Catalysis by heteropoly compounds: 34. Skeletal isomerization of n-butane over Pt- or Pd-promoted cesium hydrogen salts of 12-tungstophosphoric acid

Article abstract of DOI:10.1006/jcat.1997.1735

Skeletal isomerization of n-butane has been studied over bifunctional catalysts consisting of noble metals and an acidic cesium salt of 12-tungstophosphoric acid, Cs_2.5H_0.5PW₁₂O₄₀. At 473 K in the presence of 0.05 atm of H₂, 1 wt% Pt-Cs_2.5H_0.5PW₁₂O₄₀ (1% in weight of Pt impregnated to Cs_2.5H_0.5PW₁₂O₄₀) showed activity and selectivity comparable to 1 wt% Pt-sulfated zirconia, while at 573 K and 0.5 atm of H₂, 1 wt% Pt- and Pd-Cs_2.5H_0.5PW₁₂O₄₀ became much more active and selective for the formation of isobutane than 1 wt% Pt-sulfated zirconia and 1 wt% Pt-exchanged ZSM-5. Negative dependence of the initial reaction rate on H₂ pressure and little dependence on Pt loading level for Pt-Cs_2.5H_0.5PW₁₂O₄₀ indicate that the reaction proceeds through a bifunctional mechanism in which n-butane is rapidly dehydrogenated to n-butenes on Pt, n-butenes isomerize to isobutylene on acid sites, and then isobutylene is hydrogenated to isobutane on Pt. A remarkable effect of the presence of hydrogen was the repression of the catalyst deactivation. This is probably caused by the hydrogenation of coke and/or its precursor and a decrease in the butene partial pressure. Very interesting was the unique function of protons; protons present in proximity of Pt or Pd particles suppress the hydrogenolysis that takes place on Pt or Pd and greatly increase the selectivity to isobutane.

Full text of DOI:10.1006/jcat.1997.1735

JOURNAL OF CATALYSIS 170, 96–107 (1997)
ARTICLE NO. CA971735
Catalysis by Heteropoly Compounds
34. Skeletal Isomerization of n-Butane over Pt- or Pd-Promoted Cesium Hydrogen
Salts of 12-Tungstophosphoric Acid1
Kyutae Na, Toshio Okuhara,2,3 and Makoto Misono3
Department of Applied Chemistry, Graduate School of Engineering, The University of Tokyo, Bunkyo-ku, Tokyo 113, Japan
Received August 27, 1996; revised March 24, 1997; accepted April 21, 1997
Solid acid catalysts combined with noble metals are
called “bifunctional catalysts” and are effective for skele-
tal isomerization of n-butane and other alkanes (11–21).
Pt/chlorinated-Al2O3 has been used commercially for this
reaction (1). The addition of Pt to chlorinated-Al2O3 ac-
celerated greatly n-butane isomerization, while it little
changed the acid strength (12, 13). Recently, Pt-promoted
SO24ꢁ/ZrO2 was reported to be active for isomerization
of alkanes in the presence of H2 (14–20). H3PW12O40, a
popular heteropolyacid, is a strong acid and catalyzes ef-
fectively various reactions (22, 23). Ono and co-workers
observed that PdxH3ꢁ2xPW12O40/SiO2 catalyzed n-hexane
isomerization in the presence of H2 (24). The conversion
was nearly independent of the Pd content, while it was
low without Pd. We reported preliminarily that Pt- or Pd-
promoted Cs2.5H0.5PW12O40 was active and selective for the
n-butane isomerization (25).
The roles of the metallic component of bifunctional
catalysts have been claimed to be (i) dehydrogenation of
n-butane to butenes, which are precursors of sec-butyl car-
benium ion intermediate and (ii) formation of hydrogen
atom on the surface from a gas phase H2 (1, 11). Roles of
hydrogen atoms are possibly (i) suppression of the coke for-
mation by hydrogenation of coke precursors (11, 20) or (ii)
generation of new protonic acid sites from hydrogen atoms
(18, 21).
In the present study, we investigated n-butane isomeriza-
tion using a cesium hydrogen salt of 12-tungstophosphoric
acid promoted by noble metals, that is, Pt- or Pd-
Cs2.5H0.5PW12O40, to elucidate the roles of the metals, H2,
and protons on the activity and selectivity. The effects of
H2 pressure, the loading amount of Pt, pretreatment con-
ditions, and reaction temperature have been examined. By
comparing the catalysts with other bifunctional catalysts
such as sulfated zirconia promoted by Pt, Pt-SO24ꢁ/ZrO2,
and Pt-loaded hydrogen form zeolite, Pt-HZSM-5, the cata-
lytic features of Pt- or Pd-Cs2.5H0.5PW12O40 have been
demonstrated with emphasis on the reaction mechanism.
Skeletal isomerization of n-butane has been studied over bifunc-
tional catalysts consisting of noble metals and an acidic cesium
salt of 12-tungstophosphoric acid, Cs2.5H0.5PW12O40. At 473 K in
the presence of 0.05 atm of H2, 1 wt% Pt-Cs2.5H0.5PW12O40 (1%
in weight of Pt impregnated to Cs2.5H0.5PW12O40) showed activity
and selectivity comparable to 1 wt% Pt-sulfated zirconia, while at
573 K and 0.5 atm of H2, 1 wt% Pt- and Pd-Cs2.5H0.5PW12O40 be-
came much more active and selective for the formation of isobutane
than 1 wt% Pt-sulfated zirconia and 1 wt% Pt-exchanged ZSM-5.
Negative dependence of the initial reaction rate on H2 pressure
and little dependence on Pt loading level for Pt-Cs2.5H0.5PW12O40
indicate that the reaction proceeds through a bifunctional mech-
anism in which n-butane is rapidly dehydrogenated to n-butenes
on Pt, n-butenes isomerize to isobutylene on acid sites, and then
isobutylene is hydrogenated to isobutane on Pt. A remarkable ef-
fect of the presence of hydrogen was the repression of the cata-
lyst deactivation. This is probably caused by the hydrogenation of
coke and/or its precursor and a decrease in the butene partial pres-
sure. Very interesting was the unique function of protons; protons
present in proximity of Pt or Pd particles suppress the hydrogenoly-
sis that takes place on Pt or Pd and greatly increase the selectivity to
c
isobutane.
ꢀ 1997 Academic Press
INTRODUCTION
Skeletalisomerization ofn-butane to isobutane isofgreat
importance because isobutane is a feedstock for alkylation
with butenes to C8 alkylates and synthesis of methyl tert-
butyl ether (MTBE) (1). Solid acid catalysts such as sulfated
zirconia (SO24ꢁ/ZrO2) (2–4), zeolites (5–7), chlorinated-
Al2O3 (8), and heteropoly compounds (9, 10) have been
reported to catalyze this reaction.
1
For part 33, see Einaga, H., and Misono, M., Bull. Chem. Soc. Japan
70, (1997), in press.
2
Present address: Graduate School of Environmental Earth Science,
Hokkaido University, Sapporo 060, Japan.
3
Corresponding authors.
96
0021-9517/97 $25.00
c
Copyright ꢀ 1997 by Academic Press
All rights of reproduction in any form reserved.

BUTANE ISOMERIZATION OVER HETEROPOLY COMPOUNDS
EXPERIMENTAL
97
solution was stirred at 353 K for 3 h, the solid was filtered
and washed with deionized water, and then dried at 373 K.
One wt% Pt-SO24ꢁ/ZrO2 (125 m2 gꢁ1) was prepared by
a method as in the literature (18). Zr(OH)4 obtained from
ZrOCl2 was treated with an aqueous solution of H2SO4
(1 N), and the resulting solid was calcined at 873 K for 5 h
in air. Then the obtained SO24ꢁ/ZrO2 was impregnated with
the aqueous solution of H2PtCl6 (0.04 mol ꢂ dmꢁ3) and was
calcined at 873 K in air. This catalyst is denoted by 1 wt%
Pt-SO24ꢁ/ZrO2. 1 wt% Pt-Al2O3 (N. E. Chemcat Co., Lot
No. 137-90099, 154 m2 gꢁ1) was also used as reference.
1. Catalysts
Cesium salts, Cs2.5H0.5PW12O40 and Cs3PW12O40, were
prepared as described previously (26). These salts will
be denoted by Cs2.5 and Cs3, respectively. Four kinds of
Pt-Cs2.5H0.5PW12O40 were prepared. The first series was ob-
tained from H2PtCl6 (Kojima Chem. Co.), H3PW12O40 (Nip-
pon Colour and Chemical Co.), and Cs2CO3 (Merck Co.).
To the aqueous solution of H3PW12O40 (0.08 mol ꢂ dmꢁ3
,
,
44 ml), an aqueous solution of H2PtCl6 (0.04 mol ꢂ dmꢁ3
15 ml) was dropped at 323 K to obtain a yellow solution.
Then the aqueous solution of Cs2CO3 (0.12 mol ꢂ dmꢁ3
,
2. Skeletal Isomerization of n-Butane
36 ml) was dropped to the solution at 323 K. The resulting
milky solution was evaporated at 323 K to solid. This series
of catalyst will be denoted by Pt-Cs2.5H0.5PW12O40(A)
(abbreviated as Pt-Cs2.5(A)). The loading amount of Pt
was 1.0 wt% .
Skeletal isomerization of n-butane was carried out at
a temperature in the range from 473 to 573 K in a con-
tinuous flow reactor (Pyrex, 12 mm of inside diameter)
under an atmospheric pressure (9, 10). The reactant mix-
ture consisting of n-butane (0.05 atm), H2 (0ꢃ0.50 atm),
and N2 balance was fed to 0.5–1.0 g of catalyst. Total flow
rate was 10 cm3 ꢂ minꢁ1. Prior to the reaction, the het-
eropoly compounds were pretreated in the flow of O2 or H2
(60 cm3 ꢂ minꢁ1) at 573 K for 1 h. Pt-SO42ꢁ/ZrO2 was pre-
treated in the O2 flow (60 cm3 ꢂ minꢁ1) and then in an H2
flow (60 cm3 ꢂ minꢁ1) at 643 K for 1 h each. Pt-ZSM-5 and
Pt-Al2O3 were pretreated in the O2 flow and then in the
H2 flow at 723 K for 1 h each. The reaction products were
analyzed with an FID GC (Hitachi GC-163) equipped with
VZ-10 and SE-30 columns (9, 10, 25).
The second series was prepared from Pt(NH3)4Cl2
(Aldrich Chem. Co.) by the same method as de-
scribed above. This series will be denoted by Pt-Cs2.5H0.5
PW12O40(B) (abbreviated as Pt-Cs2.5(B)). The loading
amounts of Pt were changed from 0.1 to 1.5 wt% .
The third series was prepared by impregnating 1 wt% Pt-
Cs3PW12O40(A) with the aqueous solution of H3PW12O40,
where 1 wt% Pt-Cs3PW12O40(A) was prepared from the
aqueous solution of H2PtCl6 and was pretreated at 573 K
in a flow of O2 for 2 h. The amount of H3PW12O40 added
to 1 wt% Pt-Cs3PW12O40(A) was adjusted to obtain the
ratio of Cs+ to PW12O43ꢁ0 = 2.5. After the impregnation, it
was again treated at 573 K in the O2 flow for 2 h. This series
will be denoted by H3(imp)/Pt-Cs3(A). Besides Pt-Cs3(A),
3. Other Measurements
Infrared spectra were recorded with a Fourier transform
1 wt% Pt-Cs3PW12O40(B) (abbreviated by Pt-Cs3(B)) was infrared spectrometer (Shimadzu FT IR-8500) at room tem-
obtained from Pt(NH3)4Cl2 by the same method as de- perature. The sample was dispersed on a thin silicon plate
scribed above. Especially 1 wt% Pt-Cs3PW12O40(B) which and was set in an in situ IR cell (ca 500 cm3), which was
was pretreated at 773 K in the O2 flow for 2 h will be denoted connected to a closed circulation system (ca 250 cm3). The
by Pt-Cs3(ꢀ). sample was pretreated at 573 K in H2 (350 torr) or O2 (350
The forth series of Pt-Cs2.5H0.5PW12O40 was obtained by torr) for 2 h.
impregnation of 1 wt% Pt-Cs3(B) (after the pretreatment
The surface area was measured by BET method with
in the O2 flow at 573 K for 2 h) or 1 wt% Pt-Cs3(ꢀ) with an N2 adsorption system (Micromeritics ASAP-2000) after
the aqueous solution of H3PW12O40. The ratio of Cs+ to the pretreatment in vacuum at 573 K for 0.5 h. The surface
PW12O34ꢁ0 was adjusted to 2.5. This series is denoted by areas of Cs2.5, 1 wt% Pt-Cs2.5(A) and 1 wt% Pt-Cs2.5(B),
H3(imp)/Pt-Cs3(B) or H3(imp)/Pt-Cs3(ꢀ).
1 wt% Pd-Cs2.5(A) after evacuation at 573 K were 135,
Pd-Cs2.5H0.5PW12O40 was obtained from the aqueous so- 115, 100, and 108 m2 gꢁ1, respectively. X-ray diffraction pat-
lutions of Pd(NO3)2 (Kojima Chem. Co.), H3PW12O40, and terns were recorded with an X-ray diffractometer (Mac Sci-
Cs2CO3 by the same method described above. The loading ence MXP3) using CuKꢁ radiation. The sample was mixed
levels of Pd were 0.1–1.5 wt% . These catalysts will be ab- with Si powder as an internal standard (catalyst : Si = 1 :3
breviated as Pd-Cs2.5. 1 wt% Pt-HZSM-5 was prepared as in weight).
described previously (27). H-ZSM-5, which was obtained
Adsorption of CO on Pt-Cs2.5 and Pt-Cs3 was measured
from Na-ZSM-5 (Tosoh Corporation, HSZ-820NAA, by a pulse method. After the pretreatment of the catalyst
SiO2/Al2O3 = 23.3, 332 m2 ꢂ gꢁ1) by an ion-exchange me- (about 1 g) in the flow of O2 (60 cm3 ꢂ minꢁ1) and then in the
thod, was added to an aqueous solution of NH4NO3 to flow of H2 (60 cm3 ꢂ minꢁ1) at 573 K, a pulse of CO (0.71 cm3)
form NH4-ZSM-5 and then NH4-ZSM-5 obtained was was introduced repeatedly to the catalyst at room temper-
added to the aqueous solution of Pt(NH3)4Cl2. After the ature. The concentration of CO at the outlet of the reactor

98
NA, OKUHARA, AND MISONO
was followed by a GC or mass spectrometer (Anelva AGA-
100) directly connected to the pulse system. Solid-state 31P
NMR spectra of the heteropoly compounds were taken by
the same method as described previously (28). After the
samples were pretreated at 573 K in a vacuum, they were
enclosed into Pyrex tubes (5-mm outside diameter and 10-
mm height) to avoid the rehydration by air moisture. The
sample tube was directly set in a rotor for the solid-state
MAS NMR (JEOL JNM-GX270).
RESULTS
Figure 1 shows the time courses of n-butane isomeriza-
tion over 1 wt% Pt-Cs2.5(A) and bifunctional catalysts in
the presence of 0.5 atm H2 at 573 K; 1 wt% Pt-Cs2.5(A)
was pretreated in the O2 flow at 573 K for 2 h prior to the
reaction. The conversion for 1 wt% Pt-Cs2.5(A) decreased
initially, and reached a stationary value (43% ) after about
2 h on stream. The selectivity to isobutane for 1 wt% Pt-
Cs2.5(A) was 91% after 2 h. The yield obtained for 1 wt%
Pt-Cs2.5(A) was about 40% , which corresponds to about
80% that in the equilibrium (29). The selectivities observed
for 1 wt% Pt-SO24ꢁ/ZrO2 and 1 wt% Pt-HZSM-5 were 49%
and 34% , respectively, as shown in Fig. 1, while the activities
were higher than that of 1 wt% Pt-Cs2.5(A).
The changes of the selectivity as a function of the conver-
sion are given in Fig. 2 for 1 wt% Pt-Cs2.5(A) and 1 wt%
Pt-SO24ꢁ/ZrO2. While the selectivity to isobutane for 1 wt%
Pt-Cs2.5(A) was independent of the conversion up to 43% ,
the selectivity for 1 wt% Pt-SO24ꢁ/ZrO2 decreased as the
conversion increased. The selectivityfor 1wt% Pt-Cs2.5(A)
was always higher than that for 1 wt% Pt-SO24ꢁ/ZrO2 in
these conversions.
FIG. 2. Selectivity changes as a function of conversion over 1 wt%
Pt-Cs2.5H0.5PW12O40(A) (᭺) and 1 wt% Pt-SO24ꢁ/ZrO2 (᭡) at 573 K. Feed
gas; n-butane : H2 :N2 = 0.05 : 0.50 : 0.45.
Catalytic data at stationary states at 573 K and 0.5 atm
of H2 for Pt- or Pd-promoted catalysts are summarized in
Table 1. The catalytic activity estimated from the reaction
rate is in the order: Pt-HZSM-5 > Pt-Al2O3 > Pt-SO42ꢁ
/
ZrO2 > Pt-Cs2.5(A) > Pd-Cs2.5 > Pt-Cs2.5(B). The selec-
tivity is in the order of Pt-Cs2.5(B) ꢄ Pd-Cs2.5 >
Pt-Cs2.5(A) ꢅ Pt-SO24ꢁ/ZrO2 > Pt-HZSM-5 > Pt-Al2O3.
Consequently, the order of the rate of isobutane for-
mation is: Pt-Cs2.5(A) ꢄ Pd-Cs2.5 > Pt-HZSM-5 > Pt-
SO24ꢁ/ZrO2 > Pt-Cs2.5(B) > Pt-Al2O3.
The results at 573 K under 0.05 atm of H2 (stationary
states) are shown in Table 2. Pt-Cs2.5(A), Pt-Cs2.5(B), and
Pd-Cs2.5 are also more active and selective for the forma-
tion of isobutane than Pt-SO24ꢁ/ZrO2. The conversion for
Pt-SO24ꢁ/ZrO2 decreased rapidly with the reaction time, re-
sulting in a low conversion (3.9% ) after 2 h. Pt-HZSM-5
gave the highest conversion, but the selectivity to isobu-
tane was very low (16.4% ).
Figure 3 shows the time courses of the reaction over Pt-
Cs2.5(A) at 573 K under different H2 pressures. In the
FIG. 1. Time courses of n-butane isomerization at 573 K over Pt-
bifunctional catalysts: (᭺), 1% Pt-Cs2.5H0.5PW12O40(A); (᭡), 1 wt% Pt-
SO24ꢁ/ZrO2; and (᭹), 1 wt% Pt-HZSM-5. Feed gas; n-butane : H2 :N2 =
0.05 : 0.50 : 0.45, W/F = 40 g ꢂ h ꢂ (mol of feed gas)ꢁ1 (except for Pt-ZSM-5,
20 g ꢂ h ꢂ (mol of feed gas)ꢁ1).
FIG. 3. Changes in the conversion of n-butane with time at different
hydrogen pressures: (᭺), 0.5 atm H2; (᭹), 0.2 atm H2; (4), 0.05 atm H2;
and (᭡), in the absence of H2. Feed gas; n-butane = 0.05 atm; reaction
temperature, 573 K; catalyst, 1 g, W/F = 40 g ꢂ h ꢂ molꢁ1
.

BUTANE ISOMERIZATION OVER HETEROPOLY COMPOUNDS
99
TABLE 1
Activity and Selectivity for Skeletal Isomerization of n-Butane over Metal-Promoted Solid Acids
in the Presence of 0.5 atm of H2 at 573 K
Reaction rate
Selectivity (mol% )c
C3
Conversion
(% )
Catalyst
Ratea
Rateb
C1
C2
i-C4
C5(+)
1 wt% Pt-Cs2.5(A)d
1 wt% Pt-Cs2.5(B)e
1 wt% Pd-Cs2.5(A) f
1 wt% Pt-SO24ꢁ/ZrO2
1 wt% Pt-HZSM-5g
1 wt% Pt-Al2O3
43.0
24.5
40.0
64.6
51.7
75.8
14.5 (12.6)
8.3 (8.3)
13.6 (12.6)
22.0 (17.6)
35.2 (10.6)
25.8 (16.8)
13.3
7.9
12.7
10.6
11.9
2.1
2.0
0.9
0.6
5.8
19.8
21.9
1.8
1.5
0.6
11.3
24.2
51.0
3.8
1.6
2.8
29.4
21.2
18.8
90.8
95.3
94.1
48.6
34.0
8.3
1.6
0.7
1.9
4.9
0.8
0
Note. Reaction conditions: n-butane : H2 :N2 = 0.05 : 0.50 : 0.45; catalyst 1 g; total flow rate 10 ml minꢁ1; W/F 40 g h molꢁ1
.
a Total reaction rate/10ꢁ8 mol gꢁ1 sꢁ1 (after 5 h). The figures in the parentheses are the rate per unit surface area/10ꢁ10 mol mꢁ2 sꢁ1
b Rate of isobutane formation/10ꢁ8 mol gꢁ1 sꢁ1 (after 5 h).
.
c C1 is CH4, C2 is C2H6, C3 is C3H8, i-C4 is isobutane, and C5(+) is hydrocarbons containing more than five carbons.
d From H2PtCl6. The catalyst was pretreated at 573 K in O2 for 2 h.
e From Pt(NH3)4Cl2. The catalyst was pretreated at 573 K in O2 for 2 h.
f From Pd(NO3)2. The catalyst was pretreated at 573 K in O2 for 2 h.
g Catalyst weight 0.5 g, total flow rate 10 ml minꢁ1, W/F 20 g h molꢁ1
.
absence of H2, a significant catalyst deactivation was ob- which is about half that in the equilibrium (30). The selec-
served. In the presence of H2 (0.05–0.5 atm), the catalyst tivity to isobutane was low, 55% (conversion, 5.7% ) and
deactivation became smaller. In the case of Pd-Cs2.5, the at the same time, an appreciable amount of C3 + C5 hydro-
degree of the deactivation decreased gradually as the H2 carbons was formed (C3 + C5 = 21% ). Contrary to this, in
pressure increased, and it became nearly zero at 0.5 atm the presence of 0.05–0.5 atm H2, the selectivity to isobutane
of H2.
reached about 90% and the concentrations of butenes were
In Fig. 4, the selectivities of isobutane or C3 + C5 hy- 50–5 ppm (0.3–0.03 mol% in total products).
drocarbons (Fig. 4A) and concentration of butenes formed
during the reaction (Fig. 4B) over 1 wt% Pt-Cs2.5(A) are
In Fig. 5, the time courses at 473 K and 0.05 atm of H2 are
given for Pt-Cs2.5(A) and Pt-SO24ꢁ/ZrO2. The conversion
plotted against the H2 pressure, where the data were col- for Pt-Cs2.5(A) was nearly constant from the initial stage
lected at stationary states (after 5 h) at 573 K (see Fig. 3). In of the reaction, while the conversion for Pt-SO24ꢁ/ZrO2 de-
the absence of H2, butenes appeared at the concentration creased with time. After 5 h, the conversions and selectivi-
of 540 ppm (corresponds to 19 mol% in total products), ties for both catalysts became similar.
TABLE 2
Activity and Selectivity for Skeletal Isomerization of n-Butane over Metal-Promoted Solid Acids
in the Presence of 0.05 atm of H2 at 573 K
Reaction rate
Selectivity (mol% )c
Conversion
(% )
Catalyst
Ratea
Rateb
C1
C2
C3
i-C4
C4=
C5(+)
1 wt% Pt-Cs2.5(A)d
1 wt% Pt-Cs2.5(B) f
1 wt% Pd-Cs2.5g
35.7
18.8
10.9
3.9
12.7 (11.0)
6.4 (6.4)
3.7 (3.4)
1.3 (1.0)
49.2 (14.8)
11.3
5.4
2.6
0.9
8.0
1.0
1.3
0.3
0
0.8
0.8
0.5
2.1
6.5
5.1
7.2
14.2
15.7
66.6
89.2
85.4
72.0
74.4
16.4
<0.2e
0.7
3.7
4.6
9.8
7.6
6.4
3.2
1 wt% Pt-SO24ꢁ/ZrO2
1 wt% Pt-HZSM-5h
<0.2 f
<0.2 f
72.2
3.9
Note. Reaction conditions: n-butane : H2 :N2 = 0.05 : 0.05 : 0.90; catalyst 1 g; total flow rate 10 ml minꢁ1; W/F 40 g h molꢁ1
.
a Total reaction rate/10ꢁ8 mol gꢁ1 sꢁ1 (after 5 h). The figures in parentheses are the rate per unit surface area/10ꢁ10 mol mꢁ2 sꢁ1
b Rate of isobutane formation/10ꢁ8 mol gꢁ1 sꢁ1 (after 5 h).
.
c C1 is CH4, C2 is C2H6, C3 is C3H8, i-C4 is isobutane, C4 is butenes, and C5(+) is hydrocarbons containing more than five carbons.
d From H2PtCl6. The catalyst was pretreated at 573 K in O2 for 2 h.
e Less than 2 ppm.
f From Pt(NH3)4Cl2. The catalyst was pretreated at 573 K in O2 for 2 h.
g From Pd(NO3)2. The catalyst was pretreated at 573 K in O2 for 2 h.
h Catalyst 0.5 g; total flow rate 10 ml minꢁ1; W/F 20 g h molꢁ1
.

100
NA, OKUHARA, AND MISONO
FIG. 6. Effects of reaction temperature on n-butane isomerization:
(a) 1 wt% Pt-SO42ꢁ/ZrO2; (b) 1 wt% Pt-Cs2.5H0.5PW12O40(A), and (c)
1 wt% Pt-Cs3PW12O40(A). (᭺), conversion to isobutane; (᭹), conver-
sion to hydrogenolysis (C1ꢃC3); (᭡), selectivity to isobutane. Feed gas;
FIG. 4. Selectivities and concentration of butenes as a function of
n-butane : H2 :N2 = 0.05 : 0.50 : 0.45, W/F = 40 g ꢂ h ꢂ molꢁ1
.
H2 pressure over 1 wt% Pt-Cs2.5H0.5PW12O40 at 573 K. (A) Selectivities
to isobutane and C3 + C5 hydrocarbons. (B) Concentration of butenes
formed at the stationary states.
over Pt-SO24ꢁ/ZrO2, Pt-Cs2.5(A), and Pt-Cs3(A). In the
case of Pt-SO24ꢁ/ZrO2 (Fig. 6a), the conversion to isobu-
tane increased once, and then it decreased as the reaction
temperature increased. On the other hand, the conver-
Detailed data at the stationary states at 473 K for 1 wt%
Pt-Cs2.5(A) and 1 wt% Pt-SO24ꢁ/ZrO2 are summarized in
Table 3. The conversions at 0.05 atm of H2 for both cata-
lysts were higher than those at 0.5 atm of H2 in contrast to
those at 573 K (Tables 1 and 2). The selectivity for 1 wt%
Pt-SO24ꢁ/ZrO2 was more than 90% at both H2 pressures. On
the other hand, the selectivity to isobutane for Pt-Cs2.5(A)
was only 32% under 0.5 atm of H2, the main products be-
–
sion to C1 C3 hydrocarbons continued to increase. The
selectivities to isobutane were 96 and 49% at 473 and 573 K,
respectively. Similar trends of the conversion and selectiv-
ity for Pt-SO24ꢁ/ZrO2 have been reported by Chao et al.
(31). On the other hand, it is remarkable for Pt-Cs2.5(A)
–
ing C1 C3 hydrocarbons, whereas it was about 90% under
–
that the conversion to C1 C3 hydrocarbons was nearly inde-
0.05 atm of H2.
Figure 6 provides the dependencies in the stationary
conversion and selectivity upon the reaction temperature
pendent of the reaction temperature, while the conversion
to isobutane continued to increase with increasing the re-
action temperature (Fig. 6b). Consequently, the selectivity
for Pt-Cs2.5(A) was improved by the rise in the temper-
ature. In the case of Pt-Cs3(A) (Fig. 6c), the hydrogenol-
ysis occurred exclusively and the conversion for the hy-
drogenolysis increased monotonically with the reaction
temperature.
H2 pressure dependencies of the reaction rates for Pt-
Cs2.5(A) at 473 K are given in Fig. 7. It was found that the
initial conversion decreased as the H2 pressure increased,
while the conversion at the stationary state increased once
and decreased slightly. The selectivities to isobutane (at the
stationarystate) became slightlyhigher (85% to 90% ) when
the H2 pressure was increased from 0 to 0.05 atm, and then
decreased to 32% by the further increase to 0.5 atm. On the
other hand, as shown in Fig. 8, at 573 K, although the initial
conversion also tended to decrease with the H2 pressure, the
stationary conversion was greatly enhanced even by the in-
troduction of 0.05, atm of H2. The selectivities to isobutane
were 55, 89, 90, and 91% at 0, 0.05, 0.2, and 0.5 atm of H2,
respectively.
FIG. 5. Changes in the conversion and selectivity of n-butane iso-
merization at 473 K: (᭺), 1% Pt-Cs2.5H0.5PW12O40(A); (᭡), 1 wt%
Pt-SO24ꢁ/ZrO2. Feed gas; n-butane : H2 :N2 = 0.05 : 0.05 : 0.90, W/F =
Figure 9 shows the dependencies of the rates on H2
pressure for 1 wt% Pd-Cs2.5 at 573 K. The dependencies
are essentially the same as those for Pt-Cs2.5 at 573 K,
40 g ꢂ h ꢂ molꢁ1
.

BUTANE ISOMERIZATION OVER HETEROPOLY COMPOUNDS
101
TABLE 3
Activity and Selectivity for Skeletal Isomerization of n-Butane at 473 K
Reaction rate
Selectivity (mol% )c
H2
(atm)
Conversion
(% )
Catalyst
Ratea
Rateb
C1
C2
C3
i-C4
C5(+)
1 wt% Pt-Cs2.5(A)d
0.05
0.5
17.1
5.4
5.2 (4.5)
0.6 (0.5)
0.6
1.3
1.8
27.1
1.2
12.2
4.5
28.1
89.9
31.9
2.6
0.7
1 wt% Pt-SO24ꢁ/ZrO2
0.05
0.5
21.1
10.7
6.5 (5.2)
3.5 (2.8)
0.7
0.01
0.1
0.6
0.2
2.4
5.0
0.9
90.5
96.1
4.2
0
Note. Reaction conditions: n-butane 0.05 atm; catalyst 1 g; total flow rate 10 ml minꢁ1; W/F 40 g h molꢁ1
a Rate of isobutane formation/10ꢁ8 mol gꢁ1 sꢁ1 (after 5 h). The figures in parentheses are the rate per unit surface area/10ꢁ10 mol mꢁ2 sꢁ1
.
.
b Rate for the hydrogenolysis/10ꢁ8 mol gꢁ1 sꢁ1
.
c C1 is CH4, C2 is C2H6, C3 is C3H8, i-C4 is isobutane, and C5(+) is hydrocarbons containing more than five carbons.
d From H2PtCl6. The catalyst was pretreated at 573 K in O2 for 2 h.
except that the initial conversion decreased more in the about significantly higher conversion and selectivity than
presence of H2. The selectivities of isobutane (after 5 h) the H2-pretreatment. Similar to Pt-Cs2.5(B), the activity
were 13, 78, 87, and 96% for 0, 0.05, 0.2, and 0.5 atm of H2, of Pt-Cs3(B) was enhanced by the O2-treatment, but, re-
respectively.
gardless of the pretreatment conditions, Pt-Cs3(A) and
The changes in the activity and selectivity at the station- Pt-Cs3(B) were not selective to isobutane.
ary state with Pt loading level for Pt-Cs2.5(B) are given in
Changes in the catalysts with the pretreatment conditions
Fig. 10. The reaction was accelerated by a factor of about were examined by using IR spectroscopy. The intensities
4 only by the addition of 0.1 wt% Pt. The rate and selec- and wavenumbers of the IR bands due to the Keggin struc-
tivity were found to be nearly independent of the loading ture were independent of the ambient gas for the pretreat-
levels above 0.1 wt% . The activities of 1 wt% Pt-Cs2.5(A) ment. What is remarkable for Pt-Cs2.5(B) is the presence
and 1 wt% Pt-Cs2.5(B) were about 6 and 4 times as high of a band at 1415 cmꢁ1 due to NH4+ after the N2 and H2 pre-
as Cs2.5 (10). In Fig. 11, the effects of Pd loading level are treatments, in addition to the bands due to the Keggin an-
shown. Also in this case, the reaction rate and selectivity ion. This band was much smaller after the O2-pretreatment.
were almost unchanged upon the variation of the Pd load- In Table 5, the amounts of NH+4 which remained after
ing level.
the pretreatments are summarized. The amounts of NH4+
In Table 4, the influence of the ambient gas used for the in the table were estimated from the relative peak in-
pretreatment of Pt-Cs2.5 and Pt-Cs3 is summarized. Both tensities of NH+4 and ꢂ(W O) (985 cmꢁ1) by compar-
==
the activity and selectivity for Pt-Cs2.5(A) were little af- ing with those of the stoichiometric salt, (NH4)3PW12O40,
fected by the ambient gas. In contrast, significant changes and were expressed by the ratios of NH+4 /polyanion. The
were observed for Pt-Cs2.5(B) which was prepared from ratio NH+4 /polyanion for Pt-Cs2.5(B) was 0.20 after the
Pt(NH3)4Cl2. The O2-pretreatment of Pt-Cs2.5(B) brought
FIG. 7. Effects of H2 pressure on initial and stationary activities for
isomerization of n-butane at 473 K over 1 wt% Pt-Cs2.5H0.5PW12O40(A).
(᭺), initial conversion (after 5 min); (᭹), stationary conversion (after 5 h).
FIG. 8. Effects of H2 pressure on initial and stationary activities for
isomerization of n-butane at 573 K over 1 wt% Pt-Cs2.5H0.5PW12O40(A).
(᭺), initial conversion (after 5 min); (᭹), stationary conversion (after 5 h).
Feed gas; n-butane: 0.05 atm, W/F = 40 g ꢂ h ꢂ molꢁ1
.
Feed gas; n-butane: 0.05 atm, W/F = 40 g ꢂ h ꢂ molꢁ1
.

102
NA, OKUHARA, AND MISONO
FIG. 9. Effects of H2 pressure on initial and stationary activities for
isomerization of n-butane at 573 K over 1.5 wt% Pd-Cs2.5H0.5PW12O40
FIG. 11. Dependencies of the loading amount of Pd of Pd-Cs2.5H0.5
PW12O40 on the activity and selectivity in n-butane isomerization: (᭹),
stationary conversion of n-butane; (᭺), selectivity to isobutane. Feed gas;
n-butane : H2 :N2 = 0.05 : 0.50 : 0.45; reaction temperature, 573 K; W/F =
.
(᭺), initial conversion (after 5 min); (᭹), stationary conversion (after 5 h).
Feed gas; n-butane: 0.05 atm, W/F = 40 g ꢂ h ꢂ molꢁ1
.
40 g ꢂ h ꢂ molꢁ1
.
O2-treatment, while these ratios were 0.47 and 0.46 after
the N2 and H2 pretreatments, respectively. No peak due to
NH+4 was detected for 1 wt% Pt-Cs2.5(A) and Pt-Cs3(A),
which were prepared from H2PtCl6.
Effects of the addition of H3PW12O40 to 1 wt% Pt-Cs3(B)
and 1 wt% Pt-Cs3(ꢀ) on the activities and selectivities are
given in Table 6; 1 wt% Pt-Cs3(B) and 1 wt% Pt-Cs3(ꢀ)
were those after the O2 pretreatment at 573 and 773 K,
respectively. The Pt particle sizes of Pt-Cs3(B) and Pt-
Cs3(ꢀ) estimated from the CO adsorption were 27 and
effect of the addition of H3PW12O40 to Pt-Cs3(ꢀ) was es-
sentially the same as Pt-Cs3(B), except that the rate of
hydrogenolysis on Pt-Cs3(ꢀ) changed little, resulting in a
lower selectivity than H3(imp)/Pt-Cs3(B).
Figure 12 shows the solid-state 31P NMR spectra of these
heteropoly compounds. Pt-Cs3(A) gave a single peak at
ꢁ15.0 ppm (Fig. 12d). H3(imp)/Pt-Cs3(A) and Pt-Cs2.5(A)
showed similar patterns, in which two peaks were detected
mainly at about ꢁ15.0 ppm (ꢁ15.1 and ꢁ15.0 ppm for
Figs. 12b and 12c, respectively) and ꢁ13.4 ppm. Besides
these two peaks, a small peak at ꢁ12.2 ppm was also ob-
served for H3(imp)/Pt-Cs3(A). In the case of Cs2.5, three
main peaks were detected at ꢁ14.8, ꢁ13.3, and ꢁ12.0 ppm.
H3(imp)/Cs3, which was prepared by impregnation by us-
ing Cs3 instead of Pt-Cs3(A) and pretreated at 573 K in
˚
180 A, respectively. As shown in Table 6, both Pt-Cs3(B)
–
and Pt-Cs3(ꢀ) produced mainly C1 C3 hydrocarbons. Pt-
Cs3(B) was more active for the hydrogenolysis than Pt-
Cs3(ꢀ). When H3PW12O40 was added to Pt-Cs3(B) by the
impregnation method, the isomerization was greatly ac-
celerated, while the hydrogenolysis was nearly completely
suppressed. Consequently, the selectivity to isobutane was
significantly enhanced by the addition of H3PW12O40. The
TABLE 4
Effects of Ambient Gas at Pretreatment on Activity and Selecti-
vity for Skeletal Isomerization of n-Butane at 573 K
Conver-
sion
(% )
Selectivity (mol% )a
Pretreatment
atmosphere
Catalyst
C1
C2
C3 i-C4 C5
1% Pt-Cs2.5(A)b
O2
H2
43.0
38.0
2.0 1.8 3.8 90.8 1.6
1.7 2.2 3.5 91.0 1.6
1% Pt-Cs2.5(B)c
1% Pt-Cs3(A)b
1% Pt-Cs3(B)c
O2
H2
24.5
13.2
0.9 1.5 1.6 95.3 0.7
32.8 34.9 30.5 1.8
0
O2
H2
44.0
35.9
30.7 46.7 20.6 2.0
30.4 46.6 21.9 1.1
0
0
O2
H2
31.0
4.0
39.1 20.3 37.8 2.6 0.2
33.7 31.9 30.8 3.6
0
FIG. 10. Dependencies of the loading amount of Pt of Pt-Cs2.5H0.5
PW12O40(B) on the activity and selectivity (5 h) in n-butane isomerization:
(᭹), conversion of n-butane; (᭺), selectivity to isobutane. Catalyst: Pt-
Cs2.5H0.5PW12O40(B) (from Pt(NH3)4Cl2). Feed gas; n-butane : H2 :N2 =
a C1 is CH4, C2 is C2H6, C3 is C3H8, i-C4 is isobutane, and C5(+) is
hydrocarbons containing more than five carbons.
b Pt-Cs2.5(A) and Pt-Cs3(A): from H2PtCl6.
c Pt-Cs2.5(B) and Pt-Cs3(B): from Pt(NH3)4Cl2. Feed gas; n-butane :
0.05 : 0.50 : 0.45; reaction temperature, 573 K; W/F = 40 g ꢂ h ꢂ molꢁ1
.
H2 :N2 = 0.05 : 0.50 : 0.45; catalyst 1 g; W/F 40 g h molꢁ1
.

BUTANE ISOMERIZATION OVER HETEROPOLY COMPOUNDS
TABLE 5
103
Amount of NH+4 Remained on 1 wt% Pt-Cs2.5(B)
after the Pretreatments at 573 K
Ambient gas
at treatment
NH+4 /polyaniona
H+/polyanionb
N2
H2
O2
0.47
0.46
0.20
0.03
0.04
0.30
a NH+4 /PW12O43ꢁ0 determined by infrared spectroscopy.
b
The amount of protons after the pretreatment in the
unit of H+/PW12O43ꢁ0 (see text).
vacuum, gave a similar spectrum to that of Cs2.5, as re-
ported previously (32).
DISCUSSION
1. Reaction Paths of Skeletal Isomerization of n-Butane
FIG. 12. Solid-state 31P NMR spectra of heteropoly compounds: (a)
It is generally accepted that skeletal isomerization of Cs2.5H0.5PW12O40; (b) 1 wt% Pt-Cs2.5H0.5PW12O40(A); (c) H3(imp)/1 wt%
Pt-Cs3PW12O40(A); and (d) 1 wt% Pt-Cs3PW12O40(A). The samples were
pretreated at 573 K in vacuum.
n-butane over bifunctional catalysts proceeds according
to the following steps as shown by Eqs. [1]–[3] (1, 11): (i)
dehydrogenation of n-butane to butenes on metal sites;
(ii) isomerization of butenes to isobutylene at acid sites via probably due to the catalyst deactivation by coke. Actually
sec- and tert-butyl carbenium ions; and (iii) hydrogenation the deactivation was significant in the absence of H2 and was
of isobutylene to isobutane on metal sites. As another appreciable at low H2 pressure range. By this, the stationary
possible step for the formation of isobutane from tert-butyl conversions at low H2 pressures became low. In the high H2
carbenium ion, hydride transfer from H2 or n-butane has pressure range, the coke formation was possibly suppressed
been proposed (19).
by the rapid hydrogenation of the coke and/or its precursor
and the low concentrations of butenes.
As another possible role of H2, the generation of acidic
protons has been proposed (18, 21, 33). Hattori and
co-workers claimed for Pt-SO24ꢁ/ZrO2 that protons are gen-
erated from hydrogen atom formed on Pt, releasing elec-
tron to Lewis acid site (18). Fujimoto and co-workers pre-
sumed that protons are produced on the surface of ZSM-5
via the spillover of hydrogen atom from the surface of Pt in
a hybrid catalyst consisting of Pt/SiO2 and H-ZSM-5 (21).
Indeed, the effect of H2 was significant as shown in Fig. 3 in
the case of Pt-Cs2.5(A). However, Pt-Cs3(A), which has no
[1]
[2]
[3]
“Bimolecular mechanism” is also possible. Although the inherent acid sites, was still inactive for the isomerization
elucidation of the mechanism is not the primary purpose of even in the presence of H2 (Table 4). This shows that pro-
the present study, the mechanism will be discussed briefly tons were not formed on Pt-Cs3(A). Probably, new protons
to make clear the characteristics of the present catalytic were not generated also in the case of Pt-Cs2.5.
system.
If the dehydrogenation-hydrogenation steps are in equi-
Promoting effects of Pt or Pd (Figs. 10 and 11), as well as librium, the slow step would be in Eq. [2]. Since the protona-
the negative dependency of the initial rate on H2 pressure tion of butenes and the deprotonation are postulated to be
(Figs. 7–9) support the important roles of metals in the step rapid, the rate-determining step is probably the transforma-
of Eq. [1]. Little dependencies of the loading amount of Pt tion from sec-butyl carbenium ion to tert-butyl carbenium
and Pd (Figs. 10 and 11) indicate that the step of Eq. [1] is ion. When HF-SbF5 was used as a catalyst, the isomeriza-
in equilibrium as in the case of Pt/chlorinated-Al2O3 (12). tion of n-butane to isobutane was very slow, while n-butane-
In contrast to the initial rate, the stationary rate depended 1-13C was rapidly isomerized to n-butane-2-13C and skeletal
positively on the H2 pressure (Figs. 8 and 9). This is most isomerization of n-pentane was also fast (11). These results

104
NA, OKUHARA, AND MISONO
TABLE 6
Effects of H3PW12O40 Addition to 1 wt% Pt-Cs3PW12O40(B)
Pretreatment
Selectivity (mol% )d
temperaturea
(K)
Conversion
(% )
Catalyst
Pt-Cs3(B)
Db
(d)c
C1
C2
C3
i-C4
C5
573
773
0.33
0.05
(27)
(180)
31.0
7.1
39.1
32.9
20.3
39.8
37.8
17.4
2.6
9.9
0.2
0
Pt-Cs3(ꢀ)
H3(imp)/Pt-Cs3(B)e
573
573
0.20
0.05
(45)
(180)
39.0
32.3
1.1
0.9
0.7
1.6
2.6
12.2
94.0
80.1
1.6
5.2
H3(imp)/Pt-Cs3(ꢀ)e
a Pretreatment temperature.
b Dispersion of Pt by CO adsorption.
c
˚
The figures in the parentheses are particle sizes estimated from the dispersion in A.
d C1 is CH4, C2 is C2H6, C3 is C3H8, i-C4 is isobutane, and C5(+) is hydrocarbons containing more than five carbons.
e H3PW12O40 was impregnated with Pt-Cs3(B); Final Pt loading: 0.85 wt% . Feed gas: n-butane : H2 :N2 = 0.05 : 0.50 : 0.45;
catalyst 1 g; W/F 40 g h molꢁ1; reaction temperature 573 K.
were explained by the monomolecular mechanism through ated products such as C3 and C5 hydrocarbons, as well
protonated cyclopropane intermediates (11), in which only as isobutane. High selectivities obtained over 1 wt% Pt-
the n-butane isomerization to isobutane is required to pro- Cs2.5(B) and 1 wt% Pd-Cs2.5(A) (Table 1) suggest that the
ceed through the primary carbenium ion.
bimolecular mechanism may be less important over these
It should be emphasized that the selectivity to isobutane catalysts under these reaction conditions. Skeletal isomer-
reached 91–95% on Pt-Cs2.5(A), (B) and Pd-Cs2.5at 573K. ization of butenes to isobutylene proceeded at 573 K on
As can be seen in Table 1, C1, C2, and C3 hydrocarbons are H-ZSM-5 (38). Since the acid strength of Cs2.5 was higher
formed as byproducts for these catalysts, showing that the than that ofH-ZSM-5(26), it isprobable that the step shown
hydrogenolysis of n-butane slightly took place as the side by Eq. [2] is facile on Pt-Cs2.5 at 573 K, if the reaction took
reaction. Contrary to this, the byproducts on Cs2.5 (with- place via monomolecular mechanism. In order to determine
out Pt) were mainly C3 and C5 as reported previously (10). the contribution of each mechanism, the experiment using
According to “bimolecular mechanism” (3, 34–37), it is pos- isotopic labeled molecules is necessary.
sible that C8 and C12 hydrocarbons are formed as interme-
diate products from sec-butyl carbenium ion and butenes
2. State of Pt Particle and Selectivity
which were produced by deprotonation of sec-butyl carbe-
Whereas Pt-Cs3(A) was active for the hydrogenolysis
nium ion and are cracked into smaller hydrocarbons such of n-butane, Pt-Cs2.5(A) was very effective for the skele-
as C3, isobutane, and C5. Whether the reaction proceeds via tal isomerization, being less active for the hydrogenolysis
monomolecular or bimolecular mechanism is likely depen- (Table 4). The rate of hydrogenolysis on Pt-Cs2.5(A) is
dent on the reaction conditions. As a matter of fact, whereas about one-tenth that on Pt-Cs3(A). In order to elucidate
the fractions of C3 + C5 hydrocarbons in total products were the reason for the different catalysis, the states of Pt will be
small (less than 2% ) over 1 wt% Pt-Cs2.5 and 1 wt% Pd- discussed first. It was reported that the particle size of Pt
Cs2.5 at 573 K and 0.5 atm H2 (Table 1), the amounts of influenced greatly the activity for hydrogenolysis (39–42).
C3 + C5 hydrocarbons were not slight at 0.05 atm H2 at both Indeed, highly dispersed Pt-Cs3(B) was more active for the
573 and 473 K (Tables 2 and 3). These results suggest that hydrogenolysis than Pt-Cs3(ꢀ) (Table 6). The particle size
˚
at 573 K and 0.5 atm of H2, monomolecular mechanism is determined by the CO adsorption was about 30 A both for
preferable over these catalysts, but the bimolecular mech- 1 wt% Pt-Cs2.5(A) and 1 wt% Pt-Cs3(A). In addition, nei-
anism cannot be ignored at lower H2 pressures. The C3 and ther Pt-Cs2.5(A) nor Pt-Cs3(A) gave the XRD peak due to
C5 hydrocarbons are considered to be formed by cracking Pt crystallites. Therefore, the different catalytic activity for
of C8 species produced via bimolecular mechanism. As is hydrogenolysis between Pt-Cs2.5(A) and Pt-Cs3(A) is not
demonstrated in Fig. 4, the fraction ofC3 + C5 hydrocarbons due to the difference in the particle size.
correlated with the concentration of butenes formed during
The oxidation state of Pt may be influential in the activ-
the reaction. This is reasonable because the contribution of ity and selectivity. But as shown in Table 4, the selectivity
the bimolecular mechanism increases as the concentration for Pt-Cs2.5(A) or Pt-Cs3(A) was independent of the am-
of butenes increases.
bient gas for the pretreatment (H2 or O2). Thus the differ-
If the bimolecular mechanism operates, various kinds of ence in the selectivity between Pt-Cs2.5(A) and Pt-Cs3(A)
C8 species are formed from sec-butyl carbenium ion and may not be due to the difference in the oxidation state
butenes. These C8 species are likely to give disproportion- of Pt.

BUTANE ISOMERIZATION OVER HETEROPOLY COMPOUNDS
105
3. Roles of Protons in Selective Isomerization
to ꢃ-bonded butenes (I) or ꢄ-bonded C4 species (I0) on the
Pt surface (11, 42, 44). In the absence of proton, the hy-
drogenolysis proceeds on Pt through the intermediates (I)
or (I0). According to Anderson et al. (44), n-butane is par-
tially dehydrogenated to form multiply ꢄ-bonded species
such as 1,3-diadsorbed species (Scheme 1) on metal sur-
faces, which is considered to be a common intermediate for
the hydrogenolysis and the skeletal isomerization on it. In
the presence of protons, intermediate (I) and/or (I0) would
interact with protons and transform to sec-butyl carbenium
ion (II). If so, the hydrogenolysis on Pt may be suppressed,
because sec-butyl carbenium ion is unlikely the intermedi-
ate for the hydrogenolysis. By this unique function of pro-
tons, the selective isomerization may be attained over these
heteropoly bifunctional catalysts.
It should be noted that the O2-treatment regenerated
the protons on the surface of Pt-Cs2.5(B) (Table 5) and
enhanced greatly the selectivity to isobutane (Table 5).
Contrary to this, the H2-pretreatment did not regener-
ate protons by NH+4 decomposition and did not improve
the selectivity to isobutane (Tables 4 and 5). Therefore,
it can be concluded that the remarkable effect of the O2-
pretreatment on the selectivity is attributed to the regen-
eration of protons. The NH+4 observed on Pt-Cs2.5(B) is
likely to form from protons of H3PW12O40 and NH3 in
Pt(NH3)4Cl2 and the regeneration of protons by the O2-
pretreatment can be described as in Eq. [4] (43). Con-
trary to Pt-Cs2.5(B), the O2-pretreatment of Pt-Cs2.5(A)
changed the selectivity little, which is reasonable because
Pt-Cs2.5(A) is NH3-free:
It should be emphasized that the activity of Pt-Cs2.5(A)
for hydrogenolysis was higher than that of Pt-Cs3(A) at
473 K and 0.5 atm of H2 (Table 3 and Fig. 6), indicating
that Pt on Pt-Cs2.5(A) has a high inherent activity for the
hydrogenolysis. The fact that the rate of the hydrogenolysis
over Pt-Cs3(A) increased by a factor of about 20 with
the rise in the temperature from 473 to 573 K (Fig. 6)
resembles the results of Pt catalysts (42, 44). Thus the little
temperature dependencies of the rate of hydrogenolysis
over Pt-Cs2.5(A) (Fig. 6a) is unique and can be explained
by the mechanism in Scheme 1; as the reaction temper-
ature increased, the hydrogenolysis is more effectively
suppressed due to the rapid migration of protons. Baba
NH+4 + 3/4O2 → H+ + 1/2N2 + 3/2H2O.
[4]
As shown in Table 6, the addition of H3PW12O40 to
Pt-Cs3(A) greatly accelerated the skeletal isomerization,
but suppressed the hydrogenolysis. The remarkable change
in the selectivity may be due to the endowment of acid
sites.
We propose a mechanism for the selective forma-
tion of isobutane over Pt-Cs2.5 and H3(imp)/Pt-Cs3 in
Scheme 1. In this mechanism, n-butane is dehydrogenated
1
et al. (33, 45) claimed from H solid-state NMR that the
mobility of protons was enhanced by the presence of Ag.
Although the mobility of protons was not measured in the
present case, Pt may accelerate the migration of proton,
similarly to Ag (33, 45).
Information on protons on Pt-Cs2.5 and H3(imp)/Pt-Cs3
was given by solid-state 31P NMR. As already reported by
us (32), the chemical shift of 31P of PW12O430ꢁ is determined
by the number of the protons which are directly interact-
ing with the bridging oxygen atoms of the Keggin anion;
the peaks observed for Cs2.5 at ꢁ14.9, ꢁ13.5, ꢁ12.1, and
ꢁ10.9 ppm can be assigned to the Keggin anion having 0, 1,
2, and 3 protons, respectively. The spectrum of Fig. 12a is es-
sentially consistent with the previous one (32). Vedrine et al.
(46) observed a similar 31P spectrum of Cs2.5 and claimed
that Cs2.5 consists of Cs3 and H3PW12O40. However, since
the peak at ꢁ10.9 ppm due to H3PW12O40 (anhydrous) was
not detected for Cs2.5 (32, 46), Cs2.5 is not a mixture of Cs3
and H3PW12O40.
The single peak at ꢁ14.9 ppm observed for Cs3 (Fig. 12d)
wasin agreement with that previouslyreported for Cs3(32).
The appearance of the peak at ꢁ13.5 ppm for Pt-Cs2.5(A)
and H3(imp)/Pt-Cs3(A) is evidence for the presence of pro-
tons. The lower intensities of the peak at ꢁ13.5 ppm than
that of Cs2.5 suggest that the amounts of protons on Pt-
Cs2.5(A) and H3(imp)/Pt-Cs3(A) were smaller than that
SCHEME 1. A proposed reaction scheme.

106
NA, OKUHARA, AND MISONO
of Cs2.5. The decrease in protons by the introduction Pt
can be described in Eq. [5]; protons react with chloride ion,
which contains the starting material, H2PtCl6 to form HCl
during the preparation processes:
3. Adeeva, V., de Haan, J. W., Janchen, J., Schunemann, V., van de Ven,
L. T. M., Sachtler, W. M. H., and van Santen, R. A., J. Catal. 151, 364
(1995); Adeeva, V., Lei, G. D., and Sachtler, W. M. H., Catal. Lett. 33,
135 (1995).
4. Yaluris, G., Larson, R. B., Kobe, J. M., Gonzalez, M. R., Fogash,
K. B., and Dumesic, J. A., J. Catal. 158, 336 (1996).
5. Stocker, M., Hemmerbach, P., Rader, J. H., and Grepstad, J. K., Appl.
Catal. 25, 223 (1986).
nH+ + PtCl26ꢁ → [PtCl6ꢁn
]
(2ꢁn)ꢁ + nHCl.
[5]
4. Effect of Pt Particle Size on the Role
of Protons and Selectivity
6. Yori, J. C., Luy, J. C., and Parera, J. M., Appl. Catal. 46, 103 (1989).
7. Shigeishi, R., Garforth, A., Harris, I., and Dwyer, J., J. Catal. 130, 423
(1991).
The fact that the rate of hydrogenolysis increased as
the Pt particle size of Pt-Cs3 decreased is consistent with
the general trend of metallic catalysts (42). The selectivity
to isobutane also depended on the particle size of Pt of
H3(imp)/Pt-Cs3 (Table 6). By the addition of H3PW12O40 to
Pt-Cs3, the hydrogenolysis was more effectively suppressed
for smaller Pt particle (H3(imp)/Pt-Cs3), resulting in the
higher selectivity for the smaller Pt particle. This is probably
due to that the smaller particles (the higher dispersion) are
favorable to intimate interaction between Pt and protons.
As reported elsewhere (47), when Pt/SiO2 was physically
mixed with Cs2.5, the reaction rate increased greatly, but
the selectivity decreased. This suggests that close proximity
between the metal site and the acid site as in the case
of Pt-Cs2.5 is required for the selective isomerization of
n-butane.
8. Mechlor, A., Garbowski, E., Mathieu, M. V., and Primet, M., J. Chem.
Soc. Faraday Trans. I 82, 1893 (1986).
9. Na, K., Okuhara, T., and Misono, M., Chem. Lett., p. 1141 (1993).
10. Na, K., Okuhara, T., and Misono, M., J. Chem. Soc. Faraday Trans. 91,
367 (1995).
11. Gates, B. C., Katzer, J. R., and Schuit, G. C. A., in “Chemistry of
Catalytic Processes.” McGraw-Hill, New York, 1979.
12. Melchor, A., Garbowski, E., Mathieu, M. V., and Primet, M., J. Chem.
Soc. Faraday Trans. I 82, 1893 (1986).
13. Bernard, P., and Primet, M., J. Chem. Soc. Faraday Trans. 86, 3667
(1990).
14. Garin, F., Andriamasinoro, D., Abdulsamad, A., and Sommer, J., J.
Catal. 131, 199 (1991).
15. Hino, M., and Arata, K., Catal. Lett. 30, 25 (1995).
16. Coelho, M. A., Resasco, D. E., Sikabwe, E. C., and White, R. L., Catal.
Lett. 32, 253 (1995).
17. Jatia, A., Chang, C., Macleod, J. D., Okubo, T., and Davis, M. E., Catal.
Lett. 25, 21 (1994).
18. Ebitani, K., Konishi, J., and Hattori, H., J. Catal. 131, 257 (1991).
19. Iglesia, E., Soled, S. L., and Kramer, G. M., J. Catal. 144, 238 (1993).
20. Tanabe, K., and Yamaguchi, T., in “Successful Design of Catalysts,”
p. 99. Elsevier, Amsterdam, 1989.
CONCLUSION
21. Fujimoto, K., Maeda, K., and Aimoto, K., Appl. Catal. A 91, 81 (1992).
22. Misono, M., Catal. Rev. Sci-Eng. 29, 269 (1987); 30, 339 (1988).
23. Okuhara, T., Mizuno, N., and Misono, M., Advan. Catal. 41, 113
(1996).
24. Suzuki, S., Kogai, K., and Ono, Y., Chem. Lett., p. 699 (1984); Ono, Y.,
Taguchi, M., Gerile, Suzuki, S., and Baba, T., in “Catalysis by Acids
and Bases” (B. Imelik, C. Naccache, G. Coudurier, Y. Ben Taarit, and
J. C. Vedrine, Eds.), p. 167. Elsevier, Amsterdam, 1985.
25. Na, K., Okuhara, T., and Misono, M., J. Chem. Soc. Chem. Commun.,
p. 1422 (1993).
26. Okuhara, T., Nishimura, T., Watanabe, H., and Misono, M., J. Mol.
Catal. 74, 247 (1992).
27. Inui, T., Makino, Y., Okazaki, F., Nagano, S., and Miyamoto, A., Ind.
Eng. Chem. Res. 26, 647 (1987).
28. Lee, K. Y., Arai, T., Nakata, S., Asaoka, S., Okuhara, T., and Misono,
M., J. Am. Chem. Soc. 114, 2836 (1992).
29. Pines, H., in “The Chemistry of Catalytic Hydrocarbon Conversions.”
Academic Press, New York, 1981.
A
new bifunctional catalyst, Pt or Pd-promoted
Cs2.5H0.5PW12O40, was found to be efficient catalyst for the
skeletal isomerization of n-butane. At 573 K in the pres-
ence of H2, these heteropoly compounds were superior to
Pt-SO24ꢁ/ZrO2 and Pt-HZSM-5 with respect to the selec-
tivity. Over these heteropoly compounds, the bifunctional
mechanism has been proposed, on the basis of the little
dependence of metal loading level and the negative depen-
dence of the H2 pressure on the reaction rate. A unique role
of proton in the selective formation of isobutane has been
demonstrated. A probable interpretation is that protons
present near the metal particles interact with the interme-
diates for hydrogenolysis and direct them to isomerization.
By this, the selective isomerization proceeds over the bi-
functional heteropoly compounds.
30. Aston, J. G., and Szasy, G., J. Chem. Phys. 14, 67 (1946).
31. Chao, K., Wu, H., and Leu, L., J. Catal. 157, 289 (1995).
32. Okuhara, T., Nishimura, T., Na, K., and Misono, M., in “Acid-Base
Catalysis-II” (H. Hattori, M. Misono, and Y. Ono, Eds.), p. 419.
Elsevier/Kodansha, Amsterdam/Tokyo, 1994.
ACKNOWLEDGMENT
This work was partly supported by a Grant-in Aid for Scientific Re-
search from the Ministry of Education, Science, Culture, and Sports of
Japan.
33. Baba, T., and Ono, Y., Zeolite 7, 292 (1987).
34. Wan, K. T., Khouw, C. B., and Davis, M. E., J. Catal. 158, 311 (1996).
35. Adeeva, V., Lei, G. D., and Sachlter, W. M. H., Appl. Catal. A: General
118, L11 (1994).
REFERENCES
36. Zarkalis, A. S., Hsu, C.-Y., and Gates, B. C., Catal. Lett. 29, 235 (1994).
37. Cheung, T., D’Itri, J. L., and Gates, B. C., J. Catal. 151, 464 (1995).
38. Butler, A. C., and Nicolaides, C. P., Catal. Today 18, 443 (1993).
39. Moreno-Castilla, C., Porcel-Jime´nez, A., Carrasco-Marı´n, F., and
Utrera-Hidalgo, E., J. Mol. Catal. 66, 329 (1991).
1. Frischkorn, G. L., Kuchar, P. J., and Olson, R. K., Energy Prog. 8, 154
(1988).
2. Hino, M., Kobayashi, S., and Arata, K., J. Am. Chem. Soc. 101, 6439
(1979); Arata, K., Adv. Catal. 37, 165 (1990).

BUTANE ISOMERIZATION OVER HETEROPOLY COMPOUNDS
107
40. Bond, G. C., and Yide, X., J. Chem. Soc. Faraday Trans. I 80, 969 (1984).
41. Mills, G. A., Heinemann, H., Milliken, T. A., and Oblad, A. G., Ind.
Eng. Chem. 45, 134 (1953); Weisz, P. B., and Swegler, E. W., Science
126, 31 (1957).
44. Anderson, J. R., and Avery, N. R., J. Catal. 5, 446 (1996).
45. Baba, T., Nomura, M., and Ono, Y., in “Acid-Base Catalysis II” (H.
Hattori, M. Misono, and Y. Ono, Eds.), p. 419. Elsevier/Kodansha,
Amsterdam/Tokyo, 1994.
42. Boudart, M., Advan. Catal. 20, 153 (1969).
43. Okuhara, T., Na, K., and Misono, M., in “Science and Technology in
46. Essayem, N., Coudurier, Fournier, G. M., and Vedrine, J. C., Catal.
Lett. 34, 223 (1995).
Catalysis 1994” (Y. Izumi, H. Arai, and M. Iwamoto, Eds.), p. 245. 47. Na, K., Iizaki, T., Okuhara, T., and Misono, M., J. Mol. Catal. A 115,
Kodansha, Tokyo, 1995. 449 (1997).