Applied Catalysis B: Environmental 168-169 (2015) 164–174

Contents lists available at ScienceDirect

Applied Catalysis B: Environmental

journal homepage: www.elsevier.com/locate/apcatb

A study on characterization and methane dry reforming performance

of Co–Ce/ZrO2 catalyst
A. Ipek Paksoy a , Burcu Selen Caglayan a,b , A. Erhan Aksoylu a,∗
a
Department of Chemical Engineering, Bogazici University, 34342 Bebek, İstanbul,Turkey
b
Advanced Technologies R&D Center, Bogazici University, 34342 Bebek, İstanbul,Turkey

a r t i c l e i n f o a b s t r a c t

Article history: The overall purpose of this study is to propose an effective Co-based non-PGM bimetallic carbon dioxide
Received 12 August 2014 reforming of methane (CDRM) catalyst, and to find the optimal reaction conditions to be used in pro-
Received in revised form duction of synthesis gas. In catalyst formulation, ZrO2 , which has ability to produce surface oxygen, is
18 December 2014
used as a support. Ce is used as a promoter in order to increase oxygen storage capacity and to regulate
Accepted 21 December 2014
surface oxygen transfer. Freshly reduced and spent catalysts were analyzed by SEM–EDX, HRTEM–EDX,
Available online 2 January 2015
TPO, Raman Spectroscopy and XPS. The reaction conditions were optimised through the use of an exper-
imentally designed procedure using reaction temperature, CH4 /CO2 feed ratio and space velocity as the
Keywords:
Methane dry reforming
parameters. XPS analysis of Co 2p indicated that additional oxidation of Co was limited during CDRM
CO2 utilization which may only lead to a small decrease in activity. EM measurements showed well dispersion of Co
Catalytic H2 production and Ce on the surface. No significant metal sintering was observed. Local changes in Co/Ce ratio were
Non-PGM catalyst found to have pronounced effect on carbon formation. It is well known that methane dehydrogenation
produces surface carbon as the side product covering the active site. The comparative analysis of extent
of carbon coverage on Co and Ce sites of the freshly reduced and spent samples through metal mapping
by SEM–EDX indirectly confirmed methane dehydrogenation is the primary function of Co sites, and
there is no methane activation on Ce sites. A comparative evaluation of XPS Ce 3d obtained from freshly
reduced and spent samples on the other hand showed the surface oxygen storage/transfer function of Ce
through redox cycle. In the performance tests, high activity with stable performance was observed for
1/1CH4 /CO2 feed ratio for CDRM over Co–Ce/ZrO2 catalyst.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction excess heat from gas turbine exhaust, and as a source of CO and
H2 for flame stabilization in low temperature methane fired gas
Carbon dioxide reforming of methane (CDRM) is a catalytic turbines [3].
process that produces valuable synthesis gas from methane and The main drawbacks of CDRM are coking and metal sintering.
carbon dioxide. Compared to other reforming routes, such as Methane decomposition and carbon monoxide disproportionation
steam reforming (SR) or partial oxidation (POX), CDRM has the through Boudouard reaction are possible reasons of coke formation
advantage of utilizing carbon dioxide, which is the most ther- and deposition. The former is favored at high temperatures and
modynamically stable greenhouse gas whose concentration rise low pressures, whereas the latter is favored at low temperatures
in the atmosphere has been accepted as the most plausible rea- and high pressures [4]. The high temperature window required for
son for the recent climate change. CDRM can be directly used CDRM, and the formation of H2 O by the reversed water gas shift
without any pre-separation in producing syngas from natural gas, reaction make the catalysts sensitive to metal sintering [5].
which may have large amount of CO2 [1]. The low H2 /CO ratio Although noble metals exhibit good activity and suffer less from
obtained in CDRM is preferable in production of valuable, espe- carbon deposition, they have the drawback of high cost and lim-
cially olefinic, hydrocarbons and alcohols via further reactions, like ited availability [6,7]. As they are relatively cheap and possess high
Fischer–Tropsch. Since the reaction is highly endothermic, CDRM activity, industrial nickel-based catalysts have been preferred for
can be also used in chemical energy transmission and storage sys- reforming reactions. However, at high CDRM temperatures, they
tems [2]. Moreover, the process is suitable to be used in recovering deactivate quickly due to metal sintering and coke deposition.
In designing non-noble metal based catalysts, coke resistance
and stable performance can be achieved through the use of a
∗ Corresponding author. Tel.: +90 212 3597336; fax: +90 212 2872460. suitable metal(s)–support combination leading beneficial synergy
E-mail address: aksoylu@boun.edu.tr (A.E. Aksoylu). between metal and support [8,9]. The carbon formation in CDRM

http://dx.doi.org/10.1016/j.apcatb.2014.12.038
0926-3373/© 2014 Elsevier B.V. All rights reserved.
 A.I. Paksoy et al. / Applied Catalysis B: Environmental 168-169 (2015) 164–174 165

can be controlled by using a support, like ZrO2 , favoring surface catalyst samples in order to elucidate their micro-structural prop-
oxygen production via dissociative adsorption of CO2 ; the carbon erties. The SEM micrographs of the spent catalyst samples tested
formed on the active metal centers is cleaned by this surface oxy- under different reaction conditions were also used to observe
gen [5,10]. ZrO2 also has high thermal stability and an appreciable whether there is a change in metal dispersion and the morphology
ionic conductivity led by its readiness to form defects and surface of the deposited carbon. SEM analyses were conducted by using
oxygen vacancies [11]. Additionally, it is mentioned in literature a Philips XL 30 ESEM–FEG system, having a maximum resolution
that the use of supports with low concentration of Lewis acid sites of 2 nm. The experiments were performed at the Advanced Tech-
and/or that have basic sites, such as ZrO2 , MgO, and La2 O3, lead to nologies Research and Development Center of Boğaziçi University.
high and stable CDRM performance [12]. HRTEM analyses were carried out at the Institute of Materials
In CDRM catalysts, Co is a promising choice as a primary metal at TUBITAK-MAM by using JEOL 2100 LaB6HRTEM operating at
owing to its stable performance, abundance and relatively low cost. 200 kV.
It has been reported that deactivation of Co based catalysts may be In order to understand the nature of interaction between the dis-
due to metal oxidation under reaction conditions as well as coke persed metal species and the support, fresh and spent Co–Ce/ZrO2
deposition. Cobalt oxidation and coke deposition can be avoided catalyst samples were analyzed by XPS by using Thermo Scientific
through optimizing Co loading, using different supports, promot- K-Alpha X-ray Photoelectron Spectrometer at Advanced Technolo-
ers and reaction conditions [13,14]. Introducing a second metal(ic) gies Research and Development Center of Boğaziçi University. All
component to form a bimetallic catalyst system also improves the binding energies were referenced to the C1s line. For data analy-
anti-coking property of catalysts [15]. Ceria and ceria-based sub- sis, the peak intensities were estimated by calculating the integral
strates may enhance the catalytic performance by increasing metal of each peak, after subtraction of the S-shaped Shirley-type back-
dispersion [9]. It has been previously shown by our group that, ground, and by fitting the curve to a combination of Lorentzian
owing to its redox properties enabling release and restore of oxygen (30%) and Gaussian (70%) lines.
atoms, Ce is an effective promoter for CDRM catalysts regulating The coke deposited spent catalyst samples were analyzed by
surface oxygen transfer to active metal sites [16,17]. These oxy- temperature programmed oxidation (TPO) and by Raman spec-
gen species react with methane constituents on the surface of the troscopy. The TPO experiments were performed via an Intelligent
primary metal, which are formed upon CH4 disproportionation, Gravimetric Analyzer (Hiden Isochema) connected online to a
resulting in reduced coke formation [5]. Besides, the addition of Zr dynamic sampling mass spectrometer (Hiden Analytical) for mon-
to CeO2 increases thermal resistance, redox property and oxygen itoring the combustion products with the following procedure: (i)
storage capacity of CeO2 [18]. The spent catalyst sample (∼25 mg) was outgassed at 383 K for
The overall purpose of this initial study, which is the first paper one hour and then at 298 K overnight to eliminate humidity and
of a series, is to propose Co–Ce/ZrO2 system as an effective Co- trapped gasses, (ii) The sample was subjected to 50 mL/min flow
based non-PGM CDRM catalyst, and to find the optimal reaction of 5% O2 –95% He mixture at 1 bar, and then heated to 1023 K with
conditions to be used in production of synthesis gas leading high 2 K/min heating rate in order to complete coke oxidation. Results
and stable activity. In order to increase the oxygen storage capac- from repeated experiments showed that the TPO profiles were
ity of the catalyst, Ce was used as a promoter. The freshly reduced reproducible. Raman spectra of the spent catalysts were obtained
and spent catalyst samples were characterized by using SEM–EDX, by using a Renishaw inVia Raman microscope (Advanced Technolo-
HRTEM–EDX, TPO, Raman Spectroscopy and XPS in order to ana- gies Research and Development Center of Boğaziçi University) with
lyze dispersion of the active phases, oxidation state of Co and Ce, the following operation parameters: 514 nm 20 mW Ar+ laser as the
extent of carbon coverage of the active sites and types of the car- excitation source; laser intensity of ∼2 mW; 5 s acquisition time;
bon formed on those sites. The reaction conditions are optimized a total of 20 accumulation per spectrum. Before measurements,
through the use of an experimentally designed procedure using Raman spectrum was calibrated by using a silicon wafer peak at
reaction temperature, CH4 /CO2 feed ratio and space velocity as the 520 cm−1 . All the samples were analyzed under atmospheric con-
parameters. dition without pre-treatment with the de-focusing technique.

2. Experimental 2.3. Catalytic performance evaluation

2.1. Catalyst preparation and pretreatment Carbon dioxide reforming of methane was carried out in a fixed-
bed down-flow tubular 12 mm ID, 70 cm long quartz microreactor
5% Co–2% Ce/ZrO2 catalyst was prepared for the experiments. under atmospheric pressure over the Co–Ce/ZrO2 catalyst to see
Zirconia support was first ground and sieved to 45–60 mesh size, the effects of temperature, CH4 /CO2 feed ratio and space velocity.
and then calcined at 1073 K for 4 h in muffle furnace for high Prior to each reaction test, the catalyst was calcined and reduced
thermal stability. To obtain Co–Ce/ZrO2 catalyst, impregnation of in situ. The reaction tests were performed at the temperature inter-
aqueous precursor solution of Ce (cerium(III) nitrate hexahydrate) val of 873–973 K with CH4 /CO2 feed ratios of 1/1, 2/1, 1/2 and space
was performed first. The slurry obtained after the impregnation velocities of 30,000, 20,000 and 10,000 mL/h g-catalyst. Hewlett
step was dried at 388 K overnight. This was followed by heat treat- Packard HP5890, temperature-controlled and programmable gas
ment at 773 K for 4 h and impregnation of aqueous cobalt(II) nitrate chromatograph equipped with a thermal conductivity detector
hexahydrate solution. Upon Co impregnation, the slurry was again (TCD) and a HayeSep D column was used for analyzing the feed
dried at 388 K overnight. Prior to the performance tests, the cata- and product streams.
lysts were calcined in situ in dry air (30 mL/min) for 4 h at 773 K and
subsequently reduced in situ in H2 (50 mL/min) for 2 h at the same
3. Results and discussion
temperature.
3.1. Characterization of Co–Ce/ZrO2 catalyst
2.2. Catalyst characterization
Freshly reduced catalyst samples were characterized by
SEM, energy dispersive X-Ray (EDX) and HRTEM tests were con- SEM–EDX and HRTEM aiming to study micro-structural properties
ducted on freshly reduced (i.e., calcined and reduced) Co–Ce/ZrO2 of the catalysts and dispersion of Co and Ce.
166 A.I. Paksoy et al. / Applied Catalysis B: Environmental 168-169 (2015) 164–174

Fig. 3. XP spectrum of Ce 3d region of (a) 5% Co–2% Ce/ZrO2 catalyst tested at 973 K

with 20,000 mL/h g catalyst space velocity and CH4 /CO2 feed ratio of 2/1, (b) 5%
Co–2% Ce/ZrO2 catalyst tested at 873 K with 20,000 mL/h g catalyst space velocity
Fig. 1. Co–Ce–Zr mapping of freshly reduced 5% Co–2% Ce/ZrO2 catalyst. and CH4 /CO2 feed ratio of 1/1, and (c) freshly calcined and reduced 5% Co–2% Ce/ZrO2
catalyst (* : Ce4+ , ** : Ce3+ ).
Table 1
HRTEM-EDX analysis results for the areas in Fig. 2a and b.
887.43 (v), 892.55 (v2 ) and 902.69 (v3 ) eV and three main 3d3/2
Fig. 2 Ce (wt%) Co (wt%) Ce/Co Ce (at%) Co (at%) Ce/Co peaks at about 906.21 (u), 912.5 (u2 ) and 919.79 (u3 ) eV, which
a 6.57 1.83 3.59 1.23 0.82 1.5 are attributed to the Ce4+ state, while the peaks at about 883.01
b 5.06 3.24 1.56 1.26 1.92 0.86 (v0 ), 890.21 (v1 ), 904.67 (u0 ) and 909.12 (u1 ) eV belong to unique
photoelectron features of the Ce3+ state [19].
After deconvolution of the spectrum, the degree of ceria reduc-
The SEM–EDX mapping image (Fig. 1) shows well dispersed
tion was evaluated from the ratio of the sum of integrated peak
metal particles, Co and Ce, on the freshly-reduced catalyst sam-
areas of v0 , u0 , u1 and v1 peaks to the sum of the integrated peak
ple. HRTEM results at higher magnifications (Fig. 2), indicated local
areas of all peaks. According to the given equation (Eq. (1)),
change in Ce/Co surface ratio. Average Ce/Co surface ratio cal-
culated by using HRTEM–EDX data on different locations of two I − Ce3+
freshly reduced samples is 1.7, and the highest observed ratio is [Ce3+ ] = (1)
(I − Ce3+ + I − Ce4+ )
3.59. The detailed analysis results of the two selected locations,
shown in Fig. 2, are given in Table 1 as examples. where I-Ce3+ and I-Ce4+ represent the sum of intensities of two dou-
The Ce 3d XP spectrum of the freshly reduced and spent blets resulting from Ce2 O3 and three doublets resulting from CeO2 ,
Co–Ce/ZrO2 catalyst samples were obtained and analyzed aiming respectively [19], the amount of Ce3+ present on freshly reduced
to understand the redox ability of CeOx formations in the freshly Co–Ce/ZrO2 catalyst is estimated as 11.6%.
reduced sample and how it is changed during the reaction. XP spec- The binding energies attributed to Ce4+ and Ce3+ states for the
tra of the spent catalyst samples which were used under reaction spent catalyst samples are given in Table 2. Peaks belonging to
conditions yielding the highest, at 973 K with CH4 /CO2 feed ratio of both spent catalyst samples were observed to have a shift to lower
2, and the lowest, at 873 K with CH4 /CO2 feed ratio of 1, coke for- binding energies when compared to the peaks from the freshly
mation are compared with that obtained from the freshly reduced reduced catalyst sample. However, all binding energies found from
sample. the deconvolution analysis are in accordance with literature indi-
The Ce 3d XPS results of the freshly reduced sample are in accor- cating there is no significant possibility of alloy formation [19].
dance with the literature (Fig. 3) but peaks are slightly shifted to The deconvolution analysis indicated that the degree of ceria
higher binding energies. There are three main 3d5/2 peaks at about reduction is 13.8% for the catalyst tested at 973 K with 2/1 CH4 /CO2

Fig. 2. HRTEM images of selected regions on freshly reduced catalyst sample.

 A.I. Paksoy et al. / Applied Catalysis B: Environmental 168-169 (2015) 164–174 167

Table 2 When the surface metal site is covered even by a thin carbon
The Ce valance-binding energy (eV) relation for different reaction conditions.
layer, this site is not seen in metal mapping and the metal’s sur-
Reaction Ce valance Binding energy, face concentration, evaluated by EDX, decreases. Considering this
conditions (eV) fact, a simple methodology, which utilizes a comparative anal-
T = 873 K, Ce4+ 919.86, 912.03, 906.63, 902.84, 892.88, 886.88 ysis of Co and Ce site concentrations measured on the freshly
CH4 /CO2 = 1/1 reduced and spent samples through metal mapping (Fig. 5) and
Ce3+ 909.78, 905.07, 891.02, 882.02 SEM–EDX, was used to investigate the functions of Co and Ce
T = 973 K, Ce4+ 918.06, 908.8, 902.69, 899.29, 889.63, 884.01
sites in CDRM. The idea behind the methodology is the fact that
CH4 /CO2 = 2/1
Ce3+ 905.49, 900.86, 886.72, 880.42 carbon covers the site whose primary function is hydrogen produc-
tion from methane/methyl groups, like methane dehydrogenation,
forming carbon as the side product. A mild set of CDRM condi-
feed ratio, but it is 8.7% for the one tested at 873 K with 1/1 CH4 /CO2 tions, 873 K with the CH4 /CO2 feed ratio of 1/1 and space velocity of
feed ratio, proving Ce’s ability to store oxygen when the feed is rich 20,000 mL/h g catalyst, leading very limited carbon formation was
in oxygen source; i.e., at higher CO2 concentrations in the feed. This used in preparation of the spent sample. A comparative analysis of
eases the utilization of surface oxygen; during reaction, Ce acts as a EDX results showed that carbon formation is more pronounced on
buffer regulating surface oxygen transfer from support to Co sites Co sites. The EDX results, on the average, showed that the atomic
through redox cycle. percentage of Ce is decreased from 0.55 to 0.47 upon 6 h TOS CDRM
Fig. 4 shows the Co 2p XP spectra obtained from freshly reduced test, meaning ca. 15% of the surface Ce is covered by carbon. On the
and spent catalyst samples. The two peaks, appearing in the spec- other hand, under the same conditions, the decrease in Co atomic
trum of the freshly reduced sample (Fig. 4c) at 778.8 eV and percentage is from 4.1 to 2.28, meaning ca. 44% of Co is covered by
793.5 eV, separated by 14.7 eV, may be attributed to the Co 2p3/2 carbon. The results clearly showed carbon blockage is significantly
and Co 2p1/2 spin–orbit peaks, respectively, of the Co0 phase, indi- more severe on Co centers, confirming the primary function of the
cating the metallic nature of the reduced catalyst surface. However, Co sites is hydrogen production, whose side product is deposited
as the FWHM of the Co0 peak is <1.5 eV, in addition to the other carbon. The Co and Ce metal mapping images also confirmed this
deconvoluted peaks corresponding to the main and satellite peaks outcome. It should be noted that the carbon covered Ce sites most
of the Co2+ and Co3+ phases, the features attributed to the metal probably are the ones present in close vicinity of the Co sites on
phase are also thought to contain oxide phases [20–23]. The Co which carbon formed may extend to the neighboring Ce centers.
2p XP spectra of both spent catalyst samples informs us about the It is known that the type and amount of coke deposited on
variation of the Co oxidation state during reaction. The absence the catalyst during CDRM plays an important role in the catalyst
of Co0 2p3/2 and 2p1/2 spin–orbit peaks indicates, in addition to performance through affecting its activity and selectivity. In order
the oxide phases present on the freshly reduced catalyst surface, to understand the effect of local surface Ce concentration on pre-
there is further Co oxide formation during reaction. As can be seen vention of coke formation, SEM–EDX analyses were performed on
from Fig. 4a and b, the spectra of the spent samples used in reac- spent catalysts used in reaction tests at temperatures of 873 and
tions conducted at 973 and 873 K, respectively, involve Co 2p3/2 973 K, CH4 /CO2 feed ratios of 1/1 and 2/1 with the space velocity of
and Co 2p1/2 spin–orbit peaks and their satellite features belonging 20,000 mL/h g catalyst.
to cobalt oxide phases. It is very difficult to distinguish between Two distinct, namely coked and coke-free, regions observed for
the Co2+ and Co3+ oxide phases due to the small shifts (∼0.5 eV) each one of the spent Co–Ce catalyst samples, tested under four
between their peaks. On the other hand, as it has been reported different reaction conditions, are presented as examples (Fig. 6).
that when Co is deposited onto other oxides the XPS Co 2p lines EDX analyses (Table 3) performed for both regions on each sample
of Co2+ shift to higher BE values, and Co2+ has characteristic satel- reveal that Ce/Co surface molar ratio of coked parts is lower than
lites which are more pronounced in CoO than in Co3 O4 [20,23], the that of the coke-free parts. The results confirmed that the pres-
peaks shown in Fig. 4 are attributed to Co2+ phase. ence of ceria on the surface creates an additional storage capacity
for oxygen coming from ZrO2 support, which is known to catalyze
CO2 disintegration generating surface oxygen [24]. Ce centers go
through continuous reduction/oxidation cycle during the reaction;
they act as a buffer for surface oxygen, regulating and enhancing
oxygen transfer to Co centers [17]. This transfer removes deposited
carbon through oxidation and makes catalytic centers more carbon
resistant. It is observed from the images that formation of carbon
is favored for the regions with low Ce/Co ratio. It should be noted
that there may be also a possibility of increased Co–Ce interaction.
Many studies have shown that during CDRM, surface oxygen
(or oxygen-containing species) originating from CO2 , oxidize car-
bon species formed by CH4 decomposition on the metallic sites.
Therefore, the rate of carbon accumulation on the catalyst sur-
face is determined by the relative rates of the generation and
oxidative removal of carbon species. Excess carbon deposition will
occur when the former is faster than the latter [13,20]. SEM analyses
(Fig. 7) obtained from the spent samples tested at different temper-
ature and CH4 /CO2 feed ratios gave information about the relation
between reaction conditions, yielding changes in the relative rates
of carbon generation and removal, and the type and abundance of
Fig. 4. XP spectrum of Co 2p region of (a) 5% Co–2% Ce/ZrO2 catalyst tested at 973 K carbon deposition.
with 20,000 mL/h g catalyst space velocity and CH4 /CO2 feed ratio of 2/1, (b) 5%
High level of carbon deposition was observed on the spent sam-
Co–2% Ce/ZrO2 catalyst tested at 873 K with 20,000 mL/h g catalyst space velocity
and CH4 /CO2 feed ratio of 1/1, and (c) freshly calcined and reduced 5% Co–2% Ce/ZrO2 ple tested at 973 K with 1/1 CH4 /CO2 feed ratio, indicating that
catalyst. condition favors carbon formation more than carbon oxidation.
168 A.I. Paksoy et al. / Applied Catalysis B: Environmental 168-169 (2015) 164–174

When the CH4 /CO2 feed ratio was increased to 2/1 while keeping tion was observed. In addition to the temperature effect, another
the temperature and space velocity fixed, the highest coke deposi- reason for higher coke accumulation is the higher amount of carbon

Fig. 5. Co–Ce mapping of 5% Co–2% Ce/ZrO2 catalyst (a) Co mapping of freshly reduced catalyst, (b) Co mapping of spent catalyst sample used during the reaction performed at
873 K and CH4 /CO2 = 1/1, (c) Ce mapping of freshly reduced catalyst, (d) Ce mapping of spent catalyst sample used during the reaction performed at 873 K and CH4 /CO2 = 1/1.

Fig. 6. SEM images of Co–Ce catalyst samples used during the reactions performed at (a) 973 K and CH4 /CO2 = 1/1, (b) 973 K and CH4 /CO2 = 2/1, (c) 873 K and CH4 /CO2 = 1/1,
(d) 873 K and CH4 /CO2 = 2/1 (the regions for which EDX analyses were performed are shown by circles for coke-free regions and by rectangles for coked regions).
 A.I. Paksoy et al. / Applied Catalysis B: Environmental 168-169 (2015) 164–174 169

Table 3
Ce and Co surface atomic percentages of the spent catalysts obtained by SEM–EDX analysis.

Reaction Ce (%) coked Co (%) coked Ce/Co coked part Ce (%) Co (%) Ce/Co coke-free
conditions part part coke-free part coke-free part part

T = 973 K, CH4 /CO2 = 1/1 0.35 1.25 0.28 1.89 3.10 0.61
T = 973 K, CH4 /CO2 = 2/1 0.43 0.97 0.44 1.42 3.15 0.45
T = 873 K, CH4 /CO2 = 1/1 0.56 2.03 0.28 1.39 2.60 0.54
T = 873 K, CH4 /CO2 = 2/1 0.24 0.89 0.27 1.40 2.79 0.50

Fig. 7. SEM images of the catalyst samples used during the reaction at space velocity of 20,000 mL/h g-catalyst and at (a) 973 K, CH4 /CO2 = 1/1 (b) 973 K, CH4 /CO2 = 2/1 (c)
873 K, CH4 /CO2 = 1/1 (d) 873 K, CH4 /CO2 = 2/1.

source in the feed relative to the oxygen source, led by the increased symmetry, respectively [20,25,26]. The relative intensity of D and
methane concentration in the feed. Consequently, coke deposition G bands (ID /IG ) gives information about the degree of crystallinity;
is strongly favored when CH4 /CO2 feed ratio and reaction temper- smaller ID /IG value indicates higher crystallinity [20]. The ID /IG val-
ature are both high.
On the other hand, it was seen that coke formation decreased
with a decrease in reaction temperature in 873–973 K temperature
region. From the SEM image of the sample tested at 873 K, with
the CH4 /CO2 feed ratio of 1/1 and space velocity of 20,000 mL/h g-
catalyst, the lowest carbon deposition was observed. It should be
noted that Boudouard reaction favored at low temperatures was
thought not significant except CDRM tests conducted at 873 K.
Although the temperature parameter was still 873 K, the increased,
2/1, ratio of CH4 /CO2 in the feed caused carbon formation, indicat-
ing relatively limited carbon oxidation activity.
The structure of the deposited carbon formed on the spent cat-
alyst samples was further characterized by Raman spectroscopy
and temperature programmed oxidation (TPO). The catalyst sam-
ples yielding the highest, at 973 K with CH4 /CO2 feed ratio of 2, and
the lowest, at 873 K with CH4 /CO2 feed ratio of 1, coke formation
were chosen for analysis. The Raman spectra of both spent catalyst
samples given in Fig. 8 show two well defined bands at around 1344
and 1580 cm−1 , which are attributed to the D band, associated with
Fig. 8. Raman spectra of the catalyst samples used during the reaction at space
the disordered structural mode of crystalline carbon species, and
velocity of 20,000 mL/h g-catalyst and at (black curve) 973 K, CH4 /CO2 = 2/1; (grey
G band, corresponding to the graphitic carbon with high degree of curve) 873 K, CH4 /CO2 = 1/1.
170 A.I. Paksoy et al. / Applied Catalysis B: Environmental 168-169 (2015) 164–174

Fig. 9. TPO of the catalyst samples used during the reaction at space velocity of 20,000 mL/h g-catalyst and at (black curve) 973 K, CH4 /CO2 = 2/1; (grey curve) 873 K,
CH4 /CO2 = 1/1.

ues were calculated as 0.83 and 1.13 for the catalyst samples spent No severe activity loss was observed during the performed
at 873 and 973 K, respectively; pointing out that the coke formed experiments. The CH4 conversion results of the reaction tests car-
during reaction at 873 K has higher degree of graphitization. ried at CH4 /CO2 feed ratio of 2/1, which was expected to cause the
In TPO tests, it is widely accepted that amorphous carbon species highest activity loss due to the higher loading of CH4 , indicated
are combusted at lower temperatures, i.e., below 773 K, than crys- only 25% decrease at most in terms of methane conversion during
talline ones [20,26–28]. The TPO profile (Fig. 9) of the catalyst tested the time-on-stream (TOS) experiments (Fig. 10). This activity loss
at 973 K with 2/1 CH4 /CO2 ratio, reveals a wide and asymmetric was consistent with the carbon deposition observed in SEM anal-
peak that begins around 550 K and ends around 750 K. According ysis (Fig. 7). Main activity loss was observed for the first 15 min
to the deconvolution results this wide peak encompasses 4 peaks TOS for all temperature levels indicating coke formation was more
at 628, 652, 673 and 703 K which are all attributed to the amour- rapid than carbon oxidation reaction on the freshly reduced sam-
phous coke. On the other hand, in accordance with the Raman ple, leading to an activity decrease (Fig. 10). When the rates of
spectroscopy results, the TPO profile of the catalyst used at 873 K methane dehydrogenation and oxidation of carbon formed on Co
records a wider peak between 670 and 900 K, with 3 major peaks sites become compatible, the activity level is stabilized. The activity
at 758, 810 and 838 K. The higher oxidation temperatures of the and selectivity profiles indicate that the performance of the catalyst
deposited carbon on that sample indicates an increase in degree is very stable when the CH4 /CO2 feed ratio is 1/1 for the whole tem-
of crystallinity in the carbon structure. The peak evolved at 758 K perature range except at 873 K, which hints limited carbon removal
may be attributed to monoatomic coke, and the latter two major
peaks may be assigned to filamentous coke with different degree
of crystallization, getting closer to pure grafitic structures with the
increase in oxidation temperature [27]. Thus, a combined evalua-
tion of SEM images, Raman results and TPO profiles of spent catalyst
samples clearly confirms the formation of different carbon struc-
tures on the catalyst samples during CDRM, and indicates that the
fraction of different carbon types depends on the reaction condi-
tions.

3.2. Performance tests

A parametric CDRM performance study was conducted on

Co–Ce/ZrO2 system considering reaction temperature, CH4 /CO2
feed ratio and space velocity as the parameters. H2 /CO ratio in
the product stream was used as the measure of selectivity. As the
syngas fed to Fischer–Tropsch reactor yields valuable olefins when
its H2 /CO ratio is close to 1, obtaining H2 /CO ratio in CDRM prod-
uct closer to 1 is widely accepted advantageous. The activity of the
catalyst was studied and discussed in terms of both CH4 and CO2 Fig. 10. Time-on-stream activity profiles of the catalyst in terms of CH4 conversion
conversions, and H2 yield achieved. for CH4 /CO2 feed ratio of 2/1.
 A.I. Paksoy et al. / Applied Catalysis B: Environmental 168-169 (2015) 164–174 171

Fig. 11. Time-on-stream activity and selectivity profiles of the catalyst in terms of CH4 conversion and H2 /CO product ratio, respectively, for CH4 /CO2 feed ratio of 1/1 (filled
symbols: conversion; hollow symbols: selectivity).

ability compared to methane dehydrogenation at low temperatures conversion values (Fig. 12a and b) increased with the increase in
(Fig. 11). temperature. At CH4 /CO2 feed ratio of 1/1, e.g., methane conver-
sions at the end of 6 h TOS were 30%, 61% and 73% for the reaction
3.2.1. The effect of reaction temperature temperatures of 873 K, 923 K and 973 K, respectively (Fig. 12a). Car-
The effect of temperature on catalyst performance was ana- bon dioxide conversions, which were higher than that of methane,
lyzed for the temperature range of 873–973 K in a detailed fashion were measured as 41%, 70% and 83% at 873 K, 923 K and 973 K,
by using feeds having CH4 /CO2 ratios of 1/2, 1/1 and 2/1 for respectively (Fig. 12b). It should be noted that the percentage
20,000 mL/h g-catalyst space velocity. decline in activity with reaction time was inversely proportional
The effects of temperature and CH4 /CO2 feed ratio on CH4 and to increase in reaction temperature. Therefore, it can be concluded
CO2 conversion, CH4 /CO2 activity ratio and H2 /CO product ratio that surface oxygen generation rate increases sharply relative to
are shown in Fig. 12a–d based on the performance values obtained carbon formation with an increase in temperature. Additionally,
at the end of 6 h. TOS. For all cases studied, both CH4 and CO2

Fig. 12. The effect of reaction temperature and CH4 /CO2 feed ratio on (a) CH4 conversion (b) CO2 conversion (c) CH4 activity/CO2 activity ratio (d) H2 /CO product ratio (All
values presented were obtained at the end of 6 h. TOS).
172 A.I. Paksoy et al. / Applied Catalysis B: Environmental 168-169 (2015) 164–174

ature may indicate the supression of Boudouard reaction at higher

temperatures.
Conclusively, it was clearly observed that increasing tem-
perature resulted in increasing activity and selectivity towards
hydrogen as well as increasing coke deposition for Ce–Co catalyst
in CDRM.

3.2.2. The effect of CH4 /CO2 feed ratio

The effect of CH4 /CO2 feed ratio on the activity and selectivity
of the Co–Ce catalyst has been studied in a detailed fashion. For the
reactions carried out at the same temperature and space velocity,
CH4 activity losses over time were lower at the CH4 /CO2 feed ratio
of 1/2 and 1/1 but significantly high for the feed ratio of 2/1 (Fig. 14).
As reaction proceeded, CH4 conversion decreased by only 3% for the
feed ratios of 1/2 and 1/1, whereas the decrease was 15% for the feed
ratio of 2/1 at 973 K. The limited activity loss at lower CH4 /CO2
Fig. 13. Time-on-stream profile of H2 yield (H2 produced/CH4 consumed) obtained feed ratios can be attributed to the higher loading of the oxygen
at CH4 /CO2 feed ratio of 1/1 and 20,000 mL/h g catalyst.
source, carbon dioxide. With increased CO2 feeding rate, carbon
formed as the side product of CDRM could be easily cleaned away,
resulting in less coke deposition and therefore limited activity loss.
the activation of surface carbon formed may be more favorable at This effect is also in agreement with the results of SEM–EDX anal-
higher reaction temperatures. ysis showing the effect of CH4 /CO2 feed ratio on deposited carbon
The temperature increase resulted in an increase in the ratio of (Figs. 6 and 7).
CH4 activity to CO2 activity as the methane decomposition reaction The CH4 /CO2 feed ratio of 1/2 yielded the highest methane con-
favored at high temperatures (Fig. 12c). The highest observed ratio version values, whereas the lowest carbon dioxide conversion and
was 1.35 and it was obtained for the reaction at 973 K with CH4 /CO2 H2 /CO product ratio values were obtained for the whole tempera-
feed ratio of 2/1. Consequently, both H2 yield and H2 /CO product ture range for that feed ratio (Fig. 12a, b and d). CH4 /CO2 feed ratio
ratio (Fig. 12d) increase with the increase in temperature. Stabi- of 2/1, on the other hand, had an opposite effect on catalyst activ-
lized H2 /CO product ratio, at the end of 6 h TOS, was 0.64 at 973 K, ity and selectivity; it yielded the highest carbon dioxide conversion
0.59 at 923 K and 0.45 at 873 K for the CH4 /CO2 feed ratio of 1/1. values and H2 /CO product ratios with the lowest methane conver-
On the other hand, increase in CH4 activity relative to that of CO2 sions observed for the conditions tested. For reactions conducted
caused higher level of coke deposition which may lead to reactor at 973 K, methane conversion values were 82%, 73% and 54% for
blockage. the feed ratios of 1/2, 1/1 and 2/1, respectively (Fig. 12a). Carbon
TOS data for H2 yield (Fig. 13) for the reactions carried out dioxide conversion values, on the other hand, were 67% at 1/2 feed
between 873–973 K at CH4 /CO2 feed ratio of 1/1 reflect the effect ratio, 82% at 1/1 feed ratio and 83% at 2/1 feed ratio (Fig. 12b). At
of coke deposition. It is evident that coke may form either through the end of 6-h-on-stream, the highest H2 /CO product ratio, 0.68,
methane dehydrogenation or Boudouard reaction. At the begin- was obtained for 2/1 feed ratio whereas the lowest ratio, 0.53, was
ning of the reaction, surface concentrations of the species are recorded for the reaction with 1/2 feed ratio (Fig. 12d). It was clearly
stabilised and the effect of temperature could not be fully observed; observed that both conversion and H2 /CO product ratio values were
as an example, the yield at the end of 1st hour-on-stream was sufficiently high for the tests conducted using 1/1 feed ratio.
the highest for the test conducted at 873 K. However, with the When CH4 /CO2 feed ratio was increased, because of higher
increase in TOS, Boudouard reaction favored at low temperatures loading of hydrocarbon, CH4 activity increased while CO2 activity
may become relatively significant leading lower H2 yields at 873 K decreased (Fig. 14), leading to an increase in CH4 /CO2 activity ratio
compared to yields at higher temperatures. The slight increase (see Fig. 12c). Considering also the increase in H2 yield, based on
in “CH4 activity/CO2 activity” ratio with temperature (Fig. 12c) CH4 fed (not shown), one can explain the enhanced H2 selectivity in
supports the dominance of CH4 dehydrogenation, while relatively the product stream. However, the increase in CH4 /CO2 activity ratio
higher increase in H2 /CO product ratio with the increase in temper- caused coke deposition, as there was not enough oxygen source to

Fig. 14. CH4 and CO2 activities at 973 K for 20,000 mL/h g catalyst (filled symbols: CH4 activity; hollow symbols: CO2 activity).
 A.I. Paksoy et al. / Applied Catalysis B: Environmental 168-169 (2015) 164–174 173

Fig. 15. Activity and selectivity trends at 973 K as a function of CH4 /CO2 feed ratios and space velocities (a) CH4 conversions (b) CO2 conversions (c) H2 /CO product ratio.

clean away the formed carbon and, therefore, H2 yield, based on from 10,000 to 30,000 mL/h g-catalyst, H2 /CO product ratio only
CH4 consumed, was decreased. decreased from 0.66 to 0.64 and 0.68 to 0.67 for CH4 /CO2 feed ratios
It was clear that CH4 /CO2 feeding ratio has a serious effect on of 1/1 and 2/1, respectively. However, H2 /CO product ratio became
both activity and selectivity of the Ce–Co catalyst. Evaluating activ- 0.49 for CH4 /CO2 feed ratio of 1/2 at 30,000 mL/h g-catalyst, which
ities and coke deposition in a combined fashion indicates that 1/1 was 0.56 at 10,000 mL/h g-catalyst.
feed ratio can be taken as a suitable value for the current catalyst. Based on the results of the experiments carried out, the increase
in space velocity resulted in a decrease in both CH4 and CO2 conver-
sions and H2 /CO product ratio for all CH4 /CO2 feed ratios applied.
3.2.3. The effect of space velocity
The effect of space velocity on the CDRM performance of 3.2.4. Stability Test
Co–Ce system was investigated for three different space velocities, The 72 h stability test was performed over 5% Co–2% Ce–ZrO2
10,000 mL/h g-catalyst, 20,000 mL/h g-catalyst and 30,000 mL/h g- at 873 K with the CH4 /CO2 feed ratio of 1/1 and space velocity of
catalyst at 973 K for CH4 /CO2 feed ratio values between 0.5 and 2 20,000 mL/h g catalyst. Despite very limited coke formation, which
(Fig. 15). Similar activity trends were observed for all space veloci- was characterized by TPO and Raman analysis, the catalyst showed
ties tested (not shown). no deactivation (Fig. 16) during 72-h-on-stream.
As the space velocity increased from 10,000 mL/h g-catalyst
to 30,000 mL/h g-catalyst, both CH4 and CO2 conversion values
decreased for all feed compositions (Fig. 15a and b). The decrease
was 5% for the CH4 /CO2 feed ratio of 1/1 and most of the decrease
was observed when the space velocity was changed from 10,000
to 20,000 mL/h g-catalyst. At the end of 6th h-on-stream, methane
conversions for the feed ratio of 1/1 were recorded as 78%, 73% and
71% for the space velocities of 10,000, 20,000 and 30,000 mL/h g-
catalyst, respectively.
It was interesting to notice that the effect of space velocity
was observed more significantly as the CH4 /CO2 feed ratio was
decreased. This can be explained by the comparative analysis of
methane dehydrogenation and surface carbon oxidation rates. The
increase in oxygen source resulted in closer oxidation and dissoci-
ation rates leading stable CDRM activity.
There were no major changes observed for H2 /CO product ratio
for the CH4 /CO2 feed ratios of 1/1 and 2/1 in response to changes Fig. 16. Stability test over 5% Co–2% Ce/ZrO2 (873 K, CH4 /CO2 = 1/1, space veloc-
in space velocity (Fig. 15c). As the space velocity was increased ity = 20000 mL/h g catalyst).
174 A.I. Paksoy et al. / Applied Catalysis B: Environmental 168-169 (2015) 164–174

4. Conclusions [3] L.A. Arkatova, Catal. Today 157 (2010) 170–176.

[4] G.S. Gallego, F. Mondragon, J. Tatibouet, J. Barrault, C. Batiot-Dupeyrat, Catal.
Co–Ce/ZrO2 catalyst was characterized and tested in CDRM for Today 133–135 (2008) 200–209.
[5] S. Corthals, J. Van Nederkassel, J. Geboers, H. De Winne, J. Van Noyen, B.
different reaction conditions. The following conclusions can be Moens, B. Sels, P. Jacobs, Catal. Today 138 (2008) 28–32.
drawn according to the characterization and catalytic performance [6] Z. Hou, P. Chen, H. Fang, X. Zheng, T. Yashima, Int. J. Hydrogen Energy 31
data: (2006) 555–561.
[7] D. Liu, W.N.E. Cheo, Y.W.Y. Lim, A. Borgna, R. Lau, Y. Yang, Catal. Today 154
(2010) 229–236.
- The comparative analysis of Co and Ce site concentrations mea-
[8] R. Bouarab, O. Akdim, A. Auroux, O. Cherifi, C. Mirodatos, Appl. Catal. A: Gen.
sured on the freshly reduced and spent samples through metal 264 (2004) 161–168.
mapping and SEM–EDX indirectly confirmed that methane dehy- [9] S. Damyanova, B. Pawelec, K. Arishtirova, M.V. Martinez Huerta, J.L.G. Fierro,
drogenation is the primary function of Co sites. Appl. Catal. B: Environ. 89 (2009) 149–159.
[10] A.D. Ballarini, S.R. de Miguel, E.L. Jablonski, O.A. Scelza, A.A. Castro, Catal.
- Combined evaluation of SEM images, Raman results and TPO pro- Today 107–108 (2005) 481–486.
files of spent catalyst samples clearly showed the formation of [11] J.D.A. Bellido, E.M. Assaf, Appl. Catal. A: Gen. 352 (2009) 179–187.
different carbon structures on the catalyst samples during CDRM, [12] M. Rezaei, S.M. Alavi, S. Sahebdelfar, Z. Yan, Scr. Mater. 61 (2009)
173–176.
and indicated that the fraction of different carbon types depends [13] E. Ruckenstein, H.Y. Wang, J. Catal. 205 (2002) 289–293.
on the reaction conditions. [14] D. San-Joseı́ Alonso, J. Juan-Juan, M.J. Illan-Gomez, C. Roman-Martinez, Appl.
- XPS Ce 3d results revealed that ceria creates an additional oxy- Catal. A: Gen. 371 (2009) 54–59.
[15] J. Xu, W. Zhou, Z. Li, J. Wang, J. Ma, Int. J. Hydrogen Energy 34 (2009)
gen storage capacity resulting in an enhanced surface oxygen 6646–6654.
storage/transfer function through redox cycle and higher carbon [16] S. Ozkara-Aydınoglu, E. Ozensoy, A.E. Aksoylu, Int. J. Hydrogen Energy 34
resistance. (2009) 9711–9722.
[17] S. Ozkara-Aydınoglu, A.E. Aksoylu, Catal. Commun. 11 (2009) 1165–1170.
- Co–Ce/ZrO2 catalyst has stable activity for CH4 /CO2 feed ratio of 1. [18] L.V. Mattos, E. Rodino, D.E. Resasco, F.B. Passos, F.B. Noronha, Fuel Process.
Coke accumulation was mostly observed when temperature and Technol. 83 (2003) 147–161.
CH4 /CO2 feed ratio were both high. Decreasing space velocity is [19] P. Leppet, B. Schumacher, V. Plzak, M. Kinne, R.J. Behm, J. Catal. 244 (2006)
137–152.
beneficial for high and stable activities and high H2 /CO product
[20] M. Dominguez, E. Taboada, H. Idriss, E. Molins, J. Llorca, J. Mater. Chem. 20
ratio. (2010) 4875–4883.
[21] J.W. Chai, J.S. Pan, S.J. Wang, C.H.A. Huan, G.S. Lau, Y.B. Zheng, S. Xu, Surf. Sci.
589 (2005) 32–41.
Acknowledgements
[22] S.S.Y. Lin, D.H. Kim, M.H. Engelhard, S.Y. Ha, J. Catal. 273 (2010)
229–235.
This work is financially supported by TUBITAK through project [23] R. Watanabe, Y. Hondo, K. Mukawa, C. Fukuhara, E. Kikuchi, Y. Sekine, J. Mol.
111M144, Boğaziçi University through project BAP-M6755, and Catal. A: Chem. 377 (2013) 74–84.
by State Planning Organization of Turkey through project DPT [24] S. Wang, G.Q. Lu, Appl. Catal. B: Environ. 19 (1998) 267–277.
[25] D.S. Wragg, A. Gronvold, P. Norby, H. Fjellvag, Microporous Mesoporous
07K120630. Mater. 173 (2013) 166–174.
[26] F.F. de Sousa, H.S.A. de Sousa, A.C. Oliveira, M.C.C. Junioe, A.P. Ayala, E.B.
References Barros, B.C. Viana, J.M. Filho, A.A. Oliveira, Int. J. Hydrogen Energy 37 (2012)
3201–3212.
[1] D. Cheng, X. Zhu, Y. Ben, F. He, L. Cui, C. Liu, Catal. Today 115 (2006) [27] J. Vicente, J. Erena, C. Montero, M.J. Azkoiti, J. Bilnoa, A.G. Gayubo, Int. J.
205–210. Hydrogen Energy 39 (2014) 18820–18834.
[2] G.S. Gallego, C. Batiot-Dupeyrat, J. Barrault, E. Florez, F. Mondragon, Appl. [28] C.E. Daza, A. Kiennemann, S. Moreno, R. Molina, Appl. Catal. A: Gen. 364
Catal. A: Gen. 334 (2008) 251–258. (2009) 65–74.