Oxygen reduction on Ru1.92Mo0.08SeO4, Ru/carbon, and Pt/carbon in pure and methanol-containing electrolytes

Article abstract of DOI:10.1149/1.1393579

The oxygen reduction reaction (ORR) activity of a Ru_1.92Mo_0.08SeO₄ catalyst, a Vulcan XC72-supported Ru catalyst and, for comparison, a Vulcan XC72-supported Pt catalyst was studied with a rotating ring-disk electrode. The very similar reaction characteristics of the two Ru catalysts in pure and CH₃OH-containing H₂SO₄ electrolyte, which differ markedly from those of the Pt catalyst, indicate that the reactive centers in both Ru catalysts must be identical. They are highly selective (>95%) toward reduction to H₂O (four electron pathway), independent of the presence of methanol. In the latter case, they are 100% selective toward the ORR, i.e., completely methanol tolerant, while the ORR on Pt catalysts is accompanied by significant CH₃OH oxidation. Based on mass specific current densities, however, the Ru catalysts are significantly less active than the standard Pt catalysts. Only at methanol concentrations above 10-30 mM does their methanol tolerance make them more active than Pt/Vulcan. Implications for their use as cathode catalysts in a direct methanol fuel cell are discussed.

Full text of DOI:10.1149/1.1393579

2
620
Journal of The Electrochemical Society, 147 (7) 2620-2624 (2000)
S0013-4651(99)10-097-1 CCC: $7.00 © The Electrochemical Society, Inc.
Oxygen Reduction on Ru1.92Mo0.08SeO , Ru/Carbon, and Pt/Carbon in
4
Pure and Methanol-Containing Electrolytes
T. J. Schmidt, * U. A. Paulus,a,c,* H. A. Gasteiger,a,d,** N. Alonso-Vante, and R. J. Behm **
a, ,e,z
b
a,
a
Abteilung Oberflächenchemie und Katalyse, Univeristät Ulm, D-89069 Ulm, Germany
Laboratoire Electrocatalysis, UMR-CNRS 6503, Université de Poitiers, F-86022 Poitiers, France
b
The oxygen reduction reaction (ORR) activity of a Ru1.92Mo0.08SeO catalyst, a Vulcan XC72-supported Ru catalyst and, for com-
4
parison, a Vulcan XC72-supported Pt catalyst was studied with a rotating ring-disk electrode. The very similar reaction character-
istics of the two Ru catalysts in pure and CH OH-containing H SO electrolyte, which differ markedly from those of the Pt cata-
3
2
4
lyst, indicate that the reactive centers in both Ru catalysts must be identical. They are highly selective (>95%) toward reduction to
H O (four electron pathway), independent of the presence of methanol. In the latter case, they are 100% selective toward the ORR,
2
i.e., completely methanol tolerant, while the ORR on Pt catalysts is accompanied by significant CH OH oxidation. Based on mass
3
specific current densities, however, the Ru catalysts are significantly less active than the standard Pt catalysts. Only at methanol
concentrations above 10-30 mM does their methanol tolerance make them more active than Pt/Vulcan. Implications for their use
as cathode catalysts in a direct methanol fuel cell are discussed.
©
2000 The Electrochemical Society. S0013-4651(99)10-097-1. All rights reserved.
Manuscript submitted October 25, 1999; revised manuscript received March 16, 2000.
ing ring-disk electrode (RRDE).18,19 Data on the ORR activity and
selectivity of these catalysts with respect to H O formation and
methanol oxidation are presented. Similar results for Pt/Vulcan XC72
are provided as reference. This is followed by an absolute assessment
of the activities on the basis of mass specific current densities. It is
Direct methanol fuel cells (DMFCs) offer an attractive way for
conversion of chemical energy into electrical energy for various sta-
2
2
1
tionary and mobile applications. The DMFC performance is, how-
ever, limited by a number of basic problems, including kinetic
restraints for methanol electro-oxidation and methanol permeation
crossover) from the anode to the cathode side through the common-
(
demonstrated that (i) Ru1.92Mo0.08SeO and Ru/Vulcan XC72 exhib-
4
ly used Nafion-type polymer electrolyte membranes. The former re-
sults in high overpotentials on the anode side even on state-of-the-art
it practically identical reaction characteristics, indicating the exis-
tence of identical reactive centers in both catalysts, and that (ii) the
activity of the Ru-containing catalysts is significantly lower than that
of the standard Pt catalyst in pure acid solution.
2
anode catalysts like PtRu. The latter is concomitant with cathode
performance losses due to the formation of “mixed potentials” on the
3-5
Pt-containing cathode electrocatalysts.
To avoid crossover, the
development of less methanol permeable membranes, e.g., acid-
doped polybenzimidazole (PBI) or polyacrylamide (PAA), would
be highly desirable.
Experimental
1,6
For the measurements, we used an unsupported semiconducting
transition metal chalcogenide Ru1.92Mo0.08SeO catalyst, which was
prepared according to Ref. 13 [mean particle size between 3 and
nm, determined by high resolution transmission electron micros-
copy (HRTEM) ] and two commercially available carbon-supported
Pt and Ru catalysts, respectively (20% Pt/Vulcan XC72, mean parti-
cle size determined by HRTEM, dPt ϭ 3.7 Ϯ 1 nm; 20% Ru/Vul-
can XC72, mean particle size determined by X-ray diffraction
XRD), dRu ϭ 4.4 nm; both from E-TEK Inc., Natick, MA). The E-
4
Another approach is to allow methanol crossover but to use an
oxygen reduction selective cathode catalyst, i.e., to supress methanol
oxidation on the fuel cell cathode. This route is particularly attractive
4
13
7
if methanol passing through to the cathode is recovered. In the last
years, several transition metal compounds have been proposed as
oxygen reduction reaction (ORR) selective catalysts: metal-contain-
18
3
,8
ing porphyrin systems,
chevrel phase-type compounds (e.g.,
(
9
,10
Mo Ru Se ),
transition metal sulfides (e.g., Mo Ru S ,
or other transition metal chalcogenides, e.g.,
In contrast to Pt cathode catalysts, which are
unselective for the ORR in the presence of methanol,
4
2
8
x
y
z
TEK catalysts were pretreated in a tube furnace at 300ЊC, first in 10%
1
1,12
Mo Rh S ),
x
y
z
O /Ar, then in pure H (each step for 30 min) in order to remove pos-
2
2
11,12
(
Ru1ϪxMo ) SeO .
18
x y
z
sible surface contaminants. XRD data were acquired after this pre-
treatment and showed a pure hexagonal close packed Ru phase con-
sistent with metallic Ru. The chalcogenide catalyst was used unsup-
ported and without further pretreatment.
3,15
all the afore-
mentioned systems are characterized by their nearly 100% selectivi-
ty toward oxygen reduction. Fundamental questions relevant for their
application as cathode catalyst, such as the absolute mass-specific
activity in the ORR or the extent of H O formation (two electron
The electrochemical measurements were conducted in a ther-
mostated standard three compartment electrochemical cell using an
interchangeable ring-disk electrode setup with a bipotentiostat and
rotation control (Pine Instruments). The Pt ring electrode was poten-
tiostated at 1.2 V vs. reversible hydrogen electrode (RHE) where the
detection of peroxide is diffusion limited. The collection efficiency,
2
2
pathway) as compared to reduction to H O (four electron pathway) in
2
the presence of methanol, are still open. Another interesting system
for the ORR in a methanol background are supported Ru catalysts,
since pure Ru is nearly inactive toward methanol oxidation even at
16
elevated temperatures (60-80ЊC), but shows acceptable activity
19
N, was previously determined to be N ϭ 0.19 Ϯ 5%. A saturated
17
toward oxygen reduction at least in alkaline solution.
calomel electrode (SCE), separated from the working electrode com-
partment by a closed electrolyte bridge in order to avoid chloride
contamination, was used as reference electrode. All potentials in this
study, however, refer to that of the RHE.
In this work, we studied the ORR kinetics of the transition metal
chalcogenide Ru1.92Mo0.08SeO and Ru/Vulcan XC72 both in the ab-
4
sence and presence of methanol by measuring with a thin-film rotat-
The rotating thin-film electrodes were prepared as described in
*
*
*
* Electrochemical Society Student Member.
* Electrochemical Society Active Member.
Present address: Paul-Scherrer-Insitut, Renewable Energies, Ch-5232 Villigen PSI,
1
8,20
2
recent studies.
In summary, 70 ␮gcatalyst/cm were attached onto
c
a glassy carbon substrate (Sigradur G, Hochtemperaturwerkstoffe
GmbH) with a thin (<0.2 ␮m) Nafion film cast from an aqueous
solution prepared according to the method described in Ref. 21.
Owing to the thin catalyst agglomerate (<1 ␮m) and the thin Nafion
film, mass-transport resistances in the Nafion film and the catalyst
Switzerland.
*
*
Berkeley, California 94720, USA.
*
d
Present address: GM Global R&D, GAPC, Honeoye Falls, NewYork 14472, USA.
Present address: Lawrence Berkeley National Laboratory, University of California,
e
z
E-mail: tjschmidt@lbl.gov
Downloaded on 2014-03-26 to IP 134.208.103.160 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, 147 (7) 2620-2624 (2000)
2621
S0013-4651(99)10-097-1 CCC: $7.00 © The Electrochemical Society, Inc.
agglomerate were negligible up to ca. 1.5 A/mgmetal for the carbon-
1
8,20
supported catalysts and ca. 0.5 A/mg Ru for the chalcogenide.
Table I. Activity of Ru1.92Mo0.08SeO toward the ORR in
dependence of the cycling time and positive potential limit
following the experimental protocol described in the text.
4
2
The resulting electrodes with noble metal loadings of 40 ␮g /cm
Ru
2
for the chalcogenide catalyst and 14 ␮gmetal/cm for the Pt and Ru
catalyst, respectively, were immersed in deaerated (argon, Westfalen
N6.0) 0.5 M H SO (Merck suprapure) electrolyte under potential
E limit
i at 0.7 V
(mA/cm )
2
4
2
control at 0.1 V.
(min)
(V)
For the oxygen reduction experiments, the electrolyte was satu-
rated with oxygen (Westfalen N5.7). Current densities given in this
1
.7
15
90
15
90
15
90
0.8
0.8
0.9
0.9
1.0
1.0
study are normalized either to the geometric area of the electrode
1.7
1.8
1.6
1
1
2
(
0.283 cm ) or to the noble metal loading on the electrode (see
above). All measurements were carried out at 60°C.
.0
.4
Results and Discussion
Base voltammetry and ORR on Ru1.92Mo0.08SeO .—This section
4
reports on the base voltammetry and the activity toward oxygen
This procedure was then repeated applying more positive potential
limits of 0.9 and 1.0 V, respectively. The resulting ORR currents at
reduction on a Ru1.92Mo0.08SeO electrode. The base voltammo-
4
gram of Ru1.92Mo0.08SeO (Fig. 1, 20 mV/s, 60°C, dashed line) is
4
0
0
.7 V are summarized in Table I. Up to a positive potential limit of
.9 V, the current densities for the ORR at 0.7 V remain stable.
characterized by a large pseudocapacitance between 0.04 and
0
.85 V. At negative potentials, the absence of a distinct Hupd region
When applying a higher positive potential limit (1.0 V), the activity
decreases drastically. This finding, however, is in strong contrast to
the instability of the base voltammetry (Fig. 1) with a positive
potential limit of only 0.85 V. One can conclude that the dissolving
Se and Mo-containing species (Epositive Յ 0.85 V) are likely to not
represent the active center of the catalyst for the ORR. On the other
hand, the active centers are deactivated when applying a positive
potential, Epositive > 0.9 V, consistent with the onset of Ru dissolu-
(
as, e.g., compared to Pt electrodes) indicates no or only very weak
hydrogen adsorption. In the anodic sweep, a constant increase in the
voltammetric currents can be observed. In the negative-going sweep,
a broad oxide reduction wave centered at ϳ0.4 V is visible. Cycling
the electrode for more than 1 h (Fig. 1, solid and dotted line) up to
0
.85 V, however, lead to a significant decrease of the pseudocapac-
itance. According to ex situ X-ray photoelectron spectroscopy (XPS)
data, these changes can be attributed to the partial dissolution of
SeO and MoO compounds. This dissolution process takes place
mainly within the first 30 min of cycling, whereas in the second
3
1
3
2
2
tion at these high electrode potentials. These findings also strong-
ly indicate that the Ru centers in the chalcogenide play a central role
2
3
2
3
in the ORR mechanism.
0 min the voltammetry remains more or less stable.
In order to test whether the dissolution process affects the ORR
A closer look on the ORR activity of the Ru1.92Mo0.08SeO cata-
4
lyst is depicted in Fig. 2. The solid line in Fig. 2a represents the
activity, the ORR activity was determined in a next step as a func-
tion of cycling time and positive potential limit. These measure-
ments were carried out according to the following experimental pro-
tocol (60ЊC): After cycling, a freshly prepared electrode between
2
oxygen reduction on this chalcogenide catalyst at 60ЊC (40 ␮g /cm ,
Ru
5
mV/s, 1600 rpm). As obvious by comparison with the ORR cur-
2
rents on Pt/Vulcan (dotted line, 14 ␮g /cm ), the diffusion-limited
Pt
0
.04 and 0.8 V (20 mV/s) for 15 min in O -free electrolyte, oxygen
2
reduction currents were recorded in O -purged electrolyte (5mV/s,
2
1
600 rpm). In a subsequent step, after replacing O in the electrolyte
2
by Ar and cycling the electrode in the same potential range for
another 90 min, the oxygen reduction currents were recorded again.
Figure 2. (a) Oxygen reduction and (b) fractional H O formation (see
2
2
2
Figure 1. Variation of base voltammetry of Ru1.92Mo0.08SeO with cycling
Eq. 1) on Ru1.92Mo0.08SeO (40 ␮g /cm ) in 0.5 M H SO (solid line) and
4
4 Ru 2 4
time. After: five min (dashed line), 35 min (solid line), and 65 min (dotted
in 0.5 M H SO /0.5 M methanol (dashed line). Dotted line: ORR on Pt/Vul-
2 4
2
2
line). (20 mV/s, 0.5 M H SO , 60ЊC, 40 ␮g /cm .)
can (14 ␮g /cm ). E
ϭ 1.2 V, 60ЊC, 5 mV/s, 1600 rpm.
2
4
Ru
Pt
ring
Downloaded on 2014-03-26 to IP 134.208.103.160 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

2
622
Journal of The Electrochemical Society, 147 (7) 2620-2624 (2000)
S0013-4651(99)10-097-1 CCC: $7.00 © The Electrochemical Society, Inc.
current is not reached over the applied potential range, indicating
mixed kinetic diffusion-controlled ORR at potentials <0.6 V (note
that the catalyst agglomerate thickness is the same for both catalysts,
so that mass transport related artifatcs can be ruled out).
2
Compared to the ORR on Pt/Vulcan (14 ␮g /cm , and otherwise,
Pt
same conditions as above) the overpotential increases significantly
(see below). A quantitative comparison of the activity on both cata-
lysts, however, cannot be carried out since different noble metal
loadings were used for both systems: this issue is discussed later. In
Fig. 2b, the fraction of peroxide, x , which is formed in the two
H2O
electron pathway on the disk electrode, is plotted. x
can be cal-
H2O
D
culated from the ring currents, I , the disk currents, I , and the N,
R
1
9
according to Eq. 1
IR
2
N
IR
xH
ϭ
[1]
O
2
2
ID
ϩ
N
According to Fig. 2b (solid line) only negligible amounts H O
2
2
(
ϳ2%) are formed on the chalcogenide catalyst within the whole
potential range. Hence, this catalyst is highly selective (ϳ98%)
toward oxygen reduction via the 4 electron pathway (i.e., O reduc-
2
tion to H O) in qualitative agreement with similar findings in
2
Ref. 24. Even at potentials <0.25 V, peroxide formation does not in-
crease. This confirms that on the chalcogenide catalyst Hupd is ab-
sent or very weak. This is in contrast to Pt/Vulcan electrodes (dotted
line in Fig. 2b), where the peroxide formation rate increases rapidly
below 0.25 V, which was attributed to site blocking by a strongly
adsorbed Hupd layer, concomitant with a blocking of paired adsorp-
Figure 3. (a) Oxygen reduction and (b) fractional H O formation (see
1
9
2 2
tion sites for O dissociation.
2
2
Eq. 1) on Ru/Vulcan (14 ␮gRu/cm ) in 0.5 M H SO (solid line) and in 0.5 M
2 4
Finally, the effect of methanol on the ORR activity and selectiv-
ity (4 electron vs. 2 electron pathway; ORR vs. methanol oxidation)
was investigated. In the dashed line in Fig. 2 the ORR current den-
H SO /0.5 M methanol (dashed line). Dotted line: ORR on Pt/Vulcan
2
4
(14 ␮g /cm2). E
ϭ 1.2 V, 60ЊC, 5 mV/s, 1600 rpm.
Pt
ring
sities as well as x
in presence of 0.5 M methanol in the elec-
H2O
increasing H O formation is similar to that of Pt/Vulcan in the H
upd
trolyte are illustrated. Over the entire potential range, the measured
ORR currents are not affected by the presence of methanol, indicat-
ing that there is no methanol effect on the ORR nor methanol oxida-
tion and that the chalcogenide catalyst is nearly 100% selective for
2
2
region, although it originates from a different adsorbate. At poten-
tials <0.5 V, the product distribution (i.e., percent H O ) is nearly
2
2
constant which also points to weak hydrogen adsorption, as in the
1
4
case of Ru1.92Mo0.08SeO . The origin of the slight maximum in the
O reduction. Similar results were reported previously. Further-
4
2
H O formation around 0.3 V is still unclear, and further measure-
more, nearly similar amounts of peroxide are collected on the ring
electrode with or without methanol in solution. In this context, it
should be mentioned that peroxide detection on the Pt ring electrode
at 1.2 V was possible, since at this potential, the electrode is rather
inactive toward methanol oxidation and an only small but very sta-
ble methanol oxidation current had to be subtracted from the perox-
ide oxidation currents (methanol oxidation currents under identical
conditions on the ring electrode were ca. one order of magnitude
lower than those corresponding to 2% H O in Fig. 2b, 3b, and 4b).
2 2
ments are planned to clear this point. Overall, however, Ru/Vulcan is
5 to 97% selective toward oxygen reduction to H O (4 electron
9
2
pathway) in the entire potential range.
It is well known that a pure Ru catalyst does not show significant
activity toward methanol oxidation, rationalized by a lack of methanol
adsorption sites on the Ru surface due to the presence of strongly
16
adsorbed oxygenated species. Therefore, we expect little changes in
the ORR kinetics on Ru with methanol in solution. That the ORR
kinetics indeed stay the same with methanol in solution can be nicely
seen in Fig. 3a (dashed line, 0.5 M methanol). Hence, Ru/Vulcan is
equally an ϳ100% selective catalyst for oxygen reduction in a
methanol background, identical to the Ru chalcogenide. The rate of
peroxide formation also stays the same with methanol in solution,
indicating an unchanged balance between oxygen reduction to water
or to peroxide (2 electron vs. 4 electron pathway). These results for
2
2
Our findings point to the same ORR mechanism in both electrolytes.
Oxygen reduction on Ru/Vulcan XC72.—For comparison, we
performed similar measurements on a Ru/Vulcan catalyst (Fig. 3;
2
1
4 ␮g /cm , 60ЊC, 5 mV/s, 1600 rpm). Analogous to the chalco-
Ru
genide diffusion limited currents are not reached over the whole
potential range, pointing again to still remaining mixed kinetic dif-
fusion control of the reaction. This also agrees well with oxygen
reduction data on massive Ru in alkaline solution and the behavior
of Ru1.92Mo0.08SeO described in the previous section. Compared to
Ru/Vulcan and Ru1.92Mo0.08SeO indicate (surprisingly) similar reac-
1
7
4
tion behavior in terms of both the selectivity with respect to the reac-
tion product of the ORR (water vs. peroxide) and the selectivity
toward oxygen reduction in a methanol background. The only signif-
icant difference in H O formation is at potentials above 0.5 V, where
4
that of Pt/Vulcan, the overpotential in the potential range between
2
0
.8 and 0.6 V (equivalent to current densities up to ϳ2.5 mA/cm )
2
2
is roughly 100 to 200 mV larger on the Ru/Vulcan electrode. A sim-
ilarly lower catalytic activity of Ru compared to Pt in 0.5 M H SO
the H O formation is larger on the Ru catalyst, which might be relat-
2
2
2
4
ed to remaining traces of Se and Mo species in the chalcogenide.
2
5
was observed by Romero et al. This can be rationalized by the
blocking of surface sites strongly adsorbed oxygenated species
Comparison of the ORR mass-specific activity.—This section,
focuses on a quantitative assessment of the activity of the Pt/Vulcan,
2
6
(
OH ) at lower potentials than on Pt electrodes, which is reducing
ad
the ORR activity. This can also be seen in the increasing amount of
peroxide formed on the Ru/Vulcan electrode at E > 0.5 V which
we attribute to the site blocking effects as discussed above for Hupd
on Pt/Vulcan, but in this case by strongly adsorbed oxygenated
Ru/vulcan, and Ru1.92Mo0.08SeO catalysts for oxygen reduction by
comparing mass-specific current densities.
Figure 4a we plotted the data from Fig. 2 and 3 on a mass-specif-
ic current density rather than a geometric current density scale in order
to compare the results on the electrodes with different metal loadings
4
species (from H O decomposition). Hence, the reason for the
2
Downloaded on 2014-03-26 to IP 134.208.103.160 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, 147 (7) 2620-2624 (2000)
2623
S0013-4651(99)10-097-1 CCC: $7.00 © The Electrochemical Society, Inc.
Figure 5. ORR on Ru/Vulcan in the presence of 0.5 M methanol (straight line)
and on Pt/Vulcan in the presence of 10 mM (dotted line) and 30 mM (dashed
line) methanol. The ORR currents on Pt/vulcan without methanol are given as
reference (dashed-dotted line). 0.5 M H SO , 5 mV/s, 1600 rpm, 60ЊC.
2
4
Implications for DMFCs.—At a typical cathode potential of
0.7 V in a DMFC, the mass-specific current density of both
Ru1.92Mo0.08SeO and Ru/Vulcan is Ϸ80 mA/mg . Assuming a
Figure 4. (a) Mass-specific Tafel plots for: Ru1.92Mo0.08SeO4 (triangles),
Ru/Vulcan (squares), and Pt/Vulcan (circles) extracted from Fig. 2a and 3a.
4
Ru
2
DMFC cathode loading of 1 mg /cm , this would result in a maxi-
Ru
2
mum current density in the DMFC of 80 mA/cm in the limits of no
2
in the previous sections (Ru .92Mo SeO : 40 ␮g /cm , Pt/Vulcan:
1
0.08
4
Ru
diffusion resistance. This represents a quite low current density even
for a DMFC, where at least three times higher current densities are
currently being obtained.
In the last part, we want to compare the ORR activities of Pt/Vul-
can and Ru/Vulcan in the presence of methanol in order to estimate
the maximum amount of methanol, for which the Pt/Vulcan catalyst
2
2
1
4 ␮g /cm , Ru/Vulcan: 14 ␮g /cm ). To calculate the mass-specif-
Pt Ru
ic current densities, i , the data in Fig. 2a and 3a were corrected for
m
the diffusion resistance to extract the kinetic current densities, i ,
k
27
20
Eq. 2, which were then divided by the noble metal loading
1
1
1
ϭ
Ϫ
[2]
still shows better kinetics than Ru/Vulcan (Fig. 5, 5 mV/s, 1600 rpm,
ik
i
id
2
1
4 ␮gmetal/cm ). Since it is not possible to extract the true kinetic cur-
In this relation, i represents the diffusion limited current densi-
rent density (Eq. 2) required to calculate mass-specific current den-
sities for the combined ORR/methanol oxidation process on Pt/Vul-
can, only data for Pt/Vulcan and Ru/Vulcan (each 14 ␮gmetal/cm ) are
d
ty. On both Ru/Vulcan and Ru1.92Mo0.08SeO , no diffusion-limited
4
2
current densities are observed. Since the diffusion-limited current
does not depend on the catalyst (note that the catalyst agglomerate
thickness was the same in all cases), we used the value obtained
from the measurement on Pt/Vulcan for all calculations. Comparing
compared in Fig. 5.
The i/E curves for oxygen reduction obtained on the Pt/Vulcan
catalyst in the presence of methanol are qualitatively identical to
those reported in the literature for a polycrystalline Pt RDE at room
the activity for Ru1.92Mo0.08SeO and Ru/Vulcan on the basis of
4
1
5
mass-specific current densities plotted in Fig. 4a, only negligible
differences are observed. Since the mean particle size of the chalco-
genide and the carbon supported Ru are both of the order of 4 nm
temperature. They represent superpositions of the ORR and meth-
anol oxidation, respectively. For O reduction on Pt/Vulcan in the
2
presence of 10 mM CH OH (Fig. 5, dotted line), the potential loss
3
(
i.e., comparable Ru dispersion) and since the catalyst particles in
with respect to oxygen reduction in the absence of methanol (Fig. 5,
dashed-dotted line) increases by only 20 to 40 mV at current densi-
ties of up to Ϸ2 mA/cm (equivalent to potentials above ca. 0.75 V).
thin film RRDE experiments are completely wetted by the elec-
trolyte (i.e., 100% utilization, see Ref. 20), this close agreement in
mass-specific current densities (A/mg ) suggests that the catalytic
activity of the Ru1.92Mo0.08SeO catalyst is predominantely due to
Ru. This, indeed, was already confirmed by in situ EXAFS meas-
urements on Ru1.92Mo0.08SeO4 where it was shown that Ru cen-
ters are essential for the ORR. The additional fact presented here that
both Ru and Ru1.92Mo0.08SeO exhibit the same tolerance toward
methanol and that the only apparent difference between the chalco-
genide and the Ru catalyst is the lower percentage of H O above
2
At higher current densities, the methanol-induced overpotential
(mainly due to a mixed potential) increases to up to ca. 150 mV due
to the fact that the maximum methanol oxidation activity is centered
Ru
4
2
3
15
about ϳ0.7 V. A comparison with the Ru/Vulcan catalyst in meth-
anol containing electrolyte (Fig. 5, solid line) clearly shows that the
activity of the Ru catalyst (and, by analogy, of the Ru chalcogenide)
is inferior to the Pt/Vulcan catalyst in the entire potential range.
After increasing the methanol concentration to 30 mM, however,
there is a clear advantage of the Ru/Vulcan catalyst over the Pt/Vul-
4
2
2
0
.6 V for the chalcogenide, even suggests that the chalcogenide de-
2
rives its activity solely from Ru. At any rate, both Ru-containing cat-
alysts in our study are clearly significantly less active than Pt/Vulcan
for the ORR as evidenced by the higher overpotential (about 0.1 to
can catalyst at cathodic current densities below 3 mA/cm (equiva-
lent to potentials over 0.6 V) owing to the large methanol-induced
potential loss of ca. 0.2 V observed for the Pt/Vulcan catalyst.
Therefore, the maximum methanol concentration for which pure
Pt/Vulcan shows still a better performance than the methanol toler-
ant catalyst Ru/Vulcan, at 60ЊC, is ϳ10-30 mM.
0
.2 V) compared to Pt/Vulcan. Considering that the dispersion of all
three catalysts is quite comparable (mean particle sizes of 3-4 nm),
the one to two orders of magnitude larger mass-specific current den-
sities of Pt/Vulcan vs. the Ru-based catalysts (Fig. 4a) indicates that
the intrinsic catalytic activity of the Pt/Vulcan catalyst toward the
ORR is also one to two orders of magnitude larger than that of either
Ru/Vulcan or the Ru chalcogenide.
For applications as cathode catalyst in a DMFC, both Ru-based
catalysts are attractive because of their methanol tolerance, in par-
ticular for high crossover rates and if the methanol can be recycled
7
from the cathode exhaust. It should be noted, however, that the
Downloaded on 2014-03-26 to IP 134.208.103.160 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

2
624
Journal of The Electrochemical Society, 147 (7) 2620-2624 (2000)
S0013-4651(99)10-097-1 CCC: $7.00 © The Electrochemical Society, Inc.
overpotential for the ORR strongly depends on the electrolyte and
4. M. K. Ravikumar and A. Shukla, J. Electrochem. Soc., 143, 2601 (1996).
5
.
T. I. Valdez and S. R. Narayanan, in Proton Conducting Membrane Fuel Cells II,
S. Gottesfeld and T. F. Fuller, Editors, PV 98-27, pp. 380, The Electrochemical
Society Proceedings Series, Pennington, NJ (1998).
the nature of the respective anions. In HClO solution, which would
4
be a better match for the situation in a DMFC, we expect the ORR
to shift anodically by 50-100 mV, also in the presence of methanol.
In this case the critical concentration of 10-30 mM methanol deter-
mined in these measurements for a better performance of Ru/Vulcan
6. J.-T. Wang, J. S. Wainright, R. F. Savinell, and M. Litt, J. Appl. Electrochem., 26,
751 (1996).
7
.
R. W. Reeve, P. A. Christensen, A. J. Dickinson, and A. Hamnett, in Proceedings of
the 3rd International Symposium on Electrocatalysis: Advances and Industrial Ap-
plications, S. Hocevar, M. Gaberscek, and A. Pintar, Editors, p. 222, National Insti-
tute of Chemistry, Ljubljana (1999).
(
and the Ru0.92Mo0.08SeO catalyst) as compared to the standard
4
Pt/Vulcan catalyst would be higher in a DMFC, in agreement with
7
recent results.
8. R. Holze, I. Vogel, and W. Vielstich, J. Electroanal. Chem., 210, 277 (1986).
9
.
N. Alonso-Vante, B. Schubert, and H. Tributsch, Mater. Chem. Phys., 22, 281
Conclusions
(1989).
1
1
0. F. J. Rodriguez, P. J. Sebastian, O. Solorza, and R. Perez, Int. J. Hydrogen Energy,
3, 1031 (1998).
1. O. Solorza-Feria, R. Rivera-Noriega, and N. Alonso-Vante, in Proceedings of the
2nd International Symposium on New Materials for Fuel Cell and Modern Battery
Systems, O. Savadogo and P. R. Roberge, Editors, p. 666, Ecole Polytéchnique de
Montréal, Montreal (1997).
We have shown that both the chalcogenide Ru1.92Mo0.08SeO cat-
4
2
alyst and a Vulcan (XC72)-supported Ru catalyst exhibit similar reac-
tion characteristics in the oxygen reduction reaction, which are
markedly different from those of the standard ORR catalyst, Pt/Vul-
can, investigated for comparison. This points to essentially identical
reactive centers in both catalysts, with only minor differences in the
1
2. R. W. Reeve, P. A. Christensen, A. Hamnett, S. A. Haydock, and S. C. Roy, J. Elec-
trochem. Soc., 145, 3463 (1998).
reaction pathway (ORR to H O vs. H O), if any. They are >95%
2
2
2
13. O. Solorza-Feria, K. Ellmer, M. Giersig, N. Alonso-Vante, Electrochim. Acta, 39,
selective toward oxygen reduction to water (4 electron pathway) as
compared to H O formation (2 electron pathway). In the presence of
1647 (1994).
1
4. N. Alonso-Vante, in Proceedings of the 1st International Symposium on New Mate-
rials for Fuel Cells and Modern Battery Systems, O. Savadogo and P. R. Roberge,
Editors, p. 658, École Polytechnique de Montréal, Montréal, Canada (1995).
5. D. Chu and S. Gilman, J. Electrochem. Soc., 141, 1770 (1994).
2
2
0.5 M methanol, the ORR reaction pathway remains unchanged; fur-
thermore, the Ru catalysts are 100% selective toward the ORR, i.e.,
completely methanol tolerant, in contrast to Pt catalysts where the
ORR is accompanied by considerable methanol oxidation. The activ-
ity of the Ru-containing catalysts is, however, low compared to that of
the standard Pt catalyst in methanol-free electrolyte and only above
that of Pt/Vulcan catalysts at methanol concentrations > 10-30 mM.
1
16. H. A. Gasteiger, N. Markovic, P. N. Ross, and E. J. Cairns, J. Electrochem. Soc.,
41, 1795 (1994).
1
1
1
7. N. M. Markovic and P. N. Ross Jr., J. Electrochem. Soc., 141, 2590 (1994).
8. T. J. Schmidt, H. A. Gasteiger, G. D. Stäb, P. M. Urban, D. M. Kolb, and R. J.
Behm, J. Electrochem. Soc., 145, 2354 (1998).
19. U. A. Paulus, T. J. Schmidt, H. A. Gasteiger, and R. J. Behm, J. Electrochem. Soc.,
Submitted.
Acknowledgments
2
0. T. J. Schmidt, H. A. Gasteiger, R. J. Behm, J. Electrochem. Soc., 146, 1296 (1999).
We want to thank J. Giallombardo (E-TEK Inc., Natick, MA) for
providing us with catalyst samples. Financial support came from the
State of Baden-Württemberg, within the joint research project Meth-
anol Fuel Cells; the Ulmer Universitätsgesellschaft; and the Fonds
der Chemischen Industrie.
21. J. Denton, J. M. Gascoyne, and D. Thompsett, Eur. Pat. EP 0 731 520 A1 (1996).
2
2. H. A. Gasteiger, N. Markovic, P. N. Ross, and E. J. Cairns, J. Phys. Chem., 98, 617
1994).
(
2
3. N. Alonso-Vante, M. Fieber-Erdmann, H. Rossner, E. Holub-Krappe, C. Giorgetti,
A. Tadjeddine, E. Dartyge, A. Fontaine, and R. Frahm, J. Phys. IV, 7, C2-C2-889
(1997).
2
4. N. Alonso-Vante, H. Tributsch, and O. Solorza-Feria, Electrochim. Acta, 40, 567
References
(1995).
1
2
.
.
S. Wasmus and A. Küver, J. Electroanal. Chem., 461, 14 (1999).
Proton Conducting Membrane Fuel Cells II, S. Gottesfeld, T. F. Fuller, Editors,
PV 98-27, The Electrochemical Society Proceedings Series, Pennington, NJ
25. T. Romero, O. Solorza, R. Rivera, and P. J. Sebastian, J.New Mater. Electrochem.
Sys. 2, 111 (1999).
26. E. Ticanelli, J. G. Beery, M. T. Paffett, and Gottesfeld, J. Electroanal. Chem. 258,
(1999), and references therein.
61 (1989).
3
.
B. Bittins-Cattaneo, S. Wasmus, B. A. Lopez de Mishima, W. Vielstich, J. Appl.
Electrochem., 23, 625 (1993).
27. A. J. Bard and L. R. Faulkner, Electrochemical Methods, Wiley & Sons, New York
(1980).
Downloaded on 2014-03-26 to IP 134.208.103.160 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).