Preparation and oxygen reduction activity of stable RuSex/C catalyst with pyrite structure

Article abstract of DOI:10.1016/j.electacta.2009.02.081

Carbon supported RuSe_x (x = 0.35-2) catalysts of controlled stoichiometry and phase are synthesized via precipitating ruthenium nanoparticles on Vulcan XC-72R, then selenizing ruthenium with hydrogen annealing. The competition for Se between so

Full text of DOI:10.1016/j.electacta.2009.02.081

Electrochimica Acta 54 (2009) 4297–4304
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Preparation and oxygen reduction activity of stable RuSe /C catalyst
x
with pyrite structure
a
a
a,∗
Ming-Yi Shen , Shun-Ping Chiao , Dah-Shyang Tsai
,
b
a
David P. Wilkinson , Jyh-Chiang Jiang
Department of Chemical Engineering, National Taiwan University of Science and Technology, Taipei 10607, Taiwan
Department of Chemical and Biological Engineering, University of British Columbia, Vancouver, BC, Canada V6T 1Z4
a
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 3 January 2009
Received in revised form 26 February 2009
Accepted 26 February 2009
Available online 9 March 2009
Carbon supported RuSex (x = 0.35–2) catalysts of controlled stoichiometry and phase are synthesized
via precipitating ruthenium nanoparticles on Vulcan XC-72R, then selenizing ruthenium with hydrogen
annealing. The competition for Se between solid-state Ru selenization and volatile selenium hydride
formation results in RuSex nanoparticles with the pyrite structure, the Ru hcp structure, or a mixture of
the two. These catalysts are methanol tolerant, and catalytically active toward oxygen reduction reaction
∼
ORR). RuSex/C (x = 2) with a pyrite structure, produced at 400 C, exhibits catalytic activity and stability
◦
(
Keywords:
Oxygen reduction
Pyrite
Ruthenium selenide
Catalyst stability
Selectivity
superior to that of RuSex/C with a ruthenium hcp structure or a mixed phase. And this pyrite RuSex/C
∼
(
x = 2) catalyst yields H2O2 less than 1.5% in the technically pertinent potential range of 0.4–0.6 V (vs.
◦
Ag/AgCl). Its stability is verified in 100 CV cycles, which shows that the catalyst annealed at 400 C is more
enduring in potential cycling over 0.65 V (vs. Ag/AgCl), compared with the RuSex/C catalyst annealed at
◦
300 C and the RuSecluster/C reference catalyst prepared by thermolysis of Ru3(CO)12 . After CV cycles, the
◦
4
00 C-annealed catalyst still exhibits a higher ORR activity than the other two catalysts.
©
2009 Elsevier Ltd. All rights reserved.
1
. Introduction
plex exhibited a hcp-type structure of Ru core decorated with Se
on the surface, while above that temperature, the complex split
into Ru metal and RuSe2 pyrite [23]. Several optimum Se/Ru val-
ues for high ORR activity level were reported, 0.3 by Zaikovskii et
al. [19], 0.15 by Cheng et al. [24] and Nagabhushana et al. [25],
and 0.6 by Colmenares et al. [20]. However, the stability of these
RuSex nanoparticles seems dubious. Cyclic voltammograms (CVs)
of RuSex/C catalyst with a Ru core were measured with continuous
cycling to 1.2 V (vs. RHE) and the response current was reported
to decrease rapidly due to Se oxidation and partial dissolution
[15,18]. The onset potential of Se stripping from Ru core at 0.85 V
(vs. RHE) was in line with the observation that the dissolution of
Se on Au electrode occurred in the potential range of 0.66–0.99 V
(vs. Ag/AgCl) [26]. The ORR activity was quickly lost as soon as the
Se removal exposed the underlying Ru core that was easily oxi-
dized.
In this work, we prepare the RuSex/C catalyst using a two-step
method with the emphasis on those stable catalysts. The Se/Ru ratio
of RuSex/C was controlled between 0.35 and 2, and the crystal struc-
ture could be Ru metal with an hcp structure, RuSe2 with a pyrite
structure, or a mixture of both. The catalytic activity was analyzed
using linear scanning voltammetry in combination with a thin film
rotating disk electrode (RDE). The catalyst of pyrite structure and
a Se/Ru ratio near 2 exhibits both high ORR activity and improved
stability in the high potential region.
Selectively catalyzing the oxygen reduction reaction (ORR) in
the presence of methanol is a desirable attribute for the cathode
catalyst of a direct methanol fuel cell owing to the methanol cross-
over problem. Four types of methanol tolerant catalysts have been
developed for this purpose, including metal-containing macrocy-
cles [1–4], nitrides or carbides of non-precious metal [5–8], Pd and
Pt-based alloys [9–14], and Rh or Ru chalcogenides [15–22]. Among
the chalcogenides, carbon supported RuSex (RuSex/C) stands out as
one of the most active ORR catalysts.
The active center of RuSex/C catalyst toward the ORR is believed
to be the Ru–Ru bond, which is protected by one or two Se lay-
ers covering the Ru metallic core. Hence the stoichiometric ratio
of Se/Ru is generally controlled to be less than 1. The published
X-ray diffraction (XRD) patterns of these selenium modified Ru cat-
alysts displayed definite signatures of the Ru metallic phase, not
the RuSe2 pyrite structure [17,19]. A recent work on Ru Se (CO)
4
2
11
◦
pyrolysis in helium indicated that, below 430 C, the carbonyl com-
∗
Corresponding author at: Department of Chemical Engineering, National Tai-
wan University of Science and Technology, 43 Keelung Road, Section 4, Taipei 10607,
Taiwan. Tel.: +886 2 27376618; fax: +886 2 27376644.
E-mail address: dstsai@mail.ntust.edu.tw (D.-S. Tsai).
0
013-4686/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2009.02.081

4
298
M.-Y. Shen et al. / Electrochimica Acta 54 (2009) 4297–4304
2. Experimental
Preparation of RuSex/C electrocatalyst involved two steps; the
first step was to synthesize carbon supported Ru nanoparticles
Ru/C) and the second step was to selenize Ru/C in a solid state
reaction. To uniformly disperse the Ru nanoparticles on Vulcan XC-
2R (Cabot), a proper pretreatment was carried out to remove its
(
7
adsorbed species and oxidize the carbon surface. Typically, 20 g of
XC-72R was blended into 500 ml of 65% nitric acid and boiled for 8 h
in a pyrex glass reactor. The slurry was then cooled and its pH value
was adjusted to 7.0 with 35% ammonia water. The pretreated XC-
72R was then filtered and washed with purified water and stored
in a vacuum chamber.
The xylene solvent was dehydrated with 4 Å molecular sieve
for 1 week before use in synthesis. The catalyst synthesis, unless
stated otherwise, was performed under a purified nitrogen or argon
atmosphere. In the preparation of Ru/C, 0.9 g of pretreated XC-72R
◦
was mixed with 400 ml de-aerated xylene and heated to 120 C.
Another hot solution containing 400 ml xylene and 0.9 g Ru (CO)
3
12
◦
was slowly mixed into the XC-72R slurry at 120 C using a drop-
◦
ping funnel. Xylene was removed at 50 C and 90 rpm in a reduced
pressure rotary evaporator (Heidoph Laborota 4000), and the pow-
der mixture was transferred from the evaporator using anhydrous
diethyl ether and stored in a vacuum chamber. The dried mixture
was decomposed and converted to Ru/C catalyst in a tubular fur-
Fig. 1. XRD patterns of RuSex/C-I samples which were annealed in hydrogen at
◦
◦
◦
nace at 300 C and 20 Torr under flowing hydrogen for 100 min. The
300–800 C, also including that of Ru/C annealed at 300 C. The high intensity back-
ground at low angle is the signal of Vulcan carbon support. For reference, diffraction
lines of the JCPDS file 80-0670 of RuSe2 pyrite and the 06-0663 file of Ru metal are
marked at the bottom.
cooled sample of Ru/C was removed under argon protection.
For selenization, a mixture of 0.1 g Ru/C and a calculated amount
of selenium (99.9%, SHOWA) was suspended in de-aerated xylene
◦
using a magnetic stirrer and refluxed at ∼142 C for 12 h. The pow-
of 0.2 mg cm 2, which included Ru and Se, excluding carbon or
residual oxygen. The working electrode was prepared by coating
the slurry of Vulcan supported RuSex, which was ultrasonically dis-
persed in 500 ␮L of a 0.5 wt% Nafion solution (Aldrich). CV curves
−
der of Se covered Ru/C was separated by centrifuging at 20,000 rpm
and removed using anhydrous diethyl ether, and then stored in
a vacuum chamber. Solid state reaction between Ru and Se was
carried out in a tubular furnace with flowing hydrogen at 20 Torr.
Two kinds of RuSex/C catalyst were synthesized. RuSex/C-I was pre-
pared using the powder of Ru/C covered with a large excess amount
−
1
were recorded in 0.5 M de-aerated H SO at 50 mV s , using a mul-
2
4
tichannel potentiostat (Solartron 1470E). Current–potential data for
O2 reduction were collected in a RDE configuration as the potential
∼
◦
of selenium (Se:Ru = 23:1) and annealed at 300–900 C for 1 h in
−
1
was scanned negatively from 0.7 to −0.1 V at 10 mV s and rota-
hydrogen. RuSex/C-II was prepared by annealing Ru/C covered with
◦
tion speeds of 400, 900, 1600, 2500 rpm in an O -saturated 0.5 M
less selenium (Se:Ru ≤ 12:1) at 400 C for 1 h in hydrogen so that the
2
H SO solution. Before the measurement, the electrode was cycled
Se/Ru ratio would be less than 2. Specimens of RuSex/C-I are further
denoted by their annealing temperature, for example, RuSex/C-I500
2
4
−
1
1
0 times at 50 mV s in N -saturated 0.5 M H SO between −0.2
2
2
4
◦
and 0.6 V to produce clean surfaces. Hydrogen peroxide (H O ) pro-
means the sample was annealed at 500 C. Samples of RuSex/C-II are
2 2
duction in the O -saturated 0.5 M H SO was recorded at 1600 rpm
denoted by their Se/Ru ratio measured by energy dispersive X-ray
2
2
4
in a RRDE setup, using a mirror-polished Pt ring electrode biased
at 1.0 V. In the methanol tolerance tests, the O -saturated 0.5 M
(EDX), for example, RuSe1.87/C-II.
X-ray diffraction data were recorded using a diffractometer
Rigaku D/MAX-RC), equipped with a Cu K␣ radiation source.
2
H SO solution also contained 1.0 M methanol. In the catalyst sta-
(
2
4
bility experiments, the working electrode underwent CV scans from
0.15 to 1.0 V in N -saturated 0.5 M H SO for a preset number of
The crystallite size was estimated using the Scherrer formula,
.9ꢀ/(B cos ꢁ), in which ꢀ was the wavelength and B was the full
−
0
2
2
4
cycles, then was placed in O -saturated 0.5 M H SO for measuring
width at half-maximum at the Bragg angle ꢁ. The shape and size of
nanoparticles supported by XC-72R were analyzed and the selected
area diffraction (SAD) patterns were taken for phase identifica-
tion using a transmission electron microscope (FEI Tecnai G2 F20
FEG-TEM). Statistical analysis of particle size distribution was per-
formed with an image analysis software, SigmaScan Pro5.0 (SPSS).
Elemental analysis was carried out using a field emission scanning
electron microscope (JEOL JSM-6500F) at an acceleration voltage
of 10 kV, equipped with an energy dispersive X-ray microanalyzer
2
2
4
−
1
the oxygen reduction current at 10 mV s and 1600 rpm. All sup-
2
porting electrolytes were prepared by diluting concentrated H SO4
with purified water (resistivity > 18 Mꢂ cm, ROpure ST, Barnstead).
3. Results and discussion
3.1. Structure and stoichiometry of RuSex
(
INCA, Oxford). The reported composition was an average value of
Phase and stoichiometry of the RuSex/C catalyst are manipu-
lated through adjusting the selenization conditions. XRD patterns
of RuSex/C-I and that of Ru/C are compiled in Fig. 1, which shows
pyrite-type RuSe2 is the dominant phase in the samples annealed
six different sites in one sample.
◦
Activities of the electrocatalysts were measured at 27 ± 1 C
in a thermostated three-electrode cell which included a rotating
ring-disk electrode R(R)DE (Pine instrument), an Ag/AgCl (satu-
rated KCl) reference electrode, and a platinum counter electrode
with an area of 2 cm × 2 cm. The potential values of this work are
referred to the Ag/AgCl reference electrode, if not specified other-
wise. The working electrode of glassy carbon RDE had a loading
◦
at 350–500 C. These patterns show the four most intense peaks of
◦
◦
◦
◦
RuSe2 (2 0 0), (2 1 0), (2 1 1), (3 1 1) at ∼30 , 34 , 37 , and 51 . The
signatures of ruthenium hcp crystal are evident in the pattern of
Ru/C and also in RuSex/C-I annealed at 700 C and temperatures
above. These patterns show three distinctive Ru peaks at ∼38 , 42 ,
◦
◦
◦

M.-Y. Shen et al. / Electrochimica Acta 54 (2009) 4297–4304
4299
Table 1
Compositions and crystallite sizes of Ru/C, RuSex/C-I and RuSex/C-II catalysts. The crystallite size was calculated using the full width at half the maximum peak height.
Sample
O (atom%)
Se (atom%)
Ru (atom%)
Se/Ru Ratio
Crystallite size (nm)
Ru/C
12.08
6.08
6.26
5.10
4.28
5.15
6.90
2.57
–
4.63
3.28
3.06
2.08
3.68
3.62
4.95
3.36
–
4.2 (Ru)
RuSex/C-I300
RuSex/C-I400 (also RuSe2.04/C-II)
RuSex/C-I500
RuSex/C-I600
RuSex/C-I700
5.67
6.25
4.07
2.87
0.56
0.60
0.26
1.73
2.04
1.96
0.78
0.15
0.12
0.08
3.1 (RuSe2)
5.8 (RuSe2)
8.6 (RuSe2)
12.1 (RuSe2), 4.6 (Ru)
6.9 (Ru)
RuSex/C-I800
RuSex/C-I900
7.7 (Ru)
9.7 (Ru)
RuSe0.35/C-II
RuSe0.86/C-II
RuSe1.48 /C-II
RuSe1.87/C-II
6.51
4.49
4.56
6.74
1.45
2.57
4.00
3.42
4.13
2.98
2.70
1.83
0.35
0.86
1.48
1.87
5.4 (Ru)
5.1 (Ru)
–
4.6 (RuSe2)
◦
and 44 , indexed as Ru (1 0 0), Ru (0 0 2), and Ru (1 0 1); individu-
enize Ru/C and prepared samples of RuSex/C-II with a Se/Ru ratio
0.35–2.0. Fig. 2 shows the XRD patterns of RuSex/C-II catalyst which
was annealed with less Se at 400 C. When the Se/Ru ratio of
◦
ally. The 600 C pattern displays both features of Ru hcp and RuSe
pyrite crystal, as does the 300 C pattern but with weaker inten-
2
◦
◦
sity. For identification, the diffraction lines of Ru hexagonal crystal
and RuSe2 pyrite crystal [27] are marked at the bottom of Fig. 1.
In the 600 C diffraction pattern, the Ru (1 0 1) peak at 44 emerges
RuSex/C-II is near 2, the diffraction results show the pyrite struc-
ture. When the Se/Ru ratio is less than 1, the characteristics of Ru
hcp crystal are more noticeable.
◦
◦
from the diffraction peaks of the pyrite phase, suggesting the RuSe2
Table 1 lists the crystallite sizes of RuSex/C-I and II, estimated
using XRD data. The crystallite size of 300 C-annealed Ru nanopar-
◦
◦
crystal starts to decompose in H at 600 C. Decomposition of RuSe
2
2
is also evidenced by values of the Se/Ru ratio, listed in Table 1. The
ticles is 4.2 nm, while the pyrite crystal of RuSe2 of RuSex/C-I300
is smaller, 3.1 nm. As expected, the crystal size increases with
increasing annealing temperature due to Ostwald ripening. For
pyrite-type RuSe2 crystal, the average size increases with anneal-
◦
Se/Ru ratio is ∼2.0 when the annealing temperature is 400 or 500 C,
and drops below 1.0 when the annealing temperature is equal or
◦
higher than 600 C.
◦ ◦ ◦
ing temperature, 3.1 nm (300 C), 4.2 nm (350 C), 5.8 nm (400 C),
The above results indicate that two reactions were competing for
selenium in hydrogen annealing. One reaction is fixing Se though
solid-state selenization of Ru metal, the other is reducing the Se
◦
◦
◦
8.6 nm (500 C), and 12.1 nm (600 C). Beginning at 600 C, the Ru
crystal size of RuSex/C-I also increases with temperature, 4.6 nm
◦
◦
◦
◦
content via the formation of gaseous H Se. When the annealing
(600 C), 6.9 nm (700 C), 7.7 nm (800 C), and 9.7 nm (900 C). On
the other hand, the crystallite sizes of RuSex/C-II are quite similar.
The Ru crystal size of RuSe0.35/C-II is 5.4 nm, and that of RuSe0.86/C-
II is 5.1 nm. The pyrite crystal size of RuSe1.87/C-II is 4.6 nm, and that
of RuSe2.04/C-II is 5.8 nm.
2
◦
temperature is low, for instance 300 C, the Ru selenization is rela-
tively slow and incomplete. Most of selenium on the carbon support
is driven away as the hydride, and only a small fraction of Se reacts
with Ru metal. When the annealing temperature is in the range of
◦
3
50–500 C, although a large fraction of Se is removed, the Ru sel-
TEM micrographs of RuSex/C provide more details on the parti-
cle shape and size distribution. Bright field (BF) images of Fig. 3(a)
illustrate many equiaxed Ru nanoparticles are well dispersed on a
large carbon particle of the Ru/C specimen. The size distribution of
Ru nanoparticles is a positively skewed curve, indicating the rel-
atively small particles occupy a major fraction in the distribution.
Fig. 3(b) shows the BF image of RuSe0.86/C-II nanoparticles on car-
bon particles. Despite a low Se content (Se/Ru ratio = 0.86), a few
nanoparticles near the center of the enlarged image display the
core–shell feature, which presumably means a Ru core with a sel-
enized shell [19]. The SAD pattern inset shows clear rings of Ru
metal and vague diffractions of pyrite crystal. Many nanoparticles
with core–shell features are found in the image of RuSe1.48 /C-II, as
illustrated in Fig. 3(c). The average particle sizes of RuSex/C-II are
quite similar. The ring pattern of pyrite crystal is more intense in
the SAD pattern of Fig. 3(c). Core–shell features were not found
in BF images of RuSe1.87/C-II and RuSe2.04/C-II, suggesting Ru sel-
enization was complete in these specimens. Fig. 3(d) displays the
TEM image of RuSex/C-I800. Most of the RuSex nanoparticles have
decomposed so that the SAD pattern is dominated by the Ru diffrac-
tion spots. In addition, many of the Ru particles have grown to the
size that their habit crystal planes of hcp symmetry are evident
in the BF image. The particle size distribution in Fig. 3(d), simi-
lar to other particle size distributions, is positively skewed toward
smaller particle sizes.
enization rate is fast enough to retain the stoichiometric amount of
Se, so that the pyrite phase emerges as the major phase. Therefore,
◦
the Se/Ru ratio of 400 and 500 C annealed samples is near 2, in con-
◦
◦
trast to 1.73 for 300 C. When the temperature is 600 C or above,
the pyrite crystal decomposes into Ru metal and gaseous H Se,
2
hence the Se/Ru ratio decreases rapidly. The temperature window
◦
for preparing RuSex/C with a pyrite structure is 350–500 C.
Since high temperature may cause particle coarsening and
reduce catalytic activity, and low temperature may diminish the
◦
catalyst stability, a temperature of 400 C was chosen to sel-
3.2. ORR activity and methanol tolerance
Fig. 2. XRD patterns of RuSex/C-II samples which were prepared by annealing Ru/C
with insufficient amount of Se at 400 C. The Se/Ru ratios measured by EDX are
◦
Fig. 4 shows the prepared RuSex/C catalysts are methanol toler-
marked, as well as the diffraction lines for RuSe2 pyrite and Ru metal.
ant, which is manifested in comparing the ORR currents with and

4
300
M.-Y. Shen et al. / Electrochimica Acta 54 (2009) 4297–4304
Fig. 3. Bright field image and SAD patterns of (a) Ru/C, (b) RuSe0.86/C-II, (c) RuSe1.48 /C-II, and (d) RuSex/C-I800. The particle size distribution of each specimen is also shown.
Note the Ru metallic particles exhibit clear hexagonal facets in the inset of (d).
1
−
−1
without 1.0 M methanol. Two widely different current–potential
curves, as illustrated in Fig. 4(a), indicate that the commercial Pt
catalyst catalyzes both O2 reduction and methanol oxidation. In
addition, the inset shows the Ru/C electrode catalyzes the ORR
selectively, so that its current–potential curve hardly changes in the
presence of methanol [16]. Three pairs of ORR current–potential
curves for RuSex/C-I700, I400, I300, with different particle sizes
and degrees of selenization, are plotted in Fig. 4(b). Evidently, they
are selectively active toward ORR regardless of their differences.
Fig. 4(b) also points out the degree of selenization is an important
factor for the ORR activity. The ORR current of RuSex/C-I400, with
the Se/Ru ratio near 2, is generally higher than the currents of other
RuSex/C-I catalysts. However, the activity of RuSex/C-I400 is still less
current j is expressed as j 1 = j + (Bω1/2)− , in which jk is the
k
1/2
kinetic current in the absence of mass-transfer effects, Bω
the mass-transfer limited current, B the Levich constant and ω the
rotation speed. The experimental value of B can be evaluated as
versus ω . Assuming a
four-electron reaction (n = 4) and a known electrode geometric area
is
−
1
−1/2
the reciprocal slope of the plot of j
2
(A = 0.247 cm ), the theoretical B value for Koutecky–Levich plot is
calculated as 0.62nFADO2 2
/3 1/6
ꢃ
CO2 , where F is the Faraday con-
−
5
2
−1
stant, D the diffusion coefficient of oxygen (1.80 × 10 cm s at
O2
◦
−6
−3
−1
25 C), C the solubility of O in 0.5 M H SO (1.13 × 10 mol cm
O2
2
2
4
◦
2
at 25 C) [28,29], and ꢃ the kinematic viscosity (0.01 cm s ).
The theoretical value of the Levich constant is calculated to be
0.1 mA s1 . For RuSex/C-I and II catalysts, the experimental values
of the Levich constant (the reciprocal slope), ranged from 0.12 to
0.10 mA s1 , and agree reasonably well with the theoretical result.
The mass transfer corrected Tafel plots of kinetic current jk for
RuSex/C are illustrated in Fig. 6. In the potential range of interest
for the ORR (0.54–0.62 V), RuSex/C-I catalysts exhibit two types
of catalytic behavior. RuSex/C-I700 and I800 catalysts, character-
ized by the Ru metallic phase and a low Se/Ru ratio, display a low
ORR activity and a high Tafel slope ∼130 mV. In contrast, RuSex/C-
/2
−
2
than that of commercial Pt/C. At the current density 2.4 mA cm
,
/2
the RuSex/C-I400 catalyst shows a potential difference of ∼0.07 V in
comparison with Pt/C toward oxygen reduction. Fig. 4(c) confirms
the methanol tolerance of RuSex/C-II catalysts (x = 0.35, 1.48, 1.87),
whose particle sizes are similar, but different in phase and degree
of selenization. The ORR current of RuSe1.87/C-II is higher than the
other two.
To separate the mass-transfer effects, the Koutecky–Levich
equation is applied to analyze the current–potential data at dif-
ferent rotation speeds. A typical set of ORR data for RuSex/C-I400
is plotted in Fig. 5, which shows the mass-transfer limited cur-
rent increases with increasing rotation speed. The measured disk
∼
I400 and I500 catalysts with x = 2, characterized with a distinct
pyrite phase, exhibit a high catalytic activity and a lower Tafel
slope ∼90 mV. The jk values of RuSex/C-I300 and I600 contain-
ing both pyrite and Ru metallic phases, are between those of the

M.-Y. Shen et al. / Electrochimica Acta 54 (2009) 4297–4304
4301
Fig. 5. Rotating disk electrode (RDE) polarization curves of RuSex/C-I400 for oxygen
−
1
reduction in O2-saturated 0.5 M H2SO4 at a sweep rate of 10 mV s . The inset shows
Koutecky–Levich plots at 0. 4, 0.3, and 0.2 V.
under refluxing conditions [30], reacting the reduced commercial
Ru/C with Se under refluxing conditions [15,18], and anneal-
ing HSeO3 impregnated carbon supported Ru in H2 [16,19]. The
cluster-type RuSe catalysts are believed to have a Ru metallic
core embedded in Se layers (or matrices). The Se layer could
Fig. 4. Comparison between two current–potential curves for oxygen reduction with
and without the presence of 1 M methanol, measured in O2-saturated 0.5 M H2SO4
−
1
at 10 mV s and 1600 rpm using RDE. Catalysts are (a) Johnson–Matthey 20 wt%
TM
Pt/C (HiSPEC
3000) and Ru/C (the inset), (b) RuSex/C-I300, I400, I700, and (c)
RuSe0.35/C-II, RuSe1.48 /C-II, RuSe1.87/C-II.
separate phases, and the Tafel slope of RuSex/C-I300 and I600 is near
9
0 mV. The kinetic current at 0.6 V is also plotted in the inset, reflect-
ing the influence of annealing temperature. The j values at 0.6 V
k
◦
◦
◦
−2
are 0.78 (300 C), 1.03 (400 C), and 1.17 (500 C) mA cm . Beyond
00 C, j at 0.6 V drops rapidly, since the catalyst loses high surface
◦
5
k
area and selenium protection. The high catalytic activity of pyrite
RuSe2 is reconfirmed in Fig. 6(b), which illustrates the Tafel plot of
RuSex/C-II and the kinetic current at 0.6 V. The catalytic activities
of RuSe1.87/C-II and RuSe2.04/C-II are evidently higher than those of
RuSe0.35/C-II and RuSe0.86/C-II, even though they have similar par-
ticle size. The jk values at 0.6 V for RuSe1.87/C-II and RuSe2.04/C-II
−2
are around 1.0 mA cm
.
3
.3. Stability and H O2 yield
2
Several synthesis routes were reported in the literature, includ-
Fig. 6. Mass transfer corrected Tafel plots of kinetic current for oxygen reduction:
(a) RuSex/C-I and (b) RuSex/C-II.
ing thermolysis of Ru (CO)
and reaction with dissolved Se
3
12

4
302
M.-Y. Shen et al. / Electrochimica Acta 54 (2009) 4297–4304
molysis of Ru (CO) and reaction with Se in xylene at the reflux
3
12
condition for 20 h [30]. Fig. 7 indicates the response currents of
RuSex/C-I400 and RuSex/C-I300 decay slowly in cycling, compared
with the current for RuSecluster/C. Furthermore, a salient attribute
in the CV of RuSex/C-I400 is that the peak potentials for the anodic
and cathodic reactions do not shift as the cycling proceeds. The
result implies that the number of active sites decreases, but the cat-
alytic characteristics remain the same. On the other hand, the CV
cycling of RuSecluster/C catalyst not only reduced the response cur-
rent severely, but also shifted the peak potentials toward those of
Ru/C [15,18]. The CV of RuSex/C-I300, with the degree of Ru seleniza-
tion between RuSex/C-I400 and RuSecluster/C, exhibits some shifting
in peak potentials over 100 cycles. It appears that the unbound sele-
Fig. 7. Stability of RuSex/C-I400 and RuSex/C-I300 catalysts with multiple CV scans
over a broad potential range between −0.15 and 1.0 V. Cyclic voltammograms of (a)
RuSex/C-I400, (b) RuSex/C-I300, and (c) RuSecluster/C as a function of scan number in
−1
de-aerated 0.5 M H2SO4 at 50 mV s
.
be stripped by potential cycling over 0.85 V (vs. RHE) via anodic
+
−
oxidation: Se + 3H O = H SeO + 4H + 4e . Lewera et al. demon-
2
2
3
strated that when the potential was scanned between 0.05 and
.2 V (vs. RHE), the Se layer was removed completely in 30 cycles
1
and the oxygen reduction current decreased rapidly [15,18]. The
ORR activity loss accompanied with Se dissolution at potentials
>
[
0.85 V (vs. RHE) was also confirmed by Papageorgopoulos et al.
31].
In the pyrite crystal structure, a pair of Se atoms is octahe-
drally bonded with Ru atoms and vice versa. Thermally annealed
RuSex/C with a pyrite structure should be more resistant toward
◦
Se dissolution. Figs. 7 and 8 demonstrate that the 400 C-annealed
RuSex/C is more durable in repeated potential cycling to 1.2 V. Two
families of CV curves for RuSex/C-I400 and I300 scanned between
Fig. 8. Variation in mass transfer corrected Tafel plots for the ORR kinetic current
after the indicated number of CV cycles for (a) RuSex/C-I400, (b) RuSex/C-I300, and (c)
RuSecluster/C of Fig. 7. The insets illustrate current–overpotential curves for the ORR,
−1
−
0.15 and 1.0 V at 50 mV s are plotted in Fig. 7(a) and (b), in con-
trast with the CV curves for RuSecluster/C, illustrated in Fig. 7(c).
The RuSecluster/C catalyst (Se/Ru ratio = 0.32) was prepared via ther-
−
1
recorded at 10 mV s and 1600 rpm. The overpotential is defined as the difference
between the electrode potential and the reversible potential of the O2 electrode,
◦
which is 1.225 V at 27 C.

M.-Y. Shen et al. / Electrochimica Acta 54 (2009) 4297–4304
4303
Table 2
−
2
−1
Values of the Tafel slope and the overpotential at 1 mA cm for oxygen reduction in O2-saturated 0.5 M H2SO4 at a scan rate of 10 mV s for RuSex/C-I400, RuSex/C-I300, and
RuSecluster/C. The reversible potential of the O2 electrode is estimated to be 1.225 V (27 C).
◦
RuSex/C-I400
Tafel slope Overpotential at
RuSex/C-I300
(number of CV cycles) (mV)
Tafel slope Overpotential at RuSecluster/C
Tafel slope Overpotential at
−2
−2
−2
(
number of CV cycles) (mV)
1 mA cm (V)
1 mA cm (V)
(number of CV cycles)
(mV)
1 mA cm (V)
0
5
90.3
90.6
94.1
96.1
98.0
98.8
101.8
−0.445
−0.452
−0.457
−0.459
−0.461
−0.470
−0.483
0
5
10
20
40
70
100
92.4
94.5
96.5
101.9
100.9
102.2
106.6
−0.474
−0.475
−0.477
−0.478
−0.476
−0.482
−0.488
0
5
10
20
40
70
100
116.4
122.2
125.5
137.4
146.7
–
−0.484
−0.476
−0.478
−0.494
−0.549
–
1
0
0
0
0
2
4
7
1
00
–
–
1 mA cm 2 for RuSe /C-I400 is the minimum, indicating its high
−
nium of RuSex/C-I300 is less than RuSecluster/C, but slightly more
than that of RuSex/C-I400.
x
ORR activity is enduring.
The influences of potential cycling on ORR activity are also mani-
fested in the variations in the Tafel plot, as illustrated in Fig. 8. Tafel
plots of Fig. 8 are plotted as kinetic ORR current versus reaction
overpotential, which is the difference between the disk electrode
potential and the reversible potential of the O2 electrode [32]. The
kinetic currents were calculated as j = jlimj/(jlim − j), in which j is
Besides the high activity and stability, it is of interest to quan-
tify the relative formation rates of H O (four-electron pathway) and
2
H O (two-electron pathway) for RuSe /C-I400. Fig. 9 presents the
x
2
2
ring currents and the percentage of H O2 yielded when the disk
2
potential changes from 0.7 to −0.1 V. The peroxide percentages were
k
lim
calculated as X (%) = 200(j
ring
/N)/(j + j
ring
/N), in which j
ring
is the cur-
the limiting current of disk electrode at 0.2 V. The current–potential
data for the ORR, as shown in the insets of Fig. 8, were taken after
the electrode had been scanned the indicated number of CV cycles
in Fig. 7. The insets reveal that RuSex/C-I400 and RuSex/C-I300 both
decrease in ORR activity, but the decays are considerably less than
that of RuSecluster/C. The activity of RuSecluster/C is so degraded after
rent of ring electrode, j the current of disk electrode, N (=0.37) the
collection efficiency of the ring electrode supplied by the manufac-
turer. Fig. 9 indicates that both of RuSe /C-I400 and 300 produce
x
H O less than 1.5% in the potential range of 0.4–0.6 V. Since the
2
2
peroxide causes a slow degradation of the polymeric electrolyte, its
two-electron pathway in catalysis is to be avoided. The H O yields
2
2
4
0 cycles that we could not find the limiting current, as shown
of RuSe /C-I400 and 300 catalysts are comparable to those of the
x
in the inset of Fig. 8(c). The degradation is analogous to that of
Ru/Se electrode reported by Cao et al. [18]. It is of interest to note
that a significant decrease in the CV current of RuSex/C-I300 does
not result in the kinetic ORR current decay of proportioned magni-
tude. A relatively small decay in CV current of RuSex/C-I400 causes
more ORR activity loss. Even with more catalytic activity loss, the
mass transfer corrected kinetic current of RuSex/C-I400, Fig. 8(a),
is higher than the corresponding kinetic current of RuSex/C-I300,
as shown in Fig. 8(b). A more quantitative comparison on the Tafel
excellent RuSe /C and Ru-based catalysts reported in the literature
[16,25,33,34].
x
4. Summary
The structure and catalytic activity of RuSex/C, which was pre-
pared from selenization of Ru/C in flowing hydrogen, was studied
using transmission electron microscopy, X-ray diffraction, and
rotating (ring) disk electrode voltammetry. Significantly different
from the literature reports, RuSex/C with a pyrite structure and
Se/Ru ratio near 2 exhibits a high ORR activity and superior sta-
bility, compared with the RuSex/C catalysts with Se/Ru ratio less
than 1 and the RuSecluster/C reference catalyst. The high activity
was verified by contrasting the mass transfer corrected kinetic cur-
rents of various RuSex/C catalysts. Mass transfer corrected Tafel
plots for the kinetic ORR current indicate that the Tafel slope of
−
2
slope and the overpotential value at 1 mA cm for ORR is given in
Table 2 for the three catalysts shown in Fig. 8. The value of Tafel slope
generally increases with increasing number of CV cycles. The Tafel
slope of RuSex/C-I400 is less than that of RuSex/C-I300 and that of
−
2
RuSecluster/C. The absolute value of overpotential at 1 mA cm gen-
erally increases with increasing number of CV cycles, reflecting the
activity degradation. Among the three catalysts of RuSex/C-I400,
RuSex/C-I300, and RuSecluster/C, the absolute overpotential value at
◦
the 400 C-annealed RuSe /C (x > 1) is ∼90 mV, while that for x < 1 is
x
∼
130 mV. The stability after exposure to potentials >0.85 V (vs. RHE)
was demonstrated in Tafel plots of kinetic current after multiple
CV scans over a broad potential range. RuSex/C-I400 outperformed
RuSex/C-I300 and RuSecluster/C in catalyzing the ORR after 100 CV
cycles. Both of RuSex/C-I400 and 300 produce H O2 less than 1.5%.
2
Acknowledgements
The authors thank the National Science Council of Taiwan for
supporting this work through the project of NSC 96-2221-E011-
164-MY3, and providing travel funding in NSC 97-2918-I011-001.
References
[
[
[
1] Y. Lu, R.G. Reddy, Electrochim. Acta 52 (2007) 2562.
2] R. Baker, D.P. Wilkinson, J. Zhang, Electrochim. Acta 53 (2007) 6906.
3] S. Baranton, C. Coutanceau, C. Roux, F. Hahn, J.M. Leger, J. Electroanal. Chem.
577 (2005) 223.
[
[
4] X. Wang, M. Waje, Y. Yan, J. Electrochem. Soc. 151 (2004) A2183.
5] R. Yang, K. Stevens, A. Bonakdarpour, J.R. Dahn, J. Electrochem. Soc. 154 (2007)
B893.
Fig. 9. Fractional yield of hydrogen peroxide in the ORR for RuSex/C-I400 and
RuSex/C-I300. The RRDE measurements were performed in the O2-saturated 0.5 M
H2SO4 electrolyte at 10 mV s and 1600 rpm.
[6] R. Yang, A. Bonakdarpour, E.B. Easton, P. Stoffyn-Egli, J.R. Dahn, J. Electrochem.
Soc. 154 (2007) A275.
−1

4
304
M.-Y. Shen et al. / Electrochimica Acta 54 (2009) 4297–4304
[
7] H. Meng, P.K. Shen, Electrochem. Commun. 8 (2006) 588.
[21] N. Alonso Vante, W. Jaegermann, H. Tributsch, W. Honle, K. Yvon, J. Am. Chem.
Soc. 109 (1987) 3251.
[8] K. Lee, A. Ishihara, S. Mitsushima, N. Kamiya, K. Ota, Electrochim. Acta 49 (2004)
3
479.
[22] N. Alonso Vante, H. Tributsch, Nature (London) 323 (1986) 431.
[23] W. Vogel, P. Kaghazchi, T. Jacob, N. Alonso-Vante, J. Phys. Chem. C 111 (2007)
3908.
[
9] X. Li, Q. Huang, Z. Zou, B. Xia, H. Yang, Electrochim. Acta 53 (2008) 6662.
10] L. Zhang, K. Lee, J. Zhang, Electrochim. Acta 52 (2008) 3088.
[
[
11] M.H. Shao, T. Huang, P. Liu, J. Zhang, K. Sasaki, M.B. Vukmirovic, R.R. Adzic,
[24] H. Cheng, W. Yuan, K. Scott, Electrochim. Acta 52 (2006) 466.
[25] K.S. Nagabhushana, E. Dinjus, H. Bonnemann, V. Zaikovskii, C. Hartnig, G.
Zehl, I. Dorbandt, S. Fiechter, P. Bogdanoff, J. Appl. Electrochem. 37 (2007)
1515.
Langmuir 22 (2006) 10409.
[12] W. Yuan, K. Scott, H. Cheng, J. Power Sources 163 (2006) 323.
[
13] G. Ramos-Sanchez, H. Yee-Madeira, O. Solorza-Feria, Int. J. Hydrogen Energy 33
(
2008) 3596.
[26] M.O. Solaliendres, A. Manzoli, G.R. Salazar-Banda, K.I.B. Eguiluz, S.T. Tanimoto,
S.A.S. Machado, J. Solid State Electrochem. 12 (2008) 679.
[27] JCPDS card no. 06-0663 (Ru metal), JCPDS card no. 80-0670 (RuSe2 pyrite crys-
tal), International Center for Diffraction Data, Newtown Square, PA, 1997.
[28] S. Gottesfeld, I.D. Raistrick, S. Srinivasan, J. Electrochem. Soc. 34 (1987) 1455.
[29] R.N. Itoe, G.D. Wesson, E.E. Kalu, J. Electrochem. Soc. 147 (2000) 2445.
[30] N. Alonso-Vante, Fuel Cells 6 (2006) 182.
[
[
[
14] O. Savadogo, K. Lee, K. Oishi, S. Mitsushima, N. Kamiya, K.I. Ota, Electrochem.
Commun. 6 (2004) 105.
15] A. Lewera, J. Inukai, W.P. Zhou, D. Cao, H.T. Duong, N. Alonso-Vante, A. Wieck-
owski, Electrochim. Acta 52 (2007) 5759.
16] L. Colmenares, Z. Jusys, R.J. Behm, J. Phys. Chem. 111 (2007) 1273.
17] H. Cheng, W. Yuan, K. Scott, Fuel Cells 7 (2007) 16.
[
[
18] D. Cao, A. Wieckowski, J. Inukai, N. Alonso-Vante, J. Electrochem. Soc. 153 (2006)
A869.
[31] D.C. Papageorgopoulos, F. Liu, O. Conrad, Electrochim. Acta 52 (2007) 4982.
[32] G. Lewis, M. Randall, International Critical Tables, vol. 7, McGrawHill, NY,
1930.
[33] L. Liu, H. Kim, J.W. Lee, B.N. Popov, J. Electrochem. Soc. 154 (2007) A123.
[34] A. Racz, P. Bele, C. Cremers, U. Stimming, J. Appl. Electrochem. 37 (2007)
1455.
[
19] V.I. Zaikovskii, K.S. Nagabhushana, V.V. Kriventsov, K.N. Loponov, S.V.
Cherepanova, R.I. Kvon, H. Bonnemann, D.I. Kochubey, E.R. Savinova, J. Phys.
Chem. B 110 (2006) 6881.
[20] L. Colmenares, Z. Jusys, R.J. Behm, Langmuir 22 (2006) 10437.