Oxygen reduction at platinum modified gold electrodes

Article abstract of DOI:10.1016/S0013-4686(03)00529-2

The reduction of oxygen has been studied on polycrystalline gold electrodes modified by platinum deposited spontaneously from an aqueous K ₂PtCl₆ solution via the displacement of copper or lead adlayers. The change in the surface composition and morphology has been checked by XPS, AES and AFM data. The kinetic results have shown that such electrodes may present a higher catalytic activity compared to bulk platinum electrodes during a scan where the potential is made more positive and is thus clearly expressed by an hysteresis in the CV curves. The displacement of copper and lead deposits gave similar amplitudes of the hysteresis but the modified electrodes obtained via a lead deposit present a better stability upon cycling in acid solutions. The observed behaviour can be correlated to the modification of the M-OH formation and reduction on noble metals.

Full text of DOI:10.1016/S0013-4686(03)00529-2

Electrochimica Acta 48 (2003) 3909ꢁ3919
/
www.elsevier.com/locate/electacta
Oxygen reduction at platinum modified gold electrodes
M. Van Brussel a,1, G. Kokkinidis a,1,2, A. Hubin b,1, C. Buess-Herman a,1,
*
a Faculte´ de Sciences, Service de Chimie Analytique et Chimie des Interfaces, Universite´ Libre de Bruxelles, CP 255, Boulevard du Triomphe, B-1050
Brussels, Belgium
b Department of Metallurgy, Electrochemistry and Materials Science, Vrije Universiteit Brussel, Pleinlaan 2, B-1050 Brussels, Belgium
Received 7 November 2002; received in revised form 28 February 2003; accepted 4 March 2003
Abstract
The reduction of oxygen has been studied on polycrystalline gold electrodes modified by platinum deposited spontaneously from
an aqueous K2PtCl6 solution via the displacement of copper or lead adlayers. The change in the surface composition and
morphology has been checked by XPS, AES and AFM data. The kinetic results have shown that such electrodes may present a
higher catalytic activity compared to bulk platinum electrodes during a scan where the potential is made more positive and is thus
clearly expressed by an hysteresis in the CV curves. The displacement of copper and lead deposits gave similar amplitudes of the
hysteresis but the modified electrodes obtained via a lead deposit present a better stability upon cycling in acid solutions. The
observed behaviour can be correlated to the modification of the MÃ
/OH formation and reduction on noble metals.
# 2003 Elsevier Ireland Ltd. All rights reserved.
Keywords: Au electrode; Displacement deposition; Platinum; Oxygen reduction; Catalytic activity
1. Introduction
On Pt electrodes, which are the most active pure metal
surfaces, the oxygen reduction follows predominantly
the direct four-electron mechanism. As a result, the
preparation of supported platinum electrocatalysts has
received increasing attention in order to produce cheap
and/or efficient cathodes.
So far, only a limited amount of work has been
performed on the deposition of platinum from platinum
salt solutions at open-circuit potential. Electroless
Regarding its application in various technological
areas, the reduction of oxygen is since many years the
subject of intensive research [1]. In aqueous solutions,
oxygen reduction is a multi-electron reaction that may
proceed according to two overall pathways: a direct
four-electron reduction and a peroxide pathway that
involves a two-electron reduction leading to peroxide as
intermediate which can undergo further reduction or
deposition of Pt on freshly cleaved HOPG [4ꢁ6] involves
/
the reaction of incompletely oxidized functionalities
existing at defects on the graphite surface and is there-
fore difficult to control properly. Kokkinidis et al. [7,8]
have shown recently that the deposition of Pt takes place
on freshly polished titanium at open-circuit through a
displacement reaction between Ti(0) and dissolved
Pt(IV). However, it was found that the activity of the
modified electrode for the oxygen reduction was smaller
compared to that of a smooth platinum electrode.
Platinum overlayers on an Au(1 1 1) substrate have
also been grown under UHV conditions and their
reactivity has been studied using CO as a probe
molecule [9]. Surprisingly, the Au substrate has in-
creased the Pt overlayer reactivity. The authors explain
decomposition. The high dissociation energy of the OÃ
/
O bond (494 kJ molꢀ1) is responsible for the fact that
the four-electron reduction is observed only on some
metals while hydrogen peroxide with a dissociation
energy of 146 kJ molꢀ1 can more easily be reduced on
many surfaces [2,3].
* Corresponding author. Tel.: ꢂ
/
32-2-650-2939; fax: ꢂ32-2-650-
/
2934.
E-mail address: cbuess@ulb.ac.be (C. Buess-Herman).
1
ISE Member.
2
On leave from Laboratory of Physical Chemistry, Department of
Chemistry, Aristotle University of Thessaloniki, Thessaloniki, Greece.
0013-4686/03/$ - see front matter # 2003 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/S0013-4686(03)00529-2

3910
M. Van Brussel et al. / Electrochimica Acta 48 (2003) 3909ꢁ3919
/
the results on the basis of a simple model in which the
change in the CO binding energy is directly proportional
to the shift of the d-band of the metal overlayer.
According to this model, the increased activity of the
Pt/Au(1 1 1) system should be expected for other
adsorbates and reactions.
Very recently, we have reported preliminary experi-
ments relative to the performance of a gold-supported
platinum electrocatalyst [10] for oxygen reduction. Since
spontaneous deposition of Pt can be achieved by a
displacement reaction between the platinum ions and a
metal less noble than Pt, we have adopted the procedure
where Pt is deposited on Au by replacing Cu deposits.
Such a procedure was recently applied by Brankovic et
al. [11] for the formation of submonolayers of Pt by
displacement of Cu upd monolayers on Au(1 1 1).
Recently, the electrodeposition of platinum from
PtCl26ꢀ or PtCl24ꢀ solutions has also been achieved on
gold single crystal substrates and characterized by cyclic
2.2. Deposition of Pt on Au
The deposition of Pt onto the gold disk electrode is
performed according to an experimental procedure
described previously [10].
The overpotential deposition of Cu from a solution of
0.1 M HClO4ꢂ
10ꢀ2 M CuSO4 was performed by
applying a potential pulse of controlled duration. The
various deposition times (tCu,dep) were 3, 10, 40, 60 and
/
120 s. The initial potential was Eiꢃ
/
ꢂ0.77 V and the
/
final potential Efꢃ 0.03 V vs. RHE.
/
ꢀ
/
The displacement of Cu by Pt at open-circuit poten-
tial was carried out by immersing the Cu/Au electrode in
a solution of 0.1 M HClꢂ
10ꢀ3 M K2PtCl6 during 180 s.
/
In order to gather more information on the displace-
ment reaction and on the electrochemical behaviour of
the modified surfaces, two additional procedures were
devised to determine a possible influence either from the
gold substrate or from the displaced metal. The first
procedure involves the replacement of the gold substrate
with a platinum one. The second experiment is per-
formed by using a lead adlayer deposited on gold
instead of the copper adlayer. Since the driving force
for the displacement reaction is the potential difference
between the two electrochemical systems, the displace-
ment reaction can take place provided that the redox
potential of the displaced metal is lower than the redox
potential of the hexachloroplatinate/platinum system.
For the deposition of Pt on gold by displacement of a
lead adlayer, the overpotential deposition of Pb from a
voltammetry, STM and EQCM [12ꢁ15].
/
In the present paper we have examined the surface
modification and the activity for oxygen reduction in
acid and alkaline media of several platinum modified
gold electrodes which differ by the amount of copper
deposited in the opd region on a polycrystalline gold
substrate. Better insight into the activity of the modified
electrodes is also provided by varying the substrate and
the nature of the displaced metal.
solution 0.1 M HClO4ꢂ
10ꢀ2 M Pb(ClO4)2 was also
/
2. Experimental
performed by applying a potential pulse of controlled
duration. The deposition time (tPb,dep) was 60 s while the
immersion time of the electrode in the hexachloroplati-
2.1. Electrodes and electrolytes
nate solution was 180 s. The initial potential was Eiꢃ
/
ꢂ
/
0.77 V and the final potential Efꢃ 0.33 V vs. RHE.
/
ꢀ
/
The electrochemical measurements were performed
with an Autolab (Eco Chemie) PGSTAT 30 potentiostat
equipped with a Scan-Gen module connected to a three-
electrode cell. The working electrode was gold or a
platinum rotating-disc electrode (EDI Tacussel, 2 mm in
diameter). The rotating speed was controlled with a
Controvit unit from Tacussel. A Radiometer Ag/AgCl
(saturated with KCl) electrode, connected to the work-
ing electrode compartment by a Luggin bridge and a Pt
grid served as the reference and the counter electrode,
respectively. The working electrolyte was 0.1 M HClO4
or 0.1 M NaOH saturated with oxygen at 1 atm. The cell
was thermostated at 22 8C. All potentials given in this
paper are referred to the RHE.
The electrolyte solutions were prepared from Milli-Q
water (Millipore 18 MV cm) and HClO4 or NaOH
(Merck, suprapure). The other reagents were HCl 36%
(Merck, suprapure), CuSO4 (Merck, ACS reagent),
Pb(ClO4)2 (Aldrich), O2 3.5 (Messer) and K2PtCl6
(Alpha Aesar, Johnson Matthey GmbH).
2.3. Preparation of samples for AES, XPS and AFM
studies
Gold films epitaxially grown on glass, supplied by
Arrandee GmbH, were used for the preparation of the
samples for AES, XPS and AFM. These gold substrates
were annealed at 700 8C for 6 h prior to the experiment
and kept in Milli-Q water to avoid contamination.
The AES measurements were performed on a PHI-
590 equipment. The spectra were acquired with electron
beam energies of 3 and 10 keV. Relative sensitivity
coefficients were derived from a copperꢁplatinum
/
reference sample under the same experimental condi-
tions. The detector used was a CMA analyser (resolu-
tion set at 0.3%). The operational pressure in the
analyser chamber was 10ꢀ9 Torr.
The XPS measurements were performed using a VG-
CLAM II analyser and a dual anode (Mg/Al) X-ray
source powered at 300 W. The anode used for all spectra

M. Van Brussel et al. / Electrochimica Acta 48 (2003) 3909ꢁ
/3919
3911
was the Mg anode. The spectra are reported in binding
energy mode, with a pass energy ranging from 100 to 25
eV, depending on the experiment.
The microscope used for the AFM study was a
TopoMetrix Discoverer TMX 2000, operating in con-
tact mode. The tip was TopoMetrix manufactured, (ref.
number 1520-00-A) with a spring constant of 0.03 N
mꢀ1. Resolution is 400ꢄ
400 points for all images.
/
3. Results and discussion
For the sake of clarity, all the platinum modified gold
electrodes are labelled as follows:
Pt=(Mexs)=Au
where, Me is the displaced metal and xs represents the
deposition time in seconds of the metal to be displaced.
The Pt/(Me)/Au label should be regarded as a general
label for platinum modified gold electrodes obtained by
Me adlayer displacement.
The cyclic voltammogram obtained with a modified
electrode in oxygen-free 0.1 M HClO4 (Fig. 1) shows
typical features for both polycrystalline gold and
platinum.
Fig. 2. Currentꢁpotential curves for oxygen reduction recorded (A) in
/
the negative and (B) in the positive potential sweep direction on Pt/
(Cuxs)Au rotating-disc electrodes in O2-saturated 0.1 M HClO4 (dE/
3.1. Oxygen reduction on Pt/(Cu)/Au for different
deposition times in acidic medium
dtꢃ
20 mV sꢀ1). Rotation rate 1100 rpm. Deposition time x (s): (1) 3;
(2) 10; (3) 40; (4) 60; (5) 120. The insets A and B show mass-transport
/
corrected Tafel plots.
Rotating-disc voltammograms for oxygen reduction
were recorded on Pt/(Cu)/Au modified electrodes where
the deposition time, and consequently, the amount of
copper deposited, was changed while the displacement
time was kept constant. The deposition times were 3, 10,
40, 60 and 120 s. Fig. 2A and B clearly show that the
catalytic activity for oxygen reduction does not only
depend on the potential sweep direction but also on the
amount of copper deposited on the surface. Fig. 2
clearly shows that the halfwave potential for oxygen
reduction increases as the amount of copper is increased,
until it reaches a maximum value for Pt/(Cu60s)/Au. If
the amount of copper is further increased, the halfwave
potential decreases. Using values of the kinematic
viscosity, nꢃ
/
8.93ꢄ
1.93ꢄ
oxygen concentration, cO
/
10ꢀ3 cm2 sꢀ1, [16] the O2 diffusion
10ꢀ5 cm2 sꢀ1 [16] and the
1.22ꢄ
10ꢀ6 mol cmꢀ3
coefficient, DO
ꢃ
/
/
2
ꢃ
/
/
2
[16] the Levich analysis of the rotating-disc voltammo-
grams for oxygen reduction yielded a four-electron
reduction wave for all but the Pt/(Cu3s)/Au electrode,
where the number of electrons is equal to 3.6. This latest
observation is an indication that the amount of plati-
num is in this particular case too low to achieve limiting
currents at the level of a four-electron reduction. The
curves recorded for the Pt/(Cu40s)/Au and the Pt/(Cu60s)/
Au electrodes display a remarkable hysteresis between
the forward and backward scans. The potential differ-
ence between the positive and the negative scan for the
Pt/(Cu60s)/Au electrode amounts to almost 200 mV. This
unusual behaviour is not recorded on bulk platinum.
Both the Pt/(Cu40s)/Au and the Pt/(Cu60s)/Au samples
have a higher catalytic activity when compared to
polycrystalline platinum [1], but the Pt/(Cu60s)/Au
electrode shows the most promising behaviour towards
oxygen electroreduction.
Fig. 1. Cyclic voltammogram of a Pt/(Cu60s)/Au modified electrode in
50 mV sꢀ1). Initial potential is
indicated by the arrow.
oxygen-free 0.1 M HClO4 (dE/dtꢃ
/

3912
M. Van Brussel et al. / Electrochimica Acta 48 (2003) 3909ꢁ3919
/
The limiting currents in Fig. 2 are well expressed over
number of electrons involved in the oxygen reduction
reaction. The number of electrons involved is consistent
with the Levich analysis. The values of k derived from
a wide potential range. It seems that there is no influence
of Pt distribution and loading on the limiting current.
This is due to the fact that the spherical diffusion zones
corresponding to the Pt particles overlap and as a result
of this the diffusion of oxygen on the Pt/(Cu)/Au surface
is equivalent to the planar diffusion to the whole
surface. Therefore, the area-term in the Levich equation
should be the geometric area of the Au-disc (0.0314 cm2)
and not the exposed area of the Pt clusters. This is,
however, not the case in the kinetic study where the real
area of the electrode must be taken into account.
The halfwave potentials of the reduction waves
(positive scan) on Pt/(Cu60s)/Au measured at different
rotation rates are ca. 100 mV more positive than those
measured for smooth Pt [1] and comparable (or slightly
less negative) than those estimated from the polarization
curves, reported for the most active Pt(1 1 0) face by
Perez et al. [17] and Markovic et al. [18] in a perchloric
acid solution at the same rotation rates. It should,
however, be noted that E1/2 values can be considered
only as a qualitative criterion for the catalytic activity,
since E1/2 is known to depend on the real area of the
electrodes.
the KouteckyꢁLevich analysis are reported in Table 1.
/
The indicated values of k are, however, reported using
the geometric area of the Au electrodes. If the real
surface is taken into account, the derived values of k are
smaller, and in the specific case of Pt/(Cu60s)/Au it was
reported previously that the value of k at 0.8 V vs. RHE
was 5.21ꢄ
/
10ꢀ2 cm sꢀ1 using the geometric area and
3.17ꢄ
/
10ꢀ2 cm sꢀ1 using the real surface area estimated
from the hydrogen upd in oxygen-free HClO4. It must
be emphasized that the latter value is higher than the
reported value of Pt(1 1 0) [17,18], the most active face
in the same experimental conditions.
The Tafel behaviour in the mixed kinetic-diffusion
control region was also examined by using mass-
transport corrected currents. The insets of Fig. 2 shows
the Tafel slopes obtained at 1500 rpm for the Pt/(Cu60s)/
Au electrode. The slopes were ꢀ
negative scan and ꢀ
76 mV decꢀ1 for the positive scan.
These values are very close to the value of ꢀ60 mV
decꢀ1, indicating Ptꢁ
OH formation and reduction play
/
65 mV decꢀ1 for the
/
/
/
a decisive role on the kinetics of oxygen reduction. Table
1 displays Tafel analysis data for other Pt/(Cu)/Au
electrodes.
The significant hysteresis on the polarization curves
for the Pt/(Cu40s)/Au and the Pt/(Cu60s)/Au shows that
the oxygen reduction reaction is much faster on a
reduced than on an oxidized surface of the Pt adlayer.
The Tafel data suggests that this hysteresis is related to
A kinetic analysis of the currentꢁ
made in the mixed kinetic-diffusion control region.
From the KouteckyꢁLevich analysis, kinetic currents
were calculated by means of the equation
/
potential curves was
/
1
1
1
ꢃ ꢂ
ik bv1=2
(1)
i
where i is the measured current, v is the electrode
rotation rate, ik is the kinetic current given by
the formation of PtꢁOH and its possible modification
by the substrate. The possible effect of the gold substrate
is further discussed in Section 3.4.
/
ik ꢃnFAkcO
(2)
2
To complete the study of Pt/(Cu)/Au electrodes in
acidic medium surface analysis techniques were used.
The displacement of the copper adlayer by platinum was
confirmed. However, the displacement time of 180 s is
not enough to replace the whole copper adlayer with
platinum. Auger and XPS measurements confirmed the
presence of residual copper on all samples. This
observation indicates that although the displacement
reaction is quite fast, the rate eventually drops to zero as
where A is the area of the electrode and b is the Levich
slope:
bꢃ0:62nFAcO (DO )2=3nꢀ1=6
(3)
2
2
KouteckyꢁLevich plots for various potentials yield
/
straight lines with intercepts corresponding to the
kinetic currents ik, which give the reaction rate constants
k. The slope of the straight lines allows us to assess the
Table 1
Kinetic parameters for oxygen reduction on Pt/(Cuxs)/Au in 0.1 M HClO4 (Tꢃ
/
22 8C)
tCu,dep (s)
Tafel analysis
KouteckyꢁLevich analysis (positive sweep)
/
Negative sweep
b (mV decꢀ1
Positive sweep
b (mV decꢀ1
0.65 V
0.80 V
n
)
109 j0 (A cmꢀ2
)
)
107 j0 (A cmꢀ2
)
102 k (cm sꢀ1
)
3
ꢀ
ꢀ
ꢀ
ꢀ
/
104
3.0
1.1
ꢀ
/
150
122
122
2.2
0.9
0.33
1.03
6.22
55.0
ꢁ
ꢁ/
ꢁ/
5.21
/
3.6
4.1
4.1
4.0
10
40
60
/
92
77
65
ꢀ
/
/
0.58
0.14
ꢀ
/
10.3
1.9
/
ꢀ
/
76

M. Van Brussel et al. / Electrochimica Acta 48 (2003) 3909ꢁ
/3919
3913
the reaction can only take place as long as copper is in
contact with the platinum containing solution. As the
reaction proceeds, the copper adlayer is covered by
deposited platinum until no exchange sites are left
available.
dimensions, in a particular grape-like structure. The
rather rough surface of these metallic particles probably
explains part of the increased catalytic activity.
Finally, the stability of the Pt/(Cu60s)/Au electrode
was examined by continuously cycling at 1500 rpm. Fig.
5 shows no decrease in the catalytic activity in the first
cycles, but as the cycling went on, a progressive slight
decrease in the catalytic activity was observed at the
positive branch of the hysteresis. After potential cycling,
AFM reveals that the surface morphology is completely
different (Fig. 4C). Roughness data were calculated by
means of Eq. (4) for different images.
AES experiments indicate that the amount of depos-
ited platinum increases as the displacement reaction
proceeds. Fig. 3AꢁD show a series of Auger spectra
/
recorded at different immersion times for the Pt/(Cu60s)/
Au electrode. The amount of platinum increases, how-
ever, the derived quantitative data show a decrease in
the exchange rate, which is consistent with the displace-
ment reaction mechanism.
Fig. 4A and B show the surface morphology of the Pt/
(Cu60s)/Au modified electrode, at different resolutions.
The modified surface exposes metallic particles homo-
geneously dispersed on the surface. The size distribution
of these particles is quite narrow. On Fig. 4A it appears
that most particles exhibit a more or less spherical shape
with a 100 nm radius. Upon closer inspection, (Fig. 4B)
it appears that each the particles consists in an
agglomeration of metallic particles of much smaller
ꢀ
ꢁ
N
N
X
X
1
1
Raꢃ
Zi ꢀ
Zi
(4)
j
j
N
N
iꢃ1
iꢃ1
Ra values are decreased by 9
/
50% after potential
cycling. Although the electrooxidation of the surface is
known to increase the roughness of the surface, it
appears that in this particular case, the potential cycling
led to a flattening of the metallic particles on the surface.
The decrease in catalytic activity as the cycling goes on is
most likely related to the change in surface morphology.
Fig. 3. AES spectra for the system Pt/(Cu60s)/Au at different immersion times. (A) 0 s; (B) 60 s; (C) 120 s; (D) 180 s.

3914
M. Van Brussel et al. / Electrochimica Acta 48 (2003) 3909ꢁ3919
/
Fig. 4. AFM contact mode images of Pt/(Cu60s)/Au at low resolution (A), high resolution (B), AFM contact mode images of Pt/(Cu60s)/Au after
20 mV sꢀ1) (C) and AFM contact mode image of Pt/(Pb60s)/Au.
continuous potential cycling in O2-saturated 0.1 M HClO4 (dE/dtꢃ
/
The underlying copper layer was possibly partially
dissolved during the potential cycling, which explains
the change in morphology. Metallic nanoparticles are
also known to change their shape upon potential cycling
[19].
3.2. Oxygen reduction on Pt/(Cu60s)/Au in alkaline
medium
The kinetics of oxygen reduction on Au-supported Pt
electrocatalysts were also studied in aqueous 0.1 M
NaOH solution. Fig. 6 shows currentꢁpotential curves
/
for oxygen reduction on the Pt/(Cu60s)/Au electrode as a
function of the rotation rate. Very stable polarization
curves were obtained. After 4 h of continuous cycling of
the potential at 20 mV sꢀ1 in the range between 1.67
and 0.03 V vs. RHE, the halfwave potentials remained
constant for both cathodic and anodic sweep directions.
An hysteresis is also present in the region of potentials
where the surface is covered by OHads and Oads, but the
potential difference between the forward and backward
scan is much lower than in the case of acid solutions.
The limiting current varies linearly with v1/2 (inset A
in Fig. 6). The number of electrons involved in the
reaction mechanism was estimated from the Levich
slope; it was found equal to four. Kinetic analysis in
the mixed kinetic-diffusion control region is also pre-
sented in Fig. 6 (inset B) in the form of Tafel plots
corrected for the mass-transport. Tafel slopes at low
62 mV decꢀ1 in
Fig. 5. Continuous potential cycling showing the deactivation of the
Pt/(Cu60s)/Au electrode for the reduction of oxygen in O2-saturated 0.1
currents were found to be ꢀ
the cathodic and anodic sweep direction, respectively.
At high currents, the Tafel slope increases to about ꢀ90
mV decꢀ1 in both sweep directions. The Tafel slopes
/
65 and ꢀ
/
M HClO4 (dE/dtꢃ
20 mV sꢀ1). Rotation rate 1500 rpm. The inset
shows the full voltammogram (first cycle) and its part, which is
/
/
demonstrated in the figure.

M. Van Brussel et al. / Electrochimica Acta 48 (2003) 3909ꢁ
/3919
3915
Fig. 6. Currentꢁ
Rotation rate (rpm): (1) 500; (2) 750; (3) 1100; (4) 1500; (5) 2100; (6) 3000. The insets show: (A) plot of jL vs. v1/2 (vꢃ
corrected Tafel plots (1) in the negative and (2) in the positive potential sweep direction at 1500 rpm.
/
potential curves for oxygen reduction on Pt/(Cu60s)/Au rotating-disc electrode in O2-saturated 0.1 M NaOH (dE/dtꢃ
/
20 mV sꢀ1).
/
2pf). (B) Mass-transport
obtained seem to be in line with the Tafel slopes of ꢀ
/
60
120 mV decꢀ1 reported for smooth Pt in alkaline
inspection of these particles reveals grape-like structures
similar to those reported for a copper adlayer. This
seems to indicate that the surface morphology does not
depend significantly on the nature of the displaced
metal.
and ꢀ
/
solutions, indicating some similarity between oxygen
reduction on smooth and Au-supported Pt electrocata-
lysts. Unlike the acidic medium, there was no significant
deactivation of the electrode upon cycling.
Fig. 7 shows the recorded rotating-disc voltammo-
grams for the Pt/(Pb60s)/Au modified electrode in acid
solution for oxygen reduction. The curves are similar to
those reported for the Pt/(Cu60s)/Au modified electro-
des. There is a significant hysteresis of about 200 mV
between the positive going scan and the negative going
scan. The remarkable similarity between the polariza-
tion curves recorded on the copper based electrodes and
the lead based electrodes confirms the fact that the
nature of the displaced metal does not have an
important effect on the behaviour of the electrode.
One peculiar feature of the potential hysteresis is that
the potential difference between the positive and nega-
tive going sweeps depends on the vertex potentials of the
3.3. Oxygen reduction on Pt/(Pb60s)/Au in acidic
medium
Additional experiments, in which the copper adlayer
was replaced by a lead adlayer were performed. The lead
adlayer was deposited potentiostatically at ꢀ330 mV vs.
/
RHE for 60 s. The displacement was carried out during
180 s. Fig. 4D shows a AFM image of the same surface.
The morphology is roughly the same as that obtained
for a 60 s copper deposit followed by an exchange of 180
s. Metallic particles are uniformly spread on the
substrate and have a radius of about 50ꢁ100 nm. Closer
/

3916
M. Van Brussel et al. / Electrochimica Acta 48 (2003) 3909ꢁ3919
/
Fig. 7. Comparison between the polarization curves of oxygen reduction obtained in O2-saturated 0.1 M HClO4 on Pt/(Pb60s)/Au at low positive end
potential and high positive end potential. (dE/dtꢃ
20 mV sꢀ1). Rotation rate: 1500 (rpm).
/
experimental procedure. If the positive end of the
voltammogram is decreased, the width of the hysteresis
is also decreased but the positive branch is unaffected.
On the other hand, the negative end potential of the
cyclic voltammogram has no effect on the electroche-
mical behaviour of the electrode when the cycling is
performed. Hysteresis phenomena have been reported
earlier for the formation and reduction of oxide films on
noble metals [20]. The present hysteresis for oxygen
reduction can therefore be correlated to the character-
Fig. 8. Comparison between the recorded XPS spectra of Pt/(Pb60s)/Au before and after continuous potential cycling. Mg-Anode, pass-energy: 50
eV, step 0.20 eV, dwell 50 ms.

M. Van Brussel et al. / Electrochimica Acta 48 (2003) 3909ꢁ
/3919
3917
istics of the oxide formation and reduction on the
modified electrode. If the electrode is not polarized too
positively before reducing the oxygen, the onset of
oxygen reduction in the negative going scan takes place
at a more positive potential since the amount of surface
OH is lower than in the case where the electrode is
oxidized. In the case of noble metals, the difference of
pathway in formation from that in reduction of the
oxide film is generally explained by the presence of a
place-exchange mechanism involving the metal and the
OH and O species. Such mechanism can be expected to
be modified when distinct metal atoms are present at the
electrode surface.
XPS measurements performed onto the Pt/(Pb60s)/Au
modified electrode, both before and after potential
cycling reveal that the amount of lead on the electrode
decreases as it dissolves upon cycling. Fig. 8 shows two
XPS spectra recorded before oxygen reduction and after
12 potential cycles. The amount of Pb decreases and
consequently the relative surface concentration of pla-
tinum increases. Moreover, when compared to XPS
spectra recorded on Pt/(Cu60s)/Au modified electrodes,
the amount of deposited Pt is roughly the same,
regardless if lead or copper was used. The dissolution
of the remaining metal while cycling, among other
effects, probably explains why there is a progressive
deactivation of the surface as the oxygen reaction
proceeds.
there is no significant hysteresis between the positive
and the negative going scan, the electrochemical beha-
viour for oxygen reduction being otherwise the same.
On the other hand, the replacement of the gold substrate
by a polycrystalline platinum one, yields a typical
polycrystalline platinum curve. This observation is an
indication that indeed the gold substrate plays a role in
the oxygen reduction reaction. Additional evidence
supporting this hypothesis is given by Fig. 10, which
shows an AFM image of the Pt/(Cu60s)/Pt electrode. The
morphology as well as roughness of the surface is
identical to those reported in Fig. 4A and D. The
remarkable catalytic activity can thus not be ascribed to
the surface roughness only. Possible complex bi-metallic
interactions and/or electronic effects are responsible for
this particular activity. More sensitive XPS measure-
ments are going to be carried out in order to detect a
possible surface alloying.
4. Conclusions
The study of the oxygen reduction on a polycrystal-
line gold electrode modified by the spontaneous deposi-
tion of platinum through replacement of a copper or
lead deposit has shown that such electrodes may present
an increased catalytic activity compared to bulk plati-
num electrodes. The high electrocatalytic activity is
observed only when the O2 reduction takes place during
a back potential scan and is thus clearly expressed by an
hysteresis in the CV curves. The results obtained at
constant time of immersion of the electrode into a
platinum salt solution have shown that the activity is
dependent on the amount of platinum displacing copper
and that a maximum in activity is reached. XPS and
AES data have revealed that the displacement reaction
was not complete in the various experimental conditions
used but a quantitative determination of the platinum
content was not achieved in this work since, prior to the
exchange, the deposition of copper has involved charges
which were in some instances equivalent to more than
100 monolayers.
3.4. Oxygen reduction on Pt/(Cu60s)/Pt and
polycrystalline platinum in acidic medium
Fig. 9 shows the recorded cyclic voltammograms for
oxygen reduction on polycrystalline platinum and on a
Pt/(Cu60s)/Pt electrode. On polycrystalline platinum,
The importance of the role of the displaced metal has
been investigated by performing experiments with lead
instead of copper. The displacement of copper and lead
deposits gave similar results and in particular the
amplitude of the hysteresis is not significantly dependent
on the nature of the displaced metal.
The increased activity revealed through the hysteresis
cannot be attributed to an increase of the roughness of
the deposit. Deposition of platinum on a massive
platinum electrode following exactly the same procedure
did not give rise to the presence of an important
hysteresis in the CV curves for oxygen reduction and
AFM images reveal similar deposit morphology on a
gold and a platinum substrate.
Fig. 9. Comparison between the polarization curves of oxygen
reduction obtained in O2-saturated 0.1 M HClO4 on (
/
) Pt/(Cu60s)/
*
Au and (---) Pt/(Cu60s)/Pt and a bare Pt rotating-disc electrodes (dE/
dtꢃ
20 mV sꢀ1). Rotation rate 1500 rpm.
/

3918
M. Van Brussel et al. / Electrochimica Acta 48 (2003) 3909ꢁ3919
/
Fig. 10. AFM contact mode image of Pt/(Cu60s)/Pt.
We suggest thus that the platinum modified gold
Acknowledgements
electrode must present sites which have different geo-
metric and energetic characteristics. Oxygen reduction
on bare platinum is known to involve the adsorption of
O2 which is favourably binded to the surface according
to a twofold bridge site. The presence of a gold substrate
may lead to a modification of the kinetic of the initial
stages of oxide formation and reduction.
The action ‘Research in Brussels’ is gratefully ac-
knowledged. O. Steenhaut is also acknowledged for his
technical support.
The distinct feature observed in NaOH and HClO4
solutions is indicative of the intervention of the change
References
in MÃ
/
OH formation. The deactivation observed by
[1] R. Adzic, in: J. Lipkowski, P. Ross (Eds.), Electrocatalysis, Wiley-
VCH, Inc, New York, 1998, p. 197.
cycling in acid media with the modified electrode
resulting from the copper deposit replacement is prob-
ably due to the diffusion of remaining Cu atoms through
the Pt film since it is known that copper upd layers
unlike lead cause an inhibition of the oxygen reduction
[2] R.R. Adzic, J.X. Wang, Electrochim. Acta 45 (2000) 4203.
[3] Y. Miyata, S. Asakura, Corros. Sci. 44 (2002) 589.
[4] J.V. Zoval, J. Lee, S. Gorer, R.M. Penner, J. Phys. Chem. B 102
(1998) 1166.
[5] I. Lee, K.-Y. Chan, D.L. Phillips, Ultramicroscopy 75 (1998) 69.
[6] I. Lee, K.-Y. Chan, D.L. Phillips, Appl. Surf. Sci. 136 (1998) 321.
[7] G. Kokkinidis, A. Papoutsis, D. Stoychev, A. Milchev, J.
Electroanal. Chem. 486 (2000) 48.
rate [21ꢁ
Further work is required to have a better insight into
OH and the chemisorption of O2
/23].
the formation of MÃ
/
[8] G. Kokkinidis, D. Stoychev, V. Lazarov, A. Papoutsis, A.
Milchev, J. Electroanal. Chem. 511 (2001) 20.
on these platinum modified electrodes.
Finally, the procedure used in this work may be
extended to produce multimetallic electrodes such as
[9] M.Ø. Pedersen, S. Helveg, A. Ruban, I. Stensgard, E. Lægsgaard,
J.K. Nørskov, F. Besenbacher, Surf. Sci. 426 (1999) 395.
[10] M. Van Brussel, G. Kokkinidis, I. Vandendael, C. Buess-Herman,
platinumꢁruthenium modified gold electrodes which
/
Electrochem. Commun. 4 (2002) 808ꢁ813.
/
[11] S.R. Brankovic, J.X. Wang, R.R. Adzic, Surf. Sci. 474 (2001)
can be easily obtained by displacement of a lead adlayer
deposited on gold. Such a procedure offers the advan-
tage of depositing both metals at the same time on the
substrate which may lead to a better homogeneity and a
possible higher activity of the modified surface.
L173.
[12] H.-F. Waibel, M. Kleinert, L.A. Kibler, D.M. Kolb, Electrochim.
Acta 47 (2002) 1461ꢁ1467.
/
[13] K. Uosaki, S. Ye, H. Naohara, Y. Oda, T. Haba, T. Kondo, J.
Phys. Chem. B 101 (1997) 7566.

M. Van Brussel et al. / Electrochimica Acta 48 (2003) 3909ꢁ
/3919
3919
[14] B. Gollas, J.M. Elliott, P.N. Bartlett, Electrochim. Acta 45 (2000)
3711.
[19] P. Shen, N. Chi, K.Y. Chan, D.L. Phillips, Appl. Surf. Sci. 172
(2001) 159.
[15] K.A. Friedrich, A. Marmann, U. Stimming, W. Unkauf, R.
Vogel, Fresen. J. Anal. Chem. 358 (1997) 163.
[16] R.R. Adzic, J. Wang, B.M. Ocko, Electrochim. Acta 40 (1995)
83.
[20] B.E. Conway, Prog. Surf. Sci. 49 (1995) 331.
[21] T. Abe, G.M. Swain, K. Sashikata, K. Itaya, J. Electroanal.
Chem. 382 (1995) 73.
[22] R.R. Adzic, in: W.E. O’Grady, P.N. Ross, F.G. Wills (Eds.),
Electrocatalysis, Proc. Symp., The Electrochem. Soc. Inc., Prin-
ceton, 1982, p. 419.
[17] J. Perez, H.M. Villullas, E.R. Gonzalez, J. Electroanal. Chem. 435
(1997) 179.
[18] N.M. Markovic, H. Gasteiger, P.N. Ross, J. Electrochem. Soc.
144 (1997) 1591.
[23] N. Markovic, E. Yeager, R.R. Adzic, ESM Extended Abstracts 85
(1) (1985) 919.