Platinum-based electrocatalysts for generation of hydrogen peroxide in aqueous acidic electrolytes. Rotating ring-disk studies

Article abstract of DOI:10.1149/1.1525270

The electrocatalytic activity of selenium-, and sulfur-modified polycrystalline Pt electrodes for O₂ reduction in acid media has been examined using rotating ring-disk electrode techniques. The results obtained indicated that, within a rather narrow range of coverages, both Se-, and S-modified Pt surfaces promote O₂ reduction via a two-electron pathway to yield hydrogen peroxide at close to 100% faradaic efficiency over a wide potential region. Also presented in this work is an experimental procedure for Se-modification of high area, unsupported Pt particles based on fluidized packed bed reactor principles. Such a strategy could be readily scaled up opening new prospects for the development of large scale hydrogen peroxide generation in acid media, including electrochemically based, room temperature, Nafion-based oxygen concentrators.

Full text of DOI:10.1149/1.1525270

Journal of The Electrochemical Society, 150 ͑1͒ E39-E46 ͑2003͒
E39
0013-4651/2003/150͑1͒/E39/8/$7.00 © The Electrochemical Society, Inc.
Platinum-Based Electrocatalysts for Generation of Hydrogen
Peroxide in Aqueous Acidic Electrolytes
Rotating Ring-Disk Studies
*
**,z
Yibo Mo and Daniel A. Scherson
Department of Chemistry, Case Western Reserve University, Cleveland, Ohio 44106-7078, USA
The electrocatalytic activity of selenium-, and sulfur-modified polycrystalline Pt electrodes for O2 reduction in acid media has
been examined using rotating ring-disk electrode techniques. The results obtained indicated that, within a rather narrow range of
coverages, both Se-, and S-modified Pt surfaces promote O2 reduction via a two-electron pathway to yield hydrogen peroxide at
close to 100% faradaic efficiency over a wide potential region. Also presented in this work is an experimental procedure for
Se-modification of high area, unsupported Pt particles based on fluidized packed bed reactor principles. Such a strategy could be
readily scaled up opening new prospects for the development of large scale hydrogen peroxide generation in acid media, including
electrochemically based, room temperature, Nafion-based oxygen concentrators.
© 2002 The Electrochemical Society. ͓DOI: 10.1149/1.1525270͔ All rights reserved.
Manuscript submitted March 18, 2002; revised manuscript received July 10, 2002. Available electronically December 3, 2002.
The reduction of dioxygen in aqueous electrolytes, including
Nafion,1-5 has received extraordinary attention over the past few
decades.6,7 Much of the research impetus has been driven by the
need to improve the performance of gas permeable cathodes for fuel
cell applications, including stationary energy generation,8 vehicular
propulsion,9 and, more recently, portable electronics.10,11 Not sur-
prisingly, most of the emphasis, from a fundamental viewpoint, has
focused on the search for electrocatalysts displaying both high ac-
tivity and selectivity for the four-electron reduction of O2 to yield
water as the product. Yet another electrochemical process that
involves O2 as a reactant is the electrosynthesis of hydrogen
peroxide,12-16 an environmentally benign chemical currently being
considered for use in diverse technological areas, ranging from the
industrial bleaching of paper17,18 and wastewater treatment,19 to oxi-
dation of hazardous chemicals and bioremediation.17 Although cer-
tain types of carbon are effective electrocatalysts for the 2eϪ reduc-
tion of O2 in alkaline media, in terms of potency and high
selectivity,6 materials displaying equally desirable properties in acid
electrolytes are quite rare. Notable exceptions are provided by cer-
tain transition metal macrocycles of the phthalocyanine20 and por-
phyrin types21-24 and by a few simpler inorganic and organic spe-
cies, either in solution phase or immobilized on the surface of
otherwise inert electrodes by spontaneous adsorption, chemical de-
rivatization, and other means. Unfortunately, many of these molecu-
lar electrocatalysts also promote decomposition of hydrogen
peroxide25 generating radicals that attack their often delicate struc-
tures rendering products devoid of significant activity. Direct gen-
eration of H2O2 in acid media is particularly advantageous, as it
would meet pH conditions required for Fenton’s reagent
applications26,27 both in chemical synthesis and chemical degrada-
tion without further processing. In addition, it could impact mark-
edly the development of lightweight, Nafion-based, room-
temperature electrochemical oxygen concentrators for medical and
other applications.28
The approach currently under investigation in our laboratories
relies on the use of elemental surface modifiers capable of blocking
active sites on Pt responsible for the four-electron reduction of O2 ,
while opening pathways for H2O2 generation. Based on an initial
screening of a large number of atomic species, particularly those
capable of undergoing underpotential deposition, sulfur and sele-
nium were selected as the most promising candidates.
disk electrode ͑RRDE͒ technique29 to assess the activity ͑A͒ and
selectivity ͑S͒ of Se- and S-modified polycrystalline Pt surfaces for
O2 reduction to H2O2 as a function of coverage and applied poten-
tial. The methodologies involved in the preparation and electro-
chemical characterization of modified Pt electrodes were largely
based on the reports of Wieckowski et al.30,31 for S and Feliu et al.32
for Se, and those of Stickney et al.33-35 for S- and Se-modified
single-crystal Au surfaces. As is shown, faradaic efficiencies for
H2O2 generation ͑ᒃ͒ approaching 100% can be achieved within a
rather narrow range of Se and S coverages.
Experimental
All measurements were performed with a Pt ͑disk͒/Pt ͑ring͒ ro-
tating ring-disk electrode ͑RRDE, Pine Instruments, disk area
ϭ 0.164 cm2, inner and outer ring diameters, 4.93 and 5.38 mm,
respectively, i.e., theoretical collection efficiency N ϭ 0.22), in
0.50 M H2SO4 for Se-, or 0.10 M H3PO4 ͑J.T. Baker Ultrex, diluted
with ultrapure water, Barnstead͒ for S-modified Pt surfaces, respec-
tively, at room temperature, using a conventional three-compartment
electrochemical cell. Solutions were purged with O2 for at least 1 h
before RRDE data acquisition, and continuously thereafter. A revers-
ible hydrogen electrode ͑RHE͒ in the same solution was used as
reference, and a platinized Pt foil ͑see below͒ as counter electrodes.
Prior to dynamic polarization RRDE measurements in O2-saturated
solutions, the Pt ring was platinized to prevent rapid contamination
with adventitious impurities in the solution. This procedure was per-
formed by cycling the Pt ring in a 2-3% H2PtCl6 aqueous solution at
50 mV/s between Ϫ0.1 and ϩ1.1 V s. SCE and then holding at
v
Ϫ0.10 V for 3 min. The electrode was then removed from this
solution and rinsed thoroughly with ultrapure water before further
use. Increases in the actual area of the Pt ring of up to an order of
magnitude could be achieved by this methodology.
Selenium-modified Pt surfaces were prepared by electrochemical
deposition from 1.0 mM SeO2 in 20 mM H2SO4 solutions ͑J.T.
Baker, Ultrex͒ in a separate three-compartment cell to avoid cross
contamination.34 Sulfur-modified Pt surfaces were obtained by plac-
ing a small volume of an aqueous Na2S solution in the concentration
range of 0.27-1.0 mM on the Pt disk for a specified period of time
followed by thorough rinsing with pure water.31 Such electrodes
were then immersed in O2-free 0.10 M H3PO4 or 0.50 M H2SO4 and
the platinized-Pt ring was subsequently cycled at 50 mV/s in the
range 0.03-1.49 V s. RHE to oxidize any remaining sulfur or se-
v
This work exploits the well-known virtues of the rotating ring-
lenium on the surface. Once the ring voltammogram showed fea-
tures characteristic of clean Pt, the solution was saturated with O2 ,
and dynamic polarization measurements were then initiated.
The degree of selectivity of Se-, and S-modified Pt surfaces for
the 2eϪ reduction of O2 to H2O2 as a function of coverage and
* Electrochemical Society Student Member.
** Electrochemical Society Active Member.
z E-mail: dxs16@po.cwru.edu
Downloaded on 2015-06-29 to IP 140.112.4.67 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E40
Journal of The Electrochemical Society, 150 ͑1͒ E39-E46 ͑2003͒
Figure 1. Cyclic voltammogram for a Pt disk of a Pt/Pt RRDE in a 1.0 mM
SeO in 20 mM H SO quiescent solution in the range 0.34 -1.4 V s. SHE
͑thin line͒ at a scan rate ␷ ϭ 5 mV/s. The voltammogram in thick lines was
v
Figure 2. A series of linear potential scans for deposition and stripping of Se
on the Pt disk of a Pt/Pt RRDE electrode in a 1.0 mM SeO2 in 20 mM
H2SO4 quiescent solution (␷ ϭ 20 mV/s). The scans in the negative direc-
2
2
4
obtained by first scanning in the negative direction from 1.4 down to 0.54 V
and then up to 0.74 V s. SHE under otherwise identical conditions. At this
v
tion were interrupted at various potentials Edep s. SHE in the range 0.34-
v
point the scan was interrupted for ca. 3 min and then resumed up to ϩ1.4 V
0.54 V s. SHE and, after a holding period ␶ of 2.5 or 3.5 min, resumed in
v
s. SHE.
v
the positive direction. Thin solid line: Edep ϭ 0.54 V, ␶ ϭ 3 min; thick solid
line: Edep ϭ 0.49 V, ␶ ϭ 3.5 min; dotted line: Edep ϭ 0.44 V, ␶ ϭ 3 min;
dashed line: Edep ϭ 0.39 V, ␶ ϭ 3 min; dot-dash line: Edep ϭ 0.34 V,
␶ ϭ 2.5 min.
potential, was determined from the analysis of disk (idisk) and ring
(iring) currents measured with the Pt/Pt RRDE assembly. For all
these experiments, the platinized Pt ring was polarized at 1.4 V s.
v
RHE, i.e., positive enough for the oxidation of H2O2 generated at
the disk to proceed under diffusion-limited conditions.36
was first scanned in 1.0 mM SeO2 in 20 mM H2SO4 solutions in the
negative direction at ␷ ϭ 20 mV/s, subsequently reversed at 0.54 V,
and finally stopped and held at Edep ϭ 0.73 V s. SHE for ca. 3
v
Results and Discussion
min. The electrode was removed from the solution under potential
control, washed thoroughly with ultrapure water, and then reim-
mersed in a deaerated, Se-free 0.50 M H2SO4 solution in a different
cell without potential control. No evidence for hydrogen adsorption
was found for such Se-modified electrodes cycled between 0.08 and
0.73 V ͑see curve a, in panel A, Fig. 3͒ consistent with the blocking
of active sites on the Pt surface. Moreover, a subsequent scan in the
positive direction up to 1.43 V ͑see dotted line in panel A, Fig. 3͒
Se-modified Pt electrodes.—Cyclic voltammetry curves recorded
at a scan rate ␷ ϭ 5 mV/s in 1.0 mM SeO2 in 20 mM H2SO4
solutions over the range of 0.34-1.4 V s. SHE ͑see thin line, Fig. 1͒
v
displayed two sets of peaks, labeled as A, A ͑centered at 0.64 and
Ј
1.11 V, respectively͒ and C, C ͑centered at 0.44 and 0.83 V, respec-
Ј
tively͒ similar in shape to those reported by Stickney and co-
workers for Se deposition on the three low index faces of single-
crystal Au.33,34 A stripping peak at potentials slightly more negative
yielded a well-defined peak A over virtually the same region where
Ј
than C ͑denoted as B by Lister et al.͒ has also been observed for
Pt oxidation occurs on the bare metal in Se-free 0.50 M H2SO4 , i.e.,
0.9-1.4 V ͑see thin solid line in Fig. 3͒, in agreement with the results
shown in Fig. 1.
Ј
Ј
Se on Au, although in this case ͑see feature centered at 1.07 V͒, it
may have contributions due to Pt oxide formation. Similar experi-
ments in which the scan in the negative direction was reversed at
The coverage of Se, ␪Se , was determined by subtracting the
charge obtained by integrating the stripping curve up to 1.45 V ͑see
dotted lines, Fig. 3͒ from ϩ0.73 to 1.45 V and 1.45 to 1.35 V during
the reverse scan, from the voltammetric curve for the same Pt elec-
trode over the same potential range in a solution devoid of Se, i.e.,
ca. 0.144 mC for the Pt electrode used for these experiments ͑see
thin solid line, Fig. 3͒. The oxide corrected Se-stripping charge in
this case yielded a value of 0.161 mC. If it is assumed that UPD Se
0.54 V s. SHE, stopped at 0.74 V s. SHE ͑see arrow͒, and then
v
v
resumed after ca. 3 min up to 1.4 V, yielded only peak A without any
contributions due to B ͑see thick solid lines͒. In an analogy with the
Ј
assignments of Lister et al.,34 peaks A and A are ascribed to the
Ј
surface-limited underpotential deposition ͑UPD͒ and stripping of Se,
which in this case has also contributions due to Pt oxidation and
reduction ͑see below͒, whereas C and C are attributed to the for-
Ј
mation and stripping of a second, distinct adsorbed Se phase. Cor-
is present as a zero valent species and its oxidation up to 1.45 V s.
v
relations between the shapes and magnitudes of A and C were
Ј
Ј
SHE is not kinetically hindered, standard redox potentials would
predict for the oxidation of the UPD layer ͑in the potential range in
question͒ to proceed via a 6eϪ process. On this basis, the extra
charge under the linear scan ͑dotted lines͒ in Fig. 3 would be con-
sistent with the stripping of ca. 1.7 ϫ 1014 atoms. The actual area of
Pt estimated from the charge associated with hydrogen adsorption/
desorption in Se-free solutions ͑see thin solid line, panel A, Fig. 3͒,
i.e., 220 ␮C/cm2, is about 0.230 cm2 ͑roughness factor ca. 1.4͒;
hence, the number of Se atoms per Pt atom on the surface would
correspond to a Se coverage, ␪Se , of about one-half a monolayer.
This value is in excellent agreement with that reported by Huang
et al.33 for Se on Au͑100͒ for experiments in which the scan was
obtained from a series of experiments in which the Pt electrode was
scanned at ␷ ϭ 20 mV/s from 1.45 V down to the desired potential,
E
dep , in the region 0.34-0.54 V, held at that value for a certain
period of time, ␶, in the range 2.5-3.5 min, and then scanned at the
same rate up to 1.45 V to oxidize adsorbed Se ͑see Fig. 2͒. Cursory
Ͻ
inspection of these curves reveals that in the region Edep
0.54 V,
and for ␶ in the range specified, the area under C increases as the
Ј
value of Edep becomes more negative, and its peak potential shifts
toward more positive values.
Further insight into some of the properties of Se deposits was
obtained from measurements in which the potential of the Pt disk
Downloaded on 2015-06-29 to IP 140.112.4.67 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑1͒ E39-E46 ͑2003͒
E41
Figure 4. Plots of idisk ͑upper panel͒, iring ͑middle panel͒, and
͉iring /idisk͉
͑lower panel͒ s. E for a bare ͑curves a͒ and Se-modified Pt disk of a
v
disk
Pt/Pt RRDE electrode ͑curves b-f͒ in O2-saturated 0.50 M H2SO4 at ␻
ϭ 400 rpm for ␪Se ϭ 0.15 ͑curves b͒, 0.25 ͑curves c͒, 0.34 ͑curves d͒, 0.51
͑curves e͒, 1.24 ͑curves f͒. ͑See inset Fig. 3 and also text for details͒.
Figure 3. Panel A, cyclic voltammograms of a bare Pt disk of a Pt/Pt RRDE
in the range 0.03-1.43 V s. SHE ͑thin solid line͒ and of the same electrode
v
modified by preadsorbed Se ͑see text for details͒ in the range 0.08-0.73 V
s. SHE ͑curve a, thick solid line͒ in 0.50 M H SO , ␷ ϭ 20 mV/s. Also
v
2
4
selectivity of the Se-modified surfaces were found to be optimum
͑ca. 1/2 ϭ ␪Se , see below͒. As evidenced by the data collected ͑not
shown in this panel is a linear scan in the range 0.73 to 1.43 V for a Se-
modified Pt disk surface recorded immediately after holding at 0.73 V under
otherwise identical conditions ͑see dotted line͒. Panel B, stripping curves for
the Se-modified Pt disk in panel A as a function of Edep (␶ ϭ 3 min, see
inset in this panel͒ under otherwise identical conditions as those specified in
shown͒, the differences in ␪ observed in independent experiments
Se
were about 3% and thus regarded as negligible. This is an important
observation as it indicates that the emersion/washing/immersion
procedure, at least for the most active Se-modified surfaces, does not
lead to changes in ␪Se , providing a useful strategy for preparing
Se-modified Pt in high area form, as would be required for technical
applications. It is also essential to note that the onset of O2 reduction
occurs at potentials negative to the onset of Se-stripping ͑see be-
low͒, allowing the electrocatalytic properties of such surfaces for
this reaction to be examined as a function of ␪Se in the absence of Se
in solution.
panel A. Panel C, plot of selenium coverage, ␪
s. E
.
dep
v
Se
reversed between peaks A and C, for which the onset of Au oxide
formation occurs at more positive potentials compared to Pt.
A series of experiments involving this same basic protocol were
then performed to determine ␪ as a function of Edep in the poten-
Se
tial region about peak A. Once immersed into the Se-free solution,
the Se-modified electrode was scanned in the negative direction to
assess blockage of hydrogen adsorption, and then scanned in the
positive direction up to 1.45 V. The results obtained for various Edep
Rotating disk-ring electrode dynamic polarization curves for O2
reduction.—Correlations between ␪ and electrocatalytic activity
Se
for O2 reduction were obtained using five Edep values ͑see inset, Fig.
3͒ at a single rotation rate, ␻ ϭ 400 rpm. In each case, Edep was set
by scanning the electrode in a 1.0 mM SeO2 in 20 mM H2SO4 in a
separate cell to potentials in the negative direction at 20 mV/s to the
desired value and holding for 3 min. Immediately thereafter, the
Se-modified Pt electrode was removed from the solution under po-
tential control, rinsed thoroughly with pure water, and immersed in
the O2-saturated ͑more than 30 min purging͒ Se-free 0.50 M H2SO4
without potential control.
values ͑see insert, panel B, Fig. 3͒, including a plot of ␪
s. E
,
Se v
dep
based on coulometric analyses of these data are shown in panels B
and C, in Fig. 3, respectively. In earlier studies,32 problems were
encountered for determining ␪Se on Pt͑110͒ and Pt͑100͒, the two low
index surfaces which would more closely resemble Pt͑poly͒, be-
cause the potentials required for full electro-oxidative desorption of
Se are too high to avoid compromising the single-crystalline char-
acter of their surfaces. This behavior is unlike that observed by the
same authors for Se on Pt͑111͒, which displays a sharply defined
surface bound, 4eϪ redox process, thereby allowing reliable coulo-
Plots of idisk , iring , and ͉iring /idisk͉
vs. Edisk for five different
values of Edep ͑or equivalently ␪Se), and ␶ ϭ 3 min ͑see curves b-f,
Fig. 4 and inset in Fig. 3͒, including those obtained with bare Pt
metric determination of ␪
.
Se
Efforts were also made in our studies to establish whether ␪
Se
measured in Se-containing solutions, under otherwise identical con-
ditions, differed from the values determined by the method de-
scribed above for Edep values in the region in which the activity and
͑curve a͒, are shown in Fig. 4. For ␪ ϭ 1.25, for which the strip-
Se
ping curve shows a quite prominent peak B , the electrode displayed
Ј
virtually no activity, i.e., idisk ϭ iring ϭ 0. This behavior was unlike
Downloaded on 2015-06-29 to IP 140.112.4.67 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E42
Journal of The Electrochemical Society, 150 ͑1͒ E39-E46 ͑2003͒
Figure 5. Plots of ilim ͑solid circles, left ordinate͒ and 2iring /N ͑open circles,
right ordinate͒ s. based on the data shown in Fig. at
␪
4
␻
v
Se
ϭ 400 rpm ͑see caption for details͒. The collection efficiency based on the
geometry of the RRDE assembly is 0.22.
Ͻ
Ͻ
0.50,
that
found
in
the
range
0.15
␪
i.e.,
Se
0.54 р Edep р 0.69, for which the limiting currents for
O2-reduction, ilim , measured at 0.18 V s. SHE ͑see left ordinate,
v
Fig. 5͒, as well as the amount of H2O2 generated at the disk, as
judged by the magnitude of iring ͑see right ordinate, Fig. 5͒, were
Figure 6. Plots of idisk ͑upper panel͒, iring ͑middle panel͒, and
͑lower panel͒ s. Edisk for Pt (␪Se ϭ 0.5)/Pt RRDE electrode in
O2-saturated 0.50 M H2SO4 at ␻ ϭ 400 ͑curve a͒, 900 ͑curve b͒, 1600
͉iring /idisk͉
quite significant. In fact, for ␪Se ϭ 1/2 (Edep ϭ 0.54 V͒,
͉iring /idisk͉
a
v
ca. 0.21 and ilim(␪ ϭ 0)ϭ 2ilim(␪ ϭ 0.5), which are consistent
Se
Se
with the quantitative reduction of O2 to yield H2O2 as the only
product ͑note that for bare Pt, iring is negligible͒. Prompted by its
unique performance, the overall electrocatalytic characteristics of
(␪Se ϭ 0.5)Pt surfaces were examined in more detail.
͑curve c͒, and 2500 ͑curve d͒ rpm. Ering ϭ 1.38 V s. SHE.
v
This phenomenon was explored in more detail by performing
measurements in which sulfur was gradually removed from a
S-modified Pt disk ͑19 ␮L of 1.0 mM Na2S solution for 5.5 min͒ by
scanning the potential in stages to potentials sufficiently positive for
sulfur to undergo partial oxidation to solution phase sulfate at ␷
ϭ 10 mV/s. Prior to sulfide exposure, the Pt disk was cleaned by
electrochemical cycling in O2-free 0.10 M H3PO4 . Following S
adsorption, the platinized Pt ring was cycled in O2-free 0.10 M
Plots of idisk ͑upper panel͒, iring ͑middle panel͒, and
͉iring /idisk͉
͑lower panel͒ s. E for Pt/(␪Se ϭ 0.5)Pt RRDE in O2-saturated
v
disk
0.50 M H2SO4 at ␻ ϭ 400 ͑curve a͒, 900 ͑b͒, 1600 ͑c͒, and 2500
rpm ͑d͒ are shown in Fig. 6. As indicated, iring /idisk for 0.07
͉
͉
Ͻ
Edisk
Ͻ 0.6 V, is ca. 0.21 for all ␻, affording strong evidence that O2
reduction over this voltage range proceeds predominantly via a 2eϪ
pathway to yield H2O2 . Additional support for this view was ob-
tained from the linear character of the Levich plot, ilim s. ␻1/2, for
v
which the slope was Ϫ0.167 mA cmϪ2 s1/2, i.e., about half that
obtained for O2 reduction on bare ͑Se-free͒ Pt, ͑a 4eϪ-process͒ for
which the experimental slope was Ϫ0.356 mA cmϪ2 s1/2 ͑see
Fig. 7͒.
S-modified Pt electrodes.—Shown in Fig. 8 are dynamic polar-
ization curves obtained in O2-saturated 0.10 M purified H3PO4 so-
lutions at a scan rate of 10 mV/s and ␻ ϭ 900 rpm for a Pt disk ͑of
a Pt-Pt RRDE͒, which had been previously exposed to a solution
0.27 mM Na2S for 3 min ͑curve a͒ and for 20 s ͑curve c͒. Curve b
in that figure was recorded under the same conditions for an elec-
trode prepared by the same procedure using, instead, a 1.0 mM Na2S
solution for 20 s. Corresponding idisk and iring data recorded with a
S-free Pt-Pt RRDE at the same rotation rate ͑not shown in this
figure͒ yielded, as before, large idisk with iring about two orders of
magnitude lower. These data indicate that the O2-reduction activity
is rather low for large sulfide exposure ͑curve a͒; as the exposure is
reduced however, both idisk and iring increase ͑curve c͒, signaling
significant H2O2 generation at the disk. In fact, increasing the con-
centration of sulfide, while keeping the time of exposure constant
͑20 s͒, decreased idisk , but increased iring ͑curve b͒, i.e., the selec-
tivity ͑S͒ of the S-modified Pt electrode was greatly enhanced.
1/2
Figure 7. Plots of ilim s.
␻
for a bare Pt disk ͑curve a͒ and a
v
Pt (␪ ϭ 0.5) disk ͑curve b͒ of a Pt/Pt RRDE in 0.50 M H2SO4 solutions
Se
͑see text for other conditions͒.
Downloaded on 2015-06-29 to IP 140.112.4.67 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑1͒ E39-E46 ͑2003͒
E43
Figure 9. Plots of idisk , iring , and
͉
iring /idisk
͉
s. E
for the first, 7th, and
v
disk
18th oxidative cycles for a S-modified Pt disk surface of a Pt/Pt RRDE
prepared by placing 19 ␮L of a 1.0 mM Na2S solution for 51/2 min followed
by rinsing with ultrapure water (␻ ϭ 900 rpm). ␷ ϭ 10 mV/s. The dotted
line in the right panel was collected during the scan in the negative direction.
that despite the changes in the activity ͑A͒ of S-modified Pt due to
changes in ␪S , the surface, within a certain range of ␪S , displays
very high selectivity ͑S͒ toward H2O2 generation. Also noteworthy
is the fact that cycle ten coincides with the onsets for the increase in
both idisk and iring and an overall decrease in
these results, four electrocatalytic regions ͑I-IV in Fig. 10͒, can be
͉iring /idisk͉. Based on
Figure 8. Dynamic polarization curves for O2 reduction in O2-saturated
purified 0.10 M H3PO4 solutions for a S-modified Pt disk electrode of a
Pt/Pt RRDE prepared by exposure to an aqueous solution 0.27 mM Na2S for
3 min ͑curve a͒ for 20 s ͑curve c͒ followed by rinsing with ultrapure water.
Curve b was obtained via the same procedure, except that the Pt disk elec-
trode was placed in contact with a 1.0 mM Na2S solution for 20 s (␻
ϭ 900 rpm, ␷ ϭ 10 mV/s, Ering ϭ 1.3 V s. SHE͒.
v
H PO at 50 mV/s in the range Ϫ0.07 to 1.4 V s. SHE to oxidize
v
3
4
any sulfur that may have adsorbed on the surface. Once the ring
voltammogram showed features characteristic of clean Pt, the solu-
tion was saturated with O2 . After collecting RRDE data for the
freshly prepared S-modified Pt disk surface at ␻ ϭ 900, the elec-
trode was cycled in the same solution once in the range R
ϭ 0.10-0.70 V at 10 mV/s, and a new set of polarization curves
recorded under otherwise identical conditions. This procedure was
followed for a total of 18 cycles with R ϭ 0.10-0.70 ͑cycles 2-4͒;
R ϭ 0.10-0.90 ͑cycles 5-8͒; R ϭ 0.10-1.00 ͑cycles 9 and 10͒, and
R ϭ 0.10-1.10 V ͑cycles 11-18͒. Shown in Fig. 9 are plots of idisk
,
i
ring, and iring /idisk s. E for the first, seventh, and 18th cycles.
͉
͉
v
disk
Data of this type collected for the entire series of cycles was then
used to create plots of ilim (Edisk ϭ ϩ0.10 V͒, iring (Ering ϭ 1.3 V͒,
and ͉iring /idisk͉ s. number of cycles, a parameter that correlates at
v
least qualitatively with a monotonic decrease in sulfur coverage, ␪S .
This does not represent a serious limitation, as the conclusions made
from the analysis of the data do not rely on a quantitative knowledge
of ␪S . As shown in Fig. 10, the overall activity of Pt is low for large
␪S , increases to reach a plateau up to about cycle 10 and finally
increases even further as the sulfur is oxidatively removed from the
surface. The amount of H2O2 produced, on the other hand, is small
for high ␪S , reaches two plateaus, and finally decreases after the
14th cycle. Particularly interesting, however, is the fact that, despite
variations in idisk and iring , ͉iring /idisk͉ ͑see, lower panel, Fig. 10͒ is
Figure 10. Plots of ilim (Edisk ϭ ϩ0.10 V s. SHE͒, iring, and ͉iring /idisk͉
v
very high, ca. 0.25, and independent of ␪ up to cycle ten, decreas-
data extracted from the complete series of 18 oxidation cycles for a
S
ing monotonically thereafter. It becomes evident from these data,
S-modified Pt disk electrode ͑see caption Fig. 9 for details͒.
Downloaded on 2015-06-29 to IP 140.112.4.67 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E44
Journal of The Electrochemical Society, 150 ͑1͒ E39-E46 ͑2003͒
Figure 12. Plots of ilim (Edisk ϭ ϩ0.13 V͒, iring (Ering ϭ 1.38 V͒, and
iring /idisk as a function of time for a Se-modified Pt (␪Se ϭ 0.5) disk of a
Pt/Pt RRDE in 0.50 M H2SO4 solutions at ␻ ϭ 400 rpm. Polarization of the
disk and ring was discontinued for ca. 48 min keeping the electrode rotating
and then resumed under the same initial conditions.
Figure 11. Plots of idisk , iring, and
Se-modified Pt disk electrode of a Pt ͑platinized͒/Pt RRDE polarized at
͉
iring /idisk
͉
s. time recorded for a
v
͉
͉
E
disk ϭ 0.13 V and Ering ϭ 1.38 V s. SHE in O -saturated 0.50 M H SO
v
2
2
4
at ␻ ϭ 200 ͑curve a͒, 400 ͑curve b͒, and 900 rpm ͑curve c͒.
defined as region I ͑cycles 1-3͒, where A for peroxide generation
increases steadily to reach in region II ͑cycles 4-8͒ a maximum both
on the results shown in Fig. 11 and 12, this deactivation does not
seem to be caused by Se desorption, but rather to adsorption of a
solution-phase impurity that desorbs under open-circuit conditions.
In fact, assuming a diffusion coefficient of 5 ϫ 10Ϫ6 cm2/s and also
that adsorption occurs under pure diffusion control, the onset to
deactivation ͑ca. one monolayer͒ would correspond to the contami-
nant being present at a concentration of ca. 10Ϫ8 M.
Much faster deactivation was found in the case of S-modified Pt
electrodes as illustrated by transient data of the same type as that
shown in Fig. 13 acquired between cycles 6 and 7 ͑see Fig. 10͒ at
Edisk ϭ ϩ0.10 V. Despite the gradual loss in activity, ͑see Fig. 13͒
in A and S. Based on the rather large and constant
͉iring /idisk͉, the
faradaic efficiency for this process in both these regions appears
close to 100%. As the cycling is continued, i.e., region III ͑cycles
11-14͒, the amount of H2O2 produced increases but at lower faradaic
efficiency ᒃ and finally region IV represents a gradual activation of
the Pt surface for 4eϪ reduction, i.e., lower S for H2O2 generation.
Transient experiments.—Insight into the stability of Se- and
S-modified Pt as electrocatalytic surfaces for the reduction of O2
was obtained from measurements in which idisk and iring for the Pt/Pt
RRDE were monitored as a function of time in O2-saturated acid
electrolytes at various rotation rates. Plots of ilim (Edisk ϭ 0.13 V͒,
͉
iring /idisk͉ remained virtually unchanged, i.e., electrode deactivation
does not lead to losses in specificity, as was found for Se-modified
Pt electrodes. Rather unambiguous evidence for an impurity, and not
sulfur desorption, as being responsible for deactivitation was found
upon cycling the deactivated electrode ͑cycle 8͒ into the oxide for-
mation region, a procedure that restored fully the activity and selec-
tive of the surface ͑see Fig. 10͒.
iring (Ering ϭ 1.38 V͒ and ͉iring /idisk͉ as a function of time for Se-
modified Pt (␪Se ϭ 0.5) in 0.50 M H2SO4 solutions recorded at
␻ ϭ 200 ͑curve a͒, 400 ͑curve b͒, and 900 rpm ͑curve c͒ are shown
in Fig. 11. These curves are characterized by an initial period in
which both A and S for the Se-modified Pt disk electrode are high,
followed by a region in which A decreases, but S remains virtually
unaffected. At longer times, both idisk and iring /idisk fell to very low
values signaling significant deactivation. In the case of ␻ ϭ 400
rpm ͑curve b͒ polarization of the disk and ring was discontinued at
about ca. 100 min for ca. 48 min keeping the electrode rotating, and
then resumed under the same initial conditions. As shown in Fig. 12,
the activity of the disk electrode was to a large extent restored,
although the time to deactivation was shortened considerably. Based
Conclusions and Future Prospects
The results obtained in this work have provided ample evidence
that the amount of adsorbed sulfur or selenium can have a profound
effect on the overall electrocatalytic activity and selectivity of poly-
crystalline Pt for the reduction of dioxygen in acid electrolytes. Al-
though actual coverages of the surface modifiers could only be reli-
ably determined for Se, analysis of results obtained with the RRDE
Downloaded on 2015-06-29 to IP 140.112.4.67 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑1͒ E39-E46 ͑2003͒
E45
Figure 15. Charge s. potential of a Pt foil ͑11 cm2 cross-sectional area͒ in
v
20 mM H2SO4 obtained with the minireactor in Fig. 14, while the main
compartment was purged with Ar ͑open circles͒. The data in solid circles was
obtained in an otherwise identical experiment, except that 0.4 mg Pt black
was added to the solution ͑filled circles͒.
washing and subsequent immersion in either deareated or
O2-saturated acidic solutions devoid of Se do not seem to affect the
actual surface coverage. On this basis, it would seem possible to
prepare Se-modified high area Pt surfaces of that specified ␪Se using
fluidized packed-bed-reactor-type techniques. Promising results
were obtained with a minireactor shown schematically in Fig. 14,
which consists of two chambers separated by a Nafion tube. The
outside or working electrode compartment houses a large Pt foil or
cathode current collector mounted along the inner walls of a Teflon
nipple. Prior to the actual experiments, the Nafion tube was filled
with high area Pt using two Pt wires inserted one from the bottom
and one from the top, which serve as current collectors for the
anode. The bottom of the cell is then closed, and the main chamber
filled with an acid solution through the top of the cell. A Teflon tube
pressed to fit into an orifice drilled into the wall of the nipple was
Figure 13. Plots of idisk , iring, and ͉iring /idisk͉ as a function of time following
polarization of the S-modified Pt disk electrode at ϩ0.10 V s. SHE be-
v
tween cycles 6 and 7 ͑see text for details͒. ␻: 900 rpm.
for both types of modified Pt surfaces indicated that dioxygen can be
reduced to hydrogen peroxide at nearly 100% faradaic efficiency
over a wide potential range provided the coverage of Se or S are set
within a rather narrow range. Unlike S, for which the coverage can
be changed by anodic stripping, although in an as yet noncontrol-
lable fashion, ␪ can be readily adjusted by polarizing the Pt elec-
Se
trode at the desired value within the UPD region to yield the highest
activity and selectivity. Furthermore, subsequent emersion and
Figure 16. Plot of charge s. time for a Pt foil recorded upon stepping the
v
potential from ca. 0.84 down to 0.54 V s. SHE in a 1 mM SeO ϩ 20 mM
v
2
H2SO4 , ͑dotted line͒ using the minireactor in Fig. 14. The curve ͑solid lines͒
was obtained in an otherwise identical experiment except that the 0.40 mg Pt
black was added to the same solution before application of the potential step
͑see text for details͒.
Figure 14. Schematic diagram of a fluidized-type minireactor for Se modi-
fication of high area Pt. W.E ͑working electrode͒, R.E. ͑reference electrode͒,
and C.E. ͑counter electrode͒.
Downloaded on 2015-06-29 to IP 140.112.4.67 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E46
Journal of The Electrochemical Society, 150 ͑1͒ E39-E46 ͑2003͒
2. C. R. Martin, L. D. Whiteley, D. R. Lawson, and M. N. Szentirmay, Polym. Prepr.
(Am. Chem. Soc. Div. Polym. Chem.), 29, 442 ͑1988͒.
3. A. K. Shukla, P. Stevens, A. Hamnett, and J. B. Goodenough, J. Appl. Electro-
chem., 19, 383 ͑1989͒.
4. F. A. Uribe, T. E. Springer, and S. Gottesfeld, J. Electrochem. Soc., 139, 765
͑1992͒.
5. S. Gottesfeld, I. D. Raistrick, and S. Srinivasan, J. Electrochem. Soc., 134, 1455
͑1987͒.
6. Kinetics and Mechanisms of Electrode Processes Vol. 7, M. R. Tarasevich, A.
Sadkowski, and E. B. Yeager, Editors Plenum Press, New York ͑1983͒.
7. R. Adzic, in Electrocatalysis, J. Lipkoswki and P. N. Ross, Editors, p. 197, Wiley-
VCH, Inc., New York ͑1998͒.
8. A. J. Dickman, AMD (Am. Soc. Mech. Eng.), 40, 337 ͑2000͒.
9. A. J. Appleby, Sci. Am., 281, 74 ͑1999͒.
used to connect the main reactor chamber to an external compart-
ment that houses the reference electrode. During operation, the so-
lution in the main reactor chamber which contains the high area Pt is
agitated continuously by bubbling Ar gas.
For these preliminary experiments, the reactor was filled first
with a 20 mM H2SO4 solution and a cyclic voltammogram recorded
after deaeration yielded features characteristic of clean Pt ͑not
shown͒. The potential of the Pt foil was then stepped from ϩ0.74 V
s. SHE ͑platinum-oxide-free͒ to the desired potential Edep and held
v
at that value while recording the current for ca. 80 s, a time suffi-
cient for the current to reach steady state. Immediately thereafter, the
electrode was stepped back to ϩ0.74 V for a few minutes and then
stepped to a new Edep value. Shown in Fig. 15 ͑open circles͒ are
plots of the charge obtained from integration of the steady-state
background current as a function of potential. Once the entire series
was completed, 0.40 mg Pt black was added into the cell and the
same type of experiments repeated yielding the results shown in
solid circles in the same figure.
10. C. K. Dyer, Sci. Am., 281, 88 ͑1999͒.
11. S. R. Narayanan, T. I. Valdez, and F. Clara, in Direct Methanol Fuel Cells, S. R.
Narayanan, S. Gottesfeld, and T. Zawodzinski, Editors, PV 2001-4, p. 254, The
Electrochemical Society Proceedings Series, Pennington, NJ ͑2001͒.
12. P. C. Foller and R. T. Bombard, J. Appl. Electrochem., 25, 613 ͑1995͒.
13. P. Piela and P. K. Wrona, J. Phys. Chem. B, 105, 1494 ͑2001͒.
14. A. Alvarez-Gallegos and D. Pletcher, Electrochim. Acta, 44, 853 ͑1998͒.
15. K. B. Lee, M. B. Gu, and S.-H. Moon, Eng. Life Sci., 1, 237 ͑2001͒.
16. P. Tatapudi and J. M. Fenton, J. Electrochem. Soc., 140, L55 ͑1993͒.
17. R. P. Sethi, B. K. Rao, and D. Srinivas, Chem. Ind. Dig., 14, 95 ͑2001͒.
18. F. P. Komarov, Bum. Prom., , 11 ͑1969͒.
The solution was then replaced by deaerated 1 mM SeO2 ϩ 20
mM H2SO4 solution and the potential stepped from ϩ 0.84 to
ϩ 0.54 V s. SHE while monitoring the current ͑see dotted line in
v
Fig. 16͒. Immediately thereafter, 0.40 mg Pt black ͑Alfa͒ was added
19. M. Ficek and M. Ficek, Chemik, 47, 9 ͑1994͒.
and the potential stepped from ϩ0.84 to ϩ0.54 V s. SHE yielding
v
20. J. Zagal, P. Bindra, and E. Yeager, J. Electrochem. Soc., 127, 1506 ͑1980͒.
21. N. Kobayashi, T. Matsue, M. Fujihira, and T. Osa, J. Electroanal. Chem. Interfacial
Electrochem., 103, 427 ͑1979͒.
22. A. Bettelheim, R. J. H. Chan, and T. Kuwana, J. Electroanal. Chem. Interfacial
Electrochem., 99, 391 ͑1979͒.
23. A. Bettelheim, R. J. H. Chan, and T. Kuwana, J. Electroanal. Chem. Interfacial
Electrochem., 110, 93 ͑1980͒.
24. P. A. Forshey, T. Kuwana, N. Kobayashi, and T. Osa, in Advances in Chemistry
Series, K. M. Kadish, Editor, Vol. 201, ACS, Washington, DC, ͑1982͒.
25. F. Beck, J. Appl. Electrochem., 7, 239 ͑1977͒.
26. T. Harrington and D. Pletcher, J. Electrochem. Soc., 146, 2983 ͑1999͒.
27. K. Fukatsu, S. Kokot, and D. P. Schweinsberg, Text. Res. J., 69, 769 ͑1999͒.
28. D. A. Scherson, Patent ͑2002͒, In press.
29. A. J. Bard and L. R. Faulkner, Electrochemical Methods: Fundamentals and Ap-
plications, 2nd ed., Wiley, New York ͑2001͒.
30. Y. E. Sung, W. Chrzanowski, A. Zolfaghari, G. Jerkiewicz, and A. Wieckowski, J.
Am. Chem. Soc., 119, 194 ͑1997͒.
31. Y.-E. Sung, W. Chrzanowski, A. Wieckowski, A. Zolfaghari, S. Blais, and G. Jerk-
iewicz, Electrochim. Acta, 44, 1019 ͑1998͒.
the charge s. time curve shown in Fig. 16 ͑solid line͒. As indicated,
v
the charge difference between the Pt black and the foil is larger in
the presence compared to the absence of Se in solution, and there-
fore consistent with the adsorption of Se on the Pt black surface.
Unfortunately, the amounts of Se-modified Pt that could be prepared
with this minireactor were found to be too small to carry out mean-
ingful electrocatalytic experiments. However, high activities and se-
lectivities for the reduction of dioxygen to hydrogen peroxide in
acid electrolytes were observed recently in our laboratory for high
area Pt particles embedded on the surface of a glassy carbon rotating
disk electrode modified by Se in situ,37 providing clear evidence that
the effects observed for massive Pt can indeed be reproduced for the
technically more relevant Pt blacks.
Acknowledgments
This work was supported by a contract from CAMP.
32. J. M. Feliu, R. Gomez, M. J. Llorca, and A. Aldaz, Surf. Sci., 289, 152 ͑1993͒.
33. B. M. Huang, T. E. Lister, and J. L. Stickney, Surf. Sci., 392, 27 ͑1997͒.
34. T. E. Lister and J. L. Stickney, J. Phys. Chem., 100, 19568 ͑1996͒.
35. B. W. Gregory, M. L. Norton, and J. L. Stickney, J. Electroanal. Chem. Interfacial
Electrochem., 293, 85 ͑1990͒.
Case Western Reserve University assisted in meeting the publication
costs of this article.
References
36. S. L. Gojkovic, S. Gupta, and R. F. Savinell, Electrochim. Acta, 45, 889 ͑1999͒.
37. Y. Mo, S. Sarangapani, A. Le, and D. A. Scherson, In press.
1. H. Y. Liu and F. C. Anson, J. Electroanal. Chem. Interfacial Electrochem., 158, 181
͑1983͒.
Downloaded on 2015-06-29 to IP 140.112.4.67 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).