Electrocatalytic activation of Ni for H2 evolution by spontaneous deposition of Ru

Article abstract of DOI:10.1016/j.chemphys.2005.06.042

Ni was activated by spontaneous deposition of Ru carried out in solutions of HCl containing RuCl₃. The occurrence of Ru deposition was confirmed by EDX and XPS and followed as a function of deposition time by cyclic voltammetry and Tafel lines

Full text of DOI:10.1016/j.chemphys.2005.06.042

Chemical Physics 319 (2005) 192–199
www.elsevier.com/locate/chemphys
Electrocatalytic activation of Ni for H evolution by
2
q
spontaneous deposition of Ru
I. Bianchi, E. Guerrini, S. Trasatti *
Department of Physical Chemistry and Electrochemistry, University of Milan, Via Venezian 21, 20133 Milan, Italy
Received 14 November 2004; accepted 13 June 2005
Available online 3 August 2005
Abstract
Ni was activated by spontaneous deposition of Ru carried out in solutions of HCl containing RuCl
3
. The occurrence of Ru
deposition was confirmed by EDX and XPS and followed as a function of deposition time by cyclic voltammetry and Tafel lines
ꢀ3
for H evolution in 1 mol dm KOH. Experiments were performed by varying temperature, HCl as well as RuCl concentration.
2
3
The criticity of the presence of O
2
in solution was demonstrated. Ni activation by Ru is manifested by the appearance of H
adsorption peaks in the voltammetric curve, and by the decrease of the Tafel slope for H
Ni to ca. 40 mV.
2
evolution from a 120 mV for bare
ꢀ
2005 Elsevier B.V. All rights reserved.
Keywords: Electrocatalysis; H evolution; Spontaneous deposition; Ru overlayer; Ni electrode
2
1
. Introduction
Ni is the most used electrode material for cathodes in
by direct thermal decomposition [4], or by electrodepos-
ition of preformed RuO particles from a Ni bath [5]. In
2
both cases activation is excellent, but on the whole there
are residual difficulties. In the former case (thermal
decomposition), the mechanical stability of the over-
layer can lead to loss of material by erosion due to the
evolving gas. In the latter case, the preparation of the
activated electrode involves consumption of electrical
energy which affects negatively the economic balance
of the operation.
alkaline solutions [1]. Although its electrocatalytic activ-
ity is not excellent, the metal is resistant to corrosion at
high pH values. There are, however, important prob-
lems of electrocatalytic stability. Besides exhibiting a
Tafel slope for H evolution close to 120 mV, the over-
2
potential drifts with time, especially in conditions of
intermittent electrolysis as it is the case with electrolysers
driven by solar energy conversion (solar panels) [2].
For the above reasons much research work has been
addressed to the activation of Ni cathodes [3].
In previous papers, we have shown that Ni can be
activated substantially by depositing on its surface thin
layers of catalytic oxide electrodes, mainly RuO , either
Scrutiny of the activity of metals for H evolution
2
shows that Ru metal is among the best electrocatalysts
for this reaction [3,6]. In situ activation of Ni by Ru elec-
trodeposition was demonstrated by Schmidt and Wendt
[7] during H2 evolution from 40% KOH at 120 ꢁC.
Moreover, Wieckowski and co-workers [8–10] have
shown that Ru can be spontaneously deposited on Pt
from an acidic solution of its chloride. Along the same
lines, we have investigated the possibility of spontaneous
deposition of Ru on Ni electrodes. Since Ni is attacked
by acids whereas it is passivated by alkalis, it was
decided to deposit Ru on Ni in an acidic solution and
2
q
In honour of Professor A.M. Kuznetsov on the occasion of his 65th
birthday.
*
Corresponding author. Tel.: +39 02 503 14223; fax: +39 02 503
14224.
E-mail address: sergio.trasatti@unimi.it (S. Trasatti).
0
301-0104/$ - see front matter ꢀ 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.chemphys.2005.06.042

I. Bianchi et al. / Chemical Physics 319 (2005) 192–199
193
to test the resulting modified surface for H evolution in
2.5. Non-electrochemical techniques
2
alkaline solution.
The morphology of the samples was observed by
means of a model LEO1430 scanning electron micro-
scope (SEM), equipped with an Oxford INCA 2005
microprobe used to detect the distribution of Ru on
the surface of the support.
2
. Experimental
Experiments consisted of two steps: (i) preparation of
Ru-activated Ni, and (ii) test of the electrodes for H2
evolution.
The surface chemical composition of the samples
after the deposition of Ru was analysed by X-ray photo-
electron spectroscopy (XPS) using a M-probe (Surface
Science Instruments, Mountain View, CA). Details
about the XPS instrument and the procedure of analysis
can be found elsewhere [11].
2
.1. Ni support
About 0.2 · 10 cm Ni platelets were cut from a
.25 mm thick foil. Samples were first smoothed espe-
0
cially at the edges by means of very fine emery paper,
then etched in boiling 20% HCl. After careful rinsing
and drying, only 1 cm total surface area was left open,
3. Results
2
while the rest of the samples was carefully isolated by
means of Teflon strips. Electrodes were used immedi-
ately after preparation to avoid self-passivation in open
air.
3.1. Ni surface
Bare Ni was taken as a reference to assess the entity
of activation. Two main ‘‘indicators’’ were scrutinized:
(
i) CV curve and (ii) Tafel slope for H evolution.
Fig. 1(a) shows the CV curve at 100 mV s for bare
2
ꢀ
1
2
.2. Solutions
Ni. On the anodic cycle there are the two characteristic,
irreversible peaks associated with the Ni(II)/Ni(III)
Ru deposition was carried out by immersing the pre-
treated electrodes in de-aerated acid solution (HCl) con-
transition [12–14]. On the cathodic side H evolution,
2
taining a concentration of RuCl Æ xH O (Aldrich)
3
2
whose magnitude was dictated by the need to avoid fast
multilayer deposition of the metal thus hiding the details
of the process. Unless otherwise specified, experiments
ꢀ3
ꢀ3
a
were carried out with 10 mol dm of Ru(III).
10
Solutions of RuCl were prepared by dissolving the
3
5
0
salt in little water, acidifying with HCl and waiting a
few hours before diluting to the desired concentration.
The solution was used after about a week ageing (see la-
ter Section 3.4.4).
-1000
-500
0
500
-5
10
15
-
-
ꢀ
3
Activated electrodes were tested in 1 mol dm KOH
Aldrich) by means of cyclic voltammetry (CV) and
(
quasi-stationary potentiostatic curves.
-20
During spontaneous deposition the solution was not
stirred and the electrode was not moved. During electro-
chemical experiments the solution was stirred by bub-
b
E/mV
-
-
-
-
1.2
bling purified N . Unless otherwise specified, the
2
temperature of the solution was kept at 25 ± 0.1 ꢁC by
means of a water thermostat.
1.3
1.4
1.5
2
.3. Reference electrode
In all experiments, the reference electrode was a com-
mercial saturated calomel electrode (SCE).
-
1
-0.5
0
0.5
1
1.5
2
-
2
2
.4. Electrochemical instrumentation
log(j/mA cm )
Fig. 1. (a) Cyclic voltammetry of Ni over the whole potential window
Electrochemical experiments were carried out using a
model 273A EG&G Potentiostat/Galvanostat driven by
a PC equipped with the software M270 v. 4.11.
ꢀ3
from H
2
ꢀ
evolution to incipient O
2
evolution: 1 mol dm
00 mV s . (b) Typical Tafel line for H
slope = 117 mV.
KOH,
1
1
2
evolution on bare Ni. Tafel

194
I. Bianchi et al. / Chemical Physics 319 (2005) 192–199
the only visible feature, commences at about ꢀ1.3 V
sition of the solution, if the reaction is carried out in a
test tube, Ru deposition occurs even in the presence of
O . If the reaction is carried out in an open becker no
2
(
SCE) while Erev ꢁ ꢀ1020 mV.
The extent of the roughness of the pre-treated Ni sup-
port was estimated by recording CV curves at several
scan rates (v) in a restricted potential range (0.1 V) away
from apparent faradic processes. The slope of j vs. v at a
given E gives the capacitance. Ni samples were prepared
with different surface pre-treatments: mechanical polish-
ing, emery paper, HCl etching, sandblasting. CV curves
showed that roughness increased in the order: mechani-
cal polishing < emery paper < chemical etching < sand-
blasting. If the roughness factor is taken as 1 for
mechanical polishing (mirror finishing by means of alu-
mina), the factor becomes ca. 1.7 for etching and ca. 4
for sandblasting.
Ru deposition is obtained unless the solution is deaer-
ated. Evidently, in a test tube where the access of air is
difficult, Ru deposition can occur after all dissolved O2
is consumed by cathodic reduction. Therefore, the con-
trol of the atmosphere is essential for the reproducibility
of the deposition conditions.
3.3. Evidence for Ru deposition
In preliminary experiments, after depositing Ru,
samples were subject to cyclic voltammetry (CV) and
quasi-stationary cathodic polarization to record Tafel
Fig. 1(b) shows the Tafel plot for H evolution on the
plots for H evolution. Experiments showed the pres-
2
2
same sample. A Tafel line, whose slope is close to
ence of Ru (see later on), provided the CV potential
range did not reach the region of Ni(II)/Ni(III) transi-
tion (see previous section). In the latter case the evi-
dence for the presence of Ru faded with increasing
number of cycles. This implies that Ru is slowly dis-
solved during anodic oxidation. For this reason, all
successive CV curves were confined to the potential
range ꢀ1.15–0 V (SCE).
1
20 mV, can be identified unambiguously [3,15–17].
The deviations visible at high current densities are en-
tirely attributable to uncompensated ohmic drops, as
confirmed by an analysis of the data based on a plot
of DE/Dj vs. 1/j [18].
Fig. 2 shows the variation of the open circuit poten-
ꢀ3
tial (OCP) of Ni after immersion in 0.1 mol dm HCl.
The effect of the atmosphere is striking. In the presence
Fig. 3 shows the CV curve of Ni before and after
of O , the potential stays constant at about ꢀ0.1 V
immersion in a solution containing RuCl . The activa-
2
3
(
SCE), while if O is stripped off the potential of Ni goes
tion of the electrode is evident, H evolution commenc-
2
2
down to ca. ꢀ0.35 V (SCE).
ing ca. 0.2 V earlier in the presence of Ru. If the CV
curves are scrutinized more closely (inset), a pair of
peaks is seen to emerge around ꢀ1.0 V (SCE) as Ru is
present on the surface of Ni.
3
.2. Ru deposition
Fig. 2 also shows that if the de-aerated HCl solution
The peaks in the CV curves are attributed to hydro-
gen deposition (UPD H)/ionization on Ru sites
ꢀ
3
ꢀ3
contains RuCl Æ xH O (10 mol dm ), the open cir-
3
2
cuit potential of Ni takes an intermediate position. Un-
der similar circumstances Ru is deposited on the surface
of Ni.
[19,20]. Their area is thus proportional to the surface
+
charge spent in the reaction H +e ! H and can be
ads
a direct indicator of the amount of Ru atoms present
The effect of the presence of O is clearly evident if
2
different reaction vessels are used. For the same compo-
50
1
0
0
-
1000
-500
O 2
-
-
-
-
0.05
-
50
20
-
0.1
0.15
0.2
0.25
1
1
0
-
-
100
150
2
0
-
3
-950
-750
-550
-350
-
10
N 2
-200
-250
2
-
0.3
-20
0.35
0
0.5
1.0
1.5
2.0
E vs SCE / mV
Time, t / h
Fig. 3. Cyclic voltammetry of bare Ni (1) and Ru-modified Ni (2) in
the H evolution region: 1 mol dm KOH, 100 mV s , 25 ꢁC, 4 h
2
deposition. Inset: enlargement of the CV curve of Ru-modified Ni in
the potential range where peaks develop, for various Ru deposition
times. From inner to outer: 2, 4, 8, 24 h.
ꢀ3
ꢀ1
Fig. 2. Evolution of the open circuit potential (OCP) of bare Ni after
immersion in 0.1 mol dm HCl in the presence (1) or in the absence
ꢀ3
ꢀ3 ꢀ3
(
2) of O
, as well as in the presence (3) of 10 mol dm RuCl (de-
2 3
aerated solution).

I. Bianchi et al. / Chemical Physics 319 (2005) 192–199
195
on the Ni surface exposed to the solution (i.e., propor-
tional to the Ru deposit surface area) [21].
-1000
-1 100
-1200
-1300
-1400
-1500
-1600
The inset of Fig. 3 shows the CV of Ni after different
periods of immersion in the RuCl solution. Two fea-
3
tures are evident: (i) the peak height increases with
immersion time, and (ii) the ‘‘double layer’’ current also
increases in the same direction. Fig. 4 shows the charge
associated with the peaks in the CV curves plotted as a
function of the immersion time. Charge increases non-
linearly with time, showing a sort of induction period
at short times. Fig. 4 also shows that charge increases
at constant time of deposition as the temperature of
the solution is increased.
-
3
-2
-1
0
1
2
3
Voltammetric charges (q ) were obtained by graphi-
H
-2
log (j/mA cm )
cal integration of the H adsorption peak with the exclu-
sion of the H evolution tail. No correction for double
Fig. 5. Tafel lines for H
electrodes after various times of Ru deposition. From bottom: 0, 2, 4,
2
evolution on Ni and Ru-modified Ni
2
layer charging were introduced. Thus, qH values are
not to be regarded as very accurate from a quantitative
point of view. On the other hand, rigorous quantitative
analysis of the H adsorption process was not the main
objective of the work. The aim has been to provide evi-
dence of Ru deposition on the Ni surface. Already the
features of the curves in the inset of Fig. 3 indicate the
ꢀ3
8
, 24 h, 1 mol dm KOH, 25 ꢁC.
The Tafel slope, close to 120 mV for bare Ni, is seen
to decrease progressively down to 40 mV after a
deposition time of 24 h. It is also clear that for the
experimental conditions in this work, maximum
activation of Ni is obtained after ca. 8 h of spontane-
ous deposition. Therefore, no direct proportionality is
to be expected between q (charge of Ru) and b (activ-
ity of Ru), but rather between q and j at constant
potential.
phenomenological trend of q . It is very unlikely that
H
such trend can be overturned by inaccuracies in the
quantitative determination of qH.
The issue is different as q values are used to estimate
H
surface roughness or amount of Ru deposited (see Sec-
tion 4). However, also in that case inaccuracies in the
determination of qH are very unlikely to result in phys-
ically distorted results.
Evidence for Ru deposition was acquired also by
non-electrochemical techniques: EDX and XPS.
Fig. 6(a) shows that Ru is present on the surface of Ni
with a distribution which is probably affected by the
morphology of the etched metal surface. Thus, there
are regions with different density of Ru atoms, the latter
indicated by white spots. Fig. 6(b) shows the presence of
Fig. 5 shows the electrocatalytic evidence for the
presence of Ru on the surface of Ni. The Tafel lines
were not corrected for uncompensated ohmic drops.
Nevertheless, linear sections are clearly visible at
ꢀ
2
current densities j < 10 mA cm . In terms of electro-
catalytic activity the evidence for the presence of Ru
is even more striking, especially because the slope of
Tafel lines is an intensive quantity, i.e., it does not
depend on the true surface area of the electrode.
0
characteristic XPS peaks corresponding to Ru (metal),
as well as to RuO . The latter may well be due to the
x
easy adsorption of oxygen even from the gas phase be-
fore samples were introduced into the XPS chamber,
but it can also be related to the overlapping of the po-
tential regions of hydrogen and oxygen adsorption
100
[
19–24].
8
6
4
2
0
0
0
0
0
3
.4. Variables of process
3
2
A number of variables influence the spontaneous
deposition of Ru on Ni. Focussed experiments were car-
ried out to investigate each of them.
1
3.4.1. Temperature
Specific experiments showed that the amount of Ru
0
5
10
Time / h
15
20
25
deposited on Ni increases with increasing temperature.
On the other hand, Fig. 7(A) shows that as the deposi-
tion time increases the effect of temperature decreases
since activity tends to saturation.
Fig. 4. Charge under the peaks in Fig. 3, inset, as a function of Ru
deposition time, for different solution temperatures: (1) 20, (2) 25,
3) 30 ꢁC.
(

196
I. Bianchi et al. / Chemical Physics 319 (2005) 192–199
so much that, to a first approximation, the concentra-
tion of RuCl (within the range explored in this work)
3
can be regarded as uninfluential.
3
.4.4. Effect of solution preparation
The way the solution of RuCl is prepared has a ma-
3
jor effect on the experimental parameters. There are two
limiting procedures: (a) RuCl Æ xH O is dissolved in
3
2
water, then the required amount of HCl is added and fi-
nally the given concentration is achieved by further dilu-
tion with H O. The colour of this solution resulted dark
2
yellow/brown. (b) RuCl Æ xH O is dissolved in the min-
3
2
imum amount of H O, HCl is added and the solution is
2
left at rest for a few hours before diluting to the required
concentration. The colour of this solution resulted light
yellow/red. This solution is finally aged for about a
week.
Experiments show that in the solution where RuCl is
3
left to ‘‘digest’’ HCl (aging) Ru deposition takes place in
higher amount. Also the electrocatalytic activity is
clearly higher for samples modified in aged solution.
The higher efficiency of ‘‘aged’’ solutions of the precur-
sor was observed also by Wieckowski and co-worker [8].
3
.4.5. Spontaneous vs. electro-deposition
As shown in Fig. 2, spontaneous reduction of Ru(III)
occurs at a mixed potential of ca. ꢀ0.18 V (SCE). The
efficiency of the process can be assessed (among other
indicators) by carrying out electrodeposition from solu-
tions of the same composition. In terms of charge, spon-
taneous deposition occurs with higher efficiency than
electrodeposition, i.e., it produces data placed between
electrodeposition at ꢀ0.2 and ꢀ0.3 V, respectively (at
ꢀ0.18 V electrodeposition is undetectable). The same
picture is obtained also in terms of electrocatalytic activ-
ity (current density). Both increase with the time of
Fig. 6. (a) Ru distribution on the surface of Ni as evidenced by EDX.
White spots indicate areas covered by Ru after 8 h of Ru deposition.
Magnification 4000·. (b) Typical XPS spectrum of Ru 3d5/2 and Ru
3d3/2 for the same sample.
(electro)deposition of Ru.
3
.4.2. HCl concentration
Increasing the HCl concentration has a clear effect on
3.5. Electrocatalytic activity increment
Ru deposition in terms of voltammetric charge. In par-
ticular qH increases with HCl concentration at short
deposition times, while the plots tend to converge to a
common value at long times. Fig. 7(B) shows Tafel plots
at two different deposition times. It is evident that the
concentration of HCl becomes uninfluential at longer
times.
While the CV charge of H adsorption is a measure of
the exposed surface area of the Ru overlayer, the electro-
catalytic activity is measured by the rate of H evolution
2
at a given electrode potential. Fig. 8 shows plots of the
current density for H evolution at E = ꢀ1.4 V (SCE)
2
vs. q , the CV charge for H adsorption on the same sam-
H
ple. It is evident that the rate of increase of the electrocat-
alytic activity of Ru-activated Ni electrodes decreases
with increasing amount of deposited Ru.
3
.4.3. RuCl concentration
3
The effect of RuCl concentration shows an opposite
3
trend with respect to T and HCl: it increases with
increasing deposition time. The curve of qH vs. deposi-
tion time practically overlap at short times while they di-
4. Discussion
verge at longer times. But if the activity for H evolution
is expressed in terms of current at constant potential, the
Spontaneous deposition of Ru on Ni takes unambig-
uously place in the conditions of this work. Ni is
attacked by HCl, therefore there are three possible
2
effect of RuCl concentration turns out marginal in fact,
3

I. Bianchi et al. / Chemical Physics 319 (2005) 192–199
197
A
B
-1050
-
1150
3
-
1250
3
2
2
2
-
1450
-
1450
1
a
a
1
-950
-
1000
-
1150
-
1200
1
-1350
2
1
b
-1400
-1
b
3
-
2
0
1
2
-2
-1
0
1
2
-
2
-2
log(j/mA cm )
log(j / mA cm )
ꢀ3
Fig. 7. Tafel lines for H
1) 20, (2) 25, (3) 30 ꢁC. (B) Various HCl concentrations: (1) 0.05, (2) 0.1, (3) 0.2 mol dm . Ru deposition time: (a) 4 h, (b) 24 h.
2
evolution from 1 mol dm KOH solution on Ru-modified Ni electrodes. (A) Various temperatures of Ru deposition:
ꢀ3
(
In the absence of O (de-aerated solution), experi-
mental data acquire reproducibility since the consump-
8
6
4
2
0
0
0
0
2
2
1
tion of O , if present, will depend on the volume of
2
the reaction vessel and its shape. This explains why irre-
producible deposition was observed in test tubes: dis-
solved O was consumed cathodically at the beginning
2
j
and Ru deposition started to occur as the amount of
O was severely reduced, immission of further O from
2
2
the open end of the test tube being a very slow process.
Therefore, the exchange reaction taking place on Ni can
be written as:
0
0
10
20
30
40
50
-2
q
H
/ mC cm
2
y
y
Ni þ 2xH þ RuðIIIÞ ! Ni2þ þ xH
þ
þ Ru
3
ð1Þ
2
ꢀ3
Fig. 8. Current density for H
2
evolution from 1 mol dm
KOH
3
solution at ꢀ1.4 V (SCE) (from CV curves) as a function of H
adsorption charge. Temperature: (1) 20: (2) 25 ꢁC. q values refer to
Ru deposition times of 2, 3, 4, 6, 8 h from 0.1 mol dm HCl in the
with x + y/3 = 1.
H
ꢀ3
According to Eq. (1), Ni ions should be formed in the
2
+
ꢀ3
ꢀ3
presence of 10 mol dm RuCl
3
.
anodic reaction. The presence of Ni was clearly de-
tected by adding dimethylglyoxime to the residual solu-
tion. In this respect, it is interesting to note that the
2
+
cathodic processes driven by the anodic dissolution of
the metal: (1) O reduction, (2) H evolution, and (3)
green colour expected for Ni ions is not visible at
naked eyes at these concentrations, although the brown
colour of Ru(III) species vanishes as the reaction is con-
ducted in a small volume of solution (test tubes). It is,
however, possible to estimate that the amount of Ni ions
is higher in the presence of Ru(III), which indicates a
higher dissolution rate of Ni in line with the less negative
value of the OCP with respect to the blank HCl solution.
In other words, Ru(III) species in solution accelerate Ni
dissolution.
2
2
Ru deposition. The more positive the potential of the
cathodic reaction, the more that process predominates.
Therefore, in the presence of O , Ru deposition does
2
not occur, or if it occurs, it takes place with lower effi-
ciency. Ru deposition in the presence of O is evidently
2
possible as the rate of metal reduction is higher than the
limiting current of O reduction (the reaction was con-
2
ducted in the absence of solution stirring).
The effect of the state of the precursor solution is pre-
sumably related to a different complex of Ru in the solu-
tion. This can modify the redox potential of the ionic
species in solution as well as its reducibility to metal.
Further study to identify the species in solution would
throw light on this particular aspect.
The effect of temperature is rather expected. An in-
crease in temperature accelerates the etching of Ni,
therefore also the deposition of Ru. On the other hand,
at long times Ru deposition is probably controlled by
mass transfer limitations. Thus, Ni activation attains
at a saturation value independent of temperature.

198
I. Bianchi et al. / Chemical Physics 319 (2005) 192–199
The effect of HCl concentration, combined with the
curve for such an electrode the charge under the H
adsorption peak has turned out to be ca. 0.05 C cm .
Thus, only about 1/6 of the Ru atoms deposited are in
a position to be able to adsorb H, i.e., to exchange 1e
with species in solution.
ꢀ
2
effect of Ru(III) concentration, suggests that the whole
process of Ru spontaneous deposition is governed by
the reaction of anodic dissolution. The latter is acceler-
ated as HCl concentration increases. This leads to a par-
allel increment of Ru deposition. But at constant HCl
concentration the rate of anodic dissolution is fixed
and as a consequence also the rate of Ru deposition is
imposed. Thus, the amount of Ru deposited turns out
independent of the amount of Ru in solution.
On the other hand, an ideal monoatomic layer of Ru
ꢀ
2
should require ca. 280 lC cm to adsorb a monolayer
of H [27]. The much higher experimental value of charge
is presumably related to a pronounced roughness of the
surface. An estimate can be obtained from the ratio qexp
qth = 50 mC/0.28 mC = 180.
/
The occurrence of Ni activation by Ru spontaneous
deposition is striking: Tafel plots show that the mecha-
In section 3.1, it has been reported that the roughness
factor of the pre-treated Ni support was approximately
1.7. This value can increase a little during the spontane-
ous deposition of Ru due to the joint dissolution of Ni.
However, the final value of 180 cannot be a result of the
morphology of the support only, but the initial rough-
ness is presumably exhalted by preferential deposition
of Ru on the apices of the underlying metal layer. Thus,
roughness increases with the amount of deposited Ru.
On the other hand, electrodeposition of Ru has been re-
ported to result in roughness factors as high as 400–1000
nism of H evolution changes dramatically, the Tafel
2
slope turning from ca. 120 mV for bare Ni to ca.
0 mV for Ru-activated Ni. In the former case primary
discharge is rate-determining
4
þ
H
þ e ! HadðNiÞ
.
ð2Þ
ðsolÞ
In the presence of Ru H adsorption becomes stronger so
much that the primary discharge becomes faster than
the ion+atom reaction (second electron transfer) [3,25]:
þ
H
þ e ! HadðRuÞ
ð3aÞ
ðsolÞ
[
21–23].
The non-linear dependence of the electrocatalytic
adðRuÞ þ Hþ þ e ! H2ðgasÞ
H
ð3bÞ
ðsolÞ
activity for H evolution on the extension of the actual
2
Although not explicitly provided by the authors, the
Tafel slope can be seen to change in the same way in
Fig. 8 of reference [7]. A Tafel slope of 40 mV was also
surface area, i.e., qH , suggests that as the Ru layer be-
comes thicker, part of the surface, although still accessi-
ble to H adsorption, becomes excluded from the
reaction of H2 evolution presumably because of the
occlusion of small pores or crevices.
reported long ago [26] for H evolution on Ru nuclei on
2
carbon (but in acid solution).
The deposition of Ru on Ni has been ascertained. The
increase of the CV charge associated with H deposition
as well as of the ‘‘double layer’’ charge is clear evidence
for an increase of the exposed surface area of the Ru de-
posit. This implies that the surface roughness increases
with the amount of deposited Ru. Nevertheless, the
trend is to increase non-linearly with time, which can
be understood in terms of the size of (nano)crystals of
Ru increasing with deposition time so that the surface-
to-volume ratio decreases with time.
Actually, qH is measured by CV in a potential range
where molecular H is not yet liberated. Thus, H adsorp-
2
tion/desorption can sill take place inside the pores. The
situation is of course quite different during massive H2
evolution: the internal walls of pores can work only at
the very beginning of the reaction, then they remain
excluded.
An alternative (or additional) explanation for the
non-proportionality between j and qH is that with an
increasing size of Ru crystals their surface becomes more
regular, with a lower density of steps, kinks, isolated
atoms and other active sites.
The apparent higher efficiency of spontaneous depo-
sition with respect to electrodeposition at the same po-
tential indicates that the deposit nanostructure is
presumably different in the two cases. Further study
is necessary to scrutinize this aspect in a systematic
way.
Much more difficult is the assessment of the amount
of Ru deposited. An attempt at a calculation is reported
below. As Ni is immersed in 3 ml of solution in a test
tube, the colour of the solution vanishes after approxi-
mately 30 h. All Ru in solution can be supposed to a first
approximation to be present on the surface of Ni. Since
ꢀ
3
ꢀ3
the solution was 10 mol dm , the number of Ru
2
ꢀ3
atoms deposited (per cm ) is 0.001 mol dm
0.003 Æ
ꢀ
3
18
ꢀ2
dm Æ N = 1.8 Æ 10 cm , where N is the Avogadro
A
A
constant. If the deposit were as compact as bulk Ru me-
ꢀ3
tal (density 12.2 g cm ), the amount deposited (ca.
ꢀ2
0
2
.3 mg cm ) would correspond to a layer of ca.
50 nm average thickness. On the other hand if all the
5. Conclusions
atoms were distributed on the surface, the total charge
1. Ru can be spontaneously deposited on Ni from an
HCl solution containing Ru(III). The amount of metal
deposited depends on several variables including,
1
8
18
for H adsorption would be 1.8 Æ 10 Æ e = 1.8 · 10
Æ
ꢀ
19
ꢀ2
1
.6 Æ 10
C ꢁ 0.3 C cm
.
From the voltammetric

I. Bianchi et al. / Chemical Physics 319 (2005) 192–199
199
besides time, concentration of HCl and Ru(III), tem-
References
perature, gaseous atmosphere, and procedure of prep-
aration of the solution of precursors.
[
1] H. Wendt, V. Plzak, in: H. Wendt (Ed.), Electrochemical
Hydrogen Technologies, Elsevier, Amsterdam, 1990, p. 15.
2] S. Trasatti, Int. J. Hydrogen Energy 20 (1995) 835.
3] S. Trasatti, in: H. Gerischer, C.W. Tobias (Eds.), Advances in
Electrochemical Science and Engineering, vol. 2, VCH, Weinheim,
1992, p. 1.
2
. The deposition of Ru on Ni is revealed by the appear-
ance in alkaline solution of voltammetric peaks asso-
ciated with hydrogen adsorption/desorption, as well
as by the decrease of the Tafel slope for hydrogen
evolution from ca. 120 mV for pure Ni to ca. 40 mV.
. Ru is present on the Ni surface mainly as metal, but
[
[
[4] M. Jaccaud, F. Leroux, J.C. Millet, Mater. Chem. Phys. 22 (1989)
1
05.
3
4
[
[
[
5] A.C. Tavares, S. Trasatti, Electrochim. Acta 45 (2000) 4195.
6] S. Trasatti, J. Electroanal. Chem. 39 (1972) 163.
7] T. Schmidt, H. Wendt, Electrochim. Acta 39 (1994) 1763.
partly also as RuO as revealed by XPS.
x
. Ru deposition occurs with formation of a thick, por-
ous layer. Accordingly, the distribution of Ru over
the surface of the support, as determined by EDX
[8] W. Chrzanowski, A. Wieckowski, Langmuir 13 (1997) 5974.
9] A. Crown, A. Wieckowski, Phys. Chem. Chem. Phys. 3 (2001)
290.
[
3
(
Fig. 7(A)) shows phenomena of clustering and
[
[
10] C. Vericat, M. Wakisaka, R. Haasch, P.S. Bagus, A. Wieckowski,
J. Sol. State Electrochem. 8 (2004) 794.
11] M.G. Cattania, S. Ardizzone, C.L. Bianchi, S. Carella, Colloids
Surf. A 76 (1993) 233.
12] M.A. Sattar, B.E. Conway, Electrochim. Acta 14 (1969) 695.
13] M.M. Jaksic, J. Brun, B. Johansen, R. Tunold, Int. J. Hydrogen
Energy 20 (1995) 265.
agglomeration.
5
6
. In view of the previous point, should the obtainment
of submonolayers of Ru be the objective, it would call
for a careful selection of the various experimental
variables, i.e., lower HCl concentration, lower tem-
perature and, particularly, shorter times of contact.
. The deposition of Ru increments the roughness of the
electrode surface, as manifested by the increasing vol-
tammetric charge associated with H adsorption on
Ru sites. The roughness factor increases of up two
orders of magnitude.
. Ru is stable on the Ni surface as long as the electrode
is not brought to potentials where surface oxidation
takes place. Under similar conditions, dissolution of
Ru can occur.
. Spontaneous deposition is more efficient than electro-
deposition at comparable electrode potentials.
[
[
[
14] A. Palencsar, D.A. Scherson, Electrochem. Solid State Lett. 6
(
2003) E1.
[15] J.OꢀM. Bockris, E.C. Potter, J. Chem. Phys. 20 (1952) 614.
16] I.J. Brown, S. Sotiropoulos, J. Appl. Electrochem. 30 (2000)
[
1
07.
[
[
[
17] S.A.S. Machado, L.A. Avaca, Electrochim. Acta 39 (1994) 1385.
18] N. Krstaji c´ , S. Trasatti, J. Appl. Electrochem. 12 (1998) 1291.
19] R. Lezna, N.R. De Tacconi, A.J. Arvia, J. Electroanal. Chem. 151
(1983) 193.
20] S. Had zˇ i-Jordanov, H. Angertein-Kozlowska, M. Vukovi c´ , B.E.
Conway, J. Phys. Chem. 81 (1977) 2271.
7
8
[
[
[
21] E.K. Tuseeva, A.M. Skundin, V.S. Bagotskii, Elektrokhimiya 10
(
1974) 601.
22] V.S. Bagotzky, A.M. Skundin, E.K. Tuseeva, Electrochim. Acta
1 (1976) 29.
2
[
[
23] V.S. Bagotzky, A.M. Skundin, Electrochim. Acta 30 (1985) 485.
24] V. Horvat-Radosevic, K. Kvastek, M. Vukovic, D. Cukman, J.
Electroanal. Chem. 482 (2000) 188.
Acknowledgment
[
[
25] S. Trasatti, J. Electroanal. Chem. 39 (1972) 173.
26] R.D. Giles, J.A. Harrison, H.R. Thirsk, J. Electroanal. Chem. 20
The authors are grateful to MIUR (PRIN) for finan-
cial support to this work.
(
1969) 47.
[27] V.S. Entina, O.A. Petrii, Elektrokhimiya 4 (1968) 457.