A kinetic study of catalytic methanol decomposition over nickel supported on silica

Article abstract of DOI:10.1039/b008854o

The kinetics of methanol decomposition to CO and hydrogen over nickel supported on silica was studied using a flow reaction system at 433-498 K. The active site comprised two Ni atoms that were almost covered with adsorption species such as CO, atomic hydrogen, and methoxy groups. The reaction pathway was assumed to be the dissociative adsorption of methanol to methoxy groups and hydrogen adsorbed on Ni sites; decomposition of the methoxy groups to adsorbed CO and hydrogen; and desorption of the surface CO and hydrogen species. The second process was the rate determining step, wherein the surface hydrogen species promoted the decomposition of the methoxy groups, but this contradicted previous studies of decomposition over a clean Ni surface on which hydrogen atoms in the methoxy groups probably transfer to free Ni sites. The estimated heat of adsorption of methanol was close to the activation energy for methanol desorption from a clean Ni surface. However, the heat of adsorption of CO was significantly smaller than that estimated from the adsorption on a clean Ni surface, and it may be due to co-adsorption of hydrogen. The decomposition of the methoxy groups with free Ni sites, which could proceed on a clean Ni surface, was disadvantageous due to the lack of free Ni sites, although the activation energy of the process on a clean surface was probably lower than that for the decomposition process promoted by adsorbed hydrogen.

Full text of DOI:10.1039/b008854o

View Article Online / Journal Homepage / Table of Contents for this issue
A kinetic study of catalytic methanol decomposition over nickel
supported on silica
Yasuyuki Matsumura*¤ and Naoki Tode
Osaka National Research Institute, AIST , Midorigaoka, Ikeda, Osaka 563-8577, Japan
Received 3rd November 2000, Accepted 1st February 2001
First published as an Advance Article on the web 13th March 2001
A kinetic study of catalytic methanol decomposition to carbon monoxide and hydrogen over nickel supported
on silica has been carried out with a Ñow reaction system in the temperature range of 433È498 K. The active
site probably consists of a pair of nickel atoms which can strongly adsorb hydrogen and carbon monoxide.
The reaction pathway is assumed to be: (1) dissociative adsorption of methanol to methoxy groups and
hydrogen adsorbed on nickel sites; (2) decomposition of the methoxy groups to adsorbed carbon monoxide
and hydrogen; and (3) desorption of the surface carbon monoxide and hydrogen species. In the second process,
which is the rate determining step, the surface hydrogen species promote the decomposition of the methoxy
groups, but this contradicts previous studies of decomposition over a clean nickel surface on which hydrogen
atoms in the methoxy groups probably transfer to free nickel sites. The kinetic analysis shows that almost all
the active sites are occupied by adsorption species and the number of free nickel sites is very small under the
chosen reaction conditions. Thus, the transfer of hydrogen to free sites will be a minor step, although this
process should be energetically advantageous to the decomposition step in which the methoxy groups and the
adsorbed hydrogen atoms interact.
contradiction between the reaction kinetics and the Ðndings of
surface science.
Introduction
Methanol decomposition to carbon monoxide and hydrogen
is endothermic, and the reaction is applicable to a methanol-
fueled automobile in which the heat of the exhaust gas is
recovered with the decomposition and the product gas is fed
to the engine.1 In practical application methanol decomposi-
tion is an on-site source of hydrogen and/or carbon monoxide
for chemical processes and also for fuel cells. Catalysts con-
taining nickel are generally active in methanol decomposi-
tion.2h24 In the development of a decomposition catalyst
containing nickel, which is an inexpensive metal, the clari-
Ðcation of the active sites and the reaction mechanism is indis-
pensable. In the case of nickel supported on silica the catalytic
activity does not relate to the metallic surface area of nickel
but to the amounts of carbon monoxide and hydrogen strong-
ly adsorbed on the surface at room temperature.24 Since the
strong adsorption of carbon monoxide takes place on a well-
crystallized nickel surface,25 the relation between adsorption
and activity suggests that the major active sites are present on
the ordered metallic surface.24 The reaction kinetics was
studied with metallic nickel, such as nickel wire and Raney
nickel, and it was proposed that hydrogen adsorbed on the
surface promotes the cleavage of the CÈH bond in the
methoxy group which is formed by the dissociative adsorption
of methanol.2,11 However, surface science studies showed no
such interaction between the hydrogen atom and the methoxy
group, and rather suggested the self-decomposition of the
methoxy to carbon monoxide and hydrogen on the
surface.26h28
Experimental
Nickel supported on silica was prepared by hydrolysis and
polymerization of tetraethyl orthosilicate (GR grade, Kanto
Chemical Co. Ltd) whose solution contained nickel nitrate
[Ni(NO ) É6H O, GR, Kanto] and ethanol. After drying in
3 2
2
air at 393 K the solid was heated in air for 5 h at 673 K for
removal of NO ~ anions and residual organic compounds.
3
The sample contained 40 wt.% of nickel as metal (40 wt.%
NiÈS). Nickel supported on commercial silica (Fuji-Silicia,
ID-G) was prepared by impregnation with nickel nitrate. After
the impregnation the solid was heated in air at 773 K for 5 h
(10 wt.% NiÈI).
Adsorption of hydrogen was carried out in a vacuum
system equipped with Baratron pressure gauges at room tem-
perature. Just before the adsorption, a fresh sample was
reduced with hydrogen (ca. 20 kPa) at 773 K for 1 h and
evacuated at the same temperature for 0.5 h.
The BET surface areas of the catalysts were determined by
the isotherms of nitrogen physisorption.
The catalytic experiments were performed in a Ðxed-bed
continuous Ñow reactor operated at atmospheric pressure.
The catalyst (0.05È0.40 g) was mixed with the commercial
silica (total mass, 0.50 g) and sandwiched with quartz wool
plugs in a tube reactor made of quartz glass (id 6 mm). It was
conÐrmed that neither silica nor the reactor contributed to the
reaction. After reducing the sample in a Ñow of hydrogen (20
vol.%) in argon for 1 h at 773 K, methanol was fed with an
argon carrier at 433È498 K. Since carbon monoxide and
hydrogen signiÐcantly a†ect the reaction rate, the gases were
often co-fed in order to change the reaction composition. The
total Ñow rate of the reaction gas was 1.8È7.5 dm3 h~1. The
reactant and products were analyzed with an on-stream
In this work we have studied the kinetics of methanol
decomposition over nickel supported on silica to solve the
¤ Present address: Research Institute of Innovative Technology for
the Earth, Kizu-cho, Soraku-gun, Kyoto 619-0292, Japan. E-mail:
yasuyuki=rite.or.jp
1284
Phys. Chem. Chem. Phys., 2001, 3, 1284È1288
This journal is ( The Owner Societies 2001
DOI: 10.1039/b008854o

View Article Online
Table 1 Adsorption study of Ni/SiO
2
Amount of H adsorbed
/lmol g(cat)~1
2
Surface area
of Ni
/m2 g(cat)~1
BET
surface area
/m2 g~1
Sample
Stronga
At 20 kPa
40 wt.% NiÈS
10 wt.% NiÈI
52
8
125
51
9.5
3.9
377
215
a The amount adsorbed at a pressure less than 0.01 kPa.
Yanaco G2800 gas chromatograph (Porapak T, 4 m; Ar
carrier) equipped with a thermal conductivity detector. The
methanol conversion was always kept less than 10% and the
partial pressures of the components employed here were the
averages at the inlet and outlet of the reactor.
nickel was determined from the amount of hydrogen adsorbed
at 20 kPa (Table 1) assuming stoichiometric adsorption of a
hydrogen atom on a single nickel site, the atomic cross section
of which is 0.0633 nm2 (ref. 29). The amount of hydrogen
adsorbed on 10 wt.% NiÈI at a pressure below 0.01 kPa was
signiÐcantly smaller than that for 40 wt.% NiÈS (see Table 1).
The BET surface areas of 40 wt.% NiÈS and 10 wt.% NiÈI
were 377 and 215 m2 g~1, respectively.
Results
Adsorption study of the catalysts
Adsorption of hydrogen was carried out at room temperature
with the samples reduced at 773 K for 1 h. The surface area of
Fig. 3 Plot of r~1@2P1@2 [ AP P~1@2 [ CP vs. P1@2 for methanol
M
M
H
C
H
decomposition over 40 wt.% NiÈS.
Fig. 1 Plot of r~1@2P1@2 [ BP1@2 [ CP vs. P P~1@2 for methanol
M
H
C
M H
decomposition over 40 wt.% NiÈS.
Fig. 2 Plot of r~1@2P1@2 [ BP1@2 [ CP vs. P P~1@2 for methanol
Fig. 4 Plot of r~1@2P1@2 [ AP P~1@2 [ CP vs. P1@2 for methanol
M
H
C
M
H
M
M
H
C
H
decomposition over 10 wt.% NiÈI.
decomposition over 10 wt.% NiÈI.
Table 2 Rate constants of methanol decomposition over Ni/SiO
2
Temperature
/K
A
B
C
Catalyst
/mol~1@2 h1@2 g1@2
/mol~1@2 h1@2 g1@2
/mol~1@2 h1@2 g1@2 kPa~1@2
40 wt.% NiÈS
473
453
433
0.32
0.40
0.50
4.9
11.7
31.2
0.66
1.73
8.27
10 wt.% NiÈI
498
473
448
0.50
0.86
1.47
6.9
17.9
63.7
0.77
2.44
15.7
Phys. Chem. Chem. Phys., 2001, 3, 1284È1288
1285

View Article Online
COÈNi H CO ] Ni (equilibrium constant, K )
(5)
C
The dissociative adsorption of methanol to methoxy groups
(step 1) and the adsorption of hydrogen and carbon monoxide
(steps 4 and 5) can take place on the surface of nickel
metal.25h27 When step 2 is the rate determining step and the
reaction is at steady state, the following equations can be
derived:
h \ K~1@2P1@2h
(6)
(7)
(8)
H
H
H
S
h \ K~1P h
C
C
C S
h \ K P h~1h2 \ K K1@2P P~1@2h
M
M
M H S
S
M
H
M
H
where h , h , h , and h are the fractions of Ni (free nickel
s
H
C
M
site), HÈNi, COÈNi and CH OÈNi, respectively, and
Fig. 5 Plot of r~1@2P1@2 [ AP P~1@2 [ BP1@2 vs. P for methanol
3
M
M
H
H
C
decomposition over 40 wt.% NiÈS.
1 \ h ] h ] h ] h
(9)
M
H
C
S
Reaction kinetics
Thus,
Methanol was stoichiometrically decomposed to carbon mon-
oxide and hydrogen over the nickel catalysts under the chosen
reaction conditions. Negligible amounts of methane and water
were detected as by-products. The kinetic data can be Ðtted
with the following equation, which is derived from the reac-
tion kinetics described in the Discussion section:
h \ (K K1@2P P~1@2 ] K~1@2P1@2 ] K~1P ] 1)~1 (10)
S
M
H
M
H
H
H
C
C
Since step 2 is rate determining,
r \ kh h \ kK P h2
M H
M M S
\ kK P (K K1@2P P~1@2 ] K~1@2P~1@2 ] K~1P ] 1)~2
M
M
M
H
M
H
H
H
C
C
r~1@2P1@2 \ AP P~1@2 ] BP1@2 ] CP
(I)
M
M
H
H
C
(11)
where r is the rate of carbon monoxide formation, and P , P
and P are the partial pressures of methanol, hydrogen and
carbon monoxide, respectively. The constants, A, B and C, are
M
H
The rates of formation and decomposition of CH OÈNi (steps
C
2
2 and 3) should be balanced, thus
listed in Table 2. Figs. 1 and 2 show the plots of r~1@2P1@2
M
r \ k h h2
(12)
[ BP1@2 [ CP vs. P P~1@2 for the two catalysts, and the
F F S
H
C
M H
value A is the slope of the line. Figs. 3 and 4 and Figs. 5 and 6
show the plots of r~1@2P1@2 [ AP P~1@2 [ CP vs. P1@2 and
where h is the fraction of CH OÈNi. From eqns. (11) and (12)
F
2
M
M
H
C
H
the plots of r~1@2P1@2 [ AP P~1@2 [ BP1@2 vs. P , respec-
h \ kK P k~1
(13)
M
M
H
H
C
F
M M F
tively. The values B and C are the slopes of the respective
is obtained. Since the decomposition of formaldehyde (step 3)
lines.
occurs easily on nickel catalysts,2,5 k is considered to be
F
much larger than k. Hence, h can be negligible and step 3
F
Discussion
does not contribute to the equation. Then eqn. (11) can be
transformed to
Yasumori et al.2 proposed the following reaction mechanism
for methanol decomposition over a nickel surface:
r~1@2P1@2 \ k~1@2K1@2K1@2P P~1@2 ] k~1@2K~1@2K~1@2P1@2
CH OH ] 2Ni H CH OÈNi ] HÈNi
M
M
H
M
H
M
H
H
3
3
] k~1@2K~1@2K~1P ] k~1@2K~1@2
(14)
(equilibrium constant, K ) (1)
M
C
C
M
M
When the k~1@2K~1@2 term is negligible, the equation can be
CH OÈNi ] HÈNi ] CH OÈNi ] H
M
3
2
2
expressed as
(rate constant, k) (2)
r~1@2P1@2 \ k~1@2K1@2K1@2P P~1@2
CH OÈNi ] 2Ni ] COÈNi ] 2HÈNi
M
M
H
M H
2
] k~1@2K~1@2K~1@2P1@2
(rate constant, k ) (3)
M
H
H
F
] k~1@2K~1@2K~1P
(II)
2HÈNi H H ] 2Ni (equilibrium constant, k )
(4)
M
C
C
2
H
This rate equation is consistent with the experimental one
[eqn. (I)].
We derived other rate equations assuming the rate deter-
mining steps to be
CH OÈNi]Ni]CH OÈNi]HÈNi (rate constant, k ) (15)
3
2
1
CH OÈNi ] CH OÈNi ] 1H (rate constant, k ) (16)
3
2
2
2
2
The following rate equations can be obtained at steady state:
r~1@2P1@2P~1@4
M
H
\ k~1@2K1@2P P~1@2 ] k~1@2K~1@2K~3@4P1@2
1
M
M
H
1
M
H
H
]k~1@2K~1@2K~1@4K~1P ] k~1@2K~1@2K~1@4 (17)
1
M
H
C
C
1
M
H
r~1@2P P~1@2 \ k~1P P~1@2 ] k~1K~1K~1P1@2
M
H
2
M
H
2
M
H
H
Fig. 6 Plot of r~1@2P1@2 [ AP P~1@2 [ BP1@2 vs. P for methanol
M
M
H
H
C
decomposition over 10 wt.% NiÈI.
] k~1K~1K~1@2K~1P ] k~1K~1K~1@2 (18)
2
M
H
C
C
2
M
H
1286 Phys. Chem. Chem. Phys., 2001, 3, 1284È1288

View Article Online
respectively. However, no linear relationship was obtained on
the basis of the rate expressions; hence, the mechanisms (15)
and (16) are less probable.
heat of adsorption of carbon monoxide can be estimated as
74È89 kJ mol~1. The activation energy of carbon monoxide
desorption from a nickel surface was reported26,27,31 to be ca.
130 kJ mol~1, and the values obtained were lower than that.
In step 5 carbon monoxide desorbs from a single nickel site.
However, actual adsorption of carbon monoxide on nickel is
not so simple. At room temperature two adsorption species,
i.e., bridge and linear, are known for strongly adsorbed carbon
monoxide,25,32,33 but the fraction of the bridge species
decreases with increasing coverage of carbon monoxide and
increasing temperature.34,35 Under the chosen reaction condi-
tions the temperature is so high that the bridge species will be
minor. On the surface hydrogen is co-adsorbed. Since the
ratio of the fractions of HÈNi and COÈNi relates to the value
of K~1@2K , i.e., h h~1 \ K~1@2K P1@2P~1, the coverage of
The three kinetic parameters k, K K and K~1@2K can be
M
H
H
C
obtained as A~1B~1, AB~1 and BC~1, respectively, by com-
parison between eqns. (I) and (II). These parameters at 473 K
are listed in Table 3. The activation energy of step 2 (E ) and
a
the heats of adsorption of methanol, hydrogen and carbon
monoxide (Q , Q and Q , respectively) are related to r, K ,
M
H
C
M
K and K as follows:
H
C
r \ A e~Ea@RT
(19)
(20)
(21)
(22)
K \ A eQM@RT
M
M
K \ A e~QH@RT
H
H
H
C
H C
H
C
H
C
hydrogen on the surface is usually higher than that of carbon
monoxide under the chosen reaction conditions. This suggests
that the presence of hydrogen atoms on the surface disturbs
the adsorption of carbon monoxide and, as a result, the
adsorption strength will be weaker.
K \ A e~QC@RT
C
C
where A, A , A and A are frequency factors. Thus, the
M
H
C
values of E , Q ÈQ and Q [ 0.5Q were calculated from the
a
H
M
C
H
H
H
Arrhenius plots of k, K K and K~1@2K (Table 4 and Fig.
M
C
7).
When the fraction of free nickel surface sites is small (h @
S
The activity of Ni/SiO relates empirically to the amount of
1), the term k~1@2K~1@2 in eqn. (16) will be negligible. Hence,
2
M
hydrogen or carbon monoxide strongly adsorbed.24 Thus, the
active site is considered to adsorb hydrogen strongly. The
activation energy for hydrogen desorption from a nickel
surface was reported26,27,30 to be 90È117 kJ mol~1. Assuming
that the value is the same as the heat of adsorption of hydro-
gen, the heats of adsorption of methanol can be calculated as
30È57 and 48È75 kJ mol~1 for 40 wt.% NiÈS and 10 wt.%
NiÈI, respectively. The activation energy of methanol desorp-
tion from the nickel surface was reported26h28 to be 41È67 kJ
mol~1, and the value is close to the heat of adsorption of
methanol calculated here (see Table 4). On the other hand, the
the absence of the term k~1@2K~1@2 in eqn. (II) implies that the
surface sites are almost entirely covered with the adsorption
M
species. On the basis of the rate equation [eqn. (II)], we simu-
lated the changes in reaction rate and surface fractions of
CH OÈNi, HÈNi and COÈNi with changes in the composition
3
of the reaction mixture when 50 kPa of methanol is initially
supplied with argon over 40 wt.% NiÈS at 473 K under atmo-
spheric pressure (Fig. 8). In the Ðgure the change in the com-
position is expressed by the conversion of methanol. It is
noteworthy that the rate increases with methanol conversion
in the low conversion range. At low methanol conversion
where the partial pressure of hydrogen is low, the fraction of
HÈNi increases drastically with increase in methanol conver-
sion because the rate relates to the surface fractions of both
Table 3 Kinetic parameters at 473 K
CH OÈNi and HÈNi (step 2). Yasumori et al.2 conÐrmed the
k
K~1@2K
H
3
C
promotional e†ect of hydrogen in methanol decomposition
Catalyst
/mol h~1 g~1
K
K
/kPa1@2
M
H
over nickel wire with a circulation system operated below 3
kPa, while we were unable to obtain kinetic data at a very
small contact time with the Ñow reaction system. On the other
40 wt.% NiÈS
10 wt.% NiÈI
0.64
0.05
0.065
0.065
7.5
7.3
Fig. 7 Arrhenius plot of the rate constants. Open symbols, 40 wt.%
NiÈS; solid symbols, 10 wt.% NiÈI.
Fig. 8 Relationship between the reaction gas composition and the
reaction rate for 40 wt.% NiÈS at 473 K.
Table 4 The activation energy and the heats of adsorption
E
Q [ Q
Q [ 0.5Q
C
Q
Q
Q
a
H
M
H
H
M
C
Catalyst
/kJ mol~1
/kJ mol~1
/kJ mol~1
/kJ mol~1
/kJ mol~1
/kJ mol~1
40 wt.% NiÈS
10 wt.% NiÈI
Ni surfaceb
98
123
67
60
42
29
30
30È57a
48È75a
41È67
74È88a
75È89a
130
90È117
a The values are calculated on the assumption of Q \ 90È117 kJ mol~1. b The values are taken from refs. 26È28, 30 and 31.
H
Phys. Chem. Chem. Phys., 2001, 3, 1284È1288
1287

View Article Online
hand, the study of methanol decomposition on a clean nickel
surface shows hydrogen transfer from the methoxy species to
the surface of nickel such as by mechanism (15). The activa-
tion energy of the decomposition of the methoxy group on a
clean nickel surface was reported28 to be 67 kJ mol~1, and
this reaction step is probably advantageous to step 2, the acti-
vation energy of which is considerably larger than that (see
Table 4). However, under the conditions of the Ñow reaction,
the pressure of methanol is so high that the surface fraction of
free nickel sites is negligibly small. Hence, the probability of
reaction (15) is actually small, and reaction mainly takes place
although the activation energy of the process on a clean
surface is probably lower than that for the decomposition
process promoted by adsorbed hydrogen.
References
1
2
3
National Research Council, Catalysis L ooks to the Future,
National Academy Press, Washington, DC, 1992.
I. Yasumori, T. Nakamura and E. Miyazaki, Bull. Chem. Soc.
Jpn., 1967, 40, 1372.
I. Yasumori and E. Miyazaki, Nippon Kagakukaishi, 1971, 92,
659.
between CH OÈNi and HÈNi, although the step is energeti-
4
5
S. Kasaoka and T. Shiraga, Nenryo Kyokaishi, 1980, 59, 40.
H. Niiyama, S. Tamai, J.-S. Kim and E. Echigoya, Sekiyu Gak-
kaishi, 1981, 24, 322.
A. Tada, T. Yoshino and H. Itoh, Chem. L ett., 1987, 419.
M. Akiyoshi, H. Hattori and K. Tanabe, Sekiyu Gakkaishi, 1987,
30, 156.
M. Shimizu, K. Nobori and S. Takeoka, Kagaku Kogaku Ron-
bunshu, 1988, 14, 114.
M. Akiyoshi and H. Hattori, Sekiyu Gakkaishi, 1988, 31, 239.
3
cally less advantageous. This reaction mechanism suggests
that the active site comprises two nickel sites.
The kinetic parameters for 10 wt.% NiÈI are similar to
those for 40 wt.% NiÈS except for the values of k (see Fig. 7).
The rate constant, k, relates to the number of active sites, and
the number of sites strongly adsorbing hydrogen on 10 wt.%
NiÈI is signiÐcantly smaller than on 40 wt.% NiÈS. Hence, the
di†erence in the constant, k, is mainly due to the di†erence in
the quantity of active sites that can strongly adsorb hydrogen,
while the quality of the sites for 10 wt.% NiÈI is similar to
that for 40 wt.% NiÈS.
6
7
8
9
10 Y. Nakazaki and T. Inui, Ind. Eng. Chem. Res., 1989, 28, 1285.
11 M. Shimizu and S. Takeoka, Kagaku Kogaku Ronbunshu, 1989,
15, 284.
12 O. Tokunaga, Y. Satoh, T. Fukushima and S. Ogasawara, Sekiyu
Gakkaishi, 1990, 33, 173.
13 H. Imai, T. Tagawa and K. Nakamura, Appl. Catal., 1990, 62,
Conclusions
348.
14 H. Imamura, T. Takada, S. Kasahara and S. Tsuchiya, Appl.
Catal., 1990, 58, 165.
15 A. Tada, Y. Watarai, K. Takahashi, Y. Imizu and H. Itoh, Chem.
L ett., 1990, 543.
16 T. Okubo, M. Watanabe, K. Kusakabe and S. Morooka, Key
Eng. Mater., 1991, 61&62, 71.
17 M. Morita, K. Takeda, S. Tanabe and Y. Matsuda, Nippon
Kagakukaishi, 1991, 1238.
18 M. Watanabe, T. Okubo, K. Kusakabe and S. Morooka, Ind.
Eng. Chem. Res., 1992, 31, 2633.
19 B. Chen and J. L. Falconer, J. Catal., 1993, 144, 214.
20 K. J. Yoon, K. S. Jeong and J. E. Yie, Hwahak Konghak, 1993, 31,
569.
The reaction kinetics of the decomposition of methanol to
hydrogen and carbon monoxide below 500 K over nickel sup-
ported on silica can be expressed as
r~1@2P1@2 \ AP P~1@2 ] BP1@2 ] CP ,
M
M
H
H
C
where r is the rate of carbon monoxide formation, and P , P
M
H
and P are the partial pressures of methanol, hydrogen and
carbon monoxide, respectively. The equation is consistent
C
with the rate derived from the following reaction steps: (1)
dissociative adsorption of methanol to methoxy groups and
hydrogen adsorbed on nickel sites (CH OH ] 2Ni H
21 Y. Matsumura, N. Tode, T. Yazawa and M. Haruta, J. Mol.
3
CH OÈNi ] HÈNi); (2) decomposition of the methoxy groups
Catal. A: Chem., 1995, 99, 183.
3
promoted by the
adsorbed
hydrogen (CH OÈNi
22 Y. Matsumura, K. Kagawa, Y. Usami, M. Kawazoe, H. Sakurai
and M. Haruta, Chem. Commun., 1997, 657.
3
] HÈNi ] CH OÈNi ] H and CH OÈNi ] 2Ni ] COÈNi
2
2
2
23 Y. Matsumura, K. Kuraoka, T. Yazawa and M. Haruta, Catal.
T oday, 1998, 45, 191.
24 Y. Matsumura, K. Tanaka, N. Tode, T. Yazawa and M. Haruta,
J. Mol. Catal. A: Chem., 2000, 152, 157.
] 2HÈNi); and (3) desorption of the surface carbon monoxide
and hydrogen species (2HÈNi H H ] 2Ni and COÈNi H CO
2
] Ni). The heats of adsorption of methanol and carbon mon-
oxide can be estimated on the assumption that the heat of
adsorption of hydrogen is 90È117 kJ mol~1. The heat of
adsorption of methanol estimated is close to the activation
energy for methanol desorption from a clean nickel surface.
However, the heat of adsorption of carbon monoxide is sig-
niÐcantly smaller than that estimated from the adsorption on
a clean nickel surface, and it may be due to co-adsorption of
hydrogen. The active site comprises two nickel atoms which
are almost covered with adsorption species such as atomic
hydrogen, carbon monoxide and methoxy groups. Thus, the
decomposition of methoxy groups with free nickel sites
(CH OÈNi ] Ni ] CH OÈNi ] HÈNi), which can proceed on
25 J. T. Yates and C. W. Garland, J. Phys. Chem., 1961, 65, 617.
26 S. R. Bare, J. A. Stroscio and W. Ho, Surf. Sci., 1985, 150, 399.
27 L. J. Ritcher and W. Ho, J. Chem. Phys., 1985, 83, 2569.
28 J. J. Vajo, J. H. Campbell and H. B. Christopher, J. Phys. Chem.,
1991, 95, 9457.
29 J. W. E. Coenen, Appl. Catal., 1991, 75, 193.
30 K. Christmann, O. Schober, G. Ertl and M. Neumann, J. Chem.
Phys., 1974, 60, 4528.
31 N. Takagi, J. Yoshinobu and M. Kawai, Chem. Phys. L ett., 1993,
215, 120.
32 C. E. OÏNeill and D. J. C. Yates, J. Phys. Chem., 1961, 65, 617.
33 M. Primet, J. A. Dalmon and G. A. Martin, J. Catal., 1977, 46,
25.
34 J. Lauterbach, M. Wittmann and J. Kuppers, Surf. Sci., 1992, 279,
3
2
a clean nickel surface, is disadvantageous under the actual
reaction conditions because of the lack of free nickel sites,
287.
35 J. Yoshinobu and M. Kawai, Surf. Sci., 1996, 363, 105.
1288
Phys. Chem. Chem. Phys., 2001, 3, 1284È1288