Catalysis by heteropoly compounds Part 39. The structure and redox behaviour of vanadium species in molybdovanadophosphoric acid catalysts during partial oxidation of isobutane

Article abstract of DOI:10.1039/a708629f

The states and roles of vanadium of 11-molybdo-1-vanadophosphoric acid (H₄PMo₁₁VO₄₀, PMo11V) catalyst in the partial oxidation of isobutane (2-methylpropane) were analysed by EPR, ⁵¹V and ³¹P NMR, IR spectroscopy, and redox titration, and compared with dodecamolybdophosphoric acid (H₃PMo₁₂O₄₀, PMo12) catalyst. Thermal treatment of PMo11V at 623 K in O₂ caused the elimination of V from the Keggin anion and formed undefined polymeric and a monomeric V species. It was shown based on the redox titration and EPR spectra that the average valency of V in used catalysts significantly changed with the oxygen partial pressure, while that of Mo remained almost unchanged for both PMo11V and PMo12. The selectivities to methacrylic acid (MAA) and methacrolein (MAL) extrapolated to zero conversion were similar for both catalysts, but PMo11V showed a significantly higher selectivity as the conversion increased. This is because the secondary reaction, that is, oxidative decomposition of MAL and MAA, was slower over PMo11V than over PMo12. This was consistent with the marked difference in the reaction order in oxygen pressure, and corresponded to the differences in the above redox properties.

Full text of DOI:10.1039/a708629f

View Article Online / Journal Homepage / Table of Contents for this issue
Catalysis by heteropoly compounds Part 39”
The structure and redox behaviour of vanadium species in
molybdovanadophosphoric acid catalysts during partial oxidation of isobutane
Kei Inumaru,* Akiko Ono, Hiroshi Kubo and Makoto Misono*¤
Department of Applied Chemistry, Graduate School of Engineering, T he University of T okyo,
Hongo, Bunkyo-Ku, T okyo 113, Japan
The states and roles of vanadium of 11-molybdo-1-vanadophosphoric acid (H PMo VO , PMo11V) catalyst in the partial
4
11
40
oxidation of isobutane (2-methylpropane) were analysed by EPR, 51V and 31P NMR, IR spectroscopy, and redox titration, and
compared with dodecamolybdophosphoric acid (H PMo
O
, PMo12) catalyst. Thermal treatment of PMo11V at 623 K in O
3
12 40
2
caused the elimination of V from the Keggin anion and formed undeÐned polymeric and a monomeric V species. It was shown
based on the redox titration and EPR spectra that the average valency of V in used catalysts signiÐcantly changed with the
oxygen partial pressure, while that of Mo remained almost unchanged for both PMo11V and PMo12. The selectivities to
methacrylic acid (MAA) and methacrolein (MAL) extrapolated to zero conversion were similar for both catalysts, but PMo11V
showed a signiÐcantly higher selectivity as the conversion increased. This is because the secondary reaction, that is, oxidative
decomposition of MAL and MAA, was slower over PMo11V than over PMo12. This was consistent with the marked di†erence
in the reaction order in oxygen pressure, and corresponded to the di†erences in the above redox properties.
Recently, selective oxidation of lower alkanes by catalysts
designed at the molecular level has attracted much atten-
tion.1h3 Heteropolyacids and their salts are advantageous
compounds for this purpose because of their molecular
nature: their oxidation ability and high acidity can be adjust-
ed by substitution of the constituent elements.4 It has been
reported that heteropoly compounds are active catalysts for
oxidation of ethane,5 n-butane,6 n-pentane7 and isobutyric
acid.8h12 In these reactions, the incorporation of vanadium
atoms into molybdophosphoric acid has been reported to
enhance catalytic performances. Therefore, the states and the
role of vanadium in the working state of the heteropolyacid
catalysts are of great interest. It was elucidated that the vana-
dium atom in H PMo VO is readily eliminated from the
molybdovanadophosphoric acid containing various metal
cations for this reaction.24 In this work, we attempted to elu-
cidate the structure of vanadium species and the redox behav-
iour of the catalysts during catalytic oxidation of isobutane,
by using redox titration, EPR, IR and NMR spectroscopy.
H PMo VO and H PMo
ing on the behaviour at low conversion levels to reveal con-
trasts between the two catalysts. The oxidation states of
O
were both studied, focus-
4
11
40
3
12 40
catalysts were quantitatively analyzed at various O pressures
2
in the working states, and then their relationship to the selec-
tivity was discussed.
Experimental
4
11
40
polyanion structure by thermal treatment.10,13,14 As for oxi-
dative dehydrogenation of isobutyric acid, a vanadyl salt of
molybdophosphoric acid was reported to be highly selective.10
When H PMo VO was used as a catalyst for this reaction,
Materials
H PMo
O
and H PMo VO were supplied by Nippon
3
12 40
4
11
40
4
11
40
Inorganic Colour & Chemical Co., Ltd. and used after extrac-
tion with diethyl ether and recrystallisation from water. The
a cubic phase assignable to a vanadyl salt was detected by
using in situ X-ray di†raction, where this phase was con-
sidered to be an active phase.11 As for oxidation of n-butane,
a dimer structure of two Keggin anions connected by a
VwOwV linkage was proposed as an active phase.6 E†ects of
vanadium in cationic sites as well as in the Keggin anions
have also been studied for n-butane oxidation.15,16 These
studies showed that various vanadium species are formed
during catalysis depending on the reactions and have great
inÑuence on the catalytic performance.
BET surface area of H PMo VO and H PMo
O
were
4
11
40
3
12 40
2.6 and 4.2 m2 g~1, respectively (measured after evacuation at
623 K with an ASAP-2000, Micromeritics Instrument). Here-
after, the catalysts used in this study are designated PMo11V
and PMo12, respectively, irrespective of their structural
changes during the pretreatment and the catalytic reaction.
Isobutane (99.5%, Takachiho Chemical Co. Ltd.), He
(99.995%, Nippon Sanso Co. Ltd.), O (99.9%, Suzuki Shokan
2
Co. Ltd.) and 1,4-dioxane (Tokyo Kasei) were used as
Direct oxidation of isobutane (2-methylpropane) is a chal-
lenging route for the production of methacrylic acid, to
replace the acetoneÈcyanohydrin process and two step oxida-
tion of isobutylene.1,17 After the Ðrst reports on hetero-
polyacid catalysts for this reaction,18 heteropolyacids and
their salts with various additives have been examined exten-
sively.19h23 Also for the oxidation of isobutane, vanado-
molybdophosphoric acids show performances better than
molybdophosphoric acids. For example, Mizuno et al. recent-
ly reported a fairly good catalytic performance of Cs salts of
received.
Oxidation of isobutane
Isobutane oxidation catalysed by PMo12 and PMo11V was
carried out in a tubular Ñow reactor. Typically, 1 g of PMo12
or PMo11V (sieved into 250È425 lm prior to use) was diluted
with 2 g of inert quartz powder and loaded in the reactor to
give a catalyst bed about 1È2 cm in height. The catalysts were
pre-treated in Ñowing O at 623 K for 1 h. The catalytic activ-
2
ity was measured typically under the following conditions:
¤ E-mail: M. Misono: tmisono=hongo.ecc.u-tokyo.ac.jp
” Part 38: ref. T Kengaku, Y. Matsumoto, K. Na and M. Misono,
J. Mol. Catal., in press.
partial pressures, P
balance; Ñow rate, 30 cm3 min~1; temperature, 623 K. The
\ 0.3 atm, PO2 \ 0.15 atm and He
isobutane
J. Chem. Soc., Faraday T rans., 1998, 94(12), 1765È1770
1765

View Article Online
pressures of isobutane and O were varied in the range 0.15È
2
0.6 and 0.11È0.3 atm, respectively.
Redox titration with KMnO
4
The degrees of reduction of catalysts after the catalytic reac-
tion were determined by means of redox titration with
KMnO . After the catalytic reactions, the Ñowing gas was
4
changed to He and the catalyst was rapidly cooled. Then the
catalyst (1 g) was dissolved into deionised water (1000 cm3) to
which H SO was added. The solution was titrated carefully
2
4
with a 0.01 M KMnO solution (Wako Chemical). The tem-
4
perature of the solution was precisely controlled at 343 K to
accelerate the electron transfer and to avoid the decomposi-
tion of KMnO . After KMnO was added, the colour of the
4
4
solution was pale and transparent enough to readily deter-
mine the end point of the titration. Titration of VOSO solu-
4
tions under the same conditions conÐrmed the validity of this
technique. The amounts of heteropolyacid in the solution
determined by ICP (ICAP-575, Nippon Jarrell-Ash) were used
to calculate the degree of reduction to avoid error due to
variation of crystallised water in the fresh catalysts.
EPR measurements
A reactor directly connected to an EPR tube was used to
avoid exposure of the samples to the air. After catalytic reac-
tions, the Ñowing gas was changed to He and the catalyst was
quickly cooled, followed by sealing the top and bottom of the
reactor tube. Then, the catalyst was transferred into the EPR
tube without exposure to the air and the EPR tube was sealed
by Ðring, keeping the catalysts at the bottom of the tube at
Fig. 1 Dependence of selectivities of products on isobutane conver-
sion for PMo11V(a) and PMo12(b). (…) MAA, (K) MAL, (=) AcOH,
(|) CO . Dashed lines represent the sums of MAA and MAL. Reac-
x
tion conditions: 623 K; isobutane : O : N \ 30 : 15 : 55, Ñow rate 30
2
2
cm3 min~1. Catalysts, 0.25È1.5 g.
liquid N temperature. EPR measurements (X-band) were per-
2
formed with a JES-RE-1X spectrometer (JEOL) at 100 K or
motional e†ect of V on the activity is small, while the e†ect of
pressure on the rate di†ered between the two catalysts as
room temperature. The resonance frequency, modulation
width and time constant were 9.218 GHz, 0.5 mT and 0.03 s,
respectively. DPPH and Mn2` in MgO were used for the cali-
bration of g and A values. Simulation of V hyperÐne structure
(hfs) was carried out according to the literature,25 assuming
axial symmetry.
O
2
will be described below. PMo11V showed higher selectivities
for partially oxidised products (MAA and MAL) than PMo12
at higher conversion levels. For instance, while the sum of
MAA and MAL (indicated by dashed lines in Fig. 1) at 1%
conversion was ca. 60% for both catalysts, the value was
higher for PMo11V (ca. 50%) than for PMo12 (ca. 30%) at
3% conversion. For both catalysts, the values approached ca.
80% when extrapolated to zero conversion, as shown in Fig.
1. The catalytic performance observed for PMo12 is compara-
ble with that reported in the literature, although it is less than
the acidic Cs24 and pyridine26 salts.
NMR and IR measurements
The pre-treated catalysts were transferred from the reactor
tube into a mixed solvent of 1,4-dioxane and water (7 : 3, v/v)
in an Ar-Ðlled bag to avoid absorption of moisture prior to
the dissolution. The concentrations of catalysts were adjusted
to 16 mmol dm~3. 51V and 31P NMR spectra were measured
with a JNM-EX 270 (JEOL) instrument equipped with a 10
mm tunable probe. D O was used as an internal locking
solvent for 31P NMR measurements. For comparison, a
VO ` solution was prepared by the dissolution of V O into
3.5 wt% HCl, followed by the addition of 1,4-dioxane. The
51V and 31P NMR chemical shifts are expressed with refer-
ence to VOCl and 85% H PO , respectively.
IR spectra of the catalysts were measured with an FTIR-
8300 (Shimadzu) instrument by the conventional KBr method.
The reaction orders for isobutane (0.15È0.60 atm) were 0.5
and 0.4 for PMo11V and PMo12, respectively. On the other
hand, the reaction order in O (0.11È0.30 atm) was signiÐ-
2
2
cantly di†erent between the two catalysts: 0.7 for PMo11V
and zero for PMo12.
Another di†erence between the two catalysts was found in
2
2 5
the dependence of the selectivity on O pressure, as shown in
2
Fig. 2. For PMo12 [Fig. 2(b)], MAL and MAA signiÐcantly
3
3
4
decreased, accompanying an increase in the formation of
CO , when the O pressure increased. The rate of the pro-
x
2
duction of MAA at PO2 \ 0.3 atm was one third that at
P
O2 \ 0.11 atm. Contrary to this, for PMo11V, the selectivity
Results
of MAA increased slightly at Ðrst and the rate of MAA pro-
duction increased by a factor of ca. 3 with an increase O
pressure from 0.11 to 0.3 atm [Fig. 2(a)]. The rate of MAA
formation for PMo11V was more than eight times greater
Catalytic reaction
2
Fig. 1 shows the selectivities of the oxidation of isobutane
catalysed by PMo11V and PMo12 as functions of the conver-
sion of isobutane. Products detected were methacrylic acid
(MAA), methacrolein (MAL), acetic acid (AcOH), CO and
than that for PMo12 at 0.3 atm O pressure.
2
Redox titration and IR characterisation of thermally treated
and used catalysts
CO . A trace amount of isobutene (2-methylprop-1-ene) was
2
sometimes detected. It was conÐrmed that the extent of con-
version was proportional to the contact time (W /F,
W \ catalyst weight, F \ Ñow rate) for each catalyst. As for
the catalytic activity per unit surface area, PMo11V and
PMo12 were comparable at PO2 \ 0.15, showing that the pro-
In relation to the e†ect of O pressure, the oxidation states or
the degrees of reduction (DR) of the two catalysts after cata-
lytic reaction were determined by redox titration with
2
KMnO and the results are given in Table 1. For PMo12, the
4
1766
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Article Online
Fig. 3 IR spectra of catalysts. (a) Fresh PMo11V, (b) PMo11V after
the catalytic reaction, (c) fresh PMo12 and (d) PMo12 after the cata-
lytic reaction. Reaction conditions were the same as in Fig. 1.
structural changes of H PMo VO caused by heating using
4
11
40
31P NMR spectroscopy in waterÈdioxane solution.13
Although VO ` easily exchanges with Mo of the Keggin
2
anion in aqueous solution,27 this exchange is very much sup-
pressed in waterÈdioxane solution and the states of V in the
solid state, as to whether V exists in the Keggin anion or not,
are well reproduced in solution.10 In order to conÐrm this, the
following experiments were carried out in this study. A VO `
2
aqueous solution prepared from V O was added to an
2
5
H PMo
O
waterÈdioxane solution (V/PMo
O
molar
3
12 40
12 40
ratio \ 3). Then the solvent was evaporated to dryness, and
the solid again dissolved in the same solvent. 31P and 51V
NMR of the resulting solution showed single peaks due to
PMo
O
3~ and VO `, respectively. This demonstrates that
12 40
2
their structures were unchanged after this procedure.
Fig. 2 Dependence of conversion of isobutane and selectivities of
Fig. 4(a) and (b) show 31P and 51V NMR spectra of a
products on O pressure for PMo11V(a) and PMo12(b). (L) Iso-
2
waterÈdioxane solution of PMo11V pretreated at 623 K in
butane conversion. The other symbols are the same as those in Fig. 1.
O , respectively. In Fig. 4(a) (31P NMR), a large peak at d
Reaction condition: 623 K, isobutane : O : N \ 30 : 11È30 : balance,
2
2
2
[2.91 assignable to PMo
O
3~ was observed in addition
catalyst, 1 g; Ñow rate, 30 cm3 min~1.
12 40
to a small peak at d [2.87 of PMo VO 4~. These peaks
were assigned based on the spectra of fresh PMo12 and
11
40
degree of reduction was low (on average, \0.1 electrons per
Mo atom or 0.8 electrons per polyanion) and were little inÑu-
PMo11V that gave peaks at d [2.91 and at d [2.87, respec-
enced by the O pressure. On the other hand, PMo11V was
tively (not shown). The integrated peak area of PMo
O
3~
2
12 40
reduced more extensively and the degree of reduction varied
was 3.8 times larger than that of PMo VO 4~.
11
40
substantially (from 2.1 to 0.84 e anion~1) with O pressure.
51V NMR spectroscopy provides more information about
the state of V eliminated from the Keggin anion. As shown in
Fig. 4(b), the 51V NMR spectrum consists of at least three
peaks at d [525, [542 and [555 as indicated by dashed
2
IR spectra of the catalysts are shown in Fig. 3. For PMo12,
typical IR bands assigned to MowOwMo (792, 873 cm~1),
MoxO (965 cm~1) and PwOwMo (1065 cm~1) changed
little before and after catalytic reaction [Fig. 3(c) and 3(d)].
For PMo11V, only small changes were observed; a shoulder
at 1075 cm~1 observed for the fresh catalyst [Fig. 3(a)], disap-
peared after the catalytic reaction [Fig. 3(b)], as reported in
the literature.14a,14b
lines. The peak at
d [542 is assignable to intact
PMo VO 4~, since fresh PMo11V gave a single peak at the
11 40
Solution NMR of fresh and thermally treated catalysts
In order to elucidate the structural changes of catalysts, 31P
and 51V NMR spectra of waterÈdioxane solutions of fresh and
used catalysts were measured. MarchalÈRoch et al. studied the
Table 1 Degree of reduction (DR) of PMo11V and PMo12 after
reaction as a function of the O partial pressure
2
DR by titration/e anion~1
O
pressure
/atm
2
PMo11V
PMo12
0.11
0.15
0.30
2.1
1.4
0.84
0.74
0.67
0.45
Fig. 4 31P (a) and 51V (b) NMR of an waterÈ1,4-dioxane solution of
PMo11V after treatment at 623K in O
2
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
1767

View Article Online
same position (not shown). The peak at d [525 is assigned to
a monomeric VO ` species as its chemical shift is the same as
2
that of a VO ` solution prepared from V O in the same
2
2 5
solvent. The broad peak at d [555 can probably be assigned
to polymeric species. The peak areas due to VO `,
PMo VO 4~, and the unknown species were ca. 19, 15 and
2
11
40
66% of the total V peak area, respectively, being consistent
with the 31P NMR spectra. When the solution was kept at
room temperature for 2È3 days, the 31P and 51V NMR
changed gradually in accordance with slow incorporation of V
into the Keggin anion to form PMo VO 4~. This also sup-
11
40
ports the assignments of the two peaks in the 51V NMR
(d [525 and [542). 31P and 51V NMR spectra of dioxaneÈ
water solutions were measured for PMo11V which had been
Fig. 5 EPR spectra of PMo12 after catalytic reaction at di†erent O
treated in O at various temperatures. Structural changes
2
2
pressures. (a) 0.11 atm and (b) 0.30 atm. Reaction conditions were the
were observed above 473 K, and PMo
O
3~, VO ` and the
same as in Fig. 2.
12 40
2
unknown species monotonically increased with an increase in
the treatment temperature.
[dashed line in signal (b)]. Obtained EPR parameters are
listed in Table 2. The fractions of Mo spins in the total spins
(Mo ] V) were calculated to be ca. 57, 49 and 62% for the
spectra in Fig 6(a), (b) and (c), respectively. Combining these
results with the overall DRs obtained by redox titration
(Table 1), the average valency of V and Mo were calculated.
The average valency of V changed from ca. 4.1 at 0.11 atm of
O to ca. 4.7 at 0.30 atm of O , while that of Mo remained at
EPR analyses of catalysts used at various O pressures
2
Fig. 5 and 6 respectively show EPR spectra measured at 100
K of PMo12 and PMo11V after catalytic reaction at various
O
pressures. The temperature of EPR measurement did not
2
a†ect the signal shapes in the range from 100 K to room tem-
perature. For PMo12 (Fig. 5), the signal was broad and no
hyperÐne structure (hfs) was observed. The signal shape was
little inÑuenced by the O pressure (0.11È0.3 atm); g and g
2
2
5.9È6.0.
2
A
M
values were calculated and are listed in Table 2.
EPR spectra of PMo11V after catalytic reaction (Fig. 6)
contain an hfs signal of V in addition to the signal of Mo.
Direct simulation and curve Ðtting of the spectra is difficult
because the central part of the V signal is hidden by the broad
and intense Mo signal. Hence, it was attempted to extract the
hfs of V by assuming that each spectrum can be expressed by
a linear combination of each component. As the line shape
changed little from spectrum to spectrum, this assumption
may be acceptable for the present semi-quantitative dis-
cussion. For example, subtraction of the signal in Fig. 6(b)
multiplied by a factor of 0.77 from the signal in Fig. 6(a) gave
the spectrum shown in Fig. 7(a) in which the hfs signals of V
were obviously eliminated (note that the ordinate is expanded
in comparison with that in Fig. 6). The spectrum in Fig. 7(a) is
very similar to that reported for Mo5` in thermally treated
H PMo
O
. By subtracting the signal in Fig. 7(a) from Fig.
3
12 40
6(a), the signal in Fig. 7(b) is obtained. Here the hfs of V is
clearly seen. The hfs was well reproduced by a simulation as
shown in Fig. 7(c), except for the broad superimposed signal
Fig. 6 EPR spectra of PMo11V after catalytic reaction at di†erent
pressures. (a) 0.11 atm, (b) 0.15 atm, and (c) 0.30 atm. Reaction
O
2
conditions were the same as in Fig. 2.
Table 2 EPR parameters of catalysts after reaction and comparison with catalysts in the literature
Mo
V
sample
g
g
g
g
A /G
A /G
ref.
A
M
A
M
A
M
PMo12a
PMo11Va
1.853
1.853
1.852
1.966
1.961
1.958
this work
this work
28
1.930
1.962
164
40
H PMo
evacuated at 473 K
O
3
12 40
H PMo
decomposed at 773 K
O
1.873
1.938
1.930
1.949
29
3
12 40
PMo
(VO)HPMo
O
4~
30
31
31
32
12 40
O
1.930
1.936
1.926
1.961
1.974
1.956
165
164
163
44
57
41
12 40
40
PMo VO 5~
11
H PMo VO
4
11
40
treated at 623 K in air
H PMo V O
1.936
1.920
1.946
1.968
1.965
1.953
166
207
181
74
77
78
6
5
10 2 40
calcined at 573 K
V O /Al O
33
2
5
2 3
a Measured after reaction at 623 K; isobutane: O : N \ 30 : 15 : 55; Ñow rate, 30 cm3 min~1. Estimated errors are ^0.002 for g values.
2
2
1768
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Article Online
Table 3 Relative rate constantsa
catalyst
k
k
k
k
k
0
1
2
3
4
PMo11V
PMo12
0.78
0.83
0.22
0.17
45
40
25
70
50
120
a The values are normalised so that k ] k \ 1.
0
1
nation of ca. 80% of V from PMo VO 4~ anions. The
11
40
elimination indicates the following transformation took place,
in general agreement with the results reported in the liter-
ature.10,13,14
(12/11) PMo VO 4~ ] PMo
11 40
O
3~
12 40
] (1/11) PO ] (12/11) VO
x
y
As shown in Table 2, after the catalytic reaction, the g
values (g \ 1.853, g \ 1.961) and the lineshape of the Mo
A
M
EPR signal of PMo11V were similar to those for
H PMo
O
which had been evacuated at 473 K (Table 2)28
3
12 40
and for H PMo V O reduced by H at 698 K.34 This signal
6
9 3 40
2
was assigned by Fricke and Ohlmann28 to Keggin anions
having lost some of their bridging oxygen (“signal FÏ in the
literature). On the other hand, the g values of the present
study were much di†erent from those of decomposed
Fig. 7 EPR signals of Mo (a) and V (b) which were extracted from
EPR spectra of PMo11V after catalytic reaction at di†erent O pres-
2
sure, and simulated V signal (c).
H PMo
O
at 773 K 28,29 (presumably MoO ) (Table 2),
3
12 40
3
Discussion
MoO supported on silica (g \ 1.882, g \ 1.940),35 and
3
A
M
from those of electrochemically reduced PMo
O
4~ (Table
Comparison of catalytic properties between PMo12 and
PMo11V
12 40
2).30 The IR bands characteristic of the Keggin structure were
maintained for PMo11V after catalytic reaction, except for the
disappearance of a small shoulder at 1075 cm~1. Thus, all of
these data are consistent with the idea that most of the Mo
atoms of PMo11V exist in reduced Keggin anions, although V
atoms were eliminated from the anion structure.14a,14d For
PMo12, IR spectroscopy showed that most of the Keggin
structure was maintained after catalytic reaction. The g values
of EPR signals of PMo12 were close to those of oxygen-
deÐcient Keggin anions described above.
It has been claimed in patents that PMo11V is a superior
catalyst to PMo12 for the partial oxidation of isobutane as
well as for other selective oxidations. Most of the patents con-
cerning this reaction employed reaction conditions with a
high isobutane/oxygen ratio, considering commercial pro-
cesses in which unconverted isobutane is recycled. These
studies indicate that the reaction proceeds in a stepwise
manner as is shown in Scheme 1, where the activation of iso-
butane is slow.
51V NMR spectra of solutions of catalysts after pretreat-
ment provides information about the state of V eliminated
from the polyanion structure. There were at least three kinds
of species (Fig. 4); PMo VO 4~, monomeric VO `, and
Similar reaction conditions to the patents were employed in
this study and, as shown in Fig. 1, PMo11V showed a higher
selectivity than PMo12 for MAA and MAL, and the di†er-
ence was large at high conversions, consistent with the pre-
vious reports.19h24 The sums of selectivities to MAA and
MAL extrapolated to zero conversion were ca. 80% for both
PMo11V and PMo12 at PO2 \ 0.15 atm, indicating that the
di†erence between the two catalysts in selectivities are small in
the Ðrst step of this reaction, i.e., the step from isobutane to
MAL. As described above, the di†erence in catalytic activity
per surface area was also not large at PO2 \ 0.15 atm. There-
fore, the major di†erence between the two catalysts exists in
the selectivity at higher conversions. The relative rate con-
stants (k Èk in Scheme 1) were roughly estimated from the
11
40
2
unknown species which correspond(s) to the broad peak at d
[555. Since the composition of the solution of VO ` and
2
H PMo
O
did not change upon evaporation and re-
3
12 40
dissolution, the unknown species detected must have formed
during the pretreatment but not in solution. This (these)
species is (are) presumably polymeric with or without a phos-
phorus atom, since monomeric species must be detected as
VO ` at the pH (\2) of this solution.36
2
The EPR signal of V comprises a broad intense signal,
which is indicated by a dashed line in Fig. 7(b), and a weak
but sharp signal with clear hfs. For the sharp component, the
EPR hfs parameters are very close to those for the monomeric
0
4
conversionÈselectivity correlation in Fig. 1, assuming Ðrst
order reactions with respect to organic molecules for all steps
in Scheme 1 and neglecting the small change of O pressure
by its consumption, and results are listed in Table 3. It is
evident that the oxidative decomposition of MAA and MAL
is much slower for PMo11V than for PMo12 (see k and k in
Table 3).
VO2` cation in (VIVO)HPMo
O
.31 Similar signals with
12 40
hfs were reported for V species from heteropolyanions,10,32,37
2
but di†erent from the other signals in Table 2.
The broadness of the other signal suggests the presence of
the dipoleÈdipole interaction. This broad signal was not
observed for V-free catalyst as shown in Fig. 5. Fricke et al.
also reported a similar result for H PMo V O (n \ 0È
3
4
3`n 12vn n 40
H PMo VO , EPR indi-
Structure of PMo11V and PMo12 during the catalytic reaction
3) reduced by H .34 For Ce
2
0.25
3
11
40
cated that VO2` eliminated from Keggin anions formed
dimers.38 In terms of the number of spins, the broad signal
was the major component. This species is possibly associated
with the unknown species detected by 51V NMR in dioxaneÈ
water solution of the used catalysts [Fig. 4(b)].
31P NMR spectra of waterÈdioxane solutions showed that
thermal treatment of PMo11V at 623 K caused the elimi-
Redox behaviour of catalysts and possible correlation with
catalytic performance
Scheme 1
The present results demonstrate the di†erent redox behaviour
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
1769

View Article Online
9
M. Akimoto, H. Ikeda, A. Okabe and E. Echigoya, J. Catal.,
1984, 89, 196.
of the two catalysts. When the O pressure was varied, the
degree of reduction (DR) at steady state of PMo11V changed
much more than that of PMo12. According to semi-
quantitative EPR analysis the larger change for PMo11V was
mainly due to V, since Mo was essentially hexavalent and
2
10 E. Cadot, C. Marchal, M. Fournier, A. Teze and G. Herve, in
Polyoxometalates, ed. M. T. Pope and A. Muller, Kluwer Aca-
demic Publishers, New York, 1994, p. 315.
11 Th. Ilkenhans, B. Herzog, Th. Braun and R. Schlogl, J. Catal.,
little depended on the O pressure for PMo11V and PMo12.
1995, 153, 275.
2
12 K. Y. Lee, S. Oishi, H. Igarashi and M. Misono, Catal. T oday,
1997, 33, 183.
13 C. Marchal-Roch, R. Bayer, J. F. Moisan, A. Teze and G. Herve,
T op. Catal., 1996, 3, 407.
14 (a) C. Rocchiccioli-Deltche† and M. Fournier, J. Chem. Soc.,
Faraday T rans., 1991, 87, 3913; 1992, 88, 1227; (b) B. Taouk, D.
Ghoussoub, A. Bennani, E. Crusson, M. Rigole, A. Aboukais, R.
Decressain, M. Fournier and M. Guelton, J. Chim. Phys., 1992,
89, 435; (c) M. Fournier, C. Feumi-Jantou, C. Rabia, G. Herve
and S. Launay, J. Mater. Chem., 1992, 2 971; (d) C. Rabia, M. M.
Bettahar, S. Launay, G. Herve and M. Fournier, J. Chim. Phys.,
1995, 92, 1442.
15 D. Casarini, G. Centi, P. Jiru, V. Lena and Z. Tvaruzkova, J.
Catal., 1993, 143, 325.
16 T. Bergier, K. Bruckman and J. Haber, Recl. T rav. Chim. Pays-
Bas, 1994, 113, 475.
17 M. Misono and N. Nojiri, Appl. Catal., 1990, 64, 1.
18 H. Krieger and L. S. Kirch, US Pat. 4 260 822, 1981, assigned to
Rohm & Haas.
The catalytic behaviour of the two catalysts also varied dif-
ferently with the O partial pressure. At low O pressure, the
2
2
catalytic activities of the two catalysts were similar. However,
for PMo11V, the oxidation to MAL was promoted consider-
ably by O , while the decomposition of MAL and MAA was
2
not much accelerated. By contrast, the activity of PMo12 was
little dependent on the O pressure and the decomposition of
2
MAL and MAA was faster.
The di†erent behaviours of PMo11V and PMo12 are prob-
ably related. Although there is little direct evidence to correl-
ate the behaviour, the following speculation may be made.
Since the essential trend is similar for PMo11V and PMo12,
the reactions proceed mainly on the Keggin polyanions of
dodecamolybdophosphate (PMo
steps shown in Scheme 1.
O
) following the reaction
12 40
The VO species which are present near the polyanions in
y
the case of PMo11V are more easily reduced and reoxidised
19 H. Imai, T. Yamaguchi and M. Sugiyama, Jpn. Pat., JP 63-
as shown above, so it may be possible that active oxygen
145249, 1988, assigned to Asahi Chemical.
20 S. Yamamatsu and T. Yamaguchi, Jpn. Pat., JP 2-42032, 1990,
assigned to Asahi Chemical.
21 S. Yamamatsu and T. Yamaguchi, Jpn. Pat., JP 2-42033, 1990,
species formed by the adsorption on V4` ion such as
V5`wO ~ and V5`wO~ accelerate the activation of iso-
2
butane. Another possibility is that lattice oxygen of VO acti-
assigned to Asahi Chemical.
y
vates isobutane as in the case of oxidation of n-butane by
22 K. Nagai, Y. Nagaoka and M. Ohsu, Eur. Pat., EP-0418657-A2
(JP 3-106839), 1991, assigned to Sumitomo Chemical.
23 T. Kuroda and M. Okita, Jpn. Pat., JP 4-128247, 1992, assigned
Mitsubishi Rayon.
24 (a) N. Mizuno, M. Tateishi and M. Iwamoto, Appl. Catal. A:
Gen., 1994, 118, L1; (b) N. Mizuno, M. Tateishi and M. Iwamoto,
J. Catal., 1996, 163, 87.
25 C. P. Stewart and A. L. Porte, J. Chem. Soc., Dalton T rans., 1972,
1661.
(VO) P O .39
2 2
7
The slower rates of oxidative decomposition of MAL and
MAA for PMo11V than for PMo12 are the other character-
istics presumably brought about by the presence of V species
near the polyanions. It was reported that the presence of
reducible metal cations such as Cu2` and Pd2` promoted the
reduction and reoxidation of H PMo
O
.40 If so, in the
3
12 40
case of PMo11V, the V species would make the reduction and
reoxidation of the polyanions easier and more readily trans-
form the adsorbed oxygen species, which is probably active
for the decomposition of MAA and MAL, to less active lattice
oxygen of polyanions and improve the selectivity for MAL
and MAA.
26 W. Li and W. Ueda, Catal. L ett., 1997, 46, 261.
27 P. Courtin, Rev. Chim. Miner., 1971, 8, 75.
28 R. Fricke and G. Ohlmann, J. Chem. Soc., Faraday T rans. 1,
1986, 82, 263.
29 R. Fricke and G. Ohlmann, J. Chem. Soc., Faraday T rans. 1,
1986, 82, 273.
30 C. Sanchez, J. Livage, J. P. Launay, M. Fournier and Y. Jeannin,
J. Am. Chem. Soc., 1982, 104, 3194.
31 R. Bayer, C. Marchal, F. X. Liu, A. Teze and G. Herve, J. Mol.
Catal. A: Chem., 1996, 110, 65.
This work was supported in part by a Grant-in-Aid in Pri-
ority Area from the Ministry of Education, Science, Sports
and Culture.
32 R. Fricke, H-G. Jerschkewitz and G. Ohlmann, J. Chem. Soc.,
Faraday T rans. 1, 1986, 82, 3479.
33 V. K. Sharma, A. Wokaun and A. Baiker, J. Phys. Chem., 1986,
90, 2715.
References
34 R. Fricke, H-G. Jerschkewitz and G. Ohlmann, J. Chem. Soc.,
Faraday T rans. 1, 1986, 82, 3491.
1
2
3
F. Cavani and F. TriÐro` , Catalysis 1994, 11, 246.
35 M. Che, J. C. McAteer and A. J. Tench, J. Chem. Soc., Faraday
T rans. 1, 1978, 74, 2378.
H. H. Kung, Adv. Catal., 1994, 40, 1.
Activation and Functionalization of Alkanes, ed. C. L. Hill, John
Wiley & Sons, New York, 1989.
36 M. T. Pope, Heteropoly and Isopoly Oxometalates, Springer-
Verlag, Berlin, 1983, p. 35.
4
T. Okuhara, N. Mizuno and M. Misono, Adv. Catal., 1996, 41,
113.
37 M. Akimoto, H. Ikeda, A. Okabe and E. Echigoya, J. Catal.,
1984, 89, 196.
5
6
S. S. Hong and J. B. Mo†at, Appl. Catal., 1994, 109, 117.
G. Centi, V. Lena, F. TriÐro, D. Ghoussoub, C. F. A•ssi, M.
Guelton and J. P. Bonnelle, J. Chem. Soc., Faraday T rans., 1990,
86, 2775.
38 A. Abouka•s, C. Haubtmann, J. J. Andre, C. Desquilles, M.
Dourdin, I. Mathes-Juventin-Andrieu, F. C. A•ssi and M.
Guelton, J. Chem. Soc., Faraday T rans., 1995, 91, 1025.
39 R. Contractor, J. Ebner and M. J. Mummey, in New Development
in Selective Oxidation, ed. G. Centi and F. TriÐro, Elsevier,
Amsterdam, 1990, p. 553.
7
8
(a) G. Centi, J. Lopez-Nieto, C. Iapalucci, K. Bruckman and
E. M. Serwicka, Appl. Catal., 1989, 46, 197; (b) K. Bruckman,
J. Harber and E. M. Serwicka, Faraday Discuss Chem. Soc., 1989,
87, 173.
(a) M. Otake and T. Onoda, in Proc. 7th Int. Congr. Catal.,
Kodansha, Tokyo, 1980, p. 780; (b) M. Otake and T. Onoda,
Shokubai (Catalysts), 1976, 18, 169.
40 M. Akimoto, K. Shima, H. Ikeda and E. Echigoya, J. Catal.,
1984, 86, 173.
Paper 7/08629F; Received 10th December, 1997
1770
J. Chem. Soc., Faraday T rans., 1998, V ol. 94