Volume 46 Number 2 21 January 2017 Pages 303–562

Chem Soc Rev

Chemical Society Reviews
rsc.li/chem-soc-rev

ISSN 0306-0012

REVIEW ARTICLE
Yi-Jun Xu, Hao Ming Chen et al.
Electrocatalysis for the oxygen evolution reaction: recent development
and future perspectives
 Chem Soc Rev
View Article Online
REVIEW ARTICLE View Journal | View Issue

Electrocatalysis for the oxygen evolution reaction:

recent development and future perspectives
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

Cite this: Chem. Soc. Rev., 2017,

46, 337
Nian-Tzu Suen,a Sung-Fu Hung,a Quan Quan,b Nan Zhang,b Yi-Jun Xu*b and
Hao Ming Chen*a

There is still an ongoing effort to search for sustainable, clean and highly efficient energy generation to
satisfy the energy needs of modern society. Among various advanced technologies, electrocatalysis for
the oxygen evolution reaction (OER) plays a key role and numerous new electrocatalysts have been
developed to improve the efficiency of gas evolution. Along the way, enormous effort has been devoted
to finding high-performance electrocatalysts, which has also stimulated the invention of new techniques
to investigate the properties of materials or the fundamental mechanism of the OER. This accumulated
knowledge not only establishes the foundation of the mechanism of the OER, but also points out the
important criteria for a good electrocatalyst based on a variety of studies. Even though it may be difficult
to include all cases, the aim of this review is to inspect the current progress and offer a comprehensive
insight toward the OER. This review begins with examining the theoretical principles of electrode
kinetics and some measurement criteria for achieving a fair evaluation among the catalysts. The second
part of this review acquaints some materials for performing OER activity, in which the metal oxide
materials build the basis of OER mechanism while non-oxide materials exhibit greatly promising
performance toward overall water-splitting. Attention of this review is also paid to in situ approaches to
Received 24th April 2016 electrocatalytic behavior during OER, and this information is crucial and can provide efficient strategies
DOI: 10.1039/c6cs00328a to design perfect electrocatalysts for OER. Finally, the OER mechanism from the perspective of both
recent experimental and theoretical investigations is discussed, as well as probable strategies for
www.rsc.org/chemsocrev improving OER performance with regards to future developments.

1. Introduction
a
Department of Chemistry, National Taiwan University, No. 1, Sec. 4, Roosevelt Rd,
Taipei 10617, Taiwan, Republic of China. E-mail: [email protected]
Since the first and second industrial revolutions, new inven-
b
State Key Laboratory of Photocatalysis on Energy and Environment, College of tions and technologies have brought great convenience to
Chemistry, Fuzhou University, Fuzhou, 350002, China. E-mail: [email protected] modern society. However, most advanced technologies rely on

Nian-Tzu Suen received his BSc and Sung-Fu Hung received his BSc and
MSc in Applied Chemistry from the MSc in Material Science &
National Chi Nan University with Engineering from National Tsing
Professor Hong-Yi Tang in 2006 and Hua University in 2008 and 2010.
2007, respectively. He obtained his He is currently a PhD student in
PhD degree from the University the Department of Chemistry,
of Delaware with Professor Svilen National Taiwan University, super-
Bobev in 2014 and is currently vised by Prof. Hao Ming Chen. His
working as a post-doctoral current research interests include
researcher in Professor Hao Ming the synthesis of semiconductor
Chen’s group at National Taiwan nanomaterials and interfacial
University. His research interests chemistry for photo- and electro-
Nian-Tzu Suen focus on the structure–property rela- Sung-Fu Hung catalysis.
tionship of materials and their
application in electrocatalysis.

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 337
 View Article Online

Review Article Chem Soc Rev

carbon-based fuels such as coal, natural gas, and petroleum.1

The depletion of these limited resources and their impact on
the environment have motivated a variety of researchers to
pursue alternative energy.2–4 Numerous innovative ideas have
been proposed for achieving more efficient energy conversion
or storage systems such as alkaline water electrolysis, fuel
cells, and metal–air batteries.5–11 The module design of these
systems may be complicated, but the essential ideas are
relatively simple. All of them are two-electrode systems, where Fig. 1 Polarization curves for HER (left) and OER (right). The Zc and Za are the
the cathode part involves hydrogen evolution reaction (HER) or
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

overpotentials for cathode and anode at the same current ( j), respectively.
oxygen reduction reaction (ORR) and the anode part proceeds
with oxygen evolution reaction (OER) or oxidation of some
chemical fuels. One of the important reasons that keeps these energy (higher overpotential) to overcome the kinetic barrier of
systems from being of practical use to date is the sluggish OER to occur. In the past several decades, the electrocatalytic
kinetics (high overpotential (Z), Za > Zc, Fig. 1) of the oxygen OER has been extensively studied and various catalysts have
evolution reaction. OER is a four electron-proton coupled been designed to improve electrode kinetics and stability under
reaction while HER is only a two electron-transfer reaction, different electrolyte environments. For instance, rutile-type
and hence it can be easily expected that OER requires a higher RuO2 exhibits excellent OER catalytic activity in both acidic

Quan Quan is now pursuing her Nan Zhang received her PhD
PhD degree under the supervision degree under the supervision of
of Prof. Yi-Jun Xu at the State Key Prof. Yi-Jun Xu at the State Key
Laboratory of Photocatalysis on Laboratory of Photocatalysis on
Energy and Environment, College Energy and Environment,
of Chemistry, Fuzhou University, College of Chemistry, Fuzhou
P. R. China. Her current research University, P. R. China. Her
interests include composite main research interests include
materials synthesis and their the fabrication of core–shell and
application in energy conversion carbon-based composites with
and storage such as photocatalytic potential target applications in
and photoelectrochemical redox heterogeneous photocatalysis.
Quan Quan processes. Nan Zhang

Yi-Jun Xu received his PhD degree Hao Ming Chen received his BSc
in heterogeneous catalysis at the and MSc from National Taiwan
School of Chemistry, Cardiff University in 2002 and 2004, and
University, U.K. in 2006 under the obtained his PhD degree in
supervision of Prof. Graham Chemistry from the National
J. Hutchings (FLSW FRS). From Taiwan University in 2008. He
2007 to 2009, he worked as a then worked as a postdoctoral
postdoctoral fellow at the Fritz fellow at the Department of
Haber Institute of the Max Planck Physics in National Taiwan
Society, Berlin, Germany. After University with Prof. Din Ping
that, he became a full professor Tsai and University of California
working at the State Key (Berkeley) with Prof. Peidong
Yi-Jun Xu Laboratory of Photocatalysis on Hao Ming Chen Yang. In 2013, he joined the
Energy and Environment, College Department of Chemistry at
of Chemistry, Fuzhou University, P. R. China. He is a Fellow of Royal National Taiwan University as an Assistant Professor. His current
Society of Chemistry (FRSC) and his current research interests research interests include the synthesis of nanomaterials, metallic
primarily focus on the assembly and applications of composite nanocrystals and semiconductor nanomaterials for solar energy
materials, such as graphene-based semiconductor composites, core– conversion, and developments of in situ methodology toward
shell composites and metal-based nanostructured materials in the relevant chemical reactions.
field of heterogeneous photocatalysis.

338 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

and alkaline electrolytes.12–15 However, RuO2 is highly unstable the same electrolyte are substantially lower among all electro-
in both electrolytes under high anodic potential, in which it will catalysts and thereby becoming a potential family for overall
be oxidized to form RuO4 and dissolved in solution.16 IrO2, water electrolysis.
another typical OER catalyst, shows good OER performance as Organometallics for use in OER have been studied since the
well. Nevertheless, IrO2 also suffers from a similar problem that early 1980s, and these compounds are suitable for studying
it will be oxidized to form IrO3 and dissolved during OER the mechanism owing to their tunable molecular/electronic
as well.17,18 The advantage of IrO2 is that it is more stable than structure. In this way, one could target and analyze specific
RuO2, and it can sustain at higher anodic potential in both effects toward the OER mechanism. However, their low electro-
acidic and alkaline electrolytes.19 However, both of them are catalytic activities for the OER usually discourage their develop-
made of precious metals which are not practical for large-scale ment in practice. Recently, significant progress has been revealed
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

production. by dispersing organometallics on specific substrates (graphene or

Substantial research effort has been devoted to investigating reduced graphene oxide),78 which can greatly improve their
the OER activity of metal oxide catalysts containing transition corresponding activities. In addition to those metal-based electro-
metals,20–38 and various metal oxide families such as perovskite catalysts, some non-metallic carbon based electrocatalysts have
(ABO3; A = alkaline- and/or rare-earth metals, B = transition also been investigated for OER with moderate OER activities.79–86
metals),20–23 spinel (A 0 B 0 2O4; A 0 = alkaline-earth and/or transition However, the benefits of non-metal systems still make this field
metals, B 0 = group 13 elements and/or transition metals),24–26,28 certainly worth exploring.
and the layer structure type-family (M(OH)2, MOOH and LiMO2; In recent years, more and more researchers have developed
M = Mn, Fe, Co and Ni)27,29–31 have been revealed to exhibit advanced methods and measurement techniques to alter the
remarkable activity toward OER. These transition metal-based intrinsic properties of materials or to probe the OER at the
oxides are low-cost, easily synthesized and environmentally atomic level in order to elucidate a detailed mechanism.29,31,50,65,87–99
benign. Furthermore, they are stable in alkaline solution and Through the application of nanotechnology, the size and the
display moderate conductivities, which make them act as good morphology of the compounds are tunable and even a multi-
candidates for electrocatalytic OER. In addition, these oxide sheet material is able to be exfoliated into a single-sheet
catalysts can possess rich combinations of transition metals conformation or re-arrange single-sheets with another material
in their structures especially for perovskite and spinel-type (graphene) to form an alternately stacked conformation.29,31,50,99
structures, since transition metals can have different oxidation This approach not only increases the number of active sites on
states and coordination environments (i.e., tetrahedral and the materials, but also sometimes changes their intrinsic
octahedral sites), which result in completely tunable OER properties such as electroconductivity, binding energy to the
behaviors. Furthermore, compounds belonging to the layer OER intermediate and so on. In addition, nowadays there are
structure type family have been extensively studied owing to many cutting-edge instruments that allow us to monitor the
their incredibly high OER activity. It is noted that the effect of OER in situ (operando) to confirm the actual active phase/
doping is very profound and an effective approach to promote intermediates and electrochemical behavior during OER on
the OER activity for the compounds in this family. Aside from the surface of the materials.87–89,91,93,94,97 These provide critical
the compounds in the three families, other metal oxides offer information for the OER mechanisms, which can further
good OER activity,32–38 and some of them are formed in situ support the mechanistic studies from experimental Tafel
and are suitable to perform OER in neutral electrolyte.32,39,40 analysis100–104 and theoretical density functional theory (DFT
Various efforts and systematic studies based on these metal calculation)105–110 viewpoints. The merging of in situ experi-
oxides have provided insightful observations that are helpful in ments, Tafel analyses and DFT calculations have established
understanding the fundamental mechanism of OER. On the the keystone foundation in the understanding of the OER
other hand, non-oxide catalysts like metal chalcogenides TC, mechanism, which sheds light on strategies to design electro-
TC2, T9S8, Ni3C2, T3S4 (T = Co and Ni, C = S, Se and Te),41–58 catalysts for OER.
metal pnictide Co2N, Co3N, Co4N, Ni3N, TP and T2P59–72 and Herein, we present the recent developments and break-
organometallics (metal coordination complexes)73–77 have also throughs both in experimental approaches and the theoretical
been reported as promising electrocatalytic materials for HER modeling of the electrocatalysis of the OER. We hope to provide
and/or OER, which could be good candidates for the overall a comprehensive review starting from the basic concepts of
water-splitting reaction. Nonetheless, as compared with metal electrocatalytic kinetics, including a brief discussion of the
oxide catalysts for electrocatalytic OER, only a handful of important electrochemical parameters such as overpotential
the cases for non-oxide catalysts for OER have been reported (Z), exchange current density (i0) and Tafel slope (b) that are
until recently. Metal chalcogenides have shown excellent OER commonly used to evaluate the performance of the catalysts,
activities with low overpotentials and good long-term stability in which will be addressed in Section 2. Electron transfer reactions
alkaline electrolyte. In terms of metal pnictides, they commonly for single-electron transfer and multiple-electron transfer reac-
suffer from slightly higher overpotential for OER compared to tions as well as the basic OER mechanism in acid and alkaline
that of metal chalcogenides. However, metal pnictides exhibit electrolytes are included. It is noticed that the reported values
amazing HER activities even in a strong alkaline electrolyte, in of these electrochemical parameters for the same materials can
which the combined overpotentials from both HER and OER in be very distinct from different research groups in some cases.

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 339
 View Article Online

Review Article Chem Soc Rev

This phenomenon is commonly observed due to different

sample histories (synthetic method, morphology and particle
size) and/or measurement conditions. With the aim of having a
fair assessment of the catalysts, some measurement criteria will
be addressed in this section as well. In Section 3, we will focus
on metal oxide catalysts for OER based on different structures
such as rutile (RuO2 and IrO2), perovskite, spinel, and the layer-
structure type family (LiMO2 and layer double hydroxides).
Non-oxide materials such as metal chalcogenide, metal pnictide,
organometallics and non-metal electrocatalysts that also dis- Fig. 2 (a) Polarization curves for cathode (red) and anode (blue) electrode.
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

j0 is the exchange current when ja = jc. j10 is a commonly agreed value for
played promising OER activities will be discussed in Section 4.
one to report overpotential (Z) at this current (assuming the area is 1 cm2)
Most importantly, we will present several in situ experimental value. (b) In the Z vs. log i curve, the exchange current density (i0) could be
results as examples to realize the actual active phase/intermedi- derived by extrapolating the curve to Z = 0. Slope b is the Tafel slope, where
ate during OER (Section 5); Tafel analysis, together with recently it is usually assumed that the steeper (larger) the slope is, the worse the
developed DFT calculations looking at the OER mechanism will electrode kinetics will be (slower current increased in the same given
potential comparing in the case of b1 and b2).
be associated with findings from in situ approaches to achieve a
comprehensive understanding of the OER mechanism. Finally,
we summarize all studies mentioned herein and provide a new the applied potential and E0 0 is the formal potential of the overall
perspective on material design for OER catalysis, which may be reaction. T denotes absolute temperature, R is the universal gas
helpful for other research fields (Section 6). constant, F is Faraday constant, n is the number of transferred
electrons in the reaction, and CO and CR are the concentrations
of oxidized and reduced reagents, respectively. The overpotential
2. Electrocatalytic kinetics, (Z) as illustrated in eqn (2.3) is a difference between the applied
measurement criteria and theoretical potential (E) and potential under equilibrium conditions (Eeq).
Notably, the overpotential (Z) is commonly referred to a value
modeling for oxygen evolution
that has to be applied to achieve a specified current density, and
reaction a lower overpotential of an electrocatalyst in the system indicates
2.1 Electrocatalytic kinetics its superior electrocatalytic ability for the target reaction. We
An electrocatalyst is a catalyst that is used to facilitate electro- have to note here that different current densities will be referred
chemical reactions (reactions involving charge transfer, such as to different overpotential (Z) values. It is important to state the
eqn (2.1)). It can either be modified on the surface of the current density of the reported overpotential (Z). Normally, the
electrode or act as the electrode itself. Generally speaking, the reported overpotential (Z) may be referred to a current density
main function of the electrocatalyst is to adsorb reactant on reaching to 10 mA cm2 (indicated as j10 in Fig. 2(a)).
the surface to form the adsorbed intermediate and thereby 0 RT CO
E ¼ E0 þ ln (2.2)
facilitate charge transfer between the electrode and reactant. nF CR
There are many electrocatalytic kinetic parameters such as
overpotential (Z), exchange current density (i0) and Tafel slope
Z = E  Eeq (2.3)
(b), which are utilized to fairly evaluate the performance of
electrocatalysts. These parameters are crucial and can offer 2.1.2 Exchange current density (i0). Another important
insightful information with regards to the mechanism of the indicator for electrocatalytic kinetics is the exchange current
electrochemical reaction. For this reason, before proceeding to density (i0). For a given reaction (2.1), the overall current ( j) is
the review of various electrocatalysts for OER, we provide the sum of anodic ( ja) and cathodic ( jc) currents (eqn (2.4)) and
a succinct introduction of these kinetic parameters in the the contributions from each anode and cathode ends are
following subsections. shown in eqn (2.5) and (2.6), respectively. ka and aa represent
the rate constant of the anodic-half reaction and anodic trans-
O + ne 2 R (2.1) fer coefficient, respectively. kc and ac have the same meaning in
the cathodic-half reaction.111
2.1.1 Overpotential (g). Overpotential (Z) is one of the most
important descriptors to evaluate the performance of target j = ja + jc (2.4)
electrocatalysts. In an ideal world, the applied potential for
driving a specific reaction should be equal to the potential of  
the reaction at equilibrium. In reality, it is always not the case aa nFE
ja ¼ nFka ½CR  exp (2.5)
that the applied potential commonly is much higher than that RT
at equilibrium in order to overcome the electrode kinetic  
barrier of the reaction. According to the Nernst equation,111 ac nFE
jc ¼ nFkc ½CO  exp  (2.6)
the applied potential can be expressed as eqn (2.2), where E is RT

340 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

At equilibrium (Z = 0; E = Eeq) conditions, the anodic ( ja) and constant), which implies good electrocatalytic kinetics. In
cathodic ( jc) currents are equal to each other, which results in a addition, Tafel slope (b) provides valuable and insightful
zero total net current (Fig. 2(a)). The magnitude of intercepts at information toward the mechanism of the reaction, especially
Z = 0 refers to exchange current ( j0), it is customary to divide for elucidating the rate-determining step. This can be very
exchange current ( j0) by the area of the electrode (A) and the helpful for understanding the fundamental behavior between
exchange current becomes exchange current density (eqn (2.7)). electrocatalyst and reactant. In the following subsection, we
will further demonstrate the strong correlation between Tafel
j0/A = i0 (2.7)
slope and reaction mechanism.
The magnitude of exchange current density (i0) reflects the intrinsic
2.2 Electron transfer reaction and mechanism of OER
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

bonding/charge transferring interactions between electrocatalyst

and reactant. A high exchange current density (i0) is usually a good The above-mentioned electrocatalytic parameters acquired
indication of being a good electrocatalyst for the target reaction. from experiment are important to assess the performance of
This argument has been supported by the fact that platinum (Pt) the electrocatalyst. Some electrocatalytic parameters such as
shows a very high exchange current density in HER systems among the Tafel slope even can reveal detailed information on reaction
many precious metals such as gold (Au), rhodium (Rh) and iridium mechanism. In this subsection, brief introduction for the
(Ir), which accounts for its superior electrocatalytic ability in HER. theory of single and multiple electron reactions is given and
Although the exchange current density (i0) is a good approach to the linkage between these theories and the Tafel slope will be
evaluate the electrocatalytic ability of a catalyst, it is a great presented herein.
challenge to directly find the exchange current density (i0) since 2.2.1 Single/multiple electron reaction. In a single-electron
we can only obtain overall current density (i) from the experiment transfer reaction, the transfer coefficient (a) usually refers to
(i = 0 when ia = jc = j0). One can still calculate the exchange current symmetry factor (b) and can be recognized as eqn (2.12). In
density from the Tafel equation (vide infra). most cases, the symmetry factor (b) is equal to 0.5 since the
2.1.3 Tafel equation and Tafel slope (b). For practical purposes, overpotential (Z) is usually much smaller than the re-organization
one is required to apply a high overpotential (Z) in order to have energy (l). If this assumption is valid, one can calculate the
significant magnitude of current density (i). In general, a smaller Tafel slope for a single electron reaction based on eqn (2.11)
overpotential (Z) with a faster increase in corresponding current and yield a value of 120 mV dec1. This implies that the rate-
density (i) is desired. The current density (i) and the applied determining step in the electrochemical system is controlled by
overpotential can be described according to the well-known the single-electron transfer step.
Butler–Volmer equation (eqn (2.8)).112
    1 Z
aa nFE ac nFE a¼b¼ þ (2.12)
i ¼ i0 exp þ exp (2.8) 2 l
RT RT
In a variety of electrochemical systems, the situation is much
From the Butler–Volmer equation and Fig. 2(a), under high more complex and electrochemical systems often consist of a
anodic overpotential conditions, the overall current is mainly series of consecutive reaction steps. These steps can be either
attributed to the anodic end while the contribution from the electron transfer steps or chemical steps such as association or
cathodic part is negligible. Accordingly, the Butler–Volmer dissociation reactions. The transfer coefficient, derived by
equation can be simplified as eqn (2.9) which is also known Bockris and Reddy,115 for a multiple-electron reaction is shown
as the Tafel equation.113,114 in eqn (2.13), where nb is the number of electrons that transfer
aa nFZ back to the electrode before the rate-determining step, and n is
i  i0 exp (2.9)
RT the number of rate-determining steps that have taken place in
the overall reaction. nr is the number of electrons that participate
By translating the Tafel equation to logarithm function form,
in the rate-determining step. Guidelli et al.116 have suggested
eqn (2.9) can be re-written as eqn (2.10), where the exchange current
that it is unlikely to have more than one electron transferred
density (i0) and Tafel slope (b) can be calculated accordingly.
simultaneously, and hence nb is either 1 or 0. Once the rate-
The Tafel slope (b) can be expressed as eqn (2.11), and from this
determining step is an electron transfer reaction, nb is equal to
one can understand that the definition of Tafel slope (b) is
1 while the value of nb is 0 in the case of chemical reaction.
‘‘how fast the current increases against overpotential (Z)’’ and
its value mostly depends on the transfer coefficient (a). nb
aa ¼ þ nr b (2.13)
Z n
logðiÞ ¼ logði0 Þ þ (2.10)
b Eqn (2.13) is simple yet very powerful for predicting the rate-
@Z 2:303RT determining step. For example, if the first electron transfer reaction
b¼ ¼ (2.11) is the rate-determining step, the values of both nb and n are
@ logðiÞ aF
equal to 0 while nr and b are 1 and 0.5, respectively. The transfer
In this regard, a smaller Tafel slope (b) as illustrated coefficient is calculated to be 0.5 and the corresponding Tafel
in Fig. 2(b) indicates that current density can increase faster slope is 120 mV dec1 (similar to a single electron transfer
with smaller overpotential (Z) change (i.e., faster reaction rate reaction). If the rate-determining step is the chemical reaction,

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 341
 View Article Online

Review Article Chem Soc Rev

after one-electron transfer reaction, the values of nb and n are difference, the common consensus is that the electrocatalysis
equal to 1 while the value of nr is 0. Consequently, the transfer of OER is a heterogeneous reaction, in which the bonding
coefficient is unity and the Tafel slope becomes 60 mV dec1. In interactions (M–O) within the intermediates (MOH, MO and
some systems such as OER (a four-electron transfer system), if MOOH) are crucial for the overall electrocatalytic ability.
the rate-determining step is the third electron transfer step, nb
2H2O - 2H2 + O2 (2.14)
and n are equal to 2 and 1 (nr and b are 0), respectively. This
yields a transfer coefficient of 2 and a Tafel slope of 30 mV dec1. Acidic conditions
Based on the above analysis, it is clear that different Tafel Cathode reaction:
slopes implicate different rate-determining steps. In a given
4H+ + 4e - 2H2 E0c = 0 V (2.15)
system, the transfer coefficient and the corresponding Tafel
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

slopes are closely correlated to the electrons involved in the Anode reaction:
reaction. In a consecutive reaction, a smaller Tafel slope
suggests that the rate-determining step is at the ending part 2H2O(l) - O2(g) + 4H+ + 4e E0a = 1.23 V (2.16)
of the multiple-electron transfer reaction, which is commonly a Proposed mechanism under acidic conditions
sign of a good electrocatalyst.
2.2.2 Mechanism of OER. In the case of the oxygen evolu- M + H2O(l) - MOH + H+ + e (2.17)
tion reaction, this is the half reaction of the water splitting
MOH + OH - MO + H2O(l) + e (2.18)
reaction (eqn (2.14)). It is noted that the reactions at the cathode
and anode parts for the water-splitting reaction are different 2MO - 2M + O2(g) (2.19)
under acidic (eqn (2.15) and (2.16)) and alkaline conditions
MO + H2O(l) - MOOH + H+ + e (2.20)
(eqn (2.22) and (2.23)). Many research groups have proposed
possible mechanisms for oxygen evolution reaction at the anode MOOH + H2O(l) - M + O2(g) + H + e + 
(2.21)
electrode for either acid (eqn (2.17)–(2.21)) or alkaline conditions
Alkaline conditions
(eqn (2.24)–(2.28)), and some disparities and similarities among
Cathode reaction:
these proposed mechanisms are present. Most of the proposed
mechanisms include the same intermediates such as MOH and 4H2O + 4e - 2H2 + 4OH E0c = 0.83 V (2.22)
MO, while the major difference is probably featured around the
reaction that forms oxygen. It is noted that there are two Anode reaction:
different approaches to form oxygen from a MO intermediate 4OH - 2H2 + 2H2O(l) + 4e E0a = 0.40 V (2.23)
(Fig. 3). One is the green route as illustrated in Fig. 3 through the
direct combination of 2MO to produce O2(g) (eqn (2.19)), and the Proposed mechanism under alkaline conditions
other one involves the formation of the MOOH intermediate
M + OH - MOH (2.24)
(eqn (2.20) and (2.27)) which subsequently decomposes to O2(g)
(black route in Fig. 3; eqn (2.21) and (2.28)). Despite this MOH + OH - MO + H2O(l)

(2.25)

2MO - 2M + O2(g) (2.26)

MO + OH - MOOH + e
 
(2.27)

MOOH + OH - M + O2(g) + H2O(l) (2.28)

2.3 Measurement criteria for OER

Various research groups have reported the electrocatalytic
activities of a great amount of samples based on conventional
electrochemical techniques. However, sometimes the published
data for the same material from different groups can be all
correct but very different.12 This is most likely because one group
may employ different measurement conditions from the others.
For this reason, it is necessary to establish a standard protocol of
measurement in order to make a fair comparison and evaluation
of the catalysts. McCrory et al.117 have proposed a standard flow
chart to combine many measurement techniques to ensure the
reliability of the result, where a lot of critical properties such
Fig. 3 The OER mechanism for acid (blue line) and alkaline (red line)
as elemental composition, surface area, Faradaic efficiency,
conditions. The black line indicates that the oxygen evolution involves the
formation of a peroxide (M–OOH) intermediate (black line) while another
catalytic activity and stability were suggested to be measured.
route for direct reaction of two adjacent oxo (M–O) intermediates (green) They also suggested that the environments for acid and alkaline
to produce oxygen is possible as well. electrolytes were 1 M H2SO4 and 1 M NaOH, respectively.

342 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

Moreover, the overpotential (Z) was suggested to be determined It has to be pointed out here that the evaluation of the
at a current density per geometric area of 10 mA cm2 at room electrocatalytic ability for specific catalysts should consider the
temperature under 1 atm O2, which is roughly the same current surface area and be corrected through the employment of cyclic
density at the anode part of a 10% efficiency of solar water- voltammetry (double-layer capacitance), especially for porous
splitting device under 1 sun illumination. substrates. For example, nickel foam is an excellent substrate
The flow chart from McCrory et al.117 covers many critical owing to its high surface area which leads to astonish electro-
measurements to evaluate the performance of the electro- catalytic activity once electrocatalysts are loaded on the nickel
catalysis for OER. Two measurements are worth further discus- foam. As shown in Fig. 4(a), at first glance, the performance of
sion, which are Faradaic efficiency (FE) and electrochemically Co3O4 material deposited on nickel foam is superior to that on
active surface area (ECSA) measurements. In the electrocatalysis nickel plate, FTO and a glassy carbon electrode (RDE). This
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

of OER, many studies report the value of overpotential only for result may lead to the conclusion that the Co3O4 loaded onto
electrocatalysts at a current density of 10 mA cm2, and assume nickel foam is much better than that of onto nickel plate and
the FE was 100%. Nevertheless, collected current densities in FTO. However, after the electrochemically active surface area
some cases may not contribute to the electrocatalysis of OER, (ECSA) has been corrected accordingly, the performance as
which results in overestimating the efficiency of the electrocata- illustrated in Fig. 4(b) of these samples is actually more or less
lysts for OER. This point could be very profound especially in a on the same scale. This commonly gives rise to misleading
high oxidizing potential situation while the electrocatalysts or observations in various studies.
substrates (carbon-based materials) suffered from phase trans- In addition to these measurement criteria, we should also
formation or decomposition. Therefore, it is crucial to carry out pay attention to the structural evolution of the catalyst during
the measurements of Faradaic efficiency either from H2 and O2 OER. Recently, it has been suggested by many research groups
probe as McCrory et al.117 suggested or using GC measurements that the active phase for performing OER is different from
to detect the produced O2 gas and calculate the Faradaic the original prepared samples due to occurrence of a phase
efficiency of O2 evolution. transformation during OER;51,94,95,118 this is understandable
since the OER is carried out under very high anodic potential.
Therefore, a situation where structural information of the electro-
catalysts during the reaction is lacking might lead to drawing a
conclusion based on the wrong information.

3. Oxide electrocatalysts for OER

3.1 Precious metal oxides, RuO2 and IrO2
These two precious metal oxides adopt a rutile structure, where
Ru and Ir locate in the center of octahedral site with oxygen in
the corners, each octahedron connects to each other through
sharing the corners, and a representative structure of RuO2
is shown in Fig. 5(a). Both RuO2 and IrO2 are commonly
considered as benchmark electrocatalysts for OER owing to
their high electrocatalytic activities toward OER in both acidic
and alkaline solution.12,13 However, as shown in Fig. 5(b), the
OER performances of RuO2 and IrO2 are greatly influenced by
the method of sample preparation. For instance, IrO2 film
exhibited excellent OER activity and can reach 0.1 Jgeo mA cm2

Fig. 4 The uncorrected (a) and corrected (b) electrochemically active Fig. 5 (a) Crystal structure of RuO2. (b) OER performance of Ru and Ir
surface area (ECSA) for linear-sweep voltammetry (LSV) curves of Co3O4 based electrocatalysts in acid electrolyte from different groups. Repro-
on nickel foam (orange), FTO (red), nickel plate (blue) and glass-carbon duced with permission from ref. 13. Copyright r 2014, Wiley-VCH Verlag
(RDE, green). GmbH & Co. KGaA, Weinheim.

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 343
 View Article Online

Review Article Chem Soc Rev

below Z = 0.275 V,15 while the RuO2 nanoparticles showed only

0.01 Jgeo mA cm2 at the same overpotential. As we mentioned
earlier, RuO2 is less stable than IrO2 during OER. Cherevko et al.19
have demonstrated that RuO2 is unstable under high anodic
potential and is going to deteriorate (dissolve) during OER, while
IrO2 is a more stable catalyst for OER than RuO2. The decom-
position mechanisms for RuO2 and IrO2 have been proposed.
That is, (Ru4+)O2 will transform into the hydrous compound
RuO2(OH)2 and deprotonate into a high oxidation state (Ru8+)O4
under anodic conditions,16,18 which is not stable in electrolyte
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

and can be further dissolved into solution (also accompanied

with a color change), thus leading to the deterioration of RuO2
catalysts. A similar case for IrO2 has been clarified as well.
A high oxidation state (Ir6+)O3 will be formed under high anodic
potential and will be further dissolved in the electrolyte.18
Although a similar concern in stability is present in the OER
of IrO2, many researchers have concluded that the OER perfor-
mance of IrO2 is slightly worse but much more stable than that
of RuO2. To improve the stability of RuO2, a doped bimetallic Fig. 6 (a) Crystal structure of a perovskite (ABO3). (b) The correlation
between applied potential (E) and eg electron for each perovskite
oxide system RuxIr1xO2 was proposed.119–122 This system
compound at a current density of 50 mA cm2. Reproduced with permis-
turned out to be very effective and only a small amount of sion from ref. 22. Copyright r 2011, American Association for the
Ir was incorporated into the sample, which could significantly Advancement of Science. (c) The occupation of d states of eg and t2g
suppress the deterioration without sacrificing much perfor- symmetry and the correlation to absorption energy of oxygen (eV).
mance of OER. Another route to enhance their electrocatalytic Reproduced with permission from ref. 23. Copyright r 2011, American
Association for the Advancement of Science. (d) Schematic molecular
stability is to synthesize a core–shell structure (IrO2@RuO2) and
orbitals for [Fe4+O6]. Reproduced with permission from ref. 125. Copyright
once subjected to OER electrocatalysis. It has been elucidated r 2015, Nature Publishing Group.
that this core–shell conformation can not only lower the over-
potential (B300 mV) but also increase the stability (1000th
cycling with 96.7% remaining).123 mechanism path, while the higher oxidation state of Co
could also enhance the reaction rate of the second step.20 This
3.2 Perovskite family pioneering work is the first demonstration to correlate the
Perovskites are a large family with the general formula ABO3, bonding interaction of the catalyst with OER activity. Bockris
where A is alkaline- and/or rare-earth metal and B is a transi- and Otagawa also reported a very thorough and extensive OER
tion metal. As illustrated in Fig. 6(a), the transition metal B study based on a perovskite system,21 where they discovered a
resides in the center of the octahedron and each octahedron trend (Ni > Co > Fe > Mn > Cr) of OER activity for samples
joins each other by sharing the corner to make the backbone of containing different transition metals. This trend also referred
the structure, where A can fill in the open space and complete to the Tafel slopes of the samples, while Ni, Co and Fe/Mn
the perovskite. Generally speaking, the oxidation state of transi- samples showed a similar trend on the Tafel slopes with values
tion metals (B) in a perovskite can be either 3+, 4+ or a mixture of of 40 mV dec1, 60 mV dec1 and 120 mV dec1, respectively.
3+ and 4+ depending on the oxidation state of A. Furthermore, they reported that the Tafel slopes might change
For instance, it can be (A12+)(B4+)O3, (A23+)(B3+)O3 or under higher overpotential conditions (e.g., from 70 mV dec1
(A1 )0.5(A23+)0.5(B3.5+)O3. By doping different amounts of
2+
to 130 mV dec1 for LaCoO3), which suggested that the rate-
alkaline-earth metal (A1) and rare-earth metal (A2), this allows determining steps changed at different applied potentials. They
us to design and to have desired oxidation states of transition proposed that the strength of B–OH bonding in perovskite
metals in the perovskite structure. Numerous research groups samples has great influence on their OER activity, in which
have conducted systematic studies of the perovskite for the B–OH bond strength was dominated by the electron popu-
OER and found interesting results. Matsumoto et al.20 have lations in the s* band with decreasing from Ni to Fe samples
demonstrated a thorough study on a La1xSrxFe1yCoyO3 and further corroborated with the experimental results of their
system and pointed out that the OER activity would increase OER activities.35
with increasing amount x and y. They attributed this effect to In recent years, Suntivich et al.22 reported that many perovs-
the band distribution of the d character and the higher oxida- kite samples containing different transition metals correlated the
tion state of the Co ion, suggesting that the OER electrocatalytic applied potentials at a specific current density to eg electrons.
activity was strongly related to the electrons in the d band of the In Fig. 6(b), the Ba0.5Sr0.5Co0.8Fe0.2O3d (BSCF) sample dis-
perovskite. They also indicated that the reaction rate of first played the best OER performance, which was superior to that
and third steps would be facilitated once the s* band width of LaNiO3 and LaCoO3. In terms of this observation, they drew
increased (i.e., higher x and y values) in the proposed OER two directions to design a good OER catalyst based on

344 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

their data: (1) the eg electrons should be close to unity and (2) than that of a perovskite. There are two crystallographic sites
stronger covalent interactions between transition metal and O including octahedral (Oh) and tetrahedral (Td) sites in the structure
atoms was better. Moreover, the DFT calculations carried out by which thereby lead to two types of spinel. One is normal spinel



Vojvodic and Nørskov23 revealed that the electrons on both eg with a configuration of A2þ Td B3þ Oh 2 O4 (divalent cation A2+)
and t2g resulted in the contribution to the oxygen adsorption



and the other is inverse spinel of A2þ Oh B3þ Td B3þ Oh O4
energy (Fig. 6(c)). According to the Sabatier principle, either too (divalent cation A2+ in octahedral site). As compared to the
strong (i.e., LaCuO3) or too weak (i.e., LaMnO3) adsorption transition metals in the perovskite, the transition metals in
energy is not appropriate for catalysis reactions, and moderate spinels can exhibit tetrahedral and octahedral coordinations,
conditions in between (LaCoO3 and LaNiO3) should be suitable. which results in different d-band splitting (Td: egt2g; Oh: t2geg)
This suggestion further explained a common observation that Co
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

according to crystal field theory.127

or Ni based oxides exhibited good catalytic activities toward OER. Spinel compounds are good electric conductors and
The unity of the eg electron turns out to be a very good very stable in alkaline solution under high anodic potential,
indication of explaining the OER activity for perovskite com- indicating that they appropriately act as electrocatalysts for
pounds. For instance, Lee et al.124 have suggested that phase OER. Most spinel oxides for OER are iron-based (ferrite) and
transformations occurred between BaNiO3 and BaNi0.83O2.5 cobalt-based compounds with other transition/alkaline metals
samples, which led to a fact that the electron on eg was zero dopants such as Mn, Ni, Cu, Zn and Li.26,28,128,129 In the case of
for BaNiO3 and was close to unity for BaNi0.83O2.5 and further the ferrite system, Li et al.28 have reported a trend of CoFe2O4 >
explained the superior OER activity of BaNi0.83O2.5. In the study CuFe2O4 > NiFe2O4 > MnFe2O4 for MFe2O4 (M = Co, Ni, Cu, Mn)
of Yagi et al.,125 they revealed that the molecular orbital of samples with regards to the OER electrocatalytic activities. In
[Fe4+O6] has one electron only in the eg orbital which thereby addition, N. K. Singh and R. N. Singh,25 and Al-Hoshan et al.130
results in good OER activity (Fig. 6(d)). The authors have also have shown that the incorporation of Ni in NixFe3xO4
suggested that increasing the Fe4+ population through oxygen ferrite compounds would effectively improve their OER perfor-
vacancies in La1xFeO3d could enhance its OER activity. mance. Nevertheless, the situation was slightly different in the
Indeed, both SrFeO3 and CaFeO3 samples exhibited very good cobalt-based system: introduction of Ni, Cu and Li in a
OER activities, but their performance rapidly decayed. Incor- MxCo3xO4 (M = Li, Ni, Cu) system would be beneficial to their
poration of Cu to substitute for Sr and Ca could stabilize Fe4+ OER activities except for Mn case,26,131 suggesting that the
via covalent bonding interaction during OER,125 where the incorporation of Mn in cobalt-based spinels would be harmful
synthesized CaCu3Fe4O12 sample displayed excellent stability for their OER activity. This phenomenon might be attributed to
in OER. The argument that Fe4+ in the sample could enhance the suppression of the Jahn–Teller distortion,132 which has
OER activity was also reported by Zhu et al.,126 suggesting the been suggested by a similar observation that the Mn3xCoxO4
superior OER activity of La1xFeO3d compared to that of (0 r x r 1) system could perform better activities with
LaFeO3 is due to the increasing Fe4+ population through oxygen decreasing Mn amount.132
vacancies in La1xFeO3d. Consequently, in various studies, spinel bimetallic oxides
have been demonstrated to perform better OER activities
3.3 Spinel family as compared to monometallic spinel oxides, which suggested
More than 100 compounds in the spinel family have been that different metal ions in spinel metal oxides might play
reported to date. The general formula of a spinel is A 0 B 0 2O4 different roles during OER. By replacing either Co2+ ions with
where A 0 and B 0 atoms usually are made up of group 2, group 13 Zn2+ ions or Co3+ ions with Al3+ ions to form ZnCo2O4 and
and first-row transition metal elements. The structure of a CoAl2O4 samples, Wang et al.128 demonstrated a systematic
spinel, as illustrated in Fig. 7(a), is slightly more complicated investigation to reveal the individual roles of Co2+ ions and

Fig. 7 (a) Crystal structure of spinel (AB2O4). (b) Cyclic voltammetry (CV) curves of Co3O4, ZnCo2O4 and CoAl2O4. (c) Tafel slopes for Co3O4, ZnCo2O4,
and CoAl2O4. Reproduced with permission from ref. 128. Copyright r 2016, American Chemical Society.

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 345
 View Article Online

Review Article Chem Soc Rev

Co3+ ions during OER. Their observations clearly revealed that structural transformation would occur on spinel metal oxide and
Co2+ ions and Co3+ ions were differently responsible for OER further produce a metal oxyhydroxide phase over the surface of
activity in the Co3O4 system and confirmed that the divalent catalysts,94,96 and these results have inspired many researchers
Co2+ dominated the OER activity (Fig. 7(b)).128 For this reason, to consider the layer structure-type oxide systems (Section 3.4)
OER activity of CoAl2O4 (with presence of Co2+) was similar to toward OER.
that of Co3O4 (with presence of Co2+) and was superior to that
of ZnCo2O4 (with the presence of Co3+) (Fig. 7(c)).128 These 3.4 Layer structure type family
studies conclude that the incorporation of specific metal ions The compounds belonging to this family usually possess good
in spinel bimetallic oxides allowed us to modify the electronic electric conductivity, and have been used in alkaline batteries
structure of catalysts and to further optimize the binding con- since the 1970s. Layer structure type oxides including metal
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

ditions between the intermediates and reactants during OER. hydroxides (M(OH)2; M = transition metals) and metal oxyhydr-
Various studies have also suggested that a strong coupling oxides (MOOH) commonly exhibit very high electrocatalytic
(synergistic effect) interaction between electrocatalysts and activity toward OER. The crystal structure is a typical layered-
substrates is a promising approach for promoting electrocata- stacking conformation with proton (H+) sandwiching between
lytic performance of OER. For example, Koza et al.133 suggested layers, where the transition metals are located in the center of
that electrodepositing Co3O4 film on a Au substrate (100) would the octahedron with oxygen anions on the corners as [MO6]
lead to Co3O4 grown along the (111) orientation, and it has subunits and these subunits further connect to each other via
been suggested by Chen et al.134 that Co3O4 with a (111) sharing the edges to construct 2-D layers (Fig. 8(a)). Consequently,
orientation could expose a higher density of Co ion to the layer structure-type oxides are classified as layered double
electrolyte and thereby result in a higher OER activity. In hydroxides (LDH), and this family is usually cobalt- and nickel-
addition to Au substrate, there are numerous reports revealing based compounds because of their excellent activity toward OER.
that the OER activities of spinel compounds can be improved Notice that there are many different types of metal oxyhydroxides.
through the dispersion of spinel electrocatalysts onto carbon- Taking b-CoOOH and g-CoOOH as examples, the major differ-
based substrates such as single-walled carbon nanotubes ence between b- and g-form is different intercalation species
(SWCNTs), carbon nanosheets (CNS) and graphene. In a report that causes a disparity of d-spacing among layers. In b-CoOOH
by Al-Mamun et al.,135 they have shown that the electrocatalytic (dehydrated), there are protons (H+) among layers with a
efficiency of the CoCr2O4 sample on carbon nanosheets out- d-spacing of around 4.6 Å and oxidation state of Co is trivalent.
performed that of RuO2. A similar result was reported by In the case of g-CoOOH (hydrated), water molecules and first
Ma et al.136 that the hybrid Co3O4–carbon porous nanowire arrays alkali group elements (Na+ or K+) reside between layers, in
(Co3O4 C-NA) displayed excellent OER and ORR activities, and which the d-spacing substantially increases to approximately
also indicated that catalytic activity of Co3O4 C-NA was superior to 7 Å and the oxidation state of Co is above trivalent.
that of physically mixed Co3O4 nanowire arrays with carbon. This With increasing research interest in water-splitting, these
strongly suggests that increasing electrical conductivity may not LDH compounds have been investigated and demonstrated to
be the only reason to account for the better catalytic performance exhibit remarkable OER activity. Subbaraman et al.27 have
of Co3O4 C-NA but may also be the interaction (synergistic effect) carried out a systematic study of 3d transition-metals (Ni, Co,
between Co3O4 and carbon species. Fe, Mn) hydr(oxy)oxide for water electrolyser reaction and
The aforementioned synergistic effect is particularly impor- discovered the OER activity of those samples, and found that
tant toward multicomponent composites in oxide electro- the behavior followed the same trend (Ni > Co > Fe > Mn) as
catalysts due to their inherently low electric conductivities.137 observed in perovskite systems.21 They suggested that the
As detailed and pointed out in a recent review by Ma et al.,138 there superior OER activity for the NiOOH sample was attributed to
are numerous advantages of these host–guest type composites the optimal bonding strength between Ni and OH (Sabatier
(self-supported electrocatalysts). These composited electro- principle). However, Corrigan139 has pointed out that the
catalysts bear large electroactive surface areas, and have efficient unintentionally introduced Fe impurity (1 ppm) in the electro-
electron-transfer within their structure and strong adhesion lyte (unpurified) would significantly enhance the OER activity
between electrocatalysts and substrates. All the above-mentioned of NiOOH and its activity further increased with increasing Fe
effects are essential factors to dominate the efficiency and content in NiOOH sample. The influence of Fe ions in NiOOH has
durability of electrocatalysts. Accordingly, as we proceed with been extensively investigated by Trotochaud et al. as well.140 They
discussion in other sections, one can realize that this strategy suggested a similar conclusion that the electrocatalytic ability of
(multicomponent composite) can be of benefit not only for NiOOH for OER was very sensitive to the presence of Fe ions in the
oxide electrocatalysts but also for other electrocatalysts such as system (Fig. 8(b)). The overpotential of Ni1xFexOOH significantly
metal chalcogenides, metal pnictides and organometallics. dropped by a value of 200 mV when the Fe content reached 25%.
Another important issue worth noting here is the phase This improvement, however, cannot be explained solely by the
transformation in spinel during OER. As mentioned earlier, we slight increase in conductivity due to co-precipitation with Fe.
should consider the structural evolution of the catalysts during In the Ni(OH)2 and NiOOH system, Bode et al.141 have demon-
OER, which is operated under very high anodic potential. strated the phase transformation behavior during charge/discharge
Recently, several research groups have revealed that a remarkable in alkaline solution. The b-(Ni2+)(OH)2 would be oxidized and

346 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

Fig. 8 (a) Crystal structures of LDH b-CoOOH and g-CoOOH. (b) Cyclic voltammetry measurements for different Fe content in Ni1xFexOOH and
measurements for NiOOH in purified and non-purified electrolytes. Reproduced with permission from ref. 140. Copyright r 2015, American Chemical
Society. (c) Schematic illustration of Bode scheme. Reproduced with permission from ref. 142. Copyright r 2015, American Chemical Society. (d) OER
performance of exfoliated NiFe-, NiCo and CoCo-NS (nanosheet) and non-exfoliated NiFe-, NiCo and CoCo-B (bulk). Reproduced with permission from
ref. 31. Copyright r 2015, Nature Publishing Group.

lose a proton (H+) to become b-(Ni3+)(OOH3), as shown As mentioned above, the Ni1xFexOOH material has shown
in Fig. 8(c).142 A similar situation also occurred between great potential for OER. Song and Hu have successfully pre-
a-(Ni2+)(OH)2 and g-(Ni3+)(OOH3), where the a-Ni(OH)2 would pared single-layer nanosheets (NS) of NiFe LDH material
endure an aging effect and transform into b-Ni(OH)2 slowly in through exfoliation process.31 NiFe-NS exhibited an overpoten-
alkaline solution. Under long-term overcharging, b-NiOOH tial of B300 mV with a Tafel slope of B40 mV dec1 for OER,
(dehydrated) was also transformed into g-NiOOH (hydrated). whose performance was comparable to precious metal oxides
As Trotochaud et al.140 and Klaus et al.142 suggested, the OER such as RuO2 and IrO2 (Fig. 8(d)). Long et al.145 have reported
activity of Ni(OH)2 would increase in unpurified KOH while it the OER activity of FeNi LDH by inserting graphene oxide (GO)
would decrease in a purified KOH electrolyte. The improvement and reduced graphene oxide (rGO) layers between metal oxide
in OER activity after cycling was commonly attributed to the layers to form a unique alternated stacking conformation.
phase transformation of a-Ni(OH)2 to b-Ni(OH)2 (so-called These hybrids of FeNi–GO LDH and FeNi–rGO LDH have
aging effect) and led to a conclusion that the OER activity of displayed incredibly high OER activities with overpotentials
b-NiOOH was more active than that of g-NiOOH. However, as (Z) were 210 mV and 195 mV, respectively. A similar effect of
indicated by Trotochaud et al. and Klaus et al.,140,142 the Fe doping in the CoOOH LDH system has been explored and
increase in OER activities for Ni(OH)2 and NiOOH systems revealed that doping of Fe could work as well to improve
were likely due to the existence of Fe impurities in the system, its OER activity.146 Introducing foreign elements in Co- and
which changed the long-term viewpoint of the OER activities Ni-based LDH to elevate OER activity is not exclusive for Fe;
for g-NiOOH and b-NiOOH. DFT calculations were performed other metals such as Mn and Cr dopants have been proved to
with the aim of understanding the roles of Fe and Ni in show a similar effect for improving the OER activity of
the Ni1xFexOOH system during OER,143,144 which indicated NiOOH.147 Recently, Zhang et al. have shown that the perfor-
that the incorporation of Fe would lower the overpotential mance of b-CoOOH for OER was optimized through including
of Ni1xFexOOH during OER while the Fe acted as the active W and Fe via sol–gel method.148 The adding of Fe and W in
site responsible for performing OER. Li and Selloni further G-FeCoW LDH (gelled) could alter the adsorption energy of
illustrated that OER activity of b-NiOOH (Z = 0.26 V) under Co–OH to achieve the most appropriate energy state and
the influence of Fe impurity was better than that of g-NiOOH further lower the overpotential down to 190 mV for OER.
(Z = 0.48 V), which was in agreement with the reported In addition to Co- and Ni-based LDH, the g-FeOOH was also
experimental result.143 suggested to be a potential electrocatalyst for OER, since Fe was

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 347
 View Article Online

Review Article Chem Soc Rev

considered to act as the active site in both Co(Fe)OOH and

Ni(Fe)OOH.144,146 However, the OER electrocatalytic activity of
pure g-FeOOH was very poor,149,150 in which the OER activity of
g-FeOOH would dramatically increase at overpotential (Z) above
450 mV. Under this high anodic potential, the electronic
conductivity of g-FeOOH increased. Despite the superior per-
formance of g-FeOOH at high anodic potential, it was shown
that the long-term durability decreased due to the formation
of FeO42, which would dissolve in the electrolyte.149,150 This
implies that the inferior OER activity of g-FeOOH can be
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

attributed to the intrinsic poor electronic conductivity and

instability during OER. These drawbacks could be overcome
by hybridizing FeOOH and CeO2 to establish a heterolayered
Fig. 9 (a) Contour plots for onset overpotential (Z) (left), Tafel slopes
nanotube conformation on the nickel foam substrate, and this (middle) and overpotential (Z) at 0.5 mA cm2 (right) for reported 21 metal
novel architecture improved the long-term durability as well as oxide films. Reproduced with permission from ref. 153. Copyright r 2013,
the overpotential (Z) of OER.151 American Chemical Society. (b) Proposed OER mechanism for CoPi.
In the layer structure type family, LiMO2 (M = Co and Ni) Reproduced with permission from ref. 159. Copyright r 2013, American
Chemical Society.
compounds were also investigated for OER activity. The struc-
tures of the LiMO2 family were mostly the same as that of their
oxyhydroxide form (i.e., b-CoOOH and b-NiOOH) while the connects to six oxygen anions [FeO6] while phosphorous is in a
proton (H+) was replaced by Li cation. Most of the conclusions tetrahedral environment with oxygen anions in the corner [PO4].
drawn above were also applied to these compounds. For instance, It is worth mentioning that the octahedrons connect to each
the doping of Fe in LiNi1xMxO2 (M = Mn, Fe, Co) compounds other by sharing the corner, which is different from most of the
could increase their OER performance.152 The same observation octahedra mentioned herein (which share the edge). In spite of
was also confirmed in a delithiatized LiCoO2 system, in which this divergence, the doping effect seems to have a similar influ-
the delithiatized LiCoO2 samples could be synthesized through ence on the Li(FeM)PO4 (M = Mn, Co and Ni) system, in which
an electrochemical tuning route.30 In this case, the delithiated Li(FeNi)PO4 showed the best OER performance among these
samples exhibited better OER activity compared to un-treated studied compounds and the inclusion of Co slightly improved
ones, and also found that a higher oxidation state of Co could its OER activity. Similar to the case of LiMO2 (M = Co and Ni),30
perform higher OER activity owing to the enhancement of the the oxidation state of the transition metal can be increased by
electrophilicity to adsorb O and facilitate the adsorption of extracting lithium out of the structure (delithiatization) which
OH anion to form OOH intermediate (rate-determining step) on further promoted the OER activity of the De-Li(NiFe)PO4 sam-
the surface.30 ple (Z B 270 mV at 10 mA cm2).
The metal oxides discussed above commonly exhibited
3.5 Other oxides remarkable activity in alkaline electrolyte, yet they may be
A variety of the above-mentioned research has indicated that inactive in neutral electrolyte. For some specific purposes, it
doping transition metals in electrocatalysts with an appropriate is essential to develop OER electrocatalysts in neutral solution
amount can enhance the OER activity of electrocatalysts. For (pH B 7).32,34,37,39,154,155 For this reason, CoPi has drawn great
example, in an amorphous system, a detailed investigation of attention owing to its excellent OER activity in neutral solution.32,33
this subject was carried out as shown in Fig. 9(a), in which they Kanan and Nocera have reported the in situ formation of CoPi
synthesized and evaluated the OER activities of 21 amorphous during OER in phosphate buffer solution containing Co2+ ions.
compounds with various stoichiometry ratios including Fe, The in situ formed CoPi displayed an overpotential (Z) of
Co and Ni (a-Fe100yzCoyNizOx).153 The binary a-Fe40Ni60Ox 410 mV to reach a current density of 1 mA cm2 and possessed
catalysts displayed the most excellent electrocatalytic kinetics a self-healing property that could bring on a long-term
for OER among all samples, while further increasing the stability.32 These promising qualities of CoPi have triggered
amount of Fe was counterproductive toward the OER activity. many investigations into its fundamental properties such as
The influence of incorporating Fe in a-CoOx was similar to that crystal structure, composition and mechanism toward OER.
of a-NiOx, and further doping of Co into binary a-Fe40Ni60Ox Risch et al.156 and Kanan et al.157 have analyzed the local structure
to form ternary a-Fe39Ni39Co22Ox would not improve its OER of CoPi from in situ X-ray absorption spectroscopy (XAS). In the
activity.153 Fourier-transformed extended X-ray absorption fine structure
The doping effects of Fe, Co and Ni in a given oxide system (FT-EXAFS) refinements, the refined number of Co–O (1.89 Å)
(spinel and layer structure type family) seem to be universal and scattering path was close to 6, which indicated an octahedral
can be extended to other systems. For example, a LiFePO4-based coordination environment at Co. Several possible structural
material replacing Fe with Mn, Co and Ni was employed as an models for CoPi have been proposed and suggested that the
electrocatalyst for OER.35 LiFePO4 adopts an olivine structure model consisted of edge-sharing [CoO6] motifs in molecular
type, where Fe is located in the center of the octahedron and dimension (similar to CoOOH but at short range).157 They have

348 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

also shown that the oxidation state of Co in CoPi was increased In addition to these MnOx–Pi and Co–Pi systems, many
above 3+ at an applied potential of 1.25 V (vs. NHE). McAlpin electrocatalysts with well-defined stoichiometries and extended
et al. also observed the same phenomenon through ex situ structures such as Bi2WO6,37 Mn5O834 and Co(PO3)2158,161 showed
electron paramagnetic resonance (ex-EPR) measurement. The comparable or superior OER activity to benchmark Co–Pi in a
oxidation state of Co in CoPi would be elevated and further neutral system. Dincă et al.154 and Bediako et al.40,162 have reported
suggested the presence of Co4+ in CoPi during OER.158 This a Ni–borate (Ni–Bi) system that exhibited the most superior electro-
indicated that the population of Co4+ increased and clarified the catalytic kinetics of OER (Tafel slopes in pH B 7; 120 mV dec1 for
rest state should be a mixture of Co3+ and Co4+ during OER. MnOx–Pi, 60 mV dec1 for Co–Pi and 30 mV dec1 for Ni–Bi).
Surendranath et al. have done a thorough investigation into the Similar to the Co–Pi system, the structure of as-deposited Ni–Bi
OER mechanism with regards to the CoPi system (Fig. 9(b)).159 film resembled b-NiOOH and transformed to a short-range
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

Several key conclusions can be drawn from their experimental g-NiOOH like structure (elevated oxidation state of Ni and
results: (1) phosphate buffer electrolyte plays an essential role as diminished Jahn–Teller distortion) during the anodization
a proton acceptor, while the proton transfer step is not the process.40 The steady-state Tafel slopes were 30 mV dec1 and
turnover-limiting step; (2) steady-state Tafel slope of CoPi shows 100 mV dec1 for non-anodized Ni–Bi (b-NiOOH like structure)
a value of 60 mV dec1 regardless of the thickness of the and anodized Ni–Bi (g-NiOOH like structure), respectively. It
deposited films, which indicates that it undergoes an one was worth mentioning here that this anodization process was
electron/proton step and follows the turnover-limiting chemical similar to the ‘‘overcharge effect’’ (Bode scheme).141,142 The
step (possibly the O–O bond formation to produce O2). It is phase of Ni–Bi transformed from b-NiOOH like to g-NiOOH like
interesting to note here that most of the Co-based hydroxide/ structure and exhibited a superior OER activity to support the
oxydroxides and Co3O4 exhibit a similar range of Tafel slopes (60 aforementioned finding in a-Ni(OH)2/g-NiOOH and b-Ni(OH)2/
mV dec1), which implies the mechanisms of these electrocata- b-NiOOH systems. The corresponding electrocatalytic para-
lysts for OER may be identical. meters of these selected oxide materials from precious oxides,
Inspired by the above work, numerous electrocatalysts con- perovskite family, spinel family, layer structure-type family and
taining mid first-row transition metals such as Mn, Fe and oxide materials mentioned in this section have been tabulated
Ni have been synthesized in buffer solution (pH B 7).39,154,160 and are listed in Table 1.
It appears that the mechanisms for these transitional metal-
based oxides are different from each other and the mechanisms
may be different in acid and alkaline electrolyte. For instance, 4. Non-oxide electrocatalysts for OER
d-MnO2 was electrodeposited on FTO substrate and subjected
to the measurement of corresponding OER activity in electro- 4.1 Metal chalcogenides
lytes with a pH range from 3 to 13.155 In situ UV-vis spectra The above-mentioned metal oxides (or hydroxides/oxyhyrdroxides)
indicated the emerging absorption peak at 510 nm during have shown excellent electrocatalytic abilities for OER, and moving
water oxidation was confirmed and attributed to Mn3+ from oxides to other chalcogenides (sulfide, selenides or telluride)
from the surface intermediates. The generated Mn3+ species would increase the intrinsic electroconductivity of the materials.
were unstable in acidic solution and would undergo a self- Most of the metal chalcogenides for the water-splitting reaction
disproportionation reaction to produce Mn2+ and Mn4+. It is are cobalt- and nickel-based compounds. These compounds
possible for Mn2+ to dissolve in electrolyte before oxidizing were originally utilized for HER due to their excellent electro-
back to Mn3+, which was the reason to account for its instability conductivities (fast charge transfer ability) and chemical resis-
in acidic electrolyte. Under alkaline conditions (pH > 9), this tance against acid/alkaline electrolyte (no corrosion) such
self-disproportionation reaction would be suppressed due as TC, TC2, T9S8, Ni3C2, T3S4 (T = Co and Ni, C = S, Se and
to a higher stability of Mn3+, thus explaining a better OER Te) and so on.41–57,163 These materials have been inspected
electrocatalytic ability of d-MnO2 in alkaline solution (higher for OER activity in order to search for the possibility of potential
population of Mn3+).155 A very similar system to CoPi reported bi-functionality (HER and OER) to perform the overall water-
that the OER activity of MnOx–Pi was pH dependent and further splitting reaction.41–57 The coordination environments in this
deduced different mechanisms under acidic and alkaline classification of the transition metals are mostly octahedral
conditions, while two mechanisms were competing with each except for T9S8, Ni3C2 and T3S4,50,52–55 where tetrahedrally
other in neutral electrolyte.39 The turnover-limiting step under coordinated Co and/or Ni are present in these compounds. It
acidic conditions (pH 1.0 to 3.5) was the deprotonation of is worth saying that the [TC6] octahedra connect to each other
hydroxyl group bridging two Mn4+ ions, which was kinetically through a C–C homoatomic bond [C2]2 in TC2, which results
unfavorable in acid environment and yielded an incredibly high in a divalent oxidation state of Co and Ni in these compounds
Tafel slope (653 mV dec1). Under alkaline conditions instead of ‘‘+4’’. The compounds classified into this category
(pH 11.35 to 13.30), on the other hand, the Tafel slope commonly exhibit robust properties and comparable OER
decreased to a value of 60 mV dec1 which indicated that the activities to those above-mentioned metal oxides, in which
turnover-limiting step switched to the chemical step (most the overpotentials at 10 mA cm2 for these compounds were
likely O2 evolution) and followed by one electron/proton reported in a range of 200 to 300 mV. For instance, Chen et al.41
transfer reaction (similar to CoPi system). have shown a Z of B297 mV at 11 mA cm2 with a Tafel slope of

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 349
 View Article Online

Review Article Chem Soc Rev

Table 1 Important measurement and kinetic parameters for metal oxide electrocatalysts for OER

Overpotential (mV) at Current density (mA cm2) Tafel slope

Catalysts Electrolytesa Substratea specific current density at specific overpotential (mV dec1) Ref.
Precious metal (RuO2 and IrO2)
RuO2 0.1 M HClO4 GCE — B0.01/B0.003 @ 250 mV — 13
0.1 M KOH
RuO2 0.5 M H2SO4 Ti 240 @ 1 mA cm2 — 39 122
RuO2 0.5 M H2SO4 Au — B0.003 @ 250 mV 50 123
2
RuO2 0.5 M KOH FTO 358 @ 10 mA cm — 55 99
IrO2 0.5 M KOH FTO 411 @ 10 mA cm2 — 91 99
IrO2 0.5 M H2SO4 Ir 270 @ 1 mA cm2 — 40 15
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

IrO2 0.1 M HClO4 GCE — B0.004/B0.002 @ 250 mV — 13

0.1 M KOH
IrO2 0.5 M H2SO4 Au — B0.0015 @ 250 mV 57.8 123
IrO2 0.5 M H2SO4 Ti 290 @ 1 mA cm2 — 61 122
IrO2@RuO2 0.5 M H2SO4 Au — B6 @ 250 mV 60.2 123

Perovskite family
LaFeO3 1 M KOH S. R.b >900 @ 40 mA cm2 — — 20
La0.4Sr0.6FeO3 1 M KOH S. R.b 622 @ 40 mA cm2 — 58 20
La0.2Sr0.8FeO3 1 M KOH S. R.b 502 @ 40 mA cm2 — 53 20
SrFe0.8Co0.2O3 1 M KOH S. R.b 507 @ 40 mA cm2 — 113 20
SrFe0.5Co0.5O3 1 M KOH S. R.b 412 @ 40 mA cm2 — 163 20
La0.2Sr0.8Fe0.2Co0.8O3 1 M KOH S. R.b 402 @ 40 mA cm2 — 80 20
LaNiO3 1 M NaOH S. R.b — 13 @ 300 mV 43 21
LaCoO3 1 M NaOH S. R.b — 1.6 @ 300 mV 70 21
LaMnO3 1 M NaOH S. R.b — 0.05 @ 300 mV 126 21
La0.8Sr0.2O3 1 M NaOH S. R.b — 0.03 @ 300 mV 200 21
Ba0.5Sr0.5Co0.8Fe0.2O3d 0.1 M KOH GCE 250 @ 0.05 mA cm2 — — 22
370 @ 10 mA cm2
BaNi0.83O2.5 0.1 M KOH GCE 270 @ 0.05 mA cm2 — — 124
CaCu3Fe4O12 0.1 M KOH GCE 400 @ 10 mA cm2 — 51 125
CaFeO3 0.1 M KOH GCE 390 @ 10 mA cm2 — 47 125
SrFeO3 0.1 M KOH GCE 410 @ 10 mA cm2 — 63 125
LaFeO3 0.1 M KOH GCE 500 @ 10 mA cm2 — 77 126
La0.95FeO3d 0.1 M KOH GCE 400 @ 10 mA cm2 — 48 126

Spinel family
MnFe2O4 0.1 M KOH GCE 470 @ 10 mA cm2 — 114 28
CoFe2O4 0.1 M KOH GCE 370 @ 10 mA cm2 — 82 28
NiFe2O4 0.1 M KOH GCE 440 @ 10 mA cm2 — 98 28
CuFe2O4 0.1 M KOH GCE 410 @ 10 mA cm2 — 94 28
Fe3O4 1 M KOH Ni — 0.2 @ 650 mV 43 130
Ni0.5Fe2.5O4 1 M KOH Ni — 0.31 @ 650 mV 35 130
NiFe2O4 1 M KOH Ni — 0.62 @ 650 mV 42 130
Mn3O4 1 M KOH GCE >600 @ 3 mA cm2 — 60 132
Mn2.4Co0.6O4 1 M KOH GCE 510 @ 3 mA cm2 — 51 132
Mn2.1Co0.9O4 1 M KOH GCE 490 @ 3 mA cm2 — 31 132
Co3O4 1 M KOH Au 400 @ 10 mA cm2 — 49 133
Co3O4 nanooctahedron 0.1 M KOH GCE 530 @ 10 mA cm2 — 60 134
Co3O4 nanocube 0.1 M KOH GCE 580 @ 10 mA cm2 — 60 134
CoCr2O4 1 M KOH GCE 422 @ 10 mA cm2 — 63.3 135
CoCr2O4/CNT 1 M KOH GCE 326 @ 10 mA cm2 — 51 135

Layer-structure type family

MnOOH 0.1 M KOH/LiOH PtO/AuO 550 @ 5 mA cm2 — — 27
FeOOH 0.1 M KOH/LiOH PtO/AuO 525 @ 5 mA cm2 — — 27
CoOOH 0.1 M KOH/LiOH PtO/AuO 450 @ 5 mA cm2 — — 27
NiOOH 0.1 M KOH/LiOH PtO/AuO 375 @ 5 mA cm2 — — 27
g-NiOOH 0.1 M KOH Au 700 @ 10 mA cm2 — — 141
g-Ni0.87Fe0.13OOH 0.1 M KOH Au 390 @ 10 mA cm2 — — 141
g-Ni0.75Fe0.25OOH 0.1 M KOH Au 370 @ 10 mA cm2 — — 141
NiFe-LDH 1 M KOH GCE 300 @ 10 mA cm2 — 40 31
NiCo-LDH 1 M KOH GCE 335 @ 10 mA cm2 — 41 31
CoCo-LDH 1 M KOH GCE 350 @ 10 mA cm2 — 45 31
FeNi–rGO LDH 1 M KOH Ni foam 195 @ 10 mA cm2 — 39 145
FeNi–GO LDH 1 M KOH Ni foam 210 @ 10 mA cm2 — 40 145
NiCr DH 0.1 M KOH Au 310 @ 1 mA cm2 — — 147
NiMn DH 0.1 M KOH Au 380 @ 1 mA cm2 — — 147
NiFe DH 0.1 M KOH Au 290 @ 1 mA cm2 — — 147
NiCo DH 0.1 M KOH Au 500 @ 1 mA cm2 — — 147
Ni(OH)2 0.1 M KOH Au 410 @ 1 mA cm2 — — 147

350 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

Table 1 (continued)

Overpotential (mV) at Current density (mA cm2) Tafel slope

Catalysts Electrolytesa Substratea specific current density at specific overpotential (mV dec1) Ref.
2
NiCu DH 0.1 M KOH Au 450 @ 1 mA cm — — 147
NiZn DH 0.1 M KOH Au >500 @ 1 mA cm2 — — 147
LiCoO2 0.1 M KOH CFP >520 @ 5 mA cm2 — 48 30
De-LiCoO2 0.1 M KOH CFP 380 @ 5 mA cm2 — 50 30
De-LiCo0.5Ni0.5O2 0.1 M KOH CFP 370 @ 5 mA cm2 — 42 30
De-LiCo0.5Fe0.5O2 0.1 M KOH CFP 335 @ 5 mA cm2 — 40 30
De-LiCo0.33Ni0.33Fe0.33O2 0.1 M KOH CFP 295 @ 5 mA cm2 — 48 30
De-LiCo0.33Ni0.33Mn0.33O2 0.1 M KOH CFP 395 @ 5 mA cm2 — 46 30
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

Other oxides
a-Fe36Co64Ox 0.1 M KOH FTO 270 @ 1 mA cm2 — 40 153
a-Fe40Ni60Ox 0.1 M KOH FTO 250 @ 1 mA cm2 — 34 153
a-Fe40Co39Ni22Ox 0.1 M KOH FTO 280 @ 1 mA cm2 — 40 153
a-Co41Ni59Ox 0.1 M KOH FTO 270 @ 1 mA cm2 — 73 153
CoPi 0.1 M KPi ITO 410 @ 1 mA cm2 — 62 32
MnOx 0.10 M Pi and FTO B460 mV @ 0.01 mA cm2 — 127 39
1.73 M KNO3
Ni–Bi 0.5 M KB(OH)4, H3BO3, FTO 380 mV @ 1 mA cm2 — 29 162
1.75 M KNO3
Bi2WO6 CNP concave 0.5 M Na2SO4 GCE 540 mV @ 1 mA cm2 — — 37
nanoplates
2
Mn5O8 NPs 0.3 M phosphate FTO 580 mV @ 5 mA cm — 78.7 34
buffer solution
Co(PO3)2 0.1 M phosphate GCE 440 mV @ 8 mA cm2 — 74.1 158
(NaH2PO4,
Na 2HPO4)
a b
For the samples in the same work the electrolyte and substrate only stated in the first sample have been used for the sake of clarity. Sintered
rod sample.

47 mV dec1 for NiS on a stainless steel mesh support, while a active site (cobalt vacancy, Fig. 10(a)) of CoSe2 using an exfoliation
similar OER activity for NiS microsphere was clarified as well.43 process and promoted the OER activity of CoSe2. Astonishing OER
Furthermore, the incorporation of foreign metal ions seems to performance of Fe doped NiSe2 derived oxide catalysts with an
work as well in this system for improving OER activity. A ternary
FeNiS2 ultrathin nanosheets was investigated to show an
acceptable performance,164 in which the overpotential was
290 mV at 10 mA cm2 with a Tafel slope of 46 mV dec1.
The doping of Ni in Co0.85Se can enhance electroconductivity
and promote OER activity as well, and reported overpotentials
at 10 mA cm2 for Co0.85Se and (Ni,Co)0.85Se were 324 mV and
255 mV, respectively.21 It also displayed long-term durability in
OER, where the overpotential (at 10 mA cm2) only increased
to 276 mV after operating for 24 hours. Further decoration of
LDH on the surface of (Ni,Co)0.85Se by electrodeposition can
promote the OER activity of (Ni,Co)0.85Se–NiCo LDH and lower
the overpotential to 216 mV.45 Notice that the synthesized CoSe
and NiS would transform into oxide form (or oxyhydroxide) on
the surface during OER,49,165 and this phase transformation
behavior was similar to that of spinel cases and might
happen through oxidation processes from Ni2+/Co2+ to Ni3+/
Co3+. Therefore, one has to pay careful attention to the reported
Fig. 10 (a) A schematic illustration of the exfoliation process of CeSe2 to
overpotential values at 10 mA cm2, and it may include a increase the number of active sites (Covac). Reproduced with permission
faradic current from oxidation reaction rather than actual from ref. 50. Copyright r 2014, American Chemical Society. (b) Over-
performing OER, thereby resulting in a very low overpotential. potential (Z) at 10 mA cm2 for Ni foam, NiFe LHD, NiSe-DO, NiSe2-DO
In addition to metal monochalcogenides (TC), metal dichal- and NixFe1xSe2-DO (DO represent derived oxide). Reproduced with
permission from ref. 51. Copyright r 2016, Nature Publishing Group.
cogenides (TC2) are also very promising candidates for OER.
(c) Current vs. time (I–t) of Ni3S2/NF for HER and OER (left) while the
Gao et al.166 synthesized CoSe2 nanobelts on nitrogen-doped overall water electrolysis by Ni3S2/NF bifunctional electrodes at 1.76 V was
graphene support with an overpotential of 366 mV and Tafel recorded to 150 hours (right). Reproduced with permission from ref. 53
slope of around 40 mV dec1. Liu et al.50 further increased the Copyright r 2015, American Chemical Society.

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 351
 View Article Online

Review Article Chem Soc Rev

overpotential of B195 mV and a Tafel slope of 28 mV dec1 was

achieved (Fig. 10(b)), and also confirmed that the Se content
decreased with increasing the amount of O in NixFe1xSe2
after OER from energy dispersive spectrum (EDS) analysis, while
NixFe1xSe2 was assumed to transform into metal oxyhydroxide.51
The derived oxide (NixFe1xSe2-DO) exhibited much better OER
activity than that of NiFe-LDH, suggesting that the conductive
Fig. 11 (a) Theoretical calculation of the density of states (ODS) of Ni3N
NixFe1xSe2 acted not only as the precursor for transferring into
(bulk and sheet). Reproduced with permission from ref. 65. Copyright r
the OER active phase of NiFe-LDH on the surface but also the 2013, American Chemical Society. (b) Current vs. potential curve of
conductive scaffold (like core–shell confirmation) for the purpose
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

bifunctional Ni2P water electrolysis. Inset is current vs. time (I–t) for overall
of rapid charge transfer.51 Numerous work has shown that water electrolysis by Ni2P bifunctional electrodes at constant 10 mA cm2.
hybrid materials between oxides (Fe3O4, Mn3O4 and CeO2) Reproduced with permission from ref. 63. Copyright r 2015, the Royal
Society of Chemistry.
and metal chalcogenides effectively enhance corresponding
OER activities, and synergistic effects might also play critical
roles here in affecting the OER activity.
It has to be said that the phase transformation during OER (Fig. 11(a)),65 and the increase in conductivity from modifying
cannot be applied to all metal chalcogenides. In the case of material morphology (bulk to nanosheets) greatly improved the
Ni3Se2, the OER activity of Ni3Se2 was comparable to NiFe-LDH OER activity of the Ni3N sheet compared to that of bulk Ni3N
with an overpotential of only 310 mV at 10 mA cm2, while the (Z drops more than 100 mV).65
much higher Tafel slope of 97.2 mV dec1 suggested that there Another efficient approach to enhancing electroconductivity
was a different rate-determining step from that of NiFe-LDH.52 of the catalysts is to merge catalysts with conducting substrates
The differences in crystal structure and elemental analysis of such as carbon or nickel foam.66,69 Combining with these
Ni3Se2 between the samples before and after OER were similar, conducting supports can not only further improve electro-
indicating that the OER activity of Ni3Se2 could be attributed to conductivity, but also create more active sites (more dispersive
itself without involving a major phase transformation. It is on the support). For example, CoP coupling with reduced
worth mentioning that Feng et al.53 have reported a high-index graphene oxide exhibited overpotential values of B340 mV and
facet Ni3S2 nanosheet, and this material has demonstrated 105 mV for OER and HER activities, respectively (total over-
excellent electrocatalytic ability both for HER and OER in potential was 445 mV for the water-splitting reaction).70
alkaline electrolyte with remarkable stability (over 150 hours, Doping or partially replacing elements in electrocatalysts
Fig. 10(c)). The high durability of overall water-splitting for Ni3S2 has also been shown to be an alternative way to enhance
may be attributed to the intrinsic stability of this phase. electrocatalytic performance. As we mentioned above, many
studies have shown that the doping of Fe into Co- or Ni-based
4.2 Metal pnictides electrocatalysts improves OER activity. Substitution of anions
Similar to metal chalcogenides, extensive studies have been in electrocatalysts can lead to a similar effect, and this idea has
explored for metal pnictides (nitrides and phosphides) such as been demonstrated by Duan et al.,170 in which they synthesized
Co2N, Co3N, Co4N, Ni3N, TP and T2P (T = Co and Ni) for HER CoFePO (Fe- and O-doped Co2P phase) on Ni foam substrate
owing to their good electroconductivity and are relatively inert and this electrocatalyst showed a remarkable low overpotential
to harsh electrolytes (strong acid and alkaline environments).167–169 of 362 mV (274.5 mV and 87.5 mV for OER and HER, respec-
They have shown remarkable HER activities (Z o 200 mV) in tively) at 10 mA cm2.
alkaline electrolyte, and are encouraging bifunctional candidates Although incorporating other transition metals into
for the overall water-splitting reaction. Nevertheless, on the con- metal pnictides catalysts can improve their corresponding
trary, research into their OER activities is still a growing field and OER activities, their performance rapidly deteriorated after
only handful of metal pnictides have been reported for OER.59–72 It long-term usage.71 This decay in OER performance of CoMnP
was noted that the OER mechanism for these metal pnictides may be attributed to the surface oxidation with etching phos-
might be similar to that of metal chalcogenides, in which a thin phorous of CoMnP.71 Regarding the active phase of the surface
metal oxide layer was formed on the surface and was responsible oxide, it could be great challenge to control doped catalysts for
for OER with conducting metal pnictide as core to facilitate electron OER. It was difficult to precisely control the composition,
transfer.62,63 Along this line of thinking, increasing electro- morphology and even structure of the surface oxides derived
conductivity of metal pnictides can accordingly promote their from the doped catalysts. Even the nominal composition of
OER performance. A series of cobalt nitrides (Co2N, Co3N and doped catalysts could be precisely analyzed, the composition of
Co4N) revealed that the materials were more conductive with derived oxide on the surface may be deviated and thereby result
higher amounts of cobalt and thereby more active for OER.64 The in a completely different composition from that of the initial
same conclusion could be drawn from the result of the Ni3N catalysts. Despite the uncertainty in doped metal pnictides,
sheet: the electroconductivity of Ni3N sheet was higher than that some simple binary pnictides such as CoP and Ni2P could still
of bulk Ni3N due to a greater density of states emerging for the demonstrate robust electrocatalytic ability toward OER.62,63,72
Ni3N sheet around the Fermi level from theoretical calculations A Ni2P/NiOx core–shell assembly displayed long-term stability

352 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

(over 10 hours) with overpotentials of 290 mV and 400 mV at

10 mA cm2 for OER and the whole water-splitting reaction,
respectively (Fig. 11(b)).63 As aforementioned, metal pnictides
have shown great potential in both HER and OER, and are
certainly worth exploring further.

4.3 Organometallics
The original aim of developing the molecular compounds in
this field is probably to mimic the famous ‘‘CaMn4O5 cluster’’ as
a model system to investigate oxygen-evolving complex behavior
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

in natural photosynthesis (Photosystem II, PSII).171 Oxygen-

evolving complex systems are commonly organometallics con-
taining transition metals which are usually Mn, Fe, Co, Ru and Ir.
It is worth noting that the most well-known organometallic is
probably the ruthenium blue dimer,172 in which an oxo-bridge
connecting two ruthenium metals structure is present. This
created the long-term perspective that dinuclear or multinuclear
Fig. 12 (a) Structure of M1–TPyP and proposed structural model
species are necessary to trigger OER (oxygen is probably produced for extended bimetallic (M1 and M2) M1–TPyP–M2. Reproduced with
by two M–O units, via so called ‘‘oxyl radical coupling’’).173 permission from ref. 74. Copyright r 2016, American Chemical Society.
However, many mononuclear organometallics have been shown (b) Multi-electron transfer and electron rearrangement scheme for a
to exhibit excellent OER activities, and more investigations into multinuclear cluster (top); schematic illustration of adjacent active sites
the ruthenium blue dimer for OER suggested that a single site to form intramolecular O–O bond (bottom). (c) Molecular structure of
[FeII4FeIII(m3-O)(m-L)6]3+ (left); core structure of Fe5O and chemical struc-
of Ru might be responsible for the OER active site (so-called ture the LH (right). Reproduced with permission from ref. 77. Copyright r
‘‘water nucleophilic attract’’ and produced M–OOH).169 There 2016, Nature Publishing Group.
are many outstanding reviews that have already discussed the
oxygen-evolving behavior of these organometallics in great
detail.171,173,174 To avoid unnecessary overlapping content with exhibited the highest electrocatalytic ability for OER and
those reviews, we will provide concise discussion and focus on ORR among all prepared samples (M1TPyP o M1TPyP–Fe o
the latest developments in this field. M1TPyP–Co; M1 = Fe and Co). Another bimetallic system, Ni- and
One thing that stands out between organometallics and Co-porphyrin deposited on reduced graphene oxide (rGO) to
those oxide/non-oxide electrocatalysts for OER is that organo- form a layer-by-layered architecture, was demonstrated by Sun
metallics are homogenous catalysts (some composite catalysts et al.75 in which the optimized 8 layers compound exhibited a
containing organometallics might be considered as hetero- remarkably low overpotential of B330 mV with a Tafel slope of
geneous catalysts) while those oxide/non-oxide materials are 50 mV dec1.
heterogeneous catalysts for the water-splitting reaction. The In addition to these mononuclear organometallics, some
advantage of organometallics for OER is that their structures dinuclear and multinuclear organometallics, and even cluster
and electronic states could be designed in order to investigate metal–organic compounds have been reported recently. Shee-
their mechanisms for OER. Nevertheless, the shortcomings han et al.76 have shown a very stable dinuclear Ir catalyst
of their poor stabilities and low electroconductivities usually binding to the electrode (immobilized), while the anchored
prevent them from practical usage. To overcome these draw- ligand on [Ir(pyalc)(H2O)(O)(m-O)]2 was still intact under such a
backs of organometallic compounds toward OER, several com- high-anodic potential environment. Another pentanuclear iron
posite materials including mononuclear, multinuclear and even catalyst was proposed by Okamura et al.,77 and it showed a
cluster organometallic catalysts have been recently revealed to peculiar trigonal-bipyramidal molecular shape with O located
show excellent OER activity.77,175,176 For example, a manganese in the center and Fe on the corner (Fig. 12(c)) while oxygen
corrole complex can exhibit bifunctional properties for both OER gas evolved via intramolecular (adjacent FeIVQO) O–O bond
and ORR.177 This manganese corrole complex can be further formation. A similar mechanism for oxygen evolution was
modified by replacing manganese with cobalt in the center of proposed by Ullman et al.178 as well. These experimental results
corrole and combining with multi-walled carbon nanotubes suggest that both oxyl radical coupling and water nucleophilic
(MWCNTs) for improving the resulting performance.73 Superior attraction are two possible pathways for oxygen evolution.
stability and OER activity compared to the cobalt corrole
complex/MWCNTs without this substituent were revealed, which 4.4 Non-metal compounds
was attributed to the function of p–p interaction between the The electrocatalysts for OER discussed above are metal-based
pyrene group and MWCNTs.73 Wurster et al.74 have designed materials, which inevitably introduces cost problems for large-
a bimetallic porphyrin system by incorporating Fe and Co scale applications. Searching for alternative excellent electro-
(Fig. 12(a)) and verified that the Fe in the porphyrin center catalysts as well as cheap materials for OER is not an easy task,
with chelating Co through four pyridyl groups (FeTPyP–Co) but on-going efforts are still devoted to fulfilling this purpose.

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 353
 View Article Online

Review Article Chem Soc Rev

One approach is to develop non-metal materials for OER. For

instance, Mirzakulova et al.79 reported a molecule catalyst of
N(5)-ethylflavinium ion (Et–Fl+) for O2 evolution reaction.
Although this system is not totally new, it is probably the first
demonstration of a non-metal compound to actually possess
the ability to undergo the O2 evolution reaction. Though the
OER activity of N(5)-ethylflavinium ion was poor (Z > 1 V with
a Faradaic efficiency of B30% in 0.1 M phosphate buffer
solution), it provided vital information and guidance for the
latter work in this field. They proposed that the active site for
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

initiating OER was the carbon between N and CQO (ketone

group), which carried positive charge (due to adjacent N and
ketone group) to adsorb OH and began the water oxidation
catalytic loop. This pioneering work has motivated many
researchers to search for possible non-metal electrocatalysts
for OER.
Attention is immediately turned to carbon-based materials
such as carbon black, carbon nanotubes, and graphene with
N, O and S doping, because they are good electroconductors
and chemically inert to acid and alkaline electrolytes.179 For
example, nitrogen-doped carbon (N/C) nanomaterials (carbon
black) have been revealed to exhibit remarkable activity with a
current density reached to 10 mA cm2 at 1.61 V in KOH
electrolyte (pH = 13).180 However, it was advised that carbon
black might be unstable and oxidized to produce carbon Fig. 13 (a) A schematic illustration of the surface-oxidized multi-walled
dioxide at above 0.9 V vs. RHE,181 which led to a reasonable carbon nanotubes. Reproduced with permission from ref. 83. Copyright r
2015, American Chemical Society. (b) A schematic flow chart for the
doubt for long-term usage. Carbon nanotubes and graphene/
process to fabricated of 3D g-C3N4 and NS-CNT porous composite.
graphene-like compounds are well-known to be more conduc- Reproduced with permission from ref. 86. Copyright r 2014, Wiley-VCH
tive and stable than carbon black. Cheng et al.80 systematically Verlag GmbH & Co. KGaA, Weinheim.
studied the pristine character of carbon nanotubes (single wall
and multiple walls, SWNTs and MWNTs) and the correlation
between the numbers of the walls and their corresponding OER more active sites might be able to improve its OER kinetics.
activities, where the outer wall was responsible for OER while Employment of graphitic C3N4 (g-C3N4) and graphene/carbon
the function of inner wall was to increase the rate of electron nanotube can form a non-metal composite catalyst for
transfer (tunneling effect). As a result, the OER activity should OER,85,86 excellent OER activity could be expected because
be enhanced once the numbers of walls increased. However, of substantial amount of active sites in g-C3N4 (assuming the
increased numbers walls in carbon nanotubes turned out to neighboring carbon to nitrogen was the active sites). Indeed,
have lower OER activity, which was attributed to the attenuated g-C3N4/graphene performed an overpotential of B539 mV
electron tunneling between the outer wall and neighboring (0.1 M KOH) with a small Tafel slope of B68.5 mV dec1.85
inner wall due to the diminishment of dc bias influence. Lu A 3D g-C3N4/carbon nanotubes structure prepared by
et al. further increased the OER activity of carbon nanotubes Ma et al.86 (Fig. 13(b)) displayed even higher OER activity
(MWNTs) by oxidizing the surface of the carbon nanotubes with an overpotential of B370 mV and a Tafel slope of
(Fig. 13(a)),83 which was a very effective method to improve the B83 mV dec1. Moreover, g-C3N4 is a good composite material
OER activity. The reported overpotential was around 360 mV that can mix with other metal catalysts that have shown
at 10 mA cm2; this value is comparable to that of the afore- promising catalytic activities in both OER and ORR. In the
mentioned metal oxides/chalcogenides/pnictides. very recent work of Ma et al.,182 a hybrid film of overlapped g-
In addition to carbon nanotubes, graphene or graphene- C3N4 and Ti3C2 has been successfully prepared through an
like materials have also been employed for OER activity. exfoliation process, and this composite material exhibited very
Chen et al.84 have synthesized a unique 3D-architecture reliable long-term stability toward OER and could be practi-
through the mixing of nitrogen and oxygen dual doped graphene cally applied in metal–air batteries in the future.
and carbon nanotubes, and this 3D conformation possessed Recently, it was noticed that black phosphorus exhibited a
advantages such as porous material (higher surface area and very interesting property and could be applied in various
more active sites) and exhibited an overpotential of 365 mV at fields.183 This material possessed 2D puckered layer structure
5 mA cm2 for this hybrid material while the Tafel slope was and its electroconductivity can be modified by adjusting
around 141 mV dec1. This slightly higher Tafel slope implicated the film thickness.184 Black phosphorus (BP) prepared by the
that the OER kinetics still has room to be improved, and that thermal-vaporization transformation (TVT) method exhibited a

354 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

Table 2 Important measurement and kinetic parameters for non-metal oxide electrocatalysts for OER

Overpotential (mV) for OER Overpotential (mV) for HER/ Tafel slope
Catalysts Electrolytesa Substratea at specific current density overall water-splitting reaction (mV dec1) Ref.
Metal chalcogenides
CoS 1 M KOH Ti mesh 361 @ 10 mA cm2 — 64 58
NiS 1 M KOH Ni foam 335 @ 50 mA cm2 158 @ 20 mA cm2/ 89/83 43
410 @ 10 mA cm2
CoSe0.85 1 M KOH CFC 324 @ 10 mA cm2 — 85 45
(Ni,Co)Se0.85 1 M KOH CFC 255 @ 10 mA cm2 — 79 45
(Ni,Co)Se0.85–NiCo LDH 1 M KOH CFC 216 @ 10 mA cm2 — 77 45
CoSe2 bulk 0.1 M KOH GCE 590 @ 10 mA cm2 — 108 50
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

CoSe2 ultrathin nanosheets 0.1 M KOH GCE 320 @ 10 mA cm2 — 44 50

CeO2/CoSe2 0.1 M KOH GCE 288 @ 10 mA cm2 — 44 48
NiSe2-DO 1 M KOH Ni foam 241 @ 10 mA cm2 — 32 51
NixFe1xSe2-DO 1 M KOH Ni foam 195 @ 10 mA cm2 — 28 51
Ni3S2 nanosheet 1 M KOH Ni foam 260 @ 10 mA cm2 223 @ 10 mA cm2/ — 51
530 @ 10 mA cm2
Ni3Se2 0.3 M KOH Ni foam 290 @ 10 mA cm2 — 97 52

Metal pnictides
Co2N 1 M KOH GCE 430 @ 10 mA cm2 — 80 64
Co3N 1 M KOH GCE 410 @ 10 mA cm2 — 72 64
Co4N 1 M KOH GCE 330 @ 10 mA cm2 — 58 64
Co4N 1 M KOH CC 257 @ 10 mA cm2 — 44 59
Ni3N bulk 1 M KOH GCE 490 @ 10 mA cm2 — 45 65
Ni3N nanosheets 1 M KOH GCE 350 @ 10 mA cm2 — 85 65
Ni3N nanosheets 1 M KOH CC 350 @ 52.3 mA cm2 — 41 65
CoP/rGO-400 1 M KOH GCE 340 @ 10 mA cm2 150 @ 10 mA cm2/ 66/38 70
470 @ 10 mA cm2
Co2P 1 M KOH GCE 370 @ 10 mA cm2 — 128 71
CoMnP 1 M KOH GCE 330 @ 10 mA cm2 — 61 71
Ni2P 1 M KOH Ni foam 290 @ 10 mA cm2 220 @ 10 mA cm2 59/— 63

Organometallics
FeTpyp–Feb 0.1 M NaOH Au B480 @ 0.5 mA cm2 — — 74
CoTPyP–Fe 0.1 M NaOH Au B460 @ 0.5 mA cm2 — — 74
CoTPyP–Co 0.1 M NaOH Au B420 @ 0.5 mA cm2 — — 74
FeTPyP–Co 0.1 M NaOH Au B400 @ 0.5 mA cm2 — — 74
[Ir(pyalc)(H2O)(O)(m-O)]2 0.1 M KNO3 nanoITO/ B250 @ 0.5 mA cm2 — 66 76
FTO
Co1/MWCNT rinsedc 0.1 M phosphate buffer GCE B330 @ 1 mA cm2 — — 73
Co2/MWCNT rinsedd 0.1 M phosphate buffer GCE B480 @ 1 mA cm2 — — 73
rGO/(Ni2+/THPP/Co2+/THPP)8e 1 M KOH GCE B330 @ 10 mA cm2 — 50 75

Non-metal
N/C 0.1 M KOH GCE 380 @ 10 mA cm2 — — 180
Echo-MWCNTs f 0.1 M/1 M KOH GCE 450 @ 10 mA cm2/ — 72/41 83
360 @ 10 mA cm2
G-CNTg 0.1 M KOH GCE B498 @ 5 mA cm2 — 231 84
NG-CNTh 0.1 M KOH GCE 368 @ 5 mA cm2 — 141 84
g-C3N4i 0.1 M KOH GCE 734 @ 10 mA cm2 — 120.9 85
g-C3N4/graphene 0.1 M KOH GCE 539 @ 10 mA cm2 — 68.5 85
3D g-C3N4 NS–CNT 0.1 M KOH GCE B370 @ 10 mA cm2 — 83 86
BP j 0.1 M KOH Ti 370 @ 10 mA cm2 72.9 185
a
For the samples in the same work, the electrolyte and substrate only stated in the first sample have been included for the sake of clarity. b TPyP =
5,10,15,20-tetra(4-pyridyl)-21H,23H-porphyrin. c THPP = 5,10,15,20-tetrakis(4-hydroxyphenyl)porphyrin. d Co1 = Co complex of 5,15-bis-(penta-
fluorophenyl)-10-(4)-(1-pyrenyl)phenylcorrole. e Co2 = Co complex of 5,10,15-tris(pentafluorophenyl)-corrole. f Electrochemical activation of
ho-MWCNTs. g Graphene–carbon nanotubes. h Nitrogen doped graphene–carbon nanotubes. i Graphitic carbon nitride. j Black phosphorous.

current density of 10 mA cm2 at an applied voltage of around 5. Mechanism and in situ experiments
1.6 V for OER.185 This performance was comparable to that of for OER
RuO2 and further modification to this metal-free material could
lead to even better OER activity. To this end, numerous work 5.1 In situ experiments
has shown that the non-oxide compounds exhibited excellent In order to realize the mechanism and actual active phase
OER and HER activities, and could be utilized for the overall responsible for the reaction, in situ experiments are commonly
water splitting reaction in the future. These materials together carried out to provide valuable and vital information since the
with their OER kinetic parameters mentioned in this section probed reaction is nearly under practical conditions. Never-
are listed in Table 2 for comparison. theless, very few experiments (operando) have been reported for

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 355
 View Article Online

Review Article Chem Soc Rev

Co–O stretching mode (lattice oxygen, unlabeled) from Co3O4

surface and assigned to be a Co(IV)QO intermediate (or reso-
nance structure Co(III)–O ). The decay of this band with a time
constant of 1 s proved that this intermediate state was irrele-
vant to the oxygen evolution after the first laser pulse, and
therefore possessed slower water oxidation kinetics (Fig. 15(a)).
In addition to the ATR, there are also many excellent in situ
experiments to probe the electronic state and structural infor-
mation of the active phase during OER. For instance, Louie and
Bell have provided a thorough study of the Ni–Fe oxide electro-
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

catalyst toward OER using Raman spectroscopy and pointed

out that the dramatic change of electrocatalytic ability of Ni
film (fresh vs. aged) was due to the phase transformation from
g-NiOOH to b-NiOOH (Bode scheme), which was the main
reason to account for this behavior (although the Fe impurity
Fig. 14 3-D perspective view of different in situ techniques and corres-
ponding penetration depth on the surface of material.
issue could not be entirely ruled out).88 Another case study was
presented by Trześniewski et al (Fig. 15(b)),89 where they
compared the OER activities between Ni–Bi and Ni(Fe)(OH2)/
OER to date,87–98,186,187 which is probably due to the challenges Ni(Fe)OOH by various in situ techniques and offered fruitful
in measuring the heterogeneous junction between the electrode data, in which several important findings could be concluded
(immobilized electrocatalysts) and electrolyte in liquid condi- from their data. The structural and electrochemical behavior
tion. Besides, this kind of research involves designing special of Ni–Bi and Ni(Fe)(OH2)/Ni(Fe)OOH were very similar. It
devices or apparatus to fulfil this goal, which further increases was worth mentioning that two sharp bands appeared in the
the challenges of in situ experiments. There are only several range of 1200–1400 cm1 in the in situ Raman spectra, which
techniques that can provide such information on the surface indicated the adsorption of borate anions on the surface of
of the electrocatalysts (structural and electronic states). For Ni(Fe)(OH2)/Ni(Fe)OOH (playing a role of inhibitor to the active
example, Fourier transform infrared spectroscopy (FTIR) and site); this phenomenon could be suppressed by increasing
Raman spectroscopy have been demonstrated to achieve this the pH value. Once this inhibiting effect diminished, a broad
goal.87–89,187 In addition, by utilizing higher energy irradiation band was observed in the range from 900–1150 cm1, which
such as hard X-ray or g-ray (X-ray absorption/diffraction and was attributed to the ‘‘active oxygen’’ species and could be
Mössbauer spectroscopy), in situ monitoring of electrocatalysts Ni-peroxide or Ni-superoxide.
can be realized to probe materials at a deeper level for obtaining Another technique, surface interrogation scanning electro-
vital information (i.e., chemical state, coordination environment chemical microscopy (SI-SECM) also provides key information
and crystal nature) (Fig. 14). on the surface (probably at the atomic level) of electrocatalysts
Toward this end, attenuated total reflection (ATR) equipped during OER. One could study the time-dependent decay mecha-
within a Fourier transformed infrared spectroscopy instrument nism (kinetics) of individual cations in samples based on the
has been utilized to investigate the surface behavior of redox potential window that each measurement could provide.
materials.87 Zhang et al.87 have interrogated OER mechanism Ahn and Bard have made great contributions with this technique
on Co3O4 in H218O (pH B 8.5) via this technique and revealed to water oxidation reaction by several interesting systems such
two intermediate states on the surface of Co3O4. The distinct as Co–Pi and Ni1xFexOOH.90–92 In most cobalt-related systems,
catalytic behavior of these two intermediates suggested two it was believed that the high oxidation state (CoIV) was the key
possible mechanistic pathways (fast and slow). In their rapid-scan for fast OER. Many experiments have been carried out with the
FTIR experiment, two obvious bands can be observed, which were efforts to detect or to suggest the existence of CoIV in the
at B1013 cm1 and 840 cm1. The band at 1013 cm1 suggested investigated systems. In the case of the Co–Pi system,90 it was
a possible bridged superoxide intermediate [Co(III)–(OO) –Co(III)], realized that the shortest lifetime was COIV (o0.4 s) rather than
while the band at 840 cm1 indicated the presence of Co(IV)QO that of COIII (B3.6 s) in the Co–Pi system. Nevertheless, this
intermediate (or resonance structure Co(III)–O ). After consecutive unexpected short lifetime of COIII (B3.6 s) suggested that COIII
laser pulsing to trigger hole-injection from [Ru(bpy)3]2+ into could also be an active species for OER and was worth further
Co3O4, the intensity of the 1013 cm1 band decreased with investigation. As mentioned earlier, several theoretical papers
shifting the isotopic bands to 995 cm1 (18O16O) and 966 cm1 have suggested the superior performance of Ni1xFexOOH was
(18O18O), and this intermediate (superoxide) was assumed to due to a suggestion that Ni was the slow active site for OER
be the fast catalytic process for oxygen evolution. This was based on the data of NiOOH and FeOOH from the same
subsequently supported by the detection of labeled oxygen approach (Fig. 15(c)). Furthermore, they pointed out that the
(36O2 and 34O2) after the first laser pulse (300 ms) in mass fraction of the fast active site was the fast active site of
spectroscopy. On the other hand, no such isotopic shifting Fe, which were embedded in a relatively conductive NiOOH
effect was present for the 840 cm1 band, this band was the (compared to FeOOH) matrix. By using the SI-SECM technique,92

356 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

Fig. 15 (a) Proposed catalytic cycle of Co3O4 for the OER mechanism. Reproduced with permission from ref. 87. Copyright r 2012, Nature Publishing
Group. (b) Different in situ techniques including UV-vis, Raman (SERS) and X-ray absorption to investigate Ni(Fe)OOH and Ni(Fe)–Bi system for OER.
Reproduced with permission from ref. 89. Copyright r 2015, American Chemical Society. (c) A schematic illustration of SI-SECM technique to probe
active site behavior in Ni1xFexOOH system during OER. Reproduced with permission from ref. 92. Copyright r 2015, American Chemical Society.
(d) 3-D perspective view for in situ X-ray diffraction technique setup (left) and an enlarged view of interfaces between electrode and electrolyte (right).
Reproduced with permission from ref. 94. Copyright r 2015, Nature Publishing Group. (e) An in situ X-ray absorption technique and setup for detecting
the IrOx system. Reproduced with permission from ref. 97. Copyright r 2016, American Chemical Society.

two active sites were discovered in the Ni1xFexOOH system, Fe It is well-known that CoO and Co3O4 would undergo phase
was assigned as the fast active site while close to the Fe ratio in transformation to generate CoOOH under high pH and
Ni1xFexOOH. This was another strong indication to elucidate high anodic potential environments (Pourbaix diagram).110 It
that active site Fe was responsible for the outstanding OER appeared that all of the cobalt oxide phases would eventually
performance in the Ni1xFexOOH system. transform into CoOOH which was the major active phase for
On the other hand, Chen et al.93 demonstrated another OER. This argument was supported by the work from Chou
perspective on OER activity in Ni1xFexOOH, where they et al.188 Interestingly, all of them exhibited more or less similar
carried out an in situ experiment of Mössbauer spectroscopy OER activities with overpotentials (Z) at 10 mA cm2 of 497 mV,
between Ni1xFexOOH and FeOOH for their OER activities. 497 mV and 473 mV for Co3O4, CoO and e-Co, respectively.
It demonstrated the existence of Fe4+/Fe3+ mixed states in Given the very different initial state of cobalt in these three
Ni1xFexOOH while only Fe3+ in FeOOH during OER. However, samples, the very similar electrocatalytic abilities for OER were
they concluded that those detectable Fe4+ (relatively stable) in quite remarkable, which led to the conclusion that the real
Ni1xFexOOH might not be directly responsible for OER activity active phase on the surface of these electrocatalysts were
since it was generated at anodic potential (over 1.62 V) higher identical. To further understand the phase transformation
than the on-set potential for OER (1.49 V). between CoO, Co3O4 and CoOOH, Tung et al.94 have synthe-
As compared to FTIR, Raman and SECM that conduct the sized a single crystal of Co3O4@CoO nanocubes, in which they
outer junction (electrolyte and electrocatalytic interface), employ- observed the phase transformation to CoOOH through in situ
ment of X-ray spectroscopy allow us to realize the evolution onto X-ray diffraction spectroscopy (Fig. 15(d)). It was noticed that
electrocatalysts from the viewpoint of materials.94,95,97,98,128,144 Co3O4@CoO (i.e., with a thin CoO layer on the surface of Co3O4)

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 357
 View Article Online

Review Article Chem Soc Rev

showed remarkable durability for OER activity (>1000 hours) in clear shift in the XAS spectra suggested that Ir3+ had been oxidized
alkaline electrolyte. This adapting layer (CoO) may play a to Ir4+ or an even higher state (Ir5+). This transformation was
profound role and show a better pathway to transform into reversible, and the oxidation state changed back to its initial state
CoOOH rather than that of Co3O4. Moreover, this phase trans- when applying a cathodic potential.
formation to metal oxyhydroxide (MOOH) may depend on the
electrolyte conditions. Bergmann et al.95 have performed in situ 5.2 Mechanism from Tafel analysis and DFT calculation
XRD in neutral electrolyte, Co3O4 would transform into an As discussed in Sections 2.1.3 and 2.2, a shallow Tafel slope is
amorphous CoOx(OH)y shell rather than CoOOH. usually an indication of a good electrocatalyst and the calculated
In addition to in situ X-ray diffraction spectroscopy, several Tafel slope values may provide insightful information toward
in situ X-ray absorption spectroscopy experiments were also
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

the reaction mechanism of the target system. A variety of oxide/

reported recently in order to understand the material behavior non-oxide materials discussed herein constantly exhibit
during OER. Wang et al.96 have conducted an in situ X-ray absorp- ‘‘certain’’ Tafel slope values such as 30 mV dec1, 40 mV dec1,
tion spectroscopy to individually monitor the evolution of Ni and 60 mV dec1 and 120 mV dec1. Such coincidence strongly
Co metal ions during OER in spinel Ni–Co bimetallic oxide, in implicates that some traits or patterns may exist, which can be
which the chemical state and coordinated environment of Ni metal linked to the rate-determining step. Nevertheless, OER is a very
ion remarkably changed while that of Co metal ions remained complicated system that one should pay careful attention to
identical. Furthermore, Abbott et al.97 have studied particle size, when interpreting the mechanism solely based on this value.
morphology and surface hydroxo layer on IrOx-based catalysts There are many pioneering papers associated with kinetic
through in situ XAS technique (Fig. 15(e)). They have discovered studies involving the OER reported by Bockris, Damjanovic et al.,
that the smaller particles seem to be more sensitive to the applied Krasil’shchikov, and Conway and Bourgault.100–102,189,190
potential. For instance, the initial oxidation state of Ir in IrO2 was a A comprehensive but non-exhaustive table of selected OER
mixture of Ir3+ and Ir4+. Once an anodic potential was applied, a mechanism models are organized and presented in Table 3.

Table 3 Models for the OER mechanism from Tafel analysis

Oxygen formation from oxo group coupling path Oxygen formation from terminal peroxide group path
(green route in Fig. 3) (black route in Fig. 3)
Tafel slopes (mV dec1) Tafel slopes (mV dec1)
Low overpotential High overpotential Low overpotential High overpotential
100 190
(1) The oxide path (4) Conway & Bourgault path
M + OH - MOH + e 2RT/F M + OH - MOH 2RT/F
120 mV dec1 120 mV dec1
2MOH -MO + M + H2O RT/2F MOH + OH - MO + H2O

2RT/3F 2RT/F
30 mV dec1 40 mV dec1 120 mV dec1
2MO - 2M + O2 RT/4F MO + MOH - MO2H 2RT/5F
15 mV dec1 24 mV dec1
MO2H + MOH - M + H2O + O2 RT/4F
15 mV dec1

(2) The electrochemical oxide path100 (5) Doyle and Lyons path103
M + OH - MOH + e 2RT/F MOH2 + OH - MOH + H2O —
120 mV dec1
MOH + OH - MO +

2RT/3F 2RT/F MOH - MOH + e 2RT/F
H2O + e 40 mV dec1 120 mV dec1 120 mV dec1
2MO - O2 + 2M RT/4F MOH + OH - MO + H2O RT/F
15 mV dec1 60 mV dec1
MO - MO + e 2RT/F
120 mV dec1
MO + OH - MO2H + e 2RT/3F
40 mV dec1
MO2H + OH - MO2 + 2RT/3F
H2O + e 40 mV dec1
MO2 + OH - MOH + O2 —

(3) Krasil’shchikov path189 (6) Shinagawa path104 Tafel slopes (mV dec1)
M + OH - MOH + e 2RT/F MOH - MOH + e 120 mV dec1
120 mV dec1
MOH + OH - MO + H2O RT/F MOH + OH - MO + H2O + e 30, 120 mV dec1
60 mV dec1
MO - MO + e 2RT/3F 2RT/F MO + OH - MO2H + e 21, 40, 120 mV dec1
40 mV dec1 120 mV dec1
2MO - O2 + 2M RT/4F MO2H + OH - MO2 + H2O 22, 30, 60, 120 mV dec1
15 mV dec1
MO2 - M + O2 + e 22, 120 or 40, 120 mV dec1

358 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

Bockris100 has developed an approach from a reaction kinetics determining step (PDS; DG1, DG2, DG3, DG4). Fig. 16(a) is the
viewpoint and calculated corresponding the Tafel slopes ideal case scenario of the standard free energy diagram,107
for several proposed models. Based on his calculations on indicating that no overpotential (Z) is needed for OER to occur
OER kinetics, he concluded that the low- and high-applied (DG1 = DG2 = DG3 = DG4 = 1.23 eV). The theoretical overpotential
overpotential could yield different Tafel slopes even for the (Z) under standard conditions can be calculated as follows:
same rate-determining step. Indeed, much research work has
2H2O 2 H2O + HO* + e
reported low Tafel slope values in low overpotential regions
while the Tafel slope increases in high overpotential regions for
DG1 = DGHO*  eU + kBT ln aH+ (5.1)
electrocatalysts during OER. It is noted that different models
with different consecutive reaction steps may still offer the
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

same value of Tafel slope, thereby choosing a reasonable model H2O + HO* + e 2 H2O + O* + e + H+
based on the experimental facts is the key for Tafel analysis.
Nevertheless, the Tafel slope still provides an important view on DG2 = DGO*  DGHO*  eU + kBT ln aH+ (5.2)
the rate-determining step especially in the range of onset over-
potential. For instance, it is generally accepted that the first H2O + O* + 2(e + H+) 2 HOO* + 3(e + H+)
step of OER is one electron transfer, and the corresponding
Tafel slope was 120 mV dec1 if this is the rate-determining DG3 = DGHOO*  DGHO*  eU + kBT ln aH+ (5.3)
step. The second step is more complicated, in which the
calculated Tafel slope can be 60 mV dec1 if it is a chemical
HOO* + 3(e + H+) 2 O2 + 4(e + H+)
step (i.e., M–OH M–O + H+), while it becomes 40 mV dec1 if it
is an electron–proton reaction (i.e., M–OH M–O + H+ + e). The
DG4 = DGO2  DGHO*  eU + kBT ln aH+ (5.4)
situation would be more complicated in the third, forth and
even fifth steps of those proposed models.101 After investigating many systems of metals and metal oxides
In aforementioned in situ experiments, oxo, peroxide and such as Pt, Au, Ag and TiO2, RuO2 and IrO2, they discovered that
superoxide intermediates were proposed to form during OER there is a linear relationship between each binding energy
catalytic cycles. Recently, several models including these inter- difference, (DEHO*, DEO*, DEHOO*) DEHO* and DEHOO* can be
mediates for the OER mechanism were proposed by Doyle and represented with only DEO*. In this way, the Gibbs free energy
Lyons103 and Shinagawa et al.104 In the work of Doyle and for intermediates (DGHO* and DGHOO*) and for these four
Lyons,103 they have studied OER based on hydrous iron oxide reaction steps (DG1, DG2, DG3 DG4) can be correlated to DEO*.
and included the hydroxide surfaquo groups in their model of A clear volcano plot can be established, in which either too
the OER mechanism. Recently, Shinagawa et al.104 have pro- weak or too strong bonding energy of DEO* (Sabatier principle)
vided a comprehensive study on water electrolysis from micro- would increase overpotential. Their theoretical work clearly
kinetic analysis. Their model also included intermediates such illustrated that RuO2 is the most superior electrocatalyst
oxo, peroxide and superoxide group on the surface for electro- (Z B 370 mV) among all materials in their studies, owing to
catalysts in the OER catalytic cycle. They have suggested that weak binding energy to the oxygen intermediate (O*) compared
the Tafel slope could be varied to several values even for the to the most optimized (acme) situation. In addition, their data
same rate-determining steps considering different coverage also demonstrated the overpotential of IrO2 (Z B 560 mV)
degree of intermediates. Nevertheless, it is worth mentioning would be larger than that of RuO2 due to a stronger binding
that generally speaking, the Tafel slope decreased once the rate- energy to the oxygen intermediate. These data match nicely
determining step was closer to end step of a series of reactions, with experimental results.
which was a sign of a good electrocatalyst. The predicted Man et al.107 have studied OER activities of many oxide
Tafel slopes from those proposed paths were usually helpful materials using the same approach. A universal scaling rela-
and able to account for experimental outcomes. However, tionship was revealed, where the free energy difference between
it lacked detailed information on the thermodynamic aspect HO* and HOO* was always close to 3.2 eV regardless of
(e.g., bonding interactions) between intermediates and elec- different samples (Fig. 16(a)). Therefore, the lowest overpoten-
trode catalysts, which might provide valuable information with tial (Z) should be located where O* was placed just between
regards to the overpotential of the electrodes. HO* and HOO* (i.e., DG2 = DG3 > DG1 and DG4), which is 0.37 V
Recently, this insufficiency was solved by Nørskov et al.105 ((1.6–1.23 eV)/e). This idea can be further utilized to construct
and Rossmeisl et al.,106 in which their DFT calculation results an activity volcano plot based on the binding strength between
showed an interesting and important message able to describe catalyst and O*. As one can see from Fig. 16(b), the samples
the overpotential (Z) of the ORR and OER for a given electro- (SrCoO3 and LaNiO3) near top of the volcano curve possess low
catalyst. Their idea is straightforward yet powerful. A four-step overpotential values (DG2 = DG3). The sample in the left part
proton/electron mechanism was proposed for OER, and the of the volcano curve indicates stronger O* binding strength
maximum difference of Gibbs chemical binding energy between (i.e., difficult to form HOO* and DG2 o DG3), where the
two subsequent absorbed intermediates in these four consecu- samples in the right part of the volcano have weaker O* binding
tive electrochemical reactions (eqn (5.1)–(5.4)) is potentially the strength (DG2 > DG3). Their theoretical predictions for OER

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 359
 View Article Online

Review Article Chem Soc Rev

development for the future. Electrocatalysis for oxygen evolu-

tion will play an important role in many techniques for those
renewable energy systems that require converting renewable
energy (e.g., solar energy) to storable chemical fuels (e.g., oxygen)
or employing clean energy. The key, of course, is to design
electrocatalysts as efficiently as possible which could be used
in practice. This review has surveyed not only experimental work
in this field but also the fundamental understanding of the OER.
The basic concept of electrokinetics with suggested measure-
ment criteria for OER was also mentioned and discussed. In
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

addition, key developments of electrocatalysts including metal

oxides and non-metal oxide materials that have been reported
Fig. 16 (a) Standard free energies (U = 0) of each OER step for an ideal recently were reviewed. Several in situ experimental studies have
catalyst, LaMnO3, SrCoO3 and LaCuO3 (diagrams following the described provided vital information on the OER mechanism, and the
sequence). Reproduced with permission from ref. 107. Copyright r 2011, experimental (Tafel slope analysis) and theoretical (DFT calcula-
Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. (b) The theoretical
tion) standpoints are discussed as well.
overpotential (Z) vs. standard free energy DG0O  DG0HO eV of several
investigated perovskite compounds. A volcano plot to illustrate the
There has been great interest in metal oxide materials for
activity trend was based on a scaling relation between DG0HOO  DG0O OER since these metal oxide electrocatalysts generally exhibit
and DG0O  DG0HO . Reproduced with permission from ref. 101. Copyright excellent catalytic abilities for oxygen evolution. Various metal
r 2011, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. (c) A schematic oxides belonging to the perovskite, spinel, layer-structure type
illustration of a single binding site (top) and two binding sites (bottom) for families and others allow us to interrogate the electronic state
RuO2 and Ni doped RuO2, respectively. Reproduced with permission from (oxidation number) and structural influence (coordination
ref. 108. Copyright r 2014, The Royal Society of Chemistry.
environment) of the active center for OER. These works
have established key foundations in our perspective of OER.
activity of various perovskite samples are in good agreement However, the drawback for metal oxide electrocatalysts is
with experiment data as well. However, the scaling relationship usually their poor electroconductivity. Non-metal oxide systems
set an inherent theoretical limitation: the overpotentials of with superior conductivity such as metal chalcogenides and
the OER electrocatalysts could not be lower than 0.37 V. As pnictides have been employed and showed excellent activities
suggested by Halck et al.,108 modification of the active site for both HER and OER, which resulted in very promising
(through doping) for the surface active site (linear intermediate, potentials for bifunctional electrocatalysts toward the overall
*OH or *OOH, Fig. 16(c)) to become a dimensional active site water-splitting reaction. Notably, numerous reports have sug-
(bridged intermediate, m-OH or m-OOH) may be a possible path gested that an in situ formed thin oxide layer on the surface was
to break the scaling–relationship limit. Garcı́-Mota et al.109 and responsible for metal chalcogenides and pnictides to perform
Bajdich et al.110 have undertaken considerable theoretical work the OER, while the materials themselves acted as conducting
into OER activity on Co3O4 and b-CoOOH. They suggested that scaffolds. Organometallic complexes are excellent systems to
the potential-determining step for both compounds could study for OER activity due to their flexibility in designing their
be the formation of peroxide intermediate (eqn (5.3)) on the molecular and electronic structure to study the OER mecha-
surface. Bajdich et al.110 also explored the doping influence in nism. In addition, non-metal materials bear great potential in
b-CoOOH, and incorporating Ni and Fe in b-CoOOH could electrocatalysis for the oxygen evolution reaction in practice
lower the overpotentials of those doped compounds for OER. since no metal is needed in this kind of materials.
OER mechanism analysis from theoretical work and experi- Because of the contributions from the aforementioned work,
mental results (in situ experiments and Tafel analyses) are detailed and thorough information on these materials with mecha-
complementary to each other. Theoretical calculations (DFT) nistic studies of the OER offers fresh perspectives. Designing
could provide information toward the potential-determining highly efficient, stable, cost-effective and environmentally-
step (overpotential (Z)) while the Tafel analysis could offer friendly electrocatalysts is the goal for electrocatalysis of OER.
kinetic information on electrocatalysts for OER. Moreover, In light of these works, an efficient electrocatalyst should be a
DFT calculations have demonstrated great potential especially good electroconductor and possesses moderate adsorption
for resolving active site issues of those doping systems such as energy to oxygen intermediates such as peroxide and super-
Ni1xFexOOH, thereby providing a fresh perspective to improve oxide intermediates on the surface of the electrocatalyst (not
or to design advanced electrocatalysts for OER. too strong or too weak binding energy to these oxygen inter-
mediates). In addition, a good redox (ability to lose or to gain
electrons) center may be crucial for being an effective electro-
6. Conclusions and future perspectives catalyst (unity of eg) and high active sites would be desirable for
catalytic ability. Moreover, it is important for electrocatalysts to
Searching for a renewable energy system to replace the carbon- have the ability to resist corrosion in electrolytes especially
based fuel energy system is always the goal to fulfill sustainable under high anodic potential (i.e., without phase transformation

360 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

or the transformed phase is stable and will not dissolve in may have the potential to realize the purpose of in situ
electrolyte). Based on these criteria, several key strategies could monitoring. Most importantly, many other fields can take the
be mapped out to design a good electrocatalyst for OER. First, advantages of the related progress in present water oxidation
the electroconductivity of the electrocatalyst has to be high. The reaction because of the similar operating concepts in other
electroconductivity of electrocatalysts is important, since one electrocatalytic and/or photocatalytic systems.
responsibility of electrocatalysts is to transfer charge-carriers
from the metal electrode to adsorbed intermediates. Several
approaches such as introducing foreign elements, mixing with Abbreviations
conductive material, and changing morphology show great CFP Carbon fiber paper
potential to improve the electroconductivity, which may not
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

CNT Carbon nanotube

only increase electroconductivity of the electrocatalysts but also FTO Fluorine doped tin oxide
enhance the catalytic ability from other aspects. Second, the GCE Glassy carbon electrode
electrocatalysts should be able to endure redox reactions at HER Hydrogen evolution reaction
high anodic potential in electrolyte without phase degradation. ITO Indium tin oxide
It is due to the fact that in the process of OER, those electro- LDH Layered double hydroxides
catalysts may transform into other phases and dissolve in MWCNT Multi-walled carbon nanotube
electrolyte or gradually strip from the electrode during OER. NP Nanoparticle
To improve the stability of electrocatalysts, incorporating other NS Nanosheet
elements in electrocatalysts (i.e. mixing Ir into RuO2 to form OER Oxygen evolution reaction
RuxIr1xO2) has been demonstrated to be a good strategy to SWCNT Single-walled carbon nanotube
prolong the lifetime of electrocatalysts. Additionally, the decay
of OER electrocatalysts could be attributed to the phase trans-
formation that resulted in the exfoliation of newly formed Acknowledgements
phases from the electrocatalyst (due to lattice mismatch). In
this case, one can deposit a buffer layer between the electrolyte Prof. Hao Ming Chen acknowledge support from the Ministry of
and electrocatalyst to minimize the stripping issue of the newly Science and Technology, Taiwan (Contract No. MOST 104-2113-
formed phase to enhance its OER durability. M-213-001 and MOST 104-2113-M-002-011-MY2). Prof. Yi-Jun
Doping foreign elements into structures can change the Xu acknowledges support from the Key Project of National
electronic structure of the materials and enhance electro- Natural Science Foundation of China (U1463204), the project
conductivity. In addition to the improvement of electroconduc- of NSFC (21173045, 20903023), the Award Program for Minjiang
tivity, incorporating other elements in the structure also have a Scholar Professorship, the NSF of Fujian Province for Distin-
profound effect on altering the binding strength of those inter- guished Young Investigator Grant (2012J06003), the Indepen-
mediates (oxo, peroxide and superoxide) thereby leading to dent Research Project of the State Key Laboratory of
optimized OER activity. Mixing with other conductive materials Photocatalysis on Energy and Environment (No. 2014A05), the
like graphene or carbon nanotubes to become composite 1st Program of Fujian Province for Top Creative Young Talents,
electrocatalysts has also been proven to be an effective approach and the Program for Returned High-Level Overseas Chinese
to facilitate charge-transfer behavior of electrocatalysts (extrinsi- Scholars of Fujian Province.
cally), which shows great potential especially for organometallic
complexes. These complexes could also be attached to graphene Notes and references
on the surface of an electrode through dedicated design of
organometallic complexes, in which a fast charge-transfer rate 1 S. Chu and A. Majumdar, Nature, 2012, 488, 294–303.
could be foreseen owing to strong p–p interactions. Fabrication 2 T. R. Cook, D. K. Dogutan, S. Y. Reece, Y. Surendranath,
based on the morphology of the material may improve its T. S. Teets and D. G. Nocera, Chem. Rev., 2010, 110,
inherent electroconductivity (in the case of Ni3N from bulk to 6474–6502.
nanosheets for instance) of the material as well, and this can 3 E. E. Benson, C. P. Kubiak, A. J. Sathrum and J. M. Smieja,
also increase the surface area and expose more active sites Chem. Soc. Rev., 2009, 38, 89–99.
(exfoliation process) to the electrolyte, which accordingly 4 J. A. Turner, Science, 2004, 305, 972–974.
enhances the electrocatalytic ability for OER. 5 K. Zeng and D. Zhang, Prog. Energy Combust. Sci., 2010, 36,
Notably, from the viewpoint of materials design, achieving a 307–326.
perfect electrocatalyst strongly relies on the thoughtful under- 6 S. Marini, P. Salvi, P. Nelli, R. Pesenti, M. Villa, M. Berrettoni,
standing of the mechanism toward reactions. For this reason, G. Zangari and Y. Kiros, Electrochim. Acta, 2012, 82, 384–391.
we have to emphasize here that the development of in situ 7 L. Carrette, K. A. Friedrich and U. Stimming, Fuel Cells,
approaches is paramount, since they can possibly guide us 2001, 1, 5–39.
efficiently to our goal. In this respect, some techniques such 8 A. S. Aricò, S. Srinivasan and V. Antonucci, Fuel Cells, 2001,
as scanning probe microscopy, ambient pressure X-ray photo- 1, 133–161.
electron spectroscopy and soft X-ray absorption spectroscopy 9 M. R. Palacin, Chem. Soc. Rev., 2009, 38, 2565–2575.

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 361
 View Article Online

Review Article Chem Soc Rev

10 J. Suntivich, H. A. Gasteiger, N. Yabuuchi, H. Nakanishi, 38 T. Gao, Z. Jin, M. Liao, J. Xiao, H. Yuan and D. Xiao,
J. B. Goodenough and Y. Shao-Horn, Nat. Chem., 2011, 3, J. Mater. Chem. A, 2015, 3, 17763–17770.
546–550. 39 M. Huynh, D. K. Bediako and D. G. Nocera, J. Am. Chem.
11 11F. Cheng and J. Chen, Chem. Soc. Rev., 2012, 41, Soc., 2014, 136, 6002–6010.
2172–2192. 40 D. K. Bediako, Y. Surendranath and D. G. Nocera, J. Am.
12 R. Frydendal, E. A. Paoli, B. P. Knudsen, B. Wickman, Chem. Soc., 2013, 135, 3662–3674.
P. Malacrida, I. E. L. Stephens and I. Chorkendorff, Chem- 41 J. S. Chen, J. Ren, M. Shalom, T. Fellinger and M. Antonietti,
ElectroChem, 2014, 1, 2075–2081. ACS Appl. Mater. Interfaces, 2016, 8, 5509–5516.
13 Y. Lee, J. Suntivich, K. J. May, E. E. Perry and Y. Shao-Horn, 42 P. Ganesan, M. Prabu, J. Sanetuntikul and S. Shanmugam,
J. Phys. Chem. Lett., 2012, 3, 399–404. ACS Catal., 2015, 5, 3625–3637.
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

14 T. Reier, M. Oezaslan and P. Strasser, ACS Catal., 2012, 2, 43 W. Zhu, X. Yue, W. Zhang, S. Yu, Y. Zhang, J. Wang and
1765–1772. J. Wang, Chem. Commun., 2016, 52, 1486–1489.
15 M. Vuković, J. Appl. Electrochem., 1987, 17, 737–745. 44 B. Dong, X. Zhao, G.-Q. Han, X. Li, X. Shang, Y.-R. Liu,
16 R. Kötz, H. J. Lewerenz and S. Stucki, J. Electrochem. Soc., W.-H. Hu, Y.-M. Chai, H. Zhao and C.-G. Liu, J. Mater.
1983, 130, 825–829. Chem. A, 2016, 4, 13499–13508.
17 E. Antolini, ACS Catal., 2014, 4, 1426–1440. 45 C. Xia, Q. Jiang, C. Zhao, M. N. Hedhili and H. N. Alshareef,
18 R. Kötz, H. Neff and S. Stucki, J. Electrochem. Soc., 1984, Adv. Mater., 2016, 28, 77–85.
131, 72–77. 46 Z. Wang, J. Li, X. Tian, X. Wang, Y. Yu, K. A. Owusu, L. He
19 S. Cherevko, S. Geiger, O. Kasian, N. Kulyk, J.-P. Grote, A. Savan, and L. Mai, ACS Appl. Mater. Interfaces, 2016, 8, 19386–19392.
B. R. Shrestha, S. Merzlikin, B. Breitbach, A. Ludwig and 47 M.-R. Gao, Y.-F. Xu, J. Jiang, Y.-R. Zheng and S.-H. Yu,
K. J. J. Mayrhofer, Catal. Today, 2016, 262, 170–180. J. Am. Chem. Soc., 2012, 134, 2930–2933.
20 Y. Matsumoto, S. Yamada, T. Nishida and E. Sato, 48 Y.-R. Zheng, M.-R. Gao, Q. Gao, H.-H. Li, J. Xu, Z.-Y. Wu
J. Electrochem. Soc., 1980, 127, 2360–2364. and S.-H. Yu, Small, 2015, 11, 182–188.
21 J. O. M. Bockris and T. Otagawa, J. Electrochem. Soc., 1984, 49 M. Liao, G. Zeng, T. Luo, Z. Jin, Y. Wang, X. Kou and
131, 290–302. D. Xiao, Electrochim. Acta, 2016, 194, 59–66.
22 J. Suntivich, K. J. May, H. A. Gasteiger, J. B. Goodenough 50 Y. Liu, H. Cheng, M. Lyu, S. Fan, Q. Liu, W. Zhang, Y. Zhi,
and Y. Shao-Horn, Science, 2011, 334, 1383–1385. C. Wang, C. Xiao, S. Wei, B. Ye and Y. Xie, J. Am.
23 A. Vojvodic and J. K. Nørskov, Science, 2011, 334, 1355–1356. Chem. Soc., 2014, 136, 15670–15675.
24 M. R. G. De Chialvo and A. C. Chialvo, Electrochim. Acta, 51 X. Xu, F. Song and X. Hu, Nat. Commun., 2016, 7, 12324.
1993, 38, 2247–2252. 52 A. T. Swesi, J. Masud and M. Nath, Energy Environ. Sci.,
25 N. K. Singh and R. N. Singh, Indian J. Chem., 1999, 38A, 5. 2016, 9, 1771–1782.
26 I. Nikolov, R. Darkaoui, E. Zhecheva, R. Stoyanova, N. Dimitrov 53 L.-L. Feng, G. Yu, Y. Wu, G.-D. Li, H. Li, Y. Sun, T. Asefa,
and T. Vitanov, J. Electroanal. Chem., 1997, 429, 157–168. W. Chen and X. Zou, J. Am. Chem. Soc., 2015, 137, 14023–14026.
27 R. Subbaraman, D. Tripkovic, K.-C. Chang, D. Strmcnik, 54 Y. Liu, C. Xiao, M. Lyu, Y. Lin, W. Cai, P. Huang, W. Tong,
A. P. Paulikas, P. Hirunsit, M. Chan, J. Greeley, V. Stamenkovic Y. Zou and Y. Xie, Angew. Chem., Int. Ed., 2015, 127,
and N. M. Markovic, Nat. Mater., 2012, 11, 550–557. 11383–11387.
28 M. Li, Y. Xiong, X. Liu, X. Bo, Y. Zhang, C. Han and L. Guo, 55 J. Yang, G. Zhu, Y. Liu, J. Xia, Z. Ji, X. Shen and S. Wu, Adv.
Nanoscale, 2015, 7, 8920–8930. Funct. Mater., 2016, 26, 4712–4721.
29 H. Liang, F. Meng, M. Cabán-Acevedo, L. Li, A. Forticaux, 56 I. G. McKendry, A. C. Thenuwara, J. Sun, H. Peng,
L. Xiu, Z. Wang and S. Jin, Nano Lett., 2015, 15, 1421–1427. J. P. Perdew, D. R. Strongin and M. J. Zdilla, ACS Catal.,
30 Z. Lu, H. Wang, D. Kong, K. Yan, P.-C. Hsu, G. Zheng, 2016, 6, 7393–7397.
H. Yao, Z. Liang, X. Sun and Y. Cui, Nat. Commun., 2014, 57 S. Dou, L. Tao, J. Huo, S. Wang and L. Dai, Energy Environ.
5, 4345. Sci., 2016, 9, 1320–1326.
31 F. Song and X. Hu, Nat. Commun., 2014, 5, 4477. 58 T. Liu, Y. Liang, Q. Liu, X. Sun, Y. He and A. M. Asiri,
32 M. W. Kanan and D. G. Nocera, Science, 2008, 321, Electrochem. Commun., 2015, 60, 92–96.
1072–1075. 59 P. Chen, K. Xu, Z. Fang, Y. Tong, J. Wu, X. Lu, X. Peng,
33 M. W. Kanan, Y. Surendranath and D. G. Nocera, Chem. H. Ding, C. Wu and Y. Xie, Angew. Chem., Int. Ed., 2015, 54,
Soc. Rev., 2009, 38, 109–114. 14710–14714.
34 D. Jeong, K. Jin, S. E. Jerng, H. Seo, D. Kim, S. H. Nahm, 60 Y. Zhang, B. Ouyang, J. Xu, G. Jia, S. Chen, R. S. Rawat and
S. H. Kim and K. T. Nam, ACS Catal., 2015, 5, 4624–4628. H. J. Fan, Angew. Chem., Int. Ed., 2016, 55, 8670–8674.
35 Y. Liu, H. Wang, D. Lin, C. Liu, P.-C. Hsu, W. Liu, W. Chen 61 M. Liu and J. Li, ACS Appl. Mater. Interfaces, 2016, 8,
and Y. Cui, Energy Environ. Sci., 2015, 8, 1719–1724. 2158–2165.
36 M. Xing, L.-B. Kong, M.-C. Liu, L.-Y. Liu, L. Kang and 62 J. Ryu, N. Jung, J. H. Jang, H.-J. Kim and S. J. Yoo, ACS
Y.-C. Luo, J. Mater. Chem. A, 2014, 2, 18435–18443. Catal., 2015, 5, 4066–4074.
37 Z.-P. Nie, D.-K. Ma, G.-Y. Fang, W. Chen and S.-M. Huang, 63 L.-A. Stern, L. Feng, F. Song and X. Hu, Energy Environ. Sci.,
J. Mater. Chem. A, 2016, 4, 2438–2444. 2015, 8, 2347–2351.

362 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

64 P. Chen, K. Xu, Y. Tong, X. Li, S. Tao, Z. Fang, W. Chu, 90 H. S. Ahn and A. J. Bard, J. Am. Chem. Soc., 2015, 137,
X. Wu and C. Wu, Inorg. Chem. Front., 2016, 3, 236–242. 612–615.
65 K. Xu, P. Chen, X. Li, Y. Tong, H. Ding, X. Wu, W. Chu, Z. Peng, 91 H. S. Ahn and A. J. Bard, Anal. Chem., 2015, 87, 12276–12280.
C. Wu and Y. Xie, J. Am. Chem. Soc., 2015, 137, 4119–4125. 92 H. S. Ahn and A. J. Bard, J. Am. Chem. Soc., 2016, 138,
66 X.-Y. Yu, Y. Feng, B. Guan, X. W. Lou and U. Paik, Energy 313–318.
Environ. Sci., 2016, 9, 1246–1250. 93 J. Y. C. Chen, L. Dang, H. Liang, W. Bi, J. B. Gerken, S. Jin,
67 C. G. Read, J. F. Callejas, C. F. Holder and R. E. Schaak, ACS E. E. Alp and S. S. Stahl, J. Am. Chem. Soc., 2015, 137,
Appl. Mater. Interfaces, 2016, 8, 12798–12803. 15090–15093.
68 S. Fu, C. Zhu, J. Song, M. H. Engelhard, X. Li, D. Du and 94 C.-W. Tung, Y.-Y. Hsu, Y.-P. Shen, Y. Zheng, T.-S. Chan,
Y. Lin, ACS Energy Lett., 2016, 1, 792–796. H.-S. Sheu, Y.-C. Cheng and H. M. Chen, Nat. Commun.,
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

69 B. You, N. Jiang, M. Sheng, M. W. Bhushan and Y. Sun, ACS 2015, 6, 8106.

Catal., 2016, 6, 714–721. 95 A. Bergmann, E. Martinez-Moreno, D. Teschner, P. Chernev,
70 L. Jiao, Y.-X. Zhou and H.-L. Jiang, Chem. Sci., 2016, 7, M. Gliech, J. F. de Araujo, T. Reier, H. Dau and P. Strasser,
1690–1695. Nat. Commun., 2015, 6, 8625.
71 D. Li, H. Baydoun, C. N. Verani and S. L. Brock, J. Am. 96 H.-Y. Wang, Y.-Y. Hsu, R. Chen, T.-S. Chan, H. M. Chen
Chem. Soc., 2016, 138, 4006–4009. and B. Liu, Adv. Energy Mater., 2015, 5, 1500091.
72 J. Chang, Y. Xiao, M. Xiao, J. Ge, C. Liu and W. Xing, ACS 97 D. F. Abbott, D. Lebedev, K. Waltar, M. Povia, M. Nachtegaal,
Catal., 2015, 5, 6874–6878. E. Fabbri, C. Copéret and T. J. Schmidt, Chem. Mater., 2016, 28,
73 H. Lei, C. Liu, Z. Wang, Z. Zhang, M. Zhang, X. Chang, 6591–6604.
W. Zhang and R. Cao, ACS Catal., 2016, 6, 6429–6437. 98 D. Wang, J. Zhou, Y. Hu, J. Yang, N. Han, Y. Li and
74 B. Wurster, D. Grumelli, D. Hötger, R. Gutzler and K. Kern, T.-K. Sham, J. Phys. Chem. C, 2015, 119, 19573–19583.
J. Am. Chem. Soc., 2016, 138, 3623–3626. 99 W. Ma, R. Ma, C. Wang, J. Liang, X. Liu, K. Zhou and
75 J. Sun, H. Yin, P. Liu, Y. Wang, X. Yao, Z. Tang and T. Sasaki, ACS Nano, 2015, 9, 1977–1984.
H. Zhao, Chem. Sci., 2016, 7, 5640–5646. 100 J. O. M. Bockris, J. Chem. Phys., 1956, 24, 817–827.
76 S. W. Sheehan, J. M. Thomsen, U. Hintermair, 101 A. Damjanovic, A. Dey and J. O. M. Bockris, Electrochim.
R. H. Crabtree, G. W. Brudvig and C. A. Schmuttenmaer, Acta, 1966, 11, 791–814.
Nat. Commun., 2015, 6, 6469. 102 B. E. Conway and M. Salomon, Electrochim. Acta, 1964, 9,
77 M. Okamura, M. Kondo, R. Kuga, Y. Kurashige, T. Yanai, 1599–1615.
S. Hayami, V. K. K. Praneeth, M. Yoshida, K. Yoneda, 103 R. L. Doyle and M. E. G. Lyons, J. Electrochem. Soc., 2013,
S. Kawata and S. Masaoka, Nature, 2016, 530, 465–468. 160, H142–H154.
78 N. Zhang, M.-Q. Yang, S. Liu, Y. Sun and Y.-J. Xu, Chem. 104 T. Shinagawa, A. T. Garcia-Esparza and K. Takanabe, Sci.
Rev., 2015, 115, 10307–10377. Rep., 2015, 5, 13801.
79 E. Mirzakulova, R. Khatmullin, J. Walpita, T. Corrigan, 105 J. K. Nørskov, J. Rossmeisl, A. Logadottir, L. Lindqvist,
N. M. Vargas-Barbosa, S. Vyas, S. Oottikkal, S. F. Manzer, J. R. Kitchin, T. Bligaard and H. Jónsson, J. Phys. Chem. B,
C. M. Hadad and K. D. Glusac, Nat. Chem., 2012, 4, 794–801. 2004, 108, 17886–17892.
80 Y. Cheng, C. Xu, L. Jia, J. D. Gale, L. Zhang, C. Liu, 106 J. Rossmeisl, Z. W. Qu, H. Zhu, G. J. Kroes and
P. K. Shen and S. P. Jiang, Appl. Catal., B, 2015, 163, 96–104. J. K. Nørskov, J. Electroanal. Chem., 2007, 607, 83–89.
81 Y. Zhao, Y. Ding, Y. Li, L. Peng, H. R. Byon, J. B. Goodenough 107 I. C. Man, H.-Y. Su, F. Calle-Vallejo, H. A. Hansen, J. I. Martı́nez,
and G. Yu, Chem. Soc. Rev., 2015, 44, 7968–7996. N. G. Inoglu, J. Kitchin, T. F. Jaramillo, J. K. Nørskov and
82 J. Zhang, Z. Zhao, Z. Xia and L. Dai, Nat. Nanotechnol., J. Rossmeisl, ChemCatChem, 2011, 3, 1159–1165.
2015, 10, 444–452. 108 N. B. Halck, V. Petrykin, P. Krtil and J. Rossmeisl, Phys.
83 X. Lu, W.-L. Yim, B. H. R. Suryanto and C. Zhao, J. Am. Chem. Chem. Phys., 2014, 16, 13682–13688.
Chem. Soc., 2015, 137, 2901–2907. 109 M. Garcı́a-Mota, M. Bajdich, V. Viswanathan, A. Vojvodic,
84 S. Chen, J. Duan, M. Jaroniec and S.-Z. Qiao, Adv. Mater., A. T. Bell and J. K. Nørskov, J. Phys. Chem. C, 2012, 116,
2014, 26, 2925–2930. 21077–21082.
85 J. Tian, Q. Liu, A. M. Asiri, K. A. Alamry and X. Sun, 110 M. Bajdich, M. Garcı́a-Mota, A. Vojvodic, J. K. Nørskov and
ChemSusChem, 2014, 7, 2125–2130. A. T. Bell, J. Am. Chem. Soc., 2013, 135, 13521–13530.
86 T. Y. Ma, S. Dai, M. Jaroniec and S. Z. Qiao, Angew. Chem., 111 A. J. Bard and L. R. Faulkner, Electrochemical methods:
Int. Ed., 2014, 53, 7281–7285. fundamentals and applications, Wiley, 1980.
87 M. Zhang, M. de Respinis and H. Frei, Nat. Chem., 2014, 6, 112 J. O. Bockris, A. K. N. Reddy and M. Gamboa-Aldeco,
362–367. Modern Electrochemistry 2A. Fundamentals of Electrodics,
88 M. W. Louie and A. T. Bell, J. Am. Chem. Soc., 2013, 135, Kluwer Academic, New York, 2000.
12329–12337. 113 J. Tafel, Z. Phys. Chem., 1905, 50, 641.
89 B. J. Trześniewski, O. Diaz-Morales, D. A. Vermaas, 114 G. T. Burstein, Corros. Sci., 2005, 47, 2858–2870.
A. Longo, W. Bras, M. T. M. Koper and W. A. Smith, 115 J. O. M. Bockris and A. K. N. Reddy, Modern Electro-
J. Am. Chem. Soc., 2015, 137, 15112–15121. chemistry, Plenum, New York, 1970.

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 363
 View Article Online

Review Article Chem Soc Rev

116 R. Guidelli, G. Compton Richard, M. Feliu Juan, E. Gileadi, 140 L. Trotochaud, S. L. Young, J. K. Ranney and
J. Lipkowski, W. Schmickler and S. Trasatti, Pure Appl. S. W. Boettcher, J. Am. Chem. Soc., 2014, 136, 6744–6753.
Chem., 2014, 86, 4. 141 H. Bode, K. Dehmelt and J. Witte, Electrochim. Acta, 1966,
117 C. C. L. McCrory, S. Jung, I. M. Ferrer, S. M. Chatman, 11, 1079–1087.
J. C. Peters and T. F. Jaramillo, J. Am. Chem. Soc., 2015, 137, 142 S. Klaus, Y. Cai, M. W. Louie, L. Trotochaud and A. T. Bell,
4347–4357. J. Phys. Chem. C, 2015, 119, 7243–7254.
118 L. C. Seitz, C. F. Dickens, K. Nishio, Y. Hikita, J. Montoya, 143 Y.-F. Li and A. Selloni, ACS Catal., 2014, 4, 1148–1153.
A. Doyle, C. Kirk, A. Vojvodic, H. Y. Hwang, J. K. Norskov 144 D. Friebel, M. W. Louie, M. Bajdich, K. E. Sanwald,
and T. F. Jaramillo, Science, 2016, 353, 1011–1014. Y. Cai, A. M. Wise, M.-J. Cheng, D. Sokaras, T.-C. Weng,
119 A. T. Marshall and R. G. Haverkamp, Electrochim. Acta, R. Alonso-Mori, R. C. Davis, J. R. Bargar, J. K. Nørskov,
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

2010, 55, 1978–1984. A. Nilsson and A. T. Bell, J. Am. Chem. Soc., 2015, 137,
120 L.-E. Owe, M. Tsypkin, K. S. Wallwork, R. G. Haverkamp 1305–1313.
and S. Sunde, Electrochim. Acta, 2012, 70, 158–164. 145 X. Long, J. Li, S. Xiao, K. Yan, Z. Wang, H. Chen and
121 E. Mayousse, F. Maillard, F. Fouda-Onana, O. Sicardy and S. Yang, Angew. Chem., Int. Ed., 2014, 53, 7584–7588.
N. Guillet, Int. J. Hydrogen Energy, 2011, 36, 10474–10481. 146 M. S. Burke, M. G. Kast, L. Trotochaud, A. M. Smith and
122 F. I. Mattos-Costa, P. de Lima-Neto, S. A. S. Machado and S. W. Boettcher, J. Am. Chem. Soc., 2015, 137, 3638–3648.
L. A. Avaca, Electrochim. Acta, 1998, 44, 1515–1523. 147 O. Diaz-Morales, I. Ledezma-Yanez, M. T. M. Koper and
123 T. Audichon, T. W. Napporn, C. Canaff, C. Morais, F. Calle-Vallejo, ACS Catal., 2015, 5, 5380–5387.
C. Comminges and K. B. Kokoh, J. Phys. Chem. C, 2016, 148 B. Zhang, X. Zheng, O. Voznyy, R. Comin, M. Bajdich,
120, 2562–2573. M. Garcı́a-Melchor, L. Han, J. Xu, M. Liu, L. Zheng,
124 J. G. Lee, J. Hwang, H. J. Hwang, O. S. Jeon, J. Jang, F. P. Garcı́a de Arquer, C. T. Dinh, F. Fan, M. Yuan,
O. Kwon, Y. Lee, B. Han and Y.-G. Shul, J. Am. E. Yassitepe, N. Chen, T. Regier, P. Liu, Y. Li, P. De Luna,
Chem. Soc., 2016, 138, 3541–3547. A. Janmohamed, H. L. Xin, H. Yang, A. Vojvodic and
125 S. Yagi, I. Yamada, H. Tsukasaki, A. Seno, M. Murakami, E. H. Sargent, Science, 2016, 352, 333–337.
H. Fujii, H. Chen, N. Umezawa, H. Abe, N. Nishiyama and 149 M. S. Burke, S. Zou, L. J. Enman, J. E. Kellon, C. A. Gabor,
S. Mori, Nat. Commun., 2015, 6, 8249. E. Pledger and S. W. Boettcher, J. Phys. Chem. Lett., 2015, 6,
126 Y. Zhu, W. Zhou, J. Yu, Y. Chen, M. Liu and Z. Shao, Chem. 3737–3742.
Mater., 2016, 28, 1691–1697. 150 S. Zou, M. S. Burke, M. G. Kast, J. Fan, N. Danilovic and
127 R. G. Burns, Mineralogical applications of crystal field theory, S. W. Boettcher, Chem. Mater., 2015, 27, 8011–8020.
Cambridge University Press, 1993. 151 J.-X. Feng, S.-H. Ye, H. Xu, Y.-X. Tong and G.-R. Li, Adv.
128 H.-Y. Wang, S.-F. Hung, H.-Y. Chen, T.-S. Chan, Mater., 2016, 28, 4698–4703.
H. M. Chen and B. Liu, J. Am. Chem. Soc., 2016, 138, 36–39. 152 V. Augustyn and A. Manthiram, J. Phys. Chem. Lett., 2015, 6,
129 X. Liu, Z. Chang, L. Luo, T. Xu, X. Lei, J. Liu and X. Sun, 3787–3791.
Chem. Mater., 2014, 26, 1889–1895. 153 R. D. L. Smith, M. S. Prévot, R. D. Fagan, S. Trudel and C. P.
130 M. S. Al-Hoshan, J. P. Singh, A. M. Al-Mayouf, A. A. Berlinguette, J. Am. Chem. Soc., 2013, 135, 11580–11586.
Al-Suhybani and M. N. Shaddad, Int. J. Electrochem. Sci., 154 M. Dincă, Y. Surendranath and D. G. Nocera, Proc. Natl.
2012, 7, 15. Acad. Sci. U. S. A., 2010, 107, 10337–10341.
131 J. P. Singh, N. K. Singh and R. N. Singh, Int. J. Hydrogen 155 T. Takashima, K. Hashimoto and R. Nakamura, J. Am.
Energy, 1999, 24, 433–439. Chem. Soc., 2012, 134, 1519–1527.
132 S. Hirai, S. Yagi, A. Seno, M. Fujioka, T. Ohno and 156 M. Risch, V. Khare, I. Zaharieva, L. Gerencser, P. Chernev
T. Matsuda, RSC Adv., 2016, 6, 2019–2023. and H. Dau, J. Am. Chem. Soc., 2009, 131, 6936–6937.
133 J. A. Koza, Z. He, A. S. Miller and J. A. Switzer, Chem. Mater., 157 M. W. Kanan, J. Yano, Y. Surendranath, M. Dincă, V. K.
2012, 24, 3567–3573. Yachandra and D. G. Nocera, J. Am. Chem. Soc., 2010,
134 Z. Chen, C. X. Kronawitter and B. E. Koel, Phys. Chem. 132, 13692–13701.
Chem. Phys., 2015, 17, 29387–29393. 158 J. G. McAlpin, Y. Surendranath, M. Dincă, T. A. Stich,
135 M. Al-Mamun, X. Su, H. Zhang, H. Yin, P. Liu, H. Yang, S. A. Stoian, W. H. Casey, D. G. Nocera and R. D. Britt,
D. Wang, Z. Tang, Y. Wang and H. Zhao, Small, 2016, 12, J. Am. Chem. Soc., 2010, 132, 6882–6883.
2866–2871. 159 Y. Surendranath, M. W. Kanan and D. G. Nocera, J. Am.
136 T. Y. Ma, S. Dai, M. Jaroniec and S. Z. Qiao, J. Am. Chem. Chem. Soc., 2010, 132, 16501–16509.
Soc., 2014, 136, 13925–13931. 160 M. Chen, Y. Wu, Y. Han, X. Lin, J. Sun, W. Zhang and
137 T. Ling, D.-Y. Yan, Y. Jiao, H. Wang, Y. Zheng, X. Zheng, R. Cao, ACS Appl. Mater. Interfaces, 2015, 7, 21852–21859.
J. Mao, X.-W. Du, Z. Hu, M. Jaroniec and S.-Z. Qiao, Nat. 161 H. S. Ahn and T. D. Tilley, Adv. Funct. Mater., 2013, 23,
Commun., 2016, 7, 12876. 227–233.
138 T. Y. Ma, S. Dai and S. Z. Qiao, Mater. Today, 2016, 19, 162 D. K. Bediako, B. Lassalle-Kaiser, Y. Surendranath, J. Yano,
265–273. V. K. Yachandra and D. G. Nocera, J. Am. Chem. Soc., 2012,
139 D. A. Corrigan, J. Electrochem. Soc., 1987, 134, 377–384. 134, 6801–6809.

364 | Chem. Soc. Rev., 2017, 46, 337--365 This journal is © The Royal Society of Chemistry 2017
 View Article Online

Chem Soc Rev Review Article

163 I. G. McKendry, A. C. Thenuwara, J. Sun, H. Peng, E. Rauls, H. Aldahhak, W. G. Schmidt and A. Dey, Angew.
J. P. Perdew, D. R. Strongin and M. J. Zdilla, ACS Catal., Chem., Int. Ed., 2016, 55, 2350–2355.
2016, 6, 7393–7397. 178 A. M. Ullman, C. N. Brodsky, N. Li, S.-L. Zheng and
164 J. Jiang, S. Lu, H. Gao, X. Zhang and H.-Q. Yu, Nano Energy, D. G. Nocera, J. Am. Chem. Soc., 2016, 138, 4229–4236.
2016, 27, 526–534. 179 K. Qu, Y. Zheng, S. Dai and S. Z. Qiao, Nano Energy, 2016,
165 O. Mabayoje, A. Shoola, B. R. Wygant and C. B. Mullins, 19, 373–381.
ACS Energy Lett., 2016, 1, 195–201. 180 Y. Zhao, R. Nakamura, K. Kamiya, S. Nakanishi and
166 M.-R. Gao, X. Cao, Q. Gao, Y.-F. Xu, Y.-R. Zheng, J. Jiang K. Hashimoto, Nat. Commun., 2013, 4, 2390.
and S.-H. Yu, ACS Nano, 2014, 8, 3970–3978. 181 R. Borup, J. Meyers, B. Pivovar, Y. S. Kim, R. Mukundan,
167 F. Safizadeh, E. Ghali and G. Houlachi, Int. J. Hydrogen N. Garland, D. Myers, M. Wilson, F. Garzon, D. Wood,
Published on 13 January 2017. Downloaded by University of Warwick on 5/4/2021 2:20:51 PM.

Energy, 2015, 40, 256–274. P. Zelenay, K. More, K. Stroh, T. Zawodzinski, J. Boncella,

168 P. Xiao, W. Chen and X. Wang, Adv. Energy Mater., 2015, J. E. McGrath, M. Inaba, K. Miyatake, M. Hori, K. Ota,
5, 1500985. Z. Ogumi, S. Miyata, A. Nishikata, Z. Siroma, Y. Uchimoto,
169 M. Zeng and Y. Li, J. Mater. Chem. A, 2015, 3, 14942–14962. K. Yasuda, K.-i. Kimijima and N. Iwashita, Chem. Rev.,
170 J. Duan, S. Chen, A. Vasileff and S. Z. Qiao, ACS Nano, 2016, 2007, 107, 3904–3951.
10, 8738–8745. 182 T. Y. Ma, J. L. Cao, M. Jaroniec and S. Z. Qiao, Angew.
171 J. D. Blakemore, R. H. Crabtree and G. W. Brudvig, Chem. Chem., Int. Ed., 2016, 55, 1138–1142.
Rev., 2015, 115, 12974–13005. 183 L. Li, Y. Yu, G. J. Ye, Q. Ge, X. Ou, H. Wu, D. Feng, X. H. Chen
172 S. W. Gersten, G. J. Samuels and T. J. Meyer, J. Am. Chem. and Y. Zhang, Nat. Nanotechnol., 2014, 9, 372–377.
Soc., 1982, 104, 4029–4030. 184 H. O. H. Churchill and P. Jarillo-Herrero, Nat. Nanotechnol.,
173 A. R. Parent and K. Sakai, ChemSusChem, 2014, 7, 2014, 9, 330–331.
2070–2080. 185 Q. Jiang, L. Xu, N. Chen, H. Zhang, L. Dai and S. Wang,
174 H. Dau, C. Limberg, T. Reier, M. Risch, S. Roggan and Angew. Chem., Int. Ed., 2016, 55, 13849–13853.
P. Strasser, ChemCatChem, 2010, 2, 724–761. 186 B. Seo, Y. J. Sa, J. Woo, K. Kwon, J. Park, T. J. Shin,
175 M. Schulze, V. Kunz, P. D. Frischmann and F. Würthner, H. Y. Jeong and S. H. Joo, ACS Catal., 2016, 6, 4347–4355.
Nat. Chem., 2016, 8, 576–583. 187 Y. Deng, A. D. Handoko, Y. Du, S. Xi and B. S. Yeo,
176 D. R. Kauffman, D. Alfonso, D. N. Tafen, J. Lekse, C. Wang, ACS Catal., 2016, 6, 2473–2481.
X. Deng, J. Lee, H. Jang, J.-s. Lee, S. Kumar and 188 N. H. Chou, P. N. Ross, A. T. Bell and T. D. Tilley,
C. Matranga, ACS Catal., 2016, 6, 1225–1234. ChemSusChem, 2011, 4, 1566–1569.
177 W. Schöfberger, F. Faschinger, S. Chattopadhyay, S. Bhakta, 189 A. I. Krasil’shchikov, Zh. Fiz. Khim., 1963, 37, 7.
B. Mondal, J. A. A. W. Elemans, S. Müllegger, S. Tebi, 190 B. E. Conway and P. L. Bourgault, Can. J. Chem., 1962, 40,
R. Koch, F. Klappenberger, M. Paszkiewicz, J. V. Barth, 1690–1707.

This journal is © The Royal Society of Chemistry 2017 Chem. Soc. Rev., 2017, 46, 337--365 | 365