Electrochemical and structural characterization of carbon-supported Pt-Pd bimetallic electrocatalysts prepared by electroless deposition

Article abstract of DOI:10.1016/j.electacta.2010.06.068

Electrochemical and structural characteristics of various Pt-Pd/C bimetallic catalysts prepared by electroless deposition (ED) methods have been investigated. Structural analysis was conducted by X-ray diffraction spectroscopy, X-ray photoelectron spectroscopy, scanning transmission electron microscopy, and energy dispersive X-ray spectroscopy (EDS). Monometallic Pt or Pd particles were not detected by EDS, indicating the ED methodology formed only bimetallic particles. The size of the Pt-Pd bimetallic particles was smaller than those of a commercially available Pt/C catalyst. The morphology of the Pt on Pd/C catalysts was identified and corresponded to Pd particles partially encapsulated by Pt. The electrochemical characteristics of the lowest Pd loading catalyst (7.0% Pt on 0.5% Pd/C) for the oxygen reduction reaction (ORR) have been investigated by the rotating ring disk electrode technique. The electrochemical activity was equal or lower than the commercially available Pt/C catalyst; however, the amount of hydrogen peroxide observed at the ring was reduced by the Pd, suggesting that such a catalyst has the potential to decrease ionomer degradation in applications. The Pt on Pd/C catalysts also show a higher tolerance to ripening induced by potential cycling. Therefore, catalyst suitability cannot be judged solely by its initial performance; information related to specific degradation mechanisms is also needed for a more complete assessment.

Full text of DOI:10.1016/j.electacta.2010.06.068

Electrochimica Acta 55 (2010) 7376–7384
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Electrochemical and structural characterization of carbon-supported Pt–Pd
bimetallic electrocatalysts prepared by electroless deposition
Masato Ohashi, Kevin D. Beard, Shuguo Ma, Douglas A. Blom, Jean St-Pierre,
John W. Van Zee, John R. Monnier∗
Department of Chemical Engineering, University of South Carolina, Columbia, SC 29208, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 3 May 2010
Received in revised form 25 June 2010
Accepted 28 June 2010
Available online 3 August 2010
Electrochemical and structural characteristics of various Pt–Pd/C bimetallic catalysts prepared by elec-
troless deposition (ED) methods have been investigated. Structural analysis was conducted by X-ray
diffraction spectroscopy, X-ray photoelectron spectroscopy, scanning transmission electron microscopy,
and energy dispersive X-ray spectroscopy (EDS). Monometallic Pt or Pd particles were not detected by
EDS, indicating the ED methodology formed only bimetallic particles. The size of the Pt–Pd bimetallic
particles was smaller than those of a commercially available Pt/C catalyst. The morphology of the Pt on
Pd/C catalysts was identified and corresponded to Pd particles partially encapsulated by Pt.
The electrochemical characteristics of the lowest Pd loading catalyst (7.0% Pt on 0.5% Pd/C) for the
oxygen reduction reaction (ORR) have been investigated by the rotating ring disk electrode technique.
The electrochemical activity was equal or lower than the commercially available Pt/C catalyst; however,
the amount of hydrogen peroxide observed at the ring was reduced by the Pd, suggesting that such a
catalyst has the potential to decrease ionomer degradation in applications. The Pt on Pd/C catalysts also
show a higher tolerance to ripening induced by potential cycling. Therefore, catalyst suitability cannot
be judged solely by its initial performance; information related to specific degradation mechanisms is
also needed for a more complete assessment.
Keywords:
Bimetallic catalyst
Palladium
Platinum
Spectroscopy
Oxygen reduction reaction (ORR) activity
Durability
© 2010 Elsevier Ltd. All rights reserved.
1. Introduction
stability, and catalyst composition very difficult. Electroless depo-
sition (ED), however, is a method that produces true bimetallic
Bimetallic catalysts have become increasingly more important
in applications of heterogeneous catalysis ranging from petroleum
reforming to fuel cells [1–4]. Research has shown that bimetal-
lic catalysts often exhibit notably different catalytic and chemical
properties than their corresponding monometallic components [5].
Bimetallic systems often provide enhanced selectivity, stability,
and/or activity [6,7]. Several mechanisms to explain the catalytic
properties of bimetallic catalysts have been proposed, including
geometric or ensemble effects [8], formation of bifunctional sur-
faces [9], and electronic modification of surface sites [10].
Bimetallic catalysts are typically prepared using either suc-
cessive impregnation or co-impregnation methods [11]. Unfortu-
nately, these traditional methodologies provide inadequate control
over metal placement and accordingly yield catalysts containing
both isolated, monometallic particles and bimetallic particles with
varying compositions [12–14]. This complex mixture of particles
results in poor control of the final catalyst performance and makes
any definitive correlations between catalyst performance, catalyst
catalysts since it is a catalytic or autocatalytic process whereby a
chemical reducing agent reduces a metallic salt onto specific sites
of a pre-existing catalytic surface [15,16]. This targeted deposition
results in the formation of only bimetallic catalysts, and not the
wide range of compositions associated with the traditional meth-
ods of catalyst preparation.
Improvements in catalyst preparation are critical with respect
to reducing the cost of fuel cells, especially polymer electrolyte
membrane fuel cells (PEMFC), which have been proposed as an
alternative to internal combustion engines [17]. Key to this devel-
opment is reduction in the cost of the platinum electrocatalysts
used in PEM fuel cells. There have been several approaches to mini-
mize the amount of Pt in these fuel cells. A number of design issues
must be considered since the chemical environment is corrosive
during PEMFC operation. Further, sintering and poisoning during
the operation decrease the activity of any catalysts and also must
be considered in catalyst design.
Beard et al. [18,19] have targeted the catalyst for the oxygen
reduction reaction (ORR) of the PEMFC, since more platinum is
required for this reaction than at the anode where oxidation of H2
takes place. The authors have developed an ED methodology for
Pt deposition to prepare Pt–Rh and Pt–Pd bimetallic catalysts. In
∗
Corresponding author. Tel.: +1 803 777 6813; fax: +1 803 777 8142.
E-mail address: monnier@cec.sc.edu (J.R. Monnier).
0013-4686/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2010.06.068

M. Ohashi et al. / Electrochimica Acta 55 (2010) 7376–7384
7377
both cases Pt was catalytically deposited on Rh and Pd seed sites,
respectively, to give more highly dispersed, Pt-containing particles.
Thus, lower amounts of Pt might be required to achieve the same
concentration of surface Pt sites typically obtained by using higher
Pt weight loadings. In these studies the primary goal was to use
the controlled deposition of Pt to form smaller particles. In a more
recent study, Beard et al. [20] prepared a series of Pt–Co/carbon
bimetallic catalysts in which the Pt was deposited as a shell over
the Co core to give both a more efficient use of the Pt component
as well as to form Co-modified, Pt sites that have been reported to
be more active than Pt alone for the ORR [21,22].
Recently, a number of studies about the electrochemical activ-
ity of Pt–Pd bimetallic catalysts for ORR have been published.
The ORR activity of single crystal palladium surfaces were stud-
ied by Hoshi et al. [23] and Kondo et al. [24]. In comparison
to the platinum surfaces where activities were in the order of
(1 0 0) < (1 1 0) ≈ (1 1 1) [25], the corresponding order for palladium
surfaces was (1 1 0) < (1 1 1) < (1 0 0) in perchloric acid solution. Fur-
ther, the activity of palladium (1 0 0) was higher than platinum
(1 1 1). According to Zhang et al., commercially available palladium
catalysts (E-TEK 10% and 20% Pd/C) have higher ORR activity than
the corresponding platinum catalyst (E-TEK 10% Pt/C) when the
surface of palladium is covered by a platinum monolayer [26]; this
observation has been reproduced and analyzed by Hoshi et al. [23].
Wang et al. have used computational methods to determine effects
of various monolayer coverages of Pt on Pd core structures and con-
cluded that the optimal coverage of Pt on Pd (1 1 1) surfaces is on
the order of two monolayers [27].
JCPDS reference spectra using JADE (Materials Data, Inc.) software.
The radiation source was Cu K␣ (ꢀ = 1.5405 a˚ ) at operating condi-
tions of 30 kV and 15 mA. All spectra were taken at a scan rate of
1.5◦/min and sampling width of 0.05◦.
XPS measurements were performed with a Kratos AXIS Ultra
DLD XPS system equipped with a hemispherical energy analyzer
and a monochromatic Al K␣ source. A catalyst pretreatment cell,
attached to the UHV system permitted samples to be pretreated in
flowing H2 (UHP grade) at elevated temperatures before being ana-
lyzed by XPS. These samples were then transferred into the main
chamber for XPS analysis without exposure to air. The monochro-
matic Al K␣ source was operated at 15 keV and 225 W, incident
at 45◦ with respect to the surface normal. The pass energy was
fixed at 40 eV for the detailed scans. All the samples were present
as powders supported on a gold-plated stainless steel stub for the
XPS measurements. XPS measurements were first conducted on
the samples prior to in situ H2 treatment and then transferred to
the catalysis cell for a specified reduction treatment. Peak fitting
was conducted using XPSPEAK 4.0 using the sum of Gaussian and
Lorentzian functions as the model function. The spin–orbit splitting
of Pd 3d was fixed at 5.32 and 5.29 for Pd2+ and Pd0, respectively,
and the peak area ratio of Pd 3d3/2 to Pd 3d5/2 was constrained to
0.69 based on the ionization cross sections calculated by Scofield
[33]. All spectra were deconvoluted into Pd2+ and Pd0 spectra using
XPSPEAK 4.0 software assuming Pd peaks for oxidized (Pd2+) and
metallic Pd were located at binding energies (BE) of 336.7 and
337.1 eV for the Pd 3d5/2 orbitals, respectively [34].
STEM images were taken using a JEOL 2100F 200 kV FEG-
STEM/TEM equipped with a CEOS Cs corrector on the illumination
system. The geometrical aberrations were measured and controlled
to provide less than a ␲/4 phase shift of the incoming electron
wave over the probe-defining aperture of 15.5 mrad. High angle
annular dark-field (HAADF) STEM images were acquired on a Fis-
chione Model 3000 HAADF detector with a camera length such that
the inner cut-off angle of the detector was 50 mrad. The scanning
acquisition was synchronized to the 60 Hz AC electrical power to
minimize 60 Hz noise in the images and a pixel dwell time of 16 ␮s
was chosen. A solid-state Si(Li) X-ray detector from Oxford Instru-
ments was used to collect the energy dispersed X-ray (EDS) data.
Quantification was carried out via the Inca TEM software which
uses theoretical k-factors for its thin-film microanalysis algorithm.
The preparation of the Pt–Pd/C samples used in this work has
been described previously [19] but is summarized briefly here. The
carbon support, Vulcan XC-72 (254 m2/g, Cabot Corporation) was
cleaned and oxidized in 5 M HNO3 for 2 h at 50 ◦C. After rinsing
in de-ionized water and drying in vacuo, the carbon support was
placed in a pH 14 bath for 1 h at 50 ◦C to deprotonate the surface
carboxylic acid groups to form surface carboxylate anionic sites.
The support was then impregnated with a solution of Pd(C2H3O2)2
dissolved in CH2Cl2 and the excess solvent was removed by rotary
evaporation at 30 ◦C. After impregnation, the monometallic Pd/C
catalysts were reduced in a NaBH4 bath at 25 ◦C with a 10-fold
molar excess of NaBH4. Platinum was then electrolessly deposited
on the Pd/C catalyst using an aqueous ED bath composed of H2PtCl6,
dimethylamine borane (DMAB), and sodium citrate. The bath com-
ponents were in a 1:5:5 molar ratio for PtCl62−:DMAB:citrate at
an initial pH of 11 (controlled by NaOH). Deposition temperature
was maintained at 80 ◦C and deposition was conducted for 30 min.
Previous work [18,19] had shown that deposition was essentially
completed within this period. The maximum, theoretical Pt load-
ing was 8.4% Pt for all compositions based on the initial amount of
Other studies for Pt–Pd catalysts have been conducted sepa-
rately by Li et al. [28,29] for partial coverages of Pt on a Pd core. A
mixture of Pd and Pt compounds were decomposed and deposited
on the surface of carbon particles by a thermal reduction process. H.
Li et al. obtained higher ORR activity with (Pt–Pd)/C mixture than
Pt/C but with a higher ratio of Pt:Pd [29]. A different technique to
prepare palladium and platinum mixtures on the electrode surface
has been developed by Ye and Crooks [30]. Both Pd and Pt met-
als were reduced on the glassy carbon electrode from metal ions
exchanged into and encapsulated by a dendrimer template. They
observed higher activity for the Pt–Pd system (relative to pure Pt)
and concluded there may be electrochemical enhancement of Pt
by Pd in the bimetallic systems. Finally, in a different study, Peng
and Yang [31] observed that in an argon-blanketed HClO4 solution,
the agglomeration of Pt particles was suppressed by addition of
Pd. In a similar inert gas phase environment, Gu et al. [32] used
voltage cycling to study the accelerated degradation of ORR by Pt
agglomeration for a Pt/carbon catalyst. For the case of Pt alone, ORR
degradation was easily observed by repeated voltage cycling.
In this paper, we present a detailed analysis of the structure of
the bimetallic Pt–Pd compositions previously prepared by Beard et
al. [19] to better determine the distribution of the Pt component,
specifically whether the Pt–Pd particles exist as a mixture, a solid
solution, or as a Pd core-Pt shell composition, all of which have
ramifications for performance in PEMFCs. The analytical methods
used in this study include X-ray diffraction spectroscopy (XRD),
X-ray photoelectron spectroscopy (XPS), scanning transmission
electron microscopy (STEM), and energy dispersive X-ray spec-
troscopy (EDS). Finally, the ORR electrochemical trends of selected
Pt–Pd/C catalysts are determined by using a rotating ring disk elec-
trode (RRDE) and the results are correlated with the structural
composition.
2−
PtCl6 in the ED bath.
2. Experimental
A series of Pd/C (0.5, 1.0, 2.5, and 5.0 wt%) catalysts was prepared
and 7.0–8.4 wt% Pt (determined by AA, Perkin Elmer model 3300)
was electrolessly deposited on each of the different Pd/C samples
to provide Pt–Pd samples with a wide range of Pd weight loadings
Powder XRD measurements were made using a Rigaku MiniFlex
II bench-top system to examine the lattice structures of the Pt–Pd/C,
Pd/C, and carbon components. The XRD patterns were compared to

7378
M. Ohashi et al. / Electrochimica Acta 55 (2010) 7376–7384
Table 1
at relatively constant amounts of Pt. This preparation resulted in
atomic Pt:Pd ratios of 7.6, 4.6, 1.8, and 0.9 for Pd loadings of 0.5,
1.0, 2.5, and 5.0 wt%, respectively.
Compositions and averages of particle diameters of Pt–Pd bimetallic compositions.
Diameters determined by TEM analysis (Hitachi Model H-8000). Data taken from
Beard et al. [19].
Because XRD and XPS indicated that during storage at ambient
conditions the Pd0 in both the Pd/C and Pt–Pd/C catalysts slowly
underwent oxidation to Pd2+, presumably by formation of a surface
layer of PdO, samples were re-reduced with DMAB prior to XRD
and STEM analyses. The DMAB treatment procedure was the same
Sample #
Sample detail
Pt:Pd (moles)
Average particle
diameter (Å)
1
2
3
4
7.0% Pt on 0.5% Pd/C
8.4% Pt on 1.0% Pd/C
8.4% Pt on 2.5% Pd/C
8.4% Pt on 5.0% Pd/C
1:0.13
1:0.22
1:0.55
1:1.09
35
32
30
32
as used for the initial preparation of the Pt–Pd/C catalysts, with the
2−
exception that PtCl6
was not present in the solution. Thus, the
bath composition was a 1:1 molar ratio of DMAB:citrate at an initial
pH of 11 (controlled by NaOH). The pretreatment was conducted
at 80 ◦C for 30 min and was repeated two more times at the same
conditions to ensure complete reduction of Pd2+ to Pd0.
bility test, the Pt on Pd/C catalyst was carefully removed from the
disk and analyzed by STEM.
Electrochemical characterization was conducted using a rotat-
ing ring disk electrode (RRDE) with a glassy carbon disk and
platinum ring. The disk area was 0.2475 cm2 and the ring had a
collection efficiency of 37%. Four types of catalytic films were pre-
pared on the disk from two different catalysts; 7.0% Pt on 0.5% Pd/C
and commercially available E-TEK 20% Pt/C, and two different load-
ings of Pt mass, 20 ␮g and 2.4 ␮g. The catalyst films were fixed with
a 10 or 5 mL aliquot of a 20:1 isopropyl alcohol:nafion solution and
subsequent evaporation of the alcohol at room temperature. The
electrochemical cell consisted of a five necked flask with a Princeton
Applied Research mercury sulfate electrode as reference and plat-
inum wire as counter-electrode. The electrolyte was 0.1 M HClO4. A
Pine Instruments AFCBP1 bipotentiostat was connected to the elec-
trochemical cell. The ORR activity was performed in the electrolyte
with saturated oxygen at 1600 rpm and positively swept at 5 mV/s.
Cyclic voltammetry (CV) and background currents were measured
in the electrolyte saturated with argon. A durability test was also
performed in the argon-saturated electrolyte. This test consisted of
30,000 cyclic sweeps from 1.0 to 0.6 V at 100 mV/s and 1600 rpm.
CV and ORR activity measurements were taken before and after the
cyclic sweeps [31]. All voltages were converted and are reported
relative to the standard hydrogen electrode (SHE). After the dura-
3. Results and discussion
As shown previously [19], increasing the loading of Pd from 0.5%
up to 5.0% favored nucleation of additional Pd particles on the car-
bon support, rather than Pd particle growth. Thus, if ED of PtCl6
2−
occurred on all particles at approximately the same rate, the diam-
eters of the final Pt–Pd particles should be smaller for Pt deposited
on the 5% Pd/carbon sample since approximately 8 wt% Pt was
deposited in all cases. The data in Table 1 summarized these results
from our earlier work. Unfortunately, the resolution of the older
HRTEM was not adequate to unambiguously confirm the smaller
particle diameters for Pt deposited on the higher weight loadings
of Pd on the carbon support, although the trend in Table 1 does
corroborate this conclusion. The structure or morphology of the
electrolessly deposited Pt on the Pd surface was not determined.
Thus, it is not known whether the particles exist as a Pt shell over
a Pd core, a Pt-enriched bimetallic surface, or a Pt and Pd surface
mixture comparable to that of the bulk. Catalytic activities for both
fuel cell and conventional applications may be dependent on which
of these compositions is predominant.
In order to determine the bulk compositions of the Pt–Pd par-
ticles, powder XRD spectra for the 0.5%, 1.0%, 2.5%, and 5.0% Pd/C
Fig. 1. XRD spectra of Pt on Pd/C and Pd/C for the different Pd loadings [0.5% (bottom right), 1.0% (top right), 2.5% (bottom left), 5.0% (top left)] and carbon support (Vulcan
XC-72).

M. Ohashi et al. / Electrochimica Acta 55 (2010) 7376–7384
7379
were chosen since they gave the most intense XRD spectra for Pd
species. However, similar behavior was also observed for the 2.5%
Pd/C and 8.4% Pt–2.5% Pd/C samples. For both the 5% Pd/C and 8.4%
Pt–5% Pd/C samples, the PdO peak at 35◦ disappeared after DMAB
treatment. Simultaneously, the peak intensities at 40◦ for both sam-
ples increase sharply, indicating the addition of Pd0 peak intensity
to the Pt0 component. In fact, after normalization to the carbon
peak area at 25◦, the integrated area for this 40◦ peak after DMAB
treatment is equal to the sum of the peak areas for the 8.4% Pt–5%
Pd/C (before DMAB) and the 5% Pd/C (after DMAB). This indicates
that the increase of the peak area for the 8.4% Pt–5% Pd/C sample
after DMAB treatment was due to reduction of PdO to Pd0. Since
there was virtually no XRD-detectable Pd0 in the 5% Pd/C (before
DMAB), we conclude that the area of the peak at 40◦ was from
basically only Pt0 for the 8.4% Pt–5% Pd/C (before DMAB). This con-
clusion indicates the relative ease of oxidation of Pd in this series
of Pt–Pd bimetallic particles. Further, the observation that PdO can
be reversibly reduced back to Pd0 argues against the existence of a
solid solution of Pt and Pd, which should not be expected to have
reversible PdO redox characteristics.
Fig. 2. XRD spectra of 8.4% Pt on 5.0% Pd/C and 5.0% Pd/C before and after DMAB
treatment.
catalysts both before and after deposition of Pt are shown in Fig. 1.
The XRD spectrum for the XC-72 carbon support was also taken
and is shown as a reference. As indicated in the figure, the peaks at
approximately 25◦ and 44◦ are from carbon and the peaks at 40◦ and
46◦ are from the (1 1 1) and (2 0 0) planes, respectively, for both Pt0
and Pd0. Because both Pt0 and Pd0 exist in face centered cubic lat-
tices and the unit cell dimensions are 3.92 a˚ for platinum and 3.89 a˚
for palladium [18], it is not possible to distinguish between these
two metals. It also means that XRD is not able to determine whether
a solid solution, or alloy, is formed when Pt is deposited on Pd,
since solid solutions typically exhibit XRD reflections intermediate
between the reflections of the two separate metals. An additional
broad peak is also observed at 35◦ for the 2.5% Pd/C and 5.0% Pd/C
samples, which corresponds to PdO. This peak was not observed
for either the 1.0 wt% Pd/C or 0.5 wt% Pd/C samples because of the
lack of sensitivity at lower Pd loadings. The preferential formation
of PdO at ambient conditions for the supported Pd/C catalysts indi-
cates oxidation of Pd to PdO is quite facile. Even after deposition of
Pt on 5% Pd/C, there is still a shoulder of PdO at 35◦. The appear-
ance of the strong peak at 40◦ indicates existence of metallic Pt, and
possibly Pd0, although some of the Pd remains oxidized. Since elec-
troless deposition was conducted in a solution containing DMAB as
a reducing agent and because ED requires Pd to exist in the metallic
state to catalytically activate the DMAB, the existence of PdO after
Pt deposition indicates accessibility of O2 to the Pd surface. This
suggests that the Pt–Pd bimetallic particles exist either as a physi-
cal mixture of Pt and Pd aggregates in a bimetallic particle or as an
O2-permeable shell of Pt over a Pd core. While the existence of a
solid solution of Pd and Pt cannot be ruled out, the existence of an
extended lattice of PdO suggests otherwise.
Since the XRD peak at 40◦ for the Pt–Pd samples (before DMAB
treatment) corresponded to essentially only Pt0 and for the Pd/C
samples (after DMAB treatment), the 40◦ peak represented Pd0,
the Scherrer equation can be used (shown below) to calculate aver-
age Pt and Pd particle diameters. These results are summarized in
Table 2. For the Scherrer equation, ꢀ (Å) is the wave length of Cu
K␣ radiation (1.5405 a˚ ), B is FWHM (full width of peak intensity at
half maximum) in radians for the 40◦ peak, ꢁB is the peak position
in degrees, and t (Å) is the particle size.
0.9ꢀ
B cos ꢁB
t =
(1)
Results for Pd/C indicate that Pd particle sizes are approximately
3 nm for all samples, regardless of Pd weight loadings, confirming
that nucleation of additional Pd sites is favored over Pd particle
growth as Pd weight loadings on carbon are increased. Equiva-
lent Pt particle sizes for approximately equal weight loadings of Pt
deposited on Pd nuclei are consistent with higher concentrations of
Pd particles as Pd weight loadings increase. In general, equivalent
Pt diameters decrease as the concentration of Pd nuclei increases.
X-ray photoelectron spectroscopy (XPS) has also been used to
examine the surface and near surface regions of the Pt–Pd bimetal-
lic particles. Seah and Dench [35] have estimated the photoelectron
escape depth to be approximately 6–7 monolayers for elements
at binding energies of approximately 330 eV. Thus XPS gives both
surface and near-surface composition results.
Fig. 3 shows the XPS spectra of the Pd 3d orbitals for the 0.5%,
1.0%, 2.5%, and 5.0% Pd/C catalysts before and after Pt deposition.
Deconvolution of the different Pd/C spectra reveals a predominance
of Pd2+, implying that the surface Pd is easily oxidized for samples
stored at ambient conditions. Ho et al. [36] has shown that at room
temperature oxygen adsorption on Pd surfaces is restricted primar-
ily to the surface region with minimal penetration into the bulk. The
data in Table 3 reveal Pd0 fractions are quite similar for all Pd/C cat-
The 5% Pd/C and corresponding Pt–Pd/C catalyst were exposed
to DMAB at 80 ◦C to better understand the effects of re-reduction on
these compositions. The XRD spectra for these samples are shown
in Fig. 2 and are compared with the spectra of the 5% Pd/C and 8.4%
Pt–5% Pd/C samples before DMAB treatment. The 5% Pd/C samples
Table 2
Summary of XRD analysis of Pt–Pd/C samples before DMAB treatment and Pd/C samples after DMAB treatment.
#
Sample
Analyzed peak
Position 2␪(◦)
FWHM 2␪(◦)
Effective particle size (Å)
1
2
3
4
5
6
7
8
7.0% Pt on 0.5% Pd/C
8.4% Pt on 1.0% Pd/C
8.4% Pt on 2.5% Pd/C
8.4% Pt on 5.0% Pd/C
0.5% Pd/C
1.0% Pd/C
2.5% Pd/C
5.0% Pd/C
Pt0
Pt0
Pt0
Pt0
Pd0
Pd0
Pd0
Pd0
39.7
39.7
39.6
39.6
39.3
39.3
39.3
39.3
3.5
3.0
4.4
4.7
2.8
2.8
2.8
2.8
24
28
19
18
30
30
30
30

7380
M. Ohashi et al. / Electrochimica Acta 55 (2010) 7376–7384
Fig. 3. XPS spectra of Pd 3d orbitals (5/2 and 3/2 spin states) before (bottom plot in each chart) and after ED of approximately 7–8% Pt (top plot of each chart) at different
Pd weight loadings [0.5% (top left), 1.0% (top right), 2.5% (bottom left), 5.0% (bottom right)]. All spectra have been deconvoluted into Pd0 (dashed line) and Pd2+ (solid line)
regions.
alysts, again indicating similar Pd particle sizes for these samples.
Similar analyses for the analogous Pt–Pd/C bimetallic compositions
(with no further reduction after ED) in Fig. 3 indicate the existence
of both Pd0 and Pd2+. The data in Table 3 show that the fraction of
Pd0 decreases with increasing Pd weight loading, suggesting that
some of the Pt exists as a shell over a Pd core and that the Pt shell
becomes thinner as the Pd loading increases. If a protective and
continuous Pt shell is formed after Pd reduction in the ED bath, Pd
should be more likely to remain in the metallic state. Further, the
increase in Pd2+ 3d peak intensity (as a function of Pd weight load-
ing) after Pt deposition is consistent with the concentration of Pd
surface sites available for the ED of Pt, further suggesting that at
least some of the Pt exists as a shell over the Pd core. Thus, Pd 3d
peaks were significantly more suppressed by Pt for the 0.5% and
1.0% Pd/C samples than for the 2.5% and 5.0% Pd/C samples.
normalized to C1 s intensities. Values are 62%, 56%, 28% and 8%,
respectively, for Pt on 0.5%, 1.0%, 2.5% and 5.0% Pd/C. The decrease
in peak intensity (compared to the Pd 3d peak intensities for the
analogous Pd/C samples) is highest for the 0.5% Pd/C precursor
and lowest for the 5.0% Pd/C sample, which is consistent with the
hypothesis that thickness of the electrolessly deposited Pt varies
inversely with Pd weight loading. This correlation does not nec-
essarily prove that Pt is exclusively deposited as a shell over a Pd
core, but could also be explained as incomplete layers of Pt on the
Pd surface. For the latter case, the exposed Pd fraction would be
lowest for the 0.5% Pd/C sample and highest for the 5.0% Pd/C.
STEM analyses were conducted on 8.4% Pt–5% Pd/C before
and after the DMAB treatment as well as 8.4% Pt–1% Pd/C before
DMAB treatment to better determine the morphology of the Pt–Pd
particles. Fig. 4 shows HAADF images of the samples at low mag-
nifications. For all three samples, the Pt–Pd particles are well
distributed over the carbon surface. As expected, there were fewer
particles forthe 1% Pd/C weightloadingthan forthe5% Pd/C loading.
Comparison between the 8.4% Pt–5% Pd/C before and after DMAB
treatment shows the absence of obvious changes in particle size
distribution. Thus, it is concluded that the reversible, partial oxi-
dation of the Pd component has no effect on the distribution of
particle sizes.
Finally, the decrease in Pd 3d peak intensity after deposition of
Pt on different Pd/C samples is calculated. Pd peak intensities are
Table 3
Fraction of metallic Pd at the surface/near surface of Pt–Pd/C catalysts and corre-
sponding Pd/C precursor catalysts. No in situ pretreatments before XPS analysis.
0
Pd
#
Sample
0
2+
Pd +Pd
1
2
3
4
5
6
7
8
7.0% Pt on 0.5% Pd/C
8.4% Pt on 1.0% Pd/C
8.4% Pt on 2.5% Pd/C
8.4% Pt on 5.0% Pd/C
0.5% Pd/C
1.0% Pd/C
2.5% Pd/C
5.0% Pd/C
0.72
0.64
0.55
0.35
0.16
0.14
0.12
0.18
High resolution micrographs and EDS analyses of 8.4% Pt–5%
Pd/C (after DMAB reduction) and 8.4% Pt–1% Pd/C (before DMAB)
are shown in Fig. 5. Calibrated EDS peak intensities of Pt and Pd
for each analyzed particle (enclosed in rectangular box) give the
empirical formula listed with each micrograph. The empirical for-
mulae are in very good agreement with PtPd1.1 and PtPd0.22 bulk
compositions for 8.4% Pt–5% Pd/C and 8.4% Pt–1% Pd/C, respectively,

M. Ohashi et al. / Electrochimica Acta 55 (2010) 7376–7384
7381
Fig. 4. HAADF images for Pt on 5.0% Pd/C before and after DMAB treatment and Pt on 1.0% Pd/C before DMAB treatment.
Fig. 5. HAADF images, EDS spectra, and molar ratio of Pt and Pd for 8.4% Pt–5.0% Pd/C after DMAB treatment and 8.4% Pt–1.0% Pd/C before DMAB treatment. White indicator
arrows note regions of exposed Pd surfaces.

7382
M. Ohashi et al. / Electrochimica Acta 55 (2010) 7376–7384
side of Fig. 5 shows particle morphologies for the 8.4% Pt–1% Pd/C
composition. The particle imaged in the upper right corner appears
to be only partially encapsulated (see arrow denoting exposed
Pd surface), while the particle in the lower right corner shows
encapsulation of the (dark contrasted) Pd core by a thick shell of
electrolessly deposited Pt.
Results pertaining to the different analytical methods used in
this study suggest the existence of both partial and fully encap-
sulated Pd core structures by electrolessly deposited Pt as the
most likely morphologies of the Pt–Pd bimetallic particles. From
both XRD and XPS analyses, the relatively facile Pd redox behav-
ior indicates an O2 pathway(s) to the Pd component. This can be
accomplished by either a partially or fully covered structure as long
as the Pt shell is O2-permeable. XPS results also clearly indicate
that thicker Pt shells (continuous or partial) stabilize the Pd core
against oxidation to PdO and that the fraction of Pd detected by
XPS decreases as the Pd loading also decreases from 5% to 0.5%. The
most definitive proof for the existence of both complete and partial
encapsulation by Pt is shown by HRTEM/EDS data.
Fig. 6. Mass transfer corrected Tafel plots for 7.0% Pt on 0.5% Pd/C and E-TEK 20%
Pt/C, each at (a) 20 ␮g and (b) 2.4 ␮g Pt loadings on the RRDE disk, both before and
after the 30,000 cycle durability test. ꢀ Initial (0 cycle) of Pt–Pd/C, ꢁ after 30,000
cycles of Pt–Pd/C, ꢀ initial (0 cycle) of Pt/C, and ♦ after 30,000 cycles of Pt/C.
indicating the controlled deposition of Pt in the preparation step.
Further, for each image, both bright and dark parts are observed,
indicating at least partial coverage of Pt over the Pd core. The dark
inner portion of each particle indicates the presence of a lower
atomic weight species, Pd in this case. Surrounding the Pd core
is a brighter shell of Pt (higher atomic weight). The spectrum in
the upper left corner of Fig. 5 for 8.4% Pt–5% Pd/C reveals a thin Pt
shell over a Pd core, while the lower left spectrum suggests only
partial encapsulation of the Pd core by Pt. The exposed Pd portion
of the particle is indicated by the pointer arrow. The right hand
Fig. 7. Ring current plots for 7.0% Pt on 0.5% Pd/C and E-TEK 20% Pt/C, each at (a)
20 ␮g and (b) 2.4 ␮g Pt loading on the RRDE disk, both before and after the 30,000
cycle durability test. ꢀ Initial (0 cycle) of Pt–Pd/C, ꢁ after 30,000 cycles of Pt–Pd/C,
ꢀ initial (0 cycle) of Pt/C, and ♦ after 30,000 cycles of Pt/C.
Fig. 8. Disk current plots for 7.0% Pt on 0.5% Pd/C and E-TEK 20% Pt/C at (a) 2.4 ␮g
Pt loading, (b) 20 ␮g Pt loading of 7.0% Pt on 0.5% Pd/C, and (c) 20 ␮g Pt loading of
E-TEK 20% Pt/C at different time intervals during the 30,000 cycle durability test.

M. Ohashi et al. / Electrochimica Acta 55 (2010) 7376–7384
7383
Since others have reported that the highest activity [28–30]
exists when Pd cores are partially covered by Pt, the ORR electro-
chemical activity for 7.0% Pt on 0.5% Pd/C (which gives a very clear
indication of partial coverage of Pt on a Pd core) was measured
and compared with the commercially available Pt/C catalyst (E-
TEK 20% Pt/C). Fig. 6 shows mass transport-corrected polarization
curves for (a) 7.0% Pt on 0.5% Pd/C 20 ␮g Pt loading and (b) 7.0%
Pt on 0.5% Pd/C at 2.4 ␮g Pt loading; these results are compared
to E-TEK 20% Pt/C. At the beginning of the test, the performance
of the Pd-based catalyst is less than or equal to that for the com-
mercially available catalyst in the kinetic regime (∼0.85–0.95 V vs.
SHE). The Tafel slopes for the Pt/C and the Pt on Pd/C samples vary
between 65 and 85 mV/decade. Thus, the oxygen reduction reaction
mechanism for these catalysts is the same [26]. This information,
however, is not sufficient to conclude that the Pd-based catalysts
do not offer any advantage, since it only represents the initial per-
formance. PEMFC catalysts are subjected to decay mechanisms,
including dissolution and particle coalescence, during their oper-
ational lifetime [37]. Further, formation of side reactions products
such as H2O2 (hydrogen peroxide) results in degradation of other
PEMFC materials (including catalyst ionomers) [38–41]. Therefore,
an accurate catalyst assessment requires knowledge of the main
reaction product distribution as well as data obtained under suit-
able aging conditions.
Oxygen reduction leads to H2O2 as a side product; its migration
into the ionomer further leads to radical formation in the presence
of metallic impurities. Radicals attack the ionomer leading to its
decomposition [37]. Fig. 7 illustrates the amount of H2O2 detected
at the ring for two different Pt loadings of the Pt on Pd sample before
and after 30,000 test cycles. For both Pt loadings, the amount of per-
oxide formed is significantly lower for the Pd-based catalyst than
for the commercial Pt catalyst at potentials lower than 0.7 V vs. SHE.
Fig. 9. Cyclic voltammogram of 20 ␮g Pt loading of (a) 7.0% Pt on 0.5% Pd/C and (b)
E-TEK 20% Pt/C during the 30,000 cycle durability test.
Fig. 10. HAADF images for 7.0% Pt on 0.5% Pd/C before and after the 30,000 cycle durability test. The sample is the same shown in Fig. 9(a).

7384
M. Ohashi et al. / Electrochimica Acta 55 (2010) 7376–7384
Both Pt and Pd not only produce H2O2 during oxygen reduction
[38,43], but are also active catalysts for hydrogen peroxide decom-
position [44–47]. Vetter et al. measured decomposition rates for Pt
catalysts at various temperatures and determined the first order
decomposition rate constant at 20–30 ◦C in pH 7 solutions to be
3.6 × 10−4 s−1 [44]. By comparison, values for Pd catalysts are up to
thirty times larger and range from 1 × 10−3 to 10 × 10−3 s−1 in pH
2–7 solutions [45–47]. Thus, Fig. 7 data are consistent with these
trends since the amount of H2O2 decomposed by the Pt on Pd/C
catalyst is larger (smaller amount detected) than with the Pt/C cat-
alyst. On this basis, it is expected that ionomers will show better
durability at extended lifetimes.
Fuel cell electrocatalysts are typically aged by electrode poten-
tial cycling [37]. Fig. 6 data indicate that potential cycling did
not appear to affect the catalyst surface area since the change in
performance is marginal. However, as reported by others the cat-
alyst might not be fully utilized on the RRDE [42]. Specifically,
as the catalyst surface area declines or becomes deactivated, the
reaction front can move deeper within the catalyst layer, thus
maintaining cell performance in the kinetic regime. Thus, more
sensitive measurements are needed to electrochemically assess
any changes in catalyst surface area. Because incomplete cata-
lyst utilization is due to mass transport-related effects, limiting
conditions may reveal a change in the catalyst surface area. The lin-
ear sweep voltammograms displayed in Fig. 8 show that for each
catalyst loading the limiting current is either not affected (Pt on
Pd/C) or is affected (Pt/C) by potential cycling. Hydrogen adsorp-
tion/desorption (∼0–0.3 V vs. SHE) and Pt oxidation/reduction
(0.7–0.8 V vs. SHE) surface processes probed by cyclic voltammetry
also offer another method to assess catalyst surface area changes.
Fig. 9 indicates that surface processes are virtually unaffected for
Pt on Pd/C, but are affected for Pt/C by potential cycling. These
electrochemical results are corroborated by the HAADF images in
Fig. 10 for the Pt on Pd/C catalyst, which show that particle sizes are
not affected by cycling. Therefore, the results from electrochemical
analyses support the hypothesis that Pt on Pd/C catalysts are more
durable than Pt/C catalysts as previously observed [31]. Specifically,
the Pt on Pd/C catalyst advantage is maintained during cycling and
corroborates the absence of a change observed with linear sweep
voltammograms, cyclic voltammograms, and HAADF images.
Acknowledgement
The authors gratefully acknowledge NSF Grant# EEC0324260
(International Research and Education in Engineering IREE supple-
ment) for funding this work.
References
[1] J.P.M. Treichel, in: J.I. Kroschwitz, M. Howe-Grant (Eds.), Kirk-Othmer Ency-
clopedia of Chemical Technology, vol. 21, John Wiley & Sons, New York, 1991,
p. 335.
[2] T. Toda, H. Igarashi, H. Uchida, M. Watanabe, J. Electrochem. Soc. 146 (1999)
3750.
[3] L. Xiong, A.M. Kannan, A. Manthiram, Electrochem. Commun. 4 (2002) 898.
[4] L. Xiong, A. Manthiram, J. Electrochem. Soc. 152 (2005) A697.
[5] J.A. Rodriguez, Surf. Sci. Rep. 24 (1996) 225.
[6] J.A. Rodriguez, M. Kuhn, J. Catal. 154 (1995) 355.
[7] W.M.H. Sachtler, Appl. Surf. Sci. 19 (1984) 167.
[8] R. Massard, D. Uzio, C. Thomazeau, C. Pichon, J.L. Rousset, J.C. Bertolini, J. Catal.
245 (2007) 133.
[9] M.E. Davis, R.J. Davis, Fundamentals of Chemical Reaction Engineering, 1st ed.,
McGraw–Hill, New York, 2003.
[10] W. Juszczyk, Z. Karpinski, Appl. Catal. A 206 (2001) 67.
[11] J.T. Wrobleski, M. Boudart, Catal. Today 15 (1992) 349.
[12] G. Ertl, H. Knozinger, J. Weitkamp (Eds.), Handbook of Heterogeneous Catalysis,
VCH Verlagsgesellschaft, Weinheim, 1997, pp. 13, 35–38, 191–195, 257–261,
365–367.
[13] L. Guczi, Stud. Surf. Sci. Catal. 29 (1986) 547.
[14] B.D. Chandler, L.H. Pignolet, Catal. Today 65 (2001) 39.
[15] I. Ohno, Mater. Sci. Eng. A 146 (1991) 33.
[16] S.S. Djokic, in: B.E. Conway, R.E. White (Eds.), Modern Aspects of Electrochem-
istry, vol. 35, Kluwer Academic/Plenum Publishers, Fort Saskatchewan, AB,
Canada, 2002, p. p. 51.
[17] F. BarBir, PEM Fuel Cells: Theory and Practice, Elsevier Academic Press, MA,
2005.
[18] K.D. Beard, M.T. Schaal, J.W. Van Zee, J.R. Monnier, Appl. Catal. B: Environ. 72
(2007) 262.
[19] K.D. Beard, J.W. Van Zee, J.R. Monnier, Appl. Catal. B: Environ. 88 (2009)
185.
[20] K.D. Beard, D.C. Borrelli, A.M. Cramer, D.A. Blom, J.W. Van Zee, J.R. Monnier, ACS
Nano 3 (2009) 2841.
[21] U.A. Paulus, A. Wokaun, G.G. Scherer, T.J. Schmidt, V. Stamenkovic, N.M.
Markovic, J.P.N. Ross, Electrochim. Acta 47 (2002) 3787.
[22] U.A. Paulus, A. Wokaun, G.G. Scherer, T.J. Schmidt, V. Stamenkovic, V. Rad-
milovic, N.M. Markovic, J.P.N. Ross, J. Phys. Chem. B 106 (2002) 4181.
[23] N. Hoshi, M. Nakamura, S. Kondo, Electrochem. Commun. 11 (2009) 2282.
[24] S. Kondo, M. Nakamura, N. Maki, N. Hoshi, J. Phys. Chem. C 113 (2009) 12625.
[25] N. Markovic, H. Gasteiger, P.N. Ross, J. Electrochem. Soc. 144 (1997) 1591.
[26] J. Zhang, Y. Mo, M.B. Vukmirovic, R. Klie, K. Sasaki, R.R. Adzic, J. Phys. Chem. B
108 (2004) 10955.
[27] J.X. Wang, H. Inada, L. Wu, Y. Zhu, Y.-M. Choi, P. Liu, W.-P. Zhou, R.R. Adzic, J.
Am. Chem. Soc. 131 (2009) 17298.
[28] H. Li, G. Sun, N. Li, S. Sun, D. Su, Q. Xin, J. Phys. Chem. C 111 (2007) 5605.
[29] X. Li, Y. Zhu, Z. Zou, M. Zhao, Z. Li, Q. Zhou, D.L. Akins, H. Yanga, J. Electrochem.
Soc. 156 (2009) B1107.
4. Conclusion
A series of Pt on Pd/C catalysts has been synthesized by ED
and characterized using a variety of methods. Results have shown
that ED is selective and creates only bimetallic structures of Pt
and Pd. More specifically, the Pd core is either fully or partially
encapsulated by Pt. The initial performance of these electrocata-
lysts for oxygen reduction was either similar to or less than for
commercially available catalysts. However, this level of evaluation
is insufficient for catalyst selection. Thus, the amount of perox-
ide generated and the performance loss during potential cycling
were also measured and revealed that Pt on Pd/C catalysts are
much more durable and generate considerably less H2O2 byprod-
uct, which should reduce ionomer degradation due to oxidation by
H2O2. A quantitative approach that takes into account both initial
performance and durability parameters needs to be part of catalyst
evaluation and selection. The hypothesis that a Pt on Pd/C cata-
lyst can reduce ionomer degradation should be evaluated further.
The correlation of electrochemical activity and durability with Pt
content on a Pd core may provide valuable insight into the mecha-
nism(s) of enhanced stability and activity. In addition, the effects of
Pd on other catalyst degradation mechanisms remain to be investi-
gated (contamination, carbon corrosion, etc.) to more fully evaluate
its potential in electrochemical applications.
[30] H. Ye, R.M. Crooks, J. Am. Chem. Soc. 129 (2007) 3627.
[31] Z. Peng, H. Yang, J. Am. Chem. Soc. 131 (2009) 7542.
[32] Y. Gu, J. St-Pierre, A. Joly, R. Goeke, A. Datye, P. Atanassov, J. Electrochem. Soc.
156 (2009) B485.
[33] J.H. Scofield, J. Electron Spectrosc. Relat. Phenom. 8 (1976) 129.
[34] K. Persson, K. Jansson, S.G. Jaeras, J. Catal. 245 (2007) 401.
[35] M.P. Seah, W.A. Dench, Surf. Interface Anal. 1 (1979) 2.
[36] Y.-S. Ho, C.-B. Wang, C.-T. Yeh, J. Mol. Catal. A: Chem. 112 (1996) 287.
[37] R. Borup, J. Meyers, B. Pivovar, Y.-S. Kim, R. Mukundan, N. Garland, D. Myers,
M. Wilson, F. Garzon, D. Wood, P. Zelenay, K. More, K. Stroh, T. Zawodzinski, J.
Boncella, J.E. McGrath, M. Inaba, K. Miyatake, M. Hori, K. Ota, Z. Ogumi, S. Miyata,
A. Nishikata, Z. Siroma, Y. Uchimoto, K. Yasuda, K. Kimijima, N. Iwashita, Chem.
Rev. 107 (2007) 3904.
[38] U.A. Paulus, T.J. Schmidt, H.A. Gasteiger, R.J. Behm, J. Electroanal. Chem. 495
(2001) 134.
[39] H. Hori, M. Murayama, T. Sano, S. Kutsuna, Ind. Eng. Chem. Res. 49 (2010)
464.
[40] N. Ohguri, A.Y. Nosaka, Y. Nosaka, Electrochem. Solid State Lett. 12 (2009)
B94.
[41] A.P. Young, J. Stumper, S. Knights, E. Gyenge, J. Electrochem. Soc. 157 (2010)
B425.
[42] Y. Sun, J. Lu, L. Zhuang, Electrochim. Acta 55 (2010) 844.
[43] C.M. Sanchez-Sanchez, A.J. Bard, Anal. Chem. 81 (2009) 8094.
[44] T.A. Vetter, D.P. Colombo Jr., J. Chem. Ed. 80 (2003) 788.
[45] Y. Voloshin, J. Manganaro, A. Lawal, Ind. Eng. Chem. Res. 47 (2008) 8119.
[46] V.R. Choudhary, S.D. Sansare, A.G. Gaikwad, Catal. Lett. 84 (2002) 81.
[47] V.R. Choudhary, A.G. Gaikwad, React. Kinet. Catal. Lett. 80 (2003) 27.