Role of support in CO2 reforming of CH4 to syngas over Ni catalysts

Article abstract of DOI:10.1006/jcat.1996.0280

The dynamic changes in activity and selectivity for the CO₂ reforming of CH₄ over unsupported and supported Ni catalysts were investigated. The reduced NiO has low initial conversions of CH₄ and CO₂, which decrease rapidly with time. The mixture of powders of NiO and Al₂O₃ has low initial conversions of CH₄ and CO₂ and a low yield to CO, which, however, increase with reaction time, due most likely to Ni migration to Al₂O₃. In contrast, the reduced 1 wt% Ni/Al₂O₃ catalyst (reduced from NiO/Al₂O₃) has high initial conversions and CO yield, which slowly decrease with time due to carbon deposition. The reduced 13.6 wt% Ni/Al₂O₃ catalyst has high initial conversions and CO yield, but, because of carbon deposition, the reactor was completely plugged in 6 h. The reduced 1 wt% Ni/SiO₂ catalyst has moderate initial activity and CO yield, which decay fast with reaction time. The reduced 13.6 wt% Ni/SiO₂ has a high initial activity and CO yield, which decay slowly with reaction time. The reduced 13.6 wt% Ni/TiO₂ has a moderate initial activity and CO yield, which decay with reaction time. Carbon deposition was found to occur over all these catalysts, and the sequence of carbon deposition was: 13.6 wt% Ni/Al₂O₃ > 13.6 wt% Ni/SiO₂ > the reduced mixture of NiO and Al₂O₃(13.6 wt% Ni) > 13.6 wt% Ni/TiO₂ gt; 1 wt% Ni/Al₂O₃ > 1 wt% Ni/SiO₂.

Full text of DOI:10.1006/jcat.1996.0280

JOURNAL OF CATALYSIS 162, 230–238 (1996)
ARTICLE NO. 0280
Role of Support in CO2 Reforming of CH4 to Syngas over Ni Catalysts
Eli Ruckenstein1 and Yun Hang Hu2
Department of Chemical Engineering, State University of New York at Buffalo, Amherst, New York 14260
Received November 29, 1995; revised April 5, 1996
Reaction [3] provides a high CO selectivity and a more
suitable H2/CO ratio. In the CO2 reforming of CH4, the
carbon deposition sequence for various metals, reported by
Rostrup–Nielsen (15), is: Ni ꢂ Pt > Ru. Ashcroft et al. (8)
and Rostrup–Nielsen (16) demonstrated that, over nickel-
supported catalysts and for a CO2/CH4 molar ratio of 1 : 1,
carbon deposition cannot be avoided. The effect of the sup-
port on carbon deposition was also investigated by Tang
et al. (17). Ashcroft et al. (8) suppressed the carbon de-
position in CO2 reforming by replacing Ni with platinum-
group catalysts; this provided a CO yield between 69% for
Ru and 89% for Ir from a stoichiometric feed of CO2 and
CH4. It is, however, worthwhile to develop improved sta-
ble and effective nickel-based catalysts because of the high
cost of the noble metals. Rostrup–Nielsen (18) noted that
carbon-free steam reforming of methane can be achieved
using a partly sulfur-passivated nickel catalyst. However, al-
though sulfur inhibits carbon formation, it also reduces the
activity of the catalyst. Recently, we found that the reduced
NiO/MgO catalysts have high activities and selectivities, as
well as excellent stability (19, 20). In those cases the sup-
port (MgO), with which NiO forms a solid solution, inhibits
carbon deposition.
The dynamic changes in activity and selectivity for the CO2 re-
forming of CH4 over unsupported and supported Ni catalysts were
investigated. The reduced NiO has low initial conversions of CH4
and CO2, which decrease rapidly with time. The mixture of powders
of NiO and Al2O3 has low initial conversions of CH4 and CO2 and
a low yield to CO, which, however, increase with reaction time, due
most likely to Ni migration to Al2O3. In contrast, the reduced 1 wt%
Ni/Al2O3 catalyst (reduced from NiO/Al2O3) has high initial conver-
sions and CO yield, which slowly decrease with time due to carbon
deposition. The reduced 13.6 wt% Ni/Al2O3 catalyst has high initial
conversions and CO yield, but, because of carbon deposition, the
reactor was completely plugged in 6 h. The reduced 1 wt% Ni/SiO2
catalyst has moderate initial activity and CO yield, which decay fast
with reaction time. The reduced 13.6 wt% Ni/SiO2 has a high initial
activity and CO yield, which decay slowly with reaction time. The
reduced 13.6 wt% Ni/TiO2 has a moderate initial activity and CO
yield, which decay with reaction time. Carbon deposition was found
to occur over all these catalysts, and the sequence of carbon depo-
sition was: 13.6 wt% Ni/Al2O3 > 13.6 wt% Ni/SiO2 > the reduced
mixture of NiO and Al2O3(13.6 wt% Ni) > 13.6 wt% Ni/TiO2 > 1
c
wt% Ni/Al2O3 > 1 wt% Ni/SiO2.
ꢀ 1996 Academic Press, Inc.
1. INTRODUCTION
In this article, we present a number of observations re-
garding the dynamic changes in activity and CO yield, dur-
ing the CO2 reforming of methane over Ni unsupported and
supported catalysts with the emphasis on the effect of the
nature of the support.
Syngas constitutes a very important feed in petrochem-
istry. While the dominant commercial method employed to
produce synthesis gas is the steam reforming of methane,
CH4 + H2O → CO + 3H2, 1H = 225.4 kJ/mol, [1]
this process has poor selectivity for CO and a too high
H2/CO product ratio, unsuitable for the methanol and the
Fischer–Tropsch syntheses. Two other processes, namely,
the catalytic partial oxidation of methane (1–7) and CO2
reforming (8–14) can be employed to produce syngas from
methane:
2. EXPERIMENT
2.1. Catalyst Preparation
NiO catalyst preparation. NiO was prepared by the de-
composition of Ni nitrate in air at 800ꢃC for 1.5 h.
NiO-supported catalyst preparation. NiO/Al2O3 and
NiO/SiO2 (containing after reduction 1 wt% Ni) and
NiO/Al2O3, NiO/SiO2, and NiO/TiO2 (containing after re-
duction 13.6wt% Ni) catalystswere prepared byimpregnat-
ing the support with an aqueous solution of nickel nitrate.
The obtained paste was dried at room temperature in air,
then decomposed and calcined at 800ꢃC in air for 1.5 h.
CH4 + 1/2O2 → CO + 2H2, 1H = ꢁ38 kJ/mol. [2]
CH4 + CO2 → 2CO + 2H2, 1H = 247 kJ/mol. [3]
1 To whom correspondence should be addressed.
2 Permanent address: Department of Chemistry, Xiamen University,
Xiamen 361005, People’s Republic of China.
230
0021-9517/96 $18.00
c
Copyright ꢀ 1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

CO2 REFORMING OF CH4
231
2.2. Catalytic Reaction
sis of gases during the O2 pulses at 790ꢃC was carried out
with an on-line mass spectrometer. The amount of carbon
deposited was calculated from the amount of CO and CO2
produced.
The catalytic reaction was carried out under atmospheric
pressure, at 790ꢃC, in a flow system, using a vertical quartz
tube (2 mm inside diameter) as reactor. The catalyst pow-
der (weight: 0.02 g) was held on quartz wool. The molar ra-
tio of the reactant gases, with a GHSV = 60000 cm3/g/h, was
CO2/CH4 = 1 : 1. The catalyst was reduced in H2 (20 ml/min)
at 500ꢃC for 14 h, followed by an increase in temperature to
790ꢃC at a rate of 20ꢃC/min, and also in H2 (20 ml/min). The
analysis of the reactants/products mixtures was performed,
with an in situ gas chromatograph equipped with a Pora-
pak Q column. The yield to CO was calculated from the
expression: [CO/(CO + CO2 + CH4)] ꢄ 100% .
3. RESULTS
3.1. CO2 Reforming Reaction of CH4 over
Unsupported NiO Catalysts
Figure 1 shows that the reduced NiO provides very low
initial conversions and CO yield, which decrease rapidly
with time. The unreduced NiO has high initial conversion
and CO yield, which decrease catastrophically with time.
The reduced mixture of powders of NiO and ꢀ -alumina
(0.2 wt ratio NiO/Al2O3) has very low initial conversions of
CH4 (15% ) and CO2 (19% ) and a low yield to CO (17% )
(Fig. 2). As the reaction time increases, the conversions
and yield increase. After 10 h, they reach maximum values
(about 45, 42, and 40% ), which do not change at longer
times (Fig. 2).
2.3. CO Chemisorption and Temperature-Programmed
Desorption (TPD)
The exposed Ni metal surface area and hence the disper-
sion of the catalyst were measured by CO pulse adsorption
at room temperature, assuming a stoichiometry of 1/1. The
catalyst (0.25 g), held on quartz wool in a vertical quartz
tube (6 mm inside diameter), was reduced in H2 (40 ml/min)
at 500ꢃC for 3 h, followed by an increase in temperature
to 790ꢃC at a rate of 20ꢃC/min, also in H2 (40 ml/min). At
790ꢃC, the reduced catalyst was purged with He (40 ml/min)
for 20 min. Then, the catalyst was allowed to cool at room
temperature. The analysis of gases during the adsorption of
CO at room temperature was carried out with an on-line
massspectrometer (HP Quadrupole, 5971seriesmassselec-
tive detector) equipped with a fast-response inlet capillary
system. For the mechanical mixture of NiO and Al2O3, the
dispersion was measured both before and after the CO2
reforming reaction of CH4.
In contrast, the unreduced NiO catalyst has higher initial
conversions of CH4 and CO2 yield to CO, which first rapidly
decrease and then increase somewhat with time (Fig. 2).
3.2. CO2 Reforming of CH4 over 1 wt% Ni/Al2O3
and 1 wt% Ni/SiO2 Catalysts
Figure 3 shows that the reduced 1 wt% Ni/Al2O3 catalyst
has high initial conversions of CH4 and CO2 (78 and 90% )
and yield to CO (84% ), which decrease slowly as the re-
action time increases. After about 7 h, the conversions of
CH4 and CO2 and the yield of CO become 73, 90, and 81% ,
respectively; after about 30 h, they decrease to 60, 75, and
70% , respectively. Compared with the 1 wt% Ni/Al2O3, the
reduced 1 wt% Ni/SiO2 has lower initial conversions of CH4
and CO2 (52 and 72% ) and yield to CO (62% ). The con-
versions and the yield to CO decrease with time. After 7 h,
they become 35, 40, 37% , respectively. As shown in Fig. 3,
the deactivation of the 1 wt% Ni/SiO2 is faster than that of
the 1 wt% Ni/Al2O3.
After the CO adsorption, a TPD experiment was car-
ried out at a rate of 10ꢃC/min. The analysis of gases during
the desorption of CO was performed continuously with the
on-line mass spectrometer.
2.4. BET Surface Area and Pore Size
Distribution Measurements
The unreduced 1 wt% Ni/Al2O3 and 1 wt% Ni/SiO2 cata-
lysts provide initial conversions of CH4 and CO2 and yield
to CO of about 10% , which increase with time, the increase
being faster for the 1 wt% Ni/Al2O3 catalyst than for the
1 wt% Ni/SiO2 catalyst (Fig. 4).
The surface area and the pore-size distribution were de-
termined by nitrogen adsorption, using a Micromeritics
ASAP 2000 instrument. The surface area was calculated us-
ing the BET method, while the pore-size distribution curve
wasobtained from the adsorption branch ofthe N2 isotherm
by the BJH method (21). The sample was degassed at 200ꢃC
in high vacuum before measurements.
3.3. CO2 Reforming of CH4 over Reduced 13.6 wt%
Ni Supported Catalysts
The 13.6% Ni/Al2O3 reduced catalyst has very high initial
conversions of CH4 and CO2 (91 and 98% ) and yield to CO
2.5. The Determination of Carbon Deposition
The deposited carbon was determined by O2 pulse–MS. (95% ) (Table 1). It was, however, impossible to obtain the
The catalyst (0.02 g) was reduced in H2 and the reaction relationship between the CO yield and reaction time due to
carried out for 6 h, both as described in Section 2.2. Then, the difficulty of maintaining constant feed flow. This hap-
the catalyst was purged with He at 790ꢃC for 1 h. The analy- pened because the pressure drop in reactor increased, due

232
RUCKENSTEIN AND HU
FIG. 1. The relationship between reaction time and (a) CO yield, (b) CH4 conversion, and (c) CO2 conversion over NiO. Reaction conditions:
total pressure, 1 atm; T = 790ꢃC; CH4/CO2 = 1 : 1; GHSV = 60000 cm3/g/h.
to the fast carbon deposition, with reaction time. After 6 h, 3.5. CO Turnover Number and Amount
the reactor was completely plugged. In contrast, this did not
occur over the reduced 13.6 wt% Ni/SiO2 or Ni/TiO2. The
Ni/SiO2 provides very high initial conversions of CH4 and
CO2 (89 and 97% ) and yield to CO (93% ) (Table 1), which
decrease slowly with reaction time (Fig. 5). The Ni/TiO2
catalyst has relatively low initial conversions of CH4 and
CO2 (41 and 67% ) and yield to CO (53% ) (Table 1), which
decrease with reaction time faster than for the Ni/SiO2 cat-
alyst (Fig. 5).
of Carbon Deposited
The initial CO turnover numbers (TON) are listed in
Table 1. The 1wt% Ni/SiO2 hasthe veryhigh initialturnover
number of 1721 sꢁ1. (Because the amount of CO chemisor-
ped is very small, there is, however, a large error in the
determination of the number of active sites.) The sequence
of TON is: 1 wt% Ni/SiO2 > 1 wt% Ni/Al2O3 > 13.6 wt%
Ni/TiO2 > 13.6 wt% Ni/SiO2 >13.6 wt% Ni/Al2O3.
Both the 1 wt% Ni/Al2O3 and Ni/SiO2 catalysts exhibit
low carbon depositions (Table 1). For 13.6 wt% Ni-based
catalysts, the sequence of carbon deposition is: Ni/Al2O3 >
Ni/SiO2 > Ni/TiO2 (Table 1).
3.4. BET Surface Area and Pore Size Distribution
The surface areas of the unreduced catalysts are listed in
Table 1. The 1 wt% and 13.6 wt% Ni/SiO2 have the high
surface areas of 447 and 362 m2/g, respectively, the 1 and
13.6 wt% Ni/Al2O3 catalysts have the intermediate surface
areas of 75 and 63 m2/g, respectively, and the 13.6 wt%
3.6. CO TPD on Reduced Ni Unsupported
and Supported Catalysts
Figure 7 shows that the reduced mixture of NiO and
NiO/TiO2 and pure NiO catalysts have the very low surface Al2O3 (0.2 wt ratio NiO/Al2O3) has a sharp CO peak at
areas of 6 and 3 m2/g, respectively. The pore size varied for 80ꢃC and two wide CO overlapped peaks around 300ꢃC.
13.6 wt% Ni/SiO2 between 20 and 150 A with an average At about 420ꢃC, a CO2 peak appears. The reduced 1 wt%
˚
˚
˚
of 52 A and for 13.6 wt% Ni/Al2O3 between 20 and 1000 A Ni/Al2O3 catalyst has a similar TPD curve: one sharp CO
with an average of 130 A. NiO and 13.6 wt% Ni/TiO2 have peak at 80ꢃC and two wide CO overlapped peaks at a tem-
very few pores (Fig. 6).
perature over 150ꢃC; no CO2 peak was, however, detected
˚

CO2 REFORMING OF CH4
233
FIG. 2. The relationship between reaction time and (a) CO yield, (b) CH4 conversion, and (c) CO2 conversion over NiO catalyst diluted with
Al2O3 via mechanical mixing (0.2 wt ratio NiO/Al2O3). Reaction conditions: total pressure, 1 atm; T = 790ꢃC; CH4/CO2 = 1 : 1; GHSV = 60000 cm3/g/h.
(Fig. 7). In contrast, very small amounts of CO and CO2 reaction, but after reduction, the dispersion was 0.065% ,
were detected for the 1 wt% Ni/SiO2 catalyst (Fig. 7). The while after reaction it was 0.18% . The increase in activity
TPD results also show that the curves for CO desorption with reaction time is obviously due to the increase in the
from 13.6 wt% Ni/Al2O3 and 13.6 wt% Ni/SiO2 (Figs. 8 number of active sites. Two explanations are possible: either
and 9) are very different from those for 1 wt% Ni/Al2O3 new sites, which consist of Ni and C, are generated during
and 1 wt% Ni/SiO2. There are five overlapping CO peaks reaction (22), or Ni atoms are migrating to the surface of
and four overlapping CO2 peaks in the TPD curves for the Al2O3 particles. Because the experiments carried out with
13.6 wt% Ni/Al2O3. This indicates that the increase in Ni reduced NiO particles alone (Fig. 1) show that the activ-
content results in the increase of the CO disproportiona- ity decreases with time, the second explanation is the more
tion. The 13.6 wt% Ni/SiO2 catalyst has a TPD curve similar likely one. It is possible that the carbon deposited on Ni
to that of Ni/Al2O3. The TPD curve of the 13.6 wt% Ni/TiO2 during reaction stimulates the migration of Ni atoms to the
catalyst has two CO peaks at 98 and 200ꢃC, respectively, and Al2O3 surface. Indeed, the deposited carbon has the ten-
one CO2 peak at 240ꢃC (Fig. 10).
dency to diffuse into the bulk of NiO (23), thus weakening
the bonds among the Ni atoms and making the latter more
mobile.
4. DISCUSSION
For both the unreduced pure NiO and the unreduced
mixture of NiO and Al2O3, the initial conversions of CH4
and CO2 and yield to CO are high; they first decrease fast
and then slowly with time (Figs. 1 and 2). This can be ex-
plained as follows: The reduction of NiO by CH4 provides
Ni0 active sites for the reforming process. For this reason,
the initial activity is high. The rapid sintering decreases,
however, the number of active sites, and hence the acti-
vity decreases with time. A more rapid sintering occurs at
4.1. Effect of Al2O3 in Mechanical Mixture with NiO
on CO2 Reforming of CH4
For the reduced mixture of NiO and Al2O3, the initial
conversions of CH4 and CO2 and yield to CO are very low;
however, they increase fast with reaction time, attain maxi-
mum values, and then remain constant (Fig. 1). The number
of Ni sites increased after the reforming reaction. Before

234
RUCKENSTEIN AND HU
FIG. 3. The relationship between reaction time and (a) CO yields, (b) CH4 conversion, and (c) CO2 conversion over reduced 1% Ni/Al2O3 and
1% Ni/SiO2 catalysts. Reaction conditions: total pressure, 1 atm; T = 790ꢃC; CH4/CO2 = 1 : 1; GHSV = 60000 cm3/g/h.
shorter times because then the Ni particles are smaller; this tion indicates that the reduced 1 wt% Ni/SiO2 has a much
explains the rapid decay of activity at shorter times.
smaller Ni metal surface area (0.01 m2/g) than the reduced
1 wt% Ni/Al2O3 (0.11 m2/g) (Table 1). This suggests that
chemical interactions occur between NiO and SiO2 in the
NiO/SiO2 (containing 1 wt% Ni) calcined at high tempera-
tures. de Lange and Vissar (25) suggested as early as 1946
that chemical interactions between silica support and NiO,
with the formation of silicate, is the reason for the difficult
reducibility of silica-supported NiO catalysts. This was later
supported byseveralinvestigations(26–28). The smallnum-
ber ofactive sitesin 1wt% Ni/SiO2 catalyst can be explained
as follows: SiO2 has a very high surface area over which a
small amount of NiO is dispersed. As a result, many of the
NiO molecules are in direct contact with SiO2. The high
temperature (800ꢃC) employed during the catalyst prepa-
ration leads to the reaction between the two, with the gen-
eration of a nickel silicate, which is difficult to reduce. This
4.2. The Interactions between NiO and Al2O3
or SiO2 Support
In contrast to the mechanical mixture of NiO and Al2O3,
the unreduced 1 wt% Ni/Al2O3 catalyst has low initial
activity, which increases slowly with reaction time (Fig. 4).
This occurs because, during reaction, the unreduced 1 wt%
Ni/Al2O3 is reduced by CH4, and the number of Ni0
active sites increases. Comparing Figs. 1, 2, and 4, one
can conclude that the alumina-supported NiO catalyst is
reduced during reaction with greater difficulty than the
unsupported one. This can be attributed to the formation
of a nickel aluminate (24).
While the initial activities are comparable, the increase of
activity is much slower over the unreduced 1 wt% Ni/SiO2
catalyst than over the unreduced 1 wt% Ni/Al2O3 (Fig. 4).
This indicates that the reduction of NiO is more difficult
over 1 wt% Ni/SiO2 than over 1 wt% Ni/Al2O3. Although
the BET measurements show that the unreduced 1 wt%
Ni/SiO2 has a surface area (447 m2/g) much larger than the
unreduced 1 wt% Ni/Al2O3 (71 m2/g), the CO chemisorp-
hinders the reduction of Ni2+
to Ni0. When the NiO content
increases, large NiO particlesare formed, and the fraction of
NiO that does not contact SiO2 can be reduced easily to Ni.
Indeed, the Ni metal surface area of the 13.6 wt% Ni/SiO2
is much greater than that of 1 wt% Ni/SiO2 (Table 1).
Table 1 shows that the reduced 1 wt% Ni/SiO2 has a very
high turnover number (1721 sꢁ1), much larger than that

CO2 REFORMING OF CH4
235
FIG. 4. The relationship between reaction time and (a) CO yield, (b) CH4 conversion, and (c) CO2 conversion over unreduced 1% Ni/Al2O3 and
1% Ni/SiO2 catalysts. Reaction conditions: total pressure, 1 atm; T = 790ꢃC; CH4/CO2 = 1 : 1; GHSV = 60000 cm3/g/h.
of 1 wt% Ni/Al2O3 (230 sꢁ1). Because the reduced 1 wt% deposition is greater over 1 wt% Ni/Al2O3 than over 1 wt%
Ni/SiO2 has a very small number of active sites with very Ni/SiO2.
high turnover number, a smalldecrease in their number dur-
ing reaction, due to carbon deposition or sintering, results
in large decrease in activity. This explains why the deacti-
4.3. Effect of Ni Content on the Activity of Ni-Supported
Catalysts for CO2 Reforming of CH4
vation is faster over the reduced 1 wt% Ni/SiO2 than over
The reduced 1 wt% Ni/Al2O3 catalyst has high initial con-
the reduced 1 wt% Ni/Al2O3 catalyst, although the carbon versions of CH4 and CO2 and yield to CO, which decrease
TABLE 1
The Surface Area, Carbon Deposition, CO Yield, and CO TON of Various Catalysts
BET surface
area (m2/g)
Average pore
Ni surface
area (m2/g)
CH4 initial
conver. (% )
CO2 initial
conver. (% )
CO initial
yielda (% )
CO TONb
(sꢁ1
Cc
(g/gcat)
˚
Catalyst
diameter (A)
)
d
NiO + Al2O3
—
447
75
63
362
8
—
53
130
129
52
0.06
0.01e
0.11
1.76
1.57
0.19
15
52
78
91
89
41
18
72
90
98
97
67
17
62
84
95
93
53
83
1721e
230
15
18
80
3.3 ꢄ 10ꢁ3
1% Ni/SiO2
2.3 ꢄ 10ꢁ4
1% Ni/Al2O3
3.9 ꢄ 10ꢁ4
f
13.6% Ni/Al2O3
13.6% Ni/SiO2
13.6% Ni/ TiO2
1.9 ꢄ 10ꢁ1
8.0 ꢄ 10ꢁ4
—
a The initial yield for GHSV = 60,000 cm3/g/h, T = 790ꢃC, CH4/CO2 = 1 :1.
b CO turnover number.
c Amount of carbon deposited over the catalyst during CO2 reforming of methane at 790ꢃC for 6 h.
d The mixture of NiO and Al2O3 (containing 13.6 wt% Ni).
e The error is very large because the amount of CO chemisorbed is extremely small.
f The carbon deposition was so fast that the reactor was completely plugged in 6 h.

236
RUCKENSTEIN AND HU
FIG. 7. Temperature-programmed desorption spectrum of CO over
Ni from the reduced mixture of NiO and Al2O3, reduced 1 wt% Ni/Al2O3,
and Ni/SiO2 catalysts (—) CO, (- - -) CO2.
FIG. 5. The relationship between reaction time and CO yield over
reduced 13.6% Ni/SiO2 and Ni/TiO2 catalysts. Reaction conditions: total
pressure, 1 atm; T = 790ꢃC; CH4/CO2 = 1 : 1; GHSV = 60000 cm3/g/h.
Even though sintering occurs with some ease, the large
amount of Ni present in the reduced 13.6 wt% Ni/SiO2 pro-
vides a larger number of Ni sites than the 1 wt% Ni/SiO2
catalyst. Indeed, both the CO chemisorption and TPD
curve show that the 13.6 wt% Ni/SiO2 adsorbs a much larger
amount of CO than the 1 wt% Ni/SiO2 (Figs. 7 and 9). Simi-
lar to the 13.6 wt% Ni/Al2O3 catalyst, the 13.6 wt% Ni/SiO2
catalyst exhibits a much higher carbon deposition than the 1
wt% Ni/SiO2 (Table 1). The activity of the 13.6 wt% Ni/SiO2
catalyst does not decay, however, as fast as that of the 1 wt%
Ni/SiO2 catalyst. This happens probably because the initial
number of sites is much larger for the 13.6 wt% catalyst,
and the large amount of carbon deposited partially inhibits
sintering.
Although the TPD CO curves show that the reduced
13.6 wt% Ni/Al2O3 and 13.6 wt% Ni/SiO2 have compa-
rable CO disproportionation abilties, the former catalyst
exhibits at 790ꢃC a much higher carbon deposition than
the latter. This indicates that the CO dispropotionation
does not constitute the main mechanism for carbon de-
position on Ni-based catalysts at 790ꢃC. This observation
slowly with reaction time (Fig. 3). The relatively small
amount of carbon deposited during reaction (Table 1) may
explain the slow decrease in activity with time. Instead, over
the reduced 13.6 wt% Ni/Al2O3 catalyst, the carbon depo-
sition was so fast that the reactor was plugged in 6 h. As
shown by the TPD curves, the decomposition of CO to CO2
is much higher over the reduced 13.6 wt% Ni/Al2O3 than
over the reduced 1 wt% Ni/Al2O3 (Figs. 7 and 8). Therefore,
the increase of CO disproportionation is partly responsible
for the increase in carbon deposition with the increase of
Ni content in Ni/Al2O3 catalysts. In addition, the increase
of Ni content also increases the decomposition of CH4, re-
sulting in an increase in carbon deposition. As shown later,
the latter mechanism is the main one at 790ꢃC.
FIG. 8. Temperature-programmed desorption spectrum of CO over
FIG. 6. The pore size distribution of 13.6 wt% Ni-based catalysts
the reduced 13.6 wt% Ni/Al2O3 catalyst (—) CO, (- - -) CO2.
Ni/SiO2, Ni/Al2O3, and Ni/TiO2.

CO2 REFORMING OF CH4
237
duced mixture of NiO and Al2O3 (13.6 wt% Ni) > 13.6 wt%
Ni/TiO2 > 13.6 wt% Ni/SiO2 ꢅ 13.6 wt% Ni/Al2O3. The
turnover number is much smaller for the catalysts con-
taining a larger amount of Ni. Comparing the Ni/Al2O3 or
Ni/SiO2 catalysts for different Ni loadings, one can conclude
that the reaction occurs more readily on the smaller crys-
tallites, which are more likely to be present in the catalysts
that contain smaller amounts of Ni. The crowding of Ni sites
appears to affect negatively the CO turnover number for
the CO2 reforming of CH4.
5. CONCLUSION
FIG. 9. Temperature-programmed desorption spectrum of CO over
the reduced 13.6 wt% Ni/SiO2 catalyst (—) CO, (- - -) CO2.
The CO2 reforming of CH4 over Ni-supported catalysts
is profoundly affected by the nature of the support. SiO2,
Al2O3, and TiO2 have been used as supports and the ac-
tivities and CO yields compared. For completeness, the
unsupported NiO catalyst and the unsupported NiO cata-
lyst mixed mechanically with Al2O3 were also investigated.
While the reduced 1 wt% Ni/Al2O3 has high initial activity
and selectivity, which decay slowly with reaction time, the
reduced 13.6 wt% Ni/Al2O3 catalyst has high initial activity
and selectivity, but the fast carbon deposition plugsthe reac-
tor in about 6 h. In contrast, 1 wt% Ni/SiO2 has a moderate
initial activity and selectivity, which decay rapidly with re-
action time, while the 13.6 wt% Ni/SiO2 has high initial acti-
vity and selectivity, which decay slowly with reaction time.
The sequence of the initial CO yields of various supports
(all containing 13.6 wt% Ni and prepared via the reduc-
tion ofsupported NiO) is:Ni/Al2O3 > Ni/SiO2 > Ni/TiO2. In
most catalysts, the CO yield decreases with reaction time.
However, interestingly, the activity and selectivity of the
reduced NiO catalyst mixed mechanically with Al2O3 in-
crease rapidly with time at the beginning and moderately
later. The behavior of these catalysts is explained in terms
of sintering, carbon deposition, metal–support interactions,
and, in the case of NiO mixed mechanically with Al2O3, Ni
migration to Al2O3.
is consistent with the thermodynamic conclusion of Clar-
idge et al. (29), regarding the partial oxidation of methane,
that, at low temperatures (<600ꢃC), the CO disproportion-
ation is the main mechanism of carbon deposition, while at
high temperatures (>600ꢃC) the CH4 dissociation consti-
tutes the main mechanism.
4.4. Ni/TiO2
The CO yield is relatively small initially and decreases
quickly with time over the reduced 13.6 wt% Ni/TiO2 cata-
lyst (Fig. 5). The initial yield is relatively small because, dur-
ing reduction, TiOx molecules migrate over the surface of
Ni particles (30). This migration occurs because it leads to a
decrease of the surface free energy of the system. This phe-
nomenon was called “strong metal support interactions”
(30), and, indeed, the strong interactions between TiOx and
Ni are responsible for the decrease of the free energy of the
system (31). The decayofthe yield with time ismostlydue to
carbon deposition. Indeed, carbon deposition was detected
for this catalyst (Table 1).
4.5. CO Turnover Number for the CO2 Reforming of CH4
As shown in Table 1, the sequence of initial turnover
numbers is 1 wt% Ni/SiO2 > 1 wt% Ni/Al2O3 > the re-
REFERENCES
1. Hickman, D. A., and Schmidt, L. D., Science 259, 343 (1993).
2. Ashcroft, A. T., Cheetham, A. K., Ford, J. S., Green, M. L. H., Grey
C. P., Murrell, A. J., and Vernon, P. D. F., Nature 344, 319 (1990).
3. Choudhary, V. R., Rajput, A. M., and Prabhakar, B., J. Catal. 139, 326
(1993).
4. Hu, Y. H., Au, C. T., and Wan, H. L., Chin. Sci. Bull. 40(4), 303
(1995).
5. Dissanayake, D., Rosynek, M. P., Kharas, K. C. C., and Lunsford, J. H.,
J. Catal. 132, 117 (1991).
6. Dissanayake, D., Rosynek, M. P., and Lunsford, J. H., J. Phys. Chem.
97, 3644 (1993).
7. Choudhary, V. R., Rajput, A. M., and Prabhakar, B., Angew. Chem.
Int. Ed. Engl. 33, 2104 (1994).
FIG. 10. Temperature-programmed desorption spectrum of CO over
the reduced 13.6 wt% Ni/TiO2 catalyst (—) CO, (- - -) CO2.
8. Ashcroft, A. T., Cheetham, A. K., Green, M. L. H., and Vernon,
P. D. F. , Nature 352, 225 (1991).

238
RUCKENSTEIN AND HU
9. Richardson, J. T., and Paripatyadar, S. A., Appl. Catal. 61, 293
(1990).
22. Zhang, Z., and Verykios, X. E., J. Chem. Soc., Chem. Commun. 1, 71
(1995).
10. Yamazaki, O., Nozaki, T., Omata, K., and Fujimoto, K., Chem. Lett.
10, 1953 (1992).
23. Zdansky, E. O. F., Nilsson, A., and Martensson, N., Surf. Sci. 310, L583
(1994).
11. Gadalla, A. M., and Sommer, M. E., Chem. Eng. Sci. 44, 2825
(1989).
12. Gadalla, A. M., and Bower, B., Chem. Eng. Sci. 43, 3049 (1988).
13. Rostrup-Nielsen, J. R., and Hansen, J-H.B., J. Catal. 144, 38 (1993).
14. Kim, G. J., Cho, D. S., Kim, K. H., and Kim, J. H., Catal. Lett. 28, 41
(1994).
15. Rostrup-Nielsen, J. R., J. Catal. 33, 184 (1974).
16. Rostrup-Nielsen, J. R., Stud. Surf. Sci. Catal. 36, 73 (1988).
17. Tang, S. B., Qiu, F. L., and Lu, S. J., Catal. Today 24(3), 253 (1995).
18. Rostrup-Nielsen, J. R., J. Catal. 85, 31 (1984).
19. Ruckenstein, E., and Hu, Y. H., Appl. Catal. 133(1), 149 (1995).
20. Hu, Y. H., and Ruckenstein, E., Catal. Lett. 36(3 & 4), 145 (1996).
21. Barret, E. P., Joyner, L. G., and Halenda, P. H., J. Am. Chem. Soc. 73,
373 (1951).
24. Ruckenstein, E., and Lee, P. S., J. Catal. 86, 457 (1984).
25. de Lange, J. J., and Visser, G. H., Ingenieur 58, 24 (1964).
26. Mirodatos, C., Barthomeuf, D., Dalmon, J. A., and Garbowski, E. D.,
Zeolites 2, 125 (1982).
27. Schuit, G. C. A., and van Reijen, L. L., Adv. Catal. 10, 242 (1960).
28. Singley, W. J., and Carriel, J. T., J. Am. Chem. Soc. 75, 778 (1953).
29. Claridge, J. B., Green, M. L. H., Tsang, S. C., York, A. P. E., Ashcroft,
A. T., and Battle, P. D., Catal. Lett. 22, 299 (1993).
30. Stevenson, S. A., Dumesic, J. A., Baker, R. T. K., and Ruckenstein,
E., “Metal–Support Interactions in Catalysis, Sintering and Redisper-
sion.” Van Nostrand, New York, 1987.
31. Ruckenstein, E., in “Strong Metal–Support Interactions” (R. T. K.
Baker, S. J. Tauster, and T. A. Dumesic, Eds.), p. 152. Am. Chem. Soc.,
Symposium Series 298, Washington, DC, 1986.