ChemInform Abstract: Nano-Architecture and Material Designs for Water Splitting Photoelectrodes

Article (PDF Available)inChemical Society Reviews 41(17):5654-71 · July 2012with207 Reads
DOI: 10.1039/c2cs35019j · Source: PubMed
Abstract
This review concerns the efficient conversion of sunlight into chemical fuels through the photoelectrochemical splitting of water, which has the potential to generate sustainable hydrogen fuel. In this review, we discuss various photoelectrode materials and relative design strategies with their associated fabrication for solar water splitting. Factors affecting photoelectrochemical performance of these materials and designs are also described. The most recent progress in the research and development of new materials as well as their corresponding photoelectrodes is also summarized in this review. Finally, the research strategies and future directions for water splitting are discussed with recommendations to facilitate the further exploration of new photoelectrode materials and their associated technologies.
5654 Chem. Soc. Rev., 2012, 41, 5654–5671 This journal is
c
The Royal Society of Chemistry 2012
Cite this:
Chem. Soc. Rev
., 2012, 41, 5654–5671
Nano-architecture and material designs for water
splitting photoelectrodes
Hao Ming Chen,
a
Chih Kai Chen,
a
Ru-Shi Liu,*
a
Lei Zhang,
b
Jiujun Zhang*
b
and
David P. Wilkinson
c
Received 19th January 2012
DOI: 10.1039/c2cs35019j
This review concerns the efficient conversion of sunlight into chemical fuels through the
photoelectrochemical splitting of water, which has the potential to generate sustainable hydrogen
fuel. In this review, we discuss various photoelectrode materials and relative design strategies with
their associated fabrication for solar water splitting. Factors affecting photoelectrochemical
performance of these materials and designs are also described. The most recent progress in the
research and development of new materials as well as their corresponding photoelectrodes is also
summarized in this review. Finally, the research strategies and future directions for water splitting
are discussed with recommendations to facilitate the further exploration of new photoelectrode
materials and their associated technologies.
1. Introduction
Sunlight is an inexpensive, non-polluting, abundant and endlessly
renewable source of clean energy. The amount of energy that
strikes the Earth yearly in the form of sunlight is approximately
ten thousand times the total energy that is consumed on this planet,
so converting solar energy into an easily usable form has
attracted considerable interest in the last several decades.
1
With
respect to solar energy conversion, four major technologies are
currently available for converting sunlight into useful forms of
energy such as electricity and hydrogen, as discussed below.
1.1 Technologies for converting sunlight
Photovoltaics. Photovoltaic technology, used to directly
convert sunlight into electricity, has been recognized as one
of the most viable clean energy technologies for generating
electrical power on a large scale. During the generation of
electricity from solar energy using photovoltaics, the first
process is the collection of sunlight and inducement of a
charge-separated state for electron–hole pair generation. The
photon absorption process that occurs in a silicon semiconductor
film causes the migration of excitons to a p–n junction in which
a
Department of Chemistry, National Taiwan University, Taipei 106,
Taiwan. E-mail: rsliu@ntu.edu.tw; Fax: +886-2-33668671;
Tel: +886-2-33661169
b
Institute for Fuel Cell Innovation, National Research Council of
Canada, Vancouver, BC, Canada V6T 1W5.
E-mail: jiujun.zhang@nrc.gc.ca; Fax: +1 604 221 3001;
Tel: +1 604 221 3087
c
Department of Chemical and Biochemical Engineering,
University of British Columbia, Vancouver, BC, Canada V6T 1Z3.
E-mail: dwilkinson@chml.ubc.ca; Fax: +1 604 221 3019;
Tel: +1 604 221 3019
Hao Ming Chen
Hao Ming Chen is currently a
postdoctoral research fellow at
National Taiwan University.
He received his MS degree in
2004 and his PhD in chemistry
from National Taiwan University
in 2008, where he worked on
the synthesis of nanocrystals
for particular applications.
His current research interests
include the synthesis of
nanomaterials, metallic nano-
crystals and semiconductor
nanomaterials for solar energy
conversion.
Chih Kai Chen
Chih Kai Chen is currently a
PhD student at the Department
of Chemistry, National Taiwan
University, supervised by
Prof. Ru-Shi Liu. He received
his BS degree in chemistry in
2009 from National Central
University. He is working on
the synthesis of semiconductor
nanomaterials, quantum dots,
and nanomaterials for solar
energy conversion.
Chem Soc Rev
Dynamic Article Links
www.rsc.org/csr TUTORIAL REVIEW
Published on 04 July 2012. Downloaded by National Research Council Canada on 26/04/2016 20:15:27.
View Article Online
/ Journal Homepage
/ Table of Contents for this issue
This journal is
c
The Royal Society of Chemistry 2012 Chem. Soc. Rev., 2012, 41, 5654–5671 5655
charge separation occurs, producing an electromotive force.
When this p–n junction is electrically connected to an external
load, a photo-current flows through the load, generating power,
realizing the photo-to-electricity conve rsion process.
2
In the pre-
sent development, the energy conversio n efficienc y of commer-
cially available silicon photovoltaics is around 18%. In addition to
the improvement in efficiency of the photo-electricity conver-
sion, close attention must be paid to the storage and distribu-
tion of solar energy for practical applications. The most feasible
technology involves converting the electrical energy into some
form of chemical fuel such as hydrogen or other small, low-
carbon organics, which can be easily stored and transported.
Ru-Shi Liu
Ru-Shi Liu is currently a
professor at the Department
of Chemistry, National Taiwan
University. He received his
Bachelor’s degree in Chemistry
from Shoochow University
(Taiwan) in 1981. He
received his Master’s degree
in nuclear science from the
National Tsing Hua University
(Taiwan) in 1983. He
obtained two PhD degrees in
chemistry one from National
Tsing Hua University in 1990
and one from the University of
Cambridge in 1992. He
worked at Materials Research Laboratories at the Industrial
Technology Research Institute from 1983 to 1985. He was an
Associate Professor at the Department of Chemistry of National
Taiwan University from 1995 to 1999, when he was promoted to
a professorship in 1999. His research concerns the field of
Materials Chemistry. He is an author or coauthor of more than
400 publications in scientific international journals. He has also
been granted more than 80 patents.
Lei Zhang
Lei Zhang is a Research
Council Officer at National
Research Council of Canada
Institute for Fuel Cell Inno-
vation. She received her first
MSc, majoring in Materials
Chemistry from Wuhan
University, China, in 1993 and
her second MSc in Materials/
Physical Chemistry from
Simon Fraser University,
Canada, in 2000. Ms Zhang’s
main research interests include
PEM fuel cell electrocatalysis,
catalyst layer/electrode
structure, metal-air batteries
and supercapacitors. Ms Zhang has co-authored 50 refereed
journal papers and holds 3 US patent applications. Ms Zhang is
an adjunct Professor of Federal University of Maranhao, Brazil,
and Zhengzhou University, China, respectively. She is also
an active member of the Electrochemical Society and the
International Society of Electrochemistry.
Jiujun Zhang
Jiujun Zhang received his
BSc and MSc from Peking
University in 1982 and 1985,
and his PhD in Electro-
chemistry from Wuhan
University in 1988, and he is
now a Senior Research
Officer and PEM Catalysis
Core Competency Leader at
the National Research Council
of Canada Institute for Fuel
Cell Innovation (NRC-IFCI).
Dr Zhang holds several
adjunct professorships,
including one at the University
of Waterloo and one at the
University of British Columbia. His research is mainly based
on fuel cell catalysis development.
David P. Wilkinson
David P. Wilkinson received
his BASc degree in Chemical
Engineering from the University
of British Columbia (UBC) in
1978 and his PhD degree in
Chemistry from the University
of Ottawa in 1987 where his
graduate work was done with
Dr Brian Conway. He has
over 20 years of industrial
experience in the areas of fuel
cells and advanced rechargeable
lithium batteries. In 2004
Dr Wilkinson was awarded a
Tier 1 Canada Research Chair
in Clean Energy and Fuel
Cells in the Department of Chemical and Biological Engineering
at the University of British Columbia. He is currently the
Director of the Clean Energy Research Centre (CERC) at the
university. He maintains a joint appointment with the University
and the Canadian National Research Council Institute for Fuel
Cell Innovation. Prior to his university appointment Dr Wilkinson
was the Director, and then Vice President of Research and
Development at Ballard Power Systems Inc., involved with the
research, development and application of fuel cell technology.
Dr Wilkinson is one of the leading all-time fuel cell inventors by
issued US patents. Dr Wilkinson’s main research interest is in
electrochemical power sources and processes to create clean and
sustainable energy.
Published on 04 July 2012. Downloaded by National Research Council Canada on 26/04/2016 20:15:27.
View Article Online
5656 Chem. Soc. Rev., 2012, 41, 5654–5671 This journal is
c
The Royal Society of Chemistry 2012
However, the major challenge faced by the current technology
in making photovoltaics that can be competitive with fossil fuels
is its high cost, which is two to five times higher than that of
conventional energy technologies.
3
Photo-materials frequently
used in photovoltaics are crystalline silicon wafers with thick-
nesses of 180–300 mm, but both the materials and their asso-
ciated processing are expensive. To overcome this challenge of
high material cost, various types of technologies with increasing
absorbance of visible light have been explored, including dye-
sensitized solar cells and organic photovoltaics.
4,5
In contrast
to silicon photovoltaics, these systems use relatively cheap
materials and may be considerable for future utilization.
Thin-film photovoltaics. To reduce the cost of photovoltaics,
thin-film photovoltaics have been explored due to the reduced
use of expensive silicon-based photo-active materials. Some
thin-film solar cells with photo-active films in the thickness
range of 1–2 mm, fabricated on cheap substrates such as glass,
stainless steel or plastic, have been investigated and validated.
In terms of improving the photo-electricity conversion efficiency,
the semiconductors that are used to harvest solar energy in these
thin-film solar cells are typically made of GaAs, CdTe CuInSe
2
,
as well as amorphous/polycrystalline silicon as a substitute for
expensive conventional crystalline silicon.
3
Wet-chemical photosynthesis. In order to explore new sunlight
conversion technology by learning from nature, wet-chemical
photosynthesis technology, sometimes called artificial photo-
synthesis, has been investigated. The goal is to achieve efficient
and stable energy conversion, which utilizes only earth-abundant
materials and operates under mild conditions. In the long-term,
photosynthetic technology should be able to generate solar fuels
by mimicking chemical reactions occurring in natural biological
systems. Photosynthesis is a wet chemical process that transforms
carbon dioxide into organic compounds, particularly sugars
(chemical fuels) using the energy from sunlight. The harvesting
and storing of solar energy via chemical methods, which occurs
naturally in photosynthesis, is an attractive strategy for meeting
the challenges of solar energy applications.
Photoelectrolysis. Combining the concepts of photovoltaics
and wet-chemical photosynthesis, photoelectrolysis is the ability to
split water through solar energy to produce hydrogen (a chemical
fuel). In photoelectrolysis, a solid photoelectrode is adopted to
convert the energy of sunlight into splitting water and generating
chemical energy and the process can be assisted electrochemically.
It has been recognized that hydrogen is playing a critical role in the
development of green energy economy since it is an ultimate form
ofcleanenergyandcanbeusedinfuelcellsandotherprocesses.
However, currently hydrogen is primarily formed by steam
reforming, in which fossil fuels are consumed and carbon
dioxide is generated.
6
Photoelectrolysis can be utilized to split
water into hydrogen and oxygen without any emission of
byproducts. Nevertheless, the conversion efficiency today remains
lower than that of photovoltaics, and is limited mainly by the
low performance of the photoelectrodes.
The earliest work on photoelectrolysis using a photoelectro-
chemical cell that contained a photoanode and a photocathode
on which anodic/cathodic reactions occur in the splitting
of water was carried out by Honda and Fujishima in 1970s,
using TiO
2
as the photoelectrode.
7
In their work, both electrons
and holes were generated when the semiconductor electrode
(TiO
2
) was irradiated by UV illumination, and the photo-
generated electrons reduced water to form hydrogen on a Pt
counter-electrode as the holes oxidized water to form oxygen on
the surface of the TiO
2
-based electrode, which was maintained
at a certain electrode potential. However, most natural sunlight
is visible light, and so the use of visible light to irradiate a
photoelectrode and generate hydrogen (and oxygen) seems more
practical than the use of UV light. Hence, research of visible-
light effective photoelectrodes for water splitting using new
materials for both anodic–cathodic processes should focus on
the goal of realizing an efficient photoelectrochemical cell that
can simultaneously drive both the hydrogen generation and
water oxidation reaction under visible light radiation. Such a
photoelectrochemical cell requires semiconductor materials
that can provide rapid charge transfer at the semiconductor/
electrolyte interface, long-term stability, and efficient harvesting
of a wide range of the solar spectrum. Accordingly, a multi-
junction configuration, which uses p- and n-type semiconductors
with various band gaps, and surface-bound electrocatalysts, has
dominated the development of efficient photoelectrochemical
cells.
8
Although substantial efforts have been made in the last
several decades to develop photoelectrochemical cells and
photocatalysts using various semiconductor materials,
6,9–12
no
breakthrough in either the material design of the photoelectrode
or material stability has yet been achieved. With recent advances
in nanotechnology, the research field has begun attracting much
more attention for many applications. Over the past decade,
nanomaterials have been used for catalysis, information storage,
biomedicine, and electronics.
9–11,13–15
Nanomaterials and their
designs should be promising materials for constructing photo-
electrodes because of their extremely small feature size and large
surface area. It is believed that nano-structures of photo-
materials can provide beneficial effects such as band structural
modification, quantum dot sensitization, plasmonic associa-
tion, as well as the domination of crystal facets.
Although several reviews for photocatalytic splitting
of water based on electrolyte solutions have recently been
published,
1,6,9–12,16
a comprehensive review with focuses on
photoelectrochemical cells using solid photoelectrodes for
water splitting and their associated photoelectrode materials
and corresponding photoelectrode design strategies is definitely
required to facilitate research in this area.
In this tutorial review, we discuss the development of
semiconductor materials and their effect on activity, strategies
for improving the photoelectrochemical performance from the
viewpoint of absorbance or carrier transportation, and the
potential utilization and limitation of structure design. We
provide several examples of practical improvement in photo-
activity to highlight their concepts and reproducibility, and to
emphasize how advanced nanotechnology has improved the
development of photoelectrodes for solar water splitting.
1.2 Photoelectrodes and evaluation of photoelectrochemical
activity
Under standard conditions, the change in free energy (DG)
associated with the conversion of one molecule of H
2
O into H
2
Published on 04 July 2012. Downloaded by National Research Council Canada on 26/04/2016 20:15:27.
View Article Online
This journal is
c
The Royal Society of Chemistry 2012 Chem. Soc. Rev., 2012, 41, 5654–5671 5657
and 1/2 O
2
is 237.2 kJ mol
1
, which corresponds to an
electrolysis cell voltage (DE
o
) of 1.23 V per electron trans-
ferred. The photoelectrochemical method for water splitting,
as displayed in Fig. 1, involves two electrodes the anode and
the cathode. The anode is the photoactive material (or semi-
conducting material)-based electrode, which is irradiated by a
light beam during water splitting. The cathode is the counter-
electrode also called a photocathode although it is not
irradiated by light. In order to drive this reaction, the photoactive
material on the photoanode must absorb radiant light to make
its electrode potential higher than 1.23 V, so the water molecule
can be oxidized to form O
2
and protons (H
+
) and protons can be
simultaneously reduced to form H
2
at the cathode. As a result of
the sluggish kinetics for water splitting, energy is lost during the
transfer of electrons at the photoanode/electrolyte interface.
Accordingly, the energy that is required for photoelectrolysis is
frequently given as 1.7–2.4 eV. On the photoanode, this photo-
induced process must generate four electron–hole pairs for
generating O
2
. If the photoanode is irradiated by light that has
energy greater than the bandgap (E
BG
) of the photoactive
material, then the electrons of the valence band will be excited
into the conduction band while the holes remain in the valence
band. These photo-generated electrons will then pass through the
external wire and reach the surface of the cathode to react with
protons, forming H
2
, and the holes at the photoanode will
oxidize the H
2
O(O
2
) to produce O
2
.
In addition to the bandgap requirement, the band levels,
another factor that commonly dominates the ability for water
splitting, has to be considered. The splitting of water requires
that the bottom level of the conduction band must be located
at a more negative potential than the reduction potential of
H
+
/H
2
, and that the top of the valence band is more positive
than the oxidation potential of H
2
O/O
2
. With respect to the
fundamental requirements of the conduction and valence
bands, the band edge positions must straddle the reduction
and oxidation potentials of water. Some semiconductor
materials are able to reduce, but not oxidize water. On the
other hand, some metal-oxide materials such as Fe
2
O
3
are able
to oxidize, but not reduce the proton. For such materials, an
external bias is essential for reduction of the proton and to
assist photocurrent in the photoelectrochemical cell. In this
case, the onset potential of the photoresponse from the IV
curve always shifts to a higher potential region.
17
Based on
electrochemical principles, the water splitting reaction can
only be driven when the irradiation energy exceeds 1.23 eV
(approximately wavelength of 1000 nm), indicating that the
energy of the light must be larger than the bandgap to separate
the electrons and holes. In practical operation, the minimum
thermodynamic energy requirements plus the overpotential
loss require at least 1.7–1.9 eV for the photoelectrochemical
splitting of water, which corresponds to absorption at an onset
wavelength of 730 nm.
18,19
Furthermore, the intensity of the
solar spectrum dramatically falls off below 350 nm, so the
upper limit on the bandgap is about 3.5 eV. Towards this end,
the optimum value of the semiconductor bandgap should
ideally be between 1.9 and 3.5 eV, which is within the visible
range of the solar spectrum. In practical cases, the flux of solar
photons in the wavelength range from 680 nm to 280 nm
(1.8 eV–4.4 eV) represents 27.5% of the total solar photon
flux, which is the maximum efficiency that has been predicted
in various investigations based on the bandgap of the semi-
conductor material, the solar spectrum, and various losses.
20
However, this efficiency of energy conversion using a single
bandgap material is too low to satisfy the requirements for an
actual application, even if a perfect photocatalytic material
can be developed. As a result, a more efficient configuration
has to be developed using a system with multiple bandgaps
(see Section 3). Most requirements of semiconductor materials
for water splitting in a photoelectrochemical cell are similar to
those of other photocatalysts, but various other factors must be
considered here. For example, the modification of co-catalyst
that commonly provides active sites on the surface of photo-
active materials can effectively improve the kinetics of the carrier
transportation in the photochemical cell. Regarding this, RuO
2
and Pt nanoparticles have been identified to be effective oxygen
evolution and hydrogen evolution co-catalysts, respectively.
Since the transfer of carriers across the n-type semiconductor/
electrolyte interface must be sufficiently rapid to compete with
the accumulation of electrons/holes at the surface, thereby this
modification would maintain the bending of band and lead to a
concomitant decrease in electron–hole recombination.
Another critical factor is resistance to photocorrosion or
electrochemical stability, which may limit the usefulness of
several photocatalytic materials. Most metal oxide semi-
conductor materials are thermodynamically unstable, meaning
that the photogenerated holes may oxidize themselves rather
than water (photocorrosion and/or anodic photodecomposition).
These undesired photodecompositions in various photoactive
materials commonly depend on the pH value of the electrolyte
and frequently limit their utilization under certain conditions.
TiO
2
and SnO
2
are highly stable over a wide range of
pH values in aqueous environments upon illumination, while
the stability of hematite strongly depends on the presence of
dopants, pH values, and oxygen stoichiometry. Many non-
oxide semiconductor materials may either dissolve or form a
thin oxide film upon their surface, preventing the electron
transfer through the interface between the semiconductor/
electrolyte interface. Photocorrosion and/or anodic decompo-
sition is expected to be markedly inhibited if the transfer
of carriers for water oxidation through the interface is
faster than any competing reaction. Therefore, the develop-
ment of semiconductor materials with excellent stability against
Fig. 1 Diagram of the basic principles of water splitting for a
photoelectrochemical cell with an n-type semiconductor photoanode
where oxygen is evolved and a photocathode (Pt sheet) where hydrogen
is evolved.
Published on 04 July 2012. Downloaded by National Research Council Canada on 26/04/2016 20:15:27.
View Article Online
5658 Chem. Soc. Rev., 2012, 41, 5654–5671 This journal is
c
The Royal Society of Chemistry 2012
photocorrosion/anodic decomposition becomes a critical issue
for future applications.
Some important factors other than bandgap energy must
be considered for improving the water splitting reaction,
including the charge separation, charge mobility and the life-
time of photogenerated electron–hole pairs. These factors also
critically affect the photoactivity of semiconductor materials.
Structural and electronic properties of co-catalysts on the
surface of the photoelectrodes also strongly affect both the
generation and the separation of electron–hole pairs. In
the general case, highly crystalline materials with a low density
of defects are beneficial for the water splitting reaction,
because defects serve as recombination centers for the photo-
generated electrons/holes.
In terms of evaluating photo-active materials, the incident-
photon-to-current-conversion efficiency (IPCE) of materials
can be obtained experimentally, and is used to specify the
photoresponse as a function of the wavelength of incident
light. Eqn (1) shows yields of the IPCE at the desired bias
voltage.
IPCE ð%Þ¼
1240J
l I
100% ð1Þ
where J represents the photocurrent density (mA cm
2
); l is
the wavelength of incident light (nm), and I is the intensity of
the incident light (mW cm
2
). However, the most important
factor is the solar energy conversion efficiency. The overall
efficiency of solar energy conversion can be determined from
the following expression:
21
Zð%Þ¼J
E
o
rex
jE
means
E
aoc
j
I
100% ð2Þ
where Z is the solar energy conversion efficiency, J is the
photocurrent density (mA cm
2
), I is the incident light
intensity (mW cm
2
), E
0
rex
is the potential required for water
splitting, E
means
is the electrode potential of the working
electrode under illumination, and E
aoc
is the electrode
potential of the working electrode at open circuit conditions
under the same illumination. The term E
0
rex
is usually taken to
be 1.23 V for the water splitting reaction. It should be noted
that the redox potential for water splitting is strongly depen-
dent on the pH value of the electrolyte solution, because the
band level of semiconductor material can shift with a change
in pH value. In general, a higher pH value condition of
photocurrent measurement can lead to a shift in voltage
toward a lower region.
22
In the case of practical measurement,
the band edge position for many kinds of semiconductor
materials cannot simultaneously straddle the reduction and
oxidation potentials of water. Therefore, sacrificial reagents
which act as electron donors or hole scavengers such as
alcohol, sulfide ions, and silver ions are often employed to
evaluate the photocatalytic activity for water splitting.
It should be noted that the above formula cannot correctly
evaluate energy conversion efficiency when sacrificial reagents
are utilized to estimate the photoactivity, because photo-
generated electrons/holes would react with electron donors
or hole scavengers rather than water. Consequently, the
evaluation of photoactivity in a photoelectrochemical cell
commonly uses photocurrent instead of conversion efficiency
in many cases.
2. Photoelectrode materials
Since photoelectrode materials are crucial to the water splitting
process, chemists and other scientists are always seeking the
best photo-active materials, which exhibit a smaller semi-
conductor bandgap, and so cover a large fraction of the
sunlight spectrum, a conduction/valence band energy that
straddles the water oxidization and reduction potentials,
stability in an aqueous environment, and a high conversion
efficiency for splitting of water, as well as low cost. Unfortu-
nately, no single material has yet been found that satisfies all
of these requirements, although combinatorial methods have
been utilized to quickly search for and optimize materials.
4,17
Transition metal oxides, including TiO
2
and ZnO, exhibit
good stability but only absorb a small fraction of incident
sunlight because they have a large bandgap. However, the
combination of either a photoanode and a photocathode or a
photoelectrode and a photovoltaic device has also be investi-
gated with a view to constructing a tandem cell that facilitates
overall water splitting associated with multi-photon absorp-
tion events.
23
A solar energy conversion efficiency over 12%
has been achieved using IIIV semiconductor materials, but
their cost and stability remain major disadvantages.
24
Even
though the efficiency of either photocatalysts or a photo-
electrochemical cell for solar energy conversion is still low,
research on water splitting is always an important topic in
solving problems associated with energy and the environment.
A wide range of semiconductor materials have been developed
for usage in photoelectrodes for the water splitting reaction.
Table 1 presents several representative photo-active materials.
Since the medium for photoelectrochemical measurement has a
significant effect on the stability of materials in terms of photo-
corrosion and photodecomposition,
25
Table 1 also lists the
corresponding media in which the measurements were made.
For example, the photoelectrode that was made of TiO
2
can be
operated under neutral or basic conditions while ZnO can be
measured under neutral conditions. The properties of the
medium strongly affect the performance of the photoelectrode.
For instance, the pH value of the electrolyte clearly affects both
the redox potential of the water and the position of the
conduction band/valence band of the semiconductor materials.
TiO
2
was the first reported photoelectrode for water
splitting under UV irradiation.
7
It was adopted to produce
hydrogen and/or oxygen from aqueous solution. Since that
first report, various studies have focused on reducing the
probability of recombination, increasing the surface area,
and enhancing the carrier mobility, to improve the photo-
conversion efficiency of the photoelectrochemical cell.
26,27
In
this regard, catalyst nanoarchitecture has been regarded as an
important approach to deliver these demands. A mesoporous
structure has been adopted to reduce the probability of charge-
carrier recombination because such a structure is composed of
smaller particle size. One-dimensional (1D) nanostructures
(such as nanowires and nanotubes) can provide the advantages
of a high surface area and rapid diffusion in a single direc-
tion, yielding a low recombination of electron–hole pairs.
Published on 04 July 2012. Downloaded by National Research Council Canada on 26/04/2016 20:15:27.
View Article Online
This journal is
c
The Royal Society of Chemistry 2012 Chem. Soc. Rev., 2012, 41, 5654–5671 5659
Table 1 Semiconductor materials with nanostructures as photoelectrodes for water splitting
Materials Structure Bandgap Electrolyte Conditions Performance Ref.
TiO
2
(nanotube) 1 M KOH I = 100 mW cm
2
, l = 320–400 nm,
50 W metal hydride lamp
j =13mAcm
2
at 0 V vs. Ag/AgCl, Z = 6.8% 21
TiO
2
(nanotube) 1 M KOH I =95mWcm
2
, l = 320–400 nm,
UV illumination
j =26mAcm
2
at 0 V vs. Ag/AgCl Z = 16.25% 28
TiO
2
(nanorod) Anatase 3.27 eV 0.5 M NaClO
4
, pH = 7.0 I = 100 mW cm
2
, l = 350–500 nm
1000 W Xe lamp
j = 0.018 mA cm
2
at 1.0 V vs. Ag/AgCl, Z = 0.1% 26
TiO
2
(mesoporous) Anatase 3.34 eV 0.1 M NaOH I = 1000 mW cm
2
, l = 350–500 nm,
150 W Xe lamp (AM 1.5)
j = 3.5 mA cm
2
at 0 V vs. Ag/AgCl, Z = 0.36% 29
TiO
2
(nanowire) Anatase/Rutile 3.00 eV 1.0 M NaOH, pH = 13.8 I = 100 mW cm
2
, l = 350–700 nm,
AM 1.5 G simulated lamp
j = 2.6 mA cm
2
at 0.216 V vs. Ag/AgCl, Z = 1.05% 27
TiO
2
(branched) Rutile 2.95 eV 1.0 M KOH, pH = 14 I = 100 mW cm
2
, l = 300–700 nm,
Xe lamp AAA solar simulated
j = 0.85 mA cm
2
at 0.65 V vs. Pt, Z = 0.49% 48
ZnO (nanocoral) Wurtzite 0.5 M Na
2
SO
4
, pH = 6.8 I = 125 mW cm
2
, l = 350–750 nm,
Tungsten-halogen lamp
j = 0.25 mA cm
2
at 1.2 V vs. RHE 31
ZnO (thin film) Wurtzite 3.3 eV 0.5 M NaClO
4
, pH = 7.4 I = 230 mW cm
2
, l = 350–800 nm,
1000W Xe lamp
j = 0.142 mA cm
2
at 1.0 V vs. Ag/AgCl 30
ZnO (branch) Wurtzite 3.54 eV 0.5 M NaSO
4
, pH = 7.0 I = 100 mW cm
2
, 350–800 nm,
AM 1.5 G
j = 0.12 mA cm
2
at 0.31 V vs. Ag/AgCl 32
WO
3
(mesoporous) Monoclinic 2.5 eV 3 M H
2
SO
4
I = 100 mW cm
2
, l = 300–700 nm
AM 1.5 G simulated lamp
j = 2.9 mA cm
2
at 1.23 V vs. RHE 36
WO
3
(nanowire) Orthorhombic 2.9 eV 0.5 M H
2
SO
4
I = 100 mW cm
2
, l = 300–700 nm,
AM 1.5 G Xe lamp
j = 1.2 mA cm
2
at 0.82 V vs. SCE, Z = 0.33% 35
WO
3
(mesoporous) Monoclinic 2.66 eV 1 M H
2
SO
4
I = 370 mW cm
2
, l = 300–700 nm,
150 W Xe lamp (AM 1.5)
j = 9 mA cm
2
at 1.8 V vs. SCE 41
WO
3
(mesoporous) Monoclinic 2.58 eV 1 M H
2
SO
4
I = 100 mW cm
2
, l = 300–700 nm,
150 W Xe lamp (AM 1.5)
j = 3.7 mA cm
2
at 1.3 V vs. Pt 42
WO
3
(flaskers) Monoclinic 2.51 eV 0.1 M Na
2
SO
4
I = 100 mW cm
2
, l = 400–700 nm,
AM 1.5
j = 1.43 mA cm
2
at 1.3 V vs. Pt 43
BiVO
4
(nanowire) Monoclinic 0.5 M Na
2
SO
4
I = 100 mW cm
2
, l = 400–700 nm,
AM 1.5
j = 0.4 mA cm
2
at 1.0 V vs. Pt 40
BiVO
4
(nanostructure) Monoclinic 2.45 eV 0.5 M Na
2
SO
4
I = 100 mW cm
2
, l = 400–700 nm,
100 W Xe lamp (AM 1.5)
j = 0.91 mA cm
2
at 1.0 V vs. Ag/AgCl 33
BiVO
4
(scheelite) Monoclinic 2.5 eV 0.1 M K
2
SO
4
I = 100 mW cm
2
, l = 420–700 nm,
300 W Xe lamp (AM 1.5)
j = 1.2 mA cm
2
at 1.5 V vs. Ag/AgCl, Z = 0.0014% 34
Fe
2
O
3
(mesoporous) Hematite 2 eV 1 M NaOH I = 130 mW cm
2
, l = 420–700 nm,
AM 1.5 simulated lamp
j = 1.07 mA cm
2
at 1.2 V vs. RHE 46
Fe
2
O
3
(nanotube) Hematite 1.97 eV 1M KOH I =87mWcm
2
, l = 400–700 nm,
300 W solar simulator (AM 1.5)
j = 1.41 mA cm
2
at 0.4 V vs. Ag/AgCl, Z = 0.84% 44
Fe
2
O
3
(thin film) Hematite 2.06 eV 1 M NaOH, pH=13.6 I = 100 mW cm
2
, l = 300–600 nm,
450 W Xe lamp (AM 1.5)
j = 0.55 mA cm
2
at 1.43 V vs. RHE 38
Fe
2
O
3
(mesoporous) Hematite 2.15 eV 1M NaOH, pH = 13.6 I = 100 mW cm
2
, l = 300–600 nm,
450 W Xe lamp (AM 1.5)
j = 0.56 mA cm
2
at 1.23 V vs. RHE 37
Fe
2
O
3
(porous) Hematite 2.1 eV 1 M NaOH, pH = 13.6 I = 100 mW cm
2
, l = 350–650 nm,
450 W Xe lamp (AM 1.5)
j = 2.34 mA cm
2
at 1.43 V vs. RHE 45
Fe
2
O
3
(nanorod) Hematite 2.1 eV 1 M NaOH I =87mWcm
2
, 450 W Xe lamp
(AM 1.5)
j = 0.52 mA cm
2
at 1.23 V vs. RHE 39
Published on 04 July 2012. Downloaded by National Research Council Canada on 26/04/2016 20:15:27.
View Article Online
5660 Chem. Soc. Rev., 2012, 41, 5654–5671 This journal is
c
The Royal Society of Chemistry 2012
Mor et al.
21
formed TiO
2
nanotube arrays by anodizing a Ti
sheet and measured an efficiency of 4% under UV irradiation
(l = 320–400 nm). Subsequently, Paulose et al.
28
prepared
TiO
2
nanotubes with a length of 45 mm by an anodizing
method, achieving a photocurrent of 26 mA cm
2
and a
conversion efficiency of 16.25% under UV irradiation (l =
320–400 nm). Hartmann et al.
29
reported that mesoporous
TiO
2
films prepared by the sol–gel method were approximately
ten times more efficient for the water splitting reaction than
their counterparts obtained from crystalline TiO
2
nanoparticles
owing to their thick and continuous pore walls and lower
recombination rate. ZnO may become an alternative to TiO
2
for fabricating photoelectrodes, because of its various advantages,
including higher carrier mobility, greater chemical and thermal
stability, and lack of toxicity. Furthermore, the conduction band
of ZnO is at a more negative position than that of TiO
2
, such that
the photogenerated electrons are better able to reduce protons,
resulting in greater solar energy conversion efficiency. In addition,
it has been demonstrated that the morphology of ZnO can be
easily controlled by adjusting the synthetic conditions resulting in
various variously shaped ZnO nanostructures.
30
Ahn et al.
31
produced a ZnO nanocoral and measured a photocurrent of
0.25 mA cm
2
under a tungsten-halogen lamp (l = 350–750 nm).
Qiu et al.
32
prepared ZnO nanotetrapods using a multi-step
growth process, and measured a photocurrent of 0.12 mA cm
2
and a conversion efficiency of 0.045% under irradiation from a
solar simulator (l = 350–800 nm). Although both TiO
2
and ZnO
have a high performance in the UV region, their wide bandgap
nature cannot capture solar light in the visible region, in which
photons of sunlight are most abundant, so they result in a low
energy utilization. Therefore, materials with a narrower bandgap
are required.
With respect to narrow bandgap materials, BiVO
4
,
33,34
WO
3
,
35,36
and Fe
2
O
3
37–39
are very attractive, although they
do present some challenges. For example, BiVO
4
thin films
comprised of ordered arrays of pyramidal-shaped nanowires
have been successfully fabricated by seed-mediated growth in an
aqueous BiVO
4
suspension, and a photocurrent of 0.4 mA cm
2
was measured under AM 1.5 irradiation (l = 400–700 nm;
100 mW cm
2
).
40
Tungsten trioxide (WO
3
) and hematite
(a-Fe
2
O
3
) are in effect semiconductor oxides which serve as
oxygen-evolving anodes for absorbing photons from the blue
regions of the solar spectrum, because they have band gaps
of 2.2 and 2.6 eV, respectively. Cristino et al.
41
prepared
mesoporous WO
3
films using an anodization method, and
measured a photocurrent of 9 mA cm
2
under AM 1.5
irradiation (l = 300–700 nm; 370 mW cm
2
). Kim et al.
42
prepared a WO
3
photoelectrode with a mesoporous nano-
structure using polyethyleneglycol as a surfactant, and their
photoelectrochemical properties revealed a photocurrent of
3.7 mA cm
2
under AM 1.5 irradiation (l = 300–700 nm).
Su et al.
43
successfully prepared vertically aligned WO
3
nano-
flakes using a solvothermal technique, and nanowires or
nanoflakes could be fabricated by adjusting the composition
of the solution. A photocurrent of 1.43 mA cm
2
and an IPCE
(incident-photon-to-current-conversion efficiency) as high as
60% at l = 400 nm were measured.
a-Fe
2
O
3
, which has a small bandgap, low cost, and high
chemical stability, is one of the most promising metal oxide
semiconductor materials for the splitting of water using solar
energy. Mohapatra et al.
44
reported an anodization method
for growing an ultrathin wall of Fe
2
O
3
nanotubes with a
thickness of 5–7 nm and a length of 3–4 mm on Fe foil. A
photocurrent density of 1.41 mA cm
2
was obtained using the
hematite nanotube arrays, and a maximum solar-to-hydrogen
conversion efficiency of 0.84% was reached under AM 1.5
conditions. Later, Brillet et al.
45
elucidated a solution-based
strategy and fabricated a porous electrode by encapsulation
with an SiO
2
confinement scaffold, yielding a high water
oxidation photocurrent of 2.34 mA cm
2
under AM 1.5
illumination. This high photocurrent density was attributed
to the activation of the nanostructured hematite photoanode
with control over wide range of particle sizes in the porous
film. Generally, hematite has a relatively poor photoelectro-
chemical performance, which is attributable to its short charge
carrier diffusion length as well as the slow kinetics for the
oxidation of water by the valence band holes.
46
In summary, a photoelectrode that is made of single-component
photo-active materials cannot deliver the required performance in
terms of either photocurrent or conversion efficiency to meet
the demands of daily life. There are several limitations on their
performance which include low absorbance in the visible
region, poor charge-carrier transportation, poor collection of
photogenerated electrons, and limited chemical stability in an
electrolyte under illumination. Accordingly, the following
section will review various strategies for improving photo-
electrochemical performance. Extraneous materials can be
adopted to increase the absorbance of solar light and/or to
promote the collection and transportation of charge carriers.
3. Material designs and nanostructural architecture
to improve photoelectrochemical activity
As discussed in the previous section, many semiconductor
materials can be used as photoelectrode materials for splitting
water under UV or visible light irradiation.
6,9
To design ideal
photo-active materials for the efficient conversion of solar
energy, increasing the light absorbance and improving the
charge-carrier transportation are two of the main approaches
in the recent development of solar water splitting. In this regard,
nano-materials and their designs are particularly promising in
achieving these goals. Nano-scale materials have unique proper-
ties that differ from those of corresponding bulk materials, and
therefore, it is likely that nanostructural photoelectrode
materials will attract considerable attention in the future
development of the solar splitting of water. Scheme 1 shows
various strategies for meeting the demands of the solar splitting of
water using nanotechnology. A selection of these strategies will be
described in detail below, and related concepts will be discussed.
3.1 Single bandgap photo-active materials
Increasing the absorbance of sunlight should be the primary
means of maximizing the spectral range of sunlight. This
increase can be achieved by using a semiconductor with
branched structure, which can absorb many photons because
of its large absorption cross-section and its intrinsic properties.
Modification of the band structure, such as by doping, has
been found in many photo-active materials to increase the
Published on 04 July 2012. Downloaded by National Research Council Canada on 26/04/2016 20:15:27.
View Article Online
This journal is
c
The Royal Society of Chemistry 2012 Chem. Soc. Rev., 2012, 41, 5654–5671 5661
absorbance of the solar spectrum, in which a larger part of
solar light can be absorbed owing to shifting of the absorption
edge toward longer wavelengths. The band structural modifi-
cation primarily involves introducing foreign ions (cations or
anions) into the semiconducting material to change its electronic
and optical properties, and the corresponding performance
strongly depends on the modification of the band structure level.
3.1.1 Branched structure. Controlling the morphology of
nanomaterials is a common strategy for tailoring their crystal
facets and optimizing their performance in the case of various
semiconductor photomaterials, because heterogeneous reactivity
depends strongly on the surface atomic configuration and bonding
environment that can be altered by controlling crystalline facets.
Recently , Pan et al.
47
explored a set of TiO
2
anatase crystals with
predominant {001}, {101}, or {010} facets. The {001} facet
exhibited a lower reactivity than the {101} facet in photooxidation
for generating OH radicals and evolving hydrogen. However, this
observation does not support the conventional understanding that
facets with a higher percentage of non-coordinated atoms are
normally more reactive in heterogen eous reactions. The order of
photoreactivity of the facets of TiO
2
with the anatase structure in
the photoo xidation and reduction reactions for generating OH
radicals and evolving hydrogen is {010} > {101} > {001}. Early
studies tended to focus on semiconductor nanoparticle films owing
to their large surface area to volume ratios. However, nanoparticle
films suffer from a higher loss due to electron–hole recombination
since the electron mobility in a nanoparticle film is around two
orders of magnitude lower than that of a bulk single crystal
because of electron trapping/scattering at grain boundaries.
48
Defects and grain boundaries frequently become trapping and
recombination centers, which increase the probability of electron–
hole recombination. Hence, one-dimensional (1D) well-defined
nanostructures, such as nanorods/wires and nanotubes, have
attracted substantial attention since one-dimensional nano-
structures exhibit excellent properties in terms of charge separa-
tion, charge transport, and light absorption. Nanorods and
nanotubes have a lower defect density than nanoparticles, which
can facilitate the charge transfer process. Some one-dimensional
nanostructures are polycrystalline because they have a high
growth rate, and post-annealing treatment may be able to
convert such polycrystalline structures into crystalline ones.
Branched nanostructural semiconductor materials have
been developed to simultaneously provide both a rapid charge
transfer pathway for carrier collection and a large surface
area, to improve photocatalytic performance. Recently, Cho
et al.
48
described a hierarchically branched TiO
2
nanorod
structure for use in photoelectrochemical devices, in which
the material simultaneously offered a large contact area with
the electrolyte, excellent light-trapping characteristics, as well
as a highly conductive pathway for the collection of charge
carriers. Fig. 2a presents the structure of a branched nanorod,
and Fig. 2b displays a typical scanning electron microscopic
(SEM) image of a branched-nanorod sample, whose short
needle-shaped branches were grown uniformly over the whole
surface of the nanorods. Fig. 2c presents a high-resolution
transmission electron microscopic (HRTEM) image of the