A novel dual thin-layer flow cell double-disk electrode design for kinetic studies on supported catalysts under controlled mass-transport conditions

Article abstract of DOI:10.1016/j.electacta.2003.09.050

We present a novel dual thin-layer flow cell double-disk electrode design, which due to its high collection efficiency in combination with a reasonable time resolution allows highly sensitive measurements of electro-active species generated during electrochemical reactions on the working electrode. Most important, it allows for rapid changes of the electrolyte during the measurements, which in combination with a thin-film electrode, where an inert glassy carbon substrate is covered by a thin film of carbon-supported noble metal catalyst, makes it particularly useful for kinetic studies on electrocatalytic fuel cell reactions under realistic conditions (continuous reaction, continuous electrolyte flow) and on realistic materials (supported catalysts). The performance of the set-up and its suitability for kinetic studies, in particular for transient experiments involving rapid electrolyte exchange, are demonstrated in potentiodynamic and transient potentiostatic measurements for the oxygen reduction reaction on carbon-supported Pt catalysts and, for comparison, on Pt-free Vulcan carbon supports.

Full text of DOI:10.1016/j.electacta.2003.09.050

Electrochimica Acta 49 (2004) 1297–1305
A novel dual thin-layer flow cell double-disk electrode
design for kinetic studies on supported catalysts
under controlled mass-transport conditions
Z. Jusys, J. Kaiser, R.J. Behm∗
Department of Surface Chemistry and Catalysis, University of Ulm, D-89069 Ulm, Germany
Received 26 May 2003; received in revised form 6 August 2003; accepted 22 September 2003
Abstract
We present a novel dual thin-layer flow cell double-disk electrode design, which due to its high collection efficiency in combination
with a reasonable time resolution allows highly sensitive measurements of electro-active species generated during electrochemical reactions
on the working electrode. Most important, it allows for rapid changes of the electrolyte during the measurements, which in combination
with a thin-film electrode, where an inert glassy carbon substrate is covered by a thin film of carbon-supported noble metal catalyst, makes
it particularly useful for kinetic studies on electrocatalytic fuel cell reactions under realistic conditions (continuous reaction, continuous
electrolyte flow) and on realistic materials (supported catalysts).
The performance of the set-up and its suitability for kinetic studies, in particular for transient experiments involving rapid electrolyte ex-
change, are demonstrated in potentiodynamic and transient potentiostatic measurements for the oxygen reduction reaction on carbon-supported
Pt catalysts and, for comparison, on Pt-free Vulcan carbon supports.
© 2003 Elsevier Ltd. All rights reserved.
Keywords: Electrocatalysis; Flow cell; Thin layer cell; Supported Pt catalyst; Oxygen reduction; H2O2 formation
1. Introduction
the species formed on the rotating disk electrode at the sur-
rounding ring electrode, which is set at such a potential
that the reaction intermediates can be reduced (or oxidized)
without interference with other electrochemical reactions. A
typical example is the detection of H2O2 formed during O2
reduction on the disk electrode by subsequent oxidation of
the H2O2 to O2 on the ring at potentials where other reac-
tions such as anodic O2 evolution can be ruled out [6].
Other hydrodynamic techniques available in electrochem-
istry, which have been extensively applied and described
theoretically, include the Wall-Jet (or Impinging-Jet) disk
electrode (WJDE) technique [7–11] and the (channel) flow
cell technique [10–12]. In the WJDE design a continuous
flow of electrolyte impinging at the static electrode surface
results in a laminar flow of electrolyte at the electrode sur-
face under certain conditions (flow rate, diameter of the
inlet capillary and the electrode, nozzle-to-electrode dis-
tance), in contrast to the dynamic (rotating) electrode in
the RDE configuration. In contrast to the R(R)DE tech-
nique, however, both the WJDE and the (channel) flow cell
method were rarely used for studying electrocatalytic pro-
cesses. The few examples for electrocatalytic WJDE studies
Kinetic and mechanistic model studies of many electrocat-
alytic reactions depend crucially on a controlled mass trans-
port to the electrode. This is particularly true for reactions
involving dissolved gaseous species such as H2 oxidation or
O2 reduction etc., because of the low concentrations of the
reactants in the millimolar range. One important example is
model studies on fuel cell (FC) reactions, which should be
performed under conditions typical for fuel cell operation
(continuous reaction, continuous mass transport, elevated
temperatures, realistic catalysts) in order to yield results rel-
evant for FC applications. This is most commonly achieved
using the rotating disk electrode (RDE) technique [1–4]. Re-
active reaction intermediates and reaction products gener-
ated during an electrocatalytic reaction can be determined in
these experiments in a rotating ring-disk electrode (RRDE)
set-up [5]. Due to a well-defined laminar flux of electrolyte
over the ring-disk electrode assembly this allows to detect
∗
Corresponding author.
E-mail address: juergen.behm@chemie.uni-ulm.de (R.J. Behm).
0013-4686/$ – see front matter © 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2003.09.050

1298
Z. Jusys et al. / Electrochimica Acta 49 (2004) 1297–1305
include CO oxidation on Pt electrodes [13–15] and studies
on the effect of CO on H2 oxidation [16]. The thin-layer
flow cell technique was applied recently for methanol ox-
idation studies at elevated temperatures on Pt/Vulcan [17]
and Pt-Ru electrodes [18].
In both techniques reactive products/intermediates can
be detected via a second (detector) ring-shaped electrode
or via a second stationary electrode (detector) placed in
the flow channel downstream [19], which was used for
various electro-analytical applications (see [12,20,21] and
references therein). Being an open (flow-through) system
and due to a rapid exchange of electrolyte at the electrodes
combined with high collection efficiency, dual electrode
channel flow cells are advantageous for transient kinetic
measurements, which was equally demonstrated for numer-
ous electro-analytical applications [12]. However, to the
best of our knowledge, this hydrodynamic technique has not
been applied so far for studying electrocatalytic reactions
related to low temperature fuel cells, specifically, oxygen
reduction reaction pathways.
The RRDE and WJDRE set-ups have the advantage of
a rather short time required for the reactive species to be
transported from the disk electrode (generator) to the ring
electrode (‘response time’), but may suffer from moderate
collection efficiencies. Furthermore, RRDE does not allow
fast electrolyte exchange under well-controlled conditions,
which can become important, e.g., for studies of reactions
on sensitive electrode surfaces. This can be accomplished in
other channel flow designs such as the dual-band electrode
[12,22] or the radial thin-layer ring-disk configuration [23],
which can reach rather high (close to 100%) collection ef-
ficiencies at flow rates of submicroliter per second. At high
collection efficiencies these designs suffer, however, from
the generally very long times (minutes range) required for
electrolyte transport from the first ‘generator’ electrode to
the second one (‘collector’). The resulting low time resolu-
tion makes them less suitable for kinetic studies.
Fig. 1. Schematic drawing of the dual thin-layer double-disk electrode
flow cell design, showing the glassy carbon substrate with the supported
catalyst thin film at its inner side (‘generator’) (1), the Kalrez gaskets
(2), the Kel-F body of the cell with the flow channels (3), and the second
polycrystalline Pt electrode (‘collector’) (4).
in our configuration compared to typically about 20% in
the RRDE configuration. The sensitivity for detection of
electro-active products is increased accordingly.
In the following we will first, after a brief description
of the DTLFC-DDE set-up and experimental details, char-
acterize this set-up by cyclic voltammetry and calibrate
it versus a Fe(III)/Fe(II) couple. In another part we will
demonstrate the advantages and limitations of the approach
compared to a standard RRDE configuration for the quan-
titative evaluation of the ORR pathways on Pt/Vulcan fuel
cell catalyst. We will show that because of the advantages
described above—small electrolyte volume, fast electrolyte
exchange under controlled conditions, and high collection
efficiency—the DTLFC-DDE approach introduced here is
a highly advantageous electro-analytical tool for mechanis-
tic and kinetic studies of electrocatalytic reactions under
well-controlled conditions, both on massive model elec-
trodes and state-of-the-art supported catalysts. This makes it
particularly well suited for studies of electrocatalytic reac-
tions under polymer electrolyte fuel cell relevant conditions
and on fuel cell relevant materials.
2. Experimental
In the present communication we report on a novel de-
sign of a dual thin-layer flow cell with a double-disk elec-
trode (DTLFC-DDE), which is based on the dual thin-layer
flow cell developed originally by Jusys et al. [24] for si-
multaneous differential electrochemical mass spectrometry
(DEMS) and electrochemical quartz microbalance (ECQM)
measurements and applied later by Jusys et al. for DEMS
measurements on supported catalysts ([25] and references
therein), which combines a high collection efficiency and
reasonable time resolution of the order of a few seconds.
This set-up allows rapid changes from one electrolyte to
another one under well-controlled conditions, which is typ-
ical for thin-layer cells due to their small volume in the
microliter range [26], and therefore is perfectly suited for
kinetic studies involving electrolyte exchange, as required,
e.g. for transient studies of the removal of a pre-adsorbed
adlayer. Furthermore, this approach compares favorably to
the RRDE because of the high product collection efficiency
at the collector [22,23], which will turn out to be ca. 80%
2.1. Dual thin-layer double-disk electrode flow cell
The experiments were performed in the dual thin-layer
flow cell used for DEMS studies in previous publications
(see [25] and references therein), which is based on the
dual thin-layer flow cell design described in [24], modi-
fied correspondingly for the present electrochemical mea-
surements (see Fig. 1). The cell itself consists of a circular
Kel-F plate (5-mm thick) with capillaries for the electrolyte
inlet and outlet, whose openings are centered on the oppo-
site sides of the Kel-F plate. The connection between both
compartments is made by four connecting capillaries posi-
tioned at the inner edge of two Kalrez gaskets (ca. 50-␮m
thick), against which the working electrode or ‘generator’,
a mirror-polished glassy carbon disk 9 mm in diameter with
a 6-mm diameter thin Pt/Vulcan film, and the ‘collector’, a
mirror-polished polycrystalline Pt disk 15 mm in diameter,
exposed area 1.33 cm2, were pressed to form two intercon-
nected thin-layer compartments.

Z. Jusys et al. / Electrochimica Acta 49 (2004) 1297–1305
1299
The incoming electrolyte first hits the center of the work-
ing electrode and spreads over its surface in the first thin-
layer compartment, until it reaches the four outlet capillaries.
During the residence time at the electrode the electro-active
species can react and the reaction products are accumulated
within the small electrolyte volume in the first compartment
(ca. 5 ␮l), resulting in significant reaction-induced changes
in the concentration of both educt and product. In the sec-
ond thin-layer compartment the electro-active species can
react at the second electrode, which is biased to a potential
suitable for oxidation/reduction of these species. The elec-
trolyte leaves the second thin-layer compartment through
the outlet capillary positioned in the center of the oppo-
site Kel-F wall. This way the dual thin-layer cell operates
as a continuous flow-through reactor, with the supporting
electrolyte serving as a carrier for both reacting species
and products. A quantitative characterization of the elec-
trochemical properties of the set-up will be presented in
Section 3.1.
ing electrolyte containing 1 mM K3[Fe(CN)6] (Merck, p.a.).
All measurements were conducted at room temperature.
3. Results and discussion
3.1. Electrode characterization and calibration of the
set-up versus a Fe(III)/Fe(II) redox couple
First both electrodes were characterized by cyclic voltam-
metry (CV) in 0.5 M H2SO4 solution. A representative CV
of the thin-film Pt/Vulcan electrode, which is mounted in the
first compartment of the dual thin-layer flow cell, is shown in
Fig. 2a (solid line). It agrees well with those reported in the
literature (see, e.g. [29,30]). Assuming a density of Pt surface
atoms of 1.3 × 1015 per cm2, a hydrogen adsorption charge
for polycrystalline Pt of 0.21 mC/cm2 [31], and a coverage
of Had at the onset of bulk hydrogen evolution at 0.06 V
(RHE) of 0.77 ML [32], we calculate a real (active) Pt sur-
face area of 7 cm2 from the hydrogen adsorption charge on
the Pt nanoparticle surface (1.14 mC), after subtracting the
double-layer charging current. (Note that the double layer
2.2. Further experimental details
The thin-film electrode was prepared as described previ-
ously [27]. In short, an aliquot of an aqueous suspension of
the conditioned Pt/Vulcan catalyst (E-TEK, 20 wt.% Pt, con-
ditioning 30 min in 10%O2 in N2 at 300 ◦C, 30 min in H2 at
300 ◦C, at a gas flow rate 60 Nml/min and 30 Nml/min, re-
spectively, atmospheric pressure) was pipetted onto a glassy
carbon disk (Sigradur G, Hochtemperatur Werkstoffe), dried
and fixed by a thin (ca. 0.1 ␮m) Nafion film, which was pre-
pared by pipetting Nafion solution on top of the dried cata-
lyst layer. The geometric area of the electrode is 0.28 cm2,
the Pt loading 28 ␮g cm−2. The average particle size and
the particle size distribution of the Pt/Vulcan catalyst had
been characterized recently by X-ray diffraction (XRD) and
transmission electron microscopy (TEM), yielding an aver-
age particle size of 3.7 nm for the conditioned catalyst and
a dispersion of 26% [28].
Two Pt wires connected through separate ports at the in-
let and outlet of the cell served as counter electrodes. The
reference electrode (saturated calomel electrode, SCE) was
connected through a Teflon capillary at the outlet of the cell
(all potentials are quoted versus that of the reversible hydro-
gen electrode (RHE)). The potentials of both generator and
collector were controlled using a standard bi-potentiostat
from Pine Instruments. The potential scan rate applied at
the Pt/Vulcan generator in potentiodynamic experiments was
10 mV/s.
The electrolyte flow (flow rate ca. 4 ␮l/s) was driven by
the hydrostatic pressure of the electrolyte in the supply bot-
tle. The supporting electrolyte (0.5 M sulfuric acid) was pre-
pared using Millipore Q water and ultrapure sulfuric acid
(Merck, suprapur), and purged by high purity Ar (MTI Gase,
N 6.0) or saturated with O2 (MTI Gase, N 5.7) in the elec-
trolyte supply bottle. The collection efficiency of the set-up
was calibrated using a Fe3+/Fe2+ redox couple in support-
Fig. 2. Cyclic voltammograms recorded separately on a thin-film Pt/Vulcan
electrode (20 wt.% Pt, Pt loading 28 ␮g/cm2) (Fig. 2a, solid line), on
a disperse Vulcan XC72 support electrode with similar carbon loading
(loading 112 ␮g/cm2) (Fig. 2a, dash-dotted line) and on a bare glassy car-
bon stub (Fig. 2a, dotted line, 10-fold magnified) in the first compartment
of the dual thin-layer flow cell (v = 10 mV/s), and on a polycrystalline Pt
electrode (1.33 cm2 accessible surface area) in the second compartment
of the flow cell (Fig. 2b) (scan rate v = 50 mV/s, Ar-saturated 0.5 M
H2SO4 solution, electrolyte flow rate ca. 4 ␮l/s, room temperature).

1300
Z. Jusys et al. / Electrochimica Acta 49 (2004) 1297–1305
charging current results mainly from the pseudocapacitive
current of the carbon Vulcan support in the negative-going
scan (Fig. 2, dash-dotted line)). This value agrees well with
the results of our previous studies on the same catalyst [28],
and is within the range of the Pt surface area of 5 cm2 cal-
culated from the total Pt loading in the electrocatalytic area,
using the dispersion of ca. 26% (the ratio of surface atoms
to the total number of the atoms) of the catalyst nanopar-
ticles calculated from the TEM data [27,28], which sug-
gests essentially full utilization of the Pt surface atoms. (It
should be noted that the dispersion is calculated by assum-
ing a spherical shape of the nanoparticles, which is known
to be not precise [33], and a density of Pt surface atoms of
1.22 × 1015 atoms/cm2 [27]. The resulting deviations, how-
ever, would not significantly affect the above statement.)
The CVs of the second electrode (e.g. see Fig. 2b), a
mirror-polished polycrystalline Pt disk with an exposed ge-
ometric surface area of 1.33 cm2, which is mounted in the
second compartment of the dual thin-layer flow cell, also
show good agreement with the well-known current–voltage
response of clean polycrystalline Pt electrodes [34]. Apply-
ing the same procedure as described above the real surface
area of this electrode was calculated to be 1.86 cm2, which
corresponds to a roughness factor of ca. 1.4.
Fig. 3. Cyclic voltammogram of a thin-film Pt/Vulcan electrode (Pt load-
ing 28 ␮g/cm2) in the first compartment ((a), solid line) and the related
current signal recorded simultaneously on a polycrystalline Pt electrode
at E = 1.16 V in the second compartment (b) of a dual thin-layer flow
cell in Ar-saturated 0.5 M H2SO4 solution containing 1 mM K3[Fe(CN)6]
(electrolyte flow rate ca. 4 ␮l/s, potential scan rate 10 mV/s, room temper-
ature). The dotted line in (a) indicates the double-layer charging corrected
signal, obtained by averaging the positive-going and negative-going scan
of the CV. The right-hand axis in (b) indicates the collection efficiency.
Finally, based on the absence of any significant dis-
tortions, the CVs presented in Fig. 2 demonstrate that
the electrochemical measurements can be carried out in
the DTLFC-DDE configuration without significant Ohmic
drops.
For the quantitative determination of the electro-active
species produced during a continuous (bulk) electrochemi-
cal reaction on the first electrode (generator), the collection
efficiency of the second electrode (collector) has to be cali-
brated, similar to RRDE measurements. This was done us-
ing the Fe(III)/Fe(II) redox process, comparing the amount
of Fe(II) formation in the first electrode with that of Fe(II)
oxidation on the second electrode. Fig. 3a shows the CV
of the Pt/Vulcan electrode in 1 mM K3[Fe(CN)6] contain-
ing 0.5 M H2SO4 solution. The broad feature in the cur-
rent response to the potential of the Pt/Vulcan electrode
(Fig. 3a, solid line) for ferrocyanide reduction is due to large
pseudo-capacitive double-layer charging currents character-
istic for Pt/Vulcan electrodes (see Fig. 2a). Assuming a sym-
metric double-layer contribution (see Fig. 2a, dash-dotted
line), the latter can be subtracted by simply averaging the
currents in the negative-going and positive-going scans. The
resulting curve (Fig. 3a, dotted line) shows a mass trans-
port limited current of ca. −0.05 mA for Fe(III) reduction to
Fe(II) on the generator (Fig. 3a, dotted line). The hysteresis
in the negative- and positive-going scans reflects the time
required for transferring the species generated at the first
electrode to the second electrode (about 3 s). A constant,
mass-transport-limited current of ca. 0.038 mA is obtained
on the collector at generator potentials negative of 0.5 VRHE
(Fig. 3b).
To avoid the significant pseudo-capacitive contributions
encountered on the Pt/Vulcan electrode in potentiodynamic
experiments (Fig. 3a), we also performed measurements at a
constant potential for ferrocyanide reduction on the genera-
tor. In these experiments the generator potential was stepped
from 0.96 to 0.46 VRHE, while the collector potential was
kept constant at 1.16 VRHE (Fig. 4). After the initial over-
shoot in the Faradaic current, which is at least largely due to
double-layer charging upon stepping the electrode potential,
the current drops to a nearly constant value of −0.05 mA
(Fig. 4a), which is purely mass-transport limited. This value
for ferrocyanide reduction is equal to that estimated from the
CV data in Fig. 3a. The current on the collector increases
accordingly after the potential step, reaching a mass trans-
port limited value of 0.038 mA after about 5 s (ca. 90% after
3 s) (Fig. 4b). Also this value agrees well with the potentio-
dynamic data in Fig. 3b.
Fe(II), which is reached at potentials negative of 0.6 VRHE
The current measured simultaneously at the polycrystalline
Pt collector at a constant electrode potential of 1.16 VRHE
.
,
which is sufficient for electro-oxidation of electrogenerated
Fe(II) species, but not for oxygen evolution, is shown as a
function of the potential of the first electrode in Fig. 3b. The
Fe(II) electro-oxidation current to Fe(III) on the collector
(Fig. 3b) closely follows the reduction current of Fe(III) to

Z. Jusys et al. / Electrochimica Acta 49 (2004) 1297–1305
1301
a non-ideal flux of electrolyte over the electrode surface in
our thin-layer cell, as described before in Section 2. Though
a channel-flow from the inlet to each of the four connecting
capillaries is achieved, stagnation regions can occur between
the connecting capillaries, especially at the periphery of the
electrode, so that a uniform, radial flow cannot be attained
in our design. Because of the reduced flow in the stagnation
regions the experimental value of the mass-transport-limited
current is lower than that predicted theoretically according
to Eq. (1), which assumes a uniform flux of the reactant over
the whole electrode surface. This situation can be partly im-
proved increasing the number of the connecting capillaries
[37], which, however, causes an increase in the total volume
and hence in the time constant of the cell. For a more detailed
description of the electrolyte flow in thin-layer channel elec-
trochemical reactors see [22,38] and references cited therein.
Similar to the definition for the RRDE configuration [5]
the collection efficiency N for a one-electron electrochemical
reaction is given by the ratio of the collector current to the
(mass-transport-limited) current at the generator,
−Icoll
N =
.
(2)
Ilim,gen
Fig. 4. Simultaneously recorded current transients upon a potential step
(0.96 → 0.46 V) of the Pt/Vulcan (Pt loading 28 ␮g/cm2) electrode in
the first compartment. (a) current on the Pt/Vulcan working electrode
and (b) on the polycrystalline Pt electrode at E = 1.16 V in the second
compartment of a dual thin-layer flow cell in Ar-saturated 0.5 M H2SO4
solution containing 1 mM K3[Fe(CN)6] (electrolyte flow rate ca. 4 ␮l/s,
room temperature). The right-hand axis in (b) indicates the collection
efficiency.
From the data of Figs. 3 and 4 the collection efficiency for
the dual thin-layer cell double-disk electrode set-up is 0.76
(or 76%). This value is about four-fold higher than that of
a typical RRDE configuration (typically ca. 20%, depend-
ing on the design dimensions of the ring-disk electrode [5]),
and about similar to those reported for a radial thin-layer
ring-disk configuration [23,39] or a dual-band electrode con-
figuration [22]. Due to the large area of the collector our
double-disk electrode approach achieves a high collection
efficiency at a time constant of ca. 3 s, and, therefore, is ad-
vantageous compared to both an RRDE set-up or a thin-layer
ring-disk/dual band electrode configuration for sensing small
amounts of electro-active species generated at the working
electrode, such as peroxide in the oxygen reduction reac-
tion, at moderate potential scan rates (10 mV/s or below),
as will be demonstrated later in Section 3.2. The proposed
DTLFC-DDE configuration can hardly be used, however,
for the detection of short-lived intermediates. In such cases
an RRDE configuration, with the almost instantaneous de-
tection at the ring, would be the suitable instrument.
Following Weber and Long [26] the mass-transport-limited
current Ilim in a thin-layer channel flow cell is given by
ꢀ
ꢁ2/3
D · A
Ilim = k · n · F · c · u1/3
(1)
b
where k is the cell constant (k = 1.467 for a channel-flow
thin-layer cell), n is the number of electrons, F is the Fara-
day constant, c is the concentration of the species involved
in the electrochemical reaction, u is the electrolyte flow rate,
D is the diffusion coefficient of the reacting species, A is the
electrode surface area, and b is the thickness of the thin-layer
gap, respectively. A linear dependence of the rates for fer-
rocyanide reduction and ferricyanide oxidation on u1/3 was
indeed found in earlier experiments on a polycrystalline Pt
electrode in the first thin-layer compartment [35], which sug-
gests that a laminar flow regime is attained in the thin layer
of electrolyte at the electrode surface.
Inserting the corresponding values of k = 1.467, n = 1,
F = 96,480 A s/mol, c = 1 × 10−3 mol/cm3, u = 3 ×
10−3 cm3/s, D = 7.7 × 10−5 cm2/s [36], A = 0.28 cm2 and
b = 40 × 10−3 cm, the mass-transport-limited current for
ferrocyanide reduction can be calculated to be ca. −0.1 mA
according to Eq. (1). This value is about twice of that deter-
mined experimentally for ferrocyanide reduction (Figs. 3a
and 4a) The deviation can be rationalized taking into account
3.2. Oxygen reduction reaction studied in the dual
thin-layer double-disk electrode set-up
For demonstrating the performance of the DTLFC-DDE
set-up, we present measurements of the oxygen reduction
reaction (ORR) on a Pt/Vulcan electrode (Fig. 5a), where
we detect simultaneously the hydrogen peroxide oxidation
on the polycrystalline Pt collector at 1.16 V (Fig. 5b), which
allows us to evaluate the peroxide yields as a function of the
electrode potential. This reaction had been studied in our
group on the same catalyst in thin-film RRDE measurements
previously [40].

1302
Z. Jusys et al. / Electrochimica Acta 49 (2004) 1297–1305
for ferrocyanide reduction reaction can be used also for the
quantitative evaluation of the O2 reduction pathways.
The turnover frequency or TOF number, defined as the
number of O2 molecules reduced per Pt surface atom and
per second, can be calculated from
ꢀ
ꢁ
−i NA
TOF =
(3)
nF Ns
where i is the oxygen reduction current, n is the num-
ber of electrons (4 for O2 reduction to water), F is the
Faraday constant, NA the Avogadro number (6.02 × 1023
molecules/mol), and Ns is the number of Pt surface atoms
(1.3×1015 atoms/cm2, multiplied by the real Pt surface area
of the catalyst, as derived found from hydrogen UPD on the
Pt/Vulcan electrode, see Section 3.1.). The mass-transport-
limited current of ca. −0.5 mA corresponds to a turn over
frequency of about 0.07 molecules site−1 s−1. Note that at
the higher flux in a typical RRDE measurement consid-
erably higher values are obtained for the mass-transport-
limited current and the corresponding TOF values (The ca.
−1.5 mA/cm2 of geometric area obtained here correspond
roughly to that found for a rotation rate of 100 rpm in an
RRDE configuration [40]).
The hysteresis between the negative-going (Fig. 5a, solid
line) and positive-going (Fig. 5a, dashed line) scan in the
potential region 0.8–1.0 V is generally explained by the
quasi-reversible formation/reduction of PtO, (see, e.g. [42]).
At a positive potential limit of 0.96 V (Fig. 5a, gray line)
it follows the ORR in the positive-going scan (Fig. 5a,
dashed line), indicating that under these conditions PtO for-
mation is negligible. The inset in Fig. 5a shows a Tafel
Fig. 5. Simultaneously recorded cyclic voltammograms of a Pt/Vulcan
electrode (Pt loading 28 ␮g/cm2) and of a Vulcan carbon electrode (loading
112 ␮g/cm2) in the first compartment (a), and the related current signals
recorded simultaneously on a polycrystalline Pt electrode at E = 1.16 V in
the second compartment (b) of a dual thin-layer flow cell. The right-hand
axis in (b) indicates the peroxide yield on the Pt/Vulcan electrode. Dotted
line: base voltammogram of the Pt/Vulcan electrode in Ar-purged 0.5 M
H2SO4 solution; dashed line, full line: voltammogram of the Pt/Vulcan
electrode in O2-saturated 0.5 M H2SO4 solution (positive-going scan:
dashed line, negative-going scan: full line), gray line: similar CV with a
positive potential limit of 0.96 V; and dash-dotted line: voltammogram of
the Vulcan carbon electrode in O2-saturated 0.5 M H2SO4 solution: (elec-
trolyte flow rate ca. 4 ␮l/s, potential scan rate 10 mV/s, room temperature).
Inset Fig. 5a: Tafel plot for the ORR current during the positive-going
scan.
plot of the kinetic O2 reduction current (defined as Ikin
=
Ilim × I/(Ilim − I) per one milligram of the catalyst [43]
in the region 0.75–0.95 V during the positive-potential scan,
which is in good agreement with that reported previously
in RRDE measurements [40], despite the lower value of
the mass-transport-limited current obtained under present
conditions in the DTLFC-DDE configuration. The slope
of approximately 60 mV per decade in the close-to-linear
part of the kinetic ORR current in the “low-current” re-
gion (E > 0.87 V), which is usually interpreted as evidence
of a first single-electron transfer to the adsorbed O2 as the
rate-determining step [44], and the transition to a ca. 120 mV
per decade slope in the “high current” region (E < 0.87)
agree well with the literature data for the ORR in an R(R)DE
configuration on polycrystalline [45], single-crystal [46] and
nanoparticle [40,47] Pt electrodes.
The CV recorded on the Pt/Vulcan catalyst in O2 satu-
rated solution (Fig. 5a, full line, dashed lines) largely re-
sembles the base CV in sulfuric acid (Fig. 5a, dotted line)
plus a superimposed O2 reduction current, which is trans-
port limited to ca. −0.5 mA in the potential region negative
of the formation/reduction of PtO. The theoretical value for
the mass-transport-limited current, calculated according to
Eq. (1) for channel-flow with an electron number of n =
4, a diffusion coefficient for O2 diffusion of D = 1.93 ×
10−5 cm2/s, and an O2 concentration in the O2 saturated so-
lution of c = 1.26 × 10−6 mol/cm3 [41], is ca. −1.2 mA.
Again the difference to the lower experimental value is ex-
plained by a non-ideal flux of the electrolyte at the electrode,
as discussed in Section 3.1. It is important that the differ-
ence between the theoretical and experimental value for the
transport-limited current for O2 reduction is the same as that
found for ferrocyanide reduction, suggesting the same trans-
fer characteristics. Therefore the collection efficiency found
Fig. 5b shows the current measured on the collector at
1.16 V in parallel to the cyclic voltammograms plotted in
Fig. 5a. The collector signal is essentially zero over the en-
tire potential range for the base CV in Ar-saturated 0.5 M
sulfuric acid solution (dotted lines in Fig. 5). In O2-saturated
0.5 M H2SO4 solution the collector current increases signifi-
cantly at potentials negative of 0.6 V for the positive potential
limit of 0.96 V (Fig. 5b, gray line). Qualitatively, the current
at the collector follows the well-known peroxide formation

Z. Jusys et al. / Electrochimica Acta 49 (2004) 1297–1305
1303
features found at the ring electrode during the ORR on poly-
crystalline Pt and Pt/Vulcan electrodes in RRDE measure-
ments [40,47,48]. The maximum currents reached on the col-
lector at 0.06 V are similar to those measured on the ring in
a RRDE configuration at rotation rates of 1600 rpm, though
the O2 reduction current on the generator in our configura-
tion is only about one-third of that on the RRDE disk. This
underlines the high sensitivity and excellent signal-to-noise
ratio of the DTLFC-DDE set-up for peroxide detection due
to the ca. 4-fold higher collection efficiency than in a stan-
dard RRDE set-up (see Section 3.1).
posed) on the Pt nanoparticles. These additional side reac-
tions complicate a direct comparison of the catalytic activ-
ity of carbon-supported fuel cell catalysts and massive no-
ble metal model electrodes surfaces with respect to peroxide
formation.
To better separate the complex features in peroxide for-
mation rate as a function of the electrode potential occurring
in potentiodynamic experiments (Fig. 5b) and to determine
the time resolution of the DTLFC-DDE set-up in experi-
ments involving a change in electrolyte, we performed tran-
sient measurements of the ORR on Pt/Vulcan and Vulcan
carbon thin-film electrodes at a constant electrode potential
(0.06 V), switching from a (stagnant) Ar-saturated 0.5 M sul-
furic acid electrolyte to an O2-saturated 0.5 M sulfuric acid
electrolyte flow (Fig. 6). On both electrodes, the ORR cur-
rents attain constant, mass-transport-limited values ca. 5 s
after the O2-saturated electrolyte starts to reach the respec-
tive electrodes (90% of the limiting value in ca. 2 s) (Fig. 6a).
These time constants reflect the time required to completely
replace the electrolyte film on the electrode by O2-containing
solution. In order to better illustrate the time resolution we
also show the signal for the Pt/Vulcan electrode on a 10-fold
The hydrogen peroxide yield in the ORR, X(H2O2), is
quantified via [40]
(2 × Icoll/N)
X(H2O2) =
,
(4)
(Igen + Icoll/N)
where Icoll and Igen are the currents on the collector and
generator, respectively, and N is the collection efficiency
determined for ferrocyanide reduction (N = 0.76, see
Section 3.1) (the resulting value is multiplied by 100 for
the yield in percentage). The corresponding scale on the
right-hand axis of Fig. 5b shows that the maximum perox-
ide formation yield achieved at 0.06 V is about 4%. This
value is in an excellent agreement with the 5–6% perox-
ide yields at this potential detected on the same Pt/Vulcan
catalyst in RRDE measurements (at 1600 rpm and 60 ◦C),
taking into account that part of the peroxide formed on the
Pt/Vulcan electrode can be decomposed during the transfer
to the collector (ca. 3 s) in our set-up, while in the standard
RRDE configuration the transport times are much smaller
and the peroxide detection is almost instantaneous.
Since O2 is known to be reduced mainly to peroxide on
carbon at potentials negative of 0.3 V [40,49], we also per-
formed test experiments on a Pt-free Vulcan thin-film elec-
trode (Fig. 5, dash-dotted lines). The CV of the Vulcan
thin-film electrode shows an exponentially increasing oxy-
gen reduction current at potentials negative of 0.4 V (Fig. 5a,
dash-dotted line), along with a simultaneous increase in per-
oxide formation (Fig. 5b, dash-dotted line). Note that the
peroxide formation current at the negative potential limit
(0.06 V) on Vulcan carbon is ca. 7 times higher (scaled
down by one order of magnitude in Fig. 5b) compared to
that on the Pt/Vulcan catalyst (Fig. 5b, full line/dashed line),
even though the oxygen reduction current on Vulcan is only
about one third of that (Fig. 5a). This suggests a high cur-
rent efficiency for oxygen reduction to peroxide on Vul-
can carbon (ca. 50%, according to Eq. (4)), in agreement
with previous RRDE data [50]. In the presence of catalyti-
cally active Pt nanoparticles on the Vulcan carbon support,
however, the ORR proceeds preferably via the four elec-
tron pathway (on Pt), resulting in a decrease in peroxide
yield by 90% compared to Vulcan carbon (Fig. 5b). Assum-
ing that on Pt/Vulcan the ORR occurs on both carbon and
Pt and that both reaction channels contribute to peroxide
formation, this indicates that part of the H2O2 formed on
the carbon support is further reduced to water (or decom-
Fig. 6. Current transients recorded simultaneously after starting to flow
O2-saturated electrolyte through the flow cell on the Pt/Vulcan electrode
((a), solid black line) (Pt loading 28 ␮g/cm2) and on a Vulcan carbon
electrode ((a), dash-dotted black line) (loading 112 ␮g/cm2) at E = 0.06 V
in the first compartment, and on a polycrystalline Pt electrode ((b),
corresponding solid lines) at E = 1.16 V in the second compartment
of a dual thin-layer flow cell (electrolyte flow rate ca. 4 ␮l/s, room
temperature). The right-hand axis in (b) corresponds to the peroxide yields
on the Pt/Vulcan and carbon Vulcan electrodes, respectively. The initial
response of the current signal on the Pt/Vulcan electrode is more clearly
resolved on a tenfold magnified time scale (full gray lines marked with
dots which represent the individual measurements at the acquisition rate
of one point per second). For better comparison with the electrochemical
current signal on Pt/Vulcan in Fig. 6a (gray line) the collector current in
Fig. 5b (gray line) is multiplied by a factor of 3.

1304
Z. Jusys et al. / Electrochimica Acta 49 (2004) 1297–1305
expanded time scale (gray line with dots). The dots indicate
the individual measurements at an acquisition rate of one
point per second. The collector current is in addition multi-
plied by a factor of three to obtain signals of similar mag-
nitude in Fig. 6a and b. The difference in the time response
between the two signals, i.e., the smoother increase of the
signal in Fig. 6b, reflects the additional smear out occurring
during the electrolyte transport from the first to the second
compartment. Likewise, the larger time constant in this ex-
periment compared to that in the potential jump experiment
can be explained by the time required to build up a constant
flow situation at the Pt/Vulcan electrode after opening the
electrolyte valve, while in the latter case the potential jump
occurs under constant flow conditions and the time resolu-
tion is determined only by the transfer of reactive species
to the collector. (The response time of the potential is much
lower than that for reaching steady-state flow conditions.)
The ORR current on the Vulcan carbon electrode
(Fig. 6a, dash-dotted line) is ca. 2.5 times lower than that
on Pt/Vulcan (Fig. 6a, black line). On the other hand, the
hydrogen peroxide oxidation current on the Vulcan carbon
electrode (Fig. 6b, dash-dotted line) is about 3 times higher
than that on the Pt/Vulcan electrode (Fig. 6b, black line).
In combination, these two results suggest different reaction
pathways on both electrodes. The peroxide yields obtained
at 0.06 V on the respective generator electrodes (Fig. 6b),
which are calculated according to Eq. (4), are about 6.5%
on the Pt/Vulcan electrode (black line) and 42.5% on the
Vulcan carbon electrode (dash-dotted line), respectively.
The slightly different values for peroxide yields derived in
the potentiodynamic measurements on the two electrodes
(Fig. 5) most likely result from both the time delay in
peroxide detection and pseudo-capacitive contributions.
Overall, the potentiostatic measurements demonstrate that
the time resolution of the DTLFC-DDE set-up is sufficient
for performing transient kinetic measurements which in-
volve electrolyte exchange. In combination with the high
collection efficiency this method is therefore particularly
suitable for mechanistic and kinetic studies of electrocat-
alytic reactions where electrolyte exchange is mandatory.
A typical example is the study of the chemical CO oxida-
tion by O2 in the anode of a polymer electrolyte fuel cell
(PEFC) and the simultaneous ORR on COad-covered anode
catalysts, which is highly interesting for the understanding
of the reactions going on during ‘air-bleed’ operation of PE-
FCs (see [25,51]).
flow conditions electro-active species formed at the gener-
ator are detected at the potential-controlled collector with
a delay of ca. 3 s and a collection efficiency of 76%. Us-
ing thin-film electrodes it allows time-resolved studies of
electrocatalytic reactions on supported catalysts or model
catalysts, including single-crystal electrodes, under defined
electrolyte flow conditions, which makes it particularly at-
tractive for model studies on fuel cell reactions under real-
istic conditions (continuous reaction, continuous electrolyte
flow) and on realistic materials (supported catalysts) under
nevertheless well-defined conditions. Most important, and
its main advantage in comparison to the commonly used
RRDE configuration, however, is its ability to quickly ex-
change the electrolyte during measurements, which allows
kinetic studies of electrocatalytic reactions that require a
change in electrolyte. The time resolution in transient ex-
periments involving a change in electrolyte, defined by the
time required to reach 90% of the final signal, is about 2 s.
The performance of this set-up is tested in potentiody-
namic and potentiostatic measurements of the ORR on sup-
ported Pt/Vulcan catalysts and, for comparison, on Pt-free
Vulcan carbon support. The good agreement with exist-
ing RRDE data demonstrates the general suitability of the
set-up. The higher sensitivity allowed to clearly identify
a maximum around 0.78 V in the collector current in the
negative-going scan of a CV in O2-containing electrolyte,
which is attributed to H2O2 formation on partly oxide cov-
ered Pt catalyst under these conditions. The significantly
higher rate for H2O2 formation at potentials negative of
0.3 V, with a H2O2 formation yield of more than 40% on
Pt-free Vulcan carbon, compared to that on the Pt/Vulcan
catalyst indicates that on the latter catalyst part of the H2O2
formed on the carbon support is further reduced to H2O
or decomposed. Transient experiments changing from an
Ar-saturated electrolyte to a flow of O2-saturated electrolyte
under potential conditions, where the electrodes are active
for O2 reduction and H2O2 oxidation, respectively, illustrate
the time resolution of the set-up and its suitability for tran-
sient kinetic studies of such type.
Acknowledgements
We gratefully acknowledge financial support by the
state of Baden-Württemberg, within the ‘Zukunftsoffensive
Junge Generation’, by the Federal Ministry of Education,
Research and Technology (BMBF, Grant 01SF0053) and
by the Deutsche Forschungsgemeinschaft (Be 1201/8-4).
4. Summary
We have presented a novel double-disk electrode design,
which is based on a dual thin-layer flow cell. Due to its
high collection efficiency in combination with a reasonable
time resolution it allows the highly sensitive determination
of electro-active species generated during electrocatalytic re-
actions on the working electrode. Under present continuous
References
[1] V.G. Levich, Physicochemical Hydrodynamics, Prentice Hall, Eagle-
wood Cliffs, NJ, 1962.
[2] T.J. Schmidt, H.A. Gasteiger, in: W. Vielstich, A. Lamm, and
H.A. Gasteiger (Eds.), Handbook of Fuel Cells—Fundamentals

Z. Jusys et al. / Electrochimica Acta 49 (2004) 1297–1305
1305
Technology and Applications, Vol. 2. Electrocatalysis, John Wiley
& Sons, New York, 2003, p. 316.
[25] Z. Jusys, J. Kaiser, R. J. Behm, J. Electroanal. Chem. 554/555 (2003)
427.
[3] N.M. Markovic, in: W. Vielstich, A. Lamm, H.A. Gasteiger (Eds.),
Handbook of Fuel Cells—Fundamentals Technology and Applica-
tions, Vol. 2. Electrocatalysis, John Wiley & Sons, New York, 2003,
p. 368.
[4] P.N. Ross, in: W. Vielstich, A. Lamm, and H. A. Gasteiger (Eds.),
Handbook of Fuel Cells—Fundamentals Technology and Applica-
tions, Vol. 2. Electrocatalysis, John Wiley & Sons, New York, 2003,
p. 465.
[26] S.G. Weber, J.T. Long, Anal. Chem. 60 (1988) 903A.
[27] T.J. Schmidt, H.A. Gasteiger, G.D. Stäb, P.M. Urban, D.M. Kolb,
R.J. Behm, J. Electrochem. Soc. 145 (1998) 2354.
[28] Z. Jusys, J. Kaiser, R.J. Behm, Langmuir 19 (2003) 6759.
[29] T.J. Schmidt, H.A. Gasteiger, R.J. Behm, J. Electrochem. Soc. 146
(1999) 1296.
[30] B. Rush, J.A. Reimer, E.J. Cairns, J. Electrochem. Soc. 148 (2001)
A137.
[5] W.J. Albery, M.L. Hitchman, Ring-Disc Electrodes, Clarendon Press,
Oxford, 1971.
[6] A.N. Frumkin, L. Nekrasov, V.G. Levich, J. Ivanov, J. Electroanal.
Chem. 1 (1953) 84.
[7] J. Yamada, H. Matsuda, J. Electroanal. Chem. 44 (1973) 189.
[8] W.J. Albery, J. Electroanal. Chem. 191 (1985) 1.
[9] P. Laevers, A. Hubin, H. Teryn, J. Vereeken, J. Appl. Electrochem.
25 (1995) 1017.
[31] V.S. Bagotzky, Y.B. Vassilyev, Electrochim. Acta 12 (1967) 1323.
[32] T. Biegler, D.A.J. Rand, R. Woods, J. Electroanal. Chem. 29 (1971)
269.
[33] A.I. Frenkel, C.W. Hills, R.G. Nuzzo, J. Phys. Chem. B 105 (1997)
12689.
[34] H. Angerstein-Kozlowska, B.E. Conway, W.B.A. Sharp, J. Elec-
troanal. Chem. 43 (1973) 9.
[35] Z. Jusys, H. Baltruschat, unpublished.
[10] W.J. Albery, C.C. Jones, A.R. Mount, in: R.G. Compton, A. Hamnett
(Eds.), Chemical Kinetics, Vol. 29, Elsevier Publications, Amsterdam,
1986.
[11] C.M.A. Brett, A.M.F.C. Oliveira Brett, in: C.H. Bamford, R.G.
Compton (Eds.), Chemical Kinetics, Vol. 26, Elsevier Publications,
Amsterdam, 1986.
[36] P. Laevers, A. Hubin, H. Teryn, J. Vereeken, J. Appl. Electrochem.
25 (1995) 1023.
[37] H. Wang, T. Löffler, H. Baltruschat, J. Appl. Electrochem. 30 (2001)
759.
[38] I.E. Henley, K. Yunus, A.C. Fisher, J. Phys. Chem. B 107 (2003)
3878.
[12] P.R. Unwin, R.G. Compton, in: C.H. Bamford, R.G. Compton
(Eds.), Chemical Kinetics, Vol. 26, Elsevier Publications, Amsterdam,
1986.
[39] B. Soucaze-Guillous, W. Kutner, Electroanalysis 9 (1997) 32.
[40] U.A. Paulus, T.J. Schmidt, H.A. Gasteiger, R.J. Behm, J. Elec-
trochem. Soc. 495 (2001) 134.
[13] M. Bergelin, M. Wasberg, J. Electroanal. Chem. 449 (1998) 181.
[14] M. Bergelin, J.M. Feliu, M. Wasberg, Electrochim. Acta 44 (1998)
1069.
[15] M. Bergelin, E. Herrero, J.M. Feliu, M. Wasberg, J. Electroanal.
Chem. 467 (1999) 74.
[16] U. Koponen, T. Peltonen, M. Bergelin, T. Mennola, M. Valkiainen,
J. Kaskimies, M. Wasberg, J. Power Sources 86 (2000) 261.
[17] T.H. Madden, N. Arvindan, E.M. Stuve, J. Electrochem. Soc. 150
(2003) E1.
[41] N.M. Markovic, H.A. Gasteiger, P.N. Ross, J. Phys. Chem. 99 (1995)
3411.
[42] S. Gottesfeld, I.D. Raistrick, S. Srinivasan, J. Electrochem. Soc. 134
(1987) 1455.
[43] S.L. Gojkovic, S.K. Zecevic, R.F. Savinell, J. Electrochem. Soc. 145
(1998) 3713.
[44] M.R. Tarasevich, A. Sadkowski, E. Yeager, in: J.O.M. Bockris et al.
(Eds.), Comprehensive Treatise in Electrochemistry, Plenum Press,
New York, 1983.
[18] N. Wakabayashi, H. Uchida, M. Watanabe, Electrochem. Solid State
Lett. 5 (2002) E62.
[45] A. Damjanovic, M.A. Genshaw, J.O. Bockris, J. Electrochem. Soc.
114 (1967) 466.
[19] H. Gerischer, I. Mattes, R. Braun, J. Electroanal. Chem. 10 (1965)
553.
[46] N. Markovic, H.A. Gasteiger, P.N. Ross, J. Electrochem. Soc. 144
(1997) 1591.
[20] W.J. Albery, C.M.A. Brett, J. Electroanal. Chem. 148 (1983) 201.
[21] R.G. Compton, A.C. Fisher, M.H. Latham, C.M.A. Brett, A.M.C.F.
Oliveria Brett, J. Appl. Electrochem. 22 (1992) 1011.
[22] A. Paixaoa, R.C. Matos, M. Bertotti, Electrochim. Acta 48 (2003)
691.
[47] U.A. Paulus, A. Wokaun, G.G. Scherer, T.J. Schmidt, V. Stamenkovic,
V. Radmilovic, N.M. Markovic, P.N. Ross Jr, J. Phys. Chem. B 106
(2002) 4181.
[48] T.J. Schmidt, U.A. Paulus, H.A. Gasteiger, R.J. Behm, J. Electroanal.
Chem. 508 (2001) 41.
[23] K. Toda, S. Oguni, Y. Takamatsu, I. Sanemasa, J. Electroanal. Chem.
479 (1999) 57.
[49] K. Kinoshita, Electrochemical Oxygen Technology, John Wiley &
Sons, New York, 1992.
[24] Z. Jusys, H. Massong, H. Baltruschat, J. Electrochem. Soc. 146
(1999) 1093.
[50] U.A. Paulus, Diploma Thesis, University of Ulm, 1998.
[51] Z. Jusys, J. Kaiser, R.J. Behm, submitted for publication.