Oscillatory peroxodisulfate reduction on Pt and Au electrodes under high ionic strength conditions, caused by the catalytic effect of adsorbed OH

Article abstract of DOI:10.1021/jp013275p

Four electrochemical oscillations of different types (named oscillations ?±, ?2, ?3, and ??), which are different from previously reported Frumkin-type oscillations in low ionic strength electrolytes, appear for peroxodisulfate (S₂O₈^2-) reduction on Pt and Au electrodes under high ionic strength conditions. Impedance measurements for oscillation ?3 that appears in a region of the most positive potentials have shown that it is classified as hidden negative differential resistance (hidden-NDR or HNDR) oscillators. Detailed electrochemical measurements have suggested that the NDR arises from a decrease (with a negative potential shift) in the surface coverage of adsorbed OH, which acts as a catalyst for the dissociative adsorption of S₂O₈^2- (the first step of its reduction). The NDR is hidden by the desorption (with the negative potential shift) of adsorbed sulfate (SO₄^2-) produced by the dissociative adsorption and reduction of S₂O₈^2-, which works as a site-blocking agent. Mathematical simulation based on this model has reproduced the appearance of oscillation ?3. The result supports our previously proposed concept of autocatalytic effect of adsorbed OH and catalytic effect of adsorbed halogen on the dissociative adsorption of H₂O₂, showing the generality of the electrocatalytic effect of adsorbed OH and halogen.

Full text of DOI:10.1021/jp013275p

J. Phys. Chem. B 2002, 106, 2287-2293
2287
Oscillatory Peroxodisulfate Reduction on Pt and Au Electrodes under High Ionic Strength
Conditions, Caused by the Catalytic Effect of Adsorbed OH
Shuji Nakanishi, Sho-ichiro Sakai, Michiru Hatou, Yoshiharu Mukouyama, and
Yoshihiro Nakato*
Department of Chemistry, Graduate School of Engineering Science, Osaka UniVersity, and
Research Center for Solar Energy Chemistry, Osaka UniVersity, Toyonaka, Osaka 560-8531, Japan
ReceiVed: August 24, 2001; In Final Form: NoVember 29, 2001
Four electrochemical oscillations of different types (named oscillations R, â, γ, and δ), which are different
from previously reported Frumkin-type oscillations in low ionic strength electrolytes, appear for peroxodisulfate
(S2O82-) reduction on Pt and Au electrodes under high ionic strength conditions. Impedance measurements
for oscillation γ that appears in a region of the most positive potentials have shown that it is classified as
hidden negative differential resistance (hidden-NDR or HNDR) oscillators. Detailed electrochemical
measurements have suggested that the NDR arises from a decrease (with a negative potential shift) in the
surface coverage of adsorbed OH, which acts as a catalyst for the dissociative adsorption of S2O82- (the first
step of its reduction). The NDR is hidden by the desorption (with the negative potential shift) of adsorbed
sulfate (SO42-) produced by the dissociative adsorption and reduction of S2O82-, which works as a site-
blocking agent. Mathematical simulation based on this model has reproduced the appearance of oscillation γ.
The result supports our previously proposed concept of autocatalytic effect of adsorbed OH and catalytic
effect of adsorbed halogen on the dissociative adsorption of H2O2, showing the generality of the electrocatalytic
effect of adsorbed OH and halogen.
2-
Introduction
on electrochemical oscillations in S2O8 reduction, to which
the EQCM and FTIR techniques will be applicable because of
Electrochemical reactions with nonlinear kinetics show a
variety of interesting dynamic behavior such as periodic and
chaotic oscillations.1-4 Mechanistic studies of electrochemical
oscillations have made rapid progress recently, resulting in
classification on the basis of the oscillation mechanisms.4,5 Very
recently, growing attention has been paid to the observation and
understanding of spatiotemporal patterns arising from nonlinear
kinetics, such as traveling chemical waves6-10 and Turing
patterns.11
2-
2-
the large mass of adsorbed S2O8 and SO4 and strong IR
absorption of adsorbed SO42- at 1100-1200 cm-1. The purpose
2-
of this paper is to find an oscillation mechanism in the S2O8
-
reduction system, similar to that found in the H2O2-reduction
system, which enables to use the above atomic level techniques
for the further confirmation of the mechanism.
The electrochemical oscillation in the S2O82- reduction was
first reported by Frumkin et al.,21 and since then, a lot of
investigations have been done on this system.22-27 However,
most studies have been made using electrolytes of low ionic
strength, in which the origin of the oscillations (or the negative
differential resistances) comes from the electrostatic repulsion
We have reported thus far12-19 that H2O2 reduction on Pt
electrodes in acidic solution shows at least five oscillations of
different types, depending on the electrode potential and the
current density, the atomic level flatness of the electrode surface,
and the presence or absence of halogen ions in solution. The
finding of such new oscillations led to the improvement of the
classification of electrochemical oscillations.15,16 Moreover, the
studies have revealed new mechanisms for electrochemical
2-
of a negatively charged electrode to electroactive anions, S2O8
(so-called the Frumkin effect). In this work, we used electrolytes
of high ionic strength, in which the Frumkin effect is neglected,
2-
and found that the S2O8 reduction on Pt or Au electrodes
shows four oscillations of different types (named oscillations
R, â, γ, and δ). Detailed studies have revealed that oscillation
γ, appearing in the most positive potentials, arises from the
catalytic effect of adsorbed OH on the dissociative adsorption
of S2O82-, similarly to oscillations C and E in the H2O2-
reduction system.
reactions themselves, such as an autocatalytic effect of adsorbed
12,17
OH on the dissociative adsorption of H2O2
and a catalytic
effect of adsorbed halogen on the same reaction,16,19 which will
be important not only for oscillation studies but also for
exploration of new active electrode materials for fuel cells. A
similar autocatalytic effect of adsorbed OH is reported for an
oscillation in H2O2 reduction on Ag electrodes.20
Experimental Section
Unfortunately, modern atomic level techniques for further
confirmation of the above new mechanisms, such as electro-
chemical quartz crystal microbalance (EQCM) and FTIR
techniques, are difficult to apply to adsorbed OH that is the
key intermediate of the H2O2 reduction. Thus, we started studies
Polycrystalline Pt (99.9% in purity) and Au (99.99%) disks
of about 6.0 mm in diameter and single-crystal Au (99.999%)
disks of about 2-3 mm in diameter were used as the working
electrode. The poly-Pt and Au disks were polished with 0.06
µm alumina slurry and immersed in a hot (HNO3 + H2O2)
solution for 10 min to remove surface contamination. Just before
* Corresponding author. Tel: +81-6-6850-6235. Fax: +81-6-6850-6236.
E-mail: nakato@chem.es.osaka-u.ac.jp.
10.1021/jp013275p CCC: $22.00 © 2002 American Chemical Society
Published on Web 02/09/2002

2288 J. Phys. Chem. B, Vol. 106, No. 9, 2002
Nakanishi et al.
measurements, cyclic potential scans were repeated between
-0.50 and +1.50 V vs SCE in 0.5 M HClO4 (M ) mol/dm3)
for about 30 min for further cleaning of the electrode surface.
Single-crystal Au(111), -(100), and -(110) electrodes with
atomically flat surfaces were prepared by the method of Clavilier
et al.28 The details of the preparation method were described in
a previous paper.17
Current density (j) vs potential (U) curves and j vs time (t)
curves were measured with a potentiogalvanostat (Nikko-
Keisoku, NPGS-301) and a potential programmer (Nikko-
Keisoku NPS-2), using a Pt plate (10 × 10 mm2) as the counter
electrode and a saturated calomel electrode (SCE) as the
reference electrode. The data were recorded with an X-Y
recorder (Riken-Denshi F-35C) or a data-storing system (in-
struNET, GW Instruments) with a sampling frequency of 1 kHz.
The electrolyte solutions were prepared using special grade
chemicals and pure water, the latter of which was obtained by
purification of deionized water with a Milli-Q water purification
system. The electrolyte solutions were kept stagnant during
measurements. The ohmic drops in the solution between the
working electrode and the reference electrode were not corrected
in this work.
Impedance measurements were carried out with a Solartron
1260 impedance analyzer combined with a Solartron 1287
electrochemical interface potentiostat. The amplitude of modula-
tion potential was 10 mV. Just before the measurements, the
electrode potential was held at a potential at which the
measurement was done, for about 180 s, to accomplish a steady
2-
distribution of the S2O8 concentration.
Results
Figures 1 and 2 summarize four electrochemical oscillations,
2-
named oscillations R, â, γ, and δ, observed for the S2O8
reduction on poly-Au (Figure 1) and poly-Pt (Figure 2)
electrodes. Figure 3 summarizes, for reference in later discus-
sion, reported oscillations for the H2O2 reduction on Pt.15,17
Figure 1a shows a j-U curve for poly-Au in 0.5 M HClO4
without S2O82-, which is included to show the potential region
of formation and disappearance of surface Au oxide. No current
peaks corresponding to hydrogen adsorption and desorption are
observed for Au, contrary to the case of Pt (cf. Figure 2a). When
0.5 M Na2S2O8 is added to the solution (Figure 1b), a cathodic
Figure 1. Current density (j) vs potential (U) curves for poly-Au under
(a, b) potential-controlled and (c, d) current-controlled conditions.
Electrolyte: (a) 0.5 M HClO4; (b-d) 0.5 M HClO4 + 0.5 M Na2S2O8.
Scan rate: (a) 100 mV/s; (b) 10 mV/s; (c) 1 mA/s; (d) 10 mA/s.
negative potentials from 0 to -0.6 V or further negative poten-
tials. Hydrogen evolution starts at about -0.8 V in Figure 1b.
Under current-controlled conditions for Au electrodes, three
oscillations (oscillations â, γ, and δ) appear, as shown in Figure
1c,d. It is to be noted that oscillation γ appears in a region of
the most positive potentials, i.e., in a potential region of reaction
1. On the other hand, oscillation â appears in a potential region
of reaction 2 and hydrogen evolution, similarly to oscillation B
in the H2O2-reduction system which also appears in a region of
hydrogen evolution (Figure 3).12,13,16 Zelin et al.32 reported an
2-
current due to the S2O8 reduction starts at about 0.4 V vs
SCE, just a potential at which the reduction of surface Au oxide
is nearly completed (Figure 1a). According to the work by
Mu¨ller et al.29,30 and Samec et al.26,31 using rotating Au elec-
trodes, S2O82- reduction on Au proceeds via two pathways. The
2-
first one is initiated by the dissociative adsorption of S2O8
probably followed by immediate reduction of adsorbed SO4
radicals,
,
•-
2-
oscillation for the S2O8 reduction on Ag in 1.0 M NaOH,
which may correspond to oscillation â in the present work.
Oscillation δ appears between oscillations γ and â when the
scan rate of j is made rapid (Figure 1d). Oscillation δ was
unstable under a constant current density and weakened in
amplitude with time, finally disappearing.
•-
S2O82- f 2SO4
(1a)
ad
2-
•-
SO4 ad + e- f SO4
(1b)
ad
2-
2-
This pathway gives a current near the onset of the S2O8
reduction in a region from +0.4 to 0 V (see Figure 1b). The
other pathway is a direct reduction of S2O82- via no adsorption,
Figure 2 shows j-U curves for poly-Pt. When 0.7 M S2O8
is added (Figure 2b), the S2O82--reduction current starts at about
0.2 V, at which the reduction of surface Pt oxide is nearly
completed (Figure 2a), and increases (in the absolute value) with
the decreasing potential, in a manner similar to the case of Au
electrodes. (Note that the current peaks in Figure 2a are not
observed in Figure 2b owing to a large difference in the current-
density scale.) Hydrogen evolution starts at about -0.4 V, and
in a narrow potential region just before it, oscillation R appears
2-
•-
2-
S2O8 aq + e- f SO4 aq + SO4
(2a)
(2b)
aq
SO4 aq + e- f SO4
•-
2-
aq
and the corresponding current is observed in a region of more

Oscillatory Peroxodisulfate Reduction
J. Phys. Chem. B, Vol. 106, No. 9, 2002 2289
Figure 2. j-U curves for poly-Pt under (a), (b), and (d) potential-
controlled and (c) current-controlled conditions. Electrolyte: (a) 0.5
M HClO4; (b) 0.5 M HClO4 + 0.7 M Na2S2O8; (c, d) 0.5 M HClO4 +
0.5 M Na2S2O8. Scan rate: (a) 100 mV/s; (b, d) 10 mV/s; (c) 1 mA/s.
An external resistance of 120 Ω is added for (d).
Figure 3. Reported five oscillations, called oscillation A-E, observed
for the H2O2 reduction on Pt in an acidic solution under (a-d) potential-
controlled and (e, f) current-controlled conditions. Electrode: poly-Pt
except that single-crystal Pt(111) is used for (b). Oscillation E is
observed only for Pt(111). Electrolyte: 0.3 M H2SO4 containing (a)
0.7 M H2O2, (b) 1.0 M H2O2, (c, e) 1.2 M H2O2 + 1.0 × 10-3 M KCl,
and (d, f) 0.7 M H2O2 + 1.0 × 10-4 M KBr.
(Figure 2b). This behavior of oscillation R is the same as that
of oscillation A in the H2O2-reduction system (Figure 3). We
reported previously12,13 that oscillation A arises from an NDR
due to the suppression of the H2O2 reduction by the forma-
tion of upd-H (underpotential-deposited hydrogen) of a nearly
full coverage. Oscillation R in Figure 2b may thus appear by
the same mechanism as oscillation A, a slight negative shift
of the potential region of oscillation R from that of the
upd-H in Figure 2a being attributed to the ohmic drop in
solution. This explanation is supported by the fact that oscillation
R appears only for Pt electrodes on which upd-H is formed and
not for Au.
Under current-controlled conditions for Pt electrodes, oscil-
lation γ is observed, similar to the case of Au, though
oscillations â and δ are not observed. Oscillation γ for Pt
appears in the most positive potentials. In addition, oscillation
γ appears under potential-controlled conditions if an external
resistance of about 120 Ω is added (Figure 2d). We can thus
say that oscillation γ is of a character similar to oscillation C
in the H2O2-reduction system, which is also observed under both
current- and potential-controlled conditions in a region of the
most positive potentials (Figure 3).
We made detailed studies on oscillation γ because it appears
initially and thus its mechanism may be understood most easily.
Figure 4 shows a waveform of oscillation γ as a potential
oscillation at -10 mA cm-2 in 0.7 M Na2S2O8 + 0.5 M HClO4.
The stable and periodic oscillation continued for about 10 min
or more. Figure 5 shows an impedance diagram for poly-Au in
0.7 M Na2S2O8 + 0.5 M HClO4 at -0.16 V vs SCE, near which
oscillation γ appears (Figure 1c). The diagram clearly shows
the presence of an HNDR (i.e., negative differential resistances
in a region of intermediate frequencies and positive differential
resistances in low frequencies), indicating that oscillation γ is
an HNDR oscillator. Oscillation C is also classified into HNDR
oscillators.15
In accordance with the above result, the j-U curves under
potential-controlled conditions, measured at a high scan rate of
100 mV/s, showed an NDR in a region of 0.2-0.0 V (Figure
6), though the NDR is not clear at a low scan rate (Figure 1b).
The NDR at a high scan rate was observed prominently for

2290 J. Phys. Chem. B, Vol. 106, No. 9, 2002
Nakanishi et al.
Figure 7. Effect of addition of SO42- to the solution on the j-U curves
for poly-Au under potential-controlled conditions. Electrolyte: 0.5 M
HClO4 + 0.7 M Na2S2O8 + x M Na2SO4, where x is indicated in the
figure. The scan rate is 100 mV/s.
Figure 4. Potential (U) vs time (t) curve at -10 mA cm-2 for poly-
Au in 0.5 M HClO4 + 0.7 M Na2S2O8.
to that for oscillation C. Why the NDR is observed only at high
scan rates and for atomically flat Au(111) (Figure 6) will be
explained in the next section.
Oscillation γ is classified into an HNDR oscillator, as
mentioned earlier (Figure 5). To clarify what is the NDR-hiding
species, we investigated the effect of addition of SO42- ions to
2-
the solution, which are a product of the S2O8 reduction
(reaction 1). Figure 7 shows that the NDR, observed in +0.05
to -0.05 V in the absence of SO42-, is gradually hidden with
the increasing concentrations of added SO42-, accompanied by
a decrease in the S2O82--reduction current, which strongly
2-
suggests that adsorbed SO4 is an NDR-hiding species.
Mathematical Simulation
Figure 5. Impedance diagram for poly-Au at -0.16 V vs SCE in 0.5
M HClO4 + 0.7 M Na2S2O8.
The experimental results indicate that (1) oscillation γ appears
in a potential region where the S2O8 reduction proceeds via
2-
its dissociative adsorption (reaction 1), (2) oscillation γ is an
HNDR oscillator, (3) the NDR is likely to arise form a catalytic
2-
effect of adsorbed OH on the dissociative adsorption of S2O8
,
and (4) the NDR-hiding species is most probably adsorbed
SO42- produced by reaction 1. Thus, there is a good parallelism
between oscillation γ and previously reported oscillation C,
which also appeared from the NDR due to the autocatalytic
effect of adsorbed OH on the dissociative adsorption of H2O2,
with adsorbed halogen as the NDR-hiding species.
The observation of the NDR for oscillation γ only at high
scan rates and for atomically flat Au(111) (Figure 6) can be
explained, on the basis of the reported mechanism15,17 in the
H2O2 reduction system, as follows: First, the surface coverage
(θOH) of adsorbed OH, acting as a catalyst for reaction 1a,
decreases with a negative potential scan, which leads to a
decrease in the S2O82--reduction current with the negative
potential scan, i.e., the appearance of an NDR. The observation
of the NDR only at high scan rates (or the disappearance of the
NDR at low scan rates) can be attributed to slow desorption of
Figure 6. Dependences of the j-U curves on (a) the crystal faces and
(b) the surface atomic flatness for single-crystal Au electrodes under
potential-controlled conditions. Electrolyte: 0.5 M HClO4 + 0.7 M
Na2S2O8. The scan rate is 100 mV/s.
2-
adsorbed SO4 produced by reaction 1. Namely, at low scan
rates, the decrease in the amount of adsorbed OH due to the
decrease in θOH is compensated by an increase in empty surface
sites by desorption of adsorbed SO42-. The observation of the
NDR only for flat Au(111) can be attributed to the effective
autocatalytic effect of adsorbed OH on this face,17,19 as explained
in more detail later (Figure 10).
We created a mathematical simulation for oscillation γ on
the basis of the above model. We took into account the catalytic
effect of adsorbed OH on reaction 1a by expressing the rate
constant k1a as follows:
atomically flat Au(111) but not for Au(100) and Au(110) (Figure
6a). The atomically roughened Au(111), prepared by repeated
electrochemical oxidation and rereduction,17,33 also showed no
NDR (Figure 6b). Quite the same behavior was observed for
the NDR in the H2O2-reduction system, from which oscillation
C as well as oscillation E appears. Namely, the NDR in the
H2O2-reduction system, originating from an autocatalytic effect
of adsorbed OH on the dissociative adsorption of H2O2, was
prominent for atomically flat Pt(111) but not for Pt(100) and
Pt(110) and atomically roughened Pt(111).17 This fact strongly
suggests that the NDR in Figure 6 appears by a mechanism
similar to that for the above NDR in the H2O2-reduction system,
or in other words, oscillation γ appears by a mechanism similar
k1a ) k1a0 + úθOH
(3)
Here k1a0 is a normal rate constant and ú is a proportional
constant. The following reactions were considered in addition

Oscillatory Peroxodisulfate Reduction
J. Phys. Chem. B, Vol. 106, No. 9, 2002 2291
Figure 8. (a) θOH vs E, calculated by eq 12, with a ) 45 V-1 and E50′
) +0.2 V vs SCE. (b) j vs U calculated by eqs 7-12 with k1a0 ) 4.5
× 10-3 cm s-1, ú ) 5, and a scan rate of 2 mA/s, under the condition
that the calculation is carried out within a range of j from 0 to -10
mA cm-2. (c) Improved j vs U calculated by adding contributions of
b
reaction 2 and hydrogen evolution. Other parameter values are CSO
)
Figure 9. Calculated E-t, θ1-t, and θ2-t curves at j ) -10 mA
0.7 × 10-3 mol cm-3, δ ) 0.015 cm, D ) 5.0 × 10-6 cm2 s-1, A )
0.28 cm2, CDL ) 2.0 × 10-5 F cm-2, N0 ) 2.2 × 10-9 mol cm-2, RΩ
) 10 Ω, T ) 300 K, Ri ) 0.25 for i ) 1b and 4 and 0.50 for others,
cm-2. The parameter values are the same as in Figure 8.
n ) 1, k1b0 ) 6.5 × 10-4 mol cm-2 s-1, E1b0 ) 1.0 V vs SCE, k20
)
1.0 × 10-8 cm s-1, E20 ) 0.4 V, k40 ) 3.0 × 10-9 mol cm-2 s-1, and
E40 ) 0.5 V.
to the aforementioned reactions 1a,b:
2-
2-
SO4 ad f SO4
(4)
aq
OHad + H+ + e- T H2O (in the presence of S2O82-) (5)
Figure 10. Schematic illustration of a catalytic effect of adsorbed OH
2-
for the dissociative adsorption of S2O8
.
The rate constant for electrochemical reaction i, ki, was given
by the Butler-Volmer equation with the conventional nota-
tions:12,14
thickness of a diffusion layer for S2O82-, and N0 the total amount
of surface sites/unit area.
2-
The site-blocking effect of adsorbed SO4 (slow species),
ki ) ki0 exp[-RnF(E - Ei0)/RT]
(6)
together with those of adsorbed SO4•- and OH, was taken into
account by introducing such terms as (1 - θ1 - θ2 - θOH) in
the rate equations. The rate of reaction 5 was not explicitly
considered in eq 8, because reaction 5 is actually a complex
reaction, affected by a certain action of S2O82- ions (as discussed
later). The calculation of j-U curves with the neglect of reaction
5 reproduced well the observed curves, as mentioned later,
indicating that reaction 5 makes no substantial contribution to
the current.
The differential equations for a current balance, mass bal-
ances, and rate equations were given in the same way as in
previous work,12,14-19
I ) jA ) (U - E)/RΩ ) IC + IF ) ACDL(dE/dt) + IF (7)
IF ) AF(-k1bθ1)
(8)
The θOH vs E relation can be expressed, if we can assume
that reaction 5 is in equilibrium at every (true) electrode potential
(E), in the form12
(δ/2) dCSOs/dt ) (D/δ)(CSOb - CSOs) -
k1aCSOs(1 - θ1 - θ2 - θOH)2 (9)
θOH ) 1/[1 + exp{-a(E - E50′)}]
(12)
N0 dθ1/dt ) 2k1aCSOs(1 - θ1 - θ2 - θOH)2 - k1bθ1 (10)
where a equals (RT/nF) and E50′ refers to the apparent
equilibrium potential for reaction 5 in the presence of S2O8
2-
N0 dθ2/dt ) k1bθ1 - k4θ2
(11)
.
E50′ will be somewhat more negative than the (true) equilibrium
potential, E50, for reaction 5 in the absence of S2O82-. In the
present work, E50′ was estimated from the reported experiment34
discussed later and a was taken as a parameter. Figure 8a shows
a θOH vs E relation calculated with a ) 45 V-1 and E50′ )
•-
where θ1 and θ2 are the surface coverages of adsorbed SO4
and SO42-, respectively, and CSOs and CSOb the concentrations
2-
of S2O8 at the electrode surface and in the solution bulk,
respectively. D is the diffusion coefficient for S2O82-, δ the

2292 J. Phys. Chem. B, Vol. 106, No. 9, 2002
Nakanishi et al.
2-
+0.2 V vs SCE. The θOH changes from nearly one to nearly
zero in a range of E from about 0.3 to 0.1 V.
The effect of addition of SO4 ions to the solution (Figure
7) shows that adsorbed SO42- acts as the NDR-hiding species.
Oscillation γ appears without any addition of SO42-, because
Figure 8b shows a j-U curve calculated by use of eqs 7-12
under the condition that the calculation is carried out within a
range of j from zero to -10 mA cm-2. The curve reproduces
the appearance of oscillation γ as a potential oscillation. If the
above condition is removed, computers cannot finish calcula-
tions, not giving any j-U curves, because reactions 1, 4, and 5
cannot give current densities higher (in the absolute value) than
about 10 mA cm-2. This is because, at more negative poten-
tials than the calculated potential region in Figure 8b, θOH
approaches zero and k1a becomes quite small (k1a ≈ k1a0), leading
to small j.
adsorbed SO42-, produced by the S2O8 reduction (reaction
2-
1b), can act as the NDR-hiding species, as mentioned before.
This is an interesting example, for, in most cases, adsorbed
species as the NDR-hiding species are supplied by the adsorption
of solution species added.5
On the basis of the results of mathematical simulation (Figures
8 and 9), oscillation γ can be explained as follows. In a high-
potential state of a potential oscillation at a constant current
density, the surface coverage of adsorbed OH (θOH) is large
and, thus, k1a . k1a0; i.e., the S2O82- reduction occurs efficiently,
which leads to an increase in the coverage (θ2) of adsorbed
Figure 8c shows an improved j-U curve obtained by adding
2-
2-
2-
contributions of reaction 2, the second pathway for the S2O8
SO4 as a product of the S2O8 reduction owing to slow
2-
reduction, and hydrogen evolution.
desorption of adsorbed SO4 (slow species). The increase in
θ2 leads to a decrease in vacant surface sites at which the
H+ + e- T Had
2Had f H2
(13)
(14)
2-
dissociative adsorption of S2O8 occurs, thus leading to a
2-
negative shift of the electrode potential (E) to promote the SO4
desorption and keep a constant current density (j). The negative
shift in E, in turn, leads to a decrease in θOH and hence a
decrease in k1a ()k1a0 + θOH), which leads to a further negative
shift in E to keep a constant j. Here is a positive feedback
mechanism, and this causes a transition to a low-potential state.
In this case, eqs 8-10 are modified as follows, together with
the addition of another differential eq 15 for the coverage (θH)
of adsorbed hydrogen,
2-
IF ) AF{-k1bθ1 - 2k2CSOs (1 - θ1 - θ2 - θOH - θH) -
k13CHs (1 - θ1 - θ2 - θOH - θH) + k-13θH} (8′)
In a low-potential state, the desorption of adsorbed SO4
becomes fast, which leads to an increase in vacant surface sites
for the S2O82- reduction and to a positive shift in E. The positive
shift in E leads to an increase in θOH and k1a and to a further
positive shift in E to keep a constant j. Here is also a positive
feedback mechanism, which leads to a transition to a high-
potential state.
(δ/2) dCSOs/dt ) (D/δ)(CSOb - CSOs) - k1aCSOs (1 - θ1 -
θ2 - θOH - θH)2 - k2CSOs (1 - θ1 - θ2 - θOH - θH) (9′)
The NDR observed for the S2O82- reduction strongly depends
on the crystal faces of single-crystal Au electrodes and the
atomic flatness at the surface (Figure 6). The NDR arising from
the autocatalytic effect of adsorbed OH in the H2O2-reduction
system showed quite the same behavior,17 as already mentioned.
We can assume that the dissociative adsorption of negatively
N0 dθ1/dt ) 2k1aCSOs(1 - θ1 - θ2 - θOH - θH)2 - k1bθ1
(10′)
s
N0 dθH/dt ) k13CH (1 - θ1 - θ2 - θOH - θH) -
2
k-13θH - 2k14θH (15)
2-
charged S2O8 occurs rapidly on positively polarized metal
where k13 and k-13 are the rate constants for the forward and
backward processes of reaction 13, respectively, and CHs is the
surface concentration of H+ ions. The j-U curve in Figure 8c
is in good agreement with the observed curves (Figures 1c and
2c), except the appearance of oscillation â. Figure 9 shows a
calculated E-t curve at a constant current density of -10 mA
cm-2, together with θ1-t and θ2-t curves. The calculated E-t
reproduces the essential feature of the waveform of oscillation
γ as a potential oscillation (Figure 4).
atoms in the neighborhood of adsorbed OH (Figure 10), similar
to the case of dissociative adsorption of H2O2.17 The difference
in the ú value in eq 3 among crystal faces of Au electrodes can
be attributed to the difference in the number of the positively
polarized metal atoms around adsorbed OH, as reported.17,19
Finally, let us consider briefly oscillations other than oscil-
lation γ. Oscillation R is of a character similar to oscillation A
in the H2O2-reduction system, as mentioned before. This implies
that the NDR for oscillation R arises from the suppression of
2-
the S2O8 reduction by the formation of upd-H. However, a
Discussion
2-
question remains on how the upd-H suppresses the S2O8
reduction, because, in a potential region of oscillation R, the
S2O82- reduction is reported to proceed via no adsorption.31 For
oscillations â and δ, impedance measurements revealed that they
were of an HNDR-type, though further details are unclear at
present. In solutions of high ionic strengths in this work,
oscillations due to the Frumkin effect26,27 are not observed. On
the other hand, oscillations R, â, γ, and δ will in principle be
observed in solutions of low ionic strengths, though the
appearance of oscillations depends on detailed conditions in
various factors. In fact, oscillation R is observed even in 1.0
mM Na2S2O8.
The experimental results and mathematical simulation have
shown that oscillation γ is an HNDR oscillator and the NDR
arises from the catalytic effect of adsorbed OH on the dissocia-
tive adsorption of S2O82-. Here arises a question whether the
adsorbed OH can exist even in a potential region of oscillation
γ, which is considerably more negative than the current peak
for the reduction of surface Au or Pt oxide (Figures 1a and 2a).
It is reported by surface-enhanced Raman spectroscopy (SERS)34
that, in the presence of 1-20 mM S2O82-, oxygen-containing
adsorbed species (most probably attributable to adsorbed OH)
survive on Au down to potentials 0.3-0.4 V more negative than
2-
2-
the reduction peak of surface oxide. The strong oxidant, S2O8
,
In conclusion, this work has revealed that the S2O8
is likely to induce chemical (or electrochemical) oxidation of
the Au surface leading to formation of adsorbed OH, though
the detailed mechanism is unknown.
reduction on Pt and Au in high ionic strength electrolytes shows
four oscillations of different kinds, named oscillations R, â, γ,
and δ. Detailed studies on oscillation γ showed that adsorbed

Oscillatory Peroxodisulfate Reduction
J. Phys. Chem. B, Vol. 106, No. 9, 2002 2293
(15) Mukouyama, Y.; Nakanishi, S.; Chiba, T.; Murakoshi, K.; Nakato,
Y. J. Phys. Chem. B 2001, 105, 7246.
(16) Mukouyama, Y.; Nakanishi, S.; Konishi, H.; Ikeshima, Y.; Nakato,
Y. J. Phys. Chem. B 2001, 105, 10905.
(17) Nakanishi, S.; Mukouayma, Y.; Karasumi, K.; Imanishi, A.; Furuya,
N.; Nakato, Y. J. Phys. Chem. B 2000, 104, 4181.
(18) Nakanishi, S.; Mukouayama, Y.; Nakato, Y. J. Electrochem. Soc.
2001, 148, E405.
(19) Nakanishi, S.; Mukouyama, Y.; Nakato, Y. J. Phys. Chem. B 2001,
105, 5751.
(20) Fla¨tgen, G.; Wasle, S.; Lu¨bke, M.; Eickes, C.; Radhakirishnan, G.;
Doblhofer, K.; Ertl, G. Electrochim. Acta 1999, 44, 4499.
(21) Frumkin, A. N. Z. Electrochem. 1955, 59, 807.
(22) Gokhshtein, A. Y.; Frumkin, A. N. Dokl. Akad. Nauk. SSSR 1960,
132, 388.
(23) Frumkin, A. N.; Petti, O. A.; Nikolaeva-Fedorovich, N. V. Dokl.
Akad. Nauk. SSSR 1961, 136, 1158.
(24) Koper, M. T. M. Ber. Bunsen-Ges. Phys. Chem. 1996, 424, 497.
(25) Fla¨tgen, G.; Krischer, K.; Ertl, G. J. Electroanal. Chem. 1996, 409,
183.
(26) Treindl, L.; Doblhofer, K.; Krischer, K.; Samec, Z. Electrochim.
Acta 1999, 44, 3963.
(27) Wolf, W.; Ye, J.; Purgand, M.; Eiswirth, M.; Doblhofer, K. Ber.
Bunsen-Ges. Phys. Chem. 1992, 96, 278.
(28) Clavilier, J.; Faure, R.; Guinet, G.; Durand, D. J. Electrochem. Soc.
1980, 107, 205.
(29) Mu¨ller, L. J. Electroanal. Chem. 1967, 13, 275.
(30) Mu¨ller, L.; Wetzel, W.; Otto, H. J. Electroanal. Chem. 1970, 24,
175.
(31) Samec, Z.; Doblhofer, K. J. Electroanal. Chem. 1994, 367, 141.
(32) Zelin, L.; Tinghua, W.; Ke, C. Chin. J. Chem. Phys. 1999, 12, 251
(in Chinese).
(33) Itaya, K.; Sugawara, S.; Sashikata, K.; Furuya, N. J. Vac. Sci.
Technol. 1990, A8 (1), 515.
OH has a catalytic effect on the dissociative adsorption of
S2O82-, leading to the appearance of oscillation γ. The present
results as well as the reported ones on the H2O2 reduction17,19
show the generality of an autocatalytic (or a catalytic) effect of
adsorbed OH on electrochemical reactions.
References and Notes
(1) Hudson, J. L.; Tsotsis, T. T. Chem. Eng. Sci. 1994, 49, 1493.
(2) Fahidy, T. Z.; Gu, Z. H. Modern Aspects of Electrochemistry; White,
R. E., Bockris, J. O’M., Conway, R. E., Eds.; Plenum: New York, 1995;
Vol. 27, p 383.
(3) Koper, M. T. M. AdVances in Chemical Physics; Prigogine, I., Rice,
S. A., Eds.; John Willey & Sons: New York, 1996; Vol. 92; p 161.
(4) Krischer, K. Modern Aspects of Electrochemistry; White, R. E.,
Bockris, J. O’M., Conway, R. E., Eds.; Plenum: New York, 1999; Vol.
32, p 1.
(5) Strasser, P.; Eiswirth, M.; Koper, M. T. M. J. Electroanal. Chem.
1999, 478, 50.
(6) Fla¨tgen, G.; Krischer, K.; Pettinger, B.; Doblhofer, K.; Junkes, H.;
Ertl, G. Science 1995, 269, 668.
(7) Kiss, I. Z.; Gaspar, V.; Hudson, J. L. J. Phys. Chem. B 2000, 104,
7554.
(8) Christoph, J.; Otterstedt, R. D.; Eiswirth, M.; Jaeger, N. I.; Hudson,
J. L. J. Chem. Phys. 1999, 110, 8614.
(9) Strasser, P.; Christoph, J.; Lin, W. F.; Eiswirth, M.; Hudson, J. L.
J. Phys. Chem. A 2000, 104, 1854.
(10) Mazouz, N.; Krischer, K.; Fla¨tgen, G.; Ertl, G. J. Phys. Chem. B
1997, 101, 2403.
(11) Li, Y. J.; Oslonovitch, J.; Mazouz, N.; Plenge, F.; Krischer, K.;
Ertl, G. Science 2001, 291, 2395.
(12) Mukouyama, Y.; Nakanishi, S.; Nakato, Y. Bull. Chem. Soc. Jpn.
1999, 72, 25736.
(13) Matsuda, T.; Hommura, H.; Mukouyama, Y.; Yae, S.; Nakato, Y.
J. Electrochem. Soc. 1997, 144, 1988.
(14) Mukouyama, Y.; Nakanishi, S.; Konishi, H.; Nakato, Y. J.
Electroanal. Chem. 1999, 473, 156.
(34) Desilvestro, J.; Weaver, M. J. J. Electroanal. Chem. 1987, 234,
237.