ATR-SEIRAS - An approach to probe the reactivity of Pd-modified quasi-single crystal gold film electrodes

Article abstract of DOI:10.1016/j.susc.2004.05.141

Quasi-single crystalline gold films of 20 nm thickness and preferential (1 1 1) orientation on Si hemispheres were modified by controlled potentiostatic deposition of Pd (sub-ML, ML, multi-L) from sulphate and/or chloride-containing electrolyte. The electrochemical properties of these model electrodes were characterised for hydrogen and (hydrogen-) sulphate adsorption as well as for surface oxide formation by cyclic voltammetry. Conditions were developed to fabricate defined and stable Pd monolayers. In situ ATR-SEIRAS (Attenuated Total Reflection Surface Enhanced Infrared Reflection Absorption Spectroscopy) experiments were carried out to describe the electrochemical double layer of Pd modified gold film electrodes in contact with aqueous 0.1 M H₂SO ₄ with focus on interfacial water and anion adsorption. Based on an analysis of the non-resonant IR background signal the potential of zero charge is estimated to 0.10-0.20 V (vs. RHE). CO was found to be weakly physisorbed in atop sites on Au(111-20 nm)/0.1 M H₂SO₄ only in CO saturated electrolyte. CO, deposited on a quasi-single crystal gold film modified with 1 ML Pd, is chemisorbed with preferential occupation of bridge sites and atop positions at step edges. Saturated CO adlayers, as obtained by deposition at 0.10 V, contain isolated water species and are covered by a second layer of hydrogen bonded water. Potentiodynamic SEIRAS experiments of CO electro-oxidation on Pd-modified gold film electrodes demonstrate clearly the existence of a pre-oxidation region. They also provide spectroscopic evidence that isolated water and weakly hydrogen bonded water are consumed during the reaction and that atop CO on defect sites is a preferential reactant. The simultaneous in situ monitoring of the potential- and time-dependent evolution of characteristic vibrational modes in the OH- and CO-stretching regions are in agreement with the Oilman ( reactant pair ) mechanism of CO oxidation.

Full text of DOI:10.1016/j.susc.2004.05.141

Surface Science 573 (2004) 109–127
www.elsevier.com/locate/susc
ATR-SEIRAS––an approach to probe the reactivity of
Pd-modified quasi-single crystal gold film electrodes
*
S.Pronkin, Th.Wandlowski
Institute for Thin Films and Interfaces ISG3, Research Centre Ju¨lich, 52425 Ju¨lich, Germany
Received 25 July 2003; accepted for publication 3 May 2004
Available online 29 July 2004
Abstract
Quasi-single crystalline gold films of 20 nm thickness and preferential (111) orientation on Si hemispheres were
modified by controlled potentiostatic deposition of Pd (sub-ML, ML, multi-L) from sulphate and/or chloride-contain-
ing electrolyte.The electrochemical properties of these model electrodes were characterised for hydrogen and (hydro-
gen-) sulphate adsorption as well as for surface oxide formation by cyclic voltammetry.Conditions were developed to
fabricate defined and stable Pd monolayers.In situ ATR-SEIRAS (Attenuated Total Reflection Surface Enhanced
Infrared Reflection Absorption Spectroscopy) experiments were carried out to describe the electrochemical double layer
of Pd modified gold film electrodes in contact with aqueous 0.1 M H2SO4 with focus on interfacial water and anion
adsorption.Based on an analysis of the non-resonant IR background signal the potential of zero charge is estimated
to 0.10–0.20 V (vs. RHE). CO was found to be weakly physisorbed in atop sites on Au(111-20 nm)/0.1 M H2SO4 only
in CO saturated electrolyte.CO, deposited on a quasi-single crystal gold film modified with 1 ML Pd, is chemisorbed
with preferential occupation of bridge sites and atop positions at step edges.Saturated CO adlayers, as obtained by
deposition at 0.10 V, contain isolated water species and are covered by a second layer of hydrogen bonded water.
Potentiodynamic SEIRAS experiments of CO electro-oxidation on Pd-modified gold film electrodes demonstrate
clearly the existence of a ‘‘pre-oxidation’’ region.They also provide spectroscopic evidence that isolated water and weakly
hydrogen bonded water are consumed during the reaction and that atop CO on defect sites is a preferential reactant.The
simultaneous in situ monitoring of the potential- and time-dependent evolution of characteristic vibrational modes in the
OH- and CO-stretching regions are in agreement with the Gilman (‘‘reactant pair’’) mechanism of CO oxidation.
ꢀ 2004 Elsevier B.V. All rights reserved.
Keywords: Infrared absorption spectroscopy; Carbon monoxide; Gold; Palladium; Metallic films; Electrochemical methods; Oxidation;
Physical adsorption; Chemisorption; Surface defects; Vibrations of adsorbed molecules
*
Corresponding author.Fax: +49 2461 61 3907.
E-mail address: th.wandlowski@fz-juelich.de (Th.Wandlowski).
URL: http://www.fz-juelich.de/isg/isg3/tw/index.php.
0039-6028/$ - see front matter ꢀ 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.susc.2004.05.141

110
S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
1. Introduction
electrochemical plating [36], sputtering [37,39] or
chemical deposition techniques [38,40].The suc-
cess of these studies is limited by one problem,
which still hampers the development of defined
structure–reactivity relations: The morphology of
the (electro-) catalytically active deposits is not
controlled and often rather arbitrary.
The establishment of structure–reactivity rela-
tions at the atomic or molecular level is one of
the major challenges of modern physical electro-
chemistry [1].Fundamental aspects of reactivity
in electrochemical systems can be approached
through the controlled modification of well-de-
fined single crystal electrode surfaces by foreign
metal overlayers [2].Such systems exhibit often
electronic, structural, chemical and catalytic prop-
erties quite different from those of the underlying
substrate or of the bulk deposited materials [3,4].
Of particular interest are pseudomorphic adlayers
obtained by controlled deposition of metal adlay-
ers on a noble metal substrate.Pd deposited on
Pt(hkl) [5–14] or Au(hkl) [15–30] electrodes repre-
sent model systems for these studies.Pd adlayers
were prepared by physical vapour deposition in
UHV [5,13], electrochemical [6,8,11,12,14,28] and
chemical [7,9,30] strategies.The electronic and
geometrical properties of pseudomorphic adlayers
can be tuned by choosing template substrates with
different electronic structures and lattice constants.
As a consequence, the (electrocatalytic) reactivity
of the metal overlayers can be significantly modi-
fied [8,9,13,29,31,32].
In this contribution we report the first in situ
surface enhanced infrared spectroscopy (SEIRAS)
[33,34] study of well defined ultrathin Pd deposits
on quasi single crystalline gold film electrodes
employing an internal attenuated total reflection
(ATR) configuration.Osawa et al.pioneered the
application of ATR-SEIRAS to equilibrium and
time-resolved studies at electrochemical interfaces
[33,34].This technique is based on a thin metal
film or metal colloid, which is used as a working
electrode; deposited onto a highly refractive prism,
such as Si or Ge.The IR beam passes through the
prism and is totally reflected at the metal/electro-
lyte interface [33–35].This in situ spectroscopic
approach provides direct access to chemical infor-
mation on structure and dynamics of the electrode/
electrolyte interface as well as on reactions taking
place there.ATR-SEIRAS investigations at (elec-
tro-) catalytically active electrified solid/liquid
interfaces were previously reported for rough Pt
[36–38] and Pt/Fe alloy [39] deposits prepared by
As an alternative approach we propose in this
paper the design of bimetallic electrodes, composed
of quasi-single crystalline template electrodes,
which are subsequently modified by controlled dep-
osition of Pt-group metals (submonolayer, mono-
layer and/or multilayer regime).This strategy
adapts the advantages of ATR-SEIRAS on coinage
metal substrates [33,34] and opens novel opportuni-
ties for in situ spectroelectrochemical studies of
structure and reactivity of Pt-group metals under
steady state and time-resolved conditions due to
the following reasons: (1) high and specific surface
sensitivity with an enhancement of the IR-response
of physisorbed and chemisorbed species up to ca.
100 times stronger than those in conventional
IRAS, (2) dominant first layer (Helmholtz region)
effect with a probing range of a few nanometers,
(3) no limitations due to mass transport and poten-
tial perturbation, e.g. the time resolution of the
experiment is only limited by the time constant of
double layer charging (<50 ls in the present config-
uration) [33,34,41], (4) application of (1)–(3) to the
fabrication and study of well-defined (electro-)
catalytic and SEIRAS-active Pt-group model
electrodes, such as Pd, Rh and Pt, employing quasi-
single crystalline coinage metal film templates.
The electrochemical preparation and character-
isation of defined Pd-modified gold film electrodes
benefits from the detailed knowledge on the initial
stages of Pd deposition on massive Au(hkl) elec-
trodes.The potentiostatic deposition of Pd on
Au(hkl) was studied with electrochemical tech-
niques, the quartz crystal microbalance (EQCM),
in situ scanning tunnelling microscopy (STM),
and surface X-ray scattering (SXS) [8,14–23,25–
29].Pd, the lattice constant of which is about
4.8% smaller than that of Au, grows epitaxially
on the low index faces of Au, and forms rather flat
overlayers [8,16–20,22,23,25].The Pd deposition
on Au(111) and Au(100) starts with a pseudo-
morphic layer in the underpotential regime, and

S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
111
proceeds subsequently at overpotentials.STM
measurements reveal up to four pseudomorphic
monolayers for Pd on Au(111).The largest terrace
size was found in 1 ML films, thicker deposits de-
crease the domain size and increase the surface
roughness [19,22].The modified Au(111) elec-
trodes approach the electrochemical properties of
massive Pd(111) for thicker deposits [8,15,26].
No Pd–Au surface alloy formation was found on
Au(111) [19], but clear evidence was reported on
Au(100) and Au(110) [20,25].Recently, the ad/
desorption of hydrogen/(hydrogen-) sulphate ions
and oxide formation/reduction of thin Pd overlay-
ers on Au(hkl) have been investigated employing
cyclic voltammetry and electrochemical transient
techniques [15,23,26].Electrocatalytic properties
were explored for the oxygen reduction [21], the
oxidation of adsorbed CO [29], formic acid [8], for-
maldehyde [23] and methanol [42].All studies
demonstrate the unique electrochemical character-
istics of the first Pd monolayer, a clear dependence
of these reactions on Pd film thickness and sub-
strate crystallography, and distinct differences to
the properties of bulk Pd(hkl) [15,26,29,43,44].
The present article aims to characterize the
electrochemical double layer and to explore the
reactivity of Pd modified quasi single crystalline
gold film electrodes employing simultaneously in
situ ATR-SEIRAS and electrochemical techniques.
The paper is organised as follows: experimental
details are summarised in Section 2.The potentio-
static formation and electrochemical characterisa-
tion of Pd-modified gold film electrodes will be
discussed next.Subsequently, in situ ATR-SEIRA
spectroelectrochemical studies to describe the dou-
ble layer properties of these model electrodes in sul-
phate containing aqueous electrolyte will be
presented.Finally, we will demonstrate the poten-
tial of our novel approach for real time electrocat-
alytic studies, choosing the adsorption and
oxidation of CO as test reactions.
ration (base pressure<2·10ꢀ7 mbar, deposition
rate 0.005–0.007 nmsꢀ1) of a 20 nm thick gold film
onto the flat side of a silicon hemisphere (25 mm
diameter).The metal coated silicon prism was
either attached to an all-glass holder for carrying
out electrochemical characterisation and deposi-
tion experiments in a hanging meniscus configura-
tion in a three electrode cell (EC) or to a vertical
spectroelectrochemical cell (IR-C) [41].The geo-
metrical area of the working electrode was either
4.9 cm2 (EC) or 1.54 cm2 (IR-C).The metal films
were electrochemically annealed during deoxygen-
ating of the electrolyte by cycling the potential in
the double layer region with 10 or 50 mVsꢀ1 to
generate well defined (111)-oriented terrace sites.
In the following we will represent these electrodes
by the abbreviation Au(111-20 nm).
2.2. Pd-deposition
The Pd adlayers were obtained by electrodeposi-
tion from (i) 0.1 M H2SO4 + 0.1 mM PdSO4 at 0.75
V or (ii) 0.1 M H2SO4 + 0.1 mM PdCl2 at 0.80 V vs.
real hydrogen electrode (RHE, see below).The
Au(111-20 nm) film electrode was brought in con-
tact with the carefully deaerated Pd-containing
electrolyte at open circuit (ꢁ1.05 V), and subse-
quently the potential was scanned with 50 mVsꢀ1
in the negative direction until the desired deposi-
tion potential was reached (for further details see
Section 3.2). The latter value was kept constant,
and simultaneously the corresponding charge flow
was monitored.After the completion of the deposi-
tion, the modified electrode was carefully rinsed
with Milli-Q water, and transferred into a Pd-free
electrolyte for surface characterisation.Overlayers
fabricated from Clꢀ-containing electrolyte were ex-
posed to an additional conditioning step, which
consists of polarisation at 0.10 V in 0.1 M H2SO4
to remove trace impurities of halide ions, and a
subsequent solution exchange.
2. Experimental
2.3. CO deposition
2.1. Quasi-single crystalline gold film electrodes
CO (4.7 N) was adsorbed on the bare or modi-
fied Au(111-20 nm) electrode surface at 0.10 V
by injection of a mixture of Ar/CO into the electro-
lyte for 30 min to achieve saturation coverage.
The quasi-single crystalline Au(111-20 nm) film
electrodes were prepared by electron beam evapo-

112
S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
Subsequently, the electrolyte was purged with Ar
to remove CO from solution for additional 30 min.
at the sample and the reference potentials, respec-
tively.Positive going bands indicate accumulation
of the species with respect to the reference
spectrum.
2.4. Electrolyte solutions
The solutions were prepared with Milli-Q water
(18 MXcm, estimated total organic content 2–3
ppb), H2SO4 (suprapure Merck), PdO (Aldrich,
99.999%) or PdCl2 (Aldrich, 99.999%). All electro-
lytes were deaerated with argon (5 N) prior to and
during each experiment.The measurements were
carried out at 23 1 ꢁC.The glassware was
cleaned in a 1:1 mixture of hot H2SO4 and
HNO3 followed by rinsing cycles with Milli-Q
water.
3. Results and discussion
3.1. Characterisation of the bare quasi-single
crystalline gold film electrodes
Fig.1 shows typical i vs. E curves of an electro-
chemically annealed Au(111-20 nm) film electrode
in 0.1 M H2SO4.The voltammograms exhibit well
defined features of preferentially (111) oriented
single crystalline electrodes, such as the lifting of
p
the (p · 3) reconstruction (P1) and the order/dis-
2.5. Electrochemical measurements
p
p
order phase transitions of a ( 3 · 7) (hydrogen)
sulphate overlayer (P2/P20).Excursion into the
oxidation region, at E > 1.30 V, shows a shoulder
OA1 assigned to an oxidation process involving
steps, and the characteristic peak OA2, which rep-
resents contributions from terrace sites [45,46].
Internal calibration experiments with massive
stepped single crystals Au (nnn-2) revealed an
equivalent miscut angle of 2–4ꢁ, which relates to
an average size of ideally smooth terraces of 19–
27 atoms [41].Based on a comparison of the
The electrochemical measurements were carried
out with an Autolab PGSTAT 30 in a three elec-
trode configuration.The reference electrode was
either a trapped hydrogen or a Hg/Hg2SO4/
Na2SO4 sat.electrode.All potentials in this paper
are quoted with respect to the real (trapped)
hydrogen electrode (RHE: ꢁ310 mV vs.SCE).
The counter electrode was a large area platinum
spiral.
2.6. SEIRAS set-up
The SEIRAS experiments were carried out in a
vertical spectroelectrochemical cell.Details of the
set-up were reported in a previous paper [41].
The IR-spectra were measured with a Bruker IF
66V/s Fourier Transform Spectrometer choosing
a spectral resolution of 4 cmꢀ1.Unpolarised radi-
ation from a globar source was focused onto the
electrode/electrolyte interface by passing through
the modified Si-hemisphere.The incident angle
was usually 70ꢁ referring to the surface normal.
The IR radiation totally reflected at the interface
was measured with a liquid nitrogen cooled detec-
tor (MCT 317, Colmar Technologies).Typically,
160 interferograms were scanned and co-added
into each single beam spectrum during a slow
potential scan.The spectra were plotted in absorb-
ance units defined as A=ꢀlog(I/Io), where I and Io
represent the intensities of the reflected radiation
Fig.1. Steady state cyclic voltammograms of a Au(111-20 nm)
film electrode in contact with 0.1 M H2SO4, scan rate 10
mVsꢀ1, double layer (dotted line) and ox/red region (solid line).
The inset shows a typical ex situ STM image of an electrode just
after e-beam deposition of a 20 nm gold film on silicon.

S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
113
charge balance in the oxidation/reduction region
[46], for a chemisorbed uracil monolayer [47] as
well as for Cu UPD [48] between carefully pre-
pared Au(111-20 nm) film electrodes and well de-
fined Au(nnn-2) stepped single crystals we
estimated the electrochemically active area.The
latter appears to be 1.6 times larger than the corre-
sponding geometrical area.
rent peaks P10 = 0.885 V and P1 = 0.93 V is
found in 0.85 V6E61.10 V. The charge con-
sumed during a typical cathodic scan is estimated
to 700 70 lCcmꢀ2 and corresponds approxi-
mately to the formation (dissolution) of a pseudo-
morphic Pd monolayer involving two electrons per
Au surface atom and considering the previously
determined scaling factor of 1.6 (for further discus-
sion see below and Fig.4 ).Peak position and
intensity do not change upon continuous cycling
in this interval.Extending the potential window
toward more negative values, e.g. E < 0.85 V, gives
rise to an additional peak P2 on the positive poten-
tial scan.The latter increases with more negative
return potentials and/or increasing waiting time
at E < 0.85 V. The comparison of these observa-
tions with previously reported results for the Pd
deposition on Au(111) massive single crystals
[19] supports the following assignment: The pair
of peaks P1/P10 represents the underpotential dep-
osition (UPD) of a Pd monolayer on Au(111-20
nm) with preferential contributions from nuclea-
tion at defect (step) sites, and P2 corresponds to
the dissolution of bulk deposited Pd.The equilib-
rium potential for Pd deposition from 0.1 M
3.2. Pd-covered quasi-single crystalline gold film
electrodes––electrochemical deposition and
characterisation
3.2.1. Deposition
The fabrication of Pd-modified Au(111-20 nm)
film electrodes was carried out in the presence and
in the absence of chloride-containing electrolyte.
Fig.2 shows typical current–potential curves for
0.1 mM PdCl2 in 0.1 M H2SO4 recorded at 1
mVsꢀ1.The starting potential was chosen at 10. 5
V, which is more positive than Epzc = 0.550 V of
the Au(111-20 nm) film electrode in contact with
0.1 M H2SO4 [41].Under these conditions we
may assume that the electrode surface is covered
with adsorbed tetrachloropalladate (PdCl42ꢀ) ions
[19].During the negative and the subsequent pos-
itive potential scan one characteristic pair of cur-
2+
H2SO4 + 0. 1 mM Pd was determined to 0.82 V
[49].This value shifts more positive with higher
concentration of Pd2+, and the deposition rate in-
creases due to faster mass transport.Based on
these experiments, we have chosen the following
conditions for the modification of Au(111-20
nm) with Pd: sub-ML and ML Pd-coverages were
obtained by immersing the electrochemically an-
nealed electrode into deaerated electrolyte, 0.1 M
H2SO4 + 0.1 mM PdCl2 at 1.05 V, scanning the
electrode potential with 50 mVsꢀ1 to a deposition
potential in the UPD range, and waiting until the
desired charge flow through the interface was con-
sumed.Pd multilayers were fabricated by diffu-
sion-controlled electrodeposition at Ed = 0.80 V.
The inset of Fig.2 shows a typical current vs.
time transient representing the formation of
approximately 1 ML Pd on Au(111-20 nm).
Emersion of the electrode in the potential range
0.80 V6E60.90 V, careful rinsing, and subse-
quent transfer into a Pd2+-free solution results
in a Au(111-20 nm) film modified irreversibly
with Pd.
Fig.2. Current vs.potential curves for Au(111-20 nm)/01. M
H2SO4 + 0.1 mM PdCl2, 1 mVsꢀ1 with negative return poten-
tials of 0.85 V (solid line) or 0.80 V (dotted line). The inset
shows a typical chronoamperometric transient for Pd deposi-
tion triggered by a fast potential scan from 1.05 V!0.80 V (50
mVsꢀ1) and subsequent waiting at 0.80 V.

114
S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
Similar experiments were also performed with
ative region I, E < 0.15 V, represents the hydrogen
absorption, which occurs at higher overpotentials
on Au(111-20 nm) as compared with massive
Pd(111), and develops more pronounced with
increasing film thickness (Fig.3B and D ).Region
II, 0.15 V6E60.40 V, is attributed to the coupled
adsorption/desorption of hydrogen and (hydro-
gen-) sulphate [15].Up to 1 ML equivalent of
deposited Pd only one single pair of peaks with
0.1 M H2SO4 + 0.1 mM PdSO4.The UPD of Pd
is rather slow and starts approximately 50 mV
more negative than in chloride-containing electro-
lyte.Repeated cycling in the UPD range yielded
irreversible changes of the corresponding current
vs.potential curves.In consequence, Pd-deposi-
tion on Au(111-20 nm) from chloride-free electro-
lyte was carried out at Ed = 0.75 V, after
immersion of the electrode at 1.05 V, and scanning
the electrode potential with 50 mVsꢀ1 to Ed.
0
EP2 = 0.30 V (positive scan) and EP2 = 0.27 V
(negative scan) develops.The corresponding total
charge is estimated to QHA ꢂ 200 20 lCcmꢀ2
.
3.2.2. Characterization
With increasing Pd film thickness the peaks P2/
P20 increase in magnitude, but remain asymmetri-
cal.Starting from 2 ML Pd on Au(111-20 nm), a
second anodic peak appears in region II at
EP3 ꢂ 0.27 V (Fig.3B and D ).The corresponding
cathodic peak was only resolved separately for Pd
overlayers fabricated in Clꢀ-containing electrolyte,
Typical voltammograms for Pd adlayers of 1
ML thickness (A, C) as well as for variable cover-
ages (B, D) on Au(111-20 nm)/0.1 M H2SO4, fab-
ricated either in chloride-containing (A, B) or
chloride-free (C, D) electrolyte, are presented in
Fig.3 .Comparison with experiments on massive
Au(111) [15,29] and Pd(111) [44,50] single crystals
suggests a distinction between four different poten-
tial regions labelled I–IV in Fig.3 .The most neg-
0
EP3 = 0.265 V (Fig.3 ), but no other significant dif-
ferences were found between the two deposition
strategies.The observed hysteresis and the peak
Fig.3. Cyclic voltammograms for Pd deposited on Au(1 1 1-20 nm)/01. M H 2SO4 from 0.1 mM PdCl2 (A, B) or 0.1 mM PdSO4, 5
mVsꢀ1 (C, D).The curves represent single potential scans after starting at 0.50 V, and scanning first in the negative direction.(A) and
(C) represent 1 ML Pd, while (B) and (D) show current–potential curves of up to 3 ML Pd in two characteristic potential ranges.The
stability ranges of the various adlayer/substrate phases are labelled I–IV.

S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
115
shapes suggest the involvement of kinetically con-
trolled adsorption/phase formation transitions.
The systematic changes of the adsorption/desorp-
tion peaks for hydrogen and (hydrogen-) sulphate
are accompanied, as demonstrated in studies with
massive single crystals [15,19,26], by a gradual
structural transition from a pseudomorphic Pd
ML with lateral strain to thicker Pd films having
bulk properties.Therefore, we attribute the two
pairs of current peaks in region II to the presence
of laterally strained (P2/P20) and bulk-like (P3/P30)
Pd.In comparison to these rather broad peaks ob-
served on Au(111-20 nm) and Au(111) [15], nar-
row adsorption spikes were reported for Pd(111)
[26,44,50] and Pd–Pt(111) [5,7,14].
The potential interval III, which extends be-
tween 0.40 and 0.80 V, reveals a wide double layer
charging region without faradic contributions,
rather independent of the Pd coverage (we investi-
gated up to 5 equivalent ML Pd).Electrochemical
and in situ STM studies with Pd(111) and Pd–
Au(111) [15,44,50] suggest that the modified film
electrode is covered with specifically adsorbed
(hydrogen-) sulphate.At this stage we also like
to stress that extended potential cycling in 0.10 V
(I)6E60.80 V (III) does not change the current
potential profiles.This observation suggests that
surface alloying of Au and Pd [51] is either rather
slow or does not take place.
reduction as well as repetitive potential cycling
demonstrate that oxide formation causes irreversi-
ble changes, most probably related to Pd dissolu-
tion.Similar observations were also reported for
Pd(111) [50] and pseudomorphic Pd ML adsorbed
on an ideally smooth Au(111) surface [29].
3.2.3. Calibration ‘‘master curve’’
The above analysis of Pd–Au(111-20 nm) film
electrodes in 0.1 M H2SO4 reveals that the charge
density due to (hydrogen-) sulphate desorption/
hydrogen adsorption, QHA, and its anodic counter
part might be chosen as quantitative measures for
the composition of the surface structure.Compar-
ing the estimated values of QHA with the deposited
amount of Pd on Au(111-20 nm) film electrodes,
QPd, results in a ‘‘calibration master curve’’ with
two distinct slopes, 0.28 0.03 and 0.028 0.004,
respectively (Fig.4 ).The intersection point was
found at QPd = 0.72 0.08 mCcmꢀ2 and
QHA = 0.20 0.02 mCcmꢀ2.This plot includes
data from experiments carried out in the electro-
chemical (hanging meniscus configuration: j, d)
or the spectroelectrochemical cell (s) for Pd mod-
ification in the absence (d, s) as well as in the
presence (j) of chloride-containing electrolyte.
The potential region IV, at E > 0.80 V, is attrib-
uted to the surface oxidation of the Pd adlayers
[26,44,50].The single scan voltammograms as
monitored during the first positive potential excur-
sions for 1 ML Pd–Au(111-20 nm)/0.1 M H2SO4
(Fig.3A and C ) exhibit a characteristic oxidation
peak P4 at EP4 ꢂ 1.02 V and a corresponding
reduction feature P40 at EP4 ꢂ 0.75 V. The posi-
0
tion of P4 is approximately 0.10 V, respective
0.20 V more negative than for the same process
on Pd(111) [29,50] or 1 ML Pd–Pt(111) [7], and
its overall morphology compares reasonably with
results reported for 1 ML Pd–Au(111)-3ꢁ miscut/
0.1 M H2SO4 [29,44,50].The potential for the
reduction of 1 ML Pd oxide is just in-between
the published values for Pd(111) (ꢁ0.70 V)
[26,44,50] and an ideal pseudomorphic Pd ML
on a perfect Au(111) single crystal (ꢁ0.81 V)
[29].The charge balance between oxidation and
Fig.4. Calibration master curve.The charge density consumed
in the hydrogen adsorption/anion desorption region II (peaks
P20/P30), QHA, is plotted vs.the charge density of Pd deposition,
QPd, for experiments carried out in an electrochemistry cell in
hanging meniscus configuration (j, d) or in the SEIRAS cell
after subsequent steps of solution exchange (s) for depositions
of Pd from 0.1 M H2SO4 + 0.1 mM PdSO4 (d, s) or 0. 1 mM
PdCl2 (j).

116
S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
The formation of a pseudomorphic Pd ML on an
ideally smooth Au(111) surface consumes a
charge of 0.444 mCcmꢀ2.Taking this reference
and considering the ‘‘electrochemical roughness’’
of our Au(111-20 nm) templates of 1.6, we con-
clude that the charge QPd = 0.72 mCcmꢀ2 at the
intersection point corresponds to approximately
one equivalent ML of Pd.The charge QHA is
slightly lower than the accordingly scaled value
for 1 ML Pd–Au(111)/0.1 M H2SO4 [15], but
agrees well with preliminary data of CO charge
displacement experiments for hydrogen (E = 0.10
V, Q+ = 0.15 0.02 mCcmꢀ2, cf. Fig.9A ) and
anion adsorption (E = 0.40 V, Q = ꢀ0.06 0.02
ꢀ
mCcmꢀ2).The small difference to the charge bal-
ance of an ideal Au(111) surface might indicate
that, under our experimental conditions, the con-
sumption of one equivalent ML charge of Pd dep-
osition does not completely cover the entire
Au(111-20 nm) electrode.The emerging growth
of a second Pd layer, even not yet detectable in
the current potential profiles for QPd ꢂ 0.72 0.14
mCcmꢀ2 (Fig.3 ), might take place at local spots.
Nevertheless, the above data and discussion
demonstrate that we can establish in our experi-
ments conditions to fabricate well-defined Pd
monolayers on Au(111-20 nm) templates.These
electrodes were employed in the in situ SEIRAS
studies to explore double layer structure and reac-
tivity in the absence and in the presence of CO as
model reactant.
Fig.5.Selected SEIRA spectra of Au(111-20 nm)/01. M
H2SO4 (grey lines) + 1 ML Pd (black lines) as a function of
potential.The spectra were acquired during a potential sweep
with 10 mVsꢀ1 between 0.10 and 0.80 V vs. RHE. The reference
potential is 0.17 V. Every plotted spectrum represents the
average of a potential interval of 50 mV.
The following observations motivate and support
this hypothesis.(i) The potential of zero charge,
Epzc, of 1 ML Pd–Au(111) and of Pd(111) in
the absence of specifically adsorbed ions have been
determined as 0.23 and 0.20 V [28].These values
are 0.32 V respective 0.35 V more negative than
Epzc of an ideal Au(111)-(1 · 1) electrode in con-
tact with 0.01 M KClO4 [52].(ii) Petrii et al.esti-
mated Epzc ꢂ 0.18 V for Pd/0.05 M Na2SO4 +
0.001 M H2SO4 [53].(iii) The Epzc of Au(111)/
0.1 M HClO4 +5·10ꢀ3 M K2SO4 assumes a value
approximately 0.09 V more negative than in bare
0.1 M HClO4.(iv) the Epzc in Au(111-20 nm)/0.1
M H2SO4 is estimated to 0.55 V [41,54].(v) The
potential of zero total charge [55], Epztc, of 1 ML
Pd–Pt(111) is approximately 0.09 V more negative
for specifically adsorbed (hydrogen-) sulphate in
0.1 M H2SO4, in comparison to 0.1 M HClO4
[11,12].
3.3. In situ SEIRA spectra of 1 ML Pd on
quasi-single crystalline gold film electrodes
3.3.1. Overview
Fig.5 shows a typical series of SEIRA spectra
for 1 ML Pd–Au(111)/0.1 M H2SO4 (black lines)
as measured in 0.10 V6E60.80 V, e.g. within
the double layer regions II and III (cf. Fig.3 ), dur-
ing a positive potential sweep with 10 mVsꢀ1.The
grey lines represent the SEIRA spectra of the bare
gold film before modification with Pd.A set of sin-
gle beam spectra acquired at 0.17 V, e.g. within the
potential region of hydrogen adsorption, was cho-
sen as reference [15].This potential is expected to
be rather close to the potential of zero charge,
Epzc, of 1 ML Pd–Au(111-20 nm)/0.1 M H2SO4.

S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
117
As the potential advances to more positive val-
ues, at EP0.20 V negative going bands start to
appear at 3440 cmꢀ1 and at 1616 cmꢀ1.These
bands are assigned to OH-stretching (tOH1) and
HOH-bending (dHOH1) modes of weakly hydrogen
bonded interfacial water [41,54,56–59].The nega-
tive sign of these bands indicates that the amounts
of the adsorbed species decrease with respect to the
reference spectrum at 0.17 V. Their intensities are
continuously increasing in negative direction until
saturation is reached at approximately E ꢂ 0.40 V.
The peak position of dHOH1 is rather independent
of the electrode potential in 0.10 V < E < 0.80 V,
while tOH1 changes to slightly higher values.At
E P 0.25 V a new positive going band appears at
1090 cmꢀ1.This band is attributed to the tSO mode
of adsorbed sulphate species [12,41,54].The peak
position of this band shifts at more positive poten-
tials to higher wavenumbers.
Simultaneously with the adsorption of sulphate
a broad, positive going band starts to appear at
tOH2 ꢁ3150 cmꢀ1.Band shape and position sug-
gest the formation of strongly hydrogen bonded
interfacial water, co-adsorbed with sulphate
[41,54,58].The corresponding dHOH2 band was
too weak to be detected within the potential region
of an ideally polarised Pd–Au(111-20 nm) elec-
trode.Surface oxidation of Pd starts to interfere
at E > 0.80 V. Nevertheless, this limited potential
range and the low values of tOH2 demonstrate that
water interacts stronger with Pd than with the bare
gold surface [59].The quantitative deconvolution
and the more detailed interpretation of the com-
plex spectra, as plotted in Fig.5 , will require sys-
tematic spectroelectrochemical investigations of
Au(111-20 nm) with variable, but defined Pd cov-
erages to approach bulk Pd properties, and the di-
rect comparison with double layer characteristics
of well-defined Pd(hkl) massive single crystals in
contact with 0.1 M H2SO4.This work is still in
progress and will be reported elsewhere [60].
Fig.6. Plots of the integrated intensities (A), and of the peak
positions (B) of the tSO mode of the adsorbed sulphate species
for Au(111-20 nm)/0.1 M H2SO4 (s) + 1 ML Pd (h) and 1 ML
Pd + CO (D) as a function of potential.(C) illustrates the
potential dependence of the intensity of the non-resonant
background signal at t = 2000 cmꢀ1 for the bare gold surface
in the absence (s, d) and in the presence (h, j) of 1 ML Pd as
extracted from SEIRA spectra such as shown in Fig.5 during a
positive (s, h) and a negative going potential scan, respectively.
For comparison, the current–potential curve of 1 ML Pd–
3.3.2. Sulphate adsorption
A more detailed analysis of the anion adsorp-
tion shows that (hydrogen) sulphate starts to ad-
sorb just in the potential range of the
characteristic pair of peaks P2/P20 (Fig.6 ).The
integrated intensity of the tSO vibration increases
Au(111-20 nm)/0.1 M H2SO4 is also added, scan rate 5 mVsꢀ1
.

118
S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
with more positive potential (Fig.6A ).However,
the rise is not as abrupt as found for 1 ML Pd–
Pt(111)/0.1 M H2SO4 [12], and does not reach a
broad potential range of saturation.Simultane-
ously, the peak position shifts to higher wavenum-
bers with dtSO/dE = 85 10 cmꢀ1 Vꢀ1.This value
is smaller than for the disordered (hydrogen) sul-
phate, 132 10 cmꢀ1 Vꢀ1, but slightly larger than
reported for the ordered ( 3 · 7) sulphate ad-
layer on the Pd-free Au(111-20 nm) electrode,
63 3 cmꢀ1 Vꢀ1 [41,54].It is important to note
that sulphate adsorbs on the Pd-modified electrode
at more negative potentials than on the bare gold
surface (cf.arrows in Fig.6A ).The absolute wave-
numbers of tSO in the present experiment and for
Au(111-20 nm)/0.1 M H2SO4 are rather close
and approximately 100 cmꢀ1 lower than the data
obtained for 1 ML Pd–Pt(111)/0.1 M H2SO4
[12].These observations demonstrate that the gold
film template still influences the double layer prop-
erties of the Pd–modified electrode.
deposition of 1 ML Pd.A shallow new minimum
develops around 0.10–0.20 V, followed by a sharp
increase of the background signal in the potential
region of peak P2/P20, and a further, but continu-
ous increase at higher potentials accompanied by
the specific adsorption of sulphate species.No sig-
nificant hysteresis is observed in 0.10 V6E60.80
V.The extension of these measurements towards
more negative potentials is hampered by the onset
of hydrogen evolution.Despite this limited poten-
tial range, the trends in the potential dependence
of the baseline intensity of ‘‘free’’ and Pd-modified
Au(111-20 nm) film electrodes in 0.1 H2SO4
strengthen the assumption introduced at the begin-
ning of this paragraph, that Epzc of the latter sys-
tem is located around 0.10–0.20 V. Additional
support for this hypothesis will be given in exper-
iments with variable Pd adlayer coverages
approaching bulk Pd properties [60].
p
p
3.4. CO-adsorption and oxidation on quasi single
crystalline gold film electrodes modified with 1
ML Pd in 0.1 M H2SO4
3.3.3. Non-resonant IR-background
A second interesting result is derived from the
IR-nonresonant background, which is represented
here by the signal intensity of the IR-baseline at
2000 cmꢀ1.In the absence of enhanced IR-bands,
e.g. without absorption from molecular vibrations,
the reflectivity changes of an electrode or its repre-
sentation in the absorbance scale as ꢀlog(I/Io) are
determined primarily by the optical constants
(refractive indices) of the substrate and of the ad-
sorbed species [33,35,61].Changes of the back-
ground reflectivity in the present experimental
approach represent predominantly the optical
properties of the metal electrode, which are
potential-dependent, since the refractive indices
of palladium and gold are significantly larger in
the mid-IR region (Pd: 3.56–17.3i at 2420 cmꢀ1
[62]; Au: 3.27–35.21i at 2000 cmꢀ1 [63]) than the
typical values for water (1.30–1.35 [64]).
3.4.1. Adsorption
We have chosen the adsorption and electro-oxi-
dation of a saturated CO adlayer to explore addi-
tional structural and reactivity properties of the
fabricated Pd-modified thin film electrodes in
acidic medium.For reference, we investigated first
the adsorption of CO on the bare gold surface.
Fig.7 illustrates the evolution of the SEIRA spec-
tra for Au(111-20 nm)/0.1 M H2SO4 upon dosing
with CO at 0.10 V. The reference spectrum was
chosen at t = 0.No significant changes were found
in the spectral ranges of tOH and dHOH modes of
interfacial water as well as below 1500 cmꢀ1.Only
one positive going band appears at 2109 cmꢀ1
FWHM = 24 2 cmꢀ1.Maximum intensity is
,
reached under our experimental conditions in CO
saturated electrolyte at tP600 s (Fig.7A ).Peak
position and FWHM do not change with time.
This band has been observed before [65–67] and
is assigned to the stretching mode of linearly
bound CO, tCOT(Au).The peak position is much
We demonstrated before, that the potential
dependent background reflectivity at 2000 cmꢀ1
for the bare Au(111-20 nm) film electrode in con-
tact with 0.1 M H2SO4 as well as for other typical
aqueous electrolytes exhibits a parabolic shape
with a clearly developed minimum around Epzc
(Fig.6C ) [47,60].Significant changes occur upon
higher than that observed on transition metals
ꢀ1
(cf.2080–2050 cm
on Pt, Pd, Rh) [68] and clo-
ser to that of free CO (2143 cmꢀ1) [57] suggesting

S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
119
Fig.7. Time dependent SEIRA spectra for the deposition of CO on Au(111-20 nm)/01. M H 2SO4 at 0.10 V during the continuous
purging of the electrolyte with a CO/Ar mixture.The reference spectrum was taken at t = 0 s.(A) shows the spectroscopic range from
1900 cmꢀ1 < t < 2300 cmꢀ1 while (B) illustrates, as examples, selected spectra in 1000 cmꢀ1 < t < 4000 cmꢀ1 during the CO/Ar
exposure and (top spectrum) after extended purging with argon only to remove the CO from the bulk of the electrolyte.
a weak adsorption of CO.Gold, which differs
from the transition metals by its electronic config-
uration of fully occupied d-orbitals (5d106s1) is in-
deed expected to exhibit a weaker co-ordination
ability for CO.Our interpretation is further sup-
ported by the following observations: (i) The
adsorption of CO on Au(111-20 nm) does not
change the intensity of the nonresonant IR-back-
ground signal.The latter is rather sensitive to
strong physisorption and chemisorption [69].(ii)
Careful purging with Ar for at least 30 min to
remove bulk solution CO is accompanied with
the quantitative desorption of interfacial CO
(Fig.7B ).
Subsequently we deposited CO onto Au(111-20
nm) modified with 1 ML Pd at 0.10 V, and carried
out a ‘‘spectroelectrochemical charge displace-
ment’’ experiment.At the potential chosen all
Pd-sites are covered with atomic hydrogen and
no anions are adsorbed. Fig.8 illustrates a series
of time-dependent SEIRA spectra, which was re-
corded simultaneously with the corresponding
electrochemical iꢀt transient (Fig.9A ).The
experimental conditions ensured the formation of
a saturated CO adlayer, which most probably
exhibits no long range order [70].The voltammet-
ric current is completely blocked in 0.10
V 6 E 6 0.75 V. The measured spectra are dis-
tinctly different from those recorded for the bare
Au(111-20 nm) film electrode in contact with 0.1
M H2SO4 (Fig.7 ).Characteristic positive going
bands evolve in the regions of OH- and CO-
stretching.No anion-related modes were found.
The CO vibrations are readily distinguished at
2106 cmꢀ1 (FWHM = 24 1 cmꢀ1), 2052 cmꢀ1
(FWHM = 35 2 cmꢀ1) and 1930 cmꢀ1 (FWHM
= 44 2 cmꢀ1).The time dependence of the corre-
sponding integral intensities, of the peak positions
and of the IR-background signal at 2300 cmꢀ1
during CO dosing and subsequent Ar purging
are plotted in Fig.9B–D .This analysis neglects
possible intensity transfer due to dipole–dipole
coupling within the CO adlayer [71].The inspec-
tion of Figs.8 and 9C reveals that the high fre-
quency mode disappears after ca.30 min upon
treatment of the electrolyte with bare Ar.This
temporary band is attributed to atop CO physi-
sorbed on the still remaining ‘‘free’’ gold sites (tCO-
T(Au), cf. Fig.7 ).A rough estimation, based on
the integrated CO intensities of Fig.9C , suggests
that in these experiments typically more than
90% of the gold electrode area are covered with
a monolayer of Pd.
Based on IRAS data of CO adsorption on
Pd(111), Pd–Au(111) or Pd–Pt(111) in UHV
[72–76] and under electrochemical conditions in

120
S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
to SEIRAS experiments measured with CO on
sputtered [37] or electrochemically deposited Pt
films [36] we did not observe bipolar bands.
The temporary evolution of the interfacial CO
vibrations at 0.10 V (Fig.9C and D ) reveals that
t
COT(Au) and tCOT(Pd) reach their maximum at
ꢁ400 s while tCOB(Pd) rises further.At low cover-
ages, CO adsorbs preferentially at step sites.A
steady state spectroscopic response is reached after
Ar purging of the electrolyte accompanied by the
complete removal of physisorbed CO from Au
sites.Simultaneously with these CO vibrations
one observes in the OH stretching region positive
going bands, one, tOH1 ꢂ 3642 4 cmꢀ1, is sharp
and the other, tOH2 ꢂ 3250 10 cmꢀ1, is rather
broad (Fig.8 ).The corresponding dHOH bending
vibrations are present, but too weak to be quanti-
tatively resolved.Nevertheless, since CO is hydro-
phobic and the sharp OH stretching band appears
at high CO coverages, these spectral features indi-
cate the coadsorption of CO with isolated interfa-
cial water and with second layer hydrogen bonded
water [47,56–58].The presence of stabilising inter-
facial water in high coverage CO adlayers was
already hypothesised for Pd(111)/CO from IRAS
[43,84] and UHV emersion experiments [85].Zhu
et al.concluded, based on high energy electron en-
ergy loss spectroscopy (HREELS) in combination
with temperature programmed desorption (TPD)
for the competitive adsorption of water and CO
on Pd(111) that (i) CO blocks water adsorption
sites and (ii) facilitates the formation of second
layer (secondary) hydrogen bonded water [86].
However, the data plotted in Figs.8 and 9 repre-
sent, to the best of our knowledge, the first real
time in situ spectroelectrochemical study of struc-
tural changes within the Helmholtz layer during
a so-called CO charge displacement experiment.
We also like to point out, that the time constant
of the spectroscopic response is significantly higher
than that of the simultaneously recorded electro-
chemical iꢀt transient.One important conse-
quence of these results comprises the critical
evaluation of the ‘‘classical’’ charge displacement
experiment, which was usually analysed according
to the following dominant mechanisms [87]:
Fig.8. Time dependent in situ SEIRA spectra for the deposi-
tion of CO on 1 ML Pd–Au(111-20 nm)/0.1 M H2SO4 at 0.10 V
during the continuous exposure of the electrolyte to a CO/Ar
mixture (black lines) and after extended purging with Ar only
(grey lines).A set of single scan spectra measured at t = 0 in the
absence of CO was chosen as reference.
solution [6,14,43,77] we assign the low frequency
mode to chemisorbed, bridge bonded CO on Pd,
which we will abbreviate in the following as
t
COB(Pd).The slight asymmetry of this band with
an extended tail region toward lower wavenum-
bers suggests contributions from edge and terrace
sites [14].The remaining third CO vibration at
ꢁ2052 cmꢀ1 is attributed to atop CO adsorbed
at step sites on the Pd surface and will be labelled
t
COT(Pd).This interpretation is based on a com-
parison of our experimental observations with
published IRAS and SERS studies employing size-
and shape-controlled Pd-nanoparticles [78–80], Pd
multilayers [81,82] or Au–Pd alloy films of defined
composition [83].Chemisorption of CO is clearly
represented by the increase of the nonresonant
IR-background intensity until saturation is
reached at t P 30 min (Fig.9B ).In comparison
Hads þ CO ! COads þ Hþ þ eꢀ
ð1Þ

S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
121
Fig.9. (A) Simultaneously with the spectral response ( Fig.8 ) monitored ‘‘charge displacement’’ transient for the deposition of CO on
1 ML Pd–Au(111-20 nm)/0.1 M H2SO4 at 0.10 V, and the time dependent evolution of the background intensity at 2300 cmꢀ1 (B), of
the integrated intensities (C) and of the band positions (D) of the observed CO stretching modes tCOT(Au), tCOT(Pd), tCOB(Pd).
dꢀ
þ CO þ ðn ꢀ dÞeꢀ ! COads þ Anꢀ
ð2Þ
ops with a maximum at E = 0.86 V. Further oxida-
A
ads
tion of the CO adlayer gives rise to a second peak
P2 at E = 1.02 V. Reversing the electrode potential
just after passing P1 causes incomplete electro-oxi-
dation of adsorbed CO, accompanied with the par-
tial reappearance of the typical features for the
hydrogen adsorption region I.However, if the
CO oxidation is extended to E > 1.02 V the follow-
ing potential scan and the characteristic hydrogen
states appear rather distorted.The data shown in
Fig.10A demonstrate that CO oxidation on 1
ML Pd–Au(111-20 nm) electrodes proceeds in
two distinct potential regions.The comparison of
our observations with recent data for CO-oxida-
tion on stepped Au(hkl) surfaces modified with
thin Pd films [29] suggests the assignment of the
potential region around P1 to CO adsorption on
defect (preferentially step) sites, while the more
positive feature P2 is attributed to the reaction at
terraces having a certain size.The latter process
This will be the topic of a forthcoming publication,
where the potential and coverage dependence of
simultaneously recorded spectroscopic and electr-
ochemical charge displacement transients will be
quantitatively analysed, and compared with the
corresponding single crystal data [60].
3.4.2. Oxidation
After the formation of a steady state saturated
CO adlayer at 0.10 V on 1 ML Pd–Au(111-20
nm) and complete removal of bulk solution CO
we employed a slow potential scan, 2.5 mVsꢀ1
,
to monitor in situ the potential dependence of
CO electro-oxidation (Fig.10 ).The current–
potential curve (Fig.10A ) reveals that the hydro-
gen adsorption is suppressed completely and the
currents remain close to zero until E ꢂ 0.75 V.
At higher potentials a rather sharp peak P1 devel-

122
S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
Fig.10. (A) shows the current–potential curve in 01. 0 V < E < 1.18 V (solid black line, 2.5 mVsꢀ1).The ‘‘distorted’’ negative-going
scan is shown as dotted line.The grey traces represent the bare supporting electrolyte without Pd and CO.(B) Evolution of the SEIRA
spectra for the electro-oxidation of a saturated CO adlayer deposited at 0.10 V on 1 ML Pd–Au(111-20 nm)/0.1 M H2SO4 during a
slow positive going potential scan and in the absence of bulk electrolyte CO.(C) and (D) illustrate the potential dependent evolution of
the CO modes.
is superimposed with the partial oxidative dissolu-
tion of Pd accompanied by an irreversible increase
of surface defects.We also like to point out that
these voltammetric data indicate that CO oxida-
tion on Pd–Au(111-20 nm) and Pd–Au(111) elec-
trodes in 0.1 M H2SO4 [29] proceeds with a lower
rate than on ideal Pt(111) [11,12] and stepped
Pt(hkl) [88–90] surfaces, despite the higher affinity
of Pd to oxygen-containing species [7].The oxy-
gen-containing species appears to be not easily
transferred to adsorbed CO.
0.1 M H2SO4 are shown in Fig.10B (black line-
s).The reference spectrum was chosen at 0.10 V
in the absence of CO.The grey traces represent
the response of the bare Au(111-20 nm) surface
in 0.1 M H2SO4.Advancing the electrode potential
towards more positive values reveals distinct
changes in four spectral regions representing CO-
and OH-stretching modes, HOH-bending and an-
ion-related vibrations at t < 1200 cmꢀ1 (Fig.10B ).
The intensities of linear and bridge bonded CO,
tCOT(Pd) and tCOB(Pd), start to decrease already
at potentials slightly larger than 0.10 V in the so-
called pre-oxidation region (cf.discussion for CO
The simultaneously recorded SEIRA spectra
for CO oxidation on 1 ML Pd–Au(111-20 nm)/

S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
123
electro-oxidation on Pt(111) in Ref. [91]).A major
and rather abrupt decrease occurs around the peak
P1 of the simultaneously measured current–poten-
tial curve, which was previously assigned to CO
oxidation on defect sites.No CO related signals
were detected at E > 1.06 V. We note that the con-
tribution of tCOT(Pd) vanishes at lower electrode
potentials than that of tCOB(Pd) (Fig.10C ).Both
vibration modes exhibit a rather similar tuning
rate dt/dE = 45 2 cmꢀ1 Vꢀ1, deviations exists
only around the oxidation peak P1.This value cor-
responds reasonably with literature data for CO
on Pd(111) (dtCOB/dE = 35 cmꢀ1 Vꢀ1 [43]) and
CO on Pd–Pt(111) (dtCOB/dE = 35 cmꢀ1 Vꢀ1 or
50 cmꢀ1 Vꢀ1 [14]), respectively.The blue shift is
attributed to the superposition of the electrochem-
ical Stark effect [92] with contributions due to
redistribution of charges within the CO molecule
[93] and dipole–dipole coupling [68,69].The linear
dependence of t on E also indicates that the struc-
tural changes within the adlayer at E < E(P1) do
not seem to modify the local environment of the
absorbed CO species dramatically! This interpreta-
tion is supported by the estimated values of the
mode, dHOH3, is found at 1614 2 cmꢀ1.These
water features correlate with the potential depend-
ent evolution of the CO-stretching modes,
t
COT(Pd) and tCOB(Pd), and indicate that both
water species are involved in the electro-oxidation
of atop and bridge bonded CO.Atop CO at defect
sites (tCOT(Pd)) appears to be consumed preferen-
tially.The third OH-vibration, tOH2, which was as-
signed to second layer hydrogen bonded water, is
rather unperturbed by these changes until, at
E P 0.55 V, sulphate species (tSO, Fig.6A ) start
to adsorb on sites of the Pd–Au(111-20 nm) sur-
face, where CO was already desorbed due to the
onset of electro-oxidation.The begin of the
(hydrogen-) sulphate adsorption is inhibited by
the presence of adsorbed CO, and appears at
approximately 0.30 V more positive with reference
to the CO-free system 1 ML Pd–Au(111-20 nm)/
0.1 M H2SO4 (Fig.6A ).The increasing surface ex-
cess of (hydrogen-) sulphate facilitates the co-
adsorption of hydrogen-bonded interfacial water,
which is represented by an intensity increase and
a slight blue shift of tOH2, especially pronounced
at E > 0.90 V, where the CO coverage drops
dramatically.
FWHM, 35 2 cmꢀ1 (tCOT(Pd)) and 45
3
cmꢀ1 (tCOB(Pd)), which do not change with poten-
tial until E approaches 1.00 V.
The potential dependent evolution of the simul-
taneously monitored vibrational modes of interfa-
cial CO and water species support the ‘‘reactant
pair’’ mechanism of electro-oxidation of adsorbed
CO in acidic medium, originally proposed by Gil-
man [94]:
The OH-stretching (tOH) and HOH-bending
(dHOH) modes of interfacial water are observed
at 3700–3000 cmꢀ1 and at 1650–1610 cmꢀ1
respectively.One may readily distinguish three
OH-stretching modes.However, their quantitative
,
H2O þ SÃ $ OHads ꢀ SÃ þ Hþ þ eꢀ
ð3Þ
ð4Þ
deconvolution is not straightforward at the present
stage of our knowledge.Therefore, we restrict the
following description and analysis to a qualitative
level choosing the CO-free spectrum at E = 0.10 V
as reference state (cf.Section 33. ).The intensity of
the high frequency vibration tOH1, previously al-
ready assigned to isolated water, decreases with
more positive potentials and vanishes at E < 0.90
V, e.g. just past the voltammetric peak P1 (Fig.
10A), indicating that these water species are desta-
bilized and/or being directly involved in the elec-
tro-oxidation of CO.Simultaneously, a negative
COads þ OHads ! CO2 þ Hþ þ eꢀ þ 2SÃ
where S* denotes a free surface site.
This mechanism assumes a Langmuir–Hinshel-
wood type reaction between CO and a surface
oxygen-containing species, adsorbed on adjacent
sites, to form CO2 (not detected in the SEIRA
spectrum due to the desorption into the bulk of
the electrolyte, which differs from IRAS results
in ‘‘thin layer configuration’’ [33,68]!).The poten-
tial- dependent evolution of the vibrational modes
attributed to interfacial water species suggests that
the oxygen-containing species could be co-ad-
sorbed isolated and/or weakly hydrogen-bonded
interfacial water.
going band appears around 3480 cmꢀ1, tOH3
,
which can be attributed to the continuous loss of
weakly hydrogen-bonded interfacial water
[54,56–58].The corresponding HOH-bending

124
S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
4. Summary and conclusions
on Pd- Au(111-20 nm), in comparison to the bare
gold surface.Analysing the potential dependence
of the non-resonant IR background signal (at
2000 cmꢀ1) we observed distinct differences be-
tween both electrodes, and suggested that the Epzc
of 1 ML Pd–Au(111-20 nm)/0.1 M H2SO4 is
located around 0.10–0.20 V (vs. RHE).
(1) In this study we have prepared and charac-
terized quasi single crystalline gold film electrodes
of 20 nm thickness and preferential (111) surface
orientation by electron beam evaporation onto a
Si hemisphere in combination with electrochemical
annealing.These samples were subsequently mod-
ified with ultrathin layers (sub-ML, ML, multi-L)
of Pd employing a potentiostatic deposition re-
gime.Similar to previous reports for Pd on
massive Au(111) single crystals, we could demon-
strate that the deposition of the first Pd ML onto
Au(111-20 nm) starts at underpotentials.We ob-
served well developed current peaks of hydrogen
adsorption/(hydrogen) sulphate desorption (and
of the respective anodic counter parts) on first
and multilayer Pd.Comparing the charges of the
characteristic hydrogen/anion ‘‘states’’, QHA, with
the deposited amount of Pd from sulphate- and/or
chloride-containing electrolyte, QPd, results in a
‘‘calibration master curve’’ with two distinct slopes
and an intersection point at QPd = 0.72 0.08
mCcmꢀ2 and QHA = 0.20 0.02 mCcmꢀ2.Based
on these data and taking into consideration the
scaling factor of 1.6, which represents the ratio be-
tween the electrochemically active and the geomet-
rical area of our template films, we could establish
conditions to fabricate defined and stable Pd
monolayers on Au(111-20 nm).
(2) In situ ATR-SEIRAS experiments with 1
ML Pd–Au(111-20 nm) electrodes demonstrated
for the first time that enhanced IR signals of physi-
sorbed and chemisorbed species can be obtained
on well defined and catalytically active Pt-group
model electrodes.Our approach is based on the
controlled modification of a SEIRAS-active, quasi
single crystalline gold template.We compared the
potential-dependent properties of interfacial water
and adsorbed (hydrogen) sulphate on Au(111-20
nm)/0.1 M H2SO4 in the absence and in the pres-
ence of 1 ML Pd.We found that weakly hydrogen
bonded water is replaced by strongly hydrogen
bonded interfacial water accompanied by an
increasing amount of adsorbed (hydrogen) sul-
phate if the potential advances to more positive
values.The onset of adsorption of the sulphate
species appears to be at more negative potentials
(31) The adsorption and electro-oxidation of
CO were chosen as test processes to probe struc-
tural and reactivity properties of the fabricated
Pd-modified thin film electrodes.Reference exper-
iments with Au(111-20 nm)/0.1 M H2SO4 revealed
that CO is weakly physisorbed from CO-saturated
electrolyte giving rise to a characteristic band, tCO-
T(Au), at 2109 cmꢀ1, which could be assigned to
the stretching mode of linearly bound CO.How-
ever, complete desorption was found in CO-free
electrolyte after prolonged purging with argon.
(32) Simultaneous spectroelectrochemical (time-
resolved in situ ATR-SEIRAS) and electro-
chemical ‘‘charge displacement’’ experiments were
carried out to monitor the deposition of a satu-
rated CO adlayer on 1 ML Pd–Au(111-20 nm)/
0.1 M H2SO4 at 0.10 V. The measured spectra
are distinctly different from those recorded for
the bare gold film.Characteristic positive going
bands evolve in the regions of OH- and CO-
stretching, which could be assigned to isolated
water (tOH1 ꢁ 3642 cmꢀ1) and second layer,
hydrogen bonded water (tOH2 ꢁ 3250 cmꢀ1) co-
adsorbed with chemisorbed CO on Pd at bridge
sites (tCOB(Pd)) and in atop positions at step edges
(tCOT(Pd)).The temporarily detected band tCO-
T(Au) (before removing CO from the bulk of the
electrolyte) allowed us to estimate the still remain-
ing ‘‘free gold sites’’ for 1 ML Pd–Au(111-20 nm)
to be less than 10% of the active electrode area.
The time dependent evolution of the i ꢀ t and of
the spectroscopic charge displacement transients
revealed that the structural changes of the adlayer
are taking place at a much longer time scale than
one might assume, just based on the electrochem-
ical data.These results suggest the need for a crit-
ical re-evaluation of the ‘‘classical’’ charge
displacement experiment, a topic that we intend
to address in a separate communication.
(33) We presented in this paper also the first
in situ ATR-SEIRAS experiments for the electro-

S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
125
oxidation of adsorbed CO on 1 ML Pd–Au(111-
References
20 nm)/0.1 M H2SO4.Our approach allowed
recording simultaneously the potential dependent
evolution of vibration modes of adsorbed CO
and anions as well as of interfacial water species.
We provide evidence that co-adsorbed, isolated
and weakly hydrogen bonded interfacial water
are disordered and eventually consumed during
the electro-oxidation of CO, and that atop CO
on defect sites appears to react first.The ‘‘CO-
free’’ surface spots will be subsequently covered
by (hydrogen-) sulphate which, in addition, facili-
tates the co-adsorption of hydrogen bonded water.
These spectroscopic observations are consistent
with the ‘‘reactant pair’’ mechanism of electro-oxi-
dation of CO in acidic medium, and provide new
insight into structural changes within the adlayer
during the course of the reaction.
(34) The ability to fabricate well defined,
SEIRAS-active Pd–Au(111-20 nm) model elec-
trodes for in situ spectroscopic and electrochemi-
cal studies demonstrates the unique chance and
opportunity to apply the advantages of in situ
ATR-SEIRAS in terms of sensitivity, time resolu-
tion and mass transport to (electro-) catalytic
applications with well defined Pt-group substrates
to develop detailed and novel structure–reactivity
correlations.The presented approach may also
provide new insight into the double layer proper-
ties of ‘‘reactive’’ materials in contact with aque-
ous electrolytes.Both topics will be addressed in
our future research program with major focus on
extending the present ‘‘demonstration studies’’
with Pd–Au(111-20 nm) to other Pt-type materials
employing spatial as well as time resolved ap-
proaches.
[1] D.M. Kolb, Surf. Sci. 500 (2000) 722.
[2] D.M. Kolb, in: R.C. Alkire, D.M. Kolb (Eds.), Advances
in Electrochem.Sci.& Eng,. vol.7, Wiley–VCH, Wein-
heim, 2000, p.107.
[3] R.R. Adzic, in: E. Gileadi, M. Urbakh (Eds.), Encyclope-
dia of Electrochemistry, vol.1, Wiley–VCH, Weinheim,
2002, p.561.
[4] A. Ruban, B. Hammer, P. Stoltze, H.L. Skiver, J.K.
Norskov, J.Mol.Catal.A 115 (1997) 421.
[5] G.Attard, A.Bannister, J.Electroanal.Chem.300 (1991)
467.
[6] J.Inukai, M.Ito, J.Electroanal.Chem.358 (1993) 307.
[7] MJ..Llorca, JM. .Feliu, A.Aldaz, J.Clavilier, J.Electro-
anal.Chem.351 (1993) 299.
[8] M.Baldauf, DM. .Kolb, J.Phys.Chem.100 (1996) 11375.
[9] R.Gomez, A.Rodes, M.Perez, JM. .Feliu, A.Aldaz,
Surf.Sci.327 (1995) 202; Surf.Sci.344 (1995) 85.
[10] N.M. Marcovic, C.A. Lucas, V. Climent, V. Stamencovic,
P.N. Ross, Surf. Sci. 465 (2000) 103.
[11] B.Alvarez, V.Climent, A.Rodes, J.M.Feliu, Phys.Chem.
Chem.Phys.3 (2001) 3269.
[12] B.Alvarez, V.Climent, A.Rodes, JM. .Feliu, J.Electro-
anal.Chem.497 (2001) 125.
[13] M.Arenz, V.Stamencovic, TJ..Schmidt, K.Wandelt,
P.N. Ross, N.M. Marcovic, Surf. Sci. 506 (2002) 287.
[14] B.Alvarez, A.Rodes, JM. .Perez, JM. .Feliu, J.Phys.
Chem.B 107 (2003) 2018.
[15] M.Baldauf, DM. .Kolb, Electrochim.Acta 38 (1993)
2145.
[16] H.Naohara, S.Ye, K.Uosaki, J.Phys.Chem.B 102
(1998) 4366.
[17] H.Naohara, S.Ye, K.Uosaki, J.Electroanal.Chem.473
(1999) 2.
[18] H.Naohara, S.Ye, K.Uosaki, Colloids Surf.A 154 (1999)
201.
[19] L.A. Kibler, M. Kleinert, R. Randler, D.M. Kolb, Surf.
Sci.443 (1999) 19.
[20] L.A. Kibler, M. Kleinert, D.M. Kolb, Surf. Sci. 461 (2000)
155.
[21] H.Naohara, S.Ye, K.Uosaki, Electrochim.Acta 45
(2000) 3305.
[22] M.Takahashi, Y.Hayashi, J.Mizuki, K.Tamura, T.
Kondo, H.Naohara, K.Uosaki, Surf.Sci.461 (2000) 213.
[23] H.Naohara, S.Ye, K.Uosaki, J.Electroanal.Chem.500
(2001) 435.
Acknowledgments
[24] M.E. Quayum, S. Ye, K. Uosaki, J. Electroanal. Chem.
520 (2002) 126.
S.P. is indebted to the Alexander von Humboldt
Foundation for a Research Fellowship.The
authors would like to thank Mr.Bierfeld and B.
Han for support in operating the evaporation
chamber.We also like to thank Juan Feliu and
Bob de Levie for many valuable comments.The
work was also supported by the Research Center
[25] L.Kibler, M.Kleinert, V.Lazarescu, DM. .Kolb, Surf.
Sci.498 (2002) 175.
[26] L.A. Kibler, A.M. El-Aziz, D.M. Kolb, J. Mol. Catal. A
199 (2003) 57.
[27] B.Alvarez, V.Climent, J.M.Feliu, A.Aldaz, Electrochem.
Commun.2 (2000) 427.
[28] A.M. El-Aziz, L.A. Kibler, D.M. Kolb, Electrochem.
Commun.4 (2002) 535.
Julich under contract E2910201.
¨

126
S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
[29] A.M. El-Aziz, L.A. Kibler, J. Electroanal. Chem. 534
(2002) 107.
[57] K.Nakamoto, Infrared and Raman Spectra of Inorganic
and Coordination Compounds, 4th ed., Wiley, New York,
1986.
[30] S.R. Brankovic, J.X. Wang, R.R. Adzic, Surf. Sci. 474
(2001) L173.
[58] P.A. Thiel, T.E. Madey, Surf. Sci. Rep. 7 (1987) 211.
[59] M.Wolf, S.Nettersheim, JM. .White, E.Hasselbrink, G.
Ertel, J.Chem.Phys.94 (1991) 4609.
[31] M.Mavrikakis, B.Hammer, JK. .Norskov, Phys.Rev.
Lett.81 (1998) 2819.
[32] A.Roudar, A.Groß, Phys.Rev.B 67 (2003) 033409.
[33] M.Osawa, Bull.Chem.Soc.Jpn.70 (1997) 2861.
[34] M.Osawain, S.Kawata (Eds)., Topics Appl.Phys.81
(2001) 185.
[60] S.Pronkin, I.Pobelov, Th.Wandlowski, in preparation.
[61] D.Diesing, S.Pronkin, K.Ataka, Th.Wandlowski, Phys.
Chem.Chem.Phys,. in preparation.
[62] J.H. Weaver, R.L. Bendow, Phys. Rev. B 12 (1975) 3509.
[63] G.P. Motulevich, A.A. Shikin, Sov. Phys. JETP 20 (1965)
560.
[35] W.Suetaka, Surface Infrared and Raman Spectroscopy––
Methods and Applications, Plenum Press, New York,
1995.
[64] U. Dummy, in: D.R. Linde (Ed.), CRC Handbook of
Chemistry and Physics, 71st Ed., CRC, Boca Raton, 1991.
[65] S.Dennis, S.Corrigan, P.Gao, L.W.Leung, M.J.Weaver,
Langmuir 2 (1986) 744.
[36] A.E. Bjerke, P.R. Griffiths, W. Theiss, Anal. Chem. 71
(1999) 1967.
[37] Y.Zhu, H.Uchida, M.Watanabe, Langmuir 15 (1999)
8757.
[66] S.C. Cheng, A. Hamelin, M.J. Weaver, J. Phys. Chem. 95
(1991) 5560.
[38] A.Miki, S.Ye, M.Osawa, Chem.Commun.(2002) 1500.
[39] M.Watanabe, Y.Zhu, H.Uchida, J.Phys.Chem.B 104
(2000) 1762.
[67] S.G. Sun, W.B. Cai, L.J. Wan, M. Osawa, J. Phys. Chem.
103 (1999) 2460.
[40] Y.X. Chen, A. Miki, S. Ye, H. Sakai, M. Osawa, J. Am.
Chem.Soc.125 (2003) 3680.
[68] T. Iwasita, F.C. Nart, in: H. Gerischer, C.W. Topias
(Eds.), Advances in Electrochem. Sci. and Engineering, vol.
4, VCH, Weinheim, 1995, p.123.
[41] Th.Wandlowski, K.Ataka, S.Pronkin, D.Diesing,
Electrochim.Acta 49 (2004) 1233.
[42] L. Kibler, M. Baldauf, D.M. Kolb, in: J. Garche (Ed.),
[69] B.N. Person, Phys. Rev. B 44 (1991) 3277.
[70] C.A. Lucas, N. Markovic, M. Ball, V. Stamenkovic, V.
Climent, P.N. Ross, Surf. Sci. 479 (2001) 241.
Electrochemie fur Energie und Umwelt, Universita¨tsverlag,
¨
Ulm, 1997, p.63.
[71] B.N.J. Persson, M. Tushaus, A.M. Bradshaw, J. Chem.
¨
Phys.92 (1990) 5034.
[43] S.Zhou, R.Gomez, MJ..Weaver, J.Electroanal.Chem.
474 (1999) 155.
[72] A.Sellidj, BE. .Koel, Phys.Rev.B 49 (1994) 8367.
[73] F.M. Hoffmann, Surf. Sci. Rep. 3 (1983) 107.
[74] M.Tu shaus, W.Berndt, H.Conrad, AM. .Bradshaw, B.
¨
[44] N.Hoshi, K.Kagaya, Y.Hori, J.Electroanal.Chem.485
(2000) 55;
N.Hoshi, M.Kuroda, Y.Hori, J.Electroanal.Chem.521
(2002) 155.
Person, Appl.Phys.A 51 (1990) 91.
[75] S. Surnev, M. Sock, M.G. Ramsey, F.P. Netzer, M.
Wiklund, M.Borg, JN. .Andersen, Surf.Sci.470 (2000)
171.
[45] S.Strbec, RR. .Adzic, A.Hamelin, J.Electroanal.Chem.
249 (1988) 291.
[46] D.Mayer, K.Ataka, Th.Wandlowski, J.Electroanal.
Chem., in preparation, 2004.
[76] A.K. Santraa, D.W. Goodman, Electrochim. Acta 47
(2002) 3595.
[47] S.Pronkin, Th.Wandlowski, J.Electroanal.Chem.550–
551 (2003) 131.
[77] M.Arenz, V.Stamenkovic, TJ..Schmidt, K.Wandelt,
P.N. Ross, N.M. Markovic, Surf. Sci. 506 (2002) 287.
[78] K.Walter, O.Seifferth, H.Kuhlenbeck, M.Ba ¨umer, H.J.
Freund, Surf.Sci.399 (1998) 190.
[48] S.Pronkin, Th.Wandlowski, unpublished, 2003.
[49] M.Baldauf, PhD. .Thesis, University of Ulm, 1996.
[50] LJ..Wan, T.Suzuki, K.Sashikata, J.Okada, J.Inukai,
K.Itaya, J.Electroanal.Chem.484 (2000) 189.
[51] B.E. Koel, A. Selldij, M.T. Paffett, Phys. Rev. B 46 (1992)
7846.
[79] J.B. Giorgi, T. Schroeder, M. Ba¨umer, H.J. Freund, Surf.
Sci.498 (2001) L71.
[80] Y.X. Jiang, S.G. Sun, N. Dung, Chem. Phys. Lett. 344
(2001) 463.
[52] A.S. Dakkouri, D.M. Kolb, in: A. Wieckowski (Ed.),
Interfacial Electrochemistry, Plenum Press, New York,
1999, p.151.
[81] G.Q. Lu, S.G. Sun, L.R. Cai, S.P. Chen, Z.W. Tian,
Langmuir 16 (2000) 778.
[82] S.Zou, MJ..Weaver, Anal.Chem.70 (1998) 2387.
[83] F.Maroun, F.Ozanam, OM. .Magnussen, RJ..Behm,
Science 243 (2001) 1811.
[53] V.Topolev, O.Petrii, Elektrochimiya 6 (1970) 852.
[54] K.Ataka, M.Osawa, Langmuir 14 (1998) 951.
[55] A.N. Frumkin, O.A. Petrii, B.B. Damaski, in: J. OÔM
Bockris, B.E. Conway, E. Yeager (Eds.), Comprehensive
Treatise of Electrochemistry, vol.1, Plenum Press, New
York, 1980, p.221.
[84] K.Yoshimi, MB. .Song, M.Ito, Surf.Sci.368 (1996) 389.
[85] R.L. Borup, D.E. Sauer, E.M. Stuve, Surf. Sci. 374 (1997)
142.
[86] X.Y. Zhu, J.M. White, M. Wolf, E. Hasselbrink, G. Ertl,
J.Phys.Chem.95 (1991) 8393.
[56] J.R. Scherrer, in: R.J.H. Clark, R.E. Hesters (Eds.),
Advances in Infrared and Raman Spectroscopy, vol.5,
Heyden, Philadelphia, 1978, p.151.
[87] J.Clavilier, R.Albalat, R.Gomez, JM. .Orts, JM. .Feliu,
J.Electroanal.Chem.360 (1993) 325.

S. Pronkin, Th. Wandlowski / Surface Science 573 (2004) 109–127
[88] N.P. Lebedeva, M.T.M. Koper, E. Herrero, J.M. Feliu, [91] W.Akemann, A.Friedrich, U.Stimming, J.Chem.Phys.
127
R.A. van Santen, J. Electroanal. Chem. 487 (2000) 37.
[89] N.P. Lebedeva, J.M. Feliu, M.T.M. Koper, R.A. Van
Santen, J.Electroanal, Chem.524–525 (2002) 242.
113 (2000) 6864.
[92] D.K. Lambert, J. Chem. Phys. 89 (1988) 3847.
[93] A.B. Anderson, J. Electroanal. Chem. 280 (1990) 37.
[94] S.Gilman, J.Phys.Chem.68 (1964) 70.
[90] N.P. Lebedeva, A. Rodes, J.M. Feliu, M.T.M. Koper,
R.A. van Santen, Phys. Chem. B 106 (2002) 9803.