Please do not adjust margins
“This document is the Accepted Manuscript version of a Published Work that appeared in final form in Journal of
the American Chemical Society, copyright © 2015 American Chemical Society after peer review and technical
editing by the publisher. To access the final edited and published work see [insert ACS Articles on Request
author- directed link to Published Work, see http://pubs.acs.org/doi/abs/10.1021/acsami.5b09816

ACS Applied Materials & Interfaces
CuO functionalized silicon photoanodes for efficient
photoelectrochemical water splitting devices
a,b

b

a

b

a

Yuanyuan Shi, Carolina Gimbert-Suriñach, Tingting Han, Serena Berardi, Mario Lanza,*
b,c
Antoni Llobet*

One of the main difficulties for the technological development of photoelectrochemical (PEC) water splitting (WS) devices
is the synthesis of active, stable and cost-effective photoelectrodes that ensure high performance. Here we report the
development of a CuO-based photoanode, which shows an onset potential for the water oxidation of 0.53 V vs. SCE at pH
9, which implies an overpotential of only 75 mV, and high stability above 10 hours. The photoanodes have been fabricated
by sputtering a thin film of Cu0 on commercially available n-type Si wafers, followed by a photoelectrochemical treatment
in basic pH conditions. The resulting CuO/Cu layer acts as: (i) protective layer to avoid the corrosion of nSi, (ii) p-type hole
conducting layer for efficient charge separation and transportation, and (iii) electrocatalyst to reduce the overpotential of
the water oxidation reaction. The low cost, low toxicity and good performance of CuO-based coatings can be an attractive

Broader Context
The exploitation of solar energy as a renewable energy source has become one of the most popular alternatives to the consumption of
fossil fuels. Nevertheless, the intermittency and location-dependency of sunlight, as well as the low power conversion efficiency and the
still high price of photovoltaic (PV) solar cells, are the main factors hindering the massive use of PV technology. Recently, the use of sunlight
to split water to produce a clean fuel (i.e. hydrogen) has become very popular; in addtion this approach allows for the storage and
transportation of energy potentially at a relatively low cost. This strategy can be achieved using photoelectrochemical devices enabling
water splitting at the surface of semiconductor electrodes. However, in most semiconductors the water splitting chemical reaction is not
fast enough, plus it decays quickly due to photocorrosion. Thus, the development of cheap passivating and catalytic coatings has become a
major challenge. In this work, we present a solar water splitting device, which uses a n-type Si photoanode, protected with a CuO/Cu film.
The photoanodes not only show excellent activity and stability, but also low toxicity and price, as they have been synthesized using an
industry-compatible method. Our work opens the door to the use of CuO films for the development of water splitting devices.
solution to functionalize unstable materials for solar energy conversion

Introduction
Photoelectrochemical (PEC) cells exploit the built-in potential
generated
at
semiconductor-liquid
junctions
under
illumination to drive uphill chemical reactions, such as the
splitting of water (Eq. 1), which can be used to produce
1
hydrogen fuel in a clean manner. The transformation of
sunlight into chemical energy via the formation of new
chemical bonds is attractive because it allows solar energy to
2
be stored and transported. The overall electrochemical water
splitting can be divided in two half reactions, the reduction and
oxidation, also named hydrogen evolution reaction (HER, Eq.2)
and oxygen evolution reaction (OER, Eq. 3), respectively:

O2 + 2H2 (ΔE = 1.23 V ) (Eq. 1)
Overall water splitting: 2H2O
0
+
H2 ↑ ( E red = 0.00 V vs. NHE)
(Eq. 2)
HER: 2H + 2 e
0
+
O2 ↑ ( E ox = 1.23 V vs. NHE)
(Eq. 3)
OER: 2H2O + 4h

where the redox potential values are reported versus the
normal hydrogen electrode, NHE. The OER is considered the
bottleneck of the overall process, being thermodynamically
and kinetically demanding. Thus, enormous research efforts
have been invested on developing efficient photoanodes for
the OER, to be coupled, in a possible water spitting PEC device,
to a dark Pt cathode, performing the HER (Fig. 1). Several kinds
of photoanodes have been used in this PEC set-up, including
metal oxide semiconductors (Fe2O3, WO3, BiVO4, …), group IIIV semiconductors (GaAs, GaP, InP, …) and amorphous silicon,
all being excellent light absorbers and in most of the cases
3-5
with good carrier mobility.
Among them, silicon is a
particularly interesting candidate due to its low cost and good
ability to absorb light. Unfortunately, silicon is only able to
2
produce modest photocurrent densities of some µA/cm ,
which decay very fast (in the minutes time scale) due to
6
premature corrosion. The reason is that the thermodynamic
oxidation potential of Si is less positive than that of water
oxidation (1.23 V vs. NHE at pH 0), thus making it susceptible

Please do not adjust margins

Please do not adjust margins
ARTICLE

Journal Name
applying a high external voltage. In this work, we report
copper-based photoanodes for water oxidation that have been
fabricated from Cu films sputtered on nSi photoanodes. The
resulting devices work at very low overpotential (η = 75 mV)
and stability (10 hours) at pH = 9.0. These findings point out
the possibility of using copper-based composites as a
promising photoanode for PEC-WS devices.

Experimental
Sample Preparation

Fig. 1. Scheme of the PEC cell developed in this work. The cell consists
on a Cu-protected nSi photoanode connected to a Pt counter electrode
through a potentiostat. When the sunlight illuminates the surface of
the photoanode, it induces charge separation. The photogenerated
holes migrate to the surface and are used to produce O2 from water. At
the same time, the electrons promoted to the conduction band move
to the counter electrode to reduce protons to hydrogen.
7

to self-oxidation when interfaced with aqueous electrolytes.
Therefore passivating the surface of silicon with a cheap
coating that can serve as corrosion-resistant and water
oxidation catalyst (WOC) has become one of the main
8-16
challenges in this field.
Among all candidate materials, TiO2
has shown outstanding performance in protecting Si and other
4
III-V materials from oxidation, but in those cases an additional
WOC layer (film) is needed to boost the PEC reaction.
17
Superstructures made of Ir/TiO2 have shown water oxidation
activity and stability in acid and basic conditions, while
4
NiOX/TiO2 achieved even larger current densities, but only at
pH 14. In fact, Ni-based coatings can be used as both anticorrosion and WOC simultaneously. For example, Kenney et
18
al. evaporated a 2 nm Ni layer on a silicon wafer (with its
native oxide), leading to inexpensive Ni/SiO2/nSi photoanodes
with extraordinary activity at pH 14 and enhanced stability at
19
pH 9.5. Mei et al. deposited Fe-modified NiOX films on np+-Si
junctions (which notably enhance charge separation) and
achieved continuous OER during 300 hours. Recently Lewis’
group overcame that performance using 35-160 nm thick NiOx
coatings on np+-Si and achieved continuous OER during more
20
than 50 days at pH 14. Even though these devices show
higher activity, the cost of these photoanodes is still high, as
the fabrication of the np+ junction requires both implantation
and thermal processes.
Due to its low cost and toxicity, copper is an attractive material
from an industrial perspective, but its use in photoanodes for
PEC WS-devices has never been reported. Recent advances
21-25
showed that copper oxides,
as well as molecular copper
26-29
complexes,
can be used to catalyse the electrocatalytic
21
water oxidation reaction. Sun and coworkers reported a
continuous OER over 10 hours at pH 9 using a copper oxide
catalyst electrodeposited on fluorine doped tin oxide (FTO),
and a similar performance was achieved by Meyer and
22
coworkers by forming a CuO film on the surface of a Cu foil,
but in both cases the water oxidation reaction was induced by

All reagents were purchased from Aldrich unless otherwise
stated. The fluorine doped tin oxide (FTO) coated glass was
purchased from XopFisica (thickness, 2.3-3.0 mm; visible
2
transmittance, 80-81.5%; resistance, 6-9 Ω/cm ). Before the
deposition and characterization tests, the FTO-coated glass
was ultrasonically cleaned in water containing a cleaning
detergent (Hellma), deionized water and ethanol, for 10
minutes each. Standard single side polished and phosphorousdoped [100] n-type silicon wafers (0.3-0.5 Ω•cm) were
purchased from ProLog. Before all the deposition processes,
the Si wafers were cleaned with acetone, ethanol and
0
deionized water in an ultrasonic bath, for 10 minutes each. Cu
films with thicknesses of 5, 10, 15, 20 and 25 nm were
deposited onto both FTO-coated glass and nSi photoanodes in
the same vacuum chamber, by means of sputtering (ATC Orion
8-HV instrument, AJA International) at 100 W, with a
deposition rate of 1.6 Å/s for 40, 80, 120, 160 and 200 s,
respectively. An ohmic contact of 20 nm titanium was
previously deposited onto the backside of the nSi wafers by ebeam evaporation (PVD75 Kurt J. Lesker). Copper tape was
used to contact the working electrodes for electrochemical
and PEC experiments.
Sample Characterization
All the samples were tested at room temperature in a standard
three-electrode configuration with saturated calomel (SCE,
saturated KCl) or Hg/Hg2SO4 (saturated K2SO4) reference electrodes
and Pt wire as the counter electrode, using CHI 660D or CHI 660E
potentiostats (CH Instruments, Inc.). The FTO samples were tested
in a 20 mL one-compartment glass cell, and the nSi photoanodes in
a 140 mL teflon cell, the areas exposed to OER were of ~1.0 and
2
~0.45 cm , respectively. In order to quantify the amount of O2
produced, a two-compartment cell (in which anodic and cathodic
sides are separated by a glass frit) was used. In this set-up, the Cucoated nSi photoanodes were used as the working electrodes,
Ag/AgCl (saturated KCl) as the reference and a Pt mesh as the
counter electrode. All the cyclic voltammetry (CV) scans were
collected with iR compensation at a scan rate of 100 mV/s, while
the current density versus time traces were obtained without iR
compensation. For comprehensive comparison, all redox potentials
reported herein will be referenced versus SCE. An oxygen sensitive
OXNP15121 Clark electrode (Unisense) was used to measure the
photoproduced O2 under air, and in turn to calculate the Faradaic
efficiency. The electrolyte consisted of 0.2 M borate buffer solution
(pH 9), obtained by mixing a 0.05 M Na2B4O7 solution and a 0.2 M

Please do not adjust margins

Please do not adjust margins
Journal Name

ARTICLE

H3BO3 solution in a ratio of 8:2 (v/v). All pH values were measured
with a HI 4222 pH-meter (Hanna Instruments) using a calibrated
Crison 5029 electrode (Crison Instruments). For the PEC
characterization of the photoanodes, illumination was provided by
150 W Xe Arc Lamps (LS-150, ABET technology and 66906, Newport
Corporation), equipped with a UV-light cut-off filter (λ < 400 nm)
2
and calibrated to 1 sun (100 mW/cm ) using a calibrated silicon
photodiode at 25 °C.
UV-Vis characterizations were performed using a Cary 50 (Varian)
spectrophotometer. The morphological characterization of the
samples was performed by means of Transmission Electron
Microscopy (TEM, JEOL, model 1011) and Scanning Electron
Microscopy (SEM, FEI Quanta 200FEG). The chemical composition
of the samples was analyzed by X-ray Photoelectron Spectroscopy
(XPS), using a KRATOS Axis ultra-DLD spectrometer equipped with a
monochromatic Mg Kα X-ray source (hν = 1283.3 eV).

suggesting that the copper oxide active species may have
similar performances.
The stability of the generated CuO/Cu/FTO electrodes was
assessed by means of successive CV scans, as well as by
controlled potential electrolysis (CPE) at 1.1 V vs. SCE (Figs. 2b
and S2a). In particular, Fig. 2b shows the remarkable stability
of the current density trace of a 15 nm-thick CuO/Cu/FTO
electrode. This sample maintains 80% of its activity after 12
hours of continuous OER. The small current decay observed is
2+
likely due to a loss of Cu in solution, as evidenced by both
electrochemical experiments on the electrolytic solution at the
end of the CPE experiment (Fig. S3) and UV-Vis spectroscopy
before and after 1 hour of CPE at 1.1 V vs. SCE (Fig. S2). The
ability of the films to oxidize water, together with their good
stability and transparency is a further evidence that the thin
sputtered Cu layer can act as an anti-corrosion and WOC film.
Copper films on nSi: photoelectrocatalytic water oxidation

Results and discussion
Copper films on FTO: electrocatalytic water oxidation
Metallic copper films of different thicknesses (5, 10, 15, 20 and
25 nm) were sputtered on FTO substrates. The resulting films
are not homogenous as shown by the TEM images reported in
Fig. S1, where it can be seen that the surface morphology
depends on their thickness; the thinner Cu films appear to be
rougher than the thicker ones. The electrocatalytic water
oxidation ability of the sputtered copper electrodes was tested
in 0.2 M borate buffer solution at pH 9. Unless otherwise
stated, all the reported electrochemical measurements were
performed in this electrolytic solution. As shown in Fig. 2a,
regardless of their thickness, all the electrodes show similar
onset potentials for the water oxidation reaction at 0.95 V vs.
SCE (η = 495 mV), ~300 mV lower than that of the bare FTO.
2
Current densities up to 7.5 mA/cm were obtained at 1.1 V vs
SCE. Similar results have been reported for other active CuO
21-22,24-25
films, prepared using electrodeposition techniques,

Fig. 2. Electrochemical water oxidation performances of
CuO/Cu/FTO electrodes. (a) CVs of bare FTO (dashed black curve)
and 5, 10, 15, 20 and 25 nm-thick Cu/FTO in 0.2 M borate buffer
solution (pH 9). All the CV scans were obtained with iR
compensation by using a SCE reference electrode and Pt counter
electrode at a scan rate of 100 mV/s. (b) Current density traces
obtained by controlled-potential electrolysis (CPE) at 1.1 V vs. SCE
(without iR compensation) for a 15 nm CuO/Cu/FTO electrode in
0.2 M borate buffer solution (pH 9).

Copper-coated nSi photoanodes were prepared by sputtering
using the same methodology and parameters employed to
prepare the Cu/FTO electrodes. The native SiO2 on the nSi
wafer was not etched, as it can serve as adhesive layer [18],
and a layer of 20 nm Ti was evaporated on the back side of the
nSi wafer, leading to Cu/SiO2/nSi/Ti photoanodes. In total, nSi
photoanodes with three different Cu0 thicknesses (5, 10 and
15 nm) were prepared, avoiding thicker films, which can
hinder the photon absorption by silicon.
The current-density versus potential (J-E) plots in Fig. 3a
obtained from CV scans for the as-grown 5 nm Cu/SiO2/nSi/Ti
photoelectrodes show very low activity and a large onset
potential (~0.71 V vs. SCE). However, these Cu-coated nSi
photoanodes can be activated by performing either (i) several

Fig. 3. Activation of the Cu/SiO2/nSi/Ti photoanodes. (a) CV scans of 5
nm Cu/SiO2/nSi/Ti before (red solid line) and after (blue solid line)
activation under 1 sun illumination (see main text for activation
details). The trace obtained in dark conditions is also reported (dashed
black line). The CVs were obtained with iR compensation using a Pt
counter electrode and Hg/Hg2SO4 reference electrode at a scan rate of
100 mV/s in 0.2 M borate buffer solution (pH 9). Insets: two 3D
schematic representations, illustrating the activation process on the
Cu/SiO2/nSi/Ti photoanodes. The activation of the photoanode may be
related to the formation of CuO on Cu layer.

Please do not adjust margins

Please do not adjust margins
ARTICLE

Journal Name

Fig. 4. (a) Overall XPS spectra of 5 nm Cu/nSi/SiO2/Ti. Fresh
sample (red), activated sample (green), degraded sample (blue).
(b)-(d) XPS binding energy regions of O 1s, Cu 2p and Si 2p,
respectively.

CV scans in a larger potential range (from -0.4 V to 0.8 V vs.
SCE), (ii) chopped light linear sweep voltammetry (LSV) in the
range from -0.4 to 0.9 V vs. SCE, or (iii) short time
amperometric current vs. voltage (I-t) experiments (Fig. S4).
After the activation process, a significant shift of the onset
potential for OER (from ~0.71 V to ~0.53 V vs. SCE) and a
remarkable increase of the current density (up to 14.5
2
mA/cm ) can be obtained (compare red and blue lines in Fig.
3). The mechanism of the activation process for the
Cu/SiO2/nSi/Ti photoanodes is complex, possibly involving
both soluble and insoluble products and multiple oxidation
I
II
III
IV
states (Cu , Cu , Cu , and/or Cu species) depending on the
22, 30,31
applied potential.
Two three-dimensional (3D) schematic
representations of the electrode configuration are shown in
the inset of Fig 3, indicating that the activation of the
photoanodes may be related to the formation of CuO on the
Cu layer.
In order to identify the changes induced during the activation
process, the chemical composition of the 5 nm Cu/SiO2/nSi/Ti
photoanodes was analyzed by means of XPS. Fig. 4 shows the
XPS peaks of O 1s, Cu 2p and Si 2p for the as-grown (red lines),
activated (green lines) and used (blue lines) samples. The term
"used" refers to a fresh 5 nm Cu/SiO2/nSi/Ti photoanode that
was exposed to 20 hours CPE at 0.8 V vs. SCE under 1 sun
illumination (Fig. S7a). As reference, the wide binding energy
regions of the samples are also exhibited in Fig. 4a. The O 1s
binding energy region of fresh samples (red trace in Fig. 4b)
shows just one component at 532 eV, corresponding to the
32
oxygen on the surface. This feature confirms the presence of
a metallic copper component, since metallic Cu is reported to
32
attract oxygen to the surface. After the activation process
(green trace in Fig. 4b), a new peak at 529 eV was observed,
32
which can be attributed to Cu-O, supporting the formation
of a CuO film at the surface of the photoanodes. In the used
sample (blue trace in Fig. 4b), the Cu-O peak at 529 eV is still

present, but a decreasing in its intensity indicates a loss of the
CuO, which may detach from the surface. On the other hand,
in the Cu 2p binding energy region, the as-grown samples
show two peaks at 932.7 eV and 952.6 eV (red trace in Fig. 4c),
respectively assigned to 2p3/2 and 2p1/2 binding energies of
33
Cu0 metal. After the activation process, the 2p peaks shift to
934.4 eV and 954.2 eV (green trace in Fig. 4c), indicating the
oxidation of copper to higher oxidation states, consistent with
the formation of CuO, as already pointed out by the analysis of
the O 1s region, and also reported for copper foils
electrochemically exposed to high positive potentials,
22
previously reported in the literature. Finally, the clear
appearance of the Si 2p peak in the spectrum of the used
sample (blue trace in Fig. 4d) further supports that the
decrease in activity observed in these photoanodes may be
related to partial loss of the CuO film. The reported XPS
analyses are fully reproducible, and similar results have been
also obtained for the 10 and 15 nm thick counterparts (Figs. S5
and S6).
Fig. 5 reports the J-E plots measured for the
CuO/Cu/SiO2/nSi/Ti photoanodes (obtained from the
activation of 10 nm Cu/SiO2/nSi/Ti material) under dark and
under 1 sun illumination (blue dashed and solid lines
respectively) and the non-photoactive 10 nm thick
CuO/Cu/FTO electrode (green solid line) at pH 9. As it can be
observed, the effect of the CuO film on the FTO produces a
remarkable decrease on the onset potential for OER (~300 mV,
from 1.25 V to 0.95 V vs. SCE). When the CuO film is formed on
the nSi/Ti substrate after the activation process, a remarkable
further onset potential reduction of ~420 mV is achieved
under illumination, leading to an absolute value of 0.53 V vs.
SCE. This represents an impressive overpotential of 75 mV with

Fig. 5. Comparison of the onset potential for the water oxidation
reaction by the different (photo) electrodes prepared in this work.
Bare FTO (red line); 10 nm-thick CuO/Cu/FTO (green line); activated 10
nmCuO/Cu/SiO2/nSi/Ti in the dark (dashed blue line) and under 1 sun
illumination (solid blue line). All the CVs were recorded using a Pt mesh
as the counter electrode and a SCE or Hg/Hg2SO4 as the reference
electrode, with iR compensation at a scan rate of 100 mV/s in a 0.2 M
borate buffer solution (pH

Please do not adjust margins

9).

Please do not adjust margins
Journal Name

ARTICLE

Fig. 6. Stability and faradaic efficiency of the activated 10 nm CuO/Cu/SiO2/nSi/Ti photoanodes. (a) Current density versus time traces obtained
by controlled potential electrolysis at 0.8 V vs. SCE on two different10 nm Cu/SiO2/nSi/Ti photoanodes in 0.2 M borate buffer (pH 9) under 1
sun illumination, showing good reducibility. (b) Oxygen evolution trace (red solid line) measured with an activated 10 nm Cu/SiO2/nSi/Tiin 0.2
M borate buffer (pH 9) at 0.85 V vs. SCE, under 1 sun illumination. From the measured and the theoretical (black dashed line) oxygen amounts,
a ~100 % faradaic efficiency can be calculated. Both (a) and (b) were obtained without iR compensation, illuminated by a simulated 1 sun and
using a Pt mesh as counter electrode. A one-compartment cell made of Teflon and a SCE reference electrode were used for tracing the current
density (a) and a two-compartments glass cell and a Ag/AgCl reference electrode were used for the oxygen detection experiment (b).

regard to thermodynamic value.
The long-term stability of the produced photoanodes was
tested by means of CPE experiments at 0.8 V vs. SCE. Fig. 6a
shows the current density traces obtained for two activated 10
nm CuO/Cu/SiO2/nSi/Ti photoanodes, showing excellent
reproducibility and negligible variability. Once activated, both
the photoanodes are stable for 10 hours at a constant current
2
density of 0.24 mA/cm under PEC conditions. The overall (≥
20 hours) CPE plots for 5, 10 and 15 nm-thick Cu/SiO2/nSi/Ti
photoanodes indicate that the activation process (i.e. the
formation of the CuO film) is needed for all Cu thicknesses, and
that thicker films need a longer activation process (Fig. S7).
The activation of the photoanodes takes place in a
characteristic two-steps process: (i) a reduction of the initial
current, probably related to a decreased conductivity
associated to the transition from metallic Cu to CuO; and (ii) a
current increase probably related to the formation of the CuO
WOC film that boosts the OER and generates large current
intensities. After that, the performance of the photoanodes
becomes stable during long periods of time (up to 10 hours).
The surface of the samples has been further analysed by SEM
(Fig. S8). While fresh 10 nm Cu/SiO2/nSi/Ti photoanodes
display very flat surfaces, the same samples after 37 hours CPE
tests appear to be covered by nanoparticles. However,
depending on the analysed region of the photoanode surface,
a partial depletion of the number of CuO nanoparticles can be
observed (e.g. in Fig. S8d), confirming that material loss is the
major cause of current drop after long-term stability tests of
this kind of photoanodes.

Finally, in order to prove that the photocurrent observed in
the CPE experiments with the CuO/Cu/SiO2/nSi/Ti
photoanodes is actually due to the water oxidation reaction, a
quantitative O2 measurement was performed, using an oxygen
gas sensor (Clark electrode). Fig. 6b shows the O2 evolution
profile obtained with an activated 10 nm CuO/Cu/SiO2/nSi/Ti,
held at a constant potential of 0.85 V vs. SCE, under simulated
1 sun illumination. A ~100% Faradaic efficiency can be
calculated considering the theoretical and actual volume of O2
produced (respectively, black dashed and red solid lines in Fig.
6b). The small delay in the experimental measurement of the
evolved O2 by the Clark electrode with regard to the measured
current is due to the slow diffusion of oxygen into the
membrane of the electrode sensor.

Conclusions
In conclusion, we have fabricated nSi photoanodes for PEC
water splitting devices using for the first time CuO as water
oxidation catalyst at the surface of a semiconductor. The
devices show high activity at an impressively low overpotential
of 75 mV. Furthermore they are stable for over 10 hours under
1 sun illumination in 0.2 M borate buffer at pH 9, with Faradic
efficiencies close to 100%. Compared to the copper-based FTO
electrode, the CuO-coated nSi photoanode shows an onset
potential decrease for the OER of 420 mV under 1 sun
illumination, which clearly highlights the advantages of this
approach.

Please do not adjust margins

Please do not adjust margins
ARTICLE

Journal Name

Finally it is important to stress here that the photoelectrodes
described in the present work are made of Copper that is a
cheap, earth-abundant and low toxicity metal. Furthermore all
the operations carried out for the construction of the
photonade are an industrially compatible and thus could be
easily scale up for commercial applications. It thus offers a
promising solution for the generation of clean and sustainable
solar fuel. The conclusions section should come in this section
at the end of the article, before the acknowledgements.

20
21
22
23
24

Acknowledgements

25

The start-up funding from Soochow University (China),
MINECO (CTQ-2013-49075-R, SEV-2013-0319; CTQ-201452974-REDC), “La Caixa” foundation and AGAUR (2014-SGR915) are acknowledged for financial support. C.G.S is grateful
to AGAUR for a "Beatriu de Pinós" postdoctoral grant. S.B. is
grateful to the ICIQ-IPMP Marie Curie COFUND Project
(291787ICIQ-IPMP). European COST actions, CM1202 and
CM1205 are also gratefully acknowledged.

26
27
28
29
30

References

31

1

32

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

N. S. Lewis and D. G. Nocera, Proc. Natl. Acad. Sci. U.S.A.,
2006, 103, 15729-15735.
J. A. Turner, Science, 2013, 342, 811- 812.
S. Hu, C. Xiang, S. Haussener, A. D. Berger and N. S. Lewis,
Energy Environ. Sci., 2013, 6,2984-2993.
S. Hu, M. R. Shaner, J. A. Beardslee, M. Lichterman, B. S.
Brunschwigand N. S. Lewis, Science, 2014, 344, 1005-1009.
R. Liu, Z. Zheng, J. Spurgeon and Xiaogang Yang, Energy
Environ. Sci., 2014, 7,2504.
A. Q. Contractor and J. O. M. Bockris, Electrochim. Acta,
1984, 29, 1427-1434.
S. Chen, and L.-W. Wang, Chem. Mater., 2012, 24, 36593666.
M. G. Walter, E. L. Warren, J. R. McKone, S. W. Boettcher, Q.
Mi, E. A. Santori and N. S. Lewis, Chem. Rev., 2010, 110,
6446-6473.
M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous
Solutions (National Association of Corrosion Engineers,
Houston, TX, ed. 2, 1974).
H. Gerischer, Faraday Discuss. Chem. Soc., 1980, 70, 137151.
K. Sivula, J. Phys. Chem. Lett., 2013,4, 1624-1633.
B. Klahr, S. Gimenez, F. Fabregat-Santiago, J. Bisquert andT.
W. Hamann, J. Am. Chem. Soc., 2012, 134, 16693-16700.
T. Hisatomi, F. Le Formal, M. Cornuz, J. Brillet,N. Tetreault, K.
Sivula and M. Gratzel, Energy Environ. Sci., 2011, 4, 25122515.
M. Liu, C.-Y. Nam, C. T. Black, J. Kamcev and L. Zhang, J.Phys.
Chem. C., 2013, 117, 13396-13402.
Z. Li, W. Luo, M. Zhang, J. Feng and Z. Zou, Energy Environ.
Sci., 2013, 6, 347-370.
Z. Zhang, R. Dua, L. Zhang, H. Zhu, H. Zhang and P. Wang, ACS
Nano, 2013, 7, 1709-1717.
Y. W. Chen, J. D. Prange, S. Dühnen, Y. Park, M. Gunji, C. E. D.
Chidsey& P. C. McIntyre, Nat.Mater., 2011, 10, 539-544.
M. J. Kenney, M. Gong, Y. Li, J. Z. Wu, J. Feng, M. Lanza & H.
Dai, Science, 2013, 342,836-840.
B. Mei, A. A. Permyakova, R. Frydendal, D. Bae, T.
Pedersen,P. Malacrida, O. Hansen,I. E. L. Stephens, P. C. K.

33

Vesborg, B. Segerand I. Chorkendorff,J. Phys. Chem. Lett.,
2014, 5, 3456-3461.
K. Sun, M. T. McDowell, A. C. Nielander, S. Hu, M. R.
Shaner,F. Yang, B. S.Brunschwig and N. S. Lewis, J. Phys.
Chem. Lett., 2015, 6, 592-598.
F. Yu, F. Li, B. Zhang, H. Li and L. Sun, ACS Catal., 2015, 5,
627-630.
J. Du, Z. Chen, S.Ye, B. J. Wiley and T. J. Meyer, Angew.
Chem. Int. Ed., 2015, 54, 2073-2078.
Z. Chen and T. J. Meyer, Angew. Chem. Int. Ed., 2013, 52,
700-703; Angew. Chem., 2013, 125, 728-731.
X. Liu, H. Jia, Z. Sun, H. Chen, P. Xu and P. Du, Electrochem.
Commun., 2014, 46, 1-4.
X. Liu, S. Cui, Z. Sun and P. Du, Electrochim. Acta, 2015, 160,
202-208.
M.-T. Zhang, Z. Chen, P. Kang and T. J. Meyer, J. Am. Chem.
Soc., 2013, 135, 2048-2051.
S. M. Barnett, K. I. Goldberg and J. M. Mayer, Nat. Chem.,
2012, 4, 498-502.
T. Zhang, C. Wang, S. Liu, J.-L. Wang and W. Lin, J. Am. Chem.
Soc., 2014, 136, 273-281.
P. Garrido-Barros, I. Funes-Ardoiz, S. Drouet, J. B. Buchholz,
F. Maseras and A. Llobet, J. Am. Chem. Soc., 2015, 137, 67586761.
G. Kear, B. D. Barkerand F. C. Walsh, Corros. Sci., 2004, 46,
109-135.
L. D. Burke, M. J. G. Ahern and T. G. Ryan, J. Electrochem.
Soc., 1990, 137, 553-56.
Y.-K. Hsu, C.-H.Yu, Y.-C.Chen and Y.-G. Lin, Electrochimica
Acta, 2013, 105, 62-68.
C. D. Wanger, W. M. Riggs, L. E. Davis, J. F. Moulder and G. E.
Muilenberg in Handbook of x-ray photoelectron
spectroscopy: a reference book of standard data for use in xray photoelectron spectroscopy, PerkinElmer Corp., Physical
Electronics Division, Eden Prairie, Minnesota, USA, 1979, p.
82.

Please do not adjust margins