The Effect of Cation Type and H+ on the Catalytic Activity of the Keggin Anion [PMo12O40]3- in the Oxidative Dehydrogenation of Isobutyraldehyde

Article abstract of DOI:10.1006/jcat.2000.2987

The oxidative dehydrogenation of isobutyraldehyde to methacrolein over [PMo₁₂O₄₀]^3--containing catalysts has been shown to proceed through bulk catalysis-type II, which depends on the rates of diffusion of the redox carriers (H⁺ and e^-) into the catalyst bulk. Variations in catalyst behaviour have been shown to change with the countercation and appear to be related to the polarizing ability of the cation, which can be represented by the ionic potential (charge/ionic radius). This, in turn, may indicate that the active site at the [PMo₁₂O₄₀]^3- ion is close to an attendant countercation. For the alkali metal ions Li⁺, Na⁺, K⁺, Rb⁺, and Cs⁺ as well as the (isoelectronic) ions of the series Cs⁺, Ba²⁺, La³⁺, and Ce⁴⁺, the studies have shown that conversion generally decreases with increasing ionic potential, while selectivity to methacrolein is less affected by changes in this property. However, the observed enhancement in selectivity to methacrolein and methacrylic acid in the case of Ce⁴⁺ is attributed to the ability of this cation to undergo reduction to Ce³⁺, with the Ce⁴⁺/Ce³⁺ couple likely acting to expedite the transfer of electrons between anions and hence into the bulk. Trends in the conversion and selectivities among the countercations Cs⁺, NH⁺₄, and (CH₃)₄N⁺ show significant increases along this series, which is consistent with progressively increasing dissipation of the formal cationic charge. A mechanism for the oxidative dehydrogenation of isobutyraldehyde is proposed. The presence of H⁺ enhances the activity of salts of [PMo₁₂O₄₀]^3-, most likely by protonation of the organic component, as little evidence from extended-Hueckel molecular orbital calculations could be found for any reduction in the HOMO-LUMO gap of the anion upon protonation, which would make it easier to reduce the anion as part of the catalytic process.

Full text of DOI:10.1006/jcat.2000.2987

Journal of Catalysis 195, 360–375 (2000)
doi:10.1006/jcat.2000.2987, available online at http://www.idealibrary.com on
The Effect of Cation Type and H+ on the Catalytic Activity
of the Keggin Anion [PMo12O40]3 in the Oxidative
Dehydrogenation of Isobutyraldehyde
Ji Hu and Robert C. Burns1
School of Biological and Chemical Sciences, The University of Newcastle, Callaghan 2308, Australia
Received March 23, 2000; revised July 4, 2000; accepted July 5, 2000
such as the oxidative dehydrogenation of isobutyric acid
and the oxidation of methacrolein, both of which yield
methacrylic acid (1–5). Methacrylic acid is, in turn, reacted
with methanol to yield methyl methacrylate, an extremely
important acrylic monomer, which is then polymerized to
give poly(methyl methacrylate). Heteropolyoxometalates
are also active acid catalysts, and processes based both on
their redox and acid–base properties have found commer-
cial applications (3–5).
The oxidative dehydrogenation of isobutyraldehyde to metha-
crolein over [PMo12O40]3 -containing catalysts has been shown to
proceed through bulk catalysis-type II, which depends on the rates
of diffusion of the redox carriers (H+ and e ) into the catalyst bulk.
Variations in catalyst behaviour have been shown to change with
the countercation and appear to be related to the polarizing abil-
ity of the cation, which can be represented by the ionic potential
(charge/ionic radius). This, in turn, may indicate that the active
site at the [PMo12O40]3 ion is close to an attendant countercation.
For the alkali metal ions Li+, Na+, K+, Rb+, and Cs+ as well as the
(isoelectronic) ions of the series Cs+, Ba2+, La3+, and Ce4+, the stud-
ies have shown that conversion generally decreases with increasing
ionic potential, while selectivity to methacrolein is less affected by
changes in this property. However, the observed enhancement in se-
lectivity to methacrolein and methacrylic acid in the case of Ce4+ is
The study of heteropolyoxometalates as oxidation–
reduction catalysts has involved primarily Keggin-based
structures, principally [PMo12O40]3 , as well as substi-
tuted species involving replacement of one or more
framework Mo(VI) by V(V) (3–5). The geometry of the
( -)[PMo12O40]3 ion consists of a central tetrahedrally co-
ordinated phosphorus atom surrounded by four groups of
three edge-sharing octahedra (i.e., Mo3O13 subunits), which
are in turn linked to each other and to the central PO4
tetrahedron by shared oxygen atoms at the vertices (6).
Some studies have also included the related Dawson struc-
ture ion [P2Mo18O62]6 , which is formed by joining two
“[PMo9O31]3 ” monomeric units, the latter being formed
by loss of three MoO3 groups, one each from three adja-
cent Mo3O13 subunits of the Keggin structure anion (7).
The majority of studies of heteropolyoxometalates as
oxidation–reduction catalysts have concerned the oxida-
tive dehydrogenation of isobutyric acid and the oxidation
of methacrolein (3–5), although the oxidative dehydro-
genation and oxidation of isobutane have recently received
considerable attention (8, 9), as has the partial oxidation of
methane using N2O as the oxidant (10–13). Methacrolein
can be prepared by the oxidative dehydrogenation of isobu-
tyraldehyde, which is formed as a by-product of the oxo
synthesis of propene to yield n-butyraldehyde. While the
oxidative dehydrogenation of isobutyraldehyde has been
investigated, it is not as well studied as that of isobutyric
acid. Thus, for example, isobutyraldehyde has been con-
verted to methacrolein over iron phosphate at 400 C with a
conversion/selectivity of 88% /68% (14), over H3PMo12O40
(containing Sb, Zn, and Cr) at 295 C, 88.3% /40.8% (also
attributed to the ability of this cation to undergo reduction to Ce3+
,
with the Ce4+/Ce3+ couple likely acting to expedite the transfer of
electrons between anions and hence into the bulk. Trends in the
conversion and selectivities among the countercations Cs+, NH4+,
and (CH3)4N+ show significant increases along this series, which is
consistent with progressively increasing dissipation of the formal
cationic charge. A mechanism for the oxidative dehydrogenation
of isobutyraldehyde is proposed. The presence of H+ enhances the
activity of salts of [PMo12O40]3 , most likely by protonation of the
organic component, as little evidence from extended-Hu¨ckel molec-
ular orbital calculations could be found for any reduction in the
HOMO–LUMO gap of the anion upon protonation, which would
make it easier to reduce the anion as part of the catalytic process.
c
2000 Academic Press
Key Words: isobutyraldehyde; heteropolyoxometalate; Keggin
structure; bulk catalysis-type II.
INTRODUCTION
Heteropolyoxometalate acids and salts are highly effec-
tive heterogeneous catalysts for the gas-phase selective ox-
idation of organic substrates. Examples include reactions
1 Dr. R. C. Burns, School of Biological and Chemical Sciences, The
University of Newcastle, Callaghan, New South Wales, Australia, 2308.
E-mail: csrb@paracelsus.newcastle.edu.au. Telephone: +61-2-49215479.
360
0021-9517/00 $35.00
c
Copyright
2000 by Academic Press
All rights of reproduction in any form reserved.

3
CATION EFFECTS ON CATALYTIC PROPERTIES OF [PMo12O40
]
361
with 6.9% methacrylic acid) (15), and over H3PMo12O40 the appropriate amounts of aqueous Cs2CO3 to solutions
(containing Zr, V, and Cs) at 310 C, 100% /76.5% (also of H3[PMo12O40], followed by removal of water at 50 C,
n
with 8.9% methacrylic acid) (16). Other studies have while the Me4N+ and Bu4N+ salts were obtained by addi-
n
included the effects of the substitution of La3+ for H+ in tion of a stoichiometric amount of Me4NCl or Bu4NBr to
H4PVMo11O40 (17) and the effect of changing the central concentrated aqueous solutions of H3[PMo12O40], followed
heteroatom in the Keggin anion from P(V) to As(V) (18). by isolation of the yellow products by filtration and drying
There has been, however, no comprehensive study yet un- at room temperature. The infrared spectra ofallcompounds
dertaken equivalent to those of the oxidative dehydrogena- and the 31P NMR spectrum of (Me4N)3[PMo12O40] and
tion ofisobutyricacid over varioussaltsofthe [PMo12O40]3
(nBu4N)3[PMo12O40] ( = 2.73 and 2.74 ppm, respec-
ion (1, 19, 20) or of the effects of added H+ (21). The present tively, from 85% H3PO4; DMSO solvent) were in agree-
study details such an investigation of the oxidative dehydro- ment with published data (25).
genation of isobutyraldehyde and allows comparisons with
Samples of (NH4)3[PMo12O40] with a low surface area
the studies of isobutyric acid under similar conditions. Salts were prepared by two methods: (1) the thermal decom-
of the [PMo12O40]3 ion were used in this work as this avoids position of (nBu4N)3[PMo12O40] at 300 C in air and (2) by
the complications that are found in the case of, for example, direct reaction of anhydrous H3[PMo12O40] with excess
[PVMo11O40]4 , which loses VO2+ upon thermal treatment gaseous NH3 (BOC) at 0 C, followed by heating at 100 C
or during catalysis with isobutyric acid (22). In this study, to ensure complete reaction. The thermal decomposition
variations in the countercations are examined, and sugges- of (nBu4N)3[PMo12O40] to (NH4)3[PMo12O40] was accom-
tions are made concerning the mechanism of the oxidative plished by static heating in an air atmosphere at 300 C for
dehydrogenation process of isobutyraldehyde and isobu- 1 week. This resulted in a yellow-brown material, which
tyric acid as well as the role played by the presence of H+. upon washing with distilled water produced a yellow com-
pound. The infrared spectra of both materials indicated
EXPERIMENTAL
that the [PMo12O40]3 ion had remained intact, while the
X-ray powder diffraction patterns were identical to that of
an authentic sample of (NH4)3[PMo12O40]. Chemical analy-
ses of the commercial sample, that obtained by thermal de-
composition of (nBu4N)3[PMo12O40], and that obtained by
reaction of anhydrous H3[PMo12O40] with NH3 gave the fol-
lowing results: % N, 2.03, 1.55, and 2.24, respectively (theo-
retical value of 2.24% ); % H, 0.69, 0.59, and 0.60, respec-
tively (theoretical value of 0.69% ). The sample obtained
by thermal decomposition also contained 0.15% C, indicat-
ing that the nBu4N+ ion had not completely decomposed
or that some carbon-based impurity was still present in the
sample.
Preparation and Characterization of Salts
of the [PMo12O40]3 Anion
Phosphomolybdic acid, H3[PMo12O40] 23H2O (water
content determined by TGA), and (NH4)3[PMo12O40] were
obtained commercially (Ajax/Univar and Aldrich, respec-
tively). The alkali metal salts of the [PMo12O40]3 ion were
prepared according to the method of Tsigdinos (23), by ad-
dition of a stoichiometric amount (3 : 1 mole ratio) of an
alkali metal carbonate to a concentrated aqueous solution
of H3[PMo12O40] and isolation of the yellow product, in the
case of the K+, Rb+, and Cs+ salts, by filtration and drying
at room temperature. The aqueous-soluble Li+ and Na+
salts were isolated by removal of water at 50 C. The in-
soluble [PMo12O40]3 salts, i.e., K+, Rb+, and Cs+, and the
NH+4 salt were stirred subsequently for 30 min with 0.1 mol
dm 3 solutions of their respective MCl (M = alkali metal)
or NH4Cl to remove any extraneous H+ included in their
structures (24). The samples were filtered, washed with a
little water, and dried at 50 C. The Ba2+ and La3+ salts
were obtained by reaction of the appropriate amounts of
their metal carbonates with H3[PMo12O40] and removal of
the water as in the case of the Li+ and Na+ salts. The Ce4+
salt was obtained by addition of a solution of cerium(IV)
sulfate hydrate (Aldrich, cerium content 30% minimum) to
one of H3[PMo12O40], followed by the addition of the ap-
propriate amount of BaCO3 to precipitate insoluble BaSO4.
The latter was removed by filtration and the orange-yellow
Ce4+ salt isolated by evaporation of water at 50 C. The
mixed cation phases (CsnH3 n)[PMo12O40] (with n nomi-
To test for the presence of H+ in soluble catalysts (which
has an enhancing effect on the oxidative dehydrogenation
of isobutyraldehyde), a UV–visible spectroscopic proce-
dure using the base form of the indicator bromothymol
blue (sodium salt, Aldrich, 90% dye content) in N,N-
dimethylformamide (DMF) (Aldrich, 99.8% , anhydrous,
dried over type 4A molecular sieves) was employed.
Any H+ that may have been retained following isolation
of the neutral salts of the [PMo12O40]3 ion or formed
during preparation by the hydrolysis reaction given below
(Eq. [1]) can be measured spectrophotometrically in solu-
tion by conversion of the base form of bromothymol blue
(
max = 636 nm, ε = 45,000 L mol 1 cm 1) into its acid
form.
[PMo12O40]3 + 3H2O
↔ [PMo9O31]3 + 3[HMoO4] + 3H+.
[1]
nally equal to 1, 2, and 2.5) were prepared by addition of In organic solvents such as acetone and ethanol, the pK1,

362
HU AND BURNS
pK2, and pK3 values for the H3[PMo12O40] range from 2.0 to two Hammett indicators. Using a scale from 0 to 4, with 0
5.3 (26) and the various protonated forms of [PMo12O40]3
representing no response to the Hammett indicators and
will be stronger acids in the more basic solvent DMF. Bro- 4 that of H3[PMo12O40], all catalysts were assigned val-
mothymol blue has pKa values of 7.2 and 12.4 in water and ues of 3, while both the Rb+ and Cs+ salts had a value
methanol, respectively (27), and 10 in DMF [estimated of 1.
from data provided in Ref. (27)] and thus will be easily
Thermogravimetric analyses were performed on a
protonated. For the determination of the amount of H+ Stanton Redcroft TG-750 instrument coupled to a Eu-
in soluble catalysts, known concentrations (2.0 10 5 mol rotherm Model 94 temperature controller. Sample masses
3
dm in [PMo12O40]3 ) of the Li+ and Na+ salts (calcined of about 10 mg were used with a heating rate of 10 C/min
at 300 C for 6 h) were prepared and the reduction in ab- in a static air atmosphere. Infrared spectra were recorded
sorbance of the base form of the indicator at 636 nm (using on a Bio-Rad FTS-7 Fourier-transform spectrophotome-
an initial concentration of the sodium salt of bromothymol ter, with samples mounted as KBr discs. UV–visible spectra
4
blue of 2.0 10 mol dm 3) was measured relative to a were obtained using a Hitachi U-2000 spectrophotometer
solution of the same concentration of the indicator that did with 1-cm cells. X-ray powder diffraction was performed
not contain the catalyst. At this wavelength any absorbance on a Philips PW1700 Automated Powder Diffractometer
from the acid form of the indicator ( max = 403 nm) and System 1, employing graphite-monochromated Cu K ra-
from the [PMo12O40]3 ion (absorbance occurs at wave- diation. Scanning electron microscopy was performed on
lengths less than 500 nm) do not interfere with the a Philips XL-30 scanning electron microscope operating
measurement. All solutions were prepared under nitrogen at 15 kV. Solution 31P NMR spectra were recorded with
and the spectra obtained using sealed cells to eliminate a Bruker AVANCE DPX-300 spectrometer operating at
water. In both cases the reduction in absorbance was used 121.49 MHz. The reference used was external 85% H3PO4.
to generate a H+ concentration and then, using the above Solid-state 31P NMR spectra were obtained with a Bruker
equation (also noting that [HMoO4] will be a strong acid MAS-300 spectrometer operating at 121.49 MHz. Samples
in DMF), this was expressed as the amount of [PMo12O40]3
( 150 mg) were contained in a partially stabilized zirco-
hydrolysed during preparation. Initial values of absorbance nia tube fitted with a Kel-F cap and were spun at a rate of
were used to avoid the possibility of any further reactions, 10 kHz. The reference used was solid (NH4)H2PO4. Surface
such as reformation of [PMo12O40]3 upon standing in so- area measurements were made by the BET method using
lution. The results indicated that both salts experienced nitrogen adsorption on a Micromeritics ASAP 2400 instru-
a similar extent of hydrolysis of [PMo12O40]3 during ment. All samples were prepared using an identical method
preparation (8% and 5% for the Li+ and Na+ salts, res- to that described for the catalysis samples (see below) and
pectively).
were degassed at 300 C prior to measurement.
All (neutral) salts used in this study, H3[PMo12O40], and
the two samples of (NH4)3[PMo12O40] prepared by the al-
ternative methods described above were also tested using
the standard approach for solid acids(28) with the Hammett
indicators 4-(phenylazo)diphenylamine (Aldrich, 97% )
and dicinnamalacetone (Aldrich, 98% ), which cover a
range of sites of different acid strengths. Samples were
tested following calcination at 300 C for 6 h or drying
over silica gel under vacuum. All catalysts showed evi-
dence of surface acidity with 4-(phenylazo)diphenylamine
(pKa + 1.5) as well as with dicinnamalacetone (pKa 3.0).
The water-soluble catalysts exhibited strong surface acidity
based on colouration with both indicators. However, the
water-insoluble species that had been cation exchanged as
described above, with the exception of the commercial sam-
ple of (NH4)3[PMo12O40], appeared to produce less intense
colouration. This was noticably so for the Rb+ and Cs+ salts,
despite their generally greater surface areas, and suggests
that there was less H+ present on the surfaces of these two
catalysts, although this does depend on the number of pro-
tons at the particular acid strength of the site. The samples
of (NH4)3[PMo12O40] prepared by thermal decomposition
of (nBu4N)3[PMo12O40] and by reaction of H3[PMo12O40]
with NH3 both showed strong surface reactions with the
Catalysis and Analysis of Products
The catalytic activities of the various salts of the
[PMo12O40]3 anions toward the gas-phase oxidative de-
hydrogenation of isobutyraldehyde were carried out using
a conventional flow fixed-bed reactor (a 30-cm long quartz
tube with an internal diameter of 4 mm). Reactions were
conducted at atmospheric pressure and at temperatures of
275 and 300 C following an initial temperature effect study
of both (NH4)3[PMo12O40] and K3[PMo12O40] at 275, 300,
325, and 350 C. The feed gas consisted of 1% isobutyralde-
hyde (Aldrich, 98% , which was freshly distilled under nitro-
gen and transferred to the catalysis system saturator using a
gas-tight syringe) and 3% oxygen, with a nitrogen balance.
The saturator was held at 0 C using an ethanol–water cold
bath to ensure a constant vapour pressure of isobutyralde-
hyde during each catalysis study. Typical catalyst masses
of 0.35 g ( 0.21 ml) or 0.20 g ( 0.13 ml) were used (after
allowing for loss of water), which when combined with a
flowrate of150cm3 min 1 gave residence timesof 0.09and
0.05 s, respectively, assuming plug flow. This corresponds
to a W/F (weight of catalyst per total flow rate) of 0.14 and
0.08 g s ml 1, respectively. Catalysts samples were prepared

3
CATION EFFECTS ON CATALYTIC PROPERTIES OF [PMo12O40
]
363
by compressing the powdered catalyst at a pressure of
1.7 ton cm 2, which was broken up and passed through
a series of sieves, with the fraction 40–80 mesh used for
the studies. The catalyst samples were loaded between
silica wool plugs in the quartz reactor tube. Samples were
calcined at 10 C above the maximum temperature to be
employed in the study for 2 h under a flow of nitrogen
containing 3% oxygen.
Following the establishment of a steady state in each
case ( 4 h), which was monitored by the IR detection and
analysis of both CO and CO2, the condensable gaseous
products, methacrolein, acetone, methacrylic acid, and un-
reacted isobutyraldehyde, were collected in an acetone–
dry ice cold bath for a period of 30 min and used as
representative of the conversion and selectivity for the
sampling time. The water produced during the reaction
was not determined. No evidence for the formation of ni-
triles, such as isobutyronitrile and methacrylonitrile, was
found with NH+4 as the cation, although species of this
type have been observed during the oxidative dehydro-
genation of isobutyric acid (16). The above products were
analysed by gas chromatography (3 wt% FFAP and 7 wt%
OV-17 on acid-washed Chromosorb-W). The carbon bal-
ance was routinely in the range 90–100% . The conversion
percentage of isobutyraldehyde was calculated as 100[1-
(mol of unreacted isobutyraldehyde/mol of isobutyralde-
hyde introduced)], while the selectivities percentages of the
various products were determined as 100(mol of compo-
nent i/total number of mol of products).
FIG. 1. Effect of temperature on the oxidative dehydrogenation of
isobutyraldehyde over (NH4)3[PMo12O40]: 1% isobutyraldehyde, 3% oxy-
1
gen; W/F 0.08 g s ml
.
and K3[PMo12O40]. The compounds are isostructural, with
very similar unit cell dimensions and surface areas (see be-
low). In both cases a steady state was not achieved for some
4 h following initiation of catalysis. This was also observed
for all of the remaining salts in subsequent studies. The
results of temperature dependence studies (conversion, se-
lectivity, and efficiency percentage) for both compounds are
given in Figs. 1 and 2 for the range 275 to 350 C. The temper-
ature for the most efficient (% ) formation of methacrolein
(MAL) plus methacrylic acid (MAA) is given by the ex-
pression
Extended-Hu¨ckel Molecular Orbital Calculations
Extended-Hu¨ckel molecular orbital calculations were
performed using the package described by Mealli and
Proserpio (29) (Version 4.0, 1994), with the extended-
Hu¨ckel parameters, i.e., Coulomb integrals, Hii (eV),
and Slater exponents ( ), taken from the values pro-
vided in the package. The Wolfsberg–Helmholtz constant
was set to 1.75. The geometry of the [PMo12O40]3 ion
used in the calculation was taken from the structure of
H3[PMo12O40] 30H2O (30, 31). For calculations on the
efficiency (% ) = [(% conversion)
(% selectivity to MAL + MAA)]/100.
NH+4 ion the N–H distance was taken as 1.02 A, while for
˚
the Me4N+ ion the N–C and C–H distances were 1.45 and
˚
1.05 A, respectively.
RESULTS AND DISCUSSION
I. Catalysis and Supporting Studies
A. Effect of Temperature over (NH4)3[PMo12O40]
and K3[PMo12O40]
Initial investigations into the appropriate conditions un-
der which to study the type of catalysis and the effects
of cation variations on the oxidative dehydrogenation of
FIG. 2. Effect of temperature on the oxidative dehydrogenation of
isobutyraldehyde over K3[PMo12O40]: 1% isobutyraldehyde, 3% oxygen;
1
isobutyraldehyde employed the salts (NH4)3[PMo12O40] W/F 0.08 g s ml
.

364
HU AND BURNS
For (NH4)3[PMo12O40] and K3[PMo12O40] the efficiencies place at the two-dimensional surface of the solid catalyst,
peaked at 305 and 340 C, respectively. For the for- with the reaction rate proportional to the surface area. In
mer salt, although methacrolein and methacrylic acid for- bulk catalysis-type II, while the reaction still occurs at the
mation decreased with increasing temperature, conversion catalyst surface, the inner bulk of the catalyst plays a role,
increased, leading to the most efficient formation of the de- with the contribution depending on the rate of the cata-
sired products at 305 C. Interestingly, the CO2 selectivity lytic reaction relative to the rates of diffusion of the redox
changed little with temperature, suggesting that complete carriers (H+ and e ) into the bulk. In the ideal (limiting)
oxidation of the products was not a major competitive path- case, the rate is proportional to the volume (or mass) of the
way over the temperature range studied. In contrast, that of catalyst, although some dependence on surface area may
CO increased markedly, which is the cause of the decrease be observed, depending on the relative rates of the above
in methacrolein selectivity with increasing temperature. processes. It is likely that the oxidative dehydrogenation of
The selectivity of acetone paralleled that of methacrolein. isobutyraldehyde operates by bulk catalysis-type II, as does
These results suggest that methacrolein, acetone, and CO the oxidative dehydrogenation of isobutyric acid, given the
are linked through a common intermediate in the mecha- similarities in their structures.
nism of formation of the product species.
One way to test whether or not surface area is im-
For K3[PMo12O40], the selectivity of methacrolein in- portant in the oxidative dehydrogenation of isobutyralde-
creased with increasing temperature until 325 C, sub- hyde, and hence establish the type of catalysis operating,
sequently followed by a slight decrease, while that of is by the preparation of the same catalyst material with
methacrylic acid progressively increased with increasing vastly different surface areas, followed by an examination
temperature. Combined with an increasing conversion at of their relative activities under identical conditions. For
higher temperatures, this led to a maximum efficiency for this purpose, (NH4)3[PMo12O40], likely formed by direct
methacrolein and methacrylic acid production at 340 C. precipitation (commercial sample, Aldrich), was compared
Again CO2 selectivity was essentially constant, but the CO with materials prepared by two other different synthetic
and acetone selectivities both decreased with increasing approaches. These included syntheses involving the ther-
temperature as methacrolein increased and the former be- mal decomposition of (nBu4N)3[PMo12O40] (32) and by di-
came effectively constant above 325 C. Once again, this rect reaction of anhydrous H3[PMo12O40] with excess NH3
suggests a link between methacrolein, acetone, and CO for- gas. For the analogous tungstate system it has been re-
mation, as indicated above for (NH4)3[PMo12O40].
ported previously (33) that anhydrous H3[PW12O40] re-
Comparison of the trends in conversion and the selec- acts quantitatively with NH3 to produce (NH4)3[PW12O40],
tivities of the products with temperature for these two but no studies have been reported for H3[PMo12O40]. The
(isostructural) catalysts, particularly for methacrolein, in- commercial sample of (NH4)3[PMo12O40], following ion
dicates different behaviour, despite the presence of the exchange (see above), had a measured surface area of
same [PMo12O40]3 ion. This suggests that the counter- 120 m2 g 1. For the thermally produced sample, the precur-
1
cation plays an active role in the conversion and selectivity sor, (nBu4N)3[PMo12O40], had a surface area of 1.5 m2 g
.
trends of compounds of the [PMo12O40]3 ion. Similar con- Following thermal decomposition, the product had an IR
clusions have been drawn regarding the oxidative dehydro- spectrum and XRD pattern identical to those of an authen-
genation of isobutyric acid over a range of [PMo12O40]3
-
tic sample of (NH4)3[PMo12O40], but with a surface area of
containing catalysts with varying countercations (1, 20). All only 5.3 m2 g 1. Similarly, the material produced by direct
subsequent studies were performed at 275 and 300 C, to reaction of the anhydrous acid (1.2 m2 g 1) with gaseous
the lower side of the temperature of maximum efficiency NH3 would not be expected to increase greatly as NH3 is
of both (NH4)3[PMo12O40] and K3[PMo12O40]. Most of the able to penetra+te the interanionic voids and react with the
other alkali metal cations behaved in a manner similar t+o
H
+ to give NH4 ions. The resulting material again had an
that of the K+ salt, while (CH3)4N+ behaved like the NH4 IR spectrum and XRD pattern identical to those of an au-
salt. Some salts appeared intermediate in behaviour be- thentic sample of (NH4)3[PMo12O40], with a surface area
tween these extremes, although these latter conclusions are of 7.0 m2 g 1. The catalysis results for all three samples of
based only on data at 275 and 300 C.
(NH4)3[PMo12O40] are given in Table 1. Some slight dif-
ferences in the conversion percentage of isobutyraldehyde
are observed, with a decreased conversion for the lower
surface area materials. However, the changes in the rates
and conversion percentages are relatively small (less than
B. Effect of Differences in Surface Area
over (NH4)3[PMo12O40]
(i) Catalysis studies. It has been established in previ-
ousstudiesthat the oxidative dehydrogenation ofisobutyric a factor of 2) compared to the changes in the surface ar-
acid over salts of the [PMo12O40]3 ion occurs by bulk cata- eas of the samples, which vary by a factor of 20. Thus,
lysis-type II (3–5). This is different from conventional het- surface area does not appear to be a major factor in con-
erogeneous catalysis over a solid, where the reaction takes trolling the rate of this reaction and strongly indicates that

3
CATION EFFECTS ON CATALYTIC PROPERTIES OF [PMo12O40
]
365
TABLE 1
Effect of Surface Area of (NH4)3[PMo12O40] on the Oxidative Dehydrogenation of Isobutyraldehyde
% Selectivitya
Catalyst
preparation
Surface area (m2 g 1); Temp.
Rate
1
1
pore volume (cm3 g
)
( C) % Conv. (104 mol min 1 g
) MAL MAA ACE CO CO2
Commercial sample
Thermal decomp.
Acid + excess NH3
120; 0.126
275
300
61
81
1.88
2.50
71
62
<1
2
6
5
10 12
18 13
5.3; 0.012
7.0; 0.028
275
300
38
48
1.17
1.48
64
66
<1
1
4
3
19 12
16 14
275
300
44
63
1.36
1.95
70
67
<1
1
4
3
15 10
16 13
1
Note. Mass of catalyst, 0.20 g; W/F = 0.08 g s ml
.
a MAL, methacrolein; MAA, methacrylic acid; ACE, acetone.
bulk catalysis-type II is operating. The observed differences ular block-like crystals of varying sizes, which are effec-
in conversion and selectivities percentages are likely at- tively identical to the original crystals of the parent species
tributable to varying amounts of residual H+ in the sam- from which they were formed. Using a calculated crys-
3
ples (caused by either incomplete removal of H+ following tallographic density for (NH4)3[PMo12O40] of 3.93 g cm
,
exchange with NH4Cl, formation of H+ during thermal de- together with an average diameter for the spherical crys-
composition, or incomplete neutralization ofH+ with NH3), tals of 0.2 m for the high-surface area material and
which has an enhancing effect on the catalytic activity of
salts of the [PMo12O40]3 ion. This is discussed in more de- tals found for both low surface area samples, the respective
tail below.
calculated surface areas are 4 and 0.2 m2 g 1. These will
5
10 10 m for the dimensions of the block-like crys-
Comparison of the selectivity percentage data for the be lower limits as they assume that the surfaces are smooth
three samples indicates that the selectivity to methacrolein and likely will be a factor of 2 or 3 larger, allowing for
varies from 62 to 71% (data for both temperatures), with no roughness. Thus, there is an 20-fold increase in surface
dependence on the differences in the surface areas of the area in going from the low to the high surface area materi-
catalysts. The high surface area commercial sample pro- als from these calculations. Both calculated values suggest
duced slightly more methacrylic acid, but this is consis- that all three samples have some internal pore structure,
tent with the subsequent oxidation of the initial product particularly the commercial sample. The pore volume, ob-
methacrolein to methacrylic acid over these types of cata- tained from BET measurements, of the high surface area
lysts, which is a reaction that is known to depend on the sample was 0.126 cm3 g 1, while those of the other sam-
1
surface area of the catalyst (21). There appears to be no ples were 0.012 and 0.028 cm3 g (Table 1). Precipitated
trends in the selectivities to acetone, CO, or CO2 with differ- (high surface area) (NH4)3[PMo12O40] is known to be a mi-
ences in surface area. The selectivity to acetone decreased croporous material and has previously been reported to
with increasing temperature, as has been observed previ- have a surface area of 193 m2 g 1, with a mean microp-
˚
ously (Fig. 1), while those to CO and CO2 showed slight ore radius of 13 A, and an average mesopore radius of
˚
variations.
35 or 45 A, depending on the pore shape (cylindrical or
slit models, respectively) (34). The pore size analysis of the
(ii) SEM, pore volume, and pore size studies. The SEM high surface area salt in the present study yielded an aver-
˚
micrographs of the three samples of (NH4)3[PMo12O40] age mesopore diameter of 39 A (cylindrical model), which
used in the above studies are shown in Fig. 3. The com- is in reasonably good agreement with the reported data,
mercial sample (probably precipitated) that had been ion- while those of the two low surface area salts were 142 and
˚
exchanged with NH4Cl (Fig. 3a) exhibited small, spherical 189 A for (NH4)3[PMo12O40] prepared by thermal decom-
crystals ranging from 0.1 to 0.5 m in diameter. This position of (nBu4N)3[PMo12O40] and by direct reaction of
morphology likely reflects the isotropic nature of crystal anhydrous H3[PMo12O40] with excess NH3 gas, respectively.
growth of the compound from solution (which has a cubic The low surface areas of the latter compounds, with their
unit cell, see below), thus resulting in an almost spherical low pore volumes and large pore diameters, reflect the orig-
shape. The range and small sizes of the crystals contribute inal crystal morphologies of the compounds from which
to the high surface area of the compound (120 m2 g 1). they were formed. However, the slight increases in sur-
In contrast, both materials obtained by nonprecipitative face areas on their formation suggest that some pore struc-
methods (Figs. 3b and 3c, respectively) show large irreg- ture may have developed during the rearrangement of the

366
HU AND BURNS
atoms upon thermal decomposition of (nBu4N)3[PMo12O40]
and upon reaction of anhydrous H3[PMo12O40] with excess
NH3. Thus, examination of the SEM images in Figs. 3b and
3c show extensive cracking and evidence of holes leading
to the inner regions of the crystals.
C. Effect of Cation Size and Charge on the Oxidative
Dehydrogenation of Isobutyraldehyde [Li+, Na+,
K+, Rb+, and Cs+; Ba2+, La3+, and Ce4+; NH4+
and (CH3)4N+ Salts]
To investigate the effects of varying the countercation on
the oxidative dehydrogenation of isobutyraldehyde (Ta-
ble 2), an initial series of compounds of the [PMo12O40]3
ion containing cations of the same charge but varying ionic
radii (Li+, Na+, K+, Rb+, and Cs+) were examined.
When precipitated from solution, the insoluble K+, Rb+,
and Cs+ salts, and the NH+4 salt, crystallize with a cu-
˚
bic unit cell, with a = 11.7 A, space group Pn3m, and
Z = 2 (35). However, the soluble Li+ and Na+ salts, fol-
lowing removal of solvent, can have different structures.
The former had an XRD pattern somewhat similar to that
of H3[PMo12O40] 23H2O, while the pattern of the latter
indicated at least two components, the major component
having the above cubic unit cell and a second, more com-
plex pattern, showing some similarities to that of the Li+
salt. All alkali metal salts are hydrated to varying extents,
as shown by the results of TGA studies, which are given in
Table 3. The Li+ salt is the most hydrated while the salts of
the other alkali metal cations, and the NH+4 salt, all have
a similar extent of hydration. Thermogravimetric analysis
results indicated that the K+, Rb+, Cs+, and NH4+ salts lost
this water before 100 C, the Na+ salt at 170–200 C, while
the Li+ salt continuously lost water up to 400 C. At 300 C
both the Li+ and Na+ salts transform over a period of 4 to
6 h completely into phases that have the above cubic unit
cell, with the Li+ salt the slower to undergo this transfor-
mation. Thus, all five (anhydrous) alkali metal salts and the
NH+4 salt are isostructural. The actual unit cell dimensions
are also given in Table 3 from the results of a Rietvelt pro-
file analysis and not surprisingly show an increase in value
with increasing size of the alkali metal cation, except for a
discontinuity between the Na+ and K+ salts. The unit cell
size of the NH+4 salt is about the same as that of the Rb+
salt, reflecting their similar cationic radii, while the origin
of the discontinuity between the Na+ and K+ salts may be
related to expansion of the lattices of the (water-soluble)
Li+ and Na+ salts following adsorption of water from the
air upon standing at room temperature.
Examination of the SEM micrographs of the Li+, Na+,
K+, Rb+, and Cs+ salts (all calcined at 300 C for 6 h) indi-
cate several features. For the soluble Li+ and Na+ salts the
crystals were, in general, large irregularly shaped blocks,
which varied in size from 2 to 20 m. However, for the
insoluble K+ and Rb+ salts, the crystals were considerably
FIG. 3. SEM images of (NH4)3[PMo12O40]: (a) commercial sample,
Aldrich (precipitated) (scale bar, 2 m); (b) synthesised by the thermal
decomposition of (nBu4N)3[PMo12O40] in air at 300 C (scale bar, 20 m);
(c) synthesised by the direct reaction of H3[PMo12O40] and excess gaseous
NH3 (scale bar, 20 m).

3
CATION EFFECTS ON CATALYTIC PROPERTIES OF [PMo12O40
]
367
smaller, with a range of 0.1 to
1
m, fairly spherical in and hence the ability to influence electron density at neigh-
shape and similar to the crystals of the commercial sample bouring atoms, and accordingly electron movements during
of (NH4)3[PMo12O40] (Fig. 3). This again likely results from reactions of the substrate and intermediate species at the
isotropic growth from solution. Their surfaces appeared to active site of a [PMo12O40]3 ion. It is a more appropriate
be free of defects. The Cs+ salt had a similar crystal shape property to express changes in catalytic activity in these pre-
to that of the K+ and Rb+ salts, but the crystal sizes were dominately ionic compounds than, for example, electroneg-
even smaller, with a range of 0.05 to 0.3 m. Thus, the sig- ativity, which is more applicable to molecular species. For
nificant differences in the crystal sizes, resulting from their the cation radii, crystal radii for 6-coordination were em-
varying solubilities, contribute to the observed differences ployed in this study (36). As noted above, the maximum
in surface area for both the Li+ and Na+ salts compared to conversion and selectivity appear to peak at the K+ salt.
those of the K+, Rb+, and Cs+ salts (Table 3), as described This may be deceptive, however, despite the fact that all
above for the NH+4 salts.
the compounds are isostructural. The three insoluble al-
For the series of alkali metal cations the catalysis results kali metal salts with K+, Rb+, and Cs+ cations were ion
are given in Table 2. These results show that an increased exchanged with their respective cations to remove as much
conversion percentage occurred at the higher temperature, adventitious H+ in their structures as possible. It has been
as expected, along with an increased selectivity percent- established that when prepared by the synthetic approaches
age to methacrolein (and methacrylic acid). The conversion generally employed, some H+ is always incorporated into
percentage generally follows the sequence Li+ < Na+ < K+, the structures of [PMo12O40]3 salts (24). In general, a sin-
with both the Rb+ and Cs+ salts less active than the K+ gle ion exchange will remove much of the H+ (24), although
salt. At first sight there appears to be little trend in ei- thisisonlypossible for insoluble materialsand hence not for
ther the conversion percentage or the selectivity percentage the Li+ and Na+ salts. Thus, the insoluble K+, Rb+, and Cs+
to methacrolein and methacrylic acid from monotonically salts may have been less active in the oxidative hydrogena-
changing the cation and hence any dependence on a prop- tion of isobutyraldehyde than expected, as the presence of
erty such as the ionic potential of the cation. The ionic
H
+ has a marked enhancing effect on the catalytic abilities
potential is given by the charge/cation radius (Z/r) and of these salts. This is investigated and discussed in more
provides a measure of the polarizing ability of the cation detail below.
TABLE 2
CatalyticOxidativeDehydrogenation ofIsobutyraldehydeUsingSaltsofComposition (C+)3[PMo12O40]
(Where C+ = H+, Li+, Na+, K+, Rb+, and Cs+) and the Mixed Cs+–H+ Species, (CsnH3 n)[PMo12O40]
(Where n = 1, 2, and 2.5)
% Selectivityb
Temp.
( C)
Rate
(104 mol min 1 g
Catalysta
H3[PMo12O40
Li3[PMo12O40
Na3[PMo12O40
K3[PMo12O40
Rb3[PMo12O40
Cs3[PMo12O40
(Cs2.5H0.5)[PMo12O40
% Conv.
)
MAL
MAA
ACE
CO
CO2
1
]
275
300
87
94
1.54
1.66
51
44
4
7
3
3
27
32
15
14
]
275
300
38
71
0.67
1.25
39
59
—
1
9
3
40
21
13
16
]
275
300
31
87
0.55
1.54
55
65
—
2
8
6
27
19
10
9
]
275
300
48
93
0.85
1.64
34
43
—
2
5
7
37
25
24
23
]
275
300
46
74
0.81
1.31
33
46
—
1
17
12
29
22
21
19
]
275
300
67
67
1.18
1.18
46
55
2
2
23
13
16
15
14
15
]
275
300
98
99
1.73
1.75
50
34
17
27
3
2
19
23
12
13
(Cs2H)[PMo12O40
(CsH2)[PMo12O40
]
]
275
300
22
46
0.39
0.81
56
54
—
2
10
7
21
22
13
15
275
300
26
60
0.46
1.06
58
62
—
1
9
5
20
20
13
12
1
a Mass of catalyst, 0.35 g; W/F = 0.14 g s ml
.
b As in Table 1.

368
HU AND BURNS
The amount of H+ present in each of the two soluble cata- catalytic activity from these species, especially in the pres-
lysts, Li3[PMo12O40] and Na3[PMo12O40], was determined ence of varying amounts of H+, are difficult to establish.
using a spectrophotometric method based on the conver- More significant variations would be expected among a se-
sion of the base form of the indicator bromothymol blue to ries of compounds involving greater changes in the formal
its acid form in anhydrous DMF. These studies showed that oxidation state and hence ionic radius of the cation, as well
the Li+ and Na+ salts had similar amounts of H+ present (8 being provided by variations in the size and distribution of
and 5% hydrolysis during preparation, respectively), which the atoms in a polynuclear cation, such as in the NH+4 and
is not surprising given their similar synthetic approaches. Me4N+ ions.
Thus, the reactivity enhancement afforded by H+ would be
To examine a greater range of ionic potential values, a se-
similar and have little effect on their actual relative reactivi- ries of compounds with increasing cationic charge and de-
ties. The differences in the catalytic abilities of the K+, Rb+, creasing crystal radius were prepared for comparison with
and Cs+ salts may simply reflect the ease by which these in- the salt Cs3[PMo12O40], such that all cations were isoelec-
soluble salts are able to undergo cation-exchange reactions tronic. These included Ba3[PMo12O40]2, La[PMo12O40], and
(with Cs+ Rb+ K+) and are thus able to replace H+ by Ce3[PMo12O40]4, which have as their cations Ba2+, La3+
,
the respective alkali metal cation in the structures of these and Ce4+. In this series the polarizing ability of the cation
salts. A reduced presence of H+ in the insoluble catalysts, increases greatly, with charge/crystal radius values of 1.34,
particularly Rb+ and Cs+, was evident from the Hammett 2.56, and 3.96, respectively (Cs+ salt, Z/r = 0.55). The cata-
indicator tests using 4-(phenylazo)diphenylamine and lysis data are given in Table 4. If the polarizing ability of
dicinnamalacetone, which exhibited a reduced presence of the countercation (as given by ionic potential) is important
the acid form on the surfaces of these insoluble catalysts in the catalytic performance of salts of the [PMo12O40]3
compared to that of the soluble catalysts.
ion, then any effects should be reflected in trends in the
As surface area has been shown not to be an important rates of reaction and in the conversion percentage of isobu-
factor in the relative catalytic activities of these species, tyraldehyde and in the selectivities for the various products.
it would seem that the reason for the observed variations The trends in the rates and conversion percentages at both
in catalytic activity of these isostructural alkali metal salts temperatures suggest an overall decrease with increasing
may be attributed to the changes in the polarizing ability ionic potential, while the trends for the selectivities indi-
(as given by ionic potential) of the cation, together with a cate little effect on the selectivity for methacrolein, except
consideration of the presence of any residual H+. Catalytic for the Ce4+ salt. In general, the rates and conversion per-
activity would therefore appear to be favoured by low ionic centage to methacrolein are reduced compared to those
potential, with Cs+ therefore likely to produce the most of Cs3[PMo12O40], although the latter compound has been
active solid catalyst. The order observed is, however, also ion-exchanged with CsCl to remove as much H+ as possi-
dependent on the extent of substitution of H+ for the al- ble. This is, however, impossible for the salts of the Ba2+
,
kali metal cation in the solid-state structure of the catalyst. La3+, and Ce4+ ions as these three compounds are water
However, changes in ionic potential do not actually vary soluble. Moreover, the actual structures of these salts have
greatly down the alkali metal series (Table 3), especially not been established, so that the arrangement of the cations
for the heavier cations, and thus discernable effects on the around the [PMo12O40]3 ion in the solid-state structures of
these compounds are currently unknown, and this may be
of particular importance in the actual processes that occur
at the catalyst surface. It should also be noted that there
TABLE 3
Crystallographic, Thermogravimetric Analysis, and Surface
is some evidence, for the alkaline earth metal cations, that
Area Data for the Alkali Metal and NH+4 Salts, (C+)3[PMo12O40]
the [PMo12O40]3 salts are difficult to synthesize (37, 38). It
(where C+ = Li+, Na+, K+, Rb+, Cs+, and NH+4 )
may be that, like the Li+ and Na+ salts, these salts (specifi-
cally Ba2+ in the present case) form only upon calcination
at higher temperatures. Indeed, the XRD patterns of these
species undergo significant changes upon calcination and
are currently under investigation.
Cation
crystal
radius
Unit cell % H2O Surface
Ionic
potentialb
dimension loss at
300 C (m2 g
)
˚
area
a
˚
1
Salt
(A)
(A)
The increased selectivities to methacrolein and metha-
crylic acid in the case of Ce3[PMo12O40]4 appear contrary
to the observed trend in catalytic behaviour of the Cs+-,
Ba2+-, and La3+-containing compounds. However, it should
be noted that of all the countercations discussed above, only
Ce(IV) is potentially capable of undergoing changes in the
oxidation state under the conditions employed in the cata-
lysis. Recently, it has been shown that for the mixed metal
Li3[PMo12O40
]
]
0.90
1.16
1.52
1.66
1.81
—
1.11
0.86
0.66
0.60
0.55
—
11.734(2)
11.936(1)
11.5975(2)
11.6687(5)
11.8412(9)
11.6575(3)
11.0
3.5
5.2
3.3
1.9
3.7
4.8
3.2
109
94
117
120
Na3[PMo12O40
K3[PMo12O40
Rb3[PMo12O40
Cs3[PMo12O40
(NH4)3[PMo12O40
]
]
]
]
a Six coordination.
b Ionic potential is given by the charge/crystal radius.

3
CATION EFFECTS ON CATALYTIC PROPERTIES OF [PMo12O40
]
369
TABLE 4
Catalytic Oxidative Dehydrogenation of Isobutyraldehyde Using S+alts of Composition
(Cn+
)
3/n[PMo12O40] (Where C+ = Cs+, Ba2+, La3+, Ce4+, NH4 , and Me4N+)
% Selectivityc
Temp.
( C)
Rate
(104 mol min 1 g
Catalysta
Cs3[PMo12O40
Ba3[PMo12O40
% Conv.
)
MAL
MAA
ACE
CO
CO2
1
]
275
300
67
67
1.18
1.18
46
55
2
2
23
13
16
15
14
15
]
275
300
28
34
0.49
0.60
15
46
—
—
17
8
55
33
13
13
2
4
La[PMo12O40
]
275
300
23
40
0.41
0.70
51
48
—
1
9
7
34
32
7
13
Ce3[PMo12O40
]
275
300
32
22
0.56
0.39
74
80
—
—
16
11
2
2
8
8
b
(NH4)3[PMo12O40
]
275
300
61
81
1.88
2.50
71
62
<1
2
6
5
10
18
12
13
b
(Me4N)3[PMo12O40
]
275
300
95
99
2.93
3.06
74
50
2
10
2
1
13
23
9
16
1
a Mass of catalyst, 0.35 g; W/F = 0.14 g s ml
b Mass of catalyst, 0.20 g; W/F = 0.08 g s ml
c As in Table 1.
.
.
1
n
species H4[PMo11VO40] under the conditions of catalysis, Pr4N+, and Bu4N+, do not result in greater increases in
the vanadium leaves the polyoxomolybdate framework and conversion percentage, as the charge on the atoms does not
results in V(IV) counterions of a partially reduced polyox- vary greatly with increasing size of the alkyl group (32).
omolybdophosphate ion (22). Indeed, the vanadyl salt of
[PMo12O40]3 , i.e., H(VO)[PMo12O40], exhibits reactivity
similar to that of the original mixed metal catalyst (22).
A second way of changing the ionic potential of a cation
is by delocalizing the charge over several atoms through
the use of a polynuclear cation. For this purpose catalysis
data for salts of the [PMo12O40]3 ion containing the NH+4
and (CH3)4N+ cations may be compared with the data for
Cs3[PMo12O40]. The data are also given in Table 4. Al-
though there are some differences in the sizes of these
ions, the main variations concern the locations and mag-
nitudes of the (net) atomic charges of the atoms that con-
stitute these cations. The net atomic charges for the atoms
of the NH+4 and (CH3)4N+ ions, based on extended-Hu¨ckel
molecular orbital calculations, are NH+4 (N, +0.15, and H,
+0.21) and (CH3)4N+ (C, +0.13; N, +0.14, and H, +0.03).
As the ionic potential is decreased or, as in the case of
NH+4 and (CH3)4N+, a progressively decreasing concentra-
tion of charge at any one atom based on the net charges of
the atoms concerned, the conversion percentage increased,
while the selectivity to methacrolein and methacrylic acid
both increased slightly compared to that of Cs3[PMo12O40].
The trends observed in the catalytic performance of the
above salts support the proposal of the effect of ionic po-
tential of the cation affecting the electron distribution of
the transition states and/or intermediates in the oxidative
dehydrogenation of isobutyraldehyde. Further increases in
D. Effect of Added H+ on the Catalytic Ability
of the (CsnH3 n)[PMo12O40] Salts
(i) Catalysis studies. To examine the effect of the pres-
ence of added H+ on catalytic ability, a series of materials
with varying amounts of H+ in their structures were synthe-
sized, with Cs+ as the alkali metal cation. The series chosen
was(CsnH3 n)[PMo12O40], where n had nominalvaluesof1,
2, and 2.5. These were examined for catalytic activity com-
pared to that of Cs3[PMo12O40] (n = 3) and also the free
acid, H3[PMo12O40] (n = 0), with the data given in Table 2.
The effect of the presence of some H+ in the structure on
the oxidative dehydrogenation reaction isevident, although
the addition of greater amounts of H+ appears detrimental
to the reaction. Thus, the maximum conversion and selec-
tivity to methacrolein and methacrylic acid were achieved
with the compound Cs2.5H0.5[PMo12O40].
(ii) X-ray diffraction studies. Examination of the XRD
patterns of the compounds (CsnH3 n)[PMo12O40], where
n = 0, 1, 2, 2.5, and 3, showed that when isolated directly
from solution, there was no evidence for the presence of
two phases for n = 2.5 and 2, but for n = 1 the XRD patterns
of both Cs3[PMo12O40] and H3[PMo12O40] (probably as the
23- or lower hydrate) were present. Thus, for n
2, the
two cations appear to form a single mixed phase when pre-
cipitated from solution, which allows for some variability
size of the cation, such as those found in the series Et4N+, in composition. Upon heating to 300 C for 4 h, it was hard

370
HU AND BURNS
to establish the presence of the anhydrous H3[PMo12O40] those of H3[PMo12O40] 23H2O and partially rehydrated
phase in (CsH2)[PMo12O40] as the diffraction pattern of the H3[PMo12O40] following thermal treatment.
dehydrated acid was not well defined.
When freshly precipitated, the samples with n = 2.5 and
2 showed no evidence of two signals, and both had a chem-
(iii) 31P solid state NMR studies. To further explore ical shift almost the same as that of Cs3[PMo12O40]. This
the possibility of the formation of a solid solution for appeared at 4.55 0.03 ppm in all of the spectra, whether
(CsnH3 n)[PMo12O40] with n = 2 and 2.5, the solid state the samples were calcined or not. Only with n = 1 was evi-
31P NMR of all compounds in the series (n = 0, 1, 2, dence for two signals apparent (at 4.18 and
4.5 ppm).
2.5, and 3) were obtained, both on freshly isolated ma- In this case two components could be observed, which is
terial and after heating to 300 C for 4 h. The spectra consistent with the XRD data described above, showing
of the calcined materials are given in Fig. 4, along with the presence of a second phase at this composition. In view
FIG. 4. Solid-state 31P NMR spectra of CsnH3 n[PMo12O40] (n = 0, 1, 2, 2.5, and 3): (a) Cs3[PMo12O40], (b) Cs2.5H0.5[PMo12O40], (c) Cs2H[PMo12O40],
and (d) CsH2[PMo12O40], all calcined at 300 C for 4 h, (e) H3[PMo12O40] 23H2O, and (f) H3[PMo12O40] that has been partially rehydrated after thermal
treatment at 300 C.

3
CATION EFFECTS ON CATALYTIC PROPERTIES OF [PMo12O40
]
371
of the closeness of the chemical shifts of Cs3[PMo12O40] and at the contact times used, as methacrylic acid would have
H3[PMo12O40] 23H2O ( 4.02 ppm, Fig. 4e) it is difficult to been observed even at low isobutyraldehyde conversions
separate the individualsignalsfor n = 2.5and 2, and solid so- if this reaction had occurred. Moreover, no isobutyric acid
lutions may even form for these compositions. This appears was directly observed as a product species. Examination
to be slightly different from the results previously obtained of the data in Figs. 1 and 2 and in Tables 2 and 4 also in-
byBlack etal. (39) on the analogousKnH3 n[PMo12O40]sys- dicates that the formation of acetone is favoured at lower
tem, where separate peaks associated with K3[PMo12O40] temperatures (275 vs 300 C) and that the presence of H+
and partially hydrated H3[PMo12O40] were able to be re- favours the formation of CO as well as reduction of acetone
solved. This may, however, depend on the extent of hydra- formation.
tion of the free acid, which varies over a range of chemical
shifts [Fig. 4f and Ref. (39)].
The mechanism of formation of the above products can
be accounted for on the basis of Scheme 1:
Upon heating the materials of composition CsnH3
n
(CH3)2CHCHO(gas) →[(CH3)2CHCHO]ads
[2]
[3]
[PMo12O40] to 300 C for 4 h, significant changes to the spec-
tra were observed. The presence of Cs3[PMo12O40] (Fig. 4a)
was evident in all of the spectra, indicating the stability of
[O]
==
[(CH3)2CHCHO]ads→ H2C C(CH3)CHO + H2O
2[O]
this phase, but broad peaks at 3.55 and
1.8 ppm devel-
==
[H2C C(CH3)CHO]ads → ( CH3)2CO + H2O + CO [4]
oped for n = 2.5, 2, and 1 (Figs. 4b–4d, respectively). These
broad peaks are consistent with the formation of partially
hydrated forms of H3[PMo12O40] (Fig. 4f). Somewhat sim-
ilar results were observed for both the KnH3 n[PMo12O40]
(39) and CsnH3 n[PW12O40] systems (40). In the former
the appearance of several peaks was interpreted in terms
of the growth of an epitaxial surface layer of essentially
dehydrated acid on K3[PMo12O40]. However, in the lat-
ter the presence of separate components was suggested,
with the partially dehydrated acid, H3[PW12O40], trapped
in the Cs3[PW12O40] structure. More recently, examination
of the CsnH3 n[PMo12O40] system (0 < n < 3) following cal-
cination at 350 C using XRD, Raman spectroscopy, and
XPS for surface chemical analysis has concluded that two
phases (Cs3[PMo12O40] and H3[PMo12O40]) exist at all com-
positions, with the hydrated acid phase coating the pure
cesium salt (41). The solid-state 31P NMR data obtained
in the present study are consistent with this latter sug-
gestion.
n[O]
[(CH3)2CHCHO]ads → xCO + yCO2 + zH2O
[5]
SCHEME 1
where [O] represents a lattice oxygen. The last reaction has
not been written stoichiometrically as the relative yields of
CO and CO2 are uncertain and, also, as [(CH3)2CHCHO]ads
in Eq. [5] can be replaced by adsorbed methacrolein or ace-
tone, the complete oxidation of both of these products may
possibly occur prior to their desorption. According to the
above scheme, the yield of CO will always be greater than
that of acetone (compare Eqs. [4] and [5]), which is the case
observed in the present studies. As H+ favours the forma-
tion of CO over acetone, presumably Eq. [5] is also facili-
tated in some way by the presence of H+ and likely favours
CO over CO2 as the formation of the latter is little affected
by H+. The above mechanism is only meant to convey a
broad, overall approach and not a detailed description of
the individual steps involved.
The above reaction scheme is similar to, but simpler
than, that of the equivalent oxidative dehydrogenation
reaction of isobutyric acid in that it does not require a
(significant) competing pathway that is dependent on
the presence of H+, i.e., on the Brønsted acidity of the
catalyst (1, 19). In the case of isobutyric acid this pathway
II. The Mechanism of the Oxidative Dehydrogenation of
Isobutyraldehyde over Salts of the [PMo12O40]3 Ion
A. Mechanism of Oxidative Dehydrogenation
of Isobutyraldehyde
Reaction of isobutyraldehyde in the presence of oxygen involves the breakdown of adsorbed isobutyric acid to give
over salts of the [PMo12O40]3 ion at temperatures from 275 propene, CO, and H2O. Thus, any trends associated with
to 350 C produced methacrolein, methacrylic acid, acetone, cation effects (and thus ionic potential) will be more easily
CO, and CO2 as the major products observed. In the present seen in the present study as any contributions from this
studies there was no evidence for the formation of isobu- competing reaction do not occur.
tyric acid, propene, or acetic acid, although the last has been
The formation of methacrolein from isobutyraldehyde
observed as a very minor product in a previous study us- over a [PMo12O40]3 -containing catalyst is an oxidative
ing H5V2Mo10O40 as the catalyst (42). The methacrylic acid dehydrogenation reaction. Thus, following adsorption
was only formed when the conversion of the isobutyralde- of the isobutyraldehyde (Eq. [2] above), the sequen-
hyde was greater than about 50% . Although isobutyralde- tial loss of two hydrogen atoms must occur, resulting
hyde can be oxidised to isobutyric acid fairly readily, it ap- in methacrolein (these are combined in Eq. [3]). The
pears that under the conditions employed in the present mechanism of the oxidative dehydrogenation of isobu-
studies this reaction does not proceed to any great extent tyric acid to yield methacrylic acid is similar. For the

372
HU AND BURNS
oxidative dehydrogenation of isobutyric acid three mech- varying alkali metal cation showed increasing promotion of
anisms have been proposed for the sequential loss of reaction as the alkali metal cation increased in size, that is,
the two hydrogen atoms, involving either heterolytic or with decreasing polarizing ability, the trend proposed in the
homolytic processes (1, 20, 43). Similar mechanisms can present studies. Similarly, the same trend was also observed
be proposed in the present case for the oxidative dehy- by Kaszelten and Moffat (13) in their studies of the par-
drogenation of isobutyraldehyde, but whether the inter- tial oxidation of methane using N2O over silica-supported
mediate is a carbocation, carbanion, or radical species H3PMo12O40 and its salts. Notwithstanding the fact that in
cannot be inferred without further investigations. How- both of these studies the catalysts were supported, unlike in
ever, some proposals concerning the effects of variations the present studies, it is likely that, as similar dependencies
in the countercations can be made. Thus, upon adsorp- on the polarizing ability of the countercation occur in these
tion of an isobutyraldehyde molecule, a certain orien- reactions, the countercations again play an active role in
tation of the molecule on the oxomolybdate surface of influencing electron movements in the transition state(s) as
the polyanion will occur. In the vicinity of the adsorbed described above.
isobutyraldehyde molecule will reside various counter-
In the investigation of the effect of H+ on the conversion
cations. Presumably, the ionic potential of these cations (or percentages and rates of reaction for the sequence of com-
even single cation) plays a role in the two dehydrogena- pounds (CsnH3 n)[PMo12O40] (n = 1, 2, 2.5, and 3) in the
tion steps that lead to methacrolein (Eq. [3] above). It is present work, the maximum enhancement was observed
likely, therefore, that any strong electric field effects, rep- with (Cs2.5H0.5)[PMo12O40]. Interestingly, similar maxima
resented by a high ionic potential, are likely to have ad- in reactivity close to this composition have been observed
verse effects on the movement of electrons and hence bond in other quite unrelated systems, but using acidic salts of
stretching and breaking in the formation of the transition [PW12O40]3 , such as in the gas-phase methanol dehydra-
states. This will, in turn, act to increase one or both of the tion to dimethyl ether at 180 C and n-butane isomeriza-
free energies of activation for loss of the hydrogen (in what- tion over (CsnH3 n)[PW12O40] at 200 C (40), which gave
ever form) and hence decrease the rate of reaction and thus the highest activity over the range n = 2.7 to 2 for the former
the extent of conversion (for a given W/F or contact time and near 2 for the latter, and in the liquid-phase isobutane/
at a set temperature). Thus, by lowering of the ionic po- 2-butene+alkylation at 80 C and 2.5 MPa over the acidic
tential or through dissipation of the cationic charge over Cs+, NH4 , and K+ salts of H3[PW12O40] (44), which yielded
several atoms, such as with an NH+4 or (CH3)4N+ ion, the a maximum of n = 2.5 for the NH+4 and K+ salts and a maxi-
maximum rate of conversion and formation of the desired mum in the range of2
n
2.5for the Cs+ salt. Correlations
product can be achieved. Changes in ionic potential might with surface acidity have been proposed to explain the re-
also be expected to influence the selectivities to the vari- sults in these studies, and it is apparent that surface acidity
ous products as these are the result of competing reactions is important in the present study as well.
(Eqs. [3]–[5] above). Thus, for the alkali metal, Ba2+ and
Assuming that the H+ is evenly distributed throughout
La3+ salts, increasing ionic potential appears to disfavour the surface and bulk of the catalyst, then the variation of
the formation of acetone while having little effect on the surface acidity with n in (CsnH3 n)[PMo12O40] reaches a
selectivity to methacrolein. The NH+4 and (CH3)4N+ salts maximum value (46 mol g 1) at n = 2.5 in the present
behave similarly. The Ce4+ salt, however, while exhibiting studies. The surface acidity, which is proportional to
low conversion as a result of its very high ionic potential, the surface area times (3 n), the actual H+ content in
has a very high selectivity to methacrolein. As noted above, (CsnH3 n)[PMo12O40], is highly dependent on surface
this is likely the result of the electron transfer capability area and is slightly lower than the analogous value found
of the Ce(IV)/Ce(III) couple, such as has been suggested for the maximum in the related series of compounds
for V(V)/V(IV) in (thermally rearranged) H4[PMo11VO40] (CsnH3 n)[PW12O40] ( 70 mol g 1) (45, 46). This is
and in H(VO)[PMo12O40] (22). Notably, in the oxidative a direct consequence of the smaller surface area of
dehydrogenation of isobutyric acid, the latter two cata- (Cs2.5H0.5)[PMo12O40] (77 m2 g 1) compared to that of
lysts exhibit similar conversion percentages to that of (Cs2.5H0.5)[PW12O40] ( 130 m2 g 1). Thus, although a cata-
H3[PMo12O40], but have significantly increased selectivities lytic reaction may occur by bulk catalysis-type II, as in the
percentage to methacrylic acid as a result of the presence of present case, if the kinetically accessible mechanism con-
vanadium (22).
tains a step that depends on H+, then variations in rate will
Comparisons with the data obtained above for the al- actually depend on surface area in an indirect way through
kali metals as countercations can be made with the stud- the changing surface acidity. This is most obvious when
ies reported by Akimoto et al. (1) on the oxidative dehy- there are large changes in surface area, as occurs with these
drogenation of isobutyric acid to yield methacrylic acid, types of compounds, i.e., the series (CsnH3 n)[PMo12O40]
although in this work the catalysts were supported using in the present case, which shows a large increase in surface
quartz sand. Interestingly, in these studies, the trend with area for compositions above n = 2. Thus, for the series

3
CATION EFFECTS ON CATALYTIC PROPERTIES OF [PMo12O40
]
373
(CsnH3 n)[PMo12O40], where n = 0, 1, 2, 2.5, and 3, the sur- lattice oxygen) from the polyoxomolybdate framework and
face areas were 1.2, 0.8, 2.9, 77, and 117 m2 g 1, respectively. involves the formation of Mo(V) [two Mo(V) for every
phosphopolyoxomolybdate unit]. The change in total sta-
B. The Site of Anionic (Lattice) Oxygen Involved in the
Oxidative Dehydrogenation of Isobutyraldehyde
(and Related) Substrates
bilization energies, 1ET (eV), for this reaction increases in
the order
Oa(+7.64) < Ob(+12.47) < Od(+12.87) < Oc(+12.97).
The mechanism represented by Eqs. [2]–[5] above de-
scribes a series of steps in the oxidative dehydrogenation
of isobutyraldehyde and concentrates on the fate of the or-
ganic substrate. However, it has been established in previ-
ous studies that the source of oxygen, i.e., [O] in Eqs. [3]–[5],
is actually from the polyoxomolybdate framework, which
is eventually replenished by the molecular dioxygen in the
catalysis input stream. Infrared studies have suggested that
it is the bridging oxygen atoms of the [PMo12O40]3 ion
that supply the oxygen for this process (47–51). Thus, re-
duction of, for example, H3PMo12O40 by H2 at 280–320 C
showed that the (Mo–O–Mo) bands decreased relative
Not surprisingly, the lowest energy requirement is that for
an inner oxygen (Oa). However, this is inaccessable and will
not be involved in the catalysis process. The next most fa-
vorable site is that of an oxygen atom linking molybdenum
atoms within a Mo3O13 subunit (Ob), followed by a termi-
nal oxygen atom (Od) and finally a bridging oxygen atom
between Mo3O13 subunits (Oc). Although the latter three
possibilities are quite similar in their energy requirements,
that of loss of an oxygen atom within a Mo3O13 subunit is
slightly favoured. Thus, the above calculations support loss
of a bridging oxygen atom during oxidative dehydrogena-
==
to the (Mo O) band, suggesting loss of bridging oxygen
tion (and also oxidation) reactions over a [PMo12O40]3
containing catalyst, with an oxygen atom within a Mo3O13
subunit likely providing the source of oxygen.
-
atoms. Notably, the (P–O) band also decreased in intensity
during this reduction process. However, it has been shown
that similar reductions in relative intensity for these same
bands also occur as a result of electrochemical reduction,
where no structural changes take place.
C. The Role of H+ and the Mechanism of Oxidative
Dehydrogenation of Isobutyraldehyde
Examination of the Mulliken population analyses ob-
tained from extended-Hu¨ckel calculations of the bonds
from molybdenum to the four types of oxygen atoms (a
measure of bond order) are given in Table 5. These values
reflect the corresponding bond distances, which are also in-
cluded in the table. It is apparent that of the different types
of oxygen atoms, with the exception of the inner oxygen
atom, a bridging oxygen atom within an Mo3O13 subunit
is the likely candidate for the source of the lattice oxygen.
This can be confirmed by calculating the change in total
stabilization energy 1ET (where ET for a particular species
is the sum of the one-electron energies as calculated using
the extended-Hu¨ckel molecular orbital method) for each of
the potential oxygen positions (Oa → Od) for the reaction
It is evident that the presence of H+ has a marked effect
in the catalytic ability of the [PMo12O40]3 ion. The pres-
ence of H+ could potentially affect the mechanism given
by Eqs. [2] to [5] in several ways and may involve either
protonation of the [PMo12O40]3 ion or protonation of the
substrate (isobutyraldehyde), the latter probably as part of
a transition state or intermediate.
The likelyprotonation siteson the [PMo12O40]3 ion were
investigated using extended-Hu¨ckel molecular orbital cal-
culations, to see the effect of protonation on the HOMO–
LUMO gap and the total stabilization energy of the system.
Protonation may decrease the HOMO–LUMO energy gap,
therefore makingit easier to reduce thisanion, which occurs
during catalysis (1).
[PMo12O40]3 → [PMo12O39]3 + [O].
[6]
Placing a single H around the [PMo12O40]3 ion (H+ plus
[PMo12O40]3 , to give an overall [HPMo12O40]2 moiety)
in tetrahedral geometries with O–H distances as short as
This reaction simply represents loss of an oxygen atom (a
˚
1.00 A, the approximate O–H distance in water, for each
TABLE 5
of the oxygen environments (with the exception of Oa,
which is an internal atom) showed no effect on the HOMO–
LUMO gap. An alternative approach to protonation would
be to delocalize the interaction over several sites, by locat-
ing the H between several oxygen atoms, so that it forms
weak hydrogen bonds with these oxygen atoms. The effect
of protonation of the [PMo12O40]3 ion on the HOMO–
LUMO gap was therefore examined by allowing a hydro-
gen to approach the anion along selected directions, such
that the hydrogen would interact with several oxygen atoms
Mulliken Population Analysis for the [PMo12O40]3 Ion
Bond
distance (A)
Population
analysis
˚
Bond
P–O
1.542
2.434
1.923
1.911
1.676
0.688
0.128
0.442
0.475
0.726
Mo–Oa (inner)
O
b (bridging O within an Mo3O13 unit)
Oc (bridging O between Mo3O13 units)
d (terminal)
O

374
HU AND BURNS
in different environments. Three cases were investigated, oxygen vacancies could be formed by a dehydration mech-
which included approaches along a C3 axis of the anion anism, through the elimination of a water molecule from
from opposite sides, such that (i) the hydrogen interacted two protonated [PMo12O40]3 ions, yielding one intact an-
equally with three bridging oxygen atoms (Ob) within a ion and the second containing a vacancy. The vacancy could
single Mo3O13 unit or (ii) with three bridging oxygen atoms then be involved in further processes, such as supplying a
(Oc) linking three Mo3O13 units and (iii) along the C2 axis of favorable adsorption site or as a site that might receive an
the anion, such that interactions occurred with two Ob and oxygen atom following gas-phase dissociation of the ox-
two Oc oxygen atoms. In the first and third approaches, in- idant. This reaction was studied at 500–590 C, which is
teraction was also possible with an internal oxygen atom Oa. above the temperature of decomposition of H3PMo12O40
While protonation yielded an increase in the total sta- by loss of constitutional water ( 400 C). However, un-
bilization energies, resulting in energy “minima” along all der the conditions of reaction of the supported catalyst
three pathways, only the third direction of approach ac- Kaszelten and Moffat were able to show that not all of the
tually showed a decrease in the HOMO–LUMO gap. This [PMo12O40]3 ions had decomposed. It thus appears that
occurred at a P–H distance of 2.3A, which isslightlyinside oxidation catalysts based on [PMo12O40]3 (and thus also
˚
the confines of the peripheral oxygen atoms of the anion. likely [PW12O40]3 ) may exhibit different types of mecha-
This decrease was only some 0.19 eV in a HOMO–LUMO nisms over different temperature regimes.
gap of 4.79 eV, which is not highly significant, even consid-
ering the approximations in the extended-Hu¨ckel MO cal-
culational methodology. Thus, it would seem unlikely that
ACKNOWLEDGMENTS
protonation is important in increasing the ease of reduction
of the anion in the catalytic process. In passing, however,
it may be noted that, in protic solvents, protonation does
occur following multiple reductions of the [PMo12O40]3
ion (6), as a way of reducing (and hence stabilizing) the
increasing anionic charge.
The authors would like to acknowledge Mr. M. Littlefair, Ms. J. Ewing,
and Ms. S. Bell of B.H.P. (Newcastle Laboratories) for performing the
surface area and pore volume and size measurements, Ms. C. Allen and
Mr. K. Grice for help in obtaining the structural data, and Mr. G. We-
ber and Mr. D. Phelan, of the University Microscopy Unit, for recording
the scanning electron microscopy images. Dr. J. Hu would also like to ac-
knowledge the University of Newcastle for the award of a UNRS and an
OPRS.
An alternative possibility may be protonation of the
isobutyraldehyde, either as part of a transition state and/or
intermediate. This would occur at an electron-rich atom of
the isobutyraldehyde molecule, probably the carbonyl oxy-
gen atom. In doing so, this would help to delocalize the
negative charge of a possible carbanion intermediate, but
not a carbocation, and less likely, a radical intermediate.
Thus, the observed enhancement of catalytic activity in the
presence of H+ may provide evidence for the formation of
a carbanion intermediate and the mechanism of Otake and
Onoda (43). Protonation of the organic substrate is also
supported by the recent studies of Langepape et al. (41),
who have shown that, in the CsnH3 n[PMo12O40] system
(0 < n < 3), H3[PMo12O40] coats the Cs3[PMo12O40]. Thus,
as the oxidative dehydrogenation of isobutyraldehyde oc-
curs by bulk catalysis-type II, with the organic substrate
adsorbing on the catalyst surface, the presence of H+ on
this surface and thus available for protonation of the sub-
strate supports this possible pathway. Currently, however,
it is not possible to differentiate between protonation of
the anion (following reduction) or the substrate (transition
state or intermediate stages) and both may play a role in
the mechanism of oxidative dehydrogenation over a poly-
oxometalate catalyst, although the latter appears the most
likely at the present time.
REFERENCES
1. Akimoto, M., Tsuchida, Y., Sato, K., and Echigoya, E., J. Catal. 72, 83
(1981).
2. Konishi, Y., Sakata, K., Misono, M., and Yoneda, Y., J. Catal. 77, 169
(1982).
3. Misono, M., in “Polyoxometalates: From Platonic Solids to Anti-
Retroviral Activity” (M. T. Pope and A. Mu¨ller, Eds.), p. 255. Kluwer,
Dordrecht, 1994.
4. Misono, M., Catal. Rev.-Sci. Eng. 29, 269 (1987); 30, 339 (1988).
5. Mizuno, N., and Misono, M., Chem. Rev. 98, 199 (1998).
6. Pope, M. T., “Heteropoly and Isopoly Oxometalates.” Springer-
Verlag, Berlin, 1983.
7. Hu, J., Burns, R. C., and Guerbois, J.-P., J. Mol. Catal. A: Chem. 152,
141 (2000).
8. Cavani, F., Etienne, E., Favaro, M., Galli, A., Trifiro`, F., and Hecquet,
G., Catal. Lett. 32, 215 (1995).
9. Cavani, F., Comuzzi, C., Dolcetti, G., Finke, R. G., Lucchi, A., Trifiro`,
F., and Trovarelli, A., in “Heterogeneous Hydrocarbon Oxidation”
(B. K. Warren and S. T. Oyama, Eds.), ACS Symposium Series 638,
Chap. 10, p. 140. Am. Chem. Soc., Washington, DC, 1996.
10. Kaszelten, S., and Moffat, J. B., J. Catal. 106, 512 (1987).
11. Kaszelten, S., and Moffat, J. B., J. Catal. 109, 206 (1988).
12. Kaszelten, S., and Moffat, J. B., J. Catal. 112, 54 (1988).
13. Kaszelten, S., and Moffat, J. B., J. Catal. 116, 82 (1989).
14. Muneyama, E., Kunishige, A., Ohdan, K., and Ai, M., J. Mol. Catal.
89, 371 (1994).
15. Ishimi, K. (Nippon Kayaku Co. Ltd.) Japan Kokai, 54,046,705 (1975).
16. Ando, N. (Japan Synthetic Rubber Co. Ltd.) Japan Kokai, 53,124,211
(1978).
17. Wu, T., Leng, Y., Yang, H., Wang, G., Zhang, H., Hua, S., Jiang, Y.,
and Zhen, K., J. Mol. Catal. 57, 193 (1989).
Kaszelten and Moffat (10–13) have also provided ev-
idence for the involvement of protons in the partial
oxidation of methane using N2O over silica-supported
H3PMo12O40 and its salts. It was proposed that in this case

3
CATION EFFECTS ON CATALYTIC PROPERTIES OF [PMo12O40
]
375
18. Wang G., Hu, J., Yang, H., and Wu, T., Gaodeng Xuexiao Huaxue
Xuebao (Chemical Journal of Chinese Universities) 11, 1249 (1990).
19. Akimoto, M., Shima, K., Ikeda, H., and Echigoya, E., J. Catal. 86, 173
(1984).
20. McGarvey, G. B., and Moffat, J. B., J. Catal. 132, 100 (1991).
21. Komaya, T., and Misono, M., Chem. Lett. 1177 (1983).
22. Cadot, E., Marchal, C., Fournier, M., Te´ze´, A., and Herve´, G., in
“Polyoxometalates: From Platonic Solids to Anti-Retroviral Activity”
(M. T. Pope and A. Mu¨ller, Eds.), p. 135. Kluwer, Dordrecht, 1994.
23. Tsigdinos, G. A., Ph.D. thesis, Boston University, 1961.
24. McGarvey, G. B., and Moffat, J. B., J. Catal. 128, 69 (1991).
25. Massart, R., Contant, R., Fruchart, J. M., Ciabrini, M., and Fournier,
M., Inorg. Chem. 16, 2916 (1977).
37. McGarvey, G. B., and Moffat, J. B., Catal. Lett. 16, 173 (1992).
38. McGarvey, G. B., Taylor, N. J., and Moffat, J. B., J. Mol. Catal. 80, 59
(1993).
39. Black, J. B., Clayden, N. J., Gai, P. L., Scott, J. D., Serwicka, E. M., and
Goodenough, J. B., J. Catal. 106, 1 (1987).
40. Essayem, N., Coudurier, G., Fournier, M., and Ve´drine, J. C., Catal.
Lett. 34, 223 (1995).
41. Langpape, M., Millet, J. M. M., Ozkan, U. S., and Boudeulle, M.,
J. Catal. 181, 80 (1999).
42. Mu¨ller, E., and Hofmann, H., Chem. Eng. Sci. 42, 1705 (1987).
43. Otake, M., and Onoda, T., in “Proceedings, 7th International Congress
on Catalysis, Tokyo, 1980” (T. Seiyama and K. Tanabe, Eds.), B3.
Elsevier, Amsterdam, 1981.
26. Kozhevnikov, I. V., Russ. Chem. Rev. 56, 811 (1987).
27. Clare, B. W., Cook, D., Ko, E. F. C., Mac, Y. C., and Parker, A. J.,
J. Am. Chem. Soc. 88, 1911 (1966).
44. Corma, A., Martı´nez, A., and Martı´nez, C., J. Catal. 164, 422 (1996).
45. Okuhara, T., Nishimura, T., Watanabe, H., Na, K., and Misono, M., in
“Studies in Surface Science and Catalysis 90: Acid-Base Catalysis II”
(H. Hattori, M. Misono, and Y. Ono, Eds.), p. 419. Kodansha, Tokyo,
1994.
46. Okuhara, T., Nishimura, T., and Misono, M., Chem. Lett. 155 (1995).
47. Mizuno, M., Katamura, K., Yoneda, Y., and Misono, M., J. Catal. 83,
384 (1983).
28. Benesi, H. A., and Winquist, B. H. C., Adv. Catal. 27, 97 (1978).
29. Mealli, C., and Proserpio, D. M., J. Chem. Educ. 67, 399 (1990).
30. Strandberg, R., Acta Chem. Scand. A 29, 359 (1975).
31. Allmann, R., Acta Chem. Scand. A 30, 152 (1976).
32. Young, B., M. Sci. Studies Thesis, University of Newcastle, 1998.
33. Highfield, J. G., and Moffat, J. B., J. Catal. 88, 177 (1984).
34. McMonagle, J. B., and Moffat, J. B., J. Colloid Interface Sci. 101, 479
(1984).
48. Eguchi, K., Toyozawa, Y., Furuta, K., Yamazoe, N., and Seiyama, T.,
Chem. Lett. 1253 (1981).
49. Tsunecki, H., Niiyama, H., and Echigoya, E., Chem. Lett. 645 (1978).
35. Boeyens, J. C. A., McDougal, J., and Smit, J. V. R., J. Solid State Chem. 50. Akimoto, M., and Echigoya, E., Chem. Lett. 1579 (1981).
18, 191 (1976).
51. Tsunecki, H., Niiyama, H., and Echigoya, E., Chem. Lett. 1163
36. Shannon, R. D., Acta Crystallogr. A 32, 751 (1976).
(1978).