Bifunctional Oxygen Electrodes
L Jöerissen, Zentrum für Sonnenenergie- und Wasserstoff-Forschung Baden-Württemberg, Ulm, Germany
& 2009 Elsevier B.V. All rights reserved.
Introduction
In many parts of the world, increasing amounts of
intermittent energy sources such as solar or wind energy,
are fed to the electric grid. In order to level out the daily
and seasonal fluctuations associated with these, storage of
electrical energy is becoming an increasingly important
issue. Therefore, all kinds of energy storage systems in-
cluding secondary batteries are considered to be relevant
for grid management.
In secondary batteries, the amount of energy stored is
directly related to the amount of active masses in the
battery. The use of oxygen as ‘active mass’ at the positive
electrode could become a viable option since oxygen is a
part of the atmosphere and therefore does not add to the
battery mass.
Understanding the aqueous electrochemistry of oxygen,
which is of practical importance for many technical pro-
cesses is fairly complex. Oxygen electrochemistry is the key
to understanding air breathing secondary battery systems.
The oxygen reduction reaction (ORR) is the key to fuel
cell systems as well as secondary batteries in the discharge
mode; it also is relevant as the compensatory process in
corrosion reactions. Oxygen evolution is relevant in the
charging of secondary batteries involving oxygen elec-
trodes, it also takes place in water electrolysis and as a
competing reaction in other technical synthesis reactions.
The use of oxygen as a reactant at the positive elec-
trode of an electrochemical energy storage system leads
to metal–air batteries or in the case of hydrogen is used at
Electricity
Storageof negative active material
Storage reservoir of oxygen
Electrolyzer
Fuelcell
Electricity
(a)
Figure 1 Concept of electrochemical power sources involving oxygen as a reactant. (a) Combination of electrolyzer and fuel cell.
(b) Secondary battery or unitized regenerative fuel cell (URFC).
356
the negative electrode, to the so-called regenerative fuel
cells. Evolution and consumption of oxygen can take
place in separate reactors such as an electrolyzer and fuel
cell (Figure 1(a)) or in the same reactor becoming a
secondary metal–oxygen battery or in the case of hydrogen
is used at the negative electrode, a so-called unitized
regenerative fuel cell (URFC) (Figure 1(b)). In the latter
case, bifunctional oxygen/air electrodes consuming
oxygen during electricity generation (e.g., battery dis-
charge, fuel cell operation) and evolving oxygen during
battery charge or electrolysis mode are frequently used.
When assuming oxygen as part of the atmosphere and
neglecting its weight, electrochemical energy storage
systems using oxygen at the positive electrode stand out
for their high specific energy and energy density. For
practical purposes, it would be beneficial if the metal
could be deposited from an aqueous electrolyte without
significant hydrogen evolution as a side reaction.
From the metals shown in Table 1, zinc, iron, cad-
mium, and metal hydride electrodes can be charged
electrically in an aqueous electrolyte. The addition of
lead, mercury, or cadmium can further suppress hydro-
gen evolution upon charge. Lithium, calcium, mag-
nesium, and aluminum systems can only be operated as a
primary system in an aqueous environment. Considering
environmental aspects, additives of lead or mercury as
well as cadmium–air systems will most probably not be
used due to the high toxicity of the heavy metals.
Typically, aqueous alkaline electrolytes are used in
metal–air batteries due to the better corrosion resistance
Electricity
Storageof negative active material
Storage reservoir of oxygen
Charge
Secondary battery
Discharge
Electricity
(b)
 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes 357

Table 1 Characteristic data of metal–oxygen cells calculated

Metal anode Electrochemical Theoretical cell voltage Theoretical specific Practical operating
equivalent of metal with O2 electrode (V) energy of metal–oxygen voltage metal–oxygen
(Ah kg1) couple (Wh kg1) couple (V)
Li 3861 3.3 12 741 2.4
Ca 1337 3.4 4547 2.0
Mg 2205 3.1 6837 1.4
Al 2980 2.7 8046 1.6
Zn 820 1.6 1312 1.1
Fe 960 1.3 1248 1.0
Cd 478 1.2 572 0.9
H6LaNi5 367 1.2 440 0.8
Reproduced from Barin Y (1989) Thermochemical Data of Pure Substances. Weinheim: VCH and Linden D (1984) Handbook of Batteries and Fuel
Cells. New York: McGraw-Hill.

of the metals and the better performance of the oxygen Table 2 Standard electrode potentials of selected oxygen
electrode in the alkaline environment. reactions in aqueous electrolytes at 25 1C
Nevertheless, when considering hydrogen as the ac- Electrochemical reaction E0/V vs SHE
tive material at the negative electrode, progress in the O2 þ 4H þ 4e -2H2O
þ 
1.229
development of fuel cells and electrolyzers involving acid O2 þ 2Hþ þ 2e-H2O2 0.695
polymer membranes makes the development of revers- O2 þ 2H2O þ 4e-4OH 0.401
ible polymer membrane fuel cells an interesting option O2 þ Hþ þ e-HO2  0.053
O2 þ H2O þ 2e-HO2 þ OH

 0.065
for energy storage. Besides energy storage, in principle O2 þ 2H2O þ 2e-H2O2 þ 2OH  0.133
metal–air cells can also be used as oxygen sensors. O2 þ e-O 2  0.284
Reproduced from Linden D (1984) Handbook of Batteries and Fuel
Cells. New York: McGraw-Hill.
The Electrochemistry of Oxygen SHE, standard hydrogen electrode.

Bifunctional oxygen electrodes must be stable under con-

ditions of oxygen reduction and oxygen evolution. Both
reactions show considerable high overpotentials. The op-
erating range of a bifunctional oxygen electrode spans from
B0.7 V versus RHE during discharge to more than 1.8 V
versus RHE during charge. In the following paragraphs,
a short overview on the electrochemistry of the oxygen
reduction and oxygen evolution reactions is given.
Oxygen Reduction
It is well known that the electrochemical reaction kin-
etics of oxygen is generally rather slow. Although it may
appear simple at first glance, the electrochemical re-
actions taking place during oxygen reduction are rather
complicated to describe due to the strong irreversibility
of individual steps and a strong dependence of their re-
action mechanism on pH. In Table 2, the standard
electrode potentials of reactions involving oxygen in
different electrolytes are given.
The reversible potentials of oxygen electrodes as they
are calculated from thermodynamic considerations are
difficult to measure experimentally, since the reactions
taking place involve many steps and intermediate species.
Furthermore, mixed potential formation at the catalyst
surface is frequently observed. In addition, the actual
species involved are dependent on many factors such as
the electrolyte and the surface chemistry of the catalyst
used and so on.
A comprehensive presentation of the reaction path-
ways involved in the reduction of oxygen has been
worked out by N. A. Anastasijevic and coworkers for
alkaline electrolytes (Figure 2) and converted into acid
electrolytes by Gattrell and MacDougall (Figure 3). Both
papers are referenced in the ‘Further Reading’ section.
The overall reaction scheme involves only single electron
transfer steps. Table 3 shows the electrochemical and
chemical reactions and equilibria relevant in Figure 2.
The reactions in acid electrolyte can be constructed
straightforwardly by adding a proton to the charged
intermediates. Reaction [V] is known to be highly ir-
reversible in acid electrolytes. The individual reactions
shown may involve more elementary reaction steps. The
oxygen reduction initially involves the adsorption of
molecular O2 on the electrode surface.
The following sequence of reactions may involve
pathways along strongly adsorbed species (reactions [III],
[IV], [IX], [X]) where the individual reaction steps are
considered irreversible. It is assumed that the strongly
adsorbed (chemisorbed) species cannot desorb directly
into the electrolyte.
A second possible path can be constructed as a series
of reactions involving only weakly adsorbed (physis-
orbed) species. This pathway involves a superoxide ion
(O2; adsorbed ) which is not stable in aqueous solution, how-
ever, can react with water and an electron through eqn [II]
358 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes

k3(e–)

k12 OH–

Ostrongly, adsorbed

k21 k13

Diffusion k4(e–) k5(e–) k6(e–)

O2, bulk O2, Surface k22 O2–, strongly adsorbed HOO– O–strongly adsorbed OH–
strongly adsorbed
K–20 k–5(e–)
k23 k24
k20
k2(e–)
k1(e–)
O2, adsorbed O2–, adsorbed HOO–adsorbed
k–2(e–)
k–25 k25

k10(2) HOO– surface

Diffusion
k11(2)
HOO–bulk

Figure 2 Reaction mechanisms of the oxygen reduction reaction in alkaline electrolytes. Reproduced with permission from
Anastsijevic NA, Vesovic V, and Adzic RR (1987) Journal of Electroanalytical Chemistry 229: 305–316.

k3(e–)

k12 H2O

Ostrongly, adsorbed
k21 k13

k5(e ) –
Diffusion O k4(e–) k6(e–)
O2, bulk 2, surface k22 HO2, strongly adsorbed HOOH strongly adsorbed OHstrongly adsorbed H2O
k–20 k–5(e–)
k24
k20 k2(e–)
k1(e–) HOOHadsorbed
O2, adsorbed HO2, adsorbed
k–2(e–)
k–25 k25

k10(2) HOOHsurface

k11(2) Diffusion

Figure 3 Reaction mechanisms of the oxygen reduction–reaction in acid electrolytes. Reproduced from Gattrell M and McDougall B
(2003) Reaction mechanisms of the O2 reduction/evolution reaction. In: Vielstich W, Lamm A, and Gasteiger HA (eds.) Handbook of
Fuel Cells  Fundamentals, Technology and Applications, vol. 2, pp. 443–464. New York: John Wiley & Sons.

or by the decomposition reaction [VII] to form a peroxide
ion HO 2; adsorbed which can desorb from the electrode
surface into the electrolyte according to eqn [VII].
Transitions from reaction pathways involving weakly
bound species to strongly bound species are possible.
More insight into oxygen adsorption has been gained
by gas-phase adsorption studies showing strong inter-
actions and dissociative adsorption on Pt-catalysts.
Interaction of the antibonding oxygen orbital with the
metal d-band has been found to be the key factor.
Atomically adsorbed oxygen species could even be
imaged by scanning tunneling microscopy. Strong non-
uniformity of the oxygen adsorbates has been observed.
Molecular modeling of oxygen reduction is in the
initial stage. Indications could be found that the pathway
for oxygen reduction varies depending on the mode of
oxygen adsorption.
For all practical considerations, oxygen reduction can
take place through parallel pathways, one leading to
water or OH through a direct four-electron reduction,
the other one leading to peroxide through a two-electron
reduction path. The overall reactions and the associated
 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes 359

Table 3 Partial reactions taking place during oxygen reduction

Electrochemical reactions taking place in alkaline electrolyte
k1 [I]
O2; adsorbed þ e - O
2; adsorbed

k2 =k2 [II]
O
2; adsorbed þ H2 O þ e - HO
2; adsorbed þ OH
k3 [III]
Ostrongly adsorbed þ e - O
strongly adsorbed

k4 [IV]
O
2; strongly adsorbed þ H2 O þ e - HO
2; strongly adsorbed þ OH
k5 =k5 [V]
HO
2; strongly adsorbed þ e

- O
strongly adsorbed þ OH


k6 [VI]
O
strongly adsorbed þ H2 O þ e - 2OH
 

Chemical reactions of adsorbates to be taken into account

k10 [VII]
2O
2; adsorbed þ H2 O - O2; adsorbed þ HO
2; adsorbed þ OH
k11 [VIII]
2HO
2; adsorbed - 2OH þ O2; adsorbed

k12 [IX]
HO 
2; strongly adsorbed - OH þ Ostrongly adsorbed

k13 [X]
O
2; strongly adsorbed - Ostrongly adsorbed þ O
strongly adsorbed

Reproduced from Anastasijevic NA, Vesovic V, Adzic RR (1987) Journal of Electroanalytical

Chemistry 229: 305–316.

Table 4 Potentials associated to oxygen reduction in alkaline and acidic media

Reaction E0/V versus SHE
Direct four-electron pathway
Alkaline solution O2 þ 2H2O þ 4e-4OH 0.401
Acid solution O2 þ 4Hþ þ 4e-2H2O 1.229
Peroxide pathway
Alkaline solution O2 þ H2O þ 2e-HO2 þ OH  0.065
Peroxide reduction HO2  þ H2 O þ 2e -3OH 0.867
Peroxide decomposition 2HO2  -2OH þ O2 –
Acid solution O2 þ 2Hþ þ 2e-H2O2 0.67 V
Peroxide reduction H2O2 þ 2Hþ þ 2e-2H2O 1.77 V
Peroxide decomposition 2H2O2-2H2O þ O2 –
Reproduced from Yeager E (1984) Electrochimica Acta 29: 1527–1537.

potentials versus standard hydrogen electrode (SHE)
taking place are given in Table 4.
The peroxide path is favored on graphite and most
other carbons, gold (with the exception of gold(100) in
alkaline solution), oxide covered metals such as nickel or
cobalt, some transition metal macrocycles, and transition
metal oxides.
The so-called direct four-electron reduction generally
involves a sequence of steps, sometimes involving even
adsorbed peroxide intermediates (Figures 2 and 3). This
reaction pathway is predominant on noble metal catalysts
such as platinum and platinum group metals, some
transition metal oxides having the perovskite or pyro-
chlore structure, some transition metal macrocycles such
as iron phthalocyanine, and silver.
In the peroxide pathway, free H2O2 can be detected in
the electrolyte, which can be used as an indicator. The
H2O2 formed in solution can be decomposed either
electrochemically or by simple chemical decomposition.
Oxygen from the chemical decomposition reaction can
immediately react at the electrode. Therefore, a total of
four electrons are exchanged in both pathways.
Oxygen Evolution
On metallic catalysts, oxygen is generally evolved from
the corresponding metal oxides which may be either thin
films as in the case of noble metals (platinum, gold,
iridium, rhodium, and ruthenium) or thick films as in the
360 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes
case of silver or nickel. In general, oxide film growth can
be considered as a parallel reaction to oxygen evolution.
Besides being catalytically active, an oxygen evolution
catalyst should be stable and conductive at the high po-
tentials where oxygen evolution takes place. Metal redox
couples are probably present at the electrode surface.
Platinum is a rather poor electrocatalyst for oxygen
evolution due to the formation of a stable oxide layer of
low conductivity.
Ruthenium metal, although being a good electro-
catalyst for oxygen evolution is strongly corroded at the
potential of oxygen evolution. However, electrically
conductive bulk noble metal oxides such as ruthenium
oxide (RuO2) or iridium dioxide (IrO2) show rather good
performance. On the contrary, iridium dioxide (IrO2)
needs to be stabilized by thermal treatment or by doping
with tantalum.
Nickel is widely used as an electrocatalyst for oxygen
evolution in alkaline electrolytes. However, at high po-
tentials lower conductivity phases involving Ni4þ are
formed which reduce the performance.
Some transition metal oxides also show good per-
formance in oxygen evolution. This is particularly the
case in some spinel-type oxides such as nickel cobolt
oxide (NiCo2O4) or copper cobalt oxide (CuCo2O4).
The reaction pathways involved in oxygen evolution
often are rather complex. Mechanisms are also likely to
change depending on the electrode potential. On metallic
surfaces, the rate-determining step is the electron
transfer from either H2O in acid electrolytes or OH in
alkaline electrolytes to form adsorbed OH radicals at the
electrode surface.
SurfaceðH2 OÞadsorbed -SurfaceðOHÞadsorbed þ Hþ þ e ½XI
Followed by two possible reactions
2SurfaceðOHÞadsorbed -SurfaceðOÞadsorbed þ H2 O ½XII
SurfaceðOHÞadsorbed -SurfaceðOÞadsorbed þ Hþ þ e ½XIII
Completed by two consecutive reaction steps
2SurfaceðOÞadsorbed -SurfaceðO2 Þadsorbed ½XIV
SurfaceðO2 Þadsorbed-O2 þ Surface ½XV
Oxygen evolution from metal oxide surfaces in alkaline
media leads to the intermediate formation of H2O2 where
the rate-determining step was found to be the breaking of
the surface-OH bond.
Surface þ OH -SurfaceðOHÞ þ e ½XVI
SurfaceðOHÞ þ OH -SurfaceðH2 O2 Þphysisorbed þ e ½XVII
Followed by
SurfaceðH2 O2 Þphysisorbed þ OH
-SurfaceðHO2  Þphysisorbed þ H2 O ½XVIII
SurfaceðH2 O2 Þphysisorbed þ SurfaceðHO2  Þphysisorbed
-H2 O þ OH þ O2 ½XIX
Changes in the oxidation state of the metal ion of the
electrocatalyst itself can be taken into account by the
following sequence:
Surface þ OH -ðSurface-OHÞ þ e ½XX
ðSurface-OHÞ-ðSurface-OHÞþ1 þ e ½XXI
2ðSurface  OHÞþ1 þ 2OH -2Surface þ O2 þ 2H2 O ½XXII
Obviously, the electronic properties determined by the d-
electrons, the composition and stoichiometry of the metal
oxide, as well as the vacancies present in the metal oxides
are important factors for the oxygen evolution reaction.
Also geometric factors in the catalyst lattice are of
importance.
Design of Bifunctional Oxygen/Air
Electrodes
Owing to the slow kinetics involved in oxygen electro-
chemistry, porous electrodes involving high-surface area
catalysts are required to achieve significant current
densities. Porous gas diffusion-type electrodes are used
for the oxygen reduction reaction. The porous structure
must be optimized to provide an extended reaction zone
at the three-phase boundary between gas, catalyst, and
electrolyte. Although the catalytic reaction site can be
covered by a thin film of electrolyte, an open, rather
hydrophobic gas transport pore system must be main-
tained in order to optimize the gas access. Often the gas
diffusion electrode itself is divided into two layers. The
layer next to the electrolyte contains the catalytic active
material as a thin, mildly hydrophobic layer having fine
porosity, whereas a highly hydrophobic gas diffusion
layer provides bigger gas transport pores. In addition, a
current collector such as a metal mesh or foam or a
graphite felt or cloth is incorporated into the gas dif-
fusion part of the electrode.
In technical electrochemistry porous catalytic elec-
trodes sintered on a metallic substrate such as titanium
are used for oxygen evolution. Good examples are the so-
called dimensionally stable anodes which are widely used
in the chlor-alkaline industry. Porous, hydrophilic cata-
lyst layers (e.g., mixtures from ruthenium oxide, iridium
dioxide, and tantalum pentoxide (Ta2O5) are sintered on
 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes
Separate electrodes,
Single electrode,
single catalytic layer
single catalytic layer
O2 O2
Evolution
Consumption
Consumption
Bifunctional
catalyst layer
Electrolyte
(a) (b)
Figure 4 Design principles of bifunctional oxygen/air electrodes.
the corrosion-resistant titanium substrate. Evolving gases
are simply removed by bubble formation.
The function of bifunctional oxygen electrodes gen-
erally involves a combination of both operating prin-
ciples. Therefore, hydrophobic catalyzed sites must be
present for oxygen reduction as well as hydrophilic sites
for oxygen evolution. Sometimes a mixture of different
catalyst materials is used for oxygen reduction and
oxygen evolution.
Different design principles can be adopted for an
electrically rechargeable electrochemical power source
involving air (Figure 4).
• consumption,
Use of separate electrodes for oxygen evolution and
• for
Use of a multilayered electrode with separate layers
Oxygen evolution and consumption,
• catalyst.
Use of single-layered electrodes with a bifunctional
Figure 4(a) shows a single-layer electrode handling both
reactions. This construction is relatively simple to
manufacture and to integrate into batteries. However, it
requires the use of highly stable catalyst materials and
conductive support materials, which must withstand the
highly oxidizing conditions upon oxygen evolution as
well as the more reducing conditions observed under
oxygen reduction at high-current density. Such elec-
trodes have been described for use in electrically re-
chargeable metal-hydride–air batteries and in zinc–air
batteries.
361
Single electrode,
multiple catalytic layers
O2 O2
Evolution
Evolution
Consumption
Electrolyte Electrolyte
(c)
In Figure 4(b) the separation of oxygen evolution and
oxygen reduction into two different layers is shown. This
construction offers the possibility of separate optimiza-
tion of oxygen-reducing and oxygen-evolving electrodes
with respect to performance and stability. However it
adds to the complexity and weight of the system. Elec-
trically rechargeable metal-hydride–air batteries using
separate electrodes have been described.
A double-layered electrode construction is shown in
Figure 4(c) and in more detail in Figure 5. Here the
hydrophilic side for oxygen evolution which can easily
have a flooded pore system is facing the electrolyte
whereas the part used for oxygen reduction is oriented
toward the air side. A current collector could be placed
between the two functional layers. Such electrodes could
provide better performance on the expense of higher
complexity during manufacturing and possibly increased
mass transfer losses. A bifunctional air electrode design
using layers of different porosity and hydrophobicity
made from sintered nickel as a base and silver as a
catalyst has been described. The electrode consisted of a
fine-grained reaction and coarse-grained transport layer.
Spacers and hydrophobing agents were used to control
the porosity and wettability of the electrode respectively.
Stability of more than 100 cycles in charge/discharge
mode has been achieved. Recently, thin-layer electrodes
using vacuum-deposited catalytic layers have been in-
vestigated showing great promise to reduce the overall
catalyst loading and electrode weight.
During charge of bifunctional oxygen/air electrodes,
oxygen is generated in a highly reactive form thus
362 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes
Electrolyte
Figure 5 Possible microstructure of a multilayered bifunctional electrode.
causing highly corrosive conditions to the catalyst as
well as to the conductive support material. Particularly
high-surface area carbon materials frequently used as
conductive support are severely attacked under the
conditions of oxygen evolution. For high endurance, it is
of utmost importance to prevent the corrosion of catalyst
and catalyst support either by the use of highly cor-
rosion-resistant carbon materials or by the use of metallic
substrates. Recently, niobium-doped titanium dioxide
(TiO2) has been proposed as a support material stable
under oxygen reduction and oxygen evolution conditions.
In multilayered electrodes, the oxygen reduction
catalyst is protected by a dual porosity design. Oxidation
of the electrode structure during charge is minimized by
applying a slight pressure to the gas side thus breaking the
electrolytic contact to the ‘oxygen reduction’ part of the
electrode.
For high-power applications, the bifunctional oxygen
electrode has to accept rapid changes in its potential
imposed by phases of battery discharge during acceler-
ation or cruising followed by charging phases during
regenerative braking. Similar situations can occur when
air base batteries or regenerative fuel cells are used in
combination with photovoltaic cells during occasionally
cloudy days. Ideally, all components of the bifunctional
electrode should withstand the conditions during charge
and discharge. Therefore, a single-layered electrode
structure with a true bifunctional catalyst and a stable
support structure should be the most efficient and cost-
effective design for an electrically rechargeable metal–air
battery or regenerative fuel cell.
Catalyst Systems
Bifunctional oxygen/air electrodes are used in electric-
ally rechargeable metal–air batteries. They are also
relevant in the so-called regenerative fuel cells. Owing to
Current collector
Optional gas diffusion layer
Porous, hydrophobic
layer
Porous, hydrophilic layer
Oxygen evolution
their high conductivity caused by the high mobility of
protons in aqueous solution, alkaline or acid electrolytes
are preferred to neutral electrolytes.
Many synthesis routes have been investigated for the
preparation of bifunctional catalyst materials. The
methods include precipitation from aqueous and non-
aqueous solvents, sol–gel processes, freeze drying,
chemical spray pyrolysis, cathodic sputtering, and elec-
trochemical deposition. It is beyond the scope of this
article to discuss the synthesis methods in detail.
Alkaline Systems
Most electrochemical systems using bifunctional oxy-
gen/air electrodes are using an alkaline electrolyte since
evolution and consumption of oxygen can be carried out
at nonnoble metal catalysts. Furthermore, the exchange
current densities of oxygen evolution/reduction are
higher than in acidic electrolytes. A wide variety of metal
oxides have been investigated for use in electrically re-
chargeable alkaline fuel cells for application as catalyst
and catalyst support.
However, in alkaline systems the air electrodes are
sensitive to carbon dioxide, which reacts with the OH
ions to form carbonates. Bifunctional oxygen electrodes
operated in a pure oxygen environment failed in the
oxygen evolution (anodic) mode after B2500 h. Whereas
oxygen air electrodes operated in carbon dioxide con-
taining gas failed even at a carbon dioxide content below
1000 ppm after B250 h during oxygen reduction (cath-
odic) mode. The carbon dioxide sensitivity of oxygen gas
diffusion electrodes in cathodic mode can be influenced
by the choice of the catalyst support material.
Noble metal containing catalysts
In alkaline systems, oxygen reduction and evolution can
take place on catalysts involving noble metals such as the
 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes
platinum group metals. Transition metal oxides or mixed
transition metal oxides have also been successfully used.
CoIrO3 and NiIrO3 have shown reasonable activity
for oxygen evolution in a high-surface area form. Fur-
thermore, high-surface area rhodium dioxide (RhO2) and
iridium dioxide show high activity for oxygen evolution.
Good activity and stability has also been observed for
lead palladium oxide (PbPdO2).
Nonnoble metal containing catalysts
Oxygen reduction in alkaline systems does not neces-
sarily require the use of pure noble metal catalysts.
Transition metal oxides or mixed transition metal oxides
have also been successfully used. Charge–discharge effi-
ciencies above 60%, current densities of 200 mA cm2,
and power densities of 1 W cm2 have been claimed.
However, only a few materials can catalyze oxygen
evolution and reduction simultaneously. The perform-
ance of transition metal oxides and mixed transition
metal oxides is often limited by their low bulk con-
ductivity and limitations in the active surface area, which
can be achieved by chemical synthesis methods. Never-
theless, properties such as electronic conductivity or ion
mobility in the mixed transition metal oxide crystals can
be tailored by their composition.
Three types of transition metal oxide catalysts are
predominant.
• Catalysts with the perovskite structure;
• Catalysts with the pyrochlore structure; and
• Catalysts with the spinel structure.
Perovskite catalysts
Perovskite-type oxides have the general formula ABO3.
Their properties can vary over a wide range depending
on the nature of A and B. The cubic perovskite crystal
lattice is a rather rugged host for a variety of mixed
transition metal oxides. The structure can even adapt
distortions from the basic cubic symmetry giving rise to
interesting properties. Substituted perovskites can gen-
erally be described by the formula A1xA0 xB1yB0 yO3.
A wide selection of materials using lanthanum and
neodymium in position A, strontium, barium, calcium in
position A0 , and nickel, cobalt, manganese, and ruthe-
nium in position B have been synthesized and shaped
into electrodes. Thermal decomposition of nitrates,
acetates, oxalates, or carbonates was carried out initially
followed by a solid state reaction at temperatures be-
tween 600 and 900 1C. The electrodes were made from
two layers. In the oxygen reduction layer a mixture of
catalyst and polytetrafluoro ethylene (PTFE) (25 wt%)
was roll bonded onto an expanded nickel metal sheet.
The oxygen evolution layer consisted of nickel powder
bonded with 5% PTFE and was placed on the opposite
side of the expanded metal. The oxygen reduction layer
363
was backed by an additional PTFE film. In these in-
vestigations La0.5Sr0.5CoO3 and lanthanum nickel oxide
(LaNiO3) showed good initial performance. However,
their performance rapidly deteriorated upon cycling.
Admixture of graphite to the oxygen reduction layer
improved the electrical and cycling performance dra-
matically. More than 280 cycles were carried out.
Lanthanum nickel oxide was investigated by several
groups as a bifunctional catalyst. It was found to be
reasonably stable. However, typical signals for the for-
mation of the NiOOH/Ni(OH)2 redox couple are evi-
dent even after short cycling. The performance of the
material is largely dependent on the preparation method.
Particularly promising results have been achieved when
using the so-called malic acid method.
A series of compounds of the type LaFexNi1xO3 has
been investigated as model compounds. Good perform-
ance and stability in long-term operation were observed
for oxygen generation. However, perovskite materials are
not very active for oxygen reduction. Degradation effects
become prominent after a short time under cathodic
operation. Nevertheless, reasonable performance for
peroxide decomposition has been observed.
The activity of transition metal oxide catalysts can be
correlated with the ability of the cations to adopt dif-
ferent valency states, particularly when they form redox
couples at the potential of oxygen reduction/evolution.
From early studies based on lanthanum cobalt
(LaCoO3), it was found that the conductivity of the
material could be modified by doping A and B positions
of the material while maintaining the perovskite struc-
ture. A variety of substituted materials having the per-
ovskite structure was investigated in low-area gas
diffusion electrodes to evaluate their stability with re-
spect to reducing conditions.
Perovskite-type oxides of the composition La1xCaxMO3
were synthesized successfully by the amorphous citrate
method. Gas diffusion electrodes were prepared by mixing
25–70 wt% catalyst (3.5–9 mg cm2) with carbon black.
Carbon black of different surface area was used for the
catalytic and the gas supply layer. Here, PTFE (15–25 wt%)
was used as a binder. Teflonized powder mixtures were
preformed from suspension, filtered, and heated at up to
280 1C. After pulverization, the powders were pressed on
opposite sides of a nickel screen and sintered at 370 1C. It
was found that a material of composition La0.6Ca0.4CoO3
shows good catalytic activity and cycling stability. However,
La0.8Ca0.2MnO3 almost showed no improvement with re-
spect to the pure carbon black.
Bifunctional air electrodes made using La0.6Ca0.4CoO3
catalyst were operated for more than 400 cycles in elec-
trically rechargeable zinc–air batteries. The surface area of
the La0.6Ca0.4CoO3 catalyst material could be optimized
by the synthesis conditions and subsequent thermal
treatment. Well-crystallized materials were found to favor
364 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes

the oxygen reduction whereas less crystalline materials are
beneficial for oxygen evolution. A material prepared by
calcination at 650 1C and subsequent quenching at room
temperature showed the most promising performance in a
bifunctional electrode. The chemical stability of these
materials has been investigated by X-ray diffraction and
X-ray spectroscopic methods. The influence of vacuum
annealing and subsequent reoxidation on the performance
of La0.6Ca0.4CoO3x-type catalysts showed different op-
tima for the oxygen reduction reaction and the oxygen
evolution reaction. Reoxidation at temperatures of 200 1C
resulted in the maximum activity for oxygen reduction
whereas the optimum annealing temperature for oxygen
generation was found to be 150 1C.
Materials with substitution on the A and B positions
synthesized through the amorphous citrate method led to
catalysts of the general composition La1–xA0 xCo1–yB0 yO3
(A0 ¼ Ca; B0 ¼ Mn, Fe, Co, Ni, Cu). When used in PTFE-
bonded gas diffusion electrodes one finds that except for
the Cu-substituted catalyst, which had significantly lower
surface area, all catalysts performed almost identical
upon oxygen reduction (Figure 6). However, the Fe-
substituted material performed best upon oxygen evo-
lution. Cycling experiments using a La0.6Ca0.4Co0.8-
Fe0.2O3 catalyzed electrode showed reasonable stability
over B240 cycles (Figure 7). The activity with respect to
700
Cu Mn Ni
650
600
B′ = Fe
550
500
E vs Hg/HgO (mV)
oxygen reduction deteriorated while the oxygen evo-
lution activity stayed almost constant. The activity of the
materials with respect to oxygen evolution could be
correlated linearly with their ability to decompose per-
oxide (HO2  ).
Substituting iron for cobalt leads to catalysts of the
general composition La1–xA0 xFe1–yB0 yO3 (A0 ¼ Sr, Ca, Ba,
La; B0 ¼ Mn). Materials prepared by the mallate method
show only slight voltage differences in the oxygen re-
duction reaction. To sustain oxygen evolution strontium
or calcium must be present in the A0 position with
strontium substitution showing better performance at
current densities above 100 mA cm2, whereas the cal-
cium-substituted material showed slightly better per-
formance upon oxygen reduction (Figure 8).
A strong composition dependence on catalyst per-
formance resulting in steep maxima was observed for
both A- and B-site doping. In the case of calcium
1.0
Oxygen evolution
E vs Hg/HgO (V)
0.5
0
–0.5
Oxygen reduction
–1.0
0 40 80 120 160 200 240
Cycle number (times)
Figure 7 Test of gas-diffusion electrode loaded with 70 wt%
La0.6Ca0.4Co0.8Fe0.2O3 for the cycles of oxygen reduction
Co (10 min) and evolution (10 min) at 100 mA cm2 (25 1C in
7 mol L1 KOH). Reproduced with permission from Shimizu Y,
Matsuda H, Miura N, and Yamazoe N (1992) Chemistry Letters
1033–1036; Figure 3.

450 (b)
800 (Oxygen evolution)
0
(a) 600
E vs Hg/HgO (mV)

–50 Ca
400
Sr
Ba
–100 B′ = Mn 200 La

Fe Ni 0
–150
(Oxygen reduction)
Co
Cu –200
–200
0 100 200 300
–400
I mA cm−2 0 100 200 300 400
I (mA cm–2)
Figure 6 Polarization curves of gas diffusion electrodes loaded
with 25 wt% La0.6Ca0.4Co0.8 B0 0.2O3 (B0 ¼ Mn, Fe, Co, Ni, Cu) in Figure 8 Cathodic and anodic polarization curves of the
7 mol L1 KOH at 25 1C. (a) Cathodic polarization (O2 reduction). oxygen electrodes loaded with La0.6A0.4Fe0.8Mn0.2O3 (A0 ¼ Ca,
(b) Anodic polarization (O2 evolution). Reproduced with Sr, Ba, La) in 7 mol L1 KOH at 25 1C. Reproduced from Shimizu
permission from Shimizu Y, Matsuda H, Miura N, and Yamazoe N Y, Nemoto A, Hyodo T, Miura N, and Yamazue N (1993) Denki
(1992) Chemistry Letters 1033–1036; Figure 1. Kagaku 12: 1458–1460; Figure 2.
 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes 365

+ 500 100 −600 150

– Ic at –300 mV
Ic at –300 mV vs Hg/HgO
vs Hg/HgO
−500
+ 400 80

Specific surface area (m2 g–1)

–

−400
Ic or Ia (mA cm−2)

100

I (mA cm–2)
+ 300 60

Ia (mA cm−2)
–
−300

+ 200 40

c
–
−200 50
Ia at + 600 mV
Ia at + 700 mV
+ 100 20
– −100

0 0 0 0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x in La1–xCaxFe0.8 Mn0.2O3 y in La 0.8 Sr0.2 Fe1−y MnyO3

Figure 9 Cathodic and anodic performances and specific Figure 10 The current densities of the electrodes loaded with
surface area of La1xCaxFe0.8Mn0.2O3 as a function of x. La0.8Sr0.2Fe1yMnyO3 as a function of y (7 mol L1 KOH, 25 1C,
Reproduced from Shimizu Y, Nemoto A, Hyodo T, Miura N, and Air flow). Ic: cathodic performance (oxygen reduction). Ia: anodic
Yamazue N (1993) Denki Kagaku 12: 1458–1460; Figure 3. performance (oxygen evolution). Reproduced from Shimizu Y,
Nemoto A, Hyodo T, Miura N, and Yamazue N (1993) Denki
Kagaku 12: 1458–1460; Figure 1.
substitution, a maximum in anodic and cathodic current
was observed at 40% Ca substitution. This maximum
correlated with the geometric surface area of the ma- Pb2Ir2O7–y and PbBiRu2O7y . However, oxides with the
terials as shown in Figure 9. pyrochlore structure show nonnegligible solubility in
A different composition dependence was found for the strongly alkaline electrolytes. Stabilization has been
system La0.8Sr0.2Fe1–yMnyO3 (Figure 10). The anodic achieved by incorporating the catalysts in a polymer
oxygen evolution performance dependence on B-site sub- electrolyte matrix. Catalysts of composition Bi2Ru2O7
stitution showed a steep maximum at 20% manganese showed limited performance for oxygen reduction;
substitution for iron. At higher levels the performance however, high activity in peroxide decomposition and
dropped below the one observed for the unsubstituted improved stability with respect to dissolution under
material, whereas the cathodic performance started to form conditions of oxygen evolution was observed.
a plateau from 40% substitution of manganese for iron. On the contrary, excellent performance for oxygen
Activity and particularly stability of bifunctional reduction/evolution has been observed for lead ruthe-
oxygen/air electrodes involving La0.6Ca0.4CoO3 catalysts nate pyrochlores when investigated in gas-fed electrodes.
are also influenced significantly by the carbon used to Substitution of iridium into the materials proved to be
support the catalyst. This was shown by comparison of beneficial for the oxygen reduction reaction. Further-
electrodes made from high-surface area Ketjen black and more, ruthenium corrosion has been suppressed by cov-
graphitized Vulcan XC 72. Electrodes made from Ketjen ering the catalyst with layers of anion conducting
black showed improved performance upon oxygen re- polymers.
duction. However, driving the electrode into oxygen Other studies involving iridium oxides, some of which
evolution led to a faster deterioration as compared to had the pyrochlore structure, revealed moderate-to-high
electrodes prepared with graphitized Vulcan XC 72. activities with respect to the oxygen evolution and lower
activity for the oxygen reduction. Materials of com-
Pyrochlore catalysts position Pb2(PbxIr2x)O7d and Nd3IrO7 showed the
Pyrochlore-type oxides have the general formula highest activities for both reactions. High-preparation
A2B2O7. Their properties can be influenced by the choice temperature increased the stability of the materials
of the A and B cations as well as by further doping of the whereas low-preparation temperature led to higher ac-
A and B positions. Pyrochlore materials of the general tivity. Addition of tantalum increased both activity and
composition A2B2xAxO7y (A ¼ Pb/Bi, B ¼ Ru/Ir) were stability.
active as bifunctional catalysts in alkaline electrolytes.
A strong dependence of catalytic activity on Spinel catalysts
the composition has been observed. As a bifunctional Besides oxides with the pyrochlore or perovskite struc-
catalyst, the optimum compositions were found to be ture, metal oxides having the spinel structure are also
366 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes
100
O2
Current density (mA cm−2)
Air
Discharge
10
1
−0.4 −0.3 −0.2 −0.1
Electrode potential (V vs Hg/HgO)
Figure 11 Polarization curves for a 25 cm2 prototype air
electrode made entirely from carbon. The electrode was
fabricated by impregnating a mixture of graphite and a mixed
catalyst made from nickel oxide and spinel-type Co3O4 into a
carbon cloth. The electrode was equipped with a Teflon film-
backing layer. Reproduced from Ross P (1985) Extended
Abstracts of the Seventh Battery and Electrochemical
Contractors Meeting, DOE Report CONF–851146-Absts 323l;
Figure 6.
showing good performance in oxygen evolution and re-
duction. Oxygen reduction at nickel cobalt oxides has
been shown with good initial activity, however, de-
terioration is observed with time. On the contrary, this
material is quite stable under the conditions of oxygen
evolution.
Pure or lithium-doped cobalt oxide (Co3O4), which
also has the spinel structure, is considered a reasonable
bifunctional oxygen catalyst (Figure 11). A bifunctional
air electrode structure using nickel oxide or spinel-type
cobalt oxide (Co3O4) as an inexpensive transition metal
oxide catalyst has been proposed to minimize cost and
weight of electrically rechargeable zinc–air batteries at
the cost of some slight performance sacrifices. The
electrode was made from graphite cloth formed from
woven graphite fibers, which were coated by graphite
powder of a specific surface area of 25 m2 g1, which was
prepared by heat treatment of an inexpensive precursor
material. The catalysts were deposited on the graphite
and the resulting powder was coated on the graphite
cloth by Teflon bonding and sintering at 350 1C. A sec-
ond Teflon film was roll bonded to the electrode back
for wet proofing. Operation at a potential versus
Hg/HgO of  130 mV at 10 mA cm2 during discharge
and þ 620 mV at 5 mA cm2 during charge could be
achieved. Electrode startup under low-current density
charge for 100 h followed by several conditioning cycles
proved to be essential. The high-current density per-
formance of the electrode was hindered by oxygen
transport limitations caused by the PTFE film. The
electrodes were operated in oxygen and in air.
Cobaltite spinels of type MxCO3xO4 have been
300 K synthesized to enhance the oxygen reduction reaction
12 mol L−1 KOH (M ¼ Co, Ni, and Mn) and the oxygen evolution reaction
(M ¼ Co, Li, Ni, Cu, and Mn). copper cobalt oxide
proved to be a better electrocatalyst than pure cobalt
Charge
oxide manganese. Further doping with Mn decreases the
bifunctional activity; however, it increased the geometric
surface area of the catalyst powder.
Electrically rechargeable metal–air systems
In metal–air cells, the negative electrode reaction consists
in the oxidation of a metal having low electronegativity
0 0.5 0.6 0.7
such as aluminum, magnesium, cadmium, zinc, iron, or
recently, metal hydrides and lithium. Metal–air systems
are typically operated as primary cells. They are used
when high specific energy is required by the application.
Zinc–air cells represent the state-of-the-art for high-
energy microbatteries used, e.g., in hearing aids. Mag-
nesium–air cells and batteries are frequently used as the
so-called reserve batteries in marine applications or as a
power source for emergency lighting in life vests. Details
on the different primary metal–air systems can be found
in the monograph by D. Linden and coworkers found in
the Further Reading section.
Owing to the redox potential of the metal electrode,
zinc, iron, cadmium, or hydrogen can be deposited from
an aqueous electrolyte with reasonable current efficiency.
Electrically rechargeable metal–air batteries are attract-
ive from a specific energy as well as from cost point of
view. In the case of negative electrodes made from iron,
metal hydrides, or zinc, also environmental aspects are in
favor of metal–air systems.
Iron electrodes have been used for a long time as the
negative electrode in the so-called iron–nickel batteries,
which were used in heavy duty applications such as
electric traction or powering of headlamps in the mining
industry. Electrically rechargeable iron–air batteries
having a capacity of 30 kWh and reaching a specific
energy of B80 Wh kg1 have been demonstrated. How-
ever, iron–air systems are limited with respect to specific
power and coulombic efficiency due to the kinetics and
the high amount of parasitic hydrogen evolution caused
by the iron electrode.
Metal-hydride–air systems using AB5-type hydrides
have been evaluated as well. A specific energy of
B100 Wh kg1 has been claimed. Designs involving
bifunctional air electrodes as well as designs using sep-
arate electrodes for oxygen evolution and oxygen re-
duction in the same cell have been proposed.
Owing to cost, environmental, and efficiency reasons,
most work has been devoted to zinc–air systems. Initial
development of zinc–air batteries had already started at
the end of the nineteenth century. In the beginning, the
so-called mechanically rechargeable batteries were de-
veloped where spent anodes and electrolytes were used
 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes
for reprocessing of the zinc outside the battery. Designs
using a zinc slurry electrolyte as well as designs requiring
the dismantling of the battery have been proposed.
Electrically rechargeable zinc–air batteries have also
been described in cylindrical and prismatic designs.
Significant work has been done at the Paul Scherrer In-
stitute at the beginning of the twenty-first century.
Electrically rechargeable zinc–air cells having a capacity
of 20 Ah have been demonstrated. Bifunctional air elec-
trodes using a La0.6Ca0.4COO3 catalyst were used. The
cells had a specific energy of B100 Wh kg1. A total of
150 cycles have been achieved. It has been found that the
carbon used as a catalyst, support materials that had
significant influence on the performance. Although high-
surface area carbon blacks resulted in electrodes capable
of sustaining a current density of 300 mA cm1 at a
polarization of B200 mV, these electrodes were signi-
ficantly inferior in durability when compared to grap-
hitized carbon support materials. Besides the bifunctional
air electrode, the cycle life is also limited by the
zinc electrode due to the tendency of dendritic metal
deposition.
Despite the attractive specific energy of metal–air
systems on a cell level, one has to take into account that
auxiliary equipment will be required for air supply and
air purification. Since most of the metal–air systems use
potassium hydroxide as an electrolyte, removal of carbon
dioxide from the air requires maintainance of electrolyte
conductivity and to prevent clogging of the air electrode
pores by precipitation of carbonate. Besides carbon di-
oxide removal, attention is paid to water management.
Depending on the relative humidity of the air fed to the
battery, the cells can either pick up or lose water to the
environment.
In summary, metal–air batteries, particularly zinc–air
batteries are well established as primary cells. However,
no major technical application for electrically recharge-
able zinc–air batteries has been established.
Acid Electrolytes
Many studies have been published concerning oxygen
electrocatalysts in acid media, particularly for use in
regenerative fuel cells. Most studies concluded that Pt/Ir
or Pt/IrO2 is in a molar ratio of 1:1. Also mixed oxides
from rhodium and ruthenium in a molar ratio of 1:1 or
iridium and rhodium in a molar ratio of 1:2 were
promising electrocatalysts.
A study using combinatorial chemistry methods was
carried out to identify an optimum bifunctional catalyst
composition from a combination of platinum, iridium,
rhodium, ruthenium, and osmium. A library of model
electrodes of different compositions was prepared by
dispersing aqueous solutions of the respective solutions
on a Teflon-coated Toray carbon paper using a plotter.
367
The salts were reduced by sodium borohydride. Elec-
trochemical experiments were carried out at BpH 3.
The electrodes were simultaneously conditioned by an
initial polarization to a potential of 1600 mV (vs RHE) in
an electrolyte containing a fluorescence indicator. The
oxygen evolution activity was measured by stepping the
potential from 1200 mV (vs RHE) in steps of 50 mV until
visible fluorescence was observed. Oxygen reduction
screening was carried out in using a different electrolyte
and fluorescence indicator. After conditioning the library
at þ 600 mV (vs RHE), the potential was stepped down
from þ 900 mV (vs RHE) in steps of 50 mV until visible
fluorescence was observed. High activity for oxygen re-
duction and evolution as well as stability toward dis-
solution was found in rather narrow range of the
platinum and ruthenium-rich section of the Pt/Ru/Ir
catalyst library. A material with composition Pt4.5Ru4Ir0.5
was found to be particularly active. This was also proven
by a test of bulk catalyst materials. An activity plot of this
material in comparison to other catalysts is shown in
Figure 12. It is evident that the Pt4.5Ru4Ir0.5 material
outperforms the commonly used platinum for oxygen
reduction and platinum/iridium materials for oxygen
evolution. The material also showed corrosion resistance
under the test conditions used. Similar work has been
carried out using reduced titanium oxides (e.g., Ebonex,
titanium oxide (Ti4O7)) or niobium-doped titanium
oxide as catalyst support material.
0.8 Pt
Pt4.5Ru4lr0.5
0.7 Pt4Ru4lr1
PtRu
Ptlr
0.6
Current density (mA cm−2)
0.5
0.4
0.3
0.2
0.1
0
400 600 800 1000 1200 1400 1600 1800
Potential versus RHE (mV)
Figure 12 Polarization curves of bifunctional oxygen electrode
in electrolysis (oxygen evolution) and fuel cell (oxygen reduction)
modes for different catalysts in 0.5 mol L1 H2SO4. RHE,
reversible hydrogen electrode. Reproduced with permission from
Chen G, Delafuente DA, Sarangapani S, and Mallouk TE (2001)
Catalysis Today 67: 341–355; Figure 4.
368 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes
Unitized regenerative proton-exchange
membrane fuel cells
A unitized regenerative fuel cell (URFC) is an energy
storage system based on hydrogen and oxygen, which can
be operated similarly to a secondary battery. The system
electrolyzes water in charge mode, hydrogen and when
appropriate, oxygen are subsequently stored. In discharge
mode, the system is operated as a fuel cell using up the
stored reactants. Significant advantages of the specific
energy can be expected from the use of an URFC
(400–1000 Wh kg1) in comparison to other secondary
batteries (220–250 Wh kg1 for advance lithium–polymer
batteries). The URFC also offers the advantage of in-
dependent sizing of the electrochemical reactor which
provides power and the energy/reactant storage providing
the capacity. This advantage is particularly striking in ap-
plications where long charge–discharge cycles at a given
power are required.
The URFC can be operated in two distinctly different
modes as shown in Figure 13.
• same
Oxygen evolution and reduction takes place at the
electrode therefore, the use of a true bifunc-
tional electrode is mandatory.
• Anodic reactions such as oxygen evolution upon
charge and hydrogen reduction upon discharge take
place at one electrode, whereas cathodic reactions
such as hydrogen evolution upon charge and oxygen
reduction upon discharge, take place at a different
electrode. This method of operation allows catalyst
optimization for anodic and cathodic reactions, re-
spectively, which results in considerably lower losses.
However, flushing of the electrode compartments with
inert gas (e.g., nitrogen) is required when changing
from charge to discharge mode.
2H+ + 2e → H2 2H2O → O2 + 4H+ + 4e
Cathode A Anode
SPE membrane
composition
Pt-Ir catalytic
Pt catalytic
composition
Anode Cathode
(b) O2 + 4e + 4H → 2H2O
(a) H2 → 2H+ + 2e O2 + 4e + 4H+ → 2H2O
Figure 13 Modes of operation for a unitized regenerative fuel cell (URFC). (a) Oxygen evolution and reduction taking place at the
same electrode. (b) Anodic and cathodic reactions taking place at the same electrode. Reproduced with permission from Tsypkin MA,
Lyutikova EK, Fateev VN, and Rusanov VD (2000) Russian Journal of Electrochemistry 36: 545–548; Figure 1.
The membranes used in URFC are similar to the ones
used in ordinary proton-exchange membrane fuel cells
(PEMFC). Mainly polymer perfluorinated sulfonic acid
materials such as Nafion are used. Methods to apply
catalyst layers are also similar to the ones used in PEMFC.
Differences in design arise from the gas diffusion
layers and the catalyst support materials used. In
PEMFC, highly hydrophobized GDL are used in order
to provide free oxygen access to the cathode. In the case
of a URFC also water access to the catalyst layer must be
ensured during electrolysis mode. Therefore, either a
GDL providing an access path for liquid water to the
catalyst layer, or other means of transporting water to the
reaction site must be used.
Titanium-based materials have been successfully
tested as corrosion-stable catalyst support and current
collectors. Another successful path to minimize the cor-
rosion of the electrode structure involved the prevention
of water getting into contact with the carbon materials in
the GDL. A hydrophobized graphite paper has been used
as conductive fiber substrate on which a mixture of
titanium powder mixed with iridium dioxide as an oxy-
gen evolution catalyst has been applied in a microporous
layer. Fuel cell operation at 600 mA cm1 and 600 mV
has been shown. Oxygen evolution took place at
B1500 mV at the same current density. Stability over
several hours has been demonstrated.
Carbon-supported catalysts have been introduced to
minimize noble metal loading in PEMFC. However,
high-surface area carbon materials are not stable during
oxygen evolution. Electronically conducting carbides or
borides and selected titanium oxide have been investi-
gated instead.
Also in acidic electrolytes, the most prominent losses
are connected with oxygen electrochemistry.
2H+ + 2e → H2 2H2O → O2 + 4H+ + 4e
Cathode B Anode
SPE membrane
Pt-Ir catalytic
composition
composition
Pt catalytic
Cathode Anode
+
H2 → 2H+ + 2e
 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes
Unitized regenerative fuel cells using a single
bifunctional oxygen electrode
Electrodes using an internal dual layer have been used in
URFC. However, these electrodes are rather bulky and
increase the thickness of the gas diffusion path leading to
reduced performance.
To optimize transport properties, thin-layer elec-
trodes were investigated. The electrodes were made by
decal transfer methods originally developed for high-
power proton-exchange membrane fuel cells (PEMFC).
In this method, an ink consisting of catalyst powder,
ionomer solution (e.g., Nafion in the Naþ form) and a
high boiling solvent (e.g., glycerol) is spread on a PTFE
release blank. After curing the catalyst layer at elevated
temperature, it is hot pressed to the electrolyte mem-
brane (e.g., Nafion 115). Later, the PTFE film is removed
from the catalyst layer. The membrane electrode as-
sembly (MEA) was rehydrated by boiling in 0.5 mol L1
sulfuric acid. The catalyst loading was 0.4 mg cm2 Pt-
black at the hydrogen electrode and 0.4 mg cm2 of a
50:50 mixture of Pt-black and finely dispersed iridium
dioxide at the oxygen electrode. The thickness of the
catalyst layer was about 5 mm. Carbon paper containing a
PTFE content of B30 wt% was used as an electrode
backing.
The cell was operated at 80 1C at a pressure of 0.3 bar.
A current density of 400 mA cm2 at 0.6 V could be
achieved in fuel cell mode. In electrolysis mode, 1.71 V
2.2
2.0
1.8
1.6
1.4
Cell volt (V)
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0 200
Current density (mA
Figure 14 Performance of a URFC using Nafion 115 electrolyte after four cycles (numbered in the digaram) fuel cell mode: T ¼ 80 1C,
pO2 ¼ 0.3 MPa, pH2 ¼ 0.3 MPa electrolysis mode: T ¼ 80 1C, ambient pressure. Reproduced with permission from Shao Z, Yi B, and Han
M (1999) Journal of Power Sources 79: 82–85; Figure 8.
369
were required to obtain a current density of
400 mA cm2. The cell showed performance improve-
ment within the first cycles (Figure 14).
Typically, mixtures of platinum and iridium oxide are
used in URFC. An attempt to use supported catalysts has
been reported. Mixed metal catalysts containing plat-
inum, iridium, ruthenium, osmium, and rhodium were
deposited on conductive oxide supports. Titanium oxide
(Ti4O7) containing-materials as well as Ti0.9Nb0.2O2 have
been used as active supports. The highest catalysts sta-
bility and activity was found at ternary compositions near
Pt4Ru4Ir1 where X-ray near-edge spectroscopy showed
significant interaction between catalyst and support. The
Ti4O7-containing support materials showed significant
degradation under oxygen evolution whereas
Ti0.9Nb0.1O2 as an oxide with a nondefective oxygen
lattice was apparently stable.
Unitized regenerative fuel cells using different
electrodes for oxygen evolution/consumption
Work on URFC using different electrodes for oxygen
evolution and reduction has been reported. Different
catalysts were studied for both electrodes. Platinum was
found to be a reasonably good catalyst for the cathodic
reactions, i.e., oxygen reduction and hydrogen evolution.
A single catalyst for the anodic reactions, i.e., hydrogen
consumption and oxygen evolution is more difficult to
2 1
3
4
Electrolyzer
Fuel cell
1
4
3
2
400 600 800 1000
cm–2)
370 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes
find. Again, hydrogen consumption is best performed at a
platinum catalyst. However, oxygen evolution is better
performed at an iridium electrode. Since the hydrogen
oxidation is a rather reversible reaction even at iridium, a
mixture of platinum and iridium is considered a good
‘oxidation’ electrode.
The wetting behavior of the electrodes is of consider-
able importance for the water management of the cell. In a
comparative study, it was found that a URFC using dif-
ferent electrodes for oxygen evolution and consumption
can possibly be operated with little degradation.
Conclusions
Bifunctional oxygen/air electrodes are a challenging re-
search topic. In alkaline electrolyte, potentially low-cost
transition metal oxides are good candidates for bifunc-
tional catalysis. Oxides having the pyrochlore structure
containing ruthenium or iridium show signs of dis-
solution at elevated potentials even in alkaline electro-
lytes. Most transition metal oxides, particularly the ones
having the spinel structure show higher activity with
respect to oxygen evolution. Nevertheless, graphite
shows a reasonable activity with respect to two electron
oxygen reduction. Binary oxides such as nickel oxide or
spinel-type Co3O4 show the highest long-term stability.
In connection with transition metal oxides, peroxide re-
duction is greatly enhanced. In acid electrolytes, the use
of noble metals is inevitable.
Unitized regenerative fuel cells having an acid polymer
electrolyte can be operated under conservation of anodic
and cathodic processes at the same electrodes, respectively,
thus reducing losses otherwise seen in more simpler sys-
tems where oxygen evolution and consumption takes place
at the same electrode. Recent work showed the feasibility
of metal oxide-supported catalysts for URFCs.
Nomenclature
Symbols and Units
E potential
E0 standard electrode potential
Ia anodic current density
Ic cathodic current density
T temperature
Abbreviations and Acronyms
GDL gas diffusion layer
MEA membrane electrode assembly
ORR oxygen reduction reaction
PEMFC proton-exchange membrane fuel cell
PTFE polytetrafluoroethylene
RHE reversible hydrogen electrode
SHE standard hydrogen electrode
SPE solid polymer electrolyte
URFC unitized regenerative fuel cell
See also: Batteries: Charge–Discharge Curves; Self-
Discharge; Batteries and Fuel Cells: Efficiency;
Chemistry, Electrochemistry, and Electrochemical
Applications: Oxygen; Zinc; Electrodes: Porous
Electrodes; Primary Batteries – Aqueous Systems:
Zinc–Air; Secondary Batteries – Metal-Air Systems:
Iron–Air (Secondary and Primary); Zinc–Air: Electrical
Recharge; Zinc–Air: Hydraulic Recharge; Secondary
Batteries – Zinc Systems: Zinc Electrodes: Overview.
Further Reading
Anastasijevic NA, Vesovic V, and Adzic RR (1987) Determination of the
kinetic parameters of the oxygen reduction reaction using the
rotating ring-disk electrode: Part I. Theory. Journal of
Electroanalytical Chemistry 229: 305--316.
Carbonino RE, Fierro C, Tryk D, Scherson D, and Yeager E (1988)
Perovskite-type oxides: Oxygen electrocatalysis and bulk structure.
Journal of Power Sources 22: 387--398.
Carlsson L and Öjefors L (1980) Bifunctional air electrode for metal-air
batteries. Journal of the Electrochemistry Society 127: 525.
Chen G, Delafuente DA, Sarangapani S, and Mallouk TE (2001)
Combinatorial discovery of bifunctional oxygen reduction - water
oxidation electrocatalysts for regenerative fuel cells. Catalysis Today
67: 341--355.
Conway BE (1992) A direction of study of electrocatalysis in anodic o2
evolution through characterization of chemisorption behavior of
intermediates. In: Murphy OM, Srinivasan S, and Conway BE (eds.)
Electrochemistry in Transition, pp. 161--177. New York: Plenum
Press.
Damjanovic A (1969) Mechanistic analysis of oxygen electrode
reactions. In: Conway BE and Bockris JO’M (eds.) Modern Aspects
of Electrochemistry, vol. 5, pp. 369--483. New York: Plenum Press.
Damjanovic A (1992) Progress in the studies of oxygen reduction during
the last thirty years. In: Murphy OM, Srinivasan S, and Conway BE
(eds.) Electrochemistry in Transition, pp. 107--126. New York:
Plenum Press.
Damjanovic A (1966) Electrode kinetics of oxygen reduction on oxide-
free platinum electrodes. Electrochimica Acta 11: 791--814.
De Koninck M and Marsan B (2008) MnxCu1-xCo2O4 used as
bifunctional electrocatalyst in alkaline medium. Electrochimica Acta
53: 7012--7021.
De Koninck M, Poirer SC, and Marsan B (2006) Journal of the
Electrochemical Society 153: A2103--A2110.
De Koninck M, Poirer SC, and Marsan B (2007) Journal of the
Electrochemical Society 154: A381--A388.
Gamburtsev S, Zhang W, Velev OA, Srinivasan S, and Appleby AJ
(1998) Journal of the Applied Electrochemistry 28: 545.
Gattrell M and McDougall B (2003) Reaction mechanisms of the O2
reduction/evolution reaction. In: Vielstich W, Lamm A, and Gasteiger
HA (eds.) Handbook of Fuel Cells – Fundamentals, Technology and
Applications, vol. 2, pp. 443--464. New York: John Wiley & Sons.
Goodenough JB and Manoharan R (1991) New oxide catalysts for
oxygen electrodes in scherson. In: Tryk D, Daroux M, and Xing X
(eds.) Proceedings of the Workshop on Structural Effects in
Electrocatalysis and Oxygen Electrochemistry, Proceedings of the
Electrochemical Society, vol. 92–11, pp. 523--539. Pennington, NJ:
The Electrochemical Society.
Hoare J (1968) The Electrochemistry of Oxygen. New York: John Wiley
& Sons.
Horowitz HS, Longo JM, and Horowitz HH (1983) Oxygen
electrocatalysis on some oxide pyrochlores. Journal of the
Electrochemical Society 130: 1851--1859.
 Secondary Batteries – Metal-Air Systems | Bifunctional Oxygen Electrodes
Horowitz HS, Longo JM, and Horowitz HH (1983) Oxygen
electrocatalysis on some oxide pyrochlores. Journal of the
Electrochemical Society 130: 1851--1859.
Jörissen L (2006) Bifunctional oxygen/air electrodes. Journal of the
Power Sources 155: 23--32.
Kinoshita K (1992) Electrochemical Oxygen Technology. New York:
John Wiley & Sons.
Ledjeff K, Mahlendorf F, Peinecke V, and Heinzel A (1995) Development
of electrode/membrane units for the reversible solid polymer fuel cell
(RSPFC). Electrochimica Acta 40: 315.
Linden D and Reddy TB (eds.) (2001) Handbook of Batteries and Fuel
Cells. New York: McGraw-Hill.
Mitlitsky F, Myers B, and Weisber AH (1998) Energy & Fuels 12: 56.
Nikolova V, Iliev P, Petrov K, Zhecheva E, Stoyanova R, Valov I, and
Stoychev D (2008) Journal of Power Sources 185: 727--733.
Pettersson J, Ramsey B, and Harrison D (2006) Review of the latest
developments in electrodes for unitised regenerative polymer
electrolyte fuel cells. Journal of Power Sources 157: 28--34.
Prakash J, Tryk D, Alred W, and Yeager E (1992) Transition-metal oxide
electrocatalysts for O2 electrodes: The pyrochlores. In: Murphy OM,
Srinivasan S, and Conway BE (eds.) Electrochemistry in Transition,
pp. 93--106. New York: Plenum Press.
Ross P (1985) Extended Abstracts of the Seventh Battery and
Electrochemical Contractors Meeting, DOE Report CONF–851146-
Absts 323l.
Sakai H, Iwaki T, Ye Z, Noreus D, and Lindstöm O (1995) Air-metal
hydride battery construction and evaluation. Journal of the
Electrochemical Society 142: 4040.
371
Shao Z, Yi B, and Han M (1999) Bifunctional electrodes with a thin
catalyst layer for ‘unitized’ proton exchange membrane regenerative
fuel cell. Journal of Power Sources 79: 82--85.
Song S, Zhand H, Ma X, Shao ZG, Zhang Y, and Yi B (2006)
Bifunctional oxygen electrode with corrosion-resistive gas difusion
layer for unitized regenerative fuel cell. Elecrochemistry
Communications 8: 399--405.
Shimizu Y, Matsuda H, Miura N, and Yamazoe N (1992) Bi-functional
oxygen electrode using large surface area perovskite-type oxide
catalyst for rechargeable metal-air batteries. Chemistry Letters 21(6):
1033–1036.
Shimizu Y, Nemoto A, Hyodo T, Miura N, and Yamazue N (1993) Gas
diffusion-type oxygen electrode perovskite-type oxides for
rechargeable metal-air batteries. Denki Kagaku 12: 1458--1460.
Swette LL, laConti AB, and McCatty SA (1994) Proton-exchange
membrane regenerative fuel cells. Journal of Power Sources 47:
345.
Trasatti S (1990) In: Wendt H (ed.) Electrochemical Hydrogen
Technologies. Amsterdam: Elsevier.
Trasatti S (1991) Physical electrochemistry of ceramic oxides.
Electrochimica Acta 36(2): 225–241.
Tsypkin MA, Lyutikova EK, Fateev VN, and Rusanov VD (2000) Catalytic
layers in a reversible system comprising an electrolyzing cell and a
fuel cell based on solid polymer electrolyte. Russian Journal of
Electrochemistry 36: 545--548.
Yeager E (1984) Electrocatalysts for O2 reduction. Electrochimica Acta
29: 1527--1537.