Anodically formed oxide films and oxygen reduction on electrodeposited ruthenium in acid solution

Article abstract of DOI:10.1016/j.electacta.2005.05.029

The impedance of the anodically formed hydrous Ru oxide in the system Ru|oxide film|1 M HClO₄ solution has been studied in the range of potentials where the electrode process occurs by a double electron and proton exchange between the oxide film and the solution. The results allowed us to clearly distinguish between the surface process at higher frequency and the bulk process at lower frequency. The high-frequency charging is found to be coupled to Faradaic charging at the film/solution interface. Evaluation of the impedance data at lower frequency, using diffusion equations for the finite boundary conditions, yields an effective proton diffusion coefficient to be 10 ^-10 to 10^-11 cm² s^-1. Oxygen reduction on the spontaneously oxidized ruthenium electrode was discussed on the basis of a rotating ring-disk voltammetry.

Full text of DOI:10.1016/j.electacta.2005.05.029

Electrochimica Acta 51 (2006) 1157–1164
Anodically formed oxide films and oxygen reduction on
electrodeposited ruthenium in acid solution
a,∗,1
a
a
ˇ
´
´
´
ˇ b
, R. Babic , F. Jovic , Z. Grubac
M. Metikos-Hukovic
a
Department of Electrochemistry, Faculty of Chemical Engineering and Technology, University of Zagreb, P.O. Box 177, 10000 Zagreb, Croatia
b
Faculty of Chemical Technology, University of Split, N. Tesle 10 Split, Croatia
Received 8 December 2004; received in revised form 4 February 2005; accepted 15 May 2005
Available online 1 July 2005
Abstract
The impedance of the anodically formed hydrous Ru oxide in the system Ru|oxide film|1 M HClO4 solution has been studied in the range
of potentials where the electrode process occurs by a double electron and proton exchange between the oxide film and the solution. The results
allowed us to clearly distinguish between the surface process at higher frequency and the bulk process at lower frequency. The high-frequency
charging is found to be coupled to Faradaic charging at the film/solution interface. Evaluation of the impedance data at lower frequency, using
diffusion equations for the finite boundary conditions, yields an effective proton diffusion coefficient to be 10−10 to 10−11 cm2 s−1
.
Oxygen reduction on the spontaneously oxidized ruthenium electrode was discussed on the basis of a rotating ring-disk voltammetry.
© 2005 Elsevier Ltd. All rights reserved.
Keywords: Ruthenium; Ruthenium oxides; Oxygen reduction; Rotating ring-disk electrode; Impedance spectroscopy; Cyclic voltammetry
1. Introduction
electrolytes [9–11]. The electrocatalytic reduction of oxygen
plays a major role in several industrial processes and in corro-
sion and corrosion protection [1,12]. Pt–Ru alloy is preferred
over platinum as a better promoter of the oxidation of small
organic molecules [13]. For many of the above applications,
the ruthenium or ruthenium oxide coatings appear to be more
desirable and suitable [2] than the bulk metal. Electrochemi-
cal deposition of ruthenium on various substrates and anodic
oxidation of deposited ruthenium is a well-known method
for the preparation of ruthenium oxide coating [4,5,14–25].
Conway et al. [14–16], Burke and co-workers [17,18,25] and
Michell et al. [19] frequently used cyclic voltammetry (CV)
and other polarization techniques to investigate the surface
properties of electrodeposited ruthenium. They showed that
the continuous potential cycling into the oxide region up to
1.4 V led to a modification of the oxide film accompanied
by the enhanced electrocatalytic efficiency in the evolution
There has been a growing interest in the electrochemical
properties of the ruthenium, especially due to the electro-
catalytic properties of anodes based on Ru alloys with Pt
or ruthenium oxides. RuO2 films on Ti substrate, prepared
by thermal decomposition of ruthenium salts, are known as
dimensionally stable RuO2 × TiO2 anodes of excellent sta-
bility and low overvoltage for the evolution of chlorine [1].
Ruthenium hydrous oxides, especially those formed by the
sol–gel or the electrochemical processes, distinguish them-
selves from other transition metal oxides by their excep-
tionally high-charge storage capacity [2–5], due to the high
electron and proton conductivity, together with a high degree
of reversibility of the redox processes which take place over a
large range of potentials. Several studies of oxygen reduction
on a ruthenium electrode have shown its considerable affin-
ity for the oxygen reduction in both acid [6–8] and alkaline
´
of chlorine and oxygen. Vukovic et al. [20–22] have shown
that the electrodeposited ruthenium layer contains a signifi-
cant amount of physically adsorbed and chemisorbed water,
and that the modification of the oxide film caused by the con-
tinuous cycling of the potential is due to an increase in the
∗
Corresponding author. Tel.: +385 1 4597140; fax: +385 1 4597139.
E-mail address: mmetik@marie.fkit.hr (M. Metikosˇ-Hukovic´).
ISE member.
1
0013-4686/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2005.05.029

1158
M. Metikosˇ-Hukovic´ et al. / Electrochimica Acta 51 (2006) 1157–1164
number of hydroxyl groups in the film. The electrochemical
impedance spectroscopy (EIS), which has been considered
a powerful in situ diagnostic technique in separating differ-
ent rate processes in the frequency domain, has been used in
studying the electrochemical properties of the hydrous ruthe-
nium oxide film [23,24].
The electrochemical behavior of the electrodeposited
Ru/HClO4 solution (1 mol dm−3), in the range of potentials
between hydrogen and oxygen evolution, was studied under
potentiodynamic and potentiostatic experimental conditions
using CV and EIS. In this work, we aimed at studying the
interfacial and bulk solid-state redox charging processes in
the investigated system using EIS. Potentiostatic experiments
conducted in the frequency domain of 6 decades enable us to
clearly distinguish the slower potential-dependent charging
processes, i.e. the variations of the proton transport rate with
applied potential in Ru oxide films from the faster processes
related to the charging at the film/solution interface.
at −0.13 V for 60 s in order to remove completely the possible
surface oxides. Finally, the samples were potentiodynami-
cally polarized (10 mV s−1) to the potential of 0.05 V, which
wastheinitialpotentialforCVmeasurements. Forimpedance
measurements, the samples were potentiodynamically polar-
ized (10 mV s−1) to the desired potential and after a 60 s wait
period, impedance data were collected. Measurements were
performed with the ac voltage amplitude 5 mV in the fre-
quency range from 40 mHz to 50 kHz. Both CV and EIS
measurements were performed in a standard three-electrode
electrochemical cell. A thin platinum electrode was used as
the counter electrode and a saturated Hg/HgSO4 electrode,
whose potential was 0.658 V versus SHE, as the standard
electrode. A PAR EG&G potentiostat (model 273) and a PAR
EG&G lock-in amplifier (model 5301A) controlled by a PC
were used.
2.2. RRDE measurements
The electrocatalytic affinity of the electrodeposited Ru
films towards the oxygen reduction reaction (ORR) was
investigated using the ring-disk electrode. The study was a
part of our broader investigation of the ORR on Pt, Pt–Ru
electrodes as well as on Pt–Ru/Vulcan fuel cell catalysts.
RRDE measurements were carried out using the ring-
disk electrode assembly, Model EAD 10000. Both the disk
and ring were made of platinum. The disk surface area
was 0.1257 cm2. Ruthenium deposition on the disk was per-
formed under the same condition as on the gold electrode.
During ruthenium deposition on the disk electrode, the ring
electrode was held at 0.84 V versus SHE in order to pre-
vent ruthenium underpotential deposition on it. The RRDE
collection efficiency, N, was determined to be 0.24 from the
slope of the disk currents, jD, against ring currents, jR, at
different rotation rates using 10−3 mol dm−3 K3Fe(CN)6 in
0.5 mol dm−3 K2SO4. The potentiostat used for the ring-disk
measurements was a Electrolab BPG-200 bipotentiostat with
a built-in sweep generator. A platinum foil was used as a
counter electrode and the saturated calomel electrode as the
reference electrode. The disk was polarized over the potential
range 0.80–0.10 V with a scan rate 5 mV s−1 and a rotation
rate from 1000 to 6000 rpm. The ring potential was kept at
1.2 V versus SHE to oxidize peroxide at the mass transport
limiting rate.
2. Experimental
2.1. Cyclic voltammetry and EIS measurements
A disk-shaped electrode, made from spectrographically
pure gold (Johnson & Matthey) with a geometrical sur-
face of 0.071 cm2, served as the support for deposition of
ruthenium films. Before electrodeposition, the gold electrode
was mechanically and electrochemically treated. Mechanical
treatment consisted of polishing using Al2O3 granules of fine
sizes (1, 0.3 and 0.05 ␮m) up to a mirror finish. The electrode
was degreased with ethanol and rinsed with deionized water
in an ultrasound bath. The electrochemical treatment con-
sisted of potential cycling, within the potential range from 0
to 1.8 V versus SHE, with a scan rate of 100 mV s−1, until
a reproducible potentiodynamic profile was reached for the
Au electrode.
All measurements were performed at room temperature
(21 2 ◦C) and potential values in the paper refer to the SHE.
Ruthenium deposition was performed from 2 × 10−3
mol dm−3 RuCl3 in 1 mol dm−3 HClO4 at a constant poten-
tial of 0.05 V versus SHE for a period of 1500 s. The ref-
erence electrode during deposition was Ag/AgCl electrode
immersed in 3 mol dm−3 KNO3 solution (Radiometer ana-
lytical), while the platinum foil served as the counter elec-
trode. Freshly deposited electrodes were rinsed in order to
remove the rest of the deposition solution and transferred to
an electrochemical cell containing 1 mol dm−3 HClO4 (pH
0) deaerated by continual flow of N2 (99.99%). The elec-
trodes were treated by repetitive cycling (50 cycles) over the
potential range 0.05–0.85 V versus SHE with a scan rate of
100 mV s−1 with the objective of cleaning the surface and
reaching stationary conditions. Then, the electrode was held
3. Results and discussion
3.1. Cyclic voltammetry
A family of the cyclic voltammograms of the deposited
ruthenium electrode taken at a sweep rate of 10 mV s−1 in
1 mol dm3 HClO4 is shown in Fig. 1a and b. The anodic
limit was increased progressively with each sweep from 0.80
to 1.30 V and the electrode was held at the cathodic limit
(E = −0.13 V) to allow the complete reduction between each
sweep. The CV features in figures are in agreement with
literature data [14–19,25], which indicate that the overall
electro-oxidation/electro-reduction can be split into a number

M. Metikosˇ-Hukovic´ et al. / Electrochimica Acta 51 (2006) 1157–1164
1159
0.05 V). When anodic limit potential equals 1.05 V, merg-
ing of both cathodic current maxima into one maximum is
observed in the cathodic cycle.
Fig. 1b shows that the anodic current decreases beyond
1.05 V due to the electrode passivation and formation of
hydrous RuO2. After oxidation of the deposited ruthenium
from 1.05 to 1.15 V, the cathodic sweep shows a well pro-
nounced current peak corresponding to the oxide film reduc-
tion, which shifts towards cathodic potential values with the
shift of the anodic limit to more positive potentials. Beyond
1.2 V, the anodic current starts to increase due to both oxy-
gen evolution and ruthenium dissolution, where ruthenium
species bearing Ru oxidation states up to +8 (RuO4) are
involved [27,28]. When the electrode has been oxidized to
≥1.35 V, there is almost no oxide film reduction peak on the
cathodic sweep; the oxide film reduction is only completed
in the hydrogen region. Since the current, during the forward
and reversal potential scan, is not influenced by the electrode
rotation, it is proposed that the hydrogen and oxygen species
diffuse into and out of the oxide layer. Details to this point will
be discussed along with the results of the EIS experiments.
The irreversibility in the formation and reduction of
the ruthenium oxide films, which increases with increasing
anodic potential, is linked to their intrinsic structures. XPS
results obtained by Hu et al. [5] showed that the irreducible
oxide was composed of an aggregate consisting of Ru in var-
ious oxidation states, bridged oxygen, OH and water in a
3D-like structure which is relatively ordered and compact.
Fig. 1. CVs for a plated Ru deposit measured in 1 M HClO4 at various sweep
reversals: (a) from 0.8 to 1.05 V and (b) from 1.05 to 1.3 V; ν = 10 mV s−1
.
of reactions. They will be presented in short. On the pos-
itive going potential sweep, there is the hydrogen region
which overlaps with ruthenium oxidation [14,15,18,19,26].
The slightly pronounced current plateau between ca. 0.2 and
0.7 V formed during the electrode linear polarization is due
to the formation of a surface oxide film [14–19,25,27]. Since
the corresponding cathodic branch of the CV shows two cur-
rent maxima, it seems that the anodic plateau was formed
by overlapping two stages of the surface oxidation (Ru(I)
and Ru(II)) as was suggested by Conway and co-workers
[14,15]. The oxidation state of ruthenium in the oxide is two
at the end of this potential region, i.e. ruthenium oxide exists
as hydrated RuO [19], Ru(OH)2 or RuO × H2O [28]. Fur-
ther increase in current, which occurs by increasing anodic-
potential limit, corresponds to the formation of ruthenium
oxides where ruthenium is of higher oxidation states. Burke
and co-workers [17,18] proposed that the slight increase in
oxidation current from 0.7 to ca. 1.0 V is possible due to
the formation of RuOOH (Ru2O3). The standard electrode
potential of Ru/Ru2O3 couple equals to 0.738 V [29]. Its for-
mation in that potential range on the deposited ruthenium
electrode was proposed also by other authors [20,22,28,30].
Fig. 1a shows that the oxide layers formed on the ruthenium-
deposited electrode in the potential range 0.8–1.0 V give
cathodic sweeps with two current maxima. The first, slightly
pronounced current peak shifts to more cathodic values with
increasing anodic-potential limit, while the second and more
pronounced current peak remains at the same potential (ca.
3.2. Electrochemical impedance spectroscopy
A typical set of impedance spectra recorded on the
deposited ruthenium electrode potentiostatically oxidized
at potentials between 0.3 and 1.1 V, presented as Bode
plots, is shown in Fig. 2. The modeling procedure for the
impedance frequency responses is given in the appropriate
description of relevant surface and bulk redox processes
involved in the system investigated. The Faradaic process
Fig. 2. Bode plots for a plated Ru deposit in 1 M HClO4 at specified
potentials.

1160
M. Metikosˇ-Hukovic´ et al. / Electrochimica Acta 51 (2006) 1157–1164
of the electrochemical reactivity of hydrous oxides in the
investigated potential range varies strongly, depending on
the extent of coupling between the internal redox system
of the oxide and the redox system at the oxide/solution
interface. The activity of this redox system depends, in turn,
on the rate of double electron–proton injection process into
a hydrous Ru oxide according to Eq. (1):
RuO(x+δ)(H2O)(y−δ)+2δH+ + 2δe− ⇔ RuOx(H2O)y, (1)
i.e. a hydrogen atom is incorporated into a hydrous oxide
by injection of an electron from the ohmic contact and a
proton from solution, thus lowering the formal oxidation
state of Ru by one unit. The reverse electro-oxidative process
involves the double ejection of an electron and a proton from
a hydrous oxide. While this process is very fast for surface
sites located in easily accessible regions, the proton incorpo-
ration in regions with difficult access creates a slow diffusion-
controlled process. Slow diffusion-controlled proton penetra-
tion has been shown to occur using diverse techniques such
as potential pulse [31,32], current pulse [33] spectroelectro-
chemistry [34] and impedance analysis [23,24].
In order to model the impedance data for the above pro-
cess (Fig. 2), the following important aspects of the potential
dependence of the Ru oxide film must be considered: (i) the
pseudocapacitive nature of the electrode process occurring at
the surface of Ru oxide under the conditions of our experi-
ments according to the relation (1); (ii) the interfacial transfer
processes, one is an electron transfer at the electrode/film
(M/F) and the other is an ion transfer at the solution/film
interface (F/S); (iii) the charge transport within the film. As
resistance does not appear to depend on the oxide film prop-
erties, the impact of the RF on the observations is ignored in
the present discussion.
Fig. 3. The electrical equivalent circuit used to fit the impedance spectra.
(ii) the χ2 error is suitably low (χ2 ≤ 10−4) and the errors
associated with each element are up to 5%. Many EEC were
examined, but only the one shown in Fig. 3 obeyed these cri-
teria. Also, this EEC fits well to the impedance data obtained
at various potentials in 1 mol dm−3 HClO4 solution. Fig. 2
depicts the overlay of the experimental (at the selected poten-
tials) and the calculated Bode data shown in Table 1, using
the EEC in Fig. 3. The EEC in Fig. 3 is also applicable to the
thin film-modified electrodes (oxides and conducting poly-
mers) if the conventional semi-infinite Warburg impedance
is replaced with a finite Warburg term [36] or a finite length
transmission line [23,37–40]. The meaning of the parameters
in the EEC is as follows:
- Rꢀ represents the serial sum of all resistance at the F/S
interface, i.e. the charge transfer resistance (Rct), the oxide
film resistance (RF), the proton transfer resistance (Rpt) and
the electrolyte resistance (Rel), in which the Rel dominates
the serial sum.
- CF/S is the total film/solution interfacial capacitance (CF/S).
Two approaches were taken in this work to understand
and interpret the impedance data given in Fig. 2. The first
one involved identifying the best-fit electric equivalent circuit
(EEC), based on Boucamp’s software [35], in order to estab-
lish the values of circuit elements. The intention of this work
was then to correlate their values with cyclic voltammetry
data as a function of the applied potential. The usual guide-
lines for the selection of the best-fit EEC were followed: (i)
a minimum number of circuit elements are to be employed;
Itrepresentstheparallelsumofthepseudocapacitance(C )
and the double layer capacitance (Cdl).
␾
- Serial combination of the surface resistance R corre-
␾
sponding to the surface redox process and the transmission
line-like impedance (ZD) represents the impedance due
to a slow charging/discharging process in the bulk of the
oxide film, occurring through the Faradaic charge trans-
fer and restricted diffusion of protons within the bulk of
the film.
Table 1
Impedance parameter values of the deposited ruthenium electrode at different potentials in 1 mol dm−3 HClO4
E (V)
103 × QF/S (ꢀ−1 cm2 sn)
n
102 × R (ꢀ cm2)
RD (ꢀ cm2)
102 × τD (s)
103 × CT (F cm−2
)
␾
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
1.08
10.4
10.6
10.1
7.5
5.9
5.1
4.0
4.2
4.8
0.80
0.81
0.82
0.82
0.82
0.82
0.84
0.83
0.83
4.8
9.3
12.4
12.1
7.0
1.30
0.98
0.69
0.74
3.11
5.08
8.73
27.49
67.01
0.19
0.11
0.11
0.23
1.05
1.59
3.19
8.82
12.04
11.9
11.7
11.7
10.6
9.3
8.2
7.7
7.4
6.6
6.0
18.0
50.0
368.0
Rꢀ = 0.41 ꢀ cm2.

M. Metikosˇ-Hukovic´ et al. / Electrochimica Acta 51 (2006) 1157–1164
1161
The overall impedance of the equivalent circuit as a func-
tion of frequency was calculated according to Eq. (2):
structural parameters are the concentration of the protons as
diffusing species and their diffusion coefficient. The oxide
films with higher thickness, lower concentration of protons
and lower diffusion coefficient will have higher values of
these parameters. An increase of these parameters generally
denotes an increase in the inhibition of the film equilibra-
tion, once the electrode has been polarized. Table 1 shows
that these values, particularly those of RD and τD increase
(except from 0.3 to 0.4 V) with increasing Ef. Oxide films
grown on Ru by constant potential anodization reach a thick-
ness of up to 30 nm [23]. Taking into account this thick-
ness and applying the relation τD = L2/D, where L is the
thickness of the oxide layer, the diffusion coefficient of pro-
tons (D) can be estimated. The values obtained are between
10−10 and 10−11 cm2 s−1 depending on Ef. With increas-
ing Ef, the proton diffusion coefficient decreases due to the
change in film structure that becomes less amorphous and
more compact at higher anodic potentials. The other fac-
tor that influences the change in D values is the change in
a film thickness with increasing Ef. Optical measurements
[23] as well as potentiodynamic [17] measurements indi-
cate that the film thickness increases with increasing Ef. For
the oxide film on ruthenized gold, Rishpon and Gottesfeld
[23] calculated an effective diffusion coefficient of protons
to be between 10−11 and 10−12 cm2 s−1. For thermally pre-
pared RuO2 films, the experimentally determined values for
diffusion coefficient of protons differ significantly, depend-
ing upon the layer preparation. Using ac impedance studies
Raistrick and Sherman [43] and Tsai and Rajeshwar [44] esti-
Rꢀ + (R + ZD)
␾
Z(ω) =
(2)
[1 + (R + ZD)jωCdl]
␾
where ω is the angular velocity and j is (−1)1/2. It should be
noted that EIS was unable to distinguish between the dou-
ble layer capacitance (Cdl) and the pseudocapacitance (C )
␾
chargingmodesoftheRuoxidefilms. Thus, inEq. (2), instead
of Cdl, the value of CF/S is introduced.
The high-frequency impedance (ZHF) is given by Eq. (3):
Rꢀ + 1
ZHF(ω) =
(3)
(jω)nQF/S
where the frequency-independent impedance parameters
QF/S and n, describing the impedance of a constant phase
element, are introduced instead of the film/solution interfa-
cial capacitance (CF/S) due to the non-ideality of the solid
electrode surface response [41].
The frequency dependence of the so-called transmission
line-like impedance elements is given by Eq. (4):
ZD(ω) = RD(jωτD)−1/2ctnh (jωτD)1/2
(4)
where τD represents the diffusion time constant related to the
diffusion resistance, RD, and the internal capacitance, CD,
τD = RDCD.
AccordingtoEqs. (2)–(4), theoverallelectrodeimpedance
is described by six frequency-independent parameters, three
of which (Rꢀ, QF/S and n) describe the fast electrode pro-
cess dominating at higher frequencies, while the other three
mated a diffusion coefficient of protons to be 10−10 cm2 s−1
.
Conway et al., investigating a self-discharge of a thermally
prepared electrode, estimated the proton diffusion coefficient
to be around 10−12 cm2 s−1, while Weston and Steele [32],
using current transient measurements, obtained the values
between 10−14 and 10−15 cm2 s−1. Generally, the charging of
the grain surface, which requires only the transport of protons
along aqueous pores, may be associated with a much higher
proton diffusion coefficient than the charging of the bulk of
oxide grains. The proton diffusion coefficient estimated in
this work correlates well with values given in literature. Our
results suggest that the main reason for the observed increase
of RD and τD is a change in the oxide structure and a decreas-
ing concentration of the protons inside the oxide layer due to
the lower content of the proton-donating species such as H2O
and OH. The values of RD and τD can be grouped into three
regions: the first one with very low values of RD and τD in the
potential range from 0.3 to 0.6 V, where the oxide layer exists
as hydrated RuO. An increase in RD and τD is observed in
the potential region from 0.7 to 0.9 V, where hydrated RuO
undergoesfurtheroxidationtoRuOOH(Ru2O3). Finally, oxi-
dation of Ru(III) oxide layer to RuO2 is accompanied with
a significant increase in all three impedance parameters, RS,
RD and τD.
(R , RD and τD) describe the slow electrode process that
␾
dominates at low frequencies. Their values, as a function
of film formation potential (Ef), are given in Table 1. The
high-frequency resistance Rꢀ, that is equal to 0.41 ꢀ cm2,
does not depend on the electrode potential and presents the
electrolyte resistance. The high-frequency capacitance is pre-
sented by a constant phase element (CPE) which generates an
impedance ZCPE = [QF/S(jω)n]−1. The CPE has no physical
meaning, but allows an evaluation that indicates the degree
of system complexity. Hence, the CPE represents the devia-
tions observed from the ideal capacitor, such as slope in plot
log |Z| versus log f smaller than −1 and a phase angle in
plot θ versus log f smaller than −90◦, respectively. The value
of n associated with CPE, greater than 0.80, indicates that
CPE behaves as a capacitor, hence it can be assumed that
∼
QF/S
CF/S. Data in Table 1 show that CF/S values are rather
=
high (4–11 mF cm−2) and decrease with increasing anodic
potential. Since CF/S values are much higher than ordinary
Cdl values for plane, compact oxide electrodes [42], it may
∼
be assumed that CF/S C . The decrease of C values with
=
␾
␾
increasing Ef points to a certain extent of surface area reduc-
tion due to the formation of less hydrous and more compact
oxide films.
An insight into the content of proton-donating species
could also be obtained from the values of total capacitance,
The values of R , τD and RD depend on the thickness
␾
and the structure of the oxide film. The most important
CT, equal to: CT = C + CD. The CT values were calculated
␾

1162
M. Metikosˇ-Hukovic´ et al. / Electrochimica Acta 51 (2006) 1157–1164
under the assumption that QF/S ≈ C because QF/S ꢀ Cdl,
mixed activation–diffusion control at all rotation rates exam-
ined. The ring current, jR, which corresponds to the diffusion-
controlled oxidation of peroxide is very small indicating that
the O2 reduction on the disk proceeds as a four-electron pro-
cess. The amount of hydrogen peroxide can be calculated
using the detected ring currents and disk currents normalized
for the collection efficiency, N:
␾
while CD values were calculated from the τD and RD values.
Thevalueof CT isalsopresentedin Table1 asafunctionof Ef.
The CT decreases with increasing Ef suggesting a loss of reac-
tion sites caused by the formation of the oxide films in which
the oxidation state of Ru increases. Our results also show that
the Faradaic pseudocapacitance (C ) dominates the total film
␾
capacitance of oxide layers prepared at lower potential val-
ues; when increasing the film formation potential, its fraction
in the total capacitance decreases. All impedance results indi-
cate that anodic oxidation at lower potential values produces
more porous and more open oxide films, i.e. the films that are
more hydrous or amorphous, at least in their surface region.
Oxide films formed at higher anodic potentials are character-
ized by a reduced surface area, they are less hydrous and/or
their structure is more crystalline.
100jR
%H2O2 =
(5)
jR + NjD
The calculation shows that the amount of H2O2 decreases
with increasing cathodic potential: at ∼0.6 V, it approaches
6% and drops to ∼2% at 0.1 V. Thus, the main reaction at the
disk electrode involves the four-electron overall reduction.
For a rotating disk electrode experiment involving an elec-
trode process, which is first order in reactant, the observed
current is given by Koutecky–Levich Eq. (6) [45]:
3.3. Oxygen reduction
1
1
1
=
+
(6)
j
jk
jL
Rotating ring-disk experiments were performed to mea-
sure the amount of H2O2 species formed as an intermediate
or to confirm that O2 is directly reduced to water in a four-
electron process. Fig. 4 shows a typical result obtained on a
rotating ruthenium-deposited electrode as a function of the
potential and rotation rate. A platinum ring, potentiostated at
E = 1.2 V, was used to monitor the formation of the hydrogen
peroxide. The S-shaped disk current, jD, and its dependence
upon the rotation rate indicates that the reaction is under
where jk is the kinetic current and jL is the corresponding
diffusion limiting current equal to Bω1/2, where B is the con-
stant. The plots of 1/j versus ω1/2, assuming a uniform current
distribution, should yield parallel straight lines with the inter-
cepts corresponding to jk and the slope yielding the value
of B.
The parameter B can be calculated from Eq. (7) [45]:
B = 0.62nFD2/3
ν
−1/6c(O2)
(7)
where F is the Faraday constant (96,490 C mol−1), D the
molecular diffusion coefficient of oxygen in 1 mol dm−3
perchloric acid (1.1 × 10−5 cm2 s−1), ν the kinematic vis-
cosity (0.01 cm2 s−1) and c(O2) is the bulk concentration of
dissolved oxygen (1.6 × 10−6 mol cm−3) [46]. When n = 4,
B = 0.132 mA cm−2 (rpm)−1/2
.
The linearity and parallelism of the plots obtained when
data from Fig. 4 were plotted as 1/j versus ω−1/2 confirm
that the kinetics are first order with respect to oxygen and
that Eq. (6) is applicable. The average experimental value
of B obtained from these plots is 0.13 mA cm−2 (rpm)−1/2
,
which is in good agreement with the calculated value. This
indicates that the oxygen reduction proceeds by an overall
four-electron transfer process, i.e. O2 + 4H+ + 4e− → 2H2O.
The Tafel slope could be obtained from the mass transfer-
corrected plot E versus log [jjL/(jL − j)] [47]. The value for
Tafel slope that equals −124 mV dec−1 is obtained from this
plot in the potential range from 0.77 to 0.64 V. The obtained
value is not far from −118 mV dec−1 (−2.3 × 2RT/F), which
corresponds to the mechanism that involves one-electron
discharge-determining step. Such data allow us to con-
clude that the product of the rate-determ−ining step in the
ORR is the electro-reduced adsorbed (O2
)
species, i.e.
ads
(O2)ads + e− → (O2
)ads. Between 0.6 and 0.4 V, the slope
−
is −190 mV dec−1 and below 0.4 V, the current approaches
limiting values. The obtained values are close to those found
on platinum in acid media [46]. However, Tafel slopes higher
Fig. 4. Rotating disk-ring currents during the ORR on a plated Ru deposit at
specified rotation rates (rpm) in 1 M HClO4; ν = 5 mV s−1. Ring potential,
E = 1.2 V vs. SHE.

M. Metikosˇ-Hukovic´ et al. / Electrochimica Acta 51 (2006) 1157–1164
1163
than −120 V dec−1 are often observed in reactions through
2Z1m2
k2 =
k3 =
(10)
some adsorbed layers at the surface. It is quite clear that
a thin layer of ruthenium oxide, which exists in the men-
tioned potential range, could cause an increase in the Tafel
slope.
(I1N + 1)
Z2Nm1
(11)
(I1N + 1)
To gain more insight into the kinetics of oxygen reduction
on ruthenium-deposited electrode, the RRDE measurements
were analyzed on the basis of the simple model proposed by
Damjanovic et al. [48].
where Z1 = 0.62 D2/3(O2) ν−1/6; Z2 = 0.62 D2/3(H2O2) ν−1/6
,
D(H2O2) = 6.8 × 10−6 cm2 s−1; I1 and m1 are the intercepts
and slope of the plot of jD/jR versus ω−1/2, respectively; m2
is the slope of the plot jDL/(jDL − jD) versus ω−1/2 and N
is the electrode collection factor. The results showed that
when jD/jR versus ω−1/2 is plotted at different electrode
potentials (in the range from 0.4 to 0.6 V), a linear behavior
is obtained with both the intercept and slope being depen-
dent upon the electrode potential. At potentials below 0.4 V,
the reproducibility of the experiments was relatively poor,
which prevented us from analyzing such data. According
to Damjanovic et al. [48], the type of jD/jR versus ω−1/2
dependence obtained suggests that all three reactions occur
in the model. Within the above mentioned potential range, k1
The model assumes: (i) no catalytic decomposition of
hydrogen peroxide; (ii) adsorption/desorption of hydrogen
peroxide is fast and in equilibrium; (iii) the rate constant
for the hydrogen peroxide oxidation is negligible. The disk-
ring data analysis reported by Hsueh et al. [47] could be
used to derive a mathematical expression which would per-
mit the calculation of the rate constants for the intermedi-
ate steps for the ORR. The main feature of the analysis is
plotting jDL/(jDL − jD) versus ω−1/2 at a constant potential,
where jDL is the limiting current at the disk electrode. This
plot is linear and adopts the same form, independently of
the reaction scheme, provided that no catalytic decomposi-
tion of hydrogen peroxide occurs at the electrode surface.
Fig. 5 shows that there is a linear dependence of the plot
jDL/(jDL − jD) versus ω−1/2 upon the electrode potentials.
Such dependence suggests that the ORR proceeds without
catalytic decomposition of the hydrogen peroxide and that the
model proposed by Damjanovic et al. [48] could be applied.
The rate constants according to Hsueh et al. [46] are given by
Eqs. (9)–(11):
exponentially decreases from 2.5 × 10−2 to 4 × 10−3 cm s−1
,
while k2 does not depend upon the potential and is equal to
5 × 10−4 cm s−1. Since the k1 is much larger than k2 over the
potential range explored, oxygen is mainly reduced to water
through the four-electron transfer reaction. The value of k3
is of the order of 10−3 cm s−1 and slightly increases with the
potential. It is larger than k2, indicating that the formation of
hydrogen peroxide is a slow process in the series of the ORR
steps.
4. Conclusions
Oxide films anodically formed on electrodeposited ruthe-
nium are investigated by impedance spectroscopy. The results
are discussed in terms of the formation and reduction of the
phase oxides in the anodic film by mechanism involving the
double ejection and injection of electrons and protons, with
corresponding changes in the valence state of the ruthenium
atom. The time-dependent information obtained from in situ
EIS measurements showed that the processes in the system
Ru/oxide film/solution are kinetically limited by diffusion of
protons within the hydrous oxide film. Using equations for
the finite boundary conditions, which is equivalent from an
electrical point of view to the impedance of a finite trans-
mission line, an effective diffusion coefficient of 10−10 to
10−11 cm2 s−1 is obtained. The high-frequency charging is
found to be coupled to Faradaic charging at the film/solution
interface.
m2Z1(I1N − 1)
k1 =
(9)
(I1N + 1)
The ORR was investigated on sponataneosly oxidized
electrodeposited ruthenium using ring-disk technique. The
ORR proceeds mostly via four-electron transfer reaction to
water with the formation of a small percentage of hydro-
gen peroxide. The rate constants for the oxygen reduc-
tion according to the Damjanovic reaction scheme were
evaluated.
Fig. 5. The jDL/(jDL − jD) as a function of rotation rate at different potentials:
(1) 0.400 V, (2) 0.425 V, (3) 0.450 V, (4) 0.475 V, (5) 0.500 V, (6) 0.525 V,
(7) 0.550 V, (8) 0.575 V, (9) 0.600 V.

1164
M. Metikosˇ-Hukovic´ et al. / Electrochimica Acta 51 (2006) 1157–1164
[33] S. Ardizzone, A. Carugati, G. Lodi, S. Trasatti, J. Electrochem. Soc.
References
129 (1982) 1689.
[1] K. Kinoshita, Electrochemical Oxygen Technology, Wiley, NY,
1992.
[34] G.W. Jang, E.W. Tsai, K. Rajeshwar, J. Electroanal. Chem. 263
(1989) 383.
[2] S. Trasatti, Electrochim. Acta 36 (1991) 225.
[3] B.E. Conway, J. Electrochem. Soc. 138 (1991) 1539.
[4] T. Liu, W.G. Pell, B.E. Conway, Electrochim. Acta 42 (1997)
3541.
[5] C.-C. Hu, H.-R. Chiang, C.-C. Wang, J. Solid State Electrochem. 7
(2003) 477.
[6] S. Duron, R. Rivera-Noriega, P. Nkeng, G. Poillerat, O. Solorza-
Feria, J. Electroanal. Chem. 566 (2004) 281.
[7] H. Inoue, S.R. Brankovic´, J.X. Wang, R.R. Adzˇic´, Electrochim. Acta
47 (2002) 3777.
[8] N.A. Anastasijevic´, Z.M. Dimitrijevic´, R.R. Adzˇic´, Electrochim. Acta
31 (1986) 1125.
[9] J. Prakash, H. Joachin, Electrochim. Acta 45 (2000) 2289.
[10] N.A. Anastasijevic´, Z.M. Dimitrijevic´, R.R. Adzˇic´, Electrochim. Acta
37 (1992) 457.
[11] N.A. Anastasijevic´, Z.M. Dimitrijevic´, R.R. Adzˇic´, J. Electroanal.
Chem. 199 (1986) 351.
[35] B. Boukamp, Solid State Ionics 20 (1980) 31.
[36] C. Ho, I.D. Raistrick, R.A. Huggins, J. Electrochem. Soc. 127 (1980)
343.
[37] C.M. Elliot, A.B. Kopelove, W.J. Albery, Z. Chen, J. Phys. Chem.
95 (1991) 1743.
[38] S.H. Glarum, J.H. Marshal, J. Electrochem. Soc. 127 (1980) 1467.
[39] T. Amemiya, K. Hashimoto, A. Fujishima, J. Phys. Chem. 97 (1993)
9736.
[40] G. Inzelt, G. Lang, J. Electroanal. Chem. 378 (1994) 39.
[41] J.R. Macdonald (Ed.), Impedance Spectroscopy, Wiley, NY,
1987.
[42] S. Levine, A.L. Smith, Discuss. Faraday Soc. 52 (1967) 1290.
[43] I.D. Raistrick, R.J. Sherman, Proc. Electrochem. Soc. 142 (1987)
582.
[44] E.W. Tsai, K. Rajeshwar, Electrochim. Acta 36 (1991) 27.
[45] Y.V. Pleskov, V.Y. Filinovskii, The Rotating Disk Electrode, Plenum
Press, NY, 1976.
[12] R.R. Adzˇic´, in: J. Lipowski, P.N. Ross (Eds.), Electrocatalysis,
Wiley-VCH, NY, 1998, p. 197.
[46] K.-L. Hsueh, E.R. Gonzales, S. Srinivasan, Electrochim. Acta 28
(1983) 691.
[13] M. Metikosˇ-Hukovic´, R. Babic´, J. Piljac, J. New Mater. Electrochem.
Syst. 7 (2004) 179.
[14] S. Hadzˇi-Jordanov, H. Angerstein-Kozlowska, M. Vukovic´, B.E.
Conway, J. Phys. Chem. 81 (1977) 2271.
[47] K.-L. Hsueh, D.T. Chin, S. Srinivasan, J. Electroanal. Chem. 153
(1983) 79.
[48] A. Damjanovic, M.A. Genshav, J.OꢁM. Bockris, J. Chem. Phys. 45
(1966) 4057.
[15] S. Hadzˇi-Jordanov, H. Angerstein-Kozlowska, M. Vukovic´, B.E.
Conway, J. Electrochem. Soc. 128 (1978) 1471.
[16] V. Birss, R. Myers, H. Angerstein-Kozlowska, B.E. Conway, J. Elec-
trochem. Soc. 131 (1984) 1502.
Editor’s Note
[17] L.D. Burke, J.K. Mulcahy, J. Electroanal. Chem. 73 (1976) 207.
[18] L.D. Burke, T.O. OꢁMeara, J. Chem. Soc. Faraday Soc. 68 (1972)
839.
After correcting the proofs and with the corrected paper
available on the web, the Authors of this work realized that
appropriate reference to two previous papers was missing.
The two papers are:
[19] D. Michell, D.A.J. Rand, R. Woods, J. Electroanal. Chem. 89 (1978)
11.
ˇ
[20] M. Vukovic´, D. Cukman, J. Electroanal. Chem. 474 (1999) 167.
ˇ
[21] D. Marijan, D. Cukman, V. Vukovic´, M. Milun, J. Mater. Sci. 30
ˇ
´
´
A V. Horvat-Radosevic, K. Kvastek, M. Vukovic, D.
(1995) 3047.
ˇ
Cukman, J. Electroanal. Chem. 482 (2000) 188.
[22] M. Vukovic´, T. Valla, M. Milun, J. Electroanal. Chem. 356 (1993)
ˇ
´
B K. Kvastek, V. Horvat-Radosevic, J. Electroanal. Chem.
81.
[23] J. Rishpon, S. Gottesfeld, J. Electrochem. Soc. 131 (1984) 1960.
[24] V. Horvat-Radosˇevic´, K. Kvastek, M. Vukovic´, D. Cukman, J. Elec-
511 (2001) 65.
ˇ
These two references are to be introduced as follows:
troanal. Chem. 482 (2000) 188.
[25] L.D. Burke, O.J. Murphy, J. Electroanal. Chem. 96 (1979) 19.
[26] V.S. Bagotzky, A.M. Skundin, E.K. Tuseva, Electrochim. Acta 21
(1976) 29.
-p. 1 right, line 10 . . ..[4,5,14–25 + A + B]
-p. 2 left, line 6 . . ..[23,24 + A + B]
[27] R. Ko¨tz, S. Stucki, D. Scherson, D.M. Kolb, J. Electroanal. Chem.
172 (1984) 211.
[28] R.O. Lezina, N.R. de Tacconi, A.J. Arvia, J. Electroanal. Chem. 151
(1983) 193.
[29] M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solu-
tions, Pergamon Press, NY, 1966.
-p. 4 left, line 14 . . ..a hydrous oxide [19,23]. While. . ..
-p. 4 left, line 21 . . ..impedance analysis [23,24 + A + B]
-p. 4 right, line 9 . . ..in Fig. 3, which was developed for
electrodeposition of Ru [24]
-p. 4 right, line 14 . . ..EEC is as follows [24]:
[30] D. Galizzioli, F. Tantardini, S. Trasatti, J. Appl. Electrochem. 4
(1974) 57.
The Authors apologize for the inconvenience.
[31] K. Doblhofer, M. Metikosˇ, Z. Ogumi, H. Gerischer, Ber. Bunsenges.
Phys. Chem. 82 (1978) 1046.
[32] J.E. Weston, B.C.H. Steele, J. Appl. Electrochem. 10 (1980) 49.
S. Trasatti
Editor-in-Chief