Surface and bulk aspects of mixed oxide catalytic nanoparticles: Oxidation and dehydration of CH3OH by polyoxometallates

Article abstract of DOI:10.1021/ja904957d

The molecular structures and surface chemistry of mixed metal oxide heteropolyoxo vanadium tungstate (H_3+xPW_12-xV _xO₄₀ with x = 0, 1, 2, and 3) Keggin nanoparticles (NPs), where vanadium is incorporated into the primary Keggin structure, and supported VO_x on tungstophosphoric acid (TPA, H₃PW ₁₂O₄₀), where vanadium is present on the surface of the Keggin unit, were investigated with solid-state magic angle spinning ⁵¹V NMR, FT-IR, in situ Raman, in situ UV-vis, CH₃OH temperature-programmed surface reaction (TPSR), and steady-state methanol oxidation. The incorporated VO_x unit possesses one terminal V=O bond, four bridging V-O-W/V bonds, and one long V-O-P bond in the primary Keggin structure, and the supported VO_x unit possesses a similar coordination in the secondary structure under ambient conditions. The specific redox reaction rate for VO_x in the Keggin primary structure is comparable to that of bulk V₂O₅ and the more active supported vanadium oxide catalysts. The specific acidic reaction rate for the WO_x in the TPA Keggin, however, is orders of magnitude greater than found for bulk WO₃, supported tungsten oxide catalysts, and even the highly acidic WO₃-ZrO₂ catalyst synthesized by coprecipitation of ammonium metatungstate and ZrO(OH)₂. From CH ₃OH-TPSR and in situ Raman spectroscopy it was found that incorporation of vanadium oxide into the primary Keggin structure is also accompanied by the formation of surface VO_x species at secondary sites on the Keggin outer surface. Both CH₃OH-TPSR and steady-state methanol oxidation studies demonstrated that the surface VO_x species on the Keggin outer surface are significantly less active than the VO _x species incorporated into the primary Keggin structure. The presence of the less active surface VO_x sites in the Keggins, thus, decreases the specific reaction rates for both methanol oxidation and methanol dehydration. During methanol oxidation/dehydration (O₂/CH ₃OH = 2.17, T = 225°C), in situ UV-vis diffuse reflectance spectroscopy revealed that vanadium oxide is primarily present as the V ⁺⁵ cation, which reflects the Mars-van Krevelen redox mechanism and rapid reoxidation by molecular O₂. The bulk TPA Keggin structure becomes more disordered and less thermally stable as the vanadium content increases. Although surface polyaromatic carbon forms during methanol oxidation on the Keggin surfaces, its influence on the reaction kinetics seems minimal as the carbon content diminishes as the vanadium oxide content increases and the reaction temperature is raised. No relationships were found between the electronic structure (UV-vis E_g values) and TOF_redox and TOF_acid (TOF = turnover frequency) kinetics, which reflect the complexity of H_3+xPW_12-xV_xO₄₀ Keggins. The overall catalytic performance of the H_3+xPW _12-xV_xO₄₀ Keggin materials results from a complex interplay among the presence of redox vanadium (as secondary surface VO_x species and substituted VO_x sites in the primary Keggin NP structure), structural disorder of the Keggin NPs, exposed surface acid and redox sites, and coke deposition. These new insights reveal that the Keggin heteropolyoxometallates are much more complex than originally thought and that care must be taken in using Keggins as model mixed metal oxide NPs in catalytic kinetic and theoretical studies because their surface and bulk structures are dynamic under the reaction conditions.

Full text of DOI:10.1021/ja904957d

Published on Web 10/06/2009
Surface and Bulk Aspects of Mixed Oxide Catalytic
Nanoparticles: Oxidation and Dehydration of CH3OH by
Polyoxometallates
Lingaiah Nakka,† Julie E. Molinari, and Israel E. Wachs*
Operando Molecular Spectroscopy & Catalysis Laboratory, Chemical Engineering Department,
Lehigh UniVersity, Bethlehem, PennsylVania 18015
Received June 18, 2009; E-mail: iew0@lehigh.edu
Abstract: The molecular structures and surface chemistry of mixed metal oxide heteropolyoxo vanadium
tungstate (H3+xPW12-xVxO40 with x ) 0, 1, 2, and 3) Keggin nanoparticles (NPs), where vanadium is
incorporated into the primary Keggin structure, and supported VOx on tungstophosphoric acid (TPA,
H3PW12O40), where vanadium is present on the surface of the Keggin unit, were investigated with solid-
state magic angle spinning 51V NMR, FT-IR, in situ Raman, in situ UV-vis, CH3OH temperature-programmed
surface reaction (TPSR), and steady-state methanol oxidation. The incorporated VOx unit possesses one
terminal VdO bond, four bridging V-O-W/V bonds, and one long V-O-P bond in the primary Keggin
structure, and the supported VOx unit possesses a similar coordination in the secondary structure under
ambient conditions. The specific redox reaction rate for VOx in the Keggin primary structure is comparable
to that of bulk V2O5 and the more active supported vanadium oxide catalysts. The specific acidic reaction
rate for the WOx in the TPA Keggin, however, is orders of magnitude greater than found for bulk WO3,
supported tungsten oxide catalysts, and even the highly acidic WO3-ZrO2 catalyst synthesized by
coprecipitation of ammonium metatungstate and ZrO(OH)2. From CH3OH-TPSR and in situ Raman
spectroscopy it was found that incorporation of vanadium oxide into the primary Keggin structure is also
accompanied by the formation of surface VOx species at secondary sites on the Keggin outer surface.
Both CH3OH-TPSR and steady-state methanol oxidation studies demonstrated that the surface VOx species
on the Keggin outer surface are significantly less active than the VOx species incorporated into the primary
Keggin structure. The presence of the less active surface VOx sites in the Keggins, thus, decreases the
specific reaction rates for both methanol oxidation and methanol dehydration. During methanol oxidation/
dehydration (O2/CH3OH ) 2.17, T ) 225 °C), in situ UV-vis diffuse reflectance spectroscopy revealed
that vanadium oxide is primarily present as the V+5 cation, which reflects the Mars-van Krevelen redox
mechanism and rapid reoxidation by molecular O2. The bulk TPA Keggin structure becomes more disordered
and less thermally stable as the vanadium content increases. Although surface polyaromatic carbon forms
during methanol oxidation on the Keggin surfaces, its influence on the reaction kinetics seems minimal as
the carbon content diminishes as the vanadium oxide content increases and the reaction temperature is
raised. No relationships were found between the electronic structure (UV-vis Eg values) and TOFredox and
TOFacid (TOF ) turnover frequency) kinetics, which reflect the complexity of H3+xPW12-xVxO40 Keggins.
The overall catalytic performance of the H3+xPW12-xVxO40 Keggin materials results from a complex interplay
among the presence of redox vanadium (as secondary surface VOx species and substituted VOx sites in
the primary Keggin NP structure), structural disorder of the Keggin NPs, exposed surface acid and redox
sites, and coke deposition. These new insights reveal that the Keggin heteropolyoxometallates are much
more complex than originally thought and that care must be taken in using Keggins as model mixed metal
oxide NPs in catalytic kinetic and theoretical studies because their surface and bulk structures are dynamic
under the reaction conditions.
etc.).2 Each metal oxide cation possesses one terminal MdO
bond, four bridging M-O-M bonds, and one long M-O-P
bond. The negatively charged mixed metal oxide POM cluster
is charge balanced by surface protons. The molecular structural
chemistry of POM catalysts can be tuned by the substitution of
different cations or by exchanging the POM surface protons
with surface cations.3 For example, one or more of the primary
Introduction
Polyoxometallates (POMs) are viewed as ideal subnanometer
mixed oxide particles possessing a dimension of ∼0.8 nm.1 The
POM Keggin cluster consists of a central cation (e.g., P, Si, Ti,
etc.) that is surrounded by 12 metal oxide cations (e.g., W, Mo,
† Permanent address: Catalysis Laboratory, I&PC Division, Indian
Institute of Chemical Technology, Hyderabad-500607, India.
(1) Pope, M. T.; Muller, A. Angew. Chem., Int. Ed. Engl. 1991, 30, 34.
(2) Tsigdinos, G. A. Top. Curr. Chem. 1978, 76, 1.
(3) Mizuno, N.; Hann, W.; Kudo, T. J. Catal. 1998, 178, 391.
9
15544 J. AM. CHEM. SOC. 2009, 131, 15544–15554
10.1021/ja904957d CCC: $40.75  2009 American Chemical Society

Oxidation and Dehydration of CH3OH by POMs
A R T I C L E S
W6+ cations in the H3PW12O40 ·nH2O Keggin structure can be
substituted by V5+ cations, and the charge imbalance is
compensated by additional surface protons or even surface
inorganic cations in the surrounding secondary structure.4,5
Keggin clusters have generated much interest as catalysts
since their constitutional elements can readily be varied and
tuned for acidic and redox catalytic applications.6-8 For
example, vanadium oxide addition to the acidic H3PW12O40 ·
nH2O Keggin enhances its catalytic partial oxidation proper-
ties because vanadia introduces redox character to the
Keggins will be chemically probed by CH3OH temperature-
programmed surface reaction (TPSR) spectroscopy and steady-
state methanol oxidation. In situ Raman spectroscopy will be
employed to determine the surface and bulk Keggin molecular
structural chemistry during methanol oxidation. In situ UV-vis
will be employed to determine the corresponding electronic
structures of the mixed metal oxide POMs. These studies will
provide new fundamental insights into the subnanometer-sized
mixed metal oxide Keggin POMs.
H3+xPW12-xVxO40 · nH2O POM.9 Vanadium-containing H3+x
-
Experimental Section
PMo12-xVxO40 ·nH2O POMs have been investigated as mixed
metal oxide catalysts for partial oxidation of alcohols and
alkanes.10-12 The ability of mixed metal oxide POMs to perform
selective oxidation reactions has stimulated much interest in this
class of catalytic materials and motivated a deeper fundamental
understanding of their molecular structural chemistry and its
relationship to their catalytic performance. The establishment
of molecular structure-catalytic performance relationships is
made all the more challenging because it has been shown that
the structure of the POM catalysts can change under the reaction
conditions and may not necessarily correspond to the ideal
crystallographic structure of the originally prepared Keggin-
type material.13,14 Thus, in situ molecular structural character-
ization of mixed metal oxide POMs under different environ-
mental conditions must be collected to be able to establish their
fundamental molecular structure-activity relationships.15 In
addition to the bulk structures of POMs in reactive environ-
ments, there currently is no information about the surface
composition of POMs under partial oxidation/dehydration
reaction conditions.
The objective of this investigation is to examine both the
surface and bulk properties of the subnanometer-sized Keggin
POMs during the partial oxidation/dehydration of CH3OH over
H3+xPW12-xVxO40 ·nH2O mixed metal oxide Keggins, where it
will be converted to HCHO over surface redox sites and to
CH3OCH3 (DME, dimethyl ether) over surface acid sites, to
determine their molecular structure-performance relationships.
Unlike the more investigated molybdenum-based POMs
(H3+xPMo12-xVxO40),10-12 the tungsten-containing POMs have
not received much attention as partial oxidation catalysts because
of their stronger acidic character.16 In the present investigation,
the vanadium-containing H3+xPW12-xVxO40 (x ) 0, 1, 2, and 3)
will be examined for the influence of vanadia upon the Keggin
surface composition, bulk structure, and surface reactivity during
methanol oxidation/dehydration. The surface chemistry and
composition of the V-containing tungstophosphoric acid (TPA)
Catalysts. The H3PW12O40 and H3+xPW12-xVxO40 (x ) 1, 2, 3)
POMs were purchased from Aldrich Chemical Co. and Nippon
Inorganic Color and Chemical Co., respectively. The vanadium-
containing tungstophosphoric acid, H3+xPW12-xVxO40 (x ) 1, 2, 3),
catalysts will be denoted by TPAV1, TPAV2, and TPAV3,
respectively. In addition, vanadia was also introduced onto the
surface or secondary structure of TPA, and this catalyst is denoted
as VOTPA.17 An aqueous solution containing an appropriate
amount of VO2+ ions was prepared by reaction of crystalline V2O5
with oxalic acid at about 100 °C with maintenance of the VO2+
/
TPA molar ratio equal to 2. After complete dissolution of the V2O5
solid, the solution was cooled to room temperature, and the required
amount of TPA was added. The excess water was then removed
by evaporation at 100 °C, and the dried mass was subsequently
calcined at 300 °C in air for 3 h.
Solid-State 51V NMR. Solid-state magic angle spinning (MAS)
51V NMR spectra were recorded under ambient conditions on a
General Electric model GN-300 NMR spectrometer at 78.89 MHz.
The instrument was equipped with a Nicolet 2090-IIIA high-speed
digital oscilloscope and a 5 mm MAS NMR Doty Scientific probe.
The measurements were carried out on static samples with a simple
one-pulse excitation of 1 µs width, a preacquisition delay of 10
µs, a dwell time of 0.5 µs, and a relaxation delay of 4 s. Prior to
Fourier transformation, the signal-averaged free induction decays
were multiplied by an exponential function equivalent to 500 Hz
line broadening to decrease noise in the spectra. All chemical shifts
were referenced against liquid VOCl3.
Raman Spectroscopy. The in situ Raman spectra of the POMs
were collected with a Horiba-Jobin Yvon LabRam-HR spectrometer
equipped with a confocal microscope, 2400/900 grooves/mm
gratings, and a notch filter. The visible laser excitation at 532 nm
(visible/green) was supplied by a Yag doubled diode pumped laser
(20 mW). The scattered photons were directed and focused onto a
single-stage monochromator and measured with a UV-sensitive
LN2-cooled CCD detector (Horiba-Jobin Yvon CCD-3000 V). The
powdered samples, ∼5-10 mg, were loosely spread onto a small
ceramic boat inside the environmental cell (Linkam T-1500) and
maintained below the confocal microscope. Dehydration of the
catalyst samples was carried out by heating the environmental cell
to different temperatures under O2/He or CH3OH/O2/He flow (30
mL/min). After the desired treatment temperature was reached, the
samples were dehydrated for 30 min before the Raman spectrum
was recorded. The W-O-W bending modes at ∼220-250 cm-1
were employed as an internal standard to normalize the intensity
of the Raman spectra since these vibrations originate from the intact
Keggin unit and are least sensitive to the environmental conditions.
FT-IR Spectroscopy. The IR spectra were obtained under
ambient conditions with a Bio-Rad FTS-40A FT-IR spectrometer
equipped with a DTGS detector. The IR spectra were recorded with
a resolution of 2 cm-1 using 250 signal-averaged scans using the
KBr disk method.
(4) Okuhara, T.; Mizuno, N.; Misono, M. AdV. Catal. 1996, 41, 113.
(5) Neumann, R.; Dror, I. Appl. Catal., A 1998, 172, 67.
(6) Misono, M. Chem. Commun. 2001, 1141.
(7) Mizuno, N.; Misono, M. Chem. ReV. 1998, 98, 198.
(8) Hill, C. L.; Posser-McCarthy, C. M. Coord. Chem. ReV. 1995, 143,
407.
(9) Bru¨ckman, K.; Tatibouet, J. M.; Che, M.; Serwicka, E.; Haber, J. J.
Catal. 1993, 139, 455.
(10) Liu, H.; Iglesia, E. J. Phys. Chem. B. 2003, 107, 10840.
(11) Centi, G.; Lena, V.; Trifiro, F.; Ghoussoub, D.; Aissi, C. F.; Guelton,
M.; Bonnelle, J. P. J. Chem. Soc., Faraday Trans. 1990, 86, 2775.
(12) Marchal-Roch, C.; Bayer, R.; Moisan, J. F.; Teze´, A.; Herve´, G. Top.
Catal. 1996, 3, 407.
UV-Vis Diffuse Reflectance Spectroscopy (DRS). Solid
catalyst samples were examined using a Varian Cary 5E UV-vis
spectrophotometer equipped with the Harrick Praying Mantis
(13) Watzenberger, O.; Emig, G.; Lynch, D. T. J. Catal. 1990, 124, 247.
(14) Lee, K. E.; Melsheimer, J.; Berndt, S.; Mestl, G.; Schlo¨gl, R.; Ko¨hler,
K. Appl. Catal., A 2001, 214, 125.
(15) Ressler, T.; Timpe, O.; Girgsdies, F.; Wienold, J.; Neisius, T. J. Catal.
2005, 231, 279.
(17) Casarini, D.; Centi, G.; Jiru, P.; Lena, V.; Tvaruzkova, Z. J. Catal.
1993, 143, 325.
(16) Kozhevnikov, I. V. Chem. ReV. 1998, 98, 171.
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15545

A R T I C L E S
Nakka et al.
attachment (model DRA-2) for diffuse reflectance spectroscopy.
The reference laser was passed through a 1.5 absorbance unit filter
to minimize background noise. Approximately 30 mg of each finely
ground sample was loaded into the cup of the Harrick cell (model
HVC-DR2 with a high-pressure dome). A magnesium oxide white
reflectance standard baseline was collected under ambient condi-
tions. The spectrum of each catalyst sample was collected over the
wavelength range of 200-800 nm under ambient conditions, under
dehydrated conditions at 225 °C, and under in situ reaction
conditions during methanol oxidation (O2/CH3OH ) 2.17). The
UV-vis DRS spectra were analyzed using the Kubelka-Munk
function F(R¥) and evaluated for the band gap value (Eg) using
methods described by Gao18 and Ross-Medgaarden.19
CH3OH-TPSR Spectroscopy. The CH3OH-TPSR spectroscopy
measurements were performed on an Altamira (AMI-200) system
equipped with an online quadrupole mass spectrometer (Dycor
Dymaxion DME200MS). About 200 mg of sample was typically
loaded into a U-shaped quartz tube and initially treated at 250 °C
(ultra zero grade O2/He, 30 mL/min) for ∼1 h to remove any
combustible impurities that may be present. To ensure that the
catalysts remain in a fully oxidized state, the pretreated samples
were first cooled in flowing air to 110 °C and then the gas stream
was switched to an ultra-high-purity He flow with further cooling
to 100 °C. After flushing with continuously flowing He for another
1 h at 100 °C to remove any physically adsorbed oxygen or potential
background gases, a CH3OH/He gas mixture (2000 ppm methanol)
feed was introduced at 30 mL/min for CH3OH chemisoprtion and
maintained for ∼40 min. Previous work demonstrated that the
adsorption temperature of 100 °C almost completely minimizes the
presence of physically adsorbed methanol on the catalyst samples
since the physically adsorbed CH3OH desorbs below this temper-
ature.20 After methanol adsorption, the POM catalysts were again
purged at 100 °C with an ultra-high-purity (UHP) He flow for an
additional 1 h to remove any residual physically adsorbed methanol.
The CH3OH-TPSR spectroscopy experiment was then performed
with a heating rate of 10 °C/min in the flowing UHP He, and the
desorption products were monitored with the online mass spec-
trometer. The m/e values used to detect the different desorption
products were 31 (CH3OH), 30 (H2CO), 45 (CH3OCH3-DME),
76 ((CH3O)2CH2-DMM), 44 (CO2), and 28 (CO). For desorbing
products that gave rise to several fragments in MS, additional m/e
values were also collected to confirm the identity of the desorbing
Figure 1. Solid-state MAS 51V NMR spectra of VOTPA and TPAV1
catalysts under ambient conditions (the asterisk indicates the central band).
oxidation to redox products (HCHO, MF, and DMM) and methanol
dehydration to DME were determined by normalizing the steady-
state reaction rates per gram by the number of redox and acid sites
per gram of catalyst, respectively, which were determined from
CH3OH-TPSR.
+
products (e.g., m/e ) 45 for CH3OCH2 and m/e ) 15 for the
associated CH3+ cracking fragment). The number of exposed redox
and acid sites per gram (Ns) of each catalyst was determined from
the TPSR spectra. The area under the curve for the formaldehyde
peak was integrated to determine the number of redox sites, and
the area under the curve for the dimethyl ether peak was integrated
to determine the number of acid sites. The MS signals for HCHO
and DME were calibrated against quantitative thermogravimetric
analysis (TGA) measurements with exclusive redox and acid
catalysts, respectively.
Results
Solid-State MAS 51V NMR under Ambient Conditions. Solid-
state MAS 51V NMR studies were undertaken to determine the
local coordination of the VOx sites in the V-containing TPA
mixed oxide Keggins under ambient conditions where the POMs
are in a hydrated state from background moisture. The MAS
51V NMR spectra of the catalysts are shown in Figure 1. The
solid-state MAS 51V NMR chemical shift for the supported 20%
V2O5/Al2O3 under ambient conditions was taken as a reference
since prior NMR studies have demonstrated that only hydrated
VO6 species are present for this sample under ambient condi-
tions.21 The solid-state MAS 51V NMR chemical shift for
supported 20% V2O5/Al2O3 under ambient conditions exhibits
a central band at -362 ppm characteristic of VO6 sites. The
chemical shifts for TPAV1 and VOTPA occur at -385 and
-246 ppm, reflecting the VO6 coordination of the hydrated VOx
sites in these V-containing TPA Keggins. The VO6 coordination
for the VOTPA catalyst is consistent with the VO6 unit
substituting for one of the WO6 units in the TPA Keggin
structure. The VO6 coordination for the hydrated VOTPA
CH3OH Oxidation/Dehydration. Steady-state methanol oxida-
tion was performed in an isothermal fixed-bed differential reactor
at atmospheric pressure. In a typical experiment about 30 mg of
catalysts was held between two glass wool beds and pretreated under
O2/He flow at 250 °C for 30 min before passing the gaseous
reactants at the desired reaction temperature. The volume composi-
tion of the gaseous reactant feed was 6% CH3OH, 13% O2, and
balance He, with a total flow rate of ∼100 mL/min. The methanol
conversion and reaction products were analyzed using an online
gas chromatograph (HP 5890 series II) equipped with TCD and
FID detectors. A Carboxene-1000-packed column and a CP-sil 5CB
capillary column were used in parallel for TCD and FID, respec-
tively. The catalytic turnover frequencies (TOFs) for methanol
(18) Gao, X.; Wachs, I. E. J. Phys. Chem. 2000, 104, 1261.
(19) Ross-Medgaarden, Wachs, I. E. J. Phys. Chem. C 2007, 111, 15089.
(20) Briand, L. E.; Farneth, W. E.; Wachs, I. E. Catal. Today 2000, 62,
219.
(21) Eckert, H.; Wachs, I. E. J. Phys. Chem. 1989, 93, 6796.
9
15546 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Oxidation and Dehydration of CH3OH by POMs
A R T I C L E S
P-Oa band vibration is not observed in the Raman spectrum,
which is most probably related to its weaker Raman cross
section compared to the strong WdOt vibrations. The P-Oa
vibration, however, is not generally observed with Raman
spectroscopy for polyoxometallates in the solid form, but such
a weak Raman band has been detected for molybdophosphoric
acid in aqueous solution.25 The Raman and IR vibrational
assignments are summarized in Table 1.
The supported VOTPA catalyst exhibits the main infrared
bands (Figure 3) characteristic of the Keggin anion at 1080
(P-Oa asymmetric), 981 (WdOt asymmetric), 888 (W-Ob-W
asymmetric), and 803 (W-Oc-W asymmetric) cm-1. When
vanadium is incorporated into the primary Keggin structure in
TPAV1, a shift is observed and the P-Oa band splits (see also
Table 2). Such a shift is not observed for the supported VOMPA
sample where vanadium is presented in the secondary structure
as a surface VOx species. The TPAV1 catalyst with vanadium
in the Keggin primary structure also shows a shoulder band at
1097 cm-1. This infrared band is due to the asymmetric vibration
of the P-Oa-V bond.26 The splitting of the P-Oa band is due
to the introduction of a cation other than W into the Keggin
structure, which induces a decrease in the W-Ot stretching
frequency, leading to splitting of the P-Oa band,27 thus
confirming the incorporation of vanadium into the primary
Keggin structure for TPAV1. The incorporation of vanadium
into the primary TPA structure leads to a shift and splitting of
the main FT-IR band due to the reduced structural symmetry
as shown by the vibrational bands listed in Table 2.17 A clear
shift in the P-Oa vibration (Oa denotes an oxygen atom bound
to three W atoms and to P) is noticed from 1080 cm-1 for TPA
to 1069 cm-1 for the TPAV3 sample. This shift in the P-Oa
band position reflects the incorporation of vanadium into the
primary structure of the tungstophosphoric acid Keggin.
In Situ Raman Spectroscopy under Dehydrated Con-
ditions. The in situ Raman spectra of the dehydrated POM
Keggins at 200 °C are contained in Figure 4 and are normalized
against the TPA Keggin W-O-W bending modes at ∼220-250
cm-1. No additional Raman bands from the VdOt vibrations
are observed because of the stronger Raman features of the
WdOt vibrations that are broadened by the presence of VOx.
The dehydrated Raman spectra exhibit a shift of the WdOt
vibration from 1011 to 1022 cm-1 due to the loss of crystal
water.28 Such ModOt vibrational shifts upon dehydration have
previously been reported for H3SiMo12O40 and are simply related
to desorption of water and removal of its associated hydrogen
bonding (MdOt · · ·HsOH).29 There is no significant shift in
the position of all other bands upon dehydration since dehydra-
tion primarily affects the terminal WdOt vibrations.
Figure 2. Raman spectra of V-containing tungstophosphoric acid,
H3+xPW12-xVxO40 (x ) 0, 1, 2, and 3), and VOTPA (supported VOx/
H3PW12O40) polyoxometallates under ambient conditions. The Raman
spectra were normalized against Keggin W-O-W bending modes at
∼220-250 cm-1
.
catalyst is consistent with the very low pH at the PZC of this
sample (PZC ≈ 0.5).22
Raman and FT-IR Spectroscopy under Ambient Con-
ditions. The Raman spectra of the H3+xPW12-xVxO40 (x ) 0, 1,
2, and 3) and VOTPA mixed oxide Keggins under ambient
conditions are presented in Figure 2, and their assignments are
given in Table 1. The V-free TPA Raman spectrum exhibits
strong WdOt symmetric and asymmetric stretching modes at
1011 and 990 cm-1, respectively. The asymmetric stretch is a
consequence of the presence of vibrationally coupled adjacent
WdOt bonds in the Keggin structure. The weaker Raman bands
at 890 and 530 cm-1 are characteristic of the asymmetric
stretching vibration of bridging W-Ob-W (Ob denotes a corner-
sharing bridging oxygen atom) and symmetric stretching of
bridging W-Oc-W (Oc denotes an edge-sharing bridging
oxygen atom), respectively. The Raman band at ∼240 cm-1
represents the bending mode of the bridging W-O-W bonds
of the intact Keggin structure. These Raman spectra agree with
those reported in the literature.23,24 The vanadia-substituted TPA
and VOTPA mixed oxides also exhibit Raman bands that are
similar to those of the TPA Keggin, but the resolution of the
WdOt vibrations at 1011 and 990 cm-1 decreases as the vanadia
content is increased. The similar Raman bands for TPA, TPAV,
and VOTPA reveal that the Keggin structure remains intact upon
incorporation of vanadia into the primary and secondary
structures.
The addition of the surface VOx species to the TPA in
supported VOTPA results in enhanced Raman bands at ∼984
and ∼930 cm-1 as well as a slight shift of the WdOt vibration
to lower wavenumbers that reflects perturbation of the WOx
vibrations in the primary Keggin structure by the presence of
surface VOx species. The shift from 1022 to 984 cm-1 appears
The corresponding FT-IR spectrum of tungstophosphoric acid
under ambient conditions is presented in Figure 3 and shows
bands at 1080, 981, 888, 797, and 595 cm-1 that are related to
asymmetric vibrations of P-Oa, WdOt, W-Ob-W, and
W-Oc-W and the bending mode of Oa-P-Oa, respectively.
The FT-IR band assignments are given in Table 1. The oxygen
atom bridging the central P and W atoms, P-Oa-W, is denoted
as Oa. Comparison of the Raman and IR bands reveals that IR
gives rise to a stronger band than Raman at 1080 cm-1 related
to the P-Oa asymmetric stretch. The corresponding symmetric
(25) Mestl, G.; lkenhans, I. T.; Spielbauer, D.; Dieterle, M.; Timpe, O.;
Kro¨hnert, J.; Jentoft, F.; Kno¨zinger, H.; Schlo¨gl, R. Appl. Catal., A
2001, 210, 13.
(26) Shinach, S.; Matsushita, M.; Yamaguchi, K.; Mizuno, N. J. Catal.
2005, 233, 81.
(27) Rocchiccioli-Deltcheff, C.; Fournier, M. J. Chem. Soc., Faraday Trans.
1991, 87, 3913.
(22) Deo, G.; Wachs, I. E. J. Phys. Chem. 1991, 95, 5889.
(23) Busca, G. J. Raman Spectrosc. 2002, 33, 348.
(24) Caliman, E.; Dias, J. A.; Dias, S. C. L.; Prado, A. G. S. Catal. Today
2000, 107-108, 816.
(28) Brown, G. M.; Noe-Spirlet, M. R.; Busing, W. R.; Levy, H. A. Acta
Crystallogr. 1977, B33, 1038.
(29) Banares, M. A.; Hu, H.; Wachs, I. E. J. Catal. 1995, 155, 249.
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15547

A R T I C L E S
Nakka et al.
Table 1. Comparison of FT-IR and Raman Spectra for Tungsten- and Vanadium-Containing Polyoxometallate Catalysts under Ambient
Conditionsa
FT-IR
Raman
sample
band position (cm-1
)
assignment
band position (cm-1
)
assignment
H3PW12O40
1080 (s)
P-Oa asymmetric stretch
WdOt asymmetric stretch
1011 (s)
990 (m)
930 (w)
890 (vw)
WdOt symmetric stretch
WdOt asymmetric Stretch
W-Ob-W asymmetric stretch
984 (s)
888 (s)
797 (m)
595 (w)
W-Ob-W asymmetric stretch
W-Oc-W asymmetric stretch
Oa-P-Oa asymmetric bending
530
W-Oc-W symmetric stretch
W-O bending
W-O-W bending
325 (vw)
225 (m)
242 (m)
148 (w)
lattice phonon
H3PW11V1O40
1097 (vw)
1078 (s)
(P-Oa-V) asymmetric stretch
P-Oa asymmetric stretch
1011 (s)
990 (m)
930 (w)
890 (vw)
WdOt symmetric stretch
WdOt asymmetric stretch
981 (s)
WdOt asymmetric stretch
886 (s)
797 (m)
595 (w)
W-Ob-W asymmetric stretch
W-Oc-W asymmetric stretch
Oa-P-Oa bending
W-Ob-W asymmetric stretch
530 (w)
325 (vw)
W-Oc-W symmetric stretch
W-O bending
225 (vw), 242 (m)
148 (w)
W-O-W bending
lattice phonon
a Key: Oa, oxygen atom bound to three W atoms and to P; Ot, terminal oxygen atom; Ob, corner-sharing bridging oxygen atom; Oc, edge-sharing
bridging oxygen atom; s, strong; m, medium; w, weak; vw, very weak.
Figure 3. FT-IR spectra of hydrated TPA, VOTPA, and TPAV1 catalysts
Figure 4. In situ Raman spectra of dehydrated V-containing tungstophos-
under ambient conditions.
phoric acid, H3+xPW12-xVx (x ) 0, 1, 2, and 3), and VOTPA (supported
VOx/H3PW12O40) polyoxometallates (200 °C). The Raman spectra were
Table 2. Comparison of FT-IR Stretching Frequencies of the Main
normalized against Keggin W-O-W bending modes at ∼220-250 cm-1
.
Characteristic Bands for TPA and Vanadium-Containing
Tungstophosphoric Acid (TPAV1, TPAV2, and TPAV3) Catalysts
under Ambient Conditionsa
bonds become vibrationally coupled.30 Thus, the new Raman
band at ∼984 cm-1 is related to the symmetric WdOt stretch
of tungsten oxide interacting with surface VOx species, and that
at ∼930 cm-1 is either the associated asymmetric stretch or the
perturbed asymmetric W-O-W stretch of the TPA Keggin.
The in situ Raman spectra of TPA, TPAV, and VOTPA
POMs as a function of calcination temperature from 300 to 600
°C in flowing O2/He were also examined (see the Supporting
Information, Figures SI-1-SI-5). The Raman spectra of these
POMs at 300 °C are similar to their dehydrated spectra at 200
°C shown in Figure 4, with only a small difference in the
spectrum of TPAV3, which demonstrates that the POM Keggin
IR stretching frequency (cm-1
)
sample
P-Oa
WdOt
W-Ob-W
W-Oc-W
TPA
1080
1078
1073
1069
984
981
980
978
890
886
881
876
800
797
795
795
TPAV1
TPAV2
TPAV3
a Key: Oa, oxygen atom bound to three W atoms and to P; Ot,
terminal oxygen atom; Ob, corner-sharing bridging oxygen atom; Oc,
edge-sharing bridging oxygen atom.
to be related to perturbation of the vibrational coupling of
adjacent WdOt bonds since the WOt vibration for supported
tungsten oxide catalysts also shifts from ∼990 to 1020 cm-1 as
the surface WOx species become polymerized and the WdOt
(30) Kim, T.; Wachs, I. E. J. Catal. 2008, 255, 197.
9
15548 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Oxidation and Dehydration of CH3OH by POMs
A R T I C L E S
structures are thermally stable up to ∼300 °C. The vanadia-
free TPA POM is thermally stable in this temperature range,
and the bands thermally broaden as 600 °C is approached. In
contrast, the V-containing POMs begin to decompose at
temperatures of 400-500 °C. Decomposition of the V-contain-
ing Keggin structures is accompanied by the appearance of a
small band at ∼1034 cm-1 that is characteristic of the VdO
vibration of the surface VOx species in the secondary structure31,32
and is accompanied by broad bands at 690 and 790 cm-1 that
become sharper with increasing temperature from formation of
crystalline WO3 nanoparticles (NPs).33 These observations
demonstrate that incorporation of vanadia into the TPA primary
and secondary structures leads to destabilization of the TPA
Keggin during thermal treatments at elevated temperatures.
Thus, the presence of surface VOx species enhances the
destabilization of the TPA Keggin at elevated temperatures. All
the TPAV and the supported VOPTPA catalysts, however, are
thermally stable below 300 °C.
The WdOt and VdOt bond lengths in the Keggins were also
determined from established W-O and V-O bond length-
vibration diatomic correlations.28,31,34 The symmetric WdOt
vibration occurs at ∼1022 cm-1 for the dehydrated H3PW12O40,
H3+xPW12-xVxO40, and supported VOx/H3PW12O40 catalysts. The
vibration at 1022 cm-1 corresponds to a WdOt bond length of
0.1698 nm, and the vibration at 534 cm-1 corresponds to a bond
length of 0.2039 nm. The symmetric VdOt vibration for the
dehydrated H3+xPW12-xVxO40 and supported VOx/H3PW12O40
catalysts is not detectable because it is overshadowed, and
possibly vibrationally coupled to WdOt, by the strong sym-
metric WdOt vibration. If it is assumed that the symmetric
VdOt vibration also occurs at 1022 cm-1, this results in a VdOt
bond length of 0.1585 nm. The surface VOx species expelled
from the primary Keggin structure give rise to a band at ∼1034
cm-1 that corresponds to a VdOt bond length of 0.1579 nm.
These WdOt and VdOt vibrations and short bond lengths
correspond to monooxo tungsten oxide and vanadium oxide
molecular structures, resepctively.35
In Situ UV-Vis DRS under Dehydrated Conditions. The in
situ UV-vis DRS spectra of TPA, TPAV, and the supported
VOTPA catalyst samples are presented in Figure 5 plotted as
the Kubelka-Munk function using the methods described by
Gao18 and Ross-Medgaarden.19 The ligand to metal charge
transfer (LMCT) transition for the W6+ cation appears between
250 and 300 nm, and the LMCT of the V+5 cation appears
between 350 and 450 nm. Introduction of vanadium into the
TPA Keggin introduces a new band between 350 and 450 nm
with an Eg value that is ∼1 eV lower than that for TPA (see
Table 3 for exact values). The Eg values for the TPAV catalyst
samples are comparable, and that for the VOTPA Keggin is
only slightly lower.
Figure 5. In situ UV-vis spectra of dehydrated TPA and V-containing
tungstophosphoric acid catalysts (T ) 225 °C).
Table 3. UV-Vis DRS Edge Energy Values (eV) for TPA and the
Vanadium-Containing Tungstophosphoric Acid (VOTPA, TPAV1,
TPAV2, and TPAV3) Catalysts under Different Environmental
Conditions
edge energy (eV)
in situ during MeOH oxidation
at 225 °C
catalyst
dehydrated at 225 °C
TPA
3.22
2.05
2.30
2.34
2.28
3.23
1.93
2.25
2.25
2.08
VOTPA
TPAV1
TPAV2
TPAV3
from the reaction of surface CH3* and CH3O* intermediates.20,36
The CH3OH-TPSR spectrum of the vanadium-free TPA Keggin
catalyst exclusively yields CH3OCH3 with Tp ) 182 °C, which
reflects the surface Bronsted acid sites present on this POM
Keggin.37 The addition of vanadium oxide to the TPA Keggin
does not change the CH3OCH Tp value, indicating that the same
surface Bronsted acid sites are also present for the TPAV
Keggins. The incorporation of vanadium oxide into the TPA
Keggin structure results in the formation of H2CO at a Tp ≈
205 °C from the surface redox sites (see Figure 6). The HCHO/
DME formation ratio during CH3OH-TPSR increases linearly
from TPAV1 to TPAV3 as indicated in Table 4.
The supported VOTPA catalyst, however, indicates a sig-
nificantly less active V-containing redox site. The CH3OH-TPSR
spectrum for the supported VOTPA catalyst is also shown in
Figure 6 and yields significant amounts of H2CO with a HCHO/
DME formation ratio intermediate between those of TPAV1
and TPAV2. The surface VOx species on TPA, however,
significantly affect the kinetics for breaking the surface methoxy
C-H bond by shifting the H2CO Tp value from 205 to 245 °C.
This reveals that the vanadium oxide redox sites incorporated
into the primary Keggin structure in the TPAV catalysts are
significantly more active than the surface VOx redox sites present
on VOTPA for surface methoxy C-H bond breaking. A closer
examination of the H2CO/CH3OH-TPSR spectra in Figure 6
CH3OH-TPSR Spectroscopy. The CH3OH-TPSR experiments
were conducted to chemically probe the surface chemistry of
the TPA, TPAV, and VOTPA catalysts, and the results are
presented in Figure 6. It is well-known that methanol chemi-
sorption at 100 °C on oxides forms surface CH3O* intermediates
that dehydrogenate to form H2CO on surface redox sites and
deoxygenate on surface acid sites to form CH3OCH3 (DME)
(31) Hardcastle, F. D.; Wachs, I. E. J. Phys. Chem. 1991, 95, 5031.
(32) Gao, X.; Jehng, J.-M.; Wachs, I. E. J. Catal. 2002, 209, 43.
(33) Scheithauer, M.; Grasselli, R. K.; Knozinger, H. Langmuir 1998, 14,
3019.
(36) Burcham, L. J.; Briand, L. E.; Waches, I. E. Langmuir 2001, 17, 617.
(37) Moffat, J. B. Metal-Oxygen Cluster: The Surface and Catalytic
Properties of Heteropolyoxometalates; Kluwer Academic-Plenum:
New York, 2001.
(34) Hardcastle, F. D.; Wachs, I. E. J. Raman Spectrosc. 1995, 26, 397.
(35) Lee, E. L.; Wachs, I. E. J. Phys. Chem. C 2007, 111, 4410.
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15549

A R T I C L E S
Nakka et al.
Figure 6. CH3OH-TPSR spectra for V-containing tungstophosphoric acid catalysts.
Table 4. Relative H2CO/DME Ratio during CH3OH-TPSR from the
Tungsta-Based Polyoxometallate Catalysts
Table 5. Steady-State Methanol Oxidation Data over the
Tungsta-Based Polyoxometallate Catalysts
HCHO/DME ratio
(arbitrary units)
selectivity
TOF (s-1
)
activity
conversion (mol/(g of
(%) catalyst s)) DME HCHO acid
catalyst
Tp of HCHO (°C)
Tp of DME (°C)
temperature
(°C)
catalyst
redox
TPA
182
190
185
180
182
0.0
2.1
4.2
8.8
3.1
TPA
225
18
22
8
9.6
14.7
7.4
100
82 18 7.6
90 10 0.46 0.010
100 1.9 0.00
0
3.9
0.0
TPAV1
TPAV2
TPAV3
VOTPA
205
205
206
245
TPAV1 225
TPAV2 225
TPAV3 225
VOTPA 250
0.32
5.6
23
3.0
12.5
0
88 12 0.12 0.0054
0.046 0.0021
VOTPA 225 (extrapolated)
reveals that surface VOx sites are also present in the TPAV
catalysts since they also exhibit a significant H2CO formation
tail in the 245 °C region. The TPAV1 catalyst has only a minor
amount of surface VOx species, but the TPAV2 and TPVA3
catalysts contain appreciable quantities of surface VOx sites as
judged from the trailing tails of the HCHO/CH3OH-TPSR
spectra.
CH3OH Oxidation/Dehydration. The steady-state methanol
oxidation/dehydration reaction over TPA, TPAV, and supported
VOTPA catalysts was also examined in the 200-250 °C
temperature range where the TPAV catalysts are known to be
thermally stable (see the Supporting Information, Figures SI-
1-SI-5). The reaction results for TPA and the TPAV catalysts
at 225 °C are presented in Table 5. The dominant reaction
product from the TPA and TPAV catalysts is DME from the
surface acid sites, and H2CO is the main product from surface
redox sites. Small amounts of methyl formate and dimethoxyme-
thane were also formed by secondary reaction of formaldehyde
with surface redox and acid sites, and carbon oxide products
were not observed. The redox selectivity, TOF values, and
TOFredox/TOFacid ratio do not monotonically increase with
increasing vanadium content of the TPA. The TPAV3 catalyst
with the highest vanadium content does not even form HCHO
9
15550 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Oxidation and Dehydration of CH3OH by POMs
A R T I C L E S
Figure 7. In situ Raman spectra of vanadium-containing tungstophosphoric acid catalysts during methanol oxidation/dehydration reaction conditions (T )
225 °C, O2/CH3OH ) 2.17). The Keggin W-O-W bending modes at ∼220-250 cm-1 are employed as the internal standard.
In Situ Raman Spectroscopy during CH3OH Oxidation/
Dehydration. To better understand the state of these V-
containing TPA-based catalysts during steady-state CH3OH
oxidation, in situ Raman spectroscopy studies were conducted
under the methanol oxidation/dehydration reaction conditions.
The in situ Raman spectra of the TPA, TPAV, and supported
VOTPA catalysts during methanol oxidation/dehydration as a
function of reaction temperature in the 200-250 °C range,
where the TPA, TPAV, and supported VOTPA catalyst samples
are thermally stable (see the Supporting Information, Figures
SI-1-SI-5), are presented in Figure 7. The Raman spectrum
for the TPAV1 catalyst exhibits new bands at ∼1350 and ∼1600
cm-1 as well as the characteristic bands of the Keggin TPA
structure under methanol oxidation/dehydration conditions. The
new Raman bands at ∼1350 and ∼1600 cm-1 correspond to
polyaromatic coke and reveal that carbon deposition on the TPA
catalyst is taking place during methanol oxidation.40 The amount
of carbon deposited on the TPAV catalyst decreases with
increasing content of vanadium as shown for TPAV1, TPAV2,
and TPAV3, and there is almost no coke present for TPAV3
during methanol oxidation. The amount of surface carbon further
during steady-state reaction, and DME is the only reaction
product. The TOF values are highest for the TPAV1 catalyst
and decrease with increasing vanadia content for both the surface
redox and acid sites, especially for the TOFredox kinetics. Similar
activity and selectivity trends have also been found for
vanadium-containing molybdenum POMs (MPAs).9,38,39
The supported VOTPA catalyst was found to be barely active
for methanol oxidation at 225 °C, requiring reaction studies at
250 °C for appreciable methanol conversion (see Table 5). The
TOF for supported VOTPA was experimentally determined at
250 °C and extrapolated to 225 °C, which is also reported in
Table 5. The significantly lower catalytic activity of the
supported VOTPA catalyst further confirms that surface VOx
sites in the secondary structure of the TPA Keggin are
significantly less active than the VOx sites incorporated into the
primary Keggin structure and is consistent with the HCHO/
CH3OH-TPSR studies reported above. Thus, at the reaction
temperature of 225 °C the surface VOx species in the secondary
structure are essentially poisons that shut off catalytic activity.
(38) Liu, H.; Iglesia, E. J. Phy. Chem. B 2003, 107, 10840.
(39) Liu, H.; Iglesia, E. J. Catal. 2004, 223, 161.
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15551

A R T I C L E S
Nakka et al.
decreases with increasing reaction temperature. Surprisingly,
only a small amount of carbon is deposited on the acidic TPA
catalyst. The VOTPA catalyst exhibits modest amounts of
carbon deposition and is almost free of coke at 250 °C.
and the rapid reoxidation kinetics of the vanadia sites by gas-
phase molecular O2.
Discussion
The corresponding in situ Raman bands from the Keggin unit
provide structural information about the TPA Keggins under
the methanol oxidation reaction conditions (200-250 °C). The
characteristic Keggin Raman bands are present under reaction
conditions for TPA, TPAV1, and TPAV2 catalysts and reflect
the structural integrity of the Keggin unit for these mixed metal
oxide catalysts during methanol oxidation in this temperature
range. The TPAV2 catalyst, however, also exhibits a broad
shoulder on the low-wavenumber side of the major WdOt
vibrational band. Further increasing the vanadium content to
TPAV3 significantly broadens the Keggin vibration at 1022
cm-1 and also reveals the formation of a new broad band at
∼970 cm-1. The weakening and broadening of the primary
Keggin vibrations for TPAV2 and TPAV3 reflects the structural
disorder introduced by the presence of increasing amounts of
vanadium in the primary Keggin unit. Such structural disorder
of the primary Keggin unit is not observed for the TPAV1 and
VOTPA catalysts where vanadium oxide has been substituted
into the Keggin unit and is present as surface VOx species,
respectively.
Bulk or Primary Keggin Structures. The H3+xPW12-xVxO40
Keggins retain the primary TPA structure upon incorporation
of vanadium oxide. Broadening of the Raman bands indicates
that the incorporated vanadium oxide is creating structural
disorder in the primary Keggin structure, most likely by
occupying different primary sites. Solid-state MAS 51V NMR
confirms that the incorporated vanadia possess VO6 coordina-
tion, which reflects substitution of VO6 for WO6 in the TPA
Keggin structure. UV-vis DRS reveals that the Keggin Eg value
decreases by ∼1 eV by incorporation of vanadia into the TPA
Keggin. When the Eg values for the dehydrated TPAV and
VOTPA Keggins are evaluated using the V-O-V correlation
(N ) 14.03 - 3.95Eg),18 the correlation indicates that four
bridging V-O-V bonds are present for every V-containing
TPA Keggin. However, the V-containing TPA Keggins mostly
contain V-O-W and not V-O-V bonds because of the
minority concentration of vanadium oxide. This indicates that
the correlation published by Gao et al.18 also holds up when
the vanadium cation is bonded to other metal oxide cations such
as tungsten oxide.
Dehydration of the Keggins results in removal of crystal water
and shifting of the WdOt vibrations from 1011 to 1022 cm-1
.
The presence of the surface VOx species on the TPA support
in supported VOTPA, however, gives rise to additional sharp
Raman bands at ∼970 and 870 cm-1 during methanol oxidation
at elevated temperatures. These vibrational bands have been
shown above to occur when surface VOx species perturb the
asymmetric vibrations of the WdOt and W-Ob-W function-
alities. The shift of the WdOt vibration from 1022 to 970 cm-1
is related to perturbation of the vibrational coupling of adjacent
WdOt bonds. The enhanced 870 cm-1 band is probably related
to perturbation of the W-Ob-W vibration by the surface VOx
species and/or surface methoxy species. The presence of the
∼970 cm-1 band in the Raman spectra of TPAV2 and TPAV3
also reflects the presence of surface VOx species in the secondary
structure of the Keggin units in these samples. The enhanced
perturbations of the TPA Keggin WdOt and W-Ob-W
vibrations during methanol oxidation reveal that the surface VOx
species are well dispersed on the TPA and TPAV supports.
These in situ Raman findings are in agreement with the above
CH3OH-TPSR spectra (see Figure 6 and discussion above),
indicating that surface VOx species are present on the outer
surface or secondary structure of the TPAV Keggin units,
especially for TPAV2 and TPAV3.
Similar shifts in the infrared spectra of WdOt bonds of
H3PW12O40 upon dehydration were previously proposed to result
from proton movement from the terminal WdOt bond to the
bridging W-O-W bond,41 but such changes in the WdOt
vibrations are routinely observed for supported tungsten oxide
catalysts and just reflect the removal of hydrogen-bonded water
from the WdOt bond upon dehydration.42,43 The Raman bands
of the dehydrated V-containing Keggins are also broadened
relative to those of TPA due to structural disorder, and the
broadening increases with increasing vanadium content. The
addition of vanadium oxide to the TPA Keggin results in thermal
decomposition of the H3+xPW12-xVxO40 Keggins in the 400-500
°C range, with greater instability paralleling increasing vanadium
oxide content. The origin of this thermal instability is most
probability related to the much lower Tammann temperature
of vanadium oxide (∼200 °C) compared to tungsten oxide
(∼600 °C), which is responsible for the enhanced mobility of
vanadium oxide. Decomposition of the V-containing TPA
Keggins is accompanied by expulsion of VOx, exhibiting a
VdOt vibration at ∼1034 cm-1, and formation of poorly ordered
WO3 NPs (Raman vibrations at ∼690 and ∼790 cm-1).
During methanol oxidation/dehydration the in situ Raman
bands of the primary Keggins further broaden with vanadium
oxide content relative to those of the dehydrated catalysts. This
may be related to the enhanced mobility of vanadium oxide by
formation of mobile surface V-OCH3 complexes44 and the
coordination of surface methoxy species. The corresponding in
situ UV-vis DRS spectra show that the vanadium oxide sites
remain essentially fully oxidized during methanol oxidation/
dehydration at 250 °C. Thus, the primary structures of the
In Situ UV-Vis DRS during Methanol Oxidation/Dehy-
dration. The in situ UV-vis spectra were also obtained during
methanol oxidation/dehydration at 225 °C (see the Supporting
Information, Figures SI-6-SI-9). There is an indication of only
a minor reduction in the V5+ LMCT and a slight increase in
the V3+/V4+ d-d transition region, most probably related to
the small carbon deposition that also gives rise to a UV-vis
signal in this region, indicating that the vanadium oxide sites
in the TPA Keggin are primarily present as V5+ during methanol
oxidation/dehydration at 225 °C (O2/CH3OH ) 2.17). This
behavior is consistent with the Mars-van Krevelen mechanism
(41) Lee, K. Y.; Mizuno, N.; Okuhara, T.; Misono, M. Bull. Chem. Soc.
Jpn. 1989, 62, 1731.
(42) Kim, D. S.; Ostromecki, M.; Wachs, I. E. J. Mol. Catal. A 1996, 106,
93.
(43) Kim, T.; Burrows, A.; Kiely, C. J.; Wachs, I. E. J. Mol. Catal. 2007,
246, 370.
(40) Li, C.; Stair, P. C. Catal. Today 1997, 33, 353.
(44) Wang, C.-B.; Cai, Y.; Wachs, I. E. Langmuir 1999, 15, 1223.
9
15552 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009

Oxidation and Dehydration of CH3OH by POMs
A R T I C L E S
V-containing TPA Keggins remain intact during methanol
oxidation, but are structurally disordered by the introduction of
vanadium oxide and the reaction conditions.
The surface redox sites in the TPAV Keggins break the
surface methoxy C-H bond at ∼205 °C. This is consistent with
corresponding CH3OH-TPSR results for bulk V2O5 (Tp ) 200
°C)46 and supported V2O5 catalysts (Tp ≈ 180-190 °C),30,49
which indicates comparable redox kinetics for breaking the
surface methoxy C-H bond for the V-containing TPA Keggins,
bulk V2O5, and supported vanadium oxide catalysts. The surface
VOx species present in the secondary Keggin structure of
supported VOTPA catalysts, however, exhibit a Tp ≈ 245 °C,
indicating much lower redox activity of the surface VOx species
compared to the VOx sites incorporated in the primary TPA
Keggin structure.
The VOx sites essentially remain fully oxidized during the
methanol oxidation/dehydration reactions with an O2/CH3OH
ratio of 2.17 (see the Supporting Information, Figures SI-6-SI-
9). This indicates that the methanol oxidation kinetics follows
a Mars-van Krevelen mechanism where the reduced vanadium
oxide sites are rapidly reoxidized by gas-phase molecular O2.
Surface or Secondary Keggin Structures. The V-containing
TPA Keggins also possess significant surface VOx sites in their
initial synthesized state, which are distinct from expelled surface
VOx species on defective Keggins (Raman band at ∼1034 cm-1)
observed at elevated temperatures (see the Supporting Informa-
tion, Figures SI-2-SI-4). The presence of surface VOx species
in the supported VOTPA Keggin perturbs the vibrational
coupling of adjacent WdOt bonds and shifts the symmetric
WdOt vibration toward 970 cm-1. The W-Ob-W vibration at
∼870 cm-1 is also perturbed by the surface VOx and methoxy
species. These vibrations have previously been assigned to
defective Keggin structures,41 but the current study has shown
that expulsion of VOx from the TPAV Keggins is not ac-
companied by enhanced Raman bands at ∼970-990 and ∼870
cm-1 (see the Supporting Information, Figures SI-2-SI-4). It,
thus, appears that the surface VOx species are perturbing the
symmetric WdOt vibration. These perturbations by the surface
VOx species are further enhanced during methanol oxidation
and may be related to the greater mobility of surface V-OCH3
complexes and the presence of surface W-OCH3 species (see
Figure 7).
Polyaromatic carbon deposition on the surface of the TPAV
Keggins also takes place during steady-state methanol oxidation
(see Raman bands at ∼1360 and ∼1600 cm-1 in Figure 7).40
The extent of carbon deposition is related to the amount of
vanadium oxide in the TPAV Keggin since the V-free TPA
exhibits a negligible amount of carbon deposition and the
amount of surface carbon decreases with vanadium content
(TPAV1 > TPAV2 > TPAV3). Thus, the vanadium oxide sites
appear to be involved in both the carbon formation and carbon
combustion reactions. The amount of surface carbon present
during steady-state methanol oxidation at 225 °C, however, must
be minimal since the UV-vis DRS background in the 400-800
nm region was not significantly perturbed by the deposited
carbon at the reaction temperature of 225 °C (see the Supporting
Information, Figures SI-6-SI-9).45
Structure-Catalytic Activity Relationships. There are two
distinct VOx redox sites in the TPAV catalysts: VOx sites in
the primary TPA Keggin structure and VOx sites on the surface
or secondary TPA Keggin structure (see the above discussion
in the “Surface or Secondary Keggin Structures” section). The
steady-state methanol oxidation to formaldehyde studies at 225
°C revealed that VOx sites in the primary Keggin structure
exhibit a TOF that is ∼160 times greater than that of the surface
VOx sites. The continuous decrease in TOFredox in going from
TPAV1 to TPAV2 to TPAV3 reflects the increasing concentra-
tion of surface VOx sites in the secondary structure of the
Keggins with increasing vanadium oxide content. The catalytic
poisoning effect of surface VOx species in the secondary
structure of V-containing Keggin catalysts was not previously
fully appreciated in the literature.9,10,38,39
The minor amount of surface carbon deposition may also
affect the catalytic activity, and the amount of surface carbon
decreases with vanadium oxide content, which follows the
opposite of the activity trend, and its quantitative effect on the
TOFredox is not known. The contribution of the surface carbon
to the methanol oxidation kinetics over the H3+xPW12-xVxO40
Keggins, however, can be neglected in a first approximation
because of the inverse relationship between TOFredox and
vanadium oxide content and its minor amounts.
Comparison of the TOFredox kinetics with the UV-vis DRS
Eg values reveals that a relationship between the electronic
structure and TOFredox kinetics for the H3+xPW12-xVxO40 Keggins
does not exist. Although the addition of vanadium oxide to the
TPA Keggins introduces redox sites and lowers the Eg value,
which reflects enhanced electron delocalization, there is no
correlation with the TOFredox kinetics. The UV-vis Eg values
for the TPAV1, TPAV2, and TPAV3 Keggins are comparable,
but their corresponding TOFredox kinetics decrease by orders of
magnitude. The supported VOTPA catalyst possesses a lower
Eg value than the TPAV catalysts and is even less active for
methanol oxidation to formaldehyde, which is the opposite of
what would be anticipated for its lower Eg value and greater
electron delocalization. The current UV-vis DRS and steady-
state methanol oxidation kinetics demonstrate that correlations
between Eg and TOFredox for Keggins should be carefully viewed.
Comparison of the methanol oxidation to formaldehyde
reaction rates for the H3+xPW12-xVxO40 and supported VOx/
H3PW12O40 Keggins with unsupported V2O5 and supported V2O5
catalysts provides interesting insights into the catalytic activity
of these V-containing Keggin catalysts. The steady-state metha-
nol oxidation specific reaction rate for TPAV1 (TOFredox ) 0.3
s-1 at 225 °C) is comparable to that found for bulk V2O5
Surface Chemistry. The surface acid sites present in the
H3PW12O40 Keggin are extremely active for breaking the surface
methoxy C-O bond involved in the formation of CH3OCH3
compared to surface acid sites present on other tungsten oxide-
based catalysts. The DME/CH3OH-TPSR of Tp ≈ 180 °C
reflects much faster surface kinetics than found for unsupported
WO3 (Tp ≈ 260 °C),46,47 supported WO3/ZrO2 with monolayer
surface WOx coverage (Tp ≈ 290 °C),47 and even the highly
acidic WO3-ZrO2 catalyst (Tp ≈ 216 °C) produced by reacting
ammonium metatungstate with ZrO(OH)2.48,49 The supported
WO3/Al2O3 catalyst system (Tp ≈ 206 °C), however, is slightly
more active than the other supported WO3 catalysts, but is still
less active than the H3PW12O40 Keggin (Tp ≈ 180 °C).
(45) Tinnemans, S. J.; Kox, M. H. F.; Sletering, M. W.; Nijhuis, T. A.;
Visser, T.; Weckhuysen, B. M. Phys. Chem. Chem. Phys. 2006, 8,
2413.
(48) Ross-Medgaarden,; Knowles, W. V.; Kim, T.; Wong, M. S.; Zhou,
W.; Kiely, C. J.; Wachs, I. E. J. Catal. 2008, 256, 108.
(49) Kim, T. Ph.D. Thesis, Lehigh University, 2007.
(46) Badlani, M.; Wachs, I. E. Catal. Lett. 2001, 75, 1.
(47) Kim, T.; Wachs, I. E. J. Catal. 2007, 246, 370.
9
J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009 15553

A R T I C L E S
Nakka et al.
(TOFredox ) 1.1 s-1 at 230 °C),46 supported V2O5/CeO2 (TOFredox
) 1.0 s-1 at 230 °C),50 supported V2O5/ZrO2 (TOFredox ) 0.17
s-1 at 230 °C),50 and supported V2O5/TiO2 (TOFredox ) 0.11
s-1 at 230 °C).50 The TPAV1 TOFredox value, however, is much
under ambient conditions. Introduction of the VOx units into
the primary H3+xPW12-xVxO40 Keggin structure causes structural
disorder and lowers the thermal stability of the Keggin clusters,
which is much less pronounced for the surface VOx species or
vanadium oxide in the secondary structure.
greater than found for the supported V2O5/Al2O3 (TOFredox
)
0.007 s-1 at 230 °C)48-50 and supported V2O5/SiO2 (TOFredox
) 0.002 s-1 at 230 °C) catalysts.50 Thus, the specific reaction
rate for TPAV1, which contains minimal amounts of surface
VOx species in the secondary structure of the H3+xPW12-xVxO40
Keggin, is comparable to the most active V-containing catalysts.
The surface VOx species in the secondary structure of the
H3+xPW12-xVxO40 Keggin and coordinated to the Keggin WOx
sites, however, possess an extremely low specific redox activity
(TOF ) 0.002 at 225 °C) that is comparable to that of the least
active supported V2O5/SiO2 catalyst (TOFredox ) 0.007 at 230
°C).
Comparison of the steady-state methanol dehydration kinetics
by the H3+xPW12-xVxO40 Keggins with those for bulk WO3 and
supported WO3 catalysts provides additional insights into the
catalytic acidic properties of the W-containing Keggins. The
dehydration kinetics by the H3PW12O40 Keggin (TOFacid ) 3.9
s-1 at 225 °C) is significantly faster than that for bulk WO3
(TOFacid ) 0.03 s-1 at 230 °C)47,49 and the supported WO3
catalysts (TOFacid ≈ 0.02-0.08 s-1 at 230 °C for WO3/ZrO2,
WO3/TiO2, WO3/Nb2O5, and WO3/Al2O3).47 The methanol
dehydration TOFacid for H3PWo12O40 is even 2 orders of
magnitude faster than that found for the highly active ZrOx-
stabilized WO3 nanoparticles present in the coprecipitated
WO3-ZrO2 catalysts (TOFacid ) 0.05 s-1 at 225 °C).48 Thus,
the surface acid sites present in the H3PWo12O40 Keggin are
highly acidic for methanol dehydration, and their kinetics are
significantly faster than those of other W-containing catalysts
(bulk WO3, supported WO3 catalysts, and even the highly acidic
WO3-ZrO2 catalyst synthesized by AMT and ZrO(OH)2 co-
precipitation). The steady-state methanol dehydration TOFacid
values track the large variation in the DME/CH3OH-TPSR
trends reported above (180 °C < Tp < 290 °C), which suggests
that breaking of the surface methoxy C-O bond dominates the
steady-state catalytic activity trend.
The structurally disordered H3+xPW12-xVxO40 clusters remain
intact during methanol oxidation/dehydration in the 200-250
°C range. The VOx units in the primary and secondary structures
are essentially fully oxidized as V5+ during the methanol
oxidation/dehydration reaction (O2/CH3OH ) 2.17), which
reflects the operation of the Mars-van Krevelen redox mech-
anism and the rapid reoxidation by gas-phase molecular O2. The
vanadium oxide site in the primary Keggin structure is ∼160
times more active for methanol oxidation to formaldehyde at
225 °C than the vanadium oxide site in the secondary structure.
The decreasing TOFredox kinetics with increasing vanadium oxide
content in the H3+xPW12-xVxO40 Keggins is a consequence of
the increasing amount of vanadium oxide in the Keggin
secondary structure as the V content is increased. The presence
of the surface VOx species on the Keggin cluster also retards
the TOFacid kinetics for methanol dehydration because the
surface VOx species block access to WOx acid sites and titrate
acidic surface hydroxyls.
The methanol dehydration kinetics (TOFacid) over the
H3PW12O40 Keggin is ∼100 times faster than for bulk WO3 and
supported tungsten oxide catalysts, which reflects the strong
acidic character of the protons in the TPA Keggin cluster. The
presence of surface VOx species can reduce the TOFacid kinetics
by as much as a factor of ∼100. The methanol oxidation to
formaldehyde kinetics (TOFredox) over H4PW11V1O40 is com-
parable to that for the more active vanadium oxide catalysts
(bulk V2O5 and supported V2O5/CeO2, V2O5/ZrO2, and V2O5/
TiO2). The TOFredox kinetics for the supported VOx/H3PW12O40
Keggin catalyst is comparable to that of the least active
supported vanadium oxide catalyst (V2O5/SiO2).
No correlations were found between the electronic Eg values
and the methanol oxidation TOFredox and methanol dehydration
TOFacid kinetics of the H3+xPW12-xVxO40 Keggins. A minor
amount of surface carbon is also present on the Keggins during
methanol oxidation/dehydration, which decreases with increasing
vanadium oxide content, and does not appear to affect the
kinetics. These new insights reveal that mixed metal oxide
Keggins are much more complex than originally thought and
that care must be taken in employing Keggins as model mixed
metal oxide NPs in kinetic and theoretical studies.
Introduction of VOx sites into the H3PW12O40 Keggin also
modifies its acidic character. Surface VOx sites on TPA decrease
the TOFacid kinetics by ∼100-fold from site blocking and
titration of acidic surface hydroxyls. Incorporation of VOx into
the TPA Keggin does not affect the TOFacid kinetics for TPAV1,
which has only a minimal amount of surface VOx sites. The
TOFacid kinetics for TPAV2 and TPAV3, however, are lower
than that for TPAV1 because of the greater presence of surface
VOx sites. As for the methanol oxidation TOFredox kinetics, the
methanol dehydration TOFacid kinetics over the Keggins do not
correlate with the UV-vis Eg value, as well as the amount of
carbon deposition, since the TOFacid values are essentially the
same for TPA and TPAV1.
Acknowledgment. L.N. thanks the Department of Science &
Technology, New Delhi, India, for financial assistance in the form
of a BOYSCAST fellowship. We thank Prof. Mark A. Barteau,
Department of Chemical Engineering, University of Delaware, for
providing some of the Keggin catalyst samples. The Lehigh
University researchers were supported by NSF Nanoscale Integrated
Research Team (NIRT) Grant 0609018.
Conclusions
The H3+xPW12-xVxO40 Keggins contain VOx sites in both the
primary and secondary structures. The vanadium oxide incor-
porated into the primary structure possesses VO6 coordination
and has substituted for WO6 units in the Keggin. The vanadium
oxide in the secondary structures possesses the same structure
Supporting Information Available: Information related to the
in situ Raman spectra of the catalysts obtained at different
temperatures and the in situ UV-vis DRS carried out under
the reaction conditions. This material is available free of charge
via the Internet at http://pubs.acs.org/.
(50) Wachs, I. E.; Chen, Y.; Jehng, J.-M.; Briand, L. E.; Tanaka, T. Catal.
Today 2003, 78, 13.
JA904957D
9
15554 J. AM. CHEM. SOC. VOL. 131, NO. 42, 2009