Effects of NH4/+, Cs+, and H+ counterions of the molybdophosphate anion in the oxidative dehydrogenation of isobutyric acid

Article abstract of DOI:10.1006/jcat.1999.2733

Mixed NH₃-Cs salts of the molybdophosphoric acid (Cs_x(NH₄)_3-x[PMo₁₂O₄₀]) were prepared by solid-state cationic exchange using the NH₃ salt impregnated with the appropriate amount of Cs nitrate at incipient wetness. Structural characterization of the solids revealed that the exchange was quantitative and that NH₃ and Cs cations were randomly distributed in the lattice. Catalytic activity of these salts in the oxidative dehydrogenation of isobutyric acid increased with increases in the amount of Cs ≤ 3 Cs per heteropolyanion. Between 3 and 3.1 Cs, the catalytic activity suddenly fell and, simultaneously, the major product switched from methacrylic acid (MAA) to acetone. This abrupt change in the activities for x ~ 3 corresponded with the strong decrease of the rates of formation of MAA and propylene, while the rate of formation of acetone remained almost unmodified. Departure of NH₃ was total for x ≥ 2 and partial for x < 2, thus, catalysts formed at steady state had to be considered as acid NH₃-Cs salts when x < 2 and acid Cs salts when 2 < x < 3. Very few protons were needed to obtain an active and selective catalyst in MAA.

Full text of DOI:10.1006/jcat.1999.2733

Journal of Catalysis 190, 173–181 (2000)
doi:10.1006/jcat.1999.2733, available online at http://www.idealibrary.com on
Effects of NH+4 , Cs+, and H+ Counterions of the Molybdophosphate
Anion in the Oxidative Dehydrogenation of Isobutyric Acid
C. Marchal-Roch, N. Laronze, R. Villanneau, N. Guillou, A. Te´ze´, and G. Herve´
´
´
Institut de Re´activite´, Electrochimie et Microporosite´, UMR 8637, Universite´ de Versailles, 45 Avenue des Etats-Unis, 78035 Versailles Cedex, France
E-mail: Herve@chimie.uvsq.Fr
Received August 4, 1999; revised October 1, 1999; accepted October 13, 1999
the cesium salt has a large surface area (about 200 m2/g) and
is microporous (10, 11). The influence of the counterions on
the heterogeneous catalytic activity is then rather complex.
Conflicting results have been obtained for the oxidative
dehydrogenation of isobutyric acid by cesium salts of the
molybdophosphoric acid (12–14). According to Akimoto,
cesium can enhance the reactivity of the oxygen atoms of
the polyanion (12) and the tricesium salt is very active. But,
on the other hand, Ai and Misono have reported a decrease
of the activity for acid cesium molybdophosphates with in-
creasing amount of cesium (13, 14) and the tricesium salt is
fairly inactive.
McGarvey et al. have studied the catalytic activity of
ammonium, cesium, and mixed ammonium–cesium molyb-
dophosphates obtained by cationic exchange in aqueous so-
lution in the oxidative dehydrogenation of isobutyric acid
(IBA) (15, 16). They concluded that the catalytic behavior
depends on the Cs+/NH4+ ratio and that even small quanti-
ties of ammonium can activate the catalyst.
Mixed ammonium–cesium salts of the molybdophosphoric acid
(Csx(NH4)3 x[PMo12O40]) have been prepared by solid-state
cationic exchange using the ammonium salt impregnated with the
suitable quantity of cesium nitrate at incipient wetness. Structural
characterization of the solids showed that the exchange is quanti-
tative and that ammonium and cesium cations are randomly dis-
tributed in thelattice. Catalyticactivityofthesesaltsin theoxidative
dehydrogenation of isobutyric acid increases as the amount of ce-
sium increasesup to3Csper heteropolyanion. Between 3and 3.1Cs,
the catalytic activity abruptly falls and, simultaneously, the major
product switches from methacrylic acid to acetone. This abrupt
change in the activities for x 3 corresponds to the strong decrease
of the rates of formation of methacrylic acid and propene, whereas
the rate of formation of acetone remains almost unmodified. Un-
der the conditions of reaction, departure of ammonia was total for
x
2 and partial for x < 2. Therefore, catalysts formed at steady
state have to be considered as acid ammonium–cesium salts when
x < 2 and acid cesium salts when 2 < x < 3. So, very few protons are
needed in order to obtain an active and selective catalyst in MAA.
c
2000 Academic Press
In an attempt to understand these large effects and
somewhat conflicting results, we prepared the mixed
ammonium–cesium and tricesium salts by a different
method, with the aim of using true stoichiometric salts, i.e.,
without residual protons, and we tested them for the oxida-
tive dehydrogenation of isobutyric acid.
Key Words: catalysis; heteropolyanion; ammonium–cesium
molybdophosphate; oxidative dehydrogenation; isobutyric acid.
1. INTRODUCTION
Utilization of Keggin heteropolyacids Hn[XM12O40] in
acid or mild oxidation catalyzed reactions has been largely
developed in recent years (1–4). Acidic and redox prop-
erties depend on the nature of the constitutive metal-
lic elements, usually tungsten, molybdenum, or vanadium:
tungstic compounds are more acid but less oxidizing than
2. EXPERIMENTAL
2.1. Preparation of the Samples
(NH4)3[PMo12O40]. Two solutions were prepared.
molybdenum ones, and partial substitution of vanadium (A) Ten grams of the molybdophosphoric acid were dis-
for tungsten or molybdenum enhances the oxidizing abil- solved in 75 ml of a solution of water/dioxane in equal
ity (5, 6). The Brønsted acidity can be controlled by the volumes. Then the heteropolyacid was neutralized by a so-
substitution of metallic cations for protons. For example, lution of ammonium hydrogenocarbonate, 0.9 M, in wa-
Cs3[PW12O40] does not show any acid catalytic activity for ter/dioxane in equal volumes. (B) Five grams of ammonium
dehydration of alcohols (4, 7–9). On the other hand, pro- nitrate were dissolved in 35 ml of the water/dioxane mix-
gressive Cs+/H+ substitution leads to important modifica- ture. Solution B was added dropwise to solution A under
tions of the crystalline structure and of the texture of the stirring. A bright yellow precipitate of ammonium molyb-
solid: H3[PW12O40] has a low surface area (<10 m2/g), but dophosphate appeared slowly, and the mixture was gently
173
0021-9517/00 $35.00
c
Copyright
2000 by Academic Press
All rights of reproduction in any form reserved.

174
MARCHAL-ROCH ET AL.
stirred for about 2 h. The solid was collected by filtration 2.3. Characterization Methods
through a fine frit, washed with cold water, and dried at
120 C for 1 h.
Nitrogen adsorption–desorption isotherms were ob-
tained using a Micromeritics ASAP 2010 apparatus. The
samples were evacuated at 150 C for 3 h. The results were
analyzed by the ASAP 2010 V3.00 program. Morphologies
of the catalysts were investigated by SEM using a JEOL
JSM-5800LV microscope.
Before characterization and catalytic testing, the ammo-
nium salt was heated under N2 flow (20 ml/min) using the
following procedure: (a) the temperature was raised to
250 C over 1 h (3.83 C/min) and maintained for 20 h at
this temperature; (b) the temperature was raised again to
350 C at 0.42 C/min and maintained for 30 h at this tem-
perature; (c) the sample was cooled to room temperature.
Analysis, % exp (theor.): Mo 56.97 (57.08), N 2.08 (2.07),
P 1.67 (1.54).
The X-ray diffraction patterns were recorded at room
temperature with a Siemens D 5000 diffractometer using
˚
Cu K radiation ( = 1.5418 A) and were fitted by means
of the Socabim program PROFILE. The unit cell parame-
ters were determined with the program DICVOL91 (17a)
Cs3[PMo12O40]. Two solutions were prepared. (A) Ten and were refined by means of the program NBS* AIDS83
grams of molybdophosphoric acid were dissolved in 110 ml (17b). X-ray thermodiffractometry was performed under
of a solution of water/dioxane (2/3). Then the heteropoly- air in an Anton Paar HTK16 high-temperature device of
acid was neutralized by addition of a solution of lithium a Siemens D 5000 diffractometer ( – mode) using Co K
˚
hydrogenocarbonate, about 0.5 M, until pH 4.5–5.5 (29 ml). radiation ( = 1.7903 A) and equipped with an M Braun
The solution of lithium hydrogenocarbonate was obtained linear position-sensitive detector (PSD). Samples were de-
by bubbling CO2 in a solution of lithium hydroxide, 0.5 M, posited directly on the heating platinum support connected
and it was titrated by HCl. (B) Three grams of cesium ni- to a thermocouple. Patternswere recorded from 30to 610 C
trate were dissolved in 50 ml of the water/dioxane mixture every 20 C, with a temperature ramp of 1.8 C min 1. The
(1/1) at 60 C. Solution B was added dropwise to solution temperature was stabilized 2 min before every measure-
A at 40 C under stirring. A yellow precipitate of cesium ment; each data collection lasted 570 s.
molybdophosphate appeared immediately. The solid was
Thermogravimetric analysis was carried out under air
collected by filtration through a fine frit, washed with cold flow(50ml/min) with a Perkin–Elmer electrobalance TGA-
water, and dried at room temperature for 2 days.
7 with a heating rate of 5 C/min up to 600 C. The infrared
spectra (KBr pellets) were recorded on a Fourier-trans-
formed Nicolet 550 apparatus.
Csx(NH4)3 x[PMo12O40], (x = 1, 2, 3, 3.1, 3.3, 3.5).
A
series of mixed ammonium–cesium salts was prepared by
thermal treatment of a solid mixture of ammonium molyb-
dophosphate and cesium nitrate: five grams of the insolu-
ble (NH4)3[PMo12O40] salt (2.48 10 3 mol), first dehydrated
for 1 h at 100 C, was impregnated at incipient wetness by
(x 2.48 10 3) mol cesium nitrate dissolved in 5 ml water.
The solid was then dried for 90 min at 120 C and ground us-
The solution 31P NMR spectra were recorded on a
Brucker AC 300 apparatus at 121.5 MHz. Chemical shifts
are referenced to 85% H3PO4.
3. RESULTS
ing a pestle and a mortar. The thermal treatment presented 3.1. Preparations of the Mixed Ammonium–Cesium Salts
above for (NH4)3[PMo12O40] allowed the Cs+/NH4+ cationic
Ammonium or cesium 12-molybdophosphates are usu-
exchange with simultaneous decomposition of ammonium
ally obtained by addition of ammonium or cesium chlo-
nitrate beginning at 210 C. The results of elemental analysis
ride (nitrate) to the solution of the heteropolyacid and
and thermogravimetric analysis agreed with the expected
filtration of the solid. An alternative method is the addi-
formula and revealed that a quantitative exchange was ob-
tion of ammonium (cesium) carbonate in a stoichiometric
tained whatever the x value.
amount and evaporation of water. In the first case, it is
difficult to eliminate the protons quantitatively, and salts
of mean composition M3 xHx[PMo12O40] (x 0.3) are cur-
2.2. Oxidative Dehydrogenation of Isobutyric Acid
Reactions were carried out at 320 C using a continu- rently obtained (18). In the second method, it is our ex-
ous flow pyrex microcatalytic reactor at atmospheric pres- perience that the basicity of CO23 is enough to partially
sure. The feed gas consisted of 2% IBA, 5.2% O2, and N2 decompose the [PMo12O40]3 anion that is stable only at
balance. The total flow rate was 180 ml/min. The catalyst pH < 2 in aqueous solution. In order to avoid hydrolysis of
(1 g) was introduced in the reactor and reached the temper- the heteropolyanion, bases of lower strength can be used.
ature of reaction in 30 min under an air flux of 180 ml/min. Another way is to utilize mixed water/organic solvent such
All reaction products were analyzed by on-line gas-phase as water/dioxane since hydroorganic solvents are known
chromatography with molecular sieves 5A, Porapak Q, and to stabilize the heteropolyanions. Hydrogenocarbonate ion
AT 1200 with 1% H3PO4 columns. The carbon, oxygen, and was chosen to neutralize acidity, and ammonium or cesium
hydrogen balances were in the range 90–100% .
nitrate were then added to obtain the desired salts. The

NH+4 /Cs+/H+ SALTS OF PMo12O430 FOR ODH OF IBA
175
FIG. 2. Infrared spectra of (NH4)3PMo12O40 (a), (NH4)3PMo12O40
impregnated by 3CsNO3 (b), (NH4)3PMo12O40 impregnated by 3CsNO3
after thermal exchange (c), and (NH4)3PMo12O40 after reaction (d).
FIG. 1. Thermogravimetric curves recorded under air flow of
(a) (NH4)3PMo12O40, (b) (NH4)3PMo12O40 + 3CsNO3 dried at 100 C, and
(c) (NH4)3PMo12O40 + 3CsNO3 after thermal exchange.
stability of the PMo12O340 was checked by 31P NMR during
the neutralization by(NH4)HCO3. Itschemicalshift slightly
increases from 3.10 to 2.99 ppm at the neutralization
(3 equivalents). It is only above neutralization (3.5 equiv-
alent) that other weak peaks appear at 0.51 ppm (15% )
and 1.85 ppm (2% ) when an excess of hydrogenocarbon-
ate was introduced. Therefore it is possible to neutralize
the heteropolyacid in this solvent without any decomposi-
tion. This method assures the decoupling of neutralization
and precipitation steps in order to obtain the required tri-
ammonium or tricesium salts without residual protons as
indicated by elemental analysis.
decomposition of the heteropolyanion was observed and a
quantitative substitution of cesium for ammonium can be
obtained as shown by infrared spectroscopy (Fig. 2), RX,
TGA, and elemental analysis.
3.2. Characterization of the Catalysts before Reaction
Ammonium, cesium, and mixed ammonium–cesium salts
crystallize in a cubic lattice, as previously reported for the
pure salts (10). Progressive substitution of Cs+ for NH4+
leads to a linear change of the lattice parameter (Fig. 3)
according to Ve´gard’s law, suggesting a random distribution
of the cations in the lattice.
The mixed ammonium–cesium salts were obtained by
solid-state cationic Cs+/NH+4 exchange which occurred dur-
ing a thermal treatment of the ammonium salt impregnated
by cesium nitrate. The exchange is promoted by the de-
composition of ammonium nitrate at temperatures higher
than 210 C. Figure 1 shows the thermogravimetric curves
of (NH4)3[PMo12O40] (a) and of the (NH4)3[PMo12O40] +
3CsNO3 mixture (b). Ammonia was eliminated at tem-
peratures higher than 410 C for the pure ammonium salt.
The weight loss observed between 150 and 270 C on curve
(b) was attributed to the decomposition of ammonium
Evolution of the X-ray powder patterns of (NH4)3
[PMo12O40] with temperature reveals a decomposition in
and
MoO3 at about 410 C when ammonia is elimi-
nated. In contrast, as is well known, the cesium salt is stable
up to 600 C. All the mixed salts Csx(NH4)3 x[PMo12O40]
with x < 3 present one cubic phase until 410 C. At this
nitrate since the relative intensity of the bands at 1409
dNH+4 ) and 1385 cm
( dNO3 ) on the IR spectrum
1
(
strongly decreased after a thermal treatment at 270 C. Ac-
tually, a treatment temperature of 350 C was required in
order to observe a complete Cs+/NH4+ exchange. But, if
the treatment was performed directly at 350 C, some lines
of molybdenum trioxide appeared on the X-ray powder
diffraction pattern, showing a partial decomposition of the
heteropolyanion. The treatment was then performed in two
steps, first, at 250 C in order to eliminate more than 80%
of the ammonium nitrate and, second, at 350 C in order to
FIG. 3. Lattice parameter of Csx(NH4)3 xPMo12O40 as a function of
complete the cationic exchange. Under these conditions, no xCsNO3 before (a) and after (b) reaction.

176
MARCHAL-ROCH ET AL.
FIG. 4. X-ray diffraction patterns recorded every 20 C from 30 to
590 C under static atmosphere and using Co K radiation of Cs1.5(NH4)1.5
PMo12O40. (The asterisk indicates the line of Pt, 2 scale.)
FIG. 5. Surface area of Csx(NH4)3 xPMo12O40 as a function of x be-
fore and after reaction.
temperature molybdenum oxide phases appear in decreas-
ing proportion as x increases. Therefore, the stability of the
heteropolyanion is not increased by the partial substitution
of Cs+ for NH+4 (Fig. 4).
ization in water/dioxane solution also forms spherical
aggregates of 1- to 3- m diameter. On the other hand, the
tricesium salt obtained by precipitation in pure aqueous
solution, studied for comparison, shows a quite different
aspect: fine particles (0.25 m) form irregular large plates
(>5 m).
The ammonium salt obtained in water/dioxane solvent
has a surface area of 120 m2/g, which is slightly lower than
the values observed when this salt was obtained from pure
aqueous solution (150 to 180 m2/g). Thermal treatment at
250 C then 350 C does not modify the surface area. Varia-
tion of the surface area of the solids obtained by Cs+/NH4+
thermal exchange with the quantity of Cs+ is shown in Fig. 5.
All the samples have a surface area of 10–15 m2/g. There-
fore, the Cs+/NH+4 exchange, even with small quantities of
cesium, leads to a collapse of the porous structure.
The morphology of the catalysts was observed by elec-
tron microscopy. The ammonium salt (Fig. 6) is composed
of spherical aggregates of 2.5- m average diameter. After
thermal exchange the resulting mixed ammonium–cesium
salts and the tricesium salt present the same morphology.
The cesium salt obtained by precipitation after neutral-
3.3. Catalytic Oxidative Dehydrogenation
of Isobutyric Acid
The catalytic activity and selectivities of (NH4)3
[PMo12O40] and Cs3[PMo12O40] obtained by precipitation
from the water/dioxane solution, reported in Table 1, are
in good agreement with those previously reported (14, 16).
Conversion of the triammonium salts decreased continu-
ously with time from 60 to 47% after 18 h. For the tricesium
salt the conversion remained very low, about 8% . The se-
lectivities of these two salts differed strongly: the major
product was methacrylic acid (48% ) for the ammonium
salts but was acetone (55% ) for the cesium salt.
TABLE 1
Catalytic Activities after 18 h of Reaction for 1 g of Hetropolysalts
Selectivities
IBA
conversion
MAA
48
Acetone
29
Propene
19
AcOH
1
COxa
(NH4)3[PMo12O40
(NH4)3[PMo12O40
]
]
47
3
+ xCsNO3 treated
x = 1
45
55
73
18
15
20
48
45
57
50
6
28
29
23
30
80
87
20
22
16
15
8
1
1
2
0
0
0
3
3
4
5
6
3
x = 2
x = 3
x = 3.1
x = 3.3
x = 3.5
Cs3[PMo12O40
5
6
]
8
14
55
17
0
13
a
The selectivities to COx are obtained after the subtraction of the contributions of the reaction of
formation of acetone propene and acetic acid.

NH+4 /Cs+/H+ SALTS OF PMo12O430 FOR ODH OF IBA
177
FIG. 6. SEM images of (NH4)3PMo12O40 (a), (NH4)3PMo12O40 + 3CsNO3 after thermal exchange (b), Cs3PMo12O40 obtained by precipitation in
water (c), and in water/dioxane (d).
The mixed ammonium–cesium salts obtained by thermal not modified for x = 1 but increased to 73% for x = 3. Selec-
exchange reached a steady state after 12 h of reaction. The tivity to MAA was constant (46% ) up to x = 2 and reached
results obtained after 18 h are presented in Fig. 7. Conver- 57% for x = 3. Conversely, the selectivity to acetone and
sion of IBA, which was 47% for the ammonium salt, was propene decreased between x = 2 and x = 3. Acetic acid
and COx were obtained in small proportions.
A drastic decrease in the measured catalytic activity
was observed when the composition was close to Cs+/
[PMo12O40]3 = 3, corresponding to the stoichiometric salt.
The conversion fell from 73% for x = 3 to 15% for x =
3.3. At the same time, selectivities were strongly modified,
56 to 6% in MAA and 23 to 80% in ACE. Several prepara-
tions and experiments were performed to confirm this ob-
servation.
3.4. Characterization of the Catalysts after Reaction
Elemental analysis showed that the amount of ammo-
nium cations was smaller after catalytic testing than in the
initial solids (Table 2). In agreement, the intensity of the
NH+4 vibration at 1400 cm 1 on the IR spectrum decreased
(Fig. 2). The loss of ammonia was very important during the
reaction since the percentage of ammonia release reached
43% for x = 0, 50% for x = 1, and 100% for x = 2. Catalysts
correspondingto x 2did not contain anymore ammonium
ion at the steady state.
The surface areas of all the samples were between 2 and
4 m2/g after reaction, whatever the Cs+/NH+4 initial com-
FIG. 7. Catalytic activity at 320 C of Csx(NH4)3 xPMo12O40 after 18 h
position (Fig. 5). Electron microscopy showed that the
catalytic reaction did not modify the morphology of the
as a function of x: IBA conversion (a) and selectivity to methacrylic acid
(b), acetone (c), propene (d), carbone oxides (e), and acetic acid (f).

178
MARCHAL-ROCH ET AL.
TABLE 2
ammonium, which could be correlated to the progressive
decrease of the conversion of IBA with time. Since protons
must remain after the release of ammoni+a, the catalyst has
to be considered either as a mixed NH4 /H+ salt or as a
mixture of ammonium salt and acid. X-ray patterns display
only the cubic phase characterizing the ammonium salt, but
with a slightly greater unit cell parameter, perhaps because
of the partial reduction of the heteropolyanion. It should
be noted that an acid phase eventually formed could be
difficult to identify by X-ray diffraction since acid is often
poorly crystallized after catalytic reactions.
Composition of Csx(NH4)3 x[PMo12O40] Determined by Elemen-
tal Analysis (1), Thermogravimetric Analysis (2), and Elemental
Analysis after Reaction (3)
x
0
1
2
3
NH+4 /HPA (1)
NH+4 /HPA (2)
NH+4 /HPA (3)
2.92
3.20
1.74
1.89
2.04
0.87
0.87
1.00
/
/
/
/
Note. (1) Under the limit of analysis.
It is noteworthy that the loss of ammonium occurs at
lower temperature under reaction conditions other than N2
or air flux. Formation of isobutyronitrile or methacryloni-
trile has been reported in the presence of the ammonium
salt (16), but none of these products were detected in this
work. It is known that ammonium partially reduces the het-
eropolyanion on evacuation (19), so the loss of ammonia
can be due to its oxidation catalyzed by the heteropolyan-
ion itself. Such a redox reaction could explain the deep blue
color of the ammonium salt, revealing the reduction of the
catalyst.
particles. X-ray diffraction patterns of all the catalysts after
reaction showed only the cubic-phase isotype with ammo-
nium and cesium salts, but with a lattice parameter signifi-
cantly increased, except for x close to 3 (Fig. 3).
The catalystswere blue after reaction when the initial am-
monium proportion was high, showing that they were sig-
nificantly reduced. This color disappeared progressively as
the cesium proportion increased, and the active exchanged
salt corresponding to x = 3 was yellow-green after reaction
and was then not reduced at the steady state. The nonactive
samples corresponding to x > 3 were also yellow-green.
For the mixed ammonium–cesium salts, the proportion of
ammonium lost during the reaction increases from 40% for
x = 0 to 100% for x > 2. So, the catalysts after reaction have
to be considered as Cs+/NH+4 /H+ salts when x < 2 and as
4. DISCUSSION
In a previous study, McGarvey et al. (11, 16) studied Cs+/H+ salts when 2 < x < 3. X-ray patterns show only the
the catalytic behavior of mixed ammonium–cesium molyb- cubic phase with a lattice parameter systematically greater
dophosphates for the oxidative dehydrogenation of isobu- than before catalytictest (Fig. 3). Thisvariation corresponds
tyric acid. These salts were obtained by ionic exchange be- to the increase of the Cs+/NH4+ ratio following the reaction.
tween ammonium and cesium cations in aqueous solution, For x > 2, the lattice parameter is that of the tricesium salt.
starting from the insoluble ammonium or cesium salts. Self- Once again, formation of acid cesium salts or of a mixture
exchanges have been also performed to substitute residual of tricesium and acid phases should have occurred.
protons of initial ammonium or cesium salts in order to
The oxidative dehydrogenation of isobutyric acid is a
approach the stoichiometric composition in NH+4 or Cs+. bulk type (II) reaction (20). McGarvey et al. did not men-
Nevertheless protons were still present as shown by ele- tion any influence of the surface area on the reactivity for
mental analysis (16). A conclusion of this study was that mixed ammonium and cesium salts (16). Moreover, a corre-
the tricesium salt is a little active but that small quantities lation between the catalytic activity and the surface acidity
of ammonium cations enhance greatly both the catalytic calculated from the value of the surface area before re-
activity and the selectivity to methacrylic acid.
action has been established for Cs+–H+ mixed salts of the
In the present work, ammonium, cesium, and mixed [PMo11VO40]4 heteropolyanion (21). The catalysts studied
ammonium–cesium salts were obtained by a new method. in this work have a low surface area before catalytic test,
First, stoichiometric triammonium or tricesium salts were about 10 m2/g whatever the Cs+/NH4+ ratio, and a lower sur-
obtained after neutralization of the heteropolyacid by a face area at the steady state ( 2 m2/g). Since all the catalysts
base of moderate strength in dioxane/water solvent. In this have quite the same low surface area, variations of the cata-
medium, the neutral heteropolyanion is stable and solid tri- lytic activity and selectivities are essentially dependent on
ammonium or tricesium salts can be obtained by addition other factors.
of the suitable nitrate.
A decrease of the reduction state of the catalysts at the
The catalytic activity of the triammonium salt decreases steady state was observed as the cesium fraction increased,
with time, from 60 to 47% after 18 h reaction. Character- suggesting that cesium enhances the rate of oxidation of
ization after reaction revealed a release of about 40% of the reduced catalysts. The rate of oxidation by O2 of acid–
the initial ammonium quantity, a decrease of the surface cesium molybdophosphates previously reduced by H2 was
area, an increase of the unit cell parameter, and a reduced studied by Misono (14). The tricesium salt showed the
state of the catalyst. The first important point is the loss of higher rate of oxidation, and the reversibility of the reaction

NH+4 /Cs+/H+ SALTS OF PMo12O430 FOR ODH OF IBA
179
SCHEME 1
reached 97% . In agreement with these results, we observed acid or acetone. Otake and Onoda have proposed a het-
that the active salt corresponding to x = 3 was fully oxidized erolytic process (23) in which the proton in the position
at the steady state.
to the carboxyl group is first eliminated by nucleophilic at-
The catalytic activity decreases abruptly when the com- tack by an oxygen atom of the heteropolyanion. Echigoya
position of the catalyst approaches that of the neutral tri- has proposed a homolytic process (12). Later, McGarvey
cesium salt. McGarvey et al. (16) have reported that for the et al. suggested a heterolytic process involving the initial ab-
mixed ammonium–cesium salts, only small quantities of the straction of a hydride ion to yield a carbocation (16) which
ammonium ions are required to maximize the methacrylic can yield methacrylic acid by abstraction of a proton or
acid selectivity. Results of the present study are in agree- acetone by nucleophilic attack of an oxygen atom.
ment with this conclusion if the initial composition of the
The strong decrease of the rate of formation of metha-
catalyst is considered. Actually, the cations in the active crylic acid when protons are quantitatively eliminated
catalysts are Cs+ and H+ since ammonia is quantitatively shows that methacrylic acid formation, but not acetone for-
eliminated during the reaction if x > 2. Therefore, the con- mation, needs acid sites, as was also previously reported
clusion is that very few protons, not ammonium cations, by Ai (24) for the study of the oxidative dehydrogenation
are sufficient to obtain an active catalyst in the oxidative of isobutyric acid over the H5[PMo10V2O40] heteropoly-
dehydrogenation of isobutyric acid.
acid. Thus, two parallel routes of activation of the isobu-
When all the protons are eliminated (x > 3) acetone be- tyric molecule yielding methacrylic acid and acetone can
comes the most abundant product, essentially because of be proposed. It can be observed that, at the temperature
the strong decrease of the rate of formation of methacrylic of the reaction (320 C), elimination of constitutional water
acid and also of propene, likely due to the disappearance occurs, at least partially, and generates coordinatively un-
of the corresponding catalytic sites. The decrease of the saturated molybdenum cations. Thus, molybdenum atoms
propene formation rate is evidently due to the lack of should have both Lewis acid and oxidizing properties. The
Brønsted acid sites since it is obtained by an acid-catalyzed possibility exists for formation of an organometallic inter-
mechanism (Scheme 1). Several redox mechanisms have mediate (Lewis acid properties of Mo, Scheme 2) of limited
been proposed in order to explain the concurrent formation stability at the temperature of reaction, whereby a con-
ofmethacrylicacid and acetone. Theyinvolve the formation certed transfer of electrons (oxidizing properties of Mo)
of a common intermediate that can yield either methacrylic yields methacrylic acid. In this second step, the formation of
SCHEME 2

180
MARCHAL-ROCH ET AL.
SCHEME 3
the double bond involves the attack of the hydrogen atom
An almost linear decrease of the catalytic activity with
by an oxygen atom of the heteropolyanion (basic site). With increasing cesium content has been reported for acidic-
the tricesium salt, no Lewis acid site could be generated cesium salts of [PMo12O40]3 prepared by partial neutral-
and methacrylic acid would not be obtained. In contrast, ization of the protons by addition of cesium carbonate and
in the first step of oxidation of isobutyric acid in acetone evaporation of water to dryness (14). It is now established
(Scheme 3), both oxygen basic sites and oxidizing proper- that this method of preparation leads to the formation of
ties of molybdenum can be implied. The breaking of the tricesium salt particles embedded in a bulk acid phase for
Mo–O bond is then assisted by a decarbonylation process x < 2, whereas tricesium salt particles appear to be coated
in the second step.
The role of cesium counterions of heteropolyanions in
with a thin film of the acid for higher cesium contents (22).
So, it appears that the catalytic activity of such acidic ce-
acid and redox catalytic reactions has been largely stud- sium salts decreases as the amount of the surface layer acid
ied. It is generally assumed that cesium enhances the ther- phase decreases. A very different behavior was observed
mal stability of the heteropolycompound. This conclusion with the ammonium-cesium salts obtained by thermal ex-
is not right for mixed cubic ammonium–cesium salts since change since the catalytic activity increases with cesium
crystallized MoO3 appears always at about 410 C (Fig. 4) content up to 3 and falls just after. The catalysts studied in
whatever the x value if x < 3. In addition, owing to their this+work might be at the steady state a solid solution with
low electronegativity and large size, the interaction of Cs+ NH4 /Cs+/H+ cations for x < 2 and Cs+/H+ cations when
cations with oxygen atoms is likely to be weak, and oxygen x > 2. It can be suggested that the particular catalytic behav-
atoms keep basic properties which play an important role in ior of these catalysts is related to the random distribution
the transformation of the isobutyric molecule (Schemes 2 of the cations in the solid and, likely on the surface.
and 3). The last step in the two catalytic redox reactions
involves reoxidation of the catalyst. Perhaps the most im-
portant role of cesium is to enhance the rate of oxidation of
REFERENCES
the reduced catalyst by dioxygen, solids being almost fully
oxidized at the steady state.
1. Misono, M., Catal. Rev. Sci. Eng. 29(2-3), 269 (1987).
2. Misono, M., and Nojiri, N., Appl. Catal. 64, 1 (190).
3. Kozhevnikov, I. V., Catal. Rev. Sci. Eng. 37(2), 311 (1995).
4. Okuhara, T., Mizuno, N., and Misono, M., Adv. Catal. 41, 113
(1996).
5. Pope, M. T., “Heteropoly and Isopoly Oxometalates,” Inorganic
Chemistry Concepts, Vol. 8. Springer–Verlag, New York, 1983.
6. Marchal, C., Davidson, A., Thouvenot, R., and Herve´, G., J. Chem.
Soc. Faraday Trans. 1 89, 3301 (1993).
In conclusion, with nearly stoichiometric cesium salts,
some heteropolyanions deficient in oxygen would be well
dispersed on the surface layer and would be oxidized at the
steady state, increasing the selectivity for the oxidation of
isobutyric acid to methacrylic acid.
Now it is possible to examine the apparent discrepancy
between some results in the literature, keeping in mind that
a very small proportion of protons, which are very difficult
to detect, can induce high reactivity for cesium salts. The
high catalytic activity observed by Akimoto (12) for the tri-
cesium salt would be related to salts with traces of protons.
In other works (13, 14), true stoichiometric salts were likely
obtained since protons were quantitatively neutralized by
cesium carbonate.
7. Okuhara, T., Kasai, A., Hayakawa, N., Yoneda, Y., and Misono, M.,
J. Catal. 83, 121 (1983).
8. Tatematsu, S., Hibi, T., Okuhara, T., and Misono, M., Chem. Lett., 865
(1984).
9. Ebeid, F., Ali, L., and Ali, A., Bull. Soc. Chim. Fr. 128, 644 (1991).
10. McGarvey, G. B., and Moffat, J. B., J. Colloid Interface Sci. 101, 51
(1988).
11. McGarvey, G. B., and Moffat, J. B., J. Catal. 130, 483 (1991).
12. Akimoto, M., Tsuchida, Y., Sato, K., and Echigoya, E., J. Catal. 72, 83
(1981).

NH+4 /Cs+/H+ SALTS OF PMo12O430 FOR ODH OF IBA
181
13. Ai, M., Appl. Catal. 4, 245 (1982).
19. Albonetti, S., Cavani, F., Trifiro, F., Gazzano, M., Koutyrev, M., Aissi,
14. Misono, M., Mizuno, N., and Komaya, T., in “Proceedings, 8th In-
F. C., Aboukais, A., and Guelton, M., J. Catal. 146, 491 (1994).
ternational Congress on Catalysis, Berlin, 1984,” Vol. 5. Dechema, 20. Mizuno, N., Watanabe, T., Mori, H., and Misono, M., J. Catal. 123, 157
Frankfurt-am-Main, 1984.
(1990).
15. McGarvey, G. B., and Moffat, J. B., J. Catal. 128, 69 (1991).
16. McGarvey, G. B., and Moffat, J. B., J. Catal. 132, 100 (1991).
21. Lee, K. Y., Oishi, S., Igarashi, H., and Misono, M., Catal. Today 33, 183
(1997).
17. (a) Boultif, A., and Loue¨r, D., J. Appl. Crystallogr. 24, 987 (1991); 22. Langpape, M., Millet, J. M. M., Ozkan, U. S., and Boudeulle, M.,
(b) Mighell, A. D., Hubbard, C. R., and Stalik, J. K., “NBS AIDS80:
A Fortran Program for Crystallographic Data Evaluation,” National
Bureau of Standards (U.S). Technical Note 1141, 1981. [NBS AIDS83
is an expanded version of NBS AIDS80.]
J. Catal. 181, 80 (1999).
23. Otake, M., and Onoda, T., in “Proceedings, 7th International Congress
on Catalysis, Tokyo, 1980” (T. Seiyama and K. Tanabe, Eds.), p. 780.
Elsevier, Amsterdam, 1981.
18. Highfield, J. G., and Moffat, J. B., J. Catal. 98, 245 (1986).
24. Ai, M., J. Catal. 98, 401 (1986).