Professional Documents
Culture Documents
C using feed compositions relevant to those of reformate gas streams, both in the
absence and in the presence of water. It has been found that, under conditions of solo-CO methanation,
all the investigated catalysts are able to completely and selectively convert CO at temperatures around
340
C, with the conversion of CO being somewhat higher for the Ru/platelet sample. For hydrogenation
of CO
2
alone, catalytic performance is not affected by the nature of the carbon nanobers used as support.
In combined hydrogenation of CO/CO
2
mixtures, catalytic performance for all the investigated catalysts is
poor since they promote the undesired reverse watergas shift reaction. However, addition of 30% water
vapour in the feed inhibits the reverse watergas shift, thereby enhancing CO hydrogenation. Results of
kinetic measurements show that the turnover frequency of CO conversion becomes 23 times higher
in the presence of steam over Ru/shbone and Ru/platelet samples over the whole temperature range
examined, whereas in the case of Ru/ribbon catalyst temperatures higher than 250
C are required in
order to achieve higher turnover frequency values. Carbon dioxide hydrogenation is not affected by the
presence of steam. For all experimental conditions investigated, selectivity toward methane increases
with increasing temperature at the expense of higher hydrocarbons and is enhanced with the addition
of water vapour in the gas mixture.
2010 Elsevier B.V. All rights reserved.
1. Introduction
One the major problems for the application of low temperature
polymer electrolytemembranefuel cells (PEMFCs) as power source
for electrically operated vehicles is the delivery of nearly-CO-free
feed gas, which becomes problematic whenever H
2
is generated
from fuels. As the FC anodes can be poisoned even by traces impu-
rities of CO, this gas has to be removed to a level below 50ppm for
PtRu anode electrocatalyst [13] and below 10ppm for Pt anode
electrocatalysts [46].
Theoretically, there are several methods to reduce carbon
monoxide to the levels acceptable for a fuel cell [7]. It is feasible to
separate hydrogen by diffusion thorough a CO ltering membrane
but themembraneis expensiveandit usuallyrequires acompressor
owingtohighpressure[8]. Themost studiedsystemfor theremoval
of nal traces of COover the last years has been the selective oxida-
tion of COin a H
2
-rich atmosphere (PROX) since it has the ability to
= 206kJ/mol (1)
0926-860X/$ see front matter 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.09.026
36 V. Jimnez et al. / Applied Catalysis A: General 390 (2010) 3544
This approach has the advantage that no oxidizing and/or inert
gases are mixed with the reformate stream and that the methane
produced, which is inert to the PEM fuel cell, can be utilized in
residential PEFC systems. However, depending on the operating
conditions and catalyst employed, reaction (1) may run in parallel
with the undesired methanation of CO
2
, which consumes signi-
cant quantities of valuable hydrogen (Eq. (2)), as well as with the
reverse watergas shift (RWGS) reaction (Eq. (3)), which shift CO
2
to CO [17].
CO
2
+4H
2
CH
4
+2H
2
O, H
= 165kJ/mol (2)
CO
2
+H
2
CO + H
2
O, H
= 41.1kJ/mol (3)
Consequently, it is important to develop selective CO metha-
nation catalysts characterized by high activity at sufciently low
temperatures, able to retard both the CO
2
and the RWGS reactions.
Methanation of COover different carried metal catalysts includ-
ing Ni [1821], Ru [18,2224] and Rh [21,2426], has been widely
investigated for the production of CH
4
from syngas [4,2729] and
recently also from a viewpoint of residual CO removal for PEM
FC applications [3036]. Ruthenium catalysts dispersed on metal
oxide carriers have been found to exhibit high activity for the solo-
methanation of CO [18,19,37] or CO
2
[20,21] as well as for the
co-methanation of CO/CO
2
mixtures [19,25,33]. In a recent study
[25] the effects of the nature of the metallic phase on the per-
formance of Al
2
O
3
-supported Ru, Rh, Pt and Pd catalysts for the
methanation of CO, CO
2
and their mixtures both in the absence and
in the presence of water in the feed has been studied. It has been
found that Ru and Rh are much more active hydrogenation cata-
lysts as comparedtoPt andPd, whichpromote the undesiredRWGS
reaction. Furthermore, the nature of the support may play a crucial
role in the mechanism of CO/CO
2
hydrogenation reactions, since
metal-support interactions can modify the catalytic properties of
the metallic phase [26,37,38]. Supports as carbon nanobers could
be an alternative to the classical ones (alumina, silica, TiO
2
, etc)
widely studied on the literature [25,16,38] due to their excellent
characteristics as the high purity of the material, high mechani-
cal strength and the mesopore nature which result in low internal
mass-transfer resistances [39].
In the present study, the catalytic performance of Ru cata-
lyst on selective CO
x
methanation using different types of carbon
nanobers (platelet, shbone and ribbon) has been investigated.
The inuence of water vapour in the feed was also studied.
2. Experimental
2.1. Support/catalyst preparation
CNFs synthesis was carried out by CVD method. The xed-bed
reactor, consisting on a quartz tube of 9cm diameter and 100cm
length, was locatedina horizontal electric furnace (JHHornos) with
an effective heating zone of 80cm. Thermocouple type K was used
for monitoring the temperature of the bed. Hydrogen and ethane
owrates werecontrolledbymass controllers (Brooks Instruments,
model 5850). Supported catalyst was taken in a quartz boat, which
was kept inside the heating zone during the experiment. Carbon
nanobers were grown at atmospheric pressure at different tem-
peratures: 450
C
(ribbon type). In each synthesis run 5g of the prepared catalyst
(Ni/SiO
2
) was placed in the centre of the reactor and activated by
heating (10
Cmin
1
) in a ow dry 20% (v/v) H
2
/He at the desired
reaction temperature. The reduced activated catalyst was thor-
oughly ushed with dry He for 1h before introducing the H
2
/C
2
H
4
ratio desired feed. The growth time was 1h and the space veloc-
ity 25,000h
1
. Separation of CNFs and catalyst particles to recover
the carbon material was carried out by repeated dissolution of the
support in hydrouoric acid (70%) for 15h under vigorous stirring,
ltering and washing with deionised water [40]. Upon treatment of
the catalystcarbon mixture, the metal component of the catalyst
was generally transferred into the solution.
Methanation catalysts were prepared employing the wet
impregnation method [25] by using Ru(NO)(NO
3
)
3
(Alfa Products)
as metal precursor salt andthe different types of CNFs obtained. The
resulting slurry was heatedat 1
Cmin
1
upto 90
Cunder continu-
ous stirring and maintained at that temperature until nearly all the
water evaporated. The solidresidue was driedat 110
Cfor 24hand
then reduced at 400
C in H
2
ow for 2h. The metal loading of the
catalysts thus prepared was 0.5wt.%.
2.2. Support/catalyst characterization
Surface area/porosity measurements were carried out using a
Micromeritics ASAP 2010 sorptometer apparatus with N
2
at 77K
as the sorbate. The samples were outgassed at 453K under vac-
uum(6.610
9
bar) for 16hprior toanalysis. Specic surface areas
were determined by the multi point BET method, total pore vol-
ume and sizes were evaluated using the standard BJH treatment
and micropore volume were evaluated using the t-plot equation.
The crystallinityof CNFs, the meancrystallite size andRuspecies
was determined using XRD analyses. These analyses were carried
out on a Philips XPert instrument using nickel ltered Cu Kradia-
tionthrougha primarymonochromator. The samples were scanned
at a rate of 0.02
step
1
over the range 5
2 80
Ctoavoidthe spill
over phenomenon [45]. Previously, the sample was pre-treated by
heating at 15
Cmin
1
in argon owup to 250
M
S
M
(5)
where d
M
is the mean crystallite diameter,
M
is the density of Ru
(12.3gcm
3
) and S
M
is the exposed surface area per gramof metal.
Temperature-programmed oxidation (TGA/TPO) was used to
determine the CNFs crystallinity. Analyses were performed on
10mg samples using a Perkin-Elmer TGA7 termogravimetric ana-
lyzer with a ow of 50cm
3
min
1
of 20% (v/v) O
2
/He mixture and
with a heating rate of 5
Cmin
1
up to 1000
C.
V. Jimnez et al. / Applied Catalysis A: General 390 (2010) 3544 37
Fig. 1. Representative images of the different type of carbon nanobers (CNFs): (a) SEM picture of general CNFs, (b) TEM picture of platelet CNFs, (c) TEM picture of shbone
CNFs and (d) TEM picture of ribbon CNFs.
Support morphology was examined by scanning electron
microscopy (SEM) using a JEOL JSM-6535F SEM unit with an
accelerating voltage of 10kV. The samples were deposited on a
standard aluminium SEM holder and coated with gold (Balzers
Union, Model MED 010). Transmission electron microscopy (TEM)
analysis employed was a Philips Tecnai 20T, operated at an accel-
eration voltage of 200keV. Samples were prepared by ultrasonic
dispersion in acetone with a drop of the resultant suspension evap-
orated onto a holey carbon supported grid.
TheRumetal loadingwas determinedbyatomic absorption(AA)
spectrophotometry, using a SPECTRA 220FS analyzer. Samples (ca.
0.5g) were treated in 2cm
3
HCl, 3cm
3
HF and 2cm
3
H
2
O
2
followed
by microwave digestion (523K).
2.3. Catalytic activity
Catalytic performance tests and kinetic measurements have
been carried out using an apparatus, consisting on a ow mea-
suring and control system, the reactor and an on-line analysis
system [25]. The ow system is equipped with a set of mass-ow
controllers (MKS) and a set of valves, which allows introduction
of the gas mixture to the reactor or to a by-pass loop, through
Table 1
Physicochemical properties of the CNFs and the reduced Ru catalysts.
CNFs support Ru-based CNFs catalysts
Platelet Fishbone Ribbon Platelet Fishbone Ribbon
BET surface area (m
2
/g) 286.0 202.0 68.0 183.6 175.9 93.9
Micropore area (m
2
/g)
a
57.0 (20%) 25.0 (13%) 2.0 (3%) 11.7 (6%) 20.6 (12%) 28.7 (31%)
Mesopore area (m
2
/g) 229.0 177.0 66.0 171.9 155.4 65.2
Micropore volume (cm
3
/g)
b
1.194 (0.62) 0.546 (0.62) 0.052 (0.64) 0.476 (0.67) 0.514 (0.66) 0.116 (0.67)
Mesopore volume (cm
3
/g)
c
0.409 (8.7) 0.977 (10.1) 0.246 (16.7) 0.403 (9.9) 0.378 (10.9) 0.232 (11.4)
Average pore size (nm) 6.9 9.3 14.1 6.1 8.8 9.7
npg
d
7.7 8.6 12.3
Weight loss temperature range
(
C)
e
380585 (519) 430618 (546) 545715 (640)
NaOH added (mmol/g) 0.75 0.56 0.19 0.78 0.58 0.40
Ru dispersion (%) 67.7 63.5 58.5
dRu,TEM (nm)
f
1.04 1.48 1.93
dRu, H
2
(nm)
g
1.41 1.50 1.63
a
In brackets: percentage of micropore area with respect to the total surface area.
b
Cumulative pore volumes obtained using HorvathKawazoe method. In brackets: mean micropore size in nm.
c
Cumulative pore volumes obtained using BJH method. In brackets: mean mesopore size in nm.
d
Number of graphene planes in the crystallites Lc/d
002
where Lc is the average crystal domain size along a direction perpendicular to the basal planes in a graphitic-type
structure, and d
002
is the average interlayer spacing.
e
In brackets: temperature at which the maximum of the oxidation temperature peak appears.
f
Average diameters of Ru particles determined by counting around 200 particles on the TEM images using the equation: dRu =
n
i
d
3
i
/
n
i
d
2
i
where n
i
is the number
of particle of diameter d
i
.
g
Average diameter of Ru particles measured using H
2
chemisorption technique using the expression dM =6/MSM where, M is the density of Ru (12.3gcm
3
) and SM is
the exposed surface area per gram of metal [20].
38 V. Jimnez et al. / Applied Catalysis A: General 390 (2010) 3544
1.0 0.8 0.6 0.4 0.2 0.0
0
50
100
150
200
250
300
350
Fishbone CNFs
Ru/Fishbone CNFs
N
2
v
o
l
u
m
e
a
d
s
o
r
b
e
d
/
d
e
s
o
r
b
e
d
(
c
m
3
/
g
)
N
2
v
o
l
u
m
e
a
d
s
o
r
b
e
d
/
d
e
s
o
r
b
e
d
(
c
m
3
/
g
)
1.0 0.8 0.6 0.4 0.2 0.0
0
50
100
150
200
250
300
350
Platelet CNFs
Ru/Platelet CNFs
N
2
v
o
l
u
m
e
a
d
s
o
r
b
e
d
/
d
e
s
o
r
b
e
d
(
c
m
3
/
g
)
1
0.000
0.025
0.050
0.075
0.100
0.125
10 1
0.0
0.1
0.2
0.3
0.4
0.5
M
e
s
o
p
o
r
e
V
o
l
u
m
e
(
d
V
/
d
l
o
g
D
)
Pore diameter (nm)
Platelet CNFs
1
0.000
0.015
0.030
0.045
0.060
0.075
10 1
0.1
0.2
0.3
0.4
M
e
s
o
p
o
r
e
V
o
l
u
m
e
(
d
V
/
d
l
o
g
D
)
Pore diameter (nm)
Fishbone CNFs
1
0.0000
0.0025
0.0050
0.0075
10 1
0.00
0.05
0.10
0.15
0.20
0.25
M
e
s
o
p
o
r
e
V
o
l
u
m
e
(
d
V
/
d
l
o
g
D
)
M
i
c
r
o
p
o
r
e
V
o
l
u
m
e
(
c
m
3
/
g
)
M
i
c
r
o
p
o
r
e
V
o
l
u
m
e
(
c
m
3
/
g
)
M
i
c
r
o
p
o
r
e
V
o
l
u
m
e
(
c
m
3
/
g
)
Pore diameter (nm)
Ribbon CNFs
1.0 0.8 0.6 0.4 0.2 0.0
0
20
40
60
80
100
120
140
160
180
Ribbon CNFs
Ru/Ribbon CNFs
Relative pressure (P/P
0
)
Relative pressure (P/P
0
)
Relative pressure (P/P
0
)
Fig. 2. N
2
adsorptiondesorption isotherms and micropore/mesopore volume distribution of the catalysts and the supports.
stainless steel tubing. When desired, water is introduced to the
system with the use of an HPLC pump (Marathon Scientic Sys-
tems), vaporized in a stainless steel evaporator maintained at
170
C and mixed with the gas stream coming from the mass-
ow controllers. The resulting gas mixture is then fed to the
reactor through stainless steel tubing maintained at 150
C by
means of heating tapes. The reactor consists of a 40cmlong quartz
tube (6mm OD) with an expanded 1cm long section in the mid-
dle (8mm ID), in which the catalyst sample is placed. Reaction
temperature is measured in the middle of the catalyst bed by
means of a K-type thermocouple placed within a quartz capil-
lary well, which runs through the cell. The reactor is placed in
an electric furnace, the temperature of which is controlled using
a second K-type thermocouple placed between the reactor and
the walls of the furnace. A pressure indicator is used to mea-
sure the pressure drop in the catalyst bed. The analysis system
consists on a gas chromatograph (Shimadzu) equipped with two
packed columns (Porapak-Q Carbosieve) and two detectors (TCD,
FID) and operates with He as the carrier gas. The response fac-
tors of the detectors were determined with the use of gas streams
of known composition (Scott specialty gas mixtures). Reaction
gases (He, 15% CO/He, CO
2
, H
2
) are supplied from high-pressure
gas cylinders (Messer Griesheim GMBH) and are of ultra-high
purity.
V. Jimnez et al. / Applied Catalysis A: General 390 (2010) 3544 39
50 40 30 20 10 0
Ru (101)
Ru (002)
Ru (100)
2
b
50 40 30 20 10 0
Ru (101)
Ru (002)
Ru (100)
2
c a
50 40 30 20 10 0
Ru (101)
Ru (002)
Ru (100)
I
n
t
e
n
s
i
t
y
(
a
.
u
.
)
2
Fig. 3. XRD pattern for Ru catalysts supported on different types of CNFs: (a) platelet CNFs, (b) shbone CNFs and (c) ribbon CNFs.
The catalytic performance of the preparedsamples for the selec-
tive methanation of CO has been investigated in the temperature
range of 200500
C
for 1h with the reaction mixture. Conversions of reactants and
selectivities toward products were then measured at that tem-
perature using the analysis system described above. Selectivity to
hydrogenation product i(S
i
) was calculated using the following
expression:
S
i
=
C
i,out
/v
i
i
C
i,out
/v
i
(6)
where C
i,out
is the outlet concentration of the product i and v
i
is the
number of carbon atoms of product i. It should be noted that the
system was left at each temperature for about 1h under isother-
mal conditions in order to achieve steady state conditions. In all
cases, data points are overages of at least three measurements. All
experiments were performed at near atmospheric pressure.
Results obtained, along withmeasurements of metal dispersion,
were used to determine the turnover frequencies (TOFs) of carbon
monoxide and carbon dioxide, dened as moles of CO or CO
2
con-
verted per surface Ru atom per second. Details of the methods and
procedures employed can be found elsewhere [25].
3. Results and discussion
3.1. Support characterization
IntherepresentativeTEMimages showedinFig. 1areclearlydif-
ferentiated the different structures of the CNFs used in this study as
catalyst supports. Carbon nanobers (CNFs) are an allotrope of car-
bon that has crystalline structure (Fig. 1a). Different types of CNFs
can be obtained depending on their synthesis temperature. Platelet
type structures (hexagonal planes perpendicular to the ber axis)
were produced at a synthesis temperature of 450
C).
The graphitic character of the different supports was eval-
uated by XRD, where the npg values (Table 1) suggested that
ribbon CNFs were the most graphic carbon nanobers. Thus, L
c
val-
ues increased in the following order: platelet <shbone <ribbon,
and d
002
values in the following one: ribbon<shbone <platelet.
Higher values of L
c
may reduce the adsorptive sites on CNFs sur-
face which will be veried by N
2
adsortion [46]. Carbon structural
order was also measured by means of TPOanalyses. All CNFs exhib-
ited only a single oxidation peak from the DTA curve (not shown),
indicating high product purity [40]. Table 1 presents the maxi-
mum weight loss temperature showing that an increase in the
CNFs synthesis temperature lead to more crystalline structures
(ribbon>shbone >platelet).
Nitrogen adsorption/desorption isotherms and pore size dis-
tribution are presented in Fig. 2. The isotherms can be clearly
assigned to the type IV IUPAC classication, characteristic of meso-
porous materials, where the mean size (see Table 1) falls within the
accepted mesoporous (250nm) range. Also, a well dened hys-
teresis (H3 type according to IUPAC classication) was observed
indicating the existence of a substantial volume of mesopores
where irreversible capillary condensation occurred [47]. Textural
parameters dataareshowninTable1. It is observedthat for increas-
ing CNFs synthesis temperature, the specic BET surface area and
the total pore volume of the resulting carbon material decreased,
in agreement with the results obtained by others authors [48].
Their corresponding pore size distribution curves are also showed
in Fig. 2. Average mesopore size increased in the following order:
platelet <shbone <ribbon, being the principal mesopore size in
the range 1530nm, accompanied with a secondary mesopore
size of 35nm. Larger mesopores have been attributed to inter-
stices between interlaces laments whereas smaller ones have
been associated with the CNFs surface roughness. Obtained results
were indicative of the higher amount of adsorption sites present in
platelet CNFs [46]. Thus, platelet CNFs had the largest BET surface
area and the smallest average pore diameter because of the pres-
ence of a higher amount of 35nm pores. Nevertheless, shbone
and ribbon CNFs had larger average pore diameters than platelet
ones because of much of their pores resulting mainly from aggre-
gation.
3.2. Catalyst characterization
The physicochemical characteristics of the synthesized Ru cat-
alysts are summarized in Table 1.
XRD patterns for the activated catalysts are presented in Fig. 3.
Peaks at 2 26
, 42.3
and 44.1
correspond,
respectively, to the (100), (002) and (101) planes of metallic
ruthenium (JCPDS-ICDD Card No. 06-0663), which are consistent
with an exclusive hexagonal geometry and establish the presence
of Ru
0
[49,50]. These results indicate that after activation of cat-
alysts with H
2
ow previous to the methanation reaction all the
metallic phase was in the Ru
0
form.
BET surface area and micropore/mesopore volume values of
Ru platelet and shbone CNFs (Table 1) decreased as compared
with the parent CNFs (before metal incorporation). This effect is
shown in Fig. 2. However, ribbon CNFs surface area and micro-
pore volume signicantly increase after metal incorporation. To
explain these results is necessary to have into account the posi-
tion of the Ru particles on the different supports. Thus, the surface
area decrease that took place with the platelet and shbone type
CNFs has been ascribed to the occupation of the adsorptive sites
present on the CNF surface by highly dispersed Ru nanoparticles
(note that the highest Ru dispersion values (Table 1) were obtained
with the platelet and shbone type CNFs). A high dispersion of
Ru nanoparticles on the CNFs is normally associated to a strong
metal-support interaction between the Ru nanoparticles and the
graphitic edges of the CNFs, as has been previously observed for a
CNF-supported Pd catalyst [51]. On the other hand, the micropore
surface area increase that was observed in ribbon CNFs could be
ascribed to the presence of additional oxygen containing groups
and defects created on the CNF surface during the Ru deposi-
tion treatment. These groups would lead to new adsorption sites
during N
2
physisorption [52]. The presence of oxygen-containing
groups, which could signicantly inuence the adsorptive and sub-
sequent catalytic behaviour [53], can be conrmed by means of
standard acidbase titrations. In such titrations the acid groups
would be neutralized by NaOH. The quantities of acidic sites, which
are presented as millimoles of NaOHneeded to neutralize the acid-
ity groups of pKi 7, are shown in Table 1. Results showed that
the total acidic groups determined by titrations decreased in the
sequence: platelet >shbone >ribbon. It is interesting to note that
the presence of oxygen-containing groups, after incorporation of
ruthenium, was the same in platelet and shbone CNFs, but it was
almost two times higher on the surface of ribbon CNFs.
Finally, the average diameter of ruthenium particles measured
by H
2
-chemisorption and TEM are shown in Table 1. It can be
observed that the results obtained by both techniques were quite
similar. Ru particle size varied with the type of CNFs support. Par-
ticle size in platelet CNFs was the lowest if compared with that
of shbone and ribbon CNFs. As expected, the larger the Ru par-
ticles were, the lower metal dispersion was observed. In Fig. 4 is
shown the representative TEM images of different types of CNFs
with Ru.
3.3. CO and CO
2
hydrogenation
The effect of the type of CNFs (platelet, shbone and ribbon)
used as support on catalytic performance for the solo-CO and CO
2
hydrogenationwas investigatedover Rucatalysts of thesamemetal
content (0.5wt.%). Results obtainedaresummarizedinFig. 5, where
conversions of CO (X
CO
) and CO
2
(X
CO
2
) are plotted as functions of
reaction temperature. It is observed that conversion of CO (Fig. 5a)
increases withincreasing temperature above ca. 240
Candreaches
100% at 340
C
and conversion reaches 50% at 500
C to 400
C are
C
2
H
6
and C
2
H
4
. At temperatures above 400
C, where conversion
of COis complete (Fig. 5a), formationof higher hydrocarbons is sup-
pressed and the only hydrogenation product formed is CH
4
(Fig. 6a
and b). Ruthenium catalyst supported on ribbon CNF was found
to be more selective toward CH
4
(Fig. 6c) than Ru supported on
platelet and shbone CNFs, with S
CH
4
being typically higher than
96% in the whole temperature range, while the only by-product
formed is C
2
H
4
. The above ndings are in good agreement with
results reported previously over noble metal catalysts supported
on Al
2
O
3
[25].
Qualitatively, similar results were obtained for the CO
2
hydro-
genation reaction. As it can be seen in Fig. 6 (open symbols)
selectivity toward methane takes more or less the same values
(for each sample investigated) with that obtained for CO hydro-
genation, with Ru/ribbon catalyst exhibiting the highest S
CH
4
. It
has been proposed over noble metal catalysts supported on metal
oxides [17,25,26,5663] that under conditions of CO
2
hydrogena-
tionappreciable amounts of CO
2
are convertedtoCOvia the reverse
WGS reaction. It is believed that carbon monoxide acts as an inter-
mediate, which is further hydrogenated to methane [59]. However,
this is not the case for the results of the present study since no
CO was detected under CO
2
hydrogenation conditions. In pre-
vious investigations [17,25,64] it was found that the occurrence
and the onset of the RWGS reaction under conditions of solo-CO
2
methanation and methanation of CO/CO
2
mixtures depends sig-
nicantly on the nature of metal and the support. For example, it
was found that, for solo-CO
2
methanation reaction, selectivity to
CO, at the expense of CH
4
, at a given temperature decreases in the
order of Pt >PdRu>Rh, with Rh catalyst producing CO only at
temperatures higher than 400450
C. This
behaviour can be explained by considering that, under the present
conditions, CO hydrogenation (Eq. (1)) runs in parallel with the
RWGSreaction(Eq. (3)), whichbecomes important aboveca. 275
C.
At temperatures higher than 275
C, i.e., until
X
CO
reaches its maximum value, and then progressively increases
withincreasingtemperature(Fig. 7). This behaviour reects thefact
that COinteracts more strongly withthe catalyst surface, compared
to CO
2
, and is in good agreement with previous studies [17,25].
The Ru catalysts supported on platelet and shbone CNFs are
practically inactive for the selective methanation of CO under the
present experimental conditions, reaching a maximumof X
CO
of 7%
at 250
C
to 275
C) to 53%
Table 2
Representative results of kinetic measurements obtained over the examined catalysts for the hydrogenation of CO and CO
2
mixture, in the absence and in the presence of
30% H
2
O in the feed. Mass of catalyst: 150mg; particle diameter: 0.18<dp <0.25mm; feed composition: 1% CO, 15% CO
2
, 50% H
2
, 0% or 30% H
2
O (balance He); total owrate:
200cm
3
/min.
Catalyst (0.5% Ru) Reaction
CO/CO
2
/H
2
CO/CO
2
/H
2
/H
2
O
S
CH
4
at 250
C (%) TOF
CO
at 250
C (s
1
) TOF
CO
2
at 330
C (s
1
) S
CH
4
at 250
C (%) TOF
CO
at 250
C (s
1
) TOF
CO
2
at 330
C (s
1
)
Platelet 93.2 0.017 0.56 100 0.058 0.51
Fishbone 93.1 0.023 0.39 100 0.043 0.38
Ribbon 91.8 0.080 0.26 100 0.070 0.50
V. Jimnez et al. / Applied Catalysis A: General 390 (2010) 3544 43
400 350 300 250 200
-200
-100
0
20
40
60
80
100
a
0.5% Ru/platelet
X
C
O
,
X
C
O
2
(
%
)
Temperature (C)
H
2
O Content (%)
0 30
, X
CO
, X
CO
2
400 350 300 250 200
-200
-100
0
20
40
60
80
100
b
0.5% Ru/fishbone
X
C
O
,
X
C
O
2
(
%
)
Temperature (C)
H
2
O Content (%)
0 30
, X
CO
, X
CO
2
400 350 300 250 200
-200
-100
0
20
40
60
80
100
c
0.5% Ru/ribbon
X
C
O
,
X
C
O
2
(
%
)
Temperature (C)
H
2
O Content (%)
0 30
, X
CO
, X
CO
2
Fig. 8. Effect of addition of water vapour in the feed (30%) on the catalytic per-
formance of Ru (0.5wt.%) catalysts supported on (a) platelet, (b) shbone and
(c) ribbon for co-methanation of CO/CO
2
mixture. Solid symbols: CO conver-
sion; open symbols: CO
2
conversion. Mass of catalyst: 150mg; particle diameter:
0.18<dp <0.25mm; feed composition: 1% CO, 15% CO
2
, 50% H
2
030% H
2
O (balance
He).
(at 300
C
for Ru/ribbon catalyst increases from 91.8 to 100% with increas-
ing the concentration of water in the feed from 0 to 30% (Table 2).
It has been reported that S
CH
4
always increases with increasing
X
CO
in a manner which does not depend on whether this is due
to an increase of reaction temperature [17,25], to the effect of
metal crystallite size [17], to the effect of the oxide support [17]
or to variations of space velocity [17,25]. Furthermore, Inui et al.
[67] demonstrated that the formation of higher hydrocarbons has
a strong retarding effect on CO methanation, which is released
with increase of CO conversion. It may then be suggested that
the observed increase of S
CH
4
with the addition of water vapour
(Table 2) is related to the increase of X
CO
(Fig. 8).
Results of kinetic measurements (Table 2) showed that, for
all catalysts examined, TOF of CO
2
conversion remains practically
unaffected by the presence of water, whereas TOF of COconversion
increases with increasing water content from 0 to 30%. In partic-
ular, TOF of CO over Ru/shbone catalyst increases by a factor of
2 (at 250
C (not
shownfor clarity). Theeffect is morepronouncedfor theRu/platelet
catalyst, for which the specic activity for the CO hydrogenation
reaction becomes more than three three times higher (at 250
C)
with the addition of water (Table 2).
Results of thepresent studyshowthat the0.5%Ru/CNFs catalysts
investigated exhibit similar catalytic activity with that obtained in
previous studies over the well known Ru/Al
2
O
3
catalyst conducted
under the same experimental conditions [25]. However, it has been
reported that catalytic performance for the title reaction could be
further improved provided that catalyst characteristics are opti-
mized and operating conditions are properly selected [17,25]. For
example, it has been found that catalytic activity of Ru/TiO
2
and
Ru/Al
2
O
3
catalysts is greatly enhanced with increase of Ru load-
ing or crystallite size [17,33]. Furthermore, results presented in
Figs. 58 were obtained with a relatively high space velocity (SV)
of ca. 48,800h
1
, compared to that usually used (5000h
1
) in fuel
cell applications. It was found over 0.5% Ru/Al
2
O
3
catalysts that
decreasing SV from 48,800h
1
to 12,200h
1
resulted in a shift of
the X
CO
curves toward lower temperatures by ca. 80
C, whereas
the X
CO
2
curve was not affected signicantly [17]. Thus, catalytic
performance of the investigated 0.5% Ru/CNFs catalysts could be
further improved by decreasing SV and/or increasing Ru loading
(i.e. crystallite size), leading to completely methanation of CO and
making them suitable for use in the selective methanation of CO
for fuel cell applications.
44 V. Jimnez et al. / Applied Catalysis A: General 390 (2010) 3544
4. Conclusions
Results of the present study show that catalytic performance
and selectivity to reaction products over Ru catalysts supported
on different types of CNFs depend strongly on the experimen-
tal conditions used (e.g. reaction temperature, solo-methanation
of CO or CO
2
, co-methanation of CO/CO
2
mixtures, addition of
water vapour etc.). Under solo-COmethanationconditions, COcon-
version reaches 100% at 340