Small Science - 2022 - Li

REVIEW
www.small-science-journal.com
Recent Progress on Semiconductor Heterojunction-Based

Photoanodes for Photoelectrochemical Water Splitting
Shengnan Li, Weiwei Xu, Linxing Meng, Wei Tian,* and Liang Li*
as aviation fuel was put forward, and it

For photoelectrochemical (PEC) water splitting, the utilization of semiconductor was first verified in practice in 1960.
heterojunctions as building blocks for photoanodes allows for high light Subsequently, the first hydrogen fuel cell
absorption, effective charge separation, and superior redox capability, greatly vehicle was released in 1972. Since the
1970s, in the face of environmental pollu-
boosting the solar energy conversion efficiency. This review mainly focuses on
tion and other crises, hydrogen energy
the construction of heterojunction photoanodes, improvement strategies of research has been conducted extensively
carrier transmission, and their application in PEC water splitting. First, a series of all around the world. There are currently
carrier dynamics characterization methods are introduced to reveal the principle a variety of hydrogen production technol-
and significance of promoting carrier transport in heterojunctions. Then, from ogies, dominated by the following three
the perspective of the mechanism of promoting the separation and transport of technical routes. The first is that fossil
energy, including coal and natural gas,
charge carriers, several strategies are summarized and analyzed, including the
is reformed to produce hydrogen. The sec-
micro/nanostructure, energy band structure, photothermal effect, piezoelectric ond is that hydrogen is produced from
effect, pyroelectric effect, ferroelectric effect, and intermediate layer. Finally, the industrial byproduct gases such as those
challenges and opportunities for heterojunction photoanodes in PEC reactions from coke ovens, the chlor-alkali process,
are explained clearly, which points the way forward for the development of and propane dehydrogenation. The third
heterojunctions. technology is the production of hydrogen
by water electrolysis.[6] To achieve a low-
cost, pollution-free, and simple hydrogen
production method, solar energy photoly-
1. Introduction sis of water for hydrogen production was developed.[7,8]
However, the photocatalytic hydrogen production rate is
Currently, the social demands for energy are increasing.[1] extremely low due to the low light conversion efficiency.[9,10]
Traditional fossil energy sources such as coal and oil are no lon- Therefore, a compromise method is adopted between electro-
ger sufficient to support current production and living needs.[2] catalysis and photocatalysis. Through the combination of solar
At the same time, the massive consumption of fossil fuels also energy and electric energy, the cost can be reduced, and the
causes considerable harm to the environment. Therefore, it is hydrogen production rate can be increased, making photoelec-
urgent to seek renewable clean energy to relieve the status quo trochemical (PEC) water splitting favored by researchers.[11–13]
of the energy crisis and environmental pollution.[3] Hydrogen After generations of researchers sparing no exploration efforts,
energy is considered one of the most promising green energy the field of PEC water splitting has harvested fruitful results.[14]
sources in the 21st century.[4] Humans became interested in In 1972, Japanese scientists Fujishima and Honda first
the application of hydrogen energy 200 years ago. The first observed the phenomenon of photocatalytic water splitting on
single-cylinder hydrogen internal combustion engine was TiO2 photoanodes under ultraviolet irradiation.[15] A typical PEC
invented in 1807.[5] In 1938, the hypothesis of using hydrogen water splitting system consists of a photoanode, a photocathode,
and an electrolyte. The photoanode is usually based on an n-type
S. Li, W. Xu, L. Meng, W. Tian, L. Li semiconductor material with a light response. When excited by
School of Physical Science and Technology light, it can generate electron–hole pairs. In the case of an exter-
Jiangsu Key Laboratory of Thin Films nal circuit, the electrons generated in the photoanode flow to the
Center for Energy Conversion Materials & Physics (CECMP) photocathode, and hydrogen ions in the electrolyte accept elec-
Soochow University
Suzhou 215006, P. R. China trons from the photocathode to produce hydrogen.[16] The spe-
E-mail: wtian@suda.edu.cn; lli@suda.edu.cn cific reaction process is as follows[17]
The ORCID identification number(s) for the author(s) of this article Photocathode∶ 2Hþ þ 2e ! H2 (1)
can be found under https://doi.org/10.1002/smsc.202100112.
© 2022 The Authors. Small Science published by Wiley-VCH GmbH. This Photoanode∶ 2H2 O þ 4hþ ! O2 þ 4Hþ (2)
is an open access article under the terms of the Creative Commons
Attribution License, which permits use, distribution and reproduction The total reaction equation of water splitting is
in any medium, provided the original work is properly cited.
DOI: 10.1002/smsc.202100112 2H2 O ! O2 þ 2H2 (3)
Small Sci. 2022, 2, 2100112 2100112 (1 of 23) © 2022 The Authors. Small Science published by Wiley-VCH GmbH
26884046, 2022, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/smsc.202100112 by INASP/HINARI - PAKISTAN, Wiley Online Library on [23/09/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.advancedsciencenews.com www.small-science-journal.com
The generation of PEC water splitting is a three-step proce- of the heterojunctions, which will facilitate the photogenerated
dure: 1) Light absorption, in which the semiconductor is excited electrons and holes to reach their own reaction sites to take part
by light irradiation that is larger than the semiconductor bandgap in the reaction. Therefore, the review will have certain signifi-
generating electron and hole pairs. 2) The separation and recom- cance as a reference for future heterojunction modification work.
bination of the photogenerated electron and hole pairs in the
semiconductor. 3) The occurrence of the surface redox reac-
tions.[18,19] Light absorption can be adjusted by regulating the
2. Definition of a Heterojunction
morphology, changing the sample thickness, and modulating
the bandgap.[20–23] Regarding the separation and transport of A heterojunction is the interfacial area at the contact point of two
electrons and holes, there are currently a series of solutions, such semiconductors. An ideal PEC photoelectrode should utilize as
as element doping,[24,25] introducing defects,[26,27] and construct- much of the sunlight as possible and have a strong redox poten-
ing heterojunctions.[28–30] The oxygen evolution reaction (OER) tial, but the relationship between these two features is contradic-
on the photoanode surface involves a complex four-electron tory. A single semiconductor cannot meet these demands. A
four-proton transfer process, and its kinetic process is very heterojunction is composed of different semiconductors.[40–42]
slow, which can generally be greatly alleviated by loading a A wide bandgap semiconductor is selected to combine with a
cocatalyst.[31–33] Among the above three influencing factors, small bandgap semiconductor to absorb a wider range of light
the competition of separation and recombination between photo- while it is expected to maintain a strong redox potential.[43,44]
generated electrons and holes is crucial for the PEC process. Heterojunctions more easily meet the requirements of wide light
Appropriate amount of element doping and defects mainly absorption and strong redox ability than a single semiconductor.
change the conductivity and conductivity type of the sample to
Therefore, the construction of heterojunctions is quite extensive
promote charge transfer.[34,35] According to the position where
in the PEC field.[45] According to the energy band alignment of
the introduced impurities and defects generate energy levels,
the heterojunctions formed by the two semiconductors, they can
they can be divided into shallow level and deep level. Shallow
be classified into straddling-gap heterojunctions (type-I),
energy level means that a donor energy level closes to the con-
staggered-gap heterojunctions (type-II), and broken-gap hetero-
duction band or an acceptor energy level closes to the valence
junctions (type-III).[46,47] In addition, there is a Z-scheme heter-
band (VB) is generated in the semiconductor, and basically all
ojunction, and the charge transfer pathway exhibits a
ionization is achieved at room temperature, thereby improving
Z-shape.[48,49] In light of the conductivity type of the semiconduc-
the conductivity of the semiconductor and changing the conduc-
tor, heterojunctions can be divided into n–n junctions, p–p junc-
tivity type.[36] The deep energy level refers to the donor energy
tions, and p–n junctions.[50]
level far away from the conduction band and the acceptor energy
Generally, the construction of heterojunctions should meet
level far away from the VB introduced in the semiconductor,
the following conditions: similar crystal structures, lattice spac-
which requires a large ionization energy, so there is no shallow
energy-level impurity that affects the carrier concentration and ings, and thermal expansion coefficients. Constructing hetero-
conductivity type. The deep energy-level defects can act as effec- junctions has been widely used as an effective means to
tive recombination centers, reducing the carrier lifetime, known promote carrier separation and transport. Heterojunctions have
as a nonradiative recombination center, affecting the luminous been widely developed to improve the PEC performance in pho-
efficiency, and increasing the resistivity of the material as a com- toelectric catalysis. In addition, the specific surface area of photo-
pensating impurity. Therefore, the type of defect states intro- anodes may also be increased by constructing heterojunctions,
duced is crucial for the performance of PEC water which could enlarge the contact region between the photoanode
splitting.[37,38] The construction of heterojunctions meets the and the electrolyte, strengthen light harvesting by compounding
needs of ideal PEC photoanodes with a wide light absorption with a narrow gap semiconductor, and achieve performance
range, effective charge separation, and strong redox ability. improvements.[51] However, constructing a heterojunction is
Different semiconductor materials are combined to boost the not a perfect solution that is profitable and harmless. There
effective separation and transport of electrons and holes assisted are still unavoidable problems in the process of constructing het-
by the built-in electric field at the interface. This review article erojunctions, which are not conducive to carrier transmission
takes several strategies by adjusting the band structure to pro- and separation, thereby affecting the PEC performance.[52]
mote carrier transport mainly from the point of view of construct- Due to the lattice mismatch of the two semiconductors forming
ing heterojunctions.[39] heterojunctions, the bonding at the interface may not be tight
This review first introduces the definition of heterojunctions enough or there may be a large number of defects, which is unfa-
and their advantages in acting as photoanodes for PEC water vorable for carrier transmission and separation at the interfaces
splitting. Second, several strategies for characterizing charge car- of heterojunctions.[53] At the same time, it is essential to take the
rier dynamics are presented. Then, we specifically introduce energy band positions of the two semiconductors into account
micro–nanostructures, type-II heterojunctions, Z-scheme heter- while constructing heterojunctions. The matched energy band
ojunctions, gradient doping, various physical effects (photother- structure forms a cliff-like junction at the interface, which can
mal, piezoelectric, pyroelectric, and ferroelectric effects), and the bring high photocurrent and OER performance. However, with
introduction of intermediate layers to enhance carrier transfer in a poorly matched band structure, the interface barrier will form a
heterojunctions. By applying the described carrier transport spike-like junction, which is not conducive to charge transfer.
dynamics strategies to the heterojunctions, the carrier transport These problems need to be avoided and solved one by one to
path will be further improved based on the intrinsic advantages obtain a relatively perfect heterojunction. Therefore, it is
necessary to provide relevant strategies for the carrier transmis- is basically in nanoseconds, and the pulse frequency is not high
sion problem of the heterojunction. (104–105 Hz). Therefore, the data acquisition time is long, and
the light intensity may drift during the measurement process,
which affects the efficiency and accuracy of the measurement.
3. Carrier Dynamics Characterization Methods However, pulsed lasers can give picosecond pulses, and the pulse
frequency can be very high, but its price is relatively expensive. In
There are three basic steps in a photocatalytic reaction: the semi-
terms of detectors, photomultiplier tubes (PMTs) and micro-
conductor catalyst absorbs light larger than its bandgap produc-
channel plate detectors (MCPs) are generally used. The latter
ing photogenerated carriers, the carriers migrate to the catalyst
has a faster response time and less interference. Theoretically,
surface, and the photogenerated carriers contact the reactants
the instrument using pulsed laser and MCPs can detect the fluo-
performing the catalytic conversion. The efficiency of converting
rescence lifetime of 10-20 ps. Furthermore, the main features of
light energy to chemical energy is equal to the product of the
TCSPC test are as follows: 1) Intuitive view of fluorescence decay
efficiencies of the following three processes: light capture, photo-
curve. 2) High sensitivity, and low fluorescence intensity can be
generated charge separation and migration, and the high-
solved by prolonging the acquisition time. 3) The error of the
efficiency surface catalytic reaction. How to achieve efficient syn-
random distribution of photons obeys the Poisson distribution,
ergy of these three processes is a very critical scientific problem.
which is helpful for the data processing and judgment. 4) Direct
First, light capture determines the theoretical efficiency of the
and convenient for the recording of TRPL spectrum.
maximum energy conversion of the catalyst. Second, the separa-
5) Deconvolution calculation is often required due to low mea-
tion of photogenerated charges is a very complex process of mul-
sured fluorescence lifetime. This requires that the excitation sig-
tiple time scales and multiple space scales. The recombination
nal of the instrument itself must be accurately measured,
process of the bulk, surface, and interface photogenerated car-
otherwise the calculated results will deviate greatly from the
riers is inevitable mainly because the reactions involve excited
actual. 6) Complicated analysis of fluorescence lifetime from a
states that occur in the femtosecond to millisecond time range.
mathematical point of view. 7) Data acquisition is time-
Therefore, it is necessary to study the dynamics of photogener-
consuming.
ated carriers.[54] In general, the characterization methods of the
carrier transport dynamics are mainly based on the degree of
recombination between photogenerated carrier, the transmission 3.2. Transient Absorption Spectrum
rate or the transmission efficiency and other parameters.
Transient absorption analysis refers to the material being excited
3.1. Time-Resolved Photoluminescence Spectroscopy from the ground state to the excited state by a beam of pumped
light with high energy, and then the relaxation process of the
Time-resolved fluorescence analysis is known as transient fluo- excited state particles is detected by applying a beam of probe
rescence analysis. When the semiconductor material absorbs light with low energy, which can be used to evaluate carrier
enough photons, the VB electron will be excited to a higher state, dynamics.[57] In the ultrafast transient absorption analysis spec-
and the corresponding hole will remain in the VB, but the excited trum, the abscissa is the relative value of the probe light, and the
state electron will also move back to a lower energy level, and hole ordinate ΔA represents the difference between the absorption
recombination and radiation emit fluorescence.[55] In the short spectra measured by the probe light irradiated on the sample
time after the material is excited instantaneously, the fluores- before and after the addition of pumped light.[58] Regarding
cence intensity reaches its maximum value and then decays expo- the delay time of the pumped light, the resolution can reach pico-
nentially. The time in which the fluorescence intensity decays to seconds and even femtoseconds. The relaxation time is obtained
1/e of the initial time is defined as the fluorescence lifetime. by fitting the curve. The longer the relaxation time is, the longer
Therefore, the time-resolved fluorescence spectrum can be is the electron lifetime and the better is the separation effect of
obtained through testing, and the fluorescence lifetime can be photogenerated electrons and holes.
analyzed and fitted, thereby assisting the analysis of the lifetime The detection light of nanosecond transient absorption spec-
of the excited electrons.[56] trum (TAS) usually adopts a noncollimated “xenon lamp,” and
At present, the pulse method for time-resolved photolumines- the efficiency of detection light is low. It is necessary to increase
cence spectroscopy (TRPL) is the most commonly used tech- the pump and xenon lamp power or use the μs pulse xenon lamp
nique, and time-correlated single photon counting (TCSPC) is to achieve instantaneous high light intensity in order to obtain
currently the main application technology of TRPL measure- the TA signal. Furthermore, high-energy pumping and high-
ment. The principle is that the probability of detecting an emitted power detection also lead to solid samples that are easily ablated.
photon at a certain time t is proportional to the fluorescence In addition, the time resolution is mainly determined by the
intensity at that time point. Time resolution is an important probe light and pump light pulse times. Currently, the pulse time
parameter of fluorescence lifetime measuring instrument, which of high-power lasers is usually 5–10 ns, which limits the time
determines the shortest fluorescence lifetime that can be mea- resolution of TAS to 7–10 ns. However, pulsed white light laser
sured by the instrument, which is jointly determined by the exci- as the detection light can greatly improve the detection efficiency.
tation light source and the detector. Moreover, the pulsed white light laser is a real collimated laser,
In the TCSPC, the choice of the excitation light source is very and the detection spot is also smaller. For solid thin film samples,
important, either various gas flash lamps or pulsed lasers can be the scanning and moving sample holder can be used to achieve
used. The cost of the flash lamp is relatively low, the pulse given spatial resolution. In short, as the charge transfer or energy
transfer in semiconductors is generally on the order of femtosec- The TPC response measures the electronic transmission and
ond and picosecond, TAS is one of the best methods to study the collection in the semiconductors. The appearance of a positive
carrier relaxation process of semiconductors in real time. spike under light illumination indicates that electrons accumu-
late on the surface of the semiconductor and that the photocur-
rent decays to a stable value, which means that the electrons
3.3. Transient Photovoltage/Transient Photocurrent
accumulated on the surface are transferred to the electrolyte.[63]
A rapid decay certificates that there is a serious recombination of
Before the laser pulse with nanosecond width (5 ns) is emitted,
electrons and holes, and a high steady-state photocurrent indi-
the laser host transmits the pulse signal to the oscilloscope
cates excellent separation efficiency.[64] If a negative spike
through the trigger signal line, and the oscilloscope is ready
appears after the light illumination is removed, it suggests that
to measure and record the data after receiving the pulse signal:
there is back electron and hole recombination.[65]
when the laser is irradiated on the photoanode absorbing laser
TPC/TPV can be used to study the kinetic information of pho-
photons, the free charges will rapidly drift and diffuse to the elec-
togenerated charge separation, including drift and diffusion pro-
trodes on both sides after a series of photoelectric conversion pro-
cesses during photogenerated charge separation. Drift processes
cesses. At this time, if the photoanode and the external circuit
generally refer to rapid separation processes that occur within
form a loop, there will be a current response, and the magnitude
particles while diffusion processes generally refer to charge
of the current response is related to the directionally moving
transport between particles over a long period of time. In addi-
charges in the photoanode, as well as the impedance of the loop.
tion, by analyzing the shape, time scale, symbol, and intensity of
The input impedance of the oscilloscope has two types: 50 ohm
the TPV spectrum, a series of information such as the speed,
and 1 M ohm. When the input impedance of the oscilloscope is
separation direction, separation mode, and separation degree
adjusted to 1 M ohm, the photoanode is equivalent to being con-
of photogenerated charge separation and recombination can
nected in series with a large resistor, which is equivalent to an
be obtained. TPC/TPV are effective means to study the behavior
open circuit. This measurement method is generally called tran-
of photogenerated charges on the nanosecond time scale. They
sient photovoltage (TPV) measurement. When the input imped-
have the advantages of noncontact and lossless, and can directly
ance of the oscilloscope is adjusted to 50 ohm, the photoanode is
reflect the kinetic information such as the separation direction,
equivalent to being connected in series with a small resistor,
separation efficiency, and charge lifetime of photogenerated
which is similar to a short circuit. This measurement method
charges. For photocatalytic reactions, the separation and trans-
is generally called transient photocurrent (TPC) measurement.
port process of photogenerated electron–hole pairs can be ana-
The TPV spectrum provides kinetic information on the photo-
lyzed by TPC/TPV, which can explore the mechanism of
generated charge separation of different samples. The photovolt-
charge recombination and provide guidance for improving the
age response includes two components: rising and decay. The
high efficiency of charge separation.
rising part in TPV corresponds to the accumulation of electron
concentration of the conductive substrate on the electrode
(resembling capacitor charging). TPV decay refers to the recom- 3.4. Intensity Modulated Photocurrent/Photovoltage
bination process of electrons leaving the conductive substrate Spectroscopy
(resembling capacitor discharge).
The working principle in the TPV is the Dember effect.[59] The intensity modulated photocurrent/photovoltage spectros-
When the semiconductor is exposed to light due to the diffusion copy (IMPS/IMVS) measurement system mainly consists of
length difference between photogenerated electrons and holes, a three parts: 1) Light source system—xenon lamps, UV lamps,
Dember photovoltage will be generated. At the same time, the light-emitting diodes (LBD), etc. are commonly used as light
light intensity in contact between the semiconductor surface sources (dual light sources can be used, that is, one light source
and the body gradually weakens, so the diffusion rate of the pho- generates steady-state incident light, and the other light source
togenerated carriers, which changes from a high concentration to generates modulated light. A DC voltage can also be used to
a low concentration, is also different. For this reason, the photo- superimpose a modulated voltage to drive a light source at the
generated electrons and holes partially drift to realize the estab- same time, and the superposition of steady-state light and mod-
lishment of a built-in electric field. The stronger the light ulated light can also be realized). 2) Light source drive system—
irradiates the semiconductor, the more photogenerated electrons the main equipment is a frequency response analyzer and a
are generated. When the stronger the light irradiates the semi- potentiostat, which generates a voltage or current signal to drive
conductor, more photogenerated electron–hole separation will be the light source to generate a modulated light signal. 3) Test and
produced, and the higher the photovoltage will be.[60] Similarly, analysis system—the frequency response analyzer is used to
when the light intensity remains the same, the high photovoltage compare and analyze the relationship between the amplitude
represents the good separation capability of photogenerated elec- and phase of the input signal and the output signal, and finally
trons and holes. The decay process occurs when the light is the test results are output by the microcomputer.
removed, which is devoted to characterizing the recombination IMPS and IMVS are frequency domain detection technolo-
of the photogenerated electrons and holes in the heterojunc- gies. The input signal is generally composed of a beam of stable
tions.[61] A long electron recombination lifetime indicates that background light and a beam of sinusoidally modulated light that
the recombination rate is low. In contrast, a short electron recom- has a small disturbance.[66] The output signal is the correspond-
bination lifetime suggests that the recombination rate of the pho- ing steady-state photocurrent with modulation. The IMPS and
togenerated electrons and holes is high.[62] IMVS tests are similar to impedance tests.[67] During impedance
testing, a constant voltage or current signal is applied to the test 4. Effective Strategies to Promote Carrier
system, and an alternate current (AC) signal with a certain ampli- Transport in Photoelectrodes
tude is superimposed at the same time to control the frequency
change. The frequency of the measured AC signal is consistent To promote charge carrier transfer in heterojunctions, several
with the frequency of the applied AC signal, but the phase angle effective strategies have been adopted, including constructing
is offset, and the impedance value is related to the frequency that micro–nanostructures, modulating the energy band structure,
can be measured. In contrast, the light beam intensity shining on utilizing pyroelectric, piezoelectric, pyroelectric, and ferroelectric
the photoelectrode changes rather than the amplitude of the volt- effects, and introducing an intermediate layer.
age or current signal. The first quadrant of the IMPS represents
the hole transfer and recombination process. The second, third, 4.1. Micro–Nanostructures
and fourth quadrants represent the diffusion of electrons and the
response process of the resistance–capacitance circuit.[68] The Heterojunctions with different micro–nanostructures exhibit dif-
electron transmission time (τd) and the electron diffusion coeffi- ferent PEC performance characteristics. The following aspects
cient (Dn) can be calculated.[69,70] IMVS can be qualitatively ana- determine the PEC performance of photoanodes: light absorp-
lyzed: 1) Direct transition recombination between the conduction tion, bulk carrier separation, and surface charge carrier injec-
band (CB) electrons and VB holes inside the semiconductor.[71] tion.[29,72] First, the most intuitive function of constructing a
2) Recombination through the surface states on the semiconduc- micro–nanostructure is to alter the specific surface area of the
tor surface, yielding the electron lifetime (τn) and electron diffu- material, as shown in Figure 1a,b. The larger the specific surface
sion length (Ln) that can be quantitatively calculated. Therefore, area is, the larger is the contact region between the photoanode
IMPS/IMVS can be used to characterize τn, Ln, and recombina- and electrolyte, which would result in better PEC perfor-
tion degree in a heterojunction to obtain the transport properties mance.[73] Yang and co-workers[74] grew nanoporous BiVO4
of photogenerated electrons and holes in a heterojunction. coated onto SnO2 nanorods, which increased the contact area
IMPS/IMVS is a measurement method in the frequency between BiVO4/SnO2 and electrolyte and was beneficial to the
domain, which studies the system under test in a wide modula- improvement of the performance of the photolysis of water
tion frequency range and can distinguish dynamic processes and the rate of hydrogen production. Second, the construction
with different speeds. At the same time, electrons and holes of micro–nanostructures improves the light capture capabilities
appear in different phases in opposite directions, and the of photoanodes.[73] Compared with conventional bulk planar
dynamic process of electrons and holes can be distinguished. structures, nanorods, nanosheets, nanobowl arrays, and other
Therefore, IMPS/IMVS can obtain more dynamic microscopic structured photoanodes exhibit better light refraction capabili-
process information than other conventional measurement ties, which lengthens light traveling, extends the interaction
methods. between the light and the photoanode, and improves light
Figure 1. Depiction of the incident light processes, carrier transfer, and interfacial injection process of a) planar-structured and b) nanorod arrays.
c) Schematic for charge transport and light scattering in the TiO2 NR@NB array electrode. Reproduced with permission.[80] Copyright 2016, Wiley-
VCH. d) Schematic for spatial charge transfer pathway. Reproduced with permission.[81] Copyright 2019, Wiley-VCH.
harvesting efficiency. Tian et al.[75] grew In2S3 nanosheets on is insufficient driving force to separate the photogenerated elec-
WO3 nanowalls to enhance the light refraction and contact area trons and holes.[78] The micro–nanostructure better meets this
with electrolytes compared to WO3 nanowalls, which caused requirement than the bulk structure, which allows for effective
WO3/In2S3 heterojunctions to possess PEC potential. Third, utilization of photogenerated charge carriers. In addition, the
when the diffusion length is larger than the size of the material, ordered micro–nanostructure provides an ordered channel for
holes can smoothly arrive at the photoanode/electrolyte interface carrier transmission and inhibits bulk recombination.[79] As
and thus be involved in the oxidation reaction before hole recom- shown in Figure 1c, the construction of a TiO2 nanorod/nano-
bination, which is conducive to PEC water splitting. The diffu- bowl array by Wang and co-workers[80] not only enhanced light
sion length is used to characterize the hole transport absorption but also promoted the transport and separation of the
performance in the photoanode. For micro–nanostructured photogenerated electrons along the rutile nanorods to the anatase
materials, the distance from photogenerated holes to the surface nanobowls. Dong et al.[81] constructed a 1D Ta3N5 nanorod/
of the semiconductor/electrolyte is much less than that of bulk BaTaO2N (BTON) nanoparticle heterojunction, as shown in
materials, which is beneficial to increase the probability of Figure 1d. The nanorod morphology promoted the directional
photogenerated minority carrier reaching the interface, inhibit- transport of carriers, and 0D BTON nanoparticles had a short
ing carrier recombination and improving PEC performance. pathway for photogenerated holes to reach the surface, which
Considering that BiVO4 has a short diffusion length of was beneficial to the oxidation reaction of water splitting.
70 nm, Pan et al.[76] constructed a SnO2/TiO2/BiVO4 layered Table 1 presents a summary of the performance of different
nanosheet@hollow microsphere array, and the thickness of elec- micro–nanostructured heterojunctions.
trodeposited BiVO4 was 20 nm. Because the thickness of BiVO4 The micro–nanostructures with different shapes increase the
is smaller than its hole diffusion length, the bulk separation of specific surface area, which are advantageous in light capturing
the SnO2/TiO2/BiVO4 photoanode was greatly improved, and and maximizing contacting area with the electrolyte. In addition,
this design helped the photogenerated holes reach the photoa- the ordered array structure provides a fast transport channel for
node/electrolyte interface. Fourth, the width of the depletion photogenerated carriers. As micro–nanoengineering contributes
layer has something to do with the characteristics of the material to the light absorption, carrier transport, and separation capabili-
itself, along with the temperature and bias voltage. Therefore, a ties of photoanodes, it is desirable to design and construct photo-
change in the morphology will change the width of the depletion anodes with special shapes for water splitting.
layer. Compared with bulk materials, the proportion of depletion
layers in nanomaterials is significantly increased, thereby pro-
moting effective separation of the carriers. Cho et al.[77] con- 4.2. Band Structure
structed CdS/CdO core–shell nanorods with different shell
thicknesses and found that the shell thickness had a certain rela- Constructing a heterojunction is a universal and effective means
tionship with the width of the space charge region. Once the to promote carrier separation. At present, type-II and Z-scheme
thickness of the CdO shell was shorter than the width of the heterojunctions have attracted the attention of researchers due to
depletion layer, 0.125 and 0.25 mM CdS/CdO photoanodes their advantages, and many studies have been carried out on
had a built-in field within the thickness of the CdO shell. them.[82–84] However, it is not easy to select two types of semi-
Shell thickness values of the 0.5 and 1 mM CdS/CdO photoan- conductors that are suitable for energy bands, so the search for a
odes exceeded the width of the depletion layer, and there was a method to regulate and control the energy bands brooks no delay.
serious recombination phenomenon in the bulk. Fifth, when the Doping can adjust the energy band up or down by introducing
thickness of the micro–nanostructure is within the range of the donor impurities or acceptor impurities, and gradient doping is
space charge layer, the photogenerated carriers can be effectively equivalent to constructing a continuous gradient energy band
separated. If light absorption occurs in the deep bulk phase, there structure, which effectively solves the energy band problem.
Table 1. Summary of PEC performance for heterojunctions with micro–nanostructures.
Photoanodes Micro–nanostructures J (1.23 VRHE) Vonset Testing condition H2 evolution rate Ref.
2
NiFe/BiVO4/SnO2 Nanoporous/Nanorod 5.61 mA cm / 1.0 M K-Bi buffer / [74]
WO3/In2S3-P Nanowall/Nanosheet 1.61 mA cm2 0.02 VRHE 0.1 M Na2SO4 / [75]
SnO2/TiO2/BiVO4 Nanosheet@Microsphere 5.03 mA cm2 / 0.5 M Na2SO4/0.1 M Na2SO3 / [76]
3.1 mA cm2 0.5 M Na2SO4
CdS/CdO Core-shell nanorods 4.35 mA cm2 (0 VSCE) -1.40 VSCE 0.35 M Na2S/0.25 M Na2SO3 / [77]
R-TiO2/A-TiO2 Nanorod@Nanobowl 1.24 mA cm2 / 1 M NaOH / [80]
Ta3N5/BaTaO2N Nanorod/Nanoparticle / / / 27.3 μmol h1 [81]
In2S3/F-Fe2O3 Nanoparticle/Nanorod 2.21 mA cm2 / 1 M KOH / [95]
2 1 1
FH-TiO2 Nanosheet/Nanorod 2.24 mA cm / 6 M KOH/10% v/v methanol 1.441 mmol g h [96]
0.566 mmol g1 h1
NFO/TiO2/ZFO/SnO2 Helix@Dendrite 1.0 mA cm2 0.76 VRHE 1 M NaOH / [168]
4.2.1. Type-II Heterojunctions As most heterojunctions cannot be formed at the same time,
the preparation methods are mostly in a stepwise manner, that is,
On the basis of band alignment of the two semiconductors, they a semiconductor is first grown on a conductive substrate, and
can be classified into type-I, type-II, and type-III heterojunc- then another semiconductor is prepared.[92–95] For example,
tions.[85] The structures of the three types of heterojunctions Gao et al.[96] grew TiO2 nanorods on fluorine-doped tin oxide
are shown in Figure 2a–c. In the type-I heterojunction, both elec- (FTO) via a hydrothermal method, and the obtained samples
trons and holes are transferred from semiconductor A to semi- were subjected to secondary hydrothermal treatment and then
conductor B. In this case, the photogenerated electron–hole pairs grew TiO2 ultrathin nanosheets in situ on TiO2 nanorods. The
accumulate in semiconductor B and result in serious carrier construction of this facet heterostructure promoted the efficiency
recombination. In the type-II heterojunction, the photogenerated of charge separation, prolonged the life of carriers, and was con-
electrons in semiconductor B are transported to semiconductor ducive to improving the performance of hydrogen production. Ye
A, while the photogenerated holes on semiconductor A are and co-workers[93] designed a B-C3N4/Mo-BiVO4 photoanode by
transferred to semiconductor B, realizing spatial separation three-step processes, which increased the separation capability
between photogenerated electron–hole pairs. In the type-III of photogenerated carriers at the interface of B-C3N4 and
heterojunction, there is no directional transport of photogener- Mo-BiVO4.
ated carriers between semiconductor A and semiconductor B. The stepwise preparation of heterojunctions not only compli-
In addition, a narrow bandgap semiconductor is compounded cates the preparation process but is also likely to destroy the first
on a wide bandgap semiconductor to construct a heterojunction, prepared semiconductor during the preparation of the second
which reduces the bandgap of photoelectrodes, enhances the semiconductor. Therefore, researchers are devoted to developing
absorption of sunlight, and facilitates the extraction of photogen- a one-step process for fabricating heterojunctions.[97,98] For
erated carriers. The advantage of efficient photogenerated example, Cao et al.[99] grew a 3D porous pyramidal In2O3/
electron–hole pairs separation makes constructing heterojunc- In2S3 heterojunction array by an ion exchange–induced synthesis
tions an effective way to improve carrier dynamics. In the past strategy, obtaining the best performance at 1.23 V versus
few decades, many researchers have discovered that a large num- reversible hydrogen electrode (RHE) consisting of all In2S3
ber of type-II heterojunctions can effectively separate electrons photoanodes and long water splitting stability. Hou and co-
and holes and improve the performance of PEC water workers[100] synthesized a mesoporous In2O3–x/In2S3 2D lateral
splitting, such as WO3/BiVO4,[86] WO3/CdIn2S4,[87] ZnO/ heterostructure by in situ oxidation, as shown in Figure 2d, and
ZnS,[88] etc.[89–91] the final performance was 21 times that of the In2S3 atomic
Figure 2. Illustration of the band alignment a) type-I, b) type-II, and c) type-III. d) Schematic rendition of the In2O3–x/In2S3 heterostructure charge transfer
tunneling mechanism. Reproduced with permission.[100] Copyright 2018, Wiley-VCH.
layers. Dong et al.[81] used a one-step ammonia thermal route to choice for promoting photogenerated carrier separation. As
prepare Ta3N5/BaTaO2N heterostructures and modulated the name implies, the Z-scheme heterojunction refers to the
ammonia thermal parameters for better PEC performance. charge transfer pathway in a “Z"-shaped structure under illumi-
Meng et al.[101] built a CdIn2S4/In2S3 bulk heterojunction inside nation. Z-scheme heterojunctions can be divided into three
CdIn2S4 nanosheets by Ni-phthalocyanine solution directly types: ionic Z-scheme heterojunctions,[102] all-solid-state Z-
mixed with the CdIn2S4 precursor solution and achieved a sepa- scheme heterojunctions,[103] and direct Z-scheme heterojunc-
ration efficiency of up to 90% at 1.23 V versus RHE. tions.[104] The development history and specific carrier transport
The type-II heterojunction promotes the separation and trans- mechanisms of the three Z-scheme heterojunctions are shown in
port of carriers owing to the matched band structure and the Figure 3a–d.
built-in electric field, but the type-II heterojunction achieves this The ionic Z-scheme heterojunction first appeared in 1979. In
goal by sacrificing its redox capability. In addition, due to the exis- Figure 3b, this heterojunction requires a redox electron mediator
tence of electrostatic interactions, photogenerated electron holes for charge transport, which speeds up the recombination rate of
in the original photocatalyst will inhibit the interface transfer of photogenerated electrons in the CB of semiconductor A and pho-
photogenerated electrons and holes in other catalysts. Therefore, togenerated holes in the VB of semiconductor B, promoting the
it is necessary to explore new types of heterojunctions to solve efficient separation of photogenerated carriers. The commonly
these drawbacks. used redox electron mediators are [Co(bpy)3]3þ/2þ, Fe3þ/Fe2þ,
IO3/I, NO3/NO2, and [Co(phen)3]3þ/2þ. In light of the lim-
4.2.2. Z-Scheme Heterojunction itation of the applicable environment of the ionic state Z-scheme
heterojunction, the all-solid-state Z-scheme heterojunction has
To effectively separate electrons and holes, a type-II heterojunc- gradually replaced it. The all-solid-state Z-scheme heterojunction
tion sacrifices the potential, which will adversely affect PEC water is composed of a solid electronic conductor and two semiconduc-
splitting. Z-scheme heterojunctions with different carrier trans- tors, as shown in Figure 3c. The electronic conductor forms an
fer pathways than type-II heterojunctions have become a new ohmic contact with the two semiconductors, which promotes the
Figure 3. a) The development history of the Z-scheme heterojunction. Illustrations of the band alignment, b) traditional Z-scheme, c) all-solid-state
Z-scheme, and d) direct Z-scheme. e) Transfer pathway of the g-C3N4/rGO/PDIP Z-scheme heterojunction. Reproduced with permission.[105]
Copyright 2021, Wiley-VCH. f ) Depiction of the band alignment and electron transport process in the BP/RP direct Z-scheme heterojunction.
Reproduced with permission.[108] Copyright 2019, Wiley-VCH.
recombination of photogenerated electrons in the CB of semi- to characterize the carrier transport of the photoanode. The carrier
conductor A and photogenerated holes in the VB of semiconduc- lifetime of ZnO/TiO2 was longer than that of TiO2, indicating that
tor B. The commonly used electronic conductors are Au, Ag, Pt, the construction of heterojunctions helped electron transport
Cu, Cd, W, and rGO. As shown in Figure 3e, Chen et al.[105] suc- between the bulk phase and the interface. Xu et al.[109] compared
cessfully constructed a graphitic carbon nitride/rGO/perylene the TRPL of ZnIn2S4, ZnIn2S4/CdS, and ZnIn2S4/CdS/ZnO sam-
diimide polymer (g-C3N4/rGO/PDIP) Z-scheme heterojunction ples to prove that the successful construction of the Z-scheme het-
through a simple multistep method, which achieved high- erojunction promoted carrier separation and that the ZnO layer
throughput charge transfer and effective photocatalytic water inhibited surface recombination and improved OER dynamics.
splitting. Through the TAS spectrum of g-C3N4/rGO/PDIP, The current direct Z-scheme heterojunction usually has two
the average lifespan of the electrons trapped by PDIP was short- serious problems. 1) Due to the lack of mediators and the exis-
ened, and the average lifetime of holes captured by PDIP was tence of weak van der Waals interlayer interactions, the recom-
prolonged, which proved that charge transfer occurred in a bination of photogenerated electrons in the CB of semiconductor
timely and spatial manner in g-C3N4/rGO/PDIP and successfully A and photogenerated holes in the VB of semiconductor B is very
constructed a Z-scheme heterojunction. Fu et al.[106] built an α- low. For example, Wang et al.[110] used two different ultrathin
Fe2O3/Au/TiO2 ternary photoanode. By coating Au nanoparticles polymers, aza-conjugated microporous polymers (CMP) and
on the Fe2O3 surface, the photogenerated holes could be effi- C2N nanosheets, to form a van der Waals Z-scheme heterostruc-
ciently extracted. The presence of amorphous TiO2 on the surface ture. As the two polymers in the Z-scheme heterojunction were
accelerated the surface water oxidation reaction. Unfortunately, combined through van der Waals forces, the carrier recombina-
these mediators also generate reverse reactions and light- tion at the interface increased, and the conversion efficiency of
shielding effects, which bring about a greatly reduced number solar energy to hydrogen (STH) was 0.23%. Once RGO was intro-
of electron and hole pairs under light excitation. Furthermore, duced as a solid electronic mediator into the mental-free Z-
it is a rigorous problem for Z-scheme heterostructures with elec- scheme heterostructure, the STH value was further increased
tron mediators to maintain long-term stability. to 0.40%. 2) Cocatalysts and sacrificial agents are used to improve
Inspired by nature, the direct Z-scheme heterojunction mimics the surface catalytic ability beneficial to water redox reactions,
photosynthesis and is used to improve light conversion efficiency. specifically for the slow four-electron involved OER. Wang and
As shown in Figure 3d, when two different semiconductors are in co-workers[110] adopted Pt and Co(OH)2 cocatalysts loaded onto
contact, the free photogenerated electrons in the CB of semicon- the C2N releasing H2 and aza-CMP nanosheets releasing O2,
ductor A migrate to the CB of semiconductor B until the Fermi respectively, to further improve the photocatalytic performance.
level reaches a balance. Accordingly, positive charges accumulate Regardless of whether theoretically or experimentally, the
in semiconductor A, and negative charges gather in semiconduc- straightforward characterization of direct Z-scheme heterojunc-
tor B, forming a built-in electric field. In the meantime, band edge tions is still very challenging. Low et al.[111] reported that in situ
bending occurs at the interface between semiconductor A and irradiated X-ray photoelectron spectroscopy (ISI-XPS) characteri-
semiconductor B. Under light irradiation, the built-in electric field zation, hydroxyl radical generation test, and density functional
accelerates the recombination of the free electrons in the CB of theory (DFT) simulation had been applied to confirm the forma-
semiconductor A and free holes in the VB of semiconductor B. tion of a direct Z-scheme heterojunction obtained by compound-
In the Z-scheme heterojunction, an oxidation reaction of water ing TiO2 with CdS. In ISI-XPS, the Cd characteristic peaks
splitting eventuates on the VB of semiconductor A, and the other underwent a negative shift, and the Ti characteristic peaks gen-
half reaction of water splitting occurs on the CB of semiconductor erated a positive shift, indicating that electrons were transferred
B, thereby realizing the efficient detachment of photogenerated from TiO2 to CdS via a Z-scheme channel. In the hydroxyl radical
carriers. That is, the VB of semiconductor A with a relatively test, as TiO2 was combined with OH/H2O to generate hydroxyl
strong redox ability and the CB of semiconductor B with a com- groups but not CdS, hydroxyl groups were generated at the TiO2.
paratively strong redox ability are retained in the Z-scheme heter- Therefore, the detection of whether TiO2/CdS generated
ojunction. As there is no mediator, the charges are directly hydroxyl groups was used as a criterion for judging Z-scheme
transmitted through the interface, which shortens the transmis- heterojunctions. Wang et al.[112] used electron spin resonance
sion distance, benefits the occurrence of carrier recombination, (ESR) and surface photovoltaic spectroscopy (SPV) to verify
and improves the photocatalytic potential of the photoelectrode. the Z-scheme carrier transfer mechanism of Sv-ZnIn2S4/
In addition, the direct Z-scheme heterojunction avoids the occur- MoSe2 for photocatalytic hydrogen production. The SPV signal
rence of reverse reactions and light shielding phenomena, of Sv-ZnIn2S4/MoSe2 was lower than that of Sv-ZnIn2S4, indicat-
increases the density of effective photogenerated carriers, and ing that the holes reaching the surface of Sv-ZnIn2S4/MoSe2
enhances its photocorrosion resistance. The direct Z-scheme black were reduced. Given that the CB of MoSe2 did not reduce O2
phosphorus/monolayer Bi2WO6 photoanode prepared by Hu and to ·O2, the presence of ·O2 signal in Sv-ZnIn2S4/MoSe2 was
co-workers[107] showed strong photocatalytic performance in H2 detected by ESR, which indicated that Sv-ZnIn2S4 accumulated
production from water splitting and NO removal for air purifica- a large number of photogenerated electrons, proving the forma-
tion. In Figure 3f, Liu et al.[108] designed and fabricated a black tion of a Z-scheme heterojunction.
phosphorus/red phosphorus (BP/RP) heterophase junction, and Unlike the type-II heterojunction that obtains effective bulk
the appropriate band structure with staggered alignment could separation at the expense of redox capability, the Z-scheme het-
help to possess the notable separation and transfer capacity of pho- erojunction ensures efficient separation of photogenerated elec-
togenerated electrons and holes. The evidence was provided by trons and holes while retaining a strong redox potential, which
TAS via a direct Z-scheme pathway. Zhou et al.[61] used TPV decay demonstrates the coexistence of high-efficiency bulk phase
separation and strong redox capability. Therefore, the direct doping. Uniform doping needs to consider the amount of dop-
Z-scheme heterojunction has advantages in the field of PEC ing. Otherwise, excessive uniform doping can easily form a
water splitting. However, reasonably designing and directly char- recombination center, making the width of the depletion layer
acterizing the premium direct Z-scheme heterostructure are still smaller and affecting the bulk separation.[114,115] Gradient dop-
challenging. ing not only has the advantages of uniform doping but can also
form a gradient energy band structure, which can expand the
width of the depletion layer and promote the bulk separation
4.2.3. Gradient Doping of carriers in the semiconductor. Hufnagel et al.[116] proved,
with the help of IMPS, that Sn doping only occurred on the
The introduction of alien elements into the semiconductor is
hematite surface (surf-Sn) and that the top 5 nm hematite layer
called doping. According to the semiconductor conductivity for-
(top-Sn) showed higher charge transfer (ηct) efficiency than Sn
mula
doping that was evenly distributed throughout the whole hema-
σ ¼ ne μe þ nh μh (4) tite layer (bulk-Sn) and the sample without Sn doping
(undoped). Moreover, top-Sn introduced a gradient band that
n is the carrier concentration and μ is the carrier mobility. facilitated the efficiency of charge transfer and separation,
If the semiconductor is doped with alien elements, the enhancing PEC performance. Meng et al.[117] doped SnS2 nano-
carrier concentration n increases, thereby increasing the sheets with In and Zn/In to form a gradient energy band, which
conductivity of the semiconductor.[113] In addition, the orbital increased the carrier concentration and improved the carrier
hybridization of the introduced ions and semiconductors can separation efficiency.
change its electronic structure, CB, and VB. According to the The types of doping elements can be divided into metal doping
doping positions, it can be divided into interstitial or substitu- and nonmetal doping.[118–120] The introduction of suitable metal
tional doping. Substitutional doping needs to meet two condi- ions into the semiconductor can generate a brand-new energy
tions. The radius of the doped element and the element to be level in the forbidden band to expand the response range of visi-
replaced are not very different, and the valence electron shell ble light and facilitate charge carrier separation. In Figure 4a–c,
structures are similar. In line with the distribution of the dop- Xiao and co-workers[121] showed that gradient Mg doping with
ing amount, it can be divided into uniform doping and gradient the substitution of Ta5þ by Mg2þ in Ta3N5, which formed a
Figure 4. a) Schematic for pristine Ta3N5, Mg(L):Ta3N5, and Mg(H):Ta3N5. Schematics for band bending of b) homogenous and c) gradient Mg:Ta3N5.
Reproduced with permission.[121] Copyright 2020, Springer Nature. d) Schematic of pristine CdS and Grad-O CdS sample charge transfer mechanisms.
Reproduced with permission.[122] Copyright 2018, Elsevier. e) Diagrams of the preparation process and band bending of a grad-P: Fe2O3/Co-Pi photo-
anode. Reproduced with permission.[123] Copyright 2017, Royal Society of Chemistry. f ) Schematic transfer tunneling and band alignment of the Pd
nanoparticles/black TiO2/b-Si photocathode for water reduction. Reproduced under the terms of the CC-BY 4.0 license.[127] Copyright 2018, The
Authors, published by Springer Nature.
gradient energy band and suppressed the carrier recombination and cause upward bending of the energy band in a wider area
caused by the defect states, thereby improving the charge sepa- (Figure 4e), thereby promoting charge separation along the radial
ration efficiency, prolonged lifetimes as evidenced by TRPL spec- direction of the hematite nanobeam.
troscopy, when compared to homogeneous Mg doping. Zhang Similar to doping, introducing vacancies can destroy the peri-
et al.[35] reported that linear electron energy loss spectroscopy odic arrangement of crystals and produce lattice distortions, which
(EELS) scanning implied that the Ta doped in hematite lattices cause defect states to be generated in the bandgap, expand the
must be in the form of substitution, and gradient Ta doping of range of light response, and promote the separation of photogen-
the core–shell structure of hematite homojunction nanorods erated charge carriers.[124–126] As depicted in Figure 4f, Zheng
enhanced the photogenerated current density and reduced the et al.[127] used a simple method to construct a crystalline TiO2
turn-on voltage. Nonmetal doping does not form acceptor or protective layer with gradient oxygen defects on a silicon-based
donor energy levels in the forbidden band, and it is not easy photocathode. Conductive atomic force microscopy (C-AFM) veri-
to form recombination centers. Instead, it narrows the forbidden fied that the concentration of oxygen defects determined the chan-
band by shifting the top of the VB of semiconductor photocata- nels by which the current passed through the TiO2 layer and
lysts. Therefore, it has a wider range of solar spectral response. affected the distribution and number of transport channels.
Our group[75] produced S vacancies and gradient oxygen doping Gradient oxygen defects were the major factors promoting and
in In2S3 by PVP treatment, which not only increased the charge enhancing the PEC performance of the photovoltaic device.
carrier concentration and reactive sites but also formed a gradi- Gradient doping can not only improve the conductivity and
ent band structure that promoted the transfer and separation of charge carrier concentration but also avoid the serious recombi-
carriers. Yu et al.[122] doped 3D CdS dendritic nanorods with oxy- nation problems caused by uniform heavy doping through gra-
gen and distributed them along the radial gradient of the nano- dient modulation of the doping concentration. In addition, a
rods so that a continuous band-bending structure was formed in gradient energy band structure is constructed to widen the deple-
the nanorods, which promoted the migration of photogenerated tion layer and promote effective directional transport and sepa-
electrons from the surface to the core and Ti substrate, as illus- ration of carriers.
trated in Figure 4d. Luo et al.[123] constructed gradient P doping Table 2 shows the performance comparison of the above three
to increase the conductivity of the Fe2O3 nanoarray photoanode tuned heterojunction energy bands.
Table 2. Summary of PEC performance for heterojunctions with different energy band structures.
Sample Energy band structure J (1.23 VRHE) Vonset Testing condition H2 evolution rate Ref.
FH-TiO2 Type-II 2.24 mA cm2 / 6 M KOH/10% v/v methanol 1.441 mmol g1 h1 [96]
1 1
0.566 mmol g h
NiFeOx/B-C3N4/Mo-BiVO4 Type-II 5.93 0.3 mA cm2 0.34 VRHE PPB (pH ¼ 7) 77.5 μM h1 [93]
In2O3/In2S3 Type-II 8.2 mA cm2 0.02 VRHE 0.25 M Na2S/0.35 M Na2SO3 / [99]
In2O3–x/In2S3 Type-II 1.28 mA cm2 0.75 VRHE 1.0 M KOH / [100]
Ta3N5/BaTaO2N Type-II / / / 27.3 μmol h1 [81]
2
CdIn2S4/In2S3 Type-II 2.98 mA cm 0.02 VRHE 0.5 M Na2SO4 / [101]
g-C3N4/rGO/PDIP Z-scheme / / / 15.80 μmol h1 [105]
α-Fe2O3/Au/TiO2 Z-scheme 1.05 mA cm2 / 1 M NaOH 18.67 μmol cm2 h1 [106]
1
BP/MBWO Z-scheme / / / 21 042 μmol g [107]
BP/RP Z-scheme / / / 0.33 mmol g1 h1 [108]
BN ZnO/TiO2 Z-scheme 2.75 mA cm2 / 0.2 M Na2SO4 45.6 μmol cm2 h1 [61]
ZnIn2S4/CdS/ZnO Z-scheme 3.48 mA cm2 0.03 VRHE 0.5 M Na2SO4 / [109]
aza-CMP/C2N Z-scheme / / / 5.0 μmol g1 [110]
aza-CMP/RGO/C2N Z-scheme / / / 10.0 μmol g1 [110]
TiO2/CdS Z-scheme / / / 11.9 mmol h1 m2 for CH4 [111]
Sv-ZnIn2S4/MoSe2 Z-scheme / / / 63.21 mmol g1 h1 [112]
Zn/In:SnS2 Gradient band structure 0.23 mA cm2 0.3 VRHE 0.5 M Na2SO4 / [117]
NiCoFe-Bi/Mg:Ta3N5/Nb Gradient band structure 8.5 mA cm2 0.40 VRHE 1.0 M KOH / [121]
NiFe(OH)x/Ta:Fe2O3@Fe2O3 Gradient band structure 3.22 mA cm2 0.55 VRHE 1.0 M KOH / [35]
WO3/In2S3-P Gradient band structure 1.61 mA cm2 0.02 VRHE 0.1 M Na2SO4 / [75]
Grad-O CdS/Ti Gradient band structure 6.0 0.1 mA cm2 / 1.0 M Na2SO4/1.0 M N2H4·H2O / [122]
grad-P:Fe2O3/Co-Pi Gradient band structure 2.0 mA cm2 / 1 M KOH / [123]
Pd/b2-TiO2/b-Si Gradient band structure 8.3 mA cm2 (0 VRHE) 0.32 VRHE 1.0 M NaOH / [127]
4.3. Regulating Heterojunction Photoelectrodes with Second, heat is restricted to the surface of the catalyst owing
Photothermal, Piezoelectric, Pyroelectric, and Ferroelectric to local heating, eliminating the influence of electrolyte temper-
Effects ature. When the photothermal materials are irradiated by NIR
light, the photoanodes are heated in situ, and the increase in tem-
Although type-II and Z-scheme heterojunctions are extensively perature causes the thermal movement of molecules in the mate-
used, we can use semiconductors with special properties, such rial to increase significantly, thereby promoting electron
as photothermal and electrical properties, to further enhance transmission, accelerating reaction kinetics, and promoting
the separation and transport of carriers in the heterojunc- gas escape. He et al.[131] showed that NiOOH/FeOOH/Co3O4/
tions.[128] The photothermal effect mainly reduces the reaction BiVO4 combined standard sunlight and NIR irradiation to
barrier and accelerates carrier transport through in situ heating. expand light absorption, and in situ heating promoted charge
The piezoelectric effect, pyroelectric effect, and ferroelectric transfer while increasing the water oxidation kinetics, as shown
effect enhance the performance of PEC water splitting by in Figure 5. At the same time, they also proved that the photo-
enhancing the built-in electric field of the heterojunction. thermal effect has universal significance for the improvement of
photoanode performance. In addition, the measured tempera-
ture is much lower than the actual temperature. Increasing
4.3.1. Photothermal Effect
the temperature in situ in the nanoscale region can reduce
the reaction barrier and promote the photocatalytic water split-
Sunlight is an electromagnetic wave. According to the thermal
ting reaction. Zhang et al.[132] showed that Ni3S2 was heated
effect of the electromagnetic wave, the inside of a material is neu-
in situ with the assistance of the photothermal effect and pro-
tral and has two types of charges with the same amount of elec-
vided evidence that the reaction potential decreased as the lumi-
tricity. The alternating electric field in the propagation of
nescence time increased. Jin and co-workers[133] proved that the
electromagnetic waves makes the positive and negative charges
temperature increase of the Co3O4 electrode caused by NIR irra-
move in the material. This process is accompanied by collision
diation promoted the electrolyte to penetrate into the photoelec-
and friction, showing a “hot” phenomenon. As the frequency of
trode, and greater amounts of active material were exposed to the
infrared rays is close to the frequency of the material, it is easier
electrolyte, which may be due to the large electrochemical active
to cause resonance of the material, and the thermal effect is sig-
surface areas (ECSAs).
nificant. For example, Zhang et al.[129] used the photothermal
In short, the advantages of using photothermal materials to
effect of Cu2S to enhance the light absorption of visible light
construct heterojunctions are significant. On the basis of the het-
and near-infrared (NIR) light radiation and to profitably increase
erojunction itself, the addition of the photothermal effect brings
the rate of OER due to the in situ heating of Cu2S.
about a widening of the light absorption range, speeding up
Generally, there are a number of photothermal materials:
charge transport, improving OER reaction kinetics and other
organic composites, noble metals, carbon-based materials, and
advantages, and realizing a further improvement in the perfor-
transition metal semiconductors. The photothermal effect is
mance of the heterojunction.
mostly based on the principle of surface plasmon resonance.
The electromagnetic field of the surface plasmon is localized
in a range far smaller than the wavelength scale, and the energy 4.3.2. Piezoelectric Effect
of the incident light is concentrated in the deep subwavelength or
even nanometer-scale spatial range through the surface plasmon, According to the structural symmetry, the crystal structure can be
which makes the energy density of the light and local electric divided into 7 crystal systems and 32 point groups, 20 of which
field strength enhanced, forming so-called “hot spots,” indicating do not have center symmetry, and their electric dipole moments
that surface plasmons can significantly enhance the interaction can be changed due to elastic deformation, so they have piezo-
of light and matter. The characteristics of surface plasmon polar- electric such as ZnO, CdS, C3N4, MoS2, MoSe2, GaN, InGaN,
itons are closely related to materials, morphology and structure. ABO3, etc. Ten point groups with a unique polar axis in a piezo-
When the resonance wavelength of surface plasmon polaritons is electric body can exhibit spontaneous polarization, that is, elec-
close to the wavelength of incident light, the photothermal effect tric polarization exists even in the absence of an electric field.
is the strongest and the photothermal conversion efficiency is the They generate electric charges due to heating, so they are called
highest. Moreover, the free electron concentration, mobility, and pyroelectrics including ZnO, CdS, Bi2S3, ABO3, etc. Among
interband/intraband transition of carriers are three important these polar crystals, the crystal whose spontaneous polarization
factors that affect the properties of surface plasmons. Hot car- direction is changed by applying an external voltage is a ferroelec-
riers will be generated under light excitation and can excite tric, for example, ABO3. In addition to conventional oxides and
the electronic energy level or vibrational state of the surface plas- sulfides that can be used as photoanode, the flexible choice of
mon or adsorbed molecules before generating heat, making ABO3 structure of A and B cations makes the energy band struc-
them in a more active excited state, thereby greatly speeding ture tunable, making it one of the choices for photoanode.
up the chemical reaction rate. The electrons or holes are reso- Piezoelectric materials have a noncentrosymmetric crystal struc-
nated by light shining on the semiconductor, resulting in nano- ture. Due to their specific crystal structure, the electric dipoles
scale local heating.[130] Introducing the photothermal effect into generated in the material body change with the strain of the
the catalytic process has two obvious advantages. First, a catalyst material.[134] Generally, piezoelectric materials can be divided
with a photothermal material can realize self-heating under into three types of materials: ferroelectric materials, pyroelectric
visual or NIR light without an external heating device. materials, and piezoelectric materials, and their relationships are
Figure 5. a) Temperature changes of the electrolyte solution and Co3O4 with time. b) Temperature–time graphs of BVO, CB, and NFCB photoanodes
under AM1.5G illumination with or without simultaneous NIR light irradiation. Digital images of NFCB composite photoanode. O2 evolution process c,d)
without e,f ) with NIR light irradiation under the same current density. Reproduced with permission.[131] Copyright 2021, Wiley-VCH.
shown in Figure 6a.[135] The introduction of pyroelectric materi- of the nanorods. When the solution flow direction was perpen-
als and ferroelectric materials is described in Section 4.3.3 dicular to the nanorods, the nanorods were obviously bent, which
and 4.3.4, respectively. significantly improved the performance. The physical mecha-
Piezoelectric materials will not have a polarization electric nisms of the vertical flow of the solution are shown in
field under any external force. Once a tensile stress is applied Figure 6e–f. When the solution flowed perpendicularly to the
to the piezoelectric material, a polarization field emerges inside ZnO nanorods, the nanorods bent, and a built-in field was gen-
the piezoelectric material, and the orientation of the polarization erated inside the nanorods. The built-in field helped the photo-
field for applying compressive stress is the opposite. The specific generated holes arrive at the surface of the ZnO to cause
schematic diagrams are shown in Figure 6b–d.[136] oxidation reactions. When the solution flowed vertically, the
The presence of the piezoelectric effect redistributes the bending of the ZnO–WO3–x nanorods not only generated a
charges at the interface of the heterojunction and changes its built-in electric field in ZnO but also changed the band structure
built-in field and space charge region. Zhang and co-workers[137] at the interface between ZnO and WO3–x. As positively polarized
constructed a Pt/ZnO/Co-Pi photoanode, in which the piezoelec- electrons accumulated on the surface of ZnO, the Fermi level of
tric effect of ZnO under the action of mechanical energy effec- ZnO was close to that of CB, which further strengthened the
tively enhanced the electric field intensity, prevented transmission mechanism of the Z-scheme heterojunction and
photogenerated charges from recombining, and enhanced the improved the behavior of the PEC. As ultrasound is generally
water splitting performance of Pt/ZnO/Co-Pi. Therefore, build- used as the external pressure, the direction of the piezoelectric
ing a heterojunction in PEC will change the interface band bend- field cannot be well controlled. Therefore, simple linear sweep
ing, increase the strength of the built-in field, and expand the voltammetry cannot describe the piezoelectric effect very well.
depletion layer by drawing on the support of the piezoelectric Xu and co-workers[139] utilized a Au4/BaTiO3 heterostructure
potential in the piezoelectric material, which further promotes by measuring the degradation of organic dyes under auxiliary
the separation and transport of electrons and holes. Chen piezoelectric potential to prove that ultrasonic excitation inducing
et al.[138] agitated the solution to bend ZnO–WO3–x nanorods a piezoelectric field enhances the photocatalytic effect. The
to generate a piezoelectric electric field in ZnO and found that amount of sonophotocatalytic degradation of Au4/BaTiO3 in
as the agitation rate increased, the PEC activity of the Z-scheme 75 min was greater than that of photocatalytic degradation and
heterojunction improved. Additionally, the bending of the sonocatalytic degradation.
ZnO–WO3–x nanorods was controlled by changing the direction Therefore, piezoelectric materials are used to construct heter-
of the solution flow. When the direction of the solution flow was ojunctions, and ultrasound is used to impart external force to
parallel to the nanorods, the performance of the nanorods was semiconductors, causing the occurrence of polarization
not significantly changed due to the insignificant deformation fields because of the redistribution of polarization charges.
Figure 6. a) The relationship between piezoelectric materials, pyroelectric materials, and ferroelectric materials. Piezoelectric semiconductors b) without
strain, c) under tensile stress, and d) under compressive stress. e,f ) Schematic band bending diagrams of ZnO and ZnO/WO3–x with the piezoelectric
effect. Reproduced with permission.[138] Copyright 2019, Elsevier.
Reasonable utilization of a polarized electric field can effectively water, a photoanode can use the pyroelectric effect to make
improve the carrier transport of semiconductor heterojunctions. the thermally generated charges participate in the redox reactions
of water. The voltage or current generated by the pyroelectric
4.3.3. Pyroelectric Effect effect has the following relationship[143]
dT
When the spontaneous polarization in a material changes with I ¼ Pc ⋅ A ⋅ (5)
dt
temperature, the surface of the material generates positive and
negative polarization charges gathering in different areas along a P c ⋅ h ⋅ ΔT
V ¼ (6)
certain direction.[140] This characteristic is defined as the pyro- ε0 ε33
electric effect, and a material with this pyroelectric effect is a
pyroelectric material. The discharge effect can be divided into where A is the surface area, dT/dt refers to the temperature
the first-level pyroelectric effect, the second-level pyroelectric change rate, Pc denotes the pyroelectric coefficient, h is the sam-
effect, and the third-level pyroelectric effect.[141] The so-called ple thickness, ΔT represents the temperature change, ε0 is
first-level pyroelectric effect means that the material maintains defined as the vacuum dielectric constant, and ε33 represents
a constant strain, that is, a pyroelectric effect is derived from the relative dielectric constant of the material. The pyroelectric
the polarization caused by temperature changes. The second- current is in direct proportion to the temperature change rate
level pyroelectric effect is a pyroelectric effect that occurs when and surface area, and the pyroelectric voltage is positively related
the material is subjected to a constant stress. This is the polari- to the temperature change and the thickness of the sample.
zation caused by the deformation of the material. The third-level Therefore, as long as the pyroelectric material is thick enough
pyroelectric effect is the polarization induced by the gradient and the temperature changing rate is fast, the polarization poten-
temperature of the material and the shear stress. tial that meets the requirements of the photolysis water can be
When a pyroelectric material is heated, its electric dipoles obtained, the reaction rate of the photolysis of water can be accel-
vibrate violently, and the spontaneous polarization is weakened erated, and the performance can be improved. As shown in
so that the free charge generated on the surface is redistributed Figure 7d, Zhang et al.[144] constructed the CdS-2-mercaptoben-
and electrons are released outward. In contrast, when a pyroelec- zimidazole (2MBI) heterojunction to improve its hydrogen pro-
tric material is cooled, the electric dipole vibrates slightly along duction capacity under cooling and heating cycles, which was
the polarization axis to increase the spontaneous polarization and attributable to the pyroelectric effect of CdS, enhancing the
absorb electrons from the outside.[142] The specific diagram is built-in field and broadening the width of the depletion layer
shown in Figure 7a–c. In the process of the PEC photolysis of to promote the transport and separation of carriers. Dai
Figure 7. Piezoelectric semiconductors a) without temperature change, b) under warming up, and c) cooling down. d) Schematic description of the
energy band and pyroelectric catalysis hydrogen evolution. Reproduced with permission.[144] Copyright 2020, Elsevier.
et al.[145] depended on the photothermal effect of carbon nano- when the parallel dipoles are not so strongly bound to each other,
tubes to provide temperature oscillation to the poly(vinylidene the polarization direction is easily reversed by the electric field,
fluoride-co-hexafluropropylene) microfiber to generate a polar- which is called ferroelectrics.[148] When the action of the polari-
ized electric field on the surface of the microfiber. TPC, EIS, zation field belongs only to the ferroelectric material itself, once
and TRPL photoluminescence proved that the existence of a pyro- the electrode and the substrate at the interface cannot provide
electric field greatly inhibited carrier recombination, which was enough free charge to compensate for the polarization charge,
the key to increasing the hydrogen production rate. For the first a depolarization electric field opposite to the polarization field
time, Li and co-workers[146] prepared FeS2/Bi2S3 for photocata- will be generated, which can also be understood as an internal
lytic hydrogen production by combining the photothermal effect electric field.[149] The appearance of this polarization field will
of FeS2/Bi2S3 and circulating condensed water to provide tem- affect the band bending degree. The accumulation of positive
perature vibration to the pyroelectric semiconductor Bi2S3. In polarity charges impels downward bending of the energy band,
this way, the direction of the Bi2S3 pyroelectric field was con- and the collection of negative polarity charges induces upward
trolled, and FeS2/Bi2S3 stimulated the photogenerated electrons bending of the energy band, as shown in Figure 8a,b.[150]
of FeS2 to migrate to the surface of Bi2S3 under the influence of Therefore, if an applied external electric field is consistent with
the positive polarization field, which boosted the decomposition the polarization field direction of the ferroelectric material, the
of water to produce hydrogen. built-in field intensity will be enhanced, and if an applied external
The pyroelectric catalytic process requires the following indis- electric field is reversed with its own polarization field direction,
pensable conditions: one is a pyroelectric catalyst with sufficient the built-in electric field intensity will be reduced.[151] However,
thickness, another is a temperature gradient, and the third is that the compounds of semiconductor materials and ferroelectric
the working temperature is below the Curie temperature.[147] materials are influenced not only by the built-in electric field
Only when there are temperature fluctuations in the pyroelectric but also by band bending at the interface. At the same time, when
body and they are below the Curie temperature, it is possible for positive charges appear at the interface, they will cause the energy
the pyroelectric free charges with catalytic activity to be gener- band to bend downward, resulting in an increase in the depletion
ated. At the same time, the thickness of the catalyst is positively layer in the ferroelectric. When negative charges appear at the
correlated with the pyroelectric potential difference, which also interface, the energy band bends upward, resulting in a decrease
affects the oxidation–reduction reaction. Therefore, while con- in the depletion layer in the ferroelectric. Wu and co-workers[152]
structing heterojunctions, materials with pyroelectric effects compounded SrTiO3 on the surface of TiO2 due to lattice distor-
are selected, or some modification methods (such as element tion to induce the iron polarization of SrTiO3, which applied a
substitution, construction of ultrathin structures, etc.) are used positive polarization field to improve performance and applied
to change the symmetry of the materials to have pyroelectric a negative polarization field to decrease performance. When pos-
properties, which contribute to the catalytic ability of the itive poling was applied, the built-in electric field was consistent
heterostructure. with the transfer direction of charge, and the driving force was
enhanced. When negative poling was applied, as the SrTiO3 is
4.3.4. Ferroelectric Effect only 10 nm thick, the band bending direction was not changed,
but the built-in electric field intensity was weakened. In the case
Pyroelectric crystals can be classified into two categories. One is a of positive and negative polarization, the distribution of polariza-
pyroelectric crystal whose energy is a minimum value, and the tion charges, the built-in electric field and energy band bending,
interaction between parallel dipoles is very strong in most cases, and the corresponding performance characteristics are presented
and it is difficult to change its direction. Another type is that in detail in Figure 8c–e. However, Zhang et al.[153] prepared a
Figure 8. a) Schematic diagram of the depolarization field generated by the polarization charges of the ferroelectric material without compensation.
b) Schematic of the depolarization field generated by the polarization charges of the ferroelectric material with partial compensation. Descriptions of band
bending with c) positive poling, d) no poling, and e) negative poling. Reproduced with permission.[152] Copyright 2017, Wiley-VCH. f ) Schematic depic-
tions of J–V curves and g) electron transfer diagrams of 3 wt% BTO/TiO2/QDs with positive poling, no poling, and negative poling. Reproduced with
permission.[153] Copyright 2021, Elsevier.
BaTiO3/TiO2/graded quantum dots (QDs) photoelectrode, and was significantly increased, proving that the piezoelectric and
the photocurrent density values on applying positive and negative pyroelectric effects played an important role in the heterojunc-
electric fields were reduced and improved, respectively, as shown tion. Liu et al.[155] added a BaTiO3 layer between Ag2O and
in Figure 8f–g. Because the BaTiO3 nanoparticles were 208 nm, TiO2, which applied forward polarization to generate a piezoelec-
when the BaTiO3/TiO2/graded QDs photoelectrodes were nega- tric field and enhanced the photogenerated carrier lifetime.
tively polarized, the band bending direction and the increase in Table 3 summarizes the catalytic performance of heterojunctions
the depletion layer in the BaTiO3 body were beneficial to the sep- with special physical effects.
aration and transport of the photogenerated electrons. However,
when the BaTiO3/TiO2/graded QDs photoelectrodes were posi- 4.4. Intermediate Layer
tively polarized, the band bending direction changed greatly,
which suppressed the performance. Generally, heterojunctions have two fundamental problems: lat-
Because the piezoelectric effect, pyroelectric effect, and ferro- tice mismatch and energy band mismatch, which are extremely
electric effect all promote the enhancement of the built-in electric detrimental to charge transport in the bulk.[156] Therefore, adding
field in the heterojunctions and increase the width of the deple- an intermediate layer between the two heterojunctions can
tion layer, the application of a variety of effects in the heterojunc- relieve these problems. Generally, the heterojunction photoelec-
tion can achieve shocking results. For example, Zhao et al.[154] trode has an interface between FTO and the semiconductor, a
applied ultrasonic vibration and temperature changes to the contact interface between the heterojunction interfaces, and
Ag2O/BaTiO3 heterojunction, and the pollution degradation time an interface between the photoanode and the electrolyte. This
Table 3. Summary of PEC performance for heterojunctions aided by different physical effects.
Sample Physical effects Light condition J (1.23 VRHE) Testing Catalytic activity Ref.
condition
Cu2S/Fe2O3-anneal Photothermal UV–vis–NIR 1.19 mA cm2 1 M KOH / [129]
NiOOH/FeOOH/Co3O4/ Photothermal AM1.5G þ NIR 6.34 mA cm2 0.5 M KPi 93% of Faraday efficiency for O2 [131]
BiVO4 evolution
Ni2S3/FeNi Foam Photothermal IR 492.2 mA cm2 (1.55 1 M KOH / [132]
VRHE)
Ni-Co3O4 Photothermal 808 nm laser of 5 W cm2 441.8 mA cm2 (1.6 1 M KOH / [133]
VRHE)
Pt/ZnO/Co-Pi Piezoelectric AM1.5G 0.80 mA cm2 0.1 M Na2SO4 / [137]
90 kHz ultrasonic vibrations 0.1 M KPi
ZnO-WO3–x Piezoelectric AM1.5G 3.38 mA cm2 1 M Na2SO4 / [138]
1000 rpm magnetic stir
Au4/BaTiO3 Piezoelectric AM1.5G and ultrasonic / / MO degradation rate of 97% in [139]
75 min
CdS-2MBI Pyroelectric 0.05 mW m2 light 25–55 C / / H2 yield of 4.3 μmol g1 per [144]
(12 C min1) thermal cycle
PVDF-HFP/CNT/CdS Pyroelectric 280 W Xenon lamp / / H2 evolution rate of [145]
451 μmol g1 h1
FeS2/Bi2S3 Pyroelectric AM1.5G / / H2 evolution rate of [146]
22.7–33.9 C (2.24 C min1) 16.8 mmol g1 h1
TiO2–SrTiO3 Ferroelectric AM1.5G 1.43 mA cm2 1 M NaOH / [152]
10 V poling voltage
BaTiO3/TiO2/graded QDs Ferroelectric AM1.5 G 15.3 mA cm2 0.25 M Na2S H2 evolution rate of [153]
10 V DC voltages (0.5 VRHE) 0.35 M Na2SO3 96.5 mL cm2 d1
Ag2O/BaTiO3 Piezoelectric AM1.5G 3.48 mA cm2 0.1 M Na2SO4 MO degradation rate of nearly [154]
Pyroelectric Ultrasonic vibrations (0.5 VSCE) 100% in 60 min
20–50 C (1 C min1)
TiO2/ BaTiO3/Ag2O Piezoelectric 150 W Xenon lamp 1.8 mA cm2 1 M NaOH / [155]
Ferroelectric A bias of 0.3 V (0.8 VAg/Cl)
section mainly focuses on adding intermediate layers at layer and rutile TiO2 nanorods, bringing about enhanced photo-
these three interfaces to achieve PEC performance generated carrier separation performance and better electron
improvements. storage efficiency at the TiO2/FTO interface, as well as an excel-
When complex defect states are generated at the interface lent electron transfer rate from defect states in the CdS QDs to
between the photoanode and the FTO, the defect states become the CB of the TiO2 in TiO2/CdS heterostructure.
recombination centers of photogenerated pairs of electrons and As the CB and VB band positions of the two semiconductors
holes, which seriously affect the photocatalytic performance of are quite different from each other, which is unfavorable to the
photoanodes. At the FTO/BiVO4 interface, the defect states transportation of photogenerated electrons and holes, an inter-
trapped a large number of electrons because the defect states mediate layer is added at the interface of the heterojunction to
acted as deep-level defects in the FTO and the accumulated elec- achieve a graded transition to promote the bulk phase separation
trons flattened the energy band, which promoted the back contact of the carriers. Cao et al.[159] showed that the stepped band struc-
of the photogenerated holes and resulted in serious recombina- ture between ZnO, CdS, and ZnFe2O4 improved the separation
tion. As shown in Figure 9a, the introduction of SnO2 into FTO/ efficiency of photogenerated charges. Pan et al.[76] achieved a
BiVO4 by Liang and co-workers[157] passivated the defects at the double type-II heterojunction by adding a TiO2 layer between
interface, inhibited the back transport of photogenerated holes, SnO2 and BiVO4 to promote the separation of charges and car-
and achieved effective separation between photogenerated elec- riers in the bulk. In addition to the impact of band mismatching,
trons and holes, thereby improving its PEC performance. In there may also be interface contact problems. Therefore, optimiz-
addition, the lattice mismatch between the FTO and the semicon- ing interface contact and reducing recombination can also be
ductor makes the interface contact between the semiconductor achieved through the use of an intermediate layer.[160] In
and the FTO poor. A seed layer can be introduced for the transi- Figure 9b, Luo and co-workers[161] deposited Al2O3 on the inter-
tion, which can not only provide a bridge for carrier transport but face between Si/SiOx and Ni/NiOx by atomic layer deposition,
also achieve directional growth of crystals. Chen et al.[158] con- which passivated the interface defects and effectively reduced
trolled the band arrangement between the anatase TiO2 seed the pinning effect of the Fermi level.
Figure 9. a) Schematic depiction of the band energy positions after the introduction of SnO2. Reproduced with permission.[157] Copyright 2011, American
Chemical Society. b) Band structure of the n-Si/SiOx/Al2O3/Ni/NiOx/NiOOH photoanode. Reproduced with permission.[161] Copyright 2019, Wiley-VCH.
c) Schematic diagrams showing the introduction of ZIF-GHJ as a photogenerated hole extraction layer at the photoanode/electrolyte interface.
Reproduced with permission.[166] Copyright 2021, Wiley-VCH. d) An energetics schematic showing the photogenerated carrier transport process of
composite photoanodes without a mediator and with a mediator. Reproduced with permission.[167] Copyright 2019, Wiley-VCH.
In view of the photoanode/electrolyte interface, the carriers transfer of holes to the photoanode/electrolysis interface and fur-
may accumulate on the surface to cause serious interface recom- ther inhibited the recombination of the carriers. Ning et al.[167]
bination and affect performance due to electron back transfer found that there was a large potential difference between the VB
and sluggish transfer dynamics.[162–165] Therefore, an intermedi- of reduced BiVO4 and the potential of splitting water to produce
ate layer is added to promote carrier transport for the water split- oxygen. Therefore, in Figure 9d, 5,10,15,20-tetrakis(4-carboxy-
ting redox reaction. Tang et al.[166] found that TiO2 forced phenyl)porphyrin-Co(CoPy) acted as a transport channel, similar
electrons to be transported to the electrolyte interface, causing in action to that of a volleyball setter, to advance the transfer rate
serious carrier loss. By constructing a Co3O4/TiO2 heterojunc- of photogenerated holes reaching the surface of the photoanode,
tion, the back-transport phenomenon of the TiO2 electrons and IMPS certified that BiVO4/CoPy/FeNi(OH)x was longer than
was suppressed, and the matching type-II heterojunction band BiVO4/FeNi(OH)x in accordance with τd ¼ (2πfmax)1, in which
structure promoted the transport of the holes to the electrolyte. CoPy facilitated hole transmission from BiVO4 to FeNi(OH)x and
As shown in Figure 9c, considering that ZIF-CoxZn1–x had a avoided the accumulation of carriers at the photoanode/electrol-
more negative VB position than Co3O4, the driving force for hole ysis interface.
extraction in the ZIF-CoxZn1–x/Co3O4/TiO2 multigraded hetero- Therefore, in light of the unique problems of the selected
junction was enhanced, which was conducive to the rapid semiconductors, such as large energy band differences, interface
Table 4. Summary of PEC performance for heterojunctions with intermediate layers.
Sample Intermediate layer J (1.23 VRHE) IPCE Testing condition Ref.

FTO/SnO2/BiVO4 Seed layer / 46% at 450 nm, 1.63 VRHE 0.15 M K2SO4 [157]
TiO2-SC-c/CdS Seed layer 6.80 mA cm2 (1.0 VRHE) / 0.25 M Na2S/0.35 M Na2SO3 [158]
ZnFe2O4/CdS/ZnO Protection layer 3.88 mA cm2 (0 VAg/AgCl) / 0.5 M Na2S [159]
2
SnO2/TiO2/BiVO4/FeOOH/NiOOH Hole blocking layer 3.1 mA cm 85% at 350 nm, 1.23 VRHE 0.5 M Na2SO4/0.1 M Na2SO3 [76]
n-Si/SiOx/Al2O3/Ni/NiOx/NiOOH Passivation layer 28 mA cm2 70% at 420–800 nm 1 M Na2SO4/0.1 M KPi [161]
1.23 VRHE
ZIF-CoxZn1–x/Co3O4/TiO2 Hole extraction layer 2.91 mA cm2 / 1 M NaOH [166]
R-BiVO4/CoPy/FeNi(OH)x Passivation layer 4.75 mA cm2 / 0.2 M Na2SO4 [167]
Hole extraction layer
defects, and other problems, it is easy to cause carrier losses, and transition, etc. to enhance the photogenerated carrier separation
the introduction of an intermediate layer can be used as a tran- and transport capability in the heterojunctions. To monitor car-
sition, thereby realizing the effective transmission and separa- rier transmission, characterization methods, such as TRPL, TAS,
tion of carriers and improving the performance of photoanode TPV, TPC, IMPS, and IMVS, are provided in this review for char-
PEC water splitting. Table 4 briefly summarizes some hetero- acterizing the behavior of carriers in heterojunctions.
junctions with different functional interlayers. It is not common for the construction of photoelectrodes to
Regarding the four strategies for promoting carrier transport use only a single semiconductor. In contrast, the combination
in the above analysis, the following four points need to be clari- of multiple semiconductors has become the main trend of photo-
fied: first, micro–nanostructures are diverse and each material electrode construction. Therefore, the successful construction of
has several conventional morphologies due to its intrinsic crys- an effective heterojunction has become an urgent problem to be
tallographic structure, so it is a certain challenge to synthesize solved. The construction of efficient heterojunctions must take
new micro–nanostructures to some extent. Moreover, the nano- into account the factors of lattice constants, thermal expansion
structure is selectively fine-tuned according to the specific char- coefficients, and band positions, which depend on the material
acteristics of the material. For example, the hole diffusion length itself. Lattice mismatch and different thermal expansion coeffi-
of BiVO4 is about 70 nm, so the synthesized BiVO4 is often a cients tend to introduce a large number of interface states during
hollow structure rather than a bulk structure. Second, the con- the construction of heterojunctions, resulting in nonradiative
struction of type-II and Z-scheme heterojunctions requires the recombination. If the epitaxial layer below the critical thickness
selection of semiconductors that match the energy bands as is grown by atomic layer epitaxy, the lattice mismatch can be
far as possible, while the element selection of gradient doping compensated by elastic strain. Therefore, first, it is necessary
and conventional preparation methods have certain limitations. for the construction of heterostructures to correctly select semi-
Third, semiconductors with specific structural characteristics conductor materials and growth techniques. Second, in situ
have special physical effects. Fourth, the selection of the interme- growth and one-step synthesis can serve as the preferred options
diate layer needs to be determined according to the specific situ- for the construction of heterojunctions. Finally, type-II and
ation, and the preparation method should avoid introducing Z-scheme heterostructures are the most advantageous hetero-
more defect states. junction types at present. It is vital to select semiconductors with
suitable energy bands for compound heterostructures. However,
semiconductors that are not fully adapted can be modified and
5. Conclusion regulated by introducing heteroatoms or defects so as to meet the
requirements of energy band matching. This review has
In conclusion, we summarized some strategies that can be used launched a systematic and detailed discussion and summary
to improve carrier transport in heterojunctions. First, some on the carrier transport problem of heterojunctions.
neatly arranged nanotopographies are chosen as promising can- Due to the current serious problems of energy shortages and
didates for photoelectrodes because they can increase light environmental pollution, PEC production capacity and pollutant
absorption and improve the carrier transport channel. Second, degradation show great advantages over many other solutions.
by constructing band-matched heterojunctions, such as type-II The key contents of PEC water splitting are light absorption,
and Z-scheme heterojunctions, bulk carrier transport is pro- the competition between the separation and recombination of
moted. In addition, effective bulk phase separation of heterojunc- photogenerated electron–hole pairs, and the reaction kinetics
tions can also be achieved with the help of gradient doping. Then, of the OER. With the discovery of nanomaterials with 2D layered
by utilizing the photothermal effect, piezoelectric effect, and structures, vertical and in-plane heterostructures received great
other effects to further accelerate the carrier bulk transfer rate, attention from researchers. The construction of vertical hetero-
improved photoelectric catalytic performance was achieved. junctions mainly uses mechanical transfer technology to stack
Finally, we introduced the intermediate layer, which plays the layered 2D materials in the vertical direction, and the layers
roles of electron transport layer, hole blocking layer, energy-level are combined by van der Waals forces, which has nearly perfect
interfaces. Therefore, van der Waals heterojunctions provide an [6] Z. Li, M. Hu, P. Wang, J. Liu, J. Yao, C. Li, Coord. Chem. Rev. 2021, 439,
ideal platform for the exploration and understanding of hetero- 213953.
junctions and may be another possible route for photoelectroca- [7] Y. Qiu, Z. Pan, H. Chen, D.. Ye, L. Guo, Z. Fan, S. Yang, Sci. Bull.
talysis. An in-plane heterojunction refers to a composite material 2019, 64, 1348.
in which two 2D layered materials are bonded by chemical bonds [8] M. Faraji, M. Yousefi, S. Yousefzadeh, M. Zirak, N. Naseri, T. H. Jeon,
in a plane. The bonding force of the chemical bond is much W. Choi, A. Z. Moshfegh, Energy Environ. Sci. 2019, 12, 59.
[9] C. Jiang, S. J. A. Moniz, A. Wang, T. Zhang, J. Tang, Chem. Soc. Rev.
larger than the van der Waals force, so in-plane heterojunctions
2017, 46, 4645.
are superior to the vertical heterojunctions in driving the sepa-
[10] H. Wu, H. L. Tan, C. Y. Toe, J. Scott, L. Wang, R. Amal, Y. H. Ng, Adv.
ration and migration of photogenerated charges. However, the Mater. 2020, 32, 1904717.
construction of planar heterojunctions is currently limited to sev- [11] H. Huang, B. Pradhan, J. Hofkens, M. B. J. Roeffaers, J. A. Steele, ACS
eral 2D nanomaterials due to strict lattice matching constraints Energy Lett. 2020, 5, 1107.
and the difficulty of selecting suitable material components.[169] [12] A. Ghobadi, T. G. Ulusoy Ghobadi, F. Karadas, E. Ozbay, Adv. Opt.
Therefore, exploring the possibility of new 2D material hetero- Mater. 2019, 7, 1900028.
junction structures is crucial for promoting the development [13] M. Volokh, G. Peng, J. Barrio, M. Shalom, Angew. Chem., Int. Ed.
of materials field and improving solar energy conversion 2019, 58, 6138.
efficiency. [14] R. Siavash Moakhar, S. M. Hosseini-Hosseinabad, S. Masudy-Panah,
However, in order to obtain a universal modification method A. Seza, M. Jalali, H. Fallah-Arani, F. Dabir, S. Gholipour, Y. Abdi,
for the heterojunction interface problem, we need to pay more M. Bagheri-Hariri, N. Riahi-Noori, Y. F. Lim, A. Hagfeldt,
attention to the physical and chemical properties of the material, M. Saliba, Adv. Mater. 2021, 33, 2007285.
[15] A. Fujishima, K. Honda, Nature 1972, 238, 37.
which can be exploited to improve the lattice mismatch, thermal
[16] Y. Wang, W. Tian, C. Chen, W. Xu, L. Li, Adv. Funct. Mater. 2019, 29,
expansion coefficient mismatch, and energy band mismatch. Of
1809036.
course, mining new nanomaterials and preparation methods is [17] Y. Yang, S. Niu, D. Han, T. Liu, G. Wang, Y. Li, Adv. Energy Mater.
also an important part of current heterojunction engineering. 2017, 7, 1700555.
Therefore, the exploration based on the mechanism of improving [18] Q. Wang, K. Domen, Chem. Rev. 2020, 120, 919.
the heterojunction interface problem will become the main direc- [19] Y. He, T. Hamann, D. Wang, Chem. Soc. Rev. 2019, 48, 2182.
tion of future research, rather than simply selecting materials for [20] S. M. Thalluri, L. Bai, C. Lv, Z. Huang, X. Hu, L. Liu, Adv. Sci. 2020, 7,
trial and stacking. 1902102.
[21] X. Zhang, H. Guo, G. Dong, Y. Zhang, G. Lu, Y. Bi, Appl. Catal., B
2020, 277, 119217.
[22] M. Kim, B. Lee, H. Ju, J. Y. Kim, J. Kim, S. W. Lee, Adv. Mater. 2019, 31,
Acknowledgements
1903316.
The authors acknowledge the support from the National Natural Science [23] X. Zhou, P. Gai, P. Zhang, H. Sun, F. Lv, L. Liu, S. Wang, ACS Appl.
Foundation of China (grant nos. 52025028, 52072254, 51772197, and Mater. Interfaces 2019, 11, 38993.
51872191) and the Priority Academic Program Development of Jiangsu [24] H. Zhao, X. Li, M. Cai, C. Liu, Y. You, R. Wang, A. I. Channa, F. Lin,
Higher Education Institutions. D. Huo, G. Xu, X. Tong, Z. M. Wang, Adv. Energy Mater. 2021, 11,
2101230.
[25] B. Zhang, H. Zhang, Z. Wang, X. Zhang, X. Qin, Y. Dai, Y. Liu,
Conflict of Interest P. Wang, Y. Li, B. Huang, Appl. Catal., B 2017, 211, 258.
[26] M. Sun, R. T. Gao, J. He, X. Liu, T. Nakajima, X. Zhang, L. Wang,
The authors declare no conflict of interest. Angew. Chem., Int. Ed. 2021, 60, 17601.
[27] Z. Wu, Y. Zhao, W. Jin, B. Jia, J. Wang, T. Ma, Adv. Funct. Mater. 2021,
31, 2009070.
[28] L. Ran, S. Qiu, P. Zhai, Z. Li, J. Gao, X. Zhang, B. Zhang, C. Wang,
Keywords L. Sun, J. Hou, J. Am. Chem. Soc. 2021, 143, 7402.
charge carrier transfer, heterojunctions, photoanodes, [29] Y. Pihosh , V. Nandal, T. Minegishi, M. Katayama, T. Yamada, K. Seki,
photoelectrochemical (PEC) water splitting M. Sugiyama, K. Domen, ACS Energy Lett. 2020, 5, 2492.
[30] Y. Wu, X. Liu, H. Zhang, J. Li, M. Zhou, L. Li, Y. Wang, Angew. Chem.,
Received: October 18, 2021 Int. Ed. 2021, 60, 3487.
Revised: January 24, 2022 [31] S. Liu, R. Gao, M. Sun, Y. Wang, T. Nakajima, X. Liu, W. Zhang,
Published online: March 13, 2022 L. Wang, Appl. Catal., B 2021, 292, 120063.
[32] B. Zhang, X. Huang, Y. Zhang, G. Lu, L. Chou, Y. Bi, Angew. Chem.,
Int. Ed. 2020, 59, 18990.
[1] T. R. Cook, D. K. Dogutan, S. Y. Reece, Y. Surendranath, T. S. Teets, [33] F. Li, H. Yang, Q. Zhuo, D. Zhou, X. Wu, P. Zhang, Z. Yao, L. Sun,
D. G. Nocera, Chem. Rev. 2010, 110, 6474. Angew. Chem., Int. Ed. 2021, 60, 1976.
[2] J. Qi, W. Zhang, R. Cao, Adv. Energy Mater. 2018, 8, 1701620. [34] Y. Yu, Y. Huang, Y. Yu, Y. Shi, B. Zhang, Nano Energy 2018,
[3] R. E. Blankenship, D. M. Tiede, J. Barber, G. W. Brudvig, G. Fleming, 43, 236.
M. Ghirardi, M. R. Gunner, W. Junge, D. M. Kramer, A. Melis, [35] H. Zhang, D. Li, W. J. Byun, X. Wang, T. J. Shin, H. Y. Jeong, H. Han,
T. A. Moore, C. C. Moser, D. G. Nocera, A. J. Nozik, D. R. Ort, C. Li, J. S. Lee, Nat. Commun. 2020, 11, 4622.
W. W. Parson, R. C. Prince, R. T. Sayre, Science 2011, 332, 805. [36] G. Liu, M. Wang, H. Wang, R. E. A. Ardhi, H. Yu, D. Zou, J. K. Lee,
[4] J. Lin, W. Wang, G. Li, Adv. Funct. Mater. 2020, 30, 2005677. Nano Energy 2018, 49, 95.
[5] P. AR, IJIRAE 2014, 2, 2349. [37] J. H. Kim, J. S. Lee, Adv. Mater. 2019, 31, 1806938.
[38] M. Ma, K. Zhang, P. Li, M. S. Jung, M. J. Jeong, J. H. Park, Angew. [71] P. Xu, C. L. Gray, L. Xiao, T. E. Mallouk, J. Am. Chem. Soc. 2018, 140,
Chem., Int. Ed. 2016, 55, 11819. 11647.
[39] S. Singh, H. Chen, S. Shahrokhi, L. Wang, C. Lin, L. Hu, X. Guan, [72] S. Cao, Y. Wu, J. Hou, B. Zhang, Z. Li, X. Nie, L. Sun, Adv. Energy
A. Tricoli, Z. Xu, T. Wu, ACS Energy Lett. 2020, 5, 1487. Mater. 2019, 10, 1902935.
[40] J. Su, T. Hisatomi, T. Minegishi, K. Domen, Angew. Chem., Int. Ed. [73] R. Tang, S. Zhou, Z. Zhang, R. Zheng, J. Huang, Adv. Mater. 2021,
2020, 59, 13800. 33, 2005389.
[41] D. Lee, A. Kvit, K. Choi, Chem. Mater. 2018, 30, 4704. [74] J. W. Yang, I. J. Park, S. A. Lee, M. G. Lee, T. H. Lee, H. Park, C. Kim,
[42] H. Zhu, Q. Yang, D. Liu, D. Liu, W. Zhang, Z. Chu, X. Wang, S. Yan, J. Park, J. Moon, J. Y. Kim, H. W. Jang, Appl. Catal., B 2021, 293,
Z. Li, Z. Zou, J. Phys. Chem. Lett. 2020, 11, 9184. 120217.
[43] D. Chen, Z. Liu, S. Zhang, Appl. Catal., B 2020, 265, 118580. [75] W. Tian, C. Chen, L. Meng, W. Xu, F. Cao, L. Li, Adv. Energy Mater.
[44] Y. Fu, Y. Lu, F. Ren, Z. Xing, J. Chen, P. Guo, W. Pong, C. Dong, 2020, 10, 1903951.
L. Zhao, S. Shen, Solar RRL 2019, 4, 1900349. [76] Q. Pan, A. Li, Y. Zhang, Y. Yang, C. Cheng, Adv. Sci. 2020, 7,
[45] Y. Gu, A. Wu, Y. Jiao, H. Zheng, X. Wang, Y. Xie, L. Wang, C. Tian, 1902235.
H. Fu, Angew. Chem. Int. Ed. 2021, 60, 6673. [77] K. Cho, Y. Sung, Nano Energy 2017, 36, 176.
[46] J. Low, J. Yu, M. Jaroniec, S. Wageh, A. A. Al Ghamdi, Adv. Mater. [78] B. Dong, J. Cui, Y. Qi, F. Zhang, Adv. Mater. 2021, 33, 2004697.
2017, 29, 1601694. [79] N. N. Vu, S. Kaliaguine, T. O. Do, Adv. Funct. Mater. 2019, 29,
[47] M. Xiao, B. Luo, Z. Wang, S. Wang, L. Wang, Solar RRL 2020, 4, 1901825.
1900509. [80] W. Wang, J. Dong, X. Ye, Y. Li, Y. Ma, L. Qi, Small 2016, 12, 1469.
[48] Q. Xu, L. Zhang, B. Cheng, J. Fan, J. Yu, Chem 2020, 6, 1543. [81] B. Dong, J. Cui, Y. Gao, Y. Qi, F. Zhang, C. Li, Adv. Mater. 2019, 31,
[49] Z. Wang, C. Li, K. Domen, Chem. Soc. Rev. 2019, 48, 2109. 1808185.
[50] X. An, Y. Wang, J. Lin, J. Shen, Z. Zhang, X. Wang, Sci. Bull. 2017, 62, [82] S. Vanka, B. Zhou, R. A. Awni, Z. Song, F. A. Chowdhury, X. Liu,
599. H. Hajibabaei, W. Shi, Y. Xiao, I. A. Navid, A. Pandey, R. Chen,
[51] S. Yi, Z. Wang, H. Li, Z. Zafar, Z. Zhang, L. Zhang, D. Chen, Z. Liu, G. A. Botton, T. W. Hamann, D. Wang, Y. Yan, Z. Mi, ACS
X. Yue, Appl. Catal., B 2021, 283, 119649. Energy Lett. 2020, 5, 3741.
[52] S. Lee, J. Song, Y. R. Jo, K. S. Choi, J. Lee, S. Seo, T. L. Kim, H. W. Jang, [83] F. F. Abdi, L. Han, A. H. Smets, M. Zeman, B. Dam, R. van de Krol,
C. Jeon, B. J. Kim, B. Kim, S. Lee, ACS Appl. Mater. Interfaces 2019, 11, Nat. Commun. 2013, 4, 2195.
44069. [84] D. Zhao, Y. Wang, C. Dong, Y. Huang, J. Chen, F. Xue, S. Shen,
[53] Z. Yan, H. Sun, X. Chen, H. Liu, Y. Zhao, H. Li, W. Xie, F. Cheng, L. Guo, Nat. Energy 2021, 6, 388.
J. Chen, Nat. Commun. 2018, 9, 2373. [85] J. Low, J. Yu, M. Jaroniec, S. Wageh, A. A. Al-Ghamdi, Adv. Mater.
[54] L. Zhang, J. Ran, S. Z. Qiao, M. Jaroniec, Chem. Soc. Rev. 2019, 48, 2017, 29, 1601694.
5184. [86] I. Grigioni, L. Ganzer, F. V. A. Camargo, B. Bozzini, G. Cerullo,
[55] T. Kirchartz, J. A. Márquez, M. Stolterfoht, T. Unold, Adv. Energy E. Selli, ACS Energy Lett. 2019, 4, 2213.
Mater. 2020, 10, 1904134. [87] C. Pei, Y. Chen, L. Wang, W. Chen, G. Huang, Appl. Surf. Sci. 2021,
[56] J. Xu, X. Li, Z. Ju, Y. Sun, X. Jiao, J. Wu, C. Wang, W. Yan, H. Ju, J. Zhu, 535, 147682.
Y. Xie, Angew. Chem., Int. Ed. 2019, 58, 3032. [88] X. Zhao, J. Feng, J. Liu, J. Lu, W. Shi, G. Yang, G. Wang, P. Feng,
[57] J. Li, H. Chen, C. A. Triana, G. R. Patzke, Angew. Chem., Int. Ed. 2021, P. Cheng, Adv. Sci. 2018, 5, 1700590.
60, 18380. [89] J. J. Deng, X. X. Lv, K. Q. Nie, X. L. Lv, X. H. Sun, J. Zhong, ACS Catal.
[58] Z. Wei, W. Wang, W. Li, X. Bai, J. Zhao, E. C. M. Tse, D. L. Phillips, 2017, 7, 4062.
Y. Zhu, Angew. Chem., Int. Ed. 2021, 60, 8236. [90] F. M. Pesci, M. S. Sokolikova, C. Grotta, P. C. Sherrell, F. Reale,
[59] Z. Wang, Y. Qi, C. Ding, D. Fan, G. Liu, Y. Zhao, C. Li, Chem. Sci. K. Sharda, N. Ni, P. Palczynski, C. Mattevi, ACS Catal. 2017, 7, 4990.
2016, 7, 4391. [91] D. E. Schipper, Z. Zhao, A. P. Leitner, L. Xie, F. Qin, M. K. Alam,
[60] W. Guan, Y. Li, Q. Zhong, H. Liu, J. Chen, H. Hu, K. Lv, J. Gong, Y. Xu, S. Chen, D. Wang, Z. Ren, Z. Wang, J. Bao, K. H. Whitmire, ACS
Z. Kang, M. Cao, Q. Zhang, Nano Lett. 2021, 21, 597. Nano 2017, 11, 4051.
[61] T. Zhou, J. Wang, S. Chen, J. Bai, J. Li, Y. Zhang, L. Li, L. Xia, [92] T. H. Jeon, G. h. Moon, H. Park, W. Choi, Nano Energy 2017, 39, 211.
M. Rahim, Q. Xu, B. Zhou, Appl. Catal., B 2020, 267, 118599. [93] K. H. Ye, H. Li, D. Huang, S. Xiao, W. Qiu, M. Li, Y. Hu, W. Mai, H. Ji,
[62] Z. Liu, K. Deng, J. Hu, L. Li, Angew. Chem., Int. Ed. 2019, S. Yang, Nat. Commun. 2019, 10, 3687.
58, 11497. [94] J. Zhang, H. Ma, Z. Liu, Appl. Catal., B 2017, 201, 84.
[63] S. Corby, L. Francas, S. Selim, M. Sachs, C. Blackman, A. Kafizas, [95] H. Chai, L. Gao, P. Wang, F. Li, G. Hu, J. Jin, Appl. Catal., B 2021,
J. R. Durrant, J. Am. Chem. Soc. 2018, 140, 16168. 305, 121011.
[64] Y. Liu, M. Xia, L. Yao, M. Mensi, D. Ren, M. Grätzel, K. Sivula, [96] C. Gao, T. Wei, Y. Zhang, X. Song, Y. Huan, H. Liu, M. Zhao, J. Yu,
N. Guijarro, Adv. Funct. Mater. 2021, 31, 2010081. X. Chen, Adv. Mater. 2019, 31, 1806596.
[65] B. Li, W. Peng, J. Zhang, J. C. Lian, T. Huang, N. Cheng, Z. Luo, [97] L. Meng, X. Zhou, S. Wang, Y. Zhou, W. Tian, P. Kidkhunthod,
W. Q. Huang, W. Hu, A. Pan, L. Jiang, G. F. Huang, Adv. Funct. S. Tunmee, Y. Tang, R. Long, Y. Xin, L. Li, Angew. Chem., Int. Ed.
Mater. 2021, 31, 2100816. 2019, 58, 16668.
[66] D. Cardenas-Morcoso, A. Bou, S. Ravishankar, M. García-Tecedor, [98] L. Cai, J. Zhao, H. Li, J. Park, I. S. Cho, H. S. Han, X. Zheng, ACS
S. Gimenez, J. Bisquert, ACS Energy Lett. 2019, 5, 187. Energy Lett. 2016, 1, 624.
[67] X. Huai, L. Girardi, R. Lu, S. Gao, Y. Zhao, Y. Ling, G. A. Rizzi, [99] S. Cao, Y. Wu, J. Hou, B. Zhang, Z. Li, X. Nie, L. Sun, Adv. Energy
G. Granozzi, Z. Zhang, Nano Energy 2019, 65, 104020. Mater. 2019, 10, 1902935.
[68] M. Antuch, P. Millet, A. Iwase, A. Kudo, Appl. Catal., B 2018, 237, 401. [100] J. Hou, S. Cao, Y. Sun, Y. Wu, F. Liang, Z. Lin, L. Sun, Adv. Energy
[69] G. Liu, S. Ye, P. Yan, F. Xiong, P. Fu, Z. Wang, Z. Chen, J. Shi, C. Li, Mater. 2018, 8, 1701114.
Energy Environ. Sci. 2016, 9, 1327. [101] L. Meng, M. Wang, H. Sun, W. Tian, C. Xiao, S. Wu, F. Cao, L. Li,
[70] L. Meng, J. He, X. Zhou, K. Deng, W. Xu, P. Kidkhunthod, R. Long, Adv. Mater. 2020, 32, 2002893.
Y. Tang, L. Li, Nat. Commun. 2021, 12, 5247. [102] A. J. Bard, J. Photochem. 1979, 10, 59.
[103] H. Tada, T. Mitsui, T. Kiyonaga, T. Akita, K. Tanaka, Nat. Mater. [137] S. Zhang, Z. Liu, M. Ruan, Z. Guo, L. E, W. Zhao, D. Zhao, X. Wu,
2006, 5, 782. D. Chen, Appl. Catal., B 2020, 262, 118279.
[104] J. Yu, S. Wang, J. Low, W. Xiao, Phys. Chem. Chem. Phys. 2013, 15, [138] Y. Chen, L. Wang, R. Gao, Y. Zhang, L. Pan, C. Huang, K. Liu,
16883. X. Chang, X. Zhang, J. Zou, Appl. Catal., B 2019, 259,
[105] X. Chen, J. Wang, Y. Chai, Z. Zhang, Y. Zhu, Adv. Mater. 2021, 33, 118079.
2007479. [139] S. Xu, L. Guo, Q. Sun, Z. L. Wang, Adv. Funct. Mater. 2019, 29,
[106] Y. Fu, C. Dong, W. Zhou, Y. Lu, Y. Huang, Y. Liu, P. Guo, L. Zhao, 1808737.
W. Chou, S. Shen, Appl. Catal., B 2020, 260, 118206. [140] L. Xiao, X. Xu, Y. Jia, G. Hu, J. Hu, B. Yuan, Y. Yu, G. Zou, Nat.
[107] S. Hou, X. Dai, Y. Li, M. Huang, T. Li, Z. Wei, Y. He, G. Xiao, F. Xiao, Commun. 2021, 12, 318.
J. Mater. Chem. A 2019, 7, 22487. [141] C. Wang, N. Tian, T. Ma, Y. Zhang, H. Huang, Nano Energy 2020, 78,
[108] F. Liu, R. Shi, Z. Wang, Y. Weng, C. M. Che, Y. Chen, Angew. Chem., 105371.
Int. Ed. 2019, 58, 11791. [142] S. Zhang, B. Zhang, D. Chen, Z. Guo, M. Ruan, Z. Liu, Nano Energy
[109] W. Xu, W. Tian, L. Meng, F. Cao, L. Li, Adv. Energy Mater. 2021, 11, 2021, 79, 105485.
2003500. [143] H. You, Y. Jia, Z. Wu, F. Wang, H. Huang, Y. Wang, Nat. Commun.
[110] L. Wang, X. Zheng, L. Chen, Y. Xiong, H. Xu, Angew. Chem. Int. Ed. 2018, 9, 2889.
2018, 57, 3454. [144] M. Zhang, Q. Hu, K. Ma, Y. Ding, C. Li, Nano Energy 2020, 73,
[111] J. Low, B. Dai, T. Tong, C. Jiang, J. Yu, Adv. Mater. 2019, 31, 1802981. 104810.
[112] X. Wang, X. Wang, J. Huang, S. Li, A. Meng, Z. Li, Nat. Commun. [145] B. Dai, J. Fang, Y. Yu, M. Sun, H. Huang, C. Lu, J. Kou, Y. Zhao,
2021, 12, 4112. Z. Xu, Adv. Mater. 2020, 32, 1906361.
[113] S. Wei, C. Wang, X. Long, T. Wang, P. Wang, M. Zhang, S. Li, J. Ma, [146] M. Li, J. Sun, G. Chen, S. Yao, B. Cong, P. Liu, Appl. Catal., B 2021,
J. Jin, L. Wu, Nanoscale 2020, 12, 15193. 298, 120573.
[114] L. Meng, D. Rao, W. Tian, F. Cao, X. Yan, L. Li, Angew. Chem., Int. Ed. [147] X. Xu, L. Xiao, Y. Jia, Z. Wu, F. Wang, Y. Wang, N. O. Haugen,
2018, 57, 16882. H. Huang, Energy Environ. Sci. 2018, 11, 2198.
[115] W. Qiu, S. Xiao, J. Ke, Z. Wang, S. Tang, K. Zhang, W. Qian, [148] L. Yang, Y. Xiong, W. Guo, M. Zhou, K. Song, P. Xiao, G. Cao, Nano
Y. Huang, D. Huang, Y. Tong, S. Yang, Angew. Chem., Int. Ed. Energy 2018, 44, 63.
2019, 58, 19087. [149] S. Dunn, P. M. Jones, D. E. Gallardo, J. Am. Chem. Soc. 2007, 129,
[116] A. G. Hufnagel, H. Hajiyani, S. Zhang, T. Li, O. Kasian, B. Gault, 8724.
B. Breitbach, T. Bein, D. Fattakhova-Rohlfing, C. Scheu, [150] H. Wu, H. Ling, Z. Zhang, Y. Li, L. Liang, G. Chai, Acta Phys. Sinica
R. Pentcheva, Adv. Funct. Mater. 2018, 28, 1804472. 2017, 66, 167702.
[117] L. Meng, S. Wang, F. Cao, W. Tian, R. Long, L. Li, Angew. Chem., Int. [151] S. Singh, A. L. Sangle, T. Wu, N. Khare, J. L. MacManus-Driscoll, ACS
Ed. 2019, 58, 6761. Appl. Mater. Interfaces 2019, 11, 45683.
[118] R. Shi, H. F. Ye, F. Liang, Z. Wang, K. Li, Y. Weng, Z. Lin, W. F. Fu, [152] F. Wu, Y. Yu, H. Yang, L. N. German, Z. Li, J. Chen, W. Yang,
C. M. Che, Y. Chen, Adv. Mater. 2018, 30, 1705941. L. Huang, W. Shi, L. Wang, X. Wang, Adv. Mater. 2017, 29, 1701432.
[119] J. Zheng, Y. Lyu, C. Xie, R. Wang, L. Tao, H. Wu, H. Zhou, S. Jiang, [153] M. Zhang, F. Li, D. Benetti, R. Nechache, Q. Wei, X. Qi, F. Rosei,
S. Wang, Adv. Mater. 2018, 30, 1801773. Nano Energy 2021, 81, 105626.
[120] K. Zhang, G. Zhang, J. Qu, H. Liu, Small 2018, 14, 1802760. [154] W. Zhao, Q. Zhang, H. Wang, J. Rong, L. E, Y. Dai, Nano Energy
[121] Y. Xiao, C. Feng, J. Fu, F. Wang, C. Li, V. F. Kunzelmann, C. Jiang, 2020, 73, 104783.
M. Nakabayashi, N. Shibata, I. D. Sharp, K. Domen, Y. Li, Nat. Catal. [155] Z. Liu, L. Wang, X. Yu, J. Zhang, R. Yang, X. Zhang, Y. Ji, M. Wu,
2020, 3, 932. L. Deng, L. Li, Z. L. Wang, Adv. Funct. Mater. 2019, 29,
[122] Y. Yu, Y. Huang, Y. Yu, Y. Shi, B. Zhang, Nano Energy 2018, 43, 236. 1807279.
[123] Z. Luo, C. Li, S. Liu, T. Wang, J. Gong, Chem. Sci. 2017, 8, 91. [156] X. Sun, D. Tiwari, D. J. Fermin, Adv. Energy Mater. 2020, 10, 2002784.
[124] Z. Zhang, I. Karimata, H. Nagashima, S. Muto, K. Ohara, [157] Y. Liang, T. Tsubota, L. P. A. Mooij, R. van de Krol, J. Phys. Chem. C
K. Sugimoto, T. Tachikawa, Nat. Commun. 2019, 10, 4832. 2011, 115, 17594.
[125] Y. Liu, C. Xiao, Z. Li, Y. Xie, Adv. Energy Mater. 2016, 6, 1600436. [158] Y. L. Chen, Y. H. Chen, J. W. Chen, F. Cao, L. Li, Z. M. Luo, I. C. Leu,
[126] Q. He, Y. Wan, H. Jiang, Z. Pan, C. Wu, M. Wang, X. Wu, B. Ye, Y. C. Pu, ACS Appl. Mater. Interfaces 2019, 11, 8126.
P. M. Ajayan, L. Song, ACS Energy Lett. 2018, 3, 1373. [159] S. Cao, X. Yan, Z. Kang, Q. Liang, X. Liao, Y. Zhang, Nano Energy
[127] J. Zheng, Y. Lyu, R. Wang, C. Xie, H. Zhou, S. P. Jiang, S. Wang, Nat. 2016, 24, 25.
Commun. 2018, 9, 3572. [160] P. Wang, F. Li, X. Long, T. Wang, H. Chai, H. Yang, S. Li, J. Ma, J. Jin,
[128] S. Zhang, D. Chen, Z. Liu, M. Ruan, Z. Guo, Appl. Catal., B 2020, Nanoscale 2021, 13, 14197.
284, 119686. [161] Z. Luo, B. Liu, H. Li, X. Chang, W. Zhu, T. Wang, J. Gong, Small
[129] Y. Zhang, Y. Huang, S. S. Zhu, Y. Y. Liu, X. Zhang, J. J. Wang, Methods 2019, 3, 1900212.
A. Braun, Small 2021, 17, 2100320. [162] K. Zhang, B. Jin, C. Park, Y. Cho, X. Song, X. Shi, S. Zhang, W. Kim,
[130] X. Hu, J. Huang, F. Zhao, P. Yi, B. He, Y. Wang, T. Chen, Y. Chen, H. Zeng, J. H. Park, Nat. Commun. 2019, 10, 2001.
Z. Li, X. Liu, J. Mater. Chem. A 2020, 8, 14915. [163] M. Zhong, T. Hisatomi, Y. Kuang, J. Zhao, M. Liu, A. Iwase, Q. Jia,
[131] B. He, S. Jia, M. Zhao, Y. Wang, T. Chen, S. Zhao, Z. Li, Z. Lin, H. Nishiyama, T. Minegishi, M. Nakabayashi, N. Shibata,
Y. Zhao, X. Liu, Adv. Mater. 2021, 33, 2004406. R. Niishiro, C. Katayama, H. Shibano, M. Katayama, A. Kudo,
[132] Y. Zhang, Y. Wang, H. Jiang, M. Huang, Small 2020, 16, 2002550. T. Yamada, K. Domen, J. Am. Chem. Soc. 2015, 137, 5053.
[133] B. Jin, Y. Li, J. Wang, F. Meng, S. Cao, B. He, S. Jia, Y. Wang, Z. Li, [164] H. Chai, P. Wang, T. Wang, L. Gao, F. Li, J. Jin, ACS Appl. Mater.
X. Liu, Small 2019, 15, 1903847. Interfaces 2021, 13, 47572.
[134] J. Wu, W. Wang, Y. Tian, C. Song, H. Qiu, H. Xue, Nano Energy 2020, [165] T. Wang, X. Long, S. Wei, P. Wang, C. Wang, J. Jin, G. Hu, ACS Appl.
77, 105122. Mater. Interfaces 2020, 12, 49705.
[135] M. B. Starr, X. Wang, Nano Energy 2015, 14, 296. [166] R. Tang, L. Wang, M. Ying, W. Yang, A. Kheradmand, Y. Jiang, Z. Li,
[136] Z. Liu, X. Yu, L. Li, Chinese J. Catal. 2020, 41, 534. Y. Cui, R. Zheng, J. Huang, Small Sci. 2021, 1, 2000033.
[167] X. Ning, B. Lu, Z. Zhang, P. Du, H. Ren, D. Shan, J. Chen, Y. Gao, [169] Y. Gong, J. Lin, X. Wang, G. Shi, S. Lei, Z. Lin, X. Zou, G. Ye, R. Vajtai,
X. Lu, Angew. Chem., Int. Ed. 2019, 58, 16800. B. I. Yakobson, H. Terrones, M. Terrones, B. K. Tay, J. Lou,
[168] J. H. Kim, I. Y. Choi, J. H. Kim, J. Kim, Y. K. Kim, J. K. Kim, J. S. Lee, S. T. Pantelides, Z. Liu, W. Zhou, P. M. Ajayan, Nat. Mater.
Small 2021, 17, 2103861. 2014, 13, 1135.
Shengnan Li received her bachelor's degree in physics from Jiangsu Normal University in 2020 and is
currently studying for a master's degree at Soochow University. Her research focuses on the design and
synthesis of photoanodes for PEC water splitting cells.
Wei Tian received his Ph.D. degree from Waseda University, Japan, in 2014. Afterward he joined the
National Institute for Materials Science (NIMS), Japan, as a postdoctoral researcher. Currently, he is a
professor in Soochow University, China. His research interests include the controlled fabrication and
characterization of low-dimensional semiconducting nanostructures and their application in
optoelectronics devices.
Liang Li received his Ph.D. degree from Institute of Solid State Physics, Chinese Academy of Sciences in
2006. From 2007 to 2012, he worked in National University of Singapore (NUS), Singapore, National
Institute of Advanced Industrial Science and Technology (AIST), Japan, National Institute of Materials
Science (NIMS), Japan, and the University of Western Ontario (UWO), Canada. Since August 2012, he is
a full professor in Soochow University, China. His research group (http://ecs.suda.edu.cn) focuses mainly
on energy conversion materials for solar cells, photodetectors, and electrochemical batteries.

Small Science - 2022 - Li

Uploaded by

Copyright:

Available Formats

Small Science - 2022 - Li

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Small Science - 2022 - Li

Uploaded by

Copyright:

Available Formats

REVIEW

Recent Progress on Semiconductor Heterojunction-Based

as aviation fuel was put forward, and it

Table 1. Summary of PEC performance for heterojunctions with micro–nanostructures.

Table 4. Summary of PEC performance for heterojunctions with intermediate layers.

Sample Intermediate layer J (1.23 VRHE) IPCE Testing condition Ref.

You might also like