A heterojunction strategy to improve the visible light sensitive water splitting performance of photocatalytic materials

Khurshida Afroz a, Md Moniruddin a, Nurlan Bakranov a, Sarkyt Kudaibergenov bc and Nurxat Nuraje *a
aDepartment of Chemical Engineering, Texas Tech University, Lubbock, TX79409, USA. E-mail: nurxat.nuraje@ttu.edu
bDepartment of Physics and Engineering Profile, Kazakh National Technical University, Almaty, Kazakhstan
cInstitute of Polymer Materials and Technology, Kazakhstan

Received 5th May 2018 , Accepted 28th June 2018

First published on 28th June 2018


Photocatalytic water splitting is a promising path for generating hydrogen which will decrease the dependency on conventional fossil fuels to generate power and provides an environmentally benign way to store solar energy. Designing photocatalysts is one of the key challenges in making photocatalytic water splitting efficient and economically viable. Heterojunction formation using different functional materials in a single photo-catalyst has the potential of broadening light harvesting properties, improving chemical stability, enhancing the photoexcited charge separation, and thus boosting water splitting efficiency. This article gives an overview on the heterojunction strategy for improving photocatalytic water splitting performance. Recent developments in different visible light-responsive heterojunction systems are discussed for metal oxide–metal oxide and metal oxide–non-metal oxide heterojunctions. Along with experimental observation and the proposed charge separation mechanism, synthesis techniques of heterojunction materials are also included in this article. In addition to recent progress in the heterojunction-based photocatalytic system, this review article also provides future directions for photocatalytic water splitting.


image file: c8ta04165b-p1.tif

Nurxat Nuraje

Nurxat Nuraje received his PhD from CUNY in 2008. He was a postdoctoral associate at MIT from 2008 to 2013 and a research scientist at MIT from 2013 to 2015. Then, he joined Texas Tech University as an Assistant Professor in January of 2015. For his PhD research, he has been awarded the ‘‘Graduate Student Silver Medal’’ by the MRS and the ‘‘Rose Kfar Rose’’ award by CUNY. In 2015, because of his outstanding research, he has been selected as the winner of the “Joseph Wang Award” by Cognizure Publishing. He was the recipient of both 2017 and 2018 US Airforce Faculty Summer Fellowship and recognized as the most influential faculty member by the college of engineering of Texas Tech University. Dr Nuraje's research interests are primarily in nanomaterials and biomaterials and their applications in renewable energy.


1 Introduction

Researchers have meticulously researched solar energy conversion and storage processes to efficiently utilize the tremendous amount (4.3 × 1020 J h−1) of energy hitting the earth's surface as sunlight.1 It is stated that the amount of solar energy reaching the earth's surface in one hour is sufficient to meet the global annual energy consumption (4.1 × 1020 J).2 Solar energy can be converted directly into electricity or stored in as a chemical bond.3 Photocatalytic water splitting is a promising, ecofriendly way to store untapped solar energy via chemical bonding as a form of hydrogen fuel. Hydrogen is a green energy source when it is used in fuel cells for power generation because water is the only byproduct.4 In addition, hydrogen also offers high energy density and it can be easily stored and transported using existing technology.5

In the photocatalytic water splitting process, a photoactive material is used in aqueous solution to harness the photonic energy from light, which then drives the thermodynamically unfavorable water splitting reaction.3 Researchers have devoted their efforts to exploring different photoactive materials which meet kinetic and thermodynamic criteria for efficient photo-conversion of solar energy through water splitting.6 An efficient photocatalyst will have characteristics including a suitable band gap, favorable conduction band (CB) and valence band (VB) edge positions, high photo-corrosion resistivity, and a suitable structure for charge transfer.7 Many research studies have been conducted on photocatalytic water splitting using a single photocatalyst material.3 However, most single photocatalyst materials are either UV-active or suffer from significant photo-corrosion. Photoexcited charge recombination is another major drawback of the single photocatalyst based water splitting system.8 Forming heterostructures with different functional materials in a single photocatalyst shows great promise in overcoming the single photocatalyst material drawbacks. The heterojunction approach allows different favorable properties from each participating compound to be combined, extending the absorption range of the visible spectrum,9 reducing photoexcited electron hole recombination,10 and increasing the photo-corrosion stability, thus improving water splitting efficiency.11

Design strategies and the fabrication of heterostructures with different functional materials are the key challenges to achieving the desired characteristics for a water splitting photocatalyst. In designing a heterojunction, the energy level of each coupling material needs to have nearly matched overlapping band structures.12 At the heterojunction interface, when photo-generated charges migrate from one material to another material, the oxidation and the reduction ability of these transported holes and electrons decrease respectively.13 Especially in type I heterojunctions (Fig. 6), holes transfer to the less positive valence band of a material while the electrons transfer to the less negative conduction band position.14 As a result, the photoexcited charge separation improves but the redox ability of photoexcited charges decreases after transfer through the heterojunction. In a heterojunction structure for water splitting, the participating materials need to meet strict conditions, such as a narrow band gap that facilitates visible light absorption and proper energy level matching, which facilitates good charge transfer along with improved oxidation and reduction ability.13 The fabrication of heterojunctions with the proper combination of one H2 evolving and one O2 evolving photocatalyst is important for overall water splitting under visible light irradiation.15

There are several excellent review papers on different types of photocatalysts for particulate and photoelectrochemical water splitting systems. We refer readers to early published tutorial review articles.16,17 These review articles provide very nice explanations on water splitting spanning the fundamentals to specific material applications. Some review articles focus on specific metal oxide-based nanomaterials such as TiO2,18 Fe2O3,19 or general metal oxide-based systems.20,21 Also several review articles22–24 nicely summarize non-metal oxide-based photocatalytic systems. Particulate photocatalytic systems4,25–27 are summarized by Chen,4 Kudo,16 and Moniz.17 Moniruddin et al.28 discussed perovskite-based material applications in water splitting. There are also some review articles on Z scheme-based14,29 water splitting. Some review articles focus solely on the photoelectrochemical-based30–32 water splitting system. Hisatomi et al.33 discussed the catalytic and kinetic aspects of photocatalytic water splitting. In this review article, we summarize recent developments in heterojunction visible light-active photocatalytic systems based on our literature search of research progress of the past 10 years since the heterojunction system is a promising strategy to improve overall water splitting efficiency. The heterojunction strategy not only improves light harnessing capability but also enhances charge separation, which are both considered critical issues for overall water splitting efficiency. Recently Co–P,34 a heterojunction photocatalyst, showed the highest efficiency (STH efficiency 5.4%) in overall water splitting, which is the champion photocatalyst so far for a heterojunction-based particulate system.

This review article starts with the fundamentals of various heterojunction strategies applied in water splitting. Recent progress of different visible light-active heterojunction photocatalysts, including metal oxide–metal oxide and metal oxide–non-metal oxide groups, in photocatalytic water splitting systems is discussed. In the Metal oxide–metal oxide section, metal oxy nitrides along with metal oxides are included. Metal sulfides and phosphides are included in the non-metal oxide group. The synthesis of heterojunctions is an important step in investigating their performance in water splitting. The synthesis of heterojunction materials is included in a separate section. Recently reported visible light sensitive heterojunction systems along with their corresponding synthesis method, hydrogen/oxygen evolution rate, apparent quantum yield, current density, incident photon to current conversion efficiency and measurement conditions are summarized and tabulated [Tables 1 and 2] in this article. Especially, Table 1 provides a recent update on the photocatalytic performances of particulate-based heterojunction photocatalysts, whereas Table 2 summarizes the performances of electrode-based heterojunction photocatalysts for water splitting. Finally, the Summary and outlook section provides some future directions to improve the performances of heterojunction photocatalysts.

Table 1 Photocatalytic performance of heterojunction particles
Material Synthesis method Hydrogen evolution Oxygen evolution Apparent quantum yield Sacrificial agent Stability Ref.
ZnIn2S4/g-C3N4 g-C3N4 by thermal polymerization, ZnIn2S4 hydrothermal deposition 14.1 μmol h−1 with 15 wt% ZnIn2S4 Triethanolamine-mine (TEOA) Stable for a 24 h run 106
Cu2ZnSnS4–Pt/ Cu2ZnSnS4–Au Solvothermal method 1.02 mmol h−1 g1 catalyst 0.1 M Na2S and Na2SO3 Stable within a 12 h run 107
NiO/NaTaO3 doped with lanthanum Solid state reaction 19.5 mmol h−1 9.7 mmol h−1 56% at 270 nm Stable for more than 400 h 108
g-C3N4/InVO4 Hydrothermal process 212 μmol h−1 g−1 with 20% InVO4 4.9% at 420 nm Methanol Stable during a 200 h run 13
MsTa2O6XNY/TaON 1.25 mol h−1 6.8% at 420 nm 20 vol% CH3OH buffered with 0.15 g La2O3 109
In0.9Ni0.1TaO4 with NiOy and with RuO2 Solid state reaction 16.6 μmol h−1 with NiOy and 8.7 μmol h−1 with RuO2 8.3 μmol h−1 with NiOy and 4.3 μmol h−1 with RuO2 0.66% at 420 nm Pure water Stable during a 400 h run 71
ZnRh2O4/Ag/s-BiV2O11 Melting-slow cooling to row s-BiVO11, ZnRh2O4 by a solid state reaction 0.02 μmol h−1 under 545 nm irradiation 0.01 μmol h−1 under 545 nm 29 irradiation-n 0.036% at 500 nm Pure water Stable during a 120 h run 80
ZnRh2O4/Ag/Ag1−XSbO3−Y Solid state reaction 0.1682 μmol after 24 h 0.0841 μmol after 24 h Pure water Stable during a 144 h run 90
Ta2O5/g-C3N4, Pt cocatalyst Solid state reaction method 36.4 μmol h−1 g−1 with 7.5% TO/CN 20 vol% methanol Stable during a 20 h run 92
ZnO/TiO2 nanotube TiO2 by anodizing titanium foil, ZnO by chemical bath deposition 44 μL cm−2 240 min−1 1 M NaOH (pH = 13.6) 110
NiS2/CdLa2S4 Hydrothermal method 2.5 mmol h−1 g−1 with 2 wt% NiS2 1.6% at 420 nm 0.35 M Na2S and 0.25 M Na2SO3 Constant activity in an up to 20 h run 89
Gd2Ti2O7/In2O3 Solid state reaction 869.7 μmol in 3 h 10 vol% methanol Stable during a 9 h run 63
g-C3N4/CdS quantum dot with Pt co catalyst Heterojunction by the chemical impregnation method 17.27 μmol h−1 with 30 wt% CdS QD 25 vol% methanol No decrease in activity during a 28 h run 104
CaFe2O4/TiO2 TiO2 nanosphere by the solvothermal method, CaFe2O4 by the polymerizable complex method, heterojunction by solid state dispersion 2100 μmol h−1 g−1 in 1 wt% CaFe2O4 with 10 wt% methanol (solar light), for pure water 92 μmol h−1 g−1 with 2 wt% CFTO Both for 10 wt% methanol and pure water Stable during a 25 h run 111
Fe2O3/TiO2 Sol–gel and calcination route using Papilio paris butterfly wings as the template 217.6 μmol h−1 with 2 wt% TiO2 0.94% at 447 nm 0.35 M Na2S/0.25 M Na2SO3 Stable during a 7 h run 112
ZnO/ZnS/g-C3N4 g-C3N4 by direct solid state thermal decomposition, ZnO by chemical precipitation, ZnO/ZnS by adding ZnO on Na2S and then heating and drying 1205 μmol g−1 H2 at 4 h 0.25 M Na2S/0.25 M Na2SO3 105
N doped TiO2/SrTiO3 Sol–gel method 136 μmol g−1 h−1 25% methanol 73
SrTiO3[thin space (1/6-em)]:[thin space (1/6-em)]La, Rh/Au/BiVO4[thin space (1/6-em)]:[thin space (1/6-em)]Mo particulate sheet Particle transfer method. SrTiO3[thin space (1/6-em)]:[thin space (1/6-em)]La, Rh and BiVO4[thin space (1/6-em)]:[thin space (1/6-em)]Mo were embedded on a gold layer 80 μmol cm−2 at 331 K and 10 kPa after 4 h 40 μmol cm−2 after 4 h at 331 K and 10 kPa 33% at 419 nm at 331 K temperature and 10 kPa pressure Pure water Without surface modification water splitting decreased with time, and after surface modification with Cr2O3 the same activity was maintained for 10 h 75
SrTiO3[thin space (1/6-em)]:[thin space (1/6-em)]La, Rh/C/BiVO4[thin space (1/6-em)]:[thin space (1/6-em)]Mo particulate sheet SrTiO3[thin space (1/6-em)]:[thin space (1/6-em)]La, Rh and BiVO4[thin space (1/6-em)]:[thin space (1/6-em)]Mo were produced by a solid state reaction and then SrTiO3[thin space (1/6-em)]:[thin space (1/6-em)]La, Rh/C/BiVO4[thin space (1/6-em)]:[thin space (1/6-em)]Mo by the particle transfer method 8.6 μmol h−1 cm−2 at 288 K and 5 kPa 4.4 μmol h−1 cm−2 at 288 K and 5 kPa 26% at 419 nm Pure water High activity for at least 6 h 76
NiS/C3N4 NiS was deposited on C3N4 by the hydrothermal method 48.2 μmol h−1 1.9% at 440 nm 15 vol% triethanolamine H2 evolution decreased gradually over time 103
NiS nanoparticle/CdS nanorod CdS by the solvothermal method and NiS by the hydrothermal method 1131 μmol h−1 g−1 6.9% at 420 nm 0.35 M Na2S and 0.25 M Na2SO3 No decrease in activity after 9 h 113
Co/P Solvothermal method 131.6 μmol h−1 at 353 K 42.55% at 430 nm and at 353 K Pure water No decrease in activity after a 40 h run 34


Table 2 Photocatalytic performance of heterojunction films
Material Synthesis method Hydrogen or oxygen evolution Apparent quantum yield/IPCE/photo-conversion efficiency System Current density Stability Ref.
WO3/BiVO4 Solvothermal deposition of WO3 nanorods on glass, and then spin coating of BiVO4 Nanorod WO3/BiVO4 showed 31% IPCE where planar WO3/BiVO4 showed 9.3% IPCE at 420 nm 0.5 M sodium sulfate solution 0.8 mA cm−2 at +1 V for the planar heterojunction and 1.6 mA cm−2 for the nanorod Nanorod heterojunction more stable than the planar WO3/BiVO4 film 48
CuO/CdS/TiO2/Pt CuO by electrodeposition, CdS by chemical bath deposition, TiO2 by atomic layer deposition, Pt by photoelectron deposition 100% faradaic efficiency (hydrogen) Less than 5% IPCE at a 500 nm wavelength 1 M phosphate-ate buffer (K2HPO4/KH2PO4) 1.68 mA cm−2 at 0 V (RHE) Highly stable over a 30 minute run 3
WO3/W·BiVO4 nanowire Flame vapor deposition and drop casting 79% faradaic efficiency with 0.5 M Na2SO4 (oxygen) IPCE 60% at 300–450 nm 0.5 M potassium phosphate electrolyte 3.1 mA cm−2 at 1.23 V Stable within a 1 h run 69
α Fe2O3/graphene/BiV1−xMoxO4 α Fe2O3 by the hydrothermal method, GO synthesized by chemical exfoliation deposited by spin coating, BiV1−xMoxO4 by spin coating 95% faradaic efficiency (oxygen) Photo-conversion efficiency 0.53% a −0.4 V vs. Ag/AgCl 0.01 M Na2SO4 1.97 mA cm−2 at 1 V vs. Ag/AgCl Stable during a 12 h run 10
ZnO/TiO2/CuO tree-like structure ZnO electrochemical deposition, TiO2 and CuO by the hydrothermal method 0.8 μmol cm−2 h−1 (hydrogen) 0.1 M KH2PO 1.4 mA cm−2 at 0.3 V vs. Ag/AgCl Stable during a 4 h run 114
Se/BiVO4 Se deposition galvanostatic-ally, BiVO4 by chemical vapor deposition 0.5 M Na2SO4 2.2 mA cm−2 at 1.3 V Stable current density in an up to 40 minute run 56
Rutile TiO2/α-Fe2O3 α-Fe2O3 by the solvothermal method, rutile TiO2 by the hydrothermal method 1 M NaOH (pH = 13.6) 0.3 mA cm−2 at 0.4 V 65
BiVO4/WO3/SnO2 Sol–gel process, BiVO4 by spin coating Internal quantum yield 80% 0.5 M phosphate buffer solution with H2O2 as a hole scavenger 3.1 mA cm−2 at 1.23 V (RHE) Stable for a 20 minute run 64
g-C3N4/WS2 Gas–solid reaction method 101 μmol h−1 g−1 with 0.01 wt% WS2 from 25 vol% methanol solution (hydrogen) 0.5 M Na2SO4 for the film 0.11 μA cm−2 (film) No decrease in activity during a 9 h run 115
WO3/BiVO4 Combination of glancing angle deposition (GLAD) of WO3 and electrochemical deposition of BiVO4 102 μmol h−1 cm−2 (1 sun, 25 °C), 281 μmol h−1 cm−2 (3 sun, 50 °C) for hydrogen; 51 μmol h−1 cm−2 (1 sun, 25 °C), 140.5 μmol h−1 cm−2 (3 sun, 50 °C) for oxygen IPCE more than 90% across 300–500 nm wavelengths pH 7 6.72 mA cm−2 under 1 sun at 1.23 V (RHE) Stable 68
GaTe/ZnO Two step chemical vapor deposition 1.5 μmol h−1 cm−2 (hydrogen) 0.1 mol L−1 Na2SO4 (pH = 7) −2.5 mA cm−2 at −0.39 V vs. RHE Constant activity for a 50 minute run 57
Si/ZnO nanowire (NW) Metal assisted chemical etching to grow Si NW and hydrothermal growth of ZnO NW 0.25 M Na2SO4 buffered with PBS with pH = 7.2 8 mA cm−2 at −1.5 V vs. Ag/AgCl RE Stable for a 200 second run 9
Ag3PO4/TiO2 TiO2 array by anodizing Ti foil in ethylene glycol and Ag3PO4 was deposited by the sequential chemical bath deposition method 1 M KOH solution 2.34 mA cm−2 at 0 V 84
CdS/Au/TiO2 TiO2 on FTO by the hydrothermal method, Au by photo-reduction, CdS quantum dot by chemical bath deposition IPCE 85% at 375 nm. Photoconversion efficiency 2.8% at −0.56 V vs. Ag/AgCl 0.25 M Na2S and 0.35 M Na2SO3 (pH 12) 4.07 mA cm−2 at 0 V (Ag/AgCl) 98
Pt/In2S3/CdS/Cu2ZnSnS4 Cu2ZnSnS4 by electrodeposition of Cu, Sn, and Zn followed by sulfurization, CdS and In2S3 by chemical bath deposition, Pt by photoelectro-deposition 0.477 μmol min−1 H2 at 0 V for 20 minutes. 96% faradaic efficiency for hydrogen evolution IPCE 45–50% from 400 to 700 nm 0.2 mol dm−3 Na2HPO4/NaH2PO4 −9.3 mA cm−2 at 0 V 99
WO3/Bi2S3 WO3 on FTO by the hydrothermal process, Bi2S3 seed layer by the SILAR method, Bi2S3 by chemical bath deposition IPCE 68.8% and photo-conversion efficiency of 1.7% 0.1 M Na2S and 0.1 M Na2SO3 (pH 12) 5.95 mA cm−2 at 0.9 VRHE Photocurrent density decreased slowly during a 600 s stability test 100
ZnO/Fe2O3 nanosheet ZnO by electrochemical fabrication, Fe2O3 by spin coating 0.01 M KCl solution 1.5 μA cm−2 66
ZnO/ZnS/CdS/CuInS2 Ion exchange and hydrothermal process IPCE 57.7% at 480 nm 0.5 mol L−1 Na2S and 0.5 mol L−1 Na2SO3 10.5 mA cm−2 at 0 V versus Ag/AgCl 93
CaFe2O4/TaON Both materials were deposited by electrophoretic deposition No H2 or O2 evolution without co-catalysts IPCE 30% at 400 nm 0.5 M NaOH solution 1.26 mA cm−2 at 1.23 V vs. RHE Photocurrent density decreased to half after 3 h 5
p-GaInP2/TiO2/(OOCpy)Co(dmgH)2-(Cl)/TiO2 p-GaInP2 was grown on a GaAs substrate by an atmospheric pressure organometallic vapor phase epitaxy process. Electron beam deposition of a Ti layer, and then dipping the electrode in 5 × 10−3 M (HOOCpy)Co(dmgH)2(Cl) ethanol solution for 24 h, and then atomic layer deposition of a TiO2 layer Faradaic efficiency close to 100% for hydrogen evolution. Turn over number (TON) was over 139000 and turn over frequency (TOF) was greater than 1.9 s−1 for 20 h IPCE up to 80% across 470 to 680 nm Argon purged 0.1 M NaOH solution (pH 13) 9 mA cm−2 at 0 V versus RHE During a 20 h run a decrease in activity from 10 to 5 mA cm−2 in the first 4 h and then no change for 16 h 96
ZnO/α-Fe2O3 core–shell nanowire with an Fe2PO5 protective layer ZnO on FTO by the hydrothermal method, wet-chemical route to form an Fe2O3 shell layer, PH3 annealing at 300 °C form an Fe2PO5 layer IPCE 55% at around 400 nm 1 M NaOH Highest 2.3 mA cm−2 at 1.23 V vs. RHE 41.67% decrease from the initial photocurrent for a 10 minute run 67
g-MoSx/MoOx/c-TiO2p GaInP2 p-GaInP2 by metal–organic vapor phase epitaxy, a-TiO2 by atomic layer deposition, MoS2 by electrochemical cathodic deposition, annealing at 450 °C form an intermediate MoOx layer 94% faradaic efficiency for hydrogen evolution after 6 h, after 20 h TON 367000 and TOF 5.7 s−1 were obtained IPCE around 70% across the visible light range Ar purged 0.5 M H2SO4 (pH 0.3) 11 mA cm−2 at 0 V vs. RHE Retention of 80% of the initial photocurrent during a 20 h run 95
WO3/g-C3N4 WO3 on FTO by the hydrothermal method, g-C3N4 by electrophoretic deposition 33% IPCE at 330 nm 0.5 M Na2SO4 electrolyte 0.82 mA cm−2 at 1.23 V vs. RHE from the type II heterojunction 101


1.1 Photocatalytic water splitting principles

Thermodynamically, water splitting is an unfavorable reaction and requires more than 1.23 eV energy to drive the reaction forward.16,35 In an artificial technique, water oxidation is five times slower than the hydrogen evolution process17 because every oxygen molecule requires four holes, making water oxidation more challenging than H2 reduction.36 Photocatalytic water splitting consists of several key steps. In designing photocatalysts for this process, it is of the utmost importance to understand the mechanism of each step and identify the key parameters influencing the process efficacy. Some review articles nicely describe the fundamental steps of water splitting and discuss the key parameters that determine the efficiency of this process.37,38 The major steps involved in the photocatalytic water splitting process are: (a) photon absorption, (b) photoexcited charge separation, (c) charge diffusion and transport, (d) catalytic reaction on the catalyst's active site, and (e) mass transfer (Fig. 1).39 The water splitting reaction initiates with photon absorption via a photocatalyst, where numerous photoexcited electrons and holes are generated in the CB and VB, respectively. These photoexcitation processes occur when the photonic energy is higher or equivalent to the band gap energy of the photocatalyst materials and happen on the femtosecond time scale.40,41 The photoexcited charges relax quickly at their respective band edge positions on the femto- to the pico-second time scale and then transport to the catalytic active site on the nano- to the micro-second time scale to perform the water redox reaction.41 For the hydrogen evolution reaction (HER) and the oxygen evolution reaction (OER) to occur on the catalyst's active sites, the energy level of the CB needs to be more negative than the hydrogen evolution potential (EH2/H2O, 0 V vs. NHE) and the VB energy level has to be more positive than the water oxidation potential (EO2/H2O, 1.23 V vs. NHE), respectively.42 Theoretically, the minimum band gap of a material for water splitting is 1.23 eV. However, in practice relatively high energy is required due to overpotential involved in the water splitting reaction.43 The suitable band alignment and high stability of the photocatalytic material in an aqueous solution are also important for the water splitting reaction. The low activation barrier of the catalyst surface is preferable for the water splitting reaction.44 Sometimes cocatalysts can be added to the photocatalyst materials which catalyze the water redox reaction faster than the photocatalyst itself. The photon absorption, charge separation, and charge transportation significantly depend on the crystal structure as well as electronic configuration of the photocatalytic material.45 The physiochemical properties of a photocatalyst also have a strong effect on the photocatalytic water splitting reaction.26
image file: c8ta04165b-f1.tif
Fig. 1 Schematic of the major steps in the photocatalytic water splitting process. Reprinted with permission from ref. 39.

The photocatalytic performance of a material is determined by the solar to hydrogen (STH) energy conversion for overall water splitting. The STH efficiency is defined as:

 
image file: c8ta04165b-t1.tif(1)
where rH2 is the hydrogen production rate, ΔG is the Gibbs free energy, PSun is the energy flux of sunlight, and AGeometric is the area of the reactor. Theoretical rH2 can be calculated from the photon's number in the solar spectrum at a different quantum efficiency (QE).
 
image file: c8ta04165b-t2.tif(2)

Thus, theoretical STH at a different quantum efficiency can be estimated using eqn (1) and (2).

For sacrificial agent-assisted water splitting, incident photon to current conversion efficiency (IPCE) and photo-conversion efficiency are used which are defined by the following two equations:

 
image file: c8ta04165b-t3.tif(3)
where J is the photocurrent density at 0 V at a certain wavelength (λ), and Ilight represents irradiance intensity at λ
 
image file: c8ta04165b-t4.tif(4)
where jp denotes the photocurrent density at the measured potential, V is the bias potential vs. RHE, and Io is the power density of incident light.

Fig. 2 shows the solar to hydrogen efficiency and photon number as a function of wavelength according to the data from the standard AM 1.5G spectrum.39 The maximum theoretically attainable STH efficiency is 48% at the integrated wavelength from UV to 1000 nm, whereas STH efficiency in the UV region (up to 400 nm) is only 3.3% at a QE of 100%. According to DOE, the targeted baseline of STH efficiency is set to 10% based on the economic feasibility of the existing hydrogen market. At this efficiency, a photoelectrode needs to produce hydrogen at a rate of ∼154 μmol cm−2 h−1 and a photoelectrochemical current of ∼8.3 mA cm−2, which corresponds to a photon consumption rate of ∼260 nm−2 s−1 on a flat surface.46 A large reactor surface area is required at low STH (UV region) to meet these criteria, but a large reactor area increases the cost of hydrogen production. Therefore, designing photocatalysts that can absorb solar light up to the 600–700 nm (∼1.8–2.0 eV) wavelength is the key to achieving the targeted efficiency.


image file: c8ta04165b-f2.tif
Fig. 2 Theoretical STH and photon number of AM 1.5G solar light integrated from a low to the respective wavelength at different QEs (30%, 60% and 100%) as a function of wavelength.39

A large portion of reported photocatalysts are not suitable for the overall water splitting reaction and only a few of them are visible light sensitive.17,47 The photocatalytic performance of the materials is limited due to the quick recombination of photogenerated charges, photo-corrosion, and a lack of appropriate band gap energy for absorbing visible light.48 In Fig. 3, the band positions of different materials are shown. The band gap of the material should be less than 2.8 eV to effectively absorb visible light (λ > 420 nm).49 Band engineering is one of the strategies used to develop the visible light sensitivity of water splitting materials. Low band gap photocatalytic materials are usually more susceptible to photo-corrosion than larger band gap materials, but higher band gap materials are not capable of harnessing visible light.48 Heterojunction based composite materials can be optimized to get appropriate band gap alignment for visible light sensitive water splitting. These heterojunction based photocatalysts are advantageous over single material-based photocatalysts since they harness more visible light than the individual materials. In addition, heterojunction formation with suitable band alignment also facilitates an enhanced charge separation. The heterojunction combines different favorable properties from different materials and can be used to achieve the targeted STH efficiency.


image file: c8ta04165b-f3.tif
Fig. 3 Band position of different photocatalytic materials. Reprinted with permission from ref. 50.

1.2 Photoelectrochemical and particulate heterojunction systems for water splitting

In photoelectrochemical water splitting, photoexcited electron–hole pairs are generated before electrons transfer from the anode to the cathode through an external circuit.51 At the interface of the electrode and electrolyte, electron transfer continues until the Fermi level of the semiconducting electrode is in the same position as the electrolyte redox potential.52 Sometimes an external bias voltage is also applied to overcome the potential difference between the two electrodes.21

Generally, photoelectrochemical (PEC) cells consist of either a single photoelectrode (photoanode or photocathode) along with a counter electrode or two photoelectrodes (photoanode and photocathode).53Fig. 4(a) and (b) present single photoelectrochemical cell systems and Fig. 4(c) depicts a dual photoelectrochemical cell system. In Fig. 4(a), after light illumination the photoexcited electron from the conduction band of a n-type photoanode transfers to the counter electrode through an external circuit for the HER whereas the respective photogenerated hole is consumed in the OER.21 In a p-type photocathode (Fig. 4(b)), the photoexcited electrons are directly utilized to produce hydrogen from the electrolyte and the OER occurs at the counter electrode.


image file: c8ta04165b-f4.tif
Fig. 4 Different types of photoelectrochemical cells, (a) the semiconductor photoanode with the counter electrode, (b) the semiconductor photocathode with the counter electrode, and (c) both the photoanode and the photocathode used in a dual band PEC cell. Reprinted with permission from ref. 21.

For this type of construction, the conduction band position of the electrode needs to be more negative than the water reduction potential. Instead of using a counter electrode, both a photoanode and photocathode are applied to construct a dual band photoelectrochemical cell (Fig. 4(c)).21

Unlike the particulate photocatalytic system, photoelectrochemical water splitting is dependent on efficient light absorption, band gap, and band edge position of semi conductive materials. Heterojunction-based photoelectrodes provide improved charge separation and a broader wavelength of light harvesting with improved stability.52 The application of narrow band gap semiconductors is preferred for designing heterojunction-based PEC cells since they can harness a large portion of the solar spectrum. In the design strategy of heterojunction PEC cells, relative band edge alignment is critical to improving charge transfer for overall water splitting.54 When a photoanode is constructed with an n–n type heterojunction (Fig. 5(a) and (b)), photoexcited electrons go to the counter electrode for water reduction through an external circuit. For this type of design, the conduction band position of the semiconductor attached with the conductive layer needs to be lower than the conduction band position of the semiconductor on the electrolyte side. In the electrolyte side semiconductor, holes accumulate to oxidize water and generate oxygen. However, p–p type heterojunctions (Fig. 5(c)) work as a photocathode when photoexcited electrons transfer to the electrolyte side semiconductor and reduce water to hydrogen. Finally, p–n type heterojunctions can be used as a photoanode (Fig. 5(d)) or photocathode (Fig. 5(e)). In both cases, electrons and holes transfer in opposite directions and better change separation occurs compared to a single semiconductor electrode.


image file: c8ta04165b-f5.tif
Fig. 5 Different types of heterojunctions in photoelectrodes with band alignment and the mechanism of possible charge transfer. Reprinted with permission from ref. 52.

One of the most important advantages of a heterojunction structure is that it allows for preferential band alignment to achieve visible light responsive photocatalytic water splitting.55 Based on band alignment, there are three main types of heterojunction structures between binary dissimilar particulates. In type I heterojunctions, material A has a smaller band gap than material B.56 Holes and electrons transfer from material B to material A due to a more negative CB and a more positive VB of material B.17 An example of a type I heterojunction is shown in Fig. 6, and photo-generated hole and electron flow occurs from TiO2 to Fe2O3.


image file: c8ta04165b-f6.tif
Fig. 6 Band alignment for three different types of heterojunctions. Redrawn with permission from ref. 56.

In type II heterojunctions, the CB position of material B is more negative than that of material A, but material A has a more positive VB position. As a result, holes and electrons transfer in opposite directions. When an electron transfers from material B to A, the hole transfers from A to B. The WO3–BiVO4 heterojunction is one example of this type of system. This kind of charge flow results in effective charge separation and therefore improves photocatalytic activity.17 In this type of band alignment, the Z scheme can also be established when electrons from the CB of material A combine with the holes from the VB of material B. Water oxidation occurs in material A, where photo-excited electrons from material B contribute to hydrogen production. In Fig. 6, BiVO4-g-C3N4 represents this type of Z scheme heterojunction.56

Type III heterojunctions have a similar band structure to type II heterojunctions. The difference between the VB and CB of two materials is larger in type III heterojunctions, however, which increases photocatalytic performance by providing a higher driving force for charge transfer.57 The photogenerated electron from material A combines with the hole of material B, and the electron from material B and the hole from material A participate in the reduction and oxidation of water respectively. The FeS2 and Sb2Se3 system represents this type of heterojunction in Fig. 6.17

In spite of several challenges, various effective, viable, and visible light-responsive sulfide, oxide, and oxynitride photocatalysts have been synthesized for photocatalytic water splitting applications.11

1.3 Metal oxide–metal oxide heterojunction

Inspired by the pioneering work of Fujishima and Honda in 1972, who used titania for photocatalytic water splitting with UV irradiation,58 many materials have been studied to improve photocatalytic water splitting efficiency.59,60 Developing a highly efficient and sturdy photoanode in aqueous solution is one of the main challenges for photoelectrochemical (PEC) water splitting. Good resistance against corrosion, low electric resistance, flat band potential, and suitable band gap of metal oxides make them a potential highly efficient photoanode for PEC cells.61 Photocatalytic materials need to absorb a large portion of visible light to become highly efficient. Lower cost, higher stability, and good visible light sensitivity make simple metal oxides (Fe2O3, Al2O3, Ga2O3, Ta2O5, CoO, Co3O4, WO3, Cu2O, BiVO4, Bi2O3, PbO, SnO, CuO, and ZrO2) an attractive option for photocatalytic water splitting.62 Oxynitrides like TaON and LaTiO2N also perform efficiently as oxidation photocatalysts for water splitting.5 Some binary and ternary metal oxides with complex structures are also suitable photocatalysts for water splitting. Many perovskite-based materials like NaTaO3, SrTiO3, PbTiO3, KTaO3, and CaTa2O6 have shown good performance in water splitting.28 Layered perovskite materials (Sr2Nb2O7, Sr2Ta2O7, and La2Ti2O7) perform better than bulk perovskite materials (SrTiO3 and LaTiO3) as they have inter-layer space that can be used for reaction sites.63 Further improvement of optical and interfacial properties as well as charge separation and transportation is possible by connecting dissimilar materials via a heterojunction.64

Hematite (Fe2O3) has a suitable band gap (2–2.2 eV) for visible light absorption and offers several benefits including high photocurrent density, good aqueous stability and photo-corrosion resistance, naturally abundant, and less expensive than other commonly used photo active materials.65 In spite of these advantages, its limitations include a short excited-state lifetime, low pH instability, low carrier mobility, and a weak absorption coefficient for high utility of Fe2O3.65 To improve its water splitting efficiency, surface modification, nanostructure control, elemental doping, and forming heterojunctions with other photocatalysts with a suitable energy band structure were investigated thoroughly. The heterojunction strategy shows favorable effects to improve photocatalytic water splitting performance.65 Other materials with a lower conduction band position than Fe2O3 will thermodynamically drive photogenerated electrons in space and work well for charge separation when coupled with Fe2O3.65 Unusual electron transfer from the CB of Fe2O3 to the higher position CB of rutile and anatase TiO2 was reported by Luan et al.65 Because of the slightly less negative CB position of rutile TiO2 (−0.05 eV) compared to anatase TiO2 (−0.16 eV), better electron transfer was obtained in the rutile TiO2–Fe2O3 heterojunction. Because of increased charge transfer, under the same operating conditions rutile TiO2–Fe2O3 showed 1.6 times higher photocurrent density than anatase TiO2–Fe2O3. The highest reported photocurrent density for rutile TiO2–Fe2O3 was 0.3 mA cm−2 at 0.4 V vs. the Ag/AgCl electrode from 1 M NaOH aqueous solution. Hou et al. used graphene as a fast electron transfer mediator in the heterojunction of an α-Fe2O3 nanorod/BiV1−XMoXO4 system and compared its photocatalytic activity with and without graphene (Fig. 7). Graphene inserted into the heterojunction showed a 1.4 times higher photocurrent density than the system without graphene. Improved photoelectrochemical properties as a result of window effect was observed within α-Fe2O3 cores and a BiV1−XMoXO4 shell where an effective photo excited carrier separation in α-Fe2O3 nanorod/graphene/BiV1−XMoXO4 interface occurred.10 Nanorod array (NA) geometry showed some special advantages over other various structures. When these NAs were aligned vertically (e.g., TiO2, ZnO, and Fe2O3) they provided a large surface area and contributed to a small diffusion length which was perpendicular to the charge-collecting substrate. This is an effective structure for a smaller amount of loss due to charge recombination.10 Its outstanding charge carrier mobility enables graphene to act as an electron mediator for photo-excited electrons which leads to effective charge separation in this system.10 In the core/shell heterojunction structure of an α-Fe2O3 nanorod array (NA) with graphene sheets, graphene played a role of an electron mediator for the photocatalytic water splitting application.10 Reduced graphene oxide (RGO) sheets developed a bridge-like connection and cover the α-Fe2O3 NA.10 This special structural feature provided an intimate interface connecting the nanorod array and the reduced graphene oxide which played a role in improving the transportation of photo-excited electrons. Both α-Fe2O3 NA and BiV1−XMoXO4 simultaneously absorbed photons which contributes to an increase in light absorption ability. It could be said that the ‘window effect’ was exhibited by the α-Fe2O3-NA/RGO/BiV1−XMoXO4 heterojunction.10 This system showed 0.53% photo-conversion efficiency at −0.04 V and good stability during a photocurrent density test in 0.01 M Na2SO4 solution. However, this system suffered from the low conduction band position of Fe2O3 in the visible light region. An external bias was applied to obtain oxygen evolution. Fe2O3-coated ZnO heterojunction systems also showed visible light photoactivity and high photocurrent density due to improved charge separation. Johnson et al. recently reported the effect of the morphology of ZnO and the polarization field in the ZnO/Fe2O3 system.66 ZnO nanorods and nanosheets were prepared with different precursor concentrations and different electrochemical deposition times. On top of the ZnO structure, Fe2O3 was deposited by spin coating. Along with favorable Fermi level alignment, the spontaneous polarization field built in the ZnO nanosheets contributed to improved charge separation and increased current density.66 The non-centrosymmetric structure of the ZnO nanosheets with a polar direction along the c-axis resulted in an opposite direction fixed induced polarization charge.66 ZnO nanorods did not show this type of polarization effect and showed a photocurrent density 4 times lower than that of the ZnO nanosheets because of quick charge recombination.66 The ZnO/Fe2O3 heterojunction photocatalytic performance can be further improved by applying a Fe2PO5 protective layer. Recently, Qin et al. applied a protective layer of Fe2PO5 using PH3 gas on a ZnO/α-Fe2O3 core–shell nanowire.67 In this system, the Fe2PO5 layer helped to activate the hematite. Thus, the PH3-treated ZnO/α-Fe2O3 electrode showed higher photocurrent density (2.3 mA cm−2 at 1.23 V) than untreated ZnO/α-Fe2O3 (1 mA cm−2 at 1.23 V) by suppressing surface recombination and protecting the ZnO from photo-corrosion. In addition, the average transport time of photoexcited electrons to diffuse to the electrode was reduced from 0.9 ms to 0.38 ms after PH3 treatment, which indicated the significant reduction of charge recombination in the treated electrode. An incident photon to current conversion efficiency (IPCE) of 55% was reported for the PH3-treated system in 0.1 M NaOH solution. The major limitation of this system is ascribed to the low conduction band position of Fe2O3. Another heterojunction system, In2O3/Gd2Ti2O7, demonstrated high photocatalytic activity compared to pure In2O3 or Gd2Ti2O7 under visible light. A 60% Gd2Ti2O7 loaded In2O3/Gd2Ti2O7 sample produced 869.7 μmol H2 from 10 vol% methanol solution after 3 h where pure In2O3 evolved only 98.6 μmol H2 under the same operating conditions.63 The reason for the improved photocatalytic activity was the well-matched band alignment between In2O3 and Gd2Ti2O7 which reduced charge recombination at the In2O3/Gd2Ti2O7 heterojunction interface.63 Although the heterojunction strategy was applied to improve hydrogen evolution over single particle systems (In2O3 and Gd2Ti2O7), this system has limited light absorption due to large band gaps (2.6 and 3.2 eV for In2O3 and Gd2Ti2O7, respectively).


image file: c8ta04165b-f7.tif
Fig. 7 Mechanism for charge transfer through an α-Fe2O3-NA/RGO/BiV1−XMoXO4 heterojunction. Reprinted with permission from ref. 10.

Although BiVO4 has a small band gap (2.4 eV) which facilitates visible light absorption, the high rate of recombination of photoexcited charges has decreased its electron transport properties.68 This high charge recombination rate results from the small diffusion length (Ld around 70 nm), which is ultimately responsible for the poor photocurrent density of BiVO4. Co-catalysts, such as Co, RhO2, and NiOOH/FeOOH, help to improve its performance. Doping with Mo and W helps to develop the transport properties and increase the photocurrent density of BiVO4. Rao et al. reported natural doping of W in a thin BiVO4 shell which facilitated efficient light absorption in a WO3/W–BiVO4 heterojunction photoanode.69 This doping occurred naturally during the annealing process because of the close contact between WO3 and the BiVO4 layers, and the doping concentration decreased gradually from the contact surface to the outer BiVO4 layer. This gradient W doping introduced an electric field, which is the reason for efficient radial electron transfer from the W[thin space (1/6-em)]:[thin space (1/6-em)]BiVO4 shell to the conductive WO3 core. The long length (average length 2.5 μm) of WO3 nanowires provided a higher surface area of the BiVO4 shell which introduced efficient light absorption.69 Among light absorption (ηabs), charge separation (ηsep), and charge transfer (ηtrans) efficiencies, ηsep was especially improved in that system compared to other WO3/W[thin space (1/6-em)]:[thin space (1/6-em)]BiVO4 structures. High ηsep (77%) resulted from a thin (60 nm) W[thin space (1/6-em)]:[thin space (1/6-em)]BiVO4 layer because photogenerated holes needed to travel a shorter path than their diffusion length (70–100 nm).69 For this system, the obtained current density was 3.1 mA cm−2 in 0.5 M potassium phosphate electrolyte solution with 60% IPCE across the 300–450 nm range.69 Triple planar heterojunctions (TPHs) fabricated with BiVO4/WO3/SnO2 were able to improve photocurrent density (∼3.1 mA cm−2 at 1.23 V vs. RHE) in 0.5 M phosphate buffer solution and had an internal quantum efficiency (IQE) of around 80%.64 The bottom layer of WO3/SnO2 helped the efficiency of charge transfer, decreased interfacial resistance, and developed close contact by forming a disordered heterojunction layer. The top BiVO4/WO3 heterojunction layer increased carrier density by improving photon absorption and charge separation. The construction of a SnO2 seed layer over ITO and FTO can improve the attachment of the catalyst layer on the substrate. The catalyst seed layer over ITO or FTO is also effective for providing better contact between the catalyst and the substrate. The interaction between the seed and the catalyst layer also needs to be considered for effective charge transfer and better performance. The structure of the BiVO4/WO3/SnO2 double heterojunction and its performance are shown in Fig. 8. A further improvement in photo-catalytic performance was achieved by CoOx deposition and by performing an etching treatment with a reactive ion.64 The Co3O4/BiVO4 heterojunction system was reported to increase photocurrent density with improved charge separation because of the formation of an internal electric field inside the p–n junction.70 It is possible to drive photogenerated electrons from a small band gap-containing material to a large band gap-containing material by making a couple with large and small band gap semiconductors with a more negative CB position.48 The heterojunction formation of BiVO4, a small band gap photocatalyst, with WO3 (which has a relatively less negative CB and decent corrosion resistivity) showed good photoactivity under visible light irradiation.48 The WO3/BiVO4 nanorod heterojunction showed 1.6 mA cm−2 photocurrent density in 0.5 M sodium sulfate solution with 31% IPCE at a 420 nm wavelength.48 Another strategy was applied to construct a layer of WO3 as an extremely thin layer absorber (ETA), thinner than the diffusion length (Ld) of BiVO4 in the heterojunction of BiVO4/WO3 to enhance the photocatalytic performance. This ETA layer resolved the negative effect of the short Ld of BiVO4 and improved the charge collection probability. Photon absorption was also enhanced by this ETA, because of improved light scattering through the optical path. Pihosh et al. were able to get nearly 90% (6.72 mA cm−2) of maximum theoretical photocurrent (7.5 mA cm−2) density from pure water by tuning the electronic thickness and optical optimization of ETA.68 Inserting a thinner absorber layer than the diffusion length is an effective strategy to improve the performance of materials with a large diffusion length.


image file: c8ta04165b-f8.tif
Fig. 8 (a) Structure of the BiVO4/WO3/SnO2 double heterojunction and (b) its performance in water splitting. Reprinted with permission from ref. 64.

Sometimes doping one material and making a heterojunction with another material improves photocatalytic efficiency. Zou et al. reported a RuO2 heterojunction with In0.9Ni0.1TaO4 as a visible light-sensitive material able to evolve 8.7 mmol h−1 H2 and 4.3 mmol h−1 O2.71 When NiOy was inserted instead of RuO2, an increased hydrogen and oxygen evolution of 16.6 mmol h−1 and 8.3 mmol h−1, respectively, were obtained from pure water.71 The reported apparent quantum yield was 0.66% at 420 nm for this system.71 STH for this system was not reported. Ni-doping narrowed the band gap energy, which increases the photocatalytic activity by facilitating electron excitation from the valence to the conduction band position. No gas was formed when the light was turned off, which indicates that the production of gas was conducted with visible light absorption.71 Under visible light irradiation, metal-doped strontium titanate SrTiO3 (dopant: Ta, Rh or Ni) and metal oxides like RbPb2Nb3O10 and SnNb2O6 can split water with the aid of a sacrificial agent.72 TiO2 and SrTiO3 are both UV-responsive materials. However, doping TiO2 with nitrogen and making a heterojunction with SrTiO3 makes it an efficient solar light responsive photocatalyst.73 Along with particle agglomeration prevention, finer crystals and a larger specific surface area in this heterojunction73 improved reduction by favorably shifting the Fermi level of the catalyst system. Up to 5 wt% SrTiO3, photocatalytic ability increased, but a further increase in SrTiO3 caused the material to behave more like SrTiO3. Since SrTiO3 is only responsive in the UV region, photocatalytic efficiency decreased.73 This system showed 136 μmol g−1 h−1 H2 evolution from 25% methanol solution.73 SrTiO3 can also be modified as a visible light-responsive material by doping with appropriate materials. For example, Rh-doped SrTiO3 reduced its band gap and gave a response in the visible light range.74 Rh and La-doped SrTiO3 and Mo-doped BiVO4 embedded into a gold layer were reported to split pure water under visible light in the presence of a RuCl3 co-catalyst.75 RuCl3 was deposited as a co-catalyst, with an optimum addition efficiency of 0.2 μmol.75 RuCl3 provided more active sites for the reaction, but further loading blocked the light absorption site for the catalyst. Fig. 9 illustrates the synthesis method and the pure water splitting mechanism for the SrTiO3[thin space (1/6-em)]:[thin space (1/6-em)]La, Rh/Au/BiVO4[thin space (1/6-em)]:[thin space (1/6-em)]Mo catalyst. Here, the thick gold layer did not act as a surface plasmonic material, but acted as a conductive medium to facilitate the charge transfer in this system.75 This system decomposed pure water at a 2[thin space (1/6-em)]:[thin space (1/6-em)]1 ratio (H2[thin space (1/6-em)]:[thin space (1/6-em)]O2) under ambient conditions, but the photocatalytic activity of this system drastically dropped at elevated pressure due to the undesirable backward reaction. To address this issue, the SrTiO3[thin space (1/6-em)]:[thin space (1/6-em)]La, Rh/Au/BiVO4[thin space (1/6-em)]:[thin space (1/6-em)]Mo particulate sheet was modified with a Cr2O3 and TiO2 layer that helped to maintain catalyst activity and negate the pressure effect to a certain extent by reducing the backward reaction. After this modification, this catalyst showed a 33% apparent quantum yield at 419 nm and reached 1.1% STH efficiency at an elevated temperature and pressure of 331 K and 10 KPa, respectively. Furthermore, as part of the interface modification, a less expensive carbon film was introduced instead of Au to form SrTiO3[thin space (1/6-em)]:[thin space (1/6-em)]La, Rh/C/BiVO4[thin space (1/6-em)]:[thin space (1/6-em)]Mo particulate photocatalyst sheets.76 This system demonstrated an enhanced STH efficiency of 1.2% at a temperature of 331 K and a pressure of 10 kPa with an AQY of 26% at 419 nm for unassisted pure water splitting. In addition, activated carbon is also less prone to facilitate the oxygen reducing backward reaction than Au and results in higher photocatalytic performances than SrTiO3[thin space (1/6-em)]:[thin space (1/6-em)]La, Rh/Au/BiVO4[thin space (1/6-em)]:[thin space (1/6-em)]Mo particulate sheets. The effects of varying pH, temperature, and pressure on the water splitting activity of the SrTiO3[thin space (1/6-em)]:[thin space (1/6-em)]La, Rh/C/BiVO4[thin space (1/6-em)]:[thin space (1/6-em)]Mo system were investigated.76 Applying a particulate sheet to generate hydrogen from water is a scalable photocatalytic water splitting process. Here, the main challenge is to reduce the reverse reaction at elevated pressure. Inserting a membrane in this type of system to separate evolved hydrogen and oxygen could be an effective way to improve the photocatalytic performances.77 More studies regarding surface modification are required to prevent the backward reaction and make this process scalable.


image file: c8ta04165b-f9.tif
Fig. 9 (a) Synthesis of a SrTiO3[thin space (1/6-em)]:[thin space (1/6-em)]La, Rh/Au/BiVO4[thin space (1/6-em)]:[thin space (1/6-em)]Mo particulate sheet by the particle transfer method. (b) Overall water splitting mechanism with the synthesized material. Reprinted with permission from ref. 75.

Oxynitrides of transition metals and lanthanides such as TaON,78 LaTiO2N, and CaTaO2N79 are good hydrogen evolving photocatalysts under visible light irradiation. Environmentally toxic heavy metals (Pb) and rare and expensive transition metals (Rh, Nb, Ta or La) have limited the practical utility of these oxynitrides and transition metal oxides for being used in large-scale photocatalytic water splitting.72 Kim et al. reported an improvement in photocurrent with a TaON/CaFe2O4 heterojunction.5 Bare TaON showed a very small amount of photocurrent density (4 μA cm−2 at 1.23 V vs. RHE) which was caused by the poor contact of the electrode and particles. Photocurrent density improved with TaCl5 treatment, which contributed to fill the void between the electrode and particles. Photocurrent density was also improved by extending the electrophoretic deposition time from 1 to 2 minutes, since it increased the agglomeration of TaON particles on the FTO substrate and simultaneously helped to rearrange them.5 It was also found that heterojunction formation with CaFe2O4 enhanced the photocurrent density by 5 times (1.24 mA cm−2) compared with that of bare TaON.5 The IPCE of bare TaON was 5% at 500 nm, whereas for the TaON/CaFe2O4 heterojunction it was 30% at 400 nm.5 The maximum STH efficiency for the heterojunction system was 0.053% at 1 V. This system is still subject to self-oxidation of TaON and the backward reaction between hydrogen and oxygen. These findings indicated that the intimate contact of TaON and CaFe2O4 was essential for increased photocatalytic performance.5 The CaFe2O4 layer not only improved the separation of photo-generated electrons and holes, but also increased the charge carrier density by absorbing photons from visible light.5

Two photocatalytic materials can be coupled to form ‘Z-scheme’ heterojunction structures for water splitting applications.80 Sayama and coworkers first reported heterojunctions of a hydrogen evolution photocatalyst strontium titanate co-doped with chromium and tantalum, and an oxygen evolution photocatalyst tungsten trioxide with Pt co-catalyst deposited.81 This Z-scheme arrangement was capable of overall water splitting in NaI aqueous solution. In the Z scheme system, Ag may work as a mediator to transfer photo-excited electrons from the conduction band of the oxygen evolution catalyst to the valence band of the hydrogen evolution catalyst.80 A ZnRh2O4/Ag/Ag1−XSbO3−Y system was able to utilize visible light up to 545 nm while a ZnRh2O4/Ag/Bi4V2O11 system was able to function up to 700 nm irradiation. Mainly Ag1−XSbO3−Y and Bi4V2O11 determine the absorption properties in their respective systems.82 Normal solid-state reactions can be used for the synthesis of polycrystalline Bi4V2O11 (p-Bi4V2O11) powder to make a heterojunction with ZnRh2O4/Ag/p-Bi4V2O11.82 Kobayashi et al. synthesized a single crystal of Bi4V2O11 (s-Bi4V2O11) using a pulverization technique which resulted in a powdered ZnRh2O4/Ag/s-Bi4V2O11 photocatalyst system.80 This single crystal Bi4V2O11 had higher crystallinity and anisotropy than polycrystalline Bi4V2O11, which improved the mobility and separation of the photogenerated electrons and holes. Moreover, this single crystalline Bi4V2O11 mediated ZnRh2O4/Ag/s-Bi4V2O11 was able to utilize the visible spectrum up to 740 nm.80 The scarcity of the Rh component is a huge drawback for utilizing the favorable properties of ZnRh2O4/Ag/s-Bi4V2O11 on a large scale. Excess Ag in the system created another drawback in ZnRh2O4/Ag/s-Bi4V2O11 mediated pure water splitting because excess Ag worked as a sacrificial agent for oxygen evolution. To remove excess Ag, a nitric acid treatment is required which might damage ZnRh2O4 and s-Bi4V2O11. The damage of ZnRh2O4 and s-Bi4V2O11 leads to reduced photocatalytic activity.80 This problem can be solved by replacing Ag with Au, which does not act as a sacrificial agent and is highly stable. Thus Au inserted in the ZnRh2O4/Au/s-Bi4V2O11 system was capable of splitting pure water.80 This system demonstrated a hydrogen production rate of 0.02 μmol h−1 and an oxygen production rate of 0.01 μmol h−1 with an AQY of 0.036% at a 545 nm wavelength. Although this system was able to split pure water up to 720 nm, no STH value was reported for this system. Au and Ag have another unique property of surface plasmonic resonance that improves photocatalytic efficiency.

The surface plasmon resonance (SPR) of noble metals like Au and Ag can be utilized for increasing the photocatalytic activity of different materials.2 Plasmon resonance is confined within an interface which is actually a charge density wave.2 Surface plasmon resonance is developed by the collective oscillation of free conductive electrons which is induced by photons.83 Surface electrons naturally oscillate against the restoring positive force exerted by the nucleus, and SPR is observed when this frequency is matched with the frequency of incident photons.44 Different metals exhibit this resonance at different photon wavelengths. With the addition of the metal, the intensity of the plasmonic resonance is also dependent on the size and shape of the nanoparticle. Manipulating the size, shape, and composition of plasmonic metals, it is possible to fabricate a photocatalytic system which can absorb any wavelength of the solar spectrum.44 The utilization of the plasmonic resonance characteristics of novel metals is a good strategy for fabricating visible light-responsive photocatalysts. For large nanoparticles (Ag > 50 nm) the plasmonic resonance decays by scattering resonant photons radiatively while for comparatively small nanostructures (Ag < 30 nm) the resonance decreases with the formation of energetic charge carriers.44 SPR is capable of converting photonic energy with heat energy, amplifying the electromagnetic field and scattering electromagnetic radiation. These mechanisms are elaborately discussed in various review papers.85–87

SPR improves the local electric field close to the photocatalyst which ultimately enhances the electron hole pair generation rate.2 The exciton diffusion length can be increased by improving the electron and hole lifetime on the metal–photocatalyst interaction surface. Along with an efficient charge transfer mechanism, these noble metals enhance the absorption of visible light with their SPR.62 Cao et al. reported an improved photocatalytic activity of m-BiVO4 combined with Au nanoparticles. In 5.33 mM Na2S2O8 solution, Au incorporated on m-BiVO4 nanosheets showed 5862 μmol h−1 g−1 O2 evolution while bare m-BiVO4 evolved only 1579 μmol h−1 g−1 of O2.62 m-BiVO4 showed low photocatalytic activity because of poor charge separation.62 But, the charge separation was dramatically enhanced with the addition of Au, which prevented charge recombination before taking part in the reaction.62 This system is still at the half reaction stage and the conduction band position of BiVO4 is not favorable for H2 evolution. Ingram et al. reported an increased photocatalytic activity of N-doped TiO2 under visible light irradiation when it was combined with plasmonic Ag particles.88 Luchao et al. thoroughly investigated the TiO2 particle size dependence for plasmon-induced charge separation and the recombination dynamics in a plasmonic Au–TiO2 system.83 Within 50 fs, the electron injection from gold nanoparticles to the conduction band of TiO2 happened and charge recombination reduced.83 This decrease in charge recombination was a strong function of the TiO2 particle diameter.83 A time-resolved IR probe technique was used to investigate charge transfer and charge recombination processes. TiO2 particles with a larger diameter needed longer diffusion lengths and resulted in a longer charge recombination time compared with smaller diameter particles.83 Yu et al. reported a promising TiO2/Ag3PO4 electrode where Ag3PO4 reduced to a Ag particle under visible light irradiation and facilitated electron and hole transport by surface plasmonic resonance.84Fig. 10 shows that electrons generated in the Ag3PO4 layer are transferred to the CB of TiO2 through the Ag particle layer. In 1 M KOH solution, the photocurrent density was 5 times higher (2.34 mA cm−2 at 0 V) than that for a pure TiO2 electrode.84 This type of in situ generated layer provides a strong connection between the two different catalyst layers and thus facilitates charge transfer. An in situ generated conductive layer is widely utilized in the solid state Z scheme system.


image file: c8ta04165b-f10.tif
Fig. 10 (a) Illustration of electron and hole flow in the TiO2/Ag3PO4 electrode and (b) charge separation mechanism with band alignment. Reprinted with permission from ref. 84.

Co-catalyst loading is another effective strategy to improve water splitting efficiency. Loading an appropriate co-catalyst onto a catalyst surface provides an electron sink-like support that improves photo-induced charge separation and also acts as a catalytic active site for generating hydrogen and oxygen molecules.89 Co-catalysts can also be inserted into the heterojunction to further improve the oxidation or reduction efficiency of photocatalysts. Besides co-catalyst loading, in most of the catalyst systems some sacrificial agents are used. Some photocatalysts like bismuth-yttrium-tungsten ternary oxide (BiYWO6), zinc-germanium oxynitride ((Zn1.44Ge) (N2O0.44)), and niobium-substituted silver-tantalum oxide (AgTa0.7Nb0.3O0.3) are capable of splitting pure water under visible light illumination.90 The ZnRh2O4/Ag/Ag1−XSbO3−Y system showed 0.1682 μmol H2 and 0.0841 μmol O2 evolution from pure water after 24 h of visible light irradiation.90 No AQY or STH was reported for this system. Nanoparticulated CoO was reported to split water under visible light irradiation without the help of any co-catalyst or sacrificial agents.91 This CoO catalyst showed 5% STH efficiency but suffered from quick deactivation (active for around 1 h).91 Making suitable heterojunctions with other materials may improve the stability of nanocrystalline CoO.

1.4 Metal oxide–non-metal oxide heterojunction

The applications of noble metal co-catalysts for H2 evolution are becoming infeasible because they are expensive and not readily available in nature.89 As an alternative to using noble metals, people have been using transition metal sulfide co-catalysts like MoS2, NiS, CdIn2S4, CdLa2S4, and ZnIn2S4 for H2 evolution because they are cheap and highly abundant in nature.89 These co-catalysts are great candidates for H2 evolution.89 Different types of heterojunctions between metallic and non-metallic components were reported which improve photocatalytic activity and increase the hydrogen production rate from water splitting.92 Some of them include Se/BiVO4,56 g-C3N4/TiO2, g-C3N4/CdS, g-C3N4/MoS2, g-C3N4/Cu2O, g-C3N4/WO3, g-C3N4/Ni(OH)2, g-C3N4/ZnIn2S4, g-C3N4/Cd0.5Zn0.5S7, g-C3N4/SrTiO3, g-C3N4/Sr2Nb2O7, and g-C3N4/ZnFe2O4.92

Se is a non-metallic photo-responsive material and when used to make a heterojunction with a comparatively large band gap-containing material, the combined effect is beneficial for both reducing photo-generated electron hole recombination and suppressing photo-corrosion of the Se layer.56 Se is a promising alternative as a photo-absorber in place of the traditional semiconducting Si component.56 Various advantages like small grain size, higher surface area, and increased roughness have heightened its preference as a good photocatalyst. A Se layer played a vital role in trapping photo-generated hole, absorbing light, and improving charge separation in the Se/BiVO4 film (Fig. 11).56 Nasir et al. reported a Se/BiVO4 film where both Se and BiVO4 worked as a dual absorption layer which significantly increased the density of photo-induced charges.56 Maximum reported current density was 2.2 mA cm−2 at 1.3 V from a 0.5 M Na2SO4 solution.56 Sun et al. fabricated nanowire heterojunctions of a Si backbone and ZnO branches which were capable of enhancing light absorption and photocurrent generation (Nanoscale, 2012, 4, 1515–1521). This fabrication drastically increased the surface area and reduced the radius of curvature which contributed to effective charge separation and improved gas evolution from the water splitting reaction. Si-based materials usually face high charge recombination on the electrolyte surface and have a low hydrogen evolution rate. The gas evolution performances of one Si-based material was improved through the fabrication of a two step solution phase integration of ZnO branches on a Si core to create an effective p–n junction for water splitting (Nanoscale, 2012, 4, 1515–1521). In a Si/ZnO heterojunction, ZnO had a lower CB position than Si, which drove photo-excited electrons towards the ZnO layer to the reduction level of water, where photo-generated holes proceed toward Si in the opposite direction. Different etching times were used to grow different lengths of nanowires. Longer Si nanowires obtained from 15 minutes of etching produced 50–100% higher photocurrent than shorter Si nanowires obtained with 5 minutes of etching. Longer wires resulted in better light absorption and provided an increased surface area for ZnO branches (Nanoscale, 2012, 4, 1515–1521).


image file: c8ta04165b-f11.tif
Fig. 11 Charge flow within the Se–BiVO4 heterojunction structure. Reprinted with permission from ref. 56.

Yu et al. reported a system fabricated with ZnO/ZnS/CdS/CuInS2, which exhibited a photocurrent density of 10.5 mA cm−2 and an IPCE efficiency of 57.7% at 480 nm in the presence of 0.5 M Na2S and 0.5 M Na2SO3 aqueous solutions.93 This high photocurrent density resulted from the combined effect of large visible light absorption of CuInS2 and CdS, excellent charge carrier supply from CdS, charge recombination prevention within the metal oxide and CuInS2 interface by the ZnS buffer layer, and action of ZnO as a pathway for photogenerated electron transfer.93 ZnO underwent photo-corrosion by self-oxidation; an atomic layer deposition of a very thin TiO2 layer can protect it from photo-corrosion.94 GaInP2 has excellent electronic and diffusion properties but suffers from poor corrosion resistance. Gu et al. described graded MoSx, MoOySz, MoOx, and TiO2 layers to protect the GaInP2 electrode in highly acidic environments.95 A MoOx–MoSx hybrid improved the conductive and diffusive properties of the MoSx layer. Moreover, MoOx–TiO2 mixed metal oxides provided strong acid resistance because of their Bronsted acidity properties.95 Here, annealing the MoSx/a-TiO2–GaInP2 catalyst at 400 °C helped form a MoOx layer and transformed TiO2 from amorphous to a crystalline state. This system exhibited an IPCE efficiency of 70% and a faradaic efficiency of 94% in the presence of 0.5 M H2SO4 aqueous solution (pH 0.3). GaInP2 also showed high photocurrent density under strongly basic conditions (pH 13).96 A covalently attached molecular cobalt catalyst and an amorphous TiO2 protective layer were used in the p-GaInP2 electrolyte interface in a highly basic environment. At pH 13 it showed an enhanced IPCE efficiency of 80% and 100% faradaic efficiency in a 0.1 M NaOH aqueous solution. The truancy of oxide and surface hydride layers made it difficult to attach a molecular (HOOCpy)Co(dmgH)2(Cl) catalyst onto the p-GaInP2 surface. The insertion of an amorphous TiO2 layer by atomic layer deposition helped to form covalent bonding with the carboxylic group of the molecular cobalt catalyst.96 The covalently attached molecular cobalt reduced the kinetic barrier for the HER and worked as a highly active HER catalyst. CdS is used in different photocatalysts as a quantum dot (QD). Some other semiconductor QDs such as CdTe, CdSe, PbS, and CuInS2 can be used in photocatalysts to improve their photoactivity. Sun et al. reported that CdS QDs attached to a TiO2 photocatalyst showed a H2 evolution rate of 2200 μmol h−1 W−1 within the visible light spectrum.97 Moreover, the addition of a plasmonic metal at the CdS/TiO2 interface can improve photocatalytic activity. Li et al. incorporated Au, as a plasmonic photosensitizer, at the interface of the CdS/TiO2 heterojunction. This system showed an improved photocurrent density of 4.07 mA cm−2 without applying any bias voltage in a 0.25 M Na2S and 0.35 M Na2SO3 aqueous solution. Here, Au not only acted as a photo-sensitizer but also worked as a conductive bridge that facilitated the charge transfer.98 The addition of a CdS-decorated In2S3 layer in the Pt/Cu2ZnSnS4 electrode system also enhanced its photocatalytic activity.99 The Pt/In2S3/CdS/Cu2ZnSnS4 electrode demonstrated 1.63% HC-STH at 0.31 VRHE, which was the highest reported STH efficiency for Cu2ZnSnS4-based heterojunction systems.99 CdS-based heterojunctions still suffer from poor stability.

Recently a WO3-based photoelectrochemical system was reported by Wang et al.100 After the deposition of WO3 on FTO via hydrothermal methods, they deposited a seed layer of Bi2S3via a successive ionic layer adsorption and reaction (SILAR) method before chemical bath deposition of a Bi2S3 layer.100 This seed layer increased contact between the WO3 and Bi2S3 layers. For the same synthesis conditions (Bi2S3 deposited for 4 h), increased photocurrent density (4.3 mA cm−2 at 0.9 V vs. RHE) was observed compared to the WO3/Bi2S3 electrode without the seed layer (3.1 mA cm−2 at 0.9 V vs. RHE).100 This system had a good photocurrent density but suffered from poor corrosion resistance in 0.1 M Na2S and 0.1 M Na2SO3 aqueous solution.100 Wang et al. reported two different electron transfer mechanisms by either WO3 or g-C3N4 initial deposition on FTO.101 When WO3 was first deposited, a type II (first charge transfer mechanism of the type II heterojunction in Fig. 6) heterojunction was formed. Initially g-C3N4 deposition formed a Z scheme (second charge transfer mechanism of the type II heterojunction in Fig. 6) charge transfer mechanism. The initial WO3 deposition showed better photocurrent density (0.82 mA cm−2 at 1.23 V vs. RHE) than the initial g-C3N4 deposited sample (0.22 mA cm−2 at 1.23 V vs. RHE).101

Developing a g-C3N4-based heterojunction is a good approach for improving the photocatalytic activity of other materials since g-C3N4 alone can split pure water.102 The large band gap of Ta2O5 has confined its function within the UV range; coupling with small band gap g-C3N4 initiates the absorbance of visible light.92 The more negative conduction band potential of g-C3N4 helped to transfer photogenerated electrons to the conduction band of Ta2O5 through the heterojunction interface. These gathered electrons in the CB of Ta2O5 then participated in photocatalytic hydrogen production via water reduction. Methanol worked as a sacrificial agent to tap photo-generated holes from the valence band of g-C3N4.92 This Ta2O5/g-C3N4 heterojunction thus simultaneously improved visible light absorption and reduced photoinduced electron hole pair recombination, which resulted in increased hydrogen evolution compared to pure Ta2O5 or g-C3N4 alone.92 NiS-loaded C3N4 was reported to efficiently evolve H2 under visible light irradiation.103 This system produced H2 at a rate of 48.2 μmol h−1 with an AQY of 1.9% in 15 vol% triethanolamine aqueous solution. During the H2 evolution reaction, water passed through a reactor outside to maintain low temperatures since at higher temperatures, Ni2+ ions work as a catalyst to decompose C3N4.103 Optimum NiS loading in that system was found to be 1.25 wt%. Higher loading decreased photocatalytic efficiency by blocking pores of the C3N4 catalyst.103 A g-C3N4 sheet also worked as a promising support for the in situ growth of other nanomaterials.13 For example, Hu et al. confirmed the uniform dispersion of InVO4 nanoparticles (20 nm) on g-C3N4 sheets via SEM and TEM imaging.13 This uniformly dispersed surface helped to get intimate contact, which improved photocatalytic activity by accelerating charge transfer and reducing the photoinduced electron hole recombination rate.13 The optimized InVO4 nanoparticle dispersed (20 wt%) g-C3N4/InVO4 heterojunction showed an enhanced hydrogen production compared to pure g-C3N4 at 212 μmol g−1 h−1 and had an AQY of 4.9% at 420 nm in 20 vol% methanol solution. Here, g-C3N4 acted very significantly to form g-C3N4/InVO4 nanocomposites. The probable growth mechanism of g-C3N4/InVO4 depending on various g-C3N4 mass fractions was discussed.13 Sometimes, quantum dots are introduced into g-C3N4 to increase its photoactivity.104 A g-C3N4-based heterojunction showed an enhanced charge separation but its applications are limited to sacrificial agent-assisted water splitting systems.

Ternary heterojunctions including g-C3N4 are also useful for getting higher photocatalytic hydrogen generation from water splitting. Dong et al. reported a double Z-scheme ZnO/ZnS/g-C3N4 ternary heterojunction which was able to produce hydrogen at a rate of 301.25 μmol g−1 h−1 in a 0.25 M Na2S and 0.25 M Na2SO3 aqueous solution.105 High surface area and multi-stage charge transfer contributed to better electron hole transfer and separation which resulted in better photocatalytic activity. The charge transfer mechanism through the ZnO/ZnS/g-C3N4 system and hydrogen evolution using different systems are shown in Fig. 12. In the proposed mechanism [Fig. 12(b)] a double Z-scheme is established when electrons from the CB of ZnO transfer to the VB of ZnS, and electrons from the CB of ZnS accumulate in the VB of g-C3N4. These electron transfers ultimately increase the electron density in the CB of g-C3N4.105


image file: c8ta04165b-f12.tif
Fig. 12 (a) Performance of the synthesized material for water splitting and (b) the proposed mechanism for hydrogen production using ZnO/ZnS/g-C3N4. Reprinted with permission from ref. 105.

2 Synthesis method

The fabrication of photoelectrodes for PEC water splitting involves synthetic, etching, and cutting processes. In this section we briefly introduce the synthesis techniques of composites for photocatalyst preparation. Some of the synthesis methods consist of complicated steps, such as lithography or e-beam evaporation116 while others employ only simple and versatile forms such as hydrothermal, sol–gel,117 layer by layer (LBL),118 and dip coatings.46 Each system has its own shortcomings and advantages. For example, in contrast to the high vacuum film deposition method, the sol–gel technique has a number of advantages such as low-temperature processing, easy large surface coating, and possible formation of homogeneous multicomponent films.119 But the use of this method is limited by the list of certain materials.120 To produce complex composites which are not feasible by a single synthesis method, a combination of several synthetic techniques into one fabrication process is being exploited. Proceeding from the fact that different methods allow obtaining various structures, it becomes possible to tune the architecture of coatings using a wide range of materials. In the literature,121 authors synthesized a TiO2/TiSi2 core/shell heterojunction nanostructure by combining both chemical vapor deposition (CVD)122 and atomic layer deposition (ALD) routes. Various nanoarchitectures of multicomponent systems were also produced by using several methods or steps. As an example,123 four separate steps were utilized to obtain a system of a platinum-tipped cadmium sulfide rod with an embedded cadmium selenide seed, improving the catalytic performance in the visible range. Some of the common methods for the synthesis of nanostructured materials are discussed in more detail below.

2.1 Electrospinning

Electrospinning is a technique used for the fabrication of continuous nanofibers124 by applying an electric field to create a charged jet of polymer solution.125 This method creates a 1D material with a diameter ranging from a few micrometers to a few nanometers.126 For example, hierarchical structures of Bi2WO6/TiO2 (BWO/TiO2)127 and CuO/SnO2 have been produced by combining electrospinning and other techniques.128 Controllable fabrication of nanostructures via electrospinning can be realized by changing several parameters:129

(1) Intrinsic properties of the solution: type of polymer, polymer chain, and viscosity.

(2) Strength of the applied electric field.

(3) Temperature.

(4) Humidity.

In addition to the above parameters, the ratio of polymer, precursor, and solvent involved in the synthetic process influences the morphology and diameter of low dimensional fibers. Furthermore, authors130 fabricated core–shell heterojunction structures of (CdS)1/(ZnO)1, (CdS)0.5/(ZnO)1, and (CdS)1/(ZnO)0.5 by optimal design of the above parameters in electrospinning and compared their photocatalytic performance. The best result was obtained from (CdS)1/(ZnO)1 and explained by reduced photoinduced e/h+ recombination. The combination of the electrospinning method with the hydrothermal method opens up opportunities for growing nanoobjects directly onto micro/nanofibers. For example, SrTiO3 nanocubes can be grown hydrothermally onto electrospun long TiO2 fibers to form SrTiO3/TiO2 nanofiber heterostructures. This fabricated heterostructure provided fast separation of photogenerated charge carriers due to direct contact between SrTiO3 and TiO2, thus dramatically enhancing photocatalytic activity.131 Despite the many advantages of the electrospinning process, it has some challenges, including the production of low dimensional fibers in c axes.132 The ability to produce aligned nanofibers is more interesting in PEC terms because vertically aligned nanoobjects provide a high area of interfacial contact and highly efficient pathways for generated charges.133 It is essential to increase the speed of the electrospinning process without affecting nanofiber properties for an effective manufacturing process.

2.2 Hydrothermal

The hydrothermal synthesis of nanomaterials proceeds at high temperature, where solid materials react in aqueous solutions or vapors.134 Due to its simplicity, hydrothermal synthesis has become very popular in the development of photoelectrodes as an easy route to obtain nanostructures in relatively large amounts.135 Active layers of inorganic materials are hydrothermally deposited on electrodes by many researchers working in the field of photocatalysis. Various structures with a high surface area such as nanotubes of titanium dioxide,136 ZnO nanorods,137 crystalline powders of BiVO4,138 and TiO2 nanowires can be obtained in one hydrothermal step without further treatment.139 Through multiple hydrothermal synthesis steps, more complex hierarchical structures can be obtained. The hydrothermal fabrication of 3D hierarchical architectures with a core of TiO2 nanobelts and a shell of MoS2 nanosheets has been demonstrated in ref. 140. The obtained TiO2/MoS2 heterostructure performed well and demonstrated stable hydrogen production. Limin Song and co-workers141 applied the hydrothermal method without any surfactants or templates to produce Fe2O3 nanotubes, which act as a good visible light photocatalyst and are also suitable for other electronic nanodevices.142 The broadening of the spectral sensitivity of wide bandgap photocatalysts obtained by hydrothermal synthesis is also feasible by depositing narrow bandgap semiconductor materials over them with other methods. Plasmonic heterostructures as the photoanode for a water splitting system were prepared by the photoreduction of Au3+ to Au0 directly onto hydrothermally obtained wide bandgap semiconductors.137 Wang X. W. and Cheng H. M. synthesized ZnO/CdS core–shell nanorods, which exhibit highly stable and efficient photocatalytic hydrogen evolution by two step hydrothermal routes. First, ZnO nanorods were obtained by a modified hydrothermal method, where 0.1 M zinc acetate and 0.6 M hexamethylene tetramine dissolved in 50 mL DI water was treated for 10 h in an autoclave at 95 °C. The CdS shell was deposited onto the hydrothermally obtained ZnO nanorods by suspending them in 20 mL cadmium acetate solution under ultrasonic conditions for 1 h, followed by heating in H2S gas at different temperatures.143 Despite its simplicity, hydrothermal synthesis possesses some disadvantages, such as a long reaction duration and difficulty in obtaining a uniform photocatalyst size.136

2.3 Sol–gel

The sol–gel process includes two steps, which are hydrolyzation and condensation. Reactive species are initially hydrolyzed to become colloidal sol. With the assistance of a catalyst, the sol solution further reacts to form an infinite network of gel particles through the condensation process. The gel can be further treated to obtain the desired material by a heating treatment.144 For example, nanocrystalline ZnO, CdS, and SnO2 particles have been synthesized by a simple sol–gel method.145

2.4 Solid state reactions

Solid state reactions offer a good opportunity to synthesize unique nanostructured materials which are usually difficult to obtain by conventional synthesis techniques.146 During solid state reactions, a phase change occurs after mixing necessary component substances. The powder mixture is further treated at a high temperature. For obtaining an AI2O3–TiO2 composition via a solid state reaction, the powder mixture of small particle size was isothermally treated at around 1580 K.147 To synthesize CaFe2O4, powders of Fe2O3 and CaCO3 were mixed and then heated to 1100 °C for 2 h.148

3 Summary and outlook

One decade ago, photocatalytic pure water splitting under visible light irradiation was a dream reaction. Tremendous research into photocatalytic and photoelectrochemical water splitting has turned that dream into reality. To compete with conventional fuel prices, the efficiency of photocatalytic water splitting needs to improve now. According to the US Department of Energy, photocatalytic water splitting can compete with conventional fuel if the water splitting efficiency (STH) reaches around 10%.149 To attain this efficiency, photocatalysts are required to absorb wavelengths between 600 and 700 nm with a 40–60% apparent quantum yield. The heterojunction strategy helps to synthesize highly efficient photocatalytic materials at low cost. Deep insight into the heterojunction interface is needed to build up new, efficient photocatalysts. To create heterojunctions, one needs to proceed strategically. First, band alignment should be investigated to determine which heterojunction system will give a thermodynamically favorable charge transfer. Then, a study of charge transfer dynamics with appropriate characterization tools can provide a fundamental understanding of the performance of the heterojunction system and help to design an efficient photocatalyst system. Thus, various research studies are going on to improve the efficiency of photocatalytic water splitting. The improvement of the stability of different photocatalysts in aqueous medium is still one of the main concerns in photocatalytic water splitting. Surface modifications of different catalysts, crystal structure study, utilization of plasmonic resonance of metallic components in heterojunction systems, and making conductive bridges with appropriate materials help to improve the efficiency of photocatalytic water splitting. Some recent studies, including those on the polarization field effect and particulate sheets in heterojunction systems, have opened the door to further improvements in the photocatalytic system. Particulate photocatalyst sheet-based systems have shown promising results for designing a scalable water splitting system. Various types of materials could be explored to design particulate photocatalyst sheets for obtaining a higher performance from this type of system. Light harnessing capability, effective charge separation, and improvements in the stability of photocatalysts remain as the key challenges to enhancing photocatalytic water splitting efficiency.

In future studies, we believe the overall pure water splitting efficiency using photoelectrochemical cells still needs to be improved. Designing heterojunction nanostructured materials leveraging the advantages of different material properties will lead to the achievement of a high overall pure water splitting efficiency. In heterojunction design, the inclusion of conductive bridging (Au, Ag, graphene, and carbon) and an internal electrical field (ferroelectric materials and polarized materials) still needs to be extensively studied for effective charge separation and efficient charge transportation. Instead of two-component heterojunction systems, some multicomponent heterojunction systems show better activity towards the visible light driven water splitting process. A proper understanding of the interactions between different layers is necessary to successfully design more effective multicomponent heterojunction systems. The stability issue of single photocatalysts can be further improved in the future through designing new heterojunction systems. The excellent electronic and diffusion properties of metal sulfides and phosphides can be utilized more by forming effective heterojunctions with a protective layer around them. Different inexpensive molecular catalysts can be investigated more to fabricate the protective layer. Finding the optimal co-catalyst for a heterojunction photocatalyst system needs to be fully understood. The elimination of expensive and rare co-catalysts is necessary to obtain an economically feasible photocatalytic system. Mixed metal oxide co-catalysts with different compositions should be further investigated for improved photocatalyst system performance. The fabrication of photoelectrodes still needs to be improved to obtain stable photoelectrochemical systems.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors greatly acknowledge financial support from Texas Tech University. This work was supported by ACS PRF (57095-DNI7). The authors want to thank Robin Dupre for her contribution to fixing grammatical errors in the manuscript.

References

  1. M. Moniruddin, K. Afroz, Y. Shabdan, B. Bizri and N. Nuraje, Appl. Surf. Sci., 2017, 419, 886–892 CrossRef.
  2. Z. Liu, W. Hou, P. Pavaskar, M. Aykol and S. B. Cronin, Nano Lett., 2011, 11, 1111–1116 CrossRef PubMed.
  3. W. Septina, R. R. Prabhakar, R. Wick, T. Moehl and S. D. Tilley, Chem. Mater., 2017, 29, 1735–1743 CrossRef.
  4. X. Chen, S. Shen, L. Guo and S. S. Mao, Chem. Rev., 2010, 110, 6503–6570 CrossRef PubMed.
  5. E. S. Kim, N. Nishimura, G. Magesh, J. Y. Kim, J.-W. Jang, H. Jun, J. Kubota, K. Domen and J. S. Lee, J. Am. Chem. Soc., 2013, 135, 5375–5383 CrossRef PubMed.
  6. J. Su, X.-X. Zou, G.-D. Li, X. Wei, C. Yan, Y.-N. Wang, J. Zhao, L.-J. Zhou and J.-S. Chen, J. Phys. Chem. C, 2011, 115, 8064–8071 CrossRef.
  7. A. Paracchino, V. Laporte, K. Sivula, M. Grätzel and E. Thimsen, Nat. Mater., 2011, 10, 456–461 CrossRef PubMed.
  8. S. Y. Reece, J. A. Hamel, K. Sung, T. D. Jarvi, A. J. Esswein, J. J. Pijpers and D. G. Nocera, Science, 2011, 334, 645–648 CrossRef PubMed.
  9. K. Sun, Y. Jing, C. Li, X. Zhang, R. Aguinaldo, A. Kargar, K. Madsen, K. Banu, Y. Zhou and Y. Bando, Nanoscale, 2012, 4, 1515–1521 RSC.
  10. Y. Hou, F. Zuo, A. Dagg and P. Feng, Nano Lett., 2012, 12, 6464–6473 CrossRef PubMed.
  11. Q. Xiang, J. Yu and M. Jaroniec, J. Phys. Chem. C, 2011, 115, 7355–7363 CrossRef.
  12. Y. Tian, B. Chang, J. Lu, J. Fu, F. Xi and X. Dong, ACS Appl. Mater. Interfaces, 2013, 5, 7079–7085 CrossRef PubMed.
  13. B. Hu, F. Cai, T. Chen, M. Fan, C. Song, X. Yan and W. Shi, ACS Appl. Mater. Interfaces, 2015, 7, 18247–18256 CrossRef PubMed.
  14. P. Zhou, J. Yu and M. Jaroniec, Adv. Mater., 2014, 26, 4920–4935 CrossRef PubMed.
  15. S. Tanigawa and H. Irie, Appl. Catal., B, 2016, 180, 1–5 CrossRef.
  16. A. Kudo and Y. Miseki, Chem. Soc. Rev., 2009, 38, 253–278 RSC.
  17. S. J. Moniz, S. A. Shevlin, D. J. Martin, Z.-X. Guo and J. Tang, Energy Environ. Sci., 2015, 8, 731–759 RSC.
  18. M. Ni, M. K. Leung, D. Y. Leung and K. Sumathy, Renewable Sustainable Energy Rev., 2007, 11, 401–425 CrossRef.
  19. K. Sivula, F. Le Formal and M. Grätzel, ChemSusChem, 2011, 4, 432–449 CrossRef PubMed.
  20. N. Nuraje, R. Asmatulu and S. Kudaibergenov, Curr. Inorg. Chem., 2012, 2, 124–146 CrossRef.
  21. T. Hisatomi, J. Kubota and K. Domen, Chem. Soc. Rev., 2014, 43, 7520–7535 RSC.
  22. S. Ye, R. Wang, M.-Z. Wu and Y.-P. Yuan, Appl. Surf. Sci., 2015, 358, 15–27 CrossRef.
  23. S. Cao and J. Yu, J. Phys. Chem. Lett., 2014, 5, 2101–2107 CrossRef PubMed.
  24. J. Wen, J. Xie, X. Chen and X. Li, Appl. Surf. Sci., 2017, 391, 72–123 CrossRef.
  25. K. Maeda, J. Photochem. Photobiol., C, 2011, 12, 237–268 CrossRef.
  26. K. Maeda and K. Domen, J. Phys. Chem. Lett., 2010, 1, 2655–2661 CrossRef.
  27. S. Chen, T. Takata and K. Domen, Nat. Rev. Mater., 2017, 2, 17050 CrossRef.
  28. M. Moniruddin, B. Ilyassov, X. Zhao, E. Smith, T. Serikov, N. Ibrayev, R. Asmatulu and N. Nuraje, Materials Today Energy, 2018, 7, 246–259 CrossRef.
  29. K. Maeda, ACS Catal., 2013, 3, 1486–1503 CrossRef.
  30. S. Chen, S. S. Thind and A. Chen, Electrochem. Commun., 2016, 63, 10–17 CrossRef.
  31. G. Wang, Y. Ling, H. Wang, L. Xihong and Y. Li, J. Photochem. Photobiol., C, 2014, 19, 35–51 CrossRef.
  32. M. G. Walter, E. L. Warren, J. R. McKone, S. W. Boettcher, Q. Mi, E. A. Santori and N. S. Lewis, Chem. Rev., 2010, 110, 6446–6473 CrossRef PubMed.
  33. T. Hisatomi, K. Takanabe and K. Domen, Catal. Lett., 2015, 145, 95–108 CrossRef.
  34. B. Tian, B. Tian, B. Smith, M. Scott, R. Hua, Q. Lei and Y. Tian, Nat. Commun., 2018, 9, 1397 CrossRef PubMed.
  35. J. Tang, J. R. Durrant and D. R. Klug, J. Am. Chem. Soc., 2008, 130, 13885–13891 CrossRef PubMed.
  36. M. Graetzel, Acc. Chem. Res., 1981, 14, 376–384 CrossRef.
  37. K. Takanabe, ACS Catal., 2017, 7, 8006–8022 CrossRef.
  38. Z. Hu, B. J. Deibert and J. Li, Chem. Soc. Rev., 2014, 43, 5815–5840 RSC.
  39. K. Takanabe, in Solar Energy for Fuels, Springer, 2015, pp. 73–103 Search PubMed.
  40. H. Wang, L. Zhang, Z. Chen, J. Hu, S. Li, Z. Wang, J. Liu and X. Wang, Chem. Soc. Rev., 2014, 43, 5234–5244 RSC.
  41. N. J. Turro, V. Ramamurthy and J. C. Scaiano, Photochem. Photobiol., 2012, 88, 1033 CrossRef.
  42. T. Jafari, E. Moharreri, A. S. Amin, R. Miao, W. Song and S. L. Suib, Molecules, 2016, 21, 900 CrossRef PubMed.
  43. J. Liao, B. Sa, J. Zhou, R. Ahuja and Z. Sun, J. Phys. Chem. C, 2014, 118, 17594–17599 CrossRef.
  44. S. Linic, P. Christopher and D. B. Ingram, Nat. Mater., 2011, 10, 911 CrossRef PubMed.
  45. Y. Yuan, J. Lv, X. Jiang, Z. Li, T. Yu, Z. Zou and J. Ye, Appl. Phys. Lett., 2007, 91, 094107 CrossRef.
  46. K. Kato, Y. Torii, H. Taoda, T. Kato, Y. Butsugan and K. Niihara, J. Mater. Sci. Lett., 1996, 15, 913–915 CrossRef.
  47. Z.-K. Tang, W.-J. Yin, L. Zhang, B. Wen, D.-Y. Zhang, L.-M. Liu and W.-M. Lau, Sci. Rep., 2016, 6, 32764 CrossRef PubMed.
  48. J. Su, L. Guo, N. Bao and C. A. Grimes, Nano Lett., 2011, 11, 1928–1933 CrossRef PubMed.
  49. A. Kudo, H. Kato and I. Tsuji, Chem. Lett., 2004, 33, 1534–1539 CrossRef.
  50. X. Li, J. Yu and M. Jaroniec, Chem. Soc. Rev., 2016, 45, 2603–2636 RSC.
  51. Z. Shi, X. Wen, Z. Guan, D. Cao, W. Luo and Z. Zou, Ann. Phys., 2015, 358, 236–247 Search PubMed.
  52. S. Choudhary, S. Upadhyay, P. Kumar, N. Singh, V. R. Satsangi, R. Shrivastav and S. Dass, Int. J. Hydrogen Energy, 2012, 37, 18713–18730 CrossRef.
  53. Z. Li, W. Luo, M. Zhang, J. Feng and Z. Zou, Energy Environ. Sci., 2013, 6, 347–370 RSC.
  54. P. V. Kamat, J. Phys. Chem. C, 2007, 111, 2834–2860 CrossRef.
  55. J. S. Jang, H. G. Kim and J. S. Lee, Catal. Today, 2012, 185, 270–277 CrossRef.
  56. S. N. F. M. Nasir, H. Ullah, M. Ebadi, A. A. Tahir, J. S. Sagu and M. A. Mat Teridi, J. Phys. Chem. C, 2017, 121, 6218–6228 CrossRef.
  57. Q. Li, X. Ma, H. Liu, Z. Chen, H. Chen and S. Chu, ACS Appl. Mater. Interfaces, 2017, 9(22), 18836–18844 CrossRef PubMed.
  58. A. Fujishima and K. Honda, nature, 1972, 238, 37–38 CrossRef PubMed.
  59. J. Masud, A. T. Swesi, W. P. Liyanage and M. Nath, ACS Appl. Mater. Interfaces, 2016, 8, 17292–17302 CrossRef PubMed.
  60. J. Masud, P. C. Ioannou, N. Levesanos, P. Kyritsis and M. Nath, ChemSusChem, 2016, 9, 3128–3132 CrossRef PubMed.
  61. M. Shi, X. Pan, W. Qiu, D. Zheng, M. Xu and H. Chen, Int. J. Hydrogen Energy, 2011, 36, 15153–15159 CrossRef.
  62. S.-W. Cao, Z. Yin, J. Barber, F. Y. Boey, S. C. J. Loo and C. Xue, ACS Appl. Mater. Interfaces, 2011, 4, 418–423 CrossRef PubMed.
  63. A. Nashim, S. Martha and K. Parida, ChemCatChem, 2013, 5, 2352–2359 CrossRef.
  64. J. H. Baek, B. J. Kim, G. S. Han, S. W. Hwang, D. R. Kim, I. S. Cho and H. S. Jung, ACS Appl. Mater. Interfaces, 2017, 9(2), 1479–1487 CrossRef PubMed.
  65. P. Luan, M. Xie, D. Liu, X. Fu and L. Jing, Sci. Rep., 2014, 4, 6180 CrossRef PubMed.
  66. J. Johnson, N. Bakranov, M. Moniruddin, R. Iskakov, S. Kudaibergenov and N. Nuraje, New J. Chem., 2017, 41, 15528–15532 RSC.
  67. D. D. Qin, C. H. He, Y. Li, A. C. Trammel, J. Gu, J. Chen, Y. Yan, D. L. Shan, Q. H. Wang and J. J. Quan, ChemSusChem, 2017, 10, 2796–2804 CrossRef PubMed.
  68. Y. Pihosh, I. Turkevych, K. Mawatari, J. Uemura, Y. Kazoe, S. Kosar, K. Makita, T. Sugaya, T. Matsui and D. Fujita, Sci. Rep., 2015, 5, 11141 CrossRef PubMed.
  69. P. M. Rao, L. Cai, C. Liu, I. S. Cho, C. H. Lee, J. M. Weisse, P. Yang and X. Zheng, Nano Lett., 2014, 14, 1099–1105 CrossRef PubMed.
  70. X. Chang, T. Wang, P. Zhang, J. Zhang, A. Li and J. Gong, J. Am. Chem. Soc., 2015, 137, 8356–8359 CrossRef PubMed.
  71. Z. Zou, J. Ye, K. Sayama and H. Arakawa, nature, 2001, 414, 625–627 CrossRef PubMed.
  72. M. Manikandan, T. Tanabe, P. Li, S. Ueda, G. V. Ramesh, R. Kodiyath, J. Wang, T. Hara, A. Dakshanamoorthy and S. Ishihara, ACS Appl. Mater. Interfaces, 2014, 6, 3790–3793 CrossRef PubMed.
  73. B.-S. Huang and M.-Y. Wey, J. Nanopart. Res., 2014, 16, 2178 CrossRef.
  74. K. Iwashina and A. Kudo, J. Am. Chem. Soc., 2011, 133, 13272–13275 CrossRef PubMed.
  75. Q. Wang, T. Hisatomi, Q. Jia, H. Tokudome, M. Zhong, C. Wang, Z. Pan, T. Takata, M. Nakabayashi and N. Shibata, Nat. Mater., 2016, 15, 611 CrossRef PubMed.
  76. Q. Wang, T. Hisatomi, Y. Suzuki, Z. Pan, J. Seo, M. Katayama, T. Minegishi, H. Nishiyama, T. Takata and K. Seki, J. Am. Chem. Soc., 2017, 139, 1675–1683 CrossRef PubMed.
  77. V.-H. Nguyen and J. C. Wu, Appl. Catal., A, 2018, 550, 122–141 CrossRef.
  78. M. Higashi, K. Domen and R. Abe, J. Am. Chem. Soc., 2012, 134, 6968–6971 CrossRef PubMed.
  79. M. Higashi, R. Abe, T. Takata and K. Domen, Chem. Mater., 2009, 21, 1543–1549 CrossRef.
  80. R. Kobayashi, T. Takashima, S. Tanigawa, S. Takeuchi, B. Ohtani and H. Irie, Phys. Chem. Chem. Phys., 2016, 18, 27754–27760 RSC.
  81. K. Sayama, K. Mukasa, R. Abe, Y. Abe and H. Arakawa, J. Photochem. Photobiol., A, 2002, 148, 71–77 CrossRef.
  82. K. Teramura, K. Maeda, T. Saito, T. Takata, N. Saito, Y. Inoue and K. Domen, J. Phys. Chem. B, 2005, 109, 21915–21921 CrossRef PubMed.
  83. L. Du, A. Furube, K. Yamamoto, K. Hara, R. Katoh and M. Tachiya, J. Phys. Chem. C, 2009, 113, 6454–6462 CrossRef.
  84. L. Yu, J. He, C. Huang, M. Li, Y. Zhang, X. Zhou and H. Zhu, RSC Adv., 2017, 7, 54485–54490 RSC.
  85. J. A. Schuller, E. S. Barnard, W. Cai, Y. C. Jun, J. S. White and M. L. Brongersma, Nat. Mater., 2010, 9, 193 CrossRef PubMed.
  86. P. K. Jain, X. Huang, I. H. El-Sayed and M. A. El-Sayed, Acc. Chem. Res., 2008, 41, 1578–1586 CrossRef PubMed.
  87. S. C. Warren and E. Thimsen, Energy Environ. Sci., 2012, 5, 5133–5146 RSC.
  88. D. B. Ingram and S. Linic, J. Am. Chem. Soc., 2011, 133, 5202–5205 CrossRef PubMed.
  89. Y.-P. Yuan, S.-W. Cao, L.-S. Yin, L. Xu and C. Xue, Int. J. Hydrogen Energy, 2013, 38, 7218–7223 CrossRef.
  90. R. Kobayashi, S. Tanigawa, T. Takashima, B. Ohtani and H. Irie, J. Phys. Chem. C, 2014, 118, 22450–22456 CrossRef.
  91. L. Liao, Q. Zhang, Z. Su, Z. Zhao, Y. Wang, Y. Li, X. Lu, D. Wei, G. Feng and Q. Yu, Nat. Nanotechnol., 2014, 9, 69 CrossRef PubMed.
  92. Y. Hong, Z. Fang, B. Yin, B. Luo, Y. Zhao, W. Shi and C. Li, Int. J. Hydrogen Energy, 2017, 42, 6738–6745 CrossRef.
  93. Y.-X. Yu, W.-X. Ouyang, Z.-T. Liao, B.-B. Du and W.-D. Zhang, ACS Appl. Mater. Interfaces, 2014, 6, 8467–8474 CrossRef PubMed.
  94. H. Zhao, W. Deng and Y. Li, Advanced Composites and Hybrid Materials, 2017, pp. 1–10 Search PubMed.
  95. J. Gu, J. A. Aguiar, S. Ferrere, K. X. Steirer, Y. Yan, C. Xiao, J. L. Young, M. Al-Jassim, N. R. Neale and J. A. Turner, Nat. Energy, 2017, 2, 16192 CrossRef.
  96. J. Gu, Y. Yan, J. L. Young, K. X. Steirer, N. R. Neale and J. A. Turner, Nat. Mater., 2016, 15, 456 CrossRef PubMed.
  97. W.-T. Sun, Y. Yu, H.-Y. Pan, X.-F. Gao, Q. Chen and L.-M. Peng, J. Am. Chem. Soc., 2008, 130, 1124–1125 CrossRef PubMed.
  98. J. Li, S. K. Cushing, P. Zheng, T. Senty, F. Meng, A. D. Bristow, A. Manivannan and N. Wu, J. Am. Chem. Soc., 2014, 136, 8438–8449 CrossRef PubMed.
  99. F. Jiang, T. Harada, Y. Kuang, T. Minegishi, K. Domen and S. Ikeda, J. Am. Chem. Soc., 2015, 137, 13691–13697 CrossRef PubMed.
  100. Y. Wang, W. Tian, L. Chen, F. Cao, J. Guo and L. Li, ACS Appl. Mater. Interfaces, 2017, 9, 40235–40243 CrossRef PubMed.
  101. C.-H. Wang, D.-D. Qin, D.-L. Shan, J. Gu, Y. Yan, J. Chen, Q.-H. Wang, C.-H. He, Y. Li and J.-J. Quan, Phys. Chem. Chem. Phys., 2017, 19, 4507–4515 RSC.
  102. J. Liu, Y. Liu, N. Liu, Y. Han, X. Zhang, H. Huang, Y. Lifshitz, S.-T. Lee, J. Zhong and Z. Kang, Science, 2015, 347, 970–974 CrossRef PubMed.
  103. J. Hong, Y. Wang, Y. Wang, W. Zhang and R. Xu, ChemSusChem, 2013, 6, 2263–2268 CrossRef PubMed.
  104. L. Ge, F. Zuo, J. Liu, Q. Ma, C. Wang, D. Sun, L. Bartels and P. Feng, J. Phys. Chem. C, 2012, 116, 13708–13714 CrossRef.
  105. Z. Dong, Y. Wu, N. Thirugnanam and G. Li, Appl. Surf. Sci., 2018, 430, 293–300 CrossRef.
  106. Z. Zhang, K. Liu, Z. Feng, Y. Bao and B. Dong, Sci. Rep., 2016, 6, 19221 CrossRef PubMed.
  107. X. Yu, A. Shavel, X. An, Z. Luo, M. Ibanez and A. Cabot, J. Am. Chem. Soc., 2014, 136, 9236–9239 CrossRef PubMed.
  108. H. Kato, K. Asakura and A. Kudo, J. Am. Chem. Soc., 2003, 125, 3082–3089 CrossRef PubMed.
  109. S. Chen, Y. Qi, T. Hisatomi, Q. Ding, T. Asai, Z. Li, S. S. K. Ma, F. Zhang, K. Domen and C. Li, Angew. Chem., Int. Ed., 2015, 54, 8498–8501 CrossRef PubMed.
  110. M. M. Momeni and Y. Ghayeb, J. Appl. Electrochem., 2015, 45, 557–566 CrossRef.
  111. A. K. R. Police, S. Basavaraju, D. K. Valluri, S. Machiraju and J. S. Lee, Chem. Eng. J., 2014, 247, 152–160 CrossRef.
  112. S. Zhu, F. Yao, C. Yin, Y. Li, W. Peng, J. Ma and D. Zhang, Microporous Mesoporous Mater., 2014, 190, 10–16 CrossRef.
  113. J. Zhang, S. Z. Qiao, L. Qi and J. Yu, Phys. Chem. Chem. Phys., 2013, 15, 12088–12094 RSC.
  114. Z. Yin, Z. Wang, Y. Du, X. Qi, Y. Huang, C. Xue and H. Zhang, Adv. Mater., 2012, 24, 5374–5378 CrossRef PubMed.
  115. M. S. Akple, J. Low, S. Wageh, A. A. Al-Ghamdi, J. Yu and J. Zhang, Appl. Surf. Sci., 2015, 358, 196–203 CrossRef.
  116. A. Kargar, K. Sun, Y. Jing, C. Choi, H. Jeong, G. Y. Jung, S. Jin and D. L. Wang, ACS Nano, 2013, 7, 9407–9415 CrossRef PubMed.
  117. R. J. Candal, W. A. Zeltner and M. A. Anderson, Journal of Environmental Engineering, 1999, 125, 906–912 CrossRef.
  118. M. Li, S. Ishihara, Q. M. Ji, Y. G. Ma, J. P. Hill and K. Ariga, Chem. Lett., 2012, 41, 383–385 CrossRef.
  119. L. L. Hu, T. Yoko, H. Kozuka and S. Sakka, Thin Solid Films, 1992, 219, 18–23 CrossRef.
  120. Y. I. Song, G. Y. Kim, H. K. Choi, H. J. Jeong, K. K. Kim, C. M. Yang, S. C. Lim, K. H. An, K. T. Jung and Y. H. Lee, Chem. Vap. Deposition, 2006, 12, 375–379 CrossRef.
  121. Y. J. Lin, S. Zhou, X. H. Liu, S. Sheehan and D. W. Wang, J. Am. Chem. Soc., 2009, 131(8), 2772–2773 CrossRef PubMed.
  122. G. Che, B. B. Lakshmi, C. R. Martin, E. R. Fisher and R. S. Ruoff, Chem. Mater., 1998, 10, 260–267 CrossRef.
  123. L. Amirav and A. P. Alivisatos, J. Phys. Chem. Lett., 2010, 1, 1051–1054 CrossRef.
  124. Z. M. Huang, Y. Z. Zhang, M. Kotaki and S. Ramakrishna, Compos. Sci. Technol., 2003, 63, 2223–2253 CrossRef.
  125. J. Doshi and D. H. Reneker, J. Electrost., 1995, 35, 151–160 CrossRef.
  126. A. Greiner and J. H. Wendorff, Angew. Chem., Int. Ed., 2007, 46, 5670–5703 CrossRef PubMed.
  127. M. Shang, W. Z. Wang, L. Zhang, S. M. Sun, L. Wang and L. Zhou, J. Phys. Chem. C, 2009, 113, 14727–14731 CrossRef.
  128. S. W. Choi, A. Katoch, J. Zhang and S. S. Kim, Sens. Actuators, B, 2013, 176, 585–591 CrossRef.
  129. D. Li and Y. N. Xia, Adv. Mater., 2004, 16, 1151–1170 CrossRef.
  130. G. R. Yang, W. Yan, Q. Zhang, S. H. Shen and S. J. Ding, Nanoscale, 2013, 5, 12432–12439 RSC.
  131. H. W. Bai, J. Juay, Z. Y. Liu, X. X. Song, S. S. Lee and D. D. Sun, Appl. Catal., B, 2012, 125, 367–374 CrossRef.
  132. W. E. Teo and S. Ramakrishna, Nanotechnology, 2006, 17, R89–R106 CrossRef PubMed.
  133. I. Gonzalez-Valls and M. Lira-Cantu, Energy Environ. Sci., 2009, 2, 19–34 RSC.
  134. A. Tavakoli, M. Sohrabi and A. Kargari, Chem. Pap., 2007, 61, 151–170 Search PubMed.
  135. Y. L. Pang, S. Lim, H. C. Ong and W. T. Chong, Appl. Catal., A, 2014, 481, 127–142 CrossRef.
  136. N. Liu, X. Y. Chen, J. L. Zhang and J. W. Schwank, Catal. Today, 2014, 225, 34–51 CrossRef.
  137. M. Wu, W. J. Chen, Y. H. Shen, F. Z. Huang, C. H. Li and S. K. Li, ACS Appl. Mater. Interfaces, 2014, 6, 15052–15060 CrossRef PubMed.
  138. J. B. Liu, H. Wang, S. Wang and H. Yan, Mater. Sci. Eng., B, 2003, 104, 36–39 CrossRef.
  139. Y. C. Pu, G. M. Wang, K. D. Chang, Y. C. Ling, Y. K. Lin, B. C. Fitzmorris, C. M. Liu, X. H. Lu, Y. X. Tong, J. Z. Zhang, Y. J. Hsu and Y. Li, Nano Lett., 2013, 13, 3817–3823 CrossRef PubMed.
  140. W. J. Zhou, Z. Y. Yin, Y. P. Du, X. Huang, Z. Y. Zeng, Z. X. Fan, H. Liu, J. Y. Wang and H. Zhang, Small, 2013, 9, 140–147 CrossRef PubMed.
  141. L. M. Song, S. J. Zhang, B. Chen, J. J. Ge and X. C. Jia, Colloids Surf., A, 2010, 360, 1–5 CrossRef.
  142. Y. Hou, F. Zuo, A. Dagg and P. Y. Feng, Nano Lett., 2012, 12, 6464–6473 CrossRef PubMed.
  143. X. W. Wang, G. Liu, G. Q. Lu and H. M. Cheng, Int. J. Hydrogen Energy, 2010, 35, 8199–8205 CrossRef.
  144. B. B. Lakshmi, C. J. Patrissi and C. R. Martin, Chem. Mater., 1997, 9, 2544–2550 CrossRef.
  145. F. Gu, S. F. Wang, M. K. Lu, G. J. Zhou, D. Xu and D. R. Yuan, J. Phys. Chem. B, 2004, 108, 8119–8123 CrossRef.
  146. K. Biradha and R. Santra, Chem. Soc. Rev., 2013, 42, 950–967 RSC.
  147. B. Freudenberg and A. Mocellin, J. Am. Ceram. Soc., 1987, 70, 33–38 CrossRef.
  148. E. Casbeer, V. K. Sharma and X. Z. Li, Sep. Purif. Technol., 2012, 87, 1–14 CrossRef.
  149. DOE Technical Targets for Hydrogen Production from Photoelectrochemical Water Splitting, https://www.energy.gov/eere/fuelcells/doe-technical-targets-hydrogen-production-photoelectrochemical-water-splitting, (accessed 5/3/2018, 2018).

This journal is © The Royal Society of Chemistry 2018