Anion effects in the UPD of copper on Pd/Pt(111) bimetallic electrodes

Article abstract of DOI:10.1021/jp962840z

The underpotential deposition (UPD) of copper has been used to characterize the structure of palladium films supported on Pt(111). Depending upon the nature of the anions in solution (HSO₄^-, Cl^-, Br^-), a variety of adsorption behaviors are revealed. In pure sulfuric acid, palladium islands give rise to copper stripping peaks some 100-150 mV negative of the copper desorption states associated with Pt(111). As already deduced from hydrogen electrosorption measurements, palladium islands form at coverages of palladium <1 monolayer (mL), and long-range (111) order is preserved within regions of the surface free of palladium, even at coverages close to the completion of the first monolayer. The palladium films themselves are stable up to 0.6 V (Cu²⁺/ Cu), but desorption is facilitated at 0.8 V (Cu²⁺/Cu), leaving behind a somewhat disordered Pt/Pd phase. In contrast, copper UPD in the presence of strongly adsorbing anions (Cl^-, Br^-) leads to a much narrower range of stability [0.3 V (Cu²⁺/Cu)] for the palladium film although, from inspection of copper UPD data, island growth of palladium for θ_Pd < 1 mL is again implied with significant preservation of long-range Pt(111) order. It is reported that desorption of multilayer palladium films may be catalyzed by Cl^- and Br^- anions, resulting ultimately in a highly ordered palladium monolayer. Further oxidative desorption of the ordered palladium monolayer produces new copper adsorption states associated with place exchange between the halide and the palladium monolayer and also copper UPD onto the resulting Cu-Pd surface alloy. Repetitive potential cycles between 0.1 and 0.8 V (Cu²⁺/Cu) transform all palladium islands into a Cu-Pd surface alloy. Further oxidative desorption beyond this stage generates a disordered copper-rich Cu-Pd alloy phase and palladium-free regions of Pt(111). After complete desorption of palladium, the UPD voltammogram of Pt(111)-copper is fully restored with no evidence of surface perturbation having occurred. The relevance of these findings in the context of halide-catalyzed dissolution of palladium surfaces is discussed.

Full text of DOI:10.1021/jp962840z

J. Phys. Chem. B 1997, 101, 4597-4606
4597
Anion Effects in the UPD of Copper on Pd/Pt(111) Bimetallic Electrodes
Ali Al-Akl and Gary A. Attard*
Department of Chemistry, UniVersity of Wales, Cardiff, P.O. Box 912, Cardiff CF1 3TB, U.K.
ReceiVed: September 16, 1996; In Final Form: March 24, 1997X
The underpotential deposition (UPD) of copper has been used to characterize the structure of palladium films
supported on Pt(111). Depending upon the nature of the anions in solution (HSO4 , Cl-, Br-), a variety of
-
adsorption behaviors are revealed. In pure sulfuric acid, palladium islands give rise to copper stripping peaks
some 100-150 mV negative of the copper desorption states associated with Pt(111). As already deduced
from hydrogen electrosorption measurements, palladium islands form at coverages of palladium <1 monolayer
(mL), and long-range (111) order is preserved within regions of the surface free of palladium, even at coverages
close to the completion of the first monolayer. The palladium films themselves are stable up to 0.6 V (Cu2+/
Cu), but desorption is facilitated at 0.8 V (Cu2+/Cu), leaving behind a somewhat disordered Pt/Pd phase. In
contrast, copper UPD in the presence of strongly adsorbing anions (Cl-, Br-) leads to a much narrower range
of stability [0.3 V (Cu2+/Cu)] for the palladium film although, from inspection of copper UPD data, island
growth of palladium for θPd < 1 mL is again implied with significant preservation of long-range Pt(111)
order. It is reported that desorption of multilayer palladium films may be catalyzed by Cl- and Br- anions,
resulting ultimately in a highly ordered palladium monolayer. Further oxidative desorption of the ordered
palladium monolayer produces new copper adsorption states associated with place exchange between the
halide and the palladium monolayer and also copper UPD onto the resulting Cu-Pd surface alloy. Repetitive
potential cycles between 0.1 and 0.8 V (Cu2+/Cu) transform all palladium islands into a Cu-Pd surface
alloy. Further oxidative desorption beyond this stage generates a disordered copper-rich Cu-Pd alloy phase
and palladium-free regions of Pt(111). After complete desorption of palladium, the UPD voltammogram of
Pt(111)-copper is fully restored with no evidence of surface perturbation having occurred. The relevance of
these findings in the context of halide-catalyzed dissolution of palladium surfaces is discussed.
2
1
Introduction
proximately /3 mL of copper and /3 mL of sulfate.3b,11 An
overpotential for the second stage of adsorption is associated
with the place exchange of the sulfate anions within the
The underpotential deposition (UPD) of metals (the deposition
of foreign metal ions onto a metallic substrate at potentials
positive of the Nernst potential) is one of the most extensively
studied processes in surface electrochemistry.1 Recently, a
variety of in situ and ex situ spectroscopic,2 surface imaging,3
surface diffraction,4 radiochemical,5 and thermodynamic6 meth-
odologies have been utilized in order to gain insight into the
nature of this apparently “simple” redox reaction. It has been
found that coadsorption of cations and anions within the
interphasial region constituting the UPD layer is a general
phenomenon with the relative strength of interaction between
anion-substrate, cation-substrate, and anion-cation governing
the overall structure of the coadsorbate film. Questions as to
the exact location of the charge transferred to the interface as a
consequence of metal deposition and the precise concentration
of each surface component have also been addressed.6 In the
present study, the UPD of copper will be used as a structural
probe of well-defined palladium films on Pt(111) produced as
a consequence of “forced deposition”.7 Palladium adsorption
on platinum is an attractive system for study since, on the basis
of thermodynamic and lattice strain considerations, high-quality
films should form.8 Experimentally, this prediction has been
borne out in several studies.7-10
1
coadsorbed copper-sulfate layer by /3 mL of copper. In the
presence of Cl--containing acidic aqueous electrolytes, again a
two-stage copper adsorption process is observed on Pt(111),
but from recent RRDE, XRD, and STM data,3d,12,13 it is clear
that, in this case, the first stage of adsorption corresponds to
the formation of a copper-chloride bilayer, although residual
chloride adsorbed on Pt(111) still remains adjacent to copper-
chloride bilayer islands.14 Presumably, the second stage of
copper adsorption to complete the copper monolayer13 is
associated with place exchange with copper of these residual
chloride ions bonded to Pt(111). The deposition of copper in
the presence of Br- on Pt(111) follows a similar course to that
of Cl-.15 Copper UPD onto single crystalline palladium
electrodes has also been extensively investigated16 as has copper
UPD onto thick (>5 mL) palladium films supported on Pt(100)
and Au(111).17,18 Interestingly, the CV of copper UPD on
massive palladium films mimics exactly the response of well-
defined, bulk, single-crystal palladium electrodes.16b This is
not unexpected considering the high quality of palladium films
formed on Pt single crystals reported by the present authors
and others.7-10,18
Recent investigations have also concentrated on the iodine-
catalyzed dissolution of palladium single-crystal electrodes19 in
which ostensibly “disordered” surfaces have been transformed
into highly crystalline two-dimensional lattices. This finding
has implications for ongoing research into electropolishing
technologies. The ability to monitor electrochemically the
dissolution of well-defined thicknesses of palladium supported
The UPD of copper on Pt(111) proceeds via a two-stage
adsorption process.4a In pure sulfuric acid, the initial adsorption
of copper has been shown to generate a mixed, almost coplanar,
honeycomb structure of copper/sulfate corresponding to ap-
* Corresponding author.
X Abstract published in AdVance ACS Abstracts, May 1, 1997.
S1089-5647(96)02840-4 CCC: $14.00 © 1997 American Chemical Society

4598 J. Phys. Chem. B, Vol. 101, No. 23, 1997
Al-Akl and Attard
Figure 1. UPD of copper on Pt(111) in 10-3 M Cu(2a+q) + 0.1 M H2SO4. Sweep rate ) 10 mV/s. Insert: cyclic voltammogram of Pt(111) in 0.1 M
H2SO4. Sweep rate ) 50 mV/s.
on Pt(111) appeared to the present authors a suitable strategy
for further investigating this interesting phenomenon.
for a set period of time, a well-defined coverage of palladium
could be obtained:
Pd2+(aq) + H2(g) f Pd(adsorbed) + 2H+(aq)
(1)
Experimental Section
The experimental apparatus used in the present investigation
and glassware/solution preparation has been fully described
elsewhere.20 Briefly, the electrochemical hardware consisted
of a homemade potentiostat/computer interface/voltage ramp
generator connected to a three-electrode cell (Pd/H reference
electrode in pure sulfuric acid, copper wire reference electrode
for UPD, Pt counter electrode) containing a variety of O2-free
electrolytes (0.1 M H2SO4, 0.1 M H2SO4 + 10-3 M CuSO4,
0.1 M H2SO4 + 10-3 M CuSO4 + 10-3 M NaCl, 0.1 M H2SO4
+ 10-3 M CuSO4 + 10-3 M NaBr), depending upon the type
of experiment to be performed. All solutions were prepared
from ultrapure reagents (BDH Aristar grade) and Millipore 18.2
MΩ cm-1 water. The Pt(111) bead electrode, approximately 2
mm in diameter, was manufactured using the method developed
by Clavilier.21 After flame annealing and cooling in H2/H2O,
the electrode surface was protected by a droplet of ultrapure
water and transferred to the electrochemical cell where it could
be brought into electrical union with the test solution via a
meniscus contact under potential control. Palladium deposition
was carried out by placing a droplet of aqueous palladium nitrate
solution onto the surface of a freshly flamed Pt(111) electrode.
By passing hydrogen gas over the droplet of palladium nitrate
This so-called “forced deposition” of palladium has been fully
described previously.7
All voltammetric data involving copper-free electrolytes are
referenced to a saturated palladium-hydrogen electrode in
contact with the electrolyte. UPD measurements were refer-
enced to a copper wire (Goodfellow Metals Ltd.) also in contact
with the electrolyte, and 1 mL is defined as 1.5 × 1015 atoms/
cm2, the density of Pt atoms in the surface of Pt(111).
Results and Discussion
Copper UPD in 0.1 M H2SO4. As signified by the intensity
and sharpness of the “spike” at 0.42 V, the Pt(111) electrode
used in the present study appears to be well-ordered and
essentially free of surface defects (insert of Figure 1).22 Figure
1 also shows the same electrode in a copper UPD experiment
in order to assess the degree of surface perfection using an
alternative procedure to that of hydrogen adsorption and to
estimate the surface coverage of copper formed prior to bulk
deposition. The two adsorption states of copper depicted in
Figure 1 are in excellent agreement with previous findings for
this adsorption system.11 The total Coulometric charge deduced

Copper on Pd/Pt(111) Bimetallic Electrodes
J. Phys. Chem. B, Vol. 101, No. 23, 1997 4599
Figure 2. Insert: cyclic voltammogram of irreversibly adsorbed palladium on Pt(111) in 0.1 M H2SO4. Sweep rate ) 50 mV/s. θPd ) 0.33 mL.
2+
UPD of copper on the same electrode in 10-3 M Cu + 0.1 M H2SO4. Sweep rate ) 10 mV/s.
(aq)
from integration of the copper stripping peaks amounted to 480
µC cm-2, a value consistent with the formation of an epitaxial
monolayer of neutral copper adatoms on Pt(111). However,
that this charge is the net sum of several adsorption/desorption
processes has been recognized for some time.6 Making no
assumptions as to the nature of the distribution of charge within
the adlayer, underpotential deposition of 1 mL of copper in the
presence of bisulfate may be represented formally as
eq 2, has been ignored. What is actually measured upon
integration of the copper stripping peaks is (λ + (θ2 - θ1) +
(θ1δ - θ2γ) + x - y) electrons per adsorbed copper cation.
That this quantity should correspond precisely to 480 µC cm-2
(equivalent to exactly two electrons transferred to the interface
per surface platinum atom) should be seen as fortuitous since
it will be shown later that an even greater amount of charge is
generated from the stripping of copper from Pt(111) in the
presence of chloride anions. Furthermore, in a future publication
by Wieckowski and co-workers,24 it will be demonstrated using
AES that 1 mL of copper adatoms is formed in the UPD
potential range. Nonetheless, the stripping charge for the copper
monolayer was found by these workers to be a sensitive function
of bisulfate concentrationsthe more concentrated the solution,
the greater was found to be the stripping charge. Therefore,
the present voltammetric data appear to be consistent with the
presently accepted model of a two-stage adsorption process for
copper UPD on Pt(111)13,14 in which 1 mL of copper is formed
prior to bulk deposition.
Cu2+(aq) + [θ1HSO4(1-δ)-/Ptx] + (λ + (θ2 - θ1) +
(θ1δ - θ2γ) + x - y)e ) [θ2HSO4(1-γ)-/Cu(2-λ)+/Pty] (2)
where
0 e δ e 1, 0 < θ1 < 1, 0 e γ e 1, 0 < θ2 < 1,
0 e λ e 2
depending upon the precise distribution of charge within the
region comprising the adsorbed overlayer before and after
copper adsorption. x and y are the screening charges associated
with the substrate in response to electrosorption, and θ1 and θ2
are the fractional coverages of adsorbed anions before and after
copper adsorption, respectively. Even within this very general-
ized form, the possible contribution of coadsorbed hydroxonium
ions to the total charge,23 particularly on the left-hand side of
Irreversible adsorption7,25 of increasing submonolayer amounts
of palladium on Pt(111) and its subsequent electrochemical
characterization are shown in Figures 2 and 3. In 0.1 M H2-
SO4, the presence of adsorbed palladium manifests itself in two
ways. First, a new adsorption state is observed between 0.16
and 0.17 V (Pd/H) which is very narrow and intense. This state

4600 J. Phys. Chem. B, Vol. 101, No. 23, 1997
Al-Akl and Attard
Figure 3. Insert: cyclic voltammogram of irreversibly adsorbed palladium on Pt(111) in 0.1 M H2SO4. Sweep rate ) 50 mV/s. θPd ) 0.66 mL.
2+
UPD of copper on the same electrode in 10-3 M Cu + 0.1 M H2SO4. Sweep rate ) 10 mV/s.
(aq)
has been ascribed to coupled adsorption/desorption of anions
and hydrogen.7,8,25 It grows in intensity up to a maximum of
exactly 1 mL of palladium.7,8 The second consequence of
palladium growth on Pt(111) is the decrease in the intensity of
the “anomalous” region.7,8 The presence of this state for θPd
< 1 mL implies island growth of the palladium deposit in which
significant long-range order is preserved within Pt(111) domains
free of palladium.7,8 Previous AES/LEED/XPS investigations
of palladium growth on Pt(111) in UHV8,9 have also concluded
that epitaxial palladium islands form at submonolayer coverages
followed by Frank-Van der Merwe growth at higher coverages.
Factors favoring island growth would include negligible lattice
mismatch between palladium and platinum (atomic diameters
of 2.76 and 2.77 Å, respectively26), the greater surface energy
of platinum compared to that of palladium27 and attractive lateral
interactions exhibited by transition metal adatoms adsorbed on
close-packed transition metal surfaces.8
To estimate the coverage of palladium, the area under the
clean surface “anomalous” peak was utilized. Since, for island
growth of palladium, the charge associated with the anomalous
peak bears a simple relationship to the number of platinum sites
free of palladium (80 µC cm-2 ≡ 1 mL, 40 µC cm-2 ≡ 0.5
mL, etc.7), the coverage of palladium in the insert of Figure 2
is deduced to be 0.33 mL. Independent confirmation of this
estimate comes from Figure 2 itself, showing the corresponding
copper UPD CV from the same sample. The sharp, well-
resolved peaks of copper UPD on Pt(111) between 0.30 and
0.35 V (Cu2+/Cu) give rise to 327 µC cm-2 of charge. Since
480 µC cm-2 is required to strip 1 mL of copper from Pt(111),
it is estimated that the palladium coverage in this case is (1 -
327/480) ) 0.32 mL, in good agreement with the coverage
deduced in aqueous sulfuric acid. Since long-range Pt(111)
order appears to be a prerequisite to observe the two copper
stripping peaks,22 island growth of palladium is strongly
suggested, in agreement with findings obtained in copper-free
electrolytes.7,8 In a similar fashion, taking into account copper
stripping and “anomalous” region charges leads to a palladium
coverage of 0.66 mL for Figure 3.
The presence of palladium in Figure 2 is signified by the
broad multiplet just negative of the clean surface peaks between
0.1 and 0.3 V (0.2-0.32 V for copper stripping and 0.1-0.20
V for copper adsorption). It is seen that these multiplet peaks
increase in magnitude as the palladium coverage is increased
in parallel with a decrease in the size of the copper peaks derived
from the clean surface. Charge integration over the entire
potential range of UPD, including both platinum and palladium
sites, always produced a total charge of 480 µC cm-2. This
implies that, by taking the charge under the multiplet peaks, it
is also possible to estimate the palladium coverage. For
example, in Figure 3, the charge derived from stripping of

Copper on Pd/Pt(111) Bimetallic Electrodes
J. Phys. Chem. B, Vol. 101, No. 23, 1997 4601
Figure 4. UPD of copper on bulk palladium film irreversibly adsorbed
on Pt(111) in 10-3 M Cu(2a+q) + 0.1 M H2SO4. Sweep rate ) 10 mV/s.
copper from palladium is 313 µC cm-2. Therefore, if 480 µC
cm-2 is the charge produced by desorption of copper from an
epitaxial monolayer of palladium, the coverage of palladium is
313/480 ) 0.65 mL, in good agreement with estimates made
from hydrogen and copper adsorption on domains of clean
platinum. The upper potential limit of stability of palladium
films in the presence of aqueous copper sulfate was found to
be 0.6 V (Cu2+/Cu). Potential excursions to 0.8 V and above
always led to stripping of palladium, and in the case of bulk
films, significant roughening of the surface was observed as
signified by peak broadening and increased irreversibility. It
is necessary, at this juncture, to point out also that the “multiplet”
state only appeared at submonolayer coverages of palladium
and that, for coverages of palladium in excess of approximately
1 mL, only one peak was observed. This is exemplified in
Figure 4 in which copper UPD onto palladium multilayers on
Pt(111) gave rise to a single UPD peak. It should be noted
that the copper stripping charge in Figure 4 amounts to only
452 µC cm-2, again emphasizing the “conditionality” of the
stripping charge (see eq 1) and also the significant differences
between the adsorption characteristics of bulk palladium films
and their monolayer counterparts.
2+
Figure 5. (a) UPD of copper on Pt(111) in 10-3 M Cu + 0.1 M
(aq)
H2SO4 + 10-3 M NaCl. Sweep rate ) 10 mV/s. (b) UPD of copper on
bulk palladium film irreversibly adsorbed on Pt(111) in 10-3
M
2+
(aq)
Cu + 0.1 M H2SO4 + 10-3 M NaCl. Sweep rate ) 10 mV/s. Insert:
Interestingly, the stripping peak in Figure 4 is very similar
to that reported for UPD of copper onto massive Pd(111)
electrodes16b and also multilayers of palladium on Au(111).17
For multilayers of palladium adsorbed on Pt(100), it has also
been shown that an almost identical voltammetric response to
electrosorption of copper onto bulk Pd(100) is obtained.18 Thus,
we conclude that high-quality Pd(111) films may be produced
by irreversible adsorption of palladium on Pt(111) as also
deduced from LEED.8,9 Speculations as to the origin of the
splitting of the copper UPD peak on submonolayer palladium
sites await further work using STM and stepped single crystals.29
corresponding CV of hydrogen electrosorption.
(compare with Figure 1). The total copper stripping charge is
515 µC cm-2 with approximately twice as much charge
appearing in the more positive stripping peak compared to the
peak at more negative potentials. The result of dosing sub-
monolayer amounts of palladium (0.27 mL) onto the Pt(111)
substrate is shown in Figure 6a. A new peak is observed at
0.2 V associated with copper UPD onto islands of palladium
and a second peak corresponding to the more negative of the
two copper adsorption states characteristic of Pt(111). The need
to limit the upper potential range to 0.3 V will be explained
subsequently. Increasing the coverage of palladium up to 1 mL
led to the attenuation of all voltammetric states pertaining to
the Pt(111) substrate and a maximum in the intensity of the
peak at 0.2 V. Thus, island growth of palladium is once again
indicated. In contrast to copper UPD in pure sulfuric acid,
however, palladium submonolayer islands may be assigned
unambiguously to a single, sharp peak at =0.2 V. Also, in
Copper UPD in 0.1 M H2SO4 in the Presence of 10-3
M
Chloride. In Figure 5a is shown the CV resulting from copper
UPD on Pt(111) in the presence of chloride. Agreement
between the present study and earlier work is good.4a,b,11,13
A
two-step adsorption process for copper is once again signified
by the presence of two adsorption states as found for copper
UPD in pure sulfuric acid although, for chloride, the potential
separation (130 mV) between each state is somewhat greater

4602 J. Phys. Chem. B, Vol. 101, No. 23, 1997
Al-Akl and Attard
maximum intensity, the charge under the palladium monolayer
peak corresponded to 428 µC cm-2. In Figure 5b is shown a
voltammogram of the UPD of copper onto a thick palladium
film (>4 mL) in the presence of chloride anions. The measured
copper stripping charge in this case amounted to 420 µC cm-2
.
Again, a remarkable similarity is observed between Figure 5b
and previous data on palladium single crystals and (111)
orientated supported films16b whereby, in contrast to (111)
surfaces of platinum and gold, a single copper adsorption state
is observed. Interestingly, electrode surfaces such as Rh(111),
Pd(111), and Ir(111) which display the strongest interactions
with specifically adsorbed anions30 also tend to produce a single
copper UPD state.24 Iodine on Pt(111) is also known to produce
a single copper UPD peak,31 reflecting once again that strong
anion-substrate interactions appear to preclude the possibility
of a two-stage copper adsorption process. When further
oxidative stripping of palladium from the well-ordered mono-
layer was carried out at potential excursions up to 0.6 V (Cu2+
Cu), a remarkable change in the voltammogram was realized.
/
It should be noted that these changes only occurred for θPd
e
1 mL at potentials in excess of 0.3 V (Cu2+/Cu). Figure 7a
shows the effect of repetitive potential cycling of a palladium
monolayer between 0.1 and 0.6 V (Cu2+/Cu). The adsorption
state associated with the palladium monolayer is continuously
transformed into two new states at =0.4 and 0.15 V. This
2+
(aq)
behavior was not observed in Cu /H2SO4(aq) solutions. A
model in accordance with the present data would be as follows.
At the edges of palladium monolayer islands, place exchange
of strongly adsorbed chloride with palladium takes place. This
would be consistent with the fact that chloride catalyzes the
oxidative desorption of palladium relative to bisulfate (anion-
induced weakening of Pt-Pd bonds), presumably via some
complexation between Cl- and adsorbed palladium. Since the
new state at 0.4 V is slightly negative of the more intense of
the copper UPD states found for copper adsorption on the well-
ordered Pt(111) surface, it would also be consistent with copper
adsorption onto platinum sites. It should be remembered that
the adsorption of copper onto palladium sites takes place some
200 mV more negative in potential. If the present assumptions
holds and, following repetitive cycling, palladium is stripped
off the surface to be replaced with strongly adsorbed chloride
ions at the palladium island rim, the state at 0.4 V is simply
copper UPD onto Pt sites originally occupied by place-
exchanged chloride ions. However, the electronic structure of
the Pd-Cu coplanar layer formed after place exchange of
chloride with copper will obviously be different from that of
the center of the palladium island which is still dominated by
Pd-Pd interactions. Therefore, one would predict that copper
UPD onto Pd-Cu sites would occur at a more negatiVe potential
than copper UPD on Pd-Pd regions. This is because (as has
been noted by several workers32,33) as a surface alloy becomes
enriched with one component relative to another, UPD peaks
tend to shift in potential away from the value associated with
the initially more concentrated component to that of the majority
component. Thus, since copper UPD on Cu-Pd alloys cannot
take place at more positive potentials than on pure palladium
or more negative potentials than on pure copper, one expects
the potential of the copper UPD peak to lie between these two
extremes, depending upon the precise surface composition of
the alloy. Therefore, the new peak appearing at =0.15 V as a
consequence of oxidative stripping of palladium in chloride
would be wholly consistent with copper UPD onto a Pd-Cu
alloy formed originally at potentials close to 0.4 V. The half-
width of the 0.15 V peak is comparable with that of the 0.2 V
UPD copper peak, indicating that the Cu-Pd alloy is well-
defined. Progressive nucleation and growth of the place-
Figure 6. (a) Copper UPD on 0.27 mL of palladium irreversibly
2+
adsorbed on Pt(111) in 10-3 M Cu
+ 0.1 M H2SO4 + 10-3
M
(aq)
NaCl. Sweep rate ) 10 mV/s. (b) Oxidative stripping of palladium
irreversibly adsorbed on Pt(111) in 10-3 M Cu(2a+q) + 0.1 M H2SO4 +
10-3 M NaCl. The gradual decrease in the magnitude of the copper
adsorption state corresponding to second layer palladium and the
simultaneous growth in the intensity of the first layer palladium peak
are evident. Sweep rate ) 10 mV/s.
contrast to findings in pure sulfuric acid, Figure 6b shows that,
for coverages of palladium in excess of 1 mL, a second
adsorption state of copper develops at more positive potentials,
which may be attributed to copper UPD onto second layer
palladium, as also reported in the case of hydrogen electrosorp-
tion.7,8 Oxidative desorption of palladium between 0.45 and
0.5 V gives rise to preferential desorption of second layer
palladium (decrease in intensity of peak at 0.23 V) and an
increase in the intensity of the peak at 0.2 V as more and more
of the underlying palladium monolayer was exposed. At its

Copper on Pd/Pt(111) Bimetallic Electrodes
J. Phys. Chem. B, Vol. 101, No. 23, 1997 4603
exchanged chloride sites toward the center of the palladium
monolayer island would account for the quantitative decrease
in the intensity of the 0.2 V peak and the correlated growth of
the states at 0.15 and 0.4 V. A schematic of the processes
thought to be occurring is presented in Figure 8. It should be
noted also that the total stripping charge from all three states is
fairly constant, indicating that the total number of sites for
copper UPD remains the samesit is only the relative surface
concentration of each type of copper adsorption site that is
changing [Cu2+ on Cl-Pt (0.4 V), Cu on Pd-Pd (0.2 V), and
Cu2+ on Cu-Pd (0.15 V)]. Figure 7b shows the voltammogram
of the palladium layer after extensive potential cycling in order
to generate the maximum surface concentration of the Cu-Pd
alloy. Although there is still a small amount of the 0.2 V
palladium monolayer peak visible, it is clear that the majority
of the surface has been converted to the Cu-Pd alloy phase.
Integration of the stripping peaks at 0.15 and 0.4 V leads to
values of 311 and 118 µC cm-2, respectively. Since, according
to our model, this should reflect the number of palladium (0.15
V) and copper sites (0.4 V) within the alloy (ignoring differences
due to anions as discussed in eq 2), it is possible to obtain an
estimate of the stoichiometry of the alloy phase.
Cu:Pd ) 118:311 ) 0.38 = 1:3
Thus, if one asserts that, because of the narrowness of the two
peaks at 0.15 and 0.4 V, the Cu-Pd alloy being generated as a
result of potential cycling is uniform, its composition according
to charge considerations is estimated to be Pd3Cu. Further
oxidative stripping of the Pd-Cu alloy beyond the point reached
in Figure 7b is shown in Figure 7c (potential excursion to 0.9
V). It is evident from Figure 7c that the Pd3Cu alloy phase is
gradually changed into a more copper-rich alloy phase as
oxidative stripping of palladium is continued. This is signified
by the attenuation of the peak at 0.15 V and the appearance of
a new state at 0.04 V. In parallel with these changes is the
growth in the Pt(111) copper adsorption peaks at =0.4 V.
Continued potential cycling was found to eliminate entirely the
0.04 V state with subsequent restoration of the Pt(111) volta-
mmogram features shown already in Figure 5a. This implies
that no structural transformations have occurred to the Pt(111)
substrate as a consequence of complete palladium desorption.
Furthermore, so long as the positive potential limit was
maintained at 0.5 V (Cu2+/Cu), the Pd3Cu alloy phase remained
stable, as signified by the constancy of the voltammetric
response of the electrode shown in Figure 7b. Finally, within
the constraints of a 0.5 V (Cu2+/Cu) potential limit, sweep rate
was found to have no discernible effect on the stability of the
alloy phase.
Copper UPD in 0.1 M H2SO4 in the Presence of 10-3
M
Bromide. Having already investigated the effect of chloride
on the dissolution of palladium, it was thought prudent to
examine whether or not bromide ions also had a catalytic
influence over the desorption kinetics. Previous studies4b,11,18
have concluded that a copper-bromide bilayer is formed during
the initial copper adsorption step on clean Pt(111). The second
peak at more negative potentials may, as for chloride, be related
to place exchange with copper of residual bromide ions bonded
to the Pt(111) substrate. The total copper stripping charge for
copper in the presence of bromide (not shown) was found to
be 490 µC cm-2, slightly less than the stripping charge in
chloride but greater than the charge in pure sulfuric acid. Figure
9 shows the oxidative stripping of a palladium film (θPd > 1)
in the bromide-containing copper electrolyte. Although peak
resolution is not as good as for chloride, it is still possible to
discern the preferential stripping of second layer palladium
(UPD peak at 0.17 V, negative-going sweep) and the subsequent
Figure 7. (a) Effect of potential cycles between 0.1 and 0.6 V (Cu2+
/
2+
Cu) in 10-3 M Cu
+ 0.1 M H2SO4 + 10-3 M NaCl on an
(aq)
irreversibly adsorbed palladium monolayer on Pt(111). Sweep rate )
10 mV/s. (b) Copper UPD on maximum surface coverage of copper-
palladium surface alloy obtained from treatment outlined in (a). Sweep
rate ) 10 mV/s. (c) Changes in the voltammetric profile of (b) after
repetitive cycles to 0.9 V, showing progressive desorption of palladium
from Cu/Pd alloy phase and recovery of adsorption properties of Pt-
(111). Sweep rate ) 10 mV/s.

4604 J. Phys. Chem. B, Vol. 101, No. 23, 1997
Al-Akl and Attard
Figure 8. Schematic of oxidative stripping of palladium in the presence of halide ions.
to 0.3 V, but significant oxidative currents associated with
palladium desorption could be generated at potentials >0.3 V.
This behavior is identical with that found for chloride-catalyzed
desorption whereby preferential desorption of palladium mul-
tilayers takes place prior to any stripping of the palladium
monolayer. The charge under the copper stripping peak from
a monolayer of palladium derived from stripping of second and
third layer palladium was estimated to be 466 µC cm-2. The
resemblance between the behavior of bromide and chloride
becomes even more striking when oxidative desorption of the
palladium monolayer is facilitated. The insert of Figure 9 shows
the effect of repetitive potential cycling up to 0.75 V at 50 mV/s
on the palladium monolayer. As with chloride, new copper
adsorption states are generated close to the potential of the UPD
peak on Pt(111) and simultaneously at potentials negative of
the copper UPD state associated with the palladium monolayer.
It is proposed that the same microscopic model outlined for
chloride applies in this case. That is, place exchange of bromide
ions with simultaneous desorption of a palladium bromide
complex occurs at the edges of palladium monolayer islands,
leading subsequently to the formation of a copper desorption
state at ∼0.44 V. Copper UPD onto the Cu-Pd alloy formed
as a result of bromide-copper place exchange (peak at 0.32
V) manifests itself as the adsorption state at potentials negative
of the UPD copper peak associated with the palladium mono-
layer. Eventually, after continued potential cycling, recovery
of the UPD response of the palladium-free Pt(111) substrate
was obtained. Hence, the behavior of bromide ions in control-
ling the oxidative desorption of palladium films on Pt(111) is
demonstrated to be almost identical with that of chloride.
Figure 9. Oxidative stripping of palladium irreversibly adsorbed on
2+
Pt(111) in 10-3 M Cu
+ 0.1 M H2SO4 + 10-3 M NaBr. The
(aq)
Further Discussion
gradual decrease in the magnitude of the copper adsorption state
corresponding to second layer palladium and the simultaneous growth
in the first layer palladium state are evident. Sweep rate ) 10 mV/s.
Insert: Effect of oxidative stripping of a monolayer of palladium. Sweep
rate ) 50 mV/s.
The adsorption of palladium on Pt(111) leads to the formation
of well-defined epitaxial films.7-10 In 0.1 M H2SO4, copper
UPD may be used to determine the palladium coverage for θPd
e 1 mL. However, oxidative desorption of palladium multi-
layers in copper-containing electrolytes in the absence of halide
growth in the intensity of the first layer palladium peak (0.16
V, negative going sweep). The palladium film was stable up

Copper on Pd/Pt(111) Bimetallic Electrodes
J. Phys. Chem. B, Vol. 101, No. 23, 1997 4605
ions always led to a disordered Pt/Pd phase. This behavior
upon the type of anion present in the electrolyte. Whereas in
the absence of halide ions, the palladium deposit was stable up
to 0.6 V (Cu2+/Cu), in the presence of chloride and bromide,
oxidative desorption could be observed at potentials greater than
0.3 V (Cu2+/Cu). The morphology of the palladium film after
oxidative stripping was also dependent upon the extent of oxide
formation which could be minimized by specifically adsorbed
halide ions. As reported by Soriaga and co-workers36 in the
case of iodide monolayers, the halide-induced dissolution of
palladium occurs in a layer-by-layer fashion with selective
corrosion at step sites. In addition, it is asserted that the initial
step in the dissolution process is place exchange of the halide
anion at defect sites contained within the palladium film.
contrasts strongly with that found in halide solutions.
A
reasonable explanation of this difference would be to consider
the extent to which surface oxides can form on the palladium
deposit and the ability of the palladium to form complexes with
the anions in the electrolyte phase. It is well-known that halides
possess the ability to inhibit the onset of oxide formation on
transition metal electrodes.34 One of the striking consequences
of electrochemical oxide formation on electrode surfaces is
reported to be a significant perturbation of the surface crystal-
linity.35 Thus, if oxide formation is precluded, surface roughen-
ing by this mechanism should not happen. This would be
consistent with the present halide results whereby surface
roughening via oxide formation and adsorption is prevented due
to the strong interaction of the halides with the electrode surface.
However, in the presence of the more weakly adsorbed bisulfate
anion, surface oxide formation is more likely to take place in
competition with anion adsorption with a corresponding increase
in the surface disorder upon desorbing the oxide. The ability
of strongly adsorbed anions to weaken the bonding between a
metal atom and a metallic substrate (for example, during
“electrochemical annealing”7,25) may also play in part in
facilitating the dissolution of palladium. Having discussed
possible reasons as to why chloride and bromide catalyze the
dissolution of palladium without encouraging the growth of
surface defects, it remains to explain why second layer palladium
should be removed exclusively via oxidative desorption, even
though palladium monolayer sites are available (see Figure 6b).
If the proposed model of dissolution occurring from island
boundaries into the island center is correct, it means that step
sites (at the junction between palladium first and second layers)
are highly susceptible to dissolution. This phenomenon could
be attributed to the lower PZTC exhibited by step sites versus
terraces.30 That is, at a given potential, the excess positive
charge at the step site would always be greater than at a terrace
site, and therefore, a greater interaction with adsorbed anions
would be exhibited. To obtain a similar degree of interaction
at a terrace site, the electrode potential would need to be
increased toward more positive potentials. In fact, it was found
during the stripping experiment in chloride that, if a bulk
palladium layer was desorbed by cycling between 0 and 0.5 V
(Cu2+/Cu) until a “perfect” palladium monolayer was evolved,
to cause further place exchange of the chloride with the
palladium monolayer, the overpotential had to be increased to
at least 0.6 V (Cu2+/Cu), consistent with the absence of step
defect sites in such an electrochemically formed palladium
monolayer. Thus, from a fundamental viewpoint of excess
charge at local sites, the rate of palladium dissolution at steps
must always be greater than at a terrace, giving rise to the
behavior reported above. Strong evidence in support of our
model has recently been published in which in situ STM was
used to monitor the iodine-catalyzed dissolution of palladium
single crystals. Both layer-by-layer dissolution and selective
corrosion at steps were observed.36 One further advantage of
using a thin palladium layer to observe halide-catalyzed
palladium desorption is that it demonstrates unambiguously that
the initial step in dissolution must involve a place exchange
with the halide anion. This finding is revealed only as a
consequence of the fact that a halide anion bonded to platinum
exhibits an adsorption/desorption state some 200 mV removed
from an equivalent palladium site.
Acknowledgment. The financial support of the EPSRC
(equipment grant to G.A.A.) and the Saudi Arabian Government
(studentship for A.A.A.) is gratefully acknowledged.
References and Notes
(1) (a) Kolb, D. M. In AdVances in Electrochemistry and Electro-
chemica Engineering; Gerisher, H., Tobias, C. J., Eds.; Wiley and Sons:
New York, 1978; Vol. 11. (b) Adzic, R. In AdVances in Electrochemistry
and Electrochemical Engineering; Gerisher, H., Tobias, C., Eds.; Wiley
and Sons: New York, 1985; Vol. 13. (c) Lorenz, W. J.; Hermann, H. D.;
Wuthrich, N.; Schmidt, E. J. Electrochem. Soc. 1974, 121, 1167. (d) Breiter,
M.; Knorr, C.; Volkl, W. J. Electrochem. 1955, 59, 681. (e) Tindall, G.
W.; Bruckenstein, S. Anal. Chem. 1968, 40, 1637.
(2) (a) Blum, L.; Abruna, H. D.; White, J.; Gordon II, J. G.; Borges,
G.; Samant, M. G.; Melroy, O. R. J. Chem. Phys. 1986, 85, 6732. (b)
Durand, R.; Faure, R.; Aberdam, D.; Salem, C.; Tourillon, G.; Guay, D.;
Ladouceur, M. Electrochim. Acta 1992, 37, 1977. (c) Tadjeddine, A.; Guay,
D.; Ladouceur, M.; Tourillon, G. Phys. ReV. Lett. 1991, 66, 2335.
(3) (a) Magnussen, O. M.; Hotlos, J.; Nichols, R. J.; Kolb, D. M.; Behm,
R. J. Phys. ReV. Lett. 1990, 63, 2929. (b) Sashikata, K.; Furuya, N.; Itaya,
K. J. Electroanal. Chem. 1991, 315, 275. (c) Mo¨ller, F.; Magnussen, O.
M.; Behm, R. J. Electrochim. Acta 1995, 40, 1259. (d) Matsumoto, H.;
Oda, I.; Inukai, J.; Ito, M. J. Electroanal. Chem. 1993, 356, 275.
(4) (a) Markovic, N.; Ross, P. N. Langmuir 1993, 9, 580. (b) Michaelis,
R.; Zei, M. S.; Zhai, R. S.; Kolb, D. M. J. Electroanal. Chem. 1992, 339,
299. (c) Hubbard, A. T. Chem. ReV. 1988, 88, 633. (d) Markovic, N. M.;
Gasteiger, H. A.; Lucas, C. A.; Tidswell, I. M.; Ross, P. N. Surf. Sci. 1995,
335, 91.
(5) (a) Horanyi, G.; Rizmayer, E. M.; Too, P. J. Electroanal. Chem.
1983, 152, 211. (b) Zelenay, P.; Rice-Jackson, L. M.; Gawlowski, J.;
Wieckowski, A. Surf. Sci. 1991, 256, 253. (c) Varga, K.; Zelenay, P.;
Wieckowski, A. J. Electroanal. Chem. 1992, 330, 453.
(6) (a) Shi, Z.; Lipkowski, J. J. Electroanal. Chem. 1994, 365, 303.
(b) Shi, Z.; Lipkowski, J.; Gamboa, M.; Zelenay, P.; Wieckowski, A. J.
Electroanal Chem. 1994, 366, 317. (c) Zhichao, S.; Lipkowski, J. J. Phys.
Chem. 1995, 99, 4170.
(7) (a) Clavilier, J.; Llorca, M. J.; Feliu, J. M.; Aldaz, A. J. Electroanal.
Chem. 1991, 310, 429. (b) Clavilier, J.; Llorca, M. J.; Feliu, J. M.; Aldaz,
A. J. Electroanal. Chem. 1993, 351, 299.
(8) Attard, G. A.; Price, R.; Al-Akl, A. Electrochim. Acta 1994, 39,
1525.
(9) Han, M.; Mrozek, P.; Wieckowski, A. Phys. ReV. B 1993, 48, 8329.
(10) Inukai, J.; Ito, M. J. Electroanal. Chem. 1993, 358, 307.
(11) (a) Abe, T.; Swain, G. M.; Sashikata, K.; Itaya, K. J. Electroanal.
Chem. 1995, 382, 73. (b) Shingaya, Y.; Matsumoto, H.; Ogasawara, H.;
Ito, M. Surf. Sci. 1995, 335, 23.
(12) Tidswell, I. M.; Lucas, C. A.; Markovic, N. M.; Ross, P. N. Phys.
ReV. B 1995, 51, 10205.
(13) Markovic, N. M.; Gasteiger, H. A.; Ross, P. N. Langmuir 1995,
11, 4098.
(14) Gomez, R.; Feliu, J. M.; Abruna, H. D. J. Phys. Chem. 1994, 98,
5514.
(15) Matsumoto, H.; Inukai, J.; Ito, M. J. Electroanal. Chem. 1994, 379,
223.
(16) (a) Chierchie, T.; Mayer, C. Electrochim. Acta 1988, 33, 341. (b)
Rigano, P. M.; Mayer C.; Chierchie, T. Electrochim. Acta 1990, 35, 1189.
(c) Solomun, T. J. Electroanal. Chem. 1989, 261, 229. (d) Lenz, P.;
Solomun, T. J. Electroanal. Chem. 1993, 353, 131.
(17) Baldauf, M.; Kolb, D. M. Electrochim. Acta 1993, 38, 2145.
(18) Gomez, R.; Rodes, A.; Perez, J. M.; Feliu, J. M.; Aldaz, A. Surf.
Sci. 1995, 344, 85.
Conclusion
(19) (a) Mc
Bride, J. R.; Soriaga, M. P. J. Electroanal. Chem. 1989, 303,
255. (b) Cali, G. J.; Berry, G. M.; Bothwell, M. E.; Soriaga, M. P. J.
Electroanal. Chem. 1991, 297, 523. (c) Schimpf, J. A.; Carrasquillo, A.;
Soriaga, M. P. Electrochim. Acta 1995, 40, 1203.
Copper UPD may be used to estimate the surface coverage
of irreversibly adsorbed palladium films on Pt(111). The range
of potential in which the palladium film was stable depended
(20) Evans, R. W.; Attard, G. A. J. Electroanal. Chem. 1993, 345, 337.

4606 J. Phys. Chem. B, Vol. 101, No. 23, 1997
Al-Akl and Attard
(21) Clavilier, J.; Faure, R.; Guinet, G.; Durand, R. J. Electroanal. Chem.
1980, 107, 205.
(22) Rodes, A.; El-Achi, K.; Zamakhchari, M. A.; Clavilier, J. J.
Electroanal. Chem. 1990, 284, 245.
(29) Attard, G. A.; Timothy, B. To be published.
(30) Attard, G. A.; Ahmad, A. J. Electroanal. Chem. 1995, 389, 175.
(31) White, J. H.; Abruna, H. D. J. Phys. Chem. 1990, 94, 894.
(32) Aberdam, D.; Salem, C.; Durand, R.; Faure, R. Surf. Sci. 1990,
(23) Hirota, K.; Song, M. B.; Ito, M. Chem. Phys. Lett. 1996, 250, 335.
(24) Wieckowski, A. Private communication.
(25) Attard, G. A.; Bannister, A. J. Electroanal. Chem. 1991, 300, 467.
(26) Preuss, E.; Kraul-Urban, B.; Butz, R. In Kernforsschunganlage
Julich Laue Atlas; Wiley: New York, 1973.
239, 71.
(33) Gileadi, E.; Vaskevich, A.; Rosenblum, M. J. Electroanal. Chem.
1995, 383, 167.
(34) Angerstein-Koslawska, H.; Conway, B. E.; Sharp, W. B. A. J.
Electroanal. Chem. 1973, 43, 9.
(27) Tyson, W. R. Can. Met. Q. 1975, 14, 307.
(35) Wagner, F. T.; Ross Jr., P. N. Surf Sci. 1985, 160, 305.
(36) Soriaga, M. P.; Schimpf, J. A.; Carrasquillo, A.; Abreu, J. B.;
Temesghen, W.; Barriga, R. J.; Jeng, J. J.; Sashikata, K.; Itaya, K. Surf.
Sci. 1995, 335, 273.
(28) Bauer, E. In Chemical Physics of Solid Surfaces and Heterogeneous
Catalysis; King, D. A., Woodruff, D. P., Eds.; Elsevier: Amsterdam, 1982;
Vol. IIIa.