Surface X-ray scattering studies of the growth of Pd thin films on the Pt(0 0 1) electrode surface and the effects of the adsorption of CO

Article abstract of DOI:10.1016/S0039-6028(03)00841-0

The morphology of electrochemically deposited Pd films on the Pt(0 0 1) electrode surface has been examined through the combination of cyclic voltammetry (CV) and in situ surface X-ray scattering (SXS). Analysis of SXS measurements has indicated that the Pd grows via pseudomorphic island formation, with the partial occupation of successive layers occurring at a first layer occupation of 0.8 ML. Further Pd deposition sees the formation of larger islands built onto the now complete monolayer, characteristic of pseudomorphic Stranski-Krastanov (SK) growth. In the H_UPD potential region the effect of CO on the surface expansion of the multilayer Pd film is negligible. In the hydrogen evolution region, however, the effect of the adsorption of CO has been shown to produce surface normal expansion and in-plane disorder of the Pd film. It is suggested that hydrogen permeation into the Pd film is enhanced on the CO-poisoned surface.

Full text of DOI:10.1016/S0039-6028(03)00841-0

Surface Science 540 (2003) 295–302
www.elsevier.com/locate/susc
Surface X-ray scattering studies of the growth
of Pd thin films on the Pt(0 0 1) electrode surface
and the effects of the adsorption of CO
a
M.J. Ball , C.A. Lucas , N.M. Markovic , V. Stamenkovic , P.N. Ross
a,
b
b
b
*
ꢀ
ꢀ
a
Oliver Lodge Laboratory, Department of Physics, University of Liverpool L69 7ZE, UK
b
Materials Sciences Division, Lawrence Berkeley National Laboratory, University of California, Berkeley, CA 94720, USA
Received 4 April 2003; accepted for publication 28 May 2003
Abstract
The morphology of electrochemically deposited Pd films on the Pt(0 0 1) electrode surface has been examined through
the combination of cyclic voltammetry (CV) and in situ surface X-ray scattering (SXS). Analysis of SXS measurements
has indicated that the Pd grows via pseudomorphic island formation, with the partial occupation of successive layers
occurring at a first layer occupation of 0.8 ML. Further Pd deposition sees the formation of larger islands built onto the
now complete monolayer, characteristic of pseudomorphic Stranski–Krastanov (SK) growth. In the HUPD potential
region the effect of CO on the surface expansion of the multilayer Pd film is negligible. In the hydrogen evolution region,
however, the effect of the adsorption of CO has been shown to produce surface normal expansion and in-plane disorder
of the Pd film. It is suggested that hydrogen permeation into the Pd film is enhanced on the CO-poisoned surface.
ꢀ 2003 Elsevier B.V. All rights reserved.
Keywords: Surface structure, morphology, roughness, and topography; Surface relaxation and reconstruction; Platinum; Palladium;
X-ray scattering, diffraction, and reflection; Solid–liquid interfaces; Electrochemical methods
1. Introduction
properties vary greatly from the bulk metal
counterparts and have been shown to be excellent
surfaces for catalysing the oxidation of small or-
ganic molecules and CO. In the case of CO it has
been shown that the Pt–CO bond is very strong
and thus can poison active Pt sites [1–9]. This has
led to the development of many bimetallic cata-
lysts which are able to oxidatively remove CO
from the surface at much more negative potentials
than on Pt: representative examples are Pt–Ru, Pt–
Mo. In contrast, for Pt modified with thin metal
films of Pd, despite apparently easier OH adsorp-
tion [10], the oxidation of CO occurs at higher
potentials [11,12].
Developments in surface science spectroscopic
techniques and UHV technology have, in recent
years, provided the impetus for extensive experi-
mental and theoretical study into adsorption
phenomena on single crystal metal surfaces. Of
particular interest from a catalysis/electrocatalysis
standpoint are bimetallic surfaces whose catalytic
* Corresponding author. Tel.: +44-151-794-3361; fax: +44-
151-794-3441.
E-mail address: clucas@liv.ac.uk (C.A. Lucas).
0039-6028/$ - see front matter ꢀ 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0039-6028(03)00841-0

296
M.J. Ball et al. / Surface Science 540 (2003) 295–302
The growth of Pd on Pt(h k l) single crystal
contamination. The surface was then immersed in
0.05 M H2SO4 (Baker, Ultrex) containing 10ꢁ5
surfaces has been well documented in recent years
[13–17] and on Pt(0 0 1) the formation of the sec-
ond layer was observed to commence before the
completion of the first [15]. Whereas the catalytic
significance of the Pt(0 0 1)–Pd surface has been
demonstrated with regard to the oxidation of
formic acid, for example in Ref. [18], to our
knowledge, little has been published on the effects
of CO adsorption and oxidation regarding this
surface. Electrochemically deposited Pd on the
Pt(1 1 1) surface has been shown to grow via
pseudomorphic Stranski–Krastanov (SK) growth
whereby a complete Pd monolayer is formed be-
fore the onset of three-dimensional island forma-
tion [19].
In this paper we describe the growth of elec-
trochemically deposited Pd films on the Pt(0 0 1)
electrode surface. In Section 3.1, the voltammetric
profiles from the Pt(0 0 1)–Pd surface over the
range of coverage are discussed. In order to gain
insight into the atomic structure of the Pd film,
SXS measurements from the Pt(0 0 1)–Pd ‘‘1 ML’’
and ‘‘n ML’’ surfaces (a notation used to describe
the Pd film in the high coverage regime, i.e. the
coverage equivalent of several Pd monolayers,
n P 2) are presented in Section 3.2. These two re-
gimes are initially defined by the approximate Pd
coverage inferred from the CV curves from the
Pt(0 0 1)–Pd surface, the precise level of coverage is
determined by the analysis of the crystal trunca-
tion rods (CTRs) measured from the surface. In
Section 3.3, we present data for the multilayer film
during the adsorption and oxidation of CO. These
results indicate that the adsorption of CO on the
multilayer Pd film appears to enhance the ab-
sorption of hydrogen into the Pd lattice.
M
PdO (Alfa Products), the Pd being deposited while
sweeping the potential at 20 mV sꢁ1. The amount
of Pd on the surface was controlled by the con-
tinuous change of the voltammetric features as-
sociated with Pd deposition [15,21]. Once the
approximate coverage had been attained, the Pd
modified Pt electrodes were removed, rinsed with
water and transferred to the X-ray cell containing
0.05 M H2SO4 free from Pd2þ ions. The SXS
measurements were performed on a four-circle
goniometer on beamline 7.2 at the Stanford Syn-
chrotron Radiation Laboratory (SSRL) using a
monochromatic incident X-ray beam of energy 10
keV defined by slits to be a 1 · 1 mm2 spot at the
sample. For the experiments involving CO, the
outer shell of the X-ray cell was purged with CO
(99.99% purity) instead of N2, which is usually
used to prevent contamination of the electrode by
the atmosphere. CO is able to saturate the solution
because it can diffuse through the polypropylene
film used to trap the electrolyte (further details
regarding the X-ray cell can be found in Ref. [22]).
The reciprocal space notation used is for the sur-
face tetragonal unit cell for the (0 0 1) surface,
which is related to the conventional cubic cell
by the transformations ð1; 0; 0Þ ¼ 1=2ð2; ꢁ2; 0Þ ,
t
c
ð0; 1; 0Þ ¼ 1=2ð2; 2; 0Þ , ð0; 0; 1Þ ¼ ð0; 0; 1Þ . All
t
c
t
c
electrode potentials are quoted versus a Pd/H2
reference electrode to an accuracy of 0.05 V.
3. Results and discussion
3.1. Cyclic voltammetry
The voltammetric response from the Pt(0 0 1)–
Pd surface has been well documented [15,21] and is
shown in Fig. 1(a). In agreement with previous
work, the base voltammetry of Pt(0 0 1) ðn ¼ 0Þ
shows three separate potential regions: the pure
HUPD region (0:1 V < E < 0:23 V) is followed first
by HUPD desorption/adsorption and bisulfate ad-
sorption/desorption (0:23 V < E < 0:41 V) and
then by the double layer potential region (E > 0:48
V). The voltammetric profile from the Pt(0 0 1)–Pd
ML surface (Fig. 1(a) ðn ꢂ 1Þ) sees the emergence
2. Experimental details
Before each experiment the Pt(0 0 1) single
crystals (miscut ꢀ0.2ꢁ) were prepared using the
flame annealing method and cooled in a hydrogen
atmosphere (for details regarding the flame an-
nealing method, see Ref. [20]). A droplet of ul-
trapure water was placed on the surface during
transfer to an electrochemical cell to prevent any

M.J. Ball et al. / Surface Science 540 (2003) 295–302
297
Fig. 1. (a) Cyclic voltammetry from the Pt(0 0 1) and Pt(0 0 1)–Pd surface in 0.05 M H2SO4 at a sweep rate of 20 mV sꢁ1. Subsequent to
Pd deposition, sharp peaks arise at ꢀ0.2 and 0.29 V which correspond to surface adsorption sites and step sites respectively. The 0.29 V
peak is due to adsorption/desorption at Pt and Pd step sites, which occur before the completion of the monolayer. (b) A schematic of
the Pt(0 0 1)–Pd surfaces derived from the CTR measurements at each level of Pd coverage.
of two main features. The first peak at ꢂ0.2 V
corresponds to HUPD adsorption/desorption ac-
companied by bisulfate desorption/adsorption
from the pseudomorphic Pd monolayer (see Sec-
tion 3.2). The second peak at ꢂ0.29 V also corre-
sponds to HUPD desorption/adsorption with
concurrent bisulfate adsorption/desorption, but
from the small number of Pd islands formed on the
top of the first Pd layer. A close inspection of Fig.
1(a) shows that there is also current in the poten-
tial range of 0:3 V < E < 0:45 V, which is most
likely due to HUPD desorption/adsorption from Pd-
free Pt atoms. Thus, the voltammetric studies in-
dicate that formation of the second layer of Pd
commences prior to the completion of the first.
The voltammetry from the Pt(0 0 1)–Pd n ML
surface is illustrated in Fig. 1(a). Notice that
depositing more than 1 ML of Pd produces vol-
tammetry in which the peak at 0.29 V is becoming
predominant while the peak at 0.2 V is hardly
visible. Furthermore, the fact that the pseudoca-
pacitance observed between 0:3 V < E < 0:45 V
on the surface covered by 1 ML of Pd is com-
pletely suppressed on Pt(0 0 1)–Pd modified by n
ML indicates that for n > 2 all Pt atoms are cov-
ered with Pd, as illustrated in the model of Fig.
1(b) and discussed in more detail below.
3.2. CTR measurements
The CTR data taken from the Pt(0 0 1)–Pd ML
surface with the potential held at 0.05 V are shown
as circles in Fig. 2. The solid line is a fit to the data
and the dashed line shows the scattering from an
ideally terminated Pt(0 0 1)-(1 · 1) surface, a con-
vention which is adopted throughout the paper.

298
M.J. Ball et al. / Surface Science 540 (2003) 295–302
relaxation of the topmost Pt atomic layer from the
bulk position, ePt, the Pt–Pd and Pd–Pd surface
normal spacing, dPt–Pd and dPd–Pd, and a Debye–
Waller type r.m.s. roughness of the surface of the
Pt electrode and the Pd adlayer, rPt and rPd re-
spectively. In order to detect any incommensurate
Pd structures, radial scans were performed at the
reciprocal lattice position (1, 0, 0.1). These scans
showed no evidence of peak splitting within the
resolution of the experiment, indicating that the Pd
is lattice matched in the surface plane to the Pt
substrate, i.e. the growth is pseudomorphic.
Therefore, the model assumes the Pd to be ad-
sorbed at fourfold hollow sites to continue the fcc
stacking of the bulk lattice. Including Pd atoms at
other adsorption sites gave calculated CTRs that
were not consistent with the measured data. The
results from the best fit gave an occupation of the
first Pd layer, hPd1 ¼ 0:8 ꢃ 0:1 ML with an occu-
pation of the second Pd layer of hPd2 ¼ 0:15 ꢃ 0:1
ML, as predicted by the electrochemical measure-
ments. The Pt–Pd layer spacing, dPt–Pd ¼ 2.00 0.01
ꢁ
A, corresponds to an expansion of ꢀ2% of the bulk
ꢁ
Pt(0 0 1) spacing (1.96 A) which is similar to that
Fig. 2. Non-specular CTRs from the Pt(0 0 1)–Pd ML surface
taken at an electrode potential of 0.05 V. The circles represent
the data; the solid line is a best fit to the data using a structural
model which includes pseudomorphic Pd adlayers as described
in the text. The dashed line is a calculation of the scattering
from an ideally terminated Pt(0 0 1)-(1 · 1) surface. The CTRs
are plotted as a function of the surface normal reciprocal lattice
vector cꢄ.
measured for the clean Pt(0 0 1) surface covered by
a ML of adsorbed HUPD [23]. The fractional oc-
cupation of the Pd adlayers provides the necessary
Pd step sites to produce a large contribution to the
voltammetry peak at 0.29 V. A schematic of the
Pt(0 0 1)–Pd ML surface is shown in Fig. 1(b).
Fig. 3 shows the CTRs measured from the
Pt(0 0 1)–Pd n ML surface, measured at an elec-
trode potential of 0.05 V. The model used to de-
scribe the multilayer employed a half-Gaussian
distribution of layer occupation. The Gaussian
was essentially a half distribution built on top of a
variable number of full layers, the peak of the
distribution being the layer with greatest occu-
pancy (for more details on the model see Ref. [19]).
The results from the best fit (solid line) give a
complete monolayer, with an occupation of
1.0 0.1 ML on top of which three-dimensional
island formation has taken place. The Pt–Pd layer
spacing is slightly contracted from that of the
Comparison between the data and the ideal calcu-
lation shows that the presence of adsorbed Pd
produces a significant reduction in the scattered
intensity midway between the Bragg reflections, at
the so-called ‘‘anti-Bragg’’ positions. Here the in-
tensity is dependent only on the surface layer of the
Pt and any adsorbed overlayers because the scat-
tering from the bulk Pt lattice cancels out. Since Pd
has a lower atomic form factor than that of Pt, the
presence of adsorbed Pd produces a large reduction
in the scattered intensity. This effect is similar to
that produced by a partially occupied surface Pt
layer. The structural model used to reproduce these
data included a variable number of partially filled
pseudomorphic Pd adlayers, with a coverage hPd
per Pt atom. Other structural parameters were the
ꢁ
Pt(0 0 1) bulk spacing at dPt–Pd ¼ 1.93 0.01 A,
which is close to the bulk Pd(0 0 1) spacing of 1.94
ꢁ
A. The Pd within the islands has a surface-normal
ꢁ
lattice spacing of 1.95 0.01 A i.e. also equal to the

M.J. Ball et al. / Surface Science 540 (2003) 295–302
299
surface the formation of the second layer begins
before the completion of the first and with further
deposition, the formation of three-dimensional is-
lands occurs. The SXS results provide strong evi-
dence in support of the assignment of the various
features in the voltammetric profiles recorded from
both the clean Pt(0 0 1), and the Pt(0 0 1)–Pd sur-
face at the two levels of Pd coverage: the ꢂ0.2 V
peak due to Pd terrace sites is greatly suppressed at
higher coverage; the ꢂ0.29 V due to Pd step sites
dramatically increases with higher coverage (which
suggests island formation); the feature associated
with Pt terrace sites is not present at higher cov-
erage (which suggests that the Pt surface is com-
pletely covered by Pd).
3.3. Adsorption of CO
To investigate the influence of CO adsorption
on the Pt(0 0 1)–Pd n ML surface, once the CTR
measurements had been taken, the electrolyte in
the X-ray cell was saturated with CO with the
potential being held at 0.05 V (in the HUPD region).
During saturation of the electrolyte with CO the
X-ray intensity was monitored at (1, 0, 2.73), a
position sensitive to lattice expansion in the Pd
film. No changes were observed in the X-ray in-
tensity, although HUPD is displaced from the Pd
surface by CO. This implies that there is no sig-
nificant change in the thin film structure although
it is not possible to rule out an expansion of the Pd
surface atoms. Sweeping the potential into the
hydrogen evolution region ()0.04 V), however, led
to a dramatic decrease in the X-ray intensity at (1,
0, 2.73). This is shown in Fig. 4(b) as the potential
is cycled from )0.04 to ꢀ0.6 V. It should be noted
that a similar effect was not observed in CO-free
electrolyte implying that CO adsorbed on Pd has a
special effect on the Pd film structure in the hy-
drogen evolution region.
Fig. 3. Non-specular CTRs from the Pt(0 0 1)–Pd n ML surface
taken at an electrode potential of 0.05 V. The circles represent
the data taken, the solid line is a best fit to the data using a
structural model which describes a half-Gaussian distribution
of layer occupation for pseudomorphic layers of Pd as de-
scribed in the text. The dashed line is scattering calculated from
an ideally terminated Pt(0 0 1)-(1 · 1) surface. The CTRs are
plotted as a function of the surface normal reciprocal lattice
vector cꢄ. Inset: A plot of the Pd layer occupancy derived from
the half-Gaussian distribution.
bulk Pd(0 0 1) spacing. The Pd multilayer reaches a
total of seven partially occupied Pd layers. Fig.
1(b) shows a schematic of the Pt(0 0 1)–Pd n ML
thin film structure.
In a previous publication, we have shown that
Pd films grow via pseudomorphic Stranski–Kras-
tanov growth on the Pt(1 1 1) surface with three-
dimensional islands described by a half-Lorentzian
distribution of layer occupation [19]. Here our
results show that, while it may not be true SK
growth, Pd films grow in much the same manner
on the Pt(0 0 1) surface i.e. via pseudomorphic is-
land formation. The main differences between the
two substrate orientations are that on the Pt(0 0 1)
In order to characterize the structural changes,
CTR data was measured at )0.04 V in CO-satu-
rated electrolyte. As previous results suggest that
Pd is not displaced from the Pt surface [12], in
order to fit the CTR data, the same structural
model was used as for the Pt(0 0 1)–Pd n ML sur-
face free from CO. Due to the scattering power of
CO relative to Pd and Pt, it us unnecessary to

300
M.J. Ball et al. / Surface Science 540 (2003) 295–302
Fig. 4. (a) CV from the Pt(0 0 1)–Pd n ML surface in 0.05 M H2SO4 at a sweep rate of 20 mV sꢁ1. (b) XRV at the reciprocal lattice
position (1, 0, 2.73), a position sensitive to the expansion of the lattice. (c) Rocking scans through the (1, 0, 2.73) position. The data
taken before the introduction of CO are represented by circles, the triangles are data taken subsequent to saturation of the electrolyte
with CO. The solid lines are fits of a Lorentzian lineshape to the data.
include CO in the structural model. The only pa-
rameters allowed to vary were the inter layer
spacings, dPt–Pd and dPd–Pd, and the relaxation of
the top most Pt layer. All other parameters were
fixed at the values obtained from the fit to the
Pt(0 0 1)–Pd n ML data. The results from the fit
show that the introduction of CO causes a signif-
icant expansion (ꢀ3%) of both the Pd–Pd spacing
in the Pd film and the Pt–Pd layer spacing. This
dramatic expansion of the Pd lattice is illustrated
by the plot in Fig. 5 which shows the ratio of in-
tensities along the non-specular CTRs before and
after the introduction of CO, I and ICO respec-
tively. The fit to this ratio data set is essentially a
three parameter fit, only allowing for the expansion
of the Pd lattice and relaxation of the Pt surface
layer.
has no effect on the expansion in the HUPD region.
A significant lattice expansion occurs over the
range )0.04 to 0.05 V, which coincides with the
potential region of the hydrogen evolution reac-
tion (HER). Furthermore, Fig. 4(c) shows rocking
scans performed at (1, 0, 2.73) before (circles) and
after (triangles) saturation of the electrolyte with
CO. Fits of a Lorentzian lineshape of the data are
shown as solid lines. At this reciprocal lattice po-
sition, the X-ray intensity is dominated by scat-
tering from the Pd film, therefore the FWHM of
these scans are inversely proportional to the in-
plane ordering of the Pd film. The domain size was
calculated from the values obtained for the
FWHM of the Lorentzian fit. The fits gave domain
ꢁ
ꢁ
sizes of ꢀ220 A (without CO) and ꢀ201 A (with
CO). Therefore the introduction of CO has led to a
reduction of the in-plane domain size by ꢀ8.5%,
indicative of a distortion of the Pd lattice. This
distortion is probably caused by the presence of
interstitial hydrogen. With this limited data set it is
impossible to propose a detailed mechanism for
the observed lattice expansion. It is interesting,
The CTR analysis indicates that the Pd expan-
sion is caused by CO-induced hydrogen absorp-
tion into the Pd lattice. Further evidence for this
can be seen in Fig. 4(a) and (b), which show the
CV and XRV measurements. Similarly to the
Pt(1 1 1)–Pd surface [19], the introduction of CO

M.J. Ball et al. / Surface Science 540 (2003) 295–302
301
been studied. We have shown that Pd grows epi-
taxially via the formation of three-dimensional
pseudomorphic islands, i.e. the Pd is lattice mat-
ched in-plane to the Pt substrate. These islands
give rise to characteristic features in the voltam-
metric response from the surface. The CTR results
show that for the Pt(0 0 1)–Pd 1 ML, the Pt–Pd
spacing is expanded from the bulk Pt–Pt spacing
by ꢀ2% in the presence of a monolayer of HUPD
.
This expansion is similar to that observed for the
clean Pt surface [23]. With higher Pd coverage, the
Pt–Pd spacing relaxes to the bulk Pt–Pt layer
spacing. Similarly, the Pd film adopts the bulk
Pd(0 0 1) lattice spacing. The growth mechanism of
the Pd film is similar to that observed for Pd thin
films on the Pt(1 1 1) surface [19], although for the
Pt(0 0 1)–Pd system the formation of three-
dimensional islands begins to take place prior to
the completion of the first monolayer. The ad-
sorption of CO onto the Pt(0 0 1)–Pd multilayer
surface produces significant surface normal ex-
pansion and in-plane distortion of the Pd lattice in
the potential region of hydrogen evolution. It is
postulated that this is due to enhanced absorption
of hydrogen on CO-covered Pd films.
Fig. 5. Plot of the ratio, R ¼ ICO=I, of the intensities along the
non-specular CTRs measured from the Pt(0 0 1)–Pd n ML
surface before (I) and after (ICO) the introduction of CO with
the potential held at )0.04 V, plotted as a function of the sur-
face normal reciprocal lattice vector cꢄ. These show large fluc-
tuations around the Bragg reflections at (1, 0, 1), (1, 0, 3) and (1,
1, 2) indicative of significant changes in the layer spacing within
the lattice. The solid line is a fit to the data (see text for details).
Acknowledgements
however, that chemisorbed CO, similarly to ele-
ments that act as surface poisons to the HER (such
as P, S, As, Se, Sb) [24], is able to promote hy-
drogen absorption into the Pd film. Although
there may be several explanations for this effect,
this behaviour can be associated with the modifi-
cation of physico-chemical properties of the Pd
film. As previously proposed for the Pt(1 1 1)-S
system [24], a promoting effect may related to the
local surface electronic effects which give rise to
electron transfer between adsorbed CO and HUPD
through the underlying metal.
We would like to thank the staff and adminis-
tration at SSRL, Stanford for providing out-
standing research facilities. MJB is supported by
an EPSRC studentship award and CAL ac-
knowledges the support of an EPSRC Advanced
Research Fellowship. This work was also sup-
ported by the Director, Office of Energy Research,
Materials Science Division (MSD) of the US De-
partment of Energy (DOE) under contract no.
DE-AC03-76F00098. Research was carried out in
part at SSRL, which is funded by the Division of
Chemical Sciences (DCS), US DOE.
4. Conclusions
References
The growth of electrochemically deposited Pd
films on the Pt(0 0 1) electrode surface and the
structural effects due to the adsorption of CO have
[1] B. Beden, S. Blimes, C. Lamy, J.M. Leger, J. Electroanal.
Chem. 149 (1983) 295.

302
M.J. Ball et al. / Surface Science 540 (2003) 295–302
[2] S. Sun, J. Clavilier, A. Berwick, J. Electroanal. Chem. 240
(1988) 147.
[13] J. Clavilier, M.J. Llorca, J.M. Feliu, A. Aldaz, J. Electro-
anal. Chem. 310 (1991) 429.
[3] L.-W. Leung, A. Wieckowski, M.J. Weaver, J. Phys.
Chem. 92 (1988) 6985.
[14] G.A. Attard, A. Bannister, J. Electroanal. Chem. 300
(1991) 467.
[4] J.M. Orts, A. Fernandez-Vega, J.M. Feliu, A. Aldaz, J.
Clavilier, J. Electroanal. Chem. 327 (1992) 191.
[5] B.E. Hayden, A.J. Murray, R. Parsons, D.J. Pegg, J.
Electroanal. Chem. 409 (1996) 51.
[15] M.J. Llorca, J.M. Feliu, A. Aldaz, J. Clavilier, J. Electro-
anal. Chem. 351 (1993) 299.
[16] G.A. Attard, R. Price, A. Al-Akl, Electrochim. Acta 39
(1994) 1525.
[6] C.A. Lucas, N.M. Markovic, B.N. Grgur, P.N. Ross, Surf.
Sci. 448 (2000) 65.
[17] M. Han, P. Mrozek, A. Wieckowski, Phys. Rev. B 48
(1993) 8329.
[7] C.A. Lucas, N.M. Markovic, P.N. Ross, Surf. Sci. 448
(2000) 77.
[18] M. Baldauf, D.M. Kolb, J. Phys. Chem. 100 (1996) 11375.
[19] M.J. Ball, C.A. Lucas, N.M. Markovic, V. Stamenkovic,
P.N. Ross, Surf. Sci. 518 (2002) 201.
[8] M. Ball, C.A. Lucas, N.M. Markovic, B.M. Murphy, P.
Steadman, T.J. Schmidt, V. Stamenkovic, P.N. Ross,
Langmuir 17 (2000) 5943.
[20] N.M. Markovic, M. Hanson, G. McDougal, E. Yeager, J.
Electroanal. Chem. 214 (1986) 555.
[9] N.M. Markovic, B.N. Grgur, C.A. Lucas, P.N. Ross,
Langmuir 16 (2000) 1998.
[21] G.A. Attard, R. Price, Surf. Sci. 448 (2000) 65.
[22] I.M. Tidswell, N.M. Markovic, C.A. Lucas, P.N. Ross,
Phys. Rev. B 47 (1993) 16542.
[10] B. Alvarez, V. Climent, A. Rodes, J.M. Feliu, Phys. Chem.
Chem. Phys. 3 (2001) 3269.
[23] N.M. Markovic, C.A. Lucas, B.N. Grgur, P.N. Ross, J.
Phys. Chem. B 103 (1999) 9616.
[11] B. Alvarez, V. Climent, A. Rodes, J.M. Feliu, J. Electro-
anal. Chem. 497 (2001) 125.
[24] Y. Sung, W. Chrzanowski, A. Wieckowski, A. Zolfaghari,
S. Blais, G. Jerkiewicz, Electrochim. Acta 44 (1998) 1019,
and references therein.
[12] C.A. Lucas, N.M. Markovic, M. Ball, V. Stamenkovic, V.
Climent, P.N. Ross, Surf. Sci. 479 (2000) 241.