Direct decomposition of NO into N2 and O2 over La(Ba)Mn(In)O3 perovskite oxide

Article abstract of DOI:10.1016/S0021-9517(03)00265-3

Although LaMnO₃ perovskite oxide has been reported to exhibit low activity in NO direct decomposition into N₂ and O₂, doping Ga or In at the Mn site of La(Ba)MnO₃ has been found effective of increasing the activity for NO direct decomposition into N ₂ and O₂. The activity of NO decomposition increased in the order Ba>Sr>Ca for the La site dopant, and In>Ga for the Mn site. Among the investigated dopants and compositions, the highest N₂ yield was achieved with La_0.7Ba_0.3Mn_0.8In _0.2O₃. On this catalyst, NO conversion increased with increasing reaction temperature, and at 1123 K, NO conversion into N ₂ and O₂ attained values of 75 and 41%, respectively. The high yield of N₂ and O₂ was maintained for 12 h. Coexistence of oxygen decreased the N₂ yield with P_O2 ^-0.53; however, a N₂ yield of 15% could be sustained even at 10% coexisting O₂ at 1073 K. The NO decomposition rate increased with increasing NO partial pressure and obeyed with P_NO ^1.31. O₂ temperature-programmed desorption measurements showed that oxygen desorption was greatly enhanced by In doping at the Mn site. NO TPD also showed that the amount of NO adsorbed greatly increased with In doping. Therefore, improved activity of NO decomposition with In substitution seems to be caused by the weakening adsorption of oxygen and the increased adsorption of NO. IR measurements of adsorbed NO also suggest that the major adsorption species at high temperature was NO₃^- and it seems likely that NO decomposition proceeds after removal of NO₃ ^- and/or oxygen. N₂O direct decomposition on La _0.7Ba_0.3Mn_0.8In_0.2O₃ was further studied. It was found that La_0.7Ba_0.3Mn _0.8In_0.2O₃ is highly active in the direct decomposition of N₂O even under the coexistence of O₂. Therefore, decomposition of NO on La_0.7Ba_0.3Mn _0.8In_0.2O₃ may proceed via N₂O as the intermediate species.

Full text of DOI:10.1016/S0021-9517(03)00265-3

Journal of Catalysis 220 (2003) 104–114
www.elsevier.com/locate/jcat
Direct decomposition of NO into N and O over La(Ba)Mn(In)O
2
2
3
perovskite oxide
a,∗
a
a
a
b
Tatsumi Ishihara, Makoto Ando, Kenji Sada, Keiko Takiishi, Keiji Yamada,
a
a
Hiroyasu Nishiguchi, and Yusaku Takita
a
Department of Applied Chemistry, Faculty of Engineering, Oita University, Dannoharu 700, Oita 870-1192, Japan
b
Graduate School of Engineering, Oita University, Dannoharu 700, Oita 870-1192, Japan
Received 11 February 2003; revised 4 June 2003; accepted 4 June 2003
Abstract
Although LaMnO perovskite oxide has been reported to exhibit low activity in NO direct decomposition into N and O , doping Ga
3
2
2
or In at the Mn site of La(Ba)MnO has been found effective of increasing the activity for NO direct decomposition into N and O .
3
2
2
The activity of NO decomposition increased in the order Ba > Sr > Ca for the La site dopant, and In > Ga for the Mn site. Among the
investigated dopants and compositions, the highest N yield was achieved with La Ba Mn In O . On this catalyst, NO conversion
2
0.7 0.3
0.8 0.2 3
increased with increasing reaction temperature, and at 1123 K, NO conversion into N and O attained values of 75 and 41%, respectively.
2
2
−
0.53
The high yield of N and O was maintained for 12 h. Coexistence of oxygen decreased the N yield with P
5% could be sustained even at 10% coexisting O at 1073 K. The NO decomposition rate increased with increasing NO partial pressure
. O temperature-programmed desorption measurements showed that oxygen desorption was greatly enhanced by In
2
doping at the Mn site. NO TPD also showed that the amount of NO adsorbed greatly increased with In doping. Therefore, improved activity
; however, a N yield of
2
2
2
2
O2
1
2
1
.31
and obeyed with P
NO
of NO decomposition with In substitution seems to be caused by the weakening adsorption of oxygen and the increased adsorption of NO.
−
IR measurements of adsorbed NO also suggest that the major adsorption species at high temperature was NO and it seems likely that NO
3
−
decomposition proceeds after removal of NO and/or oxygen. N O direct decomposition on La Ba Mn In O was further studied.
3
2
0.7 0.3
0.8 0.2 3
It was found that La Ba Mn In O is highly active in the direct decomposition of N O even under the coexistence of O . Therefore,
0
.7 0.3
0.8 0.2
3
2
2
decomposition of NO on La Ba Mn In O may proceed via N O as the intermediate species.
0
.7 0.3
0.8 0.2
3
2

2003 Elsevier Inc. All rights reserved.
Keywords: Nitrogen oxide; Direct decomposition; Perovskite oxide; LaMnO ; N O decomposition; Dopant effects
3
2
1
. Introduction
is well known that the oxygen formed adsorbs strongly on
the catalyst, resulting in deactivation of the catalyst. Some
catalysts, such as Cu-ZSM-5 [5], Co-ZSM-5 (which con-
tains Co in the framework [6]), La2O3 [7], Ba/MgO [8], and
LaCoO3 [9] based perovskite oxides, are active in the direct
decomposition of NO. In particular, Teraoka et al. reported
Nitrogen oxides (NOx), which are mainly formed in an
internal combustion engine, are extremely toxic to the hu-
man body and also harmful to the environment as a main
source of acid rain. At present, several methods have been
proposed for NOx removal [1–3]. Among them, selective
reduction of NOx by hydrocarbons has been studied exten-
sively, and various catalysts, in particular Cu-ZSM-5, have
been reported as active catalysts for this reaction [4]. On
the other hand, direct decomposition of NO into N2 and
O2 (2NO = N2 + O2) is the most ideal reaction for NOx
removal because the process is quite simple. However, it
that La0.8Sr0.2CoO is highly active in NO decomposition,
3
and NO conversions into N and O reached values of 72
2
2
and 40%, respectively, at 1073 K [10]. Although the reac-
tion temperature is high, high activity of NO decomposition
is expected on perovskite oxides. In addition, the tempera-
ture of the exhaust gas at the engines outlet is higher than
1
073 K. Increasing the reaction temperature is expected to
decrease the negative effects of water and sulfur compounds.
In the present study, NO decomposition over LaMnO3
perovskite oxide doped with Ga or In at the Mn site was
*
Corresponding author.
E-mail address: ishihara@cstf.kyushu-u.ac.jp (T. Ishihara).
0021-9517/$ – see front matter  2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0021-9517(03)00265-3

T. Ishihara et al. / Journal of Catalysis 220 (2003) 104–114
105
investigated. The activity of LaMnO3-based oxide in NO
decomposition is low, except for the doped SrMnO3 mixed
oxide of Sr0.6La0.4Mn0.8Ni0.2O3 [10,11]. However, increas-
ing the mobility of oxide ions in LaMnO3 by doping with Ga
or In may increase the activity of NO decomposition. This is
because the doped LaGaO3 perovskite oxide exhibits a high
oxide ion mobility [12] and the rate-limiting step of this re-
action on perovskite catalysts is considered the adsorption
of NO into the coupled oxygen vacancy sites [10], and in-
creasing the mobility of oxygen vacant by dopant seems to
be a useful method for formation of the coupled oxygen va-
cancy sites. A relationship between NO decomposition and
oxide ion conductivity is also predicted on a brownmillerite
oxide [13].
by changing the feed rate of He as a balance gas to keep the
total flow rate of reactant at 20 ml/min constant.
2.3. Characterization of the catalyst
Temperature-programmed desorption of O2 and NO was
measured with a mass spectrometer (Anelva AQA-100R) as
a detector. Standart TPD equipment consisting of a gas ad-
sorption reactor connected to a mass spectrometer and an
adsorption gas circulating line was used. Catalyst (0.3 g)
was always fixed in the quartz tube (6 mm in diameter)
with quartz wool. The catalyst was evacuated at 873 K for
1 h, exposed to oxygen or NO gas at 13.32 kPa for 0.5 h,
and then cooled to room temperature in the sample gas at-
mosphere. After evacuation at room temperature for 0.5 h,
desorption of the adsorbed oxygen or NO was measured at
a heating rate of 10 K/min. The adsorption state of NO was
also measured with an IR spectrometer (JASCO-610) with
an MCT detector. Measurement was performed with the dif-
fusion reflection unit using a KBr window and connecting
to a gas-circulating and vacuum system. After evacuation at
873 K for 3 h, background spectra of the catalyst without
NO adsorption were measured at each temperature. The ca-
talyst was exposed to NO gas at 13.32 kPa and heated for
1 h in a gas-circulating system at the measurement tempera-
ture, and the IR measurement was performed at an elevated
temperature.
2
. Experimental
2
.1. Preparation of catalyst
Doped LaMnO3 was prepared by a conventional solid-
state reaction method. The precursor of LaMnO3 was ob-
tained by evaporating the aqueous solution of a calculated
amount of La(NO3)3, Sr(NO3)2, Mn(CH3COO)2, and metal
nitrate acid. The mixtures obtained were calcined in air at
1
273 K for 3 h. The sample obtained was measured with
X-ray diffraction using a commercial diffract meter (Rigaku
Rint-2500) with a Cu-Kα line. The catalyst powder thus ob-
tained was pressed into disks, crushed, and sieved into 16 to
3
2 meshes.
3. Results and discussion
2
.2. NO, N2O, and NO2 decomposition reactions
3.1. Effects of dopant on NO decomposition activity
of LaMnO3
Direct decomposition of NO, NO2, and N2O was per-
formed with a conventional fixed-bed gas-flow reactor with
a 12-mm-diameter quartz glass tube. Gaseous mixtures of
Table 1 summarizes the activities of the examined cata-
lysts doped with Ga at the B site of perovskite oxide (ABO3)
in NO decomposition at 15 min after the reaction started.
Because NO conversion on the several catalysts examined,
gradually decreased, NO decomposition activity at 15 min
after reaction start is listed in Table 1. Although the reaction
temperature was as high as 1073 K, all examined catalysts
doped with Ga exhibited activity in NO direct decomposi-
tion, except the LaCrO3 catalyst. No N2O formation was
observed on any catalyst examined. It is considered that
the temperature range for activity is too high to form N2O.
The activity of NO decomposition decreased in the order
LaMnO3 > LaCoO3 > LaCuO3 > LaFeO3 ꢀ LaCrO3,
which is almost the same tendency as that reported ear-
lier [14]. It is clear that the yield of N2 is always much higher
than that of O2. Therefore, an observed deficit of oxygen
in the products may be connected with NO2 formation in a
cool zone after the reaction. However, even considering the
formation of NO2, the amount of oxygen formation was defi-
cient on many catalysts. On these catalysts, such as LaCuO3-
or LaFeO3-based oxides, the oxygen formed strongly ad-
sorbs on the catalyst and the activity of NO decomposition
1
% NO, 1% NO2, or 10% N2O diluted with He were fed
to the catalyst bed for NO, NO2, and N2O direct decomposi-
tion reactions, respectively. One gram of the catalyst without
dilution was always set in the reactor by using quartz wool
(
ca. 0.5 mm in catalyst height) and the feed rate of the re-
3
actant was fixed at W/F = 3.0 g s/cm , where W and F
are the catalyst weight and the gas flow rate. Produced N2,
O2, and the fed NO were analyzed with online gas chro-
matography with a molecular sieve 5A column and N2O
with a Polapac-Q column with a thermal conductivity de-
tector (TCD). Because NO2 cannot be analyzed with gas
chromatography, an observed deficit in oxygen in the prod-
ucts may be connected to NO2 formation in a cool zone after
the reaction. Therefore, in this study, the activity of the cata-
lyst in NO decomposition is discussed in terms of the N2
yield. It is also noted that the formation of N2O was not ob-
served in this study.
The effects of the coexistence of O2 were measured by
mixing 10% O2 diluted with He with the reactant gas. Con-
centrations of oxygen as well as NO or N2O were controlled

106
T. Ishihara et al. / Journal of Catalysis 220 (2003) 104–114
Table 1
NO direct decomposition into N and O over LaMO (M = Cr,Fe,Cu,Co,Mn) catalyst
2
2
3
Catalyst
BET surface
Conversion (%)
2
a
area (m /g)
Of NO
Into N
Into O
Into N O
2
Into NO
2
2
2
La Ba Cr Ga
O
O
O
O
1.9
3.1
1.0
0.9
7.5
4.2
0.8
8.0
7.0
4.5
3.8
48.9
58.4
78.6
92.3
37.5
24.4
96.8
89.7
4.2
4.1
31.3
40.7
54.1
60.2
21.8
12.8
63.7
59.7
1.8
0.1
1.5
3.9
8.9
22.8
1.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
17.6
17.7
24.5
32.1
15.7
12.6
33.1
30.0
2.4
0
.7 0.3 0.8 0.2
3
3
La Ba Fe Ga
0.7 0.3 0.8 0.2
La Ba Cu Ga
0.7 0.3 0.8 0.2
3
La Ba Co Ga
0.7 0.3 0.8 0.2
3
La Ba Mn Ga O
0.7 0.3
0.8 0.2 3
La Sr Mn In
O
3
0.7 0.3
0.8 0.2
La Ca Mn In O
0.5
0.7 0.3
0.8 0.2 3
La Ba Mn In O
24.4
14.5
0.0
0.7 0.3
0.8 0.2 3
La Ba Mn Al
O
3
0.7 0.3
0.8 0.2
In O
2
3
a
−3
Estimated based on material balance of nitrogen. Catalyst, 1 g; 1.0% NO in He; W/F = 3.0 g s cm . Temperature, 1073 K.
may be decreased by the adsorbed oxygen after a longer pe-
riod. In Table 1, the BET surface area is also summarized.
However, it is apparent that no clear relationship between
surface area and NO decomposition activity was observed.
Among the examined perovskite catalysts doped with Ga,
LaMnO3-based oxide exhibited a fairly high N2 yield, and
the formation of O2 was also observed. Therefore, in the
following, NO direct decomposition on doped LaMnO3 was
studied in detail. Because LaMnO3-based oxide exhibits the
largest BET surface area in the Ba-doped catalyst, the large
surface area might be one reason for the high activity of
LaMnO3-based oxide.
Fig. 1. XRD patterns of the La Ba Mn
InxO catalyst.
3
0.7 0.3
1−x
Effects of dopant at the Mn site of LaMnO3 were studied
further. NO decomposition activities on the catalyst doped
with Al and In, which are other elements in group 13 (or
III B) of the periodic table, are also shown in Table 1. Al-
though the activity of Al-doped La0.7Ba0.3MnO3 catalyst in
NO decomposition was slightly low, the catalyst doped with
In exhibited the highest N2 yield, which is comparable with
that of the Ga-doped catalyst. The yield of O2 was slightly
higher over the catalyst doped with In than over that doped
with Ga. Therefore, from a stability point of view, the In-
doped catalyst is more promising than the Ga-doped catalyst.
Consequently, in this study, NO decomposition on LaMnO3
doped with Ba and In at the La and Mn sites, respectively,
was investigated in detail.
In Table 1, the effects of alkaline earth cation at the La
site on the activity of NO decomposition are also shown.
Compared with the activity on the Sr- or Ca-doped cata-
lyst, the Ba-doped catalyst exhibited much higher activity in
NO direct decomposition. Consequently, barium is a better
dopant at the La site of LaMn0.8In0.2O3 for NO decomposi-
tion. In the following section, the content of dopant will be
optimized.
fraction peaks from a secondary phase close to the main
peaks, which may be In2O3, were also observed when the
amount of In was higher than 20 mol%. Therefore, it can
be said that the main phase of the sample is LaMnO3.
However, the limit of the solid solution of In at the Mn
site of LaMnO3 seems to exist between x = 0.1 and 0.2
in La0.7Ba0.3Mn1−xInxO3 and a small amount of In2O3 is
formed. The amount of In2O3 is negligibly small, because
the diffraction peak from In2O3 is weak. On the other hand,
diffraction peaks from LaMnO3 are shifted to a lower an-
gle with increasing amounts of In. Because the ionic sizes
of In3 and Mn with the six coordination number are 640
and 500 pm, respectively, expansion of the lattice parameter
with increasing In amount suggests that added In replaced
the lattice position of Mn in the LaMnO3 perovskite lattice.
Fig. 2 shows the conversion of NO into N2, N2O, and
O2 on La0.7Ba0.3InxMn1−xO3 at 1073 K as a function of
the value of x. Similar to the results reported by Teraoka
et al. [10], the activity of La0.7Ba0.3MnO3 in NO decom-
position was low in this study. The yield of N2 and O2 in-
crease with increasing In amounts and attain the maximum
at x = 0.2. Therefore, the optimum amount for In doping
seems to exist around x = 0.2. Since the amount of O2 for-
mation greatly increases with In doping, improved NO de-
composition activity is considered to result from weakening
O2 adsorption on the catalyst. Here it is also obvious that the
BET surface area is almost the same on all catalysts, and so
+
3+
3
.2. Effects of Ba and In contents on NO decomposition
activity of LaMnO3 catalyst
Fig. 1 shows the XRD patterns of the La0.7Ba0.3Mn1−x
InxO3 catalyst. It is clear that strong main diffraction peaks
could be assigned to those from LaMnO3, and weak dif-

T. Ishihara et al. / Journal of Catalysis 220 (2003) 104–114
107
Fig. 2. Conversion of NO and yield of N and O on La Ba Inx
Fig. 4. NO conversion and yield of N and O at 1073 K as a function of x
2 2
2
2
0.7 0.3
−3
Mn
O
at 1073 K as a function of x (NO: 1%; He: balance;
in La
Ba Mn In O (NO: 1%; He: balance; W/F = 3 g s cm ).
1−x
3
1−x
x
0.8 0.2
3
cat
−3
W/F = 3 gcat s cm ).
crease with increasing x and attain the maximum at x = 0.3.
Therefore, the optimum amount of Ba addition seems to ex-
ist at x = 0.3. From the above results, it can be concluded
that the highest activity of NO decomposition on Ba- and In-
doped LaMnO3 is attained at La0.7Ba0.3Mn0.8In0.2O3 among
the LaMnO3 perovskite oxides.
The temperature dependence of NO conversion into N2
and O2 on La0.7Ba0.3Mn0.8In0.2O3 is shown in Fig. 5. Obvi-
ously, NO decomposition proceeds from 873 K, and O2 for-
mation is also observed at temperatures higher than 973 K.
The amount of O2 formed is always smaller than that of N2.
However, if the formation of NO2 is considered, the amounts
of nitrogen and oxygen are almost balanced with the amount
of NO converted. Therefore, it seems likely that no accu-
mulation of oxygen on the catalyst occurred. In the follow-
ing, NO decomposition activity is discussed mainly with the
yield of N2.
increased NO decomposition activity cannot be explained by
the geometrical effects. Because excessively added In results
in the formation of In2O3, as detected by XRD, the activity
of In2O3 in NO decomposition was also investigated. The
results are shown in Table 1, where it is seen that the activ-
ity of In2O3 oxide in NO decomposition is low. Therefore,
the influence of the impurity phase of In2O3 seems to be
negligible in NO decomposition and the improved NO de-
composition activity can be assigned to the chemical effects
of In doped into the LaMnO3 lattice.
The effects of Ba amount at the La site on NO decom-
position activity were further studied, and XRD patterns of
La1−xBaxMn0.8In0.2O3 are shown in Fig. 3. XRD patterns
of the obtained La1−xBaxMn0.7In0.3O3 specimens also con-
sist of diffraction peaks from LaMnO3 and In2O3 in the
range of x = 0.2–0.4 in La1−xBaxMn0.8In0.2O3. On the
other hand, the diffraction peak from the BaMnO3 phase
appears at x = 0.5 and becomes the main peak at x = 0.7.
Therefore, LaMnO3 perovskite is still the main phase at low
Ba content; however, BaMnO3 is the main phase when x in
La1−xBaxMn0.8In0.2O3 is higher than 0.5.
Fig. 6 shows the conversion of NO and the yield of
N2, N2O, and O2 at 1073 K as a function of time on
stream. It is obvious that the conversion of N2 attains a
Fig. 4 shows the NO conversion and the yield of N2, N2O,
and O2 at 1073 K as a function of x in La1−xBaxMn0.8
In0.2O3. It is clear that NO conversion as well as N2 yield in-
Fig. 5. Temperature dependence of NO conversion and yield of N
2
and O on La Ba Mn In O
0.8 0.2 3
(NO: 1%; He: balance; W/F =
2
0.7 0.3
−3
Fig. 3. XRD patterns of La
Bax Mn In O catalyst.
3 g s cm ).
1−x
0.8 0.2
3
cat

108
T. Ishihara et al. / Journal of Catalysis 220 (2003) 104–114
Fig. 7. Effects of contact time on N yield on La Ba Mn In
O at
3
2
0.7 0.3
0.8 0.2
Fig. 6. Conversion of NO and yield of N and O at 1073 K on
2
2
1073 K (NO: 1%; He: balance).
La Ba Mn In
O as a function of time on stream (NO: 1%; He:
0.7 0.3
0.8 0.2
3
−3
balance; W/F = 3 gcat s cm ).
The effects of coexisting oxygen on the N2 formation
value of 64% at 1073 K during the initial 15 min and it
is quite stable over 12 h. Therefore, perovskite oxide of
La0.7Ba0.3Mn0.8In0.2O3 exhibits a stable activity in NO di-
rect decomposition, and considering the time dependence of
N2 formation, the high activity of NO decomposition will
be exhibited for longer periods. In any case, it is obvious
that La0.7Ba0.3Mn0.8In0.2O3 exhibits stable activity in NO
decomposition. In the previous study, it was generally con-
sidered that the catalytic activity of LaMnO3-based oxide
in NO decomposition was low; however, this study reveals
that In-doped LaMnO3 is active in NO direct decomposition,
which is comparable with that of LaCoO3-based oxide as
a well-known active NO direct decomposition catalyst [14].
The improvement in NO decomposition activity could be as-
signed to weakening of the oxygen adsorption; this will be
discussed later.
rate were further studied, and results are shown in Fig. 8.
Oxygen strongly adsorbs on the catalyst and decreases the
NO decomposition activity. Therefore, only a limited num-
ber of catalysts show stable NO decomposition activity un-
der the coexistence of oxygen. Teraoka et al. reported that
Sr0.6La0.4Mn0.8Ni0.2O can decompose NO into N under
3
2
coexistence of O up to 10% [10]. As shown in Fig. 8,
2
analogous to the results with SrMnO -based oxide, N
2
3
yield monotonically decreases with increasing O concen-
2
tration. Therefore, it seems likely that the coexisting oxygen
strongly adsorbs on the active site of La0.7Ba0.3Mn0.8In0.2O3
and, consequently, the coexisting oxygen strongly interferes
with the NO decomposition reaction. However, even under
10% O coexistence, La0.7Ba0.3Mn0.8In0.2O catalyst can
2
3
proceed in the NO decomposition reaction and the N yield
2
is sustained at 12%.
To analyze the reaction mechanism kinetically, PO and
dependences on N formation rate were estimated.
NO 2
2
3
.3. Effects of reaction condition on NO direct
P
decomposition activity
Fig. 7 shows the effects of contact time on NO decom-
position activity. It is seen that N2 yield gradually decreases
with decreasing contact time and at W/F = 1.5 gcat h/mol,
N2 yield is 35%. Although lower conversion in shorter con-
tact time is generally observed in all heterogeneous catalytic
reactions, the decrease in NO decomposition activity on this
catalyst is slightly significant as the contact time decreases.
This decrease in N2 yield with decreasing contact time might
be attributed to the low BET surface area, since the BET sur-
2
face area of La0.7Ba0.3Mn0.8In0.2O3 is as low as 8.0 m /g.
For removal of NOx in exhaust gas from an internal com-
bustion engine, high activity at much shorter contact time is
required. Therefore, increasing the BET surface area may be
the most effective method for improving the NO decompo-
sition activity on La0.7Ba0.3Mn0.8In0.2O3 catalyst at shorter
contact times. This is now under investigation and the results
will be reported in the future.
Fig. 8. Effects of coexisting oxygen on
N
formation rate on
2
La Ba Mn In
O
0.8 0.2 3
gcat s cm ).
at 1073 K (NO: 1%; He: balance; W/F =
0
.7 0.3
−
3
3

T. Ishihara et al. / Journal of Catalysis 220 (2003) 104–114
109
it is also noted that fairly close values for NO and O2 depen-
dence have been reported for La2O3 [15] and Ba/MgO [8].
3
.4. Adsorption states of NO and O2 by DRIFTS
To study the effects of In at the Mn site of the LaMnO3
catalyst on the adsorption states of NO and O2, the adsorp-
tion states of NO and O2 on La0.7Ba0.3Mn0.8In0.2O3 and
La0.7Ba0.3MnO3 were measured with diffuse reflectance FT-
IR spectroscopy (DRIFTS). The number of reports on the
NO adsorption state of Mn-based oxide has been rather lim-
ited so far, and the adsorption state on Mn-based oxide has
not been studied sufficiently [16]. Fig. 11 shows DRIFTS
spectra of absorbed NO at various temperatures for 1 h. Af-
ter introduction of NO at 873 K, absorption peaks appear
Fig. 9. N formation rate on La Ba Mn In O as a function of oxy-
gen partial pressure at 1073 K.
2
0.7 0.3
0.8 0.2 3
−1
at 2343 and 2333 cm , which could be assigned to N O
2
species. With increasing temperature, a broad absorption
band appears around 1100 and a sharp absorption band at
Fig. 9 shows the N2 formation rate as a function of oxy-
gen partial pressure. It is obvious that the N2 formation
rate monotonically decreases with increasing oxygen par-
tial pressure and the P O2 dependence is estimated to be
−1
−
1
349 cm . This could be assigned to nitrate (NO3 ) or ni-
−
trite (NO2 ) species according to the assignment of Vannice
and co-workers on La2O3 oxide [15]. Because the difference
in the IR spectra of NO2 and NO3 is not large, adsorption
species at 1363 cm could not be easily assigned only with
this experiment. However, considering that adsorption bands
appeared at 1180, 1114, and 1076 cm , it seems most likely
that the adsorption species at 1363 cm could be assigned
to the unidentate nitrate (MONO2, where M is metal). In ad-
dition, unidentate and bidentate NO3 are easily formed on
−
−
−
0.53, which is similar to the value reported for SrMnO3-
−1
based oxide. Fig. 10 shows the N2 formation rate as a
function of NO partial pressure. In contrast to the effects
of coexisting oxygen, N2 formation rate monotonically in-
creases with increasing NO partial pressure. The depen-
dence of N2 formation rate on NO partial pressure is es-
timated to be 1.31. Therefore, the rate-determining step
for NO decomposition on La0.7Ba0.3Mn0.8In0.2O3 seems
to be related to the adsorption or activation of NO. The
N2 formation rate for NO decomposition as a function of
O2 and NO partial pressures has already been reported
by Teraoka et al. for La0.4Sr0.6Mn0.8Ni0.2O3, and P NO
−
1
−
1
−
MnO2 [16]. Although further detailed study is required, we
−
tentatively assign this ionic derivative from NO as NO3 in
this study.
Adsorption bands also appeared at 1629 and 1513 cm 1,
−
−
which could be assigned to bent-type NO adsorption (NO )
considering the adsorption frequency [15]. After evacuation
and P O dependences on this catalyst are 1.04 and −0.16,
2
−1
at room temperature, the absorption band at 1629 cm dis-
appeared. Therefore, two absorption peaks could be assigned
respectively [10]. Therefore, the estimated reaction order
of N2 formation rate for NO and O2 partial pressure on
La0.7Ba0.3Mn0.8In0.2O3 can be almost the same as that re-
ported for La0.4Sr0.6Mn0.8Ni0.2O3 [10]. On the other hand,
−
to the NO species adsorbed at the different adsorption sites.
Fig. 11. DRIFTS spectra of NO absorbed on La Ba Mn In
O at
3
0
.7 0.3
0.8 0.2
various temperatures for 1 h. (a) NO adsorption at 873 K in NO (13.32 kPa);
(b) evacuation at 298 K for 1 h; (c) evacuation at 473 K for 1 h; (d) evacua-
tion at 673 K for 1 h.
Fig. 10. N formation rate on La Ba Mn In O as a function of NO
partial pressure at 1073 K.
2
0.7 0.3
0.8 0.2 3

110
T. Ishihara et al. / Journal of Catalysis 220 (2003) 104–114
−
It is considered that the NO adsorption species appearing
atures higher than 773 K on both catalysts. Therefore, ad-
sorption of oxygen on these catalysts is strong. On the other
hand, the amount of desorbed oxygen greatly increases with
In doping at the Mn site of La0.7Ba0.3MnO . Because no O
−1
at 1629 cm is weak and this nitrosyl species easily des-
orbed at low temperature. On the other hand, absorbance of
the peak at 1513 cm slightly decreased after evacuation at
73 K. Therefore, another bent-type NO seems to be a rather
stable adsorption species. In contrast, the intensity of absorp-
tion bands at 1349, 1180, 1114, 1076, and 831 cm hardly
changed after evacuation at 673 K for 1 h. Therefore, ad-
sorption of NO3 is strong and covers the catalyst surface.
−1
3
2
6
2 3
desorption was observed on In O oxide up to 1073 K, the
increased amount of oxygen is not caused by the decomposi-
tion of In O . Therefore, it is obvious that In doping is effec-
−1
2
3
tive for decreasing the adsorption strength of oxygen, result-
ing in improved activity in NO decomposition. Fig. 13 show
the NO desorption profiles from La0.7Ba0.3Mn0.8In0.2O3
−
Considering the change in the IR spectra of NO, the activa-
tion of NO proceeds through the bent-type adsorption and, at
the first step, NO species couple to form N2O and surface
and La0.7Ba0.3MnO oxides. In this figure, desorption of
3
−
N and O , which forms from the adsorbed NO, is also
2
2
oxygen (2NO = N2O + Oad). Surface N2O further decom-
poses to form N2 and surface oxygen. In the following steps,
gaseous NO or adsorbed NO reacts with surface oxygen to
shown. Desorption of NO occurs in three different temper-
ature ranges, namely, two peaks at low (around 373 K and
around 473 K) and one at high (around 873 K) tempera-
ture, from both catalysts. The desorption of O2 is observed
in a high temperature range on both catalysts. However,
desorption of N2 is not observed at high temperature on
La0.7Ba0.3MnO3, while a small amount of N2 formation is
observed on La0.7Ba0.3Mn0.8In0.2O3. On the other hand, a
small amount of N2 desorption is observed on both cata-
lysts at low temperatures (373–573 K). The amount of N2
desorption is much smaller than that of O2. N2 formation
is also observed during NO adsorption at 773 K. Although
the amount of N2 formed was not quantitatively analyzed
−
−
form NO3 or NO2 species. Here the charge neutrality for
−
−
the formation of NO3 or NO2 species can be retained by
oxidation of Mn3 cation to form Mn . The formation of
+
4+
N2 is also observed by mass spectroscopy in the gas phase.
−
Thus, formed NO3 is the strong adsorption species and
covers the active site on the La0.7Ba0.3MnO3 catalyst. Be-
cause formation of N2O was not detected in the reaction by
gas chromatograph analysis, N2O seems to be the adsorp-
tion species on the catalysts and decomposition of N2O is
too fast to observe in the gas phase. However, IR measure-
ment clearly suggests that NO direct decomposition seems to
proceed via N2O as the intermediate reaction species. N2O is
considered to be an intermediate in the NO selective reduc-
tion of NH3 and direct decomposition on La2O3 [15] and
Ba/MgO [7].
3
.5. Adsorption states of NO and O2 by TPD and reaction
mechanism
Effects of In on NO and O2 adsorption states were stud-
ied with temperature-programmed desorption (TPD) tech-
niques. Fig. 12 shows desorption profiles of oxygen from
La0.7Ba0.3Mn0.8In0.2O3 and La0.7Ba0.3MnO3 oxide. It can
be seen that desorption of oxygen occurs only at temper-
Fig. 12. O -TPD profiles (m/e = 32) from La Ba Mn In
O
and
Fig. 13. NO-TPD profiles (m/e = 30) from La Ba Mn In
O (a)
0.8 0.2 3
2
0.7 0.3
0.8 0.2
3
0.7 0.3
La Ba MnO oxide.
and La Ba MnO (b) oxides.
0.7 0.3 3
0
.7 0.3
3

T. Ishihara et al. / Journal of Catalysis 220 (2003) 104–114
111
−
in this study, the disproportionation reaction (3NO + e =
NO3 species which formed by disproportionate adsorp-
−
−
N2 + NO3 ) may occur during adsorption at 773 K, result-
tion of NO (3NO + e = N2 + NO3 ) or adsorption of NO
−
2−
ing in the formation of NO3 species on the surface. Here,
on the strongly adsorbed surface oxygen (NO + 2O
=
3
+
−
e is formed by the oxidation of Mn in the catalyst; i.e.,
NO3 + 3e), which may be formed during the primary NO
decomposition. In fact, N2 but not O2 formation was ob-
served around 473 K in NO-TPD (Fig. 13). This may suggest
3
+
4+
4+
Mn = Mn +e. Considering that Mn is stable in a per-
−
ovskite lattice, NO3 may also be related to the oxidation of
3
+
4+
2−
Mn to Mn
.
that partial NO decomposition (2NO+4e = N2 +2O ) oc-
On the other hand, a fairly large amount of O2 desorp-
tion was observed at temperatures higher than 673 K. Be-
cause it is well known that the lattice oxygen also desorbs
in the same temperature range as LaMnO3-based oxide [17],
the origin of desorbed oxygen was studied with 18O2 ad-
sorption. Fig. 14 shows the oxygen desorption profiles of
curs on this catalyst at decreased temperature. If gaseous NO
adsorbs on the surface oxygen which is formed by partial
NO decomposition, the overall reaction could be written as
disproportionate adsorption of NO. This may be a more rea-
−
sonable mechanism for the formation of NO3 . NO direct
decomposition occurs at temperatures higher than 873 K, as
shown in Fig. 9, and at the same temperature, desorption
of NO and O2 occurs. Therefore, it seems likely that NO
decomposition occurs after removal of NO3 species from
the surface, which is also related to the thermal reduction
1
8
La0.7Ba0.3Mn0.8In0.2O3 when O2 is used for the adsor-
bent. 18 18O desorption is observed around 373 K and in
the temperature range 673 to 873 K. In the same tempera-
O
−
ture range, 16
O
17
O desorption also occurs. Therefore, des-
of Mn4 to Mn . On the other hand, the In-doped cata-
lyst exhibits much larger NO desorption amounts in low and
high temperature ranges. Therefore, NO adsorption is also
improved by In doping, because the BET surface area of
both catalysts is almost the same, as shown in Fig. 2. Doped
In has positive effects in improving NO adsorption by serv-
ing the adsorption site, which might be an oxygen vacancy.
In addition, In doping has also positive effects in weakening
oxygen adsorption. Therefore, the NO decomposition activ-
ity of La0.7Ba0.3MnO3 is greatly increased by In doping at
the Mn site.
+
3+
orption of oxygen around 873 K can be assigned to the
adsorbed oxygen or surface weakly bonded oxygen. It is
also considered that a large part of oxygen adsorbing on
La0.7Ba0.3Mn0.8In0.2O3 catalyst is dissociative, and recom-
bination between the adsorbed and the lattice oxygen easily
1
8
16
occurs during desorption. However, desorption of
O
O as
well as 18
O
18
O is almost negligible, while that of O 16O is
16
still large at temperatures higher than 973 K. Therefore, oxy-
gen desorption above 973 K could be assigned to the lattice
oxygen which may be desorbed by reduction of Mn3
+
.
Because the amount of O2 desorbed is smaller in NO-
TPD than that in O2-TPD and the desorption of adsorbed
oxygen is complete at 973 K, the O2 desorbed during NO-
3.6. N2O and NO2 direct decomposition reaction as
a model reaction
−
TPD seems to form through decomposition of NO3 species
−
into NO and O2 (NO3 = NO + O2 + e) and not through
simple desorption of the lattice oxygen. Here, the formed
Although the formation of N2O was not detected in
the NO decomposition reaction on La0.7Ba0.3Mn0.8In0.2O3,
IR measurement suggests formation of N2O after expo-
sure to NO at temperatures lower than 673 K. To con-
firm the contribution of N2O as an intermediate in di-
rect NO decomposition, direct decomposition of N2O on
La0.7Ba0.3Mn0.8In0.2O3 catalyst was also investigated in
this study. The temperature dependence of N2O decompo-
sition activity on La0.7Ba0.3Mn0.8In0.2O3 catalyst is shown
in Fig. 15. It is seen that this La0.7Ba0.3Mn0.8In0.2O3 cat-
alyst is highly active for N2O decomposition as expected
and the complete decomposition of N2O is achieved at tem-
peratures as low as 773 K. In addition, the amount of N2
and O2 formed is similar to that of N2O converted. Fig. 16
shows the effects of coexisting oxygen on N2O decompo-
sition activity. Although the activity of N2O decomposition
decreased, the temperature for complete decomposition in-
creased with the concentration of oxygen. However, N2O
decomposition on La0.7Ba0.3Mn0.8In0.2O3 still proceeded
under the coexistence of 10% O2, and 100% N2O decompo-
sition was achieved at 900 K. Consequently, if N2O forms,
the decomposition of N2O occurs quickly on this catalyst,
and in NO decomposition in the fixed bed reactor, forma-
4
+
electron is also consumed by the reduction of Mn
Mn , which is the opposite reaction for the formation of
NO3 .
Considering the IR measurements, desorption of NO at
three temperatures could be assigned to the molecularly ad-
to
3
+
−
−
sorbed NO (may be weakly or strongly adsorbed NO ) and
18
Fig. 14. O isotopic TPD profile of La Ba Mn In
O
after
O
2
2
0.7 0.3
0.8 0.2
3
adsorbs at 873 K for 0.5 h.
2 2
tion of N O is not observed. However, N O is considered

112
T. Ishihara et al. / Journal of Catalysis 220 (2003) 104–114
Fig. 17. Temperature dependence of yield of N and O in NO and NO
2
2
2
direct decomposition on La Ba Mn In
balance; W/F = 3 gcat s cm ).
O
(NO or NO : 1%; He:
3 2
0
.7 0.3
0.8 0.2
Fig. 15. Temperature dependence of N O decomposition activity on
2
−3
La Ba Mn In
O
catalyst (N O: 10%; He: balance; W/F =
0
.7 0.3
0.8 0.2
3
2
−
3
3
gcat s cm ).
N2 yield is much smaller than that in NO direct decomposi-
tion. For the selective reduction of NOx with hydrocarbon,
the N2 yield is higher for NO2 than that for NO [1,19–21].
Therefore, selective reduction of NO is considered to occur
through NO2 or NO2 derivatives. Some similarity in reaction
mechanism is considered between the selective reduction
and the direct reduction in the case of Cu-ZSM-5. However,
in the case of La0.7Ba0.3Mn0.8In0.2O3, it is apparent that the
direct decomposition of NO on this catalyst does not pro-
−
ceed through NO2 or NO2 derivatives which may be NO2
−
or NO3 . Lunsford and co-workers also reported that ionic
intermediates such as NO3 are highly stable and could ac-
−
count for the negative influence on NO direct decomposition
on Ba/MgO catalyst with the coexistence of oxygen [8].
−
Because IR measurement suggests that the NO3 species
−
is a strong adsorption species, removal of NO3 and/or
surface oxygen from the active site seems to be the most
important step in the NO decomposition reaction over this
LaMnO3-based perovskite oxide. Considering the strong ad-
Fig. 16. Temperature dependence of N O decomposition activity on
2
−
sorption of oxygen and NO3 , it is quite reasonable that the
La Ba Mn In O catalyst under coexistence of oxygen (N O: 10%;
0
.7 0.3
0.8 0.2
3
2
−3
−
He: balance; W/F = 3 gcat s cm ).
removal of oxygen or NO3 to form the vacant adsorption
site is the important initial step in the NO direct decom-
position reaction. Once a vacant adsorption site is formed,
adsorption of NO smoothly occurs and NO decomposes into
N2 and O2 through N2O.
the intermediate species because decomposition of N2O is
highly active on La0.7Ba0.3Mn0.8In0.2O3. It is also noted that
Mn2O3 is active in N2O decomposition [18].
−
DRIFTS measurement suggests that NO3 species forms
−
3
.7. Reaction mechanism from kinetic analysis
at elevated temperature. To evaluate the possibility of NO3
as an intermediate species, the NO2 decomposition reaction
−
was further investigated in this study. This is because NO3
The reaction steps expected from the DRIFTS and TPD
species forms more easily from NO2 than from NO. Fig. 17
shows the temperature dependence of yield of N2 in NO2 di-
rect decomposition on La0.7Ba0.3Mn0.8In0.2O3. In this figure
N2 yield in NO decomposition is also shown for compar-
ison. It can be seen that NO2 direct decomposition occurs
on La0.7Ba0.3Mn0.8In0.2O3 catalyst at temperatures higher
than 873 K, and O2 formation is also observed. However,
data are as follows:
2
2
NO + 2[s] = 2NO[s],
NO[s] = N2O[s] + O[s],
(1)
(2)
(3)
(4)
N2O[s] = N2 + O[s] (fast step),
2O[s] = O2 + 2[s].

T. Ishihara et al. / Journal of Catalysis 220 (2003) 104–114
113
Here [s] is the surface adsorption site. Because the activity
observed for NO2 decomposition was lower than that ob-
served for NO, contribution of NO2 in NO decomposition
was not considered in this study. Since the desorption of
NO occurs easily at low temperatures, it is reasonably ex-
pected that the adsorption of NO expressed by Eq. (1) is the
rate-determining step in NO decomposition. The rate of de-
composition of NO is expressed as follows, provided that the
adsorption of NO [Eq. (1)] is the rate-limiting step:
The observed P NO and P O dependences of N2 forma-
2
tion rate on La0.7Ba0.3Mn0.8In0.2O3 catalyst are 1.3 and
−0.5, respectively. Therefore, the rate equations (6) and (8)
are almost consistent with the results observed in this study,
suggesting that the adsorption of NO on the active site
and/or the formation of N O is the rate-determining step. At
2
present, we cannot determine from the present data which
step is rate-determining. However, considering the strong
−
adsorption of oxygen and/or NO3 species and the weak ad-
−
→
←−
2
NO
−1
2
sorption of molecular NO, the adsorption of NO seems most
(
k1P − k1K2K3K PN PO )[S0]
4
2
2
rN =
ꢀ
.
(5)
likely to be the rate-determining step on this catalyst. Be-
2
−
1
1/2 1/2
−1/2 1/2 2
−
(
1 + K2K3K4
P
P
+ K4
P )
cause of the weakening adsorption of oxygen and/or NO3
,
N2
O2
O2
In doping at the Mn site is effective for increasing NO de-
composition activity. Doped In does not form an oxygen
vacancy because the valence number of In (+3) is the same
as that of Mn (+3). Therefore, weakening of oxygen and/or
Here, Ki represents the equilibrium constant of each equa-
−→
←−
tion i and k1 and k1 are the rate constants of the forward and
reverse side reactions of NO adsorption [Eq. (1)], respec-
tively. [S0] represents the total number of adsorption sites.
In this equation, the concentration of the surface N2O is as-
sumed to be 0, since reactions (2) and (3) occur quickly.
−
NO3 adsorption species may originate from the improved
mobility of oxygen vacancy in bulk.
Because the partial pressure of N2 is negligible, P N can
2
also be ignored under the experimental conditions. There-
fore, Eq. (5) can be simplified as
4. Conclusion
−
→
2
2
(
k1P )[S0]
NO
rN =
.
(6)
In conventional study, a limited number of catalysts can
proceed in NO direct decomposition. However, this study,
we focused on improving oxygen mobility in bulk by dop-
ing various cations, and we found that In doping is highly
effective for increasing NO direct decomposition activity of
an La(Ba)MnO3 catalyst. In spite of the negligible activity of
In-based oxide in NO decomposition, the fact that NO direct
2
−1/2 1/2
P )
O2
2
(
1 + K4
According to this equation, dependence of NO and O2 partial
pressure is given a value of 2 and a value between 0 and −1,
respectively.
If the surface reaction step of Eq. (2) is the rate-limiting
step, as suggested by Teraoka et al. [10] and Huang et
al. [15], then the rate equation can be expressed as
decomposition of LaMnO improved greatly with In doping
3
ꢁ
ꢃꢁ
ꢂ
is highly interesting from the viewpoint of developing new
catalysts for NO decomposition. This study also revealed
that La0.7Ba0.3Mn0.8In0.2O3 is active in N2O decomposi-
−
→
←−
2
−1 −1
2
rN = k2P − k2K
K
PN PO [S0]
2
NO
3
4
2
2
1
/2
−1/2 1/2
1
+ K1 PNO + K4
P
O2
ꢂ
tion. At present, removal of N O is another important subject
−1/2
1/2 2
2
−1
+
K
K
PN P
.
(7)
3
4
2
O2
from an environmental perspective because N O is one of
2
Again, this equation can be simplified by assuming a small
surface concentration of N2O and small partial pressure of
N2 in the gas phase:
the greenhouse gases and mainly exhausted in the medical
field [18,22]. The developed La0.7Ba0.3Mn0.8In0.2O cata-
3
lyst is expected to be applicable for N O removal.
2
−
→
←−
The rate-limiting step for NO decomposition on this cata-
2
2
(
k2P )[S0]
NO
rN =
.
(8)
lyst may be the adsorption of NO at the vacant site. Since the
2
1/2
−1/2 1/2
P )
O2
2
−
(
1 + K1 PNO + K4
adsorption site is occupied by oxygen and/or NO3 species
up to high temperatures, removal of oxygen from the sur-
face is important. Because the doped In is highly effective at
improving the mobility of oxygen vacancy in bulk, In dop-
ing In at the Mn site of LaMnO3 is considered effective in
increasing NO decomposition activity.
According to this equation, dependence of NO and O2 partial
pressure is given a value between 0 and 2 and between 0 and
−
1, respectively.
On the other hand, if the desorption of O2 [Eq. (4)] is
the rate-limiting step, which is also another reasonable rate-
limiting step considering the results of TPD, the decomposi-
tion rate is expressed as
−
→
←−
Acknowledgment
−
1
−1
−1
√
2
2
(
k4K
K
K3P
P
− k4PO )[S0]
1
2
N2 NO
2
rN =
.
(9)
2
1 + K1−1/2PNO + K1K2K3P
Assuming that the partial pressure of N2 is negligible, the
rate of N2 formation depends on the P NO power of a value
between 0 and −2 and P O between 0 and −1.
−1/2
2
(
PNO)
N2
Part of this study was financially supported by a Grant-
in-Aid for Science Promotion from the Ministry of Educa-
tion, Culture, Sports, Science, and Technology of Japan (No.
11102006).
2

114
T. Ishihara et al. / Journal of Catalysis 220 (2003) 104–114
References
[11] D. Klvana, J. Kirchnerova, C. Tofan, Korean J. Chem. Eng. 16 (1999)
70.
12] T. Ishihara, H. Matsuda, Y. Takita, J. Am. Chem. Soc. 16 (1994) 3801.
13] T. Mori, H. Yamamura, H. Ogino, H. Kobayashi, T. Mitamura, J. Am.
Ceram. Soc. 77 (1994) 2771.
4
[
[
[
[
[
[
1] M. Iwamoto, Shokubai 37 (1995) 614.
2] S. Kagawa, Y. Teraoka, Hyomen 31 (1993) 913.
3] H. Hamada, Shokubai 33 (1991) 320.
4] M. Iwamoto, H. Yahiro, S. Shundo, Y. Yu-u, N. Mizuno, Appl. Ca-
tal. 69 (1991) L15.
[
[
[
14] K. Tabata, M. Misono, Catal. Today 8 (1990) 175.
15] S.J. Huang, A.B. Walters, M.A. Vannice, J. Catal. 192 (2000) 29.
16] M. Mchida, M. Uto, D. Kurogi, T. Kijima, J. Mater. Chem. 11 (2001)
[5] M. Iwamoto, H. Yahiro, K. Tanda, N. Mizuno, Y. Mine, S. Kagawa,
9
00.
17] H.M. Zhang, N. Yamazoe, Y. Teraoka, J. Mater. Sci. Lett. 8 (1989)
95.
18] T. Yamashita, A. Vannice, J. Catal. 161 (1996) 254.
J. Phys. Chem. 95 (1991) 3727.
[
[6] Y.F. Chang, J.G. McCarty, J. Catal. 178 (1998) 408.
[7] M.A. Vannice, A.B. Walters, X.J. Zhang, J. Catal. 159 (1996) 119.
[8] S. Xie, M.P. Rosynek, J.H. Lunsford, J. Catal. 188 (1999) 24.
[9] Y. Teraoka, H. Fukuda, S. Kagawa, Chem. Lett. (1990) 1.
9
[
[19] D.B. Lukyanov, G. Sill, J.L. d’Itri, W.K. Hall, J. Catal. 153 (1995) 265.
[20] T. Ishihara, M. Kagawa, F. Hadama, Y. Takita, J. Catal. 169 (1997) 93.
[21] Y. Li, N. Armor, J. Catal. 150 (1994) 376.
[10] Y. Teraoka, T. Harada, S. Kagawa, J. Chem. Soc., Faraday Trans. 94
(1998) 1887.
[22] Y. Yuzaki, T. Tarimizu, S. Ito, K. Kunimori, Catal. Lett. 47 (1997) 173.