Partial oxidation of methane over a Ni/BaTiO3 catalyst prepared by solid phase crystallization

Article abstract of DOI:10.1039/a701383c

Ni/BaTiO3 catalyst has been prepared by solid phase crystallization (SPC) and used successfully for partial oxidation of CH4 into synthesis gas at 800 °C. In the SPC method, Ni/BaTiO3 catalyst is obtained in situ by the reaction of starting materials with nickel species homogeneously incorporated in the structure. For the starting reagents, materials of two compositional types were employed i. e., perovskite structures BaTi1_JtNi. tO3_i (0

Full text of DOI:10.1039/a701383c

View Article Online / Journal Homepage / Table of Contents for this issue
Partial oxidation of methane over a Ni/BaTiO catalyst prepared by
3
solid phase crystallization
Ryuji Shiozaki,a ArnÐnn G. Andersen,b Takashi Hayakawa,c Satoshi Hamakawa,c
Kunio Suzuki,c Masao Shimizuc and Katsuomi Takehirac*¤
a Chemical Research Department, Institute of Research and Innovation, T akada 1201, Kashiwa,
Chiba 277, Japan
b Nycomed Imaging, P.O. Box 4220 T orshov, 0401 Oslo, Norway
c National Institute of Materials and Chemical Research, T sukuba Research Center, AIST ,
Higashi 1-1, T sukuba, Ibaraki 305, Japan
Ni/BaTiO catalyst has been prepared by solid phase crystallization (SPC) and used successfully for partial oxidation of CH into
3
4
synthesis gas at 800 ¡C. In the SPC method, Ni/BaTiO catalyst is obtained in situ by the reaction of starting materials with
3
nickel species homogeneously incorporated in the structure. For the starting reagents, materials of two compositional types were
employed i.e., perovskite structures BaTi Ni O
1~x 3~d
BaTi O , with 0.3NiO. The starting materials were tested for oxidation of CH by increasing the reaction temperature from
(0 O x O 0.4) and stoichiometric structures, BaTiO , Ba TiO and
x
3
2
4
5
11
4
room temperature to 800 ¡C. The catalysts showed the highest activity for synthesis gas formation around 800 ¡C, and the highest
value was obtained at a composition of x \ 0.3 in the former catalysts. Among the latter catalysts, the highest activity was
observed over BaTiO É 0.3NiO, which was more active than BaTi Ni
O
. On the both catalysts, nickel species originally
3
0.7 0.3 3~d
incorporated in the structure were reduced to the metallic state during the reaction. The BaTiO É 0.3NiO catalyst was further
3
tested for 75 hours at 800 ¡C with no observable degradation and negligible coke formation on the catalyst. Thus, Ni/BaTiO
3
prepared in situ from the perovskite precursor, i.e., by the SPC method, was the most active and resistant to coke formation and
deactivation during the reaction. This may be due to well dispersed and stable Ni metal particles over the perovskite, where the
nickel species thermally evolve from the cations homogeneously distributed inside an inert perovskite matrix as the precursor.
lyst deactivation. Coke formation has been frequently ignored
in the literature,1h14 except in the case of Claridge et al.,16
who observed that the relative rate of coke formation follows
Introduction
Recently, intensive study has been carried out on the catalytic
partial oxidation of CH to synthesis gas.1h13 This process
has advantages over the conventional steam reforming of CH
to make synthesis gas, as the latter process is highly endo-
thermic and produces synthesis gas having a H /CO ratio P3.
The partial oxidation of CH is mildly exothermic on the
the order Ni [ Pd A Rh, Ru, Pt, Ir. Nickel catalysts are
4
highly e†ective for partial oxidation of CH to synthesis gas,
4
4
but they are unsatisfactory with respect to coke formation.17
Slagtern and Olsbye18 reported the in situ formation of and
2
the high activity of metal species for synthesis gas production
4
whole process and a†ords synthesis gas having a H /CO ratio
of ca. 2, making methanol synthesis an ideal follow-up
in LaRhO , which kept high activity for 120 h on stream at
2
3
800 ¡C and showed highly dispersed Rh metal.
process. Ashcroft and co-workers1h3 reported that the tran-
sition metals, Ni, Ru, Rh, Pd, Ir and Pt, supported on alumina
were active in this reaction at temperatures of 775 ¡C. Schmidt
and co-workers4h6 reported that Pt and Rh surfaces in a
A high degree of dispersion of metal species over the cata-
lyst may reduce coke formation.15 Metal-supported catalysts
are conventionally prepared by wet impregnation of di†erent
supports. This method is not fully reproducible and may give
rise to some inhomogeneity in the distribution of the metal on
the surface; therefore, the interest in preparation methods able
to produce homogeneous distributions of metal inside in the
structure of the precursors, which, on further calcination and
reduction, give rise to well dispersed and stable metal par-
ticles. Here, we named this method “solid phase
crystallizationÏ (SPC). A similar idea has been proposed by
Basini and co-workers19 in the preparation of new
hydrotalcite-type anionic clays containing noble metals as the
metal-coated ceramic monolith a†orded mostly H and CO
2
with almost complete conversion of CH and O with short
4
2
residence times. Some transition metal oxides combined with
alkaline earth or rare earth metal oxides, i.e. CoOÈMgO,
NiOÈCaO, CoOÈYb O or Ni/Yb O , were interestingly
found to be catalytically active at low temperatures below
700 ¡C,7h10 but hot spots have been observed in the catalyst
bed.11h13 Lunsford and co-workers14 reported that the partial
oxidation of CH over Ni/Al O involves Ðrst the oxidation
2
3
2 3
4
2 3
of a part of CH to H O and CO , followed by the reforming
precursor for a catalyst for the partial oxidation of CH to
4
4
2
2
4
reactions of CH with H O and CO .
synthesis gas.
2
2
In the steam reforming of CH to synthesis gas, deactiva-
tion of supported metal catalysts by coke formation is a
From the industrial viewpoint, Ni is preferable as the active
species compared to an expensive precious metal such as Rh,
Pd or Ru. We have reported sustainability and high activity of
Ni/(Ca, Sr)TiO in the oxidation of CH to synthesis gas,20h23
4
serious problem.15 Actually in the industrial process, a large
amount of steam was added in order to avoid coke formation.
3
4
In the partial oxidation of CH to synthesis gas, coke forma-
where the Ni catalyst supported on (Ca, Sr)TiO perovskite
4
3
tion over the catalyst frequently takes place, resulting in cata-
was prepared by the SPC method. However, the precursor
was not homogeneous and contained two types of nickel
species; one was dissolved in the Ti site in (Ca, Sr)TiO and
the other was separated as NiO from the perovskite structure.
¤ Present address: Department of Applied Chemistry, Hiroshima
University, Kagami-yama 1-4-1, Higashi-hiroshima 739, Japan.
3
J. Chem. Soc., Faraday T rans., 1997, 93(17), 3235È3242
3235

View Article Online
Nickel dissolved in the perovskite structure was reduced to
well dispersed, stable metal species, while the NiO was
reduced to large Ni metal particles during the reaction. The
well dispersed and stable nickel particles seem to be important
for operation at low coke formation on the catalyst during the
using a Shimadzu DTA 50 and TGA 50 containing an electro-
balance. Transmittance IR spectra were recorded by using a
KBr disk and a JASCO FT/IR 7000 spectrometer. X-Ray
photoelectron spectra (XPS) were obtained with a PHI-5000
spectrometer employing Mg-Ka radiation (1253.6 eV) and an
electron Ñood gun to provide charge neutralization of the
non-conducting samples. All binding energy values were refer-
enced to C 1 s (285.0 eV).
reaction.23 BaTiO can contain a higher amount of nickel in
3
the Ti site compared to CaTiO and give rise to more com-
3
pletely dispersed metal on the surface during the SPC prep-
aration. Here, we report the preparation and thermal
evolution of a catalyst containing well dispersed, stable nickel
on the surface, resulting in sustainable catalytic activity for the
Results
partial oxidation of CH .
BaTi Ni O
1—x
testing
, synthesis, characterization and catalytic
4
x
3—d
Experimental
Synthesis of precursors
The samples with the given formula (x \ 0, 0.01, 0.05, 0.1, 0.2,
0.3 and 0.4) were synthesized by the modiÐed Pechini
method.24 X-Ray di†raction patterns of powders after synthe-
Samples of composition BaTi Ni O
, 0 O x O 0.4, were
1~x
x
3~d
sis are illustrated in Fig. 1. The pattern of BaTiO (perovskite)
prepared according to the Pechini patent24 with some modiÐ-
cations of the procedure.25 Ethylene glycol, citric acid, tetra-
isopropyl orthotitanate, barium carbonate and nickel nitrate
were used as starting materials as described before.22,23 More-
over, several model compounds with related stoichiometric
BaÈTiÈO compositions and combined with nickel oxide, i.e.,
Ba TiO É 0.3NiO, BaTiO É 0.3NiO, BaTi O É 0.3NiO and
3
[JCPDS: 5-626] is observed for all the samples. The Ðrst sign
of excess NiO [JCPDS: 4-835] is weakly seen for x \ 0.3, and
is clearly visible for higher nickel concentrations. The absence
of NiO peaks at lower concentrations may be due to solubility
of Ni in the perovskite, as well as crystallites too small to give
a di†raction signal. ReÑections from Ba TiO [JCPDS: 38-
2
4
2
4
3
5 11
1481] can be seen increasing with the nickel concentration
TiO É 0.3NiO, were also prepared in the same way. The solu-
2
from x \ 0.2. Ba TiO can be formed in the presence of excess
Ba on A sites beyond the point of saturation of Ni on B sites.
After the catalytic testing, the XRD patterns still show
tion obtained from the starting materials was evaporated at
2
4
80È90 ¡C to form a viscous liquid, followed by a two-step
decomposition by heating at 200 ¡C for 5 h and 500 ¡C for 5 h
and Ðnally calcination at 850 ¡C in air for 10 h. Only
BaTiO to be the main constituent of the catalyst, as illus-
3
trated in Fig. 2. In these di†ractograms we see neither NiO
TiO É 0.3NiO was Ðnally calcined at 955 ¡C for 1 h to ensure
2
nor nickel metal [JCPDS: 4-850], except a small trace of NiO
at the highest nickel concentration (x \ 0.4). The second
phase in the di†ractograms is BaCO [JCPDS: 5-378], which
the formation of the rutile crystal structure.26 Cooling was
slowly performed in air by turning o† the furnace. Ni/c-Al O
2
3
and Ni/a-Al O catalysts were prepared by a conventional
3
2
3
is seen to increase with the original nickel content of the
impregnation method using nickel nitrate and each form of
Al O , and in situ reduction during the catalytic testing cited
2
3
below.
Catalytic testing
The powders produced by the above method were tested in a
U-shaped Ðxed bed reactor in a mixture of CH (1.0 l h~1)
4
and air (2.5 l h~1). The catalyst bed contained 300 mg of cata-
lyst material in 2 ml of quartz wool to avoid sintering and
clogging of the reactor.
The catalytic experiments were started by introducing CH
and air at room temperature, and carried out by increasing
4
the reaction temperature from room temperature to 800 ¡C at
a rate of 2.5 ¡C min~1. If not speciÐcally mentioned in the
text, gas samples for analysis of products were taken after 30
min equilibration at each temperature. Moreover, the reac-
tions for testing the catalyst life were carried out at 800 ¡C for
75 h under the same conditions. Product gases from catalytic
testing were analysed by a Shimadzu GC-8A gas chromato-
graph equipped with TCD detector. The selectivities to C
2
compounds, CO , CO and H were calculated based on the
2
2
numbers of carbon and hydrogen atom in CH .
4
After the tests, the catalytic experiments were stopped by
exchanging the reaction mixture with nitrogen before opening
the furnace and cooling the reactor.
Characterization of catalysts
The powder X-ray di†raction (XRD) patterns of catalysts were
recorded by using a MXP-18 di†ractometer (MAC Science
Co.) with Cu-Ka radiation. Scanning electron microscopy
(SEM) was carried out by using a Hitachi S 800 machine.
Transmittance electron microscopy (TEM) was carried out on
a JEM-2000FX (JEOL) electron microscope with an energy
dispersion spectrometer (EDS: Northern). Surface areas of the
catalysts were measured with a Micromeritics model 2200
instrument. Thermal analyses (TG/DTA) were carried out by
Fig. 1 X-Ray di†ractograms for several compositions of catalysts of
the type BaTi
Ni O
after calcination in air at 850 ¡C. (a) x \ 0,
1~x
x 3~d
(b) 0.01, (c) 0.05, (d) 0.1, (e) 0.2, (f) 0.3, (g) 0.4. (L) BaTiO , (G)
3
Ba TiO , (È) NiO.
2
4
3236
J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
30%) for C products (ethane and ethene). However, the main
2
product was CO . At high nickel content and at 800 ¡C, the
2
catalysts selectively produced synthesis gas. Visual observ-
ation, carbon mass balance of the gaseous products, as well as
TG of used catalysts revealed no or very low (\0.1%) coke
formation during the catalytic testing. Total carbon (&C \
CO ] CO ] unreacted CH ) in the effluent gas was 40.3
2
4
mmol h~1 which agreed well with the value calculated from
the CH gas Ñow, 1.01 h~1, regulated at 30 ¡C by the mass
Ñow controller.
4
Related compounds, synthesis, characterization and catalytic
testing
According to the phase diagram for BaOÈTiO ,27,28 several
2
mixed oxides, Ba TiO , BaTiO , BaTi O , Ba Ti
O
,
2
4
3
3
7
4
13 40
BaTi O and BaTi O , are expected to be formed in the
4
9
5 11
present conditions. We have tried to synthesize these mixed
oxides in the presence of Ni species by controlling the Ba : Ti
ratio using the citrate method mentioned above. Among these
mixed oxides, Ba TiO , BaTiO and BaTi O were obtained
2
4
3
5 11
as a single phase in the presence of NiO at a molar ratio of
1 : 0.3 by the citrate method. TiO É 0.3NiO was also prepared
2
as a related compound.
Fig. 3 illustrates the XRD patterns of the related com-
Fig. 2 X-Ray di†ractograms for several compositions of catalysts of
the type BaTi Ni O after catalytic testing at 800 ¡C in a mixture
pounds before and after catalytic testing. BaTiO É 0.3NiO
1~x
x 3~d
3
of CH and air (1 : 2.5). (a) x \ 0.1, (b) 0.2, (c) 0.3, (d) 0.4. (L) BaTiO ,
showed the XRD pattern of BaTiO and a trace of NiO
4
3
3
(K) BaCO , (È) NiO, (@) Ni metal.
3
material. Ba TiO disappeared after the catalytic testing. NiO
2
4
which was visible in di†ractograms after calcination was negli-
gible after the catalytic testing. We believe a substantial part
of nickel on the surface was reduced to its metallic form
during the catalytic testing. The nickel metal particles are
probably too small to give reasonable signals in XRD (vide
infra).
Reaction with the reactor Ðlled with quartz wool alone (i.e.,
a blank test without catalyst) showed an oxygen conversion of
less than 10% in the whole temperature range of this study.
CH conversion stayed below 1% and CO was the only
4
2
product. Samples of BaTi Ni O
(x \ 0, 0.01, 0.05, 0.1,
1~x
x
3~d
0.2, 0.3 and 0.4) were tested for CH oxidation. Representative
data from the catalytic testing of these materials are given in
4
Table 1. The oxygen conversion was high for all experiments
at 750 and 800 ¡C. The CH conversion increased with tem-
4
perature and with nickel content. For low nickel content and
low temperature, the catalysts showed a certain selectivity (ca.
Table 1 CH oxidation with air over a series of BaTi Ni O
4
1~x
x 3~d
catalysts
conversion (%)
CH
selectivity (%)
catalyst temperature
xa
/¡C
O
C
CO
CO
H
4
2
2
2
2
0
750
800
750
800
750
800
750
800
750
800
750
800
750
800
39.5
40.4
39.7
39.8
37.1
36.5
37.3
75.5
47.8
87.2
88.8
92.2
87.0
91.8
94.5
95.0
95.0
94.9
95.0
95.0
95.0
95.0
94.9
94.9
95.0
94.5
94.9
94.9
33.0 63.2
33.6 62.1
33.0 64.0
31.8 65.0
28.7 69.7
18.1 76.7
20.5 75.7
n.d. 17.3 82.8 83.9
n.d. 20.3 79.7 79.1
n.d. 10.2 89.8 87.9
n.d. 10.6 89.4 89.8
3.8
4.3
3.0
3.2
1.6
5.2
3.8
6.2
7.3
4.7
5.3
3.2
7.7
6.5
0.01
0.05
0.1
0.2
Fig. 3 X-Ray di†ractograms for several catalysts with a structure
related to the present study before and after the catalytic testing at
800 ¡C in a mixture of CH and air (1 : 2.5). (a) Ba TiO É 0.3NiO
0.3
n.d.
n.d. 12.4 87.6 89.3
n.d. 9.0 91.0 89.9
7.7 92.3 91.9
4
2
4
(before testing); (b) Ba TiO É 0.3NiO (after testing); (c)
0.4
2
4
BaTiO É 0.3NiO (before testing); (d) BaTiO É 0.3NiO (after testing);
3
3
(e) BaTi O É 0.3NiO (before testing); (f) BaTi O É 0.3NiO (after
11 11
testing); (g) TiO É 0.3NiO (before testing); (h) TiO É 0.3NiO (after
5
5
a x is the nickel content of the catalyst. 300 mg of catalyst was used,
2
2
and the feed consisted of 1 l h~1 of CH and 2.5 l h~1 of air at 1
testing). (G) Ba TiO ; (K) BaCO ; (L) BaTiO ; (|) BaTi O ; ())
4
2
4
3
3
5
11
atmosphere (CH /O \ 1.9). n.d. \ not determined.
TiO ; (J) NiTiO ; (È) NiO; (@) Ni metal.
2 3
4
2
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
3237

View Article Online
Table 2 Surface metal composition of BaTiO É 0.3NiO before and
during testing in air. The binding energies of the main peaks
3
after the catalytic test determined by XPS
of Ni 3p used for the calculation were 854.8 and 853.4 eV
3@2
and, 861.0 and 860.0 eV for the shake-up satellite peaks, for
content (%)
the sample before and after testing, respectively. The former
values can be assigned to Ni2`, but the latter values are rather
catalyst
Ba
Ti
Ni
high for Ni0. BaTiO É 0.3NiO after testing still shows the
3
BaTiO É 0.3NiO
40.2
39.5
43.5
45.3
46.3
43.5
14.5
14.2
13.0
same amount of surface Ni as before testing. Thus, XPS shows
3
before
after
BaTiO É 0.3NiO
3
strong evidence of Ni metal on the catalyst surface, and more-
over no observation of Ni metal by XRD suggests the forma-
tion of Ðne Ni metal particles with a high degree of dispersion.
The high values of the binding energy of Ni0 observed for the
sample after the testing also suggest the formation of Ðnely
dispersed Ni metal on the catalyst surface.
calculated
before the testing, while neither NiO nor Ni metal was
observed after the testing. In the case of Ba TiO É 0.3NiO,
NiO was weakly seen before testing and disappeared after
testing. When the same catalyst was used for catalytic testing
for 60 min at 800 ¡C, the XRD pattern was completely
2
4
Catalytic data for the related compounds combined with
NiO used in the CH oxidation are summarized in Table 3.
4
BaTiO É 0.3NiO showed higher activity for synthesis gas pro-
3
changed. Perovskite (BaTiO ) was the main compound,
together with BaCO ; no nickel was observed.
duction
than
BaTi Ni
O
(cf.
Table
1).
3
0.7 0.3 3~d
BaTi O É 0.3NiO showed a high yield of CO as well as a
11
small amount of
Ba TiO É 0.3NiO gave CO as the main product, but showed
a relatively high selectivity for C products (approximately the
same as for pure BaTiO perovskite). During testing at 800 ¡C,
3
5
2
BaTi O É 0.3NiO showed the peaks of BaTi O [JCPDS:
C
products at all temperatures.
5
11
5
11
2
35-805]; neither nickel species nor substantial change were
observed in the XRD patterns before or after the testing. The
2
4
2
2
attempt to make TiO É 0.3NiO as a related compound
2
3
resulted in the formation of NiTiO (illumenite) [JCPDS:
the selectivity gradually changed as the test was further
3
33-960] and TiO (rutile) [JCPDS: 21-1276] after calcination
2
extended to 60 min. During this period the catalyst increas-
ingly promoted the production of synthesis gas, but was not
for 1 h at 955 ¡C. After catalytic testing, XRD showed only
rutile and metallic nickel. In this case, Ni metal was clearly
as efficient a catalyst as BaTiO É 0.3NiO. IR Spectra of
3
observed, as in the cases of Ni/(Ca, Sr)TiO and Ni/c-
Ba TiO É 0.3NiO after testing showed a strong broad band
3
2
4
Al O .21h23 The colour of the catalyst changed during the
and a sharp band at ca. 1430 and at 859 cm~1, respectively,
2
3
testing as follows: from grey to black for both
suggesting the formation of BaCO . The same bands, much
weaker, were observed on all catalysts containing Ba before
3
Ba TiO É 0.3NiO and BaTiO É 0.3NiO and from yellow to
2
4
3
black for TiO É 0.3NiO. The pale green colour of
and after testing, probably due to surface carbonate formation
over the perovskite.
2
BaTi O É 0.3NiO did not change during testing, suggesting
5
11
that the nickel may be incorporated in the mixed oxide struc-
ture and cannot be reduced to the metallic state. The colour
change to black strongly suggests the formation of metallic
nickel species, nevertheless XRD measurement did not show
the di†raction lines of Ni metal in the samples of
Ba TiO É 0.3NiO and BaTiO É 0.3NiO after testing.
A comparison between the CH conversion of the four cata-
4
lysts, BaTiO É 0.3NiO, BaTi Ni
O
, Ba TiO É 0.3NiO
3
0.7 0.3 3~d
2
4
and TiO É 0.3NiO, at various temperatures is shown in Fig. 4.
All the curves level out in two areas. The reaction dominating
2
at lower temperatures mainly produces CO and some C
products. The level at higher temperatures is dominated by
2
2
2
4
3
The presence of Ni on the catalyst surface is also evident
synthesis gas production. TiO É 0.3NiO was active for com-
2
from XPS of BaTiO É 0.3NiO before and after catalytic
plete oxidation at the lowest temperature, followed by
3
testing. The compositions of the metals were calculated from
BaTiO É 0.3NiO and then BaTi Ni
0.7 0.3 3~d
gas production, both TiO É 0.3NiO and BaTiO É 0.3NiO
O
. For synthesis
3
the spectra by using the method reported by Penn29 and are
shown in Table 2. The amount of Ni was obtained by
summing Ni0 and Ni2`, since the catalyst bed was composed
of two zones, i.e., the oxidized zone and the reduced zone (vide
infra), and moreover the surface of the Ni metal on the cata-
lyst was possibly oxidized by the exposure of the samples
2
3
showed slightly higher total conversions, but the di†erence in
activation temperature may not be signiÐcant.
Ba TiO É 0.3NiO needed a higher temperature or a longer
time to be activated, and the point of conversion after 60 min
at 800 ¡C is included in Fig. 4.
2
4
Table 3 CH oxidation with air over a series of BaÈTiÈNi mixed metal oxide catalysts
4
conversion (%)
CH
selectivity (%)
catalyst
temperature/¡C
O
C
CO
CO
H
4
2
2
2
2
Ba TiO É 0.3NiO
700
750
800
800a
700
750
800
700
750
800
700
750
800
700
750
800
39.7
41.0
38.0
65.8
56.0
94.8
98.0
18.2
30.2
32.7
17.6
94.7
98.1
(89.2)
(93.5)
(97.1)
93.3
94.9
94.9
94.9
94.9
94.9
94.9
53.7
88.3
93.8
55.9
94.9
95.0
34.1
35.6
22.2
0.2
n.d.
n.d.
n.d.
5.2
10.2
11.3
0.5
n.d.
n.d.
È
64.2
62.7
70.0
24.5
44.1
7.5
1.7
1.6
7.8
75.3
55.9
92.5
94.4
12.4
6.4
6.0
n.d.
92.2
94.3
(90.5)
(94.9)
(96.8)
2.8
2.9
2
4
10.4
75.8
99.9
85.3
91.9
19.6
10.8
8.2
BaTiO É 0.3NiOb
3
5.6
BaTi O É 0.3NiO
82.5
83.4
82.7
99.5
7.8
5
11
TiO É 0.3NiO
0.1
2
91.4
92.4
(93.7)
(95.9)
(97.0)
5.6
thermodynamicsc
(100)
(100)
(100)
(9.4)
(5.2)
(3.2)
È
È
300 mg of catalyst was used, and the feed consisted of 1 l h~1 of CH and 2.5 l h~1 of air at 1 atmosphere (CH /O \ 1.9). a Sampled after 60
4
4
2
min. b C products were observed up to 650 ¡C. c Numbers in parentheses show the values calculated from thermodynamic equilibrium of
reactions (1)È(3). n.d. \ not determined.
2
3238
J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
Fig. 4 CH conversion as a function of temperature for the four
4
catalysts in a mixture of CH (1 l h~1) and air (2.5 l h~1) with 300 mg
4
of catalyst. (K) BaTiO É 0.3NiO, (=) BaTi Ni
(>)TiO É 0.3NiO, (…) Ba TiO É 0.3NiO. * After reaction for 1 h at
O
,
3
0.7 0.3 3~d
2
2
4
800 ¡C.
TG was performed on catalysts before and after testing as
illustrated in Fig. 5. All experiments were performed in
Ñowing air (20 ml min~1). Around 400 ¡C a weight increase,
probably due to the oxidation of Ni metal, is observed on
used catalysts, while fresh catalysts show Ñat curves in the
whole temperature range. The weight increases of the used
samples of BaTiO É 0.3NiO and Ba TiO É 0.3NiO amount to
3
2
4
about 0.7% of the total mass of each sample. Assuming oxida-
tion of Ni metal to NiO is the only reaction taking place, this
corresponds to 10% of all the nickel being in the metallic form
after testing. A substantial weight decease observed with
Ba TiO É 0.3NiO after testing above 800 ¡C may be due to
Fig. 6 Conversions and selectivities as a function of time for the
long-term test on (a) BaTiO É 0.3NiO and (b) TiO É 0.3NiO. (L) CH
3
2
4
and (…) O conversion; (=) CO and (K) H selectivity. The tem-
2
2
perature was 800 ¡C, and the feed was a mixture of CH (1 l h~1) and
4
air (2.5 l h~1); the amount of catalyst was 300 mg.
2
4
the decomposition of bulk BaCO . The sample of used
TiO É 0.3NiO showed a weight increase of 4.8%, correspond-
3
testing showed the catalyst to be unstable, as will be shown
below.
2
ing to 60% of the Ni in the reduced form. TiO É 0.3NiO
2
showed lower activity for CH combustion at low tem-
4
Long-term testing of BaTiO Æ 0.3NiO and TiO Æ 0.3NiO
3
2
peratures compared with the other catalysts. This catalyst was
The best performing catalyst, BaTiO É 0.3NiO, in this study
also able to support synthesis gas production, but long-term
3
was subjected to a continuous reaction test over 75 h at
800 ¡C [Fig. 6(a)]. There are no signiÐcant changes in the
values of conversions of CH and O and selectivities for CO,
4
2
H and CO over this catalyst for the period used in the test.
2
2
The selectivity for CO and &C showed almost constant
2
Fig. 5 TG of (…) BaTiO É 0.3NiO, (=) Ba TiO É 0.3NiO and (>)
3
2
4
TiO É 0.3NiO (ÈÈ) before and (È È È) after catalytic testing. The
2
heating rate was 10 ¡C min~1. Air was Ñowing constantly over the
Fig. 7 Temperature programmed oxidation of (…) BaTiO É 0.3NiO
3
and (=) TiO É 0.3NiO after 75 h of the catalytic testing. The heating
2
sample at a rate of 20 ml min~1. The curve for TiO É 0.3NiO after
2
the catalytic testing should be multiplied by a factor of 4 to be com-
rate was 2.5 ¡C min~1. Air was Ñowing constantly over the sample at
parable with the other data.
a rate of 20 ml min~1.
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
3239

View Article Online
values close to 5.0% and 40.3 mmol h~1, respectively. After 75
h of testing, the reactor was Ðlled with nitrogen and cooled
according to normal procedures. Finally, a temperature pro-
grammed oxidation (TPO) experiment was performed by
heating the reactor from room temperature to 950 ¡C at a rate
of 2.5 ¡C min~1, with an air Ñow of 41 ml min~1. O†-gases
in the perovskite under the oxygen-rich atmosphere and the
second is of surface Ni metal under the reducing atmosphere.
It is likely that the reaction is mainly controlled by the ther-
modynamic equilibrium of the reforming reactions [(2) and
(3)] as seen in the values in parentheses in Table 3.
The BaTiO É 0.3NiO catalyst showed the highest activity
3
were analysed as usual and the rate of CO formation is
for synthesis gas production around 800 ¡C, which is probably
2
plotted in Fig. 7. No signiÐcant amount of CO was observed
2
related to the formation of metallic nickel on the surface of the
catalyst. By XRD, trace NiO was observed before catalytic
testing, but no Ni metal was observed after testing. Oxidation
of the used catalysts by TG gave a strong indication of the
presence of metallic nickel (Fig. 5). The presence of Ni on the
surface of the catalyst is also evident from XPS measurements
(Table 2). These facts strongly suggest the formation of Ðnely
during this test with BaTiO É 0.3NiO; CO formation around
3
2
450 ¡C may be produced from coke formed on the catalyst
which corresponded to 0.02 wt.% carbon on the catalyst. The
peak observed above 750 ¡C may be due to the decomposition
of surface BaCO as already seen in the TG curve after the
3
catalytic testing (Fig. 5). The observed amount of CO corre-
2
sponded to the maximum of 3% BaCO in the catalyst after
75 h of catalytic testing.
dispersed Ni metal on the BaTiO É 0.3NiO catalyst during the
3
3
activation. TiO É 0.3NiO also showed a high activity, but was
2
In the CH oxidation over TiO É 0.3NiO [Fig. 6(b)], the
gradually deactivated during long-term testing. Its TG behav-
4
2
catalyst behaved initially like BaTiO É 0.3NiO. At long times,
iour (Fig. 5) is also reÑected in the XRD results (Fig. 3), sug-
gesting that nickel probably forms larger crystals on rutile
than on the barium compounds.
3
however, the CH conversion decreased, as did the selectivities
4
to H and CO; conversely, CO selectivity increased from 5.0
to 7.0% with time. &C (40.3 mmol h~1) showed no signiÐcant
change during the reaction. After 75 h of testing an amount of
2
2
TG of the used sample of Ba TiO É 0.3NiO shows an addi-
2
4
tional feature at temperatures above 800 ¡C, namely a weight
loss of 8.0% (Fig. 5). XRD and IR analyses of this catalyst
CO corresponding to 0.41 wt.% of the catalyst being coked
2
was measured (see Fig. 11, later). The CH oxidation over the
after testing revealed large amounts of BaCO , and we
4
3
other conventional Ni/c-Al O or Ni/a-Al O catalysts under
assume the weight loss corresponds to the thermal decomposi-
2
3
2
3
the same conditions as above has been tested. After 150 h of
reaction, 29.3 and 16.8 wt.% of coke were observed over Ni/c-
Al O22,23 and Ni/a-Al O catalyst, respectively. Compared to
tion of BaCO , releasing CO to the atmosphere. A weight
3
2
loss of 8% indicates that 35% of the catalyst after the testing
is composed of BaCO . Thus, it is likely that Ba TiO is
2
3
2
3
3 2 4
decomposed into BaTiO and BaCO in the atmosphere con-
the Ni/Al O catalysts, TiO É 0.3NiO had a greater resistance
2
3
2
3
3
to coking, and therefore the other deactivation mechanism,
taining CO .
2
i.e., sintering or oxidation of Ni metal, can be considered.
Ba TiO ] CO ] BaTiO ] BaCO
(4)
2
4
2
3
3
After the complete phase transfer from Ba TiO to BaTiO ,
2
4
3
the reduction of surface Ni species may start, resulting in the
activation of the catalyst.
Discussion
As previously mentioned for CH oxidation over Ni/Al O
Scanning electron micrographs (SEM) of Ba TiO É 0.3NiO
4
2 3
2
4
catalyst,14 the present process is also thought to involve Ðrst
before and after catalytic testing are shown in Fig. 8. The
plate-like crystals of the sample before testing [Fig. 8(a)] were
completely changed into a porous structure having many
crevices after testing [Fig. 8(b)]. This can be explained by the
phase transfer from Ba TiO to BaTiO and BaCO [eqn.
the oxidation of part of CH to H O and CO , followed by
the reforming reactions of CH with H O and CO .
4
2
2
4
2
2
CH ] 2O ] CO ] 2H O
(1)
(2)
(3)
4
2
2
2
2
4
3
3
CH ] H O ] CO ] 3H
(4)], followed by a partial decomposition of BaCO at 800 ¡C.
4
2
2
3
BaTiO É 0.3NiO showed no substantial change in the SEM
morphologies during the reaction. TiO É 0.3NiO produced
small particles (\0.1 lm) probably of Ni metal on the rutile
TiO surface suggesting the decomposition of NiTiO to Ni
CH ] CO ] 2CO ] 2H
3
4
2
2
2
Therefore, the catalyst bed may be composed of two zones in
the Ðxed bed reactor; the Ðrst zone is composed mainly of Ni
2
3
Fig. 8 Scanning electron micrographs of Ba TiO É 0.3NiO (a) before and (b) after catalytic testing
2
4
3240
J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
Table 4 Surface area and XRD data of BaÈTiÈNi mixed metal oxide
catalysts
observed phases
surface areaa
catalyst
/m2 g~1
before testing
after testing
Ba TiO É 0.3NiO
2.3
8.9
7.1
0.6
Ba TiO ] NiO
BaTiO ] BaCO
2
4
2
4
3
BaTiO
3
3
BaTiO É 0.3NiO
BaTiO ] NiOb
3
3
BaTi O É 0.3NiO
BaTi O
BaTi O
5
11
5
11
5 11
TiO É 0.3NiO
NiTiO (illumenite) TiO (rutile) ] Ni
2
3
2
] TiO (rutile)
2
a Fresh catalyst. b Traces of NiO peaks were observed.
metal and TiO as seen in the XRD measurements (Fig. 3). A
2
summary of the observations of phase transfer is given in
Table 4 together with measurements of the speciÐc surface
area before testing.
In the present study, XRD measurements simply indicate
that the solubility of Ni is 10È20% on B-sites of BaTiO (Fig.
3
1). The solubility will probably decrease upon lowering the
temperature and the partial pressure of oxygen. In this
respect, Ni may precipitate at the surface during the use of the
catalyst under reducing atmospheres. The precipitation may
produce well dispersed and stable Ni on the surface of the
perovskite, and contribute to the formation of an active and
sustainable catalyst. TEM observation of BaTiO É 0.3NiO
3
after the testing actually shows high dispersion of Ni species
(Fig. 9). Ni is seen as dark spots (A and B in Fig. 9), where
EDS shows the coexistence of Ba and Ti together with Ni. At
site C, Ni is not observed, but Ba and Ti are detected by EDS.
A TEM image of the dark spot at high magniÐcations (Fig.
10) suggests that the spot is composed of an agglomerate of
Ðne Ni particles (diameter of the particle O1 nm). A TEM
Fig. 10 Transmittance electron micrograph of BaTiO É 0.3NiO after
3
testing
EDS. Thus, Ni metal is completely separated from TiO as
2
the support in the case of TiO É 0.3NiO after testing, while
image of TiO É 0.3NiO after testing, for comparison, (Fig. 11)
2
2
BaTiO perovskite can hold Ðne Ni metal particles on the
surface. In the latter case, nickel is obviously distributed as
clearly shows the formation of larger Ni metal particles as
3
black spots (A and B) which are mainly composed of Ni by
Fig. 9 Transmittance electron micrograph of BaTiO É 0.3NiO after
Fig. 11 Transmittance electron micrograph of TiO É 0.3NiO after
3
2
testing
testing
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
3241

View Article Online
2
3
P. D. F. Vernon, M. L. H. Green, A. K. Cheetham and A. T.
Ashcroft, Catal. L ett., 1990, 6, 181.
Ðne particles, and we interpret this in terms of a more favor-
able interaction between nickel and the support. Several
binary oxides of Ba and Ni are known, even in the tem-
perature range of the present study.30h32 Such phases indicate
that there may be an increased “affinityÏ between the nickel
and the support during catalytic action. When nickel is pre-
cipitated at the surface, there will be a surplus of Ba on the
A-sites of the perovskite material. Barium is one of the ele-
ments that are known to keep the level of coking on metals
down. In the present set of experiments, coke was not formed
to any obvious extent.
R. H. Jones, A. T. Ashcroft, D. Waller, A. K. Cheetham and J. M.
Thomas, Catal. L ett., 1991, 8, 169.
4
5
D. A. Hickman and L. D. Schmidt, J. Catal., 1992, 138, 267.
D. A. Hickman, E. A. Haupfear and L. D. Schmidt, Catal. L ett.,
1993, 17, 223.
D. A. Hickman and L. D. Schmidt, Science, 1993, 259, 343.
V. R. Choudhary, S. D. Sansare and A. S. Mamman, Appl. Catal.
A, 1992, 90, L1.
V. R. Choudhary, A. M. Rajput and B. Prabhakar, Catal. L ett.,
1992, 15, 363.
V. R. Choudhary, A. M. Rajput and V. H. Rane, Catal. L ett.,
1992, 16, 269.
6
7
8
9
There are three advantages related to the present
10 V. R. Choudhary, A. M. Rajput and V. H. Rane, J. Phys. Chem.,
1992, 96, 8686.
11 D. Dissanayake, M. P. Rosynek and J. H. Lunsford, J. Phys.
Chem., 1993, 97, 3644.
12 Y-F. Chang and H. Heinemann, Catal. L ett., 1993, 21, 215.
13 S. C. Tsang, J. B. Claridge and M. L. H. Green, Catal. T oday,
1995, 23, 3.
14 D. Dissanayake, M. P. Rosynek, K. C. C. Kharas and J. H. Luns-
ford, J. Catal., 1991, 132, 117.
15 C. H. Bartholomew, Catal. Rev.-Sci. Eng., 1982, 24, 67.
16 J. B. Claridge, M. L. H. Green, S. C. Tsang, A. P. E. York, A. T.
Ashcroft and P. D. Battle, Catal. L ett., 1993, 22, 299.
17 M. Audier, A. Oberlin, M. Oberlin, M. Coulon and L. Bonnetain,
Carbon, 1981, 19, 217.
18 A. Slagtern and U. Olsbye, Appl. Catal. A, 1994, 110, 99.
19 F. Basile, L. Basini, G. Fornasari, M. Gazzano, F. TriÐro and A.
Vaccari, Chem. Commun., 1996, 2435.
20 T. Hayakawa, A. G. Andersen, M. Shimizu, K. Suzuki and K.
Takehira, Catal. L ett., 1993, 22, 307.
21 K. Takehira, T. Hayakawa, A. G. Andersen, K. Suzuki and M.
Shimizu, Catal. T oday, 1995, 24, 237.
22 T. Hayakawa, H. Harihara, A. G. Andersen, A. P. E. York, K.
Suzuki, H. Yasuda and K. Takehira, Angew. Chem., Int. Ed. Engl.,
1996, 35, 192.
23 T. Hayakawa, H. Harihara, A. G. Andersen, K. Suzuki, H.
Yasuda, T. Tsunoda, S. Hamakawa, A. P. E. York, Y. S. Yoon,
M. Shimizu and K. Takehira, Appl. Catal., A, 1997, 149, 391.
24 M. P. Pechini, US Pat., 3 330 697, 1967; M. S. G. Baythoun and
F. R. Sale, J. Mater. Sci., 1982, 17, 257.
25 A. G. Andersen, T. Hayakawa, T. Tsunoda, H. Orita, M. Shimizu
and K. Takehira, Catal. L ett., 1993, 18, 37.
26 Gmelin Handb. Anorg. Chem., 1951, 41, 228.
27 D. E. Rase and R. Roy, J. Am. Ceram. Soc., 1955, 38, 102.
28 J. J. Ritter, R. S. Roth and J. E. Blendell, J. Am. Ceram. Soc.,
1986, 69, 155.
BaTiO É 0.3NiO catalyst prepared by the SPC method. (1)
3
Precipitation from solid solution gives the optimum disper-
sion of Ni on the surface. The main fraction of Ni is dissolved
in the lattice during calcination, and is precipitated during
catalytic testing. (2) Ni has a greater “affinityÏ for B-sites in the
perovskite than for most traditional supports, like Al O and
2
3
TiO . This results in a better dispersion and a low driving
2
force for metal sintering. (3) Low coking is traditionally
achieved by promoting catalysis with alkaline or alkaline
earth metals. Abundant barium probably acts in this manner.
The main problem connected with the practical use of this
kind of perovskite catalyst in an operating plant is probably
the low surface area and the tendency to form Ðne particles
during the catalytic action. This must be solved by using a
binder or a certain support, like a-Al O . The latter study has
2
3
just been undertaken in our laboratory.
Conclusion
Several compositions of BaTi Ni O
related compounds have been prepared and tested as catalysts
, BaTiO É NiO and
1~x
x
3~d
3
for the partial oxidation of CH to synthesis gas at 800 ¡C.
4
Ni/BaTiO catalyst prepared by solid phase crystallization
3
(SPC) showed the highest activity as well as the highest sus-
tainability against coke formation which may be due to well
dispersed, stable metallic nickel particles on the BaTiO
surface. Testing this catalyst continuously over 75 h showed
3
no signiÐcant coking or other changes in the reaction pat-
terns.
29 D. R. Penn, J. Electron Spectrosc. Relat. Phenom., 1976, 9, 29.
30 J. J. Lander, J. Am. Chem. Soc., 1951, 73, 2450.
31 J. J. Lander and L. A. Wooten, J. Am. Chem. Soc., 1951, 73, 2452.
32 H. Krischner, K. Torkar and B. O. Kolbesen, J. Solid State
Chem., 1971, 3, 349.
The authors wish to thank Dr. H. Yokokawa of NIMC for
helpful discussions.
References
1
A. T. Ashcroft, A. K. Cheetham, J. S. Foord, M. L. H. Green,
C. P. Grey, A. J. Murrell and P. D. F. Vernon, Nature (L ondon),
1990, 344, 319.
Paper 7/01383C; Received 27th February, 1997
3242
J. Chem. Soc., Faraday T rans., 1997, V ol. 93