Selective oxidation of ethanol to acetaldehyde on V2O5/TiO2/SiO2 catalysts: Effect of TiO2-coating of the silica support

Article abstract of DOI:10.1006/jcat.1997.1760

The effect of coating the SiO₂ support of a V₂O₅ oxidation catalyst with different amounts of TiO₂ has been studied. X-ray photoelectron spectroscopy of the resulting V₂O₅/TiO₂/SiO₂ showed that the highest dispersion was reached for the low TiO₂ coverages. The catalysts were prepared by depositing vanadia in a quantity equivalent to one monolayer onto SiO₂, TiO₂, or TiO₂-coated SiO₂ samples, by impregnation with a vanadyl acetylacetonate solution in ethanol, drying at 348 K, and calcining at 773 K in air. This procedure decreased the surface area by up to 30%. TPR profiles of binary and ternary samples were similar, showing only a single peak, suggesting the presence of just one vanadium species. Electron spin resonance spectra of partially reduced TiO₂-coated silica carriers showed that, at very low titania content, titanium ions were present in the silica surface in small clusters. In the high content sample, TiO₂ covered the silica surface, as in a crystalline TiO₂ phase. The characteristics of supported vanadium oxide were affected by the TiO₂ distribution. These V/xTiSi catalysts were found to be active for the selective oxidation of ethanol. The titania coating of the support increased the activity of the supported vanadia as compared to that of the oxide supported on silica, but maintained a similar product distribution. This increase of activity depended on the number of titania monolayers (x), being maximum for x ≥ 1. The selectivity to acetaldehyde also increased with x, being maximum for the same composition. The combination of the two factors allowed the attainment of acetaldehyde yields of 74% per pass at temperatures as low as 473 K. The modifications in the catalytic behavior are discussed in terms of the changes of the support surface and its interaction with the supported vanadia caused by the presence of the titania coating.

Full text of DOI:10.1006/jcat.1997.1760

JOURNAL OF CATALYSIS 171, 1–13 (1997)
ARTICLE NO. CA971760
Selective Oxidation of Ethanol to Acetaldehyde
on V2O5/TiO2/SiO2 Catalysts
Effect of TiO2-Coating of the Silica Support
N. E. Quaranta,1 J. Soria, V. Corte´s Corbera´n, and J. L. G. Fierro2
Instituto de Cata´lisis y Petroleoqu´ımica, CSIC, Campus UAM, Cantoblanco, 28049 Madrid, Spain
Received May 12, 1995; revised May 27, 1997; accepted May 30, 1997
active phase on a carrier an attractive technique for the
tailoring of catalytic properties. It offers the advantage of
an increased exposure of the active phase and allows one to
modify its structural and catalytic properties by interaction
with the carrier.
Theeffect ofcoatingtheSiO2 support ofa V2O5 oxidation catalyst
with different amounts of TiO2 has been studied. X-ray photoelec-
tron spectroscopy of the resulting V2O5/TiO2/SiO2 showed that the
highest dispersion wasreached for thelowTiO2 coverages. Thecata-
lysts were prepared by depositing vanadia in a quantity equivalent
to one monolayer onto SiO2, TiO2, or TiO2-coated SiO2 samples, by
impregnation with a vanadyl acetylacetonate solution in ethanol,
drying at 348 K, and calcining at 773 K in air. This procedure de-
creased the surface area by up to 30%. TPR profiles of binary and
ternary samples were similar, showing only a single peak, suggesting
the presence of just one vanadium species. Electron spin resonance
spectra of partially reduced TiO2-coated silica carriers showed that,
at very low titania content, titanium ions were present in the silica
surface in small clusters. In the high content sample, TiO2 covered
the silica surface, as in a crystalline TiO2 phase. The characteristics
of supported vanadium oxide were affected by the TiO2 distribution.
These V/xTiSi catalysts were found to be active for the selective ox-
idation of ethanol. The titania coating of the support increased the
activity of the supported vanadia as compared to that of the oxide
supported on silica, but maintained a similar product distribution.
This increase of activity depended on the number of titania mono-
layers (x), being maximum for x ꢀ 1. The selectivity to acetaldehyde
also increased with x, being maximum for the same composition.
The combination of the two factors allowed the attainment of ac-
etaldehyde yields of 74% per pass at temperatures as low as 473 K.
The modifications in the catalytic behavior are discussed in terms
of the changes of the support surface and its interaction with the
supported vanadia caused by the presence of the titania coating.
Vanadia catalysts constitute a relevant example of the in-
fluence of this interaction. Several authors (1–3) have com-
pared the properties of vanadia supported on different car-
riers (SiO2, Al2O3, TiO2, MgO, ZrO2) and have concluded
that the nature of the dispersed surface metal oxide phase
enables the vanadia, depending on which support is used,
to become an effective catalyst for selective oxidation of
aromatics (4–7), olefins (8), and alcohols (9–11), as well as
for selective catalytic reduction (SCR) of NOx by NH3 (12).
Titania (anatase) interacts strongly with an immobilized
vanadia layer, generating a molecular dispersion of V2O5
(13, 14), but the system suffers from limited specific surface
area and low resistance to sintering. Conversely to TiO2,
interaction of V2O5 with SiO2 is weak and, therefore, the
properties of VOx species over SiO2 are modified to a lesser
extent, showing a higher tendency for thermally induced ag-
gregation, leading to a low dispersion of the active phase
(15, 16). However, the use of SiO2 as a support of vanadia
has the advantage of presenting a higher specific surface
area for coating by vanadia and a bigger resistance to sinter-
ing than can be provided by TiO2. A way to obtain a titania
surface with high, thermostable surface area and good me-
chanical properties is to apply TiO2 onto silica (17). Vanadia
supported on the mixed oxide TiO2/SiO2, either coprecip-
itated (18) or having the TiO2 directly applied to the silica
(19–24), has been developed because of its selectivity as
a catalyst for SCR of nitrogen oxides. Moreover, vanadia-
based systems such as V2O5/TiO2 (10) and V2O5/SiO2 (11)
catalysts have been studied for selective oxidation of alco-
hols. In particular, V2O5 (25) and V2O5/SiO2 (26) are very
active and selective for oxidation of ethanol to acetalde-
hyde. By a proper selection of the catalytic oxide system and
reaction conditions, ethanol can be converted oxidatively
c
ꢁ 1997 Academic Press
INTRODUCTION
The interaction between catalytically active metal oxide
particlesand oxide carriersgreatlyinfluencestheir structure
and size. This makes the deposition of monolayer(s) of an
1 On leave from CIC, CINDECA, La Plata, Argentina.
2
To whom correspondence should be addressed. Fax: (34)-1-585-4760.
E-mail: jlgfierro@icp.csic.es.
1
0021-9517/97 $25.00
c
Copyright ꢁ 1997 by Academic Press
All rights of reproduction in any form reserved.

2
QUARANTA ET AL.
to acetaldehyde, acetic acid, or ethyl acetate, all of which tion of the surface hydroxyls with the vanadyl groups of
can be used either as final products or as intermediates in the vanadium (IV) acetylacetonate complex (29). The im-
synthetic routes.
pregnation was made by adding an ethanolic solution of
In a previous paper (27), we studied the oxidation of the vanadyl acetylacetonate to the support particles. The
ethanol on supported monolayer vanadia catalysts, pre- suspension was evaporated at 348 K under continuous stir-
pared by impregnation of the acetylacetonate complex ring. The impregnates were washed repeatedly with pure
from an ethanolic solution, and compared the catalytic ethanol, dried again, and finally calcined at 773 K in air
performance of a V2O5/TiO2/SiO2 catalyst with those of for 2 h to convert vanadium to vanadium oxide. The added
V2O5/TiO2 and V2O5/SiO2 catalysts. It was found that the amount of vanadium was that calculated for a complete
coating of the silica carrier with a monolayer of TiO2 in- monolayer, taking into account the specific surface areas.
creased substantially the activity of the catalyst, approach- In the case of TiO2 and TiSi supports a value of 0.166 nm2
ing that of the one supported on pure TiO2, while it in- per center was considered, assuming that one vanadium ox-
creased the selectivity to acetaldehyde as compared with ide species is deposited on each center (28). In the case of
the latter. Accordingly, it was decided to study the effect of SiO2 the needed amount was calculated for a monolayer
varying the amount of the TiO2 in the coating of the silica thickness of 0.234 nm (4) taking the density of V2O5 as
support with the aim of understanding the interrelations 3.357 g/cm3. The catalysts will be denoted henceforth by
between TiO2 and SiO2 which can be important at low tita- their component elements (V/Ti, V/Si, and V/xTiSi, where
nium content, and their influence on the catalytic properties x denotes the number of ML of titania).
of supported V2O5, as a way to improve its performance for
the selective oxidation of ethanol.
Characterization of Catalysts and Supports
Specific surface areas. Specific surface areas were deter-
mined by the BET method from the adsorption isotherms
of nitrogen at 77 K, taking a value of 0.164 nm2 for the
cross-sectional area of the adsorbed nitrogen molecule at
that temperature.
EXPERIMENTAL
Preparation of the TiO2/SiO2 Supports
Titania-coated silica supports were prepared by homoge-
neous precipitation, as described by Geus (28). Nonporous
silica (Aerosil MOX 80 from Degussa, specific surface area
X-ray photoelectron spectroscopy (XPS). XP spectra
86 m2/g) was suspended in deionized water, and acidified were obtained with a Leybold LHS 10 spectrometer
with HCl to pH < 1. An appropriate amount (see below) provided with a hemispherical electron analyzer and a
of TiCl3 (Merck, 15% solution in HCl) was added to the Mg anode X-ray exciting source (Mg Kꢀ = 1253.6 eV,
suspension with vigorous stirring to ensure a good homo- 1 eV = 1.602 ꢃ 10ꢂ19 J). Samples were pumped to 10ꢂ5 Torr
geneity during the precipitation. Then, the suspension was (1 Torr = 133.33 Pa) before moving them into the analy-
neutralized up to a pH ca 8 by slow addition of a solution sis chamber. Pressure in this turbo-pumped main vacuum
of ammonium hydroxide (Merck, 20% in NH3) at a neu- chamber was maintained below 7 ꢃ 10ꢂ9 Torr during data
tralization rate of 0.002 mol OHꢂ/min. The precipitate so acquisition. Each spectral region was signal averaged for a
obtained was thoroughly washed twice with deionized wa- number of scans to obtain good signal-to-noise ratios. Ac-
ter and dried in air at 373 K for 48 h. The solid was finally curate binding energies (BE) were determined by reference
calcined at 823 K in air for 2 h. This procedure allowed one to the C 1s and Si 2p lines to which BE values of 284.6 and
to deposit hydrated Ti(III) oxide and, after the thermal 103.4 eV were respectively assigned (30).
treatment, to obtain TiO2 dispersed on the silica surface.
Temperature programmed reduction (TPR). The TPR
Four supports were prepared using the amounts of TiCl3
profiles were obtained with a Cahn 2000 microbalance
necessary to obtain TiO2 contents equivalent to 0.6, 1.0,
(sensitivity = 1 ꢁg). Prior to the experiment, the samples
1.5, and 2.0 monolayers (ML) on the silica (denoted here-
were heated in a He flow (100 cm3/min) at 4 K/min up to
after as 0.6TiSi, 1.0TiSi, 1.5TiSi, and 2.0TiSi, respectively).
773 K in order to clean the surface. After cooling in He flow
The ML of TiO2 was considered as the complete covering
down to 373K, He wasswitched to H2 (300cm3/min) and the
of the silica surface by a film of TiO2 0.38 nm thick (28),
experiment was begun; the sample was heated at 4 K/min up
which corresponds to the longest axis of the rutile cell.
to 900 K, while continuously recording the weight changes.
A commercial TiO2 sample (Degussa P25) with BET area
of 50 m2/g and consisting of 82% anatase and 18% rutile
Chemical analysis. Quantitative analysis of vanadium
was used as support reference.
in catalyst samples was made by atomic absorption spec-
trometry. Prior to analysis, samples weighing ca 0.1 g were
dissolved in 10 ml of HF, heated on a sand bath until com-
Preparation of Catalysts
Catalysts were prepared by depositing vanadium on the plete solution, and then diluted to 100 ml with deionized
supports by wet impregnation, by using the specific reac- water.

OXIDATION OF ETHANOL ON V2O5/TiO2/SiO2 CATALYSTS
3
Electron spin resonance spectroscopy (ESR). ESR
TABLE 1
spectra were obtained with a Bruker ER 200D spectrome-
ter operated in the X band. All the spectra were recorded
at 77 K. DPPH was used as a reference (g = 2.0036) to
calibrate the magnetic field. Samples were introduced in
a quartz cell provided with greaseless stopcocks and were
outgassed in high vacuum at 293 K, 573 K, and 773 K. Then
they were reduced under a continuous flow of hydrogen at
573 K and 773 K.
Chemical Analysis and Microgravimetric Data for Vanadia-
Containing Catalysts
Contenta
Theoretical
TPR
Analysis
b
Weight loss
(% )
Tr
Sample
V-Si
V
Ti
V
Ti
(K)
688
3.39
—
2.53
—
0.62
V-0.6TiSi
V-1.0TiSi
V-1.5TiSi
V-2.0TiSi
4.89
5.17
5.57
6.88
5.17
8.94
13.35
17.60
1.97
4.19
4.56
4.80
5.74
8.75
12.92
13.67
0.61
1.61
1.41
1.57
663
663
663
663
Catalytic Activity
Catalytic activity tests were made in a tubular, fixed bed
flow reactor at nearly atmospheric pressure in the tem-
perature range 400–700 K, with a residence time W/F =
45 g cat ꢄ h/mol of ethanol and feeding reacting mixtures
ethanol–oxygen–helium having the following composition
(in mol% ): ethanol 1.4, oxygen 27.5, and helium balance.
Ethanol (Prolabo p.a., 99.85 vol.% ), oxygen (SEO, 99.99% )
and helium (SEO, 99.99% ) were used as reactants and dilu-
ent, respectively. Catalyst samples (ca 250 mg, particle size
30–40 mesh) were diluted with SiC tips up to a bed vol-
ume of 5 ml. Reactants and products were analyzed on-line
by GC using two columns: 13X molecular sieve for O2 and
CO, and Porapak Q for the rest of products. C and O mass
balances of 100 ꢅ 5% were obtained. Conversion and selec-
tivity to products were calculated on a carbon atom basis,
expressed as mol% ratio of ethanol transformed to ethanol
fed, and ratio of ethanol transformed to each product to
total ethanol transformed, respectively. All measurements
were made after stabilizing the catalytic system for 1 h un-
der reaction conditions. Preliminary tests showed that in the
reaction conditions used, homogeneous reaction (in the ab-
sence of catalyst) started around 480 K with a conversion of
0.2 mol% , reaching 5.7 mol% at 593 K (28). As these val-
ues are around 15 to 20 times lower than those measured in
the catalyzed reaction tests, the influence of homogeneous
reaction can be considered to be negligible.
V-Ti
2.56
57.27
2.41
n.d.c
0.77
665
a Atomic percentages.
b Tr: temperature of maximum reduction rate.
c n.d. = not determined.
two washings with ethanol. Thus, while the incorporation of
a small amount of titanium produced a marked decrease of
the vanadium retained (40% ), when the titanium content
is equivalent to or a little higher than that corresponding
to the monolayer the vanadium retained (81% ) was higher
than for V/Si (74% ). For V/2.0TiSi only 70% of the vana-
dium was retained.
Using impregnation with vanadyl acetylacetonate in
ethanol, films of vanadia having thicknesses of the order
of 2–3 ML can be obtained, but vanadium not chemically
bound to the surface is removed by the further washings
with ethanol. This differs from the impregnation in toluene
solution, where crystalline V2O5 can be obtained after con-
secutive impregnation and calcination steps. The low reten-
tion degree for vanadium of the V/0.6TiSi sample (40% ),
lower than on V/Si, indicates that titanium incorporated
at the silica surface is hindering the vanadium retention,
probably because at low titanium content both vanadium
and titanium are bound at similar sites of the silica surface.
RESULTS
Surface Areas
Chemical Analysis
Specific surface areas of supports and catalysts are shown
in Fig. 1. The titania coating of silica produced an almost
linear increase of the BET areas as titania content increased
up to 1.5 ML (13.4 wt% Ti). The subsequent coating with
vanadia decreased the surface areas by a 30% or more,
producing very similar surface areas (85–92 m2/g) for the
ternary catalysts, always higher than that of V/Si (75 m2/g).
These effects were similar to those reported by Vogt et al.
(17). As the silica support is constituted of nonporous mi-
crospherical particles, the increase of specific surface areas
with titania content and the observed decrease after vana-
dium impregnation could be due to the deposition of porous
titanium oxide, but it could indicate that a small fraction of
the silica substrate was solubilized under the very low pH
The chemical analysis of the V/xTiSi catalysts (Table 1)
showed that the titanium amount retained on the silica cor-
responded to that calculated to deposit 0.6, 1.0, or 1.5 ML
on supports 0.6TiSi, 1.0TiSi, and 1.5TiSi, respectively, but in
the 2.0TiSi preparation only 78% of the theoretical amount
was retained, which corresponded to 1.6 ML. These results
indicate that the hydrated Ti(III) oxide in direct contact
with the silica surface is retained during the two washings
with deionized water, but only hydrated Ti(III) oxide not
interacting with the silica surface can be easily removed.
When vanadia was deposited onto the mixed supports, the
data in Table 1 show that there is not a direct relation be-
tween the titanium content and the vanadium retained after

4
QUARANTA ET AL.
who reported a higher reduction degree of vanadia when
supported on TiO2/SiO2 than on SiO2.
The TPR results accordingly indicate that vanadia inter-
acted mostly with titania. The presence of a single TPR
peak suggests the presence of only one vanadium species,
but having different reducibility, depending on the support.
The peak widening when titania content increased can be
due to a higher heterogeneity of the interaction between
vanadium and the support surface. This means, in other
words, that by varying the content of titanium which covers
the silica surface one can modify the active phase-support
interaction.
X-Ray Photoelectron Spectroscopy (XPS)
FIG. 1. Specific surface areas as a function of the nominal titania con-
tent. Symbols: open, xTiSi supports; full, V/xTiSi catalysts.
Prior to the incorporation of vanadia, the TiO2-coated
silica carriers were examined by XPS. The binding energy
(BE) of the Ti 2p3/2 peak at 458.5 eV was unchanged for
all the samples. The O 1s signal showed two components:
a major one at ca 532.9 eV associated with the lattice oxy-
gen of SiO2 and a minor one, whose intensity increased
(below 1) employed at the beginning of the precipitation of
Ti hydroxide from TiCl3 solution and the subsequent pre-
cipitation of an extra and porous SiO2 phase. Finally, some
of this porosity is lost during incorporation of vanadium
oxide.
–
with TiO2 content, at ca 530.8 eV due to Ti O bonds. The
titanium-to-silicon intensity ratios (ITi/ISi) calculated from
TGA and TPR
The differential thermogravimetric (DTGA) profiles ob-
tained from the dried supports before calcination at 823 K
show two peaks, one between 323 and 373 K and the other
between 508 and 513 K. They were assigned to water re-
tained from the preparation and to desorption of NH4Cl
formed during the neutralization step, respectively. This as-
signment was possible by monitoring with XPS the N 1s
and Cl 2p core levels (spectra not shown here) of the solid
precursor at different stages of thermal decomposition.
The TPR profiles of each catalyst sample showed a sin-
gle reduction peak (Fig. 2). Profiles corresponding to bulk
V2O5, VTi, and VSi samples are also included for compara-
tive purposes. A clear influence of titania and silica supports
on the maximum reduction rate temperature (Tr) was ob-
served, as reported by other authors (18). Ternary catalysts
showed a close similarity to the V/Ti sample (Tr = 665 K),
the only noticeable difference among them being that peaks
widened as Ti content increased. Weight losses observed in
the reduction step of every V/xTiSi catalyst corresponded
to 1.00 ꢅ 0.04 oxygen atoms per vanadium atom, equiva-
lent to the complete reduction of all the V2O5 in the sam-
ple to V2O3, as in the V/Ti sample (Table 1). The only
exception was V/1.0TiSi whose reduction went further at
high temperatures, possibly due to the reduction of surface
titanium in patches of TiO2 not covered by vanadia. Con-
versely, weight loss for the VSi sample was intermediate
between those corresponding to reduction to the lower ox-
ides V2O3 and V2O4. These findings agree well with those
FIG. 2. TPR profiles obtained microgravimetrically for supported
reported by Baiker et al. (18) and Rajadhyaksha et al. (21) vanadia catalysts. Labelling is included in figure.

OXIDATION OF ETHANOL ON V2O5/TiO2/SiO2 CATALYSTS
5
exposure to X-ray radiation, since the V/Ti catalyst after
XPS analysis showed a strong grey-bluish color which
contrasts with the yellow and pale yellow color of V/Si and
V/1.0TiSi samples, respectively. A higher hydroxylation
of the sample surface can diminish the effect of the X-ray
photoreduction of V2O5. In order to obtain an idea of
the dispersion of vanadia on the carriers, the V/(Ti + Si)
atomic ratios have been calculated from the intensities of
the XP spectra and corrected by atomic sensitivity factors.
Table 2 also shows that the V/(Ti + Si) atomic ratio for
V/xTiSi catalysts increased with the titanium content, and
for V/Ti it was the highest of all.
FIG. 3. XPS titanium-to-silicon intensities surface ratio of TiO2-
coated silica supports vs bulk composition (the dashed line corresponds
to the intensity ratios expected for monolayer catalysts). Symbols: open,
xTiSi samples; full, V/xTiSi catalysts. Samples A, B, C, and D refer to
X = 0.6, 1.0, 1.5, and 2.0, respectively.
XPS experiments with mixed supports and catalysts indi-
cate that incorporation of the active phase did not modify
the interaction between the initially deposited Ti and the
main support, SiO2, as the same tendency of titania dis-
persion was observed before and after impregnation with
vanadium (Fig. 3).
XP spectra increased progressively with increasing titania
loading. In order to have an idea of the dispersion degree
of TiO2 on the silica surface, a model consisting of a TiO2
layer on the silica surface with 100% dispersion has been
assumed and the theoretical ITi/ISi ratio has been calculated
(31). The theoretical ITi/ISi ratio as a function of the Ti/Si
atomic ratio in the bulk is given by a dashed line in Fig. 3.
As can be seen in Fig. 3, the titanium dispersion was indeed
very high for 0.6TiSi, fitting well with that expected for the
theoretical monolayer; the dispersion was a little lower for
1.0TiSi and decreased for samples 1.5TiSi and 2.0TiSi.
The ternary catalyst samples were also examined by
XPS, and the respective BE of V 2p3/2, Ti 2p3/2, and/or Si 2p
peaks are compiled in Table 2 for comparative purposes.
Regarding the V 2p3/2 BE values, the decrease of 0.8 eV of
the V/1.0TiSi catalyst with respect to that of the V/Si coun-
terpart agrees well with literature findings (32, 33). V/xTiSi
and V/Ti samples (the exception is V/1.0TiSi) showed an
additional shift of 1.0 eV toward lower BE, which is typical
of V4+ ions. Such a decrease can be explained in terms
of a larger extent of reduction of surface V5+ ions upon
Electron Spin Resonance (ESR)
(a) xTiSi supports. In order to obtain information on
the TiO2 component in the support, the xTiSi samples were
treated under vacuum and H2 at 573 K and 773 K and then
studied by ESR. The Ti3+ ions generated by these treat-
ments were used as labels to obtain information on the dis-
tribution and redox properties of the titanium located on
the silica surface.
Outgassing treatments at Tv = 573 K were necessary to
generate Ti3+ cations in the supports, except in the case of
0.6TiSi which needed a higher Tv, indicating that the Ti4+ in
the sample is less reducible. For Tv = 773 K, all the spectra
showed a broad Ti3+ signal (Fig. 4) with higher intensity
than for Tv = 573 K; a narrow signal with g = 2.002 due to
some carbon impurities or to electrons trapped at oxygen
vacancies was also observed. The lineshape of the Ti3+ sig-
nal, which appears with g < 2, can give some indications
on the Ti3+ environments. In the case of sample 0.6TiSi,
the relatively broad and not well-defined axial Ti3+ signal
(Fig. 4a) indicates very heterogeneous environments of the
paramagnetic centers. These titanium cations present at low
titanium content and less reducible than titanium in TiO2
clusters should be due to dispersed Ti3+ cations on the SiO2
surface. For sample 1.0TiSi, the Ti3+ signal shows a slightly
narrower lineshape and reducibility at lower outgassing
treatments probably because in this sample a part of the
titanium is well dispersed and another is forming TiO2 clus-
ters. The ESR spectrum of sample 1.5TiSi (Fig. 4c), shows a
narrower axial Ti3+ signal with an intensity lower than that
of the previous one, although the Ti content is higher. Both
facts indicate that the Ti3+ ions are located in larger TiO2
clusters with more homogeneous environments than in the
previous samples. In these clusters TiO2 reduction is prob-
ably more difficult and the formation of diamagnetic Ti3+
pairs more likely. For sample 2TiSi (Fig. 4d) the spectrum
TABLE 2
Binding Energy (BE) of Core Electrons and XPS Intensity and
Atomic Ratios for Fresh Vanadia-Containing Catalysts
XPS atomic
BE (eV)
ratios
Sample
V-Si
V 2p3/2
518.4
Ti 2p3/2
—
IV/ITi
IV/ISi
V/(Ti + Si)
—
0.333
0.046
V-0.6TiSi
V-1.0TiSi
V-1.5TiSi
V-2.0TiSi
516.7
517.6
516.7
516.7
458.5
458.5
458.5
458.5
1.345
1.721
1.658
1.933
0.241
0.354
0.381
0.543
0.032
0.047
0.051
0.072
V-Ti
516.9
458.5
0.197
—
0.182

6
QUARANTA ET AL.
tion as VO2+. Reduction at 773 K followed by contact-
ing the sample with O2 at 77 K yields some dispersed
VO2+ species formed by reoxidation of V3+ cations, with
g = 2.005, g = 2.000, A = 200 G, and A = 50 G, signal
k
⊥
k
⊥
A, and in some cases, O2ꢂ
V
5+ species are also formed, with
–
g1 = 2.023, g2 = 2.009, g3 = 2.003, A1 = 10 G, A2 = 7 G, and
A3 = 6 G. At 295 K the VO2+ signal increases and the O2ꢂ
signal disappears.
The spectra of V/0.6TiSi after thermal treatments under
vacuum and H2 show, in general, effects similar to those
observed for sample V/Si. The reduction of V5+ to V3+
takes place at Tv = 573 K under H2, with some formation of
VO2+ cations which are reduced at Tr = 773 K; a subsequent
ꢂ
O2 adsorption produced VO2+ and some O2
V
5+ species.
–
However, a small VO2+ signal is formed by outgassing at
573 K, indicating that some V5+ are reduced, and hyperfine
constants of the VO2+ signal detected after O2 adsorption
on the sample reduced at 773 K show parameters slightly
different from those of signal A; both modifications should
be due to the interaction of vanadium and titanium cations
dispersed in silica. As in the case of V/Si, most of the vana-
dium is not forming vanadium oxide clusters.
A first difference shown by the spectrum of V/1.0TiSi
after calcination is the presence of a broad signal B with
g = 1.95 that can be assigned to V4+ cations in vanadium
oxide clusters. The signal of dispersed vanadium is also
present (Fig. 5b), showing its largest intensity after re-
duction at 573 K, as in the previous cases. The broad
signal B presents its maximum intensity after outgassing at
Tv = 573 K, indicating that the clustered vanadium requires
FIG. 4. ESR spectra at 77 K of mixed supports after outgassing at
773 K: (a) 0.6TiSi; (b) 1TiSi; (c) 1.5TiSi; (d) 2TiSi.
is formed by the overlapping of at least four narrow Ti3+
signals, with g ꢆ 1.991, 1.950, 1.975, and 1.969 and g
⊥
k
close to 1.945. The relative narrow signals and lower inten-
sity indicate an improved crystallinity of the TiO2 clusters.
The signal with g = 1.991 is indicative of Ti3+ in anatase,
⊥
while the other titanium signals can be assigned to Ti3+ in
rutile.
Thermal treatments under H2 produced a broadening of
the Ti3+ signal and a marked intensity increase, except for
2.0TiSi which showed a small increase, probably because of
the formation of diamagnetic Ti3+ pairs. This is the usual
mechanism of TiO2 crystals to accommodate the oxygen
vacancies produced by reduction treatments, and it can be
taken as further evidence of an improved crystallinity of
the titanium oxide.
(b) V/xTiSi catalysts. The ESR study on V/xTiSi cata-
lysts provides information mainly on V4+ (or VO2+), an ox-
idation state present in a small part of the vanadium cations
because of its low stability. Therefore, the observed cations
should be considered mainly as labels and not from a quan-
titative point of view.
In the catalyst V/Si only a very small amount of vana-
dium cations are stabilized as dispersed VO2+ species af-
ter calcination; most of the vanadium should be present as
diamagnetic V5+. These cations are not reduced by out-
gassing treatments even at 773 K, but treatments under
FIG. 5. ESR spectra of catalysts: (a) V/0.6TiSi; (b) V/1.0TiSi; (c)
H2 at 573 K can reduce V5+ to V3+ with little stabiliza- V/2.0TiSi.

OXIDATION OF ETHANOL ON V2O5/TiO2/SiO2 CATALYSTS
7
milder reduction conditions than the vanadium dispersed
on silica. Both signals disappear after reduction at Tr = 773
K; a subsequent contact with oxygen produces again the
signal of vanadium dispersed on silica, and a very small
–
one of Oꢂ2
V
5+, but not the signal of clustered vanadium.
The spectra obtained for V/1.5TiSi are relatively similar
to those obtained for sample V/1.0TiSi, indicatingthat vana-
dium is present in the form of vanadium oxide aggregates
and of vanadium dispersed on silica; however, the contact
with O2 of the sample reduced at 773 K did not lead to the
generation of VO2+ signals, as was the case for the samples
with lower titanium content. This difference, which origi-
nated when the Ti content is higher than that correspond-
ing to the theoretical monolayer, is probably due to the
interaction of the vanadium dispersed on silica with TiO2
clusters which are covering most of the silica surface. After
the reduction at 773 K, the interaction of V3+ and reduced
titanium oxide clusters probably affect the vanadium reox-
idation mechanism, favoring the reoxidation of V3+ to V5+
when oxygen is adsorbed as O2ꢂ or O22ꢂ not observable
by ESR.
FIG. 6. Effect of catalyst support on the catalytic oxidation of ethanol
on vanadia catalysts supported on: ᮀ, SiO2; ᭿, 0.6TiSi; ᭡, 1.0TiSi; ᭹,
1.5TiSi; ᭞, 2.0TiSi; and ꢁ, TiO2. Experimental conditions are in the text.
The spectra obtained for V/2.0TiSi are similar to those
obtained for V/1.5TiSi, except that the contribution of sig-
nal A is lower and reduction at 573 K eliminates completely
the VO2+ signal from vanadium oxide clusters. This sec-
ond difference indicates that the vanadium oxide clusters
formed on better crystallized titanium oxide are more eas-
ily reduced, probably because there is a more important
interaction between the two oxides. The samples contain-
ing titanium oxide clusters showed small Ti3+ signals after
reduction at 773 K, being smallest in the case of V/2.0TiSi,
probably because in crystalline TiO2 part of the Ti3+ cations
form diamagnetic Ti3+ pairs. These Ti3+ signals disappeareꢂd
on contact with O2 without the formation of VO2+ or O2
signals, indicating as in the previous case that the forma-
tion of diamagnetic oxygen species such as O2ꢂ or O22ꢂ is
favored.
ported on pure titania; however, V/2.0TiSi showed a rather
exceptional behavior, its activity exceeding that of the V/Ti
sample. Considering the V/xTiSi catalysts, this surprising
behavior cannot be explained in terms of surface area, be-
cause the specific surface area was very similar for all of
them; however, the differences in surface area could con-
tribute to the conversion differences between V/2.0TiSi and
V/Ti. Within the 95% confidence interval, the values of ap-
parent activation energy calculated for the low conversion
range were similar for all V/xTiSi catalysts (54 ꢅ 21, 46 ꢅ 5,
59 ꢅ 10, and 56 ꢅ 5 kJ/mol for x = 0.6, 1.0, 1.5, and 2.0, re-
spectively), and similar to that for V/Si (54 ꢅ 7 kJ/mol).
Reaction Products and Their Distribution
In the conditions used, acetaldehyde (AcH) was the main
product of ethanol oxidation on every catalyst. As no H2
Effect of TiO2 Content on the Catalytic Activity
The activity of the V/xTiSi catalysts became measurable was detected among the products, the formation of AcH
already around 380 K. The variation of activity of the cata- seems to proceed via the oxidative dehydrogenation of
lysts with reaction temperature is shown in Fig. 6. As the ethanol. Figure 7 shows the variation of selectivity to AcH
oxidation reaction is exothermic, a slight increase in the as a function of the total conversion. The selectivity to
temperature of the catalyst bed (approximately 15 K for AcH of V/xTiSi catalysts at a constant conversion level was
the highest conversion levels) is produced. Therefore tem- affected by the titania content in a rather complex way,
peratures reported in Fig. 6 are adiabatic temperatures. The with V/0.6TiSi showing a higher selectivity than V/1.0TiSi.
results obtained with the catalysts supported on bare silica V/2.0TiSi showed an extremely high selectivity to AcH up
(V/Si) and pure titania (V/Ti) supports are also included for to high conversion levels, thus giving rise to yields of AcH
comparative purposes. Ethanol conversion increased with up to 72 mol% . The selectivity of V/Ti was high at low
increasing temperature. As can be seen, the activities of conversion levels, but decreased markedly above 20% con-
V/Si and V/0.6TiSi were relatively similar, but the activity version.
of the catalysts with higher titanium content was superior to
Additional information on the catalytic behavior can be
that of V/Si. In general, the activity of the catalysts contain- obtained by analyzing the nature and distribution of the mi-
ing clustered titania approached that of the vanadium sup- nor products: carbon oxides, acetic acid, ethene, and ethyl

8
QUARANTA ET AL.
minor products over four representative catalysts (V/Si,
V/Ti, V/1.0TiSi, and V/2.0TiSi) are shown as a function of
ethanol conversion. The patterns with samples V/0.6TiSi
and V/1.5TiSi were qualitatively similar to that of V/1.0TiSi.
Both carbon oxides were produced on every catalyst with
their yields increasing with the conversion (and tempera-
ture). Carbon dioxide was usually detected at lower con-
versions but its yield increased with the temperature at a
much lower rate than that of CO, which became detectable
at higher conversions. Significant amounts of acetic acid
were formed only on Ti-containing samples, especially on
V/Ti (up to a 10 mol% yield; Fig. 8c). This indicates that,
besides the higher activity, VTi showed another substan-
tial difference with the VSi sample in the ability to oxi-
dize the initially formed AcH to acetic acid. Besides, it may
explain the lower selectivity to AcH in the medium con-
version range observed with these catalysts, as compared
to that of the V/Si counterpart. At variance with the Ti-
containing catalysts, ethene was the main by-product on
VSi at low conversions, indicating the presence of active
centers for dehydration. Ethene was also detected among
the products on V/xTiSi catalysts, but in much lower yields
and at much higher conversions (the highest yield, less than
4% , was obtained at ca 90% conversion). Ethyl acetate was
also detected among the products on every catalyst, but
in much lower yields. Interestingly, it was detected even at
FIG. 7. Influence ofcatalyst support composition and totalconversion
on the selectivity to AcH. Symbols as in Fig. 6.
acetate. Diethyl ether and formic acid were also formed in
some experiments (especially on V/Si), but as their yield
never exceeded 1% nor their selectivity 2–3% , they will
not be taken into account. In Fig. 8, the distribution of
FIG. 8. Dependence of yields to minor products on ethanol conversion in the selective oxidation of ethanol on V/Ti (a), V/1.0TiSi (b), V/2.0TiSi
(c), and V/Si (d). Symbols: ᮀ, CO; ᭿, CO2; ꢁ, acetic acid; ᭡, ethyl acetate; and 4, ethene.

OXIDATION OF ETHANOL ON V2O5/TiO2/SiO2 CATALYSTS
9
temperatures and conversion levels where no acetic acid area is produced by increasing its “roughness.” At higher
was detected among the products.
Ti content, the number and size of these clusters increases,
partially covering the support surface as patches of defec-
tive TiO2. The result is a decrease of the XPS-dispersion
of titanium and the appearance of TiO2 particles (Ti3+ sig-
nals in ESR spectra) in sample 1.5TiSi (Fig. 9c). This in-
terpretation is in good agreement with the Raman results
reported by Jehng and Wachs (34), who found that only
surface TiO2 species are formed on 3% TiO2/SiO2 samples,
prepared from alkoxide precursors, while both dispersed
TiO2 and TiO2 (anatase) particles were found in a sam-
ple with a much higher TiO2 content (40% ). Once enough
titanium is used to reach a situation in which most of the sil-
ica surface is covered by three-dimensional arrangements
of TiO2, a further increase of added titanium to the silica
suspension should lead to the completion of the hydrated
Ti(III) oxide covering the silica surface and the formation
after calcination of better crystallized TiO2 (Figs. 9d and
9e), although those deposits will not be much thicker than
the more defective deposits (see the small difference in ti-
tanium content between 1.5TiSi and 2.0TiSi).
DISCUSSION
Surface Model
The results presented here show that the deposition of
titania monolayers onto silica as a way to modify its prop-
erties as a catalyst support is a complex phenomenon in
which the amount of added titanium plays an important
role. Based on these results, a simplified model of the way
increasing amounts of titania are deposited onto silica and
how the nature of this deposit influences the later deposi-
tion of vanadia onto it may be proposed.
When small amounts of TiCl3 are added to the stirred
suspension of silica, titanium ions become bonded to the
oxygen of silanol surface groups that hydrolyze the chlo-
ride. Consequently, titanium is highly dispersed at the sil-
ica surface (Fig. 9a), as already confirmed by quantitative
XPS data (Fig. 3) and the ESR spectrum of sample 0.6TiSi
(Fig. 4a). As the Ti content increases, the probability of
Ti ions to be linked in small TiO2 clusters on the silica
surface (Fig. 9b) increases. This picture is supported by
the monolayer-like dispersion found by XPS, as well as
by the highly heterogeneous environment of Ti ions shown
by ESR. ESR spectra of 1.0TiSi indicate that even in sample
1.0TiSi a substantial fraction of the silica surface remains
uncovered. As a consequence, an increase of the surface
Reaction Pathway
Acetaldehyde formation over V/xTiSi catalysts proceeds
through the oxidative dehydrogenation of ethanol. The for-
mation of acetic acid should occur as a consecutive reaction
by oxidation of the initially formed acetaldehyde. We found
that the yield of acetic acid depended on the nature of the
catalyst support. This means that the support may influence
not only the dispersion of the supported vanadia but also
its oxidizing power.
In our experimental conditions, little acetic acid was
formed on the V/Si catalyst. Oyama et al. (11) reported
a selectivity to acetic acid of ca 23% at ca 23% conver-
sion during the oxidation of ethanol on a 9.8% V2O5/SiO2
catalyst. The difference is probably due to the high vana-
dia loading of their sample, which will hinder the effects
of vanadia–silica interactions, as well as to the use of wa-
ter in their feed. In our case, the V/Ti catalyst produced
a high amount of acetic acid, and it was also produced in
significant amount on other Ti-containing catalysts. The sur-
prising effect of the near absence of acetic acid among the
products on the V/2.0TiSi catalyst until very high conver-
sions can be explained by considering the reaction temper-
atures. In Fig. 10 the yields of acetic acid on VTi, V/1.0TiSi,
and V/2.0TiSi catalysts as a function of reaction tempera-
ture are compared. As can be seen, acetic acid began to be
detected in the products on both V/xTiSi catalysts at tem-
peratures around 480 K and disappeared above 600 K, as
can be observed also in the results of Oyama et al. (11).
A similar effect when studying the oxidative dehydrogena-
tion of ethanol on MoO3/Al2O3 (35) and V2O5/Al2O3 cata-
lysts (36) has been reported. In the latter case, it was found
that the temperature at which acetic acid formation starts
FIG. 9. Schematic model of the surface of supports as a function of Ti
content: (a) 0.6TiSi; (b) 1.0TiSi; (c) 1.5TiSi; (d) 2.0TiSi precursor before
washing; and (e) final state of 2.0TiSi support.

10
QUARANTA ET AL.
the V/Si catalyst, which is corroborated by the formation of
the highest quantities of diethyl ether observed on this sam-
ple. In contrast with Ti-containingsamples, the veryhigh BE
value of the V 2p3/2 peak observed on fresh V/Si (see above)
could be interpreted as due to V5+ cations coordinated to
hydroxyl groups. These groups could be responsible for the
higher acidity of this sample revealed by the dehydration
of ethanol.
The much lower amounts of ethene found among the
products on Ti-containing catalysts agree well with the re-
sults of Bond and Flamerz (10) on isopropanol decom-
position on titania-supported vanadia. They found that,
although both V2O5 and TiO2 act principally as dehydra-
tion catalysts at 473 K, the latter is much more active
than the former; V2O5/TiO2 catalysts exhibit faster rates
for dehydrogenation than for dehydration. Nakagawa et al.
(41) reported that the rates of oxidation of isopropanol
or ethanol in the absence or presence of gaseous oxygen
over V2O5/TiO2 catalysts containing below 11 wt% of V2O5
are similar to each other. As our V/xTiSi catalysts contain
around 4–5 wt% of vanadia, one would expect to find a
similar catalytic behavior, provided the vanadia phase is
interacting mostly with the titania. This difference in the
formation of ethene constitutes an additional proof of the
FIG. 10. Dependence of the yield of acetic acid on the reaction tem-
perature. Experimental conditions are in the text.
is independent of space velocity. Thus, this seems to be a
general phenomenon and can be related to the adsorption
of the acetic acid. A possible interpretation is that, even if
it is formed on the catalyst surface at lower temperatures,
acetic acid is so strongly adsorbed that it cannot desorb at
temperatures below 480 K.
At temperatures above 600 K, acetic acid probably either
decomposes by decarboxylation or decarbonylation (as CO
is the main combustion product) or it reacts further to form
ethyl acetate. However, we did not observe an increase of
ethyl acetate selectivity as the yield of acetic acid decreased,
which points to its decomposition to carbon oxides. Further-
–
preferential interaction TiO2 V2O5 in V/xTiSi catalysts.
Summarizing, the oxidation of ethanol over VxTiSi cata-
lysts proceeds through an initial oxidative dehydrogenation
to acetaldehyde, followed by three consecutive parallel re-
actions: degradation by total oxidation to carbon oxides,
acetaldehyde condensation to give ethyl acetate, and, de-
pending on the temperature, oxidation to acetic acid.
Effect of TiO2 Coating
An increase of x in the binary xTiSi supports prior to
more, the ester was found among the products at conditions the deposit of vanadium oxide leads to a pronounced in-
(reaction temperature and conversions), where no acetic crease of the activity of the resulting catalysts, especially at
acid was detected, which could indicate that the formation low temperatures, and to changes in the distribution of the
of ester proceeds by a way different from esterification of products. There is a general agreement in the literature that
ethanol with the acid. Ethyl acetate can also be formed by the active sites for ethanol oxidation on vanadia catalysts
condensation of acetaldehyde in the presence of aluminum supported either on TiO2 (41) or on SiO2 (11) are surface
==
ethoxide catalyst (Tischtschenko reaction) (37), or by ox-
V O species and that the promoting effect of titania as a
idative condensation of acetaldehyde and the remaining support comes from the effective dispersal of V2O5 on the
alcohol, as has been reported over metallic palladium on titania surface, as well as the effect on the strength of these
==
active charcoal catalyst (Pd/C) (38) and over several transi-
V O bonds. Thus, the differences observed in the activity
tion metal oxides (39). As both xTiSi supports and V/xTiSi of the various V/xTiSi catalysts and with those of V/Ti and
catalysts show rather strong Lewis acid sites in their sur- V/Si counterparts should be explained in terms of the vari-
faces (23, 24), the condensation of acetaldehyde seems to ation of the interaction of the vanadia with the underlying
be the more probable route.
support surface.
The formation of ethene on V/Si catalyst, even at low
As stated above, surface areas of V/xTiSi catalysts are
conversions, and the pattern of its yield with total conver- very close each other and they cannot account for the ob-
sion indicate that it was formed through a parallel route on served variation of activity. Moreover, the activity of the
this catalyst. Dehydration of alcohols is a characteristic re- V/Ti sample, having a much lower surface area, was higher
action on acidic sites of transition metal oxides (40). Thus, than most of the catalysts on mixed supports. A second fac-
ethene formation indicates the presence of acidic sites on tor to consider is vanadium dispersion. Quantitative XPS

OXIDATION OF ETHANOL ON V2O5/TiO2/SiO2 CATALYSTS
11
analysis of fresh samples showed that the V/(Ti + Si) atomic most of the vanadium forming vanadium oxide clusters on
ratios of V/xTiSi catalysts increased with x in the range poorly crystallized titania particles. This type of vanadium is
x = 0.6–2.0, but not in the same proportion as the vana- reduced more easily than the vanadium dispersed on silica,
dium content. The vanadium dispersion is not too differ- as indicated by ESR. The better selectivity to acetaldehyde
ent for V/Si, V/0.6TiSi, and V/2.0TiSi, but it is lower for on V/1.5TiSi could be related to the absence of vanadium on
V/1.0TiSi and V/1.5TiSi. V/Ti presented the highest value silica that can give rise to Oꢂ2 radicals that favor total oxida-
of the V/(Ti + Si) ratio, indicative of a high vanadium dis- tion. The main difference between 1.5TiSi and 2.0TiSi is the
persion, despite showing the lowest surface area. Consid- improved crystallinity of the latter, as shown by ESR. The
ering that vanadium cations are forming part of the active better crystallinity of TiO2 in V/2.0TiSi will favor titania–
sites for ethanol conversion, different vanadium dispersion vanadia interactions able to facilitate the formation ofvana-
could explain differences in conversion, as is the case for dia layers that are more active for ethanol oxidation than
V/1.0TiSi, V/1.5TiSi, and V/2.0TiSi catalysts. The only ex- the vanadium on less crystalline TiO2, probably because of
ception to this tendency is sample V/0.6TiSi, in which the an easy vanadia reduction, in addition to the higher vana-
vanadium dispersion is higher than on V/1.0TiSi, although dium dispersion in V/2.0TiSi. Accordingly, the much higher
its activity is lower. These observations and the increase activity of V/2.0TiSi, compared to that of V/1.5TiSi, seems
of conversion with increasing titanium content suggest that to be due to the presence of these TiO2 crystallites causing
differences in the nature and properties of the vanadium a higher V dispersion and a better V reducibility.
species induced by the nature of the support are also im-
portant in explaining catalyst performance.
At low temperatures ethanol conversion on V/2.0TiSi
was equivalent to that on V/Ti, as one should expect, tak-
Nakagawa et al. (30) reported that the rates of oxidation ing into account the presence of crystalline TiO2 in both
of ethanol in the presence of O2 and of reduction by ethanol samples. However, V/2.0TiSi became much more active
in its absence were similar over V2O5/TiO2 catalysts, show- above ca 450 K. This difference of activity seems to in-
ing a direct correspondence between catalyst reducibility dicate that the nature of these small crystalline aggregates
and activity for the oxidation of the alcohol. Thus, changes of TiO2 is different from that of bulk titania. These differ-
in the reduction properties may explain the increase of ac- ences should come from the genesis of these aggregates,
tivity caused by the coating of silica with TiO2. TPR re- different from that of bulk titania. They were formed by
sults showed that the maximum reduction rate occurred the deposition of a much higher amount of TiO2 onto the
at the same temperature for every V/xTiSi catalyst, practi- silica surface, followed by removal of the outermost (and,
cally equal to that for V/Ti and 25 K lower than for V/Si. hence, less influenced by silica) layers. Thus, silica probably
Although these results cannot be directly related to cat- acts as a template that directs the crystal habits of these ag-
alytic activity because the measurements are carried out gregates, which may lead either to a preferential growth of
at higher temperatures, they indicate that interaction with some crystal planes or to a different proportion between the
TiO2 causes a higher reducibility of vanadium than its in-
teraction with silica.
This reducibility effect is corroborated by ESR results
and may be explained by the nature of the titania and vana-
dia species in the catalysts as revealed by this technique.
Vanadium in V/0.6TiSi is dispersed on the silica surface,
where also titanium is present as very dispersed entities,
as also indicated by XPS (Fig. 3), where the ITi/ISi ratio
measured for this sample is close to that corresponding to
100% dispersion. The vanadium dispersion in this sample
is similar to that on V/Si and the reduction and reoxida-
tion mechanisms (formation of O2ꢂ
V
5+
and stabilization
–
of VO2+) are also similar; the titanium incorporation does
not affect the temperature needed to obtain 50% conver-
sion, but favors the conversion at lower temperatures. In
V/1.0TiSi there are patches of silica not covered by TiO2,
but very little of the silica surface remained uncovered for
samples V/1.5TiSi and V/2.0TiSi, where V is mainly concen-
trated as vanadium oxide clusters on the TiO2 aggregates
and their surroundings.
The similar activity of V/1.0TiSi and V/1.5TiSi catalysts
FIG. 11. Influence of catalyst composition and reaction temperature
can be explained now, considering that both samples have on the yield of acetaldehyde. Symbols as in Fig. 8.

12
QUARANTA ET AL.
–
rutile and anatase phases, or which affects the Ti O bonds, our laboratory in order to optimize these conditions for a
thus influencing the interactions of these aggregates with better acetaldehyde yield.
the supported vanadia.
As a consequence, the use of a titanium content (ca
13 wt% ) that is close to 1.5 ML, but prepared with an ex-
cess of titania precursor, allows the formation of crystalline
TiO2 particles with specific characteristics which through
interactions with supported vanadium oxide modify the
surface properties of the catalyst. This different interac-
tion increases the catalyst activity and selectivity to AcH.
The combination of these factors brings about a substan-
tial increase of the acetaldehyde yield, compared to that
on catalysts supported either on silica or titania, as can be
observed in Fig. 11, reaching a value of 72 mol% at 473 K.
To our knowledge this value is one of the highest reported
in the open literature.
ACKNOWLEDGMENTS
N.E.Q. acknowledges a grant from the Spanish State Secretariat of
Science and Technology. Partial financial support from a cooperative re-
search program between CONICET (Argentina) and CSIC (Spain) is also
acknowledged.
REFERENCES
1. Murakami, Y., Inomata, M., Mori, K., Ui, T., Suzuki, K., Miyamoto,
A., and Hattori, T., in “Preparation of Catalysts III” (G. Poncelet,
P. Grange, and P. A. Jacobs, Eds.), p. 531, Elsevier, Amsterdam,
1983.
2. Kijenski, J., Baiker, A., Glinski, M., Dollenmeier, P., and Wokaun, A.,
J. Catal. 101, 1 (1986).
3. Wachs, I. E., and Hardcastle, F. D., in “Proceedings of the 9th In-
ternational Congress on Catalysis, Calgary, 1988” (M. J. Phillips
and M. Ternan, Eds.), Vol. 3, p. 1449, Chem. Inst. Canada, Ottawa,
1988.
CONCLUSION
4. Jonson, B., Rebenstorf, B., Larsson, R., and Andersson, S. L. T.,
J. Chem. Soc., Faraday Trans. I 84, 1897 (1988).
5. Bond, G. C., and Bru¨ckman, K., Faraday Disc. Chem. Soc. 72, 235
(1981).
6. Bond, G. C., and Ko¨nig, P., J. Catal. 77, 309 (1982).
7. Mori, K., Inomata, M., Miyamoto, A., and Murakami, Y., J. Phys.
Chem. 87, 4560 (1983).
8. Lopez Nieto, J. M., Kremenic, G., and Fierro, J. L. G., Applied Catal.
61, 235 (1990).
9. van Hengstum, A., van Ommen, J. G., Bosch, H., and Gellings, P. J., in
“Proceedings of the 8th International Congress on Catalysis, Berlin,
1984,” Vol. 4, p. 297, Verlag Chemie, Weinheim, 1984.
10. Bond, G. C., and Flamerz, S., Appl. Catal. 33, 219 (1987).
11. Oyama, S. T., Lewis, K. B., Carr, A. M., and Somorjai, G. A., in “Pro-
ceedings of the 9th International Congress on Catalysis, Calgary, 1988”
(M. J. Phillipsand M. Ternan, Eds.), Vol. 3, p. 1449, Chem. Inst. Canada,
Ottawa, 1988.
Nonporous SiO2 microspheres have been coated by a
thin layer of TiO2 by using the homogeneous precipitation
method from TiCl3 (28). The resulting xTiSi supports, with
x ranging between 0.6 and 2.0 ML, displayed different tita-
nium distributions. At low content, titanium was deposited
as a very dispersed TiO2 species on the silica surface. As
the content increased, it formed TiO2 clusters, starting to
cover the silica surface; for x = 1.5 ML the silica surface
is covered by these clusters. The use of larger amounts of
titanium in the suspensions did not lead to a significant in-
crease of the titanium content, although it increased titania
crystallinity after calcination. The incorporation of vana-
dium oxide onto xTiSi and the parent TiO2 and SiO2 carri-
ers leads also to a different vanadium distribution. A high
vanadium dispersion is observed for V/Si and V/0.6TiSi; 12. Bosch, H., and Janssen, F., Catal. Today 2, 369 (1987).
13. Cole, D. J., Cullis, C. F., and Hucknall, D. J., J. Chem. Soc., Faraday
Trans. I 72, 2744 (1976).
14. Gasior, M., Gasior, I., and Grzybowska, B., Appl. Catal. 10, 87 (1984).
in both cases vanadia is deposited with little clustering on
the silica surface, affected by not too strong interactions
with the dispersed titanium in 0.6TiSi. As titania content
15. Roozeboom, F., Mittelmeijer-Hazeleger, M. C., Moulijn, J. A.,
increased, vanadium was mainly concentrated on the TiO2
particles forming vanadium oxide clusters, and the disper-
sion increased with the TiO2 crystallinity.
Medema, J., de Beer, V. H. J., and Gellings, P. J., J. Phys. Chem. 84,
2783 (1980).
16. Takagi, M., Soma, M., Onishi, T., and Tamaru, K., Can. J. Chem. 58,
2132 (1980).
The results confirm that incorporation of V2O5 onto xTiSi
supports allows one to improve the catalyst performance,
which makes vanadia supported on TiO2-coated SiO2 espe-
cially suited for selective oxydehydrogenation of ethanol to
acetaldehyde. The use of a titanium content (13 wt% ) close
to 1.6 ML, but prepared with an excess of titania precursor,
allows one to form crystalline particles of TiO2 whose in-
teraction with vanadium oxide increases substantially the
catalyst activity and the selectivity to AcH. The combina-
tion of these factors brings about an important increase of
the acetaldehyde yield, as well as a marked reduction of the
reaction temperature, thereby permitting a molar yield of
72% per pass to be obtained at only 473 K in the experi-
mental conditions used here. Further work is in progress in
17. Vogt, E. T. C., Boot, A., van Dillen, A. J., Geus, J. W., Janssen, F. J.
J. G., and van der Kerkhof, F. M. G., J. Catal. 114, 313 (1988).
18. Baiker, A., Dollenmeier, P., Glinski, M., and Reller, A., Appl. Catal.
35, 365 (1987).
19. Reichmann, M. G., and Bell, A. T., Langmuir 3, 111 (1987).
20. Reichmann, M. G., and Bell, A. T., Appl. Catal. 32, 315 (1987).
21. Rajadhyaksha, R. A., Hausinger, G., Zeilinger, H., Ramstetter, A.,
Schmelz, H., and Kno¨zinger, H., Appl. Catal. 51, 67 (1989).
22. Wathoz, P., Ruwet, M., Machej, T., and Grange, P., Appl. Catal. 69, 149
(1991).
23. Galan-Fereres, M., Mariscal, R., Alemany, L. J., Fierro, J. L. G., and
Anderson, J. A., J. Chem. Soc., Faraday Trans. 90, 3711 (1994).
24. Galan-Fereres, M., Alemany, L. J., Mariscal, R., Ban˜ares, M. A.,
Anderson, J. A., and Fierro, J. L. G., Chem. Mat. 7, 1342 (1995).
25. Wang, L., Eguchi, K., Arai, H., and Seiyama, T., Chem. Lett., 1173
(1986).

OXIDATION OF ETHANOL ON V2O5/TiO2/SiO2 CATALYSTS
13
26. Allakhverdova, N. Kh., Adzhamov, K. Yu., and Alkhazov, T., Kinet.
Katal. 33, 261 (1992).
33. Horwath, B., Strutz, J., Geyer-Lippmann, J., and Horwath, E. G.,
Z. Anorg. Allg. Chem. 483, 181 (1981).
27. Quaranta, N. E., Corte´s Corbera´n, V., and Fierro, J. L. G., Stud. Surf.
Sci. Catal. 72, 147 (1992).
28. Geus, J. W., in “Preparation of Catalysts III” (G. Poncelet, P. Grange,
and P. A. Jacobs, Eds.), p. 1, Elsevier, Amsterdam, 1983.
29. van Hengstum, A. J., van Ommen, J. G., Bosch, H., and Gellings, P. J.,
Appl. Catal. 5, 207 (1983).
30. Wagner, C. D., Riggs, W. N., Davis, L. E., Moulder, J. F., and
Muilenberg, G. E., in “Handbook of X-ray Photoelectron Spec-
troscopy,” Perkin Elmer, Eden Prairie, MN, 1978.
31. Kerkhof, E. P. J. M., and Moulijn, J. A., J. Phys. Chem. 83, 1612
(1979).
34. Jehng, J.-M., and Wachs, I. E., Catal. Lett. 13, 9 (1992).
35. Requejo, F., Quaranta, N., Coronado, J. M., Soria, J., and Thomas, H.,
Stud. Surf. Sci. Catal. 82, 803 (1994).
36. Quaranta, N., Martino, R., Gambaro, L., and Thomas, H., Stud. Surf.
Sci. Catal. 82, 811 (1994).
37. Tischtschenko, A., Chem. Zentral. 77, 1309 (1906).
38. Kunugi, T., Matsuura, T., and Kono, T., Kogyo Kagaku Zashi 71, 1517
(1968).
39. Wang, L., Eguchi, K., Arai, H., and Seiyama, T., Appl. Catal. 37, 107
(1987).
40. Grzybowska-Swierkosz, B., Mat. Chem. Phys. 17, 121 (1987).
32. Fierro, J. L. G., Gambaro, L. A., Cooper, T. A., and Kremenic, G., 41. Nakagawa, Y., Ono, T., Miyata, H., and Kubokawa, Y., J. Chem. Soc.,
Appl. Catal. 6, 363 (1983). Faraday Trans. I 79, 2929 (1983).