Promotion by hydrous ruthenium oxide of platinum for methanol electro-oxidation

Article abstract of DOI:10.1016/j.jcat.2010.07.021

Pt-(RuO_xH_y)_m electrocatalysts (m being the atomic Ru/Pt ratio) supported on multi-walled carbon nanotubes, in which amorphous hydrous ruthenium oxide (RuO_xH_y) is the exclusive Ru-containing species, were prepared and comprehensively characterized by X-ray diffraction, X-ray photoelectron spectroscopy, temperature-programmed reduction, thermogravimetric analysis and transmission electron microscopy techniques. Cyclic voltammetry (CV) and chronoamperometry studies of CO stripping and methanol electro-oxidation indicated that the CO tolerance and catalytic activity of Pt improved remarkably by the co-presence of RuO _xH_y. Repeated CV pretreatments in 0.5 M H ₂SO₄ up to potentials higher than 0.46 V (vs. SCE) induced significant dissolution of RuO_xH_y, which changed the RuO_xH_y content, Pt-RuO_xH_y proximity and surface structure of Pt, and consequently altered the electrocatalytic activity of Pt in the final electrode. However, RuO_xH_y dissolution was not identified when the pretreatment potentials was set no higher than 0.46 V. Discussion on the promotional function of RuO _xH_y was made based on the peculiarity of RuO _xH_y as a mixed electron/proton conductor.

Full text of DOI:10.1016/j.jcat.2010.07.021

Journal of Catalysis 275 (2010) 34–44
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Promotion by hydrous ruthenium oxide of platinum for methanol electro-oxidation
Jun-Hong Ma a,b, Yuan-Yuan Feng b, Jie Yu b, Dan Zhao b, An-Jie Wang a, Bo-Qing Xu b,
*
a State Key Laboratory of Fine Chemicals, Dalian University of Technology, Dalian 116012, China
b Innovative Catalysis Program, Key Laboratory of Organic Optoelectronics & Molecular Engineering, Department of Chemistry, Tsinghua University, Beijing 100084, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 19 March 2010
Revised 8 June 2010
Accepted 5 July 2010
Available online 19 August 2010
Pt–(RuOxHy)m electrocatalysts (m being the atomic Ru/Pt ratio) supported on multi-walled carbon nano-
tubes, in which amorphous hydrous ruthenium oxide (RuOxHy) is the exclusive Ru-containing species,
were prepared and comprehensively characterized by X-ray diffraction, X-ray photoelectron spectros-
copy, temperature-programmed reduction, thermogravimetric analysis and transmission electron
microscopy techniques. Cyclic voltammetry (CV) and chronoamperometry studies of CO stripping and
methanol electro-oxidation indicated that the CO tolerance and catalytic activity of Pt improved remark-
ably by the co-presence of RuOxHy. Repeated CV pretreatments in 0.5 M H2SO4 up to potentials higher
than 0.46 V (vs. SCE) induced significant dissolution of RuOxHy, which changed the RuOxHy content,
Pt–RuOxHy proximity and surface structure of Pt, and consequently altered the electrocatalytic activity
of Pt in the final electrode. However, RuOxHy dissolution was not identified when the pretreatment poten-
tials was set no higher than 0.46 V. Discussion on the promotional function of RuOxHy was made based on
the peculiarity of RuOxHy as a mixed electron/proton conductor.
Keywords:
Methanol electro-oxidation
Electrocatalysis
Platinum catalyst
Hydrous ruthenium oxide
Catalyst promoter
Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction
contained actually a substantial fraction of amorphous hydrous
ruthenium oxide (RuOxHy or RuO2ꢀnH2O) by thermogravimetry
Direct methanol fuel cells (DMFCs) are regarded as the most
promising new power sources of the future, especially for mobile
and portable applications. However, one of the main challenges
to make DMFCs commercially feasible is to improve the kinetics
of the methanol oxidation reaction at the anode. The advent of
the so-called bimetallic PtRu catalysts has led to a big step forward
in improving the anode performance of DMFCs [1,2]. Ruthenium in
these PtRu catalysts plays, as an effective promoter to Pt, a crucial
role in the generation of oxygenated surface species (i.e., Ru–OH)
by promoting water dissociation [3]. It is generally understood that
these Ru–OH species are essential in lowering potentials for the
oxidative removal of CO-like poisoning intermediates from the sur-
faces of Pt-based anode catalyst [4–6].
In the last decade, many investigations have been conducted to
uncover the nature of PtRu anode catalysts. A consensus on the
chemical nature or oxidation state of active ruthenium species,
however, has not yet reached. While metallic Ru in PtRu alloy enti-
ties is often referred to as the active ruthenium species [7,8],
ex-situ and in situ physicochemical characterizations have shown
that the Ru component at the surface of a PtRu alloy catalyst is
partially oxidized due to oxygen uptake from the working environ-
ment [9–11]. According to Rolison et al., the as-received commer-
cial PtRu black catalysts (E-TEK and JM), which showed X-ray
diffraction (XRD) patterns consistent with an alloy assignment,
(TG) and X-ray photoelectron spectroscopy (XPS) analyses [12].
In addition, the electrocatalytic activity for methanol electro-oxi-
dation of PtRu blacks rich in RuOxHy was found two orders of mag-
nitude higher, in terms of exchange current density, than their
reduced (Pt0Ru0) and thermally dehydrated (Pt–RuO2) counterpart
catalysts [13]. The superior performance of the RuOxHy-rich PtRu
catalysts was attributed to the uniqueness of RuOxHy as a mixed
electron/proton conductor innately related with a Ru–OH specia-
tion [13]. The very beneficial effect of RuOxHy in PtRu catalyst on
methanol electro-oxidation was thereafter frequently confirmed
experimentally in different laboratories [13,14–20]; notwithstand-
ing, it was mentioned occasionally that the formation of RuOxHy
could be responsible for deterioration of PtRu electrocatalysts
[21–23]. In these earlier studies, however, RuOxHy in PtRu catalysts
was derived either from partial oxidation of the so-called metallic
Ru in a PtRu alloy [13,16,17,20] or from electrochemical proton-
ation of a RuO2 surface [14,18]. In a few cases, the RuOxHy was first
prepared by oxidation of ruthenium halide, followed by a reductive
deposition of Pt to make RuOxHy-containing PtRu samples [15,19],
during which a partial reduction of the preformed RuOxHy to
metallic Ru would be inevitable. Thus, it can be understood that
RuOxHy species in those earlier publications was always ‘‘contam-
inated” with other forms of ruthenium (metallic Ru and/or crystal-
line RuO2), which could significantly distort the signature of
RuOxHy species as a promoter in Pt-based anode catalysts. To bet-
ter understand the promoter function of RuOxHy species, it would
* Corresponding author. Fax: +86 10 62792122.
E-mail address: bqxu@mail.tsinghua.edu.cn (B.-Q. Xu).
0021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.07.021

J.-H. Ma et al. / Journal of Catalysis 275 (2010) 34–44
35
be essential to check with an anode material that clearly excludes
any possible involvement of ruthenium species other than RuOxHy.
adventitious C 1s line at 284.6 eV. XPS spectra of Pt 4f and Ru 3p
were deconvoluted by using a XPS peak 4.1 software. Gaussian–
Lorentzian natural line width and asymmetry parameters were
applied for identifying Pt 4f and Ru 3p components. The Shirley
background and a least-squares routine were used for peak fit
tings. Transmission electron microscopy (TEM) and energy-dispersive
X-ray spectroscopy (EDX) measurements were performed on a
JEM-2010/EDX instrument operating at 120 kV to determine mor-
phology and size distribution of platinum and ruthenium entities.
Temperature-programmed reduction (TPR) experiment was
conducted on a homemade TPR apparatus equipped with a thermal
conductivity detector (TCD). Prior to each measurement, 50 mg of
the catalyst was pretreated in an Ar flow at 298 K for 30 min, fol-
lowed by cooling to 163 K (with an ethanol/liquid nitrogen cold
trap) in flowing Ar. The pretreated sample was then switched to
a flow of 5% H2/Ar. After waiting for a sufficient long period to ob-
tain a stable baseline (the temperature of the cold trap increased to
193 K during this process), the sample temperature was pro-
grammed to ramp at a rate of 10 K minꢃ1 up to 700 K. The gases
were all carefully dried and de-oxygenated to ensure accurate
detection of hydrogen comsumptions. Thermogravimetric analysis
(TGA) was carried out on a Mettler-Toledo TG/SDTA 851e instru-
ment. Samples were loaded into alumina pans and heated from
In this work,
a series of multi-walled carbon nanotubes
(MWCNTs)-supported Pt–(RuOxHy)m (m being the atomic Ru/Pt ra-
tio, m = 0.04–0.40) anode catalysts are carefully prepared to pre-
clude the presence of any metallic Ru or anhydrous RuO2.
Electrochemical characterizations of these catalysts in relation to
methanol electro-oxidation are then conducted to gain insight into
the promotional effect of ‘‘non-contaminated” RuOxHy on Pt cata-
lyst. Our data demonstrate clearly that the ‘‘non-contaminated”
RuOxHy alone can act as a very efficient promoter of Pt for the elec-
tro-oxidation of CO and methanol. We also show that the compo-
sition (atomic Ru/Pt ratio) and pretreatment potentials are of vital
importance to the overall anode performance (activity and stabil-
ity) of these catalysts.
2. Experimental
2.1. Sample preparation
2.1.1. MWCNTs-supported RuOxHy (RuOxHy/MWCNTs)
The support MWCNT materials (i.d. 3–5 nm, o.d. 15–45 nm,
tube length up to 10 lm), prepared by CH4 decomposition over a
373 to 1073 K at 10 K minꢃ1 in flowing dry N2 (50 mL minꢃ1 .
)
Ni–MgO catalyst [24], were subjected to a pretreatment with con-
centrated nitric acid to introduce oxygen-containing functional
groups [25,26]. RuOxHy/MWCNT (Ru loading: 0.5–5.1 wt.%) sam-
ples were then prepared by the oxidative precipitation method
according to Fu et al. [27]. In a typical procedure, 0.40 g MWCNTs
was ultrasonically dispersed for 2 h in 25 mL of 5 mM RuCl3 aque-
ous solution. Subsequently, 30 mL of aqueous H2O2 solution (30%)
was added dropwise at room temperature under vigorous stirring.
The obtained suspension was further agitated for 8 h at 353 K, fol-
lowed by separation of the precipitates by filtration and extensive
washing with de-ionized water. The obtained solid sample was
then dried overnight at 383 K to give RuOxHy/MWCNTs. The filtrate
was collected and analyzed by inductively coupled plasma mass
spectrometry (ICP-MS) to determine the actual loading of RuOxHy.
A reference RuO2/MWCNT sample was obtained by calcining the
as-prepared RuOxHy/MWCNTs (Ru loading: 1.3 wt.%) in flowing
dry N2 at 773 K for 2 h.
2.3. Electrochemical measurements
A three-electrode measurement system was employed for elec-
trochemical tests. The working electrode was a piece of carbon
paper (Toray, TGPH-120) covered with a thin layer of Nafion-
impregnated catalyst. In a standard procedure, the catalyst inks
were prepared by mixing 5 mg catalyst sample with 0.25 mL iso-
propanol, followed by addition of 20 lL 5 wt.% Nafion solution (Al-
drich) and ultrasonic dispersion of the mixture for 15 min. The
obtained ink was then carefully coated onto a piece of carbon pa-
per (0.5 cm ꢄ 2.0 cm) to cover an area of 0.5 cm ꢄ 0.5 cm and dried
under an infrared lamp. The overall catalyst loading was controlled
at 1.0 0.1 mg. Saturated calomel electrode (SCE) and a 1.0 cm ꢄ
1.0 cm Pt foil were used as the reference and counter electrodes,
respectively. All experiments were done at room temperature in
N2-saturated solutions. The potentials were given with respect to
SCE unless otherwise stated. Prior to the electrochemical measure-
ment, the electrode catalyst was pretreated by repeated cyclic
voltammetry (CV) scanning in 0.5 M H2SO4 for 60 cycles at a scan
rate of 50 mV sꢃ1, in either a narrow (ꢃ0.20 to 0.46 V) or an
extended (ꢃ0.20 to 0.96 V) potential range.
2.1.2. Pt–(RuOxHy)m/MWCNTs
Pt–(RuOxHy)m/MWCNT samples were made by Pt deposition onto
RuOxHy/MWCNTs with a modified colloidal approach. Briefly, a Pt sul-
fite complex synthesized from K2PtCl6 and NaHSO3 [28] was dissolved
in de-ionized water at pH ꢁ3.0, and RuOxHy/MWCNTs and ethanol
were then added to form a suspension. Under vigorous stirring and
careful maintenance of pH (ꢂ3.0) with NaOH solution, H2O2 was
added dropwise at room temperature to oxidatively decompose the
sulfite complex, and the resulting suspension was further refluxed
for 2 h to deposit Pt nanoparticles on RuOxHy/MWCNTs. After filtra-
tion and extensive washing, the solids were finally dried at 383 K to
give the Pt–(RuOxHy)m/MWCNT catalyst. A similar preparation using
MWCNTs instead of RuOxHy/MWCNTs produced the reference Pt/
MWCNT sample. The Pt loading in Pt–(RuOxHy)m/MWCNTs as con-
firmed by ICP-MS analysis was 20 wt.%. The actual Ru loadings in
these samples are 0.4, 1.0, 2.1, 3.1 and 4.1 wt.% with m = 0.04, 0.10,
0.20, 0.30 and 0.40, respectively.
CO stripping experiment was carried out in 0.5 M H2SO4. The
electrolyte solution was initially purged with high-purity dry N2.
Adsorption of CO on the electrode catalyst was conducted by bub-
bling high purity CO through the electrolyte solution for 15 min,
followed by purging with N2 for 20 min to expel residual CO out
of the solution. The CO stripping CV curves were obtained in the
potentials between ꢃ0.20 and 0.96 V with a scanning rate of
20 mV sꢃ1. CV and chronoamperometry (CA) tests of methanol
electro-oxidation reaction were performed in 0.5 M H2SO4 +
0.5 M CH3OH solution. The stable CV curves were recorded with
a scanning rate of 20 mV sꢃ1. All the electrochemical tests were
carried out with a potentiostat/galvanostat model 263A controlled
by PowerSuite software (Princeton Applied Research).
2.2. Physicochemical characterizations
3. Results
XRD patterns were recorded on a Bruker D8-Advance X-ray dif-
fractometer using a Cu K
40 kV and 40 mA with a scan step of 0.02 deg. XPS spectra were ac-
quired with a PHI 5300 ESCA1610 SAM instrument using Mg K
radiation. The binding energies (BE) were calibrated using the
a radiation (k = 0.15406 nm) operating at
3.1. Characterization of ruthenium entities and platinum dispersion
a
Fig. 1 shows the XRD patterns of Pt/MWCNTs and Pt–(RuOx-
Hy)m/MWCNTs (m = 0.10, 0.20 and 0.40). The as-prepared

36
J.-H. Ma et al. / Journal of Catalysis 275 (2010) 34–44
ing: 5.1 wt.%), which features diffractions associated only with gra-
phitic carbons and hints that ruthenium species in this zero-
platinum RuOxHy/MWCNT sample exists also as amorphous
RuOxHy.
The effect of thermal treatment on the phase composition of Pt–
(RuOxHy)m/MWCNTs was exemplified by Pt–(RuOxHy)0.40/MWCNTs
before and after the thermal treatment in flowing N2 at 773 K
(Fig. 1d and e). In association with the intensity increase in the
Pt peaks, two new diffractions at 2h = 28° and 35° became evident
in the XRD pattern of the thermal-treated sample (Fig. 1e). Consis-
tent with the tetragonal RuO2 crystals formed from dehydration of
amorphous RuOxHy [29], these new peaks point to the existence of
RuOxHy in the as-prepared Pt–RuOxHy samples, as would have been
expected.
The surface composition and chemical states of Pt and Ru in
the as-prepared Pt–(RuOxHy)m/MWCNTs were characterized using
XPS (Fig. 2). The Pt 4f signals can be deconvoluted into three
doublets, among which the most intensive doublet with BE at
71.3 and 74.6 eV was a signature of Pt (0), while the other two
doublets (72.5 and 75.8 eV, 74.1 and 77.4 eV) were signatures
of Pt (II) and Pt (IV), respectively. A semi-quantitative analysis,
based on the integration areas of Pt in varied valence states,
showed that around 70% of Pt was in its metallic state for all
the investigated Pt–(RuOxHy)m/MWCNT samples. The Ru 3p sig-
nals of Pt–(RuOxHy)m/MWCNTs can be deconvoluted into two
peaks at 463.1 and 466.0 eV. These signals remarkably deviated
from those of Ru (0) (461.1 eV) and anhydrous RuO2 (462.4 eV)
[30] but match the XPS characteristics of hydrous ruthenium
oxide (RuOxHy) which was reported to take on a very broad peak
showing BE higher than that of its anhydrous form [12]. What
really matters to this work is that the XPS spectra for Ru 3p
clearly revealed an absence of Ru (0) in any Pt–(RuOxHy)m/
MWCNT catalysts. The inset of Fig. 2B–c shows the XPS spectrum
of Ru 3p for the reference sample RuOxHy/MWCNTs (Ru loading:
5.1 wt.%), which confirms that ruthenium species in this reference
Fig. 1. XRD patterns of (a) Pt/MWCNTs, (b) Pt–(RuOxHy)0.10/MWCNTs, (c) Pt–
(RuOxHy)0.20/MWCNTs and (d) Pt–(RuOxHy)0.40/MWCNTs. (e) XRD pattern of Pt–
(RuOxHy)0.40/MWCNTs after thermal treatment in flowing N2 at 773 K. The inset
shows the XRD pattern of a RuOxHy/MWCNTs (Ru loading: 5.1 wt.%) without Pt.
Pt–(RuOxHy)m/MWCNT samples (Fig. 1b–d) showed diffractions
associated with Pt and graphitic carbons; diffractions assignable
to the hexagonal close-packed (hcp) structure of metallic Ru or
anhydrous tetragonal RuO2 were not detected. These features were
independent of the atomic Ru/Pt ratio (m) of the Pt–(RuOxHy)m/
MWCNT sample. The diffractions characteristic of the face-cen-
tered cubic (fcc) crystallites of Pt in the Pt–(RuOxHy)m/MWCNT
samples (Fig. 1b–d) were basically unshifted in comparison with
those in the Pt/MWCNT (i.e., m = 0) sample (Fig. 1a). The average
crystallite sizes of Pt, calculated according to Debye–Scherrer
equation, were about 2.7 nm in both Pt/MWCNT and Pt–(RuOx-
Hy)m/MWCNT samples. These observations suggest that ruthenium
species in these as-prepared Pt–(RuOxHy)m/MWCNT samples most
likely exists as an amorphous phase (i.e., amorphous RuOxHy). The
inset of Fig. 1 gives the XRD pattern of RuOxHy/MWCNTs (Ru load-
Fig. 2. Pt 4f (A) and Ru 3p (B) XPS spectra of (a) Pt–(RuOxHy)0.10/MWCNTs, (b) Pt–(RuOxHy)0.20 /MWCNTs and (c) Pt–(RuOxHy)0.40/MWCNTs. The inset shows the Ru 3p XPS
spectrum of RuOxHy/MWCNTs (Ru loading: 5.1 wt.%) without Pt.

J.-H. Ma et al. / Journal of Catalysis 275 (2010) 34–44
37
Fig. 4. TG curves of (a) MWCNTs and (b) Pt–(RuOxHy)0.40/MWCNTs. (b-a) Shows the
calibrated weight loss due to the thermal decomposition of oxidic ruthenium
species (RuOxHy).
ably lower than the reduction temperature of crystalline RuO2 in
RuO2/MWCNTs (Fig. 3d). These results might suggest that our Pt–
(RuOxHy)m/MWCNT samples were made free of crystalline RuO2.
Calibration of hydrogen consumption for the reduction of ruthe-
nium species in RuOxHy/MWCNTs and Pt–(RuOxHy)m/MWCNTs dis-
closed a stoichiometry of H/Ru (molar) ꢁ 3.5, which is lower than
the reduction stoichiometry (H/Ru = 4.1) of RuO2 in RuO2/MWCNTs
(Table 1). Therefore, both the peak temperature and hydrogen con-
sumption data of our TPR experiments point to the fact that ruthe-
nium species in our as-prepared samples is in a mixed valence
state between +3 and +4, in accordance with a RuOxHy (or RuO2ꢀn-
H2O) speciation [33,34].
TGA experiment was conducted on Pt–(RuOxHy)0.40/MWCNTs to
measure the content of structural water in RuOxHy (RuO2ꢀnH2O)
(Fig. 4). To exclude the influence of functional groups on the sur-
face of MWCNTs, a calibrated TG curve (Fig. 4b-a) was obtained
by subtracting the weight loss of an equivalent MWCNTs
(Fig. 4a) from that of Pt–(RuOxHy)0.40/MWCNTs (Fig. 4b), and the
corrected value was used to determine the weight loss during
the thermal dehydration of RuOxHy. The result revealed a water
content of 3.1 wt.% for RuOxHy in Pt–(RuOxHy)0.40/MWCNTs, which
indicates that the hydrous ruthenium oxide can be expressed as a
Fig. 3. TPR profiles of (a) MWCNTs, (b) Pt/MWCNTs, (c) RuOxHy/MWCNTs (Ru
loading: 1.3 wt.%), (d) RuO2/MWCNTs (Ru loading: 1.3 wt.%) and (e) Pt–
(RuOxHy)0.10/MWCNTs. The inset shows an enlargement of profile (a).
sample was also characteristic of RuOxHy, same as those in the
Pt–(RuOxHy)m/MWCNT samples.
Fig. 3 shows the H2-TPR profiles of MWCNTs, Pt/MWCNTs,
RuOxHy/MWCNTs (Ru loading: 1.3 wt.%), Pt–(RuOxHy)0.10/MWCNTs
and RuO2/MWCNTs (Ru loading: 1.3 wt.%). The peak temperature
and H2 consumption data are listed in Table 1. A small but evident
peak at about 208 K was observed for every sample in Fig. 3, which
was an artifact caused by the ill-controlled temperature increase at
the very beginning of the temperature ramp (i.e., in the initial sev-
eral minutes following removal of the cold trap). The MWCNT sup-
port showed no hydrogen consumption up to 430 K, above which
hydrogen consumption due to reduction of oxygen-containing sur-
face species such as carboxylic groups on MWCNTs [31] and
methanation of carbon became observable (as shown by the inset
of Fig. 3). The peak at 237 K on the TPR profile of Pt/MWCNTs was
assigned to the reduction of PtOx; a similar peak was also observed
earlier in the TPR of Pt/C oxidized at 300–570 K [32]. The TPR pro-
file of RuOxHy/MWCNTs (Fig. 3c) showed two unresolved peaks at
357 and 395 K, different from the TPR profiles of RuOxHy in Refs.
[19,32] where only one peak at about 370 K was recorded. The
unresolved two peaks may arise from inhomogeneous composition
of the materials, which would result in varied interactions between
RuOxHy and MWCNTs. However, the temperature for the reduction
of RuOxHy in RuOxHy/MWCNTs was distinctly lower than that of
RuO2 in the counterpart RuO2/MWCNT sample (Fig. 3d).
RuO2ꢀnH2O of n = 4.1 or RuO6.1H8.2
.
Fig. 5 shows the representative TEM images of two RuOxHy/
MWCNT (Ru loading: 1.3 wt.% and 5.1 wt.%) samples and their
Pt-loaded counterparts, Pt–(RuOxHy)0.10/MWCNTs and Pt–(RuOx-
Hy)0.40/MWCNTs. A representative TEM image of the ruthenium-
free sample Pt/MWCNTs was also included for reference. RuOxHy
particles in RuOxHy/MWCNTs containing 1.3 wt.% Ru were uni-
formly dispersed on the support MWCNT material, with sizes
around 1–2 nm (Fig. 5A). However, the RuOxHy particles in the
other RuOxHy/MWCNTs containing 5.1 wt.% Ru were somewhat
aggregated (Fig. 5B). Consequently, the Pt–RuOxHy particles in
Pt–(RuOxHy)0.10/MWCNTs appeared also uniformly dispersed, their
small sizes (2–3 nm) seen in the TEM images (e.g., Fig. 5C) agreed
well with the XRD crystallite size for metallic Pt. In contrast, the
The TPR profile of Pt–(RuOxHy)0.10/MWCNTs (Fig. 3e) exhibited
two peaks centered at 233 and 374 K, which might be attributed
to the reduction of PtOx and amorphous RuOxHy, respectively.
The temperature for the reduction of RuOxHy in this sample agreed
with that in the Pt-free RuOxHy/MWCNTs (Fig. 3c) but was remark-
Table 1
H2 consumption and peak temperatures from TPR experiments.
Sample
Metal loading
Peak temperature (K)
H/Pt (molar ratio)
H/Ru (molar ratio)
Pt/MWCNTs
RuOxHy/MWCNTs
RuO2/MWCNTs
20.0 wt.% Pt
1.3 wt.% Ru
1.3 wt.% Ru
20.0 wt.% Pt, 1.0 wt.% Ru
237
0.6
–
–
–
357, 395
444, 500
233, 374
3.5
4.1
3.6
Pt–(RuOxHy)0.10/MWCNTs
0.6

38
J.-H. Ma et al. / Journal of Catalysis 275 (2010) 34–44
Fig. 5. Representative TEM images of (A) RuOxHy/MWCNTs (Ru loading: 1.3 wt.%), (B) RuOxHy/MWCNTs (Ru loading: 5.1 wt.%), (C) Pt–(RuOxHy)0.10/MWCNTs, (D) Pt–
(RuOxHy)0.40/MWCNTs and (E) Pt/MWCNTs. Samples C and D were obtained, respectively, by loading samples A and B with Pt.
Pt–RuOxHy particles in Pt–(RuOxHy)0.40/MWCNTs appeared as
ill-dispersed aggregates (Fig. 5D), showing a significantly larger
mean particle size (3.8 0.7 nm). These data suggest that the dis-
persive state of Pt–RuOxHy particles in Pt–RuOxHy/MWCNTs is clo-
sely related with the dispersion of the preloaded RuOxHy particles,
possibly due to strong interaction of Pt with RuOxHy. Moreover, the
Pt particles of Pt/MWCNTs (Fig. 5E) were well-dispersed on the
carbon support, and their mean particle size (2.8 0.5 nm) was
in good agreement with the XRD-derived crystallite size for metal-
lic Pt. The above results indicate that the content of RuOxHy has a
noticeable effect on the resulting Pt–RuOxHy particle size in the
CV curves in 0.5 M H2SO4 after such pretreatments. In a fixed pre-
treatment potential range, the capacitive currents of RuOxHy/
MWCNT electrodes after the pretreatment in the methanol-con-
taining electrolyte (Fig. 7b and d) were always lower than those
obtained after the pretreatment in the methanol-free electrolyte
(Fig. 7a and c). The lower capacitive currents after the pretreat-
ments in 0.5 M H2SO4 + 0.5 M CH3OH could be accounted for by a
loss of structural water in RuOxHy, due to methanol/water ex-
change [37].
The pretreatment potentials imposed an even larger impact on
the capacitive current. Compared with the cases of pretreatment in
the narrow potential range (Fig. 7a and b), the capacitive currents
decreased severely after the pretreatment in the extended poten-
tial range; this was regardless of the absence (Fig. 7c) or presence
of methanol in the electrolyte (Fig. 7d). These results provided a
qualitative and clear evidence for RuOxHy dissolution during the
anodic scanning at high potentials (0.46–0.96 V).
Therefore, an estimation of the ruthenium dissolution percent-
age was made according to the relative capacitance recorded in the
potential range of 0.15–0.75 V, the results of which are given in
Table 2. ICP-MS analysis of ruthenium dissolved in the electrolyte
produced similar numbers, which directly confirmed ruthenium
dissolution during the pretreatment in the extended potential
range. It is noteworthy, in reference to the CV curve of the
´
Pt–RuOxHy catalysts. This is in contrast to the results of Radmilovic
et al. who showed that the amount of Ru in PtRu alloy catalysts had
little effect on the sizes of PtRu particles [35].
The Pt–(RuOxHy)m/MWCNT samples were also subjected to EDX
analysis in the TEM mode. Although in most of the areas examined
the EDX signals of platinum were accompanied with those of
ruthenium, areas signifying only Pt were sometimes indeed de-
tected. Such platinum only (or ruthenium silent) areas seemed to
decrease with increasing the Ru/Pt ratio (i.e., m).
3.2. Electrochemical performance in half-cell measurements
3.2.1. RuOxHy/MWCNTs
Fig. 6 shows the CV curves of RuOxHy/MWCNTs containing
3.9 wt.% Ru in 0.5 M H2SO4 with or without methanol. A pair of
broad peaks at 0.40 V (forward scan) and 0.30 V (backward scan),
which is indicative of a capacitive behavior associated with the re-
dox and proton conduction properties of RuOxHy [36], was ob-
served in the absence of CO and methanol (Fig. 6a). This
capacitive behavior remained unchanged in the CO stripping vol-
tammogram (Fig. 6b). The voltammogram recorded in the presence
of 0.5 M CH3OH (Fig. 6c) showed an oxidation current only at
potentials above 0.85 V; in the lower potentials, it appeared the
same as the blank CV (Fig. 6a). These results demonstrate that
RuOxHy/MWCNTs itself is not active for the electro-oxidation of
either CO or methanol.
The RuOxHy/MWCNT (Ru loading: 3.9 wt.%) electrode was pre-
treated by repeated CV scanning for 60 cycles in 0.5 M H2SO4 and
0.5 M H2SO4 + 0.5 M CH3OH, respectively, in either narrow (ꢃ0.20
to 0.46 V) or extended (ꢃ0.20 to 0.96 V) potential range for under-
standing the electrochemical stability of RuOxHy. Fig. 7 shows the
Fig. 6. Cyclic voltammograms of RuOxHy/MWCNTs (Ru loading: 3.9 wt.%) in 0.5 M
H2SO4, (a) blank, (b) after CO adsorption (CO stripping curve) and (c) in the presence
of 0.5 M CH3OH.

J.-H. Ma et al. / Journal of Catalysis 275 (2010) 34–44
39
EAS data were consistently larger than those of their counterpart
electrodes pretreated in the narrow potential range (Fig. 8A),
especially at m P 0.20. These observations indicate an increased
recovery of active Pt surface and are compatible with the RuOxHy
dissolution during the pretreatment in the extended potential
range.
In separate experiments, the electrolyte solutions after the pre-
treatment in extended potential range were sampled and analyzed
by ICP-MS to quantitatively check the dissolution of RuOxHy from
the two Pt–(RuOxHy)m/MWCNT electrodes of m = 0.10 and 0.20.
The amount of ruthenium detected in the electrolyte corresponded
to ca. 70% and 55%, respectively.
A referee of this present paper was wondering if the loss of
ruthenium occurred evenly throughout a pretreatment or mainly
at the beginning and if ruthenium loss would continue to occur
during methanol oxidation. He/she kindly advised us to provide
information about kinetics of the ruthenium dissolution at varying
potentials applied. The Pt–(RuOxHy)0.20/MWCNT electrode was
cycled accordingly in 0.5 M H2SO4 solution for up to 100 cycles
with potential up-limits at Ea = 0.46, 0.66 and 0.96 V, respectively.
Fig. 9 shows the kinetics of RuOxHy dissolution by correlating the
relative content of ruthenium with the number of potential sweep
cycles. It is apparent that the rate of ruthenium dissolution de-
pends greatly on the potential up-limit and the number of cycles.
No ruthenium dissolution was detected when the potential up-
limit was set no higher than Ea = 0.46 V. The dissolution became
increasingly significant when the electrode was cycled to higher
potentials; the total loss of ruthenium was more than 30% and
50%, respectively, when the potential up-limit was increased to
Fig. 7. Cyclic voltammograms of RuOxHy/MWCNTs (Ru loading: 3.9 wt.%) in 0.5 M
H2SO4 after pretreatments in the narrow potential range (ꢃ0.20 to 0.46 V) in 0.5 M
H2SO4 (a) without or (b) with the presence of 0.5 M CH3OH, and in the extended
potential range (ꢃ0.20 to 0.96 V) in 0.5 M H2SO4 (c) without or (d) with the
presence of 0.5 M CH3OH. (e) Shows the reference cyclic voltammogram for a
‘‘pure” MWCNT support.
MWCNTs-based electrode (Fig. 7e), that the broad capacitive fea-
ture characteristic of RuOxHy in the RuOxHy/MWCNT electrode
always remained evident after the different pretreatments. This
is strong evidence supporting that the nature of RuOxHy species
in the RuOxHy/MWCNT electrode did not change during the
pretreatments.
3.2.2. Pt–(RuOxHy)m/MWCNTs
Fig. 8 shows the CV curves of Pt–(RuOxHy)m/MWCNT electrodes
in 0.5 M H2SO4. The voltammetric features in the potential range
between –0.20 and 0.10 V, usually referred to as ‘‘hydrogen
adsorption/desorption” peaks, were widely different for the inves-
tigated catalysts. Three hydrogen adsorption/desorption peaks
characteristic of polycrystalline Pt surfaces [38] were clearly ob-
served on the reference Pt/MWCNT electrode (i.e., m = 0). These
peaks became less resolved, and the electrochemically active sur-
face area of Pt (EAS, derived from the hydrogen desorption peaks)
decreased with the increase in m (or RuOxHy loading) in Pt–(RuOx-
Hy)m/MWCNT electrodes (Fig. 8, Table 3), indicative of a decrease in
the fraction of exposed Pt atoms. For Pt–(RuOxHy)m/MWCNT elec-
trodes subjected to the pretreatment in the extended potential
range, the hydrogen adsorption/desorption features changed to ap-
proach those on the reference Pt/MWCNTs (Fig. 8B). Their derived
Table 2
Effect of pretreatment potentials on the capacitive property of RuOxHy/MWCNTs (Ru
loading: 3.9 wt.%).
Pretreatment Pretreatment
C(RuOxHy/ C(RuOxHy)b CL(RuOxHy)c
potential
range
electrolyte
MWCNTs)a (%)
(%)
ꢃ1
(F gcat
)
0.20–0.46 V 0.5 M H2SO4
0.20–0.96 V 0.5 M H2SO4
136
100
82
66
0
0.5 M H2SO4 + 0.5 M CH3OH 119
18
34
56
104
0.5 M H2SO4 + 0.5 M CH3OH 83
44
a
Mass-specific capacitance of RuOxHy/MWCNT sample. The mass-specific
capacitance of MWCNTs obtained in Fig. 7e is 45 F gꢃ1
.
b
Percentage of RuOxHy evaluated by the capacitance.
Percentage of RuOxHy loss after pretreatment specified; the loss of capacitance
Fig. 8. Cyclic voltammograms of Pt–(RuOxHy)m/MWCNTs in 0.5 M H2SO4 after the
pretreatments in (A) narrow (ꢃ0.20 to 0.46 V) and (B) extended (ꢃ0.20 to 0.96 V)
potential ranges.
c
due to CH3OH/H2O exchange was included also in the RuOxHy loss.

40
J.-H. Ma et al. / Journal of Catalysis 275 (2010) 34–44
Table 3
Electrochemically active surface area, and activity for CO and methanol electro-oxidation of Pt in Pt–(RuOxHy)m/MWCNT catalysts after the pretreatment in the narrow and
extended potential ranges.
ma
Pretreatment (ꢃ0.20 to 0.46 V)
Pretreatment (ꢃ0.20 to 0.96 V)
ꢃ1
c
d
ꢃ1
ꢃ1
c
d
f
ꢃ1
EASb (m2 gPt
)
Eo (V)
Ep (V)
MSAe (A gPt
)
IAg (A mꢃ2
)
EASb (m2 gPt
)
Eo (V)
Ep (V)
If/Ib
MSAe (A gPt
)
IAg (A mꢃ2
)
0
66.0
56.4
43.6
29.3
22.1
14.7
0.36
0.30
0.25
0.24
0.27
0.33
0.60
0.52
0.48
0.43
0.47
0.51
3.5
20.4
30.9
17.0
8.8
0.05
0.36
0.71
0.58
0.40
0.20
69.0
59.6
49.2
41.0
36.3
26.2
0.44
0.35
0.28
0.26
0.22
0.25
0.60
0.60
0.59
0.58
0.54
0.56
0.59
0.63
0.69
0.69
0.68
0.71
13.2
5.9
15.8
18.3
13.6
7.7
0.19
0.10
0.32
0.45
0.37
0.29
0.04
0.10
0.20
0.30
0.40
2.9
a
b
c
d
e
f
Atomic Ru/Pt ratio of as-prepared Pt–(RuOxHy)m/MWCNTs.
Electrochemically active surface area.
Onset potential (Eo) for CO stripping in 0.5 M H2SO4 electrolyte.
Peak potential (Ep) for CO stripping in 0.5 M H2SO4 electrolyte.
Mass-specific activity obtained from the chronoamperometry measurement at 0.4 V (vs. SCE) at the time of 60 min (see Fig. 11).
Ratio of the anodic peak currents in the forward (If) and the reverse (Ib) scans of methanol electro-oxidation.
Intrinsic activity obtained from the chronoamperometry measurement at 0.4 V (vs. SCE) at the time of 60 min (see Fig. 11).
g
Ea = 0.66 and 0.96 V. With either of these two potential up-limits,
the dissolution of RuOxHy was very fast during the first 30 cycles
and then slowed down gradually on further increasing the sweep
cycles up to the 60th cycle. No further ruthenium dissolution
was detected later, which indicates that the electrode catalyst
was basically stabilized for the electrochemical study, as evidenced
by the invariant CV curves after the 60th cycle (data not shown
here). That is the basis why our pretreatments to the electrodes
were ‘‘standardized” by conducting 60 cycles in the narrow
(ꢃ0.20 to 0.46 V, or Ea = 0.46 V) and extended (ꢃ0.20 to 0.96 V,
or Ea = 0.96 V) potential ranges.
Fig. 10 shows the CO stripping voltammograms on Pt–(RuOx-
Hy)m/MWCNT catalysts, their onset (Eo) and peak potentials (Ep)
for the CO electro-oxidation are given in Table 3. Though the num-
bers of Eo and Ep were affected by the value of m, they were consis-
tently lower on the catalysts containing RuOxHy than on the
reference ruthenium-free catalyst (m = 0). When the pretreatment
was conducted in the narrow potential range (Fig. 10A), even a
very small amount of RuOxHy (e.g., m = 0.04) could effect substan-
tial lowering of Eo and Ep for the electro-oxidation of adsorbed CO.
As is seen in Fig. 10A, both Eo and Ep shifted continuously toward
lower potentials on increasing the amount of RuOxHy up to
m = 0.20. Specifically, the numbers of Eo (0.24 V) and Ep (0.43 V)
for CO oxidation on Pt–(RuOxHy)0.20/MWCNTs were lowered by
0.12 and 0.17 V, respectively, relative to those on the reference
Pt/MWCNT electrode (see also Table 3). However, the Eo and Ep
were shifted back toward some higher potentials on further in-
crease in the amount of RuOxHy up to m = 0.40, when compared
with the case of m = 0.20.
The CO stripping voltammograms shown in Fig. 10B were ob-
tained on the samples after the pretreatment in the extended po-
tential range. Compared with their counterparts pretreated in the
narrow potential range (see Fig. 10A), the promotional effect of
RuOxHy seemed to be significantly ‘‘weakened”. The Eo and Ep of
the CO stripping peaks were apparently positive-shifted (Table
3), becoming more close to those on the reference Pt/MWCNT elec-
trode when the amount of RuOxHy before the pretreatment was
low (i.e., m < 0.20). These phenomena were in consistent with the
Fig. 9. Dissolution of ruthenium from Pt–(RuOxHy)0.20/MWCNT electrode during the
pretreatments by potential sweep cycling with varying potential up-limits (Ea) in
0.5 M H2SO4. Ea = 0.46 (a), 0.66 (b) and 0.96 V (c).
Fig. 10. CO stripping voltammograms of Pt–(RuOxHy)m/MWCNTs in 0.5 M H2SO4
after the pretreatments in (A) narrow (ꢃ0.20 to 0.46 V) and (B) extended (ꢃ0.20 to
0.96 V) potential ranges.

J.-H. Ma et al. / Journal of Catalysis 275 (2010) 34–44
41
results presented in the preceding paragraphs of this section that
significant RuOxHy dissolution had happened during the pretreat-
ment in the extended potential range.
of even a very small amount of RuOxHy. For instance, the ratio on
Pt–(RuOxHy)0.04/MWCNTs was 0.63, not to mention that residual
RuOxHy in the working electrode could be significantly lower be-
cause of RuOxHy dissolution during the pretreatment. The ratio
became as high as If/Ib = 0.70 on every other Pt–(RuOxHy)m/MWCNT
catalyst (m = 0.10–0.40), indicating a further improved anti-poison
property of these RuOxHy-promoted Pt catalysts.
The electrocatalytic activities of Pt in these pretreated Pt–(RuOx-
Hy)m/MWCNT catalysts (Fig. 11B) were all lower than that in the
reference Pt/MWCNT catalyst, except for Pt–(RuOxHy)0.20/MWCNTs.
Also, the activity sequence of these Pt–(RuOxHy)m/MWCNT cata-
lysts was very different from that of their counterparts in Fig. 11A.
RuOxHy dissolution during the pretreatment in the extended
potential range could be the main cause for this difference. In fact,
our ICP-MS measurement of ruthenium content in the electrolyte
solution of the most active Pt–(RuOxHy)0.20/MWCNT catalyst in
Fig. 10B disclosed that a half (55%) of RuOxHy in the catalyst had
dissolved into the electrolyte solution during the pretreatment in
the extended potential range. Therefore, the actual RuOxHy content
in this most active Pt–(RuOxHy)0.20/MWCNTs (Fig. 10B) was
m = 0.09. Interestingly, this number agreed well with that of the
most active Pt–(RuOxHy)0.10/MWCNT catalyst in Fig. 11A, where
the catalyst pretreatment was conducted in the narrow potential
range.
3.2.3. Pt–(RuOxHy)m/MWCNTs for methanol electro-oxidation
Fig. 11 shows the CV curves of methanol electro-oxidation on
the Pt–(RuOxHy)m/MWCNT electrodes in the narrow and extended
potential ranges after the standardized pretreatments. Since the CV
curves obtained in the narrow potential range did not allow the
detection of methanol oxidation peaks either in the forward or in
the reverse scans, the current was plotted in Fig. 10A as a function
of potential in the positive scans for the sake of clarity. According
to the methanol electro-oxidation current, the activity of Pt in
the Pt–(RuOxHy)m/MWCNT catalysts changed remarkably with
the amount of RuOxHy, and the maximum activity was obtained
at m = 0.10. The activity decrease at higher m values could be as-
cribed to the loss of active Pt surface (EAS, Table 3) for the electro-
chemical reaction.
The methanol electro-oxidation CV curves obtained on Pt–
(RuOxHy)m/MWCNT catalysts after the pretreatment in the ex-
tended potential range manifested clear anodic peaks in both
forward and reverse scans (Fig. 11B). The ratio of peak currents on
the forward (If) and reverse (Ib) scans could be regarded as an indi-
cator of the ‘‘cleanness” of the catalyst surface from poisonous car-
bon-containing species [39]. This ratio (If/Ib) was given in Table 3
and was used to evaluate the anti-poison property of the Pt–(RuOx-
Hy)m/MWCNT catalysts shown in Fig. 10B. The lowest If/Ib ratio
(0.59) over the reference Pt/MWCNTs was indeed in line with a
poor anti-poison property of the unpromoted Pt in this catalyst.
However, this If/Ib ratio was enhanced significantly by the presence
Fig. 12 shows the CA performance for methanol oxidation of the
Pt–(RuOxHy)m/MWCNT catalysts under a constant potential of
0.40 V. All of the catalysts featured a pronounced current decay
in the first 10–20 min, which could arise from the accumulation
of poisonous intermediates such as species containing –CO, –CHO
and/or –CH2OH functionality [40]. The current decay slowed down
Fig. 11. Cyclic voltammograms of Pt–(RuOxHy)m/MWCNTs in 0.5 M H2SO4 + 0.5 M
CH3OH after the pretreatments in (A) narrow (ꢃ0.20 to 0.46 V) and (B) extended
(ꢃ0.20 to 0.96 V) potential ranges. Only the forward scan curve is shown in (A).
Fig. 12. Chronoamperograms of Pt–(RuOxHy)m/MWCNTs in 0.5 M H2SO4 + 0.5 M
CH3OH at 0.40 V after the pretreatments in (A) narrow (ꢃ0.20 to 0.46 V) and (B)
extended (ꢃ0.20 to 0.96 V) potential ranges.

42
J.-H. Ma et al. / Journal of Catalysis 275 (2010) 34–44
sarily have to come directly from the electrolyte solution; it can also
come from the structure of RuOxHy:
but was still observable at longer times, which might be associated
with the adsorption of SO4 anions on the catalyst surface and/or
2-
mass transport effects [26,41]. The initial current decay was much
more pronounced on the reference Pt/MWCNT electrode (m = 0),
which lost nearly 50% current in the first 20 min.
Pt—COads þ RuOxHy ! Pt þ RuOxꢃ1Hyꢃ2 þ CO2 þ 2Hþ þ 2eꢃ
Pt–(RuOxHy)0.20/MWCNTs was identified as the most efficient cata-
lyst for this catalysis since it produced the lowest Eo and Ep on the
CO stripping voltammogram. As the domain size of RuOxHy entities
increased with m in Pt–(RuOxHy)m/MWCNT samples (Fig. 5), the
weakening of RuOxHy promotion of Pt for the CO removal at
m > 0.20 could be probably due to a self-aggregation of the pro-
moter to form bigger RuOxHy particles that changed the proximity
between RuOxHy and Pt.
Anodic CV scanning up to a reasonably high potential (such as
0.96 V) has been frequently used to clean and activate the anode
catalysts for electrochemical reactions [44]. It was alerted earlier,
however, that dissolution of ruthenium species in anode PtRu/C
catalyst might take place at potentials higher than 0.46 V [45,46].
This possibility of ruthenium dissolution is demonstrated clearly
in this work by comparing the electrochemical performance of
RuOxHy/MWCNTs (Fig. 7) and Pt–(RuOxHy)m/MWCNTs (Fig. 8) after
they were subjected to the ‘‘standard” pretreatments both in the
narrow (ꢃ0.20 to 0.46 V) and in the extended (ꢃ0.20 to 0.96 V) po-
tential ranges. It is now clear that the pretreatment in the narrow
potential range induced no ruthenium dissolution, but the pre-
treatment in the extended potential range caused significant disso-
lution of RuOxHy from the electrodes. Quantitatively, the electrodes
of Pt–(RuOxHy)m/MWCNTs with m = 0.10 and 0.20 were confirmed,
respectively, to lose 70% and 55% of their RuOxHy during the pre-
treatment in the extended potential range.
When the pretreatment was carried out in the narrow potential
range, the activity sequence of Pt–(RuOxHy)m/MWCNTs in the CA
tests (Fig. 12A) always agreed with the order in the CV measure-
ments (Fig. 11A). When these Pt–(RuOxHy)m/MWCNT catalysts
were subjected to the pretreatment in the extended potential
range, however, the activity order based on CA tests (Fig. 12B)
was not consistent with that based on the CV measurements
(Fig. 11B). Despite this disagreement, Pt–(RuOxHy)0.20/MWCNTs
exhibited consistently the highest activity in both measurements.
The mass-specific activity (MSA) and intrinsic activity (IA) data
of Pt in Pt–(RuOxHy)m/MWCNTs are also listed in Table 3. These
activity data were obtained as the currents recorded at the time
of 60 min in Fig. 12. Quantitatively, the MSA and IA data of the
most active catalyst after the pretreatment in the narrow potential
range (i.e., Pt–(RuOxHy)0.10/MWCNTs) were enhanced by an order
of magnitude (9 times increase in MSA and 14 times in IA) com-
pared with the reference Pt/MWCNT catalyst (m = 0), due to the
promotional effect of RuOxHy. By contrast, the MSA and IA data
of the most active catalyst after the pretreatment in the extended
potential range (i.e., Pt–(RuOxHy)0.20/MWCNTs) were, respectively,
1.4 and 2.2 times those of the reference Pt/MWCNT catalyst. Inter-
estingly, the activity orders in terms of MSA and IA were consistent
with each other with the only exception at Pt–(RuOxHy)0.40
/
MWCNTs pretreated in the narrow potential range, the MSA of Pt
in this sample was even lower than the reference Pt/MWCNT cat-
alyst due to severe encapsulation of Pt by RuOxHy.
The electro-oxidation of methanol over the present Pt–(RuOx-
Hy)m/MWCNT catalysts would involve the following key steps:
CH3OH þ Pt ! Pt—COads þ 4Hþ þ 4eꢃ
4. Discussion
Pt—COads þ RuOxHy ! Pt þ RuOxꢃ1Hyꢃ2 þ CO2 þ 2Hþ þ 2eꢃ
The present data from comprehensive characterizations by XRD,
XPS, H2-TPR and TGA (Figs. 1–4) are consistent in showing that the
as-prepared Pt–(RuOxHy)m/MWCNT samples in this study con-
tained no metallic Ru and RuO2 crystallite but only amorphous hy-
drous ruthenium oxide, which can be expressed as either
RuO2ꢀnH2O or RuOxHy. Ruthenium in this RuOxHy species is charac-
terized as a mixed valence state between +3 and +4 and can behave
as a mixed electron/proton conductor innately related with the
Ru–OH speciation [42,43]:
Since the average sizes of Pt crystallites (ca. 2.7 nm) from the XRD
measurements did not change significantly in the as-prepared Pt–
(RuOxHy)m/MWCNT samples, the following discussion on RuOxHy
promotion of Pt in these samples is made by assuming no change
in the sizes of Pt.
In the absence of RuOxHy dissolution (i.e., when the pretreat-
ment of the electrode catalyst was conducted in the narrow poten-
tial range), the MSA and IA data of Pt in Pt–(RuOxHy)m/MWCNTs for
methanol electro-oxidation were seen to increase remarkably with
m up to m = 0.10 and then to decrease at higher m (Fig. 11A and
12A, Table 3). As is discussed on RuOxHy promotion of Pt for the
CO stripping catalysis, these data would indicate that an appropri-
ate m number or a suitable Pt–RuOxHy proximity is also required
for maximizing the activity of Pt for the methanol electro-oxida-
tion. The activity of Pt in the most active Pt–(RuOxHy)0.10/MWCNTs
was enhanced by a factor of 9 in MSA and 14 in IA, respectively,
when compared with the reference Pt/MWCNT catalyst. It should
be mentioned that the most efficient Pt–(RuOxHy)0.20/MWCNT cat-
alyst in the CO stripping measurements (Fig. 10A, Table 3) failed to
show the highest activity for methanol electro-oxidation. This fact
suggests that the suitable Pt–RuOxHy proximity required for the CO
removal catalysis in Pt–(RuOxHy)m/MWCNTs was somewhat differ-
ent from that required for the methanol electro-oxidation catalysis.
Significantly negative effect was noted when a large excess of
RuOxHy was present in the Pt–RuOxHy/MWCNT catalyst. For in-
stance, Pt–(RuOxHy)0.40/MWCNTs became even inferior to the ref-
erence Pt/MWCNT catalyst in terms of MSA, which could be at
least partly due to an increased blocking by RuOxHy of the active
Pt surface at high Ru loadings. Consequently, the presence of RuOx-
Hy would produce two opposite effects on the electrocatalysis of Pt,
as shown by the data summarized in Table 3. The positive effect is
RuIV—O2ꢃ þ Hþ þ eꢃ ꢀ RuIII—OH
CO stripping voltammetry is considered as an in situ approach to
study the promoter function in Pt-based electrocatalysts in electro-
chemical environments [17,21,26] and has often been used to eval-
uate the anti-poison or CO tolerance property of electrocatalysts.
The present CV measurements of the Pt-free RuOxHy/MWCNT sam-
ples (Figs. 6 and 7) clearly demonstrate that RuOxHy itself is inert for
hydrogen and CO chemisorptions and is completely inactive for
methanol electro-oxidation at potentials lower than 0.85 V. How-
ever, its co-presence with Pt in the Pt–(RuOxHy)m/MWCNT samples
induced significant lowering of the Eo and Ep on the CO stripping
curves (Fig. 10, Table 3), compared to the case without its presence
(i.e., on the reference Pt/MWCNTs). Thus, the following modified
bifunctional mechanism can be proposed to explain the observed
RuOxHy promotion of Pt for CO removal:
Pt—COads þ RuIII—OH þ H2O ! Pt þ RuIV—O2ꢃ þ 3Hþ þ 3eꢃ þ CO2
The peculiarity of RuOxHy as a mixed electron/proton conductor
would easily delocalize the protons and electrons. On the other
hand, the required water molecule for this reaction does not neces-

J.-H. Ma et al. / Journal of Catalysis 275 (2010) 34–44
43
reflected by its promotion on CO removal from the Pt surface and
its enhancement of Pt activity for methanol electro-oxidation,
and the negative one is characterized by a continuous decrease
in EAS of Pt with increasing m. However, an optimization in the
proximity between Pt particles and RuOxHy entities seems insuffi-
cient for the positive effect of RuOxHy, since the IA data of Pt (Ta-
ble 3) did not remain constant after the optimization at m = 0.10
but decreased continuously with further increase in m. Electronic
and perhaps other effects could also be involved in inducing the
change of IA at m > 0.10, which remained unclear at the present
stage.
The situation became quite complex when the catalyst pretreat-
ment was done in the extended potential range, which caused sig-
nificant dissolution of RuOxHy into the electrolyte solution. In this
case (see the right half of Table 3), Pt–(RuOxHy)0.20/MWCNTs
showed the highest activity for methanol electro-oxidation,
whereas Pt–(RuOxHy)0.30/MWCNTs became the most efficient cata-
lyst for the CO removal. Although after the pretreatment in the ex-
tended potential range the composition and EAS of the most active
Pt–(RuOxHy)0.20/MWCNTs (in the right half of Table 3) became, due
to RuOxHy dissolution, very close to those of the most active Pt–
(RuOxHy)0.10/MWCNTs after the pretreatment in the narrow poten-
tial range (in the left half of Table 3), the MSA and IA data of the
former catalyst were both significantly lower for the electro-oxida-
tion of methanol. In fact, most of the Pt–(RuOxHy)m/MWCNT sam-
ples pretreated in the extended potential range showed lower IA
than their counterparts pretreated in the narrow potential range.
This observation cannot be explained only by the dissolution of
RuOxHy since the number of IA for the RuOxHy-free Pt/MWCNT cat-
alyst after the pretreatment in the extended potential range was
even enhanced by a factor of 4 compared with its counterpart pre-
treated in the narrow potential range.
MWCNTs were still 1.4 and 2.2 times those of the reference cata-
lyst Pt/MWCNTs, respectively.
The electrode catalyst (Pt–(RuOxHy)0.20/MWCNTs) subjected to
the pretreatments of varied numbers of cycles with the potential
up-limit at Ea = 0.96 V (Fig. 9c) was also used to study the CO strip-
ping and methanol oxidation catalysis by CV measurements in the
extended potential range (ꢃ0.20 to 0.96 V) (Supplementary data
Figs. S1 and S2). Fig. 13 correlates the Pt activity at 0.40 V for both
CO stripping and methanol oxidation catalysis with the number of
such pretreatment cycles. With reference to the data shown in
Fig. 9, it is clear that the electrocatalytic activity of Pt for both reac-
tions well related with the remained percentage of RuOxHy after
more than 10 pretreatment cycles, which is in support of the func-
tion of RuOxHy as the efficient promoter of Pt for the electro-oxida-
tion catalysis. However, it is of interest to note in Fig. 13 that
during the first 10 cycles of the pretreatment at Ea = 0.96 V, the
Pt activity for methanol oxidation even increased to some extent
but the activity for CO oxidation decreased significantly, compared
with the case of no RuOxHy dissolution (i.e., data at the zero-cycle).
These facts suggest again that the Pt–RuOxHy proximity required
for the two reactions was different, and the pretreatment up to a
potential as high as 0.96 V could modify the property of Pt, in addi-
tion to the RuOxHy dissolution.
It has been known that the surface of a Pt anode catalyst would
undergo an irreversible oxidation process when the anodic poten-
tial limit of CV measurement was extended to higher than 0.71 V
vs. SCE (0.95 V vs. NHE) [47]. In an earlier work by Hu and Liu
[48], the application of repeated CV scanning to Pt electrodes in
ꢃ0.24 to 1.26 V was shown effective in inducing changes in the re-
dox reversibility of Pt surfaces, which resulted in significant activ-
ity improvement of Pt for methanol electro-oxidation. In this work,
the activity improvement in Pt/MWCNTs after the pretreatment in
the extended potential range for methanol electro-oxidation, as
shown in the right half of Table 3, seems to be in line with those
observations. Thus, it seems conclusive that the structure/property
change to Pt induced by the pretreatment of Pt–(RuOxHy)m/
MWCNTs in the extended potential range can be indeed an impor-
tant aspect. In addition to RuOxHy dissolution, for understanding
the change of the activity data in Table 3. However, more powerful
surface-specific and in situ measurements would be required to
gain further insights into the details of such surface structure
change.
Therefore, it seems quite reasonable that the pretreatment con-
ducted in the extended potential range induced not only ruthe-
nium dissolution but also significant change in the surface
property of Pt. The development of ‘‘hydrogen adsorption/desorp-
tion” features on Pt–(RuOxHy)m/MWCNT catalysts to approach
those on RuOxHy-free Pt/MWCNTs after such pretreatment
(Fig. 8B) can be further evidence for the change in the surface
structure/property of Pt. The residual RuOxHy after such pretreat-
ment would also have a positive effect on the formation of a better
Pt surface structure for the electro-oxidation of methanol, which
was supported by comparison of the MSA and IA data in the right
half of Table 3. For instance, after the pretreatment in the extended
We checked also in separate experiments if, after the ‘‘standard”
pretreatment of the Pt–(RuOxHy)0.20/MWCNT electrode in the ex-
tended potential range, ruthenium dissolution would continue to
occur during repeated CV measurements of the methanol electro-
oxidation reaction. The results showed that the loss of ruthenium
was as low as 5% when the CV measurement was repeated for
110 cycles in the full potential range (ꢃ0.20 to 0.96 V), suggesting
that the remaining RuOxHy promoter in the electrode was reason-
ably stabilized for methanol electro-oxidation.
potential range, the MSA and IA of the most active Pt–(RuOxHy)0.20
/
5. Conclusions
Our data demonstrate clearly that the RuOxHy (hydrous ruthe-
nium oxide) alone, without co-presence of other ruthenium enti-
ties (metallic Ru or anhydrous RuO2), can act as an efficient
promoter of Pt in Pt–RuOxHy/MWCNT materials for the electro-oxi-
dation of CO and methanol. The promotion of RuOxHy in the elect-
rocatalysis depended on the proximity and relative amount of
RuOxHy and Pt and was affected significantly by the pretreatment
potentials. Severe dissolution of the RuOxHy promoter from the cat-
alysts was identified when the electrode was pretreated by re-
peated CV scanning in the extended potential range (ꢃ0.20 to
0.96 V) but was not detected when the potential up-limit of the
Fig. 13. Correlation of Pt activity in the pretreated Pt–(RuOxHy)0.20/MWCNT
electrode for (a) CO and (b) methanol electro-oxidation with the number of
potential sweep cycles for the pretreatment with potential up-limit at Ea = 0.96.

44
J.-H. Ma et al. / Journal of Catalysis 275 (2010) 34–44
[18] K.I.B. Eguiluz, G.R. Salazar-Banda, D. Miwa, S.A.S. Machado, L.A. Avaca, J. Power
Sources 179 (2008) 42.
[19] J.L. Gomez de la Fuente, M.V. Martinez-Huerta, S. Rojas, P. Hernandez-
Fernandez, P. Terreros, J.L.G. Fierro, M.A. Pena, Appl. Catal. B: Environ. 88
(2009) 505.
[20] Q. Lu, B. Yang, L. Zhuang, J. Lu, J. Phys. Chem. B 109 (2005) 1715.
[21] H.N. Dinh, X. Ren, F.H. Garzon, P. Zelenay, S. Gottesfeld, J. Electroanal. Chem.
491 (2000) 222.
pretreatment was set no higher than 0.46 V. The RuOxHy dissolu-
tion would change not only the proximity and relative amount of
RuOxHy and Pt but also the surface structure of Pt. Further work
is needed to uncover factors governing the mobility and stability
of RuOxHy and Pt during the pretreatments.
[22] H. Kim, I. Rabelo de Moraes, G. Tremiliosi-Filho, R. Haasch, A. Wieckowski, Surf.
Sci. 474 (2001) L203.
Acknowledgments
[23] C. Bock, A. Collier, B. MacDougall, J. Electrochem. Soc. 152 (2005) A2291.
[24] P. Chen, H.B. Zhang, G.D. Lin, Q. Hong, K.R. Tasi, Carbon 35 (1999) 1495.
[25] X.M. Wu, Y. Wang, K.M. Dong, J.M. Zhou, G.D. Lin, H.B. Zhang, Acta Chim. Sin.
63 (2005) 484.
This work is supported by NSF(20773074, 20921001) and MOST
(2006AA03Z225) of China.
[26] L. Li, G. Wu, B.Q. Xu, Carbon 44 (2006) 2973.
Appendix A. Supplementary material
[27] X. Fu, H. Yu, F. Peng, H. Wang, Y. Qian, Appl. Catal. A: Gen. 321 (2007) 190.
[28] L. Li, H.X. Wang, B.Q. Xu, J.L. Li, W. Xing, Z.Q. Mao, Acta Phys. Chim. Sin. 19
(2003) 342.
[29] J.P. Zheng, P.J. Cygan, T.R. Jow, J. Electrochem. Soc. 142 (1995) 2699.
[30] B. Yang, Q.Y. Lu, Y. Wang, L. Zhuang, J.T. Lu, P.F. Liu, J.B. Wang, R.H. Wang,
Chem. Mater. 15 (2003) 3552.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2010.07.021.
[31] S. Kundu, Y. Wang, W. Xia, M. Muhler, J. Phys. Chem. C 112 (2008) 16869.
[32] S. Huang, S. Chang, C. Yeh, J. Phys. Chem. B 110 (2006) 234.
[33] S. Sarangapani, B.V. Tilak, C.P. Chen, J. Electrochem. Soc. 143 (1996) 3791.
[34] K.E. Swider, C.I. Merzbacher, P.L. Hagans, D.R. Rolison, J. Non-Cryst. Solids 225
(1998) 348.
[35] V. Radmilovic´, H.A. Gasteiger, P.N. Ross Jr., J. Catal. 154 (1995) 98.
[36] S. Trasatti, G. Buzzanca, J. Electroanal. Chem. Interfacial Electrochem. 29
(1971) 1.
[37] F. Scheiba, M. Scholz, L. Cao, R. Schafranek, C. Roth, C. Cremers, X.P. Qiu, U.
Stimming, H. Fuess, Fuel Cells 6 (2006) 439.
[38] C.H. Hamann, A. Hamnett, W. Vielstich, Electrochemistry, second ed., Wiley-
VCH, Weinheim, 1998.
[39] R. Manoharan, J.B. Goodenough, J. Mater. Chem. 2 (1992) 875.
[40] J.W. Guo, T.S. Zhao, J. Prabhuram, R. Chen, C.W. Wong, J. Power Sources 156
(2006) 345.
[41] J. Jiang, A. Kucernak, J. Electroanal. Chem. 543 (2003) 187.
[42] A. Yamada, J.B. Goodenough, J. Electrochem. Soc. 145 (1998) 737.
[43] W. Dmowski, T. Egami, K.E. Swider-Lyons, C.T. Love, D.R. Rolison, J. Phys. Chem.
B 106 (2002) 12677.
[44] H.A. Gasteiger, N. Markovic, P.N. Ross Jr., E.J. Cairns, J. Phys. Chem. 98 (1994)
617.
[45] H.A. Gasteiger, N. Markovic, P.N. Ross Jr., E.J. Cairns, J. Phys. Chem. 97 (1993)
12020.
References
[1] H.A. Gasteiger, N. Markovic, P.N. Ross Jr., E.J. Cairns, J. Electrochem. Soc. 141
(1994) 1795.
[2] T. Kawaguchi, W. Sugimoto, Y. Murakami, Y. Takasu, J. Catal. 229 (2005) 176.
[3] M.P. Hogarth, G.A. Hards, Platinum Met. Rev. 40 (1996) 150.
[4] M. Watanabe, S. Motoo, J. Electroanal. Chem. Interfacial Electrochem. 60 (1975)
267.
[5] T. Frelink, W. Visscher, A.P. Cox, J.A.R. van Veen, Electrochim. Acta 40 (1995)
1537.
[6] K.L. Ley, R. Liu, C. Pu, Q. Fan, N. Leyarovska, C. Segre, E.S. Smotkin, J.
Electrochem. Soc. 144 (1997) 1543.
[7] D. Chu, S. Gilman, J. Electrochem. Soc. 143 (1996) 1685.
[8] A. Hamnett, Catal. Today 38 (1997) 445.
[9] A. Hamnett, B. Kennedy, Electrochim. Acta 33 (1988) 1613.
[10] A.S. Arico, P. Creti, H. Kim, R. Mantegna, N. Giordano, V. Antonucci, J.
Electrochem. Soc. 143 (1996) 3950.
[11] A.H.C. Sirk, J.M. Hill, S.K.Y. Kung, V.I. Birss, J. Phys. Chem. B 108 (2004) 689.
[12] D.R. Rolison, P.L. Hagans, K.E. Swider, J.W. Long, Langmuir 15 (1999) 774.
[13] J.W. Long, R.M. Stroud, K.E. Swider-Lyons, D.R. Rolison, J. Phys. Chem. B 104
(2000) 9772.
[14] H.M. Villullas, F.I. Mattos-Costa, L.O.S. Bulhoes, J. Phys. Chem. B 108 (2004)
12898.
[46] J. McBreen, S. Mukerjee, J. Electrochem. Soc. 142 (1995) 3399.
[47] B.E. Conway, Prog. Surf. Sci. 49 (1995) 331.
[48] C.C. Hu, K.Y. Liu, Electrochim. Acta 44 (1999) 2727.
[15] L. Cao, F. Scheiba, C. Roth, F. Schweiger, C. Cremers, U. Stimming, H. Fuess, L.
Chen, W.T. Zhu, X.P. Qiu, Angew. Chem. Int. Ed. 45 (2006) 5315.
[16] S. Huang, C. Chang, K. Wang, C. Yeh, ChemPhysChem 8 (2007) 1774.
[17] J. Zeng, J.Y. Lee, Mater. Chem. Phys. 104 (2007) 336.