Selective oxidation of CO in the presence of H2, H2O and CO2 via gold for use in fuel cells

Article abstract of DOI:10.1039/b505295p

An Au/Fe₂O₃ catalyst prepared using a two-stage calcination procedure achieves target conversion and selectivity for the competitive oxidation of dilute CO in the presence of moist excess H₂ and CO₂. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b505295p

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Selective oxidation of CO in the presence of H2, H2O and CO2 via gold
for use in fuel cells{
Philip Landon,a Jonathan Ferguson,a Benjamin E. Solsona,a Tomas Garcia,a Albert F. Carley,a
Andrew A. Herzing,b Christopher J. Kiely,b Stanislaw E. Golunskic and Graham J. Hutchings*a
Received (in Cambridge, UK) 14th April 2005, Accepted 22nd April 2005
First published as an Advance Article on the web 17th May 2005
DOI: 10.1039/b505295p
selective for CO oxidation in the presence of H2 at 25 uC, but the
selectivity was lost at higher temperatures. To date, some of the
best results for this demanding reaction have been reported by
Grigorova et al.12 for a Au/CeO2 catalyst promoted with SnO2,
but even this catalyst gives significant (ca. 5%) H2 conversion at
temperatures . 70 uC with CO conversions of ca. 40–60%.
Recently, Corma and co-workers13 reported a nanocrystalline
CeO2-supported gold catalyst that gave appreciable selectivity at
60 uC for a dry CO/H2 mixture. However, it is significant that these
previous studies use dry reactants and consequently are not appli-
cable to the humid conditions prevailing in the fuel cell application,
which in addition also contains significant quantities of CO2. The
presence of excess CO2 presents an additional problem since if the
catalyst is active for the reverse water gas shift reaction then CO
will be remade, and this is a key feature that has limited success in
this field to date. An interesting development by Watanabe et al.14
was the use of Pt–Fe supported on mordenite rather than alumina,
but these catalysts are not robust under reaction conditions and
generally suffer the same problems as the Pt–Fe/alumina catalysts;
consequently they have not been commercially exploited.
An Au/Fe2O3 catalyst prepared using a two-stage calcination
procedure achieves target conversion and selectivity for the
competitive oxidation of dilute CO in the presence of moist
excess H2 and CO2.
The recently established ability of gold to oxidise CO at relatively
low temperatures1 has a potential application in power-generation
systems that combine a polymer electrolyte fuel cell (PEFC) with a
catalytic reformer,2 e.g. for electric vehicles. The reformer converts
an organic fuel to a hydrogen-rich gas feed for the PEFC, which
electrochemically oxidises the hydrogen to generate electricity and
water. However, residual CO in the reformer output poisons the Pt
anode at the relatively low PEFC operating temperature (ca.
80 uC).3 Industry has adopted two approaches in an attempt to
deal with this problem, thereby ensuring that effective performance
of the PEFC can be achieved, namely, either to make the anode of
the fuel cell more CO tolerant, or to develop catalytic systems
capable of removing even trace amounts of CO from H2 in the
presence of excess CO2 and water. We have now successfully
addressed the latter of these two approaches. Although concep-
tually simple, the oxidation of CO in the presence of excess moist
H2 and CO2, without oxidising the hydrogen or remaking CO via
the reverse Water Gas Shift reaction, is a particularly difficult
objective, which has so far only been achieved using a multistage
reactor.4 There have been no reported examples of this process
being successfully carried out in a single catalyst bed under
industrially relevant conditions. We now present a heterogeneous
catalyst for the selective oxidation of CO in the presence of H2 and
CO2 at 80 uC that gives stable catalyst performance under realistic
fuel cell conditions.
We have studied Au/Fe2O3 catalysts prepared by coprecipita-
tion.15 Initially, the catalysts were investigated for CO oxidation in
the absence of H2 (Table 1). We subsequently investigated the
addition of H2 to the reactant gases (2 : 1 H2 : CO molar ratio), but
still using dry gases in the absence of excess CO2, and this led to a
decrease in CO conversion (Table 1), which is consistent with
earlier studies.9 Drying the 5 wt.% Au/Fe2O3 at 120 uC gave a
catalyst that was active for H2 oxidation at 27 uC (Table 1).
Catalysts calcined at 550 uC, whether for 3 h or 8 h, retain
considerable CO oxidation activity but show no oxidation of H2
even at 73 uC. This activity was consistently observed for tens of
hours of catalyst testing. Calcination of the material at lower
temperature, e.g. 400 uC, gave a catalyst with activity for H2
oxidation at 73 uC (Table 1). Interestingly, calcination at 400 uC
followed by calcination at 550 uC did not reproduce the
performance of a catalyst calcined only at 550 uC. This
demonstrates the extreme sensitivity of this catalyst system to the
calcination conditions, and this will be discussed subsequently.
Calcining at 600 uC completely destroys the catalyst activity
(Table 1). We investigated alternative supports and active metals,
but no combination was as effective as the Au/Fe2O3 catalysts.15
Au/TiO2 and Au/Al2O3 catalysts were inactive for both CO and
H2 oxidation whereas the Au/Pd/Al2O3 and Pd/Al2O3 catalysts
were only active for H2 oxidation (Supplementary Table 2). 1 wt.%
Au/CeO2 was selective up to 52 uC but at higher temperature
significant H2 conversion was observed, which is consistent with
the previous literature.12
Considerable effort has been applied to the design of suitable
catalysts for the competitive oxidation of CO in the presence of H2,
and the most commonly used partial oxidation catalysts (alumina-
supported Pt, Pt/Fe, Ru and Rh) are unsuitable since they require
temperatures in the range 150–200 uC to be effective,5–7 and even
then significant conversion of H2 is observed. Haruta and co-
workers8,9 first demonstrated that supported Au nanoparticles
could be active at much lower temperatures, but again H2 was
oxidised together with CO at temperatures . 70 uC. Au/Fe2O3
catalysts have also been studied and Kahlich et al. used catalysts
calcined at 400 uC but found that significant H2 oxidation
occurred at 80 uC.10 Most recently, Qiao and Deng11 showed that
non-calcined Au/Fe2O3 catalysts, dried only at 120 uC, were very
{ Electronic supplementary information (ESI) available: details of catalyst
preparation and characterisation, and supplementary experimental data.
See http://www.rsc.org/suppdata/cc/b5/b505295p/
*hutch@cardiff.ac.uk
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 3385–3387 | 3385

View Article Online
a
Table 1 Oxidation of CO by 5% Au/Fe2O3 in the presence of H2
CO conversion (%)
H2 conv. (%)
Calcination conditions
No H2b27 uC
27 uC
50 uC
73 uC
27 uC
50 uC
73 uC
dried 120 uC, 8 h
400 uC, 3 h
400 uC, 3 h, then 550 uC, 2 h
550 uC, 3 h
550 uC, 8 h
600 uC, 3 h
100
91
68
31
—
0
100
68
37
17
12
0
92
84
71
25
19
0
90
91
80
33
37
0
7
0
0
0
0
0
40
0
0
0
0
85
17
5
0
0
0
0
a
b
0.4 vol.% CO, 0.8 vol.% H2, 16 vol.% O2, 50 mg catalyst, GHSV 5 30 000 h21
.
0.5% CO in air, 50 mg catalyst; GHSV 5 24 000 h21
.
must be operated at ca. 80 uC in a complex feed comprising CO,
O2, H2, CO2, H2O and N2, and be capable of reducing the CO
concentration from about 1% to below 50 ppm – this is equivalent
to a CO conversion of at least 99.5%. In addition, this conversion
must be achieved without the addition of excess O2 and the
competitive oxidation of H2 must be minimised. This is expressed
as selectivity which is defined as the percentage of the oxygen fed
that is consumed in the oxidation of CO, and for commercial
operation a selectivity of 50% is acceptable, since at this selectivity
minimal H2 is oxidised to water. Taken together this represents a
demanding target and this is summarised in Table 2 and
Supplementary Table 1. Tests using our supported gold catalyst
show that we can readily achieve this very demanding target when
the Au/Fe2O3 catalyst is calcined in two stages (Fig. 2). The
catalysts are exceptionally stable and have been operated
successfully for extended periods. This is the first reported study
for any catalyst that has fully met this required operational target
in a single packed bed at realistically high space velocity. In Table 2
our results are contrasted with those of previous studies using
Au/Fe2O3, and in Supplementary Table 1 with the most important
results from the literature, confirming that we are the first to report
this achievement. In addition, we have prepared two of the
leading catalyst formulations, namely Au/Fe2O3 dried at 120 uC,
and Au/CeO2, and neither of these catalysts is capable of achieving
the required target (Supplementary Table 1).
Fig. 1 Influence of the presence of water (4 wt.%) on the CO (&) and H2
(#) conversion at different reaction temperatures for a 5% Au/Fe2O3
catalyst calcined in air at 550 uC. Reaction conditions: 0.4 vol.% CO,
0.8 vol.% H2, 16 vol.% O2, 50 mg catalyst, GHSV 5 30000 h21
.
On the basis of these initial results we were encouraged to
investigate the effect of adding water vapour under these reaction
conditions. Fig. 1 shows the influence of the presence of water
(4 wt.%) in the feed at different reaction temperatures. The
presence of water leads to a decrease in CO conversion that is
recovered by increasing the reaction temperature. It should be
noted that no H2 conversion was observed even at a reaction
temperature of 80 uC, showing the beneficial effect of water on the
selectivity towards CO oxidation in the presence of H2.
Analysis of the 5 wt.% Au/Fe2O3 catalysts using electron
microscopy (Supplementary Fig. 1) shows that calcination of the
catalyst increases the Au particle size: for the series (i) dried at
120 uC, (ii) calcined at 400 uC, (iii) calcined at 400 and 550 uC, (iv)
calcined at 550 uC and (v) calcined at 600 uC; the Au particle sizes
are 3.5, 3.7, 6.7, 8.0 and 16.1 nm respectively. For the same series
of catalysts the surface areas are (i) 203, (ii) 45, (iii) 26, (iv) 16 and
(v) 11 m2 g21 respectively. The dramatic loss of activity for
catalysts calcined at temperatures greater than 550 uC is thus due
to a combination of growth in the Au particle size and the loss of
Operation in those fuel cell systems in which the hydrogen is
generated by reforming hydrocarbons or oxygenates requires
additional factors to be considered, since not only water vapour
but CO2 is present in excess.2 For successful operation as a
selective CO oxidation catalyst in a reformer–PEFC system, it
Table 2 Comparison of Au/Fe2O3 catalysts for the selective oxidation of CO in the presence of H2, H2O and CO2
Reactant mixture composition
Tcalc uC
CO
O2
H2
H2O
CO2
Conv%
Sel%
TuC
Target
0.8–1.0
0.4–1.0
0.9
0.1
4
1.25
1.25
50
50
66
50
50
50
2–10
20–24
22
22
. 99.5
99.8
. 50
51
29
100
60
80–110c
80
80
a
0.9
0.1
1
1
1
4.7
10
0
10
10
400 uC20
120 uC11
400 uC19
400 uC19
a
b
0
15
15
100
65
99.5
50
80
100
54.5
b
c
This work 400 uC + 550 uC. Activity quoted as 0.003 mol CO (g Au?s). Most PEFCs currently run at 80 uC. Figures in bold italics show
where target requirements are not achieved.
3386 | Chem. Commun., 2005, 3385–3387
This journal is ß The Royal Society of Chemistry 2005

View Article Online
markedly at higher temperatures due to the reverse WGS reaction
catalysed by the residual cationic gold present in this sample.
We conclude that the most selective catalysts for CO oxidation
in the presence of H2 comprise relatively large Au nano-crystals
(. 5 nm) supported on a reducible oxide. These initial results
demonstrate that the calcination temperature used to prepare
supported Au nano-crystals is of crucial importance. Most
importantly it is essential that residual cationic gold is reduced
during the pre-treatment since this, if present, will catalyse the
reverse WGS reaction thereby decreasing CO conversion, and it is
for this reason that previous Au formulations have failed to
achieve this challenging target. Hence, by careful manipulation of
the preparation conditions we have prepared a stable gold catalyst
that can selectively oxidise CO in the presence of H2, H2O and
CO2 and successfully operate in the demanding environment of a
fuel cell system, giving stable sustained performance. We have not
fully optimised the formulation of this catalyst and so we
anticipate further improvements in this exciting new development.
We acknowledge Johnson Matthey plc and the Engineering &
Physical Sciences Research Council (ATHENA project), and the
EU (AURICAT; Contract HPRN-CT-2002-00174) for financial
support.
Fig. 2 Variation of CO (&) and O2 (#) conversion with time on line for
a 5% Au/Fe2O3 catalyst calcined in air at 400 + 550 uC. Reaction
conditions: 80 uC, 0.9% CO, 0.9% O2, 50% H2, 22% CO2, 4.7% H2O with
the balance N2; 100 mg of catalyst were used with a total flow rate of
20 ml/min, GHSV 5 12000 h21
.
surface area. Au(4f) X-ray photoelectron spectra obtained for the
Au/Fe2O3 samples calcined at different temperatures
(Supplementary Fig. 2) show that for samples calcined at
temperatures of 400 uC and lower, the signal comprises a marked
doublet which we ascribe to the presence of cationic gold in
addition to metallic gold. It is possible that this shift in binding
energy is associated with a relatively sharp bimodal particle size
distribution but the microscopy we have carried out does not
reveal this to be the case (Supplementary Fig. 1). The spectrum for
the sample calcined at 400 uC and subsequently 550 uC shows that
all the cationic gold has now been reduced to metallic gold. The
Au(4f) spectrum observed after calcination at 600 uC shows a
significant increase in intensity, and a clear decrease in line-width,
consistent with the diffusion of Au from the pores of the catalyst,
and agglomeration to form metallic gold particles, which are
inactive for both CO and H2 oxidation.
Philip Landon,a Jonathan Ferguson,a Benjamin E. Solsona,a
Tomas Garcia,a Albert F. Carley,a Andrew A. Herzing,b
Christopher J. Kiely,b Stanislaw E. Golunskic and
Graham J. Hutchings*a
aSchool of Chemistry, Cardiff University, Main Building, Park Place,
Cardiff, UK CF10 3AT. E-mail: hutch@cardiff.ac.uk
bCenter for Advanced Materials and Nanotechnology, Lehigh
University, 5 East Packer Avenue, Bethlehem, PA 18015-3195, USA
cJohnson Matthey Technology Centre, Blount’s Court, Sonning
Common, Reading, UK RG4 9NH
Notes and references
1 G. C. Bond and D. T. Thompson, Catal. Rev. Sci. Eng., 1999, 41, 319.
2 D. L. Trimm and Z. I. Onsan, Catal. Rev. Sci. Eng., 2001, 43, 31.
3 T. R. Ralph and M. P. Hogarth, Platinum Met. Rev., 2002, 46, 117.
4 N. Edwards, S. R. Ellis, J. C. Frost, S. E. Golunski, A. N. J. van
Keulen, N. G. Lindewald and J. G. Reinkingh, J. Power Sources, 1998,
71, 123.
5 M. J. Kahlich, H. A. Gasteiger and R. J. Behm, J. Catal., 1997, 171, 93.
6 S. Ozkara and A. E. Aksoylu, Appl. Catal., A, 2003, 251, 75.
7 P. V. Snytnikov, V. A. Sobyanin, V. D. Belyaev, P. G. Tsyrulnikov,
N. B. Shitova and D. A. Shlyapin, Appl. Catal., A, 2003, 239, 149.
8 R. M. Torres Sanchez, A. Ueda, K. Tanaka and M. Haruta, J. Catal.,
1997, 168, 125.
9 M. Okumura, S. Nakamura, S. Tsubota, T. Nakamura, M. Azuma and
M. Haruta, Catal. Lett., 1998, 51, 53.
10 M. J. Kahlich, H. A. Gasteiger and R. J. Behm, J. Catal., 1999, 182,
430.
11 B. T. Qiao and Y. Q. Deng, Chem. Commun., 2003, 2192.
12 B. Grigorova, J. Mellor, A. Palazov and F. Greyling, WO Pat.,
00/59631/2000.
It is important to determine the key factors controlling the
enhanced activity and selectivity we have observed with the Au/
Fe2O3 catalyst since previous studies have used similar catalysts
and have come close to achieving the target, yet failed. The answer
lies in the need to control the activity of the catalyst for the reverse
water gas shift reaction whilst retaining CO oxidation activity.
Two recent studies on model catalysts16,17 are important in this
respect since it has been shown that metallic gold, in particular
morphologies, is very active for CO oxidation16 whereas cationic
gold is particularly effective for the water gas shift (WGS)
reaction,18 and consequently will also catalyse the reverse reaction:
CO2 + H2 A CO + H2O
This reaction leads to the formation of CO thereby ensuring the
target CO conversion cannot be achieved. Hence the design
criterion is to prepare a catalyst comprising small gold nanocrys-
tals whilst ensuring that all cationic gold is removed. In our case
we have achieved this by a two stage calcination procedure, but it
is possible that other preparation strategies are equally feasible. A
scrutiny of the results of previous studies presented in
Supplementary Table 1 shows that Avgouropoulos et al.19 came
close to achieving the target at 100 uC , with a Au/Fe2O3 catalyst
calcined at 400 uC, but required excess O2 and low CO2.
Furthermore, in their studies the CO conversion decreased
13 S. Carrettin, P. Concepcion, A. Corma, J. M. L. Nieto and V. F. Puntes,
Angew. Chem., Int. Ed., 2004, 43, 2538.
14 M. Watanabe, H. Uchida, K. Okhubo and H. Igarashi, Appl. Catal., B,
2003, 46, 595.
15 See Electronic Supplementary Information.
16 M. S. Chen and D. W. Goodman, Science, 2004, 306, 252.
17 Q. Fu, H. Saltsburg and M. Flytzani-Stephanopoulos, Science, 2003,
301, 935.
18 Y. Liu, Q. Fu and M. F. Stephanopoulos, Catal. Today, 2004, 93–95, 241.
19 G. Avgouropoulos, T. Ioannides, Ch. Papadopoulou, J. Batista,
S. Hocevar and H. K. Matralis, Catal. Today, 2002, 75, 157.
20 M. M. Schubert, A. Venugopal, M. J. Kahlich, V. Plzak and R. J. Behm,
J. Catal., 2004, 222, 32.
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 3385–3387 | 3387