REVIEWS

Earth-abundant catalysts
for electrochemical and
photoelectrochemical water splitting
Isolda Roger, Michael A. Shipman and Mark D. Symes
Abstract | Sunlight is by far the most plentiful renewable energy resource, providing Earth with
enough power to meet all of humanity’s needs several hundred times over. However, it is both
diffuse and intermittent, which presents problems regarding how best to harvest this energy and
store it for times when the sun is not shining. Devices that use sunlight to split water into
hydrogen and oxygen could be one solution to these problems, because hydrogen is an excellent
fuel. However, if such devices are to become widely adopted, they must be cheap to produce and
operate. Therefore, the development of electrocatalysts for water splitting that comprise only
inexpensive, earth-abundant elements is critical. In this Review, we investigate progress towards
such electrocatalysts, with special emphasis on how they might be incorporated into photoelec-
trocatalytic water-splitting systems and the challenges that remain in developing these devices.

WestCHEM, School of
Chemistry, University of
Glasgow, Glasgow,
G12 8QQ, UK.
Correspondence to M.D.S.
mark.symes@glasgow.ac.uk
doi:10.1038/s41570-016-0003
Published online 11 Jan 2017
Renewable energy is plentiful but tends to be intermittent
and unpredictable. As our reliance on renewable energy
grows, there will be an increasing need to store this energy
for times when the sun is not shining and the wind is not
blowing 1. One of the more attractive options for storing
this energy is to use renewable power (as and when it is
available) to split water into hydrogen and oxygen accord-
ing to equation 1 (REF. 2). The hydrogen produced by the
reaction in equation 1 can be stored and then oxidized
(by burning in air or in a fuel cell) to release energy and
regenerate water, giving a carbon-neutral fuel production
and consumption cycle.
2H2O → O2 + 2H2 (1)
The reaction in equation 1 is a thermodynamically uphill
process and requires an energy input of 286 kJ mol−1 at
room temperature and pressure. Electrochemical meth-
ods are viewed as efficient and scalable means to effect this
conversion, which would simply require inputs of water
and electricity to produce hydrogen (see BOX 1 for an
introduction to the fundamentals of the electrochemical
water-splitting process).
Solar power is by far the largest source of renewable
energy (~1.2 × 1014 kJ are received at the Earth’s surface
every second)3. Therefore, it is perhaps natural that
there is considerable interest in the coupling of solar
irradiation to electrochemical water splitting 4. Two
general routes are available for this process: the use of
conventional solar panels to run conventional electro-
lysers (‘indirect’ solar‑to‑hydrogen production) or the
development of a single device that could perform both
light absorption and water splitting (‘direct’ solar‑to‑
hydrogen production). An indirect approach benefits
from the use of conventional, tried-and-tested tech-
nologies but suffers from efficiency losses owing to the
additional step involved (electricity is first generated
in the photovoltaic cell and is then consumed in a sub-
sequent electrolysis step in a separate device)5. Hence,
in recent years, there has been a veritable explosion of
activity targeting the development of direct solar‑to‑
fuels devices, in the hope that these will eventually
become more efficient than indirect methods6,7. Such
direct solar‑to‑fuels devices can be considered to per-
form ‘artificial photosynthesis’, whereby the energy of
sunlight is captured and stored in the form of chemical
bonds in a product ‘solar fuel’ (in this case, hydrogen)8.
FIGURE  1 shows schematics of two photoelectro-
chemical solar‑to‑hydrogen artificial photosynthesis
systems, wherein the light harvesting function might
be performed by a buried photovoltaic (FIG. 1a) or a
semiconductor film immediately underneath each
electrocatalyst (FIG.  1b). Practical and scalable artifi-
cial photosynthesis systems of this type will have some
fairly stringent materials requirements. First, the current
densities that can be obtained from these devices will be
limited by the power of the incident sunlight (0.1 W cm−2
at most), which means that realistic current densities are
on the order of 10 mA cm−2 for artificial photosynthesis
systems9 (compared with ~0.5–2 A cm−2 for commercial
electrolysers10). This means that to make the same amount

NATURE REVIEWS | CHEMISTRY VOLUME 1 | ARTICLE NUMBER 0003 | 1

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Tafel slopes
of gas per unit time, a photoelectrochemical solar‑to‑ electrolytes avoid many of the potential safety issues
Tafel slopes are the gradient of hydrogen device would require around 50–200 times the that could arise if more caustic electrolytes were used;
the linear portion of a graph of electrode area of a conventional electrolyser, making the considering that artificial photosynthesis systems will
overpotential versus the use of precious metal electrocatalysts in such artificial have to cover large areas to produce useful quantities of
logarithm of the current
photosynthesis systems economically unviable. Instead, fuel on practical timescales, it seems almost inevitable
density for a given
electrocatalyst performing a the electrocatalysts in artificial photosynthesis systems that leaks will occur.
particular reaction. Hence, the will have to be both very cheap and widely available, thus In this Review, we examine recent progress towards
value of each slope gives an restricting our choice to first row transition metals and the development of earth-abundant electrocatalysts for
idea as to how the their compounds11. water splitting in the context of their potential use in
electrocatalytic performance of
a material changes over a given
Conventional electrolysis works best at either very photoelectrochemical solar‑to‑hydrogen devices. The
potential range. These slopes high or very low pH, at which the concentration of use of ‘earth abundant’ excludes Re, Ru, Os, Rh, Ir, Pd,
are often quoted in millivolts charge carriers is greatest. However, the vast majority Pt, Ag and Au. We examine, in turn, electrocatalysts for
per decade, where a decade in of known semiconductors suitable for use as light- the hydrogen evolution reaction (HER) and the oxygen
this sense relates to a decade of
absorbers degrade rapidly at extreme pH values, evolution reaction (OER), before exploring examples of
current density on the log scale.
Catalysts with lower slopes meaning that neutral electrolytes are sought for many the use of particular catalysts in photoelectro­chemical
require smaller increments of artificial photosynthesis applications. Milder pH envi- cells for performing the HER and OER. Finally, we
overpotential to give increased ronments in artificial photosynthesis systems would present examples of complete and functional photo-
current densities, making them also slow the rate of degradation of other cell compo- electrochemical solar‑to‑hydrogen devices and critically
more effective.
nents (for example, gaskets, connections and membrane appraise the remaining challenges in this field.
separators), which is vital for long-term performance This Review does not intend to be exhaustive but
(and overall system cost) 12. Moreover, pH-neutral rather aims to present a snapshot of a rapidly changing
area in as concise a manner as possible. We hope that
the discussions here prove useful both to newcomers to
Box 1 | Fundamentals of electrochemical water splitting the field as well as to those already well-versed in the
topics under consideration (through the use of compar-
An archetypal electrolysis cell for water splitting is shown in the figure below. Under
ative tables and a focus on the more recent literature).
acidic conditions, water is oxidized at the anode according to the equation:
2H2O → O2 + 4e− + 4H+. This is known as the oxygen evolution reaction (OER). The
Readers desiring more in‑depth treatises on the electro-
electrons travel through the external circuit while the protons cross into the cathode catalytic HER13,14 and OER15,16 are directed elsewhere.
compartment (to complete the electrochemical circuit). The protons and electrons then Likewise, because we focus primarily on progression
combine at the cathode in the hydrogen evolution reaction (HER): 4e− + 4H+ → 2H2. towards devices rather than purely fundamental stud-
To drive the OER and HER (to split water), the minimum theoretical voltage requirement ies, we only discuss heterogeneous catalysts that could
is 1.23 V at room temperature. However, additional energy is required to make the easily be integrated into devices such as those shown
reactions proceed at appreciable rates (that is, activation energy). The faster the rate of in FIG. 1. Therefore, we do not discuss homogeneous
water splitting (measured as charge flowing per unit area of electrode per unit time, or or molecular systems for water splitting, purely photo­
‘current density’), the greater the activation energy that must be supplied. This additional catalytic systems, or systems in which either the HER or
energy manifests as a requirement for potential bias above the 1.23 V minimum. The term
OER can only be achieved through the use of sacrificial
‘overpotential’ (often given the symbol η) describes how much additional voltage must be
applied to obtain a given current density. The role of electrocatalysts is then to reduce
reagents. Readers interested in these topics are directed
this overpotential as far as possible. Under acidic conditions (in which the to some excellent reviews on these subjects17–19.
electrochemical charge carriers are protons, as in the figure below), the best
electrocatalysts are precious metals (platinum at the cathode and IrO2 or RuO2 at the Earth-abundant HER electrocatalysts
anode). Under basic conditions, the HER and OER proceed according to the equations The best known heterogeneous catalysts for the HER are
below, in which case first row transition metals (and their alloys) and oxides make based on platinum. Because platinum is rather expensive
excellent HER and OER catalysts, respectively. and rare, there has been considerable interest in hydro-
gen evolution electrocatalysts featuring more plentiful
Basic HER: 4H2O + 4e− → 2H2 + 4OH− elements. Under very alkaline conditions (30% KOH in
Basic OER: 4OH− → O2 + 2H2O + 4e−
water), first row transition metals, such as nickel, have
e– e–
long been known to be effective HER catalysts. At only
V
modest overpotentials, alloys of nickel and cobalt with
molybdenum, in particular, display high cathodic current
densities attributed to hydrogen production20–23 (TABLE 1,
OER HER entries 1–5).
O2 2H2 It is useful to consider some of the metrics used to
describe the activity of a particular catalyst for the OER
or HER. In terms of electrochemical characterization,
TABLES 1,2 give the overpotentials required to achieve
4e– + 4H+ 4H+ 4e– a given current density (chosen as −10 mA cm−2 for
the HER and 1 mA cm−2 for the OER) and Tafel slopes
2H2O as ways of comparing different catalysts. However, it
is clearly insufficient to assume that a given current is
Anode Membrane Cathode producing a particular gas — it has to be proven that
separator the products of the reaction are indeed those claimed
Nature Reviews | Chemistry

2 | ARTICLE NUMBER 0003 | VOLUME 1 www.nature.com/natrevchem

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

a b was subsequently explained by some of the same authors,

Path of protons
through electrolyte
to complete circuit
4e– + 4H+
e–
2H2
Hydrogen Buried
evolution photovoltaic
catalyst
Figure 1 | Artificial photosynthesis systems for solar‑to‑hydrogen
a | Wireless configuration in which proton flow through the electrolyte completes the
circuit. b | Wired configuration in which the electrical circuit is completed by external
wiring. In both cases, light must pass through the catalysts to be absorbed; therefore,
thin (near-transparent) catalyst layers are desirable.
Faradaic yield
In electrochemistry, this is the
ratio between the amount of
product actually detected and
quantified, and the amount of
that product that could
theoretically have been formed
based on the charge passed in
the experiment.
Exchange current densities
Exchange current densities
reflect the intrinsic rate of
electron transfer between an
analyte in solution and the
electrode, and can thus be
viewed as a measure of the
effectiveness of a catalyst for a
particular electrochemical
reaction under a particular set
of conditions (the greater the
magnitude of the exchange
current density, the greater the
activity of the catalyst).
Electron flow through
external circuit
4H+ + 4e– + O2 4e– + 4H+ 4H+ + 4e– + O2
Protons
2H2O 2H2 2H2O
Oxygen Hydrogen Nafion Oxygen
evolution evolution membrane evolution
catalyst catalyst Semiconductor catalyst
applications.
Nature Reviews | Chemistry
and then the Faradaic yield can be determined. Hence,
quantitative gas analyses are essential. For hydrogen,
gas chromatography is used, whereas for quantitative
oxygen detection there is a choice of gas chromatogra-
phy, fluorescence-quench methods or, in some cases,
oxygen-sensitive electrodes such as the Clark electrode.
However, TABLES 1,2 (which we believe to be repre-
sentative of the field as a whole) sufficiently show that
detection of the products of electrolysis does not tend
to be addressed very rigorously (‘NA’ in this context
indicates that no analysis of the electrolysis products
was reported). It is rather unsettling to note the small
number of studies of electrocatalytic water splitting
that have actually checked that hydrogen or oxygen are
indeed produced; therefore, readers are advised to apply
a certain level of scepticism in cases where new catalysts
for the OER or HER are reported without quantification
of oxygen or hydrogen.
Among the key inspirations behind the design of
new hydrogen evolution catalysts are the so‑called vol-
cano plots that correlate HER exchange current densities
(a measure of the effectiveness of a catalyst for the HER)
for various materials with the chemisorption energy of
hydrogen on those materials (FIG. 2). This trend was first
recorded for metals by Trasatti in the 1970s24 and has been
a fertile treasure map for new catalysts ever since25,26. Such
plots reveal that there is an optimum hydrogen binding
energy (in which hydrogen is held neither too strongly
nor too weakly) at which the maximum catalytic activity
is obtained. Perhaps unsurprisingly, platinum falls near
the apex of this volcano.
In 2005, Hinnemann et al.27 exploited this relation-
ship by using density functional theory (DFT) to iden-
tify materials in which the free energy change associated
with atomic hydrogen bonding to the catalyst surface was
close to zero. Their results suggested that MoS2 would
be a suitable candidate material (which was confirmed
experimentally), despite bulk MoS2 being known to be a
rather poor HER catalyst28. This apparent contradiction
who demonstrated that hydrogen evolution occurs pref-
erentially at edge sites on MoS2, with the basal plane being
much less active29,30. Since this discovery, various transi-
tion metal sulfides and selenides have been prepared by
numerous methods (often with a view to exposing as
many edge sites as possible) and shown to have activity
for the HER31–40 (TABLE 1).
Similarly, transition metal phosphides have shown
their worth as HER catalysts. Popczun et al.41 first
demonstrated that Ni2P was an effective hydrogen evo-
lution catalyst in acidic media by noting that both Ni2P
and MoS2 are common hydrodesulfurization catalysts.
The key similarity is that both hydrodesulfurization and
HER require catalysts that reversibly bind and cleave
hydrogen. Other phosphides have since been shown
to be active42–44. Meanwhile, transition metal carbides,
borides and nitrides are also in vogue, with some of these
materials functioning in both basic and acidic media45–50
as summarized in TABLE 1 (entries 21–29).
Very recently, Fan et al.51 obtained excellent metrics
for the HER using a transition metal carbide (Ni/C) cat-
alyst prepared by heating a nickel-based metal organic
framework at 973 K. This carbonization reaction led to
the generation of carbon nanoparticles dotted with iso-
lated nickel atoms. The activity of these materials for the
HER at pH 0 was within a few millivolts of that exhibited
by platinum for current densities less than 100 mA cm−2.
Catalysts with activities that rely on highly dispersed
clusters of relatively few atoms often suffer from rapid
deterioration in performance under constant potential
because of the tendency for isolated metal centres to
agglomerate. However, in this case the activity of the
catalysts was maintained over 25 hours of continuous
operation, which is a remarkable finding.
Interrogation of volcano-type plots allows new
catalysts for the HER to be proposed by combining
catalysts on opposite slopes of the volcano to produce
new materials with intermediate hydrogen binding
energy (and hence improved catalytic performance).
Lu et al.52 recently used this strategy and DFT to pre-
dict that copper surfaces doped with titanium (to pro-
duce adsorption sites consisting of two copper centres
and one titanium centre) would have hydrogen binding
energies very similar to that of platinum. Accordingly,
the authors synthesized a range of bimetallic copper–
titanium films by arc-melting methods and tested their
activity for electrochemical reduction reactions in 0.1 M
KOH. The optimal HER performance was obtained with
films of composition Cu95Ti5, which gave current den-
sities of −10 mA cm−2 at an overpotential of only 60 mV
(note the poor performance anticipated for copper and
titanium on their own, highlighted in green in FIG. 2).
By comparison, an overpotential of around 100 mV is
required for a commercial Pt/C catalyst to achieve the
same current density, suggesting that earth-abundant
HER catalysts can outperform platinum under certain
conditions. The key challenge now for earth-abundant
HER catalysts is to prove that such excellent perfor-
mance can be maintained in the long term (continuous
operation for tens to hundreds of thousands of hours).

NATURE REVIEWS | CHEMISTRY VOLUME 1 | ARTICLE NUMBER 0003 | 3

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

Earth-abundant OER electrocatalysts exhibited by undoped nickel oxides. This echoed work by
OER electrocatalysts based on mixed metal oxides. The Corrigan in the 1980s57, which showed that the presence
other half-reaction of water splitting is the OER. The of iron (at levels as low as 0.01%) in nickel oxide films
present record performance for OER electrocatalysis produced a significant enhancement in electrocatalysis
using earth-abundant elements is claimed by Merrill of the OER.
and Dougherty 53, who reported NiFe oxides (admittedly The effects that iron doping has on the OER activ-
on platinum substrates) that require only 30 mV over- ity of nickel oxides have been studied in some detail by
potential to perform the OER at 1 mA cm−2 (or 300 mV Bell and co‑workers58, who found that the optimal OER
overpotential at 500 mA cm−2) at pH 14. Excellent OER activity of NiFe oxides was obtained at a Fe:Ni compo-
activities for NiFe oxides in basic media have also been sition of 2:3. Through electrochemical and in situ X‑ray
reported by Lu and Zhao54, Gong et al.55 and Li and absorption spectroscopy measurements, it was found
co‑workers56 (TABLE 2, entries 1–4). In the latter work, that the composition of these iron‑doped nickel oxide
the authors noted that although low iron loadings gave films was a good fit to the formula Ni(ii)1 − xFe(iii)x
enhanced OER, higher loadings served to reduce the (OH)2(SO4)x/2(H2O)y (a layered double hydroxide struc-
performance of the resulting NiFe oxides below that ture) at the catalyst rest potential. It was also observed
that the potential of the Ni(ii)/Ni(iii) couple shifted
anodically when the films contained more iron59. Using
Table 1 | Examples of earth-abundant HER electrocatalysts computational methods, the authors suggested that
Entry Catalyst η at −10 mA cm−2 Tafel slope pH Faradaic Refs raising the potential into the zone in which the OER
material (mV) (mV per yield occurred caused oxidation of Ni(ii) to Ni(iii) (whereas
decade) iron remained ferric), but that the Fe(iii) centres in these
1 NiMo 200 (100 mA cm−2) 122 14.8 NA 20 oxidized films possessed lower overpotential require-
ments for the OER than the Ni(iii) centres; hence, the
2 CoMo 170 (100 mA cm ) −2
92 14.8 NA 20
Fe(iii) centres were considered to be the sites of oxygen
3 NiMo 185 (300 mA cm ) −2
112 14.8 NA 21 evolution. Nickel and iron appear to have privileged sta-
4 NiMo 70 (20 mA cm )−2
NA 14.3 NA 22 tus when it comes to producing highly active catalysts
5 NiMo 34 (20 mA cm−2) NA 14 NA 23 for the OER under basic conditions, and, indeed, almost
all of the leading catalysts for OER in alkaline media
6 MoS2 260 50 0 NA 30
include either one or both of these metals (TABLE 2). In
7 MoS2 200 (15 mA cm−2) 40 −0.3 100%* 31 this regard, mixed NiCo and NiMo oxides have also been
8 MoS2 ~150 41 0 NA 32 shown to be competent OER catalysts60,61.
9 MoS2 170 60 0.2 NA 33
A key result in this area with important implications
in electrochemical water splitting at high pH was sup-
10 CoS2 145 51 0 NA 34 plied by Boettcher and co‑workers62,63, who prepared
11 CoS2 ~175 93 7 100%‡ 35 nickel oxide from ostensibly pure nickel salt precursor
12 CoMoSx 250 85 7 ~100% §
36 solutions. However, on careful examination of their
films, the authors discovered that significant traces of
13 WS2 ~250 60 0 NA 37
iron were present in the most active of their materials.
14 CoSe2 90 39 0 NA 38 Indeed, when iron was rigorously excluded from the
15 MoS1.0Se1.0 ~200 56 0 100%* 39 catalyst deposition baths, the overpotential requirement
16 NiSe2 ~140 49 0 NA 40 to achieve a current density of 1 mA cm−2 for the OER
increased from 300 to 470 mV. Working backwards,
17 Ni2P 130 (20 mA cm−2) 46 0 100%* 41
the authors found an optimal activity for the OER at a
18 CoP 85 (20 mA cm )−2
50 0 100%* 42 loading of 25% iron in nickel oxides (in good agreement
19 FeP 55 38 0 100%§ 43 with Bell’s results above). However, Boettcher’s group
assigned the locus of the OER activity to the resulting
20 MoP 64 NA 0 100% §
44
highly oxidizing Ni4+/3+ couple, rather than to the iron
21 CoNx 170 75 14 NA 45 centres. The same group subsequently observed similar
22 CoNx 140 30 0 NA 45 behaviour when cobalt oxides were doped with iron64.
23 NiMoNx 225 (5 mA cm−2) 35.9 1 NA 46 Perhaps even more alarming for those working in the
field was a report in 2015 demonstrating that adventi-
24 Co0.6Mo1.4N2 200 NA 1 >90%* 47
tious nickel present in several commonly used electro-
25 α‑MoB ~225 (20 mA cm−2) 55 −0.3 100%‡ 48 lytes (such as phosphate, carbonate and borate salts)
26 Mo2C 130 53 0 NA 49 were capable of giving rise to OER current densities of
27 MoC 124 43 0 NA 50 1 mA cm−2 at only 400 mV overpotential, even in the
absence of any recognized catalyst 65. This throws into
28 MoC 77 50 14 NA 50
doubt the veracity of any work that reports an OER activ-
29 Ni/C 34 41 0 100%* 51 ity of 1 mA cm−2 or less under basic conditions (at 400 mV
30 Cu95Ti5 60 110 13 NA 52 overpotential) that does not explicitly state the electro-
η, overpotential; HER, hydrogen evolution reaction; NA, no analysis. *Quantified by measuring lyte source that was used (and its purity). Therefore,
volume of gases evolved. ‡Quantified by gas chromatography. §Quantification method not stated. some caution is advised when assessing proposed OER

4 | ARTICLE NUMBER 0003 | VOLUME 1 www.nature.com/natrevchem

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

catalysts that cannot achieve demonstrably superior
performance to this baseline.
The best catalytic OER performance of a base metal
catalyst that contains no nickel was reported in 2016
by Zhang et al.66 (TABLE 2, entry 9). These authors used
DFT to predict that mixed oxides of iron, cobalt and
tungsten could display optimal binding energies of the
potential OER intermediates *OH, *O and *OOH on the
catalyst surface (where * indicates a surface-bound spe-
cies). Accordingly, the authors synthesized FeCo oxides
doped with W(vi) ions by the very carefully controlled
hydrolysis of solutions of the simple metal salts. This
control avoided phase separation of the various compo-
nents in the resulting catalyst gels, which (after drying)
displayed excellent activity for the OER for periods of
continuous operation of over 500 hours. The general-
ity of this approach presents numerous opportunities
to discover further multimetal catalysts for the OER.
Shao-Horn and co‑workers67 have similarly used com-
putational methods to guide the design of OER catalysts,
leading them to identify and then experimentally verify
the mixed-metal perovskite Ba0.5Sr0.5Co0.8Fe0.2O3 − δ as
an active OER catalyst in basic media. An alternative
approach to the identification of mixed-metal OER cat-
alysts has been reported by Gregoire and co‑workers68,
who used a high-throughput screen to produce 5,456
different oxide compositions — containing the elements
nickel, iron, cobalt and cerium — by inkjet printing. The
best material identified (Ni0.2Co0.3Ce0.5Ox) displays a cur-
rent density for the OER of 10 mA cm−2 at only 310 mV
overpotential (at pH 14). Such screening methods can
be extremely useful for identifying effective catalysts, but
they do not allow the prediction of long-term stability,
which must then be established through more traditional
means. Nevertheless, it is expected that many more OER
catalysts will be identified through such methods in the
coming years.
Many of the OER catalysts discussed up to this point
are prepared either by electrodeposition or by precipita-
tion from solutions of metal salt precursors. However,

Table 2 | Examples of earth-abundant OER electrocatalysts

Entry Catalyst material η at 1 mA cm−2 (mV) Tafel slope pH Faradaic yield Refs
(mV per
decade)
1 NiFe oxides 30 15 14 NA 53
2 NiFe oxides ~120 32 15 NA 54
3 NiFe oxides 210 31 14 ~100%* 55
4 NiFe oxides ~180 40 14 NA 56
5 NiFe oxides 300 (10 mA cm )−2
40 14 NA 58
6 NiCo oxides 370 NA 13 NA 60
7 NiMo oxides 300 47 14 NA 61
8 Ni oxides (trace) 400 57 13 >90%* 65
9 FeCoW oxy-hydroxides 191 (10 mA cm )−2
NA 14 100%* 66
10 Ba0.5Sr0.5Co0.8Fe0.2O3 − δ ~320 ~60 13 NA 67
11 NiCoCe oxides 250 ~60 14 97%* 68
12 α-Fe33Co33Ni33Ox 230 30 13 NA 69
13 Co oxides 410 60 7 100% ‡
70
14 Co oxides 390 60 9.2 100% ‡
72
15 Co oxides 530 75 9.2 NA 77
16 Co3O4 300 49 14 NA 78
17 Co oxides ~480 ~120 3.5 >95% ‡
79
18 Co oxides ~900 NA 1.6 96% (± 2%)* 81
19 CoP 345 (10 mA cm−2) 47 14 100%* 82
20 CoMnP 330 (10 mA cm )−2
61 14 96%* 83
21 Ni oxides 380 30 9.2 100%‡ 85
22 Ni oxides 450 72 9.1 NA 87
23 Ni oxides 320 (10 mA cm−2) 52 14 ~100%* 88
24 Mn oxides <300 NA 13 NA 90
25 Mn oxides 280 NA 14 NA 91
26 Mn oxides 300 114 14 ~100% ‡
92
η, overpotential; NA, no analysis; OER, oxygen evolution reaction. *Quantified by gas chromatography. ‡Quantified using a
fluorescence-quench sensor.

NATURE REVIEWS | CHEMISTRY VOLUME 1 | ARTICLE NUMBER 0003 | 5

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

2 be substitutionally labile, with the result that the Co(ii)

Pt oxides readily dissolve. However, provided an anodic

–log HER exchange current density (A cm–2)

Photosystem II
Photosystem II is a key protein
complex involved in
photosynthesis. It uses sunlight
to oxidize water, producing
high energy electrons for
subsequent chemical reactions.
4 Ir
Au
Cu
6 Fe
Sn Bi
8 Zn Ag
Pb Ga
Cd
10 Tl In
100 150
M–H bond strength (kJ mol–1)
Figure 2 | Trasatti’s HER volcano plot. The y axis plots the
exchange current densities for Nature Reviews
the hydrogen | Chemistry
evolution
reaction (HER) versus the energy of the intermediate
metal–hydrogen bond formed between absorbed H and
the electrode surface on the x axis (data are from REF. 24).
The positions of titanium and copper have been
highlighted in green.
Smith et al.69 have demonstrated the effectiveness of
photochemical deposition for the preparation of OER
electrocatalysts. Thin films of transition metal com-
plexes bearing ultraviolet (UV)‑sensitive ligands (such
as 2‑ethylhexanoate) were spin-coated on transparent
conductive oxide electrodes and then subjected to UV
irradiation to decompose the ligands. These films were
then annealed at high temperature to give the corre-
sponding oxides. By using mixtures of the metal pre-
cursor solutions, the authors were able to identify several
promising OER catalysts, of which α-Fe33Co33Ni33Ox
was the best (TABLE 2, entry 12). A particular advan-
tage of this approach is the ability to accurately control
the stoichiometry of the resulting catalysts by virtue of
controlling the ratios of metal salts in the precursor solu-
tions. Such control over catalyst composition would be
hard, if not impossible, to obtain using electrochemical
synthesis techniques.
OER electrocatalysts based on single metal oxides. The
OER electrocatalysts discussed up to now have all been
optimized for use in rather basic media (pH 13–14).
However, as previously discussed, mild pH regimes
may be advantageous in photoelectrochemical solar‑to‑
hydrogen devices. In this regard, Nocera and co‑work-
ers70–72 reported a cobalt oxy-hydroxide catalyst that
continuously performs OER over many hours at mod-
est overpotentials in neutral buffers (TABLE 2, entries
13 and 14). The local structure of the proposed active
sites in this catalyst has been likened to incomplete
cobalt‑oxo cubanes73,74, as illustrated in FIG. 3. FIGURE. 3
also shows the proposed mechanism by which the
catalyst oxidizes water: Co(iii) centres in the catalyst
resting state undergo oxidation to Co(iv) under anodic
bias, and it is these Co(iv) centres that oxidize water
to give oxygen75. As this happens, the Co(iv) centres
are themselves reduced to Co(ii) centres, which tend to
Re bias is maintained, the Co(ii) ions are rapidly reoxidized
Rh
and redeposited as Co(iii) oxides, giving the catalyst
functional (rather than structural) stability 76. Of spe-
Ni cial relevance to the development of low-cost solar‑to‑
Co
hydrogen devices, these cobalt catalysts were found to
W work effectively in both phosphate-buffered seawater
and buffered (but otherwise untreated) river water 77.
Mo
Ti This is in contrast to most commercial electrolysers,
Ta which require high purity water.
Nb
Since these reports on the OER activity of cobalt
oxides at neutral pH, numerous other groups have
measured the OER performance of cobalt‑containing
200 250 300 350 400 materials across a wide pH range ( TABLE  2, entries
16–18). In particular, heterogeneous cobalt oxides
are competent OER catalysts under very basic condi-
tions (pH 13 and above)78, but some activity for these
materials has also been observed at low pH (even as
low as pH 1.6), provided certain concentration and
cell potential conditions are met 79–81. Very recently,
cobalt-containing phosphides have emerged as prom-
ising heterogeneous catalysts for the OER82,83; the OER
activity of many more such materials will certainly be
reported in the near future.
Not long after reporting his neutral‑pH cobalt oxide
OER catalyst, Nocera described its nickel analogue84.
Detailed analysis of the performance of this catalyst for
the OER revealed that the catalytic activity improved
gradually over the first few hours of operation under
anodic bias, with Tafel slopes of 30 millivolts per decade
manifesting after sufficient anodization85. X‑ray absorp-
tion near-edge structure spectra collected on samples
after various amounts of anodization revealed a shift in
the average oxidation state of the nickel centres in the
catalyst from +3 before any anodization to +3.6 once
anodization was complete86. This was interpreted as indi-
cating that Ni(iv) was the active species in the catalytic
cycle (just as Co(iv) is the active site for the neutral‑pH
cobalt oxide water oxidation catalyst described above),
with the formation of Ni(iv) from Ni(iii) requiring a
significant energy input — manifested in the need for
anodization — on account of the reorganization energy
required to overcome the Jahn–Teller distortion in the
Ni(iii) centres. Other nickel‑only OER catalysts have
subsequently been reported87,88, but the fact remains that
nickel oxides appear to be inferior catalysts compared
with iron-doped nickel oxides.
Manganese oxides are nature’s choice of OER cat-
alyst, being found in the oxygen-evolving complex of
photosystem II in the leaves of plants89. However, until
now, manganese oxides have proven much less active as
electrocatalysts for the OER. A selection of the better-
performing manganese oxide electrocatalysts is given in
TABLE 2 (entries 24–26)90–92. Typically, the more active
manganese oxide OER catalysts exhibit a high degree of
structural disorder and feature manganese in more than
one oxidation state. Structural disorder produces smaller
manganese-oxo platelets with a large number of coor-
dinatively unsaturated and partially reduced Mn(iii)O5
moieties at the plate boundaries, which may act as hole

6 | ARTICLE NUMBER 0003 | VOLUME 1 www.nature.com/natrevchem

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

Self-repair — dissolution of CoII and redeposition under anodic bias

OH OH
HO O OH
CoIII CoIV
O O O OH OH OH OH
Ox O O Ox HO O OH O2 HO O OH
Anodic bias
Co Co Co CoIV CoIV CoII CoII
+ 2H 2O
Ox O O Ox O O O O O O
O O O OH OH OH OH
Co Co
Ox O Ox
Ox Ox

Catalyst resting state Active CoIV–CoIV Oxygen evolution and

oxidation state reduction of Co centres
Figure 3 | OER at neutral pH mediated by a cobalt‑oxide catalyst. The structure and oxygen evolution reaction (OER)
Nature Reviews
mechanism of Nocera’s cobalt oxy-hydroxide water oxidation electrocatalyst, showing the generation | Chemistry
of the reactive
Co(iv) sites and the dissolution and redeposition self-repair mechanism.

Reversible hydrogen
electrode
(RHE). The RHE is based on the
following electrochemical
half-reaction: 2e− + 2H+ → H2.
Therefore, it is a subtype of the
standard hydrogen electrode.
However, the measured
potential of the RHE does not
change with pH, which means
it can be used to make a ready
comparison between the
theoretical position of onset of
the HER (always 0 V versus RHE)
and the position of a redox
event of interest, regardless of
the pH of the solution.
traps, promoting the oxidation of nearby water mole-
cules under anodic bias93,94. Meanwhile, the role of the
multiple manganese oxidation states in catalysis is less
clear, with various groups suggesting that MnO2 (REF. 95),
Mn2O3 (REF. 91) or mixtures of these phases96 are all
required for efficient catalysis of the OER. Then again,
activity for the OER is considerably diminished below
pH 8 as Mn3+ ions disproportionate into Mn2+ and Mn4+
ions in acidic media. Takashima et al.97,98 have argued
that this is evidence for the Mn(iii) oxidation state being
implicated in the catalysis of the OER. This supposition
appears to be borne out by subsequent analysis by Huynh
et al.99,100 of the OER mechanism using manganese oxides
over the pH range 0–14.
TABLE 2 summarizes the OER catalysts discussed in
this section. Inspection of this table shows that mixed
oxides containing nickel tend to give the best perfor-
mance for the OER under basic conditions, whereas
nickel and cobalt oxides are most effective near neutral
pH. It is striking that there are currently no good base
metal catalysts for the OER under acidic conditions. The
identification of a stable catalyst that is efficacious for
the OER at low pH would be a major advance in the
field, with potential applications both in artificial photo­
synthesis systems and in more conventional polymer
electrolyte membrane electrolysers.
Photoelectrochemical catalysts
The input of electrical energy is not the only means by
which water splitting can be performed. Indeed, the use
of light energy to drive the reaction in equation 1 in a
photoelectrochemical system has proven particularly
topical of late. For a single material to split water unaided
by an external bias it must have a bandgap in excess of
1.23 eV. The material must also be able to supply addi-
tional energy for overcoming various over­potentials and
electron-transfer-induced losses; therefore, bandgaps of
between 1.6 and 2.4 eV (corresponding to wavelengths
in the visible region between ~800 and 500 nm) are
required for practical water-splitting applications101.
Various other factors (such as band-edge position,
charge-carrier diffusion lengths and so on) must then be
taken into consideration when selecting an appropriate
semiconductor light-absorber, as discussed in REF. 101.
Identifying semiconductor materials with optimal char-
acteristics that are stable under the necessary operating
conditions (aqueous solutions, possibly at variable pH)
is in fact extremely challenging. When considering these
factors, the cost of these materials is generally ignored,
and, as a result, the projected costs of many artificial
photosynthesis systems are dominated by the cost of the
light-absorbing components.
Once semiconductors with the appropriate light
absorption and stability profiles have been identified,
electrocatalysts are normally deposited on top of these
to reduce overpotentials for the OER and HER, such
as with a conventional electrode. In the following
sections, we discuss electrocatalyst/light-harvester
ensembles that have been used to drive the HER and OER
in isolation (often with the aid of an additional external
bias). Such studies are useful for optimizing a particu-
lar semiconductor–electrocatalyst combination and can
be viewed as stepping stones towards the development
of artificial photosynthesis systems for water splitting
without external bias (FIG. 1b).
Catalysts for the photoelectrochemical HER. In terms
of the photoelectrochemical HER, silicon has proven
a popular choice as a photoactive substrate on which
to deposit HER electrocatalysts, not only because it
is cheap, but, more importantly, because of its stabil-
ity under the low pH conditions used in most HER
studies. A key early result in this area was established by
Chorkendorff and co‑workers102, who deposited cata-
lytic Mo3S4 clusters onto a p‑type silicon semiconductor
substrate that was optimized to absorb red photons
(wavelengths >620 nm). The authors used photolitho­
graphy to pattern the silicon substrate, producing a
‘pillar’ architecture (somewhat reminiscent of a micros-
cale toothbrush). This design provided a large surface
area for the HER catalyst while optimizing photocur-
rent collection efficiency. Compared with planar silicon
substrates decorated with the same electrocatalyst, the
pillared Si–MoSx arrays exhibited higher HER current
densities at all potentials, giving −10 mA cm−2 at 0 V
versus the reversible hydrogen electrode (RHE). Additional

NATURE REVIEWS | CHEMISTRY VOLUME 1 | ARTICLE NUMBER 0003 | 7

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

examples35,103–107 of the use of silicon as the photoactive
substrate are given in TABLE 3.
An exception to the trend of studying the HER at low
pH can be seen in recent work by the Hu group, who
used a Mo2C catalyst in combination with an amorphous
silicon light absorber to afford photocathodes with a per-
formance that was essentially identical in 0.1 M H2SO4
and 1 M KOH (the photocathodes performed the HER
at −11 mA cm−2 in both electrolytes)108. This work was
the first report of the photoelectrochemical HER occur-
ring in strongly alkaline electrolytes using silicon‑based
photocathodes and could lead to the adoption of silicon
photocathodes across a wider pH range than that real-
ized up until now.
Cadmium chalcogenides have been used as photo-
active substrates for the deposition of earth-abundant
HER electrocatalysts109,110, but the main alternatives to
silicon are the copper oxides111–113. By way of example,
a recent report is considered by Morales-Guio et al.114,
who used a photoactive Cu2O layer to harvest visible
photons (TABLE 3, entries 15 and 16). Cu2O has an ideal
bandgap for solar‑to‑hydrogen applications (~2 eV)115,
but it tends to corrode rapidly in aqueous environments.
Hence, the authors used thin layers of aluminium‑doped
ZnO and TiO2 as barriers to protect the underlying
Cu2O. Under 1 sun irradiation (irradiance at 1 kW m−2)
at pH 0, these films mediated the HER at a current den-
sity of −5.7 mA cm−2 at 0 V versus RHE, although the
films were not stable and activity declined steadily over
a few hours of operation. However, under slightly milder
conditions (pH 4), the HER photocurrent at 0 V versus
RHE was stable at −4.5 mA cm−2 for at least 10 hours. The
effectiveness of this fabrication strategy is promising for
the use of Cu2O as a photocathode in the future, but (as
with the HER electrocatalysts discussed above) it is now
essential that the community starts to demonstrate sta-
bility over thousands of hours of continuous operation
if such materials are to be considered for use in artificial
photosynthesis systems.
Catalysts for the photoelectrochemical OER. The
first demonstration of a photoanode for OER cataly-
sis was supplied by Honda and Fujishima116 in 1972
using undoped TiO2. Current densities for the OER of
1 mA cm−2 were obtained when additional potentials of
~0.9 V versus RHE were applied under irradiation with
UV light (wavelengths of 415 nm and shorter). WO3 was
subsequently shown to behave similarly, although its
strongest absorptions are still in the UV region117. In an
effort to increase the proportion of visible light absorbed
by the photoanode, Hardee and Bard118 used a haematite
(α-Fe2O3) substrate for the OER with irradiation at wave-
lengths longer than 500 nm and an additional applied bias.
Haematite has key advantages in terms of its light absorp-
tion profile, its high abundance and low cost, although
it does exhibit high OER overpotentials. Therefore, it is
normally considered necessary to use a co‑catalyst to
improve the kinetics of oxygen evolution on haematite
(although entries 16 and 17 in TABLE 4 show that this is
not always the case). Cobalt oxides have become the OER
catalyst of choice for this purpose, because they can be
electrodeposited on haematite under mild conditions and
because they are good OER catalysts over a pH range at
which haematite is stable. Gamelin and co‑workers119–121
were among the first to deposit cobalt oxides on Fe2O3 for
the OER, with a reduction in the required overpotential of

Table 3 | Examples of earth-abundant HER photoelectrocatalytic systems

Entry Catalyst Photoactive substrate Current density at 0 V pH Faradaic yield Refs
material versus RHE and 1 sun
illumination (mA cm−2)
1 Mo3S4 p‑Si −10 0 100%* 102
2 CoS planar n+–p‑Si −11 7 100%* 35
3 W2C p‑Si −4 0 NA 103
4 MoS2 planar n+–p‑Si −17.6 0 100%* 104
5 CoPS n –p–p -Si
+ +
−35 0 NA 105
6 Co dithiolene p‑Si −3.8 1.3 80% (± 3%)* 106
polymer
7 NiMo p‑Si −15 4.5 NA 107
8 Mo2C Surface-protected Si −11.2 14 100% ‡
108
9 Co3O4 CdS −0.15 (+0.12 V) 7 NA 109
10 NiO–CoP CdSe −0.15 (+0.4 V) 6.8 81%* 110
11 NiOx Cu2O −5 6 32% (± 6%)* 111
12 Cu2MoS4 Cu2O–NiO −1.25 5 <23%* 112
14 Ni Cu2O–CuO −4.3 5 84%* 113
15 MoSx Cu2O −4.5 4 100 ‡
114
16 MoSx Cu2O −5.7 1 100 ‡
114
HER, hydrogen evolution reaction; NA, no analysis; n+, degenerately n-type; p+, degenerately p-type. *Quantified by gas
chromatography. ‡Quantified using a pressure sensor.

8 | ARTICLE NUMBER 0003 | VOLUME 1 www.nature.com/natrevchem

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

Table 4 | Examples of earth-abundant OER photoelectrocatalytic systems

Entry Catalyst Photoactive substrate Bias required to obtain pH Faradaic Refs
material 1 mA cm−2 under 1 sun yield
illumination (V versus RHE)
1 Co α-Fe2O3 1.45 13.6 NA 119
2 Co α-Fe2O3 1.1 7–13.6 NA 120
3 Co α-Fe2O3 1 13.6 NA 121
4 Co WO3 1.2 (0.5 mA cm )
−2
7 100% ‡
122
5 Co npp Si/ITO
+
1.2 7 100% ‡
123
6 Co BiVO4 1 5.6 NA 124
7 Co W‑doped BiVO4 1 5.6 NA 125
8 Co BiVO4 1.2 7 NA 126
9 Co TiO2 1.7 (0.4 mA cm )
−2
7 >95%§ 127
10 Co TiO2–g-C3N4 1 6.8 NA 128
11 Co BiVO4 and perovskite 0.5 7 100% §
129
photovoltaic
12 Ni BiVO4 1 9.2 100%‡ 130
13 Ni W‑doped BiVO4 1.3 7 NA 131
14 Ni WO3 1.2 9.2 NA 132
15 NiFe Ti‑doped α-Fe2O3 1.1 14 NA 133
16 Fe α-Fe2O3 (the oxide itself) 1.1 14 NA 134
17 Fe α-Fe2O3 (the oxide itself) 1.15 13.6 NA 135
18 Ni n‑Si ~1.6 (10 mA cm−2) 9.5 NA 136
19 NiOOH and BiVO4 0.32 7 >90% ‡
137
FeOOH
g-C3N4, graphitic carbon nitride; ITO, indium tin oxide; NA, no analysis; OER, oxygen evolution reaction; p+, degenerately p-type;
RHE, reversible hydrogen electrode. ‡Quantified using a fluorescence-quench sensor. §Quantified by gas chromatography.

several hundred millivolts being obtained (TABLE 4, entries
1–3). Seabold and Choi122 subsequently deposited cobalt
oxides on WO3 substrates and noted that the addition of
these cobalt oxides suppressed the undesirable production
of peroxo species during the photo-oxidation of water.
Since these reports, cobalt oxides have been used as OER
co‑catalysts on numerous photoactive substrates123–128, as
summarized in TABLE 4. Of these, the best performance
in terms of the current density at a given potential was
reported recently by Kim et al.129, who fabricated a photo­
anode from two comparatively cheap light absorbers
arranged in series (a single-junction perovskite solar cell
and a molybdenum‑doped BiVO4 semiconductor), on
which cobalt oxides were deposited as the OER catalyst.
The use of two light-harvesting systems in tandem greatly
reduced the external bias that was required to achieve a
given current density for the OER.
Nickel- and iron-based catalysts have also been used
as co‑catalysts for the photoelectrochemical OER130–136
(TABLE 4, entries 12–19). The best of these uses a com-
bination of nickel and iron oxides on a BiVO4 photoac-
tive substrate and gives rise to an OER current density
under illumination of 1 mA cm−2 at only a little over
300 mV overpotential137. The authors found that the best
photo­anodes were formed by first depositing a layer of
FeOOH on the BiVO4 substrate, then depositing a layer
of NiOOH on top of the FeOOH layer. This combination
resulted in superior performance relative to photo­anodes
with only the individual metal oxy-hydroxides (that is,
NiOOH on BiVO4 and FeOOH on BiVO4) or photoelec-
trodes formed in the order BiVO4–NiOOH–FeOOH.
The authors ascribed the photocurrent enhancement
for the BiVO4–FeOOH–NiOOH anode to the reduced
interfacial charge recombination between the BiVO4 and
the catalyst overlayers, and also to an optimal Helmholtz
layer potential drop at the catalyst–electrolyte junction.
However, it should be noted that additional biases
of several hundred millivolts are required (even under
illumination) to drive the OER with all the systems listed
in TABLE 4. This is a reflection of the high overpotentials
that are necessary to perform the OER (even at a current
density as low as 1 mA cm−2) and makes the examples
of functional direct solar‑to‑hydrogen devices (able to
split water using sunlight alone) in the next section even
more impressive.
Artificial photosynthesis systems
Photoelectrochemical solar‑to‑hydrogen devices. An
early example of a wireless photoelectrochemical
solar‑to‑hydrogen device (FIG. 1a) was described by
Rocheleau and co‑workers138 in 1998. Their design
called for a buried amorphous silicon photovoltaic,
with a CoMo alloy as the HER catalyst on one side and
an NiFeyOx OER catalyst on the other. A useful figure

NATURE REVIEWS | CHEMISTRY VOLUME 1 | ARTICLE NUMBER 0003 | 9

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

of merit for photoelectrochemical solar‑to‑hydrogen both in terms of materials development as well as

Solar‑to‑hydrogen
conversion efficiency
The energy that would be
released upon complete
oxidation of the hydrogen
produced by, for example, an
artificial photosynthesis system
in a given time, divided by the
energy required by the artificial
photosynthesis system to
produce that amount of
hydrogen.
systems is their solar‑to‑hydrogen conversion efficiency. In
Rocheleau’s report, an excellent photoelectro­chemical
solar‑to‑hydrogen conversion efficiency of 7.8% was
reported at pH 14 (for cells with an area of 1 cm2), with
the catalysts themselves exhibiting stability for over
7,200 hours of operation. More recently, Verlage et al.139
have reported an improved wireless photoelectrochem-
ical solar‑to‑hydrogen device (area = 1 cm2) that also
works in 1 M KOH. In this case, a tandem-junction
GaAs–InGaP light harvester was used in combination
with a NiMo HER catalyst and a nickel‑based OER cat-
alyst. Under 1 sun illumination, a solar‑to‑hydrogen
conversion efficiency of 8.6% was obtained.
In terms of producing photoelectrochemical
solar‑to‑hydrogen devices that can operate for tens or
hundreds of thousands of hours, systems that use less
corrosive electrolytes may be required to minimize degra-
dation of the photoelectrodes and other cell components.
In this regard, several devices that work at near-neutral pH
have been reported. The first such system used a (some-
what costly) buried triple-junction silicon light absorber,
a cobalt oxy-hydroxide catalyst for the OER and a ter-
nary NiMoZn HER catalyst140. Irradiation of this device
(total area of 2 cm2) with simulated solar light in borate
buffer (pH 9.2) led to water splitting with a 2.5% solar‑to‑
hydrogen conversion efficiency.
Any practical and scalable artificial photosynthesis
system will need significantly cheaper light absorbers
than triple-junction silicon, and some progress has
been made in this regard. For example, Janssen and
co‑workers141 replaced the triple-junction silicon in
REF. 140 with a triple-junction polymer solar cell and
achieved solar‑to‑hydrogen conversion efficiencies at
pH 9 of 4.9% using cells with an area of <0.1 cm2, which
decreased to 1.3% when the cell area was 1.2 cm2. The
use of organic polymer photovoltaics offers advantages
of improved flexibility and lower production costs
(because such materials are suitable for roll‑to‑roll
printing) compared with inorganic photovoltaics.
Challenges for photoelectrochemical solar‑to‑hydrogen
devices. The challenges facing the development of
affor­dable and reliable artificial photosynthesis sys-
tems for solar-driven water splitting remain signifi-
cant. For instance, in all of the examples given above,
the areas of the devices are on the order of a centime-
tre. However, arrays of several square metres would be
required to generate enough hydrogen to meet even
modest demand (for example, for powering a hydro-
gen-fuelled car for one day). Therefore, considerable
scale‑up is necessary; this will require improvements in
device performance (for example, better charge trans-
port through extended structures), and bring with it
demands for cheaper components (such as membranes,
electrocatalysts and light harvesters), greater long-term
reliability and a robust design that does not require sig-
nificant maintenance142. These requirements will each
place considerable constraints on the materials and
conditions that can be used, making the realization of
mass-producible solar‑to‑hydrogen devices challenging
engineering aspects.
Moreover, several largely unanswered questions
remain, particularly with regard to how (and even if)
hydrogen can be safely harvested from the large surface
area arrays apparently required. By assuming current
densities under irradiation on the order of 10 mA cm−2,
the permeation of both hydrogen and oxygen through
membrane separators, such as Nafion, would be signif-
icant and could lead to extensive (and dangerous) mix-
ing of the product gases143. Although methods to prevent
the build‑up of dangerous gas mixtures in electrolysers
exist, they would also significantly reduce the overall
solar‑to‑hydrogen efficiencies that could be obtained144.
Then again, collecting hydrogen efficiently from a large
area array (probably at close to atmospheric pressure) is
not necessarily a simple task to perform.
The recent development of the electron-coupled-
proton buffer (ECPB) offers solutions to some of these
challenges, in that it allows the OER and HER to be
completely decoupled from each other so that the gases
are not made in the electrochemical cell at the same
time145–147. This approach has been applied recently in
a system that produces oxygen from water under solar
irradiation in a photoelectrochemical cell, in which the
counter electrode reaction is reduction and protonation
of the ECPB rather than generation of hydrogen148. The
ECPB was then reoxidized purely electrochemically,
co‑generating hydrogen. In a scaled‑up device, hydro-
gen would be produced in a separate, more conven-
tional electrolyser, rather than in the solar-harvesting
array. This would allow the HER to proceed at much
higher current densities than those obtained for the
photo-driven OER, in turn allowing electrolysers of
much smaller dimensions than the solar harvesting
array to be used (with potential advantages in terms of
ease of gas pressurization for storage and efficient use
of the HER catalyst).
However, as the discussions above suggest, all of the
systems for using solar energy to split water that have
been trialled so far remain far from optimal. Much work
is needed across the board (in electrocatalyst optimiza-
tion, semiconductor development, device design and
balance of plant considerations) before artificial photo-
synthesis systems, such as those depicted in FIG. 1, become
a viable option for large-scale hydrogen production.
Conclusions and outlook
In this Review, we have summarized prominent electro­
catalysts for water splitting (for both the HER and
OER) that are composed of earth-abundant elements.
Throughout, we have placed emphasis on the more
recent literature and especially on examples that are of
relevance to photoelectrocatalytic solar‑to‑fuels devices.
There remains a certain mismatch between the optimal
conditions under which most of the earth-abundant
OER (photo-)electrode systems operate (pH >> 7) and
those suitable for the HER (generally pH << 7). In prac-
tice, neither of these regimes may be suitable, and mild
pH conditions (near-neutral) may be preferred (or even
required) for artificial photosynthesis on account of

10 | ARTICLE NUMBER 0003 | VOLUME 1 www.nature.com/natrevchem

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

durability and regulatory issues. Clearly, there remains
scope for significant advances to be made in the develop-
ment of both HER and OER electrocatalysts optimized
for neutral pH values.
Looking at electrocatalytic water splitting more gen-
erally, earth-abundant HER and OER catalysts that oper-
ate at very high and very low pH will certainly also be of
great utility (for example, for the production of cheaper
conventional electrolysers that run using other forms
of renewable energy). Electrocatalysis of the OER at
high pH using NiFe oxide electrodes is well established.
However, low pH water oxidation with earth-abundant
metal electrocatalysts presents enormous challenges in
terms of catalyst stability and overpotential demands.
Operation at high or low pH seems less of an issue for
the HER (TABLE 1), but considerable optimization is
almost certainly achievable here too. With regard to the
development of new catalysts, combined computational
and high-throughput methods hold great promise for
identifying highly active formulations.
What, then, are the prospects for artificial photo-
synthesis systems for solar‑to‑hydrogen applications
being deployed on a large scale in the near future? The
challenges that remain are numerous and significant.
First, any mass-producible device will need to use only
components that are abundant and readily available.
These components will then need to be assembled into
devices that are both cheap and robust. Moreover, real-
world applications demand that all aspects of hydro-
gen generation (including the efficiency and ease of
collection, storage and distribution) are addressed in a
rigorous manner: if we fail in this regard then we risk
developing catalysts and devices that are of no practical
use. We hope that this Review will go some way towards
reinforcing this need for a whole-systems approach to
those who are already experts in electrocatalytic water
splitting. However, above all, we hope that this Review
will educate and inspire those new to the field and
encourage them to make their own contributions to
this fascinating and rapidly changing area.

1. Dunn, B., Kamath, H. & Tarascon, J.‑M. Electrical
energy storage for the grid: a battery of choices.
Science 334, 928–935 (2011).
2. Lewis, N. S. & Nocera, D. G. Powering the planet:
chemical challenges in solar energy utilization. Proc.
Natl Acad. Sci. USA 103, 15729–15735 (2006).
As powerful today as when it was written 10 years
ago, this article by two of the leaders in the field
explains why renewable energy storage is so
important.
3. Blankenship, R. E. et al. Comparing photosynthetic and
photovoltaic efficiencies and recognizing the potential
for improvement. Science 332, 805–809 (2011).
4. Armaroli, N. & Balzani, V. Solar electricity and solar
fuels: status and perspectives in the context of the
energy transition. Chem. Eur. J. 22, 32–57 (2016).
A well-argued and up‑to‑date summary of our
options for solar fuels generation and the broader
techno-economic factors affecting the field.
5. Styring, S. Artificial photosynthesis for solar fuels.
Faraday Discuss. 155, 357–376 (2012).
6. Le Formal, F., Bourée, W. S., Prévot, M. S. & Sivula, K.
Challenges towards economic fuel generation from
renewable electricity: the need for efficient electro-
catalysis. Chimia 69, 789–798 (2015).
7. Joya, K. S., Joya, Y. F., Ocakoglu, K. & van de Krol, R.
Water-splitting catalysis and solar fuel devices: artificial
leaves on the move. Angew. Chem. Int. Ed. 52,
10426–10437 (2013).
8. Tachibana, Y., Vayssieres, L. & Durrant, J. R. Artificial
photosynthesis for solar water-splitting. Nat. Photonics
6, 511–518 (2012).
9. McCrory, C. C. L. et al. Benchmarking hydrogen
evolving reaction and oxygen evolving reaction
electrocatalysts for solar water splitting devices.
J. Am. Chem. Soc. 137, 4347–4357 (2015).
An article exploring how different electrocatalysts
can be compared for the OER and HER in the
context of solar‑to‑fuels devices.
10. Carmo, M., Fritz, D. L., Mergel, J. & Stolten, D. A
comprehensive review on PEM water electrolysis. Int.
J. Hydrogen Energy 38, 4901–4934 (2013).
11. Faber, M. S. & Jin, S. Earth-abundant inorganic
electrocatalysts and their nanostructures for energy
conversion applications. Energy Environ. Sci. 7,
3519–3542 (2014).
12. Nocera, D. G. The artificial leaf. Acc. Chem. Res. 45,
767–776 (2012).
13. Zou, X. & Zhang, Y. Noble metal-free hydrogen
evolution catalysts for water splitting. Chem. Soc. Rev.
44, 5148–5180 (2015).
14. Zeng, M. & Li, Y. Recent advances in heterogeneous
electrocatalysts for the hydrogen evolution reaction.
J. Mater. Chem. A 3, 14942–14962 (2015).
15. Kang, D. et al. Electrochemical synthesis of
photoelectrodes and catalysts for use in solar water
splitting. Chem. Rev. 115, 12839–12887 (2015).
16. Roger, I. & Symes, M. D. First row transition metal
catalysts for solar-driven water oxidation produced by
electrodeposition. J. Mater. Chem. A 4, 6724–6741
(2016).
17. Fukuzumi, S., Yamada, Y., Suenobu, T., Ohkubo,  31. Merki, D., Fierro, S., Vrubel, H. & Hu, X. Amorphous
K. & Kotani, H. Catalytic mechanisms of hydrogen
evolution with homogeneous and heterogeneous
catalysts. Energy Environ. Sci. 4, 2754–2766
(2011).
18. Ismail, A. A. & Bahnemann, D. W. Photochemical
splitting of water for hydrogen production by
photocatalysis: a review. Sol. Energy Mater Sol. Cells
128, 85–101 (2014).
19. Parent, A. R. & Sakai, K. Progress in base-metal water
oxidation catalysis. ChemSusChem 7, 2070–2080
(2014).
20. Fan, C., Piron, D. L., Sleb, A. & Paradis, P. Study of
electrodeposited nickel–molybdenum, nickel–tungsten,
cobalt–molybdenum, and cobalt–tungsten as hydrogen
electrodes in alkaline water electrolysis. J. Electrochem.
Soc. 141, 382–387 (1994).
21. Arul Raj, I. & Vasu, K. I. Transition metal-based
hydrogen electrodes in alkaline solution —
electrocatalysis on nickel based binary alloy coatings.
J. Appl. Electrochem. 20, 32–38 (1990).
22. McKone, J. R., Sadtler, B. F., Werlang, C. A., Lewis,  37. Voiry, D. et al. Enhanced catalytic activity in strained
N. S. & Gray, H. B. Ni–Mo nanopowders for efficient
electrochemical hydrogen evolution. ACS Catal. 3,
166–169 (2013).
23. Wang, Y. et al. A 3D nanoporous Ni–Mo electrocatalyst
with negligible overpotential for alkaline hydrogen
evolution. ChemElectroChem. 1, 1138–1144 (2014).
24. Trasatti, S. Work function, electronegativity, and
electrochemical behaviour of metals: III. Electrolytic
hydrogen evolution in acid solutions. J. Electroanal.
Chem. 39, 163–184 (1972).
25. Nørskov, J. K. et al. Trends in the exchange current
for hydrogen evolution. J. Electrochem. Soc. 152,
J23–J26 (2005).
26. Sheng, W., Myint, M., Chen, J. G. & Yan, Y. Correlating
the hydrogen evolution reaction activity in alkaline
electrolytes with the hydrogen binding energy on
monometallic surfaces. Energy Environ. Sci. 6,
1509–1512 (2013).
27. Hinnemann, B. et al. Biomimetic hydrogen evolution:
MoS2 nanoparticles as catalyst for hydrogen evolution.
J. Am. Chem. Soc. 127, 5308–5309 (2005).
This paper showed how computational methods
could be used to identify new HER catalysts (in this
case MoS2) and then demonstrated the same
experimentally.
28. Jaegermann, W. & Tributsch, H. Interfacial properties
of semiconducting transition metal chalcogenides.
Prog. Surf. Sci. 29, 1–167 (1988).
29. Jaramillo, T. F. et al. Identification of active edge sites
for electrochemical H2 evolution from MoS2
nanocatalysts. Science 317, 100–102 (2007).
30. Kibsgaard, J., Chen, Z., Reinecke, B. N. & Jaramillo, T. F.
Engineering the surface structure of MoS2 to
preferentially expose active edge sites for
electrocatalysis. Nat. Mater. 11, 963–969 (2012).
molybdenum sulfide films as catalysts for
electrochemical hydrogen production in water.
Chem. Sci. 2, 1262–1267 (2011).
32. Li, Y. et al. MoS2 nanoparticles grown on graphene: an
advanced catalyst for the hydrogen evolution reaction.
J. Am. Chem. Soc. 133, 7296–7299 (2011).
33. Li, H. et al. Activating and optimizing MoS2 basal planes
for hydrogen evolution through the formation of strained
sulphur vacancies. Nat. Mater. 15, 48–53 (2016).
34. Faber, M. S. et al. High-performance electrocatalysis
using metallic cobalt pyrite (CoS2) micro- and
nanostructures. J. Am. Chem. Soc. 136,
10053–10061 (2014).
35. Sun, Y. et al. Electrodeposited cobalt-sulfide catalyst for
electrochemical and photoelectrochemical hydrogen
generation from water. J. Am. Chem. Soc. 135,
17699–17702 (2013).
36. Tran, P. D. et al. Novel cobalt/nickel–tungsten-sulfide
catalysts for electrocatalytic hydrogen generation from
water. Energy Environ. Sci. 6, 2452–2459 (2013).
chemically exfoliated WS2 nanosheets for hydrogen
evolution. Nat. Mater. 12, 850–855 (2013).
38. Xu, Y.‑F., Gao, M.‑R., Zheng, Y.‑R., Jiang, J. & Yu, S.‑H.
Nickel/nickel(ii) oxide nanoparticles anchored onto
cobalt(iv) diselenide nanobelts for the electrochemical
production of hydrogen. Angew. Chem. Int. Ed. 52,
8546–8550 (2013).
39. Kiran, V., Mukherjee, D., Jenjeti, R. N. & Sampath, S.
Active guests in the MoS2/MoSe2 host lattice: efficient
hydrogen evolution using few-layer alloys of MoS2(1 − x)
Se2x. Nanoscale 6, 12856–12863 (2014).
40. Zhou, H. et al. One-step synthesis of self-supported
porous NiSe2/Ni hybrid foam: an efficient 3D electrode
for hydrogen evolution reaction. Nano Energy 20,
29–36 (2016).
41. Popczun, E. J. et al. Nanostructured nickel phosphide
as an electrocatalyst for the hydrogen evolution
reaction. J. Am. Chem. Soc. 135, 9267–9270 (2013).
42. Popczun, E. J., Read, C. G., Roske, C. W., Lewis, N. S. &
Schaak, R. E. Highly Active electrocatalysis of the
hydrogen evolution reaction by cobalt phosphide
nanoparticles. Angew. Chem. Int. Ed. 53, 5427–5430
(2014).
43. Jiang, P. et al. A cost-effective 3D hydrogen evolution
cathode with high catalytic activity: fep nanowire array
as the active phase. Angew. Chem. Int. Ed. 53,
12855–12859 (2014).
44. Kibsgaard, J. & Jaramillo, T. F. Molybdenum
phosphosulfide: an active, acid-stable, earth-abundant
catalyst for the hydrogen evolution reaction. Angew.
Chem. Int. Ed. 53, 14433–14437 (2014).

NATURE REVIEWS | CHEMISTRY VOLUME 1 | ARTICLE NUMBER 0003 | 11

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

45. Liang, H.‑W. et al. Molecular metal–Nx centres in
porous carbon for electrocatalytic hydrogen evolution.
Nat. Commun. 6, 7992 (2015).
46. Chen, W.‑F. et al. Hydrogen-evolution catalysts based
on non-noble metal nickel–molybdenum nitride
nanosheets. Angew. Chem. Int. Ed. 51, 6131–6135
(2012).
47. Cao, B., Veith, G. M., Neuefeind, J. C., Adzic, R. R. &
Khalifah, P. G. Mixed close-packed cobalt
molybdenum nitrides as non-noble metal
electrocatalysts for the hydrogen evolution reaction.
J. Am. Chem. Soc. 135, 19186–19192 (2013).
48. Vrubel, H. & Hu, X. Molybdenum boride and carbide
catalyze hydrogen evolution in both acidic and basic
solutions. Angew. Chem. Int. Ed. 51, 12703–12706
(2012).
49. Liao, L. et al. A nanoporous molybdenum carbide
nanowire as an electrocatalyst for hydrogen evolution
reaction. Energy Environ. Sci. 7, 387–392 (2014).
50. Shi, Z. et al. Porous nanoMoC@graphite shell derived
from a MOFs-directed strategy: an efficient
electrocatalyst for the hydrogen evolution reaction.
J. Mater. Chem. A 4, 6006–6013 (2016).
51. Fan, L. et al. Atomically isolated nickel species
anchored on graphitized carbon for efficient hydrogen
evolution electrocatalysis. Nat. Commun. 7, 10667
(2016).
52. Lu, Q. et al. Highly porous non-precious bimetallic
electrocatalysts for efficient hydrogen evolution. Nat.
Commun. 6, 6567 (2015).
53. Merrill, M. D. & Dougherty, R. C. Metal Oxide
Catalysts for the Evolution of O2 from H2O. J. Phys.
Chem. C 112, 3655–3666 (2008).
54. Lu, X. & Zhao, C. Electrodeposition of hierarchically
structured three-dimensional nickel–iron electrodes
for efficient oxygen evolution at high current densities.
Nat. Commun. 6, 6616 (2015).
55. Gong, M. et al. An advanced Ni–Fe layered double
hydroxide electrocatalyst for water oxidation. J. Am.
Chem. Soc. 135, 8452–8455 (2013).
56. Li, X., Walsh, F. C. & Pletcher, D. Nickel based
electrocatalysts for oxygen evolution in high current
density, alkaline water electrolysers. Phys. Chem.
Chem. Phys. 13, 1162–1167 (2011).
57. Corrigan, D. A. The catalysis of the oxygen evolution
reaction by iron impurities in thin film nickel oxide
electrodes. J. Electrochem. Soc. 134, 377–384
(1987).
58. Louie, M. W. & Bell, A. T. An investigation of thin-film
Ni–Fe oxide catalysts for the electrochemical evolution
of oxygen. J. Am. Chem. Soc. 135, 12329–12337
(2013).
59. Friebel, D. et al. Identification of highly active Fe sites
in (Ni, Fe)OOH for electrocatalytic water splitting.
J. Am. Chem. Soc. 137, 1305–1313 (2015).
60. Lambert, T. N. et al. Electrodeposited NixCo3 − xO4
nanostructured films as bifunctional oxygen
electrocatalysts. Chem. Commun. 51, 9511–9514
(2015).
61. Tian, J. et al. Self-supported NiMo hollow nanorod
array: an efficient 3D bifunctional catalytic electrode
for overall water splitting. J. Mater. Chem. A 3,
20056–20059 (2015).
62. Trotochaud, L., Young, S. L., Ranney, J. K. &
Boettcher, S. W. Nickel–iron oxyhydroxide oxygen-
evolution electrocatalysts: the role of intentional and
incidental iron incorporation. J. Am. Chem. Soc. 136,
6744–6753 (2014).
63. Smith, A. M., Trotochaud, L., Burke, M. S. &
Boettcher, S. W. Contributions to activity
enhancement via Fe incorporation in Ni-(oxy)
hydroxide/borate catalysts for near-neutral pH
oxygen evolution. Chem. Commun. 51, 5261–5263
(2015).
64. Burke, M. S., Kast, M. G., Trotochaud, L., Smith, A. M.
& Boettcher, S. W. Cobalt–iron (oxy)hydroxide oxygen
evolution electrocatalysts: the role of structure and
composition on activity, stability, and mechanism.
J. Am. Chem. Soc. 137, 3638–3648 (2015).
65. Roger, I. & Symes, M. D. Efficient electrocatalytic
water oxidation at neutral and high pH by adventitious
nickel at nanomolar concentrations. J. Am. Chem. Soc.
137, 13980–13988 (2015).
66. Zhang, B. et al. Homogeneously dispersed multimetal
oxygen-evolving catalysts. Science 352, 333–337
(2016).
67. Suntivich, J., May, K. J., Gasteiger, H. A.,
Goodenough, J. B. & Shao-Horn, Y. A. Perovskite oxide
optimized for oxygen evolution catalysis from
molecular orbital principles. Science 334,
1383–1385 (2011).
68. Haber, J. A. et al. Discovering Ce‑rich oxygen
evolution catalysts, from high throughput screening to
water electrolysis. Energy Environ. Sci. 7, 682–688
(2014).
69. Smith, R. D. L. et al. Photochemical route for
accessing amorphous metal oxide materials for water
oxidation catalysis. Science 340, 60–63 (2013).
70. Kanan, M. W. & Nocera, D. G. In situ formation of an
oxygen-evolving catalyst in neutral water containing
phosphate and Co2+. Science 321, 1072–1075
(2008).
The paper that ignited the field in terms of oxygen
evolution electrocatalysis under mild conditions.
71. Lutterman, D. A., Surendranath, Y. & Nocera, D. G. A.
Self-healing oxygen-evolving catalyst. J. Am. Chem.
Soc. 131, 3838–3839 (2009).
72. Surendranath, Y., Dinca, M. & Nocera, D. G.
Electrolyte-dependent electrosynthesis and activity of
cobalt-based water oxidation catalysts. J. Am. Chem.
Soc. 131, 2615–2620 (2009).
73. Risch, M. et al. Cobalt–oxo core of a water-oxidizing
catalyst film. J. Am. Chem. Soc. 131, 6936–6937
(2009).
74. Kanan, M. W. et al. Structure and valency of a cobalt-
phosphate water oxidation catalyst determined by
in situ X‑ray spectroscopy. J. Am. Chem. Soc. 132,
13692–13701 (2010).
75. McAlpin, J. G. et al. EPR evidence for Co(iv) species
produced during water oxidation at neutral pH.
J. Am. Chem. Soc. 132, 6882–6883 (2010).
76. Surendranath, Y., Kanan, M. W. & Nocera, D. G.
Mechanistic studies of the oxygen evolution reaction
by a cobalt-phosphate catalyst at neutral pH. J. Am.
Chem. Soc. 132, 16501–16509 (2010).
77. Esswein, A. J., Surendranath, Y., Reece, S. Y. &
Nocera, D. G. Highly active cobalt phosphate and
borate based oxygen evolving catalysts operating in
neutral and natural waters. Energy Environ. Sci. 4,
499–504 (2011).
78. Koza, J. A., He, Z., Miller, A. S. & Switzer, J. A.
Electrodeposition of crystalline Co3O4 — a catalyst for
the oxygen evolution reaction. Chem. Mater. 24,
3567–3573 (2012).
79. Gerken, J. B., Landis, E. C., Hamers, R. J. &
Stahl, S. S. Fluoride-modulated cobalt catalysts for
electrochemical oxidation of water under non-
alkaline conditions. ChemSusChem 3, 1176–1179
(2010).
80. Gerken, J. B. et al. Electrochemical water oxidation
with cobalt-based electrocatalysts from pH 0–14: the
thermodynamic basis for catalyst structure, stability,
and activity. J. Am. Chem. Soc. 133, 14431–14442
(2011).
81. Bloor, L. G., Molina, P. I., Symes, M. D. & Cronin, L.
Low pH electrolytic water splitting using earth-
abundant metastable catalysts that self-assemble in
situ. J. Am. Chem. Soc. 136, 3304–3311 (2014).
82. Jiang, N., You, B., Sheng, M. & Sun, Y.
Electrodeposited cobalt-phosphorous-derived films as
competent bifunctional catalysts for overall water
splitting. Angew. Chem. Int. Ed. 54, 6251–6254
(2015).
83. Li, D., Baydoun, H., Verani, C. N. & Brock, S. L.
Efficient water oxidation using CoMnP nanoparticles.
J. Am. Chem. Soc. 138, 4006–4009 (2016).
84. Dinca, M., Surendranath, Y. & Nocera, D. G. Nickel–
borate oxygen-evolving catalyst that functions under
benign conditions. Proc. Natl Acad. Sci. USA 107,
10337–10341 (2010).
85. Bediako, D. K., Surendranath, Y. & Nocera, D. G.
Mechanistic studies of the oxygen evolution reaction
mediated by a nickel–borate thin film electrocatalyst.
J. Am. Chem. Soc. 135, 3662–3674 (2013).
86. Bediako, D. K. et al. Structure–activity correlations in
a nickel–borate oxygen evolution catalyst. J. Am.
Chem. Soc. 134, 6801–6809 (2012).
87. Wu, L.‑K. Hu, J.‑M. Zhang, J.‑Q. & Cao, C.‑N. A silica
co‑electrodeposition route to highly active Ni‑based
film electrodes. J. Mater. Chem. A 1, 12885–12892
(2013).
88. Zhao, Y. et al. Ultrafine NiO nanosheets stabilized by
TiO2 from monolayer NiTi-LDH precursors: an active
water oxidation electrocatalyst. J. Am. Chem. Soc.
138, 6517–6524 (2016).
89. Barber, J. Crystal structure of the oxygen-evolving
complex of photosystem ii. Inorg. Chem. 47,
1700–1710 (2008).
90. Gorlin, Y. & Jaramillo, T. F. A. Bifunctional nonprecious
metal catalyst for oxygen reduction and water
oxidation. J. Am. Chem. Soc. 132, 13612–13614
(2010).
91. Ramírez, A. et al. Evaluation of MnOx, Mn2O3, and
Mn3O4 electrodeposited films for the oxygen evolution
reaction of water. J. Phys. Chem. C 118, 14073–14081
(2014).
92. Zhou, F. et al. Improvement of catalytic water oxidation
on MnOx films by heat treatment. ChemSusChem 6,
643–651 (2013).
93. Zaharieva, I. et al. Electrosynthesis, functional, and
structural characterization of a water-oxidizing
manganese oxide. Energy Environ. Sci. 5, 7081–7089
(2012).
94. Mattioli, G., Zaharieva, I., Dau, H. & Guidoni, L.
Atomistic texture of amorphous manganese oxides for
electrochemical water splitting revealed by ab initio
calculations combined with X‑ray spectroscopy. J. Am.
Chem. Soc. 137, 10254–10267 (2015).
95. Gorlin, Y. et al. In situ X‑ray absorption spectroscopy
investigation of a bifunctional manganese oxide
catalyst with high activity for electrochemical water
oxidation and oxygen reduction. J. Am. Chem. Soc.
135, 8525–8534 (2013).
96. Zhou, F. et al. Electrodeposited MnOx films from ionic
liquid for electrocatalytic water oxidation. Adv. Energy
Mater. 2, 1013–1021 (2012).
97. Takashima, T., Hashimoto, K. & Nakamura, R.
Mechanisms of pH‑dependent activity for water
oxidation to molecular oxygen by MnO2
electrocatalysts. J. Am. Chem. Soc. 134, 1519–1527
(2012).
98. Takashima, T., Hashimoto, K. & Nakamura, R.
Inhibition of charge disproportionation of MnO2
electrocatalysts for efficient water oxidation
under neutral conditions. J. Am. Chem. Soc. 134,
18153–18156 (2012).
99. Huynh, M., Bediako, D. K., Liu, Y. & Nocera, D. G.
Nucleation and growth mechanisms of an
electrodeposited manganese oxide oxygen evolution
catalyst. J. Phys. Chem. C 118, 17142–17152 (2014).
100. Huynh, M., Bediako, D. K. & Nocera, D. G. A.
Functionally stable manganese oxide oxygen evolution
catalyst in acid. J. Am. Chem. Soc. 136, 6002–6010
(2014).
101. Walter, M. G. et al. Solar water splitting cells. Chem.
Rev. 110, 6446–6473 (2010).
102. Hou, Y. et al. Bioinspired molecular co‑catalysts
bonded to a silicon photocathode for solar hydrogen
evolution. Nat. Mater. 10, 434–438 (2011).
103. Berglund, S. P. et al. p‑Si/W2C and p‑Si/W2C/Pt
photocathodes for the hydrogen evolution reaction.
J. Am. Chem. Soc. 136, 1535–1544 (2014).
104. Ding, Q. et al. Efficient photoelectrochemical hydrogen
generation using heterostructures of Si and chemically
exfoliated metallic MoS2. J. Am. Chem. Soc. 136,
8504–8507 (2014).
105. Cabán-Acevedo, M. et al. Efficient hydrogen evolution
catalysis using ternary pyrite-type cobalt
phosphosulphide. Nat. Mater. 14, 1245–1251
(2015).
106. Downes, C. A. & Marinescu, S. C. Efficient
electrochemical and photoelectrochemical H2
production from water by a cobalt dithiolene one-
dimensional metal–organic surface. J. Am. Chem. Soc.
137, 13740–13743 (2015).
107. McKone, J. R. et al. Evaluation of Pt, Ni, and Ni–Mo
electrocatalysts for hydrogen evolution on crystalline Si
electrodes. Energy Environ. Sci. 4, 3573–3583
(2011).
108. Morales-Guio, C. G. et al. Solar hydrogen production by
amorphous silicon photocathodes coated with a
magnetron sputter deposited Mo2C catalyst. J. Am.
Chem. Soc. 137, 7035–7038 (2015).
109. Lang, D., Cheng, F. & Xiang, Q. Enhancement of
photocatalytic H2 production activity of CdS nanorods
by cobalt-based cocatalyst modification. Catal. Sci.
Technol. 6, 6207–6216 (2016).
110. Meng, P., Wang, M., Yang, Y., Zhang, S. & Sun, L. CdSe
quantum dots/molecular cobalt catalyst co‑grafted
open porous NiO film as a photocathode for visible
light driven H2 evolution from neutral water. J. Mater.
Chem. A 3, 18852–18859 (2015).
111. Lin, C.‑Y., Lai, Y.‑H., Mersch, D. & Reisner, E.
Cu2O|NiOx nanocomposite as an inexpensive
photocathode in photoelectrochemical water splitting.
Chem. Sci. 3, 3482–3487 (2012).
112. Yang, C. et al. Engineering a Cu2O/NiO/Cu2MoS4
hybrid photocathode for H2 generation in water.
Nanoscale 6, 6506–6510 (2014).
113. Dubale, A. A. et al. Heterostructured Cu2O/CuO
decorated with nickel as a highly efficient photocathode
for photoelectrochemical water reduction. J. Mater.
Chem. A 3, 12482–12499 (2015).

12 | ARTICLE NUMBER 0003 | VOLUME 1 www.nature.com/natrevchem

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

114. Morales-Guio, C. G., Tilley, S. D., Vrubel, H.,
Grätzel, M. & Hu, X. Hydrogen evolution from a
copper(i) oxide photocathode coated with an
amorphous molybdenum sulphide catalyst. Nat.
Commun. 5, 3059 (2014).
115. Paracchino, A., Laporte, V., Sivula, K., Grätzel, M. &
Thimsen, E. Highly active oxide photocathode for
photoelectrochemical water reduction. Nat. Mater.
10, 456–461 (2011).
116. Fujishima, A. & Honda, K. Electrochemical photolysis of
water at a semiconductor electrode. Nature 238,
37–38 (1972).
The first photoelectrochemical cell for water
splitting.
117. Hodes, G., Cahen, D. & Manassen, J. Tungsten trioxide
as a photoanode for a photoelectrochemical cell (PEC).
Nature 260, 312–313 (1976).
118. Hardee, K. L. & Bard, A. J. Semiconductor electrodes: v.
The application of chemically vapor deposited iron
oxide films photosensitized electrolysis. J. Electrochem.
Soc. 123, 1024–1026 (1976).
119. Zhong, D. K., Sun, J., Inumaru, H. & Gamelin, D. R.
Solar water oxidation by composite catalyst/α-Fe2O3
photoanodes. J. Am. Chem. Soc. 131, 6086–6087
(2009).
120. Zhong, D. K. & Gamelin, D. R. Photoelectrochemical
water oxidation by cobalt catalyst (“Co–Pi”)/α-Fe2O3
Composite photoanodes: oxygen evolution and
resolution of a kinetic bottleneck. J. Am. Chem. Soc.
132, 4202–4207 (2010).
121. Zhong, D. K., Cornuz, M., Sivula, K., Grätzel, M. &
Gamelin, D. R. Photo-assisted electrodeposition of
cobalt–phosphate (Co–Pi) catalyst on hematite
photoanodes for solar water oxidation. Energy Environ.
Sci. 4, 1759–1764 (2011).
122. Seabold, J. A. & Choi, K.‑S. Effect of a cobalt-based
oxygen evolution catalyst on the stability and the
selectivity of photo-oxidation reactions of a WO3
photoanode. Chem. Mater. 23, 1105–1112 (2011).
123. Pijpers, J. J. H., Winkler, M. T., Surendranath, Y.,
Buonassisi, T. & Nocera, D. G. Light-induced water
oxidation at silicon electrodes functionalized with a
cobalt oxygen-evolving catalyst. Proc. Natl Acad. Sci.
USA 108, 10056–10061 (2011).
124. Abdi, F. F. & van de Krol, R. Nature and light
dependence of bulk recombination in Co-Pi‑catalyzed
BiVO4 photoanodes. J. Phys. Chem. C 116,
9398–9404 (2012).
125. Abdi, F. F., Firet, N. & van de Krol, R. Efficient BiVO4
thin film photoanodes modified with cobalt phosphate
catalyst and W‑doping. ChemCatChem 5, 490–496
(2013).
126. Pilli, S. K. et al. Light induced water oxidation on cobalt-
phosphate (Co–Pi) catalyst modified semi-transparent,
porous SiO2–BiVO4 electrodes. Phys. Chem. Chem.
Phys. 14, 7032–7039 (2012).
127. Khnayzer, R. S. et al. Structure and activity of
photochemically deposited “CoPi” oxygen evolving
catalyst on titania. ACS Catal. 2, 2150–2160 (2012).
128. Li, Y. et al. Efficient and stable photoelectrochemical
seawater splitting with TiO2@g-C3N4 nanorod arrays
decorated by Co–Pi. J. Phys. Chem. C 119,
20283–20292 (2015).
129. Kim, J. H. et al. Wireless solar water splitting device
with robust cobalt-catalyzed, dual-doped BiVO4
photoanode and perovskite solar cell in tandem: a dual
absorber artificial leaf. ACS Nano 9, 11820–11829
(2015).
130. Choi, S. K., Choi, W. & Park, H. Solar water
oxidation using nickel–borate coupled BiVO4
photoelectrodes. Phys. Chem. Chem. Phys. 15,
6499–6507 (2013).
131. Pilli, S. K., Summers, K. & Chidambaram, D. Ni–Ci
oxygen evolution catalyst integrated BiVO4
photoanodes for solar induced water oxidation. RSC
Adv. 5, 47080–47089 (2015).
132. Jin, T., Diao, P., Xu, D. & Wu, Q. High-aspect-ratio WO3
nanoneedles modified with nickel–borate for efficient
photoelectrochemical water oxidation. Electrochim.
Acta 114, 271–277 (2013).
133. Kleiman-Shwarsctein, A., Hu, Y.‑S., Stucky, G. D. &
McFarland, E. W. NiFe-oxide electrocatalysts for the
oxygen evolution reaction on Ti doped hematite
photoelectrodes. Electrochem. Commun. 11,
1150–1153 (2009).
134. Zeng, Q. et al. A novel in situ preparation method for
nanostructured α-Fe2O3 films from electrodeposited Fe
films for efficient photoelectrocatalytic water splitting
and the degradation of organic pollutants. J. Mater.
Chem. A 3, 4345–4353 (2015).
135. Tilley, S. D., Cornuz, M., Sivula, K. & Grätzel, M. Light-
induced water splitting with hematite: improved
nanostructure and iridium oxide catalysis. Angew.
Chem. Int. Ed. 49, 6405–6408 (2010).
136. Kenney, M. J. et al. High-performance silicon
photoanodes passivated with ultrathin nickel films for
water oxidation. Science 342, 836–840 (2013).
137. Kim, T. W. & Choi, K.‑S. Nanoporous BiVO4
photoanodes with dual-layer oxygen evolution catalysts
for solar water splitting. Science 343,
990–994 (2014).
138. Rocheleau, R. E., Miller, E. L. & Misra, A. High-efficiency
photoelectrochemical hydrogen production using
multijunction amorphous silicon photoelectrodes.
Energy Fuels 12, 3–10 (1998).
Possibly the first monolithic photoelectrochemical
solar‑to‑hydrogen device, one that still ranks as
among the most efficient.
139. Verlage, E. et al. A monolithically integrated,
intrinsically safe, 10% efficient, solar-driven water-
splitting system based on active, stable earth-
abundant electrocatalysts in conjunction with
tandem iii–v light absorbers protected by amorphous
TiO2 films. Energy Environ. Sci. 8, 3166–3172
(2015).
A recent example of a highly efficient photoelectro-
chemical solar‑to‑hydrogen device, showing how
effectively device design has been optimized since
Rocheleau’s report.
140. Reece, S. Y. et al. Wireless solar water splitting using
silicon-based semiconductors and earth-abundant
catalysts. Science 334, 645–648 (2011).
The first effective photoelectrochemical solar‑to-
hydrogen device that operated at near-neutral pH.
141. Esiner, S. et al. Photoelectrochemical water splitting
in an organic artificial leaf. J. Mater. Chem. A 3,
23936–23945 (2015).
142. Sathre, R. et al. Life-cycle net energy assessment of
large-scale hydrogen production via
photoelectrochemical water splitting. Energy Environ.
Sci. 7, 3264–3278 (2014).
143. Barbir, F. PEM electrolysis for production of hydrogen
from renewable energy sources. Sol. Energy 78,
661–669 (2005).
144. Berger, A., Segalman, R. A. & Newman, J. Material
requirements for membrane separators in a water-
splitting photoelectrochemical cell. Energy Environ.
Sci. 7, 1468–1476 (2014).
145. Symes, M. D. & Cronin, L. Decoupling hydrogen and
oxygen evolution during water splitting using a proton-
coupled-electron buffer. Nat. Chem. 5, 403–409
(2013).
146. Rausch, B., Symes, M. D. & Cronin, L. A. Bio-inspired,
small molecule electron-coupled-proton buffer for
decoupling the half-reactions of electrolytic water
splitting. J. Am. Chem. Soc. 135, 13656–13659
(2013).
147. Rausch, B., Symes, M. D., Chisholm, G. & Cronin, L.
Decoupled catalytic hydrogen evolution from a
molecular metal oxide redox mediator in water
splitting. Science 345, 1326–1330 (2014).
148. Bloor, L. G. et al. Solar-driven water oxidation and
decoupled hydrogen production mediated by an
electron-coupled-proton buffer. J. Am. Chem. Soc.
138, 6707–6710 (2016).
Acknowledgements
M.D.S. thanks the University of Glasgow for a Lord Kelvin
Adam Smith Research Fellowship.
Competing interests statement
The authors declare no competing interests.

NATURE REVIEWS | CHEMISTRY VOLUME 1 | ARTICLE NUMBER 0003 | 13

©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 O N L I N E O N LY

Subject categories
Physical sciences / Chemistry / Electrochemistry / Electrocatalysis
[URI /639/638/161/886]
Physical sciences / Chemistry / Green chemistry / Renewable energy
/ Hydrogen energy Hydrogen fuel
[URI /639/638/224/909/4086/4087]
Physical sciences / Chemistry / Energy
[URI /639/638/675]

ToC blurb

000 Earth-abundant catalysts for

electrochemical and photoelectrochemical
water splitting
Isolda Roger, Michael A. Shipman and Mark D. Symes Earth abundant catalysts?
Splitting water is an attractive means by which
energy — either electrical and/or light — is stored 2H2O 2H2 H2 H2 H2
and consumed on demand. Active and efficient
O2 + 4H+ 4e– 4H+
catalysts for anodic and cathodic reactions often
require precious metals. This Review covers base-metal Storage
catalysts that can afford high performance in a more
sustainable and available manner.
Fuel cells
©
2
0
1
7
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.