Decomposition of methane with an autocatalytically reduced nickel catalyst

Article abstract of DOI:10.1016/j.jcat.2005.05.006

Thermal decomposition of nickel acetate (T>300°C) dispersed on either silica or cordierite supports resulted in a mixture of Ni⁰/NiO that was catalytically active for the decomposition of methane to produce CO-free hydrogen, without the need for any pretreatment (i.e., using H ₂ at high temperature (>500°C) for at least 2 h). The carbon yield (g_C/g_Ni) for the catalysts supported on SiO ₂ (～200-300 m2/g) increased with an increase in the catalyst NiO mean crystallite size. XRD and FE-SEM studies confirmed the formation of graphitic-type carbon filaments during the methane decomposition.

Full text of DOI:10.1016/j.jcat.2005.05.006

Journal of Catalysis 233 (2005) 317–326
www.elsevier.com/locate/jcat
Decomposition of methane with an autocatalytically
reduced nickel catalyst
Richard A. Couttenye a, Marianela Hoz De Vila a, Steven L. Suib a,b,c,∗
a
Department of Chemical Engineering, University of Connecticut, Storrs, CT 06269-3222, USA
b
Institute of Material Science, University of Connecticut, Storrs, CT 06269-3136, USA
c
Department of Chemistry, University of Connecticut, Storrs, CT 06269-3060, USA
Received 4 March 2005; revised 5 May 2005; accepted 10 May 2005
Available online 6 June 2005
Abstract
◦
0
Thermal decomposition of nickel acetate (T > 300 C) dispersed on either silica or cordierite supports resulted in a mixture of Ni /NiO
that was catalytically active for the decomposition of methane to produce CO-free hydrogen, without the need for any pretreatment (i.e.,
◦
2
using H at high temperature (> 500 C) for at least 2 h). The carbon yield (g /g ) for the catalysts supported on SiO (∼ 200–300 m /g)
2
C
Ni
2
increased with an increase in the catalyst NiO mean crystallite size. XRD and FE-SEM studies confirmed the formation of graphitic-type
carbon filaments during the methane decomposition.
2005 Elsevier Inc. All rights reserved.
Keywords: Hydrogen production; Methane decomposition; Nickel acetate
1. Introduction
able [1]. In addition, technologies for hydrogen generation
from biomass, such as enzymatic decomposition of sugars,
steam reforming of bio-oils, and gasification, suffer from
low production rates and/or complex processing require-
ments [2].
In contrast, decomposition of methane over supported
metal catalysts (preferably nickel) [3–15] produces very
pure H2 without the formation of carbon oxides, which
eliminates the process of separating the gaseous mixtures.
However, such an advantage is outweighed by the catalyst
deactivation due to carbon deposition on the active sites,
hence requiring regeneration by conventional oxidation in
air or steam gasification [3,4,11]. Catalytic decomposition
of methane is therefore a useful application in some specific
cases, such as hydrogen fuel cells (e.g., proton-exchange
membrane, PEM), where CO-free hydrogen is required to
avoid deactivation of the platinum electrode.
The search for alternative energy sources has been pro-
pelled by the reduction of petroleum reserves around the
world and by pollution caused by the increase in energy
consumption. Compared with other options, such as solar
or wind energy, the use of hydrogen (H2) to satisfy future
energy demand emerges as an attractive choice. Hydrogen,
a clean-burning fuel, has been widely used in petroleum re-
fineries, the production of ammonia, the hydrogenation of
fats and oils, welding, the production of hydrochloric acid,
the electronics industry, fuels for rockets, and recently in
fuel cells. Worldwide, 48% of H2 is produced from natural
gas, 30% from petroleum (mostly consumed in refineries),
18% from coal, and the remaining 4% from water electroly-
sis [1].
Electrolytic production of H2, a highly inefficient process,
is only attractive when low-cost hydroelectric power is avail-
Nickel has been the most common transition metal used
for decomposition, steam reforming, and/or partial oxida-
tion of methane. For instance, Amiridis et al. [3,4] have
been using 15% Ni/SiO2 for the decomposition of methane
to produce CO-free hydrogen, and have found that fila-
*
Corresponding author. Fax: +1 860 486 2981.
E-mail address: suib@uconnvm.uconn.edu (S.L. Suib).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2005.05.006
2005 Elsevier Inc. All rights reserved.

318
R.A. Couttenye et al. / Journal of Catalysis 233 (2005) 317–326
mentous carbon started to grow when attached to a nickel
crystallite, which eventually led to catalyst deactivation.
Dicks [5], Matsukata [6], and Ermakova [7] have also re-
ported the formation of filamentous carbon during the de-
composition of methane with the use of 5% Ni/SiO2 and 90–
96% nickel with different textural promoters (SiO2, Al2O3,
MgO, TiO2, and ZrO2). Froment et al. have comprehensively
studied the filamentous carbon formation and the gasifica-
tion process during methane decomposition [8]. Takenaka et
al. [9,10] have used 40% Ni/SiO2 and 37% (Pd–Ni)/carbon
nanofibers to decompose methane into hydrogen, attain-
ing the highest hydrogen yield among those reported yet
2. Experimental
2.1. Catalyst preparation
2.1.1. Type 1
Celcor cordierite (Mg2Al4Si5O18) monoliths (as re-
ceived) from Corning with square channels and a density of
400 squares/in2 were cut into smaller pieces (L = 20 mm,
D = 9 mm) and dipped in a 0.5 M aqueous solution of
nickel(II) acetate tetrahydrate (Ni(CH3COO)2·4H2O, CAS:
6018-89-9; Aldrich Chemical Company, Milwaukee, WI) or
nickelous nitrate hexahydrate (Ni(NO3)2·6H2O; JT Baker,
Phillipsburg, NJ) for 3 h at room temperature and dried
overnight in air at 120 ◦C. The monoliths were then calcined
in air at 600 ◦C for 2 h. The final nickel content on the mono-
liths was 8 wt%. The samples were denoted Cor-A-8 and
Cor-N-8, respectively (i.e., support–salt–wt%) (A = acetate,
N = nitrate).
(16,000 molH /mol(Pd+Ni)), for the mixture 37% (Pd–Ni)
2
supported on carbon nanofibers. Choudhary et al. [11–13]
have studied the continuous production of H2 from methane
decomposition over Ni-containing metal-oxide (Ni/M =
1.0) and nickel (10%)-impregnated zeolite and SiO2 cata-
lysts. Gronchi [14] synthesized nickel-supported catalysts
on SiO2 and Al2O3 in the form of xerogels, and the in-
fluence of surface differences on the catalytic activity of
the materials was studied with the CH4 reforming reaction.
An extensive review of the multiple roles of catalysis in
the production of hydrogen was made by Armor [1]. Ar-
mor highlights the use of nickel catalyst (∼ 12–20% nickel)
supported on a refractory material, such as alpha alumina
for steam reforming of methane. Recupero et al. [15] also
used commercial Ni/Al2O3 (CRG-F) for steam reforming of
methane.
2.1.2. Type 2
Nickel catalyst supported on SiO2 (Davisil 35-60 mesh,
grade 646, type 150 A, Fisher) was prepared by conven-
tional wet impregnation with a nominal metal loading of 8
and 40 wt%. Nickel(II) acetate tetrahydrate (Ni(CH3COO)2·
4H2O, CAS # 6018-89-9; Aldrich Chemical Company, Mil-
waukee, WI) and nickelous nitrate hexahydrate (Ni(NO3)2·
6H2O; JT Baker, Phillipsburg, NJ) were used as a source
of nickel. After impregnation at room temperature, the cat-
alysts were dried overnight at 120 ◦C and then calcined, in
air, at 600 ◦C for 2 h. These samples were denoted Si-A-8,
Si-A-40, Si-N-8, and Si-N-40, respectively.
However, nickel nitrate has been the preferred nickel salt
in all of these studies. The thermal decomposition of nickel
nitrate (T > 400 ◦C) produces nickel oxide (NiO). To re-
duce the nickel from oxide to nickel metal, pretreatment
of the catalysts is required (with H2 at high temperature
(> 500 ◦C) for at least 2 h). On the other hand, many re-
search studies [16–24] have characterized in detail the ther-
mal decomposition of nickel acetate (Ni(CH3COO)2·4H2O).
There is agreement [21,23,24] that this salt decomposes
at temperatures higher than 250 ◦C through an autocat-
alytic process in which a nickel carbide (Ni3C) intermediate
species is formed. At temperatures higher than 300 ◦C, Ni3C
decomposes completely, giving rise to a mixture of Ni0 and
NiO.
2.2. X-ray diffraction
X-ray diffraction (XRD) analysis was used to identify the
nature of the powder catalyst. XRD data were collected with
a Scintag 2000 XDS diffractometer with Cu-Kα X-ray radi-
ation. Powder samples were placed on aluminum slides and
scanned at 4◦ 2θ/min. The beam voltage and beam current
were 45 kV and 40 mA, respectively.
2.3. Scanning electron microscopy and field emission
scanning electron microscopy
The morphology of the catalyst, before and after reac-
tion, was recorded by scanning electron microscopy (SEM)
and field emission scanning electron microscopy (FE-SEM).
SEM micrographs were taken with an Amray 1810 micro-
scope, and high-resolution micrographs were taken with a
Zeiss DSM982 Gemini FE-SEM microscope. Samples were
mounted on carbon tape onto aluminum sample holders for
analysis.
To the best of our knowledge, the use of this peculiar
mixture Ni0/NiO from an acetate salt for catalytic purposes
(such as the decomposition of methane) has not yet been re-
ported. The present investigation evaluated the application
of the Ni0/NiO mixture, which was obtained after thermal
decomposition of nickel acetate, for CO-free hydrogen pro-
duction by means of decomposition of methane, with the use
of two types of supports: cordierite monolith and SiO2. Cata-
lyst preparation, characterization, and catalytic performance
were reported. The relationship between nickel oxide (NiO)
mean crystallite sizes and catalyst deactivation, due to car-
bon deposition, was also studied.
2.4. Textural properties
The catalyst (types 1 and 2, as prepared) surface ar-
eas were measured, with the use of nitrogen physisorption

R.A. Couttenye et al. / Journal of Catalysis 233 (2005) 317–326
319
at liquid nitrogen temperature with a Micrometrics ASAP
Table 1
Effect of nickel precursor and support on catalysts surface area
2010 System. Before the analysis samples were outgassed at
300 ◦C under vacuum for 2 h. Surface areas were determined
by nitrogen adsorption data, with Brunauer–Emmett–Teller
(BET) methods.
a
2
Catalyst
Nickel precursor
Nickel
(wt%)
BET surface area (m /g)
B1 B2 B3
Type 1
Cor-A-8
Cor-N-8
Ni(CH COO) ·4H O
8
8
5.8
2.9
6.3
2.1
5.1
3.4
3
2
2
Ni(NO ) ·6H O
3 2
2
2.5. Catalytic activity
Type 2
Si-N-40
Si-A-40
Si-N-8
Si-A-8
Ni(NO ) ·6H O
40
40
8
178.5
249.8
273.4
285.5
169.8
230.2
277.2
294.6
173.1
255.7
266.4
280.8
3 2
2
Types 1 and 2 catalysts were tested for the decomposi-
tion of methane with a gas-flow quartz tube reactor (i.d. =
10 mm, length = 30 cm), operating at atmospheric pressure.
The reactor was kept in a tubular electric furnace controlled
by an Omega-4002A temperature controller, connected to a
thermocouple placed at the center of the reactor. Methane
(99% [Airgas]) was fed to the reactor at 40 ml/min and reg-
ulated by a MKS-247C flow controller. The catalysts (type 1:
∼ 0.6 g, or type 2: ∼ 0.040 g) were placed 17 cm from
the reactor entrance with quartz wool to set up the catalyst
bed. The gap between the catalytic bed and the reactor en-
trance was filled with glass beads (4 mm, cat # 11-312B;
Fisher) in order to increase the contact area for preheating of
the feed. Before the reaction started, the catalysts prepared
with Ni(NO3)2·6H2O as the nickel precursor were reduced
with pure hydrogen at 550 ◦C for 2 h, unless otherwise men-
tioned. During the reaction, the product stream was analyzed
by mass spectrometry (MS) with a MKS-UTI PPT quadru-
pole residual gas analyzer. The MS was calibrated before
the experiments with the use of gaseous mixtures of H2 in
CH4 of known concentrations: 5, 10, 15, 20, and 40%. Since
CH4 decomposition on nickel catalysts proceeds selectively
to carbon and H2 [3,7,9,10], methane conversion was calcu-
lated from the hydrogen concentration in the product mix-
ture. The amount of carbon deposited on the catalyst during
reaction was calculated by the weight difference from the
fresh sample. Each experiment was performed three times
with catalysts prepared in different batches (B1–B2–B3).
Ni(CH COO) ·4H O
3
2
2
Ni(NO ) ·6H O
3 2
2
Ni(CH COO) ·4H O
8
3
2
2
a
Sample nomenclature: support–salt–wt%.
The changes in methane conversion with time during the
catalytic decomposition of methane at 550 ◦C, with types 1
and 2 catalysts (from batch B1), are shown in Fig. 1. At the
beginning of the reaction, the catalysts containing 40 wt%
nickel resulted in higher conversion (13–15%) than the cata-
lysts with 8 wt% nickel loading (7–10%). However, the ini-
tial activity of the catalyst prepared with nickel acetate, with-
out reduction pretreatment (Fig. 1, curves 2, 6, and 7), was
very close to the initial activity obtained with the catalysts
that were prepared with nickel nitrate and pre-reduced with
H2 at 550 ◦C for 2 h (Fig. 1, curves 1, 4, 5). The conversion
data for the unreduced Si-N-40 catalyst are also included in
Fig. 1 (curve 3). Clearly, the conversion of methane with
this unreduced catalyst was lower than that obtained with
the pre-reduced Si-N-40 (Fig. 1, curve 1). This difference in
activity illustrated the reduction effect on the catalysts pre-
pared from the nitrate salt and the advantage of using nickel
acetate as a precursor (reduction free).
3.2. XRD characterization
Decomposition of nickel(II) acetate tetrahydrate
(Ni(CH3COO)2·4H2O) in air at 600 ◦C for 2 h resulted in
the formation of a mixture of Ni0/NiO, as indicated by the
XRD pattern shown in Fig. 2a (Ni0: 2θ = 43.7◦, 51.1◦,
75.6◦ JCPDS # 4-850; NiO: 2θ = 36.4◦, 42.5◦, 62.0◦, 78.6◦
JCPDS # 4-835). Thermal decomposition of nickelous ni-
trate hexahydrate (Ni(NO3)2·6H2O) in air at 600 ◦C for 2 h
resulted in the formation of NiO, as shown in Fig. 2b (NiO:
2θ = 36.4◦, 42.5◦, 62.0◦, 78.6◦ JCPDS # 4-835). However,
when the nickel precursors (acetate, nitrate) were impreg-
nated on SiO2 or cordierite and calcined in air at 600 ◦C
for 2 h, the main product detected by XRD was NiO, re-
gardless of the percentage weight of nickel, as indicated in
Fig. 3 (a–d) and Fig. 4b. In contrast, Fig. 3 (e and f) shows
that the oxidation state of nickel changed after reaction with
methane; that is, the NiO present in the catalysts Si-A-40
and Si-N-40 (without H2 pretreatment) was reduced to Ni0.
The same observation was valid for the Cor-A-8 (Fig. 4c)
after reaction with methane.
3. Results and discussion
3.1. Influence of nickel precursor on catalyst surface area
and activity
Table 1 summarized the effect of the nickel precursor and
type of support on the catalyst surface area. Ni(CH3COO)2·
4H2O was used as the nickel precursor and led to mate-
rials with higher surface areas than those obtained when
Ni(NO3)2·6H2O was used, regardless of the type of catalyst
(1 or 2), as indicated in Table 1. In addition, the amount of
nickel (wt%) used had an inverse effect on the catalyst sur-
face area—the higher the Ni wt%, the lower the surface area.
The difference in surface area between catalysts of types 1
and 2 was attributed to the catalyst support characteristic,
since cordierite had a lower surface area (> 0.2 m2/g) com-
pared with that of SiO2 (∼ 300 m2/g).
The difference in relative intensity of the carbon peak
(graphitic-type [3] 2θ = 26.1◦ (strongest), and 2θ = 42.6◦,

320
R.A. Couttenye et al. / Journal of Catalysis 233 (2005) 317–326
◦
Fig. 1. Changes in methane conversion during methane decomposition using B1 nickel catalysts. Experimental conditions: temperature = 550 C,
pressure = 1 atm, and flow rate = 40 ml/min.
◦
Fig. 2. XRD patterns of (a) nickel acetate, and (b) nickel nitrate, after decomposition in air at 600 C during 2 h.
53.8◦, 77.6◦) with respect to the Ni0 peaks, between the sam-
ple Si-A-40 and Si-N-40 (see Fig. 3, e and f), was higher for
the catalyst prepared with the acetate precursor, thus indi-
cating that a larger amount of carbon was deposited on this
catalyst; that is, the catalyst Si-A-40 had higher activity in
the decomposition of methane than Si-N-40 (as prepared).
Fig. 5A shows the XRD pattern for the catalyst Si-A-40 af-
ter complete deactivation (10-h reaction with methane). The
change in relative intensity of the carbon peak with respect to
the Ni0 peaks indicated that an additional amount of carbon
was accumulated on the catalyst as the reaction proceeded.
Steam regeneration [3,4,11] of the catalyst was performed
with a 50% steam in He mixture at 600 ◦C. After the regener-
ation process, the catalyst preserved its metallic nickel form;
however, some NiO appeared because of overexposure to
steam [4] (see Fig. 5B). Under the experimental conditions
small amounts of carbon also resisted gasification. The XRD
pattern of the regenerated catalyst after subsequent methane
decomposition at 550 ◦C (Fig. 5C) shows that nickel was in
metallic form (Ni0) and carbon (graphitic-type) was again
deposited on the catalyst. The relative intensity of the car-
bon peak with respect to the Ni0 peaks between the fresh
and regenerated catalysts (Fig. 5, A and C) suggested that
the activity of the regenerated catalyst was similar to the ac-
tivity of the fresh catalyst. For instance, Amiridis et al. [4]
indicated that 15 wt% Ni/SiO2 catalysts showed no signif-
icant decrease in methane conversion, even after 10 cycles
of methane decomposition–steam regeneration. The author

R.A. Couttenye et al. / Journal of Catalysis 233 (2005) 317–326
321
◦
◦
Fig. 3. XRD patterns of Ni/SiO catalysts: (a–d) after calcination in air at 600 C for 2 h, (e–f) after reaction with methane at 550 C for 40 min.
2
◦
Fig. 4. XRD patterns of Ni/cordierite catalyst: (a) bare cordierite monolith, (b) after calcination in air at 600 C for 2 h, and (c) after reaction with methane at
550 C for 17 h.
◦

322
R.A. Couttenye et al. / Journal of Catalysis 233 (2005) 317–326
◦
◦
Fig. 5. XRD patterns of catalyst Si-A-40: (A) after reaction with methane at 550 C for 10 h, (B) after steam regeneration at 600 C for 10 h, (C) regenerated
catalyst after subsequent methane decomposition at 550 C for 10 h.
◦
also highlighted the presence of small amounts of carbon af-
ter steam regeneration; however, in increase of this type of
carbon was observed from cycle to cycle.
Table 2
Catalysts carbon yield and NiO mean crystallite size
a
Catalyst Amount of catalyst
used (g)
Mass of carbon
deposited (g)
NiO mean
crystallite size (Å)
B1
B2
B3
B1
B2
B3
B1
B2
B3
3.3. Mass spectrometry
Type 1
Cor-A-8 0.59 0.61 0.60 1.13 0.63 1.68 167 125 195
Mass spectrometry analyses performed during the reac-
tion with catalysts Si-A-40 and Si-N-40, without reduction
pretreatment (Fig. 6, A and B), indicated the simultaneous
formation of carbon monoxide (CO, m/z = 28) and car-
bon dioxide (CO2, m/z = 44), immediately after exposure
to methane, for ∼ 6 min (see Fig. 6, inset). The difference
in partial pressure of CO and CO2 between the two cata-
lysts suggested that the initial amount of oxygen available,
as NiO, was lower for Si-A-40 than the amount available for
Si-N-40. It is known that nickel oxide (NiO) reacts slowly
with methane (at 700 ◦C) to produce CO2, H2O, and, even-
tually, Ni0 [25–27]. On the other hand, methane decomposes
rapidly on Ni0 to produce H2 and carbon (adsorbed to the
nickel surface). Adsorbed carbon reacts with NiO to form
CO and Ni0 [25,28]. Ni0 could also catalyze the decompo-
sition of CO to produce CO2 and carbon (i.e., Boudouard
reaction) [25]. Consequently, the immediate and simulta-
neous formation of CO and CO2 at the beginning of the
reaction indicated the presence of Ni0 in Si-A-40 and Si-N-
40 catalysts (as prepared), but with concentration below the
XRD detection limit. After the in situ reduction of nickel, the
only reaction products detected by mass spectrometry were
H2 and CH4 (i.e., carbon oxides were not detected). For in-
Type 2
Si-A-40 0.044 0.047 0.041 1.35 1.60 1.12 182 200 160
b
Si-A-40 0.040 0.040 0.040 0.87 0.93 0.98 49
56
24
62
15
Si-A-8
0.040 0.041 0.040 0.12 0.13 0.11 20
a
Sample nomenclature: support–salt–wt%.
Calcined in nitrogen at 600 C for 2 h.
b
◦
stance, Choudhary et al. [13] reported that the level of CO
formed during methane decomposition was 50 ppm when
they used 10 wt% Ni/SiO2 catalysts.
3.4. Effect of nickel oxide (NiO) mean crystallite size on
catalyst carbon yield (gC/gNi)
Table 2 shows the NiO mean crystallite size estimated
with the Sherrer equation, based on the half-width of the
diffraction peaks assigned to [200] and [111] NiO, for the
catalyst prepared with the acetate precursor. The experimen-
tal results listed in Table 2 suggest that the catalysts’ NiO
mean crystallite size was affected by (1) the nickel wt%,
(2) the type of gas used during calcination (air or nitrogen),
and (3) the catalyst support. The carbon yield (total mass of

R.A. Couttenye et al. / Journal of Catalysis 233 (2005) 317–326
323
Fig. 6. Carbon monoxide (CO, m/z = 28), and carbon dioxide (CO , m/z = 44) production monitored by mass spectrometer during decomposition of methane
2
◦
at 550 C, using samples: (A) Si-N-40 and (B) Si-A-40—without reduction pretreatment.

324
R.A. Couttenye et al. / Journal of Catalysis 233 (2005) 317–326
Fig. 7. Effect of NiO mean crystallite size on the catalysts carbon yield.
carbon (in grams) deposited on the catalyst after complete
deactivation, divided by catalyst nickel mass (in grams)) is
a common representation of the catalyst activity for decom-
position of methane [7,9,10,29]. Fig. 7 shows the effect of
NiO mean crystallite size on the catalyst carbon yield. For
type 2 catalysts, as the NiO mean crystallite size increased,
the carbon yield of the catalysts also increased (see Fig. 7).
Calcination of sample Si-A-40, under nitrogen at 600 ◦C for
2 h, resulted in a catalyst with a smaller NiO mean crys-
tallite size. Consequently, this catalyst had a lower carbon
yield, despite having 40 wt%. Ermakova et al. [7] reported
that the presence of SiO2 (10 wt% or higher) stabilized NiO
particles (with diameter < 65 nm) during reduction with H2,
and decreased sintering of the Ni0 particles being formed.
The author also indicated that as the diameter of the Ni0 par-
ticle increased (10–40 nm), the carbon yield of the catalyst
for decomposition of methane increased as well; however,
a further increase in the size of Ni0 particles above 40 nm
resulted in a sharp decrease in carbon yield. Similarly, the
experimental results shown in Fig. 7 for type 2 catalysts sug-
gest that the presence of SiO2 stabilized the NiO particles
(diameter = 2–18 nm) and the Ni0 particles (initially in the
catalyst), minimizing the sintering during the in situ reduc-
tion of NiO by methane. For instance, the mean crystallite
size of Ni0 particles formed in the catalyst Si-A-40 after
reaction with methane, for 10 h, was ∼ 25 nm (estimated
with the Sherrer equation based on half-width of the diffrac-
tion peaks assigned to (200), and (111) Ni0). In addition,
Fig. 7 showed that the carbon yield obtained with sample Si-
A-8 was higher than the yield obtained with Cor-A-8, even
though the NiO mean crystallite size was higher (∼ 16 nm)
for the catalyst prepared with cordierite (Cor-A-8). In this
case, the mean crystallite size of Ni0 particles formed in the
catalyst Cor-A-8 after reaction with methane, for 17 h, was
∼ 34 nm. This result suggested that cordierite could not sta-
bilize the NiO or Ni0 particles in the same way as SiO2, and
sintering of Ni0 particles increased.
3.5. SEM and FE-SEM characterization
The micrograph of the spent Cor-A-8 catalyst (Fig. 8a)
showed a moss-like material deposited on the catalyst sur-
face. A high-magnification micrograph revealed the pres-
ence of nickel particles (∼ 30–300 nm) located at the tip
of carbon filaments (Fig. 8b). The formation of graphitic-
type carbon filaments (as indicated by the increased inten-
sity of the XRD peak at 2θ = 26.1◦ (Fig. 4c) and FE-
SEM (Fig. 8b)) suggested that the decomposition of methane
with catalyst Cor-A-8 followed a mechanism similar to one
proposed elsewhere [3,5–8]. In particular, these authors re-
ported that carbon filament formation gave the catalyst the
capability of accommodating a significantly higher amount
of carbon than those predicted by either the site-blocking or
pore-mouth plugging models. However, because of spatial
limitations (i.e., loss of exposed nickel surface), the growth
of carbon filaments was terminated, and the catalyst was de-
activated.

R.A. Couttenye et al. / Journal of Catalysis 233 (2005) 317–326
325
(2) The use of nickel acetate as the catalyst’s nickel precur-
sor led to materials with a higher surface area than those
obtained when nickel nitrate was used, regardless of the
type of support (SiO2 or cordierite).
(3) The NiO mean crystallite size could be varied by a
change in (1) the nickel percentage weight, (2) the type
of gas used during calcination (air or nitrogen), and
(3) the catalyst support.
(4) For catalyst supported on SiO2 (∼ 200–300 m2/g), the
carbon yield (gC/gNi) attained during methane decom-
position increased with an increase in the catalyst’s NiO
mean crystallite size. SiO2 stabilized NiO particles (1.5–
20 nm) during in situ reduction by methane and mini-
mized sintering of Ni0 particles. In contrast, cordierite
could not stabilize the NiO or Ni0 particles in the same
way as SiO2, and significant sintering of Ni0 particles
occurred.
Acknowledgment
The authors are grateful to the Connecticut Global Fuel
Cell Center and the U.S. Army for financial support.
References
[1] J. Armor, Appl. Catal. Gen. 76 (1999) 159.
[2] R. Cortright, R. Davda, J. Dumesic, Nature 418 (2002) 964.
[3] T. Zhang, M. Amiridis, Appl. Catal. A: Gen. 167 (1998) 161.
[4] R. Aiello, J. Fiscus, H. Loye, M. Amiridis, Appl. Catal. A: Gen. 192
(2000) 227.
[5] A. Diks, J. Power Source 61 (1996) 113.
[6] M. Matsukata, T. Matsushita, K. Ueyama, Chem. Eng. Sci. 51 (1996)
2769.
Fig. 8. (a) SEM micrograph, and (b) FE-SEM high-resolution micrograph,
of Cor-A-8 catalyst after reaction with methane at 550 C for 17 h.
◦
4. Conclusions
[7] M. Ermakova, D. Ermakov, G. Kuvshinov, L. Plyasova, J. Catal. 187
(1999) 77–84.
[8] J.-W. Snoeck, G. Froment, M. Fowles, J. Catal. 169 (1997) 240–249.
[9] S. Takenaka, S. Kobayashi, H. Ogihara, K. Otsuka, J. Catal. 217 (2003)
79–87.
Based on the results obtained, the following conclusions
can be made:
[10] S. Takenaka, Y. Shigeta, E. Tanabe, K. Otsuka, J. Catal. 220 (2003)
468–477.
[11] V. Choudhary, S. Banerjee, A. Rajput, J. Catal. 198 (2001) 136.
[12] V. Choudhary, S. Banerjee, A. Rajput Appl. Catal. A: Gen. 234 (2002)
259.
[13] T. Choudhary, C. Sivadinarayana, C. Chusuei, A. Klinghoffer,
D. Goodman, J. Catal. 199 (2001) 9–18.
[14] P. Gronchi, P. Kaddouri, P. Centola, R. Del Rosso, J. Sol–Gel Sci.
Technol. 26 (2003) 843.
[15] V. Recupero, L. Pino, R. Di Leonardo, M. Lagana, G. Maggio, J. Po-
wer Source 71 (1998) 208.
[16] A. Da Silva, M. De Moura, L. Mercuri, A. Santo, J. Matos, An. Assoc.
Bras. Quim. 47 (1998) 133.
[17] M. Elmasry, A. Gaber, E. Khater, J. Therm. Anal. Cal. 47 (1996) 757.
[18] G. Hussein, A. Nohman, K. Attyia, J. Therm. Anal. Cal. 42 (1994)
1155.
[19] M. Mohamed, S. Halawy, M. Ebrahim, J. Anal. Appl. Pyrol. 27 (1993)
109.
(1) For the first time, the Ni0/NiO mixture obtained by ther-
mal decomposition of nickel acetate has been success-
fully used for catalytic decomposition of methane to
produce CO-free hydrogen suitable for hydrogen fuel
cells. The use of nickel acetate as a nickel precursor,
for the preparation of the catalysts, saves not only hy-
drogen (our product of interest), but also time and en-
ergy, by eliminating the pretreatment step (i.e., using
H2 at high temperature (> 500 ◦C) for at least 2 h).
In this sense, the activity of the catalyst prepared with
nickel acetate (8 and 40 wt%), supported on SiO2 or
cordierite, as prepared, was very close to the activity
obtained with the catalysts prepared with nickel nitrate
(pre-reduced with H2). These catalysts, once deactivated
by carbon deposition, were regenerated with steam at
600 ◦C. The steam regeneration produced additional H2
and preserved nickel in the metallic form.
[20] A. Gadalla, H. Yu, Thermochim. Acta 164 (1990) 21.
[21] A. Galwey, S. McKee, T. Mitchell, M. Brown, A. Bean, React. Solid. 6
(1988) 173.

326
R.A. Couttenye et al. / Journal of Catalysis 233 (2005) 317–326
[22] W. Pease, R. Segall, R. Smart, P. Turner, J. Chem. Soc. Faraday
Trans. 82 (1986) 747.
[26] K. Otsuka, Q. Lin, A. Morikawa, Inorg. Chim. Acta 118 (1986)
L23.
[27] R.K. Ungar, X. Zhang, R.M. Lambert, Appl. Catal. 42 (1988) L1.
[28] A. Slagtern, Y. Schuurman, C. Leclercq, X. Verykios, C. Mirodatos,
J. Catal. 172 (1997) 118.
[23] J. Doremieux, Compt. Rend. 261 (1965) 4426.
[24] J. Leicester, M. Redman, J. Appl. Chem. 12 (1962) 357.
[25] V.Yu. Bychkov, O.V. Krylov, V.N. Korchak, Kinet. Catal. 43 (1) (2002)
94.
[29] S. Takenaka, H. Ogihara, K. Otsuka, J. Catal. 208 (2002) 53.