Characterization and reactivity of molybdenum oxide catalysts supported on niobia

Article abstract of DOI:10.1021/jp003201y

A series of MoO₃/Nb₂O₅ catalysts with Mo loadings varying from 2.5 to 15 wt% were prepared and characterized by X-ray diffraction (XRD), temperature-programmed reduction (TPR), temperature-programmed desorption (TPD) of am

Full text of DOI:10.1021/jp003201y

4392
J. Phys. Chem. B 2001, 105, 4392-4399
Characterization and Reactivity of Molybdenum Oxide Catalysts Supported on Niobia†
Komandur V. R. Chary,* Thallada Bhaskar, Gurram Kishan, and
Kondakindi Rajender Reddy
Catalysis DiVision, Indian Institute of Chemical Technology, Hyderabad 500 007, India
ReceiVed: September 11, 2000; In Final Form: December 28, 2000
A series of MoO3/Nb2O5 catalysts with Mo loadings varying from 2.5 to 15 wt % were prepared and
characterized by X-ray diffraction (XRD), temperature-programmed reduction (TPR), temperature-programmed
desorption (TPD) of ammonia, electron spin resonance (ESR), oxygen chemisorption, and pore size distribution
measurements. X-ray diffraction patterns indicate the presence of a crystalline molybdenum phase at higher
Mo loadings on niobia. Dispersion of molybdenum was determined by the oxygen chemisorption at 623 K
by a static method on the samples prereduced at the same temperature. At low Mo loadings, i.e., <10.0%,
molybdenum oxide is found to be present in a highly dispersed state. Pore size distribution studies indicate
a decrease in average pore diameter and pore volume with increased Mo loading. ESR results suggest the
presence of Mo5+ in the reduced catalysts. TPR results suggest that the reducibility of MoO3 increases with
increased Mo loading. The reduction peaks due to niobia in TPR appeared at high temperatures (>1173 K)
and their intensity decreases with Mo loading. TPD of ammonia results suggest that acidity of the catalysts
was found to increase with increased molybdenum loading. The catalytic properties were evaluated for the
vapor-phase ammoxidation of 3-picoline to nicotinonitrile and related to oxygen chemisorption sites.
Introduction
The study of determining the dispersion of active phase in
supported metal oxide systems is an interesting topic of research
in recent years for understanding the role of the active phase
on catalytic activity/selectivity during the oxidation reactions.
A fundamental understanding of the structure-activity relation-
ships observed in heterogeneous catalytic oxidation is of basic
importance for the development of new catalytic materials and
for improving the performance of existing catalysts.35 To this
end, methods such as oxygen chemisorption have been studied
extensively in recent years to find active phase dispersion in
supported metal oxide systems.
In the present investigation we report the characterization of
MoO3/Nb2O5 catalysts by powder X-ray diffraction (XRD),
oxygen chemisorption and pore size distribution measurements
(PSD), electron spin resonance (ESR), temperature-programmed
reduction (TPR), and temperature-programmed desorption (TPD).
The catalytic properties have been evaluated for the ammoxi-
dation of 3-picoline to nicotinonitrile. We also report the relation
between dispersion of molybdenum oxide and catalytic proper-
ties of the catalysts during vapor phase ammoxidation of
3-picoline to nicotinonitrile. The purpose of this work is to
estimate the dispersion of molybdenum oxide supported on
niobia as a function of molybdenum loading and to identify
the changes in structure of the molybdena phase with increased
active phase loading and also to understand the relation between
activity/selectivity and oxygen chemisorption sites.
Catalysts containing molybdenum oxide/sulfide as an active
component have been extensively employed in the recent past
for the partial oxidation of hydrocarbons and alcohols and also
extensively used in hydroprocessing reactions for petroleum
industry.1-15 The development in petroleum refining technology
in the last three decades has raised hydroprocessing reactions
to a high level of economic importance. The catalytic properties
of the active molybdenum oxide phase can be greatly influenced
by the nature of the supported oxide and the dispersion of active
component. The most commonly used supports are alumina,16-18
silica,7,9-10,19-21,38 titania,11,15,22-24,36 and zirconia.25-27 In recent
years niobium-based materials have been employed as catalysts
in numerous catalytic applications.28-34 Niobia can be used as
support, as promoter, and as a unique solid acid. Smits33
emphasized the advantages of niobia as a catalyst support for
vanadia. These include the following: (i) Niobium is in the same
group of the periodic table as vanadium or molybdenum and is
expected to have similar properties. (ii) Niobium is much more
difficult to reduce than vanadium or molybdenum (easy reduc-
tion often causes low selectivity in selective oxidation reactions).
(iii) The addition of niobium oxide to a mixture of molybdenum
and vanadium oxides improves the activity and selectivity for
oxidation, ammoxidation, and oxidative dehydrogenation reac-
tions.33,34 Huuhtanen et al.31 reported a comparison of different
supports, and niobia appears to be promising as for as selectivity
in toluene oxidation was concerned. Matsuura et al.14 reported
that the catalytic activity during the ammoxidation of isobutane
to methacrylonitrile was improved when the Bi-Mo-based
composites were supported on Nb2O5 rather than γ-Al2O3 or
SiO2.
Experimental Section
Catalyst Preparation. A series of MoO3 catalysts with Mo
loadings ranging from 2.5 to 15 wt % supported on Nb2O5
(surface area 55 m2 g-1) were prepared by incipient wetting of
the support with aqueous ammonium heptamolybdate (Fluka
AG, Switzerland) solution at pH 8. The catalysts were subse-
quently dried at 383 K for 16 h and calcined in air at 773 K for
6 h. The niobium pentoxide hydrate (niobia HY-340 AD/1227,
† IICT Communication Number 4586.
* To whom correspondence should be addressed: e-mail kvrchary@
iict.ap.nic.in.
10.1021/jp003201y CCC: $20.00 © 2001 American Chemical Society
Published on Web 04/19/2001

Molybdenum Oxide Catalysts Supported on Niobia
J. Phys. Chem. B, Vol. 105, No. 19, 2001 4393
CBMM, Brazil) was calcined in air at 773 K for 4 h before
impregnation with ammonium heptamolybdate.
helium (50 mL/min) at 473 K for 2 h. After pretreatment, the
sample was saturated by passage of highly pure anhydrous
ammonia (75 mL/min) at 353 K and subsequently flushed at
378 K for 2 h to remove the physisorbed ammonia. TPD analysis
was carried out from ambient temperature to 1073 K at a heating
rate of 10 K/min. The ammonia concentration in the effluent
stream was monitored with the thermal conductivity detector
and the areas under the peaks were integrated by use of
GRAMS/32 software to determine the amount of desorbed
ammonia during TPD. TCD calibration was performed by an
automated experiment by passing known volumes of ammonia.
Ammoxidation of 3-Picoline. A downflow fixed-bed reactor
operating at atmospheric pressure and made of Pyrex glass was
used to test the catalysts during the ammoxidation of 3-picoline
to nicotinonitrile. About 2 g of the catalyst diluted with an equal
amount of quartz grains was charged into the reactor and was
supported on a glass wool bed. Prior to introduction of the
reactant 3-picoline with a syringe pump (B-Braun perfusor),
the catalyst was reduced at 723 K for 2 h in purified hydrogen
flow (40 mL/min). After the prereduction the reactor was fed
with 3-picoline, ammonia, and air, keeping the mole ratio of
3-picoline:H2O:NH3:air at 1:13:11:44. The reaction was carried
out at various temperatures ranging from 573 to 723 K. The
liquid products, mainly nicotinonitrile, were analyzed by the
HP 6890 gas chromatograph equipped with an FID using OV-
17 column.
X-ray Diffraction and Electron Spin Resonance. ESR
spectra were recorded at ambient temperature on a Bruker ER-
200D-SRC X-band spectrometer with 100 kHz modulation. The
reduced catalysts for the ESR study were prepared in quartz
tubes (25 cm long, 4 mm diameter). The samples were
prereduced at 623 K for 2 h in a continuous flow (40 mL/min)
of purified hydrogen. The setup was subsequently evacuated
for 1 h at 10-6 Torr. The catalyst thus prepared was transferred
to the ESR tube and sealed off under vacuum. X-ray diffraction
patterns were recorded on Siemens D-5000 X-ray diffractometer
with graphite-filtered Cu KR radiation.
Oxygen Chemisorption. Oxygen chemisorption was mea-
sured by a static method with an all-Pyrex glass system capable
of attaining a vacuum of 10-6 Torr. The details of the
experimental set up are given elsewhere.36 Before adsorption
measurements, the samples were prereduced in a flow of
hydrogen (40 mL/min) at 623 K for 2 h and evacuated at the
same temperature for 1 h. Oxygen chemisorption uptakes were
determined as the difference of two successive adsorption
isotherms measured at 623 K. The surface areas of the catalysts
were determined by the BET method with nitrogen physisorption
at 77 K, taking 0.162 nm2 as its cross-sectional area. Pore size
distribution (PSD) measurements were performed on Auto Pore
III (Micromeritics) by the mercury penetration method.
Temperature-Programmed Reduction. TPR studies were
conducted on AutoChem 2910 (Micromertitics) instrument. The
unit has a programmable furnace with a maximum operating
temperature of 1373 K. The instrument was interfaced with a
computer that performs tasks such as programmed heating and
cooling cycles, continuous data recording, gas valve switching,
data storage, and analysis.
Results and Discussion
The X-ray diffraction patterns of calcined MoO3/Nb2O5
catalysts are presented in Figure 1. In all the samples XRD peaks
due to low-temperature niobia were observed at d ) 3.95, 3.14,
2.45, 1.97, and 1.66 Å. At higher Mo loadings (above 7.5 wt
% Mo, Figure 1d), XRD peaks due to the crystalline MoO3
phase are noticed at d ) 3.26 and 3.81 Å, in addition to the
characteristic peaks of niobia. The intensity of these peaks
increases with MoO3 loading. However, at low Mo loading the
absence of crystalline MoO3 peaks cannot be ruled out as they
might be less than 40 Å in size, which is beyond the detection
capacity of the XRD technique. XRD results also suggest that
no mixed oxide is formed between MoO3 and Nb2O5. The two
low-temperature forms of Nb2O5, i.e., TT and T, have long
thought to be the same,37 because (i) they have similar X-ray
diffraction patterns and (ii) the TT phase does not always appear
with pure components as starting materials. According to Ko
and Weissman,37 TT phase may be a crystalline form of T,
stabilized by impurities. Ko and Weissman37 also reported that,
at low calcination temperature (773 K), the samples are found
to be X-ray amorphous. However, the samples calcined between
773 and 873 K show the TT phase of Nb2O5, and the samples
calcined between 873 and 973 K favor the formation of the T
phase of Nb2O5. At 1073 K calcination, the M phase of Nb2O5
is observed, and at 1273 K and above, the H phase of Nb2O5 is
observed. In the present study, the samples were calcined at
773 K and our XRD results (Figure 1) also suggest the formation
of the TT phase of Nb2O5, which is in good agreement with
the work of Ko and Weissman.37 However, although the d
spacing of TT and T phases of Nb2O5 is similar, the intensities
of d spacing are different. At d ) 3.94 Å the intensities were
found to be 100 and 84 for TT and T phases, respectively.
However, at d ) 3.13 Å the intensities were found to be 90
and 100 for TT and T phases. The intensities of d spacings in
the present XRD patterns (Figure 1) suggest that only the TT
phase of Nb2O5 is observed. XRD patterns of pure MoO3 and
pure Nb2O5 are shown in Figure 2. Pure Nb2O5 shows intense
In a typical TPR experiment about 250 mg of MoO3/Nb2O5
sample dried at 383 K for 16 h was taken in a U-shaped quartz
sample tube. The catalyst was packed on a quartz wool plug in
one arm of the sample tube. The temperature was monitored
with the aid of thermocouples located near the sample from
outside and on the top of the sample. The gas flows were
monitored by highly sensitive mass-flow controllers. Before TPR
studies the catalyst samples were pretreated by passage of ultra-
high-purity helium (50 mL/min) at 673 K for 2 h. After
pretreatment the sample was cooled to room temperature. The
reducing gas consists of 5% hydrogen and balance argon (50
mL/min), which was purified by passage through oxy-trap and
molecular sieves. The water produced during reduction was
condensed in a cold trap kept in a liquid nitrogen and 2-propanol
slurry. The hydrogen concentration in the effluent stream was
monitored with the thermal conductivity detector, and the areas
under the peaks were integrated by use of GRAMS/32 software
to determine hydrogen consumption. Tmax calibration of the TCD
was performed by stoichiometric reduction of a known amount
of high-purity Ag2O to metallic silver, a method that was found
to be more reliable and reproducible than sending the known
volumes of hydrogen pulses through the reactor.
Temperature-Programmed Desorption. In a typical experi-
ment for TPD studies, about 200 mg of oven-dried sample (dried
at 383 K for 16 h) was taken in a U-shaped quartz cell. The
catalyst sample was packed in one arm of the sample tube on
a quartz wool bed. The temperature was monitored with the
aid of thermocouples located near the sample from outside and
one on the top of the sample. The gas flows were monitored by
highly sensitive mass-flow controllers. Prior to TPD studies,
the catalyst sample was pretreated by passage of high-purity

4394 J. Phys. Chem. B, Vol. 105, No. 19, 2001
Chary et al.
Figure 2. X-ray diffractograms of pure MoO3 and pure Nb2O5
samples: (a) Nb2O5; (b) MoO3.
TABLE 1: Results of Oxygen Uptake, Dispersion, Oxygen
Atom Site Density, and Surface Area of Various MoO3/
Nb2O5 Catalysts
catalyst
composition
(wt % Mo)
surface
areaa
oxygen
uptakeb
oxygen atom
site density
dispersionc
(O/Mo)
(m2 g-1
)
(µmol g-1
)
(×1018 m-2
)
2.5
5.0
7.5
10.0
12.5
15.0
50
44
40
38
29
27
127.9
248.6
369.2
472.1
545.4
604.5
2.69
6.23
11.32
15.64
21.73
30.65
0.98
0.93
0.94
0.90
0.83
0.77
b
a BET surface area determined after oxygen chemisorption. T(reduction)
) T(adsorption) ) 623 K. c Dispersion ) fraction of molybdenum atoms
at the surface assuming Oads/Mosurf ) 1.
Figure 3. Oxygen uptake plotted as a function of Mo loading on
Niobia. (Tadsorption ) Treduction ) 623 K).
Figure 1. X-ray diffractograms of MoO3/Nb2O5 catalysts: (a) 2.5%
Mo/Nb2O5; (b) 5.0% Mo/Nb2O5; (c) 7.5% Mo/Nb2O5; (d) 10.0% Mo/
Nb2O5; (e) 12.5% Mo/Nb2O5; (f) 15.0% Mo/Nb2O5. Solid circles
indicate peaks due to MoO3.
niobia and it might be due to blocking of the pores of the support
by crystallites of niobium oxide and also molybdena as
evidenced by XRD and PSD. The oxygen chemisorption uptake
results of various catalysts are also presented in Table 1 and
the other information, such as oxygen atom site density,
dispersion, etc., derived from it are also given in Table 1. The
oxygen atom site density, defined as the number of oxygen
atoms chemisorbed per unit area of the reduced MoO3 surface,
was found to increase with increased Mo loading.
Figure 3 shows the oxygen uptake measured at 623 K for
various MoO3/Nb2O5 catalysts plotted as a function of Mo
content on niobia. The oxygen chemisorption uptakes are found
to increase with increased molybdena content. The dispersion
of molybdena was found to decrease steadily with increased
diffraction peaks at d ) 3.95 and 3.14 Å, due to the T phase of
Nb2O5. Similarly pure MoO3 show intense peaks at d ) 3.26,
3.81, and 3.46 Å. The present XRD results are in good
68
agreement with our earlier studies on Mo/TiO267 and Mo/ZrO2
catalysts, wherein crystalline MoO3 appeared above monolayer
coverage. Desikan et al.38,45 also reported the appearance of
crystalline MoO3 at higher Mo loadings in MoO3/SiO2,38 MoO3/
45
TiO2,45 and MoO3/Al2O3 catalysts.
The specific surface areas determined by nitrogen physisorp-
tion of all the catalysts are presented in Table 1. The specific
surface area decreases as a function of molybdena content on

Molybdenum Oxide Catalysts Supported on Niobia
J. Phys. Chem. B, Vol. 105, No. 19, 2001 4395
molybdena content. In Figure 3, the dashed line corresponds to
a stoichiometry of one oxygen atom per molybdenum atom.
Pure Nb2O5 was also reduced under identical conditions and
its oxygen uptake was corrected for the supported catalysts.
Dispersion of molybdena is defined as the fraction of total O
atoms to total Mo atoms in the sample. The dispersion was found
to be 98% for 2.5 wt % Mo/Nb2O5 and 77% for 15 wt % Mo/
Nb2O5. These findings are in excellent agreement with the
oxygen chemisorption results of Oyama and co-workers38 on
molybdena supported on silica and also with our earlier studies
on MoO3/TiO2.36 The decrease of Mo dispersion at higher
loadings might be due to formation of microcrystalline MoO3
in addition to niobia as evidenced from XRD results.
The reduction behavior of supported molybdena catalysts
prior to oxygen chemisorption is an interesting topic. Many
authors have reported the reduction of molybdena at 773 K
followed by oxygen chemisorption at 195 or 77 K for determin-
ing dispersion of molybdena.11,39-43 These conditions were
evaluated by Rodrigo et al.,44 who concluded that oxygen
chemisorption under these conditions does not provide a
quantitative determination of Mo dispersion because a fraction
of the reduced Mo is in the bulk. It has been shown recently
that oxygen chemisorption sites are easily generated under very
mild reduction conditions.38,45,46
We have calculated the theoretical monolayer capacity of
MoO3 supported on Nb2O5 based on the method described by
Van Hengstum et al.47 taking 0.16 wt % MoO3/m2 of surface.
Accordingly, the theoretical monolayer capacity of MoO3
supported on niobia employed in the present study having a
surface area of 55 m2 g-1 corresponds to 8.8% MoO3 or 5.86%
Mo. However, XRD results from the present work show the
presence of MoO3 crystallites from 10.0 wt % Mo (Figure 1d),
which indicates the formation of monolayer and is in good
agreement with the theoretical monolayer based on the structure
of MoO3. In the present study the leveling off of oxygen
chemisorption beyond the monolayer composition (above 6%
Mo loading) might be due to the presence of the larger
crystallites of MoO3 as evidenced from the X-ray diffraction
results (Figure 1d-f). These larger crystallites of molybdenum
oxide are preventing reduction of the catalyst with hydrogen
gas; therefore, no appreciable change in oxygen uptake is
observed for the catalysts beyond the monolayer composition.
Figure 4 represents incremental intrusion volume vs pore
diameter of various MoO3/Nb2O5 samples. All the samples show
bimodal distribution with the majority of pores present in the
large pore diameter range (>1000 Å). However, the small
amount of volume is also concentrated in the median pore
region. As the molybdena loading is increased, the population
of median pores is also found to increase in the catalysts. It has
been observed that the average pore diameter increases margin-
ally with MoO3 loading. This change might be due to blockage
of pores, which can be noticed from the decrease in the intensity
of pores. Similarly, the total pore area also decreases with
addition of MoO3. The details of total pore area, total intrusion
volume, and average pore diameter of MoO3/Nb2O5 catalysts
are reported in Table 2. The reactants and products (3-picoline
and nicotinonitrile) can easily pass through these median pores.
The increase in the median pore population and increase in
3-picoline ammoxidation conversion with molybdena loading
can be observed from Figures 3 and 9, respectively. This clearly
indicates that the median pores play a vital role during the
reaction.
Figure 4. Pore size distribution (PSD) studies of MoO3/Nb2O5
catalysts: (a) 2.5% Mo/Nb2O5; (b) 7.5% Mo/Nb2O5; (c) 12.5% Mo/
Nb2O5.
TABLE 2: Results of Pore Size Distribution Analysis for
MoO3/Nb2O5 Catalysts
wt % Mo
on Nb2O5
total intrusion
volume (mL/g)
total pore
average pore
diameter (Å)
area (m2/g)
2.5
7.5
12.5
0.3605
0.2971
0.2114
65.298
46.664
39.962
221
255
272
reduced catalysts recorded at 300 K are represented in Figure
5. In all the catalysts an axially symmetric spectrum appears at
the center of the pattern with gII ) 1.917. The intensity of the
anisotropic spectrum increased with Mo loading. The spectral
features of hydrogen-reduced samples suggest the presence of
Mo5+ and are in good agreement with earlier works.48
The TPR profile of unsupported MoO3 is presented in Figure
6. The TPR profile of pure MoO3 shows two major peaks at
1040 and 1270 K and one minor reduction peak at 1070 K. For
TPR analysis of unsupported MoO3, the reduction conditions
The ESR results of the hydrogen-reduced catalysts further
support the findings of oxygen chemisorption. The spectra of

4396 J. Phys. Chem. B, Vol. 105, No. 19, 2001
Chary et al.
Figure 5. ESR spectra of hydrogen-reduced MoO3/Nb2O5 catalysts:
(a) 2.5% Mo/Nb2O5; (b) 5.0% Mo/Nb2O5; (c) 7.5% Mo/Nb2O5; (d)
10.0% Mo/Nb2O5; (e) 12.5% Mo/Nb2O5; (f) 15.0% Mo/Nb2O5.
Figure 7. Temperature-programmed reduction (TPR) profiles of MoO3/
Nb2O5 catalysts: (a) 2.5% Mo/Nb2O5; (b) 5.0% Mo/Nb2O5; (c) 7.5%
Mo/Nb2O5; (d) 10.0% Mo/Nb2O5; (e) 12.5% Mo/Nb2O5; (f) 15.0% Mo/
Nb2O5.
the first major peak is observed at 1070 K, which corresponds
to Mo4O11 formed by reduction of MoO3. Thomas et al.49 also
noticed the following peak, which was confirmed by in situ
X-ray diffraction:
MoO3 f Mo4O11
(3)
Temperature-programmed reduction profiles of the niobia-
supported molybdenum oxide catalysts are shown in Figure 7.
TPR profiles of niobia-supported molybdenum oxide catalysts
indicates that molybdena is reducing in two stages. The first
peak Tmax value increases with increased molybdena loading
from 783 to 858 K, and also the area under the reduction peak
increases with increased molybdenum loading. The Tmax value
for the second peak also increases with increased loading from
1063 to 1113 K, and the area under the reduction peak increases
with increased molybdena loading. It is well-known that niobium
oxide is partially reducible when exposed to hydrogen at high
temperatures, and the reduction is accelerated by the presence
of supported zerovalent metals on its surface. For this reason,
metals supported on niobium oxide are known to exhibit strong
metal-support interaction (SMSI) when reduced with hydrogen
at high temperatures.51-53 The reduction of Nb2O5 is partial and
occurs around 1273 K. However, the reducibility of Nb2O5
becomes easier when it is associated with MoO3, and this can
be seen from Figure 7. On the other hand, the Tmax for MoO3
reduction in two stages is found to increase with MoO3 loading
on Nb2O5 surface. Compared to the bulk MoO3 species, the
reduction of MoO3 species on Nb2O5 becomes much easier. It
clearly indicates that there might be an interaction between
Figure 6. Temperature-programmed reduction (TPR) profile of
unsupported MoO3.
applied were similar to those for MoO3/Nb2O5 catalysts.
According to Thomas et al.49 and Arnoldy et al.,50 the reduction
of molybdena essentially can take place in two steps. The
reducibility of MoO3 is represented by the following steps:
MoO3 f MoO2
(1)
The sharp peak at 1040 K corresponds to reduction of MoO3
MoO2 f Mo (2)
(first step) and the peak at 1270 K is associated with the
reduction of MoO2 (second step). A minor peak at the edge of

Molybdenum Oxide Catalysts Supported on Niobia
J. Phys. Chem. B, Vol. 105, No. 19, 2001 4397
Figure 9. Ammoxidation of 3-picoline over MoO3/Nb2O5 catalysts
(reaction temperature 683 K).
acidic sites in the temperature region of 373-673 K are due to
moderate acid sites, and those in the region of 673-873 K are
due to strong acid sites. TPD results suggest that pure Nb2O5 is
less acidic compared to Mo/Nb2O5 catalysts. The number of
acid sites with moderate strength was found to increase with
Mo loading on niobia. However, the NH3 uptake due to strong
acid sites (Table 3) was found to be almost equal for 7.5 and
15 wt % Mo/Nb2O5 catalysts. This behavior is in agreement
with the catalytic activity beyond 7.5 wt % Mo loading, which
did not change appreciably with increased Mo loading on niobia
support. This clearly indicates that strong acid sites are
responsible for picoline ammoxidation. The TPD results suggests
that the strength of acid sites plays a crucial role in determining
the catalytic activity during ammoxidation of 3-picoline, and
the acidity of the catalysts is mainly due to the molybdena phase
since ammonia uptake increases with increased molybdena
loading. The relation obtained can be explained in terms of stable
Figure 8. Ammonia temperature-programmed desorption (TPD)
profiles of MoO3/Nb2O5 catalysts: (a) Pure Nb2O5; (b) 2.5% Mo/Nb2O5;
(c) 7.5% Mo/Nb2O5; (d) 15.0% Mo/Nb2O5.
TABLE 3: Results of Temperature-Programmed Desorption
of Ammonia for MoO3/Nb2O5
wt % Mo Tmax
1
NH3
Tmax2
NH3
S no. on Nb2O5
(K)
uptake (mL/g)
(K)
uptake (mL/g)
1
2
3
4
0.0
2.5
7.5
561
553
537
554
5.24
4.07
3.38
6.37
826
805
785
0.42
1.70
1.60
15.0
+
Bronsted-acid centers that chemisorb ammonia as NH4 and
are also of interest in the ammoxidation of 3-picoline.61 Iizuka
et al.58 showed that the surface of Nb2O5‚nH2O showed strong
acidic character (H0 e -5.6) even after heating in air at 373
K, but the strong acid sites disappeared at higher temperatures.
The active sites on the catalysts heat-treated at moderate
temperatures were ascribed mainly to Bronsted acid sites on
the basis of the IR study of adsorbed pyridine, and it is known
that the selectivity of products is very sensitive to the surface
acidic and basic properties.60
Figure 9 shows the dependence of activity and selectivity on
the molybdena loading during ammoxidation of 3-picoline to
nicotinonitrile at 683 K. The catalytic experiments were repeated
twice, and the values reported are the average of two sets of
experiments. The conversion of 3-picoline is found to increase
with Mo loading in the catalysts up to 7.5 wt % Mo loading.
Beyond this loading the activity did not change appreciably.
The selectivity toward nicotinonitrile formation is also found
to increase with Mo loading. However, at high Mo loadings
the selectivity remains unchanged. Pure Nb2O5 was also found
to be active for the nicotinonitrile formation under the experi-
mental conditions employed and corrected for the catalyst
samples. The conversion of 3-picoline by pure Nb2O5 was found
to be 3.14% and the product formed was only cyanopyridine.
The contribution of pure Nb2O5 toward ammoxidation was
subtracted from the conversion of MoO3/Nb2O5 catalysts.
MoO3 and Nb2O5. However, no such compounds are observed
from the powder X-ray diffraction patterns. Most probably the
interacting species may be present in an amorphous phase. The
TPR results thus suggest that the reducibility of molybdena
increases with increased Mo loading in MoO3/Nb2O5 catalysts.
Niobium supported on certain oxides has shown a remarkable
promoter effect for decomposition of NO,28 and niobium
pentoxide treated at relatively low temperatures, Tt (423 < Tt
< 573 K), is an effective catalyst for various reactions such as
esterification, dehydration, and isomerization reactions.54-57 The
surface acidity of niobium materials has been studied by
n-butylamine titration with Hammett indicators, by infrared
spectroscopy of adsorbed pyridine,54,58 and by volumetric and
gravimetric analysis of adsorbed NH3.59 The surface of Nb2O5‚
xH2O contains both Lewis and Bronsted acid sites.54,60 Thus it
is interesting to study the acidity of the molybdenum oxide
supported on niobia. In the present study, the acidity measure-
ments have been carried out by the ammonia TPD method.
Ammonia TPD profiles for pure Nb2O5 and some of the
representative Mo/Nb2O5 catalysts are shown in Figure 8. The
ammonia uptake by various Mo/Nb2O5 catalysts and the
temperature positions are given in Table 3. The TPD profiles
(Figure 8) suggest the strength of acid sites and their distribution
in two temperature regions, i.e., 373-673 and 673-873 K. The

4398 J. Phys. Chem. B, Vol. 105, No. 19, 2001
Chary et al.
phase, and this functionality can be titrated by the oxygen
chemisorption method reported in this work.
Conclusions
The results of oxygen chemisorption suggest that molybde-
num oxide is found to be highly dispersed on the niobia support.
Pore size distribution studies indicated decreased average pore
diameter and pore volume with increased Mo loading. TPR
results demonstrated that the reducibility of MoO3 increased
with increased Mo loading in Mo/Nb2O5 catalysts. TPD of
ammonia indicates the acidity falls into two regions, and acidity
of the catalysts was found to increase with increased molybdena
loading. The catalytic activity during 3-picoline ammoxidation
can be related to dispersion of molybdena.
Acknowledgment. CBMM, Brazil, is gratefully acknowl-
edged for providing the gift samples of Nb2O5. We thank Dr.
K. V. Raghavan, Director, IICT, for encouragement. We are
thankful to Dr. K. S. Rama Rao for helpful discussions. T.B.
and G.K. thank the Council of Scientific and Industrial Research
(CSIR), New Delhi, for the research associate (RA) positions,
and K.R.R. thank CSIR, New Delhi, for a junior research
fellowship (JRF).
Figure 10. Relationship between the amount of surface Mo on Nb2O5
and the rate of 3-picoline conversion.
Several groups62-64 studied the acid-base properties of
supported vanadia catalysts in the ammoxidation of 3-picoline.
It was found that high activity in the ammoxidation of 3-picoline
corresponds to a relatively small amount of acidic sites. A
catalyst selective in the formation of nicotinonitrile requires a
high concentration of both acidic and basic sites. The conditions
in the ammoxidation of 3-picoline are both reductive and
oxidative; i.e., the hydrocarbons consume oxygen from the
catalyst, which is then reoxidized. It can be expected that, under
steady-state conditions, the catalyst will contain a certain amount
of lower oxides formed by reduction of the originally charged
catalyst.
References and Notes
(1) Massooth, F. E. AdV. Catal. 1978, 27, 265.
(2) Grange, P. Catal. ReV. Sci. Eng. 1980, 21, 135.
(3) Knozinger, H. In Proceedings of the 9th International Congress
on Catalysis, Calgary, Canada, 1988; Phillips, M., Ternan, M., Eds; The
Chemical Institute of Canada: Ottawa, Canada, 1989; Vol. 5, p 20.
(4) Reddy, B. M.; Chary, K. V. R.; Subrahmanyam, V. S.; Nag, N. K.
J. Chem. Soc. Faraday Trans. 1 1985, 81, 1655.
(5) Ng, K. Y. S.; Gulari, E. J. Catal. 1985, 95, 33.
(6) Chung, J. S.; Miranda, R.; Bennett, C. O. J. Catal. 1988, 144, 898.
(7) Louis, C.; Tatibouet, J. M.; Che, M. J. Catal. 1988, 109, 354.
(8) Liu, Y. C.; Griffin, G. L.; Chan, S. S.; Wachs, I. E. J. Catal. 1985,
94, 108.
To find the relation between the ammoxidation activity of
3-picoline and the dispersion of molybdena, a plot of TOF versus
surface Mo content is shown in Figure 10, where TOF is defined
as the number of picoline molecules converted per second per
surface Mo. The TOF was found to be almost constant (≈7 ×
10-4 s-1) up to 7.5 wt % Mo and decreased at higher Mo
loadings. Up to 7.5 wt % Mo loading, per-site activity (constant
TOF) is constant with the increase in surface Mo sites. The
decrease in TOF beyond 7.5 wt % Mo loading indicates the
formation of bigger MoO3 particles, but not enough to attain
the bulk nature. The absence of direct correlation between TOF
and amount of surface Mo may be due to the fact that, in
addition to oxygen adsorption, ammonia adsorption may also
be responsible for this reaction. A direct correlation was also
reported earlier between oxygen uptake capacities of various
vanadia65,66 and molybdena67 catalysts and nicotinonitrile
product selectivity. As reported elsewhere in connection with
molybdenum oxide catalysts supported on alumina16 and on
zirconia,41 oxygen is chemisorbed at low temperatures selec-
tively on coordinatively unsaturated sites (CUS), generated upon
reduction, having a particular coordination environment. These
sites are located on a highly dispersed molybdena phase, which
is formed only at low molybdena loadings and remain as a
“patchy monolayer” on the support surface. At higher molyb-
dena loadings, a second phase is formed, in addition to the
already existing monolayer, and this postmonolayer phase does
not appreciably chemisorb oxygen (Figure 3). In the perspective
of the above background, the correlation shown here indicates
that the catalytic functionality of the dispersed molybdena phase
supported on niobia that is responsible for the ammoxidation
of 3-picoline to nicotinonitrile is located on a patchy monolayer
(9) Masuoka, Y.; Niwa, M.; Murakami, Y. J. Phys. Chem. 1990, 94,
1477.
(10) Zhang, W.; Desikan, A.; Oyama, S. T. J. Phys. Chem. 1995, 99,
14468.
(11) Chary, K. V. R.; Kumar, V. V.; Kanta Rao, P. Langmuir 1990, 6,
1549.
(12) Desikan, A. N.; Zhang, W.; Oyama, S. T. J. Catal. 1995, 157, 740.
(13) Miyata, H.; Mukai, T.; Ono, T.; Kubokawa, Y. J. Chem. Soc.,
Faraday Trans. 1 1988, 84, 4137.
(14) Matsuura, I.; Oda, H.; Hoshida, K. Catal. Today 1993, 16, 547.
(15) Quincy, B. R.; Houalla, M.; Proctor, A.; Hercules, D. M. J. Phys.
Chem. 1990, 94, 1520.
(16) Nag, N. K. J. Catal. 1985, 92, 432.
(17) Wang, X.; Zhao, B.; Jiang, D.; Xie, Y. Appl. Catal. A: Gen. 1999,
188, 201.
(18) Vit, Z.; Zdrazil, M. J. Catal. 1997, 171, 305.
(19) Oyama, S. T.; Zhang, W. J. Am. Chem. Soc. 1996, 118, 7173.
(20) Arena, F.; Parmaliana, A. J. Phys. Chem. 1996, 100, 19994.
(21) Kikutani, Y. J. Mol. Catal. A.: Chem. 1999, 142, 265 and references
therein.
(22) Miyata, H.; Mukai, T.; Ono, T.; Kubokawa, Y. J. Chem. Soc.,
Faraday Trans. 1988, 84, 4137.
(23) Machej, T.; Doumain, B.; Yasse, B.; Delmon, B. J. Chem. Soc.,
Faraday Trans. 1988, 84, 3905.
(24) Kim, D. S.; Segawa, K.; Soeya, T.; Wachs, I. E. J. Catal. 1992,
136, 539.
(25) Chen, K.; Xie, S.; Iglesia, E.; Bell, A. T. J. Catal. 2000, 189, 2.
(26) Calafat, A.; Avilan, L.; Aldana, J. Appl. Catal. A: Gen. 2000, 201,
215.
(27) Indovina, V. Catal. Today 1998, 41, 95.
(28) Tanabe, K. Catal. Today 1990, 8, 1.
(29) Tanabe, K.; Okazaki, S. Appl. Catal. A: Gen. 1995, 133, 191.
(30) Ross, J. R. H.; Smits, R. H. H.; Seshan, K. Catal. Today 1993, 16,
503.
(31) Huuhtanen, J.; Andersson, S. L. T. Appl. Catal. A: Gen. 1993, 98,
159.
(32) Smits, R. H. H.; Seshan, K.; Ross, J. R. H.; Kentgens, A. P. M. J.
Phys. Chem. 1995, 99, 9169.

Molybdenum Oxide Catalysts Supported on Niobia
J. Phys. Chem. B, Vol. 105, No. 19, 2001 4399
(33) Smits, R. H. H.; Seshan, K.; Ross, J. R. H.; Van den Oetelaar, L.
C. A.; Helwegen, J. H. J. M.; Anantharaman, M. R.; Brongersma, H. H. J.
Catal. 1995, 157, 584.
(34) Watling, T. C.; Deo, G.; Seshan, K.; Wachs, I. E.; Lercher, J. A.
Catal. Today 1996, 28, 139.
(35) Peeters, I.; Denier van der Gon, A. W.; Reijme, M. A.; Kooyman,
P. J.; de Jong, A. M.; van Grondelle, J.; Brongersma, H. H.; van Santen, R.
A. J. Catal. 1998, 173, 28.
(36) Chary, K. V. R.; Bhaskar, T.; Kishan, G.; Vijaya Kumar, V. J.
Phys. Chem. 1998, 102, 3936.
(37) Ko, E. I.; Weissman, J. G. Catal. Today 1990, 8, 27.
(38) Desikan, A. N.; Huang, L.; Oyama, S. T. J. Phys. Chem. 1991, 95,
10050.
(39) Parekh, B. S.; Weller, S. W. J. Catal. 1987, 40, 100.
(40) Lopez Agudo, A.; Gil-llambias, F. J.; Reyes, P.; Fierro, J. L. G.
Appl. Catal. 1981, 1, 59.
(41) Reddy, B. M.; Chary, K. V. R.; Rao, B. R.; Sunandana, C. S.;
Subrahmanyam, V. S.; Nag, N. K. Polyhedron 1996, 5, 191.
(42) Hall, W. K.; Millman, W. S. J. Catal. 1979, 59, 311.
(43) Hall, W. K.; Valyon, J. J. Catal. 1983, 84, 229.
(44) Rodrigo, L.; Marcinkowska, K.; Andot, A.; Roberge, P. C.;
Kaliaguine, S.; Stencel, J. M.; Makowsky, L. E.; Deihl, J. R. J. Phys. Chem.
1986, 90, 2690.
(50) Arnoldy, P.; de Jonge, J. C. M.; Mouljin, J. A. J. Phys. Chem.
1985, 89, 4517.
(51) Tauster, S. J.; Fung, S. C. J. Catal. 1978, 55, 29.
(52) Sankar, G.; Vasudevan, S.; Rao, C. N. R. J. Phys. Chem. 1988,
92, 1878.
(53) Weissman, J. G. Catal. Today 1996, 28, 159.
(54) Iizuka, T.; Ogasawara, K.; Tanabe, K. Bull. Soc. Chem. Jpn. 1983,
56, 2927.
(55) Chen, Z.; Iizuka, T.; Tanabe, K. Chem. Lett. 1984, 1085.
(56) Ogasawara, K.; Iizuka, T.; Tanabe, K. Chem. Lett. 1984, 645.
(57) Iizuka, T.; Fujie, S.; Ushikubo, T.; Chen, Z.; Tanabe, K. Appl. Catal.
1986, 28, 1.
(58) Okazaki, S.; Kurimata, M.; Iizuka, T.; Tanabe, K. Bull. Chem. Soc.
Jpn. 1987, 60, 37.
(59) Kurosaki, A.; Okuyama, T.; Okazaki, S. Bull. Chem. Soc. Jpn. 1987,
60, 3541.
(60) Tanabe, K.; Misono, N.; Ono, Y.; Hattori, H. Stud. Surf. Sci. Catal.
1989, 51, 61.
(61) Andersson, A. J. Catal. 1982, 76, 144.
(62) Andersson, A.; Lundin, S. T. J. Catal. 1980, 65, 9.
(63) Vejux, A.; Courtine, J. J. Solid State Chem. 1978, 23, 93.
(64) Busca, G.; Cavani, F.; Trifiro, F. J. Catal. 1987, 106, 471.
(65) Chary, K. V. R.; Kishan, G.; Bhaskar, T.; Sivaraj, Ch. J. Phys.
Chem. 1998, 102, 6792.
(66) Chary, K. V. R.; Kishan, G.; Bhaskar, T. J. Chem Soc., Chem.
Commun. 1999, 1339.
(67) Chary, K. V. R.; Bhaskar, T.; Kalyana Seela, K.; Srilakshmi, K.;
(45) Desikan, A. N.; Huang, L.; Oyama, S. T. J. Chem. Soc., Faraday.
Trans. 1992, 88, 3357.
(46) Gambaro, L. A.; Fierro, J. L. G. React. Kinet. Catal. Lett. 1981,
18, 495.
(47) Van Hengstum, A. J.; Van Ommen, J. G.; Bosch, H.; Gellings, P.
J. Appl. Catal. 1983, 5, 207.
(48) Oliver, S. W.; Smith, T. D.; Pilbrow, J. R.; Pratt, K. C.; Christov,
V. J. Catal. 1988, 111, 88.
Rajender Reddy, K. Appl. Catal. A: Gen. 2001 (in press).
(49) Thomas, R.; de Beer, V. H. J.; Moulijn, J. A. Bull. Chim. Belg.
1981, 90.
(68) Bhaskar, T.; Rajender Reddy, K.; Praveen Kumar, C.; Murthy, M.
R. V. S.; Chary, K. V. R. Appl. Catal. A: Gen. 2001(in press).