Hydrogen sorption in Pd monolayers in alkaline solution

Article abstract of DOI:10.1016/j.electacta.2009.01.051

Hydrogen adsorption/absorption at palladium monolayers (ML) deposited on monocrystalline Au(1 1 1) electrode was studied in 0.1 M NaOH solution. H charge isotherms demonstrated that adsorption started at potentials more positive than at thicker nanometric

Full text of DOI:10.1016/j.electacta.2009.01.051

Electrochimica Acta 54 (2009) 5292–5299
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Hydrogen sorption in Pd monolayers in alkaline solution
M.H. Martin, A. Lasia∗
Département de chimie, Université de Sherbrooke, 2500 blvd. de l’Universite, Sherbrooke, Québec, Canada, J1K 2R1
a r t i c l e i n f o
a b s t r a c t
Article history:
Hydrogen adsorption/absorption at palladium monolayers (ML) deposited on monocrystalline Au(1 1 1)
electrode was studied in 0.1 M NaOH solution. H charge isotherms demonstrated that adsorption started
at potentials more positive than at thicker nanometric Pd/Au(polycrystal) deposits. Due to 3-dimensional
deposit growth, absorption could be seen at all deposits thicker than 1 ML. Besides, H sorption at
Pd/Au(1 1 1) monolayers was more reversible than at nanometric Pd/Au(polycrystal) deposits. Strong
geometric and electronic effects due to the Au substrate were observed up to 5 Pd ML. Influence of ben-
zotriazole (BTA) on H sorption was also investigated. BTA blocked H adsorption above 250 mV vs. RHE.
At less positive potentials adsorbed BTA layer seemed to undergo a reorientation allowing H adsorp-
tion. Stationary and dynamic electrochemical impedance spectroscopy was used to obtain double layer
capacitance and charge transfer resistance. BTA also promoted kinetically H sorption into Pd/Au(1 1 1)
monolayer and Pd/Au(polycrystal) nanometric deposits.
Received 29 October 2008
Received in revised form 7 January 2009
Accepted 16 January 2009
Available online 24 January 2009
Keywords:
Hydrogen adsorption and absorption
Palladium monolayers
Alkaline solutions
Benzotriazole
Isotherms
Impedance spectroscopy
© 2009 Elsevier Ltd. All rights reserved.
1. Introduction
Another interesting way to get a better understanding of adsorp-
tion and absorption mechanisms is to use a poison that will block
Palladium has been intensively investigated, mainly because of
its capacity to adsorb and absorb large amount of hydrogen [1,2].
Besides, Pd is also used for its catalytic properties. Since it is an
expensive material, the interest of using it as thin films is obvi-
ous. Moreover, deposits in the monolayer range exhibit different
catalytic properties due to the interactions between the substrate
and the deposit [3,4]. Czerwinski et al. studied H sorption into Pd
thin films in both acid and alkaline media [5–12]. Following their
work, nanometric Pd films deposited at polycrystalline gold elec-
trode were studied by our group in acid media [13–17] and in
NaOH [18]. Paillier and Roué [19] also investigated H sorption at
Pd/Au(polycrystal) nanometric films in KOH. In parallel, Pd mono-
layers deposited on monocrystals demonstrated interesting prop-
erties. The first studies of Pd deposition on Pt(1 1 1) were carried
out by Clavilier et al. [20] and Attard and Bannister [21]. Kolb and
coworkers have also investigated Pd deposition on Au [22–25], Pt
[26,27], Rh [28] and other monocrystals [27] exploring the influence
of anions during deposition and showing that the Pd growth was 3-
dimensional after the first monolayer. In our earlier papers deposits
on monocrystals Pd/Au(1 1 1) and Pt(1 1 1) were studied to bet-
ter understand adsorption and absorption mechanisms in sulfuric
and perchloric acid, determine H isotherms and using impedance
measurements to evaluate kinetic effects [13,15,16,29,30]. However,
until now no studies were carried out in alkaline media.
one of the two hydrogen sorption reactions. Conway et al. [31,32]
investigated the influence of As on H insertion into Fe and steel
in alkaline media and found that the coverage ratio of metal by
adsorbed hydrogen is decreased but, at very negative potentials, H
absorption is promoted. Birry [33] studied influence of As on H sorp-
tion into Pd and found very strong poisoning effect even at very low
As concentrations but the results were not reproducible. The role
of thiourea was also studied [31,34] and a similar poisoning behav-
ior was observed. Carbon monoxide is also known for poisoning
Pd surfaces and blocking adsorption and absorption [35–40]. More
recently, our group lookedat theeffect ofcrystalvioleton Hsorption
at Pd thin films in acidic solutions [13,29], following a preliminary
work of Baldauf and Kolb [22]. It was concluded that crystal violet
inhibits H adsorption and makes absorption more reversible, which
was confirmed by observations of Bartlett et al. [41,42]. However,
crystal violet is insoluble in alkaline solutions. Another interest-
ing molecule to study H sorption in alkaline media is benzotriazole
(BTA). It is widely used and studied for its anti-corrosion protective
ability. Baldauf and Kolb [22] also demonstrated that BTA could have
poisoning properties towards H sorption into Pd even in acid media.
Elhamid et al. [43] and Amokrane et al. [44] showed that BTA blocks
H adsorption at Fe and diminish H permeation. To the best of our
knowledge, no further investigations on the influence of BTA on H
sorption in Pd thin films have been carried out.
In the present work cyclic voltammetry, H charge isotherms
and impedance spectroscopy were used to study nanometric palla-
dium films on Au(polycrystal) and Pd/Au(1 1 1) monolayers, in 0.1 M
NaOH, with and without addition of benzotriazole.
∗
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 © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2009.01.051

M.H. Martin, A. Lasia / Electrochimica Acta 54 (2009) 5292–5299
5293
2. Experimental
The 30 and 70 nm palladium films were deposited on a polycrys-
talline gold electrode, with a surface area of 0.322 cm2, from a bath
containing 0.1 M PdCl2 (Aldrich, 99%) + 0.3 M LiCl (Aldrich, 99%), at
a current density of 1 mA cm−2 [13,45].
Au(1 1 1) monocrystalline electrode had a geometric surface
area of 0.332 cm2 and was prepared by Morin and Langlois from
University of Ottawa, following the method described earlier
[46,47]. Pd monolayers were deposited on this gold monocrys-
tal, from a bath containing 0.1 M H2SO4 (Seastar Baseline) and
1 mM K2PdCl4 (Aldrich, 99.99%), by scanning electrode potential
at 0.1 mV s−1 from 0.931 to 0.841 V vs. reversible hydrogen elec-
trode (RHE), following the protocol of Duncan and Lasia [29]. The
apparent number of monolayers was estimated from the deposi-
tion charge, assuming that 440 ␮C cm−2 is needed to form a Pd
monolayer on Au(1 1 1) monocrystal [22].
The measurements were carried out in 0.1 M NaOH (Aldrich,
99.998%), and 0.1 M NaOH + 10 mM BTA (Aldrich, 99%). All solutions
were prepared using nanopure water (Millipore, Milli-Q gradient
system) and deoxygenated by bubbling argon (Praxair, Ar UHP 5.0).
A platinum gauze and an Hg | HgO in 0.1 M NaOH were used as
counter and reference electrode, respectively. The reference elec-
trode potential was E = +935 mV vs. reversible hydrogen electrode
RHE in the same solution. All the potentials are reported vs. RHE
in the same solution. The working electrode was used in a hanging
meniscus configuration.
To get more reproducible results taking into account quite fast
surface contamination in alkaline solution (contrary to the acid
solutions), before and after short series of experiments, the elec-
trode was transferred into 0.1 M H2SO4 and cycled between +100
and +500 mV at 50 mV s−1 until a correct [13,14] voltammogram
was obtained, after which the experiment was continued in alka-
line solution. Repeating such a procedure regularly allowed for
obtaining of reproducible results. The treatment in H2SO4 consisted
of series of cyclic voltammograms in the same potential range at
50 mV s−1 and polarizations at +100 mV during 30 s.
The electrochemical measurements were carried out using a
Solartron SI 1260 Frequency Response Analyzer, and a Princeton
Applied Research 273A potentiostat. The Dynamic Electrochemi-
cal Impedance Spectroscopy (DEIS) was used by applying a sum of
35 odd harmonic frequencies during the potential sweep through
PAR 273A potentiostat. The generation of the ac signal as well
as the data acquisition was carried out using Keithley KSUB3116
D/A and A/D module. The electronics for the signal amplifica-
tion/attenuation and subtraction/addition was carried out using a
homemade electronic system. The amplitude of the ac signals was
varied with frequency according to the suggestion of Popkirov and
Schindler [48] and the phases were randomized. The data genera-
tion, acquisition, and analysis program was written and provided
by D. Harrington, University of Victoria.
Fig. 1. Cyclic voltammograms at a 70 nm Pd deposit (—) and 10 ML Pd/Au(1 1 1)
deposit (
), in 0.1 M NaOH; scan rate: 10 mV s−1
.
ꢀ
and A2 in Ref. [50]). The presence of two different sites at the Pd
surface could be the reason for this phenomenon. Two peaks were
also observed in H2SO4 at Pd/Au(1 1 1) thin deposits until 10 mono-
layers [22,29]. The peak at more positive potentials was related
to hydrogen adsorption at the steps while the second peak was
representative of adsorption on terraces.
Comparison between the voltammograms at a 10 ML Pd deposit
and a 70 nm Pd film shows that adsorption at Pd monolayer deposits
begins at more positive potentials. Similar behavior was observed
in acid solutions [14,29]. The second observation is an important
increase of the reversibility of the H-absorption reaction at Pd
monolayers (peaks C1 and A1) in comparison with nanometric Pd
films. This kinetic effect could be a consequence of the substrate
influence. This was confirmed by the impedance measurements.
Fig. 2 allows for a better understanding of the nature of dif-
ferent peaks, as well as for the effect of the electrode surface
contamination. The electrode was cycled in alkaline solution after
pretreatment in 0.1 M H2SO4, see Section 2 for details. Since the
deposit was perfectly stable (reproducible cyclic voltammograms)
in sulfuric acid, the evolution in voltammograms recorded in 0.1 M
NaOH and displayed in Fig. 2 could not be due to changes in
3. Results and discussion
3.1. H sorption at Pd monolayers
Fig. 1 displays the voltammograms at a 70 nm Pd film deposited
on a polycrystalline Au and 10 monolayers (ML) Pd deposited on
Au(1 1 1) electrodes. A detailed description of voltammograms for
nanometric Pd films (tenths of nanometers thickness) was already
given in our previous work [18]. Five different peaks can be distin-
guished at Pd monolayers. Considering earlier works [18,49,50] and
the potential range at which they appear A1 and C1 were attributed
to H absorption and desorption. Peaks A2, A3 and C2 were attributed
to H adsorption and desorption. Two anodic peaks were already
Fig. 2. Evolution of cyclic voltammograms with the number of cycles, at a 10 ML
ꢀ
Pd/Au(1 1 1) film; scan rate: 10 mV s−1
.
seen at nanometric Pd films (see peaks A1 and A1 in [18] or A2

5294
M.H. Martin, A. Lasia / Electrochimica Acta 54 (2009) 5292–5299
out. However, we began our experiment from 320 mV instead of
420 mV. We finally carried out 3rd isotherm measurements in the
potential range of 220 and 60 mV, after another cleaning in H2SO4.
Fig. 3 clearly shows that the 3rd isotherm is higher than the 2nd
isotherm, which is higher than the 1st one. The final isotherm
reported for H sorption at a 7 ML Pd deposit is thus the maximal
isotherm, i.e. the 1st one between 420 and 320 mV, the 2nd between
320 and 240 mV, and the third one between 220 and 60 mV. This
setup gave us reproducible results, with clear thickness dependence
of the H oxidation charge. Moreover, these results were similar to
those obtained when carrying out very short experiments without
impedance measurements. Ideally, a cleaning would be performed
after each measurement (polarization/impedance/discharge) at a
single potential in NaOH. However, each opening of the electro-
chemical glass cell and transfer of the electrode increased the risk of
important contamination. Therefore, we considered that our setup
was the best compromise in our “race” against contamination.
Isotherms at Pd thin deposits with thicknesses between 1 and 10
ML are presented in Fig. 4, including the adsorption charge density
Qads obtained earlier [18] at thicker Pd/Au(polycrystal) deposits by
separations of adsorption and absorption contributions from the
regression analysis. As mentioned in our earlier works [13,29], it
is practically impossible to electro-deposit more than 0.8 ML Pd
on Au(1 1 1) as the first monolayer. Therefore, the isotherm for 0.8
ML was normalized, considering that only 80% of the Au electrode
surface was covered by Pd. Since there was obviously no absorp-
tion at 0.8 ML and just adsorption, isotherm at this thickness can
directly be compared with the adsorption isotherm at nanomet-
ric Pd/Au(polycrystal) films. This comparison confirms what was
observed in cyclic voltammetry, i.e. adsorption starts at more pos-
itive potentials at Pd/Au(1 1 1) monolayer deposits. This is because
the first monolayers have different properties than bulk Pd. The
substrate has multiple effects on H adsorption and absorption in
Pd. First, as the Pd deposition on gold is pseudomorphic [25], Pd
deposits have the same orientation as gold electrode, i.e. (1 1 1). It
has been demonstrated that specific crystallographic orientation
could strongly modify the electrochemical behavior of electrodes
[27]. Moreover, the interatomic distances at the first layer are differ-
ent. The interatomic distances in bulk Pd(1 1 1) is 2.75 Å while they
are 2.88 Å for Pd on Au(1 1 1) [51]. Therefore, stress appears in the
first Pd layers. These two phenomena can be grouped as the geomet-
ric effect of the substrate. Moreover, Sellidj and Koel [52] showed
that the work function of Pd monolayers deposited on Au(1 1 1) is
influenced by the substrate up to 5 monolayers. Fermi level is also
Fig. 3. Successively measured isotherms at a 7 ML Pd/Au(1 1 1) deposit, 1st isotherm
(ꢀ) between 420and 60 mV, 2nd isotherm ( )between 320 and 60 mV, 3rd isotherm
(
) between 220 and 60 mV, and the final isotherm reconstructed from the three
series (in continuous line).
the deposit’s morphology. Focusing on the peaks at more positive
potentials (C3, C2, A3 and A2), it is obvious that they decrease with
the number of cycles, with the exception of C3, which is increas-
ing. Therefore, as the hydrogen adsorption seems to be blocked
peak C3 cannot be ascribed to the adsorption process. We there-
fore considered the hypothesis of surface contamination by traces
impurities present in alkaline solution, and referred C3 to a contam-
inant adsorption. This contamination seems to act as an adsorption
blocker and an absorption promoter, i.e. the same effect as that of
BTA (see below) or crystal violet [13,29].
Fig. 3 is an example of the surface contamination illustrated
in the case of determination of the H sorption isotherms at 7 ML
Pd/Au(1 1 1). The first isotherm determination was carried out after
the same pretreatment in H2SO4 described in the last paragraph.
After the deposit was considered as stable in acid, i.e. there were no
changes in the voltammograms and results were similar to those
in the literature [13,14], it was transferred into NaOH, and the mea-
surements were carried out in the following way: polarization at
420 mV during 30 s, electrochemical impedance spectroscopy (EIS)
measurements from 20 kHz to 20 mHz, discharge at 650 mV (in the
non-faradaic potential zone), cyclic voltammetry at 200 mV s−1
,
polarization at 400 mV during 30 s, and so on every 20 mV until
E = 60 mV. These experiments took around 2 h 30 min. Similar series
of experiments but without impedance measurements gave repro-
ducible results, in a shorter experiment time (around 40 min).
The first setup with full impedance measurements gave non-
reproducible results with isotherms systematically lower than what
was obtained during the experiments without impedance measure-
ments. The fact that longer polarization times due to impedance
measurements seemed to drastically decrease the H sorption
charge confirmed again the hypothesis of surface contamination.
Our first idea to counter this problem was to decrease the frequency
range to reduce the duration of our experiments. However, this was
not sufficient to obtain reproducible results. Because the problem of
contamination does not exist in acid media, we tried to “clean” our
surface in acid solution after the first isotherm by carrying out the
same pretreatment as before in H2SO4. After 3 cyclic voltammetric
polarization sweeps in H2SO4 the same cyclic voltammetric pro-
file as that obtained during the first pretreatment and the same H
sorption charge at 100 mV were found. Following this reactivation
step, the Pd deposit was transferred back into NaOH and the same
polarization/impedance/discharge/voltammetry setup was carried
Fig. 4. Total H sorption isotherms at 0.8 ML (*), 2 ML (ꢀ), 3 ML ( ), 5 ML ( ), 7 ML
(
) and 10 ML ( ) Pd/Au(1 1 1) deposits, and H adsorption charge density ( ) at
Pd/Au(polycrystal).

M.H. Martin, A. Lasia / Electrochimica Acta 54 (2009) 5292–5299
5295
modified [27,53,54]. These electronic effects can drastically modify
adsorption properties of Pd deposits. Of course these two effects
are interrelated.
The fact that isotherms are independent of thickness at poten-
tials above 320 mV demonstrates that adsorption is predominant
in this range of potentials while the absorption is practically inex-
istent. Below 320 mV, absorption becomes a significant part of the
total H charge as it increases with the thickness. Below 120 mV,
the charge strongly increases for all deposits except 0.8 ML. This
phenomenon appears at potentials still too high to be related to H
evolution, therefore, it must be attributed to absorption. This is in
agreement with the fact that such a behavior was not visible at 0.8
ML. However, the increase below 120 mV at 2 ML is almost as strong
as for thicker deposits, even if Baldauf and Kolb [22] demonstrated
that at least 3 monolayers are normally required for absorption. This
confirms the hypothesis of a 3-dimensional growth during deposi-
tion, i.e. the third layer (and higher) starts to be formed before the
completion of the second layer [22,55]. Fig. 4 also clearly shows
that the increase in the H charge with the thickness between 0.8
and 5 ML is much more important than between 5 and 10 ML. This
is probably due to the substrate influence, which becomes less and
less important with the increase in the Pd layer thickness.
Electrochemical impedance spectroscopy was also used to
study Pd monolayers using the setup described above, i.e. a first
sequence from 420 to 60 mV, cleaning in H2SO4, new sequence
in NaOH from 320 to 60 mV, cleaning in H2SO4, and the last
series of measurements from 220 to 60 mV in NaOH. As for the
isotherms measurements, the impedance parameters were recon-
structed from these three sequences. Due to the multiple electrode
treatments in different electrolytes and the poor stability of 0.8
ML Pd/Au(1 1 1) deposit, no reliable impedance results could be
obtained for this thickness.
Fig. 6. Double layer capacitance, Cdl
Pd/Au(polycrystal) in 0.1 M NaOH.
,
at 2–17 ML Pd/Au(1 1 1) and 70 nm
series. However, because our experiments were limited to higher
frequencies, the complex plane plots only exhibited the beginning
of the semi-circle, and the capacitive wall normally observed in the
range of 0.1–0.01 Hz was never visible. Thus, the equivalent circuit
used to fit our results was just a solution resistance Rs in series with
a two branches circuit composed by the double layer capacitance
Cdl in parallel with the charge transfer resistance Rct. To take into
account the non-homogeneity of the double layer capacitance, a
constant phase element (CPE) was used for the double layer capac-
itance [56,57]. The average value of the capacitance was estimated
from the parameters T and ꢀ using equation [56]:
Because very short experimental times were required to counter
the electrode contamination and to obtain reproducible results, our
frequency range was limited to 20 kHz–2 Hz with five points per
decade. Fig. 5 shows examples of complex plane plots for H sorp-
tion at 7 ML Pd/Au(1 1 1) deposit, at four different potentials, with
their complex nonlinear least-squares fits. The equivalent circuit
used to fit our results was derived from the one presented in our
earlier works [13,14,18,29] and consisted of a solution resistance in
series with a double layer capacitance, which is in parallel with the
charge transfer resistance-low frequency capacitance connection in
ꢀ
ꢁ
1−1/ꢀ
1
Rs
1
Rct
Cdl = T1/ꢀ
¯
+
Fitting data to this model gave us maximal standard deviation of
1.5% for Rct for potentials above 340 mV, and around 0.7% for poten-
tials between 340 and 60 mV. This error is acceptable taking into
account sensitivity of this parameter to contamination.
The potential dependence of the double layer capacitance Cdl
for 70 nm Pd/Au(polycrystal) and Pd/Au(1 1 1) monolayer deposits
is displayed in Fig. 6. At monolayer deposits, the values of
Cdl in the potential range between 260 and 400 mV seem to
depend on the deposit thickness. Potential dependence of Cdl at
10 and 17 ML Pd/Au(1 1 1) deposits are quite similar to that at
70 nm Pd/Au(polycrystal), with a shift of 20 mV. For nanometric
Pd/Au(polycrystal) deposits, Cdl curves were similar at all thickness
between 10 and 90 nm. This evolution in the Cdl potential depen-
dence could be a consequence of the substrate effect. It has also
been noticed that the substrate greatly influenced sulfate adsorp-
tion at Pd in H2SO4 [30]. The same phenomenon seems to exist in
NaOH with OH adsorption.
Fig. 7 shows the potential dependence of the charge transfer
resistance, Rct, for Pd/Au(1 1 1) deposits, with thicknesses between
2 and 17 ML. For comparison, Rct at 70 nm Pd/Au(polycrystal) is also
displayed. Fig. 7 confirms that H absorption is faster at Pd/Au(1 1 1)
monolayers. Indeed, depending on the thickness, Rct is between 2
and 8 times lower at Pd/Au(1 1 1) monolayers.
The isotherms in Fig. 4 showed that H absorption appears for
all deposits thicker than 1 ML due to 3-dimensional growth. The
simulations published earlier [29] and the experimental results
demonstrate that log Rct vs. potential curves should have a W-shape
when adsorption and absorption processes coexist, with kinetic
parameters of the same order of magnitude [18,29]. In the case
when one of these reactions is much faster than the other, log Rct
vs. potential curves exhibit a single V-shape related to the fastest
Fig. 5. Experimental complex plane plots (dot) and fitting (line) at a 7 monolayers
Pd/Au(1 1 1) deposit, at four different potentials.

5296
M.H. Martin, A. Lasia / Electrochimica Acta 54 (2009) 5292–5299
Fig. 7. Charge transfer resistance Rct at 2–17 ML Pd/Au(1 1 1) in 0.1 M NaOH.
Fig. 9. Cyclic voltammograms at 2–10 ML Pd/Au(1 1 1) deposit, in 0.1 M
NaOH + 10 mM BTA; scan rate: 10 mV s−1
.
reaction, since adsorption and direct absorption reactions compete
in parallel. Focusing on 2, 3 and 5 monolayers, a single V-shape
attributed to H adsorption process is observed. Thus, it can be con-
cluded that adsorption at Pd/Au(1 1 1) deposits thinner than 5 ML
is faster than absorption between 400 and 60 mV. The fact that a
surface property (Rct of adsorption reaction) is thickness depen-
dent strongly suggests that the kinetic promotion of adsorption is
caused by the substrate effect. For thicker deposits, the W-shape
is clearly visible. The rates of adsorption and absorption are of the
same order of magnitude. Moreover, the evolution of the shape of
Rct vs. E with thickness leads to the conclusion that the kinetic pro-
motion of adsorption for deposits thinner than 5 ML is strongly
linked to the effects of the substrate. We have already demon-
strated that the substrate effects can also strongly interfere with
the thickness dependence of Rct. While Rct decreases with thick-
ness at Pd/Au(1 1 1) monolayer deposits in H2SO4 [29], it increases
at Pd/Pt(1 1 1) deposits [30]. In NaOH Rct increases with thick-
ness at Pd/Au(1 1 1) nanolayers while it decreases at nanometric
Pd/Au(polycrystal) deposits [18].
ments and impedance spectroscopy. It was concluded that the
surface was saturated by BTA at 10 mM. The protective effect in
the oxidation potential range, i.e. at potentials more positive than
500 mV, is clearly visible at both deposits. At potentials lower than
500 mV, adsorption is at least partially inhibited while absorption
is promoted.
At 70 nm Pd/Au(polycrystal) only large absorption/desorption
peaks are visible in the presence of BTA. They are shifted to more
negative potentials and are more reversible than those observed in
the absence of BTA. However, at thin layers of Pd/Au(1 1 1) a new
sharp pair of peaks around 250 mV appears on the cyclic voltam-
mograms followed, at more negative potentials, by a pair of peaks
due to H absorption/desorption. The main effect responsible for the
absorption promotion observed on cyclic voltammograms (peaks
at more negative potentials) is kinetic since the increase in the
absorption/desorption reaction reversibility is very important. This
is confirmed by impedance measurements (see later).
Fig. 9 displays voltammograms at 2–10 ML Pd/Au(1 1 1) in
NaOH + BTA. The thickness dependence of the H absorption current
below 150 mV is as expected, i.e. it increases with the thickness.
More interesting is the analysis of the cathodic and anodic peaks
around 250 mV. The cathodic peak is totally independent of the
thickness, suggesting that it represents a surface phenomenon. This
suggests a change in the BTA layer, which would start at 280 mV
(foot of the peak). BTA total desorption is improbable because an
effect on H absorption at more negative potentials can also be seen.
More probable is a reorientation of the BTA layer at the Pd surface
followed by hydrogen adsorption. BTA reorientation was already
observed on copper in the Cu oxidation potential range using FTIR
and STM [58,59]. The behavior of the anodic peak is more unusual.
The position and the shape of the peak, and thus the reversibil-
ity of the reaction, strongly depend on the thickness up to 5 ML,
increasing with the Pd layer thickness. Substrate effects can be
pointed as the main reason for this effect. It should be added that
in sulfuric acid a shift of hydrogen adsorption peaks into positive
direction upon addition of BTA was also observed [22] with the final
voltammogram similar to that observed in NaOH.
3.2. Influence of the BTA on the H sorption in Pd
Fig. 8 presents the voltammograms for 70 nm Pd/Au(polycrystal)
and 7 ML Pd/Au(1 1 1) deposits in 0.1 M NaOH, without and with
10 mM BTA. At this BTA concentration no effect of further addition
of BTA was observed on cyclic voltammetry, H isotherm measure-
EIS can provide interesting data to confirm the hypothesis
of a modification in the BTA layer. However, classical stationary
impedance was not adequate to confirm the kinetic difference
between the adsorption and desorption peaks. Thus, we completed
our study with dynamic electrochemical impedance spectroscopy
(DEIS) where sum of sinusoidal perturbations was applied during
the potential sweep followed by the FFT analysis of the potential and
current signals to obtain the impedance spectra during the sweep
and from the modeling of the impedance data the charge transfer
Fig. 8. Cyclic voltammograms at 70 nm Pd/Au(polycrystal) and 7 ML Pd/Au(1 1 1)
deposit, in 0.1 M NaOH (—) and 0.1 M NaOH + 10 mM BTA (- - -); scan rate: 10 mV s−1
.

M.H. Martin, A. Lasia / Electrochimica Acta 54 (2009) 5292–5299
5297
Fig. 10. Double layer capacitance Cdl at 70 nm Pd/Au(polycrystal) and 2–10 ML
Pd/Au(1 1 1) deposit, in 0.1 M NaOH +10 mM BTA obtained using stationary and
dynamic (DEIS) electrochemical impedance spectroscopy.
resistance and the double layer capacitance. The equivalent circuit
used to fit our results was the same as described above. Since BTA
also protects Pd from contamination, it was possible to work at
lower frequencies and observe the low frequency capacitive behav-
ior. However, the used sweep rate was 2 mV s−1 what limited the
lowest frequency at 2 Hz.
Fig. 10 displays the potential dependence of the average dou-
ble layer capacitance at 70 nm Pd/Au(polycrystal) film and at 2–10
monolayers at Au(1 1 1) in NaOH with BTA. It clearly exhibits the
typical Cdl potential evolution characteristic for the specific adsorp-
tion of an organic molecule at the Pd surface with both techniques.
This adsorption decreases the double layer capacitance at the most
positive potentials. Below 180 mV at the 70 nm Pd film and 280 mV
at Pd monolayers, Cdl slowly increases until reaching a value around
20 ␮F cm−2. This value is similar to what was obtained without BTA
at negative potentials. Moreover, the increase starts at potentials
where the cathodic peak is observed at monolayers in Fig. 9. This is
relevant to the hypothesis of a change in the BTA layer orientation.
However, a sharp peak in Cdl is expected for such a phenomenon.
This peak cannot be seen when focusing on the average Cdl, Fig. 10.
The parallel observation of T and ꢀ parameters at Pd/Au(1 1 1)
monolayers gives more information in Fig. 11. The dependence of T
parameter displays the expected peak at the potentials correspond-
ing to the voltammetric peak. At the same potential, ꢀ strongly
decreased. This diminution of ꢀ compensates the increase of T in
the calculation of Cdl, see equation above. This explains why no peak
is observed in Fig. 10. The obtained data suggest that analysis of the
CPE parameters T and ꢀ instead of the average Cdl is more sensi-
tive in the case of specific adsorption. The peaks are observed with
the stationary EIS and DEIS. However, DEIS presents two peaks:
a small one corresponding to the cathodic part of the voltammo-
gram (potentials decreasing from 420 to 80 mV), and an intense
one for the anodic part (potentials increasing from 80 to 420 mV).
These peaks perfectly match with those observed in Fig. 9. The peak
observed with EIS seems to be the mean of the two peaks obtained
by DEIS.
Fig. 11. T and ꢀ parameters obtained by EIS and DEIS, at 2–10 ML Pd/Au(1 1 1), in
0.1 M NaOH + 10 mM BTA. The arrows indicate the direction of the potential sweep
in DEIS.
als [13,29,31]. BTA addition strongly decreases the influence of the
thickness on Rct since no clear tendency was observed at Pd mono-
layers and nanometric films, and the values of Rct for both types of
deposits are of the same order of magnitude. Rct values obtained
from DEIS experiments at 3 ML Pd exhibit an important differ-
ence between adsorption and desorption reactions. The difference
between the forward (decreasing potentials) and the backward
(increasing potentials) runs is important at this deposit, while both
are quite similar at 10 ML. This confirms what was observed with
cyclic voltammetry in Fig. 9. In parallel with Rct is shown the CLF
potential dependence. CLF at Pd monolayers exhibits two distinc-
tive domains: above 240 mV, H adsorption is mainly blocked by
BTA; below 240 mV, H adsorption can be achieved since a change
in the adsorbed BTA layer occurs (see voltammograms in Fig. 9 and
double layer capacitance parameters in Fig. 11). A peak of CLF at
250 mV is also visible suggesting hydrogen adsorption.
Fig. 13 displays the total H sorption isotherms and the contri-
butions of individual processes at 30 nm Pd deposit in 0.1 M NaOH
with and without 10 mM BTA. In our previous work on Pd deposits
in NaOH [18], contributions to the adsorption and absorption were
estimated at this Pd film. The comparison between the isotherm
with BTA and the adsorbed H charge suggests that BTA blocks H
adsorption. However, the H charge In the presence of BTA is larger
than that due to H absorption only. That confirms that BTA inhibits
H adsorption only partially.
Fig. 12 presents the potential dependence of the charge trans-
fer resistance Rct and the low frequency capacitance CLF at
70 nm Pd/Au(polycrystal) 2–10 ML Pd/Au(1 1 1) deposits, in 0.1 M
NaOH + 10 mM BTA. By comparing Fig. 7 and Fig. 12, it is clear
that BTA decreased Rct by more than one order of magnitude, and
thus can be considered as a kinetic promoter for H sorption. Such
effect was also observed with other poisons and electrode materi-
In Fig. 14, the influence of BTA on H adsorption and absorption at
2, 5 and 10 Pd ML is demonstrated. The effect on ultra thin deposits

5298
M.H. Martin, A. Lasia / Electrochimica Acta 54 (2009) 5292–5299
Fig. 14. Total H sorption charge at 2 ML (ꢀ), 5 ML ( ) and 10 ML ( ) Pd/Au(1 1 1)
deposits, in 0.1 M NaOH (filled symbols) and 0.1 M NaOH + 10 mM BTA (open sym-
bols).
is similar to what is observed at nanometric films, Fig. 13. Since no
influence of the thickness can be observed at 2–10 ML Pd/Au(1 1 1)
deposits in NaOH + BTA between 240 and 280 mV, the hypothesis
of residual adsorption seems to be confirmed. Once more, 240 mV
seems to be a transition point, suggesting a rearrangement of the
BTA layer.
Results shown in Figs. 13 and 14 indicate that BTA inhibits par-
tially hydrogen adsorption but does not affect hydrogen absorption
into metal. This behavior is similar to that found in acid solutions in
the presence of another poison, crystal violet [13,14]. These results
clearly indicate that hydrogen absorption in the presence of poi-
sons must proceed through the direct mechanism. In the case of
the indirect mechanism hydrogen adsorption and absorption are
related according to: ꢁH,ads = ꢁH,abs, that is the chemical potential
of adsorbed hydrogen equals to that of absorbed one and decrease
of the surface coverage by adsorbed hydrogen leads to decrease of
ꢁH,ads and decrease of the quantity of absorbed hydrogen, what is
not experimentally observed.
Fig. 12. Charge transfer resistance Rct and low frequency capacitance CLF at 70 nm
Pd/Au(polycrystal) deposit, and 2–10 ML Pd/Au(1 1 1) deposits, in 0.1 M NaOH
+10 mM BTA. The arrows indicate the direction of the potential sweep in DEIS.
4. Conclusions
Hydrogen adsorption and absorption at Pd monolayers
deposited on Au(1 1 1) have been studied in the absence and in the
presence of benzotriazole. The following conclusions can be drawn:
1. H adsorption at Pd/Au(1 1 1) monolayers starts at more positive
potentials than at nanometric Pd/Au(polycrystal) films.
2. H adsorption and absorption at Pd/Au(1 1 1) monolayers are more
reversible than at nanometric Pd/Au(polycrystal) films.
3. These two observations are a consequence of the strong substrate
influence due to geometric and electronic effects.
4. BTA blocks partially adsorption and promotes kinetically absorp-
tion.
5. There is a modification in the adsorbed BTA layer orientation
around 240 mV at Pd/Au(1 1 1) monolayers.
6. Gold substrate also influences H sorption at Pd monolayers in
NaOH + BTA up to 5 monolayers.
7. The results obtained in the presence of poison point out to the
direct H absorption mechanism.
Acknowledgements
Fig. 13. H sorption isotherms at a 30 nm Pd/Au(polycrystal) deposit, with a real
surface area of 0.79 cm2; total H charge (ꢀ), adsorbed H charge ( ), absorbed H
charge ( ) in 0.1 M NaOH, total H charge in 0.1 M NaOH + 10 mM BTA (ꢁ).
Financial support from CRSNG (Canada) is gratefully acknowl-
edged.

M.H. Martin, A. Lasia / Electrochimica Acta 54 (2009) 5292–5299
5299
The authors acknowledge Prof. D. Harrington (University of Vic-
toria) for supplying the program for data acquisition/analysis and
Daniel Auger (Université de Sherbrooke) for constructing the elec-
tronics for DEIS.
[28] R. Hoyer, L.A. Kibler, D.M. Kolb, Surf. Sci. 562 (2004) 275.
[29] H. Duncan, A. Lasia, Electrochim. Acta 52 (2007) 6195.
[30] H. Duncan, A. Lasia, J. Electroanal. Chem. 621 (2008) 62.
[31] B.E. Conway, J.H. Barber, L. Gao, S.Y. Qian, J. Alloys Compd. 253–254 (1997) 475.
[32] S.Y. Qian, B.E. Conway, G. Jerkiewicz, J. Chem. Soc., Farad. Trans. 94 (1998) 2945.
[33] L. Birry, Ph.D. Thesis, Université de Sherbrooke, 2005.
[34] J.N. Han, S.I. Pyun, T.H. Yang, J. Electrochem. Soc. 144 (1997) 4266.
[35] M.W. Breiter, J. Electroanal. Chem. 109 (1980) 243.
References
[36] M.W. Breiter, J. Electroanal. Chem. 109 (1980) 253.
[37] M.W. Breiter, J. Electroanal. Chem. 180 (1984) 25.
[1] F.A. Lewis, The Palladium Hydrogen Systems, Academic Press, London, 1967.
[2] G. Alefeld, J. Völkl (Eds.), Hydrogen in Metals, vols. I and II, Springer-Verlag,
Berlin, 1978.
[3] M. Baldauf, D.M. Kolb, J. Phys. Chem. 100 (1996) 11375.
[4] S. Pandelov, U. Stimming, Electrochim. Acta 52 (2007) 5548.
[5] A. Czerwinski, Electrochim. Acta 39 (1994) 431.
[38] A. Czerwinski, J. Electroanal. Chem. 379 (1994) 487.
[39] A. Czerwinski, S. Zamponi, R. Marassi, J. Electroanal. Chem. 304 (1991) 233.
[40] B. Łosiewicz, L. Birry, A. Lasia, J. Electroanal. Chem. 611 (2007) 26.
[41] P.N. Bartlett, B. Gollas, S. Guerin, J. Marwan, Phys. Chem. Chem. Phys. 4 (2002)
3835.
[42] P.N. Bartlett, J. Marwan, Phys. Chem. Chem. Phys. 6 (2004) 2895.
[43] M.H.A. Elhamid, B.G. Ateya, H.W. Pickering, J. Electrochem. Soc. 144 (1997) L58.
[44] N. Amokrane, C. Gabrielli, G. Maurin, L. Mirkova, Electrochim. Acta 53 (2007)
1962.
[45] S. Szpak, P.A. Mosier-Boss, S.R. Scharber, J.J. Smith, J. Electroanal. Chem. 337
(1992) 147.
[46] A. Hamelin, B.E. Conway, J.O’M. Bockris, in: R.E. White (Ed.), Modern Aspects of
Electrochemistry, vol. 16, Kluwer/Plenum, New York, 1985, p. 1.
[47] D.F. Yang, C.P. Wilde, M. Morin, Langmuir 12 (1996) 6570.
[48] G.S. Popkirov, R.N. Schindler, Rev. Sci. Instr. 64 (1993) 3111.
[49] A.E. Bolzan, J. Electroanal. Chem. 380 (1995) 127.
[6] A. Czerwinski, R. Marassi, J. Electroanal. Chem. 322 (1992) 373.
[7] A. Czerwinski, R. Marassi, S. Zamponi, J. Electroanal. Chem. 316 (1991) 211.
[8] A. Czerwinski, Polish. J. Chem. 69 (1995) 699.
[9] A. Czerwinski, G. Maruszczak, M. Zelazowska, M. Lancucka, R. Marassi, S. Zam-
poni, J. Electroanal. Chem. 386 (1995) 207.
[10] A. Czerwinski, M. Czauderna, G. Maruszczak, I. Kiersztyn, R. Marassi, S. Zamponi,
Electrochim. Acta 42 (1997) 81.
[11] A. Czerwinski, I. Kiersztyn, M. Grden, J. Czapla, J. Electroanal. Chem. 471 (1999)
190.
[12] A. Czerwinski, I. Kiersztyn, M.S. Grden, J. Electroanal. Chem. 492 (2000) 128.
[13] L. Birry, A. Lasia, Electrochim. Acta 51 (2006) 3356.
[50] G. Denuault, C. Milhano, D. Pletcher, Phys. Chem. Chem. Phys. 7 (2005) 3545.
[51] A.M. El-Aziz, R. Hoyer, L.A. Kibler, D.M. Kolb, Electrochim. Acta 51 (2006) 2518.
[52] A. Sellidj, B.E. Koel, Phys. Rev. B 49 (1994) 8367.
[53] B. Hammer, J.K. Nørskov, Surf. Sci. 343 (1995) 211.
[54] A. Ruban, B. Hammer, P. Stolze, H.L. Skriver, J.K. Nørskov, J. Mol. Catal A 115
(1997) 421.
[14] H. Duncan, A. Lasia, Electrochim. Acta 53 (2008) 6845.
[15] C. Gabrielli, P.P. Grand, A. Lasia, H. Perrot, J. Electrochem. Soc. 151 (2004) A1937.
[16] C. Gabrielli, P.P. Grand, A. Lasia, H. Perrot, J. Electrochem. Soc. 151 (2004) A1943.
[17] C. Gabrielli, P.P. Grand, A. Lasia, H. Perrot, J. Electrochem. Soc. 151 (2004) A1925.
[18] M.H. Martin, A. Lasia, Electrochim. Acta 53 (2008) 6317.
[19] J. Paillier, L. Roué, J. Electrochem. Soc. 152 (2005) E1.
[55] C. Lebouin, Ph.D. Thesis, INP Grenoble, 2008.
[56] G.J. Brug, A.L.G. van den Eeden, M. Sluyters-Rehbach, J.H. Sluyters, J. Electroanal.
Chem. 176 (1984) 275.
[57] A. Lasia, B.E. Conway, J.O’M. Bockris, in: R.E. White (Ed.), Modern Aspects of
Electrochemistry, vol. 32, Kluwer/Plenum, New York, 2002, p. 143.
[58] M.R. Vogt, R.J. Nichols, O.M. Magnussen, R.J. Behm, J. Phys. Chem. B 102 (1998)
5859.
[20] J. Clavilier, J.-P. Ganon, M. Petit, J. Electroanal. Chem. 265 (1989) 231.
[21] G.A. Attard, A. Bannister, J. Electroanal. Chem. 300 (1991) 467.
[22] M. Baldauf, D.M. Kolb, Electrochim. Acta 38 (1993) 2145.
[23] L.A. Kibler, A.M. El-Aziz, D.M. Kolb, J. Mol. Catal. A: Chem. 199 (2003) 57.
[24] L.A. Kibler, M. Kleinert, R. Randler, D.M. Kolb, Surf. Sci. 443 (1999) 19.
[25] J. Tang, M. Petri, L.A. Kibler, D.M. Kolb, Electrochim. Acta 51 (2005) 125.
[26] R. Hoyer, L.A. Kibler, D.M. Kolb, Electrochim. Acta 49 (2003) 63.
[27] L.A. Kibler, A.M. El-Aziz, R. Hoyer, D.M. Kolb, Ang. Chem. Int. Ed. 44 (2005) 2080.
[59] Z.D. Schultz, M.E. Biggin, J.O. White, A.A. Gewirth, Anal. Chem. 76 (2004) 604.