Hydrogen electrooxidation over palladium-gold alloy: Effect of pretreatment in ethylene on catalytic activity and CO tolerance

Article abstract of DOI:10.1016/j.electacta.2012.05.043

Palladium gold alloy powder (PdAu₂) was synthesized by reduction of the [Pd(NH₃)₄][AuCl₄]₂ complex with H₂ at 90°C, and then was additionally heated in a C ₂H₄ (PdAu₂-C_x) or H₂ flow (PdAu₂-H₂) at 300°C. The samples were characterized by X-ray diffraction, X-ray photoelectron spectroscopy, transmission electron microscopy and by a set of electrochemical methods. The Pd:Au ratio on the surface of the PdAu₂ alloy suffers marginal changes after heating under C₂H₄ atmosphere, and decreases as a result of heating in a H₂ flow. Hydrogen oxidation reaction (HOR) over the catalysts was studied using rotating disk electrode at 25°C in 0.1 M HClO₄ in pure H₂ and in the presence of 1100 ppm of CO. Heat-treatment of the PdAu₂ alloy under H₂ does not improve its HOR activity and CO tolerance (normalized to Pd and total metal surface area), while the PdAu₂-C_x catalyst exhibits substantially enhanced catalytic performance in the HOR both in the absence and in the presence of CO as compared to the parent PdAu₂ alloy. The enhanced activity of the C₂H₄-treated PdAu₂ alloy is explained by the promoting action of carbon atoms incorporated into the alloy structure during treatment under C₂H₄ atmosphere. This supposition is approved by the fact that mild oxidative treatment of the PdAu₂-C_x sample degrades notably its catalytic activity.

Full text of DOI:10.1016/j.electacta.2012.05.043

Electrochimica Acta 76 (2012) 344–353
Contents lists available at SciVerse ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Hydrogen electrooxidation over palladium–gold alloy: Effect of pretreatment in
ethylene on catalytic activity and CO tolerance
Alexander N. Simonova,∗,1 , Pavel E. Plyusninb,c , Yury V. Shubinb,c , Ren I. Kvona , Sergey V. Korenevb,c
,
Valentin N. Parmona,c
a Boreskov Institute of Catalysis of Siberian Branch of Russian Academy of Sciences, Prospekt Akademika Lavrentieva, 5, Novosibirsk 630090, Russia
b Nikolayev Institute of Inorganic Chemistry of Siberian Branch of Russian Academy of Sciences, Prospekt Akademika Lavrentieva, 3, Novosibirsk 630090, Russia
c Novosibirsk State University, Pirogova Street, 2, Novosibirsk 630090, Russia
a r t i c l e i n f o
a b s t r a c t
Article history:
Palladium gold alloy powder (PdAu2) was synthesized by reduction of the [Pd(NH3)4][AuCl4]2 complex
with H2 at 90 ◦C, and then was additionally heated in a C2H4 (PdAu2–Cx) or H2 flow (PdAu2–H2) at 300 ◦C.
The samples were characterized by X-ray diffraction, X-ray photoelectron spectroscopy, transmission
electron microscopy and by a set of electrochemical methods. The Pd:Au ratio on the surface of the
PdAu2 alloy suffers marginal changes after heating under C2H4 atmosphere, and decreases as a result of
heating in a H2 flow. Hydrogen oxidation reaction (HOR) over the catalysts was studied using rotating
disk electrode at 25 ◦C in 0.1 M HClO4 in pure H2 and in the presence of 1100 ppm of CO. Heat-treatment
of the PdAu2 alloy under H2 does not improve its HOR activity and CO tolerance (normalized to Pd and
total metal surface area), while the PdAu2–Cx catalyst exhibits substantially enhanced catalytic perfor-
mance in the HOR both in the absence and in the presence of CO as compared to the parent PdAu2 alloy.
The enhanced activity of the C2H4-treated PdAu2 alloy is explained by the promoting action of carbon
atoms incorporated into the alloy structure during treatment under C2H4 atmosphere. This supposition is
approved by the fact that mild oxidative treatment of the PdAu2–Cx sample degrades notably its catalytic
activity.
Received 17 February 2012
Received in revised form 9 May 2012
Accepted 9 May 2012
Available online 17 May 2012
Keywords:
Palladium–gold alloy
Hydrogen electrooxidation
CO tolerance
Interstitial carbon
© 2012 Elsevier Ltd. All rights reserved.
1. Introduction
and theoretical effort has been made for investigating and explain-
ing the enhanced catalytic activity of various PdAu electrodes for
Changing the reactivity of metal surfaces by alloying with
another metal is a powerful tool for creating catalytic materials
with the properties strikingly different from those of the individ-
ual unalloyed components. Alteration of the strength of adsorption
of a reactant on various metals and their combinations was gen-
eralized by Hammer and Norskov in their d-band model [1]. The
general concept of the theory can be briefly summarized as follows:
an upshift or downshift of a d-band center of a metal caused by
some modification (e.g., change of the surface structure or alloying)
results in stronger or weaker adsorption on its surface, respectively,
and consequent changes in the catalytic properties of the metal [2].
Among bimetallic catalysts, palladium–gold system has been
addressed a notable attention due to its technological relevance for
the vinyl acetate synthesis [3] and application for other catalytic
reactions like, e.g., selective oxidation of alcohols [4], hydrodehalo-
genation [5] and other. At the same time, a great experimental
the hydrogen evolution and oxidation reactions (HER/HOR). Both
fundamental importance and practical relevance of discovering
highly active catalysts for the hydrogen electrode reactions [6,7]
are the main driving forces of these investigations. The most exten-
sively studied structures of the palladium–gold catalysts for the
HER/HOR are monolayers, sub-monolayer islands and nanoclusters
deposited on the surface of well-defined Au single crystal elec-
trodes (see, for example, [8] and references therein) or supported
Au nanoparticles [9,10]. Recently, two systematic studies of the
electrocatalytic activity in the HOR/HER of the PdAu alloys with var-
ied composition [11,12] were published. Both papers report unique
electrocatalytic properties of bimetallic ensembles with low con-
centration of palladium. Earlier, Schmidt et al. [13] reported that
carbon supported Pd–Au alloy nanoparticles exhibit an outstanding
CO tolerance during the HOR at low overpotentials. This property
is of utmost importance for the catalysts used at the anode of the
low-temperature proton exchange membrane fuel cells fuelled by
hydrogen contaminated with carbon monoxide.
Commonly, the alteration of electrocatalytic and adsorption
properties of palladium under the action of the Au (or other metal)
promoter is explained in terms of electronic and ensemble effects.
∗
Corresponding author. Tel.: +7 383 3269507; fax: +7 383 3304719.
E-mail address: san@catalysis.ru (A.N. Simonov).
ISE member.
1
0013-4686/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.electacta.2012.05.043

A.N. Simonov et al. / Electrochimica Acta 76 (2012) 344–353
345
The former is the mentioned above modification of the d-band cen-
ter of palladium caused by a strain of the Pd crystal lattice [14] or
specific chemical/electronic interaction of Pd with the substrate
[15,16]. An ensemble effect refers to the alteration of the catalytic
properties of PdxAuy ensembles when their chemical composition
changes [11,12,16,17].
Apart from the metallic promoters of the palladium activity for
the hydrogen electrode reactions, the promoting effect of incor-
poration of carbon atoms into the crystalline lattice of palladium
on its electrocatalytic performance for the HOR was found recently
[18]. Incorporation of C atoms into the Pd lattice leads to substan-
tial changes of the electronic state and adsorption properties of
the metal with respect to hydrogen, CO and other reagents and
consequent changes of its catalytic [19] and electrocatalytic [18]
properties.
The bench mark of the present work was an idea to conjugate the
promoting effects of alloying of palladium with gold and dissolution
of carbon atoms in the crystal lattice of the alloy on the electrocat-
alytic activity and CO tolerance of Pd during the HOR. Capability
of uptake of carbon atoms by thin films and nanoparticles of pal-
ladium [18–21] into the interstitial cites of the crystalline lattice
of the metal is well-known. Recently, it was found that gold in a
highly dispersed state is also capable of forming the Au–C solid solu-
tions with appreciable concentration of carbon at comparatively
low temperatures [22,23], in contrast to the bulk metal [24]. Taking
into account that palladium and gold are able to form a continuous
set of the stable solid solutions with a crystal structure similar to
that of the individual metals, there are no reasons to neglect the
possibility for the PdAu alloys to incorporate carbon atoms. How-
ever, to the best of our knowledge, no data on the properties of the
ternary palladium–gold–carbon solid solutions are reported for the
moment. Thus, the present paper addresses two intriguing funda-
mental problems: (1) possibility of incorporation of carbon into
the PdAu alloy, and (2) influence of the interstitial carbon on the
electrocatalytic properties of palladium alloyed with gold.
residual ammonium chloride, which presence in the samples was
controlled by X-ray diffraction (XRD) and X-ray photoelectron spec-
troscopy (XPS). To remove adsorbed chloride-ions from the surface
of the catalysts, the washed powders were additionally agitated
in the 0.1 M NaOH aqueous solution saturated with H2 and again
plentifully washed with distilled water. Finally, the samples were
dried in air at 90 ◦C.
Parts of the synthesized Pd and PdAu2 samples were treated
with ethylene at 300 ◦C to incorporate carbon atoms into the crys-
talline lattice of the samples. For that, a part of the freshly prepared
material was placed in a quartz reactor, air oxygen was removed
by flowing N2 during 30 min at ambient temperature, nitrogen was
replaced by C2H4 fed at the space velocity of ca. 5 ml min−1, the
reactor was heated at 150 ◦C for 60 min, and then the temperature
was elevated to 300 ◦C with the rate of 20 min−1. After 180 min
at 300 ◦C, ethylene was substituted by N2, the reactor was cooled
to the room temperature, and the resulting PdC0.15 or PdAu2–Cx
sample was unloaded.
Additionally, freshly prepared PdAu2 was heat-treated under
the same conditions using H2 instead of ethylene. The resulting
sample is denoted below as PdAu2–H2.
2.3. Characterization
XRD patterns of fresh and modified Pd and PdAu2 catalysts were
recorded on a DRON SEIFERT RM4 diffractometer using CuK␣ radia-
tion (ꢀ = 0.15418 nm) and graphite monochromator on the reflected
beam. The data were collected for 15 s per step with a 0.05◦ step
size in the 2ꢁ range between 15◦ and 90◦. The fcc lattice constants
of the phases in the studied samples were refined by full-profile
analysis within the whole diffraction range with the POWDERCELL
2.4 program [26]. Phase compositions of the AuxPd1−x alloys were
determined using the experimentally measured dependence of the
fcc lattice constant of the palladium–gold solid solution on the
atomic fractions of the constituent elements [27]. This dependence
strictly follows the Vegard’s law. The volume-averaged crystallite
sizes (D, Å) were determined from the mean value of the integral
breadths of the (1 1 1), (2 0 0), (2 2 0), (3 1 1) and (2 2 2) diffrac-
tion peaks, using the Scherrer equation [28]. The integral breadth
method is independent of the distribution in size and shape of
the analyzed objects [29]. Deconvolution and fitting of the X-ray
diffraction lines was based on the Pearson (PVII) function and was
performed using the WinFit 1.2.1 program [30].
Transmission electron microscopy (TEM) studies were carried
out on a JEOL JEM-2010 electron microscope with a lattice res-
olution of 0.14 nm at a 200 kV accelerating voltage. The local
elemental analysis was performed with EDX method using the
Energy-dispersive X-ray Phoenix Spectrometer EDAX equipped
with Si (Li) detector with the energy resolution not worse than
130 eV.
X-ray photoelectron spectroscopy (XPS) study was performed in
SPECS electron spectrometer equipped with PHOIBOS-150 hemi-
spherical analyzer and monochromatic Al K␣ source. The energy
scale was calibrated using the Au4f7/2 (84.00 eV) and Cu2p3/2
(932.67 eV) peaks from metallic gold and copper foils. Residual gas
pressure during the spectra acquisition was lower than 3 × 10−7 Pa.
The samples of the palladium–gold alloy powder were tightly
pressed onto an indium foil mounted at a stainless steel sam-
ple holder so that no indium was observed in the XPS spectra. A
charge neutralizer was not used due to the conductive nature of the
samples. The deconvolution of overlapped spectra (Pd3d + Au4d,
O1s + Pd3p + Au4p3/2) and quantitative data processing were done
using XPSPEAK software version 4.1 and atomic sensitivity factors
tabulated in [31]. Yet it should be noted that as far as weak O1s sig-
nal is fully overlapped with intense Pd3p3/2 peak, the validity of XPS
data concerning both the electronic state of oxygen and its surface
2. Experimental
2.1. Materials
H[AuCl4] and H2[PdCl4] were prepared by dissolving gold
(99.99%) and palladium (99.99%) metal, respectively, (both pur-
chased from Krasnoyarsk Non-ferrous Metals Plant, Russia) in
aqua regia and subsequent boiling down in water and hydrochlo-
ric acid. Complex precursors of the catalysts [Pd(NH3)4]Cl2 and
[Pd(NH3)4][AuCl4]2 were synthesized as described in Ref. [25].
Other reagents were of extra pure or analytical grade and used
as purchased without further purification. For the catalyst prepa-
ration and electrochemical measurements, high-purity gases (Ar
99.998%; N2 99.80%; He 99.999%; H2 99.999%; CO 99.999%; C2H4
99.99%) were used.
2.2. Preparation and post-treatment of the Pd and PdAu2
catalysts
Pd and PdAu2 powders were prepared by reduction of the
[Pd(NH3)4]Cl2 and [Pd(NH3)4][AuCl4]2 complexes, respectively,
with hydrogen using the following procedure. A weighed amount of
the complex was placed in a quartz crucible, which was mounted
into a quartz tube reactor. The reactor was purged with helium
to remove air oxygen at ambient temperature, than hydrogen was
substituted for He, and the temperature was elevated to 90 ◦C at the
rate of 1 min−1. After 3 h at 90 ◦C, hydrogen was replaced by helium
and the reactor was cooled to room temperature. Acquired metallic
powders were thoroughly washed with distilled water to remove

346
A.N. Simonov et al. / Electrochimica Acta 76 (2012) 344–353
concentration is rather questionable. Actually, both position and
the area of O1s peak are very sensitive to the deconvolution param-
eters, i.e., the shape of background and the fraction of Lorentzian
chosen in each case, and are not discussed further in the text.
2.4. Electrochemical measurements
Electrochemical measurements were performed with an Auto-
lab PGSTAT 30 potentiostat equipped with a Scangen module in
a three-electrode cell comprised of individual compartments for
each electrode. The counter electrode (high surface area Pt foil) and
the reference electrode (mercury/mercury sulfate) compartments
were connected to the working electrode compartment through
a glass frit and a Luggin capillary, respectively. During the mea-
surements, the whole cell was thermostated at 25 ◦C. All electrode
potentials reported in this paper are referred to the reversible
hydrogen electrode (RHE).
A glassy carbon (GC) cylinder (∅5 mm) was used as a sub-
strate for the working electrode. The GC cylinder was fixed in the
Teflon holder with a brass current collector, and its lateral sur-
face was sealed with Teflon tape so that only the flat top surface
of the cylinder was exposed to the electrolyte. The absence of
a leakage through the sealing was verified by measuring capac-
itive currents of GC in deaerated 0.1 M HClO4 or 0.1 M H2SO4.
After electrochemical characterization of bare GC electrode, an
aliquot (typically 6–12 ␮l) of the suspension of a tested catalytic
material and Vulcan XC-72 carbon (Cabot) in a mixture of iso-
propanol/water (1:1, v/v) was pipetted on the clean and dry top
surface of the GC cylinder and dried under Ar flow for 60 min.
Vulcan carbon was introduced in the suspensions to provide homo-
geneous distribution of the catalytic material over the surface of
the GC rod, i.e., to make the electrodes suitable for the rotating disk
electrode (RDE) measurements. Concentration of the tested cat-
alytic material in the suspensions was ca. 0.5 mg ml−1 for Pd and
PdC0.15 catalysts, ca. 1 mg ml−1 for PdAu2 and PdAu2–Cx catalysts
and ca. 5 mg ml−1 for PdAu2–H2 catalyst, while concentration of
Vulcan XC-72 was always 1 mg ml−1. It should be stressed, how-
ever, that the real amount of the metal powder deposited on the
GC rod is notably lower than the expected value and is unique for
each of the tested catalysts due to the difficulties in maintenance
of the homogeneity of the suspensions of high-density materials.
Therefore, comparison of the mass-weighted specific properties of
the studied catalysts was not performed in the present study. In
some experiments, the dried electrode was kept at 120 ◦C in air
in drying oven during 20–30 min and then cooled to the ambient
temperature under Ar flow for additional 20 min. The contact of
the working electrodes with the deaerated supporting electrolyte
was established at a controlled potential of 0.4 V. The employed
procedure provides acceptable reproducibility and high stability of
the catalyst layers. Prior to characterization of the catalysts treated
with ethylene, the working electrode was subjected to the ‘anodic
cleaning’ procedure described in [18]. Such treatment provides
an increase in the intensities of both HUPD (coupled to (bi)sulfate
adsorption/desorption in 0.1 M H2SO4) and the surface oxide reduc-
tion peaks and does not provoke any changes in their positions
or shapes. Typical cyclic voltammograms (CV) registered with the
electrodes prepared using the above described procedure for the Pd
and PdC0.15 samples in 0.1 M H2SO4 are exemplified in Fig. 1a. The
curve for the unmodified Pd sample exhibits well-known features
of pure palladium [9], while that for the synthesized PdC0.15 powder
is in a good agreement with the results obtained for the supported
PdC0.14/C catalyst reported earlier [18]. Taking into account the doc-
umented tendency of Pd to dissolve in acidic electrolytes at the
potentials above ca. 0.6 V vs. RHE [32], the CVs to higher potentials
(like those on Fig. 1a, as well as CO stripping and CuUPD experi-
ments) were registered for the studied electrodes at the end of a
Fig. 1. Typical cyclic voltammograms (a) and HOR RDE curves at different rotation
rates of the electrode (b) measured with the Pd (1) and PdC0.15 (2) electrodes in Ar-
saturated 0.1 M H2SO4 (a) and H2-saturated HClO4 (b) at 25 ◦C and the sweep rate of
10 (a) and 5 (b) mV s−1. Inset on (b) shows the dependence of the diffusion limited
HOR current density (jDL) on the square root of the rotation rate of the electrode and
linear approximation of this dependence. Currents are normalized to the Pd surface
area determined from the CuUPD measurements and geometric surface area of the
electrode on (a) and (b), respectively.
catalytic test or separately. Fig. 1b shows the H2 electrooxidation
curves measured with the Pd and PdC0.15 electrodes using RDE at
various rotation rates. The coincidence of the dependence of the
limiting current densities with the theoretically predicted values
(see inset on Fig. 1b) conforms the homogeneous distribution of
the catalytic material over the surface of GC and applicability of
the selected method for the preparation of electrodes for the RDE
measurements.
For the CO stripping experiments, carbon monoxide was
adsorbed at the controlled potential of 0.1 V vs. RHE during
7–10 min and then removed from the electrolyte by Ar purging
for 45 min keeping the electrode at the same potential. In the pres-
ence of dissolved carbon monoxide in the electrolyte, the measured
currents at E = 0.1 V vs. RHE were virtually zero. This vindicates neg-
ligible catalytic activity of all studied catalysts for electrooxidation
of CO at this potential.
The determination of the total metal surface area was performed
using the copper underpotential deposition (CuUPD) method uti-
lized by Rabinovich et al. [33]. The experimental procedure is
detailed elsewhere [18].
Hydrogen sorption measurements were performed by holding
the potential of the electrode at −0.02, 0.00 and 0.02 V during 500 s
and subsequent potential sweep to 0.6 V in 0.1 M HClO4.
The catalytic activity in the HOR was determined using an Auto-
lab RDE in the H2-saturated electrolyte. The measurements were
carried out at the rotation rates within the 400–3600 rpm range
under the potentiodynamic mode with the sweep rate of 5 mV s−1
.
Exchange currents of the HOR were calculated from the slope of the
RDE curves in the ‘micropolarization’ potential region [34]. Addi-
tionally, the HOR over the investigated materials was studied in
the presence of 1100 ppm of CO in the hydrogen feed in order to

A.N. Simonov et al. / Electrochimica Acta 76 (2012) 344–353
347
evaluate their CO tolerance. For that, the electrolyte was saturated
with the gas mixture H2 + 1100 ppm CO during 1 h and the measure-
ments were performed potentiostatically at E = 0.1 V at the rotation
rate of the electrode of 2500 rpm with continuous supply of the gas
mixture to the cell.
All solutions were prepared using Milli-Q purified water
(18.2 Mꢂ cm) (Millipore, USA). Extra pure H2SO4 (Acros, 96%),
HClO4 (Fluka, puriss., 60%) and CuSO4·5H2O (Aldrich, 99.999%)
were used for preparation of electrolytes. Prior to each experiment,
the glassware was cleaned by soaking in H2SO4:H2O2 (1:1, v/v)
mixture and then thoroughly washed with Milli-Q water.
3. Results and discussion
The palladium–gold alloy powder prepared by reduction of the
[Pd(NH3)4][AuCl4]2 complex with H2 under mild conditions was
utilized as a model catalyst in the present study due to the following
reasons. The materials thus prepared are pure, have a well-defined
phase composition, are easy to handle and, at the same time, pos-
sess a high enough dispersion. The latter feature of the chosen
unsupported catalysts makes them suitable for the preparation
of the electrodes with a reasonable electrochemically active sur-
face area of metals using a ‘thin layer’ method (specified in Section
2.4), which is commonly used for the carbon-supported catalysts.
The absence of the carbon support for the studied catalysts is very
important, since it excludes the possibility of the uncontrollable
incorporation of carbon atoms liberated by the support into the
crystalline lattice of metals during high-temperature treatments
[18,20,21]. Finally, according to [12], the maximal specific HOR
activity of palladium alloyed with gold is attained at the Pd con-
centration of ca. 30 at.% in the alloy, which is appreciably close to
the Pd:Au ratio of 1:2 achieved for the alloys prepared from the
[Pd(NH3)4][AuCl4]2 complex.
Fig. 2. X-ray diffraction patterns of (a) monometallic Pd sample before (1) and after
(2) treatment in C2H4 at 300 ◦C, and (b) fresh PdAu2 (1) and the same sample after
treatment in C2H4 (2) and in H2 (3) at 300 ◦C. The insets on (b) show deconvolution
of the diffraction peaks registered with the as-prepared PdAu2 sample into two
symmetric components. Solid and dashed lines indicate the positions and relative
intensities of the diffraction reflections of metallic Au (PDF #4-784) and Pd (PDF
#46-1043), respectively.
As a reasonable reference material, we employed the palla-
dium metal powder, which was prepared by reduction of the
[Pd(NH3)4]Cl2 complex under the conditions identical to those used
for the synthesis of the palladium–gold alloy.
3.1. Characterization of catalysts
(HUPD, CuUPD, formation and reduction of surface oxides), as com-
pared to the response from the Vulcan XC-72 carbon co-deposited
on the electrode surface (cf. double-layer region of Pd for curves 1
and 2 on Figs. 1a, 3a and 4a) registered after treatment of the Pd cat-
alyst with ethylene can be attributed to the partial blockage of the
surface of the material with amorphous carbon formed as a result of
decomposition of C2H4 [18]. However, as mentioned in Section 2.4,
the discussion of the changes in the mass-weighted specific sur-
face areas of the studied materials on the basis of the data acquired
for the electrodes, prepared using the procedure employed in our
work, is unfair.
Phase and surface compositions of the studied catalysts deter-
mined from XRD, XPS and electrochemical measurements are
summarized in Table 1.
3.1.1. Palladium and palladium carbide
XRD pattern of the product of reduction of the [Pd(NH3)4]Cl2
complex with hydrogen at comparatively low temperature of 90 ◦C
(see Fig. 2a, curve 1) exhibits symmetric peaks of the fcc structure
˚
with the lattice constant a = 3.893 A, which is in a fair agreement
˚
with the tabulated value for Pd metal (a = 3.890 A; PDF #46-1043).
Upon treatment with ethylene, the crystal lattice of the synthe-
sized palladium powder is notably strained due to the formation of
the PdCx phase as evidenced by the shift of the diffraction peaks
towards lower angles (see Fig. 2a, curve 2) and corresponding
increase in the lattice constant (see Table 1). The a value mea-
sured for the synthesized PdCx sample corresponds to the x value
of 0.15 [20] – the highest possible amount of C atoms dissolved in
the crystalline lattice of palladium metal reported in the literature
(see Ref. [21] and references therein). Treatment of the Pd powder
with ethylene at 300 ◦C provokes no notable changes in the mean
size of crystallites measured with XRD (see Table 1).
The differences in the HUPD region of the CVs registered with the
Pd and PdC0.15 catalysts in 0.1 M H2SO4 (Fig. 1a), namely, suppres-
sion of the peaks manifesting in the 0.15–0.30 V potential range
and attributed to co-adsorption of hydrogen and (bi)-sulfate on
Pd(1 1 1) and Pd(1 0 0) facets of the particles [35], are in agreement
with the previously reported tendencies [18]. Another distinct fea-
ture of the HUPD region of the PdC0.15 sample as compared to pure
palladium is emergence of the faint peaks at ca. 0.1 V, which are
observed both in sulfuric and perchloric acids (see Figs. 1a and 3a)
and hence are caused by the processes involving hydrogen rather
than (bi)-sulfate. The peaks commencing at potentials below 0.06 V
on the cathodic wave of CVs and attributed to the superposition
of hydrogen ad- and absorption [9] are notably narrowed after
treatment of the Pd catalyst with ethylene (Figs. 1a and 3a). This
phenomenon can be explained by almost complete suppression of
CVs of the Pd and PdC0.15 catalysts in 0.1 M H2SO4, 0.1 M
HClO4 and 5 mM CuSO4 + 0.1 M H2SO4 are exemplified in
Figs. 1a, 3a and 4a, respectively. Overall suppression of the currents,
corresponding to the processes occurring on the palladium surface

348
A.N. Simonov et al. / Electrochimica Acta 76 (2012) 344–353
Table 1
Phase and surface composition of the studied samples.
c
Sample
Preparation
XRDa
XPSb
ꢁPd
a (Å)
D (Å)
Phase
[Pd0 + Pd2+]/[Au]
[Pd2+]/[Pd0]
Pd
PdC0.15
PdAu2
Reduction of [Pd(NH3)4]Cl2 with H2
3.893(1)
3.996(1)
4.057(4)
4.011(3)
4.055(4)
4.015(3)
4.011(1)
8.6
8.5
∼7
∼7
∼7
∼7
18
Pd
PdC0.15
–
–
0.63
–
–
0.46
1.0
1.0
Treatment of Pd with C2H4 at 300 ◦C during 3 h
Reduction of [Pd(NH3)4][AuCl4]2 with H2
Au0.90Pd0.10 (10 wt.%)
Au0.67Pd0.33 (90 wt.%)
Au0.89Pd0.11 (10 wt.%)
Au0.69Pd0.31 (90 wt.%)
Au0.67Pd0.33
0.36 0.04
0.35 0.04
0.25 0.05
PdAu2–Cx
Treatment of PdAu2 with C2H4 at 300 ◦C during 3 h
Treatment of PdAu2 with H2 at 300 ◦C during 3 h
0.65
0.47
0.34
0.00
PdAu2–H2
a
a – the fcc lattice constant, D – the mean crystallite size calculated with Scherrer formula from the mean value of the integral breadths of the (1 1 1), (2 0 0), (2 2 0), (3 1 1)
and (2 2 2) diffraction peaks. For the PdAu2 and PdAu2–Cx samples, the weight fractions of the two main phases comprising the samples are indicated in the brackets.
b
[Au], [Pd0], [Pd2+] – atomic concentrations of gold, palladium in metallic and oxidized state, respectively, in the top layers of the samples.
Fraction of Pd on the surface of the samples calculated as a ratio of the Pd surface area (determined by CO-stripping) and the total metal surface area (determined by
c
CuUPD).
hydrogen absorption by palladium as a result of the formation of the
PdC0.15 phase [18,21], as approved by the results of the hydrogen
sorption measurements, performed for the Pd and PdC0.15 samples
(not shown).
3.1.2. Palladium–gold alloys
Detailed investigation of the kinetics and products of reduc-
tion of the [Pd(NH3)4][AuCl4]2 complex will be reported in the
coming publication [37]. For the present study, we employed mild
conditions for the reduction of the bimetallic complex precursor
with hydrogen, which allowed us to prepare well-dispersed unsup-
ported alloys. Asymmetric shapes of the X-ray diffraction peaks
registered for the synthesized PdAu2 catalyst suggest that the sam-
ple comprises a set of the palladium–gold solid solutions (Fig. 2b,
curve 1). Analysis of the discussed peaks reveals that they can be
authentically deconvoluted into two symmetric peaks (as exem-
plified in the inset of Fig. 2b), corresponding to the target alloy
with the Pd:Au ratio of 1:2 and to the Pd0.1Au0.9 solid solution, the
former being present in predominant amount in the catalyst (see
Table 1). Low content of the Au-rich bimetallic phase in the stud-
ied sample and findings of Hayden et al. [12] who reported zero
The CuUPD curves registered with the pristine and treated with
ethylene Pd catalyst (Fig. 4a) resemble those reported for the
carbon-supported palladium nanoparticles [18,34] and exhibit the
peaks, which can be attributed to the Cu underpotential depo-
sition on the Pd(1 1 1) and Pd(1 0 0) facets [36]. Formation of
non-stoichometric Pd carbide does not provoke any fundamental
alterations of the CuUPD process on the surface of the palladium cat-
alyst except some slight variation in the relative intensities of the
peaks, which can be arguably attributed to the preferential forma-
tion of amorphous carbon deposits on the determined crystalline
planes [18].
Fig. 3. Cyclic voltammetry for the studied samples. (a) Pd (1) and PdC0.15 (2) in 0.1 M
HClO4; (b) PdAu2 (1) and PdAu2–Cx (2) in 0.1 M HClO4, PdAu2–Cx in 0.1 M H2SO4 (3)
and CO-stripping curve for PdAu2–Cx in 0.1 M H2SO4 (4). Sweep rate 10 mV s−1. Tem-
perature 25 ◦C. Currents are normalized to the total metal surface area determined
from the CuUPD measurements.
Fig. 4. Cu underpotential deposition for the studied samples measured in 5 mM
CuSO4 + 0.1 M H2SO4 and corresponding blank curves measured in 0.1 M H2SO4. (a)
Pd (1) and PdC0.15 (2); (b) PdAu2 (1), PdAu2–Cx (2) and PdAu2–H2 (3). Sweep rate
10 mV s−1. Temperature 25 ◦C. Currents are normalized to the total metal surface
area determined from the CuUPD measurements.

A.N. Simonov et al. / Electrochimica Acta 76 (2012) 344–353
349
activity of the Pd0.1Au0.9 alloy in the HOR allow us to contend that
the electrocatalytic properties of the synthesized bimetallic sam-
ple, which are discussed in the next section, are overwhelmingly
determined by the alloy with the composition of PdAu2 rather than
by the undesired Au-rich ‘admixture’.
After treatment of the PdAu2 catalyst with ethylene under the
conditions similar to those employed for the preparation of the
PdC0.15 phase from pure palladium, no remarkable changes in posi-
tions or shapes of the peaks manifesting in the XRD-pattern of
the bimetallic sample occurred (cf. curves 1 and 2 on Fig. 2b). The
detailed consideration of both raw and deconvoluted diffraction
peaks of the C2H4-treated palladium–gold alloy reveals, however,
a faint shift of their maxima towards lower angles relative to the
positions of the peaks in the X-ray diffractogram of the untreated
sample and corresponding increase in the lattice constant for the
˚
PdAu2 phase by 0.004 A (see Table 1). It should be noted that such
shift of the diffraction peaks is reproducible under the employed
conditions. On the other hand, treatment of the starting PdAu2 cat-
alyst with hydrogen under the conditions similar to those used
for treatment of the sample with C2H4 produces no shift of the
diffraction peaks (see curve 3 of Fig. 2b and Table 1). Another
vivid distinction between the PdAu2–H2 and PdAu2–Cx samples
observed with XRD lies in the values of the average sizes of crystal-
lites measured for these catalysts. It is remarkable that after heating
in C2H4 flow, the PdAu2–Cx sample still comprises two phases with
D values close to those for the untreated bimetallic powder. On the
contrary, the PdAu2–H2 catalyst contains the sole phase with the
composition of Pd0.33Au0.67 and ca. 2.5-fold larger average crystal-
lite size as compared to the PdAu2 and PdAu–Cx samples (Table 1).
The measured increase of the lattice constant of the PdAu2 cat-
alyst after heating in ethylene is incomparable to the strain of the
crystalline lattice of the monometallic Pd powder after treatment
with C2H4 under analogous conditions (Table 1) and might be,
arguably, attributed to the redistribution of the metals between
the main PdAu2 phase and the Au-rich admixture promoted by
high temperature treatment. However, it is not obvious whether
the fcc lattice constant of the palladium–gold alloy should substan-
tially increase or not as a result of presumable incorporation of
carbon. Note that, according to Ref. [23], the lattice constant of gold
nanoparticles with the typical size of 5 and 20 nm decreases and suf-
fers minor changes, respectively, as a result of carbon uptake into
their volume, in contrast to the pronounced increase in the fcc lat-
tice constant of palladium taking place after the formation of the
PdCx solid solutions [18–21]. Presumably, this distinction is caused
by the difference in the mechanisms of dissolution of carbon atoms
in palladium and gold. Thus, the absence of the notable changes in
the fcc lattice parameter of the studied PdAu2 sample after treat-
ment with ethylene neither refutes nor approves the capability of
the palladium–gold alloys to uptake carbon atoms.
Fig. 5. Typical TEM (a) and HRTEM (b) images of the PdAu2 catalyst calcined in a
C2H4 flow at 300 ◦C during 3 h.
on their surface. The chemical composition of the surface layers of
the untreated PdAu2 powder determined by XPS is in acceptable
agreement with the phase and bulk chemical composition of the
sample (Table 1). However, under the mild reduction conditions
employed in our work for the preparation of the PdAu2 alloy, some
enrichment of the surface of the alloy with palladium occurs. As
follows from the data presented in Table 1, additional reductive
treatment of the PdAu2 catalyst with H2 at 300 ◦C results in evident
depleting of palladium in the top layers of the alloy in accordance
with the documented tendency of Au to segregate at the surface
[3]. On the contrary, heating of the palladium–gold alloy under the
same temperature regime in a C2H4 flow provokes no significant
changes in the ratio of metals in the surface layers of the sample
(Table 1). For all samples the binding energy value of Au4f7/2 peak
(84.0 eV) indicates the metallic state of gold. The positions of main
peak in the Pd3d spectra of 335.1–335.2 eV (see Fig. 6) also cor-
respond to metallic palladium. The distinct shoulder in the Pd3d
spectra of the PdAu2 and PdAu–Cx samples (Fig. 6) can be assigned
to palladium oxide (PdO), most likely formed at the surface of the
granules. It should be noted that the fraction of palladium oxide
drops down noticeably after the treatment in ethylene and it fully
disappears after reduction in hydrogen (see Table 1 and Fig. 6).
The CV curves registered with the investigated PdAu2 and
PdAu2–Cx powders in the supporting electrolytes show common
features of the palladium–gold alloys (see Fig. 3b). The HUPD poten-
tial region of the CVs measured with both catalysts exhibits broad
peaks at the potentials below ca. 0.2 V and rather low currents in the
TEM analysis of the synthesized palladium–gold alloy before and
after treatment with ethylene (see Fig. 5 for PdAu2–Cx) showed
that both catalysts represent complex sponge-like structures, con-
stituted by twisted nano-sized bimetallic strings with the typical
width of 7–12 nm. No noteworthy changes in the morphology of
the PdAu2 sample were observed as a result of its heating in a
C2H4 flow. Mean values of the interplanar spacing of the (1 1 1)
plane (d111) were extracted from the HRTEM images, taken from
various regions of the studied catalysts, and were found to be
˚
2.33 0.04 and 2.32 0.04 A for PdAu2 and PdAu2–Cx, respectively.
The measured d111 values lie between the tabulated values of the
˚
(1 1 1) interplanar spacing for Pd and Au metal (2.25 and 2.35 A,
respectively) and roughly correspond to the PdAu2 stoichiometry,
in compliance with the local EDX analysis and XRD data acquired
for the discussed samples.
Analysis of the survey XPS spectra of the studied catalysts proves
the absence of the possible impurities (like chlorine and nitrogen)

350
A.N. Simonov et al. / Electrochimica Acta 76 (2012) 344–353
Fig. 6. XPS spectra of the PdAu2 powder samples: untreated (1), and after treatment at 300 ◦C in a C2H4 (2) and H2 (3) flow. Panel (a) shows raw experimental data for the
Pd3d + Au4d region, and panel (b) exhibits the Pd3d spectra after subtraction of both Au4d spectra and Shirley background.
0.2–0.4 V interval in accordance with previously reported results
for PdAu alloys with low Pd content [11,12,17]. Juxtaposition of CVs
registered with the discussed bimetallic materials in 0.1 M HClO4
and 0.1 M H2SO4 (exemplified in Fig. 3b for PdAu2–Cx) suggests that
the peaks emerging on the CVs at the potentials below 0.4 V are
not significantly affected by the (bi)-sulfate adsorption/desorption
and are predominantly determined by HUPD process. The hydro-
gen sorption measurements approved that all bimetallic catalysts
tested in the present work are incapable of notable hydrogen
absorption in accordance with the literature data for PdAu alloys
[11,38].
Analogously to HUPD, red-ox behaviors of the PdAu2 catalyst and
that of the PdAu2–Cx sample resemble each other to a great extent
(Fig. 3b). The position of the surface reduction peaks measured
for the untreated and heated in ethylene PdAu2 alloy is substan-
tially shifted to higher potentials as compared to pure palladium,
thus approving the formation of the palladium–gold alloy. In 0.1 M
H2SO4, the currents arising from the formation/reduction of the
surface oxides on the PdAu2 and PdAu2–Cx catalysts are strongly
suppressed as compared to the curves obtained in perchloric acid
(cf. curves 2 and 3 in Fig. 3b). This can be explained by the adsorp-
tion of (bi)-sulfate on the alloy surface, which secures palladium
from oxidation [32].
The CV curves recorded with the PdAu2 sample in 5 mM
CuSO4 + 0.1 M H2SO4 (Fig. 4b) exhibit a superposition of vast peaks
arising from the copper underpotential deposition on the bimetal-
lic electrode surface, which can be hardly attributed to CuUPD on
the determined crystalline facets of either palladium or gold. Anal-
ogously to the monometallic Pd catalyst, treatment of the PdAu2
powder with ethylene does not provoke fundamental changes in
the peculiarities of CuUPD on the surface of the alloy. On the con-
trary, the CV curves measured for the PdAu2–H2 catalyst in the
presence of Cu2+ cations in the electrolyte notably differ from those
of the untreated sample and exhibit a well-pronounced character-
istic pair of peaks at ca. 0.35 and 0.40 V on the cathodic and anodic
waves, respectively, which is typical for Au(1 1 1) [9,39].
alloy with different catalytic properties towards the carbon monox-
ide electrooxidation. Both maxima manifesting on the CO-stripping
curves registered with the PdAu2 catalysts are shifted to higher
potentials with respect to the position of the CO-stripping peak
for pure Pd, consistent with the findings for PdAu alloys [12] and
Pd overlayers on Au nanoparticles [9]. By comparing the relative
intensities of the maxima on the CO-stripping curves for the PdAu2,
PdAu2–Cx and PdAu2–H2 catalysts, one can again come to the con-
clusion that the treatment of the initial sample at 300 ◦C in H2 pro-
duces some changes in the composition of the alloy surface in con-
trast to the analogous treatment of the sample with ethylene. How-
ever, one should keep in mind inevitable errors arising during the
subtraction of the background CVs from the CO-stripping curves,
which might affect the shape of the resulting difference curves.
The electrochemical data obtained for the initial and pre-treated
PdAu2 samples allows estimating the surface composition of the
studied catalysts. Taking into account ambiguity in determination
of the electrochemically active surface area of Pd alloyed with
an excess of gold from the HUPD charge due to the restrictions of
HUPD on Pd-poor bimetallic ensembles [17,40], the surface area of
palladium (SPd) was estimated from the CO-stripping charge using
the conversion coefficient of 0.42 mC cm−2. The total metal surface
area (SMetal) for the catalysts was determined by the CuUPD mea-
surements as described in Section 2.4. The fractions of palladium
on the surface of the PdAu2, PdAu2–Cx and PdAu2–H2 catalysts
The CO-stripping curve registered with the PdAu2–Cx catalyst is
exemplified in Fig. 3b, while Fig. 7 features the CO-stripping curves
corrected to the blank CVs for all studied materials. A bifurcated
shape of the carbon monoxide oxidation peaks observed for the
bimetallic catalysts (Fig. 7, curves 3–5) suggests that there are at
least two types of palladium cites on the surface of the synthesized
Fig. 7. Background corrected CO-stripping curves for the studied samples: Pd (1),
PdC0.15 (2), PdAu2 (3), PdAu2–Cx (4) and PdAu2–H2 (5). Currents are normalized to
the Pd surface area calculated from the CO-stripping charge.

A.N. Simonov et al. / Electrochimica Acta 76 (2012) 344–353
351
Table 2
of the majority of contaminants (which could, possibly, poison the
catalysts) from the surface of the materials studied in the present
work. The reason of the observed discrepancy is not fully under-
stood for the moment. Possibly, it is connected with the differences
in morphology and size of the structures comprising the Pd powder
tested here and of the highly dispersed nanoparticles supported on
carbon. Also, it might be speculated that, during deposition of Pd
in a highly dispersed state on the carbon supports, the interaction
between the metal and carbon always takes place and some inde-
finable amount of C atoms can be incorporated into the surface
layers of Pd nanoparticles even during mild temperature treat-
ments employed in Refs. [9,18,34]. As a result, some enhancement
of the electrocatalytic activity of the metal for the HOR might take
place [18].
Transformation of the model Pd catalyst to the PdC0.15 phase
promotes a notable enhancement of both electrocatalytic activity
for the HOR and CO tolerance of the metal (Table 2), consistent with
the data reported in Ref. [18]. Almost the same enhancement of the
specific catalytic activity of palladium takes place as a result of its
alloying with gold and formation of the PdAu2 solid solution. The
ca. 6-fold higher j0 [mA cm−2 (Pd)] value measured for the PdAu2
sample as compared to pure monometallic Pd catalyst is in a good
agreement with the results reported in Ref. [12]. Earlier, Shmidt
et al. [41] reported the j0 value of 0.08 mA cm−2 (metal) for the
Au(1 1 1)–Pd alloy with the surface concentration of Pd = 0.38 at.%,
which is appreciably lower than the j0 value measured for the
PdAu2 alloy sample with close surface composition studied here.
Comparison of the CO tolerance tests performed by Schmidt et al.
[13,41] and those described in the present work is complicated by
the difference in the temperature maintained in the cells during
the measurements (60 vs. 25 ◦C, respectively). Nevertheless, rough
estimates can be made with the aid of the data of Ref. [10], which
reports that the HOR current densities over CO blocked PdAu and Pd
catalysts at 0.1 V vs. RHE measured at 60 ◦C outnumber those regis-
tered at 25 ◦C approximately by a factor of 10–20. However, the HOR
current density measured by Schmidt et al. [41] over Au(1 1 1)–Pd
catalyst with the surface concentration of Pd = 0.38 at.% in the pres-
ence of 1000 ppm of CO at 60 ◦C are higher than those registered
here for the PdAu2 alloy in the presence of 1100 ppm of CO at 25 ◦C
only by a factor of ca. 2.5. The observed discrepancy in both HOR
activity and CO-tolerance of the Au(1 1 1)–Pd surface alloy and bulk
Pd–Au alloy is obviously due to the structure-sensitivity of the HOR
and CO adsorption over palladium–gold catalysts [41–43].
Exchange current densities of the HOR (j0) and the H2 oxidation current densities
in the presence of 1100 ppm of CO registered after 600 s at 0.1 V vs. RHE (jH +CO) for
the studied samples normalized to the Pd and total metal surface areas.
2
Catalyst
j0 (mA cm−2
)
jH
(␮A cm−2
)
+CO
2
Pd
Total metal
Pd
0.6 0.1
12
13
75 12
10
Total metal
0.6 0.1
12
4
26
3
Pd
PdC0.15
PdAu2
PdAu2–Cx
PdAu2–H2
0.091 0.01
0.60 0.05
0.55 0.1
1.5 0.2
0.091 0.01
0.60 0.05
0.20 0.03
0.52 0.05
0.08 0.02
1
2
1
0.6
4
0.3 0.1
3
1
calculated as ꢁPd = SPd/SMetal are consistent with the results of XPS
analysis and with the changes in the electrochemical behavior of
the palladium–gold alloy after different treatments (Table 1).
Thus, an aggregate of the data derived from XRD, XPS and
electrochemical methods indicate that (i) the surface and bulk com-
positions of the initial PdAu2 catalyst are close to each other, (ii)
treatment of the PdAu2 alloy powder with ethylene at 300 ◦C pro-
vokes slight strain of the crystal lattice of the alloy and indiscernible
changes in the surface composition of the sample, while (iii) heat-
ing of the PdAu2 catalyst in a H2 flow under analogous conditions
promotes sintering of the particles and decrease in the Pd:Au ratio
on the surface as compared to the parent alloy. Obviously, changes
occurring in the palladium–gold alloy during calcination at 300 ◦C
in the presence of ethylene and in the presence of hydrogen are
strikingly different. However, the employed characterization meth-
ods do not provide direct evidence of incorporation of carbon atoms
into the crystalline lattice of the palladium–gold alloy.
3.2. Effect of pre-treatment of the Pd and PdAu2 catalysts on their
electrocatalytic properties
Table 2 summarizes the electrocatalytic properties of the stud-
ied model catalysts in terms of the exchange current densities of the
HOR (j0) and current densities of the HOR measured in the presence
of 1100 ppm CO in the hydrogen feed at the potential of 0.1 V after
600 s (jH +CO). Both j0 and jH
values are normalized to the SPd
+CO
and SMetal surface areas. Fig. 28 exhibits typical transients of the H2
electrooxidation in the presence of 1100 ppm of carbon monoxide
in the feed measured on the studied electrodes rotated at 2500 rpm.
Activity of the pure Pd catalyst tested in the present work
was found to be lower as compared to the j0 values of
0.22 0.05 mA cm−2 (Pd) determined for the carbon-supported
palladium nanoparticles by the group of Prof. Savinova [9,34] as
well as in our earlier studies [18]. Note that XPS analysis approved
the effectiveness of the employed washing procedure for removal
2
When normalized to the total metal surface area, the activ-
ity of the PdAu2 catalyst in the HOR both in the absence and in
the presence of CO is notably inferior to the performance of pal-
ladium carbide. However, after treatment of the palladium–gold
alloy with ethylene, the situation fundamentally changes (Table 2
and Fig. 8). The exchange current density of the HOR referred to
the Pd surface area measured for the PdAu2–Cx catalyst demon-
strates the enhancement by more than one order of magnitude as
compared to the j0 [mA cm−2 (Pd)] value determined for the pris-
tine monometallic Pd catalyst. Moreover, the jH +CO [␮A cm−2 (Pd)]
2
value for the C2H4-treated PdAu2 alloy is higher than that for pure
palladium by two orders of magnitude, suggesting strong suppres-
sion of the CO adsorption on the surface of Pd alloyed with gold and
then treated with ethylene. On the contrary, treatment of the PdAu2
alloy with hydrogen at 300 ◦C does not provoke any enhancement
in the HOR activity of the catalyst but rather results in its degrada-
tion (Table 2 and Fig. 8). As discussed in Section 3.1.2, the surface of
the studied palladium–gold alloy seems to be depleted with Pd as a
result of treatment of the sample with hydrogen at 300 ◦C, and this
can be a reason of its worse performance in the HOR as compared
to the untreated sample [12].
Fig. 8. Hydrogen oxidation transients measured for Pd (1), PdC0.15 (2), PdAu2 (3),
PdAu2–Cx (4) and PdAu2–H2 (5) in 0.1 M HClO4 saturated with H2 + 1100 ppm CO
gas mixture at E = 0.1 V vs. RHE, at 25 ◦C and at the rotation rate of the electrode of
2500 rpm. Currents are normalized to the total metal surface area determined from
the CuUPD measurements.
At the same time, the composition and electrochemical prop-
erties (HUPD, CuUPD, CO stripping, red-ox behavior) of the surface

352
A.N. Simonov et al. / Electrochimica Acta 76 (2012) 344–353
of the PdAu2 catalyst suffer indiscernible changes after heating at
300 ◦C in an ethylene flow (see Section 3.1 and Table 1), while
the electrocatalytic properties and tolerance to poisoning of the
material by CO are altered profoundly as a result of such treat-
ment. There are at least two possible explanations of the registered
phenomenon. First, while keeping the constant Pd:Au ratio, the
composition of the bimetallic ensembles, which directly deter-
mines the adsorption properties of the surface of the alloy [44,45],
might be changed in a specific way in the presence of ethylene
and gain unusual adsorption and electrocatalytic properties. Also,
the employed treatment procedure might hypothetically provoke
increase in the concentration of the Au edge and step cites on
the bimetallic surface – a change quite capable of providing the
observed enhancement of the CO tolerance of the palladium–gold
alloy [42,43]. However, if this is the case, one should expect to
observe at least some changes in the HUPD, CuUPD and CO-stripping
processes on the surface of the catalyst. Second possible explana-
tion of the strong enhancement of the catalytic activity towards the
HOR and of the CO tolerance is the incorporation of C atoms from the
fragmented C2H4 molecules into the structure of the PdAu2 alloy.
Note that for the Pd and PdC0.15 catalysts, which exhibit strikingly
distinct electrocatalytic properties, no fundamental differences in
of the PdAu2 powder in a H2 flow under analogous conditions
promotes sintering of the particles and decrease in the Pd:Au ratio
on the alloy surface. Treatment of the PdAu2 alloy with ethylene
entails substantial enhancement of the electrocatalytic activity and
CO tolerance of the catalyst during the HOR, as demonstrated by
the RDE measurements at 25 ◦C. At the same time, the PdAu2 alloy
subjected to the treatment with H2 under analogous temperature
regime does not exhibit enhanced exchange current density of
the HOR and CO tolerance (both normalized to the Pd and total
metal surface area) as compared to the untreated alloy sample. The
resemblance of the HUPD, CuUPD and CO-stripping processes taking
place on the surface of the untreated and C2H4-treated PdAu2 elec-
trodes suggests that the strong differences in their electrocatalytic
properties are not the consequence of the formation of specific
PdyAuz ensembles on the surface of the alloy during heating in
C2H4, but are determined by the promoting effect of the interstitial
carbon atoms. This is additionally approved by the fact that the
activity and CO tolerance of the PdAu2–Cx catalyst abruptly drops
after the mild oxidative treatment performed at 120 ◦C in air during
just 20–30 min. The nature of the promoting effect of the presumed
embedding of carbon into the crystalline structure of the alloy
cannot be irrefutably ascertained on the basis of the experimental
data reported here. Nevertheless, the outcome of the employed
electrochemical characterization methods suggests that the con-
jectural incorporated C atoms do not produce notable changes in
the surface morphology of the PdAu2 alloy. This allows speculating
that the ‘ligand’ effect of the presumptive interstitial carbon is a
plausible root of the observed enhancement of the electrocatalytic
activity of the C2H4-treated PdAu2 alloy rather than the ‘ensem-
ble’ effect (i.e., changes in the structures of the PdyAuz surfaces
ensembles caused by incorporation of C into the alloy structure).
We realize that the characterization methods employed in our
study do not provide a direct evidence of the incorporation of car-
bon atoms into the crystalline lattice of the palladium–gold alloy,
and further thorough experimental and theoretical investigation of
the properties of the ternary Pd–Au–C system is still required. Even
though the cause of the changes in the electrocatalytic properties
of the described PdAu2 sample heated in the presence of ethylene
might not be connected with the interstitial carbon, the present
study reports a simple approach for the remarkable improvement
of the HOR catalytic activity and CO tolerance of the palladium–gold
alloys. In conclusion we would like to give a reference to a coming
publication [46], which provides further evidence of the promot-
ing effect of carbon incorporated into the palladium–gold alloys
on their catalytic activity in the HOR by the example of carbon-
supported bimetallic nanoparticles.
HUPD, CuUPD and CO-stripping were observed, analogously to the
PdAu2 and PdAu2–Cx samples.
Finally, it was demonstrated in Ref. [18] that an oxidative
treatment of PdCx nanoparticles at 300 ◦C in the 5% O2 + 95% N2
atmosphere destroys the palladium carbide structures and restores
the electrocatalytic properties of unmodified palladium. This
approach was applied in the present study as a circumstantial test
for the presence of interstitial carbon in the PdAu2 alloy subjected
to heating in an ethylene flow. Since the additional treatment of
the PdAu2–Cx catalyst in the oxidative atmosphere at 300 ◦C might
disturb the ratio of elements on its surface, the destruction of the
presumptive interstitial carbon in the sample was performed at
much lower temperature of 120 ◦C on air. As confirmed by HUPD
,
CuUPD and CO-stripping (not shown), no notable changes in the
electrochemical behavior and surface composition of the PdAu2–Cx
catalyst takes place after its’ oxidation under such conditions. At
the same time, the j0 and jH
values decrease by a factor of
+CO
ca. 2 and 4 respectively after 2the mild oxidative treatment of the
PdAu2–Cx catalyst. This observation can be hardly explained in
terms of the formation of specific PdyAuz ensembles on the surface
of the palladium–gold alloy during heating in C2H4 atmosphere,
but can be considered as a rather forcible argument in support of
the promoting effect of carbon atoms incorporated in the bimetallic
catalyst.
Acknowledgements
4. Conclusion
The authors are grateful to collaborators from the Boreskov
Institute of Catalysis Dr. Pavel A. Simonov for his help in pre-
treatment of the samples and valuable advices throughout the work
and Mr. Arkadii Ischenko for performing TEM studies. The financial
support of the work by the Russian Federation President Grants
for Young Ph.D. (no. MK-380.2011.3) and for the Leading Scien-
tific Schools (NSh 524.2012.3) and by the Russian Fund for Basic
Research (Grant no. 11-03-00668) is acknowledged.
The effect of pre-treatment of the model unsupported Pd and
PdAu2 electrocatalysts with ethylene and hydrogen on the electro-
catalytic activity in the HOR and CO tolerance of these materials was
investigated. The model unsupported Pd and PdAu2 catalysts were
preparedby reductionof the [Pd(NH3)4]Cl2 and [Pd(NH3)4][AuCl4]2
complexes, respectively, with hydrogen under mild conditions.
The formation of the PdC0.15 phase as a result of heating of the
pure palladium catalyst in a C2H4 flow at 300 ◦C was testified by
XRD and electrochemical methods. The electrocatalytic properties
of the tested PdC0.15 catalyst resemble those reported earlier for
the supported PdC0.14/C catalyst [18].
According to XRD, XPS and electrochemical methods, the sur-
face and bulk compositions of the prepared PdAu2 catalyst are close
while the treatment of the sample with ethylene at 300 ◦C provokes
slight strain of the crystal lattice of the alloy and indiscernible
changes in the Pd:Au ratio on the surface. On the contrary, heating
References
[1] B. Hammer, J.K. Norskov, Advances in Catalysis 45 (2000) 71.
[2] A. Ruban, B. Hammer, P. Stolze, H.L. Skiver, J.K. Nørskov, Journal of Molecular
Catalysis A 115 (1997) 421.
[3] M. Chen, D. Kumar, C.-W. Yi, D.W. Goodman, Science 310 (2005) 291.
[4] D.I. Enache, J.K. Edwards, P. Landon, B. Solsona-Espriu, A.F. Carley, A.A. Herzing,
M. Watanabe, C.J. Kiely, D.W. Knight, G.J. Hutchings, Science 311 (2006) 362.
[5] US Patent, 5447896, B01J21/18, 05.09.1995.

A.N. Simonov et al. / Electrochimica Acta 76 (2012) 344–353
353
[6] W. Vielstich, A. Lamm, H.A. Gasteiger (Eds.), Handbook of Fuel Cells – Funda-
mentals, Technology and Applications, John Wiley & Sons, New York, 2003.
[7] J.D. Holladay, J. Hu, D.L. King, Y. Wang, Catalysis Today 139 (2009) 244.
[8] L.A. Kibler, ChemPhysChem 7 (2006) 985.
[27] A. Maeland, T.B. Flanagan, Canadian Journal of Physics 42 (1964)
2364.
[28] B.D. Cullity, Elements of X-ray Diffraction, 2nd ed., Addison-Wesley Publishing
Company, Reading, Massachusetts, USA, 1978.
[9] P.S. Ruvinsky, S.N. Pronkin, V.I. Zaikovskii, P. Bernhardt, E.R. Savinova, Physical
Chemistry Chemical Physics 10 (2008) 6665.
[29] M. Oezaslan, F. Hasche, P. Strasser, Chemistry of Materials 23 (2011)
2159.
[10] A.N. Simonov, P.A. Pyrjaev, B.L. Moroz, V.I. Bukhtiyarov, V.N. Parmon, Electro-
catalysis, http://dx.doi.org/10.1007/s12678-012-0084-3.
[11] Y. Pluntke, L.A. Kibler, D.M. Kolb, Physical Chemistry Chemical Physics 10 (2008)
3684.
[12] F.A. Al-Odail, A. Anastasopoulos, B.E. Hayden, Physical Chemistry Chemical
Physics 12 (2010) 11398.
[13] T.J. Schmidt, Z. Jusys, H.A. Gesteiger, R.J. Behm, U. Endruschat, H. Boennemann,
Journal of Electroanalytical Chemistry 501 (2001) 132.
[14] L.A. Kibler, A.M. El-Aziz, D.M. Kolb, Journal of Molecular Catalysis A 199 (2003)
57.
[15] S. Pandelov, U. Stimming, Electrochimica Acta 52 (2007) 5548.
[16] P. Liu, J.K. Nørskov, Physical Chemistry Chemical Physics 3 (2001) 3814.
[17] F. Maroun, F. Ozanam, O.M. Magnussen, R.J. Behm, Science 293 (2001) 1811.
[18] A.N. Simonov, P.A. Pyrjaev, P.A. Simonov, B.L. Moroz, S.V. Cherepanova, D.A.
Zyuzin, V.I. Bukhtiyarov, V.N. Parmon, Journal of Molecular Catalysis A 353–354
(2012) 204.
[30] S. Krumm, Materials Science Forum 228–231 (1996) 183.
[31] D. Briggs, M.P. Seah (Eds.), Practical Surface Analysis – Auger and X-ray Photo-
electron Spectroscopy, 2nd ed., Wiley Interscience, USA, 1990.
[32] M. Tian, B.E. Conway, Journal of Electroanalytical Chemistry 581 (2005)
176.
[33] L. Rabinovich, O. Lev, G.A. Tsirlina, Journal of Electroanalytical Chemistry 466
(1999) 45.
[34] S.N. Pronkin, A. Bonnefont, P.S. Ruvinsky, E.R. Savinova, Electrochimica Acta 55
(2010) 3312.
[35] M. Hara, U. Linke, Th. Wandlowski, Electrochimica Acta 52 (2007) 5733.
[36] P.M. Rigano, C. Mayer, T. Chierchie, Electrochimica Acta 35 (1990) 1189.
[37] Y. Shubin, P. Plyusnin, M. Sharafutdinov, In situ synchrotron study of Au-Pd
nanoporous alloy formation by single-source precursor thermolysis, Nanotech-
nology (2012) under review.
[38] K. Hubkowska, M. Łukaszewski, A. Czerwinski, Electrochimica Acta 56 (2010)
235.
[19] D. Teschner, J. Borsodi, A. Wootsch, Z. Révay, M. Hävecker, A. Knop-Gericke, S.D.
Jackson, R. Schlögl, Science 320 (2008) 86.
[39] A. Kuzume, E. Herrero, J.M. Feliu, R.J. Nichols, D.J. Schiffrin, Journal of Electro-
analytical Chemistry 570 (2004) 157.
[20] P. Canton, C. Meneghini, P. Riello, A. Balerna, A. Benedetti, Journal of Physical
Chemistry B 105 (2001) 8088.
[21] W. Vogel, Journal of Physical Chemistry C 115 (2011) 1506.
[22] D. Takagi, Y. Kobayashi, H. Hibino, S. Suzuki, Y. Homma, Nano Letters 8 (2008)
832.
[23] E.A. Sutter, P.W. Sutter, Journal of Materials Science 46 (2011) 7090.
[24] H. Okamoto, T.B. Massalski, Bulletin of Alloy Phase Diagrams 5 (1984) 378.
[25] T. Nedoseykina, P. Plyusnin, Yu Shubin, S. Korenev, Journal of Thermal Analysis
and Calorimetry 102 (2010) 703.
[40] F. Hernandez, H. Baltruschat, Langmuir 22 (2006) 4877.
[41] T.J. Schmidt, V. Stamenkovic, N.M. Markovic, P.N. Ross Jr., Electrochimica Acta
48 (2003) 3823.
[42] B. Gleich, M. Ruff, R.J. Behm, Surface Science 386 (1997) 48.
[43] M. Ruff, S. Frey, B. Gleich, R.J. Behm, Applied Physics A 66 (1998) S513.
[44] C.-W. Yi, K. Luo, T. Wei, D.W. Goodman, Journal of Physical Chemistry B 109
(2005) 18535.
[45] T. Wei, J. Wang, D.W. Goodman, Journal of Physical Chemistry C 111 (2007)
8781.
[26] W. Kraus, G. Nolze, POWDERCELL 2.4, Program for the Representation and
Manipulation of Crystal Structures and Calculation of the Resulting X-ray
Powder Patterns, Federal Institute for Materials Research and Testing, Berlin,
Germany, 2000.
[46] P.A. Pyrjaev, A.N. Simonov, B.L. Moroz, D.A. Zyuzin, E.Yu. Gerasimov, I.P.
Prosvirin, T.I. Nedoseykina, V.I. Bukhtiyarov, V.N. Parmon, Journal of Molecular
Catalysis A, in preparation.