Mechanism of hydrogen adsorption/absorption at thin Pd layers on Au(1 1 1)

Article abstract of DOI:10.1016/j.electacta.2007.03.068

Hydrogen adsorption and absorption at thin palladium deposits of 0.8-10 monolayers (ML) on Au(1 1 1) was studied in 0.1 M H₂SO₄ and HClO₄ using cyclic voltammetry, ac voltammetry, and impedance spectroscopy in the absence and in the presence of poison, crystal violet. Hydrogen adsorption on palladium is more reversible in sulfuric acid than in perchloric acid but it occurs at potentials 30 mV more positive in latter. The charge-transfer resistance exhibits a minimum at ～0.27 V versus RHE and decreases with increasing in Pd deposit thickness in both acids. Adsorption capacitance at 0.8 ML Pd reaches maximum at the same potential. At other deposits the pseudo-capacitance starts to increase at lower overpotentials indicating the beginning of absorption, even at 2 ML Pd. The double layer capacitance is similar for all the deposits in sulfuric acid and it has a sharp maximum at 0.27 V versus RHE. In perchloric acid a broad maximum is observed. Crystal violet inhibits hydrogen adsorption but makes hydrogen absorption more reversible. The results suggest a fast direct hydrogen absorption mechanism that proceeds in parallel with slower hydrogen adsorption and indirect absorption.

Full text of DOI:10.1016/j.electacta.2007.03.068

Electrochimica Acta 52 (2007) 6195–6205
Mechanism of hydrogen adsorption/absorption
at thin Pd layers on Au(1 1 1)
Hugues Duncan, Andrzej Lasia∗
De´partement de Chimie, Universite´ de Sherbrooke, Sherbrooke, Que´bec, Canada J1K 2R1
Received 22 December 2006; received in revised form 21 March 2007; accepted 23 March 2007
Available online 30 March 2007
Abstract
Hydrogen adsorption and absorption at thin palladium deposits of 0.8–10 monolayers (ML) on Au(1 1 1) was studied in 0.1 M H2SO4 and
HClO4 using cyclic voltammetry, ac voltammetry, and impedance spectroscopy in the absence and in the presence of poison, crystal violet.
Hydrogen adsorption on palladium is more reversible in sulfuric acid than in perchloric acid but it occurs at potentials 30 mV more positive in latter.
The charge-transfer resistance exhibits a minimum at ∼0.27 V versus RHE and decreases with increasing in Pd deposit thickness in both acids.
Adsorption capacitance at 0.8 ML Pd reaches maximum at the same potential. At other deposits the pseudo-capacitance starts to increase at lower
overpotentials indicating the beginning of absorption, even at 2 ML Pd. The double layer capacitance is similar for all the deposits in sulfuric acid
and it has a sharp maximum at 0.27 V versus RHE. In perchloric acid a broad maximum is observed. Crystal violet inhibits hydrogen adsorption
but makes hydrogen absorption more reversible. The results suggest a fast direct hydrogen absorption mechanism that proceeds in parallel with
slower hydrogen adsorption and indirect absorption.
© 2007 Elsevier Ltd. All rights reserved.
Keywords: Kinetics and mechanism of hydrogen adsorption and absorption; Thin Pd layer; Impedance spectroscopy; Cyclic voltammetry
1. Introduction
let blocks the hydrogen adsorption but enhances the kinetics
of hydrogen absorption, evidenced by the increased reversibil-
ity of cyclic voltammetric peaks of hydrogen absorption and a
decrease of the charge-transfer resistance. The results suggested
the possibility of a direct hydrogen insertion in Pd [7,10]. It was
recommended that studies of 1–10 monolayers of palladium on
Au(1 1 1) by cyclic voltammetry and impedance spectroscopy
could help resolve the absorption mechanism. These studies
were performed in the present paper in sulfuric and perchlo-
ric acid. First studies of Pd deposition on monocrystals were
carried out by Attard and Bannister [11] and Clavilier et al.
[12], where Pd was deposited on Pt(1 1 1). Kolb and collabora-
tors have studied the initial stages of Pd deposition on Pt(1 1 1),
Au(1 1 1)and(1 0 0), andRh(1 1 1)[13–18]andtheelectrochem-
istry of these thin layers [14,19]. These studies suggested that
Pd forms one UPD layer and further layers on Au(1 1 1) remain
pseudo-morphic up to ∼10 ML. However, no impedance or other
kinetic measurements of hydrogen adsorption were performed
on those deposits and no other kinetic information is available.
Studies of very thin Pd layers should help resolve the hydrogen
absorption mechanism and the kinetics of hydrogen adsorption
reaction. The separation of these two processes is not possible
Absorption of hydrogen in palladium lattice is of great fun-
damental interest. However, there is still discussion of the
mechanism of hydrogen absorption. Two mechanisms were
proposed: (a) indirect absorption, in which the hydrogen goes
through an adsorption step [1,2] and (b) direct absorption mech-
anism [3–7]. These two mechanisms and their consequences
were discussed in our earlier papers [7–9]. Besides, there is
always presence of the underpotential hydrogen at the same
potential range as hydrogen absorption and the separation of
the adsorption and absorption processes is difficult. Moreover,
at less positive overpotentials (0 < η < 70 mV) there is an appear-
ance of the overpotentially deposited hydrogen and formation
of hydrogen in the solution, according to the Nernst equation
[7]. In our recent study [10] the effect of crystal violet on
the hydrogen adsorption and absorption in Pd membranes and
deposits on gold was investigated. It was shown that crystal vio-
∗
Corresponding author. Tel.: +1 819 821 7097; fax: +1 819 821 8017.
E-mail address: a.lasia@usherbrooke.ca (A. Lasia).
0013-4686/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2007.03.068

6196
H. Duncan, A. Lasia / Electrochimica Acta 52 (2007) 6195–6205
at thicker deposits where the adsorption reaction is obscured by
hydrogen absorption.
2. Experimental
All measurements were carried in 0.1 M H2SO4 or 0.1 M
HClO4 (Seastar Baseline) and MilliQ nanopure water. The
solution was deoxygenated by bubbling argon (Ar UHP 5.0,
Praxair) before the measurements. The reference electrode was
Hg|Hg2SO4 in 0.1 M H2SO4 but all the potentials are reported
versus the reversible hydrogen electrode (RHE) in the same
solution. The potential of the mercury sulfate electrode is
0.728 V versus RHE. Au(1 1 1) monocrystal had a surface area
of 0.332 cm2 and was prepared at the University of Ottawa
by Morin and Langlois, following the method described ear-
lier [20,21]. Pd deposits were obtained from a bath containing
0.1 M H2SO4 and 1 mM K2PdCl4 (Aldrich, 99.99%). Based on
Kibler et al. [15] results, an underpotential monolayer of Pd was
deposited by scanning electrode potential at 0.1 mV s−1 from
0.931 V to 0.841 V versus RHE and holding it at this potential
until the current dropped to zero. This ensured that only UPD Pd
was deposited. However, the typical Pd coverage was ∼0.8 ML,
as determined by the integration of the deposition charge. For
2–5 ML deposits the potential was scanned at 0.1 mV s−1 from
0.931 V to 0.726 V versus RHE and the deposition charge was
monitored and related to the thickness. For 10 ML the potential
was held at 0.7 V until the necessary charge was obtained. The
chargenecessarytodepositonemonolayerofPdis440 ␮C cm−2
[14]. The electrodes were used in the hanging meniscus config-
uration. Crystal violet chloride (Aldrich) was pr−ecipitated from
Fig. 1. Cyclic voltammograms of 0.8–5 ML of Pd on Au(1 1 1) in 0.1 M H2SO4
and HClO4 at a scan rate of 1 mV s−1. The voltammogram at bare Au(1 1 1) in
H2SO4 (top) indicates that the capacitive current is negligible.
Au(1 1 1) is not exactly layer by layer after the first monolayer. In
fact, Kibler et al. [15] noted that the formation of the third layer
of Pd starts before the completion of the second. Therefore, the
Pd deposit thickness corresponds to the equivalent of the number
of monolayer rather than an absolute number of monolayer.
In perchloric acid, the voltammetric peaks appear at slightly
more positive potentials, 0.306 V and 0.294 V, respectively. The
absence of adsorbed (bi)sulfate can explain easier hydrogen
adsorption. It can also be noticed that the hydrogen adsorption is
much less reversible than in the presence of sulfuric acid, even at
a scan rate of 1 mV s−1. However, the reversibility increases with
increase in Pd coverage. Addition of a small quantity of sulfate to
the perchloric acid solution immediately shifts the peaks by ca.
−20 mV and the reversibility increases. Baldauf and Kolb [14]
have shown that the position of the peaks is largely independent
of the concentration of sulfate between 10−1 M and 10−5 M but
it depends on H+ concentration. Plotting the cyclic voltammo-
grams with the current density, j, normalized by the sweep rate,
v, j/v, enables to visualize the effect of the sweep rate on the
reversibility of the voltammograms (see Fig. 2). With increase
in the sweep rate the hydrogen adsorption peak shifts towards
more negative potentials with a slight decrease in the intensity
while the hydrogen desorption peak shifts towards more posi-
tive potentials with a marked decrease in intensity and it spreads
over a larger range of potentials. This indicates that the hydrogen
adsorption/desorption is kinetically controlled.
perchloric acid to substitute Cl− anion by ClO4
.
The electrochemical measurements were carried using a
Solartron SI 1260 frequency response analyzer and Solartron
SI 1287 electrochemical interface. ac impedance was measured
between 0.5 V and 0.1 V vs. RHE with the amplitude of 5 mV
except in the range of 0.32–0.26 V where the amplitude was
1 mV. ac voltammetry was measured between 1 Hz and 1000 Hz
at a scan rate of 1 mV s−1 and ac amplitude of 5 mV.
3. Results
3.1. Cyclic voltammetry
Fig. 1 shows cyclic voltammograms of 0.8–5 ML Pd on
Au(1 1 1) at a low sweep rate of 1 mV s−1 in sulfuric and per-
chloric acid. At 0.8 ML Pd there is only one cathodic peak at
0.28 V in sulfuric acid. At 2 ML and greater Pd coverages a
new peak appears at 0.255 V while the first one shifts to 0.27 V.
With increasing coverage between 2 ML and 5 ML the peak at
0.255 V increases while that at 0.27 V stays the same. Baldauf
and Kolb [14] have attributed the more positive peak to hydrogen
adsorption on the perimeter of Pd islands and the more nega-
tive peak to hydrogen adsorption on terraces. No H absorption
was found at a monolayer of Pd in electrochemical conditions
[13–16,22–25] but it seems to start even at coverages as low as
2 ML, as evidenced by the increase in current at the most pos-
itive potentials. This suggests that the growth of palladium on
The monolayer of Pd on Au(1 1 1) is not very stable. Fig. 3
shows evolution of the cyclic voltammograms at 50 mV s−1 on
a newly deposited 0.8 ML Pd on Au(1 1 1). It was suggested
that initially many Pd islands are grown and they coalesce dur-
ing cycling to form a uniform layer [15]. The voltammogram
stabilizes after 20 cycles. Polarizing at 0.1 V immediately after
deposition has the same effect as cycling, but the voltammo-
´
grams stabilize after the second or third cycle. Alvarez et al. [24]

H. Duncan, A. Lasia / Electrochimica Acta 52 (2007) 6195–6205
6197
Fig. 4. Potential dependence of total charge on 0.8–5 ML Pd on Au(1 1 1) in
0.1 M H2SO4 and 0.1 M HClO4.
Fig. 2. Cyclic voltammograms of 2 ML on Au(1 1 1) in 0.1 M H2SO4 and 0.1 M
HClO4 at different sweep rates. The arrows indicate increase in the sweep rate.
5 ML in sulfuric acid and perchloric acid, respectively. Charges
found here agree with those presented by Baldauf and Kolb [14]
for 2 ML, however, they continue to increase at higher Pd cov-
erages while those reported by Baldauf and Kolb are relatively
stable between 2 ML and 5 ML. Considering that one monolayer
of adsorbed hydrogen corresponds to about 240 ␮C cm−2, an
equivalent of a full monolayer of adsorbed hydrogen is formed
at 3 ML and for thicker Pd deposits the excess charge can be
attributedtoabsorbedhydrogen. Theabsorptioncurrentseemsto
start at potentials more positive with increasing palladium thick-
ness, which is consistent with the fact that for massive palladium
adsorption is masked by absorption [14,28].
have noted the appearance of a shoulder at potentials more nega-
tive than the principal peak on 1 ML Pd on stepped Pt(1 1 1), that
is Pt(10 10 9) and Pt(11 10 10), which they attributed to hydro-
gen adsorption on steps. This shoulder exhibits similarities with
the one observed in Fig. 3 and supports the hypothesis of the
coalescence of Pd islands which reduce the number of steps and
leads to the decrease of the shoulder.
Integration of the cathodic peak was carried out between
0.5 V and 0.1 V with subtraction of the capacitive current and it
was assumed that capacitive current is independent of potential.
The results for different thicknesses in sulfuric and perchloric
acid are shown in Fig. 4. For 0.8 ML the charge up to 0.1 V
is 163 ␮C cm−2 in sulfuric acid and 152 ␮C cm−2 in perchloric
acid. This compares well with values of Baldauf and Kolb [14]
(170 ␮C cm−2 for 1 ML) and Hernandez and Baltruschat [26]
(166 ␮C cm−2 for 0.78 ML) for Pd on Au(1 1 1) in 0.1 M H2SO4.
However, it is lower than 190 ␮C cm−2, for 0.7 ML in 0.1 M
HClO4 reported by Naohara et al. [27]. The charges increase
to 205 ␮C cm−2 and 200 ␮C cm−2 for 2 ML, 244 ␮C cm−2 and
242 ␮C cm−2 for 3 ML, and 261 ␮C cm−2 and 258 ␮C cm−2 for
It was suggested in the literature that the peaks in the presence
of sulfuric acid may contain contribution due to the bisulfate
adsorption, and the charge in perchloric acid is expected to
be lower than in sulfuric acid considering the adsorption of
(bi)sulfateanioninsulfuricacidandtheabsenceofanionadsorp-
´
tion in perchloric medium. Alavarez et al. [25] integrated cyclic
voltammograms of 0.8 ML Pd on Pt(1 1 1) and found a total
(anion and hydrogen adsorption) charge of 250 ␮C cm−2 in per-
chloric acid and 265 ␮C cm−2 in sulfuric acid. They also found,
using CO displacement technique, that (bi)sulfate adsorption
charge is 90 ␮C cm−2 and in perchloric acid the anion adsorption
chargeis50 ␮C cm−2. Thechargeattributedtohydrogenadsorp-
tion is then slightly higher in perchloric acid (200 ␮C cm−2) in
comparison with 175 ␮C cm−2 in sulfuric acid. However, the
charges obtained in this work for Pd at Au(1 1 1) are very similar
in both acids.
Fig. 5 shows the effect of adding 0.1 mM crystal violet on the
voltammograms of hydrogen adsorption/absorption on 5 ML of
Pd. In the presence of crystal violet the hydrogen adsorption is
blocked. However, as found earlier [10,14], hydrogen absorption
reversibility is increased by the presence of crystal violet. This is
in agreement with the direct hydrogen absorption mechanism. It
can also be noted that hydrogen sorption (that is adsorption and
absorption) charge does not depend on the nature of the support-
ing electrolyte. An increase in reversibility of absorption in the
presence of CV is clearly visible in Fig. 6A. No absorption peak
can be seen in the presence of CV on 0.8 ML in Fig. 6A. At over-
potentials <0.07 V an increase of current due to OPD hydrogen
Fig. 3. Evolution of cyclic voltammograms of 0.8 ML Pd on Au(1 1 1) during
cycling of a new deposit in 0.1 M H2SO4 at 50 mV s−1 until stabilization.

6198
H. Duncan, A. Lasia / Electrochimica Acta 52 (2007) 6195–6205
carried by Batina et al. [29] and Wan et al. [30]. Simulation of
the ordering of CV adlayer on I-covered Au(1 1 1) [31] was also
carried out. The presence of sulphur or iodine is necessary to
obtain ordered adlayers, and this adlayer on Au(1 1 1) and Pd-
covered Au(1 1 1) is disordered and some hydrogen adsorption
is possible. It is possible that the ordering of the adsorbed layer
changes with the electrode potential allowing for some resid-
ual adsorption, however, more detailed thermodynamic double
layer studies are necessary to determine the ionic adsorption of
CV.
3.2. Impedance spectroscopy
For the deposits of one monolayer of Pd or less only hydrogen
adsorption can take place. In this case the faradaic impedance,
ˆ
Zf, consists of a charge-transfer resistance, Rct,ads, in series with
an adsorption pseudo-capacitance, Cp [32]:
Fig. 5. Cyclic voltammogram of 5 ML Pd on Au(1 1 1) in 0.1 M H2SO4 and
HClO4 with and without 0.1 mM crystal violet (CV); scan rate: 1 mV/s.
1
jωCp
ˆ
Zf = Rct,ads
+
(1)
and the total impedance is:
1
is clearly visible, Fig. 6A [7,10]. At potentials more positive than
0.24 V on 0.8 ML, but before the formation of OPD hydrogen,
current in the presence of crystal violet is higher than that of the
double layer, Fig. 6B, indicating an electrochemical process, and
suggesting that hydrogen adsorption is not entirely blocked by
crystal violet. Hydrogen absorption, evidenced by formation of
peaks, begins already at 2 ML Pd indicating that the formation of
the third layers starts before the completion of the second. Inte-
gration of the voltammograms between 0.1 V and 0.5 V yields
values of 20 ␮C cm−2, 76.4 ␮C cm−2 and 93.7 ␮C cm−2 for 0.8
ML, 2 ML and 5 ML of Pd, respectively. It should be noted that
the major increase in hydrogen absorption charge takes place
at potentials more negative than 0.2 V; for example for 2 ML
the charge between 0.5 V and 0.2 V is 20 ␮C cm−2 in the pres-
ence of CV whereas it is 154 ␮C cm−2 in the absence of CV
in sulfuric acid. This charge is only 2.2 ␮C cm−2 in the pres-
ence of CV as compared to 107 ␮C cm−2 without CV at 0.8
ML in perchloric acid. This small charge may be attributed to
hydrogen adsorption that is not blocked by CV. Scanning tun-
neling microscopy studies of the adsorption of CV and other
organic molecules on iodine or sulfur-modified Au(1 1 1) were
ˆ
Z = Rs +
(2)
ˆ
jωCdl + (1/Zf)
where Cdl is the double layer capacity and Rs the solution resis-
tance. In presence of the direct hydrogen absorption there is
an additional parallel branch consisting of the absorption resis-
ˆ
tance, Rct,abs in series with a Warburg impedance, ZW; the latter
for thin layers reduces to the connection of absorption resistance
(RW) and capacitance (CW) in series [32]. The value of RW is
directly proportional to the layer thickness and its value is very
small. At less positive overpotentials there is a hydrogen OPD
adsorption and, at negative overpotentials, hydrogen evolution.
The electrical equivalent circuit is presented in Fig. 7.
In the analysis of the impedance data the capacitances are
usually replaced by a constant phase element [33]:
1
ZCPE
=
(3)
(jω)φT
where T is the capacity parameter; ω the ac signal angular fre-
quency; φ is the parameter related to the deviation from the
Fig. 6. Cyclic voltammograms at different Pd layers on Au(1 1 1) in presence of 0.1 mM crystal violet in 0.1 M H2SO4.

H. Duncan, A. Lasia / Electrochimica Acta 52 (2007) 6195–6205
6199
Fig. 9. Potential dependence of the double layer capacity, Cdl, on potential for
0.8–10 ML Pd on Au (1 1 1) in 0.1 M H2SO4 and 0.1 M HClO4.
Fig. 7. Equivalent electrical circuit for hydrogen OPD and UPD adsorption, and
direct absorption.
curve with a minimum at ∼0.27 V is observed although at 0.8
ML Pd the curve exhibits a broad minimum. The minimum
charge-transfer resistance at 0.8 ML Pd is ∼800 ꢀ cm2; it cor-
responds to the kinetics of hydrogen adsorption process and
its value is four times larger that on a thicker Pd deposits and
on a polycrystalline Pd membrane, at which adsorption and
absorption processes occur simultaneously. A decrease of a min-
imum of the charge-transfer resistance with the increase of the
Pd deposit layer thickness is clearly visible. This decrease is
more pronounced at potentials more negative than the minimum
(between 0.1 V and 0.27 V). Formation of a maximum of Rct
at more negative potentials at thin Pd deposits (>10 nm) was
attributed earlier to the hydrogen trapping mechanism [34,35].
In perchloric acid, the charge-transfer resistance is slightly lower
than in sulfuric acid and the trend is the same, i.e. a decrease
with increasing Pd thickness. The only exception is for 0.8 ML
ideal capacitance. Experimental values of φ are in the range of
0.99–0.95.
The experimental complex plane plots displayed always a
kinetic semicircle followed by a capacitive line and the three
processes could not be separated. An example of the complex
plane plot is shown in Fig. 8.
The double layer capacitance, shown in Fig. 9, is practi-
cally independent of the deposit thickness, as expected, and
shows a sharp peak at 0.27 V which can be related to sulfate
adsorption/desorption. In the absence of sulfate adsorption (in
perchloric acid) a broader maximum is observed.
Fig. 10 shows variation of the charge-transfer resistance with
potential for various Pd deposits. In sulfuric acid a V-shaped
Fig. 8. Complex plane plot at 2 ML Pd on Au(1 1 1) at 0.2 V in 0.1 M
H2SO4 without (empty circles) and with (filled circles) 10−4 M crystal violet
(lines—approximations).
Fig. 10. Potential dependence of the charge-transfer resistance, Rct, on potential
for Pd deposits on Au(1 1 1) of 0.8–10 ML in 0.1 M H2SO4 and HClO4. Data
for 100 ␮m Pd foil from Ref. [35].

6200
H. Duncan, A. Lasia / Electrochimica Acta 52 (2007) 6195–6205
Fig. 11. Dependence of the charge-transfer resistance on potential (vs. summit
potential Es) for the hydrogen adsorption reaction under Langmuir and Frumkin
isotherm conditions, charge-transfer coefficients assumed α = 0.5, interaction
parameter g indicated in the graph.
Fig. 12. Potential dependence of the low frequency capacitance on potential at
0.8–10 ML Pd deposits on Au(1 1 1) in 0.1 M H2SO4 and 0.1 M HClO4.
sulfuric and 0.245 mF cm−2 in perchloric acid. This is in agree-
mentwiththeHadsorptionisotherms, Fig. 4, wherelowercharge
was found at 0.8 ML and suggests a more complex adsorption
isotherm.
Pd at which the charge-transfer resistance is higher in perchloric
than in sulfuric acid.
The fact that the double layer capacity is practically inde-
pendent of thickness also indicates that the surface roughness
is approximately the same and that the decrease in the charge-
transfer resistance is related to changes in the kinetics and not
to the changes in real surface area.
Impedance measurements were also carried out in the pres-
ence of 0.1 mM crystal violet. The charge-transfer resistances
are presented in Fig. 13. Instead of a V-shaped curves observed
in the absence of crystal violet the charge-transfer resistance
increases linearly (on a logarithmic scale) with the potential. At
2–5 ML Pd the charge-transfer resistance is lower at all poten-
tials than in sulfuric or perchloric acid without crystal violet. The
charge-transfer resistance decreases initially with the increase in
the Pd layer thickness but for 3 and 5 ML it is similar, as observed
earlier at thicker layers [10]. At 0.8 ML Pd Rct in the presence of
CV is lower than that observed in pure acids at potentials below
0.18 V and higher at higher potentials.
Formation of the minimum of the charge-transfer resis-
tance is predicted theoretically for the electrosorption reaction
under Langmuir isotherm conditions [32]. For the Frumkin type
isotherm the simulated dependence of Rct,ads is shown in Fig. 11.
A broad minimum, a flat minimum or even a maximum may be
observed at the surface coverage close to θ = 0.5. For the trans-
fer coefficients different from 0.5 asymmetric branches may be
observed, similar to the experimentally found at Pd monolayer.
The slopes dE/dlog Rct are quite small at more positive poten-
tials, 43–90 mV dec−1, and large at more negative potentials
−250 mV dec−1 to −420 mV dec−1. This strongly suggests the
asymmetry of transfer coefficients and a more complex adsorp-
tion isotherm.
Providing that at 0.8 ML Pd only hydrogen adsorption occurs
the decrease of the charge-transfer resistance with increase in
thickness could be interpreted as an evidence of the faster hydro-
gen absorption reaction in parallel with H adsorption.
The low frequency capacitance CLF is a sum hydrogen
adsorption, Cp, and absorption, CW, pseudo-capacitances and
the double layer capacitance, Cdl, CLF = Cp + CW + Cdl; it is plot-
ted as a function of potential in Fig. 12. There is a peak of CLF at
∼0.3 V in perchloric and at 0.27 V in sulfuric acid, correspond-
ing to hydrogen adsorption. At more negative potentials CLF first
decreases then increases as hydrogen absorption becomes more
important. As expected, CLF increases with increase in Pd thick-
ness due to increase in CW. For 0.8 ML Pd CLF = Cp + Cdl and at
the peak potential, assuming the Langmuir isotherm, the value of
Cp should be 2.04 mF cm−2 [32] whereas it is 0.285 mF cm−2 in
Fig. 13. Potential dependence of the charge-transfer resistance, Rct, on potential
for Pd deposits on Au(1 1 1) of 0.8–5 ML in 0.1 M H2SO4 with 0.1 mM crystal
violet. Results for 3 ML Pd in 0.1 M H2SO4 and 0.1 M HClO4 without crystal
violet are added for comparison.

H. Duncan, A. Lasia / Electrochimica Acta 52 (2007) 6195–6205
6201
The low frequency capacitance, Fig. 15, is lowered at high
potentials (i.e. where adsorption is the predominant reaction)
and the maxima present in sulfuric and perchloric acid are also
suppressed. However, CLF increases with decrease in potential
due to increase in absorption capacitance, CW.
3.3. ac voltammetry
Impedance spectroscopy measurements were carried out by
applying a sinusoidal perturbation at a constant potential. How-
ever, one can also use the ac voltammetry, that is sweep the
potential while applying a sinusoidal perturbation at fixed fre-
quency [36,37]. By repeating these experiments at various
frequencies it is possible to reconstruct the Nyquist and Bode
plots at each potential. The low frequency range is limited in
practice by the sweep rates used [36]. In the case of hydrogen
adsorption on Pd in sulfuric acid where the voltammetric peaks
are sharp it is difficult to make the steady-state measurements
around the voltammetric current peak and the ac voltammetry
might be useful in this case. Figs. 16 and 17 show the comparison
of the double layer capacitances and charge-transfer resistances
at 2 ML Pd on Au(1 1 1) measured by conventional impedance
spectroscopy and reconstructed using ac voltammetry. In our
experiments the potential was swept at 1 mV s−1 and the lowest
frequency used was 1.259 Hz. These two types of measurements
gave essentially the same results. One drawback of ac voltamme-
try for this system is the lack of the low frequency data which
give access to the adsorption pseudo-capacitance. This meth-
ods works for 2 ML and thicker deposits, however for 0.8 ML
the Nyquist diagrams cannot be approximated correctly which
meansachangeinthedepositstructureduringpotentialscycling.
Fig. 14. Potential dependence of the double layer capacity, Cdl, on potential for
0.8–5 ML Pd on Au(1 1 1) in 0.1 M H2SO4 with 0.1 mM crystal violet; 3 ML
Pd in 0.1 M H2SO4 and HClO4 is added for comparison.
In the presence of crystal violet only one slope of dE/dlog Rct
is observed; at 0.8 ML Pd this slope is 106 mV dec−1 and it
increases to ∼200 mV dec−1 for thicker Pd layers. Large differ-
ence between the values of Rct at 0.8 ML and other deposits is
related to the fact that hydrogen absorption is impossible at this
layer. However, the fact that there is a presence of Rct suggests
that some adsorption is still possible and that crystal violet does
not block it completely, in agreement with cyclic voltammetric
experiments.
In the presence of crystal violet the peak in double layer
capacity region disappears, Fig. 14, and Cdl is practically inde-
pendent of the Pd thickness, except at 0.8 ML where the
capacitance is lower. The suppression of the maximum is due to
the suppression of hydrogen and anion adsorption.
3.4. Simulations of the adsorption/adsorption processes
As it was mentioned above the total process taking place
at Pd involves: (a) H UPD, (b) H direct absorption, and (c) H
Fig. 15. Potential dependence of the low frequency capacitance on potential at
0.8–5 ML Pd deposits on Au(1 1 1) in 0.1 M H2SO4 with 0.1 mM crystal violet.
3 ML Pd in 0.1 M H2SO4 and HClO4 is added for comparison.
Fig. 16. Comparison between the double capacity obtained by ‘classic’
impedance and by ac voltammetry on 2 ML Pd on Au(1 1 1) in 0.1 M H2SO4.

6202
H. Duncan, A. Lasia / Electrochimica Acta 52 (2007) 6195–6205
ˆ
tance in series ZW = RW + 1/jωCW [10,32]. Simulations were
carried out for these processes assuming Langmuir isotherms;
the equilibrium surface coverages, adsorption isotherms, the
corresponding pseudo-capacitances and charge-transfer resis-
tances are displayed in Fig. 18. The parameters were chosen
in such a way that the processes are separated; hydrogen UPD
takes place at most positive potentials, EU0 PD = 0.3 V, hydro-
gen absorption at the intermediate potentials, Ea0bs = 0.1 V, and
Volmer and Heyrovsky reactions at more negative potentials,
EV0 = −EH0 = −0.05 V. Hydrogen UPD and hydrogen absorp-
tion isotherms reach maximal value of one [7,9] while that for
the hydrogen evolution lower values, dependent on the rate con-
stants [32]. In the OPD adsorption mechanism formation of
hydrogen in the solution was taken into account [7].
Two maxima related to the maximum of hydrogen UPD
adsorption and absorption as well as a shoulder for H OPD
are visible on the pseudo-capacitance plots. Two minima of the
charge-transfer resistances are observed for the H UPD and H
absorption and a linear decrease for the H OPD (with a shallow
local minimum at −0.05 V).
Fig. 17. Comparison between the charge transfer obtained by ‘classic’
impedance and by ac voltammetry on 2 ML Pd on Au(1 1 1) in 0.1 M H2SO4.
OPD; the corresponding equivalent circuit was shown in Fig. 7.
It contains three parallel branches related to the above processes
and to the double layer capacitance. At positive overpotentials
Rp,HOPD = ∞ as there is no dc current due to hydrogen evolu-
tion in the circuit, and for thin layers the Warburg impedance
reduces to a simple connection of the resistance and capaci-
Examples of the complex plane plots at various potentials
are shown in Fig. 19. The complex plane plots display one (at
overpotentials 0.25–0.4 V) or two (0.1 V) semicircles or overlap-
ping semicircles (0.15–0.20 V), followed by the capacitive line
as no dc current may flow at positive overpotentials. The low
Fig. 18. Dependence of the equilibrium UPD and OPD hydrogen surface coverage and bulk occupation X0, pseudo-capacitances and resistances on potential.
Parameters used in the simulations: rate constants of H UPD, kUPD = 10−8, H direct absorption kabs = 10−8, Volmer and Heyrovsky reactions with H2 formation in
solution [7] kV = 10−6, kH = 10−8 mol cm−2 s−1, standard potentials EU0 PD = 0.3 V, Ea0bs = 0.1 V, EV0 = −EH0 = −0.05 V, Cdl = 20 ␮F cm−2, maximal charge of
adsorption σ1 = 210 ␮C cm−2 and absorption σX = 7589 C cm3 [7], Pd layer thickness 5 ML; RLF is the low frequency resistance.

H. Duncan, A. Lasia / Electrochimica Acta 52 (2007) 6195–6205
6203
Fig. 19. Examples of the complex plane plots at different potentials simulated for the same parameters as in Fig. 18.
frequency real resistance, RLF, found at the capacitive straight
line, follows minimal values of the charge-transfer resistances
of separate processes while the low frequency capacitance fol-
lows the corresponding maximal values of pseudo-capacitances.
These variations are visible in Fig. 18. At more positive poten-
tials, where charge-transfer resistances tend to infinity and
pseudo-capacitances to zero, a simple Rs − Cdl behavior is
observed.
Of course, the parameters used for simulations are arbitrary
but they display a similar trend in function of the potential as the
experimental curves. In practice, behavior as that observed at
0.15 V and 0.2 V may be mistaken as a constant phase behavior
or as a deformation of the impedance plots due to variation of
low frequency impedances with time.
other peaks [8,35]. It was concluded that the adsorption reaction
is more reversible in sulfuric than in perchloric acid as evidenced
by cyclic voltammetry and impedance spectroscopy. The freshly
deposited Pd monolayer is unstable but stabilizes during poten-
tial cycling producing a stable voltammogram. This might be
related to the fact that Pd deposition starts at defects on Au(1 1 1)
surface as the deposition at island is more difficult. However, the
stable voltammogram is similar to that found in the literature
[14,16].
The hydrogen adsorption charge on 0.8 ML Pd (163 ␮C cm−2
in sulfuric and 152 ␮C cm−2 in perchloric acid), integrated up
to 0.1 V, is smaller than that predicted for the H monolayer
(240 ␮C cm−2) [14]. The total hydrogen charge, which includes
hydrogen adsorption, absorption and ion adsorption, reaches
equivalent of a full H monolayer only at 3 ML. Moreover, the
total peak charges at 2–5 ML are independent of the nature of
the acid. These facts may suggest that initially many small Pd
islands are formed which coalesce to larger layers. The hydro-
gen seems to adsorb preferentially on the perimeter of these
islands [14,16] and only at thicker layer on terraces. The ionic
adsorption in both acids seems be similar. Some small hydrogen
absorption is observed at potentials higher than 0.1 V for even
2 ML Pd indicating that the deposit does not grow perfectly
epitaxially and some higher islands are formed on the surface.
Crystal violet adsorbs on surface but its layer is probably dis-
ordered. It blocks hydrogen adsorption, Fig. 6, however, some
small residual H adsorption is observed even at potentials more
positive than 0.1 V. It enhances significantly the kinetics of
hydrogen absorption and should be called promoter of hydrogen
absorption.
4. Discussion
The hydrogen adsorption/absorption reactions on very thin
layers of Pd on Au(1 1 1) were studied in perchloric and sulfuric
acids in the absence and in the presence of crystal violet. There
is a strong effect of the support, Au(1 1 1), on the behavior of Pd
thin deposits. Initially, a more positive adsorption peak appears
on voltammograms at 0.8 ML Pd. In H2SO4 it reaches a maxi-
mum at 2 ML and a new, more negative adsorption peak starts
to grow with the Pd layer thickness. At potentials negative of
these peaks a current plateau due to hydrogen adsorption and
absorption into ␣-PdH phase is observed [8,35].
The adsorption peaks are influenced by the ionic adsorp-
tion as they appear at more positive potentials in HClO4, where
ionic adsorption is much weaker, than in H2SO4. Of course at
thicker deposits the absorption current becomes larger and it
overlaps with adsorption peaks and, eventually, large current
due to hydrogen absorption in ␤-PdH phase overlaps and hides
The charge-transfer resistance is the highest at a sub-
monolayer Pd deposit and decreases with increase in thickness,
indicating that hydrogen adsorption is slower and the indirect

6204
H. Duncan, A. Lasia / Electrochimica Acta 52 (2007) 6195–6205
5. The charge-transfer resistance decreases with the increase
in Pd layer thickness and it decreases in presence of crystal
violet, indicating that the direct absorption path is faster than
the indirect one. Although crystal violet is a poison for the
hydrogen adsorption it acts as hydrogen absorption promoter.
6. The hydrogen absorption mechanism might proceed through
the indirect and direct paths but in the electrochemical studies
the direct paths is much faster.
7. The total mechanism at positive overpotentials involves
hydrogen UPD and OPD adsorption, hydrogen indirect
absorption, and hydrogen formation in solution. Simulation
of the impedances of such mechanism predicts formation
of the minima of the total Rct and maxima of the total low
frequency capacitance, Fig. 18. Some parts of the predicted
behavior are observed experimentally.
Fig. 20. Mechanism of the hydrogen absorption in palladium.
absorption is faster. This fact might also be related to the differ-
ence in lattice parameters of Au and Pd as interatomic distances
between palladium atoms at Au(1 1 1) are larger than those in
Pd(1 1 1) [38]. This affects the extent of the metal–metal bond-
ing within the surface which is accompanied by a shift in the
d-band center [18].
Acknowledgement
Financial support from CRSNG (Canada) is gratefully
acknowledged.
Impedance studies using the conventional impedance spec-
troscopy and ac voltammetry gave the same results although the
ac voltammetry is limited at low frequencies.
References
[1] J.O’.M. Bockris, J. McBreen, L. Nanis, J. Electrochem. Soc. 112 (1965)
1025.
The mechanism of the hydrogen absorption is schematically
depicted in Fig. 20. There are two types of hydrogen existing on
the surface: H UPD which appears at the most positive potentials
and H OPD which could appear at much less positive potentials
η ≤ ∼70 mV [7]. The kinetics of hydrogen adsorption is slow as
found from the voltammetric and impedance measurements at
one Pd monolayer. There must be two paths of hydrogen absorp-
tion coexisting: the fast path is the direct absorption mechanism
and the other, much slower, is the indirect path. The indirect path
which passes through hydrogen adsorption might be observed
during absorption of gaseous hydrogen into Pd. As the two paths
are in parallel the faster one, that is the direct path, is practically
observed. Its kinetics is not slowed down by adsorption of crystal
violet; in fact it even becomes faster. It is not clear why crystal
violet accelerates the kinetics of hydrogen absorption. Further
studies are carried out in our laboratory.
[2] A. Lasia, D. Gregoire, J. Electrochem. Soc. 142 (1995) 3393.
[3] I.A. Bagotskaya, Zh. Fiz. Khim. 36 (1962) 2667.
[4] A.N. Frumkin, in: P. Delahay (Ed.), Advances in Electrochemistry and
Electrochemical Engineering, vol. 3, Interscience, New York, 1963,
p. 287.
[5] G. Zheng, B.N. Popov, R.E. White, J. Electrochem. Soc. 142 (1995) 154.
[6] C. Lim, S.-I. Pyun, Electrochim. Acta 39 (1994) 363.
[7] A. Lasia, J. Electroanal. Chem. 593 (2006) 159.
[8] L. Birry, A. Lasia, in: C. Gabrielli (Ed.), Proceedings of the 17ie`me Forum
sur les impe´dances e´lectrochimiques, CNRS-Universite´ Pierre et Marie
Curie, Paris, 2005, p. 1.
[9] R. Jurczakowski, B. Łosiewicz, A. Lasia, in: P. Vany´sek (Ed.), Trans-
actions of the Electrochemical Society, Impedance in Electrochemistry:
From Analytical Applications to Mechanistic Speculations, vol. 2(17), The
Electrochemical Society, 2007, p. 11.
[10] L. Birry, A. Lasia, Electrochim. Acta 56 (2006) 3356.
[11] G.A. Attard, A. Bannister, J. Electroanal. Chem. 300 (1991) 467.
[12] J. Clavilier, J.-P. Ganon, M. Petit, J. Electranal. Chem. 265 (1989) 231.
[13] R. Hoyer, L.A. Kibler, D.M. Kolb, Electrochim. Acta 49 (2003) 63.
[14] M. Baldauf, D.M. Kolb, Electrochim. Acta 38 (1993) 2145.
[15] L.A. Kibler, M. Kleinert, R. Randler, D.M. Kolb, Surf. Sci. 443 (1999)
19.
[16] J. Tang, M. Petri, L.A. Kibler, D.M. Kolb, Electrochim. Acta 51 (2005)
125.
[17] R. Hoyer, L.A. Kibler, D.M. Kolb, Surf. Sci. 562 (2004) 275.
[18] L.A. Kibler, A.M. El-Aziz, R. Hoyer, D.M. Kolb, Angew. Chem. Int. Ed.
44 (2005) 2080.
[19] L.A. Kibler, A.M. El-aziz, D.M. Kolb, J. Mol. Catal. A: Chem. 199 (2003)
57.
[20] A. Hamelin, in: B.E. Conway, J.O’.M. Bockris, R.E. White (Eds.), Modern
Aspects of Electrochemistry, vol. 16, Kluwer/Plenum, New York, 1985, p.
1.
5. Conclusions
Very thin Pd layers, 0.8–10 ML, deposited onto Au(1 1 1)
were studied in perchloric and sulfuric acids and in the presence
of crystal violet. The following conclusions can be drawn:
1. There is a very strong effect of the support Au(1 1 1) on the
H adsorption on Pd thin layers.
2. The hydrogen adsorption on Pd is thermodynamically easier
in the perchloric than sulfuric acids although the same total
charges are observed in both acids.
3. The hydrogen adsorption is more reversible in sulfuric acid
than in perchloric acid.
4. Hydrogen absorption starts at 2 ML Pd indicating that the
growth is not layer-by-layer, as 3 ML is required to have
hydrogen absorption.
[21] D.-F. Yang, C.P. Wilde, M. Morin, Langmuir 12 (1996) 6570.
[22] B. Alvarez, V. Climent, A. Rodes, J.M. Feliu, J. Electroanal. Chem. 497
(2001) 125.
[23] B. Alvarez, A. Berna, A. Rodes, J.M. Feliu, Surf. Sci. 573 (2004) 32.
´
[24] B. Alvarez, J.M. Feliu, J. Clavilier, Electrochem. Commun. 4 (2002)
379.

H. Duncan, A. Lasia / Electrochimica Acta 52 (2007) 6195–6205
6205
´
[33] A. Lasia, B.E. Conway, J.O’.M. Bockris, in: R.E. White (Ed.), Modern
Aspects of Electrochemistry, vol. 32, Kluwer/Plenum, New York, 1999, p.
143.
[34] C. Gabrielli, P.P. Grand, A. Lasia, H. Perrot, J. Electrochem. Soc. 151
(2004) A1919.
[35] C. Gabrielli, P.P. Grand, A. Lasia, H. Perrot, J. Electrochem. Soc. 151
(2004) A1937.
[36] D.A. Harrington, J. Electroanal. Chem. 355 (1993) 21.
[37] M.E. van der Geest, N.J. Dangerfield, D.A. Harrington, J. Electroanal.
Chem. 420 (1997) 89.
[25] B. Alvarez, V. Climent, A. Rodes, J.M. Feliu, Phys. Chem. Chem. Phys. 3
(2001) 3269.
[26] F. Hernandez, H. Baltruschat, Langmuir 22 (2006) 4877.
[27] H. Naohara, S. Ye, K. Uosaki, J. Electroanal. Chem. 500 (2001) 435.
[28] C. Gabrielli, P.P. Grand, A. Lasia, H. Perrot, J. Electrochem. Soc. 151
(2004) A1925.
[29] N. Batina, M. Kunitake, K. Itaya, J. Electroanal. Chem. 405 (1996) 245.
[30] L.-J. Wan, S. Shundo, J. Inukai, K. Itaya, Langmuir 16 (2000) 2164.
[31] N. Batina, A. Huerta, O. Pizio, S. Sokolowski, A. Trokhymchuk, J. Elec-
troanal. Chem. 450 (2000) 213.
[38] A.M. El-Aziz, R. Hoyer, L.A. Kibler, D.M. Kolb, Electrochim. Acta 51
(2006) 2518.
[32] A. Lasia, in: B.E. Conway, R.E. White (Eds.), Modern Aspects of Electro-
chemistry, vol. 35, Kluwer/Plenum, New York, 2002, p. 1.