Preparation and characterization of CO-tolerant Pt and Pd anodes modified with SnO2 nanoparticles for PEFC

Article abstract of DOI:10.1149/1.2775161

The PdC and PtC anodes modified with SnO2 nanoparticles for the polymer electrolyte fuel cell (PEFC) were investigated using pure and 500 ppm CO -contaminated H2 as fuel gas. Modification of the PdC anode with SnO2 nanoparticles enhanced the cell performance in pure H2, while modification of the PtC anode a little lowered the performance. The effect of SnO2 addition on the performances of the cell with the Pd anode in CO -contaminated H2 was examined. The cell voltage with the Pd SnO2 C anode in 500 ppm CO -contaminated H2 was 0.41 V at a current density of 0.2 A cm2, while that in pure H2 was 0.59 V. The Pd SnO2 C anode exhibited good tolerance to CO poisoning, since the anode adsorbed CO more weakly. Neither electrochemical oxidation of CO nor shift reaction contributed to the CO tolerance of the Pd SnO2 C anode.

Full text of DOI:10.1149/1.2775161

B1132
Journal of The Electrochemical Society, 154 ͑11͒ B1132-B1137 ͑2007͒
0
013-4651/2007/154͑11͒/B1132/6/$20.00 © The Electrochemical Society
Preparation and Characterization of CO-Tolerant Pt and Pd
Anodes Modified with SnO Nanoparticles for PEFC
2
a, ,z
a
b,
b,
Tatsuya Takeguchi, * Yuri Anzai, Ryuji Kikuchi, * Koichi Eguchi, * and
a
Wataru Ueda
a
Catalysis Research Center, Hokkaido University, Kita-ku, Sapporo 001-0021, Japan
b
Department of Energy and Hydrocarbon Chemistry, Graduate School of Engineering, Kyoto University,
Nishikyo-ku, Kyoto 615-8510, Japan
The Pd/C and Pt/C anodes modified with SnO nanoparticles for the polymer electrolyte fuel cell ͑PEFC͒ were investigated using
2
pure and 500 ppm CO-contaminated H as fuel gas. Modification of the Pd/C anode with SnO nanoparticles enhanced the cell
2
2
performance in pure H , while modification of the Pt/C anode a little lowered the performance. The effect of SnO addition on the
2
2
performances of the cell with the Pd anode in CO-contaminated H was examined. The cell voltage with the Pd/SnO /C anode in
2
2
2
5
00 ppm CO-contaminated H was 0.41 V at a current density of 0.2 A/cm , while that in pure H was 0.59 V. The Pd/SnO /C
2
2
2
anode exhibited good tolerance to CO poisoning, since the anode adsorbed CO more weakly. Neither electrochemical oxidation of
CO nor shift reaction contributed to the CO tolerance of the Pd/SnO /C anode.
2
©
2007 The Electrochemical Society. ͓DOI: 10.1149/1.2775161͔ All rights reserved.
Manuscript submitted November 13, 2006; revised manuscript received June 29, 2007.
Available electronically September 7, 2007.
Polymer electrolyte fuel cell ͑PEFC͒ is the system which can
efficiently convert the chemical energy to electricity, since the effi-
ciency of fuel cells is not restricted by Carnot’s theorem. The system
is an attractive power source as a clean power generator for domes-
tic use or automobiles. Since PEFC can generate electricity on site,
the transmission loss can be minimized. When PEFC is applied as
domestic use, the high energy utilization is expected, because the
waste heat can be utilized as air conditioners and hot water by co-
generation systems. However, there are a lot of tasks to be solved in
PEFC systems; one of the most important tasks is to lengthen the
life of the system. High reliability and durability are required to
spread the domestic-use PEFC system. At present, the cell voltage at
a constant same current density gradually decreases with operation
time; therefore, it is being aimed for the declining rate to be less
preferential oxidation of CO and the long-term stability under low
5
O /CO molar ratio. In any case, considerable amounts of precious
2
metals were used in both a shift converter unit and a CO preferential
oxidation unit, resulting in high cost. Moreover, this complicated
system containing a CO-removal unit led to low efficiency and re-
liability. Therefore, it is earnestly desired to develop CO-tolerant
anode catalysts.
Because of these backgrounds, excellent CO-tolerant catalysts
7
like Pt–Ru alloy on carbon have been developed, but Ru is also a
precious material. Therefore, it is important to develop the
CO-tolerant anode without Ru in order to cut the cost of the PEFC.
There are several CO-tolerant Pt-containing anodes, such as Pt/Sn,
8
9
10
Pt/W, Pt/Mo, Pt/Ru/Sn, Pt/Ru/W, Pt/Ru/Mo, Pt/MoO3,
11
Pt–TaO , and Pt–NbO . Even nonalloy-type anode catalysts
x
x
than 2.0 mV per 1000 h in the current density range from 0.1 to
showed the CO tolerance because the interaction between Pt and
2
0
.3 A/cm . This target is expected to be achieved under the follow-
12
metal oxides affected the nature of Pt catalyst. For instance, the
ing conditions: the entrance temperature of the cell ranges from 70
to 80°C; the fuel utilization ranges from 70 to 80%; air utilization
ranges from 30 to 60%. At the declining rate of 2.0 mV per 1000 h,
performance of the cell with the Pt/MoO anode was almost com-
x
9
parable to the PtRu ͑1:1͒ in 100 ppm CO-contaminated H . How-
2
ever, further improvement of the anode catalyst is required to spread
the domestic-use PEFC systems, since higher CO tolerance reduces
the size of the PEFC system and increase stability. Moreover, the
compact direct-methanol PEFC system is being developed, because
it does not need the heavy and bulky external fuel reformer. It is
expected to be applied especially to power electric vehicles. In this
system, various CO-tolerance PtM alloy anodes are also actively
the voltage declines by 70 mV for 40,000 h. It is corresponding to
1
10-year operation including 10 h operation a day.
The natural gas is expected as one of the promising raw materials
of fuel for the domestic use of PEFC, because the natural gas is
provided as city gas. When natural gas is reformed with steam to
produce H , a considerable amount of CO is also produced. When
2
the reformed fuel gas is fed to the anode of PEFC, CO in the re-
formed fuel gas poisons the Pt anode. To decrease the CO concen-
tration, the fuel processor of the PEFC system normally comprises a
steam reformer, a shift converter, and a CO preferential oxidation
unit. In this CO removal unit, the CO concentration can be reduced
to less than 10 ppm from 10 or 20%, but this unit is very bulky.
1
3-15
investigated.
for direct methanol and direct ethanol fuel cells.
The Pt/SnO anode catalysts were proposed for direct ethanol
An alloy of Pt and Sn is used as the anode catalyst
1
3
16
2
1
7-19
fuel cells.
The electrochemical property of Pt/SnO was inves-
2
2
0
tigated in detail. The authors developed Pt/SnO anode catalyst for
PEFC using CO-contaminated H2.
2
2
1,22
2
In this study, the CO-tolerant
Cu-based catalysts were ordinarily used as the shift-converter cata-
Pt and Pd anode catalysts modified with SnO were investigated.
lyst in plants; however, it is difficult to use the Cu-based catalyst in
the domestic-use PEFC system. The PEFC system would be fre-
quently turned on and off. When the system is turned off, Cu cata-
lysts will be easily oxidized by steam and air, and eventually deac-
tivated. Therefore, precious metal catalysts such as Pt catalysts
2
Both of Pt and Pd anode catalysts have the functions of H adsorp-
2
tion and dissociation. In expectation of the interaction between Pt or
Pd and SnO , Pt and Pd anodes were modified with SnO . To make
2
2
the active surface area larger and to increase the number of contact
points between Pt or Pd and SnO2, SnO2 nanoparticles were
supported on CeO -containing oxides were used in the PEFC sys-
2
3
4
5
23
tem. On the other hand, Pd, Au, Pt or Ru is used as the catalyst
for preferential oxidation of CO at low temperature, and Pd/SnO ,
Pt/SnO , and Ir are expected to be used for preferential oxidation of
prepared. Since Pd/SnO2 and Pt/SnO2 were known to be active for
3
the low-temperature oxidation of CO, CO poisoning of Pt-group
2
6
6
metals is expected to be avoided. The catalysts in this study were
2
CO. Among them the Ru catalyst exhibited the high performance for
prepared by the impregnation method, and catalyst metals were dis-
persed on carbon black. The electrochemical activities of the cata-
lysts were examined with a single cell in pure and 500 ppm
CO-contaminated H gas, since the higher CO tolerance the anode
catalysts have, the more merit PEFC has.
*
Electrochemical Society Active Member.
E-mail: takeguch@cat.hokudai.ac.jp
2
z
Downloaded on 2014-11-26 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 154 ͑11͒ B1132-B1137 ͑2007͒
B1133
Experimental
Anodic voltammetry.— CO-stripping voltammetry experiments
were carried out in 0.5 M H SO with usual three-electrode arrange-
2
4
Preparation of electrode catalysts.— All catalysts used in this
study were prepared by the impregnation method. First, the SnO2
nanoparticles were prepared by the following sol-gel method from
tin grains. A 3 g portion of tin, shot, 99.999% ͑Wako Pure Chemi-
cal Industries, Ltd.͒ and 10 g of citric acid ͑Wako Pure Chemical
Industries, Ltd.͒ were dissolved in the 300 mL of an 8 M nitric acid
solution ͑Kanto Chemical Co., Inc.͒. An aqueous solution of NH3
ment in an electrochemical cell using Solartron 1287 electrochemi-
cal interface. A platinum rod with an outer diameter of 6.0 mm and
Ag/AgCl ͑NaCl saturated͒ were used as counter and reference elec-
trodes, respectively. The reference electrode has a potential of E
2
3
=
+ 0.212 V vs the reversible hydrogen electrode ͑RHE͒ and all
potentials are referred to the RHE. The anode catalyst reduced at
00°C was loaded on the glassy carbon electrode, and it was used as
1
͑
Wako Pure Chemical Industries, Ltd.͒ was added to the mixed so-
a working electrode. After CO was supplied to the electrochemical
cell for 1 h, bulk CO was removed by bubbling nitrogen through it
for 15 min. The CO-stripping voltammograms were measured be-
tween −0.2 and 1.2 V ͑vs Ag/AgCl͒ with a scan rate of 20 mV/s.
lution until the pH value reached 8. Next, this solution was refluxed
at 100°C for 2 h. After cooling down, the solution was centrifuged
and washed with distilled water and ethanol to collect the Sn͑OH͒4
sol. The resulting sol containing ethanol and carbon black, Vulcan
XC72R ͑Cabot Corp.͒ were mixed with desired amounts of the
aqueous solutions of Pd͑NO ͒ ͑NH ͒ ͑Tanaka Kikinzoku Kogyo
Construction of membrane electrode assemblies (MEAs).— To
prepare the electrode catalyst paste, n-butyl acetate ͑Wako Pure
Chemical Industries, Ltd.͒, Nafion solution ͑Nafion perfluorinated
ion-exchanged resin, 5 wt % solution in a mixture of lower aliphatic
alcohols and water, Aldrich͒ and ion-exchanged water were mixed,
and then the catalyst was added to the mixture in the ultrasonic wave
bath. Finally, this mixture was put in the ultrasonic wave bath for
30 min. Next, to prepare the anode and cathode, the paste was
painted with a spatula on a peace of carbon fiber paper ͑GDL P50T
2
2
3 2
K.K.͒ containing 4.574 wt % Pd or Pt͑NO ͒ ͑NH ͒ ͑Tanaka Kikin-
2
2
3 2
zoku Kogyo K.K.͒ containing 4.552 wt % Pt. The mixed suspension
of Pd or Pt and carbon black was prepared. Each mixture was kept
on a steam bath at 80°C until the solution was evaporated to form
powders. The dried powders were heated to 300°C in 70 mL/min
flow of 100% N at heating rate of 10°C/min, and the temperature
2
was kept for 30 min. The weight ratio of Pd or Pt/SnO /carbon
2
2
black was 1 / 1 / 4. Thus, Pd/SnO /C, Pt/SnO /C, Pd/C, and Pt/C
Paper, Ballard Material Products, Inc.͒ having an area of 5 cm so
2
2
2
catalysts were prepared. Reduction was carried out in the cell under
the same operation condition at 70°C for the generation experiment.
that the loading of precious metal was 1 mg/cm . On the other hand,
2
cathodes which have 2 mg/cm of Pt were also made in the same
way.
Characterization of the anode catalysts.— The samples with
and without reduction treatment were analyzed by X-ray diffraction
Electrolyte membrane ͑Nafion 117, perfluorinated membrane,
0
.007 in. in thickness͒ was pretreated before use. The pretreatment
͑
XRD͒. Morphology of the anode catalyst without reduction treat-
process followed the conditions of standard membrane treatment
which had been recommended by NEDO PEFC R&D project; the
membrane was successively boiled in a 1 M H SO aqueous solu-
ment was observed by Hitachi H-800 transmission electron micro-
scope. A CHEMBET-3000 was used to evaluate the Brunauer-
Emmett-Teller ͑BET͒ surface area of the anode catalyst without
reduction treatment, based on the amount of N2 adsorbed at the
liquid nitrogen temperature. This apparatus was also used to deter-
mine the amount of CO adsorbed on precious metal. The ordinary
CO pulse method was employed to investigate the catalyst tempera-
ture dependence of the amount of CO adsorbed. The amount of CO
adsorbed was measured with reducing temperature. The weight of
2
4
tion for 1 h; in distilled water for 1 h; in a 1 M H O aqueous
2
2
solution for 1 h; and again in distilled water for 1 h. Thereafter, the
surface of the pretreated electrolyte membrane was washed with
ion-exchanged water. The membrane was sandwiched between the
dried electrodes, and then the component was pressed at 2 MPa and
1
30°C for 10 min. The various anodes were prepared, whereas the
cathodes of these membrane electrode assemblies ͑MEAs͒ were
each sample was 50 mg. The flow rate of H /Ar and He was
2
Pt/C in every case. These MEAs were kept under wet condition.
70 mL/min. The volume of the one pulse was 0.288 mL.
The effect of the temperature on CO adsorption was investigated
Measurement of cell performances (polarization properties).—
after the catalysts were reduced in 5% H /Ar at 100°C for 300 min.
After MEA was fastened with 3.5 N m, it was put in a single cell,
2
Then, He gas was fed, and the temperature was raised to 300°C. The
CO pulse titration method was used to estimate the CO adsorption
on the precious metal. To investigate the temperature dependence of
CO adsorption, the following method was adopted. First, 50 mg of
the sample was packed in the reactor, and He was fed to the sample.
FC05-01SP ͑Electro Chem, Inc.͒. Pure H gas and a gaseous mixture
2
of 22% O -78% N were fed to the anode and the cathode at 100
2
2
and 200 mL/min, respectively. When the CO tolerances of the anode
catalysts were examined, 500 ppm CO-contaminated H was sup-
2
plied. Both gases were humidified by bubbling in water at 73°C.
The polarization properties were measured on Autolab30-MF-SP
by using General Purpose Electrochemical System for Windows ver-
sion 4.9 ͑Autolab Eco Chemie B.V.͒. To avoid the loss of ionic
conductivity by drying the electrolyte, MEAs were kept under the
wet condition before putting in the single cell for measurement.
The sample was heated to 100°C at 10°C/min in 5% H /Ar and
2
then kept at 100°C for 30 min. The sample was flushed with He for
5
min and was heated to 300°C at 10°C/min. The 0.288 mL portion
of CO was pulsed every 2 min, until the intensity of the peak be-
came constant. Another two 0.288 mL CO pulses were added for
confirmation. Then the sample was cooled down to 270°C, and CO
was pulsed in the same way. These procedures were repeated at
Results and Discussion
2
40°C, 210°C, 180°C, 150°C, 120°C, 90°C, 60°C and 25°C in the
Cell performance.— The performances of the cells with various
Pd and Pt anodes for electrochemical reactions were compared in
Fig. 1. The cell with a Pt/C anode exhibited a higher performance
than that with a Pd/C anode, as is expected from the fact that Pt has
same way as mentioned above.
On the other hand, when the effect of the reduction temperature
was investigated, the catalysts were reduced in 5% H /Ar at 100 or
2
200°C. Then, the catalysts were cooled to 25°C, and He was fed
a higher activity for electrochemical oxidation of H than Pd. The
2
followed by CO pulse measurement.
effect of SnO addition to the Pt/C and Pd/C anodes was also ex-
2
X-ray photoelectron spectroscopy ͑XPS͒ was employed to deter-
mine the superficial composition of each catalyst with and without
reduction treatment. These spectra were acquired on a JEOL JPS-
amined. The terminal voltage of the cell with the Pt/C anode is
0.67 V at a current density of 0.2 A/cm , while that with the
2
Pt/SnO /C anode is 0.64 V. This may be caused by the lower con-
2
9
010MC using Mg K␣ radiation at 10 kV and 10 mA. The base
ductivity of SnO than C or Pt. Moreover, the number of active sites
2
−
9
−7
pressure was 6.0 ϫ 10 Torr ͑8.0 ϫ 10 Pa͒, while the operating
pressure was around 5 ϫ 10−7 Torr ͑6 ϫ 10 Pa͒. The XP spectra
were quantified by Gauss-Lorentz fitting to determine the area under
the peaks.
may decrease, since the addition of SnO partly blocks the active
2
−5
site of Pt particles.
On the other hand, the terminal voltage of the cell with the Pd/C
2
anode is only 0.55 V at 0.2 A/cm . However, the voltage with a
Downloaded on 2014-11-26 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B1134
Journal of The Electrochemical Society, 154 ͑11͒ B1132-B1137 ͑2007͒
Figure 1. Comparison of the performances of PEFCs with various Pd and Pt
anodes. Cathode: Pt/C; anode: Pt/C and Pd/C with/without SnO ; anode gas:
2
H ; cathode gas: air; cell: 70°C, humidifier: 73°C.
2
Pd/SnO /C anode rather increased to 0.59 V. Modification with
2
SnO clearly promoted the performance of a Pd/C anode, while the
2
modification is deteriorating for the Pt/C anode. In general, the
number of active sites is strongly related to the particle size of the
catalyst metals. One of the possible reasons for these effects was
that the SnO2 addition may rather increase the dispersion of Pd
particles, resulting in the increase in number of the precious metal
Figure 2. X-ray diffraction patterns of ͑SnO -modified͒ Pd and Pt anode
2
catalysts. ᭺: Pd, b: PdO, ᮀ: Pt, छ: SnO , ࡗ: C .
2
3
active site for electrochemical H oxidation at the catalyst surface.
2
II also shows the surface composition of the catalysts reduced at
00°C. The concentrations of the precious metals at the surface
were not lowered. This means that sintering or migration of the
Otherwise, the modification with SnO may change the surface na-
2
1
ture of Pd particles.
SnO to precious metal did not occur at 100°C. The data also sup-
XRD, crystallite size, and BET surface area.— Figure 2 shows
XRD patterns of the anode catalysts without reduction treatment. Pd
was partly formed during decomposition in the presence of ethanol
2
for Pd/SnO /C, while only PdO was formed for PdO/C. Formation
2
24
of any intermetallic compound like Pt Sn or alloy was not ob-
3
served from XRD for the anode catalysts. Figure 3 shows a trans-
mission electron microscope ͑TEM͒ image of Pd/SnO /C. Light
2
gray particles with diameters of around 4 nm ͑SnO ͒ and dark gray
2
particles ͑Pd͒ with diameters of 6–12 nm are uniformly dispersed on
C.
Crystallite size of the metal was determined by Scherrer’s equa-
tion using the width at half maximum of XRD peak for the sample
after reduction at 100°C. As shown in Table I, the crystallite size of
Pd particles in Pd/C was larger than that of Pt particles in Pt/C.
Since the Pd particles grow more easily than Pt particles, the number
of Pd active sites for electrochemical oxidation of H was smaller
2
than that of Pt, which is closely related to the cell performances with
Pt/C and Pd/C anodes in Fig. 1. The crystallite size of Pd ͑6.1 nm͒
in Pd/SnO /C is smaller than that ͑8.2 nm͒ of Pd/C, while the crys-
2
tallite size of Pt ͑2.8 nm͒ in Pt/SnO /C is smaller than that ͑5.3 nm͒
2
Figure 3. TEM image of Pd/SnO /C.
of Pt/C. Clearly the SnO addition suppressed the growth of Pd or Pt
2
2
particles owing to the interaction between precious metals and
SnO . However, the cell performance was not clearly related to the
Table I. BET surface areas and crystallite sizes estimated from
XRD of the catalysts.
2
crystallite size of Pd or Pt particles in Pd/SnO /C or Pt/SnO /C.
2
2
Surface composition.— XPS measurement was carried out to es-
timate the elemental composition of the surface of each catalyst. The
results are shown in Table II. The concentration of Pt or Pd at the
Surface area
Crystallite size
2
−1
Sample name
C
͑m g
͒
͑nm͒
235
195
169
165
135
135
11 ͑particle size͒
3.1 ͑SnO2͒
8.2 ͑Pd͒
5.3 ͑Pt͒
6.1 ͑Pd͒
surface of Pt/SnO /C or Pd/SnO /C was lower than nominal one
2
2
SnO /C
2
estimated from the source materials for the preparation. It suggests
Pd/C
Pt/C
that the SnO patches were covering precious metal particles, and
2
the surface of Pt and Pd particles was partially undetectable by XPS.
This was recognized as strong metal and support interaction. Table
Pd/SnO /C
2
Pt/SnO /C
2.8 ͑Pt͒
2
Downloaded on 2014-11-26 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 154 ͑11͒ B1132-B1137 ͑2007͒
B1135
Table II. Surface composition determined by XPS and nominal composition for (SnO -modified) Pd and Pt anode catalysts.
2
Nominal
composition
Surface composition ͑mol %͒
͑mol %͒
Catalyst
Pt
Pd
PdO
SnO2
Sn
Pt
Pd
SnO2
56.4
Pt/C
100
100
17.8
19.4
100
Pt/C reduced at 100°C
Pt/SnO /C
82.2
80.6
43.6
2
Pt/SnO /C reduced at 100°C
2
Pd/C
100
100
100
Pd/C reduced at 100°C
Pd/SnO /C
6.5
23.2
27.3
4.2
66.2
67.4
58.6
41.4
2
Pd/SnO /Creduced at 100°C
5.2
2
cell voltages with the Pt/C anode at a current density of 0.2 A/cm2
drastically dropped from 0.67 to 0.10 V with the addition of
500 ppm CO to H . However, the cell voltage with the Pt/SnO /C
anode dropped from 0.64 to 0.37 V at 0.2 A/cm in
CO-contaminated H2 gas as fuel. Since the cell voltage with the
port the absence of the intermetallic compounds. If these compounds
were formed, the concentration of the precious metal of Pt/SnO /C
2
or Pd/SnO /C was considerably smaller than that of Pt/C or Pd/C. It
2
2
2
2
is difficult to compare the number of active sites for Pt/C and
Pt/SnO /C, and for Pd/C and Pd/SnO /C. The surface concentration
2
2
of Pd in Pd/SnO /C was larger than that of Pt in Pt/SnO /C, though
Pt/SnO /C anode was higher than Pt/C in the presence of CO, the
2
2
2
the crystallite size of Pd was larger than that of Pt. The SnO patch
addition of SnO to Pt/C was effective in developing the CO toler-
2
2
should partly cover the surface of Pt particles, and the addition of
ance for the anode catalyst. There are two mechanisms that can
SnO to Pt/C rather suppressed the electrochemical oxidation activ-
ity. On the other hand, Pd particles were stably dispersed on catalyst
explain CO-tolerant effect. One is the change in d-band center of Pt
2
1
6
to weaken the interaction between Pt and CO, and another is called
a bi-functional mechanism. The reaction between CO adsorbed on
Pt and oxygenated species adsorbed on the second metal is pro-
surface by the addition of SnO . From these results, SnO did not
2
2
cover the surface of Pd particles, and the number of active sites for
7
Pd particles was maintained. Even after the Pd/SnO /C anode was
moted via bi-functional mechanism. In addition, it is also reported
2
2
+
reduced in H at 100°C, the 15% of surface Pd was Pd , suggesting
that the nonelectrochemical water-gas shift reaction on the anode
2
the strong chemical interaction between Pd and SnO . The stably
gas diffusion layer lowered CO content in the gas channels of the
2
2
4
dispersed Pd particles having strong interaction with SnO may con-
electrode. In this study, the mechanism of the CO tolerance will be
investigated for the Pt/SnO /C and Pd/SnO /C.
2
tribute to the enhancement of the performance as shown in Fig. 1.
2
2
Figure 5 shows the effect of SnO addition on the performances
2
Anodic voltammetry.— The authors reported that the peak poten-
tial of CO oxidation over commercial Pt/C ͑Johnson Matthey͒ was
of the cell with the Pd anode in 500 ppm CO-contaminated H . The
2
Pd/SnO /C anode catalyst exhibited higher tolerance to CO poison-
2
0
0
.70 V and CO was oxidized on Pt/SnO /C electrode catalyst at
.74 V. However, it was difficult to observe the CO oxidation on
2
2
2
ing than the Pd/C anode. The cell voltage with the Pd/SnO2/C an-
2
ode catalyst at a current density of 0.2 A/cm dropped from 0.59 to
Pd/SnO anode prereduced at 100°C in this experiment. This sug-
2
0
.41 V in 500 ppm CO-contaminated H gas, and the potential loss
2
gested that Pd/SnO anode catalysts did not adsorb CO easily.
2
resulting from the addition of CO was only 0.18 V. It indicates that
the Pd/SnO2/C anode has high tolerance to CO poisoning. The Pd/C
anode itself was higher CO tolerant at low current density around
Cell performance in the presence of CO.— Figure 4 shows the
effect of SnO addition to Pt/C in the presence of 500 ppm CO. The
2
Figure 4. Comparison of the performances of the cells using Pt anodes
Figure 5. Comparison of the performances of the cells using Pd anodes
with/without SnO in the presence of CO. Cathode: Pt/C; anode: Pt/C with/
with/without SnO in the presence of CO. Cathode: Pt/C; anode: Pd/C with/
2
2
without SnO ; anode gas: H or 500 ppm CO/H ; cathode gas: air; cell:
without SnO ; anode gas: H or 500 ppm CO/H ; cathode gas: air; cell:
2
2
2
2
2
2
7
0°C, humidifier: 73°C.
70°C, humidifier: 73°C.
Downloaded on 2014-11-26 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B1136
Journal of The Electrochemical Society, 154 ͑11͒ B1132-B1137 ͑2007͒
Table III. Electrochemical CO oxidation to CO on anodes in PEFC single cell under power generation condition.
2
Cell
voltage
͑V͒
CO conv. to CO2
on anode ͑%͒
͑at OCV͒
Current density
CO conv. to CO2
on anode ͑%͒
Catalyst
Pt/C
͑A cm−2
͒
0.1
0.2
0.2
0.2
0.2
0.3
26.4
9.2
0.8
1.1
23.2
͑0.09͒
͑0.52͒
͑0.07͒
͑0.09͒
͑0.08͒
Pt/SnO /C
0.07
0.015
0.02
0.12
2
Pd/C
Pd/SnO /C
2
Pt–Ru/C
Cathode: Pt/C; anode: Pt/C or Pd/C with/without SnO ; anode gas: 5000 ppm CO/H ; cathode gas: O ; Cell: 30°C, humidifier: 73°C.
2
2
2
2
0
.1 A/cm than the Pt/C anode. The I-V curves of the Pt-type cata-
CO resistance was confirmed for Pt–Ru/C catalyst. However, the
CO conversion on Pt/C is as high as 26.4%, and the CO conversion
lysts in Fig. 4 were bent as a result of oxidation of CO, though those
of the Pd-type catalysts in Fig. 5 remained almost straight at the
same condition. The Pt/C and Pd/C anodes obviously showed dif-
ferent electrochemical activities of CO. The author reported that CO
was electrochemically oxidized with voltage drop over 0.7 V on
Pt/C anode, which corresponds to the cell voltage less than 0.3 V.
As shown in Fig. 4, in the case of the curve of the Pt/C anode during
CO poisoning, CO was electrochemically oxidized at the cell volt-
rather decreased by the addition of SnO ͑9.2%͒. Although contri-
2
bution of the electrochemical CO oxidation was the largest for Pt/C,
Pt did not show the CO tolerance. This means that electrochemical
oxidation of CO contributes little to the CO resistance for the Pt/C
2
2
and Pt/SnO /C. Electrochemical CO oxidation did not contribute to
2
the CO tolerance for the Pd/C and Pd/SnO /C at all. This observa-
2
tion coincided with the fact that CO oxidation was not observed for
2
age less than 0.3 V. Therefore, the voltage drop over 0.07 A/cm
was lower than that over 0.3 V under 0.07 A/cm owing to the
Pd/SnO /C during CO-stripping voltammetry.
2
2
electrochemical oxidation of CO. On the other hand, for the cells
Strength of the CO adsorption on precious metal.— The effect
of temperature on CO adsorption was measured to estimate the
strength of the chemical bond between CO and precious metal.
Table IV shows the result of this experiment. The value of this table
shows the molar ratio of the CO molecules to precious metal. The
number of CO molecules adsorbed on catalysts was divided by the
number of precious metal atoms. The Pt/C catalyst adsorbed CO
more strongly than the Pd/C catalyst. The results agreed with the
data in Fig. 4 that Pt/C was more severely poisoned by CO than
Pt/SnO2/C catalyst. On the other hand, Crabb, Marshall, and
Thompsett reported that the addition of tin to Pt/C suppresses
with the Pd/SnO /C catalyst, the I-V curve was not bent, despite the
2
presence of CO. So, this CO tolerance of Pd/SnO /C was not due to
2
electrochemical CO oxidation, and it may be due to the weak ad-
sorption of CO on Pd.
To confirm the effect of electrochemical reaction, electrochemi-
cal CO oxidation was evaluated using the same cell in the presence
of CO. The adsorbed CO can be also electrochemically oxidized as
follows
+
−
Pt–CO + H O → Pt + CO + 2H + 2e
2
2
2
5
CO was electrochemically oxidized above voltage of 0.7 V vs
chemisorption of both hydrogen and CO. In this case, degradation
of the cell voltage was observed when Pt/Sn/C was used as anode
22
reversible hydrogen electrode ͑RHE͒. Therefore, the CO conver-
sion was quantitatively analyzed at the low cell voltage of 0.2 V.
The results of electrochemical CO oxidation in the presence of
catalyst. In the present investigation, the addition of SnO to Pt/C
2
and Pd/C catalysts did not cause degradation of the cell voltage.
Figure 6 shows the influence of temperature upon the amount of
CO adsorbed on each precious metal, based on Table IV. The
amount of CO adsorbed on each precious metal at 25°C was defined
as 100% for the reference. Clearly Pd/C desorbed CO more easily
than Pt/C. When the sample was heated to 300°C in flowing He,
adsorbed water on the catalyst was expected to be removed. Then,
5000 ppm CO were shown in Table III. Under open-circuit voltage
͑
OCV͒ condition CO was converted to CO only by shift reaction
2
͑
CO + H O → CO + H ͒, and CO was not electrochemically oxi-
2
2
2
dized to CO . The CO conversion at OCV on Pt/C, Pd/C, Pd/SnO
2
2
and Pt–Ru/C was almost identical and less than 0.1%, while that on
Pt/SnO /C was 0.5%. This indicated that all shift reaction scarcely
2
7
contributed to the CO tolerance. Therefore, most of CO in Table II
the bi-functional mechanism did not take place under this exami-
2
was formed not by shift reaction but by electrochemical CO oxida-
tion. Only for the Pt–Ru/C, the CO conversion was as high as
nation condition. The decrease in CO adsorption was due to weak-
ened CO adsorption. Figure 6 indicates that Pt/C strongly adsorbed
CO, since 40% of CO adsorbed on Pt/C at 25°C remained at 300°C.
On the other hand, the relative amount of CO adsorbed on
23.2%. Electrochemical CO oxidation was confirmed at higher cell
voltage of 0.3 V, the contribution of electrochemical oxidation to
Table IV. Molar ratio of adsorbed CO/precious metal.
Molar ratio of adsorbed CO to precious metal ͑%͒
Temp. ͑°C͒
Pt/C
Pt/SnO /C
Pd/C
Pd/SnO /C
2
2
2
6
9
20
50
80
10
40
70
00
5
0
0
27.0
25.5
24.4
23.0
21.4
19.7
18.8
16.8
14.4
10.4
13.3
12.5
11.7
10.6
9.6
8.5
6.9
5.1
3.5
7.0
6.2
5.5
4.8
4.1
3.3
2.5
1.8
1.2
1.1
5.4
4.8
4.2
3.6
2.7
2.0
1.4
0.7
0.5
0.3
1
1
1
2
2
2
3
1.8
Downloaded on 2014-11-26 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 154 ͑11͒ B1132-B1137 ͑2007͒
B1137
strongly affected. This fact indicates the interactions between Pt and
SnO and between Pd and SnO and each other despite low reduc-
2
2
tion temperature or under moderate reduction atmosphere like
PEFC-operating conditions. Therefore, the reduction condition that
can be encountered in the anode atmosphere of PEFC contributed to
the high CO tolerance of the SnO -containing catalysts.
2
Conclusion
The CO tolerance of the PEFC anode catalyst was improved by
addition of SnO2 nanoparticles. The SnO2 addition to the anode
somewhat decreases the electrochemical activity in pure H , but
2
clearly increases the CO tolerance of the anode catalysts. It was
revealed by CO pulse titration that the SnO -containing catalysts
2
adsorbed CO more weakly than the other catalysts without SnO2
nanoparticles. This phenomenon was derived from the interaction
between Pt or Pd atoms and SnO nanoparticles, since these cata-
lysts mixed with carbon did not exhibit intermetallic compounds.
Figure 6. Temperature dependence of CO adsorbed on several Pt or Pd
catalysts.
2
The Pd/SnO /C anode exhibited the highest tolerance to CO poison-
2
ing among the catalysts.
Pt/SnO /C was 15% at 300°C, so that the chemical bond for the
Acknowledgment
2
adsorption should be weak. In the same way, the CO adsorption on
This study was supported by the Industrial Technology Research
Grant Program in 02B61003c from the New Energy and Industrial
Technology Development Organization ͑NEDO͒ of Japan.
Pd/SnO /C was weaker than that on Pd/C. Especially, the relative
2
amount of CO adsorbed on the Pd/SnO /C at 150°C was half of that
2
at 25°C. So, the strength of CO adsorption on Pt or Pd was weak-
ened when SnO2 nanoparticles were added to the catalysts. This
behavior of the CO adsorption on these Pt/SnO /C and Pd/SnO /C
Hokkaido University assisted in meeting the publication costs of this
article.
2
2
catalysts agreed with the I-V curves in the presence of CO. There-
fore, the CO tolerance and the bonding strength between CO and
precious metal were closely related. Weakly adsorbed CO on the
precious metals may contribute to the CO tolerance. Consequently,
the high CO tolerance of these catalysts is due to the strong inter-
References
1
. M. Yoshimura, Y. Hori, Y. Tsuji, A. Yoshida, T. Yonamine, and M. Uchida, Jpn.
Kokai Tokkyo Koho, JP A 2005–216834 A ͑2005͒.
2. R. G. Herman, K. Klier, G. W. Simmons, B. P. Finn, J. B. Bulko, and T. P.
Kobylinski, J. Catal., 56, 407 ͑1979͒.
. M. M. Gadgil, R. Sasikala, and S. K. Kulshreshtha, J. Mol. Catal., 87, 297 ͑1994͒.
4. M. Haruta, S. Tsubota, T. Kobayashi, H. Kageyama, K. J. Genet, and B. Delmon, J.
3
actions between precious metal and SnO nanoparticles. The results
2
suggest that weakening of CO adsorption on precious metal was
important for improvement of CO tolerance.
Catal., 144, 175 ͑1993͒.
. M. Echigo, N. Shinke, S. Takami, S. Higashiguchi, K. Hirai, and T. Tabata, Catal.
Today, 84, 209 ͑2003͒.
5
6
7
8
. M. Okumura, N. Masuyama, E. Konishi, S. Ichikawa, and T. Akita, J. Catal., 208,
Effect of reduction temperature on the strength of CO adsorption
on precious metal.— The CO adsorption on the precious metals was
measured to estimate the effect of reduction temperature on the
bonding strength between CO and precious metal. The importance
of the intimate mixing of Pt and the second elements was claimed by
4
85 ͑2002͒.
. M. Watanabe and S. Motoo, J. Electroanal. Chem. Interfacial Electrochem., 60,
267 ͑1975͒.
. S. Mukerjee, R. C. Urian, S. J. Lee, E. A. Ticianelli, and J. McBreen, J. Electro-
chem. Soc., 151, A1094 ͑2004͒.
9. M. Götz and H. Wendt, Electrochim. Acta, 43, 3637 ͑1998͒.
0. T. Ioroi, K. Yasuda, Z. Siroma, N. Fujiwara, and Y. Miyazaki, J. Electrochem. Soc.,
150, A1225 ͑2003͒.
1. A. Ueda, Y. Yamada, T. Ioroi, N. Fujiwara, K. Yasuda, Y. Miyazaki, and T. Koba-
yashi, Catal. Today, 84, 223 ͑2003͒.
2. T. Matsui, K. Fujiwara, T. Okanishi, R. Kikuchi, T. Takeguchi, and K. Eguchi, J.
Power Sources, 155, 152 ͑2006͒.
8
Götz and Wendt. The reduction temperature of the catalyst would
1
affect CO tolerance. Table V shows the result of the effect of reduc-
tion temperature. The Pd-containing catalysts reduced at 200°C ad-
sorbed smaller amount of CO than the Pt-containing catalyst. The
amount of CO adsorbed on the Pd-containing catalyst was about half
1
1
of that on the Pt-containing catalysts. The addition of SnO nano-
13. E. Antolini, Mater. Chem. Phys., 78, 563 ͑2003͒.
4. E. Christoffersen, P. Liu, A. Ruban, H. L. Skriver, and J. K. Norskov, J. Catal.,
2
1
1
1
1
particles clearly reduced the amount of CO adsorbed. The reduction
temperature did not strongly affect the amounts of CO adsorbed on
Pt/C and Pd/C, while those on Pt/SnO /C and Pd/SnO /C were
1
99, 123 ͑2001͒.
5. A. S. Arico, Z. Poltarzewski, H. Kim, A. Morana, N. Giordano, and V. Antonucci,
J. Power Sources, 55, 159 ͑1995͒.
6. W. Zhou, Z. Zhou, S. Song, W. Li, G. Sun, P. Tsialaras, and Q. Xin, Appl. Catal.,
B, 46, 273 ͑2003͒.
7. L. Jiang, G. Sun, Z. Zhou, S. Sun, Q. Wang, S. Yan, H. Li, J. Tian, J. Guo, B. Zhou,
2
2
and Q. Xin, J. Phys. Chem. B, 109, 8774 ͑2005͒.
8. A. L. Santos, D. Profeti, and P. Olivi, Electrochim. Acta, 50, 2615 ͑2005͒.
9. M. B. de Oliveira, L. P. R. Profeti, and P. Olivi, Electrochem. Commun., 7, 703
Table V. Molar ratio of adsorbed CO/precious metal.
1
1
Molar ratio of
͑2005͒.
Catalyst
Pt/C
Reduction temperature ͑°C͒
adsorbed CO/Pt, Pd ͑%͒
20. M. Nakada, A. Ishihara, S. Mitsushima, N. Kamiya, and K. Ota, Electrochem.
Solid-State Lett., 10, F1 ͑2007͒.
1. T. Okanishi, T. Matsui, T. Takeguchi, R. Kikuchi, and K. Eguchi, Appl. Catal., A,
298, 181 ͑2006͒.
100
22.5
20.1
19.3
10.0
9.5
7.0
5.3
0.5
2
2
00
100
00
100
00
100
00
Pt/SnO /C
22. T. Matsui, K. Fujiwara, T. Okanishi, R. Kikuchi, T. Takeguchi, and K. Eguchi, J.
2
Power Sources, 155, 152 ͑2006͒.
2
2
2
3. J. Zhang and Lian Gao, Chem. Lett., 32, 458 ͑2003͒.
4. E. I. Santiago, M. S. Batista, E. M. Assaf, and E. A. Ticianelli, J. Electrochem.
Soc., 151, A944 ͑2004͒.
Pd/C
2
Pd/SnO /C
25. E. M. Crabb, R. Marshall, and D. Thompsett, J. Electrochem. Soc., 147, 4440
͑2000͒.
2
2
Downloaded on 2014-11-26 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).