation of catalysts of the two
works.
References
1
Ruiz J R, Jimenez-Sanchidrián C. Current Org Chem, 2007, 11:
1113
2
3
Gliꢅski M. Appl Catal A, 2008, 349: 133
Zhu Y Z, Liu S H, Jaenicke S, Chuah G. Catal Today, 2004, 97:
249
4
5
Liu S H, Jaenicke S, Chuah G K. J Catal, 2002, 206: 321
Shi N F, Zhang B, Xu Ch L, Zhang X B, Chen Y F, Ge Zh H. J
Chem Eng Chin Univ, 2011, 25: 61
6
7
Zhu Y Z, Chuah G K, Jaenicke S. J Catal, 2006, 241: 25
Samuel P P, Shylesh S, Singh A P. J Mol Catal A, 2007, 266:
11
8
9
Zhu Y Z, Jaenicke S, Chuah G K. J Catal, 2003, 218: 396
Xu Ch L, Zhang B, Yuan J, Lu H F, Chen Y F, Ge Zh H. Acta
Chim Sin, 2011, 69: 368
10 Busto M, Lovato M E, Vera C R, Shimizu K, Grau J M. Appl
Catal A, 2009, 355: 123
11 Hu S H, Xue M W, Chen H, Sun Y L, Shen J Y. Chin J Catal,
2011, 32: 917
12 Ciuparu D, Ensuque A, Bozon-Verduraz F. Appl Catal A, 2007,
326: 130
13 Liu D P, Quek X Y, Cheo W N E, Lau R, Borgna A, Yang Y H.
J Catal, 2009, 266: 380
14 Chen C-L, Cheng S F, Lin H-P, Wong S-T, Mou C-Y. Appl
Catal A, 2001, 215: 21
15 Damyanova S, Petrov L, Centeeno M A, Grange P. Appl Catal
A, 2002, 224: 271
16 Reddy B M, Chowdhury B, Smirniotis P G. Appl Catal A, 2001,
211: 19
17 Reddy B M, Sreekanth P M, Yamada Y, Kobayashi T. J Mol
Catal A, 2005, 227: 81
18 Wakiya N, Kuroyanagi K, Xuan Y, Shinozaki K, Mizutani N.
Thin Solid Films, 2000, 372: 156
19 López E F, Escribano V S, Panizza M, Carnasciali M M, Busca
G. J Mater Chem, 2001, 11:1891
20 Zhao J, Yue Y H, Hua W M, He H Y, Gao Z. Appl Catal A,
2008, 336: 133
21 Chary K V R, Ramesh K, Naresh D, Ramana Rao P V,
Ramachandra Rao A, Venkat Rao V. Catal Today, 2009, 141:
187
22 Wang J-H, Mou C-Y. Microporous Mesoporous Mater, 2008,
110: 260
23 Ordomsky V V, Sushkevich V L, Ivanova I I. J Mol Catal A,
2010, 333: 85
24 Ahmed M A. Fuel Process Technol, 2011, 92: 1121
25 Flego C, Carluccio L, Rizzo C, Perego C. Catal Commun,
2001, 2: 43
3 Conclusions
The MPV reduction of benzaldehyde with 2-propanol was
carried out over a series of zirconia catalysts supported on
Si-MCM-41, Al-doped MCM-41 (Si/Al = 50) mesoporous
molecular sieves, silica, ꢀ-Al2O3, and magnesium oxide, and
compared to that of pure hydrous zirconia. Differences in
catalytic activities observed for the supported zirconia catalysts
with respect to pure hydrous zirconia indicated that the carrier
affected the activity. The interaction of zirconia with the sup-
26 Garg S, Soni K, Muthu Kumaran G, Bal R, Gora-Marek K,

ZHANG Bo et al. / Chinese Journal of Catalysis, 2012, 33: 914–922
Gupta J K, Sharma L D, Murali Dhar G. Catal Today, 2009, 14:
2009, 367: 113
125
31 Li J C, Xiang L, Xu F, Wang Z W, Wei F. Appl Surf Sci, 2006,
253: 766
32 Scirè S, Crisafulli C, Maggiore R, Minicò S, Gavagno S. Appl
Surf Sci, 1998, 136: 311
33 García V, Fernández J J, Ruíz W, Mondragón F, Moreno A.
Catal Commun, 2009, 11: 240
34 Urbano F J, Aramendia M A, Marinas A, Marinas J M. J Catal,
2009, 268: 79
27 Zhang W, Tay H L, Lim S S, Wang Y S, Zhong Z Y, Xu R.
Appl Catal B, 2010, 95: 93
28 Zhao Y B, Li W, Zhang M H, Tao K Y. Catal Commun, 2002, 3:
239
29 Rodríguez Avendaño R G, De Los Reyes J A, Viveros T,
Montoya De La Fuente J A. Catal Today, 2009, 148: 12
30 Gervasini A, Messi C, Flahaut D, Guimon C. Appl Catal A,