The reactivity of surface active carbonaceous species with CO2 and its role on hydrocarbon conversion reactions

Article abstract of DOI:10.1016/j.molcata.2009.09.023

Carbon deposition on Ni-based catalysts has a significant influence on their cracking activity and selectivity and is the main reason for catalyst deactivation. To understand this behavior, pulse techniques and in situ infrared spectroscopic analysis were applied to the study of the surface carbonaceous species formation and transformation over Ni/MgAl₂O₄, Ni/MgO/γ-Al₂O₃ and Ni/γ-Al₂O₃. It was found that MgAl₂O₄ allows an effective way for CO₂ adsorption and activation through the formation of formate/carbonate type species. Carbon adspecies, mainly as CH_x (x = 1-3), are the intermediates of methane activation on Ni particles and preferably diffuse from the metal to the interference of Ni and the supports and promote the adsorbed CO₂ species to decompose and release CO through formate/carbonate type intermediates. The mechanism proposed emphasis the role of these surface species in the surface chemistry of carbonaceous reaction. The data obtained led to a satisfactory description of the working catalyst.

Full text of DOI:10.1016/j.molcata.2009.09.023

Journal of Molecular Catalysis A: Chemical 316 (2010) 1–7
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Editor choice paper
The reactivity of surface active carbonaceous species with CO2 and its role on
hydrocarbon conversion reactions
Jianjun Guoa,b, Hui Loua,∗, Liuye Moa, Xiaoming Zhenga,∗
a Institute of Catalysis, Department of Chemistry, Zhejiang University, Hangzhou 310028, PR China
b Ningbo Institute of Material Technology and Engineering, Chinese Academy of Science, Ningbo 315201, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Carbon deposition on Ni-based catalysts has a significant influence on their cracking activity and selectiv-
ity and is the main reason for catalyst deactivation. To understand this behavior, pulse techniques and in
situ infrared spectroscopic analysis were applied to the study of the surface carbonaceous species forma-
tion and transformation over Ni/MgAl2O4, Ni/MgO/␥-Al2O3 and Ni/␥-Al2O3. It was found that MgAl2O4
allows an effective way for CO2 adsorption and activation through the formation of formate/carbonate
type species. Carbon adspecies, mainly as CHx (x = 1–3), are the intermediates of methane activation on Ni
particles and preferably diffuse from the metal to the interference of Ni and the supports and promote the
adsorbed CO2 species to decompose and release CO through formate/carbonate type intermediates. The
mechanism proposed emphasis the role of these surface species in the surface chemistry of carbonaceous
reaction. The data obtained led to a satisfactory description of the working catalyst.
Received 28 August 2009
Received in revised form
30 September 2009
Accepted 30 September 2009
Available online 8 October 2009
Keywords:
Coke
Ni catalyst
Carbon dioxide
Methane reforming
FTIR
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
[4,5]. It was found that two main properties of a catalyst affect the
carbon deposition: surface structure and surface acidity. The nature
The formation of coke deposits leading to catalyst deactivation
has been a challenge for catalytic technology in many hydrocarbon
processes. The indirect and direct conversion of nature gas both on
oxidative and non-oxidative conditions are examples of such devel-
opments [1]. In these cases, the effective management of catalyst
deactivation and catalyst regeneration is the key in many hetero-
geneously catalyzed processes. The optimization of such complex
processes requires the characterization of the coke formation pro-
cess and its reactivity in order to understand the effect of the
operational variables on these deposits and, therefore, minimize
its accumulation and develop effective regeneration strategies.
Several characterization techniques [2], such as temperature-
programmed oxidation, X-ray photoelectron spectroscopy (XPS),
high-resolution electron microscopy, and Raman spectroscopy,
have been used to study coke deposits and to obtain information
regarding the deactivation mechanism and regeneration con-
ditions. Previously, we used temperature-programmed reaction
techniques and Raman spectroscopy to characterize coke species
deposited on a 5%Ni/MgAl2O4 catalyst [3], which exhibits high
activity and significant stability during dry reforming of methane
and quantity of the coke is strongly dependant on the composi-
tion of the reacting feed, reaction time and temperature. There are
three carbon species (i.e., C␣, C␤ and C␥) on the catalyst surface,
in which C␥ was graphite-like carbon species and was responsi-
ble for catalyst deactivation. The nature of the carbon species can
also be classified as polymeric, filamentous, and graphitic [6]. Poly-
meric coke is supposed to be derived from thermal decomposition
of hydrocarbons, whereas the filamentous and graphitic forms of
coke are formed on the catalyst. Coke can also be characterized
based on its reactivity with hydrogen, water and oxygen. Although
these are gross properties that do not provide detailed chemical
information, they are closely related to the actual reaction condi-
tions. Although coke characterization has been included in many
papers, there are still some disputations about the accurate mech-
anism details. Different surface species on Ni catalysts were also
identified on different supports. In situ diffuse reflectance fourier
transform infrared spectroscopy (DRIFTS) revealed that CHx can
react with OH species from the support, and CHxO species were
formed on TiO2 and Al2O3-supported transition metals [1,7]. While
on La2O3 supported transition metals, oxycarbonates were found
to be the main intermediate [8,9]. Further, it is argued that the CO2
was activated on support or Ni particles are also involved.
In the present study, the bulk structure of surface carbonaceous
species and its reactivity with various gases were further investi-
gated using X-ray diffraction (XRD), pulse experiments and in situ
∗
Corresponding authors. Tel.: +86 571 88273593; fax: +86 571 88273283.
E-mail addresses: hx215@zju.edu.cn (H. Lou), xmzheng@dial.zju.edu.cn
(X. Zheng).
1381-1169/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2009.09.023

2
J. Guo et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 1–7
the following eight compounds: H2, CH4, 13CH4, H2O, CO, 13CO,
CO2, 13CO2 within 1 s.
DRIFTS. Emphasizes was placed on the surface species formation
and transformation on the catalysts during dry reforming reactions.
The satisfactory description of the coking activity and a possible role
in hydrocarbon conversion reactions was proposed in the light of
the results obtained.
2.3. Catalyst characterization
The crystal structure of catalysts before and after reaction was
determined by XRD in a Rigaku D/max-IIIB apparatus using Cu K␣1
radiation (ꢀ = 1.54056 Å), at 45 kV and 40 mA. Diffraction peaks
recorded in a 2ꢁ range between 20◦ and 70◦ have been used to
identify the structure of the samples. The apparent crystallite
size was determined using the Scherrer formula with standard
procedure [11].
Nickel dispersion was determined by H2 chemisorption at
room temperature in a pulse experimental apparatus. The cata-
lyst was reduced by H2 at 750 ◦C for 2 h. Then the temperature
was decreased to 100 ◦C under Ar, kept there for 20 min, and
then further decreased to room temperature for the in situ H2
chemisorption measurements. The quantity of H2 adsorbed at
monolayer coverage and the mean particle sizes of Ni were
obtained by a standard procedure [12].
2. Experimental
2.1. Catalyst characterization
Magnesium aluminate spinel (MgAl2O4) was prepared in a
similar way to previous studies [4]. Stoichiometric quantities of
Mg(NO3)2·6H2O and Al(NO3)3·9H2O were dissolved separately in
distilled water. The resulting solutions and an aqueous solution of
ammonia were dropped into a flask at the same time with con-
stant stirring to maintain the pH value at about 9.5. The precipitate
was continuously stirred for an hour and then aged overnight. After
being thoroughly washed with distilled water for several times, the
resulting slurry was then dried at 120 ◦C for 15 h and calcined at
800 ◦C for 8 h in air.
Supported nickel catalyst was prepared by conventional wet
impregnation, using nitrate salt as the precursor. A given amount
of nickel nitrate solution was placed in a 100 ml beaker, the appro-
priate amount of MgAl2O4 or ␥-Al2O3 was added under continuous
stirring. The slurry was dried at 120 ◦C over night and calcined at
550 ◦C under air for 5 h.
2.4. In situ DRIFTS
DRIFTS experiments were performed on a Nicolet 740 FTIR spec-
trometer equipped with an in situ DRIFTS cell with KBr windows.
The cell was cooled by water circulating, which allowed collection
of spectra over the temperature range between 25 ◦C and 700 ◦C at
atmospheric pressure. The interaction of CO2, CH4 or the reaction
mixture with both the support and the catalyst was studied. In all
cases, the sample was reduced under a flowing H2/He mixture for
2 h and purged with He for additional 0.5 h at 700 ◦C. Then a flow
of CH4/He, CO2/He or the reaction mixture was introduced into the
DRIFTS cell. The spectra of the adsorbed specie were obtained by
subtraction of the initial spectrum registered after the reduction
step. For all the spectra recorded, a 32-scan data accumulation was
2.2. Catalytic testing
The CO2 reforming of methane was carried out at 750 ◦C and
atmospheric pressure, using 20 mg catalyst (unless otherwise men-
tioned) in a fixed-bed quartz reactor (ca. 4 mm i.d.). Activation
of the Ni-catalyst involved reductive treatment with hydrogen at
750 ◦C for 60 min. The reactant feed was CH4:CO2 = 1:1 and a flow
rate of 30 ml/min. The exit gases were analyzed with an on-line
gas chromatograph equipped with a TCD, using a TDX-01 column.
The conversions and selectivity were calculated on the basis of the
reforming reaction and the water–gas shift reaction [10].
carried out at a resolution of 4 cm−1
.
3. Results
Pulse reactions were performed in a conventional flow sys-
tem with a flow measuring and control system, a mixing chamber,
and a fixed-bed quartz reactor (4 mm i.d., 200 mm length), which
was placed in an electric oven. Before reaction, the catalysts were
reduced in situ at 750 ◦C for 60 min in H2 flow. Before the pulse data
were obtained, the catalyst was treated in situ at 750 ◦C for 30 min
in flowing Ar. The reactant was introduced with a 6-port gas sam-
pling valve under a stream of Ar carrier gas. Analysis of the gases
during the pulsed reactions was done by an on-line quadrupole
mass spectrometer (OmniStar GSD 301), which scanned peaks of
3.1. Catalyst characterization and catalytic performance
The XRD patterns of MgAl2O4 supported Ni catalysts were
quite simple [4]. The diffraction lines of Ni/MgAl2O4 catalyst vir-
tually coincided with those of the as-prepared MgAl2O4 support
and no other lines of Ni-containing species were discerned. How-
ever, NiO diffraction lines were depicted with the Ni/␥-Al2O3
and Ni/MgO–Al2O3 catalysts. This implies that the NiO in the
as-prepared Ni/MgAl2O4 catalysts was finely dispersed (with a dis-
persion of 16%, Table 1), which could be responsible for the higher
Table 1
Properties, catalytic behavior and coking of 5%Ni/␥-Al2O3 and 5%Ni/MgAl2O4 catalystsa.
Catalyst
H2 uptake (cm3 g−1
)
Ni dispersion (%)
Conversion (%)
Selectivity (%)
Metallic particle size of Ni0 (nm)
Cokeb (wt.%)
14.5
CH4
CO2
H2
CO
Before reaction
After Reaction
0.2896
3
59.1
61.0
82.9
94.2
26.5c
110d
90.0c
n.d.e
n.d.
Ni/␥-Al2O3
19.6f
1.5797
16
82.6
87.7
97.5
93.8
10.4c
20d
11.8c
n.d.
1.2
Ni/MgAl2O4
8.8f
10.1f
a
Reaction conditions: T = 750 ◦C; catalyst, 50 mg; GHSV = 50,000 ml/(g-cat h); CH4:CO2 = 1.0:1.0; P = 1 atm.
Coke deposition was quantified by thermogravimetry after 10 min on stream.
Determined by XRD.
Determined by H2 chemisorptions assuming that each surface metal atom chemisorbs one hydrogen atom, i.e., H/Nisurface = 1.
n.d., not determined.
Determined by TEM.
b
c
d
e
f

J. Guo et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 1–7
3
stability and lower carbon deposition of the Ni/MgAl2O4 catalyst
in dry reforming of methane. The mean nickel particle sizes esti-
mated by employing Scherrer’s equation following the standard
procedures are of the order of 10.4 nm and 11.8 nm before and
after 10 h on stream as indicated in Table 1. These are in accor-
dance with values derived from transmission electron microscopy
(TEM) results (Table 1), in which the Ni0 particle size increased from
8.8 nm to 10.1 nm. This indicates that Ni particle sintering occurs
under the reaction conditions. Sintering is an important route for
the deactivation of Ni-based dry reforming catalysts. However, the
particle size only increased about 15% in the first 10 h and this did
not induce deactivation, as demonstrated in Table 1. This is because
the catalyst can produce a highly dispersed active phase and a large
active surface area as bound-state Ni species after pre-reduction.
In contrast, the Ni/␥-Al2O3 catalyst has a relatively lower Ni dis-
persion and sintered severely during the reaction (the particle size
increased about 300% as indicated in Table 1). This implies that
a proper interaction exists between Ni and MgAl2O4, which con-
tribute to the high dispersion of Ni and inhibit the sintering of Ni
particles, as has been discussed [4].
Fig. 1. Spectral detail of C–H absorption band region over 5%Ni/MgAl2O4 under dif-
ferent conditions: (a) under CH4/CO2 for 10 min; (b) after purge with N2 for 20 min;
(c) after TPO for 1 min; (d) after TPO for 2 min; (e) after TPO for 5 min at 650 ◦C.
We have studied the activity and stability of the as-prepared
catalysts in differential condition. As shown in Table 1, Ni/MgAl2O4
gives superior high activity than Ni/Al2O3 toward dry reforming
of methane. The ratio of H2/CO obtained under the stated con-
ditions was found to vary between 0.78 and 1.01 for Ni/Al2O3
and Ni/MgAl2O4, respectively. Long-term stability tests indicate
that the activity of Ni/MgAl2O4 catalyst remained unchanged over
55 h on stream, while Ni/Al2O3 deactivated quickly in 10 h [4,5].
It has been established that the primary reason for deactivation
of Ni/␥-Al2O3 can be generally attributed to carbon deposition on
the Ni crystallites. As also shown in Table 1, the amount of coke
reached about 14.5 wt% on Ni/␥-Al2O3, while only 1.2% was found
on Ni/MgAl2O4 after 10 min on stream. With the time on stream,
it was found that the coking rate decreased on both catalysts, but
Ni/␥-Al2O3 deactivated quickly owing to the large amount of coke
accumulation, which plugs the reactor in around 10 h [4,5].
species. A weak shoulder at around 2890 cm−1 also suggests the
presence of some CH groups [15]. In situ oxidation of the reacted
sample was carried out subsequently. The CHx bands described
above disappeared, restoring the reference spectrum after 5 min.
3.3. Transient response reactions
Two sequences of pulsed reactions were conducted over
Ni/␥-Al2O3 and Ni/MgAl2O4 catalysts. Fig. 2 shows responses to
13CH4 pulse into 10% 12CO2/Ar steady flow with Ni/␥-Al2O3 and
Ni/MgAl2O4 at 750 ◦C. Signals of 12CO2 were omitted from the pro-
file. When 13CH4 was introduced, H2, 13CO, 12CO and 13CO2 were
generated immediately on Ni/MgAl2O4 and the signals of 12CO and
13CO2 reached a maximum at about 4 s later than those of 13CO and
H2. This indicates that CO2 and CH4 were activated separately. The
reaction process is likely preceded by rapid hydrogen abstraction
from CH4 to give adsorbed CHx and H species, followed by a slow
reaction of CO2 with adsorbed CHx species. The adsorbed H may
assist CO2 dissociation on both Ni and support surface since gener-
ation of 12CO was also improved by 13CH4 introduction, as reported
by Osaka over various catalysts [16].
3.2. Surface active carbonaceous species
Although most researchers agree that carbon formation is the
primary reason for the catalyst deactivation, disagreement exists
on the source of the carbon. In our previous study [3], the amount
of coke species deposited from CO disproportionation was found
to be too small compared to CH4 decomposition and was neg-
ligible during dry reforming of methane on Ni/MgAl2O4, which
suggested that the main source of coke during the reaction was
originated from CH4 decomposition. The nature of these carbon
species was further studied by in situ Fourier transform infrared
spectroscopy (DRIFT), which was shown in Fig. 1. When a large
amount of methane was introduced, the intense gaseous CH4
adsorption at 2900–3050 cm−1 appeared. While no CHx species
was observed under the reaction conditions even at 500 ◦C (not
shown). As a result, it hardly detect CHx (x = 1–3) species even
though CHx species was widely accepted as an intermediate species
for methane activation during CO2 reforming of methane. It should
be noted that the concentration of a species on the surface depends
on its formation and conversion rates. The importance of a sur-
face species in a catalytic cycle cannot be simply determined by
its surface concentration or coverage. Many research groups have
evidenced the existence of CHx species by CD4/H2 and CH4/D2 iso-
topic transient reactions [13] and pulsed surface reaction analysis
[14]. Indeed, a detailed spectrum of the C–H bands can be obtained
after flushing the reacting gas phase with argon after 10 min on
stream. Although weak and poorly resolved, the main bands around
2966 cm−1, 2906 cm−1, and 2871 cm−1 can be assigned to asym-
metric and symmetric stretching vibrations of –CH3 and CH2
After 13CH4 pulse passed through the catalyst bed, H2, 13CO,
12CO and a very small amount of 13CO2 were continued to evolve.
Similar results were observed over Ni/␥-Al2O3 catalyst. While the
conversion of CH4 was much lower over Ni/␥-Al2O3 than that over
Ni/MgAl2O4. This may be due to a lower dispersion of the active
metal species on Ni/␥-Al2O3 catalyst [4].
Fig. 3 shows responses to pulsing CO2 into 10% CH4/Ar steady
flow over Ni/MgAl2O4 and Ni/␥-Al2O3 catalyst. In this case, sur-
face carbonaceous species (i.e., –CH3, CH2 and CH) were formed
previously and continuous during the pulse reaction. When CO2
was introduced, formation of CO was detected immediately over
Ni/MgAl2O4, indicating of rapid CO2 dissociation over Ni and
MgAl2O4 perimeter. H2 response started to reinforce after a few
seconds. CO and H2 continued to appear after CO2 passed through
the catalyst bed. When the same reaction was conducted on Ni/␥-
Al2O3, only very small amount of CO2 reacted and a small amount
of H2 and CO produced. This is due to the very weak adsorption of
CO2 on ␥-Al2O3, unlike the case on MgAl2O4 support, where CO2
is able to be concentrated after CO2 pulses. Large amount of CHx
species were formed on both catalysts under CH4 steady flow. But
due to the weaker activation of CO2 on Ni/␥-Al2O3, the reaction of
CO2 with CHx occurred more slowly and the response of CO was
very weak as compared to that of Ni/MgAl2O4.

4
J. Guo et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 1–7
Fig. 2. Response to 13CH4 pulse into 10%CO2/Ar steady flow over 5%Ni/MgAl2O4 and
Fig. 3. Response to CO2 pulse into 10%CH4/Ar over 5%Ni/MgAl2O4 and 5%Ni/Al2O3
5%Ni/Al2O3 at 750 ◦C.
at 750 ◦C.
DRIFTS of the hydrogen-treated support after exposure to the
reacting mixture at 680 ◦C are shown in Fig. 5. When methane
was introduced into the reacting mixture, the intensity of the band
around 1500 cm−1 was reduced. After methane introduction, bands
at 2170 cm−1 and 2115 cm−1 were observed and can be assigned to
gaseous CO. The intensity of gaseous CO decreased with time on
stream and reached a constant value after 30 min. It can be seen in
Fig. 5 that the intensity of gaseous CO and carbonate species did not
3.4. DRIFTS
DRIFTS experiments were conducted in CH4, CO2 or CH4/CO2
gas flows at 680 ◦C with the objective of identifying the adsorbed
carbonaceous species under reaction conditions. Experiments were
also conducted where changes in the IR spectrum were monitored
in response to changes in the gas phase mixture induced by substi-
tuting the reactant mixture by diluted methane or carbon dioxide.
3.4.1. Surface active species on catalyst supports
DRIFTS of the MgAl2O4 support and the Ni/MgAl2O4 cata-
lyst after adsorption of CO2 and CH4/CO2 mixture are shown in
Figs. 4 and 5 separately. Apart from the gas phase contributions
of CO2 centered ca. 2349 cm−1 13CO2 centered ca. 2275 cm−1 or
,
CH4 at 1305 cm−1 and 3015 cm−1, several bands in the region
between 1200 cm−1 and 1600 cm−1 became apparent. On the sup-
port (in Fig. 4), two well-defined bands centered at 1500 cm−1 and
1293 cm−1 are observed after introduction of CO2. The IR intensities
of these two bands increase with increasing temperature and time
on stream. These bands could be attributed to the carbonate type
species adsorbed on the support at high temperature [17]. In addi-
tion, two shoulder bands centered at 1580 cm−1 and 1400 cm−1
also appeared and followed the same trends as carbonate species.
These two bands can be assi−gned to asymmetric C–O and H–C–O
stretching of adsorbed HCO2 [18]. Isolated O–H stretching mode
of alumina hydroxyl was apparent at 3609 cm−1. Meanwhile,
three well-defined bands centered at 3726 cm−1, 3702 cm−1 and
3618 cm−1 were appeared after CO2 introduction due to the Fermi
resonance of linearly adsorbed CO2 [19].
Fig. 4. DRIFT spectra of CO2 adsorption at 680 ◦C on MgAl2O4. The reference spec-
trum was that of the sample prior to gas admission.

J. Guo et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 1–7
5
Fig. 5. DRIFT spectra of CH4/CO2 reforming on pretreated MgAl2O4 with CO2 at
Fig. 7. DRIFT spectra of CH4/CO2 reforming on pre-reduced 5%Ni/MgAl2O4. The
reference spectrum was that of the sample prior to gas mixture admission.
680 ◦C. The reference spectrum was that of the sample prior to gas admission.
change as a function of time on stream after introduction of CH4 for
30 min.
of CHx, i.e., adsorbed H species, favors the formation of formate
species first. With the increase of temperature to above 600 ◦C, the
adsorption of formate species weakened, while the intensity of
carbonate species (at 1510 cm−1 and 1281 cm−1) increased accord-
ingly. Over these catalysts, the formate species may transformed
to carbonate species under a higher temperature.
3.4.2. Surface active species on catalysts
In another experiments, Ni/MgAl2O4 was reduced in situ at
700 ◦C for 1 h in flowing hydrogen. After purging with Ar for 30 min,
the gas flow was switched to 10%CO2 + Ar or 10%CH4 + 10%CO2 + Ar.
The surface species were also monitored by DRIFTS, and the results
are shown in Figs. 6 and 7. Over Ni/MgAl2O4, similar tendencies
were observed as on MgAl2O4 after introduction of CO2 (Fig. 6). The
adsorbed formate (bands at ca. 1410 cm−1 and 1551 cm−1) and car-
bonate (bands at ca. 1501 cm−1 and 1291 cm−1) species appeared
after 2 min of CO2 introduction and its intensity increased with time
on stream.
From the above experiments, no adsorbed CO and CHx species
were observed by in situ Fourier transform infrared spectroscopy
(FTIR). Kroll et al. [21] observed intense band, due to linear and
multi-bonded CO species, developed at 2012 cm−1 and 1855 cm−1
,
respectively after introduction of the reaction mixture on Ni/SiO2
catalyst at 500 ◦C. The lack of IR bands corresponding to adsorbed
CO in the spectra of our samples suggests that the rate of desorption
is greater than its formation at 680 ◦C.
In CH4–CO2 co-feeding conditions, large amount of
formate (ꢂs,c o = 1367 cm−1
,
,
ꢂas,c o = 1551 cm−1
)
and car-
4. Discussion
bonate (ꢂas,c o = 1640 cm−1
ꢂs,c o = 1432 cm−1
,
1281 cm−1
,
ıC–O = 1230 cm−1) species were formed. Efstathiou et al. [20]
found that the formate species on Al2O3 surface was stable under
CH4/CO2 reforming condition even at 500–600 ◦C. In the present
study, the formate species formed were quite similar to those over
Al2O3. The intensity of these bands indicated that the presence
One of the major problems encountered in the activation of
hydrocarbons over metal catalysts is the rapid accumulation of
coke, which leads to blocking the active sites and deactivating
the catalyst activity. However, in contrast to 5%Ni/Al2O3 cata-
lyst which exhibit rapid deactivation during CO2 reforming of
methane, no deactivation was observed during 55 h on stream over
a 5%Ni/MgAl2O4 catalyst [4]. In the 5%Ni/Al2O3 catalyst, Ni dis-
persion is very low. Based on the H2 adsorption experiments, the
Ni0 average particle size of 5%Ni/Al2O3 catalyst is of the order of
110 nm, which are approximately 5 orders of magnitude higher
than that of 5%Ni/MgAl2O4 catalyst. TEM showed that the Ni0
average particle size of 5%Ni/MgAl2O4 was 8.8 nm, which was in
agreement with the measurements from H2 adsorption results
and XRD results. In any event, the three techniques applied show
that the Ni particle size is small on the MgAl2O4 support. After
on stream for 10 h, the Ni particle size increased by only 13%
over Ni/MgAl2O4, compared to approximately 300% of Ni/Al2O3.
As have been reported [4], the particle size of MgAl2O4 was only
8.9 nm by XRD, which is comparable to 10.4 nm of the Ni particles.
A nanocomposite catalyst may be formed between MgAl2O4 and
size-comparable Ni-metal nanocrystals. The unusual interactions
between these nanocomposite catalysts may contribute to the high
dispersion of active metals and extremely stable catalysis [22].
The active oxygen species have important impact on the inhi-
bition of inert carbonaceous species formation. An “oxygen pool”
was proposed to be formed on the catalyst surface during CO2
decomposition [23]. This causes the decreasing of CO2 conversion
Fig. 6. Adsorption of CO2 over 5%Ni/MgAl2O4 at 680 ◦C. The reference spectrum was
that of the sample prior to CO2 admission.

6
J. Guo et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 1–7
and formation of CO2 instead of CO when reacted with surface
CHx species. Fig. 2 of present study 13CO2 was observed over
both Ni/Al2O3 and Ni/MgAl2O4 after pulsing 13CH4 into 10%CO2/Ar
steady flow. Over Ni/MgAl2O4, 13CO2 continued to evolve after
13CH4 pulse passed through the catalyst bed, while this is not
observed over Ni/Al2O3. These results indicate that the following
reactions occurred over these catalysts:
through surface hydroxyl group over Al:
4Al–OH + 2CO ↔ 2Al–O–COOH + Al–O–Al + H2O
2Al–OH + CO ↔ Al–O–CO–O–Al + H2O
(6)
(7)
A few papers have been published in which oxycarbonates were
detected in both Ru and Ni catalysts supported on La2O3 [8]. These
authors concluded that the formation of oxycarbonates during
reaction plays a central role in the CO2 reforming of methane. But
this was seldom addressed in the catalyst when using aluminum-
based supports. This could be ascribed to the weak adsorption
of CO2 on these compounds. When MgO was introduced into
␥-Al2O3 surface, the adsorption of CO2 was enhanced and the
specific role of carbonate species was further elucidated in our
samples. The temperature-programmed oxidation, temperature-
programmed hydrogenation and temperature-programmed CO2
reaction profiles showed that there were three carbon species (i.e.,
C␣, C␤ and C␥) on Ni/MgAl2O4 surface [3]. C␥ was responsible for
catalyst deactivation. The C␥ species was found to be the most
inactive species toward H2 and O2. In contrast, the C␥ species
was unexpectedly more active toward CO2 than C␣ and C␤ on
Ni/MgAl2O4 surface. The unique reactivity of CO2 with different
coke species could be ascribed to the carbonate, bidentate and for-
mate species formation on MgAl2O4 surface. These surface species
enhanced the oxidation of C␥ and thus contributed to the high sta-
bility of Ni/MgAl2O4 catalyst. In the dry reforming conditions with a
high temperature, the surface hydroxyl groups desorbed easily, and
CO2 decomposition was depressed. In the present study, when CO2
was adsorbed on the support, no CO was formed except carbonate
species was observed by FTIR in Fig. 6.
These conclusions may find support in the literature. Bitter et
al. [25] have suggested that Pt/ZrO2 catalysts are not efficient for
dry reforming of methane when formation of carbonates is not
possible. Recently, Zhang et al. [26] also found that La2O2CO3 and
formate species may participate in the surface chemistry to pro-
duce synthesis gas over a high stable Ni/La2O3 catalyst. From the
above proposed mechanism, the formation of carbonate and for-
mate type species is also an important intermediate for the reaction
of CHx with CO2 over Ni catalyst. Although different catalytic per-
formances usually observed considering the Ni and noble metal
catalysts (such as lower carbon deposition and hither activity), the
Ni-based catalyst can be more active and resistant to coke accu-
mulation when supported on appropriate catalyst support, which
allows a more effective CO2 activation.
13CH4(a)–Ni0 → 13CHx–Ni0 + H2(g) rapid
(1)
H–Ni0 + 12CO2–MgAl2O4 → OH–MgAl2O4 + 12CO–Ni0 rapid
(2)
12CO2(a)–MgAl2O4+13CHx–Ni0
→
12CO(g) + 13CO(g) + (x/2)H2(g) + Ni0–MgAl2O4 slow
(3)
This implies that “oxygen pool” did form over these two cata-
lysts in a large amount of CO2, and this is so prominent over
Ni/MgAl2O4 than over Ni/Al2O3. In the FTIR spectra, both MgAl2O4
and Ni/MgAl2O4 in the presence of CO2 displayed several intense
bands in the 1600–1300 cm−1 region, which can be assigned to
carbonate/bicarbonate type species. These adsorption bands were
depressed largely by the introduction of CH4 subsequently. CHx
species formation was conformed also by FTIR experiments in Fig. 1.
During the reaction, CO2 may rapidly react with MgAl2O4 to gener-
ate carbonate/bicarbonate type species, which in turn reacts slowly
with the CHx species to generate the other main products, H2 and
CO. This slow reaction occurs most likely at the metal/support
interface. As indicated in Fig. 5, there established a steady state
between the surface carbonate species and the gaseous CO pro-
duction when CH4 was introduced to carbonate/bicarbonate type
species covered catalyst. The presence of MgAl2O4 may promote
the adsorption of CO2 and formation of carbonate intermediate.
The formation and reaction equilibrium of carbonate/bicarbonate
type species sustained the high activity of Ni/MgAl2O4 catalyst.
On the other hand, Ni/MgAl2O4 and Ni/Al2O3 give similar
response to CO2 pulse into 10%CH4/Ar. CO was detected imme-
diately when CO2 was introduced. Interestingly, H2 was started
to appear after a few seconds. This implies that CO2 was firstly
adsorbed on the support and dissociated into gaseous CO and
adsorbed O species, which in turn reacted with surface CHx to
release CO:
CO2–MgAl2O4 + Ni0 → CO(g) + Ox–Ni rapid
Ni0–CHx + Ni0 − Ox → CO(g) + H2(g) + 2Ni0 slow
(4)
(5)
5. Conclusions
The continuous formation of CO and H2 evidenced this assump-
tion.
The catalyst support, MgAl2O4, plays an essential role in sur-
face carbonaceous species with CO2, including assisting nickel
to form highly and uniformly dispersed metallic nickel parti-
cles after reduction through the interaction between the active
phase Ni and the support; allowing an effective way for CO2 acti-
vation through formation of carbonate/bicarbonate type species,
which contributes to the long-term stability and specific carbon
elimination. The activation of methane through dehydrogena-
tion progressively was confirmed by the observation of CHx
species directly on Ni catalyst. The mechanistic aspect of surface
active carbonaceous species with CO2 over Ni/MgAl2O4 cata-
lysts was recommended according to these results. The surface
carbonate species may migrate to the metal–support interfacial
region or may spill over onto the metal surface to react with
CHx. This reaction provides a rationale for the stability of this
catalyst.
The exact form and properties of this “oxygen pool” were not
clearly elucidated until now although it was proposed the possible
form of adsorbed oxygen, hydroxide or carbonate. Ferreira-Aparicio
et al. [24] found that the carbonate type species, when CH4 was
introduced to the reactant flow, decreased and is replaced by
adsorbed formate species, which were depleted after carbon diox-
ide is removed from the mixture. Different trends were observed
over our catalysts. As illustrated in Figs. 3 and 6, adsorption of
CO2 over the support or catalyst produced large amount of formate
species, which can be enhanced by CH4 introduction (Figs. 5 and 7).
This implies that these species can transform from one to another.
This requires the participation of mobile oxygen species and hydro-
gen species. Surface hydroxyl groups (surface hydroxide or HCO3)
on the support may serve as carriers that allow mobility of these
species. As for alumina-based support, this can be carried out

J. Guo et al. / Journal of Molecular Catalysis A: Chemical 316 (2010) 1–7
7
Acknowledgment
[12] G. Ertl, H. Knozinger, J. Weitkamp, Hangdbook of Heterogeneous Catalysis,
Wiley, Weinheim, 1997, 442 pp.
[13] L.M. Aparicio, J. Catal. 165 (1997) 262–274.
Financial support from the Chinese National Natural Science
Foundation (Project Nos. 20343002 and 20433030) is gratefully
acknowledged.
[14] T. Osaki, H. Masuda, T. Mori, Catal. Lett. 44 (1997) 19–21.
[15] G. Socrates, Infrared and Raman Characteristic Group Frequencies-tables and
Charts, Wiley, New York, 2004, 338 pp.
[16] T. Osaki, J. Chem. Soc., Faraday Trans. 93 (1997) 643–647.
[17] A. Davydov, Molecular Spectroscopy of Oxide Catalyst Surfaces, Wiley, Chester,
2003, 134 pp.
References
[18] M.M. Schubert, H.A. Gasteiger, R.J. Behm, J. Catal. 172 (1997) 256–258.
[19] G. Heerberg, Molecular Spectra and Molecular Structure. II. Infrared and Raman
Spectra of Polyatomic Molecules, Van Nostrand, New York, 1945, 274 pp.
[20] A.M. Efstathiou, A. Kladi, V.A. Tsipouriari, X.E. Verykios, J. Catal. 158 (1996)
64–75.
[21] V.C.H. Kroll, H.M. Swaan, S. Lacombe, C. Mirodatos, J. Catal. 164 (1996) 387–398.
[22] B.Q. Xu, J.M. Wei, Y.T. Yu, J.L. Li, Q.M. Zhu, Top. Catal. 22 (2003) 77–85.
[23] A.N.J. van Keulen, K. Seshen, J.H.B.J. Hoebink, J.R.H. Ross, J. Catal. 166 (1997)
306–314.
[24] P. Ferreira-Aparicio, I. Rodriguez-Ramos, J.A. Anderson, A. Guerrero-Ruiz, Appl.
Catal. A 202 (2002) 183–196.
[25] J.H. Bitter, K. Seshan, J.A. Lercher, J. Catal. 171 (1997) 279–286.
[26] Z. Zhang, X.E. Verykios, S.M. MacDonald, S. Affrossman, J. Phys. Chem. 100
(1996) 744–754.
[1] J.H. Lunsford, Catal. Today 63 (2000) 165–174.
[2] C.A. Querini, in: J.J. Spivey, G.W. Roberts (Eds.), Catalysis, vol. 17, The Royal
Society of Chemistry, 2004, pp. 166–209.
[3] J.J. Guo, H. Lou, X.M. Zheng, Carbon 45 (2007) 1314–1321.
[4] J.J. Guo, H. Lou, H. Zhao, D.F. Chai, X.M. Zheng, Appl. Catal. A 273 (2004) 75–82.
[5] J.J. Guo, H. Lou, X.M. Zheng, React. Kinet. Catal. Lett. 84 (2005) 93–100.
[6] C.H. Bartholomew, J.B. Butt, Catalysts Deactivation, Elsevier Science Publishers,
Amsterdam, 1991.
[7] M.C.J. Bradford, M.A. Vannice, J. Catal. 173 (1998) 157–171.
[8] J.F. Múnera, S. Irusta, L.M. Cornaglia, E.A. Lombardo, D.V. Cesar, M. Schmal, J.
Catal. 245 (2007) 25–34.
[9] X.E. Verykios, Int. J. Hydrogen Energy 28 (2003) 1045–1063.
[10] L.Y. Mo, J.H. Fei, C.J. Huang, X.M. Zheng, J. Mol. Catal. A 193 (2003) 177–184.
[11] H.P. Klug, L.E. Alexander, X-ray Diffraction Procedures for Polystalline and
Amorphous Materials, 2nd ed., Wiley–Interscience, New York, 1974, 148 pp.