DOI: 10.1002/cplu.

201402238 Minireviews

A Review of the Iron–Air Secondary Battery for Energy

Storage
R. D. McKerracher,[a] Carlos Ponce de Leon,*[a] R. G. A. Wills,[a] A. A. Shah,[b] and
Frank C. Walsh[a]

ChemPlusChem 2015, 80, 323 – 335 323  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Minireviews

Recent interest in the iron–air flow battery, known since the oxygen electrodes, low-cost iron electrodes able to decrease
1970s, has been driven by incentives to develop low-cost, envi- corrosion and hydrogen evolution, new cell designs using ad-
ronmentally friendly and robust rechargeable batteries. With ditive manufacturing technologies and mathematical models
a predicted open-circuit potential of 1.28 V, specific charge ca- to improve battery performance. This Minireview considers the
pacity of < 300 A h kg1 and reported efficiencies of 96, 40 and thermodynamics and kinetics aspects of the iron–air battery,
35 % for charge, voltage and energy, respectively, the iron–air the operational variables and cell components, thereby high-
system could be well suited for a range of applications, includ- lighting current challenges and assessing recent develop-
ing automotive. A number of challenges still need to be re- ments.
solved, including: efficient and moderate-cost bifunctional

Introduction
The history of iron–air batteries
Iron is an attractive metal for a rechargeable battery owing to
its low cost, ease of oxidation and multiple oxidation states, to-
gether with the ability to be electrodeposited from an aqueous
electrolyte. The iron–air cell can be considered as a substitute
for the iron–nickel oxide alkaline cell in which the nickel elec-
trode has been substituted by a bifunctional air-breathing elec-
trode. The schematic cell in Figure 1 shows the major process- Figure 1. The unit iron–air cell with an anion membrane in an alkaline elec-
trolyte, showing the main processes occurring in: a) charge and b) discharge
es taking place under discharge and charge conditions (which
cycles.
are considered in the “Electrode and cell reactions” section).
The open-circuit cell potential of
the iron–air battery is 1.28 V,
slightly lower than that of the
iron–nickel oxide cell of 1.41 V,
but substituting the nickel with
an air-breathing electrode can
increase the energy density by
up to 100 %. The negative iron
electrode coupled with the re-
duction and evolution of
oxygen at a catalyst-coated inert Figure 2. Timeline showing major developments during the history of iron–air batteries.
positive electrode form the re-
chargeable iron–air battery.
Figure 2 is a timeline highlighting major developments Earlier work on the iron–air battery was performed by NASA,
during the history of iron–air batteries. This concept was uti- in which a 5–20 A h charge capacity battery able to achieve
lised by the Swedish National Development Corporation to < 200 cycles with an energy density of 132–154 W h kg1 was
produce a 30 kW h battery for traction applications with an developed.[2, 3] The main advantage is that no metal dendrites
1 [1]
energy density of 80 W h kg at the 5 hour rate in the 1970s. are formed during charging. On the other hand, substantial hy-
The battery was timely for vehicle applications and required drogen evolution takes place. NASA highlighted additional
low-cost and readily available materials with minimum periph- problems such as self-discharge, the occurrence of degrading
eral equipment to suit the electric vehicle environment. iron oxidation reactions and the loss of water. Despite these
limitations, Westinghouse Electric Corporation[4, 5] reported the
sustainable performance and acceptable characteristics of
a state-of-the-art iron sintered electrode of 100 cm2 geometric
[a] Dr. R. D. McKerracher, Dr. C. Ponce de Leon, Dr. R. G. A. Wills,
area. They predicted that the battery could be manufactured
Prof. F. C. Walsh
Electrochemical Engineering Laboratory in modules of 400 cm2 electrode area with a specific energy of
Engineering Sciences, University of Southampton 140 W h kg1, able to achieve 1000 cycles at a manufacturing
Highfield, Southampton SO17 1BJ (United Kingdom) cost of US$30 (kg h)1.[6] Siemens, in Germany and Matsushita,
E-mail: [email protected]
in Japan were both active in research and development of the
[b] Dr. A. A. Shah
iron–air battery in the 1970s. Interest in this technology was
School of Engineering, University of Warwick
Coventry CV4 7AL (United Kingdom) mostly driven by the 1970s oil crisis and the possibility of elec-
Part of a Special Issue on “Metal–Air and Redox Flow Batteries“. To view tric vehicles.
the complete issue, visit: http://dx.doi.org/10.1002/cplu.v80.2

ChemPlusChem 2015, 80, 323 – 335 www.chempluschem.org 324  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Minireviews
Renewed interest has been shown in the iron–air battery changing the positive nickel electrode with a bifunctional air-
technology over the last few years owing to recent advances breathing electrode. This can increase its specific energy by up
in nanotechnology and the possibility of using effective nano- to 100 %. Iron is also a very low cost, environmentally friendly
structured electrode catalysts to realise a higher energy density resource (recycling being a well-established process), although
through the use of higher surface area Fe nanoparticles. The the strong alkaline electrolyte requires adequate protective
low cost and availability of iron and atmospheric oxygen pro- equipment and safety procedures.
vide additional drivers. Two examples of recent interest are the Table 1[10–12] compares the performance characteristics of
project at the University of Southern California on rechargeable three iron–air batteries against zinc–air and lithium-ion batter-
iron–air batteries[7] and the Euro-
pean FP7 funded project, NECO-
BAUT,[8] which seeks to test the Table 1. Specific energy and number of cycles to failure of iron–air batteries compared to alternative technolo-
feasibility of an iron–air battery gies.
for automotive applications. Battery Theoretical Theoretical Practical specific Number of Specific Cost Energy Ref.
cell energy energy cycles power efficiency
voltage density density density
The scope of this Review [V] [W h kg1] [W h kg1] [W kg1]

This Minireview considers previ- Fe–air 1.28 764 80 > 1000 20 NA[a] NA [10, 11]
ous work performed on the Fe–air 1.28 764 110 > 500 150 NA NA [10, 11]
Fe–air 1.28 764 50–75 2000 NA < $100 (kW h1) 50 % [12]
iron–air battery system and indi- (expected)
cates the latest developments in Zn–air 1.65 1080 200 NA 90 NA NA [12]
manufacturing and performance Li ion – – NA NA NA > $1000 (kW h1) 90 % [12]
of components in the modern Li–air – 13300 – NA NA NA NA [10]
Al–air 2.71 8100 300–400 NA NA NA NA [10]
iron–air battery. These include
the air-breathing oxygen elec- [a] NA = not applicable.
trode, the iron electrode, the
electrolyte composition, cell sep-
arators and electrochemical cell ies. The first iron–air battery listed in Table 1 is a cell developed
design. by the Swedish Development Corporation (referred to in the
literature as “Fe air 1”) that achieved acceptable performance
levels. Westinghouse managed to build a cell with greater
The attraction of iron–air batteries in energy storage
energy density and specific power at the cost of a reduced
The iron–air battery is attractive; unlike zinc in the zinc–air bat- cycle life. The final iron–air battery listed in Table 1 is the one
tery, iron is less prone to forming dendrites with repeated reported by Narayan et al.[12] The zinc–air battery is a commer-
charge–discharge cycles from aqueous electrolytes, which can cially available battery produced by “Electric Fuel” that is suita-
cause severe changes in the shape of the electrodes during ble for use in larger passenger vehicles owing to the high
prolonged cycling.[9] In comparison to the iron–nickel battery, energy density and it is the main rival to the iron–air battery
the weight of the battery can be decreased by one third on ex- for use in electric vehicles.[13]
The advantages of the iron–air battery include its moderate
energy density of 50–75 W h kg1 and a cost below US$100
Carlos Ponce de Leon is a Senior Lec- (kW h)1.[12] For this technology to be used commercially, how-
turer at the Energy Technology Group ever, the efficiency and cycle life need to be improved. The De-
of the Faculty of Engineering and the partment of Energy in the United States has suggested that
Environment at the University of a reliable battery should have up to 5000 charge–discharge
Southampton, UK. He holds a PhD in cycles and 80 % efficiency.[14] The current cycle life of the most
electrochemistry–electrochemical engi- advanced iron–air battery stands at 2000 cycles at an efficiency
neering and an MBA. Current research of around 50 %.[12]
includes electrochemical redox flow
batteries, such as the cerium–zinc
system, iron–air, borohydride fuel cells, Acidic and alkaline electrolytes
and wastewater treatment. He super- Iron–air batteries normally use an alkaline electrolyte such as
vises a number PhD, MSc and BSc proj- KOH, which has a good electrical conductivity and is not too
ects in electrochemical engineering and collaborates with col- corrosive to the iron electrode. The low solubility of oxidation
leagues in other departments. The results of these projects have products such as Fe(OH)2 helps to avoid the formation of den-
led to over 75 peer-reviewed papers, have been presented at inter- drites (iron dendrites are instead formed under acidic condi-
national conferences and have contributed to advances in scientific tions in which iron species are more soluble).[15] Acid electro-
understanding of electrochemical reactors for energy storage lead- lytes are seldom reported, but their use could help to suppress
ing to step changes in technology. the carbonisation of the electrolyte that can clog the air-

ChemPlusChem 2015, 80, 323 – 335 www.chempluschem.org 325  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Minireviews
breathing electrode by depositing carbonates, if CO2 is not re-
moved from air entering the air electrode. A Japanese patent
suggests the use of an acid electrolyte and claims to have dep-
osition of carbonate suppressed to stabilise charge and dis-
charge cycles.[16] Another Japanese patent[17] suggests the use
of gelled products containing ketone compounds providing
high discharge capacity and a good open-circuit voltage.
High-temperature iron–air batteries have been reported that
use nanoparticles of CeO2 mixed with the FeFe3O4 electrode
at 550 8C. This battery uses high-performance electrodes with
a specific energy discharge 91.0 % of the maximum theoretical
value and a round-trip efficiency of 82.5 %.[18]

Important Features of the Iron–Air Flow Cell

Iron–air batteries normally consist of one iron electrode with
two air-breathing electrodes each side, in a plane parallel con- Figure 4. Current versus potential curves of the charge–discharge cycles of
figuration. Figure 3 a shows a single unit cell whereas Fig- a Co–Ni oxide bifunctional electrode on carbon in 12 mol dm3 KOH electro-
ure 3 b represents a six-cell configuration.[19] lyte at 27 8C.[20] The horizontal dashed lines show the overpotential needed
The air electrodes in the rechargeable iron–air battery are for the reduction and evolution of oxygen at a current density of
 20 mA cm2.
designed to achieve both the evolution and the reduction of
oxygen during the charge and discharge cycles, respectively.
The availability of an electrode material able to withstand the trode potential values at which the reduction and the evolu-
high positive potentials during oxygen evolution is one of the tion occur when the current density is  20 mA cm2 are 0.22
limitations of the development of practical metal–air batteries. and 0.66 V versus Hg/HgO, respectively. The overpotential
The electrode material must be electrochemically stable at values are 0.52 V for reduction and 0.36 V for evolution. These
high positive electrode potentials and at the same time able large distances from the reversible value show how challeng-
to catalyse the reduction and evolution of oxygen with mini- ing it is to find a catalyst that is able to decrease this overpo-
mal overpotential losses. tential. In some cases, it has been reported that oxygen evolu-
Figure 4 shows the current versus potential curve for a bi- tion can be performed by a separate electrode; however, this
functional electrode of cobalt and nickel oxides on carbon increases the volume and complexity of the cell stacks.[19] To
during charge and discharge cycles in 12 mol dm3 KOH elec- compensate for the sluggish rate at which the reduction of
trolyte at 27 8C.[20] The predicted equilibrium potential for oxygen occurs, two air-breathing electrodes are used per iron
oxygen is around 0.3 V versus Hg/HgO and the actual elec- electrode, which leads to the use of monopolar electrodes. In

Figure 3. The iron–air unit cell: a) cross section of a single cell (after Cnobloch)[19] and b) a stack of six iron–air unit cells.[19]

ChemPlusChem 2015, 80, 323 – 335 www.chempluschem.org 326  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Minireviews
a typical system, a fan and pump are used to supply air to the At 298 K, the standard cell potential E8 for reaction (4) varies
air electrode and remove any carbon dioxide in the air feed. between 1.16 and 1.28 V corresponding to a Gibbs free energy
The flow rates of the aqueous electrolyte and the air supply change between 224 and 247 kJ mol1. The predicted spe-
are related to the rate at which the current is withdrawn from cific energy density based on reaction (1) is around
the cell, and help to remove heat generated during charge 1230 W h kg1 for Fe + 2 and could increase slightly if reac-
and discharge cycles. The electrolyte circuit also flows through tion (4) is considered. In practice, the specific energy density is
a heat exchanger and purifier to facilitate electrolyte condition- nearly an order of magnitude lower at approximately
ing. 80 W h kg1, as a result of the contribution of several factors in-
Detailed descriptions of the components and manufacturing cluding cell components, overpotential inefficiencies and dis-
process of air-breathing electrodes used in the iron–air cell charge products that add mass to the cell. This will be consid-
prototypes developed in the 1970s[1–6] can be found in the lit- ered in more detail in the section “Performance”.
erature, but there is no description of the use of flow fields to It is believed that the iron electrode reaction (1) proceeds
distribute air into the porous air-breathing electrode. In these via a surface-adsorbed intermediate [Fe(OH)]ad, which is re-
prototypes, a mixture of oxygen and nitrogen gas, to avoid duced to HFeO2 , a dissolved species that reacts with OH to
carbonation of the electrolyte, was introduced into the cell produce Fe(OH)2 on the electrode surface.[22, 23]
into an acrylic chamber fitted over each air electrode mounted On recharging the battery, a second reduction plateau, rep-
in an acrylonitrile–butadiene–styrene (ABS) polymer frame. resenting the reverse of reaction (2), must compete with para-
sitic hydrogen evolution [Eq. (5)]:

Electrode and Cell Reactions 2 H2 O þ 2 e ! H2 þ 2 OH E  ¼ 0:83 V vs: SHE ð5Þ

Figure 1 shows the unit iron–air cell with an anion membrane,
and indicates the main processes occurring on a) discharge Hydrogen evolution decreases the charging efficiency such
and b) charge. The most common strategy is to use a negative that the electrode must be overcharged to reach full capaci-
iron electrode, which involves changes in the oxidation state ty[25] and decreases the cell potential on discharge, owing to
of iron from Fe0 through FeII to FeIII coupled to an oxygen posi- the ohmic resistance of gas bubbles near the iron electrode. It
tive electrode. On discharge, the electrode reactions at the is desirable to add materials such as Bi2S3, K2S or aliphatic sul-
negative electrode involve the oxidation of the metallic iron to fides to the iron electrode to increase charging efficiency (see
FeII to form the hydroxide Fe(OH)2, and some authors have the section “The negative (iron) electrode”). Sulfide compounds
suggested[21] that further oxidation occurs within this process are believed to work by decreasing the effect of iron passiva-
to form FeIII as Fe(OH)3 in a complex process relative to the for- tion on the conductivity of the iron electrode, and may even
mation of magnetite Fe3O4, which occurs at lower potential. In increase the porosity of the iron powder and promote dissolu-
fact the electrode reactions at the iron electrode are complex tion of the passivating layer.[26] In the case of aliphatic sulfides,
owing to solid-state transformation within the surface films. access of H2O to the surface of the electrode is restricted,
During discharge of the battery, metallic iron is ultimately con- which decreases the rate of parasitic hydrogen evolution.[27]
verted to magnetite (Fe3O4) and is transformed back to metal- The reaction at the air electrode converts oxygen dissolved
lic iron as the battery is recharged. The actual mechanism by in the electrolyte into OH or H2O upon discharge, and evolves
which this process occurs is still a matter of speculation, with O2 from these electrolyte species (OH or H2O) upon recharg-
several intermediate species thought to occur during the two ing. For a secondary iron–air battery, therefore, the air elec-
main stages of charge–discharge [see Eqs. (1) and (2)]:[22–24] trode must be bifunctional, that is, catalyse both the oxygen
reduction reaction (ORR) and the oxygen evolution reaction
Fe þ 2 OH Ð FeðOHÞ2 þ 2 e E  ¼ 0:88 V vs: SHE ð1Þ (OER). Additionally it must incorporate a gas diffusion layer
(GDL) to allow oxygen transport to the electrolyte/electrocata-
lyst interface.
3 FeðOHÞ2 þ 2 OH Ð Fe3 O4 þ 4 H2 O þ 2 e E  ¼ 0:76 V vs: SHE
In practice, bifunctional catalysts with low overpotentials (<
ð2Þ
250 mV) for both oxygen reduction and evolution are rare and
often contain precious metals such as Pt, Pd, Ru and Ir.[2] How-
in which SHE is the standard hydrogen electrode.
ever, most practitioners working on bifunctional electrodes use
At the positive electrode, the reduction of oxygen in an alka- non-precious metals and often use two electrocatalyst to sepa-
line electrolyte is [Eq. (3)]: rate the oxygen reduction and evolution reactions respectively,
in the same electrode configuration. Oxygen reduction in alka-
O2 þ 2 H2 O þ 4e Ð 4 OH E  ¼ 0:401 V vs: SHE ð3Þ line solution, reaction (3), is notoriously slow and can proceed
through alternative pathways owing to intermediate peroxide
The simplest overall cell reaction that takes place in the bat- formation in successive two-electron changes [Eqs. (6 a) and
tery considering reactions (1) and (2) can then be written as (6 b)]:[28, 29]
[Eq. (4)]:
O2 þ H2 O þ 2 e ! HO2  þ OH E  ¼ 0:076 V vs: SHE ð6aÞ
3 Fe þ 2 O2 Ð Fe3 O4 ð4Þ HO2  þ H2 O þ 2 e ! 3 OH E  ¼ 0:878 V vs: SHE ð6bÞ

ChemPlusChem 2015, 80, 323 – 335 www.chempluschem.org 327  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Minireviews
The direct, four-electron reduction of oxygen in reaction (3) sumes water such that the electrodes need to be overcharged
is preferred over the two-electron pathway of reactions (6 a) to reach their full capacity.[35] Another problem with the iron
and (6 b), as it occurs at a more positive standard potential electrode is its poor discharge rate as a result of the formation
and gives a more positive overall cell voltage. Additionally, the of an electrically insulating FeII hydroxide, which decreases the
peroxide formed by the two-electron reduction is reactive and cell voltage at high discharge rates.[36–39]
can contribute to the corrosion and degradation of the cell The iron electrodes have usually been made from powdered
components.[30] materials containing metallic iron or iron oxides, most com-
Generally, the four-electron pathway is favoured on plati- monly Fe3O4. These powders can be supported on conductive
num-group metals (Pt, Pd and Ni), silver and certain metal materials such as carbon to improve their available surface
oxides such as spinels, perovskites and pyrochlores.[31] The area and conductivity. The main challenges in producing a suit-
two-electron peroxide pathway is favoured on carbon materi- able iron electrode for an iron–air battery are to 1) increase uti-
als (including graphite), gold and most transition-metal lisation of iron material, 2) avoid passivation of iron (rusting)
oxides.[32] Some oxides, such as MnO2, are thought to support and 3) reduce hydrogen evolution at the electrode surface
a quasi-four-electron reaction, since MnO2 is highly catalytic to during charging.
peroxide decomposition as in reaction (6 b), this reaction oc- The best strategy to address the poor utilisation of iron is to
curring so rapidly that four electrons appear to have been make more of the solid electrode accessible to the electrolyte
transferred simultaneously.[33] by using nanosized particles of iron to increase the surface
On recharging, the reverse reaction occurs and oxygen is area. Recently, several studies have considered nanostructured
evolved in a four-electron transfer process, which may involve iron electrodes and have shown increased charge capacities
the release of two oxygen atoms coordinated to a metal atom compared with commercial Fe3O4 powder pocket-plate electro-
on the surface of the electrode.[34] The most active catalysts for des. Table 2 summarises the results achieved from these elec-
this reaction are IrO2 and RuO2 but these are expensive metals, trodes. For carbonyl iron electrodes produced by Manohar
and catalysts based on other metal oxides and their mixtures et al.,[25, 40] the high purity of the carbonyl iron was found to in-
are under development, as seen in the section “The positive crease the discharge specific capacity even though the particle
(air) electrode”. size was not much smaller than that in commercial electrodes
produced by Changhong Electric Co. (Table 2). The discharge
specific capacities in Table 2[25, 40, 41–44] are compared at the dis-
The Distinction between Iron–Air and Other
charge rate of C/5 (in which C is the full charge capacity of the
Metal–Air Flow Batteries
electrode), which is recommended by Sichuan Changong Bat-
The iron electrode in the iron–air cell does not form dendrites tery Co. for operation of its Ni–Fe batteries.
during the plating process, since the electrochemical reactions In these studies, a variety of compounds, such as Bi2S3, Bi2O3,
involve solid-phase transformations, but the shape and volume FeS and Na2S, were added to the iron electrode to reduce the
of the electrode can change during the charge and discharge rate of hydrogen evolution and increase the charging efficien-
cycles owing to the formation of the voluminous, insoluble cy of the electrode. At the C/5 discharge rate, the best results
species Fe(OH)2 and Fe3O4, which have different densities from for both capacity and efficiency were obtained with the 5 wt %
metallic Fe. There is also the possibility that parts of the elec- Bi2S3 additive, although it was found that at higher discharge
trode become inactive because of an uneven potential and rates of C/1, the electrodes with 10 % Bi2O3 + 5 % FeS additives
current distribution during the plating and dissolution cycles. retained more of their capacity and performed just as well as
The features of the iron–air cell may be summarised: the 5 % Bi2S3 electrode. The authors suggested that the pres-
ence of sulfide derivatives decreases the detrimental effect of
1) No dendrite formation or substantial shape change of the iron passivation on the electrode capacity by forming a conduc-
negative electrode (in contrast to zinc–air batteries). tive FeS layer with surface iron atoms, thereby decreasing the
2) The Fe–air cell has a lower predicted voltage than the Zn– insulating effect caused by passivation.[40] Using Bi2O3 additive
air cell (1.28 vs. 1.65 V), but is easier to recharge. alone did not significantly increase the capacity of the iron
3) The negative electrode has a long lifetime. electrodes, thus indicating that it is the sulfur atoms that are
4) The negative electrode needs less protection from corro- important rather than the presence of bismuth alone.
sion compared to Al, Li. Other iron electrodes (Table 2) have recently demonstrated
5) The Fe–air cell is safer than the Li–air cell but heavier. even greater discharge capacities when iron oxide nanoparti-
6) The Fe–air cell involves complex surface films, unlike the cles supported on conductive carbon particles were used. The
Zn–air cell, which normally involves reversible dissolution nano-Fe3O4-loaded tubular carbon nanofibre (nano-Fe3O4/
of zinc as zincate ions in solution. TCNF) electrodes produced by Ito et al. and Hang et al. are re-
ported to have yielded a specific capacity of 786 mA h g1 and
cycling efficiency of 76 % at the 30th cycle.[42, 43] The small size
and large specific surface area (> 300 m2 g1) of the particles
The Negative (Iron) Electrode
supported on conductive crystalline carbon nanofibres and
Iron electrodes suffer from low charging efficiency owing to carbon nanotubes are the most likely factors to have contrib-
the secondary reaction of hydrogen evolution, which con- uted to their increased capacity; TEM and XRD measurements

ChemPlusChem 2015, 80, 323 – 335 www.chempluschem.org 328  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Minireviews

Table 2. Properties of iron electrodes used for the iron–air battery. Data refer to 298 K.

Iron materials Particle Support materials Additives Electrode design Discharge Percentage Ref.
size [mm] specific charging
capacity [A h g1] efficiency
commercial electrode (Chang- 1–3 none none pressed pocket- 0.12 at C/5 rate, 70 at C/20 rate [25]
hong), Fe3O4 particles plate 9 mA cm2
carbonyl iron 0.5–3 none 5 % Bi2Si3 pressed powder 0.29 at C/5 rate, 96 at C/20 rate [25]
+ PE binder 3 mA cm2
carbonyl iron 0.5–3 none 10 % Bi2O3 pressed powder 0.14 at C/5 rate 92 at C/20 rate [40]
+ PE binder
carbonyl iron 0.5–3 none 10 % Bi2O3 + 5 % FeS pressed powder 0.22 at C/5 rate 93 at C/20 rate [40]
+ PE binder
carbonyl iron 0.5–3 none 5 % Bi2O3 + Na2S pressed powder 0.24 at C/5 rate not measured [40]
+ PE binder
Fe2O3 < 0.05 acetylene black K2S pressed powder 0.67 at 5 mA cm2 not measured [41]
+ PTFE
Fe2O3 < 0.05 carbon nanofibres 2 % Bi2S3 + K2S powder rolled 0.8 at 2 mA cm2 not measured [42]
with PTFE
Fe3O4 0.02 carbon nanotubes K2S powder rolled 0.79 at 0.5 mA cm2 84 [43]
with PTFE
Fe + Fe3O4 5–10 decomposed PVA carbon 1 % Bi2S3 powder spread 0.4 at C/5 rate 80 [44]
coating + carbon black with PTFE

revealed that the nanosized Fe3O4 particles were well-distribut- ciently catalysing both the ORR and the OER. A high surface
ed and bonded tightly to the tubular carbon nanofibres. The area is essential to increased catalytic activity; yet both the cat-
alkaline electrolyte contained K2S as an additive, but it appears alyst and supporting materials must be highly resistant to the
that K2S added to the electrolyte is less effective as a mecha- corrosive effect of a strongly alkaline solution, typically 5–
nism to increase the charging efficiency than solid additives 8 mol dm3 KOH.
such as Bi2S3 and FeS.[25, 40] Sulfide ions are thought to react Recently, alternatives to the expensive traditional catalysts
with Fe(OH)2 to produce FeS on the electrode surface,[40] but based on Pt, Pd (ORR) and IrO2, RuO2 (OER) and their alloys
this FeS may not be as uniformly distributed or electroactive have been considered. Mixed metal oxides, such as perov-
as solid FeS mixed directly into the electrode powder. skites, spinels and pyrochlores, are thought to promote the
To conclude this section, the source of iron (i.e., whether four-electrode oxygen reduction pathway and show good cat-
metallic or metal oxide) does not affect the capacity and uti- alytic activity for both the ORR and OER. As they are semicon-
lisation of iron as much as the size of the particles. This is re- ductive or insulating, they must be supported on a porous,
flected in Table 2 by the fact that the smaller the iron-contain- conductive material. Table 3[2, 45–56] gives a summary of the elec-
ing particles, the larger the observed discharge capacity. How- trochemical properties of some catalysts reported to have bi-
ever, the number of iron electrodes reported in the literature is functional activity towards both oxygen reduction and evolu-
too small to make a definite conclusion on the effect of parti- tion. These are relatively rare in the literature: most oxygen
cle size; there could be other factors involved (such as particle catalysts only have high activity reported towards either the
morphology and the nature of the supporting materials and ORR or the OER.
additives). The addition of bismuth sulfide or bismuth oxide Figures 5 and 6 show current versus potential curves for
with iron sulfide is highly effective at increasing the charging some of the bifunctional catalysts outlined in Table 3, with
efficiency of an iron electrode, although the effect of amounts data included for the most active noble-metal catalysts (Pt/C
of Bi2S3 greater than 5 wt % is unknown, as all electrodes have for the ORR[57] and RuO2 and IrO2 for the OER) and also for
used  5 wt % in their construction. A newly proposed alterna- some other catalysts, such as Fe3Co[58] and carbon black doped
tive to the use of Bi2S3 or FeS is to protect the surface of the with nitrogen and fluorine,[59] which have a strong activity re-
electrode with a self-assembled monolayer of alkane thiols, ported only for either the ORR or OER. Figure 5 shows results
with a reported 92 % charging efficiency for n-hexanethiol on for oxygen reduction obtained for completed gas-diffusion air
a carbonyl iron electrode.[27] This alternative has the advantage electrodes whereas Figure 6 shows results for both oxygen re-
of lower cost and better recyclability than bismuth com- duction and evolution that were obtained with the catalyst ad-
pounds, but the effect on the electrode capacity and wettabili- hered to the surface of a rotating-disk electrode in an oxygen-
ty has yet to be established. saturated solution.
The results displayed in Table 3 and Figures 5 and 6 show
that spinels or perovskites tend to have the lowest overpoten-
The Positive (Air) Electrode
tials for oxygen reduction. In particular, the perovskite
For a rechargeable secondary iron battery to be feasible, a bi- LaSr3Fe3O10 (LSFO) showed promising bifunctional activity
functional air electrode must be found that is capable of effi- when supported on a glassy carbon rotating-disk electrode,

ChemPlusChem 2015, 80, 323 – 335 www.chempluschem.org 329  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Minireviews

Table 3. Properties of recently developed bifunctional air electrodes. EORR : onset potential of oxygen reduction reaction, EOER : onset potential of oxygen
evolution reaction, jmax ORR: maximum current density for oxygen reduction at 0.2 V versus Hg/HgO using the catalyst.

Catalyst Supporting material EORR vs. Hg/HgO [V] EOER vs. Hg/HgO [V] jmax ORR at 0.2 V vs. Lifetime (number Ref.
Hg/HgO [mA cm2] of cycles)
Transition metals/simple oxides:
Ag sintered Ni powder – – 80 500 [2]
Ag–MnO2 nanorods graphite powder, carbon black 0.03 0.61 0.25 270 [45]
MnO2 + RuO2 Ketjen black 0.03 0.53 0.20 50 [46]
Ag nanoparticles glassy carbon 0.09 0.61 13 – [47]

Spinels:
Sr2 + - or Ca2 + -doped La2NiO4 Ketjen black and Super-P carbon 0.08 0.29 – 20 [48]
LaMnO3 carbon black 0.09 0.60 10 150 [49]
LaMnO3 acetylene black 0.10 0.40 51 – [50]

Perovskites:
La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) Ketjen black 0.08 0.29 200 100 [51]
LaSr3Fe3O10 (LSFO) glassy carbon 0.28 0.29 – – [52]
La0.6Ca0.4CoO3 (LCCO) Vulcan carbon 0.05 0.50 100 100 [53]

Others:
MnO2 + Ni–Fe layered double Vulcan carbon 0.05 0.5 10 50 [54]
hydroxide
NiCo2S4 glassy carbon 0.05 0.38 – – [55]
Pb2Ru2O6.5 (pyrochlore) commercial carbon GDL 0.1 0.4 20 6 [56]

Figure 5. Current density versus electrode potential curves for oxygen re-
duction at gas diffusion electrodes (GDEs) in concentrated alkaline solutions.
CB = untreated carbon black, 0.5 mV s1;[49]Ag NP = silver oxide nanoparticles,
10 mV s1;[47] LaMnO3/CB = LaMnO3 on carbon black, 0.5 mV s1;[49] LaMnO3/
AB = LaMnO3 on acetylene black (scan rate not given);[50] LCCO = -
La0.6Ca0.4CoO3, 5 mV s1;[53] Pt/C = commercial Pt on carbon GDE produced by
Johnson Matthey, at 1 mV s1.[57] The electrolyte was 6 mol dm3 KOH solu-
tion at 23 8C, except for LaMnO3/AB,[50] which was in 8 mol dm3 KOH at
60 8C.

with very low overpotentials for the oxygen reduction and

evolution reactions (Figure 6 a),[52] . Of the catalysts tested on
a fully developed gas diffusion electrode (Figure 5), the perov-
skites LCCO and LaMnO3 both achieved high ORR activity com-
pared with a commercial Pt/C electrode.[50, 53] The perovskite Figure 6. Current density versus potential curves for a) oxygen reduction
LSCF and the Sr2 + /Ca2 + -doped La2NiO4 spinel also show very and b) oxygen evolution at rotating-disk electrodes coated with various cat-
low overpotentials ( 10 mV) for the OER, and are also good alysts in 0.1 mol dm3 KOH solution at room temperature. The labels in (a)
refer to LSFO = LaSr3Fe3O10,[52] WC/Ag/C = tungsten carbide with silver on
catalysts for the ORR.[48, 51] Pyrochlores, such as Pb2Ru2O6.5, have
carbon,[77] Pt/C = platinum on carbon;[55] Fe3Co,[58] NiCo2S4,[55] and CB-
very good bifunctional properties (Table 3) but show poor sta- NF = carbon black doped with nitrogen and fluorine.[59] The labels in (b) refer
bility at negative potentials upon repeated cycling[56] and in- to Ag/MnO2,[45] Pt/C,[42] IrO2 and RuO2,[78] Pt/C 2[55] and NiCo2S4.[55]

ChemPlusChem 2015, 80, 323 – 335 www.chempluschem.org 330  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Minireviews
volve environmental and resource limitations, since they con-
tain both precious and toxic metals.
The search for new catalysts for oxygen reduction and evo-
lution must focus on the nature of the surface bonding reac-
tions with O2 and their enhancement of catalytic activity
(known as the “activity descriptor”).[60] For example, by alloying
platinum with other metals, the energy of the valance d band
of the platinum species can be shifted by the alloying metal,
thus causing stronger or weaker bonding and antibonding in-
teractions with the O2 molecule. In particular, the alloys
PtPd(111) and Pt3Ni(111) have been shown to have enhanced
bonding properties.[61, 62] Similarly, recent advances have been
made in our understanding of how some perovskite materials
have a bifunctional activity towards oxygen that is comparable
to that of the noble metals Pt and Pd. The reduction of
oxygen on perovskite compounds proceeds initially through
the peroxide route, but the perovskite retains the peroxide in- Figure 7. Charge–discharge cell potential curves for an iron–air battery at
termediate via bonding so it can be further reduced, thus re- a) 25 mA cm2 charge and 30 mA cm2 discharge of a prototype stack con-
sulting in an apparent 4 e process that is more efficient than taining 190 air and 95 iron electrodes of size approximately 10  20 cm divid-
the 2 e reduction mechanism that releases OOH , which often ed into 19 cells connected in series;[1] b) charge and discharge at
40 mA cm2 of a cell containing air electrodes of area 100 cm2 separated
occurs on other transition-metal oxides.[63] The structure of from the Fe electrode by 3–4 mm.[19]
a perovskite also enables it to catalyse the OER efficiently:
a recent study of (Ln0.5Ba0.5)CoO3d, in which Ln is a lanthanide
metal, highlighted the importance of the closeness in energy continuous line (curve a) was performed at 30 and 25 mA cm2,
of the oxygen p band to the Fermi level of the perovskite in respectively,[1] whereas the battery represented by the dashed
enhancing the rate of O2 evolution. It was found that the line of curve b) was charged and discharged at the same cur-
energy difference must be no less than 1.75 eV, otherwise the rent density of 40 mA cm2.[19] In the charge cycle curve a, the
oxide is destabilised prior to the OER.[64] cell maintains a potential of 1.6 V corresponding to the reac-
In most cases, bifunctional oxygen catalysts have limited tion of the first charging plateau that can be observed in
electrical conductivity, so a supporting material with which the Figure 8 for Fe3O4 to Fe(OH)2 (which represents the half-cell
catalyst is chemically bonded or physically mixed is essential. potential process of the iron electrode) during the first
Carbon blacks, such as Ketjen black, acetylene black and 4 hours.[35] The voltage started to increase towards 1.72 V and
Vulcan, are popular choices for the catalyst supporting materi- the cell was discharged at about 0.8 V for 4 hours before the
al, as they combine high surface area with reasonably high cell voltage dropped to 0.6 V. The cell represented by curve b
conductivity (the surface area of Vulcan XC72 is on the order was charged from 1.3 V, which increased gradually to 1.6 V
of 250 m2 g1, and the conductivity is 4.0 S cm1).[65] Occasional- during the first 2 hours then reached a plateau at about 1.64 V.
ly, other conductive powders, such as Ni powder and Ebonex, The process involved the reduction of Fe3O4 to Fe(OH)2 accord-
may be used; although these powders typically have a much
lower surface area than carbon blacks, they have a much
greater resistance to corrosion in alkaline solution at higher
potentials (> 0.5 V vs. Hg/HgO).[66, 67]

Performance
The cell potential of the iron–air battery is influenced by
a number of factors, particularly the open-circuit cell potential
o
 
Ecell ¼ Eco  Eao and the activation and concentration polarisa-
tion of the positive and negative electrodes (hact, hconc), togeth-
er with the ohmic (electronic and ionic) drop (IR) across com-
ponents k, according to [Eq. (7)]:[68]
    X    
o
Ecell ¼ Ecell  haact   hcact   k
ðIRÞk haconc   hcconc  ð7Þ

Figure 7 shows the cell voltage versus time for two typical
iron–air batteries during the charge–discharge cycles. The Figure 8. Charge–discharge half-cell potential for the iron electrode together
charge and discharge cycle for the battery represented by the with hydrogen evolution.[35]

ChemPlusChem 2015, 80, 323 – 335 www.chempluschem.org 331  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Minireviews
ing to reaction (2) and was followed by a cell potential in-
crease to 1.8 V, typically observed for the formation of metallic
iron. In the discharge cycle, the voltage activates from 0.7 to
0.8 V during the first half an hour and stays relatively constant
at approximately 0.8 V for 3 hours followed by a gradual volt-
age decline to 0.6 V. During this process, the metallic iron oxi-
dises to Fe2 + . The authors considered that after a cell potential
of 0.6 V is reached (6 hours), the iron partially oxidises to Fe3 +
and, at that point, they report 100 % charge efficiency, 40 %
voltage efficiency and 35 % energy efficiency.[19]

Electrode Materials and Cell Design Figure 9. A possible design for a gas-diffusion air electrode.[19]
To make the best use of the active materials (nanoparticulate
iron, bifunctional oxygen catalyst) described above, the elec-
trodes must be well-designed. In the case of the iron elec- oxygen to diffuse into the electrode (Figure 9). The porous hy-
trode, the surface area of iron in contact with the electrolyte drophobic layer, often termed the “gas diffusion layer” (GDL), is
must be as large as possible, to maximise utilisation of the usually composed of a microporous conductive powder treat-
active iron-containing material as seen in reaction (1) during ed with PTFE or a similar alkali-resistant hydrophobic polymer,
cell discharge. For example, the carbon nanofibre supporting to ensure that the GDL pores do not become flooded with the
material in Ref. [42] had a specific surface area of 300 m2 g1, electrolyte. Oxygen gas diffuses through these open pores, to-
and was decorated with < 50 nm sized Fe2O3 particles; both wards the hydrophilic catalyst layer, which is flooded with elec-
factors contributed towards the high specific discharge capaci- trolyte. An additional layer can be added to the back of the
ty of 800 mA h g1 obtained from these electrodes. GDL to improve the hydrophobicity and structural integrity of
As described in Table 2, the best-performing iron electrodes the electrode. This layer can be made of carbon cloth/paper,
consist of an iron oxide nanopowder, which has been well-dis- metal foams or porous PTFE, depending on the conductivity,
persed on a conductive substrate such as carbon. The conduc- density, porosity and mechanical strength required.[66, 71–74]
tive substrate helps to circumvent the problem of iron elec- The catalyst layer is often made of the catalyst powder
trode passivation, in which the build-up of electrically insulat- pressed together with a conductive powder. Carbon is often
ing iron oxides on the surface of the iron nanoparticles renders the conductive additive of choice, because of its low cost and
the iron particles less conductive.[69] Carbon is a good material low density. However, carbon has poor corrosion resistance at
for this role, as it is much less dense than iron and only makes strongly oxidising positive potentials.[75] A recent example in-
a small contribution to the overall weight of the electrode. volves an air electrode in which both the GDL and catalyst ma-
It is desirable to build a three-dimensional iron electrode, to terial were nickel-based, in the form of nickel powder and
achieve a larger volumetric energy density by storing a mass nickel foam.[66] The authors reported a good long-term stability
of iron equivalent to several flat iron sheets in a single elec- of their spinel-catalysed nickel air electrode at recharging po-
trode. This has some design implications for the iron electrode: tentials of + 0.55 V versus Hg/HgO and current densities as
because not all of the active material is at the electrode sur- high as 100 mA cm2.
face, the electrode must be porous and hydrophilic, to allow Ideally, the gas diffusion electrode should be much thinner
sufficient mass transport of OH from the electrolyte into the than the iron electrode: about 0.1 mm or thinner if possible. A
electrode. thin GDL (carbon paper and cloth are typically 100–400 mm) re-
There are several methods of building a porous 3 D elec- duces the time needed for oxygen to diffuse through the GDL.
trode: a pellet of the powdered carbon-supported iron materi- A very thin catalyst layer allows the OH in the electrolyte to
al may be hot-pressed together with a binder, such as PTFE or diffuse more rapidly towards the catalyst during oxygen reduc-
Nafion, and bonded to a metallic current collector. This is the tion, and allows oxygen bubbles to escape from the catalyst
most common method, used for all the iron electrodes report- during oxygen evolution. The overall thickness of the air elec-
ed in Table 2. An alternative method would be to load the iron trode has been shown to have a large influence on the cell
powder into a highly porous 3 D material, such as carbon felt performance in Li–air batteries.[76]
or carbon foam. Although there are, as yet, no literature re-
ports of iron electrodes constructed using this technique, this
Cell design
approach could increase the conductivity and volume/area
ratio[70] of the electrode but would have the disadvantage of The iron–air batteries reported to date have followed the typi-
adding extra mass of non-active materials to the electrode. cal parallel-plate electrode configuration used in most electro-
The air electrode has a more complex design than the iron chemical reactors.[68] This arrangement allows high mass trans-
electrode, as it requires a hydrophilic layer that contains the port rates and heat dissipation as well as the flexibility to scale
catalyst on the side facing the electrolyte, whilst having up the reactor using modules. Generally, the geometry of the
a porous hydrophobic layer on the reverse side to allow electrodes is considered first together with the paths followed

ChemPlusChem 2015, 80, 323 – 335 www.chempluschem.org 332  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Minireviews
by the fluid flow, such as the air or oxygen and the aqueous a simple treatment of the electrochemical reactions and
electrolyte. Small cells are preferred in the laboratory owing to charge transport. The model showed good agreement with
the low requirements of volume and power but in larger cells the experimental data.
the heat generated during the charge and discharge cycles Mathematical models of aqueous iron–air batteries can be
should be considered. Usually, the flow of electrolyte has been developed by considering mass, charge, energy and fluid mo-
used as a cooling system as well as ionic conductor. The reac- mentum balances, with detailed descriptions of transport phe-
tor design should be such that the gas generated during oper- nomena (through convection, diffusion and electro-migration)
ation (charge–discharge) can easily escape the reactor and in the porous electrodes and recirculation of the electrolyte.
does not increase the electrical resistance between the electro- Templates for such models are available from the fuel cell and
des. The design and construction of the iron–air cell should flow battery literature.[81, 82] Crucial to accuracy of the models
take into account the following factors: will be the description of the electrochemical reactions (multi-
step kinetics in the iron electrode and hydrogen evolution).
1) Electrode kinetics; a knowledge of the electrode kinetics is Without this level of detail it will not be possible to predict the
needed to estimate the number of unit cells required for charge–discharge performance accurately.
a particular power output, and excess metal has to be in-
cluded to compensate for overpotentials
2) Electrolyte concentration; it is important in order to choose
the material that offers the highest chemical and mechani- Summary and the Scope for Future Research
cal stability
3) Current and potential distributions and heat transfer simi- 1) The iron–air flow battery has been known since the 1970s;
larities to allow effective use of the whole electrode surface a recent upsurge of interest has been driven by enhanced
area incentives to develop moderate-cost, robust, environmen-
4) Inter-electrode gap; to minimise resistance but allow gases tally acceptable and rechargeable batteries.
to escape 2) The iron–air cell has an open-circuit cell potential of 1.28 V
5) Effective flow manifolds to distribute the flow evenly and a moderate specific charge capacity of < 300 A h kg1.
through the surface of either flat or porous electrodes Recent examples of the iron–air cell have realised 96 %
6) A calming section inside the reactor to allow the formation charge efficiency, 40 % voltage efficiency and 35 % energy
of uniform flow patterns and avoid entry effects efficiency with operation up to 3000 cycles being reported.
7) A flow regime including laminar or turbulent core flow, 3) The iron electrode supports complex reactions but does
flow distribution and the use of turbulence promoters not suffer similar shape-change problems owing to metal
8) A low pressure drop over the cell, which affects pumping deposition/stripping as in Zn–air batteries. The develop-
cost ment of robust nanostructured catalysts having sustained
performance on charge–discharge cycling without corro-
sion or hydrogen evolution problems remains an important
research endeavour. A specific capacitance of 786 mA h g1
Modelling and Simulation
and a cycling efficiency of 76 % have been reported when
The development of practical multi-physics models that ration- Fe nanoparticles have been supported on carbon nanofi-
alise the performance of the iron–air cell and simulate its re- bres.
sponse to operational variables should be considered a critical 4) There have been many approaches to synthesise more
research need in the development of modern, competitive bat- robust, lower overpotential, bifunctional air electrodes, es-
teries. The complex nature of reactions at the iron electrode, pecially those based on nanostructured mixed metal oxides
the very limited consideration of their kinetics in the literature not relying on scarce or precious metals. Combinatorial
and the moderate data available from process trials under con- methods of synthesis and testing and scalable synthesis
trolled conditions have severely restricted the development of techniques of nanostructured catalysts have a continued
mathematical models of the aqueous iron–air battery, despite role in realising improved materials in iron–air and other
their potential use in performance rationalisation, prediction batteries.
and optimisation. This situation is replicated for other metal– 5) Since the 1970s, plane parallel flow cells have been com-
air batteries. monly used. Improvements through the use of modern
Guo et al.[79] developed a detailed, physics-based model for polymers for the cell body, computational cell design and
the solid oxide iron–air redox flow battery, which was used to prototyping by, for example, 3 D printing promise to trans-
simulate the discharge behaviour at high current density. The form cell designs.
flow battery system they considered combines an FeFeO 6) The development of mathematical models to rationalise
redox couple as the energy storage unit and a regenerative performance and simulate the effect of operational varia-
solid oxide fuel cell as the electrical functioning unit in a 2 D bles has been severely hindered by the limited studies of
axial symmetric geometry. This 2 D transient, isothermal model electrode kinetics at the complex iron electrode. Such
included detailed mass balances in the gas and solid phases, models will be very dependent on an improved description
with Brinkman’s equation[80] for the momentum balance and of the electrode kinetics at the iron electrode and the

ChemPlusChem 2015, 80, 323 – 335 www.chempluschem.org 333  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Minireviews
quantitative performance of the cell under controlled op- [19] H. Cnobloch, Electrical Rechargeable Iron – Air Batteries. A Way to the
erational conditions. Electrotraction? Proceedings of the 88th Convention of the Battery
Council International Future Clean Silent Power, April 25 – 29, 1976, 88,
7) To develop more efficient iron–air batteries, critical research 39 – 48.
areas include a) more efficient and moderate-cost bifunc- [20] P. N. Ross, Proceedings of the 21st Intersociety Energy Conversion Engi-
tional oxygen positive electrodes, b) negative electrodes neering Conference, American Chemical Society, Washington, DC, 1986,
that can be manufactured economically using technologies p. 1066.
[21] M. Egashira, Secondary Batteries-Metal-Air Systems: Iron – air (Secondary
that are readily scaled, c) improved engineering in cell and Primary) in Encyclopaedia of Electrochemical Power Sources, Vol. 5,
design using additive manufacturing technology and d) ac- Elsevier, Amsterdam, 2009, pp. 372 – 375.
cessible, proven mathematical models to adequately de- [22] B. T. Hang, M. Eashira, I. Watanabe, S. Okada, J. I. Yamaki, S. H. Yoon, I.
scribe and simulate battery performance using aqueous, Mochida, J. Power Sources 2005, 143, 256 – 264.
[23] B. T. Hang, T. Watanabe, M. Egashira, I. Watanabe, S. Okada, J. I. Yamaki,
caustic electrolytes and robust electrodes. J. Power Sources 2006, 155, 461 – 469.
[24] G. P. Kalaignan, V. S. Muralidharan, K. I. Vasu, J. Appl. Electrochem. 1987,
17, 1083 – 1092.
[25] A. K. Manohar, S. Malkhandi, B. Yang, C. Yang, G. K. S. Prakash, S. R. Nar-
ayanan, J. Electrochem. Soc. 2012, 159, A1209 – A1214.
Acknowledgements [26] M. Jayalakshmi, V. S. Muralidharan, Proceedings of the Academy of India
Sciences (Chem. Sci.) 1991, 103, 753 – 761.
This work was enabled by an EU grant FP7 (NECOBAUT Grant [27] S. Malkhandi, B. Yang, A. K. Manohar, G. K. S. Prakash, S. R. Narayanan, J.
agreement no. 314159). Am. Chem. Soc. 2013, 135, 347 – 353.
[28] G. Milazzo, S. Caroli, V. K. Sharma, Table of Standard Electrode Potentials,
Wiley, Chichester, 1978.
Keywords: electrochemistry · energy storage · iron · metal–air [29] A. J. Bard, R. Parsons, J. Jordan, Standard Potentials in Aqueous Solutions,
batteries · oxygen Marcel Dekker, New York, 1985.
[30] K. Kinoshita, Electrochemical Oxygen Technology, Wiley, New York, 1992.
[31] L. Jçrissen, J. Power Sources 2006, 155, 23 – 32.
[1] L. jefors, L. Carlsson, J. Power Sources 1978, 2, 287 – 296. [32] Handbook of Fuel Cells- Fundamentals, Technology and Applications,
[2] L. Carlsson, L. jefors, J. Electrochem. Soc. 1980, 127, 525 – 528. Vol. 2 (Eds.: W. Vielstich, A. Lamm, H. A. Gasteiger), Wiley, New York,
[3] O. C. Wagner, NASA Contract. Rep., NASA-CR-97489, 1968, pp. 13 – 15. 2003, p. 443.
[4] E. S. Buzzelli, H. Moyes, L. B. Berk, D. Zuckerbrod, Review and Summary [33] F. Cheng, J. Chen, Chem. Soc. Rev. 2012, 41, 2172 – 2192.
of Iron – Air Battery Performance at Westinghouse, 1976 – 1984, 84 – [34] G. Mattioli, P. Giannozzi, A. A. Bonapasta, L. Guidoni, J. Am. Chem. Soc.
9D1-FEAIR-R2, Westinghouse R & D Center, Pittsburgh, PA (Aug. 29, 2013, 135, 15353 – 15363.
1984). [35] K. Vijayamohanan, T. S. Balasubramanian, A. K. Shukla, J. Power Sources
[5] E. S. Buzzelli, B. G. Demczyk, L. B. Berk, D. Zuckerbrod, A. Gibney, C. T. 1991, 34, 269 – 285.
Liu, P. L. Ulerich, R. E. Grimble, Iron – Air Battery Development Final [36] K. Vijayamohanan, A. K. Shukla, S. Sathyanarayana, J. Electroanal. Chem.
Report, 1983, 83 – 9D12-MOBET-Ri, Westinghouse R&D Center, Pitts- 1990, 295, 59 – 70.
burgh, PA (Mar. 2, 1983). [37] K. Vijayamohanan, A. K. Shukla, S. Sathyanarayana, J. Electroanal. Chem.
[6] E. S. Buzzelli, C. T. Liu, W. A. Bryant, Iron air batteries for electric vehicles, 1990, 289, 55 – 68.
in 13th Intersociety Energy Conversion Engineering Conference, 1978, San [38] J. Černý, K. Micka, J. Power Sources 1989, 25, 111 – 122.
Diego, California, August 20 – 25, Proceedings, Vol 1, (A79 – 10001 01 – [39] K. Micka, Z. Zabransky, J. Power Sources 1987, 19, 315 – 323.
44), Warrendale, Pennsylvania, Society of Automotive Engineers, Inc., [40] A. K. Manohar, C. Yang, S. Malkhandi, G. K. S. Prakash, S. R. Narayanan, J.
1978, p. 745 – 749. Electrochem. Soc. 2013, 160, A2078 – A2084.
[7] http://arpa-e.energy.gov/?q = arpa-e-projects/iron–air-rechargeable-bat- [41] H. Kitamura, L. Zhao, B. T. Hang, S. Okada, J. Yamaki, J. Electrochem. Soc.
tery, Access date: 14 August 2014. 2012, 159, A720 – A724.
[8] http://www.necobaut.eu/home/home.php, Access date: 14 August [42] B. T. Hang, D. H. Thang, N. T. Nga, P. T. L. Minh, E. Kobayashi, J. Electro-
2014. chem. Soc. 2013, 160, A1442 – A1445.
[9] X. Li, D. Pletcher, C. Ponce de Lon, F. C. Walsh, R. G. A. Wills, “Redox [43] A. Ito, L. Zhao, S. Okada, J. Yamaki, J. Power Sources 2011, 196, 8154 –
Flow Batteries for Energy Storage Using Zinc Electrodes”, in Advances in 8159.
Batteries for Large- and Medium-Scale Energy Storage: Applications in [44] A. S. Rajan, S. Sampath, A. K. Shukla, Energy Environ. Sci. 2014, 7, 1110 –
Power Systems and Electric Vehicles (Eds.: C. Menictas, M. Skyllas-Kazacos 1116.
T. M. Lim), Woodhead, Cambridge, 2014, Chapter 6. [45] F. W. T. Goh, Z. Lui, X. Ge, Y. Zong, G. Du, T. S. A. Hor, Electrochim. Acta
[10] C. Vincent, B. Scrosati, Modern Batteries, 2nd ed., Elsevier, Amsterdam, 2013, 114, 598 – 604.
1997. [46] K. Guo, Y. Li, J. Yang, Z. Zou, X. Xue, X. Li, H. Yang, J. Mater. Chem. A
[11] V. Barsukov, F. Beck, New Promising Electrochemical Systems for Recharge- 2014, 2, 1509 – 1514.
able Batteries, Series 3: NATO ASI Series, Springer, Dordrecht, 1996. [47] N. Sasikala, K. Ramya, K. S. Dhathathreyan, Energy Convers. Manage.
[12] S. Narayanan, G. S. Prakash, A. Manohar, B. Yang, S. Malkhandi, A. Kin- 2014, 77, 545 – 549.
dler, Solid State Ionics 2012, 216, 105 – 109. [48] K. Jung, J. Jung, W. B. Im, S. Yoon, K. Shin, J. Lee, ACS Appl. Mater. Inter-
[13] Electric Fuel. http://www.electric-fuel.com/evtech/index.html Accessed: faces 2013, 5, 9902 – 9907.
14 August 2014. [49] C. Zhu, A. Nobuta, I. Makatsugawa, T. Akiyama, Int. J. Hydrogen Energy
[14] U.S. Department of Energy, APRA-E funding announcement, Gridscale 2013, 38, 13238 – 13248.
Rampable Intermittently Dispatchable Storage, 2010,. [50] M. Yuasa, M. Nishida, T. Kida, N. Yamazoe, K. Shimanoe, J. Electrochem.
[15] R. Qiu, J. Y. Zheng, H. G. Cha, M. H. Jung, K. J. Lee, Y. S. Kang, Nanoscale Soc. 2011, 158, A605 – A610.
2012, 4, 1565 – 1567. [51] H. Ohkuma, I. Uechi, M. Matsui, Y. Takeda, O. Yamamoto, N. Imanishi, J.
[16] A. Umeda, Y. Nukuna, Kokai Tokkyo Koho, JP 2010140736 A 20100624, Power Sources 2014, 245, 947 – 952.
2010. [52] T. Takeguchi, T. Yamanaka, H. Takahashi, H. Watanabe, T. Kuroki, H. Naka-
[17] K. Hayashi, K. Tomihisa, A. Matsuda, T. Kawamura, N. Sakamoto, Kokai nishi, Y. Orikasa, Y. Uchimoto, H. Takano, N. Ohguri, M. Matsuda, T.
Tokkyo Koho, JP 2012074371 A 20120412, 2012. Murota, K. Uosaki, W. Ueda, J. Am. Chem. Soc. 2013, 135, 11125 – 11130.
[18] X. Zhao, Y. Gong, X. Li, N. Xu, K. Huang, J. Electrochem. Soc. 2013, 160, [53] S. Zhuang, H. Zhang, S. Liu, F. Tu, W. Zhang, C. Zhao, Int. J. Electrochem.
A1241 – A1247. Sci. 2014, 9, 1690 – 1701.

ChemPlusChem 2015, 80, 323 – 335 www.chempluschem.org 334  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
 Minireviews
[54] D. Kubo, K. Tadanaga, A. Hayashi, M. Tatsumisago, J. Mater. Chem. A [70] J. M. Friedrich, C. Ponce de Leon, G. W. Reade, F. C. Walsh, J. Electroanal.
2013, 1, 6804 – 6809. Chem. 2004, 561, 203 – 217.
[55] Z. Zhang, X. Wang, G. Cui, A. Zhang, X. Zhou, H. Xu, L. Gu, Nanoscale [71] B. Liu, J. Zhang, X. Wang, G. Chen, Di. Chen, C. Zhou, G. Shen, Nano
2014, 6, 3540 – 3544. Lett. 2012, 12, 3005 – 3011.
[56] J. Prakash, D. A. Tryk, E. B. Yeager, J. Appl. Electrochem. 1999, 29, 1463 – [72] S. Song, H. Zhang, X. Ma, Z. G. Shao, Y. Zhang, B. Yi, Electrochem.
1469. Commun. 2006, 8, 399 – 405.
[57] R. D. McKerracher, C. Ponce de Leon, unpublished results. [73] T. Ioroi, K. Yasuda, Z. Siroma, N. Fujiwara, Y. Miyazaki, J. Power Sources
[58] Y. Xie, H. Li, C. Tang, S. Li, J. Li, Y. Lv, X. Wei, Y. Song, J. Mater. Chem. A 2002, 112, 583 – 587.
2014, 2, 1631 – 1635. [74] T. Morimoto, K. Suzuki, T. Matsubara, N. Yoshida, Electrochim. Acta 2000,
[59] X. Sun, P. Song, Y. Zhang, C. Liu, W. Xu, W. Xing, Sci. Rep. 2013, 3, 2505. 45, 4257 – 4262.
[60] J. Suntivich, H. A. Gasteiger, N. Yabuuchi, H. Nakanishi, J. B. Goode- [75] Z. Y. Liu, J. L. Zhang, P. T. Yu, J. X. Zhang, R. Makharia, K. L. More, E. A.
nough, Y. Shao-Horn, Nat. Chem. 2011, 3, 546 – 550. Stach, J. Electrochem. Soc. 2010, 157, B906 – B913.
[61] F. H. B. Lima, J. Zhang, M. H. Shao, K. Sasaki, M. B. Vukmirovic, E. A. Ticia- [76] G. Q. Zhang, J. P. Zheng, R. Liang, C. Zhang, B. Wang, M. Hendrickson,
nelli, R. R. Adzic, J. Phys. Chem. C 2007, 111, 404 – 410. E. J. Plichta, J. Electrochem. Soc. 2010, 157, A953 – A956.
[62] V. R. Stamenkovic, B. Fowler, B. S. Mun, G. Wang, P. N. Ross, C. A. Lucas, [77] H. Meng, P. K. Sheng, Electrochem. Commun. 2006, 8, 588 – 594.
N. M. Markovic, Science 2007, 315, 493 – 497. [78] Y. Lee, J. Suntivich, K. J. May, E. E. Perry, Y. Shao-Horn, J. Phys. Chem. Lett.
[63] T. Poux, A. Bonnefont, G. Kranguven, G. A. Tsirlina, E. R. Savinova, 2012, 3, 399 – 404.
[79] M. Guo, X. Zhao, R. E. White, K. Huang, J. Electrochem. Soc. 2013, 160,
ChemPhysChem 2014, 15, 2108 – 2120.
A2085 – A2092.
[64] A. Grimaud, K. J. May, C. E. Carlton, Y.-L. Lee, M. Risch, W. T. Hong, J.
[80] M. Le Bars, M. Grae Worster, J. Fluid Mech. 2006, 550, 149 – 173.
Zhou, Y. Shao-Horn, Nat. Commun. 2013, 4, 2439.
[81] A. A. Shah, K. H. Luo, T. R. Ralph, F. C. Walsh, Electrochim. Acta 2011, 56,
[65] D. Pantea, H. Darmstadt, S. Kaliaguine, L. Summerchen, R. Christian,
3731 – 3757.
Carbon 2001, 39, 1147 – 1158.
[82] A. A. Shah, M. J. Watt-Smith, F. C. Walsh, Electrochim. Acta 2008, 53,
[66] X. Li, D. Pletcher, A. E. Russell, F. C. Walsh, R. G. A. Wills, S. F. Gorman,
8087 – 8100.
S. W. T. Price, S. J. Thompson, Electrochem. Commun. 2013, 34, 228 – 230.
[67] O. Kasian, T. Luk’yanenko, A. Velichenko, R. Amadelli, Int. J. Electrochem.
Sci. 2012, 7, 7915 – 7926. Received: July 23, 2014
[68] D. Pletcher, F. C. Walsh, Industrial Electrochemistry, 2nd ed., Chapman Revised: August 14, 2014
and Hall, London, 1990; D. Pletcher, F. C. Walsh, Industrial Electrochemis- Published online on September 26, 2014
try, 2nd ed., Softcover version, Blackie Academic & Professional, Glas-
Please note: Minor changes have been made to this article since its publica-
gow, 1993.
tion in ChemPlusChem Early View. The Editor.
[69] J. C. Rubim, J. Chem. Soc. Faraday Trans. 1 1989, 85, 4247 – 4258.

ChemPlusChem 2015, 80, 323 – 335 www.chempluschem.org 335  2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim