Determination of the chemical nature of active surface sites present on bulk mixed metal oxide catalysts

Article abstract of DOI:10.1021/jp048839e

The CH₃OH temperature programmed surface reaction (TPSR) spectroscopy was employed to determine the chemical nature of active surface sites for bulk mixed metal oxide catalysts. The CH₃OH TPSR experiments provides fundamental surface

Full text of DOI:10.1021/jp048839e

J. Phys. Chem. B 2005, 109, 2275-2284
2275
Determination of the Chemical Nature of Active Surface Sites Present on Bulk Mixed Metal
Oxide Catalysts†
Israel E. Wachs,*,‡ Jih-Mirn Jehng,§ and Wataru Ueda|
Operando Molecular Spectroscopy and Catalysis Laboratory, Department of Chemical Engineering,
Lehigh UniVersity, Bethlehem, PennsylVania 18015, Department of Chemical Engineering, National Chung
Hsing UniVersity, Taichung 402, Republic of China, and Catalysis Research Center, Hokkaido UniVersity,
N-11, S-10, Kita-ku, Sapporo 060-0811, Japan
ReceiVed: March 15, 2004; In Final Form: June 17, 2004
CH3OH temperature programmed surface reaction (TPSR) spectroscopy was employed to determine the
chemical nature of active surface sites for bulk mixed metal oxide catalysts. The CH3OH-TPSR spectra peak
temperature, Tp, for model supported metal oxides and bulk, pure metal oxides was found to be sensitive to
the specific surface metal oxide as well as its oxidation state. The catalytic activity of the surface metal oxide
sites was found to decrease upon reduction of these sites and the most active surface sites were the fully
oxidized surface cations. The surface V5+ sites were found to be more active than the surface Mo6+ sites,
which in turn were significantly more active than the surface Nb5+ and Te4+ sites. Furthermore, the reaction
products formed also reflected the chemical nature of surface active sites. Surface redox sites are able to
liberate oxygen and yield H2CO, while surface acidic sites are not able to liberate oxygen, contain either H+
or oxygen vacancies, and produce CH3OCH3. Surface V5+, Mo6+, and Te4+ sites behave as redox sites, and
surface Nb5+ sites are Lewis acid sites. This experimental information was used to determine the chemical
nature of the different surface cations in bulk Mo-V-Te-Nb-Ox mixed oxide catalysts (Mo0.6V1.5Ox,
Mo1.0V0.5Te0.16Ox, Mo1.0V0.3Te0.16Nb0.12Ox). The bulk Mo0.6V1.5Ox and Mo1.0V0.5Te0.16Ox mixed oxide catalytic
characteristics were dominated by the catalytic properties of the surface V5+ redox sites. The surface enrichment
of these bulk mixed oxide by surface V5+ is related to its high mobility, V5+ possesses the lowest Tammann
temperature among the different oxide cations, and the lower surface free energy associated with the surface
termination of VdO bonds. The quaternary bulk Mo1.0V0.3Te0.16Nb0.12Ox mixed oxide possessed both surface
redox and acidic sites. The surface redox sites reflect the characteristics of surface V5+ and the surface acidic
sites reflect the properties normally associated with supported Mo6+. The major roles of Nb5+ and Te4+ appear
to be that of ligand promoters for the more active surface V and Mo sites. These reactivity trends for CH3OH
ODH parallel the reactivity trends of propane ODH because of their similar rate-determining step involving
cleavage of a C-H bond. This novel CH3OH-TPSR spectroscopic method is a universal method that has also
been successfully applied to other bulk mixed metal oxide systems to determine the chemical nature of the
active surface sites.
Introduction
supported metal powders with large specific surface areas. This
led to the concept of structure-sensitive and structure-insensitive
catalytic reactions.1 Structure-insensitive reactions exhibit the
same TOF values for all metal crystal sizes and are independent
of specific crystallographic orientations. Structure-sensitive
reactions, however, exhibit TOF values that are dependent on
the metal crystal size and the specific crystallographic orienta-
tion. Furthermore, it was found that structure-insensitive catalytic
reactions are not strongly affected by the nature of the metal,
alloying, and poisoning and can be described by uniform surface
kinetics. In contrast, structure-sensitive catalytic reactions are
very strongly affected by the specific metal, alloying, and
poisoning and cannot be described by uniform surface kinetics.
Only kinetics based on nonuniform surface Temkin formalism
is able to satisfactorily describe structure-sensitive catalytic
reactions (e.g., ammonia synthesis from N2 and H2 over metallic
Fe catalysts).
The foundation for a revolution in metal catalysis science
took place during the 1960s and 1970s when CO, H2, O2, and
N2O selective chemisorption methods were developed to
quantitatively determine the number of exposed metal atoms
in metal catalysts.1 The seed for this development was the
pioneering studies by Emmett and Brunauer when they em-
ployed selective CO chemisorption to determine the number of
exposed metallic iron atoms and CO2 chemisorption to deter-
mine the number of exposed basic potassium oxide promoter
atoms on the surface of ammonia synthesis catalysts.2 Knowl-
edge of the number of surface metal atoms allowed, for the
first time, the systematic use of turnover frequency (TOF, the
number of molecules reacted per surface metal atom per second)
for comparison of catalytic reactions between different inves-
tigators as well as between large metallic single crystals and
† Part of the special issue “Michel Boudart Festschrift”.
‡ Lehigh University.
Djega-Mariadassou et al. stated in the early 1980s that the
progress made in the case of metal catalysts should also follow
for metal oxides,3 in particular, comparisons between the
§ National Chung Hsing University.
| Hokkaido University.
10.1021/jp048839e CCC: $30.25 © 2005 American Chemical Society
Published on Web 08/28/2004

2276 J. Phys. Chem. B, Vol. 109, No. 6, 2005
Wachs et al.
catalytic activity of large single crystals and powders with large
specific surface areas. This development, however, has been
slow in coming over the past two decades because satisfactory
selective chemisorption techniques for metal oxides have not
been developed and adopted by the catalysis researcher com-
munity. Djega-Mariadassou et al.3 and Sleight et al.4 indepen-
dently demonstrated in the 1980s that alcohols are feasible for
selective chemisorption studies of metal oxides with surfaces
of preferred orientation. However, these initial selective chemi-
of the Nb2O5 support. The solutions were slowly precipitated
by controlling the solution pH with ammonium hydroxide, dried,
and finally calcined in air at 400 °C for 1 h. Raman character-
ization confirmed that the expected crystal phases were formed.
The resultant BET surface areas were ∼3-5 m2/g.
The bulk mixed metal oxides were prepared by hydrothermal
synthesis. The starting materials were ammonium heptamolyb-
date, vanadyl sulfate, tellurium hydroxide, and niobium oxalate.
These salts were used to make the initial aqueous solution slurry.
The slurry was introduced into a stainless steel autoclave
equipped with a Teflon inner tube. The hydrothermal reaction
was carried out at 175 °C for different lengths of time for each
combination of oxides: binary Mo-V-O for 20 h, tertiary
Mo-V-Te-O for 48 h, and quaternary Mo-V-Te-Nb-O
for 48 h. The resulting solids were separated by filtration,
washed with distilled water, dried in air at 80 °C overnight,
and calcined in a flowing N2 stream at 600 °C for 2 h. The
binary M-V-O mixed oxide was calcined at 500 °C in N2
and the tertiary Mo-V-Te-O was precalcined in air at 280
°C prior to the high-temperature calcinations in N2. The resulting
bulk mixed metal oxide catalysts exhibited the orthorhombic
X-ray diffraction (XRD) structure and comparable BET surface
areas of 6-7 m2/g.
4
sorption studies on ZnO,3 MoO3, 4 and Fe2(MoO4)3 powders
were not successfully extended to other metal oxide catalytic
systems until very recently.5-8 A major hurdle in comparing
large single crystal metal oxides and metal oxide powders with
high surface areas is that the surfaces of the large single crystals
are typically prepared in ultrahigh vacuum (UHV) and, conse-
quently, do not form the same surface structures and function-
alities as metal oxide powders. For example, metal oxide
surfaces prepared in a vacuum do not terminate in M-OH
functionalities and many times cannot easily be fully oxidized
at lower O2 pressures typically employed.10,11 Metal oxides are
also much more complex catalytic materials than metals because
of their variable oxidation states, variable coordination for each
oxidation state, distribution of surface redox, basic and acidic
sites, distribution among surface Lewis and Bronsted acidic sites,
and participation of surface and bulk lattice oxygen atoms in
catalytic oxidation reactions. Thus, this multidimensional issue
for metal oxides requires not only selective chemisorption
methods that provide information about the number of active
surface sites but also information that is very rich in the chemical
nature of the active sites. The present study demonstrates how
this has been achieved for bulk Mo-V-Nb-Te-O mixed
metal oxides by employing CH3OH temperature programmed
surface reaction (TPSR) spectroscopy.
Raman Spectroscopy. The in situ Raman spectrometer
system consists of a quartz cell with a sample holder (Hastalloy
C), a triple grating spectrometer (Spex, model 1877), a CCD
detector (Jobin Yvon-Spex, ISA Inc., model Spectrum-1), and
an argon ion laser (Spectra Physics, model 165). The sample
typically contained 100-200 mg and was pressed into a disk
shape so that it could be mounted in the sample holder. The
sample holder was mounted onto a ceramic shaft that was rotated
at 1000-2000 rpm by a 115 V dc motor. A cylindrical heating
coil controlled the quartz cell temperature that was monitored
by a thermocouple placed next to the catalyst sample inside
the cell. The cell was capable of being heated to 600 °C, and
gas flow rates of ∼100 cm3/min at atmospheric pressure were
typically employed.
Experimental Section
Catalyst Preparation. The Nb2O5 (BET ) 57 m2/g) was
synthesized by air calcination of niobium oxide hydrate (Nb2O5‚
nH2O from CBMM, 99.9% purity) at 500 °C for 10 h. The
resulting Nb2O5 was also employed as an oxide support for
vanadia, molybdena, telluria, and their combinations. The
supported V2O5/Nb2O5 catalysts were prepared by incipient-
wetness impregnation of the Nb2O5 support with 2-propanol
solutions of vanadium isopropoxide (Alfa Aesar, 97% purity).
The preparation was formed inside a glovebox with continuously
flowing N2. After impregnation, the samples were kept inside
the glovebox with flowing N2 overnight. The samples were
further dried in flowing N2 at 120 °C for 1 h and calcined in
flowing air at 450 °C for 2 h. The supported MoO3/Nb2O5 and
TeO2/Nb2O5 samples were prepared by the incipient-wetness
impregnation of the Nb2O5 support with aqueous solutions of
ammonium heptamolybdate (Matheson, Coleman & Bell, 99.9&
purity) and telluric acid (H2TeO4‚2H2O, Alfa Aesar), respec-
tively. After impregnation, the samples were dried under ambient
conditions for 48 h and calcined in air at 450 °C for 2 h.
Combinations of these metal oxides on Nb2O5 were also
prepared sequentially following the above syntheses procedures.
The impregnation sequence procedure is reflected in the notation
BOx/AOx/Nb2O5, where the A component was first impregnated
onto the Nb2O5 support, dried, and calcined and then the BOx
component was impregnated onto the calcined AOx/Nb2O5
catalyst, dried, and calcined. The notation of BOx/AOx/Nb2O5
was used to describe such double impregnated samples.
CH3OH-TPSR Spectroscopy. The CH3OH-TPSR spectros-
copy studies were performed with the AMI-100 temperature
programmed system (Zeton-Altamira Instruments) equipped
with an online mass spectrometer (Dycor DyMaxion, Ametek
Process Instruments). Typically, about 25-100 mg of catalyst
was loaded in a U-type quartz tube and pretreated at elevated
temperatures to dehydrate the sample. The Nb2O5 and the niobia-
supported metal oxides were pretreated in flowing dry air at
450 °C for 1 h. The bulk, pure oxides were pretreated in flowing
dry air at 350 °C for 1 h. The bulk mixed metal oxides were
treated in flowing He at 350 °C for 1 h to prevent their full
oxidation. The supported and bulk, pure metal oxides were
subsequently cooled in flowing O2/He to ∼200 °C and then
switched to a flowing He stream to flush out any residual gas-
phase O2. In the case of the bulk mixed metal oxide catalysts,
the catalysts were cooled in a flowing He stream. For those
experiments where the effect of CH3OH-TPSR reduction was
also examined, all the catalysts were cooled in flowing He so
the reduced surface sites would not become oxidized upon
cooling (referred to as CH3OH-TPSR cycles in the paper). The
metal oxide samples were all further cooled in He inert en-
vironments to the methanol adsorption temperature of ∼50 °C.
Methanol adsorption was achieved by flowing 2000 ppm CH3-
OH/He for 30 min, and the system was then purged with flowing
He for another 30 min. Afterward, the catalysts were heated at
a constant heating rate of 10 °C/min to 400-500 °C in flowing
He (30 mL/min). The gases exiting from the quartz tube reactor
The bulk, pure metal oxides V2O5, MoO3, and TeO2 were
prepared the same way as described above, but in the absence

Active Sites on Mixed Metal Oxide Catalysts
J. Phys. Chem. B, Vol. 109, No. 6, 2005 2277
were analyzed with an online mass spectrometer, which was
linked via a capillary tube, as a function of catalyst temperature.
The following m/e ratios were employed for the identification
of the various gases: CH3OH, m/e ) 31; H2CO, m/e ) 30;
CH3OCH3, m/e ) 45 and 15; H2O, m/e ) 18; H3COOCH,
m/e ) 60; (CH3O)2CH2, m/e ) 75. The formation of CO2/CO
is not reported in the current studies because these reaction
products primarily originate from reasdsorption of formaldehyde
(via formation of surface formate species) or adsorption of
background carbonaceous molecules during the CHOH-TPSR
studies. Furthermore, formation of CO2/CO was essentially not
observed at low conversions during methanol oxidation steady-
state studies with these catalytic materials. The H2CO and
(H3CO)2CH2 mass spectrometer signals were corrected for the
minor contributions from the cracking of CH3OH in the mass
spectrometer.
spectroscopy. For the bulk metal oxides, the number of active
surface sites was determined by methanol chemisorption and
that value was multiplied by a factor of 3 to account for the
known ratio of CH3OH/M ∼ 3 due to lateral interactions in the
chemisorbed methanol monolayer.5-8,11,12 The ratio of CH3-
OH/M ∼ 3 was obtained with model supported metal oxide
systems containing monolayer surface coverage, and it was
assumed that the same stoichiometry also holds for bulk metal
oxides. The steady-state methanol oxidation kinetic expression
at low conversions and over an oxidized metal oxide surface
has been shown to be13
TOF ) krdsKads(PCH OH)1(PO )0 (s-1)
(2)
3
2
where krds is the first-order Arrhenius rate constant for the rate-
determining-step, Kads is the methanol equilibrium adsorption
Recent studies by Vohs et al. demonstrated that the same Tp,
maximum peak temperature, values can be obtained from model
single crystal and powdered supported metal oxide catalytic
systems because a significant portion of the intermediate reaction
products (e.g., H2CO) are able to leave the powdered bed
without becoming trapped on the catalyst surface and further
oxidized to CO2.11,12 Furthermore, these very significant studies
also demonstrated that the Redhead equation,1,11,12 typically
employed to obtain kinetic parameters from single-crystal
temperature-programmed studies in ultrahigh vacuum (UHV),
can also be applied to obtain quantitative kinetic parameters
from the powdered samples during the CH3OH-TPSR studies
constant, and PCH OH is the methanol partial pressure in the gas
3
phase. The reaction is zero-order in oxygen partial pressure in
the gas phase because the catalyst surfaces are fully oxidized
under reaction conditions.9 The value for the first-order surface
rate constant, krds, is independently obtained from the CH3OH-
TPSR experiments. This independent kinetic information allows
for direct determination of Kads from the steady-state methanol
oxidation TOF value, eq 2, as shown below
Kads ) TOF(PCH OH)1/krds (L/mol)
(3)
3
Furthermore, the methanol equilibrium heat of adsorption on
the catalyst, ∆Hads, can also be determined from the following
relationship
2
Ea/(RTp ) ) V/B exp(-Ea/(RTp))
(1)
where R is the gas-phase constant, B is the heating rate employed
in the temperature programmed study, Ea is the activation energy
for the surface reaction, and V is the first-order Arrhenius rate
constant pre-exponential factor. Many previous CH3OH-TPSR
studies with single crystals have shown that formation of H2-
CO from CH3OH proceeds via a first-order reaction kinetics
and a V that is typically very close to ∼1013 s-1. Consequently,
this value for V was selected to determine the activation energy,
Ea, for the first-order surface reactions of methanol from the
Redhead equation (eq 1).
∆Hads ) Eapp - Ea (kcal/mol)
(4)
where Ea is the activation energy for the first-order surface
reaction obtained from the CH3OH-TPSR experiments (see eq
1 above) and Eapp is the apparent activation energy for steady-
state methanol oxidation.14 Previous studies over many metal
oxide catalysts revealed that the Eapp for methanol oxidation to
formaldehyde was typically ∼20 kcal/mol,13,14 and consequently,
this value for Eapp was employed to determine the methanol
equilibrium heat of adsorption on the metal oxide surface.
Steady-State Methanol Oxidation. The steady-state metha-
nol oxidation studies were conducted in an isothermal fixed-
bed reactor operating under differential reaction conditions
(methanol conversions less than 10%). The reactor consists of
a 3.8 mm glass tube, and the catalyst powder, ∼25-100 mg,
was held in place by glass wool (above and below the catalyst).
The catalysts were pretreated in flowing O2 at 300 °C prior to
the methanol oxidation for the bulk, pure, and supported metal
oxide catalysts described above. The bulk mixed metal oxide
catalysts were not examined for methanol oxidation since
preliminary studies demonstrated that TeOx was volatile under
the oxidizing reaction conditions employed for methanol oxida-
tion (methanol/O2/He mixture of 6/13/81). Analysis of the
reaction products was performed with an online gas chromato-
graph (HP-5840A) containing two packed columns (Porapak R
and Carbosieve SII) and two detectors (thermal conductivity
detector and flame ionization detector). The steady-state reaction
temperature was maintained at 230 °C.
Results
CH3OH-TPSR Studies with Model Supported Metal
Oxide Catalysts. Before the studies with the bulk Mo-V-
Nb-Te-O mixed metal oxide catalytic materials were initiated,
CH3OH-TPSR studies with model supported metal oxide
catalysts were undertaken because these model systems contain
the supported active metal oxide components as a two-
dimensional metal oxide monolayer (100% dispersion) and their
surface compositions can be controlled. The model supported
metal oxide catalysts investigated consisted of molybdena,
vanadia, telluria, and their combinations supported on Nb2O5.
Niobia was selected as the support since it is the only oxide
present in the mixed Mo-V-Nb-Te-O metal oxide system
that possesses high surface areas after calcinations at 450 °C.
Raman characterization of the model niobia supported metal
oxides demonstrated that the 6% MoO3/Nb2O5, 7% V2O5/Nb2O5,
5% TeO2/Nb2O5, 1% V2O5/6% MoO3/Nb2O5, and 1% TeO2/
6% MoO3/Nb2O5 catalysts did not possess crystalline phases
and contained monolayer surface coverage of these two-
dimensional surface metal oxide species.15
The steady-state methanol oxidation catalytic data were
expressed in terms of turnover frequency, TOF, by normalizing
the reaction rate per surface metal oxide site. For the supported
metal oxides, the reaction rate was normalized per number of
active surface sites of the supported component, which assumed
100% dispersion and was independently confirmed with Raman
The CH3OH-TPSR spectrum for the Nb2O5 support is
presented in Figure 1. The formation of dimethyl ether (DME,

2278 J. Phys. Chem. B, Vol. 109, No. 6, 2005
Wachs et al.
Figure 3. CH3OH-TPSR from supported 5% TeO2/Nb2O5.
Figure 1. CH3OH-TPSR from bulk Nb2O5.
temperature shifts from 189 °C toward ∼211 °C as the surface
Mo6+ sites become reduced to lower oxidation states (formation
of formaldehyde is accompanied by extraction of one oxygen
atom from the surface MoOx species to make H2O). The shift
in Tp to higher temperatures is reflected in an order of magnitude
decrease in the Arrhenius krds rate constant. In addition, the
amount of H2CO formed continuously decreases and the amount
of CH3OH formation continuously increases. The continuous
decrease in the formation of H2CO with repeated CH3OH-TPSR
cycles reveals that reduced surface MoOx species do not
dissociatively chemisorb methanol as efficiently as the fully
oxidized surface Mo6+ redox sites. Furthermore, no DME is
formed at ∼300 °C, which indicates that all the surface Nb
acidic sites of the Nb2O5 support have been titrated by the
surface molybdena species.
The CH3OH-TPSR experiments for the model supported 7%
V2O5/Nb2O5 catalyst give the same spectra as that for the
corresponding model supported 6% MoO3/Nb2O5 catalyst, with
the exception of slightly different Tp values and ratio of H2-
CO/CH3OH. The formation of H2CO on the redox surface
vanadia sites occurs at ∼177 °C rather than ∼189 °C over the
redox surface molybdena sites, reflecting the higher activity of
the surface vanadia sites than the surface molybdena sites on
Nb2O5. Another difference between surface vanadia and mo-
lybdena species is the much higher production of H2CO and
lower production of CH3OH from the 7% V2O5/Nb2O5 catalyst.
These observations suggest that the methanol equilibrium
adsorption constant, Kads, is probably higher for surface vanadia
than surface molybdena species. The H2CO TPSR peaks also
shift to higher temperatures as the surface vanadia sites are
reduced by repeated cycles of CH3OH-TPSR, but its reduction
is more sluggish. Vohs et al. nicely demonstrated by combining
CH3OH-TPSR and XPS analysis, on single crystals of TiO2 and
CeO2 containing a surface vanadia monolayer, that H2CO
formation occurs with a Tp ∼100 °C lower from surface V5+
sites than on surface V3+ sites.11,12 This translates to at least 2
orders of magnitude in activity between the krds values of surface
Figure 2. CH3OH-TPSR from supported 6% MoO3/Nb2O5.
CH3OCH3) with a Tp ∼ 300 °C originates from surface methoxy
reactions over surface acidic sites. These are weak acidic sites
since pyridine adsorption IR measurements did not detect any
surface acidic sites on the calcined Nb2O5.16 Desorption of CH3-
OH with a Tp ∼ 150 °C is primarily due to molecularly adsorbed
methanol and is also obtained when the adsorption temperature
is further decreased toward room temperature. The high-
temperature tail of the CH3OH curve originates from surface
methoxy (*OCH3) species that become rehydrogenated by
surface hydroxyls to form CH3OH during the TPSR experiment.
No formation of H2CO from surface redox sites is observed.17
Thus, the Nb2O5 support contains weak surface Lewis acid sites
that convert the surface methoxy (*OCH3) species to DME and
also back to CH3OH. The elementary steps of the DME
formation reaction are not known in the literature, but isotopic
experiments with deuterated methanol suggest that the rate-
determining-step (rds) does not involve breaking of C-H or
O-H bonds and, thus, involves first-order scission of the C-O
bond of the surface methoxy intermediate.17
V
5+ and V3+ sites. The absence of DME formation reveals that
The CH3OH-TPSR spectra for the model supported 6%
MoO3/Nb2O5 catalyst are presented in Figure 2. The chemi-
sorbed surface methoxy species yields H2CO with a Tp of ∼189
°C and CH3OH with a Tp of ∼135 °C over surface redox sites.
Deuterium isotopic labeling studies demonstrated that the first-
order rds involves breaking of the methyl C-H bond of the
surface methoxy intermediate to form H2CO and CH3OH.17 The
reaction products of DME, DMM, and MF are not observed
during this CH3OH-TPSR experiment. As the CH3OH-TPSR
cycles are repeated without reoxidizing the surface MoOx
species (not shown for brevity), the H2CO formation peak
the surface vanadia species cover the surface acidic sites of the
Nb2O5 support.
The CH3OH-TPSR spectrum for the model supported 5%
TeO2/Nb2O5 catalyst is shown in Figure 3. The selective
oxidation reaction products from this oxidized surface are H2-
CO (Tp ∼ 263 °C) and CH3OH (Tp ∼ 144 and ∼272 °C). The
simultaneous formation of H2CO and CH3OH at 263-272 °C
reflects the redox nature of the surface TeOx species on the
Nb2O5 support.17 Desorption of CH3OH with Tp ∼ 144 °C
originates from molecularly adsorbed methanol. The absence

Active Sites on Mixed Metal Oxide Catalysts
J. Phys. Chem. B, Vol. 109, No. 6, 2005 2279
Figure 4. CH3OH-TPSR from supported 1% V2O5/6% MoO3/Nb2O5.
Figure 5. CH3OH-TPSR from supported 1% TeO2/6% MoO3/Nb2O5.
of DME formation demonstrates that the surface acidic sites
on the Nb2O5 support are also titrated by the surface telluria
species. The surface telluria species possess redox characteristics
that exhibit a significantly lower activity relative to the other
surface redox sites (V5+ > Mo6+ . Te4+).
species. From pyridine-IR chemisorption studies, however, it
was found that the calcined Nb2O5 did not possess surface Lewis
or Bronsted acid sites as probed by pyridine.16 It appears that
methanol is able to sense some weak surface acidic sites on the
Nb2O5 since it exclusively forms DME on this metal oxide
surface.
The above CH3OH-TPSR studies demonstrate that the
individual oxides among Mo, V, Te, and Nb give rise to different
reaction products, ratios of product to methanol, and specific
Tp values and that the Tp values are also dependent on the
specific oxidation states. These experiments, however, do not
provide any information about what happens when multiple
metal oxides are simultaneously present on the surface. To
address this issue, studies with supported mixed metal oxide
monolayer catalysts were also undertaken. The CH3OH-TPSR
spectrum from the model supported 1% V2O5/6% MoO3/Nb2O5
catalyst is shown in Figure 4. Interestingly, the H2CO production
at Tp ∼ 183 °C is somewhat lower than that originally observed
for 6% MoO3/Nb2O5 due to the addition of 1% V2O5 to the
surface molybdena monolayer on niobia. This reveals that the
surface chemistry of this mixed metal oxide monolayer is being
dominated by the surface V5+ species and that the surface
methoxy species are able to surface diffuse toward the surface
V5+ sites during the TPSR experiment. The rapid surface
diffusion of Mo-OCH3 and V-OCH3 was demonstrated in
previous studies.18 Furthermore, the same Tp value for 5% V2O5/
Nb2O5 and 1% V2O5/6% MoO3/Nb2O5 also reveals that the
kinetics for methoxy decomposition of the surface vanadia
species on Nb2O5 are not influenced by the presence of surface
molybdena species. The increase in reactivity is also reflected
in the much higher production of H2CO/CH3OH in the presence
of surface vanadia. To examine the kinetic behavior of the
introduction of a cation with lower activity, the model supported
1% TeO2/6% MoO3/Nb2O5 catalyst system was also investigated
with CH3OH-TPSR. The resultant CH3OH-TPSR spectrum
shown in Figure 5 is almost indistinguishable from the previous
results for the surface Mo6+ sites on the Nb2O5 support (reflected
in the same Tp values and ratio of H2CO/CH3OH). These
findings again reflect the surface mobility of the surface methoxy
intermediates and their preferred reaction pathway to H2CO
The CH3OH-TPSR spectra from bulk V2O5 are depicted in
Figure 6 as a function of several TPSR cycles without exposure
to gas-phase oxygen between each cycle. The formation of H2-
CO from the surface methoxy species always occurs with a Tp
∼ 188 °C, and the peak temperature does not change position
with each subsequent reducing cycle. The relatively constant
Tp of ∼188 °C suggests that primarily active surface V5+ redox
sites are present on the bulk V2O5 powders. The ability to
maintain the surface V5+ sites fully oxidized implies that oxygen
anions are diffusing from the bulk lattice to the surface to
reoxidize the surface V cations reduced by the CH3OH-TPSR
experiment (Mars-van Krevelen mechanism). The decrease in
the H2CO formation is due to some sintering of this thermally
fragile material during the TPSR experiment. The decrease in
the H2CO/CH3OH ratio with each CH3OH-TPSR cycle reveals
that some surface reduced VOx sites are also present and that
these reduced sites do not chemisorb methanol as readily as
the more active surface V5+ redox sites. No DME formation
was observed reflecting the absence of active surface acidic sites
for this sample under the current experimental conditions.17
The CH3OH-TPSR spectra from bulk MoO3 are shown in
Figure 7 as a function of CH3OH-TPSR cycles. Unlike bulk
V2O5 reduction cycles, the H2CO Tp for MoO3 initially occurs
at ∼196 °C and continuously increases toward 221 °C with each
reduction cycle. This change is analogous to the reduction of
surface V5+ to V4+/V3+ observed by Vohs et al.11,12 and reflects
the reduction of the surface Mo species with each cycle. The
ease of reduction of the surface Mo6+ sites to lower oxidation
states with each cycle reflects the inability of bulk lattice oxygen
of MoO3 to diffuse to the surface and fully reoxidize the reduced
Mo cations under the current experimental conditions. The
reduction in the amount of H2CO formed with each cycle is
due to sintering of this thermally fragile metal oxide and the
creation of surface reduced MoOx sites that decrease methanol
chemisorption on the defective MoO3 surface. The decrease in
the H2CO/CH3OH ratio with each cycle reflects the lower
methanol Kads on the partially reduced molybdena surface. The
significant shifts in Tp translate to about an order of magnitude
decrease in krds for the conversion of surface methoxy to H2CO
on MoO3. The lack of DME formation reveals the absence of
surface acidic sites on the surface of MoO3 for the present
catalyst and experimental conditions.
formation via the most active surface sites (V5+ > Mo6+
.
Te4+).
The CH3OH-TPSR results above for the Nb2O5 supported
metal oxide catalysts are collected in Table 1 for comparison:
Tp, krds, Kads, Ea, and H2CO/CH3OH, DME/CH3OH, or CO2/
CH3OH ratios.
CH3OH-TPSR Studies with Bulk, Pure Metal Oxides. The
CH3OH-TPSR spectrum for bulk Nb2O5 was already presented
in Figure 1 and revealed that only surface acidic sites are present
on this metal oxide and form DME from the surface methoxy

2280 J. Phys. Chem. B, Vol. 109, No. 6, 2005
Wachs et al.
TABLE 1: Quantitative Determination of Parameters from CH3OH-TPSR and Steady-State Methanol Oxidation Experiments.
rel HCHO
production
k
rds (s-1) @
TOF (s-1
)
with the
adsorption HCHO/
cycle
230 °C for
HCHO- not
Nb2O5
×10-3 at
Tp (°C)
of HCHO of DME
Tp (°C)
DME/
CH3OH CH3OH
Kads
Ea
catalyst
230 °C
(L/mol)
(kcal/mol)
6% MoO3/Nb2O5 (1st)
6% MoO3/Nb2O5 (2nd)
6% MoO3/Nb2O5 (3rd)
6% MoO3/Nb2O5 (4th)
7% V2O5/Nb2O5
5% TeO2/Nb2O5
1% V2O5/6% MoO3/Nb2O5
1% TeO2/6% MoO3/Nb2O5
V2O5 (first)
V2O5 (second)
V2O5 (third)
MoO3 (first)
MoO3 (second)
MoO3 (third)
TeO2
Nb2O5
Mo0.6V1.5Ox
Mo1.0V0.5Te0.16Ox
Mo1.0V0.3Te0.16Nb0.12Ox
18
85
189
195
210
211
177
263
183
189
188
188
188
196
213
221
432
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
299
*
*
100.0
92.3
63.3
50.1
*
*
*
1.02
0.84
0.69
0.57
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
0.20
0.13
0.05
0.04
0.6 × 102
31.5
31.9
33.0
33.1
30.7
36.7
31.1
31.5
31.5
31.5
31.5
32.0
33.2
33.8
48.7
39.3
30.4
30.5
30.4
*
*
*
0.47
1.2 × 102
0.20
1.39
0.89
3.73
1.45
1.58
1.26
1.00
0.91
0.10
0.00
0.45
0.82
0.39
1.1 × 10-3
*
*
*
0.31
*
0.20
0.22
0.22
0.22
0.12
0.04
0.02
87
18
100.0
91.5
69.2
100.0
74.6
69.3
*
*
*
*
*
273.7
*
*
99.3
*
*
0.6
7
0.72 × 10-8
8 × 10-5
0.62
0.54
0.62
0.6 × 108
0.493
*
*
5.4 × 104
173
175
173
*
*
*
146
0.24
a The TOF and Kads values were obtained from steady-state methanol oxidation studies and the remaining parameters were obtained from CH3OH-
TPSR studies.
Figure 8. CH3OH-TPSR from bulk TeO2.
Figure 6. CH3OH-TPSR from bulk V2O5, several cycles.
surface telluria species. Such a dramatic decreases in the Tp
corresponds to an increase in krds of more than 5 orders of
magnitude. A closer examination of the bulk and supported
vanadia and molybdena catalysts also shows that the Tp values
appear to slightly decrease when these metal oxides are
deposited on the Nb2O5 support (e.g., V5+ decreases from ∼188
to 177 °C and Mo6+ decreases from 196 to 189 °C). The absence
of surface acidic sites is reflected in the CH3OH-TPSR product
yield containing no DME.
The CH3OH-TPSR results for the bulk, pure Nb2O5, V2O5,
MoO3, and TeO2 metal oxides are presented in Table 1 for
comparison: Tp, krds, Kads, Ea, and H2CO/CH3OH or DME/CH3-
OH ratios.
CH3OH-TPSR Studies with Bulk Mixed Metal Oxides.
Prior to undertaking the CH3OH-TPSR experiments, the Raman
spectra of the mixed metal oxides were examined for phase
purity and structural changes. The Raman spectra of the bulk
mixed metal oxides are compared against the reference MoO3
and V2O5 crystalline Raman spectra in Figure 9. Comparison
of the Raman spectra of bulk MoO3 and V2O5 with that of bulk
Figure 7. CH3OH-TPSR from bulk MoO3, several cycles
Unlike the fully oxidized bulk and model supported V2O5
and MoO3 catalysts that gave similar Tp temperatures for H2-
CO formation, the Tp values for bulk TeO2 and supported 5%
TeO2/Nb2O5 varied by ∼170 °C. This dramatic decrease in Tp
temperatures for the formation of these reaction products reveals
that the niobia support is promoting the catalytic activity of the
M0.6V1.5Ox reveals that the binary mixed metal oxide phase is
monophasic, does not contain separate small V2O5 or MoO3
crystallites, and has a structure uniquely different from its two

Active Sites on Mixed Metal Oxide Catalysts
J. Phys. Chem. B, Vol. 109, No. 6, 2005 2281
Figure 9. Raman spectra of bulk Mo0.6V1.5Ox, Mo1.0V0.5Te0.16Ox,
Mo1.0V0.3Te0.16Nb0.12Ox, MoO3, and V2O5 metal oxides.
Figure 10. CH3OH-TPSR from bulk Mo0.6V1.5Ox.
pure metal oxide components. The Raman spectrum of bulk
Mo0.6V1.5Ox appears to be similar to a distorted MoO3 structure
since it exhibits a strong band at ∼865 cm-1, rather than 812
cm-1, and a much weaker band at ∼980 cm-1, rather than 993
cm-1. The bulk Mo0.6V1.5Ox Raman spectrum is dominated by
the bulk vibrations of molybdena rather than vanadia because
of the much higher Raman scattering cross section of molyb-
denum oxides than vanadium oxides. By analogy with the well-
known vibrations of crystalline MoO3, the band at ∼865 cm-1
is assigned to Mo-O-Mo and Mo-O-V vibrations and the
band at ∼980 is assigned to ModO and/or VdO vibrations.
The decrease in the relative intensity of the ∼980 cm-1 band
with increasing Mo/V ratio suggests that this band probably
arises mostly from VdO vibrations. As the Mo content of the
mixed oxides is increased in the tertiary and quaternary mixed
oxides, Mo/V ratio from 0.4 to 3.3, the corresponding Raman
spectra do not change very much relative to that of the binary
Mo0.6V1.5Ox because of the higher molybdenum oxide contents,
the much stronger Raman scattering cross section of the Mo-O
vibrations (Mo > V-O > Nb-O > Te-O), and the similar
bulk Mo oxide structures.
Figure 11. CH3OH-TPSR from bulk Mo1.0V0.5Te0.16Ox.
The influence of different reactive environments upon the
bulk structure of the quaternary Mo1.0V0.3Te0.16Nb0.12Ox was also
examined with in situ Raman spectroscopy, but the spectra are
not shown for brevity. The quaternary mixed oxide bulk
structure was found to be stable to dehydration in He at 350
°C, exposure to CH3OH/He at 250 °C, and various C3H8/O2
reaction environments (C3H8/O2 ratios from 6/1 to 1/1). A very
small crystalline MoO3 Raman band was only found to be
present under propane ODH conditions. However, this corre-
sponds to only a trace of crystalline MoO3 because of the very
strong Raman scattering cross section of this band. Thus, the
mixed metal oxide bulk structure appears to be relatively stable
under reactive environments that are not net oxidizing.
Figure 12. CH3OH-TPSR from bulk Mo1.0V0.3Te0.16Nb0.12Ox.
The surface of the binary bulk Mo0.6V1.5Ox mixed metal oxide
was found to only yield H2CO product from surface methoxy
species on surface redox sites at Tp ∼ 173 °C (see Figure 10),
which corresponds to the surface V5+ sites (see Table 1). The
shift of the H2CO Tp value from ∼188 to ∼173 °C from the
pure V2O5 to the bulk Mo0.6V1.5Ox mixed metal oxide, respec-
tively, reflects the promotion of V5+ sites by adjacent cation
ligands. The surface of the tertiary bulk Mo1.0V0.5Te0.16Ox mixed
metal oxide system was found to only yield H2CO at Tp ∼ 175
°C, which is also characteristic of surface V5+ redox sites
promoted by adjacent cation ligands (see Figure 11). The surface
of the quaternary bulk Mo1.0V0.3Te0.16Nb0.12Ox mixed metal
oxide was found to yield both DME (Tp ∼ 146 °C) and H2CO
(Tp ∼ 173 °C), as shown in Figure 12. The surface redox sites
for H2CO formation at Tp ∼ 173 °C also correspond to surface
V
5+ sites promoted by the other cations that were already found
for the binary and tertiary bulk mixed metal oxide catalysts.
The surface acidic sites for DME formation at Tp ∼ 146 °C
correspond to surface Mo6+ sites promoted by the other cations.
Although the supported 6% MoO3/Nb2O5 catalyst and bulk
MoO3 did not give rise to DME formation in this low-
temperature region, other supported molybdena catalysts (e.g.,
MoO3/Al2O3, MoO3/TiO2, and MoO3/ZrO2) did exhibit the
formation of DME at the same temperatures.19 The formation
of DME in this temperature range was not observed for other
supported metal oxide catalysts (e.g., V2O5, CrO3, Re2O7, WO3,

2282 J. Phys. Chem. B, Vol. 109, No. 6, 2005
Wachs et al.
Ta2O5, etc.).19 Thus, the surface V5+ sites are responsible for
redox chemical properties of all three mixed metal oxides and
only the quaternary mixed oxide system also exhibits the acidic
chemical property typically associated with surface Mo6+ sites.
The CH3OH-TPSR results for the bulk mixed metal oxides
are presented in Table 1 for comparison: Tp, krds, Kads, Ea, and
H2CO/CH3OH or DME/CH3OH ratios.
CH3OH-TPSR experiments with the molybdena catalysts also
revealed that this method is also sensitive to the simultaneous
presence of fully oxidized and partially reduced surface metal
oxide cations. Furthermore, by performing several reducing
cycles of CH3OH-TPSR it is possible to determine if the surface
cations are being reoxidized by bulk or surface lattice oxygen
via a Mars-van Krevelen reaction mechanism. CH3OH-TPSR
cyclic studies over bulk V2O5 revealed that methanol oxidation
during TPSR proceeds by a bulk Mars-van Krevelen mecha-
nism involving bulk lattice oxygen since the same Tp value was
obtained. In contrast, the CH3OH-TPSR studies over bulk MoO3
demonstrated that methanol oxidation during TPSR proceeds
with the participation of primarily surface lattice oxygen since
the Tp value increased with each reduction cycle. These findings
reveal that O2- diffusion through the bulk V2O5 lattice is much
faster than O2- diffusion through the bulk MoO3 lattice. The
current CH3OH-TPSR experiments suggest that the surfaces of
the bulk mixed metal oxide catalysts investigated in the present
study consist of fully oxidized surface V and Mo sites after
calcination in N2.
Discussion
Surface Elemental Composition of Bulk Mixed Metal
Oxide Catalysts. The CH3OH-TPSR spectra demonstrated that
each of the fully oxidized surface Mo, V, Te, and Nb oxides
possesses a different reactivity temperature window: V (Tp ∼
173-188 °C), Mo (Tp ∼ 189-196 °C), Te (∼260-430 °C),
and Nb (∼300 °C). Thus, it is possible to use this surface
chemistry knowledge to discriminate among the different surface
metal oxide species present for the mixed metal oxide catalysts.
The CH3OH-TPSR spectra, however, may sometimes also be
dominated by the most active surface sites due to the high
surface mobility of the surface *OCH3 species18 and, conse-
quently, may not provide as much information about less
reactive surface catalytic sites. This appears to be the case for
the less reactive surface Te4+ and Nb5+ oxides that are
overshadowed by the more active surface V5+ and Mo6+ sites.
Thus, to obtain a complete elemental analysis of the outermost
surface layer of mixed metal oxides, it is also necessary to
perform surface physical characterization with the surface
sensitive low-energy ion-scattering spectroscopy (LEISS)20 to
complement the CH3OH-TPSR chemical characterization. In
addition, the surface sensitive LEISS characterization investiga-
tions will also determine if various reactive environmental
treatments (e.g., H2, CH3OH, C3H8, C3H8/O2, etc.) can alter the
surface composition of mixed metal oxides.
The surface enrichment in V5+ for bulk Mo0.6V1.5Ox is not
too surprising because of the high vanadia content (71 atomic
%) of this bulk mixed oxide. The surface enrichment in V5+
for bulk Mo1.0V0.5Te00.16Ox and Mo1.0V0.3Te0.17Nb0.12Ox, how-
ever, is somewhat surprising since the bulk V atomic contents
of these mixed oxides are only 30% and 19%, respectively. The
enriched surface V5+ sites on the mixed metal oxide surfaces
is related to the high mobility of the V5+ cation and the driving
force to decrease to surface free energy of the mixed metal oxide
system. The enhanced mobility of V5+ is associated with its
much lower Tammann temperature, defined as half the melting
point, of V2O5 (209 °C) compared to the other metal oxides
TeO2 (230 °C), MoO3 (261 °C), and Nb2O5 (620 °C). The
Tammann temperature represents the temperature that surface
atoms of a material begin to diffuse and suggests that the
mobility of V5+ is greater than that the other metal cations. The
surface free energy of surfaces terminating with MdO bonds
(e.g., V2O5 and MoO3) is significantly less than surfaces
terminating with M-OH bonds (e.g., TeO2 and Nb2O5).18 Thus,
the much higher mobility of V5+ relative to the other metal
cations coupled with its ability to minimize the surface free
energy of mixed metal oxide systems is apparently responsible
for its surface enrichment of these bulk mixed metal oxides.
Surface Oxidation States of Bulk Mixed Metal Oxide
Catalysts. For those active surface cations that dominate the
CH3OH-TPSR spectra, the Tp position also reflects the cation
oxidation state. Vohs et al. nicely demonstrated that for
supported vanadia monolayer catalysts the Tp value for the
surface vanadia species varies ∼100 °C depending on the
vanadia cation oxidation state.10-12 A similar trend also occurs
for reduced surface molybdena species (see Table 1). The cyclic
Surface Chemical Nature of Bulk Mixed Metal Oxide
Catalysts. For metal oxides, it is critical to know the chemical
nature of the surface metal oxide cations, which cannot be
provided by physical characterization methods. Fortunately,
methanol oxidation over metal oxides produces different reaction
products depending on the surface chemistry of the specific
metal oxide: H2CO from surface redox sites, DME from surface
acidic sites, and CO2 from surface basic sites. The latter assumes
that CO2 is the primary product and does not originate from
overoxidation of H2CO or from surface deposition of carbon-
aceous molecules from the background.17 Surface redox sites
are able to readily give up their bridging oxygen in M-O-S
bonds and surface acidic sites are not capable of giving up their
bridging oxygen in bridging M-O-S bonds and possess either
Bronsted acidity, H+, or Lewis acidity, oxygen vacancies that
are electron deficient.16 Surface basic sites are electron rich and
readily accept molecules containing H+ or that are electron poor.
Steady-state methanol oxidation over the pure MoO3, V2O5,
TeO2, and Nb2O5 demonstrated that only redox and acidic
reaction products are obtained for low methanol conversions.21
From this information it is possible to conclude from CH3OH-
TPSR results that surface vanadia, molybdena, and telluria sites
are redox and surface niobia sites are acidic in nature. Although
higher alcohols, such as 2-propanol, can also determine the
nature of active surface sites, they are much more sensitive to
the acidic surface sites because of their more facile unimolecular
dehydration reactions compared to the more demanding bimo-
lecular CH3OH dehydration reaction to form DME.21
Surface Reactivity of Bulk Mixed Metal Oxide Catalysts.
The first-order Arrhenius rate constants for the rate-determining-
steps, krds, of the various methanol oxidation surface reactions
are quantified from the Tp values obtained from the CH3OH-
TPSR spectra and are presented in Table 1. It is clear that surface
V5+ is about an order of magnitude more active than surface
Mo6+, which in turn is ∼100 times more active than Te4+, for
methanol oxidation to H2CO. The surface Nb5+ sites are not
very active for methanol dehydration to DME and reflect the
presence of weak surface Lewis acid sites.16 The significant
effect reduction has on the surface cation oxidation state is also
apparent from the krds values. For example, surface V5+ > V4+
. V3+ and the krds spans a factor of ∼100 in going from V5+
to V3+ 11,12
Furthermore, the effect of promoting ligands on the
.
catalytic activity is directly quantifiable by the changes in the
Tp values. For example, H2CO formation from surface V5+

Active Sites on Mixed Metal Oxide Catalysts
J. Phys. Chem. B, Vol. 109, No. 6, 2005 2283
present on bulk V2O5 exhibits a Tp ∼ 188 °C and from surface
V5+ promoted by neighboring Mo6+ in the bulk mixed oxides
exhibits a Tp ∼ 173 °C, which translates to a kinetic factor of
∼3 in the first-order krds value. This promotion is most probably
related to the lower Sanderson electronegativity of adjacent
Mo6+ (2.2) and Nb5+ (1.42) cations relative to the V5+ (2.51)
and Te4+ (2.62) cations.22 The lower electronegativity of
adjacent cations would increase the partial negative charge on
the bridging V-O-M bonds (where M is either Mo or Nb).
This would further increase the basic character of the bridging
oxygen atom and enhance its affinity for accepting H atoms
from the adsorbing methanol molecule and the surface methoxy
intermediate. It also accounts for the activity enhancement
observed when the surface Mo6+, V5+, and Te4+ are supported
on Nb2O5, which possesses the lowest electronegativity among
these four.
oxidized under reaction conditions employing excess gas-phase
oxygen.24 This observation is in agreement with the conclusion
that for supported vanadia catalysts the surface lattice oxygen
from the surface vanadia species is involved in propane ODH
from oxygen isotopic labeling experiments and the zero-order
dependence on the oxygen partial pressure.25 This catalytic
reactivity pattern also seems to be applicable for propylene
selective oxidation to acrolein studies over supported vanadia
catalysts currently in progress.26 It has been shown that propane
selective oxidation to acrylic acid over the bulk Mo-V-Te-
Nb-O mixed metal oxide catalysts exhibits a zero-order
dependence on the oxygen partial pressure and first-order
dependence on the propane partial pressure,27 which implies a
Mars-van Krevelen reaction mechanism using surface or bulk
lattice oxygen from the catalyst in the rate-determining-step.
The current bulk V2O5 and MoO3 CH3OH-TPSR cyclic reduc-
tion studies revealed that it would also be possible to discrimi-
nate between surface and bulk Mars-van Krevelen reaction
mechanisms for this bulk mixed metal oxide catalyst. Thus, it
seems that the CH3OH-TPSR spectroscopy findings in the
current investigation are not only limited to methanol oxidation
studies but also can be generalized to other hydrocarbon
oxidation reactions.
Adsorption Equilibrium Constants of Bulk Mixed Metal
Oxides. Combination of the krds values obtained from the CH3-
OH-TPSR studies and the corresponding steady-state reaction
rates, TOF, allows for the direct determination of the Kads values.
This calculation assumes that the surface metal oxide sites are
fully oxidized and, thus, cannot be applied to the partially
reduced metal oxide surfaces. The methanol adsorption equi-
librium constants, Kads, are shown in Table 1 for those oxide
catalysts that have been examined for steady-state methanol
oxidation. From the current limited available data in Table 1, it
is not clear if any relationships exist between Kads and the krds
values for methanol oxidation. However, such a relationship
was found to hold for supported vanadia catalysts with different
oxide supports since both the dissociative CH3OH chemisorption
and surface CH3O* decomposition to H2CO involve transfer
of an H atom to basic oxygen.
Conclusions
The CH3OH-TPSR experiments have shown to be able to
provide fundamental surface information about the nature of
the active surface sites present on the outermost surface layer
of bulk Mo-V-Te-Nb-O mixed metal oxides. The Tp values
are sensitive to each specific cation and reflect surface com-
positional information. Furthermore, the Tp values are also
sensitive to the presence of reduced surface cations since the
Tp values shift to higher values from partially reduced surface
cations. This latter characteristic can also be applied to determine
if a Mars-van Krevelen reaction mechanism over bulk mixed
oxides proceeds via a bulk or surface lattice oxygen. The
chemical nature of the active surface sites is reflected in the
methanol oxidation reaction products formed (e.g., H2CO from
surface redox sites and DME from surface acidic sites).
Determination of the Tp values allows for the direct quantifica-
tion of the corresponding Arrhenius rate constant krds, the surface
reaction constant for the rate-determining step of the decom-
position of the surface methoxy intermediate to formaldehyde.
Changes in the Tp or krds values reflect the promotion of the
active redox sites by neighboring cation ligands that influence
the electron density on the bridging oxygen in M-O-S bonds.
Knowledge of the corresponding steady-state TOF value also
allows the determination of Kads, the equilibrium adsorption
constant. The combination of CH3OH-TPSR and LEISS studies
is ideally suited to advance our fundamental knowledge about
the chemical nature of the outermost layer of bulk mixed metal
oxide catalysts in the coming years. Molecular structural
information about the active surface sites present in the
outermost layer of mixed metal oxides still awaits the develop-
ment of novel physical spectroscopic surface characterization
methods and the use of model mixed metal oxide systems.28
Application of this approach to characterizing the surface of
the bulk Mo-V-Te-Nb-O mixed metal oxides revealed that
the active surface sites consisted of surface V5+ and Mo6+
cations and that the role of Nb5+ and Te4+ cations may primarily
be as promoter ligands. The most active and selective bulk
mixed metal oxide catalyst, the quaternary Mo-V-Te-Nb-O
system, was found to possess both surface V5+ redox and surface
Mo6+ acidic sites. Comparison of the current CH3OH-TPSR
Lattice Oxygen Transport in Bulk Metal Oxides. Cyclic
CH3OH-TPSR studies revealed the interesting observation that
the surface of bulk V2O5 was readily reoxidized by bulk lattice
oxygen after being reduced by CH3OH, but that the surface of
MoO3 was not rapidly reoxidized by the bulk lattice oxygen
and remained reduced after methanol reduction. This result is
unexpected since previous studies of bulk lattice oxygen
diffusion in V2O5 and MoO3 indicated that the oxygen mobilities
in these metal oxides were essentially identical.23,24 The origin
of this difference in oxygen diffusion measurements may be
that the current studies are only probing the outermost layer
and a limited number of layers below the surface, the surface
region, rather than solid-state diffusion of bulk lattice oxygen
in bulk metal oxides. These interesting findings suggest that
for catalytic applications the bulk lattice oxygen mobility in
the surface region may be different than the more typical bulk
lattice oxygen transport usually measured for bulk metal oxides.
Oxidative Dehydrogenation of Propane to Propylene. To
examine the applicability of the CH3OH-TPSR findings to other
oxidation reactions, the bulk and supported Mo-V-Te-Nb-O
mixed metal oxide catalysts were compared to earlier reactivity
data for propane oxidative dehydrogenation to propylene.15 For
propane ODH, the reactivity order of the fully oxidized bulk
and niobia supported catalysts followed the pattern V5+ > Mo6+
. Te4+. The propane ODH findings exactly parallel the CH3-
OH-TPSR reactivity pattern above that the surface V5+ > Mo6+
. Te4+ for methanol oxidative dehydrogenation to H2CO. This
parallel trend between propane and methanol ODH is consistent
in that the rate-determining steps in both oxidative dehydroge-
nation reactions being the breaking of a C-H bond. In situ
Raman studies reveal that the surface metal oxide species present
in the supported metal oxide catalysts are essentially fully

2284 J. Phys. Chem. B, Vol. 109, No. 6, 2005
Wachs et al.
(9) Burcham, L. J.; Deo, G.; Gao, X.; Wachs, I. E. Top. Catal. 2000,
11/12, 85.
findings with corresponding propane ODH catalytic studies
suggests that the findings from CH3OH-TPSR are also be
applicable to other hydrocarbon oxidation reactions. Thus, this
new CH3OH-TPSR spectroscopy approach to obtaining surface
information about bulk mixed metal oxide catalysts is providing
much new and needed rich chemical insights about the surface
properties of bulk mixed metal oxide catalysts that previously
were not attainable by other characterization methods.
(10) Zhang, Z.; Henrich, V. E. Surf. Sci. 1992, 277, 263. Guo, Q.; Lee,
S.; Goodman, D. W. Surf. Sci. 1999, 437, 38. Sambi, M.; Sangiovanni, G.;
Granozzi, G.; Parmiagiani, F. Phys. ReV. B 1997, 55, 7850. Wong, G. S.;
Kragten, D. D.; Vohs, J. M. Surf. Sci. 2000, 452, L293. Wang, Q.; Madix,
R. J. Surf. Sci. 2001, 474, L213.
(11) Wong, G. S.; Concepcion, M. R.; Vohs, J. M. Surf. Sci. 2003, 211.
(12) Feng, T.; Vohs, J. M. J. Catal. 2002, 208, 301.
(13) Holstein, W. L.; Machiels, C. J. J. Catal. 1996, 162, 118.
(14) Deo, G.; Wachs, I. E. J. Catal. 1994, 146, 323, 335.
(15) Zhao, Z.; Wachs, I. E. J. Phys. Chem. B 2003, 107, 6333.
(16) Datka, J.; Turek, A. M.; Jehng, J.-M.; Wachs, I. E. J. Catal. 1992,
135, 186. Jehng, J.-M.; Turek, A. M.; Wachs, I. E. Appl. Catal., A 1992,
83, 179.
Acknowledgment. This research work was sponsored by
the Department of Energy-Basic Energy Sciences (Grant
DEFG02-93ER14350). Israel E. Wachs wishes to dedicate this
paper to Michel Boudart for introducing him to the kinetics of
heterogeneous catalytic reactions in the course on “Kinetics of
Chemical Processes” in the winter of 1974, and for his
mentoring during his graduate school days as well as after
graduation from Stanford University. Michel Boudart has had
a profound influence on Wachs’ scientific career in heteroge-
neous catalysis for the past 30 years.
(17) Tatibouet, J. M. Appl. Catal., A 1997, 148, 213. Tatibouet, J. M.;
Lauron-Pernot, H. J. Mol. Catal. A: Chem. 2001, 171, 2005.
(18) Wang, C. B.; Cai, Y.; Wachs, I. Langmuir 1999, 15, 1223.
(19) Kim, T.; Wachs, I. E. To be submitted for publication.
(20) Briand, L. E.; Tkachenko, O. P.; Guraya, M.; Wachs, I. E.; Gruenert,
W. Surf. Interface Anal. 2004, 36, 238. Briand, L. E.; Tkachenko, O. P.;
Guraya, M.; Gao, X.; Wachs, I. E. J. Phys. Chem. B 2004, 108, 4823.
(21) Badlani, M.; Wachs, I. Catal. Lett. 2001, 75, 137. Kulkarni, D.;
Wachs, I. E. Appl. Catal., A 2002, 237, 121.
(22) Sanderson, R. T. J. Chem. Educ. 1988, 65, 112.
(23) Boreskov, G. In Catalysis Science and Technology; Anderson, J.
R., Boudart, M., Eds.; Springer: New York, Vol. 3, p 62.
(24) Kapoor, R.; Oyama, S. T. J. Mater. Res. 1997, 12, 467.
(25) Gao, X.; Jehng, J.-M.; Wachs, I. E. J. Catal. 2002, 209, 43.
(26) Chen, K.; Khodakov, A.; Yang, J.; Bell, A. T.; Iglesia, E. J. Catal.
2000, 186, 325. Chen, K.; Xie, S.; Iglesia, E.; Bell, A. T. J. Catal. 2000,
192, 197. Chen, K.; Bell, A. T.; Iglesia, E. J. Phys. Chem. B 2000, 104,
1292.
(27) Zhao, C.; Wachs, I. E. To be submitted for publication.
(28) Katou, T.; Vitrry, D.; Ueda, W. Chem. Lett. 2003, 32, 1028.
(29) Schlogl, R. The Impact of In Situ Analysis of Heterogeneous
Catalysts for the Improvement of Their Function. Presentation at the Meeting
of the Catalysis and Chemical Transformations Programs; Office of Basic
Energy Sciences, Department of Energy, U.S.: Rockville, MD, May 23-
26, 2004.
References and Notes
(1) Boudart, M.; Djega-Mariadassou, G. Kinetics of Heterogeneous
Catalysis; Princeton University Press: Princeton, NJ, 1984.
(2) Emmettt, P. H.; Brunauer, S. J. Am. Chem. Soc. 1937, 59, 310,
1553.
(3) Djega-Mariadassou, G.; Marques, A. R.; Davignon, L. J. Chem.
Soc., Faraday Trans. 1 1982, 78, 2447.
(4) Farneth, W.; Staley, R. H.; Sleight, A. W. J. Am. Chem. Soc. 1986,
108, 2327. Farneth, W. E.; Ohuchi, F.; Staley, R.; Chowdry, U.; Sleight,
A. W. J. Phys. Chem. 1985, 89, 2493.
(5) Briand, L. E.; Farneth, W.; Wachs, I. E. Catal. Today 2000, 62,
219.
(6) Briand, L. E.; Hirt, A. M.; Wachs, I. E. J. Catal. 2001, 202, 268.
(7) Briand, L. E.; Jehng, J. M.; Cornaglia, L. M.; Hirt, A. M.; Wachs,
I. E. Catal. Today 2003, 78, 257.
(8) Burcham, L. J.; Briand, L. E.; Wachs, I. E. Langmuir 2001, 17,
6164;. Burcham, L. J.; Briand, L. E.; Wachs, I. E. Langmuir 2001, 17,
6175.