Appearance of an Oscillation through the Autocatalytic Mechanism by Control of the Atomic-Level Structure of Electrode Surfaces in Electrochemical H2O2 Reduction at Pt Electrodes

Article abstract of DOI:10.1021/jp993961c

The influence of the atomic-level structure of electrode surfaces on electrochemical oscillations has been studied in a system of H₂O₂ reduction on Pt electrodes in acidic solutions. A current oscillation of another type, named oscillation E, has been found to appear for an atomically flat single-crystal Pt(111) electrode, in addition to previously reported oscillations, named oscillations A and B. Oscillation E does not appear for atomically flat Pt(100), Pt(110), polycrystalline Pt, and Pt(111) with atomically nonflat surfaces. Mathematical simulation by use of a model including an autocatalytic effect of adsorbed OH for dissociative adsorption of H₂O₂, as a possible explanation, has reproduced the appearance of oscillation E, as well as observed correlations between the appearance of oscillation E and the magnitudes of H₂O₂-reduction current and negative resistance. It is discussed that an efficient autocatalytic mechanism works at the atomically flat Pt(111) surface, which is responsible for the appearance of oscillation E at this surface.

Full text of DOI:10.1021/jp993961c

J. Phys. Chem. B 2000, 104, 4181-4188
4181
Appearance of an Oscillation through the Autocatalytic Mechanism by Control of the
Atomic-Level Structure of Electrode Surfaces in Electrochemical H2O2 Reduction at Pt
Electrodes
†
†
†
†
Shuji Nakanishi, Yoshiharu Mukouyama, Kouhei Karasumi, Akihito Imanishi,
‡
,†
Nagakazu Furuya, and Yoshihiro Nakato*
Department of Chemistry, Graduate School of Engineering Science, Osaka UniVersity, Toyonaka,
Osaka 560-8531, Japan, and Department of Applied Chemistry, Faculty of Engineering,
Yamanashi UniVersity, Takeda, Yamanashi 400-8510, Japan
ReceiVed: NoVember 8, 1999; In Final Form: February 14, 2000
The influence of the atomic-level structure of electrode surfaces on electrochemical oscillations has been
studied in a system of H reduction on Pt electrodes in acidic solutions. A current oscillation of another
2 2
O
type, named oscillation E, has been found to appear for an atomically flat single-crystal Pt(111) electrode, in
addition to previously reported oscillations, named oscillations A and B. Oscillation E does not appear for
atomically flat Pt(100), Pt(110), polycrystalline Pt, and Pt(111) with atomically nonflat surfaces. Mathematical
simulation by use of a model including an autocatalytic effect of adsorbed OH for dissociative adsorption of
2 2
H O , as a possible explanation, has reproduced the appearance of oscillation E, as well as observed correlations
between the appearance of oscillation E and the magnitudes of H -reduction current and “negative” resistance.
2 2
O
It is discussed that an efficient autocatalytic mechanism works at the atomically flat Pt(111) surface, which
is responsible for the appearance of oscillation E at this surface.
Introduction
Chemical and electrochemical oscillations are attractive
phenomena from a viewpoint of dynamic self-organization of
molecular systems. Electrochemical oscillations have been
reported for a variety of systems as summarized in recent
reviews.1 Recently low-index single-crystal metal electrodes
with well-defined surface structures have been used for mecha-
-3
nistic studies of oscillations.4
-10
Ertl et al. reported that the
4
oscillatory patterns for H2 oxidation on Pt in the presence of
2
+
Cu and halide ions depended on the crystallographic surface
structures of Pt electrodes, and explained it to be due to a
difference in the binding strength of halogen atoms on different
single-crystal surfaces.
We have been studying electrochemical oscillations for H2O2
reduction on Pt electrodes in acidic solutions.11-16 Interestingly,
various oscillations of different types, named oscillations A, B,
C, and D, have been found to appear. Oscillation A is observed
in a potential region just before hydrogen evolution,1 whereas
oscillation B is observed in a potential region of hydrogen
1,12
evolution.11,12
Figure 1. An electrochemical cell used for measurements of current
oscillations: WE, working electrode (single-crystal or polycrystalline
Pt disk); CE, counter electrode (Au plate); RE, reference electrode (a
Oscillations C and D are observed in the presence
-
-
-
of a small amount of halide ions such as Cl , Br , and I in
the electrolyte.1
3-15
Moreover, the oscillation period for oscil-
2 4 4
hydrogen electrode, composed of platinized Pt, 1.0 M H SO or HClO ,
lation A is modulated by adsorption of inert metal atoms, such
and 1 atm of hydrogen gas).
1
6
as Cu, Au, Ag, and Ru, to submonolayer amounts. These
results indicate that the oscillations in the “H2O2-acid-Pt
electrode” system are sensitive to the atomic-level structures
of electrode surfaces.
but also the reaction mechanism, resulting in an appearance of
another type of oscillation, named oscillation E, at the crystal-
lographically most-flat Pt(111) surface. The result is important
for deeper understanding of the mechanisms of electrochemical
reactions themselves as well as the oscillation mechanisms.
In the present paper, we report that the atomic-level structures
of the electrode surfaces affect not only the adsorption behavior
*
To whom correspondence should be addressed. Phone: +81-6-6850-
Experimental Section
6
235. Fax: +81-6-6850-6236. E-mail: nakato@chem.es.osaka-u.ac.jp.
†
‡
The electrochemical cell used is schematically illustrated in
Figure 1. A half-spherical single-crystal or polycrystalline Pt
Osaka University.
Yamanashi University.
1
0.1021/jp993961c CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/05/2000

4182 J. Phys. Chem. B, Vol. 104, No. 17, 2000
Nakanishi et al.
disk electrode (about 1-2 mm in diameter), used as the working
electrode (WE), was held such that only the flat surface was in
contact with the electrolyte by use of its meniscus. The
electrolyte was 0.3 M H2SO4 or HClO4 (pH 0.55) with or
3
without H2O2 (M ) mol/dm ). A gold plate was used as the
counter electrode (CE). A hydrogen electrode, composed of a
platinized platinum electrode, 1.0 M H2SO4 or HClO4, and
highly pure hydrogen gas, was used as the reference electrode
(RE). Either 1.0 M H2SO4 or HClO4 in the hydrogen electrode
was chosen so that the common acid was used in the WE and
RE compartments. The WE and RE compartments were
separated with a cation exchange membrane (Nafion 117).
Single-crystal Pt(111), Pt(100), and Pt(110) electrodes with
atomically flat surfaces were prepared by the method of Clavilier
1
7
et al. Single-crystal Pt spheres were prepared by heating Pt
wires (99.97% in purity) in a hydrogen flame. They were cut
in parallel to the Pt(111), Pt(100), and Pt(110) faces, and
polished with diamond slurry. They were then annealed in a
hydrogen flame for 30 min for Pt(111), and 180 min for Pt-
(100) and Pt(110). Just before measurements, all the electrodes
were annealed again in a hydrogen flame for 60 s and
immediately quenched in Ar-bubbled pure water. Single-crystal
Pt(111) electrodes with atomically nonflat surfaces were also
used for reference. The details of their preparation will be
described later. Polycrystalline Pt electrodes, having nearly the
same shape as the single-crystal Pt electrodes, were also used.
They were prepared by cutting and polishing of polycrystalline
spheres, which were obtained by heating Pt wires in a hydrogen
flame.
Aqueous electrolytes 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. Current density (j) vs potential (U) and j vs time (t)
were measured with a potentiogalvanostat (Nikko-Keisoku
NPGS-301) and a potential programmer (Nikko-Keisoku NPS-
Figure 2. Current density (j) vs potential (U) curves for a polycrys-
talline Pt (poly-Pt) electrode, obtained under potential-controlled
conditions. Electrolyte: (a) 0.3 M H
0.3 M H SO , (c) 0.7 M H + 0.3 M H
mV/s, (b, c) 10 mV/s.
2
SO
4
(pH 0.55), (b) 0.2 M H
2 2
O
+
2
4
2
O
2
2
SO . Scan rate: (a) 100
4
more negative than ca. -0.02 V. The current peaks for upd-H
are missing in Figure 2b because the j-scale in Figure 2b is
much larger than that in Figure 2a. It is to be noted that the
H2O2-reduction current is independent of the electrode potential
in a region above 0.07 V, whereas it shows the “negative slope”
in a region from +0.07 to -0.02 V. The potential-independent
H2O2-reduction current is explained by assuming that the H2O2
reduction is initiated by dissociative adsorption of H2O2
2
). They were either recorded with an X-Y recorder (Riken-
Denshi F-35C) or stored digitally at 1 kHz with a Mac ADIOS
II/16 (GW Instruments). Ohmic drops in the solution are not
corrected in the present work. Argon gas was bubbled through
the electrolyte in the working-electrode compartment for 30 min
to remove dissolved oxygen when j-U curves for adsorption
and desorption of hydrogen atoms (upd-H) were measured in
k1
2
Pt + H O
982Pt-OH
(1)
2
2
0
.3 M H2SO4 or HClO4. No bubbling of argon gas was carried
followed by electrochemical reduction of the resultant Pt-OH
out in solutions containing H2O2, because oxygen is more or
less evolved by decomposition of H2O2 on Pt. The structure of
electrode surfaces was inspected with a scanning tunneling
microscope (Nanoscope IIIa, Digital Instruments), using a Pt-
Ir wire as a tip. The tunnel current was held at 1 nA, and the
scan rate was 5.086 Hz.
k2
+
-
Pt-OH + H + e
98Pt + H O
(2)
2
and that the former (reaction 1) is a rate-determining step.15
The “negative” slope (or “negative” resistance) is caused by
the suppression of the dissociative adsorption of H2O2 on Pt by
formation of upd-H of a nearly full coverage.1
1-13
This
Results
explanation is in harmony with the fact that the negative slope
starts at a potential of the most negative end of the current peaks
for upd-H. Figure 2c shows a j-U curve in 0.3 M H2SO4
containing high-concentration (0.7 M) H2O2. Current oscilla-
Figure 2 shows current density (j) vs potential (U) curves
for a polycrystalline Pt electrode (hereafter abbreviated as a poly-
Pt electrode), observed under potential-controlled conditions.
Figure 2a is a curve in 0.3 M H2SO4 not containing H2O2 (pH
1
1
tions, called oscillations A and B, are both observed. Oscil-
lation B is less pronounced in Figure 2c than in our previous
0
.55). Two current peaks on both cathodic and anodic sides
1
1
are attributed to formation and disappearance of electrochemi-
work probably because a meniscus electrode (see Figure 1) is
used in the present work instead of a wholly immersed electrode
in the previous work.
cally adsorbed hydrogen atoms, called upd-H (under-potential
deposited hydrogen).1
1,18,19
The appearance of such two peaks
on both sides are characteristic of poly-Pt electrodes. Figure
b is a curve in 0.3 M H2SO4 containing low-concentration (0.2
M) H2O2. A current in a potential region from about 0.65 to
.00 V vs NHE is attributed to H2O2 reduction, which starts at
about 0.80 V. A hydrogen evolution current appears in potentials
Figure 3 shows j-U curves for an atomically flat Pt(111)
electrode, drawn in the same way as Figure 2. It is well-known17
that the Pt(111) electrode shows broad current peaks in a
potential region from ca. 0.35 to 0.05 V vs NHE due to the
formation and disappearance of upd-H, as shown in Figure 3a.
2
0

Electrochemical H2O2 Reduction at Pt Electrodes
J. Phys. Chem. B, Vol. 104, No. 17, 2000 4183
Figure 3. j-U curves for an atomically flat Pt(111) electrode, drawn
in the same way as Figure 2. Electrolyte: (a) 0.3 M H
2
SO
SO , (d) 1.0 M
4
. Scan rate: (a) 100 mV/s, (b, c) 10 mV/s, (d) 5
4
, (b) 0.2 M
Figure 5. (a-c) j-U curves for an atomically nonflat Pt(111) electrode
prepared by a procedure of repeated electrochemical oxidation and
rereduction, drawn in the same way as Figures 2 and 3. Electrolyte:
H
H
2
O
O
2
+ 0.3 M H
2
SO
4
, (c) 0.7 M H
2
O
2
+ 0.3 M H
2
4
2
2
+ 0.3 M H SO
2
mV/s.
(
a) 0.3 M H
.3 M H SO
illustration of the surface structure of atomically flat and nonflat Pt-
111) electrodes.
2
SO
4
, (b) 0.2 M H
2
O
2
+ 0.3 M H
2 4 2 2
SO , (c) 0.7 M H O +
0
2
4
. Scan rate: (a) 100 mV/s, (b, c) 10 mV/s. (d) Schematic
(
literature,17 clearly shows that a well-defined atomically flat Pt-
111) surface is obtained in the present work. This conclusion
(
was also confirmed by scanning tunneling microscopy (STM)
inspection, which gave an image with a clear “step and terrace”
structure, as displayed in Figure 4, similar STM images are
2
2
reported in the literature.
Figure 3b shows that a negative resistance is observed for
the atomically flat Pt(111) electrode in 0.3 M H2SO4 containing
low-concentration (0.2 M) H2O2, similarly to the case of Figure
2b. However, the negative resistance in Figure 3b shows a two-
step structure, and its potential range extends to a more positive
potential of around 0.35 V. As explained later, the negative
resistance in a region from 0.10 to 0 V is attributed to the
formation of upd-H, similarly to the case of Figure 2b, whereas
that in a region from 0.35 to 0.10 V is attributed to another
reason. In harmony with the appearance of the new negative
resistance in 0.35-0.10 V, a new current oscillation, named
oscillation E, appears in potentials of about 0.22-0.13 V in
cases where the H2O2 concentration is made high, 0.7 M (Figure
3c). The conclusion that oscillation E is distinguished from
oscillation A and truly a new oscillation is clearly shown by
the fact that oscillation A itself appears in a region of reasonable
Figure 4. An STM image of the surface of an atomically flat Pt(111)
electrode used in the present work.
Anodic and cathodic current peaks in a potential region from
ca. 0.55 to 0.35 V with sharp spikes at about 0.47 V in Figure
-
20,21
3
a are attributed to adsorption and desorption of HSO4 .
-
The observation of the upd-H and adsorbed-HSO4 peaks
characteristic of the Pt(111) electrode, in agreement with the

4184 J. Phys. Chem. B, Vol. 104, No. 17, 2000
Nakanishi et al.
Figure 7. j-U curves for an atomically flat Pt(110) electrode, drawn
Figure 6. j-U curves for an atomically flat Pt(100) electrode, drawn
in the same way as Figure 2. Electrolyte: (a) 0.3 M H SO , (b) 0.2 M
SO , (d) 1.0 M
. Scan rate: (a) 100 mV/s, (b, c) 10 mV/s, (d) 5
in the same way as Figure 2. Electrolyte: (a) 0.3 M H
2
SO
SO , (d) 1.0 M
4
. Scan rate: (a) 100 mV/s, (b, c) 10 mV/s, (d) 5
4
, (b) 0.2 M
2
4
H
H
2
O
O
2
+ 0.3 M H
2
4
SO , (c) 0.7 M H
2
O
2
+ 0.3 M H
2
4
H
H
2
O
O
2
+ 0.3 M H
2
SO
4
, (c) 0.7 M H
2
O
2
+ 0.3 M H
2
4
2
2
+ 0.3 M H SO
2
2
2
+ 0.3 M H SO
2
4
mV/s.
mV/s.
suggesting that the atomic nonflatness of the atomically nonflat
Pt(111) was restricted only to a narrow surface region (see
Figure 5d).
potentials (-0.05 to -0.07 V) in addition to oscillation E when
the H2O2 concentration is made much higher (Figure 3d).
To investigate whether the appearance of oscillation E is due
to an effect of the atomic-level flatness of the electrode surface,
atomically nonflat Pt(111) electrodes were prepared and exam-
Figures 6 and 7 show j-U curves for atomically flat Pt(100)
and Pt(110) electrodes, respectively, drawn in the same way as
Figures 2 and 3. The observed cathodic and anodic current peaks
attributed to upd-H (Figures 6a and 7a) agreed with those
reported for well-defined atomically flat Pt(100) and Pt(110),23
indicating that the electrodes in the present work had well-
defined atomic structures. The j-U curves for both the Pt(100)
and Pt(110) are rather similar to those for poly-Pt (Figure 2)
than those for atomically flat Pt(111) (Figure 3). However, Pt-
2
2
ined. Itaya et al. reported, by in situ STM experiments, that
surface roughening, such as schematically shown in Figure 5d,
is caused by a procedure of oxidation and rereduction of an
atomically flat Pt(111) surface. In the present work, the surface
roughening was carried out by shifting the electrode potential
stepwise to 1.75 V, where platinum oxide was formed and
oxygen evolution occurred. The electrode potential was kept at
(100) electrodes showed the negative resistance of a two-step
1
.75 V for 3 min, and then shifted back to 0.65 V, where the
structure (Figure 6b) though much less prominent than for Pt-
surface platinum oxide was reduced and removed. This proce-
dure was repeated three times. Parts a-c of Figure 5 show j-U
curves for an atomically nonflat Pt(111) electrode thus prepared,
drawn in the same way as Figures 2 and 3. It is seen that the
curves in Figure 5 are more similar to those for poly-Pt (Figure
(
(
111) (Figure 3b). In harmony with this result, some of the Pt-
100) electrodes showed oscillation E in addition to oscillation
A when the potential sweep was in the positiVe direction.
Figure 8 shows current (j) vs time (t) curves for atomically
flat Pt(111) and poly-Pt electrodes. Figure 8a shows the wave
shape of oscillation E at 0.13 V vs NHE. Parts b and c of Figure
8 are the wave shapes of oscillation A for atomically flat Pt-
(111) and poly-Pt, respectively. It is clear that the wave shapes
for oscillations E and A are quite different from each other. In
addition, the oscillation period for oscillation E (2-10 ms) is
much shorter than that for oscillation A (0.2-1 s).
2
) than those for atomically flat Pt(111) (Figure 3). The negative
resistance in Figure 5b appears only in a narrow potential region
0.10-0.00 V), and no oscillation E appears as seen in Figure
c, similar to the cases of parts b and c of Figure 2. The j-U
curves for the atomically flat Pt(111) such as shown in Figure
were restored when the atomically nonflat Pt(111) electrodes
giving the curves of Figure 5 were annealed in a hydrogen flame,
(
5
3

Electrochemical H2O2 Reduction at Pt Electrodes
J. Phys. Chem. B, Vol. 104, No. 17, 2000 4185
Figure 10. (a) j-U curves for atomically flat and nonflat Pt(111)
electrodes, (b) those for atomically flat Pt(111), and Pt(100), Pt(110),
and nonflat poly-Pt electrodes. Electrolyte: 0.2 M H
SO . Scan rate: 10 mV/s.
O
2 2
2
+ 0.3 M H -
4
in 0.3 M HClO4 (Figure 9a) are located in nearly the same
potentials as in 0.3 M H2SO4 (Figure 3a), in agreement with
2
0,21
Figure 8. Current density (j) vs time (t) curves, (a) and (b) for
atomically flat Pt(111) and (c) for poly-Pt. Electrolyte: (a, b) 1.0 M
the literature.
A current peak attributed to the adsorption of
ClO4 is not observed,20,21 contrary to the case of HSO4-.
Cathodic and anodic current peaks in 0.60-0.85 V (Figure 9a)
are attributed to the adsorption and desorption of OH . Though
-
H
2
O
2
+ 0.3 M H
potential: (a) 130 mV (oscillation E), (b) -30 mV (oscillation A), (c)
20 mV (oscillation A).
2
SO
4
, (c) 0.7 M H
2
O
2
+ 0.3 M H
2 4
SO . Electrode
-
21
-
-
-
the adsorption and desorption behaviors of HSO4 and ClO4
are quite different from each other, as mentioned above, the
observed negative slope and oscillation E are nearly the same
-
-
between the HSO4 - and ClO4 -based solutions (Figures 3
and 9). This indicates that the appearance of oscillation E is
not due to the effect of anion adsorption.
Figure 10 shows, comparatively, the H2O2-reduction currents
and the negative resistances for various Pt electrodes in 0.3 M
H2SO4 containing low-concentration (0.2 M) H2O2. The H2O2-
reduction current in 0.65-0.35 V for atomically flat Pt(111) is
higher in absolute value than that for atomically nonflat Pt-
(111) even in the same H2O2 concentration (Figure 10a). Also,
the H2O2-reduction current in this potential region decreases in
the order of atomically flat Pt(111), Pt(110), and Pt(100) and
nonflat poly-Pt (Figure 10b). The differences in the H2O2-
reduction current become more prominent with the increasing
scanning rate. It is also to be noted that the aforementioned
“
new” negative resistance in 0.35-0.10 V is observed promi-
nently only for atomically flat Pt(111), which shows the highest
H2O2-reduction current in 0.65-0.35 V. This result indicates
that the increase in the H2O2-reduction current in 0.65-0.35 V
for atomically flat Pt(111) has a close relation with the
appearances of the new negative resistance and oscillation E.
Figure 11 shows a phase diagram for the appearance of
oscillations A and E in the case of atomically flat Pt(111) with
respect to the H2O2 and H2SO4 concentrations. It seems that
not only the H2O2 and H2SO4 concentrations but also their ratio
are important for the appearance of these oscillations.
Figure 9. j-U curves for an atomically flat Pt(111) electrode, drawn
in the same way as Figure 2. Electrolyte: (a) 0.3 M HClO , (b) 0.2 M
+ 0.3 M HClO , (c) 0.7 M H + 0.3 M HClO . Scan rate: (a)
00 mV/s, (b, c) 10 mV/s.
4
H
1
O
2 2
4
2
O
2
4
Mathematical Simulation
We reported in previous papers12,14,15 that oscillations C and
Figure 9 shows j-U curves in 0.3 M HClO4-based solutions
instead of the 0.3 M H2SO4-based ones thus far described) for
atomically flat Pt(111) to see an effect of the kind of anions in
the electrolyte. The cathodic and anodic current peaks for upd-H
D appearing in 0.40-0.00 V in the presence of a small amount
-
-
-
(
of halide ions, such as Cl , Br , and I , could be explained by
taking into account a catalytic effect of adsorbed halogen (Pt-
X) and an autocatalytic effect of adsorbed OH (Pt-OH) for

4186 J. Phys. Chem. B, Vol. 104, No. 17, 2000
Nakanishi et al.
(RE), RΩ the solution resistance in the same place as above, IC
)
ACDL(dE/dt) the charging current, CDL the double-layer
capacitance, and IF the faradic current. Equation 9 can be
rewritten as
dE/dt ) (U - E)/AC R - I /AC
DL
(10)
DL
Ω
F
IF is expressed, from reactions 2 and 5-7, as follows:
s
s
I ) AF{-k C
+
(1 - θ - θ ) + k θ - k C θOH +
F
5
H H OH -5 H 2 H+
2
s
k θ
- k CH (1 - Θ ) + k Θ } (11)
H -7 H
+
6
OH
7
where CH s is the concentration of H+ ions at the electrode
surface and θOH, θH, and ΘH are the surface coverages of Pt-
OH, upd-H, and on-top-H, respectively. The quantities, k2, k5,
k-5, k6, k7, and k-7 are the rate constants for the ith reactions (i
+
)
2, 5, -5, 6, 7, and -7), which can be expressed by the
2 2 2 4
Figure 11. A phase diagram with respect to the H O and H SO
Butler-Volmer equations
concentrations: 0, only oscillation E appears; O, both oscillations A
and E appear; ×, no oscillation appears. Electrode: atomically flat Pt-
(111).
k ) k exp[-R n F(E - E )/RT]
i ) 2, 5, and 7 (12)
i
i0
i
i
i0
k ) k exp[(1 - R )n F(E - E )/RT]
the dissociative adsorption of H2O2. Namely, the appearance
of oscillations C and D could be explained by expressing the
rate constant (k1) for reaction 1 in the preceding section as
follows:
i
i0
i
i
i0
i ) -5, 6, and -7 (13)
where ki0 is the rate constant at E ) Ei0, Ei0 the equilibrium
redox potential for the ith reaction, Ri the transfer coefficient,
ni the number of transferred electrons, F the Faraday constant,
R the gas constant, and T the temperature. Because reactions 5
and -5 (and 7 and -7) are reverse reactions to each other, we
obtain E50 ) E-50 and E70 ) E-70.
k ) k + γ′θ θ
X OH
(3)
1
10
where k10 is a normal rate constant for reaction 1, γ′ is a
proportional constant, and θX and θOH are the surface coverages
of Pt-X and Pt-OH, respectively. The appearance of oscillation
E in 0.2-0.1 V for atomically flat Pt(111) in the present work
may be explained by taking into account a similar autocatalytic
effect of Pt-OH on the Pt(111) surface, i.e., by expressing k1
as follows:
+s
s
The time dependences of CH , CHO (the surface concentra-
tion of H2O2), θOH, θH, and ΘH are expressed as follows:
s
b
s
HO
(
δHO/2) dCHO /dt ) (DHO/δHO)(CHO - C ) -
s
2
k1CHO (1 - θ - θ ) (14)
H
OH
k ) k + γθ
OH
(4)
1
10
s
b
s
(δH
+
/2) dCH
+
/dt ) (DH
+
/δH
+
)(CH
+
- CH
+
) +
Thus, we have made mathematical simulation on the basis of
this idea.
s
s
{
-k C
+
θOH - k CH (1 - θ - θ ) + k θ +
+
2
H
5
H OH -5 H
2
s
The following reactions were considered in addition to
k θ
- k C (1- Θ ) + k Θ } (15)
6
OH
7
H+
H
-7
H
-
OH
1
5
reactions 1 and 2, in the same way as in our previous paper:
s
HO
2
s
0
OH
1
H
OH
5 H
+
N dθ /dt ) k C (1 - θ - θ ) - k C θOH
k5
+
-
2
Pt + H + e { }Pt-H (upd-H)
(5)
k θ
(16)
(17)
(18)
k-5
6
s
k6
+
-
N dθ /dt ) k C (1 - θ - θOH) - k-5θH
0
H
5
H+
H
2
Pt-OH
982Pt + O + 2H + 2e
(6)
(7)
2
s
2
k7
N
0
dΘ
H
/dt ) k
7
C
H+ (1 - Θ
H
) - k-7Θ - 2k Θ
H 8 H
+
-
Pt + H + e { }Pt-H (on-top-H)
k-7
+
where DHO and DH are the diffusion coefficients for H2O2 and
k8
+
+
H , respectively, δHO and δH the thicknesses of the diffusion
Pt-H (on-top-H) + Pt-H (on-top-H)
98H
(8)
2
+
b
+b
layers for H2O2 and H , respectively, and CHO and CH the
concentrations of H2O2 and H in the solution bulk, respectively.
+
where Pt represents schematically surface site(s) at the Pt
electrode.
+
A small contribution of the drift motion of H was neglected.
N0 schematically represents the total amount of surface Pt sites
per unit area.
The differential equations were also expressed in the same
13
way as in our previous work. From the conservation of the
current, we obtain2
4,25
Figure 12 shows j-U curves obtained by mathematical
calculations. With no or a small autocatalytic effect (γ e 7),
only a one-step negative resistance was reproduced for low-
concentration (0.2 M) H2O2 (Figure 12a) and only oscillation
A was reproduced for high-concentration (0.7 M) H2O2 (Figure
12b). On the other hand, with a large autocatalytic effect (γ g
8), the appearance of oscillation E was reproduced together with
the increase in the H2O2-reduction current in 0.60-0.40 V and
I ) jA ) (U - E)/R ) I + I ) AC (dE/dt) + I (9)
Ω
C
F
DL
F
where I is the total current, A the electrode area, j the current
density, U the (external or applied) electrode potential, E the
true electrode potential, U-E the ohmic drop between the
electrode surface and the position of the reference electrode

Electrochemical H2O2 Reduction at Pt Electrodes
J. Phys. Chem. B, Vol. 104, No. 17, 2000 4187
Figure 14. Schematic illustration of an autocatalytic mechanism in
which the dissociative adsorption of H O is accelerated on surface Pt
2 2
atoms in the neighborhood of adsorbed OH.
Discussion
The experimental results have shown that the atomically flat
Pt(111) electrodes show the electrochemical behavior for the
H2O2 reduction, quite different from that of other Pt electrodes,
such as atomically flat Pt(100) and Pt(110), atomically nonflat
Pt(111), and poly-Pt. The behavior of the latter electrodes, i.e.,
the appearance of the one-step negative resistance in a narrow
potential region from +0.10 to -0.02 V and the appearance of
only oscillation A in this potential region, can be explained by
taking into account the suppression of the H2O2 reduction by
1
1-13
the formation of upd-H as reported previously.
On the other
Figure 12. Calculated j-U curves with U scanned at a rate of 0.01
b
-3
hand, the behavior of the atomically flat Pt(111) electrode cannot
be explained by this model, and a new model has to be taken
into account.
As already mentioned in the preceding section, the appearance
of oscillation E, together with the appearance of the new
negative resistance in a potential region from 0.35 to 0.10 V
and the increase in the H2O2-reduction current in 0.65-0.35
V, can be explained by taking into account the autocatalytic
effect of adsorbed OH (Pt-OH) for the dissociative adsorption
of H2O2. This effect is not considered for our previous model
V/s. Parameter values used are as follows: CHO ) 0.2 × 10 mol
-
3
-3
-3
cm for (a) and (c) and 0.7 × 10 mol cm for (b) and (d), δHO
)
-
5
2
-1
+b
-3
-3
0
.01 cm, DHO ) 1.7 × 10 cm s , C
H
) 0.3 × 10 mol cm
,
δ
H+ ) 0.004 cm, DH+ ) 9.3 × 10 cm s , A ) 0.01 cm , CDL ) 2.0
-
5
2
-1
2
-
5
-2
-9
-2
×
10 F cm , N
0
) 2.2 × 10 mol cm , R
Ω
) 60 Ω, T ) 300 K,
2
-
-1
R
i
) 0.25 (i ) 2) and 0.50 (i * 2), n ) 1, k10 ) 4 × 10 cm s , k20
-5 -1 -2 -1 -5 -2
)
1 × 10 cm s , k50 ) 1 × 10 cm s , k-50 ) 1 × 10 mol cm
-
1
-8 -2 -1 -3 -1
s
, k60 ) 1 × 10 mol cm
s
, k70 ) 5 × 10 cm s , k-70 ) 5 ×
-6 -2
1
0- mol cm
6
-2
s
-1, k
8
) 5 × 10 mol cm s-1, E20 ) 1.37 V vs
NHE, E50 ) E-50 ) 0.08 V, E60 ) 0.67 V, E70 ) E-70 ) -0.03 V, γ
)
4 for (a) and (b) and 8 for (c) and (d).
1
3
for polycrystalline Pt. Mathematical simulation based on this
model has reproduced well the essential behavior of the
atomically flat Pt(111). According to this model, the increase
in the H2O2-reduction current in 0.65-0.35 V for Pt(111) is
attributed to the autocatalytic effect of surface Pt-OH for the
dissociative adsorption of H2O2 and the appearance of the new
negative resistance in 0.35-0.20 V is attributed to a decrease
of the Pt-OH concentration with a negative shift of the electrode
potential. Oscillation E appears on the basis of this new negative
resistance.
Figure 13. Time course of oscillation E, calculated with γ ) 8 and U
0.24 V. The other parameters are the same as those for Figure 12.
The difference in the electrochemical behavior between the
atomically flat Pt(111) and other Pt electrodes can thus be
attributed to be a difference in the efficiency of the autocatalytic
mechanism (i.e., the γ value in eq 4) between these electrodes.
In other words, the γ value is expected to decrease in the order
Pt(111) > (100) ≈ (110) g poly-Pt, as estimated from the
magnitude of the H O -reduction current in 0.65-0.35 V (Figure
10). How can this order be explained? We tentatively assume
that, as shown in Figure 14A, adsorbed OH (Pt-OH) is present
at hollow sites of surface Pt atoms and that H2O2 is adsorbed
)
the appearance of the new negative resistance in 0.35-0.10 V
Figure 12c,d). Figure 13 shows a time course of oscillation E,
calculated with γ ) 8. Compared with the observed oscillation
Figure 8a), the oscillation period is well reproduced, but there
(
(
2
2
is a difference in wave shape, which may be due to the adoption
of Nernst’s diffusion-layer model with a constant thickness in
eqs 14 and 15.

4188 J. Phys. Chem. B, Vol. 104, No. 17, 2000
Nakanishi et al.
with negatively polarized oxygen atoms of H2O2 directed to
surface Pt atoms. Under this assumption, it is expected that the
adsorption of H2O2 is accelerated on surface Pt atoms in the
neighborhood of adsorbed OH (Pt-OH) because they are more
or less positively polarized by an electronegativity difference
between the Pt atom and OH group (Figure 14A). This can be
responsible for the autocatalytic effect. The differences in the
γ value among various Pt electrodes can be attributed, in the
first approximation, to the differences in the number of “pairs
of Pt atoms” for the accelerated adsorption of H2O2. Figure 14B
illustrates possible examples of such pairs of Pt atoms by arrows
with two heads. The number of pairs of Pt atoms is 15, 12, and
This work was partly supported by a Grant-in-Aid for
Scientific Research on Priority Area “Electrochemistry of
Ordered Interfaces” (No. 09237105) from the Ministry of
Education, Science, Sports and Culture, Japan. We thank Prof.
K. Itaya and co-workers at Tohoku University for their kind
instruction of detailed techniques of STM measurements.
References and Notes
(
(
1) Hudson, J. L.; Tsotsis, T. T. Chem. Eng. Sci. 1994, 49, 1493.
2) Fahiday, T. Z.; Gu, Z. H. In Modern Aspects of Electrochemistry;
White, R. E., Bockris, J. O., Conway, R. E., Eds.; Plenum: New York,
1995; Vol. 27, p 383.
(
3) Koper, M. T. M. In AdVances in Chemical Physics; Prigogine, I.,
Rice, S. A., Eds.; John Wiley & Sons: New York, 1996; Vol. 92, p 161.
4) Krischer, K.; Luebke, M.; Wolf, W.; Eiswirth, M.; Ertl, G.
6
for atomically flat Pt(111), Pt(100), and Pt(110) surfaces,
(
respectively. Accordingly, we can expect that the autocatalytic
effect of Pt-OH is the most efficient for Pt(111) and decreases
in the order Pt(111), Pt(100), and Pt(110), in agreement with
the experiment. The corresponding number for atomically
nonflat Pt surfaces is estimated to be in the range of 6-12,
because the current peaks for upd-H for atomically nonflat Pt
are given by a simple sum of those for Pt(100) and Pt(110).
Finally it should be mentioned that Ross et al.26 reported that
the H2O2-reduction currents on low-index single-crystal Pt
electrodes showed the negative resistance in a potential region
from +0.1 to -0.3 V vs SCE, which was more prominent in
the order of Pt(110), Pt(100), and Pt(111), in agreement with
the new negative resistance in the present work. Ross et al.
explained their negative resistance to be due to the suppression
of the H2O2 reduction by the formation of upd-H, contrary to
an explanation in the present work. However, the atomically
flat Pt(111) electrodes show the negative resistance of a two-
step structure (Figure 3b), and in harmony with this result, the
Pt(111) (and sometimes Pt(100)) electrodes show two types of
oscillations (i.e., oscillations A and E). This strongly suggests
that the negative resistance for Pt(111) is composed of two
components.
In conclusion, the present work has revealed that the current
oscillations in the H2O2-acid-Pt electrode system strongly
depend on the atomic-level structure of the Pt electrode surface.
The detailed analyses of the behavior, combined with math-
ematical simulation, have suggested the presence of an auto-
catalytic mechanism of adsorbed OH (Pt-OH) for the disso-
ciative adsorption of H2O2, which depends in efficiency on the
atomic-level structure of the Pt surface. Further studies will serve
for deeper understanding of oscillation mechanisms and mech-
anisms of electrochemical reactions themselves.
Electrochim. Acta 1995, 40, 69.
(5) Raspel, F.; Eiswirth, M. J. Phys. Chem. 1994, 98, 7613.
(
6) Eiswirth, M.; Luebke, M.; Krischer, K.; Wolf, W.; Hudson, J. L.;
Ertl, G. Chem. Phys. Lett. 1992, 192, 254.
7) Raspel, F.; Nichols, R. J.; Kolb, D. M. J. Electroanal. Chem. 1990,
286, 279.
(
(8) Strasser, P.; Luebke, M.; Raspel, F.; Eiswirth, M.; Ertl, G. J. Chem.
Phys. 1997, 107, 979.
(9) Markovic, N.; Ross, P. N. J. Phys. Chem. 1993, 97, 9771.
(10) Tripkovic, A.; Popvic, K.; Adzic, R. R. J. Chim. Phys. Phys.-Chim.
Biol. 1991, 88, 1635.
(11) Hommura, H.; Mukouyama, Y.; Matsuda, T.; Yae, S.; Nakato, Y.
Chem. Lett. 1996, 391. Matsuda, T. Hommura, H.; Mukouyama, Y.; Yae,
S.; Nakato, Y. J. Electrochem. Soc. 1997, 144, 1988.
(12) Mukouyama, Y.; Nakanishi, S.; Nakato Y. Bull. Chem. Soc. Jpn.
999, 72, 2573.
1
(13) Mukouyama, Y.; Hommura, H.; Nakanishi, S.; Nishimura, T.;
Konishi, H.; Nakato, Y. Bull. Chem. Soc. Jpn. 1999, 72, 1247.
(14) Mukouyama, Y.; Konishi, H.; Nakanishi, S.; Nakato, Y. Chem. Lett.
998, 1009.
1
(15) Mukouyama, Y.; Nakanishi, S.; Konishi, H.; Nakato, Y. J.
Electroanal. Chem. 1999, 473, 156.
(16) Nakanishi, S.; Hommura, H.; Mukouyama, Y.; Matsuda, T.; Nakato,
Y. Chem. Lett. 1998, 977.
(
17) Clavilier, J.; Faure, R.; Guinet, G.; Durand, D. J. Electroanal. Chem.
980, 107, 205.
18) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Funda-
mentals and Applications, John-Wiley & Sons: New York, 1980; p 540.
1
(
(
(
19) Nichols, R. J.; Bewick, A. J. Electroanal. Chem. 1988, 243, 445.
20) Wieckowski, A.; Zelenav, P.; Varga, K. J. Chim. Phys. Phys.-Chim.
Biol. 1991, 88, 1247.
(21) Wagner, T.; Ross, P. N. J. Electroanal. Chem. 1988, 250, 301.
(
22) Itaya, K.; Sugawara, S.; Sashikata, K.; Furuya, N. J. Vac. Sci.
Technol. 1990, A8 (1), 515.
23) Furuya, N.; Koide, S. Surf. Sci. 1989, 220, 18.
(24) Koper, M. T. M.; Sluyters, J. H. J. Electroanal. Chem. 1991, 303,
73.
(
(
1.
25) Koper, M. T. M.; Sluyters, J. H. J. Electroanal. Chem. 1993, 347,
3
(26) Markovic, N. M.; Gasteiger, H. A.; Ross, Jr. P. N. J. Phys. Chem.
1995, 99, 3411.