Methanation of carbon dioxide on metal-promoted mesostructured silica nanoparticles

Article abstract of DOI:10.1016/j.apcata.2014.08.022

Metal-promoted mesostructured silica nanoparticles (MSN) have been studied for CO₂methanation under atmospheric pressure. In term of activities, high activity was observed on Rh/MSN, Ru/MSN, Ni/MSN, Ir/MSN, Fe/MSN and Cu/MSN at and above 623 K.

Full text of DOI:10.1016/j.apcata.2014.08.022

Applied Catalysis A: General 486 (2014) 115–122
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Methanation of carbon dioxide on metal-promoted mesostructured
silica nanoparticles
M.A.A. Aziza, A.A. Jalila,b, S. Triwahyonoc,∗, S.M. Sidika
a Institute of Hydrogen Economy, Universiti Teknologi Malaysia, 81310 UTM Johor Bahru, Johor, Malaysia
b Department of Chemical Engineering, Faculty of Chemical Engineering, Universiti Teknologi Malaysia, 81310 UTM Johor Bahru, Johor, Malaysia
c Department of Chemistry, Faculty of Science, Universiti Teknologi Malaysia, 81310 UTM Johor Bahru, Johor, Malaysia
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 June 2014
Received in revised form 5 August 2014
Accepted 19 August 2014
Available online 28 August 2014
Metal-promoted mesostructured silica nanoparticles (MSN) have been studied for CO2 methanation
under atmospheric pressure. In term of activities, high activity was observed on Rh/MSN, Ru/MSN,
Ni/MSN, Ir/MSN, Fe/MSN and Cu/MSN at and above 623 K. However, on an areal basis, Ni/MSN was the
most active catalyst, while Ir/MSN was the poorest catalyst. The catalysts have also been studied for elu-
cidation of the role of each metal, MSN and metal/MSN in CO2 methanation by in situ FTIR spectroscopy
studies. Firstly, CO2 and H2 was adsorbed and dissociated on metal sites to form CO, O and H atoms,
followed by migration onto the MSN surface. The dissociated CO then interacted with oxide surfaces of
MSN to form bridged carbonyl and linear carbonyl, while the presence of H atom facilitated the formation
of bidentate formate. These three species could be responsible for the formation of methane. However,
the bidentate formate species could be the main route to formation of methane. MSN support has been
found to play an important role in the mechanism. MSN support served the sites for carbonyl species
which act as precursors to methane formation. These results provided new perspectives in the catalysis,
particularly in the recycling of CO2.
Keywords:
CO2 methanation
Mesostructured silica nanoparticles
Metal site
Basic sites
Oxygen vacancy site
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
Ni catalysts have been focused on the reaction performed at high
temperatures (>473 K) [7,8]. In addition to Ni catalysts, CO2 metha-
The growing interest on greenhouse gas mitigation implies that
CO2 will be increasingly considered as a valuable feedstock for
chemical industry instead of as a waste [1]. The valorization of car-
bon dioxide emissions could be one important part of the general
strategy for reducing CO2 emissions, and push chemical and energy
companies towards a more sustainable use of the resources. There
are several ways of adding value to CO2, one of which is hydro-
genation to form methane. Provided that hydrogen is generated
from renewable energy sources such as water electrolysis [2], the
methane produced by this reaction can be sent to chemical indus-
try or can be utilized as an energy vector, which is an important
advantage considering that the infrastructure for methane storage
and transport is already present. CO2 methanation has been inves-
tigated over a variety of supported metal catalysts, including Ni
[3–11], Ru [12,13], Rh [14–17], Fe [18] and Cu [19]. Nickel catalysts
were found to exhibit high activity for the reaction [4,5]. How-
ever, most of the studies concerning the methanation of CO2 over
nation has been studied by several authors using different noble
metals such as Ru, Pd and Rh over a variety of supports. Among
them, Rh has been shown to be one of the most promising metal
catalyst that is active in low-temperature of reactions [14,16].
There is still no consensus about the mechanistic path of CO2
methanation, even though it is a comparatively simple reaction.
Thus, its reaction mechanism appears to be difficult to establish.
In general, the proposed mechanism in the previous work fall into
two main categories. The first mechanism involves the conversion
of CO2 to CO prior to methanation [20–22]. The other involves the
direct hydrogenation of CO2 to methane without forming CO as an
intermediate [23,24]. The present work tends to fall into the first
category. However, there are still different opinions on the nature
of the intermediate and the methane formation process. For exam-
ple, Betta and Shellef demonstrated the presence of formate in the
hydrogenation of CO on a Ru/Al2O3 catalyst by IR spectroscopy.
It was claimed that the formate ion was formed directly on the
alumina support and its formation did not require metallic Ru [25].
Karelovic et al. studied CO2 adsorption and dissociation on a Rh/␥-
Al2O3 system in the presence of H2 by operando-diffuse reflectance
Fourier transform infrared spectroscopy (DRIFTS) experiments.
∗
Corresponding author. Tel.: +60 7 5536076; fax: +60 7 5536080.
E-mail addresses: sugeng@utm.my, sugengtw@gmail.com (S. Triwahyono).
http://dx.doi.org/10.1016/j.apcata.2014.08.022
0926-860X/© 2014 Elsevier B.V. All rights reserved.

116
M.A.A. Aziz et al. / Applied Catalysis A: General 486 (2014) 115–122
Their results showed that the linear Rh-CO species act as precur-
sors of methane [14]. The mechanism of CO2 methanation over Ni
catalyst has been discussed by several authors [9,10,20,22]. One of
the most cited mechanism proposes that the reverse water–gas-
shift reaction, which generates CO from CO2, is the first step. This
reaction occurs on Ni surface and is followed by the methanation of
CO. Falconer and Zagli studied the mechanism of CO2 methanation
over a Ni/SiO2 catalyst, and reported that the carbon dioxide did
not adsorb significantly on silica, leading them to conclude that this
activated adsorption occurred on the nickel metal and not the sup-
port. Their conclusion was based on the capacity of CO2, which did
not adsorb significantly on SiO2 [20]. Peebles et al. also suggested
that CO2 is quickly transformed into CO on the Ni surface [22]. How-
ever, the previously proposed mechanism leaves several questions
unanswered regarding how the CO2 is adsorbed and transformed
into methane:
catalysts was 5 wt%. All metal based MSN samples were denoted
as Rh/MSN, Ru/MSN, Ni/MSN, Ir/MSN, Fe/MSN, Cu/MSN, Zn/MSN,
V/MSN, Cr/MSN, Mn/MSN, Al/MSN and Zr/MSN.
2.2. Characterization
The crystalline structure of the catalyst was determined by
X-ray diffraction (XRD) recorded on a powder diffractometer
(Bruker Advance D8, 40 kV, 40 mA) using a Cu K␣ radiation source
in the range of 2ꢀ = 1.5–90◦. The BET analysis of the catalyst
was determined by N2 adsorption–desorption isotherms using a
Quantachrome Autosorb-1 instrument. The catalyst was outgassed
at 573 K for 3 h before being subjected to N2 adsorption. Pore
size distributions and pore volumes were determined from the
sorption isotherms using a non-localized density functional theory
(NLDFT) method. In the FTIR measurements, pyrrole has been used
as a probe molecule for the characterization of basic sites [3]. The
methanation was also recorded by in situ FTIR spectroscopy to study
the surface species formed during the reaction. All the measure-
ments were performed on an Agilent Cary 640 FTIR spectrometer
equipped with a high-temperature stainless steel cell with CaF2
windows. Prior to the measurements, 30 mg of sample in the form
of a self-supported wafer was reduced in H2 stream (100 ml/min) at
773 K for 4 h and cooling to 303 K under He atmosphere. For pyrrole
adsorption, the reduced catalyst was exposed to 2 Torr of pyrrole
at 303 K for 30 min, followed by outgassing at 423 K for 30 min. All
spectra were recorded at room temperature with a spectral res-
olution of 5 cm−1 with five scans. Furthermore, the formation of
surface species during the methanation was recorded by introduc-
ing of a mixture of CO2 (4 Torr) and H2 (16 Torr) to the catalyst
at room temperature, followed by heating to 523 K. The spectra
were recorded after equilibrium conditions have been reached. For
CO2 and H2 adsorption studies, the sample was pretreated using
the same procedure as above. The adsorption of CO2 was done by
exposing 4 Torr of CO2 on sample at room temperature and sub-
sequent heating to 523 K. While, the adsorption of H2 was done
by exposing 16 Torr H2 at room temperature followed by heating
in hydrogen from 303 to 573 K. For adsorption of CO2, H2 or the
mixture of CO2 and H2 in the bare Ni catalyst, the Ni catalyst was
mixed with KBr (Ni:KBr = 1:100) to increase the transparency of the
sample.
(i) CO2 methanation over unsupported Ni exhibited low yield of
methane. Therefore, it is difficult to conclude that the CO that
adsorbed on the Ni and/or metal surface was a precursor to
methane.
(ii) The technique used to elucidate the mechanism of CO2 metha-
nation was performed on a combination of metal and support,
and it was difficult to determine the real function of the
metal and the support. Therefore, each metal, support and
metal/support combination should be tested in order to know
the role of their presence.
Therefore, in this work, a series of metal-based catalyst sup-
ported on mesostructured silica nanoparticles (M/MSNs; M = Rh,
Ru, Ni, Fe, Ir, Cu, Zn, V, Cr, Mn, Al and Zr) were prepared, and were
applied for methane production from carbon dioxide and hydro-
gen gas in the temperature range of 373–723 K under atmospheric
pressure. The effect of metal on the MSN was characterized by
X-ray diffraction (XRD), nitrogen physisorption and infrared spec-
troscopy. To elucidate the mechanistic path of CO2 methanation,
in situ FTIR spectroscopy was used to observe the surface species
during the reaction.
2. Experimental
2.1. Catalyst preparation
2.3. Catalytic activity measurements
MSN was prepared by the sol–gel method according to the
report by Karim et al. [26]. The surfactant cetyltrimethylammo-
nium bromide (CTAB; Merck), ethylene glycol (EG; Merck) and
NH4OH solution (QRec) were dissolved in water with the following
molar composition of CTAB:EG:NH4OH:H2O = 0.0032:0.2:0.2:0.1.
After vigorous stirring for about 30 min at 353 K, 1.2 mmol of
tetraethyl orthosilicate (Merck) and 1 mmol of 3-aminopropyl tri-
ethoxysilane (Merck) were added to the clear mixture to give
a white suspension solution. This solution was then stirred for
another 2 h, and the samples were collected by centrifugation
at 20,000 rpm. The as-synthesized MSN were dried at 333 K and
calcined at 823 K for 3 h in air to remove the surfactant. For unsup-
ported Ni catalyst, it was prepared by calcination of Ni(NO3)2·6H2O
at 823 K for 3 h. For metal based MSN catalysts, they were pre-
pared by impregnation of MSN powder with an aqueous solution
of the corresponding metal salt precursor (RhCl3, RuCl3, Ni(NO3)2,
IrCl3, Fe(NO3)3, Cu(CO2CH3)2, Zn(NO3)2, V(C5H7O2)3, Cr(NO3)3,
MnCl2, Al(NO3)3 and ZrOCl2) (Merck, 99%). The resulting slurry was
heated slowly at 353 K under continuous stirring and maintained
at that temperature until nearly all the water being evaporated.
The solid residue was dried in an oven at 383 K overnight before
calcination at 823 K for 3 h in air. The metal loading of the
CO2 methanation was conducted in a microcatalytic quartz reac-
tor with an interior diameter of 8 mm at atmospheric pressure at
temperature range of 373–723 K. The thermocouple was directly
inserted into the catalyst bed to measure the actual pretreatment
and reaction temperatures. The catalyst was sieved and selected in
the 20–40 ␮m fraction. Initially, 200 mg of catalyst was treated in an
oxygen stream (FOxygen = 100 ml/min) for 1 h followed by a hydro-
gen stream (FHydrogen = 100 ml/min) for 4 h at 773 K and cooled
down to the desired reaction temperature in a hydrogen stream.
After the temperature became stable, a mixture of H2 and CO2
was fed into the reactor at gas hourly space velocity (GHSV) of
50,000 ml g−1 h−1 and H2/CO2 mass ratio of 4/1. All gases were
controlled with calibrated mass flow controllers (SEC-400 MK2,
Stec Ltd., Japan). The CH4 formation rate was measured in the
temperature range of 373–723 K. The activity was monitored with
decreasing temperature and back to verify stable catalyst condi-
tions during these measurements. The composition of the outlet
gases was analyzed with an on-line 6090N Agilent gas chromato-
graph equipped with a GS-Carbon PLOT column and a TCD detector.
The lines from the outlet of the reactor to the GC were heated at
383 K to avoid condensation of the products. The moisture trap was
installed at the outlet gas line of the reactor to prevent moisture

M.A.A. Aziz et al. / Applied Catalysis A: General 486 (2014) 115–122
117
diffraction peaks at 37.3◦, 43.2◦, 62.9◦, 75.3◦ and 79.4◦ were
attributed to a face-centered cubic crystalline NiO [5]. The XRD
pattern of Ir/MSN showed the diffraction peaks at 27.5◦, 34.5◦, 38◦,
52.9◦, 56◦ and 63◦, attributed to the Ir species in an oxide form
(IrO2) [29].
A
4000
B
200
a
(100)
(110)
b
c
(200)
The porosity of catalysts was measured by the N2 physisorption
method. The location of the inflection point is related to the pores
at the mesoscale, and the sharpness of these curves reveals the uni-
formity of the mesopore size distribution [30]. Fig. 2A and B shows
the adsorption–desorption isotherms of all catalysts. All isotherm
curves exhibited two capillary condensation steps. The first step
in a relative pressure range of P/Po = 0.3–0.4 was attributed to the
nitrogen condensation that took place at the internal mesopores.
There was no hysteresis loop in this capillary condensation. The
second step above P/Po = 0.95 in the adsorption branch was due to
the presence of interparticle voids [31]. A small hysteresis loop at
this relative pressures was ascribed to the condensation of nitro-
gen within the interstitial voids or interparticle textural porosity
created by MSN particles, which indirectly reflects the size of par-
ticles, i.e. a higher partial pressure was associated with a smaller
particle size. However, almost no change in relative pressure at
P/Po = 0.95 indicates that the particle size of the MSNs was main-
tained after introduction of the metal. The inflection points were
similar but the sharpness of the steeps decreased after introduc-
tion of the metal. For instance, almost no steeps were observed on
Ru/MSN and Cu/MSN in the first condensation step indicating the
low uniformity of micropores of the catalysts. In addition, low uni-
formity of mesopores was observed only on Cu/MSN, since the steep
in the second condensation markedly decreased after introduction
of Cu.
a
d
b
c
e
d
e
f
f
g
g
0
2
4
6
8
10 30 40 50 60 70 80 90
2θ [°]
2 θ [°]
Fig. 1. (A) Low-angle and (B) wide-angle XRD patterns of (a) MSN, (b) Rh/MSN, (c)
Fe/MSN, (d) Ru/MSN, (e) Ni/MSN, (f) Ir/MSN and (g) Cu/MSN.
from entering the GC. To determine the activity and selectivity, the
products were collected after 1 h of steady-state operation at each
temperature. The conversion of carbon dioxide, selectivity of prod-
ucts, rate of methane formation and areal rate were calculated by
the following equations:
FCO ,in − FCO
,out
2
2
XCO2 (%) =
Sx(%) =
× 100%
× 100%
(1)
(2)
FCO
,in
2
FX,out
FCO ,in − FCO
,out
2
2
Textural parameters of MSNs and metal-based MSN catalysts
are summarized in Table 1. All samples possessed high surface area,
typically larger than 800 m2 g−1, except for Cu/MSN, which had a
surface area of 501 m2 g−1. The impregnation of a small amount
of metal (Rh, Fe, Ru, Ni, Ir and Cu) produced little effect on the
pore diameter. It is obvious that the nitrogen adsorption decreased
with increased of pore size. The introduction of metal on MSNs also
affected the micropore and mesopore distribution of the catalysts
(Fig. 2C and D). All catalysts showed a bimodal pore structure con-
FX,out
−1
rate(molCH mol
s
−1) =
(3)
(4)
Metal
4
mole of MTotal
FX,out
Areal rate(mmoleCH4 × m−2 s−1) =
SBET × Wcat
where XCO is the conversion of carbon dioxide (%), Sx is the
2
selectivity of x product (%) in which x is a CH4 or CO, F is a molar
flow rate of CO2 or product in mole per second, MTotal is the total
amount of metal in mole, SBET is the surface area of the catalysts
and Wcat is the weight of catalysts.
˚
sisting of framework mesopores of 20–30 A and textural mesopores
˚
in the range 100–1000 A. However, Cu/MSN showed no pore dis-
˚
tribution in the range 20–30 A, but had a pore distribution in the
˚
3. Results and discussion
range 35–45 A.
FTIR spectroscopy in the skeletal region at 1300–700 cm−1
was used to follow the variation of the structure of metal-
based MSNs. Typical absorbance bands due to siliceous material
3.1. Characterization of materials
Fig. 1A shows the small-angle XRD patterns of MSN
and metal-based MSN catalysts. The patterns exhibited three
peaks, indexed as (1 0 0), (1 1 0) and (2 0 0), which are reflec-
tions of the typical two-dimensional hexagonally ordered
mesostructure (p6mm), demonstrating the high quality of
mesopore packing [27]. The ordered MSN support structure
was disturbed slightly by the presence of metals. No shift
of peaks was observed, but the intensities decreased with
the introduction of metal. The intensity was in the order
Rh/MSN > Fe/MSN > Ru/MSN > Ni/MSN > Ir/MSN > Cu/MSN, indicat-
ing a structural degradation of MSNs. The presence of metal
crystallites on the catalysts were characterized using wide-angle
XRD (30–90◦), as shown in Fig. 1B. No diffraction peaks ascribed to
Rh, Fe and Cu species on MSNs were detected, indicating that either
the metal particles were smaller in diameter than the detection
limit (ca. 4 nm) of the diffractometer, or the large metal particles
(>4 nm) were too scarce to produce detectable diffraction signals.
XRD pattern of Ru/MSN showed the diffraction peaks of ruthenium
oxide species (RuO2) at 28◦, 35◦ and 54◦ with small diffraction
of ruthenium metal species (Ru) at 40◦ [28]. For Ni/MSN, the
Si O Si were observed for bare MSN and metal-based MSNs
(Fig. 3A). The bands at approximately 1200 and 1080 cm−1 were
ascribed to the longitudinal and transverse optical modes of
the asymmetric Si O stretching vibration. Two weaker bands at
approximately 960 and 805 cm−1 were attributed to Si OH and
the symmetric Si O stretching vibration, respectively. The band
at 960 cm−1 is attributed to the presence of transition metal
atoms near the silica framework as the stretching Si O vibra-
tion mode perturbed by the neighboring metal ions. Therefore, a
shift of the absorbance band at 960 cm−1 indicated a structural
change for the surface Si OH group caused by the presence of
metal species in the MSN, and/or due to evolution of a new band
of Si
O
M. The absorbance band of MSN at 960 cm−1 shifted
to 958, 973, 970 and 975 cm−1 for Rh/MSN, Fe/MSN, Ru/MSN
and Ir/MSN, respectively. However, no shift was observed for
Ni/MSN and Cu/MSN indicating that the Ni and Cu did not inter-
act with the MSN framework. Reduced MSNs and metal-based
MSNs possessed an intense band at 3740 cm−1 with a shoulder at
3710 cm−1 and a broad halo centered around 3650 cm−1 (Fig. 3B).
The bands at 3740 and 3710 cm−1 are unambiguously assigned to

118
M.A.A. Aziz et al. / Applied Catalysis A: General 486 (2014) 115–122
Fig. 2. N2 adsorption–desorption isotherms (A, B) and NLDFT pore size distribution (C, D) of (a) MSN, (b) Rh/MSN, (c) Fe/MSN, (d) Ru/MSN, (e) Ni/MSN, (f) Ir/MSN and (g)
Cu/MSN.
C
B
A
0.5
0.1
5%
a
3467
3740
960
3710
3650
a
a
g
g
805
1200
1080
g
3650 3550 3450 3350 3250
3800 3600 3400 3200
1300 1150 1000 850 700
Wavenumber [cm-1]
Wavenumber [cm-1]
Wavenumber [cm-1]
Fig. 3. FTIR spectra of (A) fresh and (B) reduced catalysts: (a) MSN, (b) Rh/MSN, (c) Fe/MSN, (d) Ru/MSN, (e) Ni/MSN, (f) Ir/MSN and (g) Cu/MSN. (C) Normalized IR spectra of
pyrrole adsorbed on reduced catalysts of (a) MSN, (b) Ru/MSN, (c) Fe/MSN, (d) Rh/MSN, (e) Ni/MSN, (f) Cu/MSN and (g) Ir/MSN.
terminal (isolated) and internal silanol groups, respectively. The
broad absorbance is indicative of the presence of H-bonded
hydroxyls. The result showed the decrease in intensity of a
silanol band at 3740 cm−1 upon the incorporation of metal,
suggesting a perturbation of the silica framework upon interaction
with metal [32].
The presence of basic sites on metal based MSN was evidenced
by pyrrole adsorption–desorption FTIR spectroscopy. The H-donor
Table 1
Physicochemical properties and catalytic performance. The reaction was done at 623 K with GHSV of 50,000 ml g−1 h−1 and H2/CO2 ratio of 4/1.
´
Rate × 10−2
Conversion,
Selectivity (%)a
CH4
Areal rate × 10−3
˚
SBET
(m2 g−1
Vpore
(cm3 g−1
dpore (A)
Catalyst
(molCH4 molMetal−1 s−1
)
XCO (%)
(mmolCH4 m−2 s−1
)
)
)
2
CO
Ni
MSN
3
1051
933
1004
879
976
858
501
0.016
0.758
0.910
0.934
0.798
0.723
0.805
0.642
194
28
36
37
34
33
35
32
–
–
0.5(1.5)c
0.3
99.5
95.7
85.4
9.5
100b (0)c
0b
100
100
99.9
83
0b (100)c
–
–
100b
Rh/MSN
Ru/MSN
Ni/MSN
Ir/MSN
Fe/MSN
Cu/MSN
29.71
28.42
25.39
2.81
0.97
0.57
0
0
0.1
17
8
21
0.966
0.885
1.554
0.047
0.064
0.057
4
3.3
92
79
SBET: BET surface area; Vpore: total pore volume; dpore: average pore diameter.
a
Selectivity at conversion = 20%.
Selectivity at conversion = 0.3%.
The result in the parenthesis shows the reaction in the presence of CO2 only.
b
c

M.A.A. Aziz et al. / Applied Catalysis A: General 486 (2014) 115–122
119
2
1
0
100
80
60
40
20
0
45
40
35
30
25
20
15
10
5
2
1
0
100
80
60
40
20
0
Rh/MSN
Ru/MSN
Ni/MSN
Ir/MSN
Fe/MSN
Cu/MSN
C
B
A
X [%]
X [%]
Sel CH4 [%]
Sel CO [%]
Sel CH4 [%]
Sel CO [%]
0
0
10 20 30 40 50
0
10 20 30 40 50
350 450 550 650 750
Time [min]
Time [min]
Temperature [K]
Fig. 4. Catalytic activity of (A) MSN and (B) Ni at 623 K. (C) Catalytic activity of various metal based MSN at steady state condition as a function of reaction temperature.
pyrrole properties allow C4H4NH···O bridges to form with basic
oxygen in the framework [33]. It was shown by pyrrole adsorp-
tion and desorption experiments that Si O– species could play a
role as an oxidizing center. Pyrrole adsorbed on MSNs and metal-
based MSNs verified the formation of hydrogen-bonded species of
pyrrole as observed in the IR spectra around 3467 cm−1 (Fig. 3C).
The introduction of metal, however, partially blocked the basic
sites of the catalyst and thus reduced the chemisorptions of pyr-
role on the catalyst. Each metal has different behavior on the
blocking effect of metal on the pyrrole adsorption. For example,
Rh/MSN has lower tendency to produce the blocking effect due to
its high metal dispersion. In contrast, Cu/MSN has a higher block-
ing effect due to its low metal dispersion. The basic sites have a
close relationship with the structure defects or oxygen vacancy,
which was explained in previous studies [3,34–37]. For instance,
the calcination of mixed Al2O3–MgO produces high surface area
and number of defect sites (the number of oxygen atoms asso-
ciated with Mg exhibiting a low coordination number), and this
leads to higher basicity [34]. Therefore, it may be stated that the
basic or oxygen vacancy sites can be dedicated as active sites for
the CO2 adsorption. This indicated that Ru- and Cu-based MSNs
possess highest and lowest concentration of oxygen vacancy sites,
respectively.
catalysts (Ni, Rh, Fe, Ru, Ir and Cu) were presumed to be inactive
for CO2 methanation.
For metal-based MSNs (Fig. 4C), the rate of methane forma-
tion (molCH4 molmetal−1 s−1) is plotted as a function of reaction
temperature in the range 373–723 K. S-curves results were
observed for all catalysts indicating that the rate production
of methane was not influenced by intraparticles diffusion of
the catalysts under the present reaction conditions. Further-
more, the order of the activity for CO2 methanation at 623 K
was Rh/MSN > Ru/MSN > Ni/MSN > Ir/MSN > Fe/MSN > Cu/MSN. For
Rh/MSN and Ru/MSN, the reaction started at 423 K and the conver-
sion of CO2 reached 100% at 673 K and the activity continued till
723 K.
For Ni/MSN, the reaction began at 473 K and the conversion
reached 100% at 673 K. The CO2 methanation increased rapidly
at 523 K over these three catalysts. Indeed, the Rh/MSN, Ru/MSN
and Ni/MSN could catalyze CO2 methanation at and below 473 K.
In particular, the catalytic activity of Rh/MSN was the highest,
compared to other catalysts, at and below 573 K, while Ir/MSN,
Fe/MSN and Cu/MSN were practically inactive below 573 K. In fact,
high reaction temperature (>723 K) is required in order to achieve
CO2 conversions above 10% over these latter catalysts. It is also
a common practice to compare catalyst activities based on cat-
alyst surface area rather than catalyst mass. The final column in
Table 1 provides the catalytic rates based on area (areal rates) for
all metal based MSN catalysts at 623 K. The activity order was
Ni/MSN > Rh/MSN > Ru/MSN > Fe/MSN > Cu/MSN > Ir/MSN. By con-
trast, using this measure of activity, Ni/MSN was the most active
catalyst, while Ir/MSN was the poorest catalyst. These results
suggest that Rh, Ru, or Ni metal on MSN are effective catalysts
for methanation of CO2 in the temperature range 423–573 K,
while Ir, Fe or Cu metal on MSN are active in the temperature
range 623–723 K. In addition, no activity (methane production)
was observed on Zn/MSN, V/MSN, Cr/MSN, Mn/MSN, Al/MSN and
Zr/MSN in the temperature range studied. Therefore, the charac-
terization data for these metal-based MSN catalysts are not shown
or discussed in the present study.
3.2. Catalytic performance
The results of the catalytic reactions are summarized in Fig. 4
and Table 1. For MSN (Fig. 4A), the conversion of CO2 at 623 K was
only around 0.3% with 100% selectivity to CO. Low conversion of
MSN was due to the absence of metal sites, which is crucial for H2
and CO2 dissociation. For Ni, two types of reactions occurred (Fig. 4B
and Table 1). In the absence of H2, the conversion of CO2 reached
around 1.5% with 100% selectivity for CO. No methane products
were observed due to the absence of H2. In the presence of H2, the
conversion of CO2 decreased to around 0.5% with 100% selectivity
for methane. The 67% decrease of conversion was probably due to
the competitive adsorption–dissociation of CO2 and H2 on the sur-
face of metallic Ni [38] and the absence of a MSN support, which
facilitated the formation of surface carbonyl species. In addition, the
low conversion may also be due to the absence of metal–support
interaction, which served to increase the metal dispersion [4]. Riani
et al. also reported that Ni was inactive in CO2 methanation in
which the results gave no conversion of CO2 at 573 K [39]. These
results explained the inability of unsupported metal catalysts for
CO2 methanation. Therefore, in general, the unsupported metal
3.3. Mechanistic investigation of CO2 methanation
In the following sections, FTIR spectra were recorded on Ni,
MSN, Ni/MSN, Rh/MSN and Fe/MSN catalysts under a set of defined
adsorption conditions (Fig. 5). For Ni (inset in Fig. 5A), there was a
strong band at around 450 cm−1, which corresponds to the bend-
ing vibration of Ni O bond [40]. The adsorption of CO2, H2 and

120
M.A.A. Aziz et al. / Applied Catalysis A: General 486 (2014) 115–122
Ni/MSN
MSN
0.5
0.5
f
f
a
a
3800 3600 3400 3200
3800 3600 3400 3200
Wavenumber [cm-1]
Fig. 6. The changes in the IR spectra when MSN and Ni/MSN were heated in 16 Torr
of hydrogen at (b) 303 K, (c) 323 K, (d) 423 K, (e) 523 and (f) 573 K. (a) Before exposure
to hydrogen. All samples were pretreated with hydrogen at 773 K for 4 h and cooling
to 303 K under He atmosphere.
form carbonyl while another O fills the oxygen vacancy near the Ni
sites.
The ability of Ni to dissociate H2 was evidenced by in situ FTIR
spectroscopy of H2 adsorption on Ni/MSN. From the results (Fig. 6),
the surface concentration of the hydroxyl groups increased with
increasing temperature from 323 to 573 K in the presence of hydro-
gen. A larger increase was observed on Ni/MSN compared to MSN
indicating that the supply of H atoms was higher in Ni/MSN com-
pared to MSN. The ability of Ni, Rh and Fe metal to dissociate
molecular H2 was also previously discussed by several reports
[42–44]. The MSN surface or oxygen vacancy site probably has a
low ability to adsorb and dissociate molecular hydrogen to hydro-
gen atoms compared to Ni/MSN. A similar result was observed on
a TiO2 surface in which Menetrey et al. reported that the presence
of an oxygen vacancy did not affect the adsorption–dissociation
of hydrogen molecule [45]. The interaction of molecular hydrogen
is lower on the oxygen vacancy containing TiO2 than that on the
perfect structure of TiO2. Therefore, from the results, it can be sug-
gested that the role of Ni was to dissociate H2 and CO2, followed by
spillover of CO, O and H atoms onto the MSN surface and formation
of methane and water. The detail mechanistic will be explained in
further discussion.
In situ FTIR spectroscopy of CO2 + H2 at 523 K on metal based
MSN is also presented in Fig. 5B. For Ni/MSN, Rh/MSN and Fe/MSN,
the bands at 2098, 2075, 2036 cm−1 were assigned to linear car-
bonyl, while the bands at 1978, 1940 and 1928 cm−1 were assigned
to bridged carbonyl in which both of the species were adsorbed
onto surface oxygen. The last two bands at 1627 and 1577 cm−1
were assigned to adsorbed water and bidentate formate, respec-
tively [46,47]. The negative band at 1627 cm−1 for Ni/MSN was
attributed to water which have some degree of water termination.
The CHx and CO gases vibration bands, which known to appear at
2800–3000 cm−1 and 2100–2200 cm−1 regions, were not detected
for any catalyst in this experiment. Based on these results, a plau-
sible mechanism for CO2 methanation on metal-based MSNs is
shown in Scheme 2.
Fig. 5. (A) Evaluation of the FTIR spectra of adsorbed gases (CO2, H2 and CO2 + H2)
on Ni (dotted lines) and MSN (true lines). Inset shows the FTIR spectra of Ni in range
of 1200–400 cm−1. (B) Evaluation of the FTIR spectra of adsorbed CO2 (dotted lines)
on Ni/MSN and CO2 + H2 (true lines) on Ni/MSN, Rh/MSN and Fe/MSN. All spectra of
adsorbed gases were monitored at 523 K.
Scheme 1. Plausible mechanism of CO2 methanation on Ni.
CO2 + H2 on Ni showed no significant bands. These results indicate
that the methanation reaction on Ni probably takes place in the
gas phase rather than on the catalyst surface, in agreement with
the report by Medsford [41]. Therefore, a plausible mechanism of
methanation on Ni is shown in Scheme 1.
The assumption of this phenomenon was also based on the
inability of Ni to perform high methanation activity. In the case
of Ni, as mentioned earlier, in the absence of H2, the reaction per-
formed on CO2 resulted in CO product only. In contrast, when CO2
and H2 were applied during the reaction, only a small amount of
methane was observed indicating that all CO was hydrogenated to
methane. However, the percentage conversion of CO2 on Ni was
very low (0.3%) indicating that the CO2 methanation was proba-
bly not taking place on the Ni surface. Similarly, the generation of
adsorbed species was not observed on MSN (Fig. 5A). The absence of
adsorbed species on MSN was due to the absence of Ni sites to disso-
ciate CO2 and/or H2. Under CO2 adsorption on Ni/MSN (Fig. 5B), two
types of adsorbed species were observed at 2036 and 1901 cm−1
,
Firstly, CO2 and H2 was adsorbed onto metal sites, followed
by dissociation to form CO, O and H atoms, and migration onto
the MSN surface. The CO then interacted with oxide surfaces of
the MSN to form bridged carbonyl and linear carbonyl. Biden-
tate formate was also formed through the interaction with atomic
hydrogen. Meanwhile, the O atom spilt over onto the surface of
which are attributed to linear and bridged carbonyl on the sur-
face of MSN, respectively. From these result, it is presumed that
the role of Ni was to dissociate the CO2 into CO and O adsorbed
species. The dissociated CO tends to migrate through the MSN sur-
face (spillover) and then interacted with oxide surface of MSN to

M.A.A. Aziz et al. / Applied Catalysis A: General 486 (2014) 115–122
121
4. Conclusions
Mesostructured silica nanoparticles (MSN) and a series of
metal catalysts (Rh, Ru, Ni, Fe, Ir, Cu, Zn, V, Cr, Mn, Al and Zr)
loaded on MSNs were prepared by a sol–gel and impregnation
method for methanation of CO2. The basic site of the catalysts
were characterized on the basis of pyrrole adsorption–desorption
study. The concentration of basic site altered with the type of
metals in which Ru/MSN and Cu/MSN possessed the highest and
lowest concentrations of basic sites, respectively. The results indi-
cated that surface centers containing metallic and associated basic
and/or oxygen vacancy sites are responsible for the active sites
for this process. At 623 K, the catalytic activity changed in the
order of Rh/MSN > Ru/MSN > Ni/MSN > Ir/MSN > Fe/MSN > Cu/MSN.
Low conversion of CO2 was observed on MSN, Ni, Zn/MSN, V/MSN,
Cr/MSN, Mn/MSN, Al/MSN and Zr/MSN. In term of areal basis,
Ni/MSN exhibited the most active catalyst while Ir/MSN is the poo-
rest catalyst. Based on the FTIR studies, the detailed mechanism has
been proposed for the methanation of CO2 over metal-based MSNs.
Firstly, CO2 and H2 is adsorbed and dissociated on metal sites to
form CO, O and H atoms, followed by migration onto the MSN sur-
face. The CO then interacts with oxide surfaces of the MSNs to form
bridged carbonyl and linear carbonyl, while the presence of H atom
facilitates the formation of bidentate formate. These three species
could be responsible for the formation of methane. However, the
bidentate formate species could be the main route to formation of
methane. MSN support has been found to play an important role in
the mechanism. MSN support served the sites for carbonyl species
which act as precursor to methane formation. This proposed mech-
anism creates a new perspective and better understanding of the
mechanisms of CO2 methanation.
Scheme 2. Plausible mechanism of CO2 methanation on M/MSN.
the MSN and was stabilized in the oxygen vacancy site near the
metal site. The adsorbed oxygen then reacted with atomic hydro-
gen to form hydroxyl on the MSN surface in which a further reaction
with another atomic hydrogen formed a water molecule. Finally,
the adsorbed carbon species was further hydrogenated to methane
and another water molecule. In this study, all three species (lin-
ear carbonyl, bridged carbonyl and bidentate formate) could be an
intermediate for CO2 methanation as reported by previous studies
[9–11,14,17,25]. However, the bidentate formate species could be
the main intermediate for CO2 methanation as agreement by the
previous studies [9–11]. Fujita et al. has reported that the hydro-
genation has occurred on bridged carbonyl and formate species
located on the support. Pan et al. has reported that the formate
species were found to be the main intermediate species during the
CO2 methanation. Similarly, Aldana et al. also found that formate
species were formed on the surface of catalysts and hydrogenated
to form methane.
The results provide new evidence that helps in the understand-
ing of the nature of the intermediate in the methane formation
process, and a new pathway in the mechanism of CO2 methanation
has been presented in this study. The main finding is related to the
ability of Ni to dissociates H2 and CO2 followed by migration onto
the MSN surface. Such behavior apparently conflicts with preva-
lent opinion relating strong binding of CO with metal sites (M–CO)
in which the M–CO species undergo methane formation [14–16].
Jacquemin et al. looked into the reaction mechanism of CO2 metha-
nation on Rh/␥-Al2O3 and found the dissociation of CO2 into CO
and oxygen on the surface of the catalyst, as evidenced by in situ
DRIFT experiments [16]. Karelovic et al. studied the CO2 adsorption
and dissociation on a Rh/␥-Al2O3 system in the presence of H2 by
operando-DRIFTS experiments. Their result showed the formation
of linear Rh-CO species that act as a precursors of methane [14]. In
the case of Ni catalysts, Aldana et al. studied the mechanism of CO2
methanation over Ni–Ceria/zirconia. They reported that the role of
Ni is to dissociate H2 while ceria–zirconia support was responsible
to adsorb the CO2 to form carbonates which would be hydrogenated
into formates and further into methoxy species [9]. Pan et al. pro-
posed that CO2 was adsorbed onto hydroxyl and surface oxygen
of ceria–zirconia to form bidentate and monodentate formate in
the presence of hydrogen, respectively [10]. However, they did not
discuss the detailed route of CO2 adsorption onto the catalyst sur-
face, although they did mention that CO2 was activated on the
surface oxygen of the ceria–zirconia support. Furthermore, under
the experimental conditions used in this present work, the presence
of metal carbonyls were not observed on metal (Ni, Rh and Fe). On
the other hand, it may be suggested that the route to the methane
formation was not formed via metal carbonyl. In summary, metal
sites were identified and found to be responsible for dissociating
CO2 to form CO and O, followed by migration onto the catalyst’s
surface. While, MSN support was responsible to serve the sites for
carbonyl species which act as precursor to methane formation.
Acknowledgments
This work was supported by Ministry of Science, Technology and
Innovation, Malaysia through EScienceFund Research Grant no. 03-
01-06-SF0987, MyPhd Scholarship (Muhammad Arif Ab Aziz) from
Ministry of Higher Education, Malaysia.
References
[1] M. Wang, Z. Wang, X. Gong, Z. Guo, Renew. Sustain. Energy Rev. 29 (2014) 573.
[2] G. Centi, S. Perathoner, Catal. Today 148 (2009) 191.
[3] M.A.A. Aziz, A.A. Jalil, S. Triwahyono, R.R. Mukti, Y.H. Taufiq-Yap, M.R. Sazegar,
Appl. Catal. B 147 (2014) 359.
[4] Y.-H. Huang, J.-J. Wang, Z.-M. Liu, G.-D. Lin, H.-B. Zhang, Appl. Catal. A 466
(2013) 300.
[5] G. Zhang, T. Sun, J. Peng, S. Wang, S. Wang, Appl. Catal. A 462–463 (2013) 75.
[6] G. Du, S. Lim, Y. Yang, C. Wang, L. Pfefferle, G.L. Haller, J. Catal. 249 (2007) 370.
[7] B. Lu, K. Kawamoto, Fuel 103 (2013) 699.
[8] E. Jwa, S.B. Lee, H.W. Lee, Y.S. Mok, Fuel Process. Technol. 108 (2013) 89.
[9] P.A.U. Aldana, F. Ocampo, K. Kobl, B. Louis, F. Thibault-Starzyk, M. Daturi, P.
Bazin, S. Thomas, A.C. Roger, Catal. Today 215 (2013) 201.
[10] Q. Pan, J. Peng, S. Wang, S. Wang, Catal. Sci. Technol. 4 (2014) 502.
[11] S. Fujita, M. Nakamura, T. Doi, N. Takezawa, Appl. Catal. A 104 (1993) 87.
[12] A.-X. Yin, W.-C. Liu, J. Ke, W. Zhu, J. Gu, Y.-W. Zhang, C.-H. Yan, J. Am. Chem.
Soc. 134 (2012) 20479.
[13] J.H. Kwak, L. Kovarik, J. Szanyi, ACS Catal. 3 (2013) 2449.
[14] A. Karelovic, P. Ruiz, ACS Catal. 3 (2013) 2799.
[15] A. Beuls, C. Swalus, M. Jacquemin, G. Heyen, A. Karelovic, P. Ruiz, Appl. Catal.,
B 113–114 (2012) 2.
[16] M. Jacquemin, A. Beuls, P. Ruiz, Catal. Today 157 (2010) 462.
[17] J.J. Benítez, R. Alvero, M.J. Capitán, I. Carrizosa, J.A. Odriozola, Appl. Catal. 71
(1991) 219.
[18] H.M.T. Galvis, J.H. Bitter, T. Davidian, M. Ruitenbeek, A.I. Dugulan, K.P. de Jong,
J. Am. Chem. Soc. 134 (2012) 16207.
[19] G. Zhou, T. Wu, H. Xie, X. Zheng, Int. J. Hydrog. Energy 38 (2013) 10012.
[20] J.L. Falconer, A.E. Zagli, J. Catal. 62 (1980) 280.
[21] G.D. Weatherbee, C.H. Bartholomew, J. Catal. 77 (1982) 460.
[22] D.E. Peebles, D.W. Goodman, J.M. White, J. Phys. Chem. 87 (1983) 4378.
[23] S. Fujita, H. Terunuma, H. Kobayashi, N. Takezawa, React. Kinet. Catal. Lett. 33
(1987) 179.
[24] C. Schild, A. Wokaun, A. Baiker, J. Mol. Catal. 63 (1990) 243.

122
M.A.A. Aziz et al. / Applied Catalysis A: General 486 (2014) 115–122
[25] R.A.D. Betta, M. Shelef, J. Catal. 48 (1977) 111.
[36] J.C. Lavalley, Catal. Today 27 (1996) 377.
[26] A.H. Karim, A.A. Jalil, S. Triwahyono, S.M. Sidik, N.H.N. Kamarudin, R. Jusoh,
N.W.C. Jusoh, B.H. Hameed, J. Colloid Interface Sci. 386 (2012) 307.
[27] X. Yang, D. Chen, S. Liao, H. Song, Y. Li, Z. Fu, Y. Su, J. Catal. 291 (2012) 36.
[28] J. Okal, M. Zawadzki, K. Baranowska, Appl. Catal. A 471 (2014) 98.
[29] R. Moraes, K. Thomas, S. Thomas, S.V. Donk, G. Grasso, J.-P. Gilson, M. Houalla,
J. Catal. 299 (2013) 30.
[30] J.H. de Boer, B.G. Linsen, T.J. Osinga, J. Catal. 4 (1965) 643.
[31] Z. Gao, I. Zharov, Chem. Mater. 26 (2014) 2030.
[32] N.W.C. Jusoh, A.A. Jalil, S. Triwahyono, H.D. Setiabudi, N. Sapawe, M.A.H. Satar,
A.H. Karim, N.H.N. Kamarudin, R. Jusoh, N.F. Jaafar, N. Salamun, J. Efendi, Appl.
Catal. A 468 (2013) 276.
[37] C. Drouilly, J.-M. Krafft, F. Averseng, H. Lauron-Pernot, D. Bazer-Bachi, C. Chiza-
llet, V. Lecocq, G. Costentin, Appl. Catal., A 453 (2013) 121.
[38] K.B. Kester, E. Zagli, J.L. Falconer, Appl. Catal. 22 (1986) 311.
[39] P. Riani, G. Garbarino, M.A. Lucchini, F. Canepa, G. Busca, J. Mol. Catal. A: Chem.
383–384 (2014) 10.
[40] S. Mochizuki, M. Satoh, Phys. Status Solidi B 106 (1981) 667.
[41] S. Medsford, J. Chem. Soc. 123 (1923) 1452.
[42] M.A.A. Aziz, N.H.N. Kamarudin, H.D. Setiabudi, H. Hamdan, A.A. Jalil, S. Triwahy-
ono, J. Nat. Gas Chem. 21 (2012) 29.
[43] S. Wilke, D. Hennig, R. Löber, Phys. Rev. B 50 (1994) 2548.
[44] F. Bozso, G. Ertl, M. Grunze, M. Weiss, Appl. Surf. Sci. 1 (1977) 103.
[45] M. Menetrey, A. Markovits, C. Minot, Surf. Sci. 524 (2003) 49.
[46] K. Tanaka, J.M. White, J. Phys. Chem. 86 (1982) 4708.
[47] K. Walter, O.V. Boyevskaya, D. Wolf, M. Baerns, Catal. Lett. 29 (1994) 261.
[33] J. Kucˇera, P. Nachtigall, J. Phys. Chem. B 108 (2004) 16012.
[34] J.A. Wang, O. Novaro, X. Bokhimi, T. López, R. Gómez, J. Navarrete, M.E. Llanos,
E. López-Salinas, J. Phys. Chem. B 101 (1997) 7448.
[35] M.J. Climent, A. Corma, S. Iborra, K. Epping, A. Velty, J. Catal. 225 (2004) 316.