Methanation of carbon dioxide on Ru/Al2O3: Catalytic activity and infrared study

Article abstract of DOI:10.1016/j.cattod.2015.12.010

3% Ru/Al₂O₃ catalyst is active in converting CO₂ into methane at atmospheric pressure. At 673?K and above the thermodynamic equilibrium is nearly attained. At 623?K CH₄ yield is above 85%. CO selectivity increas

Full text of DOI:10.1016/j.cattod.2015.12.010

G Model
CATTOD-9931; No. of Pages8
ARTICLE IN PRESS
Catalysis Today xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Methanation of carbon dioxide on Ru/Al2O3: Catalytic activity and
infrared study
Gabriella Garbarinoa,∗, Daria Bellottib, Elisabetta Finocchioa, Loredana Magistrib,
Guido Buscaa
a Università degli Studi di Genova, Dipartimento di Ingegneria Civile, Chimica e Ambientale (DICCA), P.zzale J.F. Kennedy, 1 I-16129 Genova, Italy
b Università degli Studi di Genova, Dipartimento di Ingegneria Meccanica, Energetica, Gestionale e dei Trasporti (DIME), Via all’Opera Pia, 15 I-16145
Genova, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
3% Ru/Al2O3 catalyst is active in converting CO2 into methane at atmospheric pressure. At 673 K and
above the thermodynamic equilibrium is nearly attained. At 623 K CH4 yield is above 85%. CO selectivity
increases by decreasing reactants partial pressure apparently more than expected by thermodynamics.
The reaction order for CO2 partial pressure is confirmed to be zero, while that related to hydrogen pressure
is near 0.38 and activation energy ranges 60–75 kJ/mol. Arrhenius plot demonstrates that only at reduced
reactant partial pressure (3% CO2) or high contact times, a contribution due to some diffusional limitation
is present. IR study shows that the H2—reduced catalyst has high-oxidation state Ru oxide species able
to oxidize CO to CO2 at 173–243 K, while after oxidation/reduction cycle the alumina surface acido-basic
sites are freed and the catalyst surface contains both extended Ru metal particles and dispersed low
valence Ru species. IR studies show that the formation of methane, both from CO and CO2, occurs when
both surface carbonyl species and surface formate species are observed. Starting from CO2, methane is
formed already in the low temperature range, i.e., 523–573 K, even when CO is not observed in the gas
phase.
Received 3 August 2015
Received in revised form 29 October 2015
Accepted 18 December 2015
Available online xxx
Keywords:
Methanation
CO2
Ruthenium on alumina
Hydrogenation
Methane
Surface species
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
have also been reported to be very active, a number of studies
show good performances of Ru/Al2O3 catalysts [9–11], essentially
The conversion of CO2 into methane (methanation)
better than Ni/Al2O3 catalysts [12]. On the other hand, Ru/Al2O3
catalysts are commercial catalysts for CO methanation for low tem-
perature applications (T < 443 K), such as the Clariant METH-150
catalyst that contains 0.3% ruthenium on alumina [13]. Interest-
ingly, supported ruthenium catalysts are also reported to be the
best for CO hydrogenation to higher hydrocarbons, i.e., the Fischer
Tropsch process [14,15]. Obviously, for the development of a per-
formant CO2 methanation process, catalysts producing methane
with high selectivity, thus with low CO and higher hydrocarbons
coproduction, must be developed.
A point closely related to reaction selectivity is that of reaction
mechanism. Until the 1970s, the mechanism of CO methanation
was supposed to occur through oxygenated intermediates [6], sup-
ported also by more recent spectroscopic studies [16]. However,
most recent studies tend to prefer a “via carbide” mechanism,
mainly based on “surface science” investigations performed on
metal monocrystals. As for the mechanism of CO2 methanation, an
CO2 + 4 H2 → CH4 + 2H2O
(1)
is one of the possible ways to utilize carbon dioxide thus reduc-
ing emissions [1], when hydrogen produced by renewable raw
materials or using waste energy is available [2]. To date, commercial
catalysts optimized for methanation of feeds primarily composed
of carbon dioxide are apparently lacking. Conventional metha-
nation catalysts, optimized to convert feeds containing primarily
carbon monoxide, have been mostly tested as CO2 methanation
catalysts and generally found active [3]. Although commercial Ni-
based CO methanation catalysts have been found to be active
also for CO2 methanation [4,5] they usually also coproduce sig-
nificant amounts of CO. Ru based catalysts have been reported
since decades to be among the best catalysts for CO2 methana-
tion [6]. Although Ru/TiO2 [7], Ru/SiO2 and Ru/ZSM5 zeolite [8]
∗
Corresponding author. Fax: +39 103536028.
E-mail address: Gabriella.Garbarino@unige.it (G. Garbarino).
http://dx.doi.org/10.1016/j.cattod.2015.12.010
0920-5861/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: G. Garbarino, et al., Methanation of carbon dioxide on Ru/Al2O3: Catalytic activity and infrared study,
Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2015.12.010

G Model
CATTOD-9931; No. of Pages8
ARTICLE IN PRESS
2
G. Garbarino et al. / Catalysis Today xxx (2016) xxx–xxx
additional central point concerns the possible role of CO, produced
by the reverse Water Gas Shift reaction (rWGS)
outlet concentrations calculated from the absorbances of CO, CO2,
CH4 and the measured inlet and outlet total flows (which allow
to take into account the variation of the number of moles dur-
ing the reaction), CO2 conversion (XCO2 ), selectivities and yields to
products, Si and Yi, have been calculated [20]. They are defined as:
CO2 + H2 → CO + H2O
(2)
as an intermediate, or/and as a competitor [17,18], the possible
intermediacy of carbide species as well as the possible role of the
support in adsorbing and activating CO2 [19].
To have more information on the surface chemistry of an active
Ru/Al2O3 catalyst and on mechanistic aspects, we deepened our
previous studies with further flow reactor experiments and by
applying IR spectroscopy methods.
FCO − FCO
out
in
2
2
XCO
=
(3);
(4);
(5);
2
FCO
in
2
Fi
Si =
Yi =
FCO − FCO
out
in
2
2
Fi
2. Experimental
FCO
in
2
where Fi is the molar flow rate of i (i.e., CO and CH4), while FCO2 is
the molar flow rate of CO2 and they were all expressed in mol/min.
From the kinetic orders determined in our previous study and
from the production rate of CH4 we calculated the kinetic constant
k [mol/(min × gcat × atm0.38)] designing for each test the Arrhenius
plot in order to have a better understanding of the controlling
regime in the different conditions [12,21].
It should be remarked that, as already reported in our previous
work [12], we did not observe any coke formation in our experi-
ments performed at the 8 h timescale. Carbon balance is 100% 1%
to our calculations and no evidence of coke was obtained from
the catalyst weight measure nor from IR, UV–vis, FE-SEM and XRD
studies.
2.1. Catalyst characterization
The catalyst used in this study is a 3% (wt%) Ru/␥-Al2O3 com-
mercial catalyst (SBET 150 m2/g) from Acta S.p.A. (Crespina, Pisa,
Italy).
IR spectroscopy experiments were performed using a Nicolet
Nexus FT instrument. Pressed disks of the pure catalysts powders
were treated “in situ” by using an infrared cell connected to a con-
ventional gas manipulation/outgassing ramp in two different ways:
by a simply reduction in H2 (500 torr) at 673 K for 1 h followed
by vacuum treatment at 773 K for 30 min. Alternatively, an oxi-
dation step in air (200 torr) at 673 K for 30 min followed by the
same reduction procedure was conducted. The catalyst disk was
first submitted to activation and then CO adsorption and reaction
experiments were carried out.
3. Results and discussion
CO adsorption was performed at 140 K by the introduction of a
known dose of the gas (10 torr) inside the low temperature infrared
cell containing the previously activated wafers. IR spectra were
recorded during evacuation upon warming at increasing tempera-
tures between 140 K and 673 K.
For the reaction experiments a mixture of H2 and CO2 (17 torr
CO2 and 170 torr on IR line and cell) and one of H2 and CO (17 torr
CO and 170 torr on IR line and cell) were put in contact with the cat-
alyst disk and the reaction was performed step-by-step in-between
523–773 K for ten minutes at each temperature; for each reaction
temperature IR spectra of both the gas phase and the catalyst sur-
faces were acquired.
3.1. CO2 methanation in flow reactor
In Fig. 1, the results of flow reactor CO2 methanation experi-
ments (line + symbols) on the 3% Ru/Al2O3 catalyst are reported
and compared with the respective thermodynamic equilibrium for
each condition. The three experiments have been performed with
the same H2/CO2 ratio of 5 at 1 atm and the same space velocity
(55,000 h−1), but with different reactant concentrations. In all cases,
no other products than methane and CO were detected. The C bal-
ance was always fully fulfilled and no evidence of carbon formation
was found using IR, UV–vis, FE-SEM and XRD techniques.
The fresh catalyst is almost not active at 523 K and is only poorly
active at 573 K (3% CO2 conversion), with a small methane yield
slightly increasing with time on stream. In the step performed at
623 K, the catalyst starts to have significant activity that definitely
increases with time on stream, due to an activation or “condi-
tioning” step. According to our previous study [12], the catalyst
is activated essentially by reduction from a partially cationic to a
metallic state. In these conditions, selectivity to methane is 100%,
except for the most diluted conditions, where CO is coproduced
in small amounts (2–5% yield). At 673 K, CO2 conversion is high
(from 80% to 84% and from 72% to 79% for 9% CO2 and 6% CO2
cases, respectively) but still growing with time on stream (Table S1),
showing that a “conditioning” effect was still in progress, except
again for the most diluted conditions, where the catalyst appears
already stable, achieving a CO2 conversion of 60% with a methane
yield of 47%. At this temperature, CO was also formed together with
methane, in particular in the most diluted conditions with a yield
of 14%. On the other hand, at the end of the experiments at 673 K
the conversion of CO2, the methane yield and selectivity depend
clearly on concentration of the reactants, being larger (83–84% for
CO2 conversion and 82–83% methane yield) when the reactants
were more concentrated. By comparison with calculated thermo-
dynamic equilibrium data (Fig. 1 and Table S2 calculated according
to [22]), CO2 conversion is still markedly lower than the equilib-
rium one. Thus, the reaction is still under kinetic control, maybe
2.2. Catalytic experiments
Catalytic experiments were carried out in a fixed-bed tubu-
lar silica glass flow reactor, operating isothermally, loaded with
700 mg of silica glass particles (60–70 mesh sieved) as an inert
material mixed with variable amounts of catalyst powder. Different
gaseous mixtures of CO2 and H2 (with excess H2) diluted with nitro-
gen were fed, with a total gas flow of 75 mlNTP/min. Temperature
was varied step by step in-between 523 K and 773 K (ascending
temperature experiment) and back down to 523 K (descending
temperature experiment). GHSV calculated as volumetric flow rate
(NTP), v [ml/h] versus catalyst volume Vcat [ml] was varied in
between 15000 and 55000 h−1. These values correspond to 5300
and 8000 h−1 if space velocity is calculated taking into account the
entire bed volume Vtot, constituted by silica glass and catalyst pow-
der. Hereinafter we will always refer to the GHSV calculated on the
catalyst volume.
Products analysis was performed on line using a Nicolet 6700
FT-IR instrument. Frequencies where CO2, CH4 and CO molecules
absorb weakly have been used (2293 cm−1 for CO2, 2100 cm−1 for
CO, 1333 cm−1 for CH4, after subtraction of baseline water absorp-
tion) with previous calibration using gas mixtures with known
concentrations, in order to have quantitative results. Produced
water was mostly condensed before the IR cell. From the inlet and
Please cite this article in press as: G. Garbarino, et al., Methanation of carbon dioxide on Ru/Al2O3: Catalytic activity and infrared study,
Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2015.12.010

G Model
CATTOD-9931; No. of Pages8
ARTICLE IN PRESS
G. Garbarino et al. / Catalysis Today xxx (2016) xxx–xxx
3
Fig. 1. CO2 hydrogenation results in terms of CO2 conversion (X) and yields to CO and methane (Yi) for different reactants partial pressure at 55,000 h−1 (82.2 mg catalyst)
in the ascending temperature (top) and descending temperature (bottom) experiments (line + symbol). Thermodynamic equilibrium displacement (line) is also reported for
each reactant compositions.
0.001
0.0005
CH4 [mol/(min*gcat)]
r
523 K
493 K
0
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
p
CO2 [atm]
Fig. 2. Methane production rate as a function of CO2 partial pressure (PCO2 ) and parametric as a function of temperature.
Please cite this article in press as: G. Garbarino, et al., Methanation of carbon dioxide on Ru/Al2O3: Catalytic activity and infrared study,
Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2015.12.010

G Model
CATTOD-9931; No. of Pages8
ARTICLE IN PRESS
4
G. Garbarino et al. / Catalysis Today xxx (2016) xxx–xxx
with a minor contribution due to diffusion. However, CO yield is
larger, in the experiment at the lowest concentrations, than that
expected by thermodynamics.
In the later steps of the experiments (higher temperatures and
descending temperatures), with all the three feeds, the catalyst per-
formance was stable with time, showing that “conditioning” was
complete and irreversible.
From the analysis reported in Fig. 3 the calculated apparent acti-
vation energy at low contact times is in the range 60–75 kJ/mol in
all cases not far from previous literature data [23]. These values are
typical for chemical kinetic regimes, thus assuring that, at least in
those conditions, diffusion limitations are negligible.
The trend of CO formation that is higher for the more diluted
conditions suggests instead that the reaction orders for CO forma-
tion might be even negative, and/or a product inhibition effect can
occur for the rWGS reaction producing CO. Alternatively, diffusion
limitations for hydrogen availability can occur in these conditions.
At 723 K, conversion with the most diluted feed was increased
with respect to the step at 673 K, while in the other cases conversion
decreased. A further decrease of conversion was found at 773 K, in
all cases (66–69% CO2 conversion). Thermodynamic calculations,
whose results are summarized in Fig. 1 and extensively reported
in Table S2, reveal that in our conditions we are only slightly
below thermodynamic equilibrium at 773 K and 723 K that implies
a decrease of CO2 conversion by increasing temperature. In agree-
ment with the expected thermodynamically-driven behavior, CO2
conversion and methane yield decrease at 773 K, and re-increase
in the decreasing temperature steps at 723 K, 673 K and 623 K. We
can note that CO2 conversion and CH4 yield observed at 673 K and
623 K in the decreasing temperature step are markedly higher than
those observed, at the same temperature, in the increasing temper-
ature experiment; thus confirming catalyst “conditioning”. In the
step at 673 K CO2 and CH4 amounts still agree to be near equilib-
rium. Also CO is formed and its concentration is definitely higher
when the experiment is conducted with lower reactant concentra-
tions. In this case, CO yield at 773 K and 723 K is markedly larger
than expected by equilibrium calculation, if the methanation and
rWGS reactions are both considered. However, CO yield could be
even higher if rWGS only would be taken into account, supposed
methanation being hindered kinetically.
By further decreasing temperature to 623 K, CO2 conversion
further increased (88–84%), as methane yield did, while CO pro-
duction decreases very much. Here, conversion increases slightly
with increasing reactant concentration. However, we are now far
from thermodynamic equilibrium, showing that kinetics governs
the system at this lower temperature. This is true, even more, at
573 K and 523 K, where catalytic activity is progressively reduced
and CO2 conversion achieves the lowest value for the descending
temperature experiment (11–20%). In spite of this, methane is still
formed at 523 K with a 12–20% yield. Thus, the “conditioned” cat-
alyst is still active at 523 K. In this low temperature step the effect
of reactant pressure on the conversion and methane yield is really
not evident.
3.2. IR study of low temperature CO adsorption
In Fig. 4, the spectra associated to CO adsorption at 140 K on
the Ru/Al2O3 catalyst simply reduced in hydrogen are reported.
A main band is formed at 140 K centered at 2151 cm−1, typically
assigned to CO interacting with the surface hydroxyl groups of alu-
mina [24,25]. Additionally a weaker band is observed at 2025 cm−1
,
with a shoulder at 2000 cm−1. Upon outgassing at 140 K the band
at 2151 cm−1 quickly disappears confirming its assignment to H-
bonded CO and shifting towards higher wavenumbers (2159 cm−1).
Instead, the band found at 2025–2000 cm−1 does not reduce its
intensity until warming to r.t. (room temperature), showing that it
is due to strongly bonded CO to zerovalent Ru◦ [24].
Interestingly, upon warming under dynamic outgassing, a sharp
band grows at 2344 cm−1, typically due to O-bonded linear CO2.
This band starts to be observed at 170 K and raises its maxi-
mum intensity at 240 K. By outgassing at higher temperature it
decreases rapidly down to disappear at 270 K, due to CO2 desorp-
tion. The weak band at 2278 cm−1 is due to the corresponding 13CO2
adsorbed species, according to the natural abundance of C isotopes.
This shows that the sample surface, in spite of the reducing pre-
treatment, still retains oxidation functionality, oxidizing CO to CO2.
This functionality is certainly associated to unreduced Rux+, whose
charge is balanced by anions such as active oxygen species. These
ruthenium oxide species reacts with CO reducing itself. Below
170 K this reaction is kinetically hindered and a band characteriz-
ing the interaction of CO with such unreduced Ru species should be
observed. This band is actually superimposed by that of CO interact-
ing with OH groups. In fact, a number of unresolved components
can be found on this band, whose maximum shifts to higher fre-
quencies upon outgassing and warming, but that shows visible
shoulders at 2165 cm−1 at the higher temperature and at 2135 cm−1
at the lowest one. Indeed, the interaction of CO with RuO2 (110)
monocrystals has been studied by Ertl and coworkers using surface
techniques, showing a CO stretching band at 2115 cm−1, and the
easy oxidation of CO to CO2 above 300 K [26]. On the other hand, CO
adsorbed on RuO2/SiO2 has been reported to adsorb at 2125 cm−1
[27]. These data confirm that simple reduction in hydrogen is not
sufficient to fully reduce the catalyst and activate it, as discussed
previously [12].
In Fig. 2, the effect of CO2 partial pressure on reaction rate is
shown for experiments performed with the same H2 partial pres-
sures, at two different temperatures: we confirm that the reaction
order with respect to CO2, determined at 493 K in differential reac-
tor hypothesis [12], essentially zero, remains still valid at 523 K
taking into account the experimental error.
In Fig. 3, the Arrhenius plots of the all catalytic tests are
reported. The experimental results obtained in the descending
temperature experiments, in the temperature range in-between
623 K and 493 K have been considered. Those experiments con-
cern “conditioned” catalysts and refer to conditions were kinetics
is determinant. The kinetic constant evaluation (k, expressed as
[mol/(min × gcat × atm0.38) was made dividing the rate of produc-
tion of methane (rCH4) by the partial pressure of H2 elevated at its
reaction order, previously determined (0.38 [12]). A linear behav-
ior can be assumed in all cases at least in the temperature region
493–623 K, except at high contact times, i.e., 30000 h−1 (at 623 K)
and 15000 h−1 (at 573 and 623 K). The evident lack of linearity of
the plot at high contact times is due to a contribution of diffu-
sional phenomena, since the high conversion, or the approaching of
the thermodynamic equilibrium. An analogous situation was found
when the lowest partial pressures of the two reactants were used
(3% CO2–15% H2) where again the diffusional effect may be relevant.
In Fig. 5 the spectra relative to a similar experiment performed
on preoxidized and later reduced Ru/Al2O3 catalyst are reported.
At 140 K, the band due to CO interacting with OH groups is still the
most intense, but is now found at 2160 cm−1. However, two other
strong adsorptions are observed. A band at 2192 cm−1, decreasing
and shifting up progressively upon warming down to 2207 cm−1
,
can be assigned to CO interacting with Al3+ cations. These species
disappear after outgassing at T > 240 K.
A strong and quite broad band, centered at 2037 cm−1, is
also observed. Its intensity decreases and its position is shifted
towards lower frequency during sample outgassing upon warm-
ing. In particular, this band, presents its maximum at 2037 cm−1
upon outgassing at 140 K and then is centered near 1990 cm−1
upon outgassing at 423 K. Above 423 K the observed features
Please cite this article in press as: G. Garbarino, et al., Methanation of carbon dioxide on Ru/Al2O3: Catalytic activity and infrared study,
Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2015.12.010

G Model
CATTOD-9931; No. of Pages8
ARTICLE IN PRESS
G. Garbarino et al. / Catalysis Today xxx (2016) xxx–xxx
5
1/T [1/K]
1/T [1/K]
0.0016
-5.0
0.0017
0.0018
0.0019
0.0020
0.0016
-5.0
0.0017
0.0018
0.0019
0.0020
-5.5
-6.0
-6.5
-7.0
-7.5
-8.0
-8.5
-5.5
-6.0
-6.5
-7.0
-7.5
-8.0
-8.5
-1
55000 h 3%CO -30%H
2
2
-1
-1
55000 h 6%CO -30%H
55000 h 7% CO -30%H
2
2
2
2
-1
-1
30000 h 6%CO -30%H
55000 h 3% CO -15%H
2
2
2 2
-1
-1
15000 h 6%CO -30%H
55000 h 9% CO -45%H
2 2
2
2
Fig. 3. Arrhenius’s plot (k [mol/(min × gcat × atm0.38)] of all the catalytic tests performed with experimental points (symbols) as a function of 1/T.
2025
2000
2151
140 K +10 torr CO
140 K 10 torr CO
0.1
140 K
2344
240 K vacuum
190 K vacuum
0.01
180 K vacuum
180 K vacuum
RT
140 K+10 torr CO vacuum
2100
2060
2020
1980
1940
2159
Wavenumbers (cm-1)
170 K vacuum
273 K vacuum
140 K
2278
RT
2400
2300
2200
2100
2000
Wavenumbers (cm-1)
Fig. 4. CO adsorption at liquid nitrogen temperature on the prereduced sample. IR surface spectra at 140 K, in contact with 10 torr of CO and upon outgassing until RT. Inset:
magnification of the low frequency region.
[24]. A triplet similar to our at 2142, 2075 and 2004 cm−1 has been
observed by Darensbourgh and Ovalles [31] on a RuCl3/Al2O3 cata-
lyst and assigned to Ru polycarbonyls, and by Elmasides et al. [32]
on a 0.5% Ru/Al2O3 sample after reduction in hydrogen and has
been assigned to polycarbonyl species on well dispersed ionic Ru.
A similar triplet was also observed by Chin et al. [33] on 1% Ru/Al2O3
after prereduction in hydrogen by in copresence of CO with oxy-
gen. These authors assigned the triplet to the superimposition of
two doublets due to dicarbonyl species on ionic and zerovalent Ru.
The residual band at 1960 cm−1 can be due to bridging CO on Ru
further decrease in intensity without clear shifting. They are
observed at 2142, 2075, 2004 and 1960 cm−1 approximately.
The band shifting from 2037 cm−1 down to ca 1990 cm−1 is
clearly due to CO adsorbed on extended Ru metal particles, shifting
down upon decreasing coverage due to the well- known coupling
effects depending on the coverage. The range of the frequency does
not allow to distinguish between flat and stepped faces such as Ru
(001) [28], Ru (0120) [29] and Ru (109) [30]. The non-shifting bands
at 2142, 2075, 2004 and 1960 cm−1 instead should be assigned
to CO interacting with isolated or clustered Ru◦ or Ruı+ centers
Please cite this article in press as: G. Garbarino, et al., Methanation of carbon dioxide on Ru/Al2O3: Catalytic activity and infrared study,
Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2015.12.010

G Model
CATTOD-9931; No. of Pages8
ARTICLE IN PRESS
6
G. Garbarino et al. / Catalysis Today xxx (2016) xxx–xxx
Fig. 5. CO adsorption at liquid nitrogen temperature on the preoxidized and reduced sample. IR surface spectra at 140 K (spectrum 24), in contact with 10 torr of CO (spectrum
1), upon outgassing in-between 140–298 K (spectra 2–15) and upon heating 673 K (spectra 16–23).
Fig. 6. CO2 + H2 reactivity in the IR cell (total pressure 170 torr, CO2 17 torr). Gas phase (A) and surface adsorbed species (B) after 10 min at the reaction temperature.
clusters. Although a more precise assignment of these bands can-
not be proposed on the basis of the reported data, it seems very
likely that the reduced Ru catalyst contains both extended metal
particles and likely more than one family of small clusters or iso-
lated Ru in the zerovalent or ionic state. In any case, the hydrogen
reduced catalyst lost its ability to oxidize CO, highly oxidized Ru
species (likely Ru4+) having been reduced.
The oxidizing and reducing pretreatments give rise to differ-
ent ruthenium species on the surface than those observed by pure
reduction in H2. This is an useful information in connection to con-
Please cite this article in press as: G. Garbarino, et al., Methanation of carbon dioxide on Ru/Al2O3: Catalytic activity and infrared study,
Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2015.12.010

G Model
CATTOD-9931; No. of Pages8
ARTICLE IN PRESS
G. Garbarino et al. / Catalysis Today xxx (2016) xxx–xxx
7
Fig. 7. CO + H2 reactivity in the IR cell (total pressure 170 torr, CO 17 torr). Gas phase (A) surface adsorbed species (B) after 10 min at the reaction temperature.
CO2 (2344 cm−1, OCO asymmetric stretching) and bicarbonate
species (1647 cm−1, 1449 cm−1, asymmetric and symmetric COO
stretching, 1228 cm−1, OH deformation) which is associated to the
weak basicity of alumina support [25]. The position of these bands
is the same observed when these species are formed on pure alu-
mina, thus indicating that they are interacting with alumina more
than with ruthenium species. On the other side, the spectra of
exposed hydroxyl groups on this catalyst are consistent with those
reported for typical ␥-aluminas, maybe slightly modified by ruthe-
nium deposition [34]. At 523 K, bands assigned to formate species
(1594 cm−1 asymmetric COO-stretching, 1394 cm−1 CH deforma-
tion, 1375 cm−1 COO-symmetric stretching) are also observed. The
position of these bands is the same observed when these species are
formed on pure alumina, thus indicating that they are not interact-
ing with ruthenium. Together, weak bands at 2050 and 1970 cm−1
are observed. These features should be due to small amounts of
strongly adsorbed CO species. They look similar to the doublets
typical of Ru dicarbonyl species as such and adsorbed on alumina
[31]. The band at 1630 cm−1, associated also to strong absorption
in the 3600–3200 cm−1 region, is due to adsorbed water.
ditioning and regeneration and to the exposure of the catalyst to
different atmpspheres as well. This behavior is likely connected not
only to oxidation/reduction of Ru species but also to the possible
removal of chlorine, residual from catalyst preparation, as reported
previously [12].
The difference between the spectra reported in Figs. 4 and 5 indi-
cates that after impregnation, and even after reduction in hydrogen,
highly charged cationic Ru species interact on alumina surface
defects where also Lewis acidic Al3+ are located. This interaction
inhibits or kills the Al3+ Lewis acid sites, which interact with CO
(band above 2180 cm−1 [34]) and not observed. Small amounts of
isolated or clustered Ru◦ are also observed probably located on the
main faces (CO bands at 2025–2000 cm−1). The further treatment
of ruthenium oxidation and reduction frees Lewis acidic Al3+ cen-
ters (CO bands at 2192–2207 cm−1) now evident, corresponding to
partial agglomeration of Ru species, producing particles, clusters
and isolated Ru atoms.
3.3. IR study of CO2 hydrogenation
This experiment shows that CH4 gas is formed before (at lower
temperature, i.e. 523 K) than CO gas (673 K), thus suggesting that
CO gas may be not an intermediate in CO2 methanation. However,
also the amount of adsorbed CO is very small. The data also suggest
that the way bicarbonate-formate-methane is a possible way to
CO2 methanation on Ru catalyst. The formation of methoxy groups
is not evident nor the formation of CHx species can be observed by
looking at the CH stretching region.
In Fig. 6A the gas phase spectra recorded upon contacting of
the prereduced catalyst disk with CO2 + H2 mixture in the IR cell
are reported. The reaction is very slow at 523 K producing very
small amounts of methane (rotovibrational bands with maxima at
3016 cm−1 and 1305 cm−1, asymmetric stretching and deforma-
tion respectively) and only a small decrease of the band of CO2
(rotovibrational bands with minima at 3716, 3613 cm−1, combi-
nation modes, and 2349 cm−1, asymmetric stretching, and with
maximum at 669 cm−1, deformation mode). At 573 K and 673 K
the reaction is faster than that achieved at 523 K (lowest tem-
perature used in the experiments), resulting in the detection of
stronger methane signals. Interestingly, only methane is found
in the gas phase at 523 K and 573 K, while very small amounts
of CO (rotovibrational band with minimum at 2140 cm−1) are
also observed at 673 K and above. The spectra of the surface
species recorded upon the experiment are reported in Fig. 6B.
Adsorption of the catalyst of the CO2 + H2 mixture at r.t. (Room
Temperature, i.e., 298 K) produces O-bonded linearly adsorbed
3.4. IR study of CO hydrogenation.
In Fig. 7A the gas phase spectra recorded upon contacting
of the prereduced catalyst with CO + H2 mixture. The reac-
tion already occurs at 523 K producing significant amounts of
methane and water. The reaction is faster at 673 K and above.
No other hydrocarbons are found, unlike previous experiments
performed in similar conditions on Ru/TiO2 experiments [15,35].
The spectra of the surface species recorded upon the exper-
Please cite this article in press as: G. Garbarino, et al., Methanation of carbon dioxide on Ru/Al2O3: Catalytic activity and infrared study,
Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2015.12.010

G Model
CATTOD-9931; No. of Pages8
ARTICLE IN PRESS
8
G. Garbarino et al. / Catalysis Today xxx (2016) xxx–xxx
iment are reported in Fig. 7B . Adsorption of the catalyst of
the CO + H2 mixture at RT produces carbonyl species absorb-
Appendix A. Supplementary data
ing at 2130 and 2077 cm−1, and bicarbonate species (1647 cm−1
1449 cm−1, asymmetric and symmetric COO stretching, 1228 cm−1
,
,
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.cattod.2015.12.
010.
OH deformation). According to the above study, the doublet
at 2130 and 2077 cm−1 could be assigned to dicarbonyls on
ionic Ru. At 523 K, formate species (1594 cm−1 asymmetric COO
stretching, 1394 cm−1 CH deformation, 1375 cm−1 COO sym-
metric stretching) are also observed, together with water
(1630 cm−1), and strong bands due to carbonyl species
(2045, 2004, 1975 cm−1). This spectrum suggests that at this tem-
perature Ru is reduced forming metal particles and clusters of
zerovalent Ru. At 673 K formate species disappeared, while weak
bands of carbonyl species are observed (2030, 1961 cm−1). The
spectra show that carbonyl species are likely intermediate in CO
methanation, although formate species may also play a role. Again,
in contrast to previous experiments performed with Ru/TiO2 cata-
lysts [15], the formation of adsorbed CHx species cannot observed
by looking at the CH stretching region. This suggests indeed that
evident CH species are precursors for higher hydrocarbons (Fischer
Tropsch synthesis) more than intermediates in methanation.
References
[1] G. Centi, S. Perathoner, Catal. Today 148 (2009) 191–205.
[2] S.K. Hoekman, A. Broch, C. Robbins, R. Purcell, Int. J. Greenhouse Gas Control 4
(2010) 44–50.
[3] W. Wang, J. Gong, Front. Chem. Sci. Eng. 5 (2011) 2–10.
[4] S. Abelló, C. Berrueco, D. Montané, Fuel 113 (2013) 598–609.
[5] G. Garbarino, P. Riani, L. Magistri, G. Busca, Int. J. Hydr. En. 39 (2014)
11557–11565.
[6] G.A. Mill, F.W. Steffgen, Chem. Rev. 8 (1973) 159.
[7] M. Matwood, R. Doepper, A. Renken, Appl. Catal. A 151 (1997) 223–246.
[8] S. Scirè, C. Crisafulli, R. Maggiore, S. Minicò, S. Galvagno, Catal. Lett. 51 (1998)
41–45.
[9] D.J. Darensbourgh, C. Ovalles, Inorg. Chem. 25 (1986) 1603–1609.
[10] J.H. Kwak, L. Kovarik, L. Szanyi, ACS Catal. 3 (2013) 2449–2455.
[11] C. Janke, M.S. Duyar, M. Hoskins, R. Farrauto, Appl. Catal. B 152 (2014)
184–191.
[12] G. Garbarino, D. Bellotti, P. Riani, L. Magistri, G. Busca, Int. J. Hydrogen Enengy
40 (2015) 9171–9182.
[13] http://www.catalysts.clariant.com/bu/Catalysis/internet.nsf/
023cfbb98594ad5bc12564e400555162/
4. Conclusion
4cf78747d693c1f0c1257ad0002d2c04?OpenDocument.
[14] G.P. van der Laan, A.A.C.M. Beenackers, Catal. Rev. Sci. Eng. 41 (1999) 255–318.
[15] J.M. González Carballo, E. Finocchio, S. García, S. Rojas, M. Ojeda, G. Busca, J.L.
García Fierro, Catal. Sci. Technol. 1 (2011) 1013–1023.
[16] V. Sanchez-Escribano, M.A. Larrubia Vargas, E. Finocchio, G. Busca, Appl. Catal.
A 316 (2007) 68–74.
[17] K. Yaccato, R. Corhart, A. Hagemeyer, A. Lesik, P. Strasser, A.F. Volpe, H. Turner,
H. Weinberg, R.K. Grasselli, C. Brooks, Appl. Catal. A 296 (2005) 30–48.
[18] A.M. Abdel-Mageed, S. Eckle, H.G. Anfang, R.J. Behm, J. Catal. 298 (2013)
148–160.
[19] F. Ussa Aldana, K. Ocampo, B. Kobl, F. Louis, M. Thibault-Starzyk, P. Daturi, S.
Bazin, A.C. Thomas, Catal. Today 215 (2013) 201–207.
[20] O. Levenspiel, Chemical Reaction Engineering, 3rd, Wiley New York, 1999, pp.
87.
[21] H.S. Fogler, Elements of Chemical Reaction Engineering, fourth, Pearson
Education International, New Jersey, 2006.
The main conclusions can be summarized as follows.
The 3% Ru/Al2O3 catalyst is active in converting CO2 into
methane at atmospheric pressure. At 673 K the performances
observed approach closely the thermodynamic equilibrium. At
623 K the maximum CO2 conversion and CH4 yield are obtained
both in the range above 85%. In the kinetically controlled regime
(T < 623 K, with the exceptions discussed in the text), the reaction
order for CO2 conversion with respect to CO2 pressure is zero, while
that related to hydrogen pressure is near 0.38. The activation energy
ranges 60–75 kJ/mol. The selectivity to CO increases by decreases
reactants partial pressure apparently more than expected by ther-
modynamics, may be due to some diffusion limitation.
[22] R.D. Reid, J.M. Prausnitz, B.E. Poling, The Properties of Gases and Liquids, 4th,
McGraw-Hill, New York, 1986.
[23] M.R. Praire, A. Renken, J.G. Highfield, K. Ravindranathan Thampi, M. Grätzel, J.
Catal. 129 (1991) 130–144.
Even after reduction in hydrogen, the catalyst shows high-
oxidation state Ru oxide species able to oxidize CO to CO2 at
173–243 K. These species are dispersed on alumina crystal defects
(edges, corners, structural defects) thus perturbing the alumina
strongest surface acido-basic sites.
After oxidation/reduction cycles, the alumina surface acido-
basic sites are freed, and the catalyst surface contains both extended
Ru particles and dispersed low valence Ru species. CO adsorption
produces terminal carbonyls on extended faces and polycarbonyls
on dispersed Ru species, and CO on alumina Lewis sites as well. CO2
adsorbs essentially on the alumina support in the form of bicarbon-
ate species as well as in a molecular species.
[24] G. Hadjiivanov, Vayssilov, Adv. Catal. 47 (2002) 307–511.
[25] T. Montanari, L. Castoldi, L. Lietti, G. Busca, Appl. Catal. A 400 (2011) 61–69.
[26] J. Wang, C.Y. Fan, K. Jacobi, G. Ertl, Surf. Sci. 481 (2001) 113–118.
[27] M. Sorlino, G. Busca, Appl. Surf. Sci. 18 (1984) 268–272.
[28] H. Pfnur, D. Menzel, F.M. Hoffmann, A. Ortega, A.M. Bradshaw, Surf. Sci. 93
(1980) 431.
[29] J. Wang, Y. Wang, K. Jacobi, Surf. Sci. 488 (2001) 83–89.
[30] T. Zubkov, G.A. Morgan Jr., J.T. Yates Jr., O. Kuhlert, M. Lisowski, R. Schillinger,
D. Fick, H.J. Jansch, Surf. Sci. 526 (2003) 57–71.
[31] D.D. Darensbourgh, C. Ovalles, Inorg. Chem. 25 (1986) 1603–1609.
[32] C. Elmasides, D.I. Kondarides, W. Grunert, X.E. Verykios, J. Phys. Chem. B 103
(1999) 5227–5239.
[33] S.Y. Chin, C.T. Williams, M.D. Amiridis, J. Phys. Chem. B 110 (2006) 871–882.
[34] G. Busca, Adv. Catal. 57 (2014) 319–404.
IR studies show that the formation of methane, both from CO
and CO2, occurs when both surface carbonyl species and surface
formate species are observed on the surface. Starting from CO2,
methane is formed already in the low temperature range, i.e.,
523–573 K, even when CO is not observed in the gas phase. This
suggests that gas-phase CO might be not an intermediate in CO2
methanation.
[35] J.M. González Carballo, E. Finocchio, S. García-Rodriguez, M. Ojeda, J.L.G.
Fierro, G. Busca, S. Rojas, Catal. Today 214 (2013).
Please cite this article in press as: G. Garbarino, et al., Methanation of carbon dioxide on Ru/Al2O3: Catalytic activity and infrared study,
Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2015.12.010