Semiconductorbased photocatalysis is gaining significant attention for its capability to directly harness solar energy to produce solar fuels like hydrogen and hydrocarbons, and for degrading various pollutants. However, the efficiency of these photocatalytic processes is still limited due to rapid electronhole recombination and poor light absorption. To address these issues, extensive research and development efforts are being made. Notably, heterojunction photocatalysts engineered for better performance have shown promise due to their ability to spatially separate photogenerated electronhole pairs, leading to higher photocatalytic activity. The fundamental principles of different heterojunction photocatalysts are thoroughly explored in this context. Recent advancements in the development of heterojunction photocatalysts for a variety of applications are reviewed and evaluated. The article concludes with a concise overview and perspectives on the challenges and future trends in the field of heterojunction photocatalysts.
Review www.advmat.de www.advancedsciencenews.com Heterojunction Photocatalysts Jingxiang Low, Jiaguo Yu,* Mietek Jaroniec, Swelm Wageh, and Ahmed A Al-Ghamdi applications.[8–10] However, the practical applications of photocatalysis are still limited by its low photocatalytic activity.[11–14] In simple terms, a photocatalytic reaction on a semiconductor includes at least five main steps: i) light absorption by the semiconductor, ii) formation of photogenerated electron–hole pairs, iii) migration and recombination of the photogenerated electron–hole pairs, iv) adsorption of reactants and desorption of products, and v) occurrence of redox reactions on the semiconductor surface (see Figure 1) Among them, the recombination of electron–hole pairs plays a negative role in the photocatalytic processes.[15–17] During photocatalytic reaction, the photogenerated electron–hole pairs can either transfer to the photocatalyst surface and initiate redox reactions, or recombine and create useless heat To better understand this process, let us illustrate it by a somewhat similar situation: the effect of the gravitational force on a man jumping off the ground (see Figure 2a,b) The expressions for both the gravitational force acting on a man jumping off the ground and the Coulomb force acting between the electron and the hole have similar forms: Semiconductor-based photocatalysis attracts wide attention because of its ability to directly utilize solar energy for production of solar fuels, such as hydrogen and hydrocarbon fuels and for degradation of various pollutants However, the efficiency of photocatalytic reactions remains low due to the fast electron–hole recombination and low light utilization Therefore, enormous efforts have been undertaken to solve these problems Particularly, properly engineered heterojunction photocatalysts are shown to be able to possess higher photocatalytic activity because of spatial separation of photo generated electron–hole pairs Here, the basic principles of various heterojunction photocatalysts are systematically discussed Recent efforts toward the development of heterojunction photocatalysts for various photocatalytic applications are also presented and appraised Finally, a brief summary and perspectives on the challenges and future directions in the area of heterojunction photocatalysts are also provided Introduction A fast-growing industry and the rising global population in recent years are the key factors contributing to the energy shortage and environmental pollution Thus, to assure a longterm and sustainable development of human society, there is an urgent need for the development of environmentally friendly and renewable technologies for green energy production and environmental remediation Among the various proposed technologies, semiconductor-based photocatalysis has great potential because it directly utilizes solar energy both for the production of valuable chemical fuels, such as hydrogen and hydrocarbon fuels, and for the degradation of harmful pollutants.[1–6] Since the pioneering work on photocatalysis by Honda and Fujishima in 1972,[7] many semiconductors have been investigated and developed for various photocatalytic J X Low, Prof J G Yu State Key Laboratory of Advanced Technology for Materials Synthesis and Processing Wuhan University of Technology 122 Luoshi Road, Wuhan 430070, P R China E-mail: yujiaguo93@whut.edu.cn; jiaguoyu@yahoo.com Prof J G Yu, Prof S Wageh, Prof A A Al-Ghamdi Department of Physics Faculty of Science King Abdulaziz University Jeddah 21589, Saudi Arabia Prof M Jaroniec Department of Chemistry and Biochemistry Kent State University Kent, Ohio 44242, USA DOI: 10.1002/adma.201601694 1601694 (1 of 20) wileyonlinelibrary.com Mm R2 (1) qe q h r2 (2) Fg = G Fc = k In Equation (1) and (2), Fg is the gravitational force, G is the gravitational constant, M is the Earth’s mass, m is the mass of the man, R is the distance between the center of the Earth and the man, Fc is the Coulombic force, k is the Coulomb constant, qe and qh are the charge magnitudes of the electron and the hole, respectively, and r is the distance between the centers of these charges When the man (cf the electron) jumps from the ground (cf from the valence band) into the air (cf the conduction band), he will fall back to the ground rapidly (cf recombine with the hole) because of the action of gravitational force (cf the Coulombic force between the electron and the hole) In order to keep the man off the ground (cf to separate the photogenerated electron–hole pairs), a stool (cf the conduction band of the semiconductor B) is provided (see Figure 2c,d); then, the aforementioned man will land again on the stool and be kept off the ground (cf the electron and hole pairs can be separated) Moreover, it should be noted that the value of the Coulomb constant (8.99 × 109 N m2 C−2) is much larger than that of the gravitational constant (6.67 ì 1011 N m2 kg2) Thus, â 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Adv Mater 2017, 29, 1601694 www.advmat.de www.advancedsciencenews.com Figure 1. Schematic illustration of the typical photocatalytic processes on a semiconductor the aforementioned man will fall back to the ground on the timescale of seconds, while electron–hole pairs will recombine much faster, i.e., in the range of nanoseconds Therefore, the separation of photogenerated electron–hole pairs on the surface of a photocatalyst is very difficult to achieve Although preventing electron–hole recombination is a very challenging task, it can be accomplished by the proper design of a photocatalyst Various strategies have been proposed to efficiently separate the photogenerated electron–hole pairs in semiconductor photocatalysts, for instance by doping,[18,19] metal loading,[20,21] and/or introducing heterojunctions.[22,23] Among the proposed strategies, engineering heterojunctions in photocatalysts has been proved to be one of the most promising ways for the preparation of advanced photocatalysts because of its feasibility and effectiveness for the spatial separation of electron–hole pairs (see Figure 2d) Basically, five different heterojunctions have been mainly investigated by our group at Wuhan University of Technology (WUT) as well as by others and proved to be efficient for enhancing the activity of photocatalysts; these are conventional type-II heterojunctions,[24–27] p–n heterojunctions,[28–30] surface heterojunctions,[31–34] direct Z-scheme heterojunctions,[35,36] and semiconductor–graphene (SC–graphene) heterojunctions.[37,38] In the past 15 years, many studies have been published by our group on the development of advanced heterojunction photocatalysts for various applications As shown in Table 1, the conventional type-II heterojunction mechanism was studied by our group in 2001 to explain the high photocatalytic activity of anatase–brookite TiO2.[24,25] Then, the p–n heterojunction was found to be more effective for enhancing the photocatalytic activity than the aforementioned conventional type-II heterojunction.[39] Thereafter, two new heterojunction concepts were firstly proposed by our group, which were the direct Z-scheme heterojunction in 2013[35] and the surface heterojunction in 2014,[31] to further improve the photocatalytic activity Recently, two-dimensional (2D) graphene nanosheets have been widely used by our group to prepare advanced sc–graphene heterojunction photocatalysts with enhanced activity due to the fascinating properties of graphene, such as its high conductivity, large specific surface area, and high photostability Adv Mater 2017, 29, 1601694 Jiaguo Yu received his B.S and M.S in chemistry from Central China Normal University and Xi’an Jiaotong University, respectively, and his Ph.D in materials science in 2000 from Wuhan University of Technology In 2000, he became a Professor at Wuhan University of Technology He was a postdoctoral fellow at the Chinese University of Hong Kong from 2001 to 2004, a visiting scientist from 2005 to 2006 at the University of Bristol, and a visiting scholar from 2007 to 2008 at the University of Texas at Austin His current research interests include semiconductor photocatalysis, photocatalytic hydrogen production, CO2 reduction to hydrocarbon fuels, and so on Mietek Jaroniec received his M.S and Ph.D from M Curie-Sklodowska University (Poland) in 1972 and 1976; afterward, he was appointed as a faculty at the same University Since 1991, he has been Professor of Chemistry at Kent State University, Kent, Ohio (USA) His research interests include interfacial chemistry and the chemistry of materials, especially adsorption at the gas/solid and liquid/solid interfaces and nanoporous materials At Kent State, he has established a vigorous research program in the area of nanomaterials, such as ordered mesoporous silicas, organosilicas, inorganic oxides, carbon nanostructures, and nanostructured catalysts/photocatalysts, focusing on their synthesis, characterization, and environmental and energy-related applications © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim wileyonlinelibrary.com (2 of 20) 1601694 Review Jingxiang Low obtained his B.Eng (Hons) from Multimedia University, Malaysia in 2011 and his M.S in materials science from Wuhan University of Technology He is a now a Ph.D candidate under the supervision of Prof Jiaguo Yu at the State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology His current research includes photocatalytic H2 production and CO2 reduction www.advmat.de www.advancedsciencenews.com Review This review represents an appraisal of the recent efforts in engineering various heterojunction photocatalysts and highlights their application in photocatalysis, including hydrogen production, CO2 reduction, and pollutant degradation The above-mentioned five important types of heterojunctions in photocatalysts are reviewed and discussed Finally, the current status, opportunities, and future directions of these heterojunction photocatalysts are presented We hope that this review can provide some useful guidelines and shed some light toward the development of highly efficient heterojunction photocatalysts for different applications Conventional Heterojunctions Figure 2. Schematic illustration of: a) the effect of gravitational force on a man who jumps off the ground, b) electron–hole recombination on a single photocatalyst, c) use of a stool to keep a man off the ground, and d) electron–hole separation on a heterojunction photocatalyst A heterojunction, in general, is defined as the interface between two different semiconductors with unequal band structure, which can result in band alignments.[47,48] Typically, there are three types of conventional heterojunction photocatalysts, those with a straddling gap (type-I), those with a staggered gap (type-II), and those with a broken gap (type-III) (see Figure 3) For the type-I heterojunction photocatalyst (see Figure 3a), the conduction band (CB) and the valence band (VB) of semiconductor A are respectively higher and lower than the corresponding Table 1. List of heterojunction photocatalysts studied by our group at WUT for various photocatalytic applications Sample Heterojunction type Application Year Ref Anatase–brookite TiO2 Conventional type-II Acetone degradation 2001 [24] Anatase–brookite TiO2 Conventional type-II Acetone degradation 2003 [25] Ag–multiphase TiO2 Conventional type-II Methyl orange (MO) degradation 2005 [40] SnO–TiO2 Conventional type-II Rhodamine B (RhB) degradation 2008 [41] NiO–TiO2 p–n p-Chlorophenol degradation 2010 [42] BIOI–TiO2 p–n MO degradation 2011 [28] NiS–CdS p–n Hydrogen production 2013 [29] {001} and {101} TiO2 Surface CO2 reduction 2014 [31] N-doped {001} and {101}TiO2 Surface CO2 reduction 2015 [32] g-C3N4–TiO2 Direct Z-scheme HCHO decomposition 2013 [35] Anatase–rutile TiO2 Direct Z-scheme Hydrogen production 2014 [43] Ag2CrO4–graphene oxide Direct Z-scheme MB degradation 2015 [36] CdS–WO3 Direct Z-scheme CO2 reduction 2015 [44] SC–graphene Hydrogen production 2011 [37] Reduced graphene oxide–CdS g-C3N4–graphene SC–graphene Hydrogen production 2011 [45] Graphene–TiO2 nanosheets SC–graphene Hydrogen production 2011 [38] CdS–graphene SC–graphene CO2 reduction 2014 [46] 1601694 (3 of 20) wileyonlinelibrary.com © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Adv Mater 2017, 29, 1601694 www.advmat.de www.advancedsciencenews.com Review Figure 3. Schematic illustration of the three different types of separation of electron–hole pairs in the case of conventional light-responsive heterojunction photocatalysts: a) type-I, b) type-II, and c) type-III heterojunctions bands of semiconductor B.[49] Therefore, under light irradiation, the electrons and holes will accumulate at the CB and the VB levels of semiconductor B, respectively Since both electrons and holes accumulate on the same semiconductor, the electron–hole pairs cannot be effectively separated for the type-I heterojunction photocatalyst Moreover, a redox reaction takes place on the semiconductor with the lower redox potential, thereby significantly reducing the redox ability of the heterojunction photocatalyst For the type-II heterojunction photocatalyst (see Figure 3b), the CB and the VB levels of semiconductor A are higher than the corresponding levels of the semiconductor B Thus, the photogenerated electrons will transfer to semiconductor B, while the photogenerated holes will migrate to semiconductor A under light irradiation, resulting in a spatial separation of electron–hole pairs.[50–52] Similar to the type-I heterojunction, the redox ability of the type-II heterojunction photocatalyst will be also reduced because the reduction reaction and the oxidation reaction take place on semiconductor B with lower reduction potential and on semiconductor A with lower oxidation potential, respectively As shown in Figure 3c, the architecture of the type-III heterojunction photocatalyst is similar to that of the type-II heterojunction photocatalyst except that the staggered gap becomes so extreme that the bandgaps not overlap.[53,54] Therefore, the electron–hole migration and separation between the two semiconductors cannot occur for the type-III heterojunction, making it unsuitable for enhancing the separation of electron–hole pairs Among the aforementioned conventional heterojunctions, it is obvious that the type-II heterojunction is the most effective conventional heterojunction to be used for improving photocatalytic activity because of its suitable structure for spatial separation of electron–hole pairs In the past several decades, enormous efforts have been made Adv Mater 2017, 29, 1601694 to prepare different type-II heterojunction photocatalysts, such as TiO2/g-C3N4,[55] BiVO4/WO3,[56] g-C3N4–WO3,[57] g-C3N4– BiPO4,[58] and so on, for enhancing the photocatalytic activity Generally, type-II heterojunction photocatalysts exhibit good electron–hole separation efficiency, wide light-absorption range, and fast mass transfer.[59] For example, Zhou et al prepared a SnO2/TiO2 type-II heterojunction photocatalyst by an electrophoretic-deposition (EPD)– calcination method for photocatalytic RhB degradation.[41] Specifically, a commercial TiO2 was first electrophoretically deposited on F-doped SnO2-coated glass, followed by calcination at 200, 300, 400, 500 and 600 °C to obtain crystallized SnO2/TiO2 type-II heterojunction photocatalyst films It was found that all the prepared samples show good photocatalytic activities due to the fast electron–hole separation through the type-II heterojunction between the TiO2 and the SnO2 Particularly, the sample prepared at 400 °C exhibited the highest photocatalytic activity among all the samples studied This high activity could be attributed to the optimal crystallinity and specific surface area, which can reduce the number of recombination centers on the sample and provide larger surface area with active sites for photocatalytic reaction Meanwhile, Wetchakun et al reported a hydrothermal synthesis of BiVO4/CeO2 type-II heterojunction photocatalysts for the photocatalytic degradation of methylene blue (MB), methyl orange (MO), and a mixture of MB and MO.[60] It was found that the the BiVO4 and the CeO2 possessed different isoelectric points, located at pH values of 4.56 and 7.33 respectively The isoelectric-point difference for these two semiconductors was shown to be beneficial for adsorbing both cationic and anionic dyes simultaneously Specifically, the BiVO4 and the CeO2 can preferentially respectively adsorb cationic MB and anionic MO © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim wileyonlinelibrary.com (4 of 20) 1601694 www.advancedsciencenews.com Review www.advmat.de Figure 4. Schematic illustration of the photocatalytic-activity enhancement mechanism of: (a) Ag/AgCl/pCN type-I, and b) Ag/AgBr/pCN type-II heterojunction photocatalysts under light irradiation Reproduced with permission.[61] Copyright 2014, Elsevier Moreover, the morphology tuning of heterojunction phoduring degradation reactions As a result, the BiVO4/CeO2 tocatalysts is also critical for optimizing the performance of composite exhibited higher photocatalytic-degradation activity type-II heterojunction photocatalysts For example, Shen and toward the mixture of MB and MO as compared with the single co-workers reported visible-light-active CdS/ZnO core/shell BiVO4 or CeO2 photocatalysts, which were not active toward the nanofibers with a type-II heterojunction (see Figure 5a,b), with photocatalytic degradation of the anionic MO and the cationic good photocatalytic activity toward hydrogen production.[10] The MB, respectively, due to the electrostatic repulsion between their surface charges and the charges of the dye molecules This type-II heterojunction significantly facilitated the electron–hole remarkable activity of the composite photocatalyst is attributed separation efficiency between the ZnO and the CdS Meanto the enhanced electron–hole separation efficiency and strong while, the core–shell structure of the CdS/ZnO featured a large electrostatic attraction between the composite and the dye contact interface, which can further enhance the separation molecules This study suggests that the proper coupling of two efficiency of electron–hole pairs As a result, the photocatalytic different semiconductors can not only enhance the electron– hydrogen-production efficiency of ZnO/CdS was higher than hole separation efficiency but can also afford photocatalysts with that of the single ZnO and CdS compounds good adsorption ability toward both anionic and cationic dyes Furthermore, type-II heterojunction photocatalysts can Ong et al systematically investigated the photocatalytic be also created between two different phases of a semiconactivity of type-I Ag/AgCl/g-C3N4 and type-II Ag/AgBr/gductor.[62–64] For instance, the mixed phase of TiO2, i.e., P25 C3N4 heterojunction photocatalysts for photocatalytic CO2 (consisting of anatase and rutile TiO2), has been commercialreduction.[61] It was found that both Ag/AgCl/g-C3N4 and Ag/ ized and has attracted wide attention for various photocatalytic applications due to its efficient charge-carrier separation In AgBr/g-C3N4 exhibited good photocatalytic CO2-reduction perdetail, the CB and VB levels of anatase TiO2 are higher than formance toward CH4 production due to the presence of the heterojunction, which may improve the charge-carrier sepathe corresponding levels of rutile TiO2 Therefore, a type-II ration in these photocatalysts Notably, the photocatalytic CO2 reduction activity of Ag/ AgBr/g-C3N4 toward CH4 production was much higher than that of Ag/AgCl/g-C3N4 This was ascribed to the formation of the type II-heterojunction instead of the type-I heterojunction in the case of Ag/AgBr/gC3N4, which can lead to the spatial separation of electrons and holes by accumulating them in the Ag/AgBr and the g-C3N4, respectively (see Figure 4) This study proved that the type-II heterojunction in photocatalysts is more effective than the type-I heterojunc- Figure 5. a,b) Scanning electron microscopy (SEM) (a) and transmission electron microscopy tion for improving their photocatalytic CO2- (TEM) (b) images of CdS/ZnO core/shell nanofibers with a type-II heterojunction Reproduced reduction activity with permission.[10] Copyright 2013, The Royal Society of Chemistry 1601694 (5 of 20) wileyonlinelibrary.com © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Adv Mater 2017, 29, 1601694 www.advmat.de www.advancedsciencenews.com Review Figure 7. Schematic illustration for the transfer and separation of charge carriers on an anatase/rutile/Ag heterojunction photocatalyst under light irradiation Figure 6. Comparison of the rate constant of anatase–brookite composites: with 80% anatase and 20% brookite (A), with 92% anatase and 8% brookite (B), with 100% anatase (C); as well as P25, in the photocatalytic degradation of acetone in air Reproduced with permission.[24] Copyright 2001, The Royal Society of Chemistry heterojunction photocatalyst can be formed by combining the anatase and rutile TiO2 in P25, which results in an enhancement of the electron–hole separation Also, in 2001 our group reported an anatase–brookite dual-phase type-II heterojunction photocatalyst by hydrolysis of titanium tetraisopropoxide in water or 1:1 ethanol–H2O solution.[24] It was found that the content of brookite in the composite was reduced in the presence of ethanol because the latter can suppress the hydrolysis of the titanium alkoxide and inhibit the rapid crystallization of the TiO2 nanoparticles into brookite by adsorbing on the surface of the TiO2 Notably, the photocatalytic activity of the sample consisting of anatase and brookite toward degradation of acetone was higher than that of one consisting of pure anatase (see Figure 6) This is due to the formation of a type-II heterojunction between the brookite and the anatase, which can greatly enhance the separation efficiency of the electron–hole pairs In order to further enhance the photocatalytic activity of the multi-phase type-II heterojunction photocatalysts, an Ag–TiO2 multiphase composite was designed and synthesized by Yu et al.[40] It was found that the silver precursor, AgNO3, has a great influence on the crystallization of the TiO2 The phase composition of the TiO2 changed with changing AgNO3 concentration Particularly, brookite TiO2 started to appear and steadily grow, starting from a AgNO3 concentration of 0.02 M, because the AgNO3 facilitates and catalyzes the formation of brookite TiO2 Meanwhile, the rutile TiO2 began to be formed at a AgNO3 concentration of 0.03 M Notably, the phase-transformation temperature of TiO2 from anatase to rutile was significantly reduced from 700 °C to 500 °C due to the presence of AgNO3 Thus, AgNO3 can suppress the growth of anatase domains and thus increase the total boundary energy of TiO2 by promoting the phase transformation of anatase to rutile It Adv Mater 2017, 29, 1601694 was found that the optimal Ag–TiO2 (0.05 M of AgNO3) sample exhibited very high photocatalytic activity for MO degradation This is due to the formation of multiphase type-II heterojunctions, such as anatase/rutile, anatase/brookite and rutile/ brookite, which can greatly enhance the electron–hole separation efficiency Moreover, Ag loading can also further enhance the charge-carrier separation in the Ag–TiO2 composite Taking anatase/rutile/Ag as an example (see Figure 7), the photogenerated electrons and holes can be spatially separated in rutile and anatase TiO2, respectively, through the type-II heterojunction Meanwhile, the electrons on the rutile TiO2 can further migrate to Ag nanoparticles (NPs) through the Schottky junctions between the Ag and the rutile TiO2, thus achieving high electron–hole separation efficiency Although type-II heterojunction photocatalysts possess good electron–hole separation efficiency, there are still some problems that limit their practical applications For example, the reduction and oxidation reactions on the type-II heterojunction photocatalysts take place on the semiconductors with the lower reduction and oxidation potentials, respectively, thereby greatly suppressing their redox ability (see Figure 3b) Moreover, the migration of electrons from semiconductor A to the electronrich CB of semiconductor B or the corresponding migration of holes from semiconductor B to the hole-rich VB of semiconductor A are physically unfavorable because of the electrostatic repulsion between electron and electron or hole and hole, respectively Therefore, the development of more effective heterojunctions in photocatalysts is urgently needed p–n Heterojunctions Although the type-II heterojunction can ideally separate electron–hole pairs in space, the achieved enhancement in the electron–hole separation across a type-II heterojunction is not sufficient to overcome the ultrafast electron–hole recombination on the semiconductor Thus, a p–n heterojunction photocatalyst concept was proposed, which is able to accelerate the electron– hole migration across the heterojunction for improving the photocatalytic performance by providing an additional electric field.[65–67] Specifically, an effective p–n heterojunction photocatalyst can be obtained by combining p-type and n-type semiconductors Before light irradiation, the electrons on the n-type © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim wileyonlinelibrary.com (6 of 20) 1601694 www.advancedsciencenews.com Review www.advmat.de Figure 8. Schematic illustration of the electron–hole separation under the influence of the internal electric field of a p–n heterojunction photocatalyst under light irradiation semiconductor near the p–n interface tend to diffuse into the p-type semiconductor, leaving a positively charged species (see Figure 8).[68–71] Meanwhile, the holes on the p-type semiconductor near the p–n interface tend to diffuse into the n-type semiconductor, leaving a negatively charged species The electron–hole diffusion will continue until the Fermi level equilibrium of the system is achieved As a result, the region close to the p–n interface is charged, creating a “charged” space or the so-called internal electric field.[72–74] When the p-type and n-type semiconductors are irradiated by incident light with an energy equal to or higher than their bandgap value, both p-type and n-type semiconductors can be excited, generating electron–hole pairs The photogenerated electrons and holes in the p-type and n-type semiconductors will migrate under the influence of the internal electric field to the CB of the n-type semiconductor and the VB of the p-type semiconductor, respectively, which results in the spatial separation of the electron–hole pairs It should be noted that this electron–hole separation process is also thermodynamically feasible because both the CB and the VB of the p-type semiconductor are normally located higher than those of the n-type semiconductor in a p–n heterojunction photocatalyst.[75,76] As a result, the electron–hole separation efficiency in p–n heterojunction photocatalysts is faster than that of type-II heterojunction photocatalysts due to the synergy between the internal electric field and the band alignment.[42] For example, our group has demonstrated that the design of NiS/CdS nanorods with p–n heterojunctions can greatly improve the photocatalytic hydrogen-production efficiency of the CdS.[29] NiS NPs with sizes of 10–30 nm were uniformly dispersed on the surface of CdS nanorods, thereby providing larger contact interface for creating p–n heterojunctions between the NiS and the CdS (see Figure 9a) The formation of p–n heterojunctions can facilitate charge transfer between the NiS and the CdS and suppress the charge-carrier recombination (see Figure 9b,c) Meanwhile, the CdS nanorods can enhance electron transport because of their one-dimensional (1D) structure As a result, the photocatalytic hydrogen-production rate on the NiS/CdS nanorods with p–n heterojunctions, having wt% NiS (56.6 µmol h−1) is much higher than that of the pure CdS (2.8 µmol h−1) and wt% Pt-loaded CdS (36.3 µmol h−1) (see Figure 9d) However, a further increase in the NiS loading on the CdS causes a decrease in the photocatalytic activity Figure 9. a) SEM image of wt% NiS-loaded CdS b,c) Schematic illustration of the charge-carrier separation on the NiS/CdS nanorods with p–n heterojunctions (b) and across the NiS/CdS p–n heterojunction (c) d) Comparison of the photocatalytic activity of CdS with different NiS loadings: Ni0 (0 wt% NiS), Ni0.5 (0.5 wt% NiS), Ni1 (1 wt% NiS), Ni3 (3 wt% NiS), Ni5 (5 wt% NiS), Ni10 (10 wt% NiS), wt% Pt–CdS, and pure NiS under visible-light irradiation Reproduced with permission.[29] Copyright 2013, The Royal Society of Chemistry 1601694 (7 of 20) wileyonlinelibrary.com © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Adv Mater 2017, 29, 1601694 www.advmat.de www.advancedsciencenews.com In 2014, our group proposed the concept of the surface heterojunction to explain the unique electron–hole separation phenomenon observed on the crystal facets of a single semiconductor.[31] It is well-known that the different crystal facets on a single semiconductor can have different band structures.[31,78] Since a heterojunction is formed by combining two semiconducting materials with different band structures, it is possible to create a hetFigure 10. a,b) SEM images of flower-like NiO (a), and a flower-like NiO/TiO2 p–n heterojuncerojunction between two crystal facets of a tion photocatalyst (b) with the inset showing high resolution Reproduced with permission.[42] single semiconductor, namely a surface hetCopyright 2010, Wiley-VCH erojunction.[31,79,80] In our work, a series of anatase TiO2 samples with different ratios of the exposed {001} because the shielding effect of the NiS may reduce the number of active sites on the surface and thus the light-absorption and {101} facets was reported These samples were prepared ability of the CdS by a F−-ions-assisted hydrothermal method The photocatalytic Furthermore, Feng and co-workers reported a MoS2/CdS activity of the sample with 55:45 ratio of the exposed {001} and {101} facets was much higher than that of the samples domip–n heterojunction photocatalyst for enhancing photocatanated either by exposed {001} or {101} facets According to denlytic hydrogen-production efficiency,[77] which was simply presity functional theory (DFT) calculations (see Figure 11a), the pared by a one-pot solvothermal method It was shown that a observed enhancement in the photocatalytic activity was due to “V-shaped” Mott–Schottky plot can be obtained by performing the formation of surface heterojunctions between the {001} and electrochemical testing, which suggests the formation of p–n {101} facets In fact, the basic principle of the surface heteroheterojunctions by the combination of the p-type MoS2 and the junction is similar to that of the type-II heterojunction, in which n-type CdS It was found that the n-type CdS is uniformly disthe CB and VB levels of the {001} facets are higher than the corpersed on the surface of nanosheets of the p-type MoS2, which responding levels of the {101} facets of anatase TiO2 Thus, elecresults in a larger contact interface between the CdS and the MoS2, being beneficial for accelerating charge transfer and septrons and holes can be spatially separated on the {101} facets for reduction reactions and on the {001} facets for oxidation aration Therefore, the MoS2/CdS p–n heterojunction photocatreactions, respectively (see Figure 11b) This finding enables the alyst exhibited an excellent electron–hole separation efficiency design of heterojunction systems into the surface of single nandue to its large contact interface and rapid electron–hole sepaoparticle In addition, the fabrication cost of the resulting hetration through the formed p–n heterojunctions Consequently, erojunction photocatalysts can be greatly reduced because only the 2D MoS2/CdS p–n heterojunction photocatalyst exhibited a one semiconductor is used Notably, the redox potential loss of hydrogen-production rate of 137 µmol h−1, which is 10 times a type-II heterojunction photocatalyst can be also minimized higher than that obtained for pure CdS This work shows by creating a surface heterojunction, because the difference in that an enlarged contact interface on the p–n heterojunction the band structures between the {001} and {101} facets of TiO2 photocatalysts is beneficial for enhancing their photocatalytic performance is small Therefore, the photocatalytic CO2 reduction activity of The morphology of the p–n heterojunction photocatalysts anatase TiO2 with an optimal ratio of exposed {001} and {101} can be tuned for achieving large specific surface area and an facets was 3.5 times higher than that of commercial TiO2, i.e., abundant number of surface active sites, which has been P25 (Figure 11c) Furthermore, it was observed that the overproved to be effective for further improvement of the photocataexposed {101} or {001} facets on the TiO2 caused an overflow lytic activity of p–n heterojunction photocatalysts Recently, our effect of holes and electrons (see Figure 12a,c), respectively, and group showed that a flower-like NiO/TiO2 p–n heterojunction consequently reduced the photogenerated electron–hole separation efficiency Apparently, the fabrication of anatase TiO2 with photocatalyst exhibited an exceptional photocatalytic activity toward degradation of p-chlorophenol.[42] The flower-like NiO/ the optimal ratio of the exposed {001} and {101} facets is crucial for enhancing the photocatalytic activity of the anatase TiO2 (see TiO2 p–n heterojunction was afforded by a simple hydrothermal method Specifically, TiO2 NPs were uniformly deposited on Figure 12b).[31,81–84] the surface of flower-like NiO particles, which resulted in the Selective deposition of oxidation and reduction co-catalysts formation of a NiO/TiO2 p–n heterojunction photocatalyst (see on a semiconductor has been widely applied to confirm the type of oxidation and reduction sites thereon.[85] Recently, Liu et al Figure 10a,b) The resulting photocatalyst featured a large specific surface, which assured the abundance of surface active reported that the spatial separation of electrons and holes on the sites for photocatalytic reactions and enhanced the adsorption different facets of a single anatase TiO2 crystal can be confirmed ability for dye molecules In addition, it was shown that the forby photo-deposition of Pt NPs Specifically, Pt NPs were selecmation of p–n heterojunctions in NiO/TiO2 can improve the tively loaded on the surface of the electron-rich {101} facets due to the spatial separation of electrons and holes on the anatase electron–hole separation rate As a result, the NiO/TiO2 photo TiO2.[86] Moreover, they found that an optimal ratio of the {001} catalyst exhibited a superior photocatalytic activity toward the degradation of p-chlorophenol and {101} facets is also an important issue for photocatalytic Adv Mater 2017, 29, 1601694 © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim wileyonlinelibrary.com (8 of 20) 1601694 Review Surface Heterojunctions Review www.advmat.de www.advancedsciencenews.com photocatalytic activity Specifically, 1, 2, and mL of H2PtCl6 solution were used to prepare a series of TiO2/Pt photocatalysts, named as TP1, TP2, and TP4, respectively Interestingly, Pt NPs were selectively deposited on the surface of the {101} and {010} facets of anatase, which was due to the formation of surface heterojunctions Particularly, {001} and {110} have a higher CB value, while {101} and {010} have a lower VB value The electrons and holes tend to migrate to the {101} and {010} facets for reduction reactions and to {001} and {110} facets for oxidation reactions, respectively Therefore, the Pt ions are accumulated on the {101} and {010} facets and reduced into Pt NPs by photogenerated electrons during the photoreduction process Due to the synergistic effect of surface heterojunctions and Schottky junctions, the photocatalytic activity of the resulting TP1 toward the degradation of phenol was much higher than that of pure TiO2 However, a further increase in the loading of Pt NPs (e.g., samples TP2 and TP4) led to the light-shielding effect, which may significantly reduce the photocatalytic activity of these samples Furthermore, Li et al reported BiVO4 with co-exposed {010} and {110} facets for the spatial separation of photogenerated electrons and holes.[87] It was found that the {010} and {110} facets exhibited different band structures, in which the CB of the {110} facets is higher than that of the {010} facets, while Figure 11. a) Density of states (DOS) plots of the {101} and {001} facets of anatase TiO2, where O 2P, Ti 3d, and TDOS are the partial DOS of O2P, partial DOS of Ti 3d, and total DOS, the VB of the {010} facets is lower than that of the {110} facets Therefore, electrons and respectively b) Electron–hole separation on the surface heterojunction of anatase TiO2 with an optimal ratio of the exposed {001} and {101} facets (55:45) c) Comparison of the photocata- holes tend to migrate to the {010} and {110} lytic CH4 production activity of P25, HF0 (anatase TiO2 with a ratio of the exposed {001} and facets, respectively, resulting in their spa{101} facets equal to 11:89), HF3 (anatase TiO2 with the ratio of 49:51), HF4.5 (anatase TiO2 tial separation In order to further improve the ratio of 55:45), HF6 (anatase TiO2 with the ratio of 72:28), and HF9 (anatase TiO2 with the the photocatalytic activity of such samples, [31] ratio of 83:17) Reproduced with permission Copyright 2014, American Chemical Society reduction (Pt) and oxidation (MnOx) cocatalysts were loaded onto BiVO4 by the photo-deposition method It was found that Pt and MnOx cohydrogen production by water splitting Namely, the photocatalytic hydrogen-production activity of the sample with an catalysts are preferably loaded on the {010} and {110} facets optimal ratio of the exposed {001} and {101} facets and 0.5% (see Figure 13a,b), respectively This is because Pt and MnOx loading of Pt was ca times higher than that of the corresponding sample dominated by {101} facets Furthermore, Gao et al prepared anatase TiO2 with exposed {101}, {010}, {001}, and {110} facets by using a simple hydrothermal method to create surface heterojunctions on the anatase TiO2 for photocatalytic phenol degradation.[33] Transmission electron microscopy (TEM) images of these samples suggested that the anatase TiO2 with exposed {101}, {010}, {001}, and {110} Figure 12. Schematic illustrations of: a) overflow effect of holes on the anatase TiO2 dominated facets was successfully fabricated Then, difby {101} facets, b) spatial separation of electrons and holes on anatase TiO2 with an optimal ferent amounts of Pt NPs were loaded on the ratio of the exposed {001} and {101} facets, and c) overflow effect of electrons on anatase as-prepared anatase TiO2 to form Schottky TiO2 dominated by {001} facets Reproduced with permission.[31] Copyright 2014, American junctions for further improvement of the Chemical Society 1601694 (9 of 20) wileyonlinelibrary.com © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Adv Mater 2017, 29, 1601694 www.advmat.de www.advancedsciencenews.com Review Figure 13. a,b) SEM images of BiVO4 with deposited Pt (a) and MnOx (b) Reproduced with permission.[87] Copyright 2013, Nature Publishing Group can be easily formed on the photogenerated electron-rich surface of the {010} facets and the photogenerated hole-rich surface of the {110} facets, respectively By adjusting the loadings of the reduction and oxidation co-catalysts, the electrons on the {010} facets and the holes on the {110} facets of the BiVO4 can further migrate to Pt and MnOx, respectively, which results in accelerating the electron–hole separation Due to the spatial separation of the electron–hole pairs and the proper loading of reduction and oxidation co-catalysts, the resulting photocatalyst exhibited a high photocatalytic activity for water oxidation Direct Z-Scheme Heterojunctions Although all the above-mentioned heterojunction photocatalysts are efficient for enhancing electron–hole separation, the redox ability of the photocatalyst is sacrificed because the reduction and oxidation processes occur on the semiconductor with the lower reduction and oxidation potentials, respectively.[88–91] In order to overcome this problem, the Z-scheme photocatalytic concept was proposed by Bard et al in 1979 to maximize the redox potential of the heterojunction systems.[92] A conventional Z-scheme photocatalytic system is composed of two different semiconductors, photocatalyst I (PS I) and photocatalyst II (PS II), and an acceptor/donor (A/D) pair (see Figure 14) PS I and PS II are not in physical contact During the photocatalytic reaction, photogenerated electrons migrate from the CB of the PS II to the VB of the PS I through an A/D pair via following redox reactions: A + e− → D (3) D + h+ → A (4) Specifically, A is reduced into D by reacting with the photogenerated electrons from the CB of the PS II After that, the D is oxidized into A by the photogenerated holes from the VB Since electrons accumulate on the PS I, with the higher reduction potential, and holes accumulate on the PS II, with the higher oxidation potential, a spatial separation of electron–hole pairs and an optimal redox ability can be achieved However, conventional Z-scheme photocatalysts can only be constructed Adv Mater 2017, 29, 1601694 in the liquid phase, thereby limiting their wide application in photocatalysis In 2006, Tada et al proposed the concept of an all-solidstate Z-scheme photocatalyst, which consisted of two different semiconductors (PS I and PS II) and a solid electron mediator between them.[93] As shown in Figure 15, electrons on the VB of the PS II are firstly excited to the CB under light irradiation, leaving holes on the VB Then, the photogenerated electrons on the PS II migrate to the VB of the PS I via an electron mediator (such as Pt, Ag, and Au), and are further excited to the CB of PS I As a result, photogenerated holes and electrons are accumulated in the PS II, with a higher oxidation potential, and in the PS I, with a higher reduction potential, respectively, which results in the spatial electron–hole separation and optimization of the redox potential Moreover, all-solid-state Z-scheme photocatalysts can be used in solution, gas, and solid media, thereby extending their photocatalytic applications.[94–96] However, electron mediators required for improving the migration path for electrons in the all-solid-state Z-scheme photocatalysts are expensive and rare, which limits large-scale applications of these photocatalysts In 2013, our group proposed a direct Z-scheme heterojunction photocatalyst concept A direct Z-scheme photocatalyst was prepared by combining two different semiconductors without an electron mediator.[35] As shown in Figure 16, the construction of this direct Z-scheme heterojunction photocatalyst is identical to that of conventional all-solid-state Z-scheme Figure 14. Schematic illustration of electron–hole separation on the conventional Z-scheme photocatalytic system under light irradiation © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim wileyonlinelibrary.com (10 of 20) 1601694 www.advancedsciencenews.com Review www.advmat.de Figure 15. Schematic illustration of the electron–hole separation on allsolid-state Z-scheme photocatalysts under light irradiation heterojunction photocatalysts, except that the rare and expensive electron mediators are not required in this system.[35,97–99] Similarly, electrons and holes are spatially separated on the semiconductor with the higher reduction potential and oxidation potential of the direct Z-scheme heterojunction photocatalyst, respectively.[100–102] Furthermore, the fabrication cost of this direct Z-scheme heterojunction photocatalyst is low and comparable to that of conventional type-II heterojunction systems.[103] Also, it shows other advantages; for instance, its redox potential can be optimized for specific photocatalytic reactions Moreover, the charge transfer on the direct Z-scheme heterojunction photocatalyst is physically more favorable than that on the type-II heterojunction photocatalyst because of the electrostatic attraction between electrons and holes In particular, in the case of the direct Z-scheme photocatalysts, the migration of photogenerated electrons from the CB of the PS II to the photogenerated hole-rich VB of the PS I is easier, due to the electrostatic attraction between the electrons and the holes In contrast, for conventional type-II heterojunction photocatalysts, the migration of photogenerated electrons from the CB of semiconductor A to the photogenerated electron-rich CB of semiconductor B is apparently harder due to the electrostatic repulsion between electrons (see Figure 3b) Due to the aforementioned advantages, direct Z-scheme heterojunction photocatalysts have recently attracted a lot of attention For instance, our group reported the photocatalytic decomposition of the major indoor pollutant formaldehyde on a g-C3N4– TiO2 direct Z-scheme heterojunction photocatalyst.[35] Specifically, a series of the samples was prepared by calcining mixtures of P25 titania with different amounts of urea As can be seen from Figure 17a, the sizes of the TiO2 NPs were ca 30 nm More interestingly, the TiO2 NPs were covered by g-C3N4 Therefore, an intimate contact between the TiO2 and the g-C3N4 was created, enabling a rapid transport of charge carriers across the contact interface The actual loading of g-C3N4 on TiO2 was further confirmed by TGA analysis The pure TiO2 exhibited no weight loss on the TG curve, while the g-C3N4 was fully decomposed at 600 °C Furthermore, the U100 sample exhibited about 12% loss at 600 °C, indicating that the actual loading of the g-C3N4 on the TiO2 was ca 12% Then, the presence of the direct Z-scheme heterojunction was also confirmed by radicaltrapping experiments Specifically, an increase in the photoluminescence intensity of 2-hydroxy-terephthalic acid (•OH-TA) 1601694 (11 of 20) wileyonlinelibrary.com Figure 16. Schematic illustration of electron–hole separation on a direct Z-scheme heterojunction photocatalyst under light irradiation at about 425 nm with increasing irradiation time is visible (see Figure 17b), indicating the generation of the •OH radical If a type-II heterojunction is formed by coupling the g-C3N4 and the TiO2, instead of a direct Z-scheme heterojunction, no increase in the PL intensity should be observed because the photogenerated holes are accumulated on the valence band of g-C3N4, which does not have sufficient oxidation power for producing •OH radicals (see Figure 17c,d) The observed production of •OH radicals confirms the accumulation of photogenerated holes on the valence band of the TiO2, suggesting the formation of the g-C3N4–TiO2 direct Z-scheme heterojunction As a result, the photocatalytic formaldehyde-decomposition activity of the optimized g-C3N4–TiO2 photocatalyst (12% g-C3N4-loaded TiO2) is 2.1 times higher than that of commercial TiO2, i.e., P25 (see Figure 18a,b) Moreover, it was shown that the photocatalytic activity of the g-C3N4-TiO2 direct Z-scheme heterojunction is greatly dependent on the amount of g-C3N4 loaded on the TiO2 An overloading of g-C3N4 on the TiO2 led to a decrease in the photocatalytic activity (see Figure 18c,d) This is due to the shielding effect of the g-C3N4 on the TiO2, which causes a reduction in the light-absorption ability of the TiO2 and inhibits the reaction between the photogenerated holes in the TiO2 and reactants In 2014, Katsumata et al fabricated a WO3/g-C3N4 Z-scheme heterojunction photocatalyst for hydrogen production under visible-light irradiation by a simple calcination method.[104] In this direct Z-scheme heterojunction system, the photogenerated electrons migrated from the WO3 to the g-C3N4, while the photogenerated holes were kept on the WO3 during the photocatalytic reaction Thus, the reduction and oxidation reactions took place on the g-C3N4, having higher reduction potential, and on the WO3, having higher oxidation potential, respectively, thereby optimizing the redox ability of the composite photocatalyst As a result, the photocatalytic hydrogenproduction rate of the WO3/g-C3N4 Z-scheme heterojunction photocatalyst is much higher than that of the pure WO3 and g-C3N4 components Very recently, our group reported the synthesis of a hierarchical CdS–WO3 direct Z-scheme heterojunction photocatalyst in the form of hollow spheres (CdS–WO3 HSs) for photocatalytic CO2 reduction.[44] According to the photocurrent characterization, the electron–hole separation efficiency of CdS–WO3 HSs is higher than that of the pure CdS and WO3 HSs This © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Adv Mater 2017, 29, 1601694 www.advmat.de www.advancedsciencenews.com Review Figure 17. a) TEM image of g-C3N4–TiO2 direct Z-scheme heterojunction photocatalyst b) •OH-TA PL spectral changes of the same photocatalyst in a 0.002 M NaOH solution in the presence of 0.0005 M terephthalic acid under UV irradiation c) Schematic illustration of the band structures of g-C3N4 and TiO2 together with the OH−/•OH potential d) The •OH radicals cannot be formed on the conventional g-C3N4–TiO2 heterojunction photocatalyst Reproduced with permission.[35] Copyright 2013, The Royal Society of Chemistry Figure 18. a) Comparison of the photocatalytic HCHO decomposition activity of U0 (pure TiO2), U20 (TiO2 calcined with 20 wt% urea), U100 (TiO2 calcined with 100 wt% urea), U200 (TiO2 calcined with 200 wt% urea), U500 (TiO2 calcined with 500 wt% urea), and pure g-C3N4 samples b–d) Schematic illustrations of the electron–hole separation on the g-C3N4-TiO2 direct Z-scheme heterojunction photocatalysts (b), g-C3N4-TiO2 with optimal loading of g-C3N4 (c), g-C3N4-TiO2 with overloaded amount of g-C3N4 (d) Reproduced with permission.[35] Copyright 2013, The Royal Society of Chemistry Adv Mater 2017, 29, 1601694 © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim wileyonlinelibrary.com (12 of 20) 1601694 www.advancedsciencenews.com Review www.advmat.de Figure 19. Schematic illustration of the charge-transfer pathway of the photogenerated electron–hole pairs on CdS–WO3 direct Z-scheme heterojunction photocatalyst under visible-light irradiation is due to the spatial separation of the electrons and holes on the CB of the CdS and the VB of the g-C3N4, respectively (see Figure 19) Moreover, the formation of the CdS–WO3 HS direct Z-scheme heterojunction can also promote electron accumulation on the surface of the CdS, which is beneficial for the multielectron CO2-reduction reaction Therefore, the main product of the photocatalytic CO2 reduction reaction on the CdS–WO3 HS Z-scheme heterojunction photocatalyst was CH4 (see Figure 20a) As a result, the photocatalytic CO2 reduction activity of the optimized CdS–WO3 HSs (0.1 µmol h−1 of CO2 were reduced to CH4 on the catalyst with mol% CdS) was 100 and 10 times higher than the corresponding values obtained for the pure WO3 HS and CdS (see Figure 20b), respectively However, an overloading of CdS led to a decrease in the photocatalytic CO2 activity, which can be also related to the observed decrease in the number of surface active sites on the WO3 HSs and the light-shielding effect Furthermore, Wong and co-workers reported that a low-cost natural mineral can be used for constructing a direct Z-scheme heterojunction photocatalyst.[105] Specifically, natural magnetic pyrrhotite (NP) mineral, after thermal treatment at 600 °C, was used as a photocatalyst It was shown that the thermally treated NP (TNP) was composed of hematite (Fe2O3) and pyrite (FeS2) Notably, the resulting TNP exhibited good photocatalytic activity for Escherichia coli disinfection in comparison to that of the initial NP This activity was attributed to the formation of the direct Z-scheme heterojunction between the Fe2O3 and the FeS2, greatly suppressing electron–hole recombination and optimizing the redox ability of the system In addition, during thermal treatment, oxygen or sulfur vacancies can be generated and utilized as electron traps prohibiting the electron–hole recombination More interestingly, the presence of Fe2O3 in the TNP direct Z-scheme heterojunction photocatalyst resulted in its strong magnetic behavior, thereby enabling the simple magnetic recovery of the photocatalyst, which is beneficial for longterm application More recently, Xu et al showed that the direct Z-scheme heterojunction can also be constructed between two different semiconductor phases They successfully prepared an anatase/ rutile biphase electrospun TiO2 as a direct Z-scheme heterojunction photocatalyst through the rapid cooling of electrospun TiO2 calcined at 500 °C (see Figure 21a,b).[43] As can be seen from Figure 21b,c, TiO2 nanofibers with both rutile and anatase phases were successfully prepared Normally, the prepared TiO2 nanofibers are dominated by anatase due to the fast transformation rate of rutile into anatase during slow cooling Notably, the proposed rapid-cooling method was shown to be effective for suppressing the transformation rate of rutile to anatase, and thus the rutile TiO2 phase can be preserved after the cooling process Moreover, it was shown that during the photo-deposition process of Pt NPs, the Pt NPs were favorably deposited on the surface of rutile instead of anatase (see Figure 21b), suggesting that the photogenerated electrons tend to migrate to the CB of the rutile phase due to the formation of the direct Z-scheme heterojunction Furthermore, the electron–hole separation efficiency of the sample prepared by a rapid cooling method was better than that of the TiO2 nanofiber prepared by slow cooling because, in the former case, the direct Z-scheme heterojunctions on the TiO2 nanofibers can greatly reduce the electron–hole-recombination rate Since the reduction and oxidation reactions occur on the semiconductors with the higher reduction and oxidation potentials, respectively, the redox ability Figure 20. a) Gas-chromatography spectra for the CO2 reduction on CdS–WO3 hollow spheres (HS) at different irradiation times b) Photocatalytic CO2 reduction activity of C0 (pure WO3 HS), C1 (1 mol% of CdS loaded on WO3 HS), C2 (2 mol% of CdS loaded on WO3 HS), C5 (5 mol% of CdS loaded on WO3 HS), C10 (10 mol% of CdS loaded on WO3 HS), C20 (20 mol% of CdS loaded on WO3 HS), C100 (pure CdS), and N5 (5 mol% CdS–WO3 nanoparticle composite) for CH4 production Reproduced with permission.[44] Copyright 2015, Wiley-VCH 1601694 (13 of 20) wileyonlinelibrary.com © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Adv Mater 2017, 29, 1601694 www.advmat.de www.advancedsciencenews.com Review Figure 21. a,b) TEM (a) and HRTEM (b) images of the mixed-phase composite of anatase and rutile TiO2 with photo-deposited Pt NPs c) X-ray diffraction (XRD) patterns of anatase–rutile composites: I (pure anatase), II (with 72% of anatase and 28% of rutile), III (with 55% of anatase and 45% of rutile), and IV (pure rutile) d) Comparison of the photocatalytic-hydrogen-production activity of the anatase–rutile composites I, II, III, and IV Reproduced with permission.[43] Copyright 2014, Elsevier of the composite can be also optimized As a result, the photocatalytic hydrogen-production efficiency of TiO2 nanofibers prepared by the rapid-cooling method (324 µmol h−1) is higher than that of the TiO2 nanofibers prepared under slow-cooling conditions (188 µmol h−1) (see Figure 21d) Semiconductor/Graphene Heterojunctions Graphene, a 2D single layer sheet of sp2-hybridized carbon atoms with hexagonally packed structure, has attracted great attention due to its extraordinary physical properties including superior charge transport, unique optical properties, high thermal conductivity, large theoretical specific surface area, and good mechanical strength.[106–108] Since the discovery of singlelayer graphene nanosheets by Geim and Novoselov in 2004,[109] graphene has been known as a perfect candidate for various applications, including solar cells,[110–112] supercapacitors,[113,114] batteries,[115,116] photocatalysis,[117–119] etc Particularly, tremendous efforts have been made for coupling graphene with other semiconductors to fabricate heterojunction photocatalysts with improved photocatalytic activity The ultrahigh electron conductivity of graphene allows the flow of electrons from the semiconductor to its surface, assuring efficient electron–hole separation Moreover, the potential of graphene/ graphene− (−0.08 V vs standard hydrogen electrode (SHE), pH = 0) is normally lower than the conduction-band potential of Adv Mater 2017, 29, 1601694 the photocatalyst, thereby enabling the fast electron migration from the photocatalyst to the graphene.[38] The use of graphene in photocatalytic applications was firstly proposed by Zhang et al in 2010.[120] It was shown that the photocatalytic activity of P25 titania toward MB degradation was greatly enhanced by adding graphene, which was attributed to the rapid separation of charge carriers across the P25–graphene heterojunction Furthermore, the large specific surface area of graphene is also beneficial for providing a greater number of surface active sites for photocatalytic reactions In 2011, our group firstly reported the CdS–graphene heterojunction system for photocatalytic hydrogen production.[37] Specifically, a CdS/reduced graphene oxide (CdS/RGO) composite was prepared by the simple solvothermal method Prior to the reaction, graphene oxide (GO) and cadmium acetate were mixed in dimethyl sulfoxide (DMSO) Since GO has a negatively charged surface at pH = 7, the Cd+ ions from the cadmium acetate in a DMSO solution can be adhered to the surface of GO because of the electrostatic attraction between the Cd+ ions and the GO Then, during the solvothermal process, H2S is produced from the DMSO and reacts in situ with the Cd+ ions to create CdS NPs on the GO surface Simultaneously, in the presence of DMSO, GO can be reduced to RGO during the solvothermal process The in situ growth of CdS nanoparticles on RGO can not only create an intimate contact between the CdS and the RGO, assuring rapid migration of electrons from the CdS to the RGO, but also limits the © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim wileyonlinelibrary.com (14 of 20) 1601694 www.advancedsciencenews.com Review www.advmat.de Figure 22. a) TEM images of the graphene/TiO2 NS at low resolution and high resolution (inset) b) Light-absorption spectra of G0 (pure TiO2 NS), G0.2 (0.2 wt% graphene on TiO2 NS), G0.5 (0.5 wt% graphene on TiO2 NS), G1.0 (1.0 wt% graphene on TiO2 NS), G2.0 (2.0 wt% graphene on TiO2 NS), and G5.0 (5.0 wt% graphene on TiO2 NS) c) Comparison of the photocatalytic-hydrogen-production activity of the G0, G0.2, G0.5, G1.0, G2.0, G5.0, and P1.0 (1.0 wt% Pt on TiO2 NS) d) Schematic illustration of the photocatalytic mechanism on graphene/TiO2 NS Reproduced with permission.[38] Copyright 2013, The Royal Society of Chemistry growth of CdS and consequently enlarges the specific surface area of the CdS/RGO composite As a result, the photocatalytic hydrogen-production activity of the CdS/RGO composite was 4.87 times higher than that of CdS with added Pt After that, our group further investigated the effect of graphene on the photocatalytic activity of TiO2 by preparing graphene/TiO2 nanosheet composites (graphene/TiO2 NS).[38] As shown in the TEM images of the sample (see Figure 22a), the TiO2 NSs were face-to-face dispersed on the surface of the graphene, enabling the intimate contact between the TiO2 NSs and the graphene Therefore, electrons can rapidly migrate from the TiO2 NSs to the surface of the graphene The lightabsorption ability of the TiO2 NSs was also greatly enhanced by adding graphene The graphene/TiO2 NS composite exhibited a broad background absorption in the visible-light region (see Figure 22b), which can be attributed to the eV bandgap of graphene Although this enhanced light absorption by graphene cannot generate any active charge carriers for redox reaction, the absorbed light can be used for producing heat and creating a unique photothermal effect around the photocatalyst surface This effect is favorable for accelerating charge transport.[118] A similar photothermal effect can be also observed in the case of the ZnxCd1−xS/graphene,[121] Cu2O/graphene,[122,123] g-C3N4/ graphene,[124] Bi2WO6/graphene,[125] and other composites Therefore, the transient photocurrent density of grapheneloaded TiO2 NSs is higher than that of pure TiO2 NSs, indicating high electron–hole separation efficiency on the former NSs As a result, the 1.0 wt% graphene-loaded TiO2 NSs exhibited the highest photocatalytic activity (36.8 µmol h−1) among 1601694 (15 of 20) wileyonlinelibrary.com all the sample studied, due to the rapid migration of electrons from the TiO2 NSs to the graphene (see Figure 22c,d) However, if the graphene content is higher than 1.0 wt% in the composite, the photocatalytic activity of such composites is drastically reduced, because the graphene overloading leads to the light-shielding effect, which greatly suppresses the light absorption of TiO2 NS Therefore, it should be noted that finding the optimal content of graphene on the semiconductor is crucial for enhancing its photocatalytic activity In addition, a large π–π conjugation on the graphene surface can be also utilized for the adsorption of different reactants during the photocatalytic reaction For example, our group prepared a Bi2WO6/graphene/Ag (Bi2WO6/G/Ag) composite for photocatalytic RhB degradation by a hydrothermal–photoreduction method.[20] It was shown that the addition of graphene can significantly enhance adsorption of the aforementioned dye pollutant In particular, the adsorption ability of the prepared samples was tested by analyzing the concentration changes of the RhB solution in the presence of the different samples, including Bi2WO6, Bi2WO6/Ag, Bi2WO6/G, and Bi2WO6/G/Ag, under dark conditions It was shown that the samples with added graphene exhibited higher RhB adsorption ability (see Figure 23a) due to a large π–π conjugation on the graphene surface, which could easily adsorb RhB molecules through the strong π–π interactions between the graphene and the RhB molecules (see Figure 23b) Dye pollutants other than RhB, such as methylene blue, methyl violet, methyl green, and so on, can be adsorbed on graphene as well.[126–128] © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Adv Mater 2017, 29, 1601694 www.advmat.de www.advancedsciencenews.com Review Figure 23. a) RhB concentration changes with time in the presence of Bi2WO6, Bi2WO6/Ag, Bi2WO6/G, and Bi2WO6/G/Ag under dark conditions, where C0 and C represent the initial and actual (at a given time) concentrations of the RhB solution b) Schematic illustration of the π–π interactions between graphene and RhB molecules Reproduced with permission.[20] Copyright 2014, The Royal Society of Chemistry Meanwhile, the adsorption ability of the graphene toward CO2 molecules was also confirmed by Yu et al.[46] Particularly, a CdS-nanorods/RGO composite was prepared for photocatalytic CO2 reduction The CdS nanorods were uniformly dispersed on the surface of RGO nanosheets, which was beneficial for creating a greater number of surface active sites (see Figure 24a) As shown in Figure 24b, G0.5 exhibited a smaller semicircle in the Nyquist plot than pure CdS, indicating faster interfacial electron transfer on the G0.5 sample Moreover, according to the nitrogen adsorption–desorption measurements, no significant changes in the specific surface area were observed for the CdS-nanorods/RGO composites in comparison with pure CdS nanorods However, the CO2-adsorption ability of the CdS-nanorods/RGO was shown to be higher than that of the pure CdS nanorods This was attributed to the π–π conjugation interaction between the CO2 molecules and the RGO Moreover, the π–π conjugation interaction between RGO and CO2 can also lead to the destabilization and activation of CO2 molecules to facilitate their reduction Thus, the photocatalytic CO2-reduction activity of CdS-nanorods/RGO toward CH4 Figure 24. a) TEM image of optimized CdS-nanorods/RGO (G0.5) (Note: the weight percentage of RGO to CdS nanorods was varied from 0, 0.1, 0.25, 0.5, 1.0 and 2.0 wt%, the corresponding samples were labeled as G0, G0.1, G0.25, G0.5, G1.0 and G2.0, respectively.) b) Nyquist plots for G0 and G0.5 c) Comparison of photocatalytic CO2 reduction activity of the G0, G0.1, G0.25, G0.5, G1.0, G2.0, N0.5 (0.5 wt% graphene on CdS nanoparticles), P0.5 (0.5 wt% Pt on CdS nanorods), and RGO for CH4 production d) Schematic illustration of the mechanism for the enhancement of the photocatalytic activity of the CdS-nanorods/RGO Reproduced with permission.[46] Copyright 2014, The Royal Society of Chemistry Adv Mater 2017, 29, 1601694 © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim wileyonlinelibrary.com (16 of 20) 1601694 www.advancedsciencenews.com Review www.advmat.de Figure 25. Schematic illustration of the advantageous contact between both components in 2D–2D composite photocatalysts in comparison to 0D–2D and 0D-1D composites Reproduced with permission.[12] Copyright 2014, The Royal Society of Chemistry production was higher than that of pure CdS nanorods (see Figure 24c,d) due to the high electron–hole separation efficiency, strong CO2 adsorption, and unique CO2 destabilization Taking into account the 2D structure of graphene, its potential can be fully utilized in constructing graphene–semiconductor 2D–2D heterojunction photocatalysts Namely, the contact interface in a 2D–2D (face contact) heterojunction photocatalyst is larger than that in the zero-dimensional–2D (0D–2D) (point contact) and 1D–2D (line contact) heterojunction photocatalysts, which facilitates the migration of charge carriers across the the graphene–semiconductor heterojunction (see Figure 25).[12] Recently, our group reported a low-cost 2D–2D graphene–g-C3N4 composite prepared by an impregnation–chemical-reduction method for photocatalytic hydrogen production.[45] In the aforementioned composite, 2D layered g-C3N4 was intimately coupled with graphene nanosheets to form a 2D–2D layered composite (see Figure 26a,b) Moreover, based on the nitrogen adsorption–desorption isotherms and UV–vis absorption spectra, the specific surface area and lightabsorption ability of g-C3N4 were enhanced by adding graphene Meanwhile, the electron–hole recombination on g-C3N4 was also greatly suppressed after graphene addition due to the fast electron–hole separation across the graphene–g-C3N4 heterojunction and a large contact interface of the 2D–2D graphene– g-C3N4 structure Therefore, the photocatalytic activity of g-C3N4 increases with increasing graphene loading The photocatalytic activity of an optimized graphene–g-C3N4 composite (1.0 wt% graphene) toward hydrogen production in methanol aqueous solution was times higher than that of g-C3N4 with Pt as a co-catalyst (see Figure 26c) However, a further increase in graphene loading to 2.0 wt% and 5.0 wt% resulted in a smaller and smaller photocatalytic activity due to the overloading of graphene on the g-C3N4, which can cause shielding effects and greatly inhibits the light absorption ability of the g-C3N4 Meanwhile, Liang et al carefully investigated the photocatalytic activity of 1D–2D single-wall carbon nanotubes (SWCNT)/ TiO2 nanosheets and 2D–2D graphene/TiO2 nanosheets toward CO2 reduction.[129] It was shown that both SWCNT and graphene can significantly enhance the photocatalytic CO2-reduction activity of the TiO2 nanosheets because of the improved electron–hole separation efficiency and enlarged specific surface area However, the photocatalytic activity of graphene/TiO2 nanosheets was higher than that of SWCNTs/TiO2 because, in the former, a 2D–2D structure was formed with a large contact interface for electron migration In addition to the above-presented simple coupling of graphene nanosheets with a single semiconductor, graphene can be utilized to further improve the photocatalytic activity of 1601694 (17 of 20) wileyonlinelibrary.com Figure 26. a,b) TEM images of graphene nanosheets (a) and graphene– g-C3N4 composite (b) c) Comparison of the photocatalytic activity of GC0 (pure g-C3N4), GC0.25 (0.25 wt% graphene added to g-C3N4), GC0.5 (0.5 wt% graphene added to g-C3N4), GC1.0 (1.0 wt% graphene added to g-C3N4), GC2.0 (2.0 wt% graphene added to g-C3N4), GC5.0 (5.0 wt% graphene added to g-C3N4), and N-doped TiO2 for photocatalytic hydrogen production Reproduced with permission.[45] Copyright 2011, American Chemical Society heterojunction photocatalysts For example, Babu et al reported a Cu2O/TiO2 p–n heterojunction photocatalyst with incorporated graphene oxide (GO) for enhancing its photocatalytic activity toward hydrogen production.[130] The electron–hole separation rate of Cu2O–TiO2 was higher than that of TiO2 due to presence of the p–n heterojunction The aforementioned electron–hole separation in this Cu2O–TiO2 p–n heterojunction photocatalyst was further improved by adding GO to facilitate the migration of electrons from the space-charge region of the Cu2O/TiO2 p–n heterojunction to GO Incorporation of GO to this p–n heterojunction photocatalyst enhanced its photocatalytic hydrogen efficiency under light irradiation, which was 14 times higher than that of a pure TiO2 Moreover, Zhao et al showed that the photocatalytic hydrogen-production activity of the WO3/g-C3N4 direct Z-scheme heterojunction photocatalyst can be further improved by the loading of RGO as a metal-free electron mediator.[131] Zhao’s group demonstrated that RGO can improve the flow of electrons from WO3 to g-C3N4 and suppress reverse migration of electrons to WO3 Therefore, the photocatalytic activity of the WO3/g-C3N4 Z-scheme heterojunction photocatalyst was greatly enhanced after the incorporation of RGO Furthermore, Kuai et al reported a low-cost direct Z-scheme heterojunction CdS/RGO/TiO2 photocatalyst[132] and © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Adv Mater 2017, 29, 1601694 www.advmat.de www.advancedsciencenews.com Conclusions and Future Perspectives In the past several decades, many studies have been reported for the preparation of various heterojunction photocatalysts Here, a concise appraisal of the current achievements in the field of heterojunction photocatalysts is presented, including the fundamental aspects in their design, synthesis, characterization, and applications, demonstrating that this research field is important, exciting, and highly rewarding However, the practical applications and commercialization of heterojunction photocatalysts require further substantial progress in the engineering of highly efficient heterojunction photocatalysts Future research directions in this field should be focused on the following aspects i) Significant challenges still remain in the development of facile, efficient, and economic methods for preparing highquality heterojunction photocatalysts at the large scale for practical applications Moreover, a further advancement is needed in controlling their morphology, contact interface, crystallization, and hierarchical assembly Note that further progress in the preparation of heterojunction photocatalysts is possible in conjunction with substantial advancements in nanoscience and nanotechnology ii) The migration pathway of the photogenerated electrons and holes in the heterojunction photocatalysts requires further systematic studies It was shown that the photogenerated electrons and holes can be spatially separated on the heterojunction photocatalysts; however, up to now, there has been no direct evidence to show the actual migration pathway of electron–hole pairs at the heterojunction interface This issue is important for confirming the formation of different types of heterojunction photocatalysts and should be further investigated by more-powerful characterization tools iii) Studies regarding theoretical calculations and modeling methods should attract much more attention To achieve a deeper understanding of the mechanism and charge-migration kinetics in the heterojunction photocatalysts, further advancements in theoretical calculations are highly desirable to shed some light on the true picture of the photocatalytic processes in the heterojunction photocatalysts iv) Further development of new photocatalyst materials for the design and fabrication of heterojunction photocatalysts is one of the key research goals The existing photocatalytic materials feature various drawbacks such as high cost, large bandgaps, low active surface area, etc It is of Adv Mater 2017, 29, 1601694 great significance to find cost-effective and advanced materials to prepare heterojunction photocatalysts for practical applications A perfect material for engineering a heterojunction photocatalyst should fulfill several requirements, such as visible-light activity, high solar-conversion efficiency, proper bandgap structure for redox reactions, high photostability for long-term applications, and scalability for commercialization In conclusion, we have summarized recent work related to heterojunction photocatalysts and their application in photo catalysis The preparation and investigation of heterojunction photocatalysts provides a meritorious platform for accelerating their practical applications We hope that this review can stimulate further exploration of the heterojunction systems in photo catalysis, solar cells, batteries, and other important research areas Acknowledgments This study was partially supported by the 973 program (2013CB632402), NSFC (21433007, 51320105001, 21573170, 51372190 and 51272199), the Fundamental Research Funds for the Central Universities (2015-III034), the Self-determined and Innovative Research Funds of SKLWUT (2015-ZD-1), and the Natural Science Foundation of Hubei Province of China (No 2015CFA001) Received: March 29, 2016 Revised: November 4, 2016 Published online: February 21, 2017 [1] G C. Xie, K. Zhang, B D. Guo, Q. Liu, L. Fang, J R. Gong, Adv Mater 2013, 25, 3820 [2] A. Kudo, Y. Miseki, Chem Soc Rev 2009, 38, 253 [3] J X. Low, B. Cheng, J G. Yu, M. Jaroniec, Energy Storage Mater 2015, 3, 24 [4] H. Tong, S. Ouyang, Y. Bi, N. Umezawa, M. Oshikiri, J. Ye, Adv Mater 2012, 24, 229 [5] U I. Gaya, A H. Abdullah, J Photochem Photobiol., C 2008, 9, [6] A. Fujishima, X. Zhang, D A. Tryk, Surf Sci Rep 2008, 63, 515 [7] I. Fujishima, K. Honda, Nature 1972, 238, 37 [8] J. Xuan, W J. Xiao, Angew Chem., Int Ed 2012, 51, 6828 [9] W G. Tu, Y. Zhou, Z G. Zou, Adv Mater 2014, 26, 4607 [10] G. Yang, W. Yan, Q. Zhang, S. Shen, S. Ding, Nanoscale 2013, 5, 12432 [11] X. Li, J. Yu, M. Jaroniec, Chem Soc Rev 2016, 45, 2603 [12] J X. Low, S W. Cao, J G. Yu, S. Wageh, Chem Commun 2014, 50, 10768 [13] J. Yang, D. Wang, H. Han, C. Li, Acc Chem Res 2013, 46, 1900 [14] P V. Kamat, J Phys Chem C 2007, 111, 2834 [15] F E. Osterloh, Chem Mater 2007, 20, 35 [16] Y. Qu, X. Duan, Chem Soc Rev 2013, 42, 2568 [17] A. Kubacka, M. Fernández-García, G. Colón, Chem Rev 2011, 112, 1555 [18] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Science 2001, 293, 269 [19] U. Diebold, Surf Sci Rep 2003, 48, 53 [20] J X. Low, J G. Yu, Q. Li, B. Cheng, Phys Chem Chem Phys 2014, 16, 1111 [21] V. Subramanian, E E. Wolf, P V. Kamat, J Am Chem Soc 2004, 126, 4943 © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim wileyonlinelibrary.com (18 of 20) 1601694 Review showed that this heterojunction photocatalyst can significantly promote the separation and accumulation of electrons and holes in the conduction band of the CdS and the valence band of the TiO2, respectively Moreover, the presence of RGO can accelerate the migration of electrons from the TiO2 to the CdS Therefore, the CdS/RGO/TiO2 Z-scheme heterojunction photo catalyst showed an exceptional photocatalytic CO2-reduction performance toward CH4 production, which was higher than that of CdS, CdS/TiO2, and CdS/RGO These studies clearly prove that combining graphene with other heterojunction photocatalysts can greatly enhance their photocatalytic performance, which is worthy of further exploration Review www.advmat.de www.advancedsciencenews.com [22] M S. Akple, J X. Low, S. Wageh, A A. Al-Ghamdi, J G. Yu, J. Zhang, Appl Surf Sci 2015, 358, 196 [23] Y. Bessekhouad, D. Robert, J. Weber, J Photochem Photobiol., A 2004, 163, 569 [24] J C. Yu, J G. Yu, W K. Ho, L Z. Zhang, Chem Commun 2001, 1942 [25] J G. Yu, J C. Yu, M K.-P. Leung, W K. Ho, B. Cheng, X J. Zhao, J C. Zhao, J Catal 2003, 217, 69 [26] M. Xie, X. Fu, L. Jing, P. Luan, Y. Feng, H. Fu, Adv Energy Mater 2014, 4, 1300995 [27] S. Cho, J.-W. Jang, J. Kim, J S. Lee, W. Choi, K.-H. Lee, Langmuir 2011, 27, 10243 [28] G P. Dai, J G. Yu, G. Liu, J Phys Chem C 2011, 115, 7339 [29] J. Zhang, S Z. Qiao, L F. Qi, J G. Yu, Phys Chem Chem Phys 2013, 15, 12088 [30] M. Wang, L. Sun, Z. Lin, J. Cai, K. Xie, C. Lin, Energy Environ Sci 2013, 6, 1211 [31] J G. Yu, J X. Low, W. Xiao, P. Zhou, M. Jaroniec, J Am Chem Soc 2014, 136, 8839 [32] M S. Akple, J X. Low, Z Y. Qin, S. Wageh, A A. Al-Ghamdi, J G. Yu, S W. Liu, Chin J Catal 2015, 36, 2127 [33] S. Gao, W. Wang, Y. Ni, C. Lu, Z. Xu, J Alloys Compd 2015, 647, 981 [34] D. Lu, G. Zhang, Z. Wan, Appl Surf Sci 2015, 358, 223 [35] J G. Yu, S H. Wang, J X. Low, W. Xiao, Phys Chem Chem Phys 2013, 15, 16883 [36] D F. Xu, B. Cheng, S W. Cao, J G. Yu, Appl Catal., B 2015, 164, 380 [37] Q. Li, B D. Guo, J G. Yu, J R. Ran, B H. Zhang, H J. Yan, J R. Gong, J Am Chem Soc 2011, 133, 10878 [38] Q J. Xiang, J G. Yu, M. Jaroniec, Nanoscale 2011, 3, 3670 [39] Y. Chen, J C. Crittenden, S. Hackney, L. Sutter, D W. Hand, Environ Sci Technol 2005, 39, 1201 [40] J G. Yu, J F. Xiong, B. Cheng, S W. Liu, Appl Catal., B 2005, 60, 211 [41] M H. Zhou, J G. Yu, S W. Liu, P C. Zhai, L. Jiang, J Hazard Mater 2008, 154, 1141 [42] J G. Yu, W G. Wang, B. Cheng, Chem Asian J 2010, 5, 2499 [43] F Y. Xu, W. Xiao, B. Cheng, J G. Yu, Int J Hydrogen Energy 2014, 39, 15394 [44] J. Jin, J G. Yu, D P. Guo, C. Cui, W K. Ho, Small 2015, 11, 5262 [45] Q J. Xiang, J G. Yu, M. Jaroniec, J Phys Chem C 2011, 115, 7355 [46] J G. Yu, J. Jin, B. Cheng, M. Jaroniec, J Mater Chem A 2014, 2, 3407 [47] K. Vinodgopal, P V. Kamat, Environ Sci Technol 1995, 29, 841 [48] K. Ranjit, B. Viswanathan, J Photochem Photobiol., A 1997, 108, 79 [49] H. Zhou, Y. Qu, T. Zeid, X. Duan, Energy Environ Sci 2012, 5, 6732 [50] R. Marschall, Adv Funct Mater 2014, 24, 2421 [51] S J. Moniz, S A. Shevlin, D J. Martin, Z.-X. Guo, J. Tang, Energy Environ Sci 2015, 8, 731 [52] H. Wang, L. Zhang, Z. Chen, J. Hu, S. Li, Z. Wang, J. Liu, X. Wang, Chem Soc Rev 2014, 43, 5234 [53] J K. Hyun, S. Zhang, L J. Lauhon, Annu Rev Mater Res 2013, 43, 451 [54] W. Shi, N. Chopra, Nanomater Energy 2013, 2, 158 [55] K. Li, S. Gao, Q. Wang, H. Xu, Z. Wang, B. Huang, Y. Dai, J. Lu, ACS Appl Mater Interfaces 2015, 7, 9023 [56] S J. Hong, S. Lee, J S. Jang, J S. Lee, Energy Environ Sci 2011, 4, 1781 [57] L Y. Huang, H. Xu, Y P. Li, H M. Li, X N. Cheng, J X. Xia, Y G. Xu, G B. Cai, Dalton Trans 2013, 42, 8606 [58] C. Pan, J. Xu, Y. Wang, D. Li, Y. Zhu, Adv Funct Mater 2012, 22, 1518 1601694 (19 of 20) wileyonlinelibrary.com [59] W. Chen, T.-Y. Liu, T. Huang, X.-H. Liu, G.-R. Duan, X.-J. Yang, S.-M. Chen, RSC Adv 2015, 5, 101214 [60] N. Wetchakun, S. Chaiwichain, B. Inceesungvorn, K. Pingmuang, S. Phanichphant, A I. Minett, J. Chen, ACS Appl Mater Interfaces 2012, 4, 3718 [61] W.-J. Ong, L K. Putri, L.-L. Tan, S.-P. Chai, S.-T. Yong, Appl Catal., B 2016, 180, 530 [62] D O. Scanlon, C W. Dunnill, J. Buckeridge, S A. Shevlin, A J. Logsdail, S M. Woodley, C R A. Catlow, M J. Powell, R G. Palgrave, I P. Parkin, Nat Mater 2013, 12, 798 [63] R. Su, R. Bechstein, L. Sø, R T. Vang, M. Sillassen, B r. Esbjörnsson, A. Palmqvist, F. Besenbacher, J Phys Chem C 2011, 115, 24287 [64] P. Deák, B. Aradi, T. Frauenheim, J Phys Chem C 2011, 115, 3443 [65] Y.-P. Yuan, L.-W. Ruan, J. Barber, S C J. Loo, C. Xue, Energy Environ Sci 2014, 7, 3934 [66] L. Li, P A. Salvador, G S. Rohrer, Nanoscale 2014, 6, 24 [67] D. Jiang, X. Li, Mater Focus 2015, 4, 14 [68] L. Ye, J. Liu, L. Tian, T. Peng, L. Zan, Appl Catal., B 2013, 134–135, 60 [69] Z. He, Y. Shi, C. Gao, L. Wen, J. Chen, S. Song, J Phys Chem C 2014, 118, 389 [70] S. Chen, S. Zhang, W. Liu, W. Zhao, J Hazard Mater 2008, 155, 320 [71] J Q. Liu, L L. Ruan, S B. Adeloju, Y C. Wu, Dalton Trans 2014, 43, 1706 [72] J. Cao, X. Li, H. Lin, S. Chen, X. Fu, J Hazard Mater 2012, 239, 316 [73] J. Lin, J. Shen, R. Wang, J. Cui, W. Zhou, P. Hu, D. Liu, H. Liu, J. Wang, R I. Boughton, Y. Yue, J Mater Chem 2011, 21, 5106 [74] Y. Peng, M. Yan, Q.-G. Chen, C.-M. Fan, H.-Y. Zhou, A.-W. Xu, J Mater Chem A 2014, 2, 8517 [75] C. Chen, W. Cai, M. Long, B. Zhou, Y. Wu, D. Wu, Y. Feng, ACS Nano 2010, 4, 6425 [76] Z. Zhang, C. Shao, X. Li, C. Wang, M. Zhang, Y. Liu, ACS Appl Mater Interfaces 2010, 2, 2915 [77] J. Zhang, Z. Zhu, X. Feng, Chem.–Eur J 2014, 20, 10632 [78] W.-J. Ong, L.-L. Tan, S.-P. Chai, S.-T. Yong, A R. Mohamed, Nanoscale 2014, 6, 1946 [79] C P. Sajan, S. Wageh, A A. Al-Ghamdi, J G. Yu, S W. Cao, Nano Res 2016, 9, [80] K. Zhou, Y. Li, Angew Chem., Int Ed 2012, 51, 602 [81] H. Lin, L. Li, M. Zhao, X. Huang, X. Chen, G. Li, R. Yu, J Am Chem Soc 2012, 134, 8328 [82] Z. Zheng, B. Huang, J. Lu, X. Qin, X. Zhang, Y. Dai, Chem.–Eur J 2011, 17, 15032 [83] M. Guo, L. Li, H. Lin, Y. Zuo, X. Huang, G. Li, Chem Commun 2013, 49, 11752 [84] G. Liu, H G. Yang, J. Pan, Y Q. Yang, G Q. Lu, H.-M. Cheng, Chem Rev 2014, 114, 9559 [85] J L. Giocondi, G S. Rohrer, MRS Proc 2000, 654, [86] C. Liu, X. Han, S. Xie, Q. Kuang, X. Wang, M. Jin, Z. Xie, L. Zheng, Chem Asian J 2013, 8, 282 [87] R. Li, F. Zhang, D. Wang, J. Yang, M. Li, J. Zhu, X. Zhou, H. Han, C. Li, Nat Commun 2013, 4, 1432 [88] H J. Yun, H. Lee, N D. Kim, D M. Lee, S. Yu, J. Yi, ACS Nano 2011, 5, 4084 [89] A. Iwase, Y H. Ng, Y. Ishiguro, A. Kudo, R. Amal, J Am Chem Soc 2011, 133, 11054 [90] P. Zhou, J G. Yu, M. Jaroniec, Adv Mater 2014, 26, 4920 [91] K. Maeda, ACS Catal 2013, 3, 1486 [92] A J. Bard, J Photochem 1979, 10, 59 [93] H. Tada, T. Mitsui, T. Kiyonaga, T. Akita, K. Tanaka, Nat Mater 2006, 5, 782 © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Adv Mater 2017, 29, 1601694 www.advmat.de www.advancedsciencenews.com Adv Mater 2017, 29, 1601694 [114] H.-J. Choi, S.-M. Jung, J.-M. Seo, D W. Chang, L. Dai, J.-B. Baek, Nano Energy 2012, 1, 534 [115] J W. Jo, J U. Lee, W H. Jo, Polym Int 2015, 64, 1676 [116] N. Gao, X S. Fang, Chem Rev 2015, 115, 8294 [117] T F. Yeh, C Y. Teng, L C. Chen, S J. Chen, H S. Teng, J Mater Chem A 2016, 4, 2014 [118] J X. Low, J G. Yu, W K. Ho, J Phys Chem Lett 2015, 6, 4244 [119] Q J. Xiang, B. Cheng, J G. Yu, Angew Chem., Int Ed 2015, 54, 11350 [120] H. Zhang, X J. Lv, Y M. Li, Y. Wang, J H. Li, ACS Nano 2010, 4, 380 [121] J. Zhang, L F. Qi, J R. Ran, J G. Yu, S Z. Qiao, Adv Energy Mater 2014, 4, 1301925 [122] P D. Tran, S K. Batabyal, S S. Pramana, J. Barber, L H. Wong, S C J. Loo, Nanoscale 2012, 4, 3875 [123] X. An, K. Li, J. Tang, ChemSusChem 2014, 7, 1086 [124] G. Liao, S. Chen, X. Quan, H. Yu, H. Zhao, J Mater Chem 2012, 22, 2721 [125] E P. Gao, W Z. Wang, M. Shang, J H. Xu, Phys Chem Chem Phys 2011, 13, 2887 [126] G. Ramesha, A V. Kumara, H. Muralidhara, S. Sampath, J Colloid Interface Sci 2011, 361, 270 [127] Y. Fu, H. Chen, X. Sun, X. Wang, Appl Catal., B 2012, 111, 280 [128] A. Farghali, M. Bahgat, W. El Rouby, M. Khedr, J Alloys Compd 2013, 555, 193 [129] Y T. Liang, B K. Vijayan, O. Lyandres, K A. Gray, M C. Hersam, J Phys Chem Lett 2012, 3, 1760 [130] S G. Babu, R. Vinoth, D P. Kumar, M V. Shankar, H.-L. Chou, K. Vinodgopal, B. Neppolian, Nanoscale 2015, 7, 7849 [131] G. Zhao, X. Huang, F. Fina, G. Zhang, J T. Irvine, Catal Sci Technol 2015, 5, 3416 [132] L. Kuai, Y. Zhou, W. Tu, P. Li, H. Li, Q. Xu, L. Tang, X. Wang, M. Xiao, Z. Zou, RSC Adv 2015, 5, 88409 © 2017 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim wileyonlinelibrary.com (20 of 20) 1601694 Review [94] X L. Yin, J. Liu, W J. Jiang, X. Zhang, J S. Hu, L J. Wan, Chem Commun 2015, 51, 13842 [95] H. Wang, Z. Ye, C. Liu, J. Li, M. Zhou, Q. Guan, P. Lv, P. Huo, Y. Yan, Appl Surf Sci 2015, 353, 391 [96] Q. Huang, Q. Zhang, S. Yuan, Y. Zhang, M. Zhang, Appl Surf Sci 2015, 353, 949 [97] X. Chen, X. Huang, Z. Yi, Chem Eur J 2014, 20, 17590 [98] W.-K. Jo, T. Sivakumar Natarajan, ACS Appl Mater Interfacees 2015, 7, 17138 [99] Y. Liu, G. Ji, M A. Dastageer, L. Zhu, J. Wang, B. Zhang, X. Chang, M A. Gondal, RSC Adv 2014, 4, 56961 [100] S. Chen, Y. Hu, L. Ji, X. Jiang, X. Fu, Appl Surf Sci 2014, 292, 357 [101] J F. Zhang, Y F. Hu, X L. Jiang, S F. Chen, S G. Meng, X L. Fu, J Hazard Mater 2014, 280, 713 [102] N. Tian, H. Huang, Y. He, Y. Guo, T. Zhang, Y. Zhang, Dalton Trans 2015, 44, 4297 [103] A. Kudo, H. Kato, I. Tsuji, Chem Lett 2004, 33, 1534 [104] H. Katsumata, Y. Tachi, T. Suzuki, S. Kaneco, RSC Adv 2014, 4, 21405 [105] D. Xia, Y. Li, G. Huang, C C. Fong, T. An, G. Li, H Y. Yip, H. Zhao, A. Lu, P K. Wong, Appl Catal., B 2015, 176–177, 749 [106] D. Chen, H. Zhang, Y. Liu, J. Li, Energy Environ Sci 2013, 6, 1362 [107] H. Chang, H. Wu, Energy Environ Sci 2013, 6, 3483 [108] N. Li, M. Cao, C. Hu, Nanoscale 2012, 4, 6205 [109] K S. Novoselov, A K. Geim, S V. Morozov, D. Jiang, Y. Zhang, S V. Dubonos, I V. Grigorieva, A A. Firsov, Science 2004, 306, 666 [110] X. Li, Z. Lv, H. Zhu, Adv Mater 2015, 27, 6549 [111] Z. Yang, J. Ren, Z. Zhang, X. Chen, G. Guan, L. Qiu, Y. Zhang, H. Peng, Chem Rev 2015, 115, 5159 [112] M. Batmunkh, M J. Biggs, J G. Shapter, Adv Sci 2015, 2, 1400025 [113] F. Bonaccorso, L. Colombo, G H. Yu, M. Stoller, V. Tozzini, A C. Ferrari, R S. Ruoff, V. Pellegrini, Science 2015, 347, 10