Catalytic performances and stability of three Sb-Mo-O phases in the selective oxidation of isobutene to methacrolein

Article abstract of DOI:10.1021/jp983232h

We investigated the possible role of binary oxides associating Sb and Mo in the synergetic effects between physically mixed MoO₃ and α-Sb₂O₄ in the selective oxidation of isobutene to methacrolein. The catalytic behavior of three Sb-Mo-O phases-Sb₂MoO₆ on one hand, a mixture of Sb₂Mo₁₀O₃₁ and Sb₄MO₁₀O₃₁ on the other hand - was investigated. The samples were characterized using X-ray diffraction (XRD), confocal laser Raman microscopy, specific area measurements, and scanning electron microscopy. Additional in situ high-temperature XRD was also performed. At similar conversions of isobutene, the selectivities to methacrolein for the Sb-Mo-O phases were always lower than those of mixtures of MoO₃ with α-Sb₂O₄ with the same atomic compositions. A second result was that the three Sb-Mo-O phases were unstable in the conditions of reaction and tended to decompose to simple molybdenum and antimony oxides. An improvement of the catalytic performances of the Sb-Mo-O catalysts occurred in parallel to their decomposition. Some apparent cooperation effects were observed for mechanical mixtures of the Sb-Mo-O phases with MoO₃ α-Sb₂O₄. These effects are actually related to the decomposition of the Sb-Mo-O phases and are due to the cooperation between MoO₃. and α-Sb₂O₄ formed at their surface. The spontaneous decomposition of the Sb-Mo-O phases shows that these phases cannot form in operandi (during the catalytic reaction) starting from the simple oxides. Consequently, this shows that they cannot be the real most performant phases in the mixtures of MoO₃ with α-Sb₂O₄ and hence that they cannot be at the origin of the synergism between the simple oxides.

Full text of DOI:10.1021/jp983232h

10542
J. Phys. Chem. B 1998, 102, 10542-10555
Catalytic Performances and Stability of Three Sb-Mo-O Phases in the Selective Oxidation
of Isobutene to Methacrolein
E. M. Gaigneaux,*,†,‡,§ M. Dieterle,‡ P. Ruiz,† G. Mestl,‡ and B. Delmon†
Unite´ de catalyse et chimie des mate´riaux diVise´s, UniVersite´ catholique de LouVain, Place Croix du Sud 2/17,
B-1348 LouVain-la-NeuVe, Belgium, and Department of Inorganic Chemistry, Fritz-Haber-Institut der
Max-Planck-Gesellschaft, Faradayweg 4-6, D-14195 Berlin, Germany
ReceiVed: July 31, 1998; In Final Form: October 6, 1998
We investigated the possible role of binary oxides associating Sb and Mo in the synergetic effects between
physically mixed MoO3 and R-Sb2O4 in the selective oxidation of isobutene to methacrolein. The catalytic
behavior of three Sb-Mo-O phasessSb2MoO6 on one hand, a mixture of Sb2Mo10O31 and Sb4Mo10O31 on
the other handswas investigated. The samples were characterized using X-ray diffraction (XRD), confocal
laser Raman microscopy, specific area measurements, and scanning electron microscopy. Additional in situ
high-temperature XRD was also performed. At similar conversions of isobutene, the selectivities to
methacrolein for the Sb-Mo-O phases were always lower than those of mixtures of MoO3 with R-Sb2O4
with the same atomic compositions. A second result was that the three Sb-Mo-O phases were unstable in
the conditions of reaction and tended to decompose to simple molybdenum and antimony oxides. An
improvement of the catalytic performances of the Sb-Mo-O catalysts occurred in parallel to their
decomposition. Some apparent cooperation effects were observed for mechanical mixtures of the Sb-Mo-O
phases with MoO3 or R-Sb2O4. These effects are actually related to the decomposition of the Sb-Mo-O
phases and are due to the cooperation between MoO3 and R-Sb2O4 formed at their surface. The spontaneous
decomposition of the Sb-Mo-O phases shows that these phases cannot form in operandi (during the catalytic
reaction) starting from the simple oxides. Consequently, this shows that they cannot be the real most performant
phases in the mixtures of MoO3 with R-Sb2O4 and hence that they cannot be at the origin of the synergism
between the simple oxides.
1. Introduction
contribute to the synergetic effects observed.1,2 The main
objective of the experiments presented here was to evaluate the
role that such binary phase, if really formed, could play in the
improvement of the performances observed for the mixtures of
MoO3 with R-Sb2O4.
A conspicuously strong cooperation effect was found in the
selective oxidation of isobutene to methacrolein when using an
MoO3 sample developing preferentially the (010) faces me-
chanically mixed with R-Sb2O4. Synergetic enhancements of
both the isobutene conversion and the selectivity to methacrolein
were observed showing that new selective sites had been created
at the surface of MoO3 in the presence of R-Sb2O4. The
phenomenon was explained by the occurrence of a remote
control mechanism between the two oxides. According to this
mechanism, R-Sb2O4 irrigates the catalyst with spillover oxygen.
This reacts with the surface of MoO3 crystals where it induces
the facetting of the nonselective (010) faces to more active and
more selective (100) steps. As strongly suggested by the
characterization of the catalysts after reaction, the model applied
adequately because the two oxides incorporated in the mixture
remained completely separate during the catalytic reaction and
because there was no evidence of mutual contamination.
However, when comparing carefully the XRD patterns of the
mixture before and after catalysis, some faint indications led
us to wonder whether a phase containing Sb, Mo, and O
simultaneously, in particular Sb2Mo10O31, could have been
formed in a small amount during the reaction and hence could
Weng and Delmon listed the different possible origins of the
synergetic effects between physically mixed simple phases.3
Among these, the formation of binary phases during the reaction
has been frequently cited. On one hand, a new phase could
present performances higher than those of the starting simple
oxides in the mixture. On the other hand, even if formed in
small amount, the new phase could cover the surface of the
starting material and hence could extensively influence the
overall activity of the catalyst. Enhancement of the perfor-
mances would thus be obtained. The reaction of Bi2O3 with
MoO3 to form Bi-Mo-O binary compounds constitutes a
classical example of the formation of a more active and more
selective new phase.4 In principle, this line of interpretation
could be extended to other types of mutual contaminations, like
solid solutions and surface contamination layers formed from
the simple oxides.3 In addition to this, and even if it is not
directly evoked in the review of Weng and Delmon,3 one could
also imagine another situation where a binary phase could play
a beneficial role on the performances of the catalyst by a
mechanism other than by exerting its own activity. Starting
from a mixture of simple oxides, this would be the formation
of a possibly inactive and/or nonselective binary phase during
the catalytic reaction, which could thus not be directly respon-
* Corresponding author. Fax: +32 10/47 36 49. E-mail: gaigneaux@cata.
ucl.ac.be.
† Universite´ catholique de Louvain.
‡ Fritz-Haber-Institut der Max-Planck-Gesellschaft.
§ “Charge´ de recherches” research fellow for the Fonds National de la
Recherche Scientifique (FNRS) of Belgium.
10.1021/jp983232h CCC: $15.00 © 1998 American Chemical Society
Published on Web 12/02/1998

Catalytic Behavior of Three Sb-Mo-O Phases
J. Phys. Chem. B, Vol. 102, No. 51, 1998 10543
sible for the enhanced performances of the mixture but which
could cooperate with one (or more) of the remaining nontrans-
formed initial simple oxides. In this case, the improvement of
the performances would be due to this cooperation. This
situation is mechanistically equivalent to that where synergisms
were detected for mixtures containing a simple oxide and an
oxide phase containing several metals.5-8 The only difference
is that the binary phase would form in operandi rather than
being synthesized separately and mixed with the simple oxides
afterward.
conditions with those of mixtures of simple oxides with
comparable compositions. (ii) Evaluation of the extent to which
a cooperation between the Sb-Mo-O phases and simple MoO3
or R-Sb2O4 could take place. The approach followed was that
classically used for the detection of the catalytic cooperations
between separate phases, namely, comparing the catalytic
performances of mechanical mixtures of a Sb-Mo-O phase
with MoO3, or with R-Sb2O4, with those of the constituting
phases tested individually. (iii) Evaluation of the stability of
the Sb-Mo-O phases in the conditions of reaction. This
consisted in observing the evolution with time on stream of the
performances of the Sb-Mo-O phases and in characterizing
them before and after catalytic tests with different durations
using X-ray diffraction, confocal laser Raman microscopy,
scanning electron microscopy, and specific area measurements.
In situ X-ray diffraction performed in the presence of air in a
high-temperature cell was also carried out for this purpose.
In the case of the mixtures of MoO3 with R-Sb2O4, extensive
investigations have been carried out in order to evaluate the
possibility to form binary oxides or contamination phases (solid
solutions or surface contamination layers) and to measure their
stability under the conditions of catalytic oxidation reactions.
It was shown that surface contamination layers of Sb ions on
MoO3 or of Mo ions on R-Sb2O4 did not form in the catalytic
conditions. On the contrary, when prepared artificially, these
layers were unstable and their transformation led to a decon-
tamination under the reaction conditions, namely the breaking
of the layer and its sintering to crystallites.3,9-13 The situation
was similar when considering the formation of bulk mixed
phases containing Sb, Mo, and O simultaneously. Except for
the work mentioned before using a mixture containing (010)-
oriented MoO3,1 the formation of a Sb-Mo-O phase was never
detected in the used mixtures of MoO3 and R-Sb2O4. Attempts
to synthesize a binary phase by coprecipitation of molybdenum
and antimony salts or by solid-state reactions of MoO3 and
R-Sb2O4 were unsuccessful.9,14,15 But the evaluation of the
stability of these phases in the catalytic conditions was never
undertaken experimentally. Weng et al. considered that it was
almost certain that they would decompose during catalytic
selective oxidation.3,9 Their conclusion was based on the work
of Parmentier et al.16-19 These authors reported (i) that the
preparation of the Sb-Mo-O phases requires conditions that
are far from those of the catalytic activity measurements (for
Sb2MoO6, calcination at 500 °C in reduced pressure of argon;
for Sb2Mo10O31 or Sb4Mo10O31, calcination around 500 °C in
nitrogen with a small amount of hydrogen in the presence of
high concentrations of water) and (ii) that the Sb-Mo-O phases
decompose spontaneously in air between 350 and 400 °C. The
criticism against the argumentation of Weng et al. would be
that the gas mixture used for the catalytic reaction, even if
presenting a composition closer to air than to hydrogen diluted
in nitrogen, in particular presenting an almost identical con-
centration in oxygen, also contained 10% of isobutene. It is
very difficult to evaluate the extent to which isobutene would
maintain the atmosphere in the reactor reductive enough to allow
the formation of a Sb-Mo-O phase. The main difficulty in
this evaluation is that the reductive potential is actually
determined by the dynamics of the reduction-oxidation cycle
to which the surface of the solid is subjected, and not by the
overall static value obtained through a traditional thermodynamic
evaluation. The assumption of Weng et al. that Sb-Mo-O
phases could not form must therefore be reexamined, and this,
experimentally.
Although only Sb2Mo10O31 was suspected to form in the
mixtures of MoO3 with R-Sb2O4, we investigated additionally
Sb2MoO6 and Sb4Mo10O31. Sb2MoO6 is a molybdate with a
monoclinic crystallographic symmetry. In Sb2Mo10O31 and
Sb4Mo10O31, crystallized in orthorombic and hexagonal sym-
metries, respectively, the average experimental oxidation levels
of Mo are 5.44 and 5.06.16-19
The performances of the catalysts reported in what follows,
in particular their selectivities to methacrolein, are low compared
to those of actual industrial catalysts used for the selective
oxidation of isobutene. As an example the formulation proposed
by Blangenois et al. leads to a selectivity to methacrolein higher
than 90%, while those of the catalysts studied here only reach
47% at the maximum.20 One could wonder about the signifi-
cance and the interest, at the industrial point of view, of an
investigation on catalysts such as in this contribution. As an
answer, we have to mention that our objective is not to challenge
the existing industrial catalysts. Commercial catalysts for
selective oxidation reaction of isobutene or similar are often
composed of intimately mixed Sb and Mo. Our objective is to
throw light on the way the combination of these two elements
can bring about improvement of performances, precisely to
clarify whether they do so by being associated in a mixed oxide
or, on the contrary, by remaining in two separate simple oxide
phasesswhich is a very controversial point. Our strategy has
been to study model catalysts with architectures as controlled
as possible to make the interpretation of their performances as
simple as possible: single Sb-Mo-O phases and mechanical
mixtures of simple oxides were selected. This choice was made
at the detriment of possibly higher catalytic performances of
catalysts with less defined structures. Nevertheless, all the
conclusions drawn in the present study can be directly transposed
for a better understanding of the working mechanism of the
real industrial catalysts. This fully justifies the interest to carry
out this investigation.
2. Experimental Section
2.1. Preparation of the Catalysts. 2.1.1. Preparation of
the Sb-Mo-O Phases. Three phases containing Sb, Mo and
O simultaneously have been synthesized by solid-state reactions
following the procedures described by Parmentier et al.16-19
To give a clear view of the role(s) of the Sb-Mo-O phases
in the synergetic effects measured in the selective oxidation of
isobutene to methacrolein for mixtures of MoO3 with R-Sb2O4,
three main experimental lines were followed. (i) Determination
of the extent to which the own reactivity of the Sb-Mo-O
phases could explain the higher performances observed for the
mixtures of MoO3 with R-Sb2O4. This consisted in comparing
the catalytic performances of the binary phases under appropriate
2.1.1.1. Sb2MoO6. Commercial Sb2O3 (14.475 g, Janssen
Chimica, 99+%) was ground in an agate mortar with com-
mercial MoO3 (7.197 g, BDH Chemicals, 99.5+%) in a 1:1
molar ratio. To maintain the stoichiometry of the system
unchanged, namely, to avoid any O supply from the atmosphere

10544 J. Phys. Chem. B, Vol. 102, No. 51, 1998
Gaigneaux et al.
Figure 1. XRD pattern (λ ) 1.5418 Å, step size ) 0.04°, step time )
6 s) of Sb2MoO6; the asterisks * indicate the positions of the standard
reflection lines of this phase (file 33-1491 JCPDS21).
Figure 2. XRD pattern (λ ) 1.5418 Å, step size ) 0.04°, step time )
6 s) of “ SbmMo10O31”; the positions of the standard reflection lines of
Sb2Mo10O31 (JCPDS file 33-0105) and of Sb4Mo10O31 (JCPDS file 33-
0104) are indicated with the asterisks * and circles O, respectively.21
and to avoid any O and Mo loss due to MoO3 sublimation during
the synthesis, the mixture was evacuated in a glass tube down
to 10-4 mbar. The tube was then filled with 670 mbar of Ar,
sealed, and thereafter maintained at 500 °C for 24 h.
TABLE 1: Compositions (in wt %) of the Mechanical
Mixtures Prepared
mixture
phase 1
phase 2
proportions (wt %)
1
2
3
4
Sb2MoO6
Sb2MoO6
SbmMo10O31
SbmMo10O31
MoO3
R-Sb2O4
MoO3
50/50
50/50
50/50
50/50
The X-ray diffraction (XRD) pattern of the resulting orange
powder presented all the reflection lines with relative intensities
as listed in the Sb2MoO6 standard 33-1491 JCPDS file.21 No
other line was present. In particular, no line corresponding (i)
to nonreacted MoO3, or any other molybdenum oxide or
suboxide possibly formed during the synthesis, (ii) to nonreacted
Sb2O3, or (iii) to any other Sb, Mo and O-containing phases
listed in the JCPDS database, was detected.21 Figure 1 presents
the XRD pattern of the Sb2MoO6 sample. The specific area
R-Sb2O4
5
6
7
MoO3
MoO3
MoO3
R-Sb2O4
R-Sb2O4
R-Sb2O4
25/75
50/50
75/25
using the sample synthesized as described before, hereafter noted
“SbmMo10O31”. Figure 2 shows the XRD pattern of SbmMo10O31.
SbmMo10O31 developed an 0.17 m2 g-1 specific area.
developed by Sb2MoO6 was 0.38 m2 g-1
.
2.1.1.2. Sb2Mo10O31 and Sb4Mo10O31. Sb2Mo10O31 was
synthesized starting from a mixture of Sb2MoO6 (3.774 g,
prepared as described before) and MoO3 (11.227 g, BDH
Chemicals, 99.5+%) with a 1:9 molar ratio. The mixture was
heated at 10 °C min-1 to 520 °C in N2 (Air Liquide, 99.8%)
and thereafter maintained at the same temperature under a
flowing (50 mL min-1) H2/N2 mixture (1/99 vol.), saturated
with 570 mmHg of water (corresponding to the water vapor
pressure in equilibrium with liquid water at 92 °C). The system
was afterward cooled in N2. The high proportion of H2O with
respect to H2 in the preparation atmosphere was fixed so that
the global reductive potential of the system was tuned at a very
low value and allowed to trigger the mild reduction of Mo
required for the formation of the binary phase. To complete
the reaction, the obtained solid was finely ground and submitted
to a second calcination cycle following the same procedure.
Purple flakes were obtained. XRD analysis showed that
Sb2Mo10O31 (JCPDS standard file 33-0105) was the main
crystallographic phase formed. Some Sb4Mo10O31 (JCPDS
standard file 33-0104) was also present.21 No other crystalline
Sb-Mo-O mixed phase was detected, but small amounts of
several MoO3-x phases were found. This is logical considering
that Sb got incorporated in the Mo-poor Sb4Mo10O31 phase
rather than in the desired Sb2Mo10O31 Mo-rich one. The Mo
excess resulting from this effect led to the formation of MoO3-x
phases in small amounts.
2.1.2. Preparation of Simple Oxides: Molybdenum Trioxide
and Antimony Oxide. MoO3 was prepared by calcination in
air of ammonium heptamolybdate tetrahydrate (Merck, p.a.) at
500 °C over 20 h; R-Sb2O4 was synthesized in the same
conditions starting from Sb2O3 (UCB Chemicals, p.a.). XRD
analysis showed that the obtained samples were, respectively,
pure molybdite (JCPDS standard file 05-0508) and cervantite
(JCPDS standard file 11-0694) crystallographic phases, respec-
tively.21 The specific areas of MoO3 and R-Sb2O4 were 0.48
and 0.50 m2 g-1, respectively.
2.1.3. Preparation of Mechanical Mixtures. “Mechanical
mixtures”, also called “physical mixtures” in this paper, were
prepared by vigorously mixing a suspension in n-pentane of
adequate amounts of powders synthesized as described above.
Ultra-Turrax agitation was applied for 10 min at 6000 rpm.
Afterward, the interdispersion of the phases was completed
ultrasonically over 10 min. The n-pentane was then removed
under vacuum at room temperature before the remaining solid
was dried gently in air at 80 °C overnight. Table 1 lists all the
mechanical mixtures prepared. Two series can be distinguished:
(i) mixtures of a binary Sb-Mo-O phase with a simple oxide
(mixtures 1 to 4), (ii) mixtures of two simple oxides (mixtures
5 to 7). To compare rigorously the performances of the mixtures
with those of the pure phases, these were submitted to exactly
the same “mixing” procedure before being catalytically tested.
An X-ray diffraction analysis was done before the catalytic
test. The patterns obtained for the pure phases having been
submitted to the mechanical mixture procedure never showed
any difference with those of the corresponding “nonmixed”
samples. Also, the patterns of the mixtures corresponded
perfectly to the theoretical patterns reconstituted by the simple
ponderated addition of those measured for the pure constituting
phases.
Attempts to prepare Sb2Mo10O31 or, alternatively, Sb4Mo10O31
(following the same procedure as for Sb2Mo10O31, but starting
from a mixture of Sb2MoO6 with MoO3 in a 1:4 molar ratio
and fixing the calcination temperature at 500 °C and the water
vapor pressure at 532 mmHg) with higher purities were
unsuccessful. The investigation of the catalytic behavior of
Sb2Mo10O31 and Sb4Mo10O31 was therefore made simultaneously

Catalytic Behavior of Three Sb-Mo-O Phases
J. Phys. Chem. B, Vol. 102, No. 51, 1998 10545
2.2. Catalytic Activity Measurements. In the absence of
special indication (about reaction temperature, mass of catalyst
tested, or duration of reaction) in the following, catalytic tests
were carried out in what is described hereafter as the “classical
conditions”. The isobutene oxidation was performed in a fixed
bed reactor at 420 °C. The composition of the gas mixture was
isobutene/O2/He ) 1/2/7 vol. The total gas flow was 30 mL
min-1. The mass of catalyst tested was 500 mg. The samples
were used as powders without pressing them into pellets and
were diluted in an equal volume of small glass balls (0.49 mm
< diameter < 0.72 mm) previously checked to be inactive. The
unreacted isobutene and the methacrolein produced were
analyzed by on-line gas chromatography. For all the catalytic
tests shown after, CO2sa derivative of isobutene and/or of
methacrolein through a total oxidation processswas the only
product detected in addition to methacrolein: no other partially
oxygenated derivatives of isobutene or of methacrolein (alcohols,
aldehydes, ketones, or carboxylic acids) nor other deep or total
oxidation products (e.g., formic acid or CO) were dectected.
After 3 h of reaction, the catalysts were cooled to room
temperature at a rate of 7.5 °C min-1 in the same flow as during
the reaction. Catalytic performances were expressed in terms
of isobutene conversion (%C) and yield and selectivity to
methacrolein (%Y(meth) and %S(meth), respectively). For
reasons of clarity, and because the investigation of the mech-
anism of its formation was not the main point of our discussion,
we did not report the yield and selectivity to CO2. However as
being the only other product of reaction than methacrolein, the
selectivity to CO2 is the complement to %S(meth) to reach a
100% selectivity. Following the basic recommendations dealing
with the kinetics of consecutive reactions, the comparison of
selectivities to methacrolein on the different catalysts was done
at comparable low conversion of isobutene. The comparison
of the conversion of isobutene and yield to methacrolein was
made after normalizing the measured values to the surface
(expressed in m2) developed by the mass of catalyst used for
the tests. In the following, we call these normalized values
“intrinsic conversion and yield”. When necessary, an “intrinsic
selectivity” was calculated by reporting the intrinsic yield to
the intrinsic conversion.
by catalytic activity measurements at 420 °Csfor these experi-
ments, the performances at 420 °C correspond to the initial
values measured immediately after the reactor stabilized at this
temperature.
The performances of the mixtures of the binary Sb-Mo-O
phases with the simple oxides were also measured under the
classical conditions in order to detect possible synergetic effects.
For that purpose, the performances measured were compared
with theoretical values calculated on the basis of the perfor-
mances obtained when the phases constituting the mixture were
tested individually and assuming that the phases in the mixture
behaved independently without any cooperation (eqs 1). The
performances reported for this section correspond to the values
arithmetically averaged on the totality of the 3 h of test.
theoretical conversion of isobutene:
(1a)
%Cth ) [%C obtained with A ×
mass fraction of A in the mixture] +
[%C obtained with B × mass fraction of B in the mixture]
theoretical yield in methacrolein:
(1b)
%Yth(meth) ) [%Y(meth) obtained with A ×
mass fraction of A in the mixture] +
[%Y(meth) obtained with B ×
mass fraction of B in the mixture]
(1c)
theoretical selectivity to methacrolein:
%Sth(meth) ) Yth(meth)/%Cth × 100
2.3. Characterization. The catalysts were characterized
before and after the catalytic tests by X-ray diffractometry
(XRD) on a Kristalloflex Siemens D-5000 equipment. Analyses
were carried out in the continuous coupled θ/2θ scan reflection
mode using the KR1,2 radiation of Cu (λ ) 1.5418 Å) for 2θ
angles varying from 10 to 80°. The fluorescence contribution
was eliminated from the diffracted beam using a curved graphite
monochromator. The scan rate was 0.4 deg min-1 with a step
size of 0.04° and a step time of 6 s.
Attention was directed to the comparison of the catalytic
performances of the Sb-Mo-O phases with those of mixtures
of MoO3 with R-Sb2O4. Potentially, Sb2MoO6 contacted with
O2 could decompose to a mixture of MoO3 and R-Sb2O4 with
a composition close to the mechanical mixture containing 25
wt % of MoO3 and 75 wt % of R-Sb2O4. This was selected for
the comparison with Sb2MoO6. Similarly, the performances of
the SbmMo10O31 sample were compared with those of the
mixture containing 75 wt % of MoO3 and 25 wt % of R-Sb2O4.
To reach conversions of isobutene similar to those on the binary
phases, the masses of mixtures of MoO3 with R-Sb2O4 tested
were adapted from 500 mg (corresponding to the “classical
conditions”) to 400 mg. For this section, the performances
reported correspond to the values measured immediately when
the reactor stabilized at 420 °Cswe called them “initial
performances”.
Additional XRD measurements were performed at high
temperature in an in situ XRD cell interfaced with an HTK 10
High Temperature Attachment. Analyses were done in air with
the same parameters as described above. Two analysis se-
quences were realized focusing (i) on the evolution of the XRD
pattern as a function of the temperature and (ii) on the evolution
of the XRD pattern at 420 °C as a function of time, respectively.
For cycle i, XRD patterns were measured consecutively at 30,
200, 300, 400, 500, and 600 °C and again at 30 °C. The
temperature set for the analysis was reached with a rate of
1 °C s-1. The sample was maintained at the analysis temper-
ature during 1 h before the recording of the pattern started. The
total time of analysis for cycle i was 22.5 h. Cycle ii consisted
of a first measurement of the XRD pattern of the sample at 30
°C followed by seven consecutive measurements at 420 °C and
an additional analysis at 30 °C. No time delay was introduced
in the sequence between the moment the analysis temperature
was reached and the beginning of the pattern measurement. The
total time of analysis for cycle ii was 19.8 h. The two cycles
were applied to Sb2MoO6, to SbmMo10O31 and to their mechan-
ical mixtures with R-Sb2O4.
Another objective was to explore the stability of the Sb-
Mo-O phases during the catalytic reaction. Attention was
particularly directed to the evolution of their catalytic perfor-
mances as a function of the time on stream. In addition to the
tests performed in the “classical conditions”, the pure Sb-
Mo-O phases were tested at 450 °C over 14 h (for Sb2MoO6
and SbmMo10O31) and over 8 days (for SbmMo10O31 only). In
both cases, the reaction at 450 °C was preceded and followed
For both types of XRD analyses, only the results obtained
with the pure Sb, Mo, and O-containing phases are presented
and discussed hereafter. The interpretation of the results with

10546 J. Phys. Chem. B, Vol. 102, No. 51, 1998
Gaigneaux et al.
TABLE 3: Comparison of the Intrinsic Conversion of
Isobutene and Intrinsic Yield to Methacrolein at 420 °C on
the Sb-Mo-O Phases (Initial Performances) and the
Corresponding Mixtures of MoO3 with r-Sb2O4
TABLE 2: Comparison, at Comparable Small Conversion of
Isobutene, of the Selectivity to Methacrolein at 420 °C on
the Sb-Mo-O Phases (Initial Performances) and the
Corresponding Mixtures of MoO3 with r-Sb2O4
mass tested,
mg
%Y
%S
catalyst
%C m-2
%Y(meth) m-2
catalyst
%C (meth) (meth)
Sb2MoO6
22.4
36.2
49.4
30.5
0.0
16.4
6.7
MoO3 + R-Sb2O4 25/75 (wt %)
Sb2MoO6
500
400
500
400
4.3
7.2
4.2
5.9
0.0
3.3
0.6
1.5
0.0
45.4
13.7
25.2
SbmMo10O31
MoO3 + R-Sb2O4 25/75 (wt %)
MoO3 + R-Sb2O4 75/25 (wt %)
7.7
SbmMo10O31
MoO3 + R-Sb2O4 75/25 (wt %)
the mixtures was impossible because of the simultaneous
presence of several phases in the samples with most of the peaks
overlapping, and because of the much higher crystallinities of
the simple oxides compared to those of the Sb-Mo-O phases
(whose peaks were masked by the intense reflections of the
simple oxides).
Confocal laser Raman microscopy was performed with a
Dilor Labram I spectrometer equipped with a computerized
X-Y translation stage. The analysis was made on the Sb-
Mo-O phases before and after tests over 14 h at 450 °C. The
Raman spectra were excited using a He laser (632.8 nm, Melles
Griot) operated at 15 mW power. The spectrometer resolution
was 2 cm-1. A notch filter was applied to cut off the laser line
and the Rayleigh scattering up to 100 cm-1. The spectrometer
was equipped with a CCD detector (1024 × 298 diodes), which
was Peltier-cooled to 240 K in order to reduce thermal noise.
Though the spectrometer function and the characteristics of the
diode array are unknown, it was assumed that the recorded
spectra (band positions and band profiles) really represented
the sample. For each sample, Raman spectra were recorded
using the scanning multichannel technique (SMT) at about 750
different spots (i.e., 30 spots × 25 spots) regularly dispersed in
a sample area of 40 × 30 µm2. The spatial resolution of each
spot was approximately 0.7 µm. This technique gives a
complete picture of the composition of the samples. The spectra
shown hereafter have been checked not to be perturbed by the
occurrence of a thermal degradation of the samples under the
laser beam during the analysis. In the discussion of the LRS
spectra, the objective was not to assign exhaustively all the bands
of the spectra, but we mainly focused on the interpretation of
the differences between the spectra measured for the corre-
sponding samples before and after the catalytic reaction.
Specific area measurements SBET were performed on the
Sb-Mo-O phases after the catalytic tests during 14 h at
450 °C using a Micromeritics ASAP 2000 device. The analysis
was based on the adsorption of Kr at the liquid nitrogen
temperature.
Figure 3. Evolution with time of the conversion of isobutene at
420 °C on Sb2MoO6 and SbmMo10O31.
Figure 4. Evolution with time of the selectivity to methacrolein at
420 °C on Sb2MoO6 and SbmMo10O31.
Scanning electron microscopy (SEM) was performed on a
Hitachi S-570 microscope using a 15 kV accelerating voltage.
We checked the occurrence of the presented features in many
different places so that all the SEM micrographs shown hereafter
are representative of the whole samples. Analysis was made
on Sb2MoO6, fresh and after test at 450 °C, and on SbmMo10O31,
fresh, after test at 420 °C, after test at 450 °C over 14 h, and
after test at 450 °C over 8 days.
Figure 5. Evolution with time of the conversion of isobutene at
450 °C on Sb2MoO6 and SbmMo10O31.
series of catalysts. On one hand, Sb2MoO6 presented a lower
intrinsic conversion of isobutene than the corresponding mixture
of MoO3 with R-Sb2O4, but SbmMo10O31 was intrinsically more
active than its comparable mixture of simple oxides (Table 3).
On the other hand, both Sb-Mo-O phases exhibited lower
selectivities to methacrolein than the mixtures of MoO3 with
R-Sb2O4 (Table 2).
3.1.2. EVolution of the Catalytic Performances of Sb-Mo-O
Phases with the Time on Stream. Figures 3 and 4 show the
evolution with time on stream of the isobutene conversion and
selectivity to methacrolein at 420 °C for Sb2MoO6 and
SbmMo10O31. Figure 5 gives the same information for the
conversion in the tests performed at 450 °C over 14 h. Table
4 compares the selectivity to methacrolein obtained at 420 °C
3. Results
3.1. Catalytic Activity Measurements. 3.1.1. ActiVity at
420 °C of the Sb-Mo-O PhasessComparison with the
Mechanical Mixtures of MoO3 with R-Sb2O4. Table 2 compares
the initial selectivity to methacrolein at similar low conversion
of isobutene obtained at 420 °C on the binary Sb-Mo-O phases
and on the mixtures of MoO3 with R-Sb2O4 with comparable
atomic compositions. Table 3 compares the intrinsic conversion
of isobutene and the intrinsic yield to methacrolein for the two

Catalytic Behavior of Three Sb-Mo-O Phases
J. Phys. Chem. B, Vol. 102, No. 51, 1998 10547
TABLE 4: Comparison, at Comparable Small Conversion of
Isobutene, of the Selectivity to Methacrolein at 420 °C on
the Sb-Mo-O Phases Before and After Catalytic Reaction
at 450 °C for 14 h
TABLE 7: Observed and Theoretical Intrinsic Conversion
of Isobutene, Intrinsic Yield, and “Intrinsic” Selectivity
(Calculated by Reporting Intrinsic Yield to Intrinsic
Conversion) to Methacrolein on the Mixtures of Sb-Mo-O
Phases with Simple Oxides at 420 °Ca
catalyst
%C
%Y(meth)
%S(meth)
%C
%Y(meth) %S(meth)
m-2
“intr.”
Sb2MoO6
before 450 °C
after 450 °C
before 450 °C
after 450 °C
3.3
5.6
3.7
3.2
0.6
1.4
0.6
0.7
17.1
24.4
17.7
22.0
catalyst
m-2
Sb2MoO6
SbmMo10O31
MoO3
24.4
3.0 12.5
SbmMo10O31
54.1
21.9
9.4
8
1.5
0.0
14.7
6.7
R-Sb2O4
0.0
TABLE 5: Comparison of Intrinsic Conversion of Isobutene
and Intrinsic Yield to Methacrolein at 420 °C on the
Sb-Mo-O Phases Before and After Catalytic Reaction at
450 °C 14 h
Sb2MoO6 + MoO3 50/50 (wt %)
24.7 (23.1) 5.4 (2.3) 22.1 (9.7)
Sb2MoO6 + R-Sb2O4 50/50 (wt %) 18.4 (16.9) 4.2 (1.5) 22.7 (9.0)
SbmMo10O31 + MoO3 50/50 (wt %) 25.6 (38.0) 3.6 (4.7) 14.2 (12.4)
SbmMo10O31 + R-Sb2O4 50/
50 (wt %)
26.9 (31.8) 3.9 (4.0) 14.7 (12.6)
catalyst
%C m-2
%Y(meth) m-2
Sb2MoO6
before 450 °C
after 450 °C
before 450 °C
after 450 °C
17.3
41.2
43.1
18.5
3.0
10.1
7.6
MoO3 + R-Sb2O4 50/50 (wt %)
26.5 (15.7) 12.4 (0.7) 46.7 (4.7)
a The theoretical values (in parentheses) were calculated assuming
the absence of cooperation on the basis of the mass-ponderated
conversions and yields to methacrolein of the pure phases tested alone
(see eq 1).
SbmMo10O31
4.1
TABLE 6: Observed and Theoretical Conversion of
Isobutene, Yield, and Selectivity to Methacrolein on the
Mixtures of Sb-Mo-O Phases with Simple Oxides at
420 °Ca
theoretical values that should have been measured if the phases
in the mixtures had behaved independently without any coop-
eration. Table 7 displays the same information considering the
intrinsic performances of the catalysts. As a reference, the
activity of a mixture of MoO3 and R-Sb2O4 50/50 (wt %) is
mentioned. Both Tables 6 and 7 display the arithmetically
averaged performances of the complete 3 h test.
Cooperation effects were detected when Sb2MoO6 was mixed
both with MoO3 and with R-Sb2O4. Only small enhancements
of the intrinsic conversion of isobutene were observed for the
mixtures (Table 7), but important improvements of the selectiv-
ity to methacrolein were obtained (Table 6). The synergetic
effect was slightly higher when Sb2MoO6 was mixed with
R-Sb2O4 than with MoO3.
catalyst
%C
%Y(meth) %S(meth)
Sb2MoO6
SbmMo10O31
MoO3
4.6
0.6
0.7
0.4
0.0
12.5
14.7
6.7
4.6
5.3
2.4
R-Sb2O4
0.0
Sb2MoO6 + MoO3 50/50 (wt %)
Sb2MoO6 + R-Sb2O4 50/50 (wt %)
SbmMo10O31 + MoO3 50/50 (wt %)
5.3 (4.9) 1.2 (0.5) 22.1 (9.4)
4.1 (3.5) 0.9 (0.3) 22.7 (8.3)
4.2 (4.9) 0.6 (0.5) 14.2 (10.4)
SbmMo10O31 + R-Sb2O4 50/50 (wt %) 4.5 (3.5) 0.7 (0.3) 14.7 (9.8)
MoO3 + R-Sb2O4 50/50 (wt %) 6.5 (3.8) 3.0 (0.2) 46.7 (4.6)
a The theoretical values (in parentheses) were calculated, assuming
the absence of cooperation, on the basis of the mass-ponderated
conversions and yields to methacrolein of the pure phases tested alone
(see eq 1).
Both mixtures of SbmMo10O31 with MoO3 and with R-Sb2O4
exhibited only synergetic improvement of the selectivity to
methacrolein (Table 6). The cooperation was more pronounced
when using R-Sb2O4 in the mixture than when using MoO3.
The magnitude of the cooperative effects with SbmMo10O31
nevertheless remained weaker than that observed with Sb2MoO6
and MoO3 or R-Sb2O4.
The most pronounced synergy observed in this series of
catalytic tests, considering both intrinsic activity and selectivity
to methacrolein, was measured for the mixture of MoO3 with
R-Sb2O4.
3.2. Characterization. 3.2.1. ConVentional X-ray Diffrac-
tion. When comparing the XRD pattern obtained for the fresh
Sb2MoO6 and those obtained after 3 h of catalytic test at
420 °C and after 14 h at 450 °C, the intensities of the reflection
lines did not present significant differences. But five small
peaks that could not be assigned to Sb2MoO6 appeared on the
patterns obtained after tests with intensities increasing with
reaction time. These peaks appeared at 21.8, 23.8, 30.3, 33.7,
and 47.7°.
Attempts to assign the five new peaks to combinations of
another Sb-Mo-O phase with an antimony oxide were
unsuccessful. But the assignment of all the peaks was possible
using a combination of R-Sb2O4 (JCPDS file 11-0694) with
Mo9O26 (JCPDS file 05-0441). Combinations of other simple
oxides were not fitting so closely. The conclusion is that a part
on the binary phases before and after having worked at 450 °C
for 14 h. Table 5 compares the intrinsic conversion of isobutene
and intrinsic yield to methacrolein on the same samples.
At 420 °C, despite the fluctuation of the performances likely
originating from the difficulty to measure low conversions,
Sb2MoO6 presented an overall slight increase of the conversion
with the time of reaction with a simultaneous clearly marked
increase of the selectivity to methacrolein. SbmMo10O31 ex-
hibited a more pronounced increase of conversion than Sb2MoO6
with a slight increase of the selectivity to methacrolein.
At 450 °C, the conversion of isobutene measured for
Sb2MoO6 increased continuously throughout the 14 h of reaction
(Figure 5). The selectivity to methacrolein remained constant
at 32%. When comparing the performances at 420 °C, it turned
out that the selectivity to methacrolein (Table 4) and the intrinsic
conversion (Table 5) were enhanced after that Sb2MoO6 was
reacted at 450 °C. SbmMo10O31 presented a constant conversion
at 450 °C both during the 14 h experiment (Figure 5) and the
8 days experiment. For both tests, the selectivity to methacrolein
stabilized at 23.7%. Comparing the performances of this oxide
before and after the reaction at 450 °C, an important decrease
of the intrinsic conversion was observed (Table 5), but the
selectivity to methacrolein increased significantly (Table 4).
3.1.3. Synergetic Effects between Sb-Mo-O Phases and
Simple Oxides. Table 6 presents the catalytic performances
measured at 420 °C for the mechanical mixtures of Sb-Mo-O
phases with simple oxides (MoO3 or R-Sb2O4). The perfor-
mances obtained with the pure phases used alone in the reaction
are also shown. The values in parentheses correspond to the
of Sb and Mo from Sb2MoO6 segregates and recrystallizes
21
during the catalytic reaction to R-Sb2O4 and Mo9O26
.
Figure
6 shows the evolution of the XRD patterns of Sb2MoO6 with
the time of catalytic reaction.

10548 J. Phys. Chem. B, Vol. 102, No. 51, 1998
Gaigneaux et al.
Figure 6. XRD patterns of (a) fresh Sb2MoO6, (b) Sb2MoO6 after 3 h
of catalytic test at 420 °C and (c) after 14 h of catalytic test at 450 °C.
The arrows pointing up indicate the five peaks appearing after catalytic
tests: * for Mo9O26, O for R-Sb2O4.
Figure 7. XRD patterns of (a) fresh SbmMo10O31, (b) SbmMo10O31 after
3 h of catalytic test at 420 °C, (c) after 14 h, and (d) after 8 days of
catalytic test at 450 °C. The arrows # pointing down indicate the
disappearing peaks of Sb4Mo10O31, the arrows O pointing up indicate
the appearing peaks of R-Sb2O4.
Considering the other Sb-Mo-O oxides, all the peaks
corresponding to Sb4Mo10O31 got progressively extinct with
increasing time of reaction. These were the peaks at 13.9, 18.4,
26.0, 28.1, 36.1, 36.9, 42.8, 49.7, 53.4, and 53.9°. The peaks
corresponding to Sb2Mo10O31 remained unchanged in position,
but the sample exhibited a tendency to amorphization. Two
new peaks appeared at 51.5 and 54.2°. These correspond to
R-Sb2O4 (JCPDS file 11-0694).21 No other peak corresponding
to another Sb-Mo-O phase or to an MoO3-x phase was
detected. The conclusion is that Sb4Mo10O31 recrystallizes to
R-Sb2O4 and very likely to Sb2Mo10O31, which seems to be
relatively stable. Figure 7 shows the evolution of the XRD
patterns of the SbmMo10O31 sample as a function of the duration
of the catalytic test.
3.2.2. In situ X-ray Diffraction. When Sb2MoO6 was
maintained in air at 420 °C, all the reflection lines corresponding
to this phase progressively vanished. They were completely
extinct after the measurement of three patterns (about 6 h). In
the same way, four new peaks progressively appeared at 25.6,
27.2, 28.8, and 33.7°, showing that MoO3 and R-Sb2O4
crystallized during the experiment. The same phase segregation
was observed when the experiments were conducted at increas-
ing temperatures. The patterns measured at low temperatures
(200 and 300 °C) did not differ from that of the fresh Sb2MoO6,
but those measured at 500 and 600 °C only presented the peaks
typical of MoO3 and R-Sb2O4. At 400 °C, the pattern was
composed of a combination of Sb2MoO6, MoO3, and R-Sb2O4.
Figure 8 shows the evolution with time of the XRD pattern of
Sb2MoO6 maintained at 420 °C.
Similar results were obtained with SbmMo10O31. When the
sample was maintained at 420 °C, the lines corresponding to
Sb4Mo10O31, mainly 25.8, 28.0, 35.9, and 36.8° completely
vanished after the first scan (about 2 h). Simultaneously,
reflections assigned to MoO3, with some R-Sb2O4 lines, ap-
peared. The lines of Sb2Mo10O31 remained unchanged after the
completion of the whole analysis cycle.
Figure 8. Evolution with time of the XRD pattern of Sb2MoO6
maintained at 420 °C. The arrows pointing up indicate the new peaks
appearing during the analysis: * for MoO3, O for R-Sb2O4. + indicates
the peaks corresponding to the Pt sample holder.
MoO3 with R-Sb2O4 was detected at 400 °C. When reaching
500 °C, the lines corresponding to Sb2Mo10O31 also started to
vanish. They got completely extinct at 600 °C. This concerned
the lines at 21.8, 24.3, 26.1, 26.5, 30.4, 34.3, 34.6, 43.8, 44.6,
and 48.5°. No other lines than those of MoO3 and R-Sb2O4
were detected. Figure 9 shows the evolution of the XRD
patterns of SbmMo10O31 with the temperature of analysis.
The conclusion of this section is that, in air, both Sb-Mo-O
phases tend to decompose to a mixture of MoO3 and R-Sb2O4.
When looking at the evolution of the patterns with the
temperature, the decomposition of Sb4Mo10O31 to a mixture of

Catalytic Behavior of Three Sb-Mo-O Phases
J. Phys. Chem. B, Vol. 102, No. 51, 1998 10549
Figure 9. Evolution with temperature of the XRD patterns of
SbmMo10O31. Arrows pointing down indicate the disappearing peaks,
respectively, ∧ for Sb2Mo10O31 and # for Sb4Mo10O31; arrows pointing
up indicate the appearing peaks of MoO3 * and R-Sb2O4 O; + indicates
point the peaks of the Pt sample holder.
The decomposition starts between 300 and 400 °C for Sb2MoO6
and Sb4Mo10O31, and between 400 and 500 °C for Sb2Mo10O31.
3.2.3. Confocal Laser Raman Microscopy (LRS). Confocal
Raman mapping of a 30 × 40 µm2 area of fresh Sb2MoO6 gave
identical spectra. This confirmed the XRD observation (section
2.1.1.1) that the fresh Sb2MoO6 sample was homogeneous and
contained only one crystalline phase. One representative LRS
spectrum of the fresh Sb2MoO6 is given in Figure 10a. Bands
were measured at 126, 137w, 151s, 157sh, 168s, 180, 205sh,
213s, 221sh, 248, 276vs, 289s, 310sh, 317, 337w, 369vw, 383sh,
393s, 435w, 477, 543vw, 594w, 713s, 746, 813vvs, 845sh, and
937 cm-1
.
After 14 h of catalytic reaction at 450 °C, the surface of
Sb2MoO6 appeared less homogeneous than the fresh material.
Two equally well represented types of spectra were detected
by confocal Raman mapping of the reacted sample. The first
type of spectrum (Figure 10b, spectrum A) exhibited the pattern
found for the fresh material. Even if presenting all the strongest
bands measured for the fresh sample, these spectra were
generally less intense and had a lower signal-to-noise ratio than
those from fresh Sb2MoO6. Thus, the weak bands at 137w,
157sh, 205sh, 310sh, 369vw, 383sh, 435w, and 543vw could
not be resolved from the noise. The other type of Raman
spectrum (Figure 10b, spectrum B) had low intensity and low
signal-to-noise ratio. Only the most intense peaks of the fresh
material were observed as broad bands at 210, 273, 389, 716,
and 812 cm-1. In addition to these, some bands that were not
observed for the fresh Sb2MoO6 were measured at 191, 234,
Figure 10. LRS spectra representative of the Sb2MoO6 fresh (a) and
after catalytic reaction at 450 °C over 14 h (b, the “S’s” indicate the
position of the bands assigned to R-Sb2O4 and the “M’s” those of the
bands assigned to MoO3).
263, 457, 495sh, 660, and 880 cm-1
. According to the
assignment given by Mestl et al., these new bands indicated
the presence of sligthly reduced R-Sb2O4 (bands at 191, 263,
and 457 cm-1) and MoO3-x (bands at 234, 495, 660, and 880
cm-1) species at the surface of the sample.22-24 Figure 10b
shows the two types of Raman spectra recorded for the reacted
Sb2MoO6. Our conclusion is that Sb2MoO6 decomposes during
catalysis at least partially to a mixture of MoO3-x and R-Sb2O4.
Three types of Raman spectra were recorded for the fresh
SbmMo10O31 sample according to the 750 spots analyzed. The
first type of spectrum (Figure 11a, spectrum A) represented
about 90% of the area measured in confocal Raman mapping
and was therefore assigned to the Sb2Mo10O31 fraction of the
sample. These spectra presented bands at 132vw, 140w, 170vw,

10550 J. Phys. Chem. B, Vol. 102, No. 51, 1998
Gaigneaux et al.
TABLE 8: Specific Area Measurements SBET for the
Sb-Mo-O Phases before and after Catalytic Reaction at
450 °C for 14 h
SBET before
SBET after reaction
(m2 g-1)
catalyst
reaction (m2 g-1)
Sb2MoO6
SbmMo10O31
0.38
0.17
0.27
0.35
and 741 cm-1. This perfectly fits with the pattern assigned to
MoO2 by Spevack and McIntyre.25 Sb2MoO6 and MoO2 were
not detected in SbmMo10O31 by XRD. However, we can
reasonably assume that traces of Sb2MoO6 might remain
in SbmMo10O31. As already proposed in section 2.1.1.2 for
the presence of crystalline MoO3-x phases, the presence of
MoO2 could originate from the formation of binary phases,
like Sb4Mo10O31 and Sb2MoO6, containing more Sb than
Sb2Mo10O31, with consequently the simultaneous formation
of molybdenum oxide species.
After the catalytic reaction, the Raman mappping of the
reacted SbmMo10O31 did not show any Raman spectra presenting
the typical features of Sb2MoO6 or of MoO2. Raman spectra
exhibiting the bands assigned to Sb2Mo10O31 were still observed.
However, this only concerned about 50% of the area measured.
The Sb2Mo10O31 spectra exhibited a lower signal-to-noise ratio
than the unreacted material with the consequence that the
weakest bands of the spectrum were not or less resolved (Figure
11b, spectrum A). In addition, about 50% of the measured area
exhibited a Raman pattern that was not observed for the fresh
SbmMo10O31 sample. This presented bands at 112w, 125w,
153s, 195, 214w, 240, 280s, 287sh, 334, 362sh, 376, 466w,
504vw, 665, 817vs, 878w, and 995s cm-1 with high intensities
and a high signal-to-noise ratio (Figure 11b, spectrum B). These
features, in particular the strong band at 995 cm-1 (ModO
stretch involving Mo atoms in an octahedral coordination), are
very typical of MoO3. However, the presence of a small signal
at 878 cm-1 (ModO stretch involving Mo atoms in a more
tetrahedral coordination) indicated the possibility that a fraction
of the MoO3 formed after reaction was very slightly reduced to
a MoO3-x species.22 As the Raman spectra exhibiting the bands
of MoO3-x represent 50% of the area measured on the reacted
sample, it is precluded that this originates exclusively from the
reoxidation of the MoO2 and from the decomposition of the
Sb2MoO6 fractions of the fresh material: as each of these phases
represented only 5% of the area measured on the fresh sample,
then if MoO3-x only originated from the decomposition of
Sb2MoO6 and the reoxidation of MoO2, one would thus obtain
only 10%, at the maximum, of the area measured on the reacted
sample with a Raman spectrum typical of MoO3-x. Even if
we did not detect any Raman spectrum with the features typical
of a simple Sb oxide (likely because being formed in 10 times
smaller amount than MoO3-x), the conclusion is thus that the
formation of simple Mo and Sb oxides mainly results from the
decomposition during catalysis of SbmMo10O31 to simple oxides.
3.2.4. Specific Area Measurement. Table 8 shows the SBET
values measured for the pure Sb-Mo-O phases before and
after the catalytic activity measurement performed at 450 °C
over 14 h. Sb2MoO6 presented a lower SBET value after the
test than before, but for SbmMo10O31 the SBET value measured
after the test was higher than before.
Figure 11. LRS spectra representative of the SbmMo10O31 fresh (a)
and after catalytic reaction at 450 °C over 14 h (b).
183vw, 192, 213, 225, 238, 253vw, 268s, 286, 302, 308sh, 331,
352, 376sh, 393s, 437sh, 466vs, 505w, 551vw, 658w, 697w,
714sh, 782, 816s, 878w, and 884 cm-1. About 5% of the area
analyzed exhibited the LRS pattern with all the bands observed
for Sb2MoO6 (Figure 11a, spectrum B). The third type of
Raman spectrum (Figure 11a, spectrum C) represented about
5% of the measured area on SbmMo10O31. It exhibited very
weak bands at 202, 227, 347, 362, 457, 466, 495, 568, 589,
3.2.5. Scanning Electron Microscopy. Figure 12 shows SEM
views of the fresh Sb2MoO6 sample and of the sample after the
catalytic test at 450 °C. Before being catalytically tested (Figure
12a), pure Sb2MoO6 was composed of small crystallites with
sizes varying between 500 nm and 10 µm and without any well-
defined morphology. After 14 h of test at 450 °C (Figure 12b),

Catalytic Behavior of Three Sb-Mo-O Phases
J. Phys. Chem. B, Vol. 102, No. 51, 1998 10551
Phases. Our objective was to explore the role that binary phases
associating Sb, Mo, and O could play in the synergetic effects
observed with mechanical mixtures of MoO3 and R-Sb2O4. The
first point to investigate was whether such Sb-Mo-O phases
could possess very high catalytic performances, and hence
whether their possible formation during the test could explain
the enhancement of selectivity to methacrolein detected for the
oxidation of isobutene on mixtures of simple oxides. On one
hand, Sb2MoO6 appeared intrinsically less active than the
mixture of MoO3 with R-Sb2O4 with the similar atomic
composition, but SbmMo10O31 was intrinsically more active than
the corresponding mixture. On the other hand, at comparable
low conversions of isobutene, the selectivities to methacrolein
of both Sb-Mo-O phases considered in this work were always
lower than those of the mixtures of MoO3 with R-Sb2O4 with
similar atomic compositions: 0.0% with Sb2MoO6 but 45.4%
with the corresponding mixture and 13.7% with SbmMo10O31
but 25.2% with the corresponding mixture. An even bigger
gap appears when comparing the selectivities of the Sb-Mo-O
phases with that of the mechanical mixture containing the
oriented (010)-MoO3 and R-Sb2O4 where the formation of
Sb2Mo10O31 was suspected to happen in operandi. This mixture
exhibited a 36.7% selectivity to methacrolein obtained at a much
higher conversion of isobutene (57.1%).1 Hence, our conclusion
is that the presence of Sb-Mo-O phases, in case they would
form in operandi starting from the simple oxides mixtures, could
not explain the high selectivities to methacrolein measured for
these mixtures.
4.1.2. Cooperation between Binary Sb-Mo-O Phases and
Simple Oxides. Another possible role of the Sb-Mo-O phases
in the synergism existing between MoO3 and R-Sb2O4 is the
existence of a cooperation between one of the binary phases
and one of the simple oxides. Qualitatively, the results indicate
that such effects might exist. For all the mixtures containing a
binary phase and a simple oxide, catalytic synergisms were
found when considering the selectivity to methacrolein. In
addition, synergetic effects were also observed for the intrinsic
conversion of isobutene on the mixtures containing Sb2MoO6
and MoO3 or R-Sb2O4. But the synergisms detected with these
mixtures were always weaker than those measured when mixing
the simple oxides directly. It is therefore unlikely that the
synergism between the simple oxides is exclusively, or even
significantly, due to the formation of Sb-Mo-O phases during
the catalytic reaction. Nevertheless, one cannot completely
discard this hypothesis without further arguments, mainly
because very few attempts to improve the synthesis procedure
of the Sb-Mo-O phases and to increase their specific areas
had been made in order to maximize the cooperations (as it
was done previously for the mixtures of MoO3 with R-Sb2O4).
Therefore, further discussion of these results cannot be dissoci-
ated from the observation, on one hand, of the evolution with
time on stream of the catalytic performances of the Sb-Mo-O
phase and, on the other hand, of their progressive recrystalli-
zation under the conditions of reaction or under conditions close
to the conditions of reaction.
Figure 12. Scanning electron microscopy views of Sb2MoO6 (a) fresh
(length of the bar ) 15 µm) and (b) after catalytic test at 450 °C for
14 h (length of the bar ) 30 µm).
the sample was mainly composed of large compact aggregates
of small crystallites. The aggregates, with sizes between 15
and 30 µm, exhibited an imperfect prismatic morphology.
Fresh SbmMo10O31 was mostly composed of large crystals
of 20-40 µm exhibiting an octagonal prismatic morphology.
The thickness of the prisms varied in a large extent. The
crystallographic faces were flat and smooth. With increasing
time and temperature of the catalytic activity measurement, a
recrystallization progressively took place on the crystals. For
the sample tested 3 h at 420 °C, most of the crystals presented
fragmented faces. Newly formed small crystallites with ill-
defined morphologies were already distinguishable on some big
crystals. After 14 h at 450 °C, the crystals having undergone
the recrystallization were more numerous. The morphology of
the new crystallites was more defined and corresponded to thin
platelets. After 8 days of catalytic test, the recrystallization had
further progressed as all the starting crystals had recrystallized
and as the size of the new crystallites had increased. A sequence
of four pictures showing the progress of the recrystallization is
shown in Figure 13.
4.1.3. EVolution of the Performances and of the Morphology
of the Binary Sb-Mo-O Phases with Time on Stream. In all
the catalytic tests shown in this work, the performances of the
pure Sb-Mo-O phases improved with the time of reaction.
For both Sb-Mo-O phases, an increase of the conversion of
isobutene took place with the time on stream. In the case of
SbmMo10O31, this improvement of the conversion is very likely
related to its tendency to have its specific area increasing during
the catalytic reaction (Table 8). But the increase of conversion
4. Discussion
4.1. Possible Role of the Sb-Mo-O Phases. 4.1.1.
SelectiVity to Methacrolein of the Pure Binary Sb-Mo-O

10552 J. Phys. Chem. B, Vol. 102, No. 51, 1998
Gaigneaux et al.
Figure 13. Scanning electron microscopy view of the progressive recrystallization of SbmMo10O31 with the duration of catalytic reaction: (a) fresh
(length of the bar ) 30 µm), (b) after 3 h of test at 420 °C (length of the bar ) 5 µm), (c) after 14 h of test at 450 °C (length of the bar ) 6 µm),
and (d) after 8 days of test at 450 °C (length of the bar ) 500 nm).
observed for Sb2MoO6 cannot be due to such an effect because
the specific area was shown to decrease during the reaction.
These conclusions are confirmed when comparing the intrinsic
activities at 420 °C before and after the catalysts were
maintained at 450 °C for a long time. In the case of Sb2MoO6,
an important increase of the intrinsic conversion was measured.
But for SbmMo10O31, an important decrease of the intrinsic
conversion of isobutene was obtained after the reaction at
450 °C (Table 5). The evolution of the selectivity to methac-
rolein with time on stream and the comparison of the selectivities
obtained at 420 °C before and after the reaction at 450 °C are
easier to interpret because these were always measured at
comparable low conversion and were not influenced by the
variation of the specific area of the catalysts. For both binary
phases, an increase of the selectivity to methacrolein with time
on stream was observed. Also, these were higher after that the
catalysts were maintained at 450 °C for 14 h. In both cases,
these results show that, under the conditions of reaction, the
Sb-Mo-O phases progressively transformed to new structures
exhibiting other catalytic performances. In the case of Sb2MoO6,
the new structure was intrinsically more active and more
selective than the fresh binary phase; in the case of SbmMo10O31,
the new structure was intrinsically less active but was more
selective to methacrolein than the fresh material.
These behaviors have to be further interpreted in view of the
other changes of characteristics that the catalysts undergo. LRS
and XRD analyses of the catalysts before and after catalytic
tests as well as the in situ XRD measurements (in the presence
of air in the high-temperature XRD cell) showed that the Sb-
Mo-O compounds were unstable under the conditions of
reactions. The part of the discussion that follows is intended
to confirm that the above effects are due to the decomposition
of the Sb-Mo-O phases to simple oxides.
The LRS analyses indicated that Sb2MoO6 dissociated to
some extent to a mixture of MoO3-x and R-Sb2O4. The XRD
analyses revealed an identical behavior for the bulk of Sb2MoO6
that decomposed, at least partially, to R-Sb2O4 and Mo9O26
crystallites. Similarly, in the in situ XRD analyses, Sb2MoO6
decomposed completely to a mixture of R-Sb2O4 and MoO3.
The phenomenon started between 300 and 400 °C, namely,
below the temperature of the catalytic tests. Sb2MoO6 trans-
formed to different simple oxides in the experiments that
compare the samples before and after catalysis and in the in
situ XRD measurements. This was most likely the consequence

Catalytic Behavior of Three Sb-Mo-O Phases
J. Phys. Chem. B, Vol. 102, No. 51, 1998 10553
of the different amount of oxygen available for the Sb-Mo-O
to undergo the oxidation associated with the decomposition.
When a sufficient amount of oxygen was supplied like in the
in situ experiments, MoO3 was formed. In the catalytic reaction,
the concentration of oxygen in the reaction gas was the same
as in air, but part of it was used for the oxidation of isobutene.
The amount of oxygen available for the reoxidation of the binary
phase being lower, a molybdenum suboxide was formed because
Mo could not fully reoxidize. Both sets of results thus reflect
the instability of Sb2MoO6 in the presence of oxygen and its
tendency to decompose to a mixture of simple oxides.
There is therefore an unambiguous connection between (i)
the aforementioned improvement with the time on stream of
the performances of the Sb-Mo-O phases and (ii) the partial
decomposition of these phases to mixtures of simple MoO3-x
and R-Sb2O4. The formation of MoO3-x and R-Sb2O4 crystals
at the surface of SbmMo10O31 and Sb2MoO6 samples is the
phenomenon responsible for the increases of conversion of
isobutene and of selectivity to methacrolein. The decomposition
of the binary phases also fully explains the difference observed
when comparing the performances of the catalysts before and
after having been reacted at high temperature for a long time
(Tables 4 and 5). Sb2MoO6 is initially intrinsically less active
and less selective than the mixture of MoO3 with R-Sb2O4 with
the same atomic composition (Tables 2 and 3). Hence, after it
has decomposed to simple oxides after a long duration test, it
is logical that the performances of Sb2MoO6 have improved.
The initial intrinsic activity of SbmMo10O31 is higher than that
of the corresponding mixture of MoO3 with R-Sb2O4, but its
selectivity to methacrolein is lower. Hence, it is logical to
observe, after its partial decomposition to simple oxides, a
decrease of the intrinsic activity with simultaneous increase of
the selectivity of the catalyst. These results further confirm that
the synergetic improvement of the selectivity to methacrolein
obtained when using mixtures of MoO3 with R-Sb2O4 cannot
be explained by the formation of Sb-Mo-O phases.
Similar differences between ex situ and in situ XRD experi-
ments were noted for SbmMo10O31. In the conditions of reaction,
Sb4Mo10O31 transformed to R-Sb2O4 and Sb2Mo10O31. The
latter only underwent a slight amorphization. In the in situ
analyses, both phases decomposed to R-Sb2O4 and MoO3. The
transition occurred between 300 and 400 °C for Sb4Mo10O31,
and between 400 and 500 °C for Sb2Mo10O31. Although our
XRD results do not give a direct proof, the detection of a strong
MoO3-x signal in the LRS measurement of the used SbmMo10O31
constitutes an undisputable argument to conclude that even in
the conditions of reaction Sb2Mo10O31 particles decomposed
during catalysis to simple oxides. This behavior might explain
the amorphization detected by XRD for the reacted sample. The
simple oxide crystallites formed were certainly very small with
likely weak crystallinity and were only present in small amount.
This might explain why the conventional ex situ XRD analyses
did not unequivocally detect new reflection lines corresponding
to simple oxides. The sequence of four SEM pictures showing
the progressive recrystallization of SbmMo10O31 was difficult
to interpret and did not allow one to determine the nature of
the small crystallites formed at the surface of the SbmMo10O31
crystals. Their platelet morphology does not correspond to that
classically found for R-Sb2O4 crystals. But, it exactly corre-
sponds to that of MoO3 crystals. Moreover, SEM pictures
obtained after 8 days of reaction showed that the surface of all
crystals was modified. This precluded the possibility that only
the crystals of the Sb4Mo10O31 phase had transformed with, in
this case, the newly formed crystallites observed corresponding
to Sb2Mo10O31. This additional result shows that Sb2Mo10O31
was also undergoing the recrystallization to simple oxides. Our
conclusion is thus that SbmMo10O31 is unstable in the conditions
of reaction and that both Sb2Mo10O31 and Sb4Mo10O31 tend to
decompose and to recrystallize at the surface to an MoO3-x
phase and R-Sb2O4.26
Our observations of the decomposition of the Sb-Mo-O
phases under the conditions of reaction go in parallel with the
conclusions reported by Parmentier et al., namely, that the binary
phases were unstable in the presence of oxygen.16-19 Differ-
ences only appeared when comparing the temperatures at which
the phases were observed to recrystallize. Parmentier et al.
reported the decomposition to simple oxides at 400 °C for
Sb2MoO6, at 300 °C for Sb4Mo10O31, and at 350 °C for
Sb2Mo10O31. We do not have any explanation for these
differences. Parmentier had only reported the transition of the
Sb-Mo-O phases in air. No attempt to evaluate the occurrence
of the phenomenon in slightly oxygen-deficient conditions such
as in catalytic oxidation reactions had been done. The work of
Parmentier was not sufficient to totally prove the instability of
the Sb-Mo-O phases or to predict the occurrence of their
decomposition to simple oxides under the catalytic conditions.
The present investigation leads us to conclude that this
decomposition actually takes place.
4.2. Role of the Cooperation between the Simple Oxides
Formed at the Surface of the Decomposed Sb-Mo-O
Phases. As our results show the formation of R-Sb2O4 and of
MoO3-x, it is necessary to focus further on the role played by
the simple oxide phases formed at the surface of the binary
phases on the improvement of their performances with time on
stream. We base the following discussion on the cooperation
between separate (not mutually contaminated) MoO3 and
R-Sb2O4 as already reported elsewhere.3
4.2.1. Influence on the EVolution with Time on Stream of
the Performances of the Binary Phases. On this base, it is
possible to explain more in detail the extent to which the
improvements of the catalytic performances of the Sb-Mo-O
samples took place as a function of the time on stream. For
Sb2MoO6, as shown by the in situ XRD analyses, the decom-
position to a mixture of MoO3-x and R-Sb2O4 occurred relatively
easily. During the catalytic test at 420 °C, the isobutene
conversion remained low during the first moments of reaction
likely because only small amounts of the simple oxides were
formed. But, on the other hand, the selectivity to methacrolein
reached rapidly its highest value because this already cor-
responded to that obtained with a mixture of MoO3-x and
R-Sb2O4 with a high proportion (67 wt %) of R-Sb2O4, which
is in the range reported to be highly selective.3 For longer times
of reaction, the formation of MoO3-x and R-Sb2O4 progressed
more deeply in the bulk of Sb2MoO6 crystals. This was
evidenced by the SEM picture and by the appearance of XRD
peaks corresponding to the simple oxides. With the number of
simple oxides crystallites formed increasing, the conversion of
isobutene consequently also continuously increased. The
selectivity, however, remained constant because the relative
proportion of the simple oxides formed remained dictated by
the stoichiometry of the starting antimony molybdate, namely,
remained constant, with a high Sb content.
A similar mechanism can be proposed for SbmMo10O31. The
situation is slightly more difficult to interpret because the
decomposition temperature was higher than for Sb2MoO6. The
fact that the evolution of its performances with time was slower
corresponded to the higher stability at higher temperature. One

10554 J. Phys. Chem. B, Vol. 102, No. 51, 1998
Gaigneaux et al.
can reasonably assign the rapid increase of the conversion of
isobutene to the surface decomposition of SbmMo10O31 to simple
oxides. For longer times of reaction, the conversion, however,
did not increase anymore even at a higher temperature. Our
hypothesis is that the recrystallization could not progress deeply
in the bulk. This would be due to the fact that the decomposition
of SbmMo10O31 necessitates much oxygen. The decomposition
of the bulk was thus impossible (or very slow) because of the
likely small flow of oxygen diffusing into the core of the large
initial SbmMo10O31 crystals (or because of the fact that bulk
diffusion in the corresponding structure is very slow). The same
weak oxygen concentration in the bulk probably also existed
when considering Sb2MoO6. However, in this case, its decom-
position was not stopped because it only required modest
amounts of oxygen. The conclusion that the SbmMo10O31
crystals only recrystallized in the surface is consistent with the
SEM observations and could explain why no simple oxides XRD
peaks were detected after the catalytic reaction. Considering
the selectivity to methacrolein obtained with SbmMo10O31, it
only improved in the first moments of the reaction for similar
reasons as for Sb2MoO6. A difference was that the improvement
of selectivity is low. This is logical. On one hand, SbmMo10O31
initially exhibited some selectivity to methacrolein. On the other
hand, the mixture of MoO3-x and R-Sb2O4 formed contained
only a small relative proportion (18 wt %) of R-Sb2O4, and it
is reported that such a composition present lower selectivities
to methacrolein than compositions containing more R-Sb2O4.3
This explains the modest improvements of selectivity observed.
Figure 14. Schematic interpretation of the catalytic behavior of the
Sb-Mo-O phases in the catalytic selective oxidation of isobutene to
methacrolein.
5. Conclusion
The presence of an Sb-Mo-O phase cannot explain the
synergetic effects observed in mechanical mixtures of MoO3
with R-Sb2O4, neither by its own activity nor by the occurrence
of cooperative phenomena. The selectivities to methacrolein
of the phases containing Sb, Mo, and O simultaneously, namely,
Sb2MoO6 on one hand and Sb2Mo10O31 and Sb4Mo10O31 on the
other hand, are lower than those of mechanical mixtures
composed of MoO3 and R-Sb2O4 with comparable atomic
compositions. The binary phases decompose under the condi-
tions of the catalytic tests to a mixture of simple molybdenum
and antimony oxides. The consequence of this decomposition
is that the catalytic performances, in particular the selectivity
to methacrolein, of the Sb-Mo-O catalysts improve with the
formation of MoO3 and R-Sb2O4 ongoing. The apparent
synergetic effects between the Sb-Mo-O phases and the simple
oxides also originate from the decomposition of the former to
simple oxides. The catalytic properties must be attributed to a
cooperation between MoO3 and R-Sb2O4. Our results are in
agreement with the predictions of Weng and Delmon and Bing
et al. concerning the in operandi instability of the Sb-Mo-O
phases.3,9,14,28 There is a substantial difference between the
behavior of the Sb-Mo-O phases in the conditions of reaction
and in experiments roughly mimicking these. The present work,
however, shows that there is a link between the model
experiments in the absence of isobutene and results of real
catalytic experiments.
4.2.2. Influence on the Apparent Cooperation between Binary
Phases and Simple Oxides. In light of the unstability of the
Sb-Mo-O phases, of their natural tendency to decompose to
MoO3-x and R-Sb2O4 in contact with air, and of the existence
of cooperation between these two simple oxides, one can also
explain the occurrence or the absence of synergetic effects in
the mixtures containing Sb-Mo-O phases and a simple oxide.
The mixture that exhibited the less intense synergetic behavior
was that containing SbmMo10O31 and MoO3. Only a weak
synergism was observed for the selectivity to methacrolein
(Table 6). When decomposing during the reaction, SbmMo10O31
formed a mixture of MoO3-x and R-Sb2O4 with a very small
proportion of R-Sb2O4. When taking into account the MoO3
initially present in the mechanical mixture, the resulting overall
proportion of R-Sb2O4 is small, even negligible, compared to
that of MoO3 admixed and could thus not bring about a
significant cooperative effect. The mixture of SbmMo10O31 with
R-Sb2O4 resulted after decomposition in a mixture containing
more R-Sb2O4. Hence, even if no synergism was observed for
the conversion, the synergism for the selectivity to methacrolein
was higher than for the mixture of SbmMo10O31 with MoO3.
For both mixtures involving Sb2MoO6, cooperative effects were
detected. The decomposition of Sb2MoO6 gave rise to a mixture
of MoO3 and R-Sb2O4 with a composition 33/67 wt %. The
additional supply of one of the two simple oxides resulted in
mixtures with overall compositions still suitable for developing
cooperations. It is however interesting to point out that the
mixture of Sb2MoO6 with R-Sb2O4, with a higher overall content
of R-Sb2O4, exhibited also a more important synergism than
the mixture of Sb2MoO6 with MoO3. The results are fully
consistent with the interpretation that the special activity of Sb-
Mo-O catalysts was due to a cooperation between the simple
oxide phases formed in operandi at their surface, namely,
R-Sb2O4 and MoO3 (or a slightly reduced MoO3-x phase), rather
than to the real performances of binary phases. Figure 14
summarizes this conclusion.27
More generally, these conclusions support the interpretation
of the cooperation between MoO3 (or slightly reduced MoO3-x
)
and R-Sb2O4 in selective oxidation reactions via a remote control
mechanism, as this considers that the synergetic effects have
to be understood on the basis of phases remaining completely
separate and mutually noncontaminated under the conditions
of the catalytic reaction.3,27
Acknowledgment. E.M.G. thanks Professor R. Schlo¨gl
(Fritz-Haber-Institut der Max-Planck-Gesellschaft, Berlin, Ger-
many) for the invitation to perform the confocal Raman
experiments in his laboratory. The Max-Planck-Gesellschaft
(Germany) is gratefully acknowledged for the financial support
of this stay. We thank Professor M. Parmentier (Ecole Nationale
Supe´rieure d’Agronomie et des Industries Alimentaires, Van-
doeuvre-Le`s-Nancy, France) for the very helpful advice provided
for the synthesis of the Sb-Mo-O phases.
References and Notes
(1) Gaigneaux, E. M.; Ruiz, P.; Delmon, B. Catal. Today 1996, 32,
37.
(2) Gaigneaux, E. M.; Ruiz, P.; Wolf, E. E.; Delmon, B. Appl. Surf.
Sci. 1997, 121/122, 552.
(3) Weng, L. T.; Delmon, B. Appl. Catal. A 1992, 81, 141.
(4) Kohlmuller, R.; Badaud, J. P. Bull. Soc. Chim. Fr. 1969, 10, 3434.
(5) Xiong, Y. L.; Zeyss, S.; Weng, L. T.; Ruiz, P.; Delmon, B. To be
submitted.

Catalytic Behavior of Three Sb-Mo-O Phases
J. Phys. Chem. B, Vol. 102, No. 51, 1998 10555
(6) Weng, L. T.; Xiong, Y. L.; Ruiz P.; Delmon, B. Catalysis Science
Technology; Kodansha: Tokyo, Japan, 1991; Vol. 1, p 207.
(7) Weng, L. T.; Sham, E.; Doumain, B.; Ruiz, P.; Delmon, B. Stud.
Surf. Sci. Catal. 1990, 55, 797.
(17) Parmentier, M. Doctoral Thesis, Universite´ de Nancy I, Nancy,
France, 1977.
(18) Parmentier, M.; Gleitzer, Ch.; Courtois, A.; Protas, J. Acta
Crystallogr. Sect. B 1979, 35, 1963.
(8) Qiu, F. Y.; Weng, L. T.; Sham, E.; Ruiz, P.; Delmon, B.
Proceedings of the 2nd Conference on SpilloVer; Karl-Marx-Universita¨t:
Leipzig, Germany, 1989; p 136.
(9) Zhou, B.; Sham, E.; Machej, T.; Bertrand, P.; Ruiz, P.; Delmon,
B. J. Catal. 1991, 132, 157.
(10) Xiong, Y. L.; Weng, L. T.; Zhou, B.; Yasse, B.; Sham, E.; Daza,
L.; Gil-Llambias, F.; Ruiz, P.; Delmon, B. Stud. Surf. Sci. Catal. 1991, 63,
537.
(11) Xiong, Y. L.; Castillo, R.; Daza, L.; Ruiz, P.; Delmon, B. Stud.
Surf. Sci. Catal. 1991, 63, 425.
(12) Weng, L. T.; Zhou, B.; Yasse, B.; Doumain, B.; Ruiz, P.; Delmon,
B. Proceedings of the 9th International Congress on Catalysis; Chemical
Institute of Canada: Ottawa, Canada, 1988; Vol. 4, p 1609.
(13) Gaigneaux, E. M.; Tsiakaras, P. E.; Herla, D.; Ghenne, L.; Ruiz,
P.; Delmon, B. Catal. Today 1997, 33, 151.
(14) Zhou, B. Doctoral Thesis, Universite´ catholique de Louvain,
Louvain-la-Neuve, Belgium, 1988.
(15) Germain, J. E.; Perez, R. Bull. Soc. Chim. Fr. 1975, 3, 735.
(16) Parmentier, M.; Courtois, A.; Gleitzer, Ch. Bull. Soc. Chim. Fr.
1974, 1, 75.
(19) Parmentier, M.; Gleitzer, Ch.; Tilley, R. J. D. J. Solid State Chem.
1980, 31, 305.
(20) Blangenois, N.; Ruiz, P.; Delmon, B. Universite´ catholique de
Louvain, Unite´ de catalyse et chimie des mate´riaux divise´s: Louvain-la-
Neuve, Belgium. Eur. Pat. 93870054.9, March 24, 1993.
(21) 1996-JCPDS International Centre for Diffraction Data.
(22) Mestl, G.; Ruiz, P.; Delmon, B.; Kno¨zinger, H. J. Phys. Chem.
1994, 98, 11269.
(23) Mestl, G.; Ruiz, P.; Delmon, B.; Kno¨zinger, H. J. Phys. Chem.
1994, 98, 11276.
(24) Mestl, G.; Ruiz, P.; Delmon, B.; Kno¨zinger, H. J. Phys. Chem.
1994, 98, 11283.
(25) Spevack, P. A.; McIntyre, N. S. J. Phys. Chem. 1992, 96, 9029.
(26) Gaigneaux, E. M.; Ruiz, P.; Delmon, B. J. Chem. Soc., Chem.
Commun. submitted.
(27) Gaigneaux, E. M. Doctoral Thesis, Universite´ catholique de
Louvain, Louvain-la-Neuve, Belgium, 1997.
(28) Zhou, B.; Machej, T.; Ruiz, P.; Delmon, B. J. Catal. 1991, 132,
183.