Influence of chloride ions on the stability of PtNi alloys for PEMFC cathode

Article abstract of DOI:10.1016/j.electacta.2011.03.043

The dependence of the rate of Ni dissolution from PtNi alloys on the chloride concentration was studied electrochemically in 0.5 M HClO₄ at room temperature. Electrodeposited PtNi catalysts were subjected to extensive potential cycling between

Full text of DOI:10.1016/j.electacta.2011.03.043

Electrochimica Acta 56 (2011) 7235–7242
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Influence of chloride ions on the stability of PtNi alloys for PEMFC cathode
K. Jayasayee∗, J.A.R. Van Veen, E.J.M. Hensen, F.A. de Bruijn1
Schuit Institute of Catalysis, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
The dependence of the rate of Ni dissolution from PtNi alloys on the chloride concentration was studied
electrochemically in 0.5 M HClO4 at room temperature. Electrodeposited PtNi catalysts were subjected
to extensive potential cycling between 20 mV and 1.3 V at various Cl− concentrations and the cyclic
voltammograms (CV) response and the oxygen reduction reaction (ORR) activity of the catalysts were
determined at different intervals. Energy dispersive X-ray spectroscopy (EDS) and inductively coupled
plasma spectroscopy (ICP) analyses were carried out to determine the elemental composition of the
alloys and the amount of dissolved Ni at different stages of the potential cycling. It was found that the
presence of Cl− increases the rate of Ni dissolution and by this accelerates the dealloying process relative
to potential cycling in chlorine-free solutions. Dealloying is most pronounced during the initial stages of
potential cycling. Already a small amount of Cl− is sufficient to dissolve the majority of the non-noble
metal from the alloys. Even then, under oxygen reduction conditions, the blockage of Pt surface by Cl− is
less pronounced for the alloys than for pure Pt catalysts, leading to marginally improved ORR activity for
the PtNi alloys at low Cl− concentrations. From a practical point of view, the effect of chloride ion leakage
from a commercially available saturated KCl reference electrode on the electrocatalytic activity was also
investigated.
Received 7 December 2010
Received in revised form 21 February 2011
Accepted 14 March 2011
Available online 22 March 2011
Keywords:
PtNi alloys
Chloride ion
PEMFC contamination
ORR activity
Fuel cells
© 2011 Elsevier Ltd. All rights reserved.
1. Introduction
HClO4 [12–14]. Studies have shown that the adsorption of anions
on Pt surfaces occurs even at impurity concentrations of less than
Air is the most practical and feasible oxidant for the proton
exchange membrane fuel cell (PEMFC) cathode and, accordingly,
impurities in air will be of major concern for the long-term perfor-
mance of PEMFC systems. Airborne impurities such as SO2, NO2,
H2S, and NO have already been identified as poisons for the Pt
catalyst with irreversible damage to the cell performance even if
present in minute quantities [1–4]. Similarly, anions like chloride
that are introduced into the membrane and electrode assembly
(MEA) through the water-cooling systems/humidifiers or residual
Cl− from catalysts prepared using chloride precursors can contam-
inate the Pt surface [3,5–9].
The adsorption of halide ions on Pt surface increases in the order
of F− < Cl− < Br− < I− [4,10,11]. Platinum single-crystal studies have
revealed the strong adsorption of chloride ions to the Pt crystal
surface planes, thus hindering the oxygen reduction reaction (ORR),
especially in electrolytes without strongly adsorbing anions such as
10−6 M [10,11,15]. Schmidt et al. reported an increased H2O2 pro-
duction with the addition of chloride ions, which could severely
degrade the membrane [6]. Yadav et al. found that the dissolution
of Pt under potential cycling depends on Cl− concentration [16]. In
general, the loss of active Pt surface area for the ORR, a change in the
reaction pathway, enhanced H2O2 formation and an enhanced Pt
dissolution are some of the major known impacts of chloride ions
in PEMFC.
To study the kinetics of the catalysts under different condi-
tions, electrochemical experiments with a liquid electrolyte are
frequently used to model the real systems. Commercially available
reference electrodes, which are invariably a part of these electro-
chemical cells, often contain chloride-bearing species. These anions
could leak into the bulk electrolyte and interfere with the work-
ing electrode depending on the operating temperature, working
potential and ion concentration.
Platinum alloyed with Co, Ni, Cu and other transition metals
are generally found to have enhanced ORR activity over pure Pt/C
[17–21]. It has been proposed that this enhancement is due to the
formation, upon leaching, of alloy particles with a Pt-skin or Pt-
shell, with distinctly different electronic properties compared to
those of pure Pt [22–27]. Strasser et al. [26] found that the Pt–Pt
interatomic distance in the Pt-rich shell on top of a Pt–Cu alloy
core is shorter than one would expect based on the composition of
∗
Corresponding author at: Department of Chemical Engineering and Chemistry,
P.O. Box 513, 5600 MB Eindhoven, The Netherlands.
Tel.: +31 40 2474916; fax: +31 40 2455054.
E-mail addresses: kaushik.jayasayee@gmail.com, k.jayasayee@tue.nl
(K. Jayasayee).
1
Present address: Nedstack Fuel Cell Technology BV, P.O. Box 5167, 6802 ED
Arnhem, The Netherlands.
0013-4686/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2011.03.043

7236
K. Jayasayee et al. / Electrochimica Acta 56 (2011) 7235–7242
this shell, leading to a compressive strain that leads to an enhanced
ORR activity. The formation of Pt–OH intermediates, which are con-
sidered to block the active Pt site for ORR, is generally found to
be delayed in Pt alloys, thus enhancing the ORR activity [28]. In
this work, we will show that chloride ions have a large negative
impact on the ORR activity of both Pt and PtNi alloys and at the
same time accelerate the dissolution of Ni from PtNi alloys. Elec-
trodeposited Pt, Pt75Ni25 and Pt10Ni90 (atomic ratios) were exposed
to electrolytes containing chloride at various concentrations under
potential cycling. EDS, ICP and XPS were employed to monitor the
changes in the Pt:Ni atom ratio in the catalyst layer and in the elec-
trolyte. As the anion impurities from standard reference electrodes
such as saturated calomel electrode (SCE) or Ag/AgCl electrode
could interfere with the electrochemical measurements and affect
the Pt ORR activity [29] from a practical point of view, the conse-
quence of chloride ion leakage from the reference electrode is also
investigated.
2.2. Elemental analysis
Energy dispersive X-ray spectroscopy (EDS) was carried out for
all the fresh alloys using a Quanta 3D FEG instrument (FEI Company)
to check the Pt:Ni atom ratios initially, so as to tune the deposi-
tion conditions to the desired elemental ratio. Various spots were
selected on the catalyst layer and the concomitant atom% of Ni and
Pt were averaged. The change in the Pt:Ni ratio for the alloys after 15
and 1000 potential cycles was also monitored. Likewise inductively
coupled plasma (ICP) analysis was performed on the electrolyte to
deduce the amount of Pt and Ni dissolved from the alloys during
the initial dealloying in both blank HClO4 and 10−4 M HCl added
HClO4.
Surface compositions of selected samples were analyzed using
a Kratos AXIS Ultra X-ray photoelectron spectrometer (XPS),
equipped with a monochromatic Al K␣ X-ray source and a delay-
line detector (DLD). Spectra were obtained using the aluminum
anode (Al KR) 1486.6 eV operating at 150 W, with survey scans
at a constant pass energy of 160 eV and with region scans at a
constant pass energy of 40 eV. Depth profiles were obtained by
sputtering, while rotating the sample, using an Argon pressure of
3 × 10−8 mbar Argon, while the emission current was set to 15 mA
with beam energy of 4 kV. Moreover, the presence of Cl− on the Pt
surface before and after exposing to chloride environment was also
evaluated using XPS. The electrodes, after exposure to 15 poten-
tial scans, were held at 0.95 V for a few seconds to prevent any
adsorbed Cl− to escape from the surface. Then the electrode was
carefully removed form the electrochemical cell with the potential
of 0.95 V still applied. The electrode was cleaned gently with deion-
ized water and then transferred quickly to the vacuum chamber for
XPS measurements.
2. Experimental
PtNi alloys were electrochemically deposited from a solution
containing 1.5 g l−1 H2PtCl6·6H2O (Merck), 25 g l−1 NiCl2·6H2O
(Merck) and 30 g l−1 NaCl (Merck) at pH 2.5 on a finely polished
Au rotating disk electrode (RDE). The catalysts were deposited
galvanostatically during a fixed time of 550 s. The current of 300
and 900 ␮A cm−2 was applied respectively to deposit Pt75Ni25 and
Pt10Ni90. For comparison, pure Pt was deposited at 300 ␮A cm−2
for 500 s from a solution of 1.5 g l−1 H2PtCl6·6H2O (Merck) at pH
2.5. Prior to deposition, the Au RDE was electrochemically cleaned
in 0.5 M H2SO4 (Merck) by potential scanning between 0 V and
1.7 V until a stable cyclic voltammogram (CV) was obtained. The
electrochemical cell was flushed with Ar for about 45 min before
the catalyst deposition. As the catalyst precursors contain chloride
ions, as a precautionary step, the electrode was washed with copi-
ous amount of deionized water to remove the likely inclusion of
chloride in the catalyst layer.
2.3. Effect of Cl−-containing reference electrode
To substantiate the influence of chloride ions leaked from a Cl−-
containing reference electrode, a few experiments were conducted
on PtNi alloys at room temperature and at 80 ◦C. A commer-
cially available reference electrode (Red Rod Technology from
Radiometer Analytical) saturated with KCl and stable in the acidic
environment up to 100 ◦C was used for this purpose. The reference
electrode used has a double junction and was placed in a separate
compartment in the electrochemical cell to minimize Cl− con-
tamination. The resultant CVs and the spectroscopic results were
compared to those of the above-described experiments.
2.1. Electrochemical characterization
The electrolytes with Cl− concentrations of 10−3 M and 10−4 M
in 0.5 M HClO4 were prepared from concentrated HCl (32%) and
HClO4 (65%) (both from Sigma–Aldrich) by adding the required
amount of HCl to 0.5 M HClO4. For all the electrodes (Pt, Pt75Ni25
and Pt10Ni90), the potential was scanned between 20 mV and 1.3 V
in blank 0.5 M HClO4, 10−3 M HCl in 0.5 M HClO4 and 10−4 M HCl
in 0.5 M HClO4 separately at room temperature under argon atmo-
sphere. A three-compartment electrochemical cell was used with
the working electrode and the reference electrode compartments
separated by a glass frit.
3. Results and discussion
3.1. Electrochemical behavior in blank 0.5 M HClO4
The hydrogen adsorption/desorption profiles of the PtNi alloys
after 15 CVs were clearly different from that of pure Pt (Fig. 1a), sug-
gesting that alloying Ni to Pt influences the adsorption of hydrogen
on Pt. Note that, contrary to what the literature leads one to expect
(see Section 1), there is no difference in the Pt–OH onset potential
between Pt and its Ni alloys. This could be a particle-size effect: the
Pt electrodeposit has somewhat larger crystallites (4.5 1.2 nm)
than the alloys (3.4 1.3 nm, and 3.6 0.9 nm, respectively, for
Pt75Ni25 and Pt10Ni90 as determined by TEM [30]). Thus, the
cathodic shift in the Pt–OH onset potential to be expected for
smaller particles [31] could be compensated for by the alloy effect.
The ORR polarization curves of the catalysts after 15 CVs, nor-
malized to the geometric surface area (0.2826 cm2), are shown in
Fig. 2a. It has to be noted, however, that the electrodeposits are
quite porous, so that the measured (kinetic) current has to be nor-
malized to the electrochemical active Pt surface area (ECSA) to
As the current work deals with the effect of Cl− impurities, care
was taken not to use calomel and other types of reference electrodes
that can leak chloride ions. Instead, a reversible hydrogen electrode
(RHE) was constructed from a Pt wire that was kept under a sat-
urated hydrogen atmosphere. A platinized platinum flag was used
as a counter electrode.
The electrodes were initially scanned for 15 CVs to ensure a sta-
ble surface. The electrodes were again scanned for 1000 potential
cycles to validate their electrochemical stability and modification in
their ORR activity under Cl− contamination. Hence, the ORR activity
of the alloys at room temperature was determined by recording the
polarization curves between 0.2 V and 1.1 V at a rotation speed of
1000 rpm under oxygen-saturated electrolyte after the completion
of 15 and 1000 CVs respectively.

K. Jayasayee et al. / Electrochimica Acta 56 (2011) 7235–7242
7237
Table 1
Specific activities of the alloys at different conditions.
Electrode
Specific activity (␮A cm−2
Cl− free at 0.95V
)
10−4 M HCl at 0.9V/0.85V
10−3 M HCl at 0.9V/0.85V
Cl—containing ref.
electrode at 0.95V/0.9V
Pt (CVs)
15
1000
112
26
5.1/35.5
–/3.1
–/6.9
Dissolved
7.6/51
0.2/5.3
Pt75Ni25 (CVs)
Pt10Ni90 (CVs)
15
1000
488
48
7.5/53.3
0.4/7.8
–/6.9
Dissolved
5.3/68
0.7/10.9
15
1000
388
66
7.7/58
1.8/19.4
0.3/9.2
Dissolved
10.4/154
1.6/18.4
arrive at the specific ORR activity. ECSA values were calculated by
integrating the area under the Hupd region of the CVs and they
vary between 3 cm2 and 6.5 cm2 for the present catalysts. The
thus obtained values for the specific ORR activities are collected
in Table 1. It is seen that at this stage the PtNi catalysts are 3–4
times more active than their Pt-only counterpart. Again, this is in
contrast to the literature position that Pt alloys are better ORR cata-
lysts than Pt because of the Pt–OH onset potential shifting to higher
potentials, which is clearly not the case here. We surmise that sub-
surface Ni here accelerates the rate-determining step in the ORR,
rather than freeing up more sites to perform the reaction on (see
further below).
The cyclic voltammograms of the present catalysts after 1000
cycles are presented in Fig. 1b. Clearly, the leaching of Ni has now
progressed to a stage that the hydrogen adsorption/desorption pro-
files of Pt and of PtNi are quite similar. However, in the Pt-oxide
region small differences are still observed. In particular, the Pt–OH
onset potential of Pt10Ni90 is somewhat cathodic to that of Pt itself.
This cannot now be a particle size effect, as the aged Pt and Pt10Ni90
electrodeposits have crystallites of approximately the same size
(5.3 2.2 and 5.1 2.3 nm, respectively; for aged Pt75Ni25 a value
of 4.4 2.3 nm was found [30]).
The ORR polarization curves after 1000 CVs are shown in Fig. 2b.
The derived values of the specific ORR activities are given in Table 1.
It is seen that both Pt and its Ni alloys have deactivated to a consid-
erable extent. For Pt this is not so usual, and it is presumably due
to a change in morphology. Comparison of Fig. 1a and b shows that
extensive potential cycling has a distinct influence on the structure
of the platinum surface. The sharp weakly adsorbed hydrogen oxi-
dation peak around 0.15 V, present in the fresh platinum, becomes
much smaller after 1000 CVs. This sharp peak can be assigned
to Pt (1 1 0) [32]. At the same time, a peak at the cathodic limit
of the CV scan, which is generally attributed to the oxidation of
adsorbed molecular hydrogen, becomes apparent in the platinum
electrode after 1000 CVs. Such peak is visible in Pt (1 0 0) single
100
a
15CVs
50
0
-50
Pt
Pt75Ni25
Pt10Ni90
-100
0.2
0.4
0.6
0.8
1.0
1.2
E, V vs RHE
100
50
b
1000CVs
0
-50
-100
Pt
Pt75Ni25
Pt10Ni90
0.2
0.4
0.6
0.8
1.0
1.2
E, V vs RHE
Fig. 2. Polarization curves of the alloys for the ORR after 15 potential scans (a) and
1000 potential scans (b) in chloride-free O2 saturated 0.5 M HClO4; 5 mV s−1 scan
rate; 1000 rpm.
Fig. 1. CV profiles of the alloys after 15 potential scans (a) and 1000 potential scans
(b) in chloride-free Ar saturated 0.5 M HClO4; 50 mV s−1 scan rate.

7238
K. Jayasayee et al. / Electrochimica Acta 56 (2011) 7235–7242
crystal electrodes as well. When looking closely to the ORR sweep
of the platinum electrode after 1000 CV scans in Fig. 2b, and com-
paring them to those in single crystal electrodes as published in
[33], the observed ORR does well correspond to that on Pt (1 0 0).
From both [32,33] it is known that the ORR activity on Pt (1 0 0)
is far lower than that on Pt (1 1 1) and Pt (1 1 0). The observed
change in the hydrogen adsorption profile, the profile of the ORR
and the decreased activity match indicate a structural change upon
extensive potential cycling.
As to the deactivation of the PtNi electrodeposits, they also may
have suffered from some morphology change, but this is not so
readily apparent from the CVs after 15 and 1000 CVs, and the main
cause for the deactivation will have been extensive Ni leaching.
Still, they are about twice as active as Pt, after 1000CVs, and again,
this is not due to an anodic shift of the Pt–OH onset potential. Thus,
the effect of Ni in the present case appears to remain an effect on
the ORR itself, rather than on the availability of Pt active sites. It
is noted that similar results have been obtained for PtCo and PtCu
electrodeposits [30].
a
100
50
10-4M HCl
15CVs
0
-50
-100
Pt
Pt75Ni25
Pt10Ni90
0.2
0.4
0.6
0.8
1.0
1.2
E, V vs RHE
b
10-4M HCl
1000CVs
50
0
3.2. Electrochemical behavior in HCl-containing 0.5 M HClO4
Addition of a small amount of HCl (10−4 M) to the elec-
trolyte leads to a substantial change in the hydrogen adsorp-
tion/desorption behavior of Pt. The CV after 15 cycles is shown in
Fig. 3a. The emergence of two sharp, highly reversible peaks instead
of one is commonly observed for Pt in the presence of strongly
adsorbing anions [10]. For the alloys, such a distinct Hupd region
in the CV profiles does not emerge (Fig. 3a), indicating the contin-
ued presence of subsurface Ni, despite the strong Ni leaching in
Cl-containing electrolyte (see below, Table 2). On the other hand,
the shrinkage of the Hupd region and the increase in the double-
layer region that is a hallmark of Pt under chloride-ion adsorption
[10] was observed for both Pt and the Pt alloys with Pt showing the
largest effect.
-50
Pt
Pt75Ni25
Pt10Ni90
-100
0.2
0.4
0.6
0.8
1.0
1.2
E, V vs RHE
c 100
10-4 M HCl
Pt Ni
10 90
50
The formation of the oxygen layer begins at the lowest poten-
tial for Pt10Ni90, followed by Pt75Ni25 and Pt itself. This implies that
the suppression of Pt–OH formation by adsorbed chloride increases
in the same sequence. We believe that this effect is mainly due
to subsurface Ni influencing the subtle balance in the competition
between OHad and Clad for the Pt surface [14], with Ni leading to a
relative favoring of OH over Cl. An additional effect could be the for-
mation of NiClx species in solution, which would lower the effective
Cl− concentration, and, hence, the Cl surface coverage. That such an
effect may play a role is shown in Fig. 3c: upon changing the Ni-
containing electrolyte for a fresh one, an anodic shift of the Pt–OH
onset potential is observed.
0
-50
With dissolved Ni
-100
Without dissolved Ni
0.2
0.4
0.6
0.8
1.0
1.2
E, V vs RHE
Addition of a larger amount of HCl (10−3 M) to the electrolyte
leads to an, on the face of it, simpler picture (see Fig. 4a). The Hupd
region still shows a difference between Pt and its Ni alloys, indi-
cating that some subsurface Ni still survives after 15 cycles in this
medium, but the Cl− concentration is now so large that Clad dom-
inates over OHad to quite high potentials, thus almost obliterating
the differences between Pt and PtNi in the oxide region.
Turning now to the ORR activity of the stabilized electrodes
(after 15 CVs), a comparison between Figs. 2a, 5a, and 6, already
show, as expected, the enormous decrease in performance in the
presence of chloride ions, due to their blocking of active sites [14].
The various values of the specific activity are collected in Table 1.
Note that the potentials at which these values are quoted are dif-
ferent for the different chloride concentrations.
Fig. 3. CV profiles of the alloys after 15 potential scans (a) and 1000 potential scans
(b) in Ar saturated 10−4 M HCl + 0.5 M HClO4; 50 mV s−1 scan rate. (C) CV profiles
of Pt10Ni90 illustrating the effect of dissolved Ni in Ar saturated 10−4 M HCl + 0.5 M
HClO4; 50 mV s−1 scan rate.
at a higher potential than desorption. In the anodic scan, the poten-
tial range in which the surface is devoid of strongly adsorbed OH
extends to higher potentials, and results in a higher ORR activity.
In the presence of chloride, its adsorption strongly inhibits the
formation of adsorbed hydroxyl ions (as discussed above). While
the irreversibility of the adsorption of OH species is associated with
the formation of platinum oxide at higher potential, the nature of
the adsorbed chloride does not change with potential, and does not
show any hysteresis effect. Combining both effects, in the presence
of chloride the effect of its adsorption dominates, leading to the
virtual absence of hysteresis in the ORR activity.
From 5a and 6, in the presence of chloride, the hysteresis in ORR
between anodic and cathodic sweeps tends to decrease consider-
ably. In the absence of chloride (Fig. 2a), the hysteresis in the ORR
sweeps is caused by strongly adsorbed OH species, which them-
selves show irreversible adsorption, i.e., the adsorption takes place
From Table 1 it can be inferred that in the 10−4 M HCl case the
advantage in specific ORR activity of PtNi over Pt has been reduced

K. Jayasayee et al. / Electrochimica Acta 56 (2011) 7235–7242
7239
Table 2
Elemental compositions of the alloys at different conditions.
Electrolyte
Pt:Ni-XPSa
Pt75Ni25
15CVs
Pt:Ni-EDSb
Pt75Ni25
15CVs
Ni dissolved – ICP (mg l−1
)
Pt10Ni90
15CVs
Pt10Ni90
15CVs
Pt10Ni90
15CVs
Pt75Ni25
15CVs
1000CVs
1000CVs
Cl-free
92:8
100:0
100:0
51:49
86:14
100:0
83/17
86/14
86/14
86/14
87/13
91/9
Dissolved
91/9
42/58
76/24
78/22
71/29
85/15
90/12
Dissolved
89/11
0.07
0.07
–
0.16
0.38
–
10–4 M HCl
10–3 M HCl
Red Rod at RT
Red Rod at 80 ◦
C
100:0
100:0
Bulk Pt:Ni ratio of fresh (a) Pt75Ni25 = 77:23 and (b) Pt10Ni90 = 15:85; RT: room temperature.
a
Surface concentration.
Bulk concentration.
b
to about a factor of 1.5. That is, although (Fig. 3a) the respective
electrodes differ in their affinity of Cl and OH, the combined affect
of these two adsorbates almost cancels out the ORR performance
difference noted in Cl-free electrolyte. In the 10−3 M HCl case the
differences in specific activity have become even smaller, as Cl
adsorption dominates.
has become more dominant than after 15 CVs, and this is in agree-
ment with the increased (1 0 0) character of the Pt surface [14]. A
small additional effect here may be an increased Cl/electrode sur-
face area ratio, since the ECSA decreases by about 50% during the
cycling. Nevertheless, the Pt10Ni90 surface retains some of its higher
oxophilicity as compared to the other two surfaces, and this must,
at least partly, be due to the continued presence of some subsurface
Ni. Another part may be, as noted above, the effective decrease of
the concentration of free chloride ions due to the formation of NiClx
species in solution.
The specific ORR activities of the electrodes after 1000 CVs in
the presence of 10−4 M chloride are collected in Table 1. As before,
it is seen that Pt has deactivated to a considerable extent, and this
is again ascribed to its surface having become more predominantly
Pt (1 0 0). While Pt75Ni25 is about two times as active now as Pt
itself, Pt10Ni90 has lost much less of its initial activity, ending up as
After 1000 CVs in the 10−4 M HCl containing electrolyte, the
Hupd signatures of Pt and PtNi have become much more similar
than before the durability test (Fig. 3b). As observed in the Cl-free
case, the Hupd region of the Pt CV has become much more (1 0 0)-
like. The hydrogen ad- and desorption profiles of Pt75Ni25 follow
those of Pt closely, indicating that very little Ni remains in sub-
surface positions, while the somewhat more important deviations
observed for Pt10Ni90 suggest the retention of more subsurface Ni
in this case. For all three electrodes the Pt–OH onset potential has
shifted anodically, indicating that the adsorption of chloride ions
a 100
a
-3
10-4M HCl
15CVs
15CVs
10 M HCl
Pt
0
Pt75Ni25
Pt10Ni90
50
0
-1
-2
-3
-4
-5
-50
-100
Pt
Pt75Ni25
Pt10Ni90
0.2
0.4
0.6
0.8
1.0
1.2
0.2
0.4
0.6
0.8
1.0
E, V vs RHE
E, V
b
10-4M HCl
1000CVs
Pt
Pt75Ni25
Pt10Ni90
0
10-3M HCl
b
~ 300CVs
-1
-2
-3
-4
-5
50
0
Pt
Pt75Ni25
Pt10Ni90
-50
0.2
0.4
0.6
E, V
0.8
1.0
0.2
0.4
0.6
E, V
0.8
1.0
1.2
Fig. 5. Polarization curves of the alloys for the ORR after 15 potential scans (a) and
1000 potential scans (b) in O2 saturated 10−4 M HCl + 0.5 M HClO4; 5 mV s−1 scan
rate; 1000 rpm.
Fig. 4. CV profiles of the alloys after 15 potential scans (a) and 1000 potential scans
(b) in Ar saturated 10−3 M HCl + 0.5 M HClO4; 50 mV s−1 scan rate.

7240
K. Jayasayee et al. / Electrochimica Acta 56 (2011) 7235–7242
100
10-3M HCl
a
15CVs
0
-1
-2
-3
-4
-5
50
0
-50
Pt
Pt75Ni25
Pt10Ni90
RHE
with Cl--containing
reference electrode
-100
0.2
0.4
0.6
0.8
1.0
0.2
0.4
0.6
0.8
1.0
1.2
E, V vs RHE
E, V vs RHE
100
50
b
Fig. 6. Polarization curves of the alloys for the ORR after 15 potential scans (a) and
1000 potential scans (b) in O2 saturated 10−3 M HCl + 0.5 M HClO4; 5 mV s−1 scan
rate; 1000 rpm.
0
about six times as active as Pt. So, the superior ORR activity of aged
Pt10Ni90, as compared with that of aged Pt, continues to assert itself
(cf. Table 1 for the Cl-free case), and this must, at least partly, be, it
is surmised, an effect of the presence of subsurface Ni.
-50
-100
-150
-200
The durability of the electrodes in the 10−3 M chloride case could
not be studied, since they were already completely dissolved after
some 300 cycles, exposing the Au substrate (Fig. 4b). This evidences
that chloride ions enhance the dissolution of Pt.
with RHE
with Cl--containing
reference electrode
0.2
0.4
0.6
0.8
1.0
1.2
3.3. PtNi composition variation
E, V
It is well known that the electrochemical potential cycling under
the relevant potential range for PEMFC dissolves the less noble
metal of the alloy, while enriching the surface layers with the more
noble metal. From Table 2 it becomes apparent that the dissolution
rate of Ni from the alloys in HClO4 solution is significantly lower
than in the chloride-containing electrolyte. In Pt75Ni25 and Pt10Ni90
about 17 and 58 atom % Ni remained, respectively, after 15 CVs in
chloride free HClO4. The same experiment in the presence of Cl−
gave values of 14 and 22 atom % Ni after the same number of CVs.
The presence of Cl− preferentially increases the dissolution rate of
Ni from the alloys during the initial potential cycling, irrespective
of the concentration of HCl (10−3 or 10−4 M). The rate of Ni dis-
solution decreases during subsequent potential cycling and this is
understood by the formation of a Pt skin. The amount of Ni present
in the alloys after 1000 CVs was fairly similar irrespective of the
initial Pt:Ni concentration and the constituents in the electrolyte.
Still, there are indications (e.g., Fig. 3b) that a bit more subsurface
Fig. 7. CV profiles of the Pt10Ni90 after 1000 potential scans in room temperature
(a) and 15 potential scans at 80 ◦C (b) in Ar saturated 0.5 M HClO4; 50 mV s−1 scan
rate.
the amount of dissolved Ni was similar in blank HClO4 as well as
HClO4 with 10−4 M HCl. For Pt10Ni90, the difference in the Ni disso-
lution rate was well recognizable when dealloyed in the presence
and absence of Cl− in the electrolyte, that is, the amount of Ni dis-
solved in the presence of 10−4 M HCl is more than twice the amount
dissolved in blank HClO4.
3.5. Impact of chloride ion leakage from KCl-saturated reference
electrodes
The above results indicate that Cl− contamination in PEMFC
even in the range of ppm levels will invariably affect the elec-
trocatalytic activity of Pt and Pt alloys besides enhanced Pt and
Ni dissolution. The CV and RDE profiles of PtNi alloys in 10−4 M
HCl suggest a distinct although minimal enhancement in their ORR
activity compared to Pt but in higher Cl− concentration the perfor-
mance of all the catalysts was alike.
A comparison of the stable CV of Pt10Ni90 measured using a RHE
and the KCl-saturated reference electrode (Red Rod) at room tem-
perature and at 80 ◦C is shown in Fig. 7a and b. Features similar
to what discussed in the previous sections were noticed for the
electrode exposed to the chloride-containing reference electrode.
The specific activity shown in Table 1 reveals that the performance
of the catalysts measured using the chloride-containing reference
electrode is very similar to what was found in 10−4 M HCl. The Cl−-
leakage rate from the reference electrode would be more at higher
temperatures and hence a greater effect of Cl− was witnessed for
the same alloy at 80 ◦C. Hence, care must be taken with the choice of
reference electrode when studying the alloying effect of Pt electro-
catalysts. Even a small leakage of Cl− (<5 ppm) from the reference
Ni is retained in Pt10Ni90 than in Pt75Ni25
.
3.4. ICP analysis
From EDS analysis, it was found that the Ni dissolution takes
place predominantly during the initial stages of potential cycling.
So, to complement the above findings, the amount of Pt and Ni
dissolved in the electrolyte was analyzed at the end of 15 CVs by
ICP elemental analysis. In all the cases, be it in blank HClO4 or in
the presence of Cl−, the exact quantity of Pt could not be estab-
lished. The amount of Pt found in the electrolyte after 15 CVs was
below the detection limit of the ICP (0.002 mg l−1). The very low
concentrations of Pt in the electrolyte might be due the redeposi-
tion of dissolved Pt on the catalyst surface during the subsequent
potential scans. However, as understood from the CVs (Fig. 4b) the
dissolution of Pt was high in chloride-rich electrolyte. On the other
hand, the amount of dissolved Ni from the alloys could be deter-
mined accurately and the data are given in Table 2. For Pt75Ni25
,

K. Jayasayee et al. / Electrochimica Acta 56 (2011) 7235–7242
7241
−
Fig. 8. XPS spectra of Cl and ClO4 2p (a) and Ni 2p (b) in Pt10Ni90 after 15 CVs under various Cl− concentrations.
electrode will drastically increase the non-noble metal dissolution
rate and influence the ORR activity.
HCl), however, the complete electrodeposit dissolves upon cycling
a few 100 times. The ORR activity of Pt and PtNi alloys is severely
impacted by the presence of chloride in the electrolyte, decreasing
about two orders of magnitude at 10−4 M Cl−. After 1000 cycles the
ORR activity of PtNi is still somewhat higher than that Pt, indicat-
ing that, especially in the case of Pt10Ni90, some Ni is retained in
subsurface positions, but it is clear that the presence of Ni does not
mitigate the effect of Cl− on electrode performance.
3.6. Electrode surface analysis by XPS
As chloride-based precursors were used for the electrodepo-
sition of alloys, it is crucial to analyze the catalysts layer for the
presence of remnant chloride. The EDS and XPS analyzes could not
detect any chloride on the bulk and surface of the electrodes and
the respective results are shown in Fig. 8a.
In all, it can be concluded that even minimal amounts of Cl−
are deleterious for the ORR on Pt and PtNi and hence needs to be
avoided completely from the fuel cell systems.
The XPS data confirm the presence of adsorbed Cl− on the sur-
face of electrodes exposed to Cl−-containing electrolyte (Fig. 8a).
The peak at ∼199.0 eV and at ∼208.0 eV belong to Cl 2p position,
while the former peak represents Cl in chloride ion while the lat-
ter corresponds to Cl in ClO4−. This also suggests the adsorption
of weakly adsorbing perchlorate ions onto the platinum surface in
addition to strongly adsorbing Cl ions at more anodic potentials (in
this case a standby potential of 0.95 V is applied).
The Pt:Ni surface compositions of the fresh alloys and poten-
tially cycled (15 CV) in blank, chloride-containing electrolyte and
blank electrolyte with KCl saturated reference electrode at 80 ◦C
are shown in Table 2 and Fig. 8b. While the surface of Pt75Ni25 is
depleted of Ni atoms upon exposing to the 10−4 M HCl-containing
electrolyte, Pt10Ni90 retained about 10 atom% Ni under the same
conditions. However, the HCl concentration of 10−3 M in the elec-
trolyte completely leaches out the Ni atoms from the surface of both
the alloys. The absence of Ni atoms on the surface of the electrodes
that were open to KCl saturated reference electrode could also be
partially due, of course, to the high operating temperature (80 ◦C)
besides the chloride-ion effect.
Acknowledgements
This work was part of the Dutch EOSLT Consortium PEMFC, Con-
tract Nos. EOSLT 06005 and EOSLT 07005, supported by the Ministry
of Economic Affairs.
Ad Wonders, Ajin Cheruvathur and Adelheid Elemans-Mehring
(all from Eindhoven University of Technology) are gratefully
acknowledged for their technical support.
References
[1] Y. Nagahara, S. Sugawara, K. Shinohara, J. Power Sources 182 (2008) 422.
[2] W. Schmittinger, A. Vahidi, J. Power Sources 180 (2008) 1.
[3] A.J. Steinbach, C.V. Hamilton Jr., M.K. Debe, ECS Trans. 11 (2007) 889.
[4] R. Mohtadi, W.K. Lee, J.W. Van Zee, J. Power Sources 138 (2004) 216.
[5] A.B. Ofstad, M.S. Thomassen, J.L. Gomez de la Fuente, F. Seland, S. Moller-Holst,
S. Sunde, J. Electrochem. Soc. 157 (2010) B621.
[6] T.J. Schmidt, U.A. Paulus, H.A. Gasteiger, R.J. Behm, J. Electroanal. Chem. 508
(2001) 41.
[7] M.S. Mikkola, T. Rockward, F.A. Uribe, B.S. Pivovar, Fuel Cells 7 (2007) 153.
[8] F.J. Luczak, US. Patent 4447506, 1984.
[9] M. Watanabe, M. Uchida, S. Motoo, J. Electroanal. Chem. 229 (1987) 395.
[10] M.W. Breiter, Electrochim. Acta 8 (1963) 925.
4. Conclusion
[11] B.E. Conway, D.M. Novak, J. Mozota, J. Electrochem. Soc. 128 (1981) C116.
[12] V. Stamenkovic, N.M. Markovic, P.N. Ross, J. Electroanal. Chem. 500 (2001) 44.
[13] N.M. Markovic, H.A. Gasteiger, B.N. Grgur, P.N. Ross, J. Electroanal. Chem. 467
(1999) 157.
[14] N.M. Markovic, P.N. Ross, Surf. Sci. Rep. 45 (2002) 121.
[15] A. Zolfaghari, B.E. Conway, G. Jerkiewicz, Electrochim. Acta 47 (2002) 1173.
[16] A.P. Yadav, A. Nishikata, T. Tsuru, Electrochim. Acta 52 (2007) 7444.
[17] H.A. Gasteiger, S.S. Kocha, B. Sompalli, F.T. Wagner, Appl. Catal. B: Environ. 56
(2005) 9.
In this work we show that the presence of even small amounts
of Cl− in the electrolyte will rapidly increase the dissolution rate of
Ni from PtNi alloys. It is found that the dissolution takes place pre-
dominantly during the first few electrochemical cycles. Thereafter,
as follows from the elemental compositions, the electrode surface
is enriched with Pt (Pt-skin) that prevents further Ni dissolution
from the electrode bulk. At higher chloride concentrations (10−3 M

7242
K. Jayasayee et al. / Electrochimica Acta 56 (2011) 7235–7242
[18] K. Jayasayee, V.A. Dam, T. Verhoeven, S. Celebi, F.A. de Bruijn, J. Phys. Chem. C
113 (2009) 20371.
[26] P. Strasser, S. Koh, T. Anniyev, J. Greeley, K. More, C.F. Yu, Z.C. Liu, S. Kaya, D.
Nordlund, H. Ogasawara, M.F. Toney, A. Nilsson, Nat. Chem. 2 (2010) 454.
[27] T. Toda, H. Igarashi, H. Uchida, M. Watanabe, J. Electrochem. Soc. 146 (1999)
3750.
[28] V. Stamenkovic, B.S. Mun, K.J.J. Mayrhofer, P.N. Ross, N.M. Markovic, J. Ross-
meisl, J. Greeley, J.K. Norskov, Angew. Chem. Int. Ed. 45 (2006) 2897.
[29] K.J.J. Mayrhofer, S.J. Ashton, J. Kreuzer, M. Arenz, Int. J. Electrochem. Sci. 4 (2009)
1.
[19] S. Mukerjee, S. Srinivasan, J. Electroanal. Chem. 357 (1993) 201.
[20] U.A. Paulus, A. Wokaun, G.G. Scherer, T.J. Schmidt, V. Stamenkovic, V. Rad-
milovic, N.M. Markovic, P.N. Ross, J. Phys. Chem. B 106 (2002) 4181.
[21] R. Srivastava, P. Mani, N. Hahn, P. Strasser, Angew. Chem. Int. Ed. 46 (2007)
8988.
[22] R.R. Adzic, J. Zhang, K. Sasaki, M.B. Vukmirovic, M. Shao, J.X. Wang,
A.U. Nilekar, M. Mavrikakis, J.A. Valerio, F. Uribe, Top. Catal. 46 (2007)
249.
[30] K. Jayasayee, PhD Thesis, Eindhoven University of Technology, The Netherlands,
2011.
[23] K. Sasaki, Y. Mo, J.X. Wang, M. Balasubramanian, F. Uribe, J. Mcbreen, R.R. Adzic,
Electrochim. Acta 48 (2003) 3841.
[31] K.J.J. Mayrhofer, B.B. Blizanac, M. Arenz, V.R. Stamenkovic, P.N. Ross, N.M.
Markovic, J. Phys. Chem. B 109 (2005) 14433.
[24] V. Stamenkovic, T.J. Schmidt, P.N. Ross, N.M. Markovic, J. Phys. Chem. B 106
(2002) 11970.
[32] N.M. Markovic, R.R. Adzic, B.D. Cahan, E.B. Yeager, J. Electroanal. Chem. 377
(1994) 249.
[25] V. Stamenkovic, T.J. Schmidt, P.N. Ross, N.M. Markovic, J. Electroanal. Chem. 554
(2003) 191.
[33] V. Komanicky, H. Iddir, K.C. Chang, A. Menzel, G. Karapetrov, D. Hennessy, P.
Zapol, H. You, J. Am. Chem. Soc. 131 (2009) 5732.