Full text of DOI:10.1016/S0378-7753(01)00957-0

Journal of Power Sources 105 (2002) 52–57
Composite anode for CO tolerance proton exchange membrane fuel cells
Hongmei Yu*, Zhongjun Hou, Baolian Yi, Zhiyin Lin
Fuel Cell R&D Center, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan Road, Dalian 116023, PR China
Received 5 April 2001; received in revised form 31 July 2001; accepted 25 September 2001
Abstract
Fuel of proton exchange membrane fuel cells (PEMFC) mostly comes from reformate containing CO, which will poison the fuel cell
electrocatalyst. The effect of CO on the performance of PEMFC is studied in this paper. Several electrode structures are investigated for CO
containing fuel. The experimental results show that thin-film catalyst electrode has higher specific catalyst activity and traditional electrode
structure can stand for CO poisoning to some extent. A composite electrode structure is proposed for improving CO tolerance of PEMFCs.
With the same catalyst loading, the new composite electrode has improved cell performance than traditional electrode with PtRu/C
electrocatalyst for both pure hydrogen and CO/H2. The EDX test of composite anode is also performed in this paper, the effective catalyst
distribution is found in the composite anode. # 2002 Elsevier Science B.V. All rights reserved.
Keywords: CO tolerance; Composite electrode; PEMFC; Electrode structure
1. Introduction
depends strongly on the CO concentration and cell tempera-
ture will make a notable impact on fuel cell performance. At
higher temperature, less CO adsorption will happen, less cell
performance drop will occur [3].
Proton exchange membrane fuel cells (PEMFCs) have
been considered as high efficiency, low pollution power
generators for stationary and transportation applications.
Much attention has been devoted to the implementation
of the fuel cell for transportation applications in current
researches, which develop the economically viable low-
temperature fuel cells.
PEMFCs with pure hydrogen as fuel can deliver high
power densities to automobiles. It is known that the pre-
paration and storage of pure hydrogen as transportation fuel
is too difficult for PEMFC to play the role of automobile
power, as the fuel chosen for the electric vehicle needs to be
readily available. Though reforming technologies are well
established, when reformate is used, CO in fuel gas leads to a
drastically decreased power density. Although many
attempts have been made in the last few decades to develop
non-noble metal electrocatalyst for replacing the platinum
catalyst in PEMFC and other low temperature fuel cells,
platinum remains as the perfect [1]. Tests in PEMFCs
indicate that more than about 10 ppm of CO in the gas
stream will decrease cell performance especially in the range
between 60–100 8C [2]. The performance of PEMFC will
continue to decrease and remain unsteady variation when
CO is fed in. The situation in which performance of PEMFC
There are several ways to overcome CO poison problem
in PEMFC. The first way is to bleed very low levels of
oxidant into the fuel feed stream, for instance, O2 [4,5] and
H2O2 [6,7]. The second way is advanced purification of
reformate gas by fuel processor [8]. The third way is to
develop new CO tolerant electrocatalyst. Regarding the first
way, when oxidant is added to the fuel stream, the utilization
of the fuel will certainly be decreased, and the safety
problem should also be considered and the second method
will make the fuel cell system much more complex and
expensive. Therefore, most of the researchers believe that
new CO tolerant electrocatalyst will be the most hopeful
way to solve the CO poisoning problem in PEMFC.
Nowadays, development of CO tolerance electrocatalyst
has been concentrated on PtM (M is usually a transition
metal) bimetallic catalysts. For example, PtRu binary cat-
alyst shows evidently CO tolerance. The oxidation potential
of CO on PtRu is reduced due to bifunctional mechanism.
The performance of PEMFCs will be improved when fuel
stream contains CO [9–11], but Pt-Ru alloy is not as active as
Pt when pure hydrogen is served as the fuel [1].
Based on previous experience, the structure of electrode
plays an importance role in improving cell performance.
However, there is less effort on CO tolerance electrode
structure improving. As we know that the electrode structure
* Corresponding author. Tel.: þ86-411-4671991x615.
E-mail address: hmyu@dicp.ac.cn (H. Yu).
0378-7753/02/$ – see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 8 - 7 7 5 3 ( 0 1 ) 0 0 9 5 7 - 0

H. Yu et al. / Journal of Power Sources 105 (2002) 52–57
53
relates to diffusion process and reaction dynamics, we have
an opportunity to decrease the poisoning of PEMFC anode.
Compared with new electrocatalyst development, the
improvement of electrode structure will be much more
attractive for PEMFC popularization.
When the electrode structure is modified, the loading of
the precious metal is reduced and the cell performance will
not drop. The electrode structure can be designed to make
the poisonous CO to react at a separate layer with CO active
electrocatalyst in advance, and make the main hydrogen to
react at another layer with traditional electrocatalyst plati-
num. Then there is possibility for us to have a good result on
the electrode structure for CO tolerance study.
Fig. 1. Schematic structure of a composite anode: (1) gas diffusion layer;
(2) outer catalyst layer; (3) inner catalyst layer; (4) Nafion membrane.
mounted in single cell with stainless steel end plates, and
mesh flow field is used. The active area of MEA is 5 cm2.
In PEMFC performance test, both fuel and oxygen are
2. Experimental
humidified at 80 8C. The operating pressures PO and Pfuel
2
are all 0.2 MPa. Cell Temperature is 80 8C. The stoichio-
metric coefficient for hydrogen and oxygen in reaction (1)
are 1.25 and 2, respectively. The polarization data is
obtained when the vibration of cell voltage is not more than
1 mV in 1 min. Each experiment operates without any
oxygen injection:
2.1. Composite electrode structure
In H2/CO fuel stream, as diffusion coefficients of H2 and
CO are different, it is possible to design a special composite
electrode structure according to the fuel component. The
anodes can be designed with different electrocatalyst com-
ponents, different contents and different pore distribution. In
this way, the designation of structure can help to solve the
CO poisoning problem. Several electrode structures are
tested in this paper. The electrode structures and electro-
catalyst components are listed in Table 1. Two kinds of
electrocatalyst 20% Pt/C and 30% PtRu/C are from Johnson-
Matthey Inc. The gas diffusion layers are made of carbon
paper (SGL TECHNIK, PE704). The proton exchange
membrane is Nafion 112 (DuPont Company) and 5% Nafion
solution is purchased from DuPont Company.
O2 þ 4Hþ þ 4eÀ ! 2H2O
(1)
2.3. EDX analysis
The electrodes are cut to expose their cross-sections. Then
the cross-sections are tested by Oxford Instruments X-ray
Microanalysis 1350 to have EDX results. The element
distribution status is shown in EDX analysis.
3. Results and discussions
In Table 1, the traditional method is a one proposed by
Lindstrom [12] and the main difference of these anodes is
the inner catalyst layer. The transfer method is described in
[13], in which the thin layer is hydrophilic as there is no
PTFE in it. The E4 and E5 are prepared by the method in
[14]. The composite anode structure is shown in Fig. 1.
3.1. Effect of different catalyst layers on CO poisoning
problem
Many researches disclose that the best electrocatalyst in
PEMFC is platinum, but when there is more than 10 ppm
CO in the fuel, the platinum will be poisoned and the fuel
cell performance will decline dramatically. Electrocatalyst
in the anode side of E1 and E3 is platinum. When pure
hydrogen is served as fuel, E1 and E3 have the similar
performance although the total platinum amount in E3 is
only 1/15 of that of E1. Since when the mass specific activity
2.2. Single cell test
Prepared Nafion 112 membrane, anode and cathode
(Pt/C from Johnson-Matthey, Pt loading: 0.5 mg/cm2) are
hot pressed at 10 MPa, 140 8C for 1 min to obtain mem-
brane electrode assembly (MEA). Then the above MEA is
Table 1
Several electrode structures and electrocatalyst loading
Anode
Inner catalyst layer
Outer catalyst layer
E1
E2
E3
E4
E5
–
Traditional method; Pt/C, 0.3 mg/cm2 Pt
Traditional method; PtRu/C, 0.3 mg/cm2 PtRu
–
Traditional method; PtRu/C, 0.2 mg/cm2 PtRu
Traditional method; PtRu/C, 0.28 mg/cm2 PtRu
–
Transfer method; Pt/C, 0.02 mg/cm2 Pt
Traditional method; Pt/C, 0.1 mg/cm2 Pt
Transfer method; Pt/C, 0.02 mg/cm2 Pt

54
H. Yu et al. / Journal of Power Sources 105 (2002) 52–57
Fig. 2. Cell performance of E1 and E3 when H2 and 50 ppm CO/H2 serve as fuels.
is considered, the utilization of electrocatalyst E3 is higher
than that of E1:
much more higher than that of E1 although total electro-
catalyst loading of E3 is less. It indicates that the transfer
method could make high catalyst utilization in PEMFC.
From Figs. 2 and 3, it can be seen that each electrode
structure has its own characteristic on pure hydrogen fuel
and CO containing fuel. The thinner catalyst layer made by
transfer method has high catalyst utilization but low CO
tolerance, the traditional method is better when CO contain-
ing fuel is fed in. It is hopeful to combine the two electrode
structures to improve the CO tolerance of PEMFC.
i
is ¼
(2)
mc
where is is the specific current density (mA/(cm2 mg)), i the
current density (mA/cm2) and mc the electrocatalyst loading
(mg). SEM experiment shows that the thickness of the
catalyst layer in E1 is about 40 mm, and the catalyst layer
of E3 is less than 5 mm [13]. This means that the catalyst
layer made by transfer method has less gas transfer distance
than that of the traditional method with the same Nafion
solution content. Then we can obtain high cell performance
with low catalyst loading. But when 50 ppm CO/H2 is served
as fuel, cell performance of E3 drops much more than that of
E1 (shown in Fig. 2). Because the diffusion distance and the
total electrocatalyst loading in E3 are all less than that of E1,
CO will diffuse easily to the surface of the electrocatalyst, it
makes the performance of PEMFC to decrease significantly.
So when CO containing fuel is fed into a single cell, E3 will
be poisoned more easily than E1 as a result of thinner
diffusion distance and less electrocatalyst loading in E3.
In Fig. 3, it is shown that the mass specific activity of E3 is
3.2. Composite anode study
At present, Pt–Ru alloy is the most active electrocatalyst
for solving CO poisoning problem, but Pt–Ru alloy is not as
active as Pt for pure hydrogen, e.g. it makes lower power
density than platinum when pure hydrogen is served as fuel.
In Fig. 4, it can be seen that E2 is less active than E1 when
pure H2 is served as fuel. When 50 ppm CO/H2 is used as
fuel, the performance of E2 is better than that of E1. The
bifunction mechanism of PtRu enhances the CO tolerance of
PEMFC.
To develop a composite anode, it is important to utilize
PtRu to electrocatalyze the oxidation of CO and use Pt to
catalyze the hydrogen oxidation reaction (HOR). As hydro-
gen diffuses faster than CO in gas diffusion layer, the inner
catalyst layer should have higher platinum loading, and the
outer catalyst layer should be rich in PtRu. Two kinds of
composite anode are made: E4 and E5. Fig. 4 shows the
PEMFC performance of E4 for pure hydrogen and 50 ppm
CO/H2.
In Fig. 4, it is seen that with the same noble electrocatalyst
loading, the composite anode E4 has almost the same
performance as that of E1 and is better than E2 when the
fuel is pure hydrogen. But when 50 ppm CO/H2 is served as
fuel, the E4 cell performance also drops seriously. It is only a
little better than that of E1. Fig. 5 shows the EDX result of
the cross-section of E4. Since the electrocatalyst is sprayed
Fig. 3. Mass specific activity comparison of E1 and E3.

H. Yu et al. / Journal of Power Sources 105 (2002) 52–57
55
Fig. 4. Cell performance of E1, E2 and E4.
or brushed onto a gas diffusion layer directly in the tradi-
tional method, some of the electrocatalyst immerse into the
gas diffusion layer. In Fig. 5, it can be seen that there is no
distinct interface between the inner catalyst layer and the
outer catalyst layer, and in the gas diffusion layer some
electrocatalyst exits. This reduces the utilization of electro-
catalyst especially the inner catalyst layer seriously. As the
inner catalyst layer is combined with the Nafion membrane
with Nafion solution, Nafion makes effective three-dimen-
sional reaction zone [15]. It is possible that the inner catalyst
layer with Nafion solution has high electrocatalyst utiliza-
tion than outer layer. From the EDX test on element dis-
tribution, it is disclosed that the traditional electrode
preparation method cannot provide an effective electroca-
talyst distribution in a composite anode and at the same time,
the mass specific activity of the traditional method is not as
high as that of the transfer method.
Figs. 6 and 7 show the fuel cell performance with E5 and
E2 for pure H2 and CO containing fuel, respectively. It can
Fig. 5. EDX result of composite anode E4.
Fig. 6. Comparison between E5 and E2 in pure hydrogen.

56
H. Yu et al. / Journal of Power Sources 105 (2002) 52–57
Fig. 7. Comparison between E5 and E2 in 50 ppm CO/H2.
be seen that with pure hydrogen and 50 ppm CO/H2, cell
performance of E5 is better than that of E2. In the case,
where CO/H2 is served as fuel, E5 is much better than E2
with about 100 mV improvement on cell performance when
the operating current density is over 500 mA/cm2.
Fig. 8 shows the EDX results of E5. Although there is
electrocatalyst immersion into gas diffusion layer, the cat-
alyst distribution in Fig. 8 differs from that in Fig. 5. We can
see that most of PtRu is distributed near to the gas diffusion
layer, while platinum concentrates in the inner catalyst
layer with its thickness less than 5 mm and adjacent to
Nafion membrane. Such a catalyst distribution can electro-
catalyze the reaction for both H2 and CO when CO/H2 fuel
is fed into PEMFC. As Nafion solution could make the
three-dimensional reaction zone in electrode. Because the
very thin inner catalyst layer containing Nafion solution,
then electrochemical reaction will take place in outer cat-
alyst layer with Nafion conducting protons, while the thin
inner catalyst layer maintains high platinum utilization for
hydrogen electro-oxidization. From structure point of view,
it is important to maintain the distinct interface between
inner and outer catalyst layer.
This means that the composite anode ensure both high
activity of hydrogen and decrease of CO poison to the
platinum. It indicated that the new composite anode E5
improves the diffusion status of hydrogen to enhance CO
tolerance of PEMFC with its performance improved when
pure hydrogen is served as fuel.
The improvement of PEMFC performance mainly comes
from the successful anode structure reformation. In the thick
outer catalyst layer, PtRu/C serves as a CO barrier, and
electrocatalyst Pt/C maintains high activity to HOR
although the platinum loading is low in the inner catalyst
layer.
4. Conclusions
The preparation method of anode catalyst layer is very
important to the performance of PEMFCs with pure hydro-
gen and CO containing fuel. The thin-film catalyst layer
made by transfer method exhibits high electrocatalyst uti-
lization and low CO tolerance with the same electrocatalyst,
while it is different from the traditional method. The advan-
tages of both electrodes are combined in the composite
anode design. When H2/CO acts as the fuel, the PEMFCs
exhibit less performance drop with the new composite ele-
ctrode structure, while pure hydrogen acts as the fuel, the
fuel cell posses almost the same performance as that of Pt/C
electrocatalyst. It is important for catalyst layer preparation
that the inner catalyst layer should be rich in platinum and
Fig. 8. EDX result of composite anode E5.

H. Yu et al. / Journal of Power Sources 105 (2002) 52–57
57
[3] J. Divisek, H.-F. Oetjen, V. Peinecke, V.M. Schmidt, U. Stimming,
Electrochim. Acta 43 (24) (1998) 3811–3815.
that the outer catalyst layer should be rich in CO tolerance
electrocatalyst, e.g. PtRu/C. The outer catalyst layer should
have a distinct boundary with the inner catalyst layer. In this
way, the effective gradient of the electrocatalyst could be
formed and CO could be electro-oxidized in the outer
catalyst layer, then HOR is able to have high activity in
the inner catalyst layer. The composite anode can improve
the CO tolerance of PEFMC with neither increasing the
amount of electrocatalyst loading nor oxygen injection.
[4] S. Gottesfeld, US Patent 4910099 (1990).
[5] S. Gottesfeld, J. Pafford, J. Electrochem. Soc. 135 (1998) 2651–2652.
[6] V.M. Schmidt, H.F. Oetjen, J. Divisek, J. Electrochem. Soc. 144 (9)
(1997) L237.
[7] R.J. Bellows, E. Marucchi-Soos, R.P. Reynolds, Electrochem. Solid-
State Lett. 1 (2) (1998) 69–70.
[8] T.V. Choudhary, D.W. Goodman, Catal. Lett. 59 (1999) 93–94.
[9] M. Watanabe, M. Uchida, S. Motoo, J. Electroanal. Chem. 229
(1987) 395–406.
[10] V.M. Schmidt, J. Power Source 49 (1994) 299–313.
[11] S.D. Lin, Ting-Chou Hsiao, J. Phys. Chem. B 103 (1999) 97–103.
[12] R.W. Lindstrom, Electrocatalytic gas diffusion electrode employing
thin carbon cloth layer, US Patent 4647359 (1987).
[13] Z.G. Shao, B.L. Yi, M. Han, CN ZL, Appl. 99112826.5 (1999).
[14] H.M. Yu, B.L. Yi, Z.J. Hou, CN ZL, Appl. 01110538.0 (2001).
[15] E. Antolini, L. Giorgi, A. Pozio, E. Passalacqua, J. Power Sourc. 77
(1999) 136–142.
References
[1] G.J.K. Acres, G.A. Hards, R.J. Potter, T.R. Ralph, D. Thompsett,
G.T. Burstein, G.J. Hutchings, Catal. Today 38 (1997) 393–400.
[2] H.-F. Oetjen, V.M. Schmidt, U. Stimming, F. Trila, J. Electrochem.
Soc. 143 (12) (1996) 3838–3842.