Industrialization has meant that the exhaustion of fossil energy sources and the deterioration of the environment have become serious bottlenecks to constrain further economic and social development. The demand for energy necessitates the search for renewable and clean alternative resources to supplement and eventually replace our dependence on fossil fuels [1]. Solar energy is a free, abundant, and renewable clean energy and is one of the most promising options to ease the energy and environmental strain. Therefore, the harvest and conversion of solar energy into a usable energy by using photocatalytic processes is highly desirable [2, 3]. The incident solar energy at the surface of the Earth is over 1.3×105 TW, which far exceeds the current global energy consumption rate (~16.9 TW in 2013). Since Honda and Fujishima [4] discovered photoelectrochemical (PEC) water-splitting over a TiO2 electrode in 1972, solar H2 production from semiconductor-based photocatalytic water splitting has become a promising method to convert solar energy into clean and carbon-neutral chemical energy. It has been proposed that the photocatalytic H2 production technology process could be used for commercial application when the solar energy conversion efficiency is above 10% [5].
The key factors influencing the efficiency of solar fuel synthesis include: (1) light absorption, (2) e-/h+ separation and transport, and (3) the photo-electron chemical reaction. Thus, the design of efficient photocatalysts for producing molecular H2 from water splitting by harvesting solar energy is one of the foremost challenges in the development of a solar hydrogen economy. In the past 40 years, great efforts have been made to synthesize new materials and enrich the photocatalytic theory [6-13]. In the past 10 years, the funding for clean energy research has been increasing in China and from other governments around the world, which has greatly promoted the development of photocatalytic hydrogen production. Many different kinds of semiconductor materials have been researched and developed as photocatalytic hydrogen production materials, such as: ion-doped TiO2 with visible-light catalytic activity [14-17]; doping of indium-tantalum-oxide with nickel (In0.9Ni0.1TaO4) inducing direct splitting of water under visible-light irradiation with a quantum yield of approximately 0.66% [18]; highly efficient hydrogen production over lanthanum-doped NaTaO3 photocatalysts with a quantum yield of approximately 56% at 270 nm [19]; cadmium sulfide [20-23] and (AgIn)xZn2(1-x)S2 [24] with a high hydrogen production efficiency in sacrificial agent solution; GaN-ZnO solid solution with high activity for photocatalytic H2 generation [25-28]; and metal-free polymeric photocatalyst g-C3N4 that can produce hydrogen from water under visible-light irradiation [9, 29].
A schematic illustration of the photocatalytic H2 production on a semiconductor is shown in Fig. 1. When the photocatalyst is illuminated with a light source, the electrons in the valence band (VB) of the photocatalyst are excited to the conduction band (CB) while the holes are left in the VB, which creates the electron-hole (e--h+) pairs. The photogenerated electrons and holes, which do not undergo recombination, migrate towards the surface of the photocatalyst, where they act as the reducing agent and oxidizing agent to result in the reduction and oxidation reactions, respectively. The photocatalytic splitting of water into hydrogen and oxygen is a thermodynamically "uphill" reaction, involving a large positive change in the Gibbs energy (∆G0 = 237 kJ/mol). To achieve an overall water splitting, both the reduction and oxidation potentials of water should lie within the band gap of the photocatalyst. The bottom of the CB should be more negative than the reduction potential of H+/H2 (0 V vs. normal hydrogen electrode (NHE)), while the top of the VB must be more positive than the oxidation potential of O2/H2O (1.23 V vs. NHE). Therefore, the minimum photon energy thermodynamically required to drive the reaction is 1.23 eV, which corresponds to a wavelength of ∼1000 nm, in the near-infrared region. To absorb the visible-light irradiation from sunlight, Eg should be lower than 3.0 eV ( > 400 nm). Hence, the band gap of the photocatalyst must match up with the CB and VB potentials, which is a critical factor to facilitate both the reduction and oxidation of H2O by photoexcited electrons and holes.
In practical applications, a more negative CB potential and a more positive VB potential are thermodynamically beneficial for the reduction and oxidation of reactants, respectively. However, this will also widen the bandgap of the photocatalyst, resulting in a poor solar-light absorption power conversion efficiency. Therefore, it is difficult for a single-component photocatalyst to exhibit both a strong redox ability and wide light-absorption range at the same time. Moreover, in the one-phase photocatalysts, the photoexcited electrons in the CB can easily return to the VB and recombine with the photogenerated holes, which seriously hinders the effective use of solar energy [30]. Therefore, a great deal of effort has been devoted to enhancing the efficiency of the photocatalytic process of one-phase materials, such as doping, sensitizing, morphology control, crystal lattice control and surface modification [31-36]. Among these approaches, the formation of semiconductor heterostructures has attracted vast attention because their performance for photocatalytic H2 generation can be enhanced by combining semiconductors with different electronic structures that are capable of harvesting visible light and can regulate the electron transfer, thus drastically expanding the research scope [37-43].
Depending on the difference in transfer mechanism of photogenerated charge carriers, the heterostructured photocatalysts can be divided into two categories as follows. (1) Heterojunctions with a classic carrier transfer mechanism (Fig. 2(a)) formed from wide band gap semiconductors-A with a lower CB energy level and narrow band gap semiconductors-B with a higher conduction band energy level. The narrow band gap semiconductors can effectively expand the light response range of heterojunctions owing to their wider light absorption range. The photogenerated electrons of B can transfer to A, which leads to the dramatic increase in the number of photogenerated electrons, and consequently the photocatalytic H2 generation performance is enhanced. Meanwhile the bottom of the CB of A and B must be more negative than the H+/H2 reduction potential. (2) Heterojunctions with the Z-scheme charge transfer mechanism (Fig. 2(b)). The artificial heterogeneous Z-scheme system mimics natural photosynthesis in green plants, which is different from the heterojunction-type charge transfer mechanism mentioned above. Under solar light irradiation, the photogenerated electrons in B with a relatively negative CB can recombine with the holes in A with a relatively positive VB at the heterostructure interface. As a result, the electrons-holes that have stronger reduction/oxidation abilities are maintained at different active sites, which enhances the photocatalytic efficiency [44]. The formation of a heterojunction can be proved as followed: (Ⅰ) Estimating the characteristics of heterojunctions with XPS, HRTEM, EDS, and so forth; (Ⅱ) Especially for a p-n heterojunction, the I-V curve and Mott-Schottky curve are usually useful.
Although there are many excellent reviews on various aspects of heterostructured photocatalysts published during the last few years [43, 45-48], there is an urgent requirement to consolidate the recent research activities in this rapidly developing area. We believe that a comprehensive review on this subject is timely to promote further developments in this exciting and still emerging area of research. In this review, we mainly focus on a systematic introduction of the basic principles of various types of heterogeneous catalyst systems and their recent progress and applications in photocatalytic hydrogen production. Finally, the opportunities and challenges in the development of heterostructured photocatalysts are also provided. We hope this review will provide a good reference and inspire new ideas to design highly efficient and stable heterojunction photocatalysts for photocatalytic H2 production.
As shown in Fig. 2(a), photogenerated electrons transfer from one semiconductor with a high CB to another semiconductor with a low CB, and then take part in the photochemical reaction. The heterojunctions with a classic carrier transfer system are mainly applied to photocatalytic water splitting, photodegradation, photovoltaic cells and other fields. Based on the bandgaps and the electronic affinity of semiconductors, the semiconductor heterojunctions with a classic carrier transfer system for hydrogen production can be divided into three different cases: type-Ⅰ, type-Ⅱ, and pn-type heterojunctions [44, 49].
As shown in Fig. 3, for the type-Ⅰ heterojunction, the VB level of semiconductor-A is lower than that of semiconductor-B, while the CB level of semiconductor-A lies above the CB level of semiconductor-B. When the type-Ⅰ heterojunction is irradiated, driven by the potential difference, photogenerated electrons transfer from semiconductor-A to the CB of semiconductor-B and photogenerated holes transfer from A to the VB of semiconductor-B, thus electrons and holes accumulate on semiconductor-B. Li et al. [50] reported that type-Ⅰ CdS/ZnS core-shell heterojunction exhibited a better photocatalytic H2 production performance compared with a bare CdS nanocrystal. As can be seen from Fig. 4, the co-catalysts play an important role in the separation and transfer of the carriers in type-Ⅰ core/shell heterojunctions. During the charge transfer process, the dual co-catalysts promote the transfer of electrons and holes from the core to the outer surface and provide active sites for photocatalytic reactions. As a result, the electrons and holes can be transferred from the core and spatially separate in the different co-catalysts, the photocatalytic activity as well as the stability of CdS nanocrystals for photocatalytic H2 evolution are enhanced. The photoactivity enhancement of such type-Ⅰ heterojunctions might be ascribed to the following reasons. The first is the fast transfer of photogenerated electrons from the core to the shell; the second is the photogenerated electrons leak from the unwrapped core surface; the third is the shell passivation that can inhibit photocorrosion.
The type-Ⅱ heterojunction has been widely applied in the photocatalysis field [51]. As shown in Fig. 5, in a type-Ⅱ band alignment, an electron-accepting semiconductor and a hole- accepting semiconductor with matching band potentials are tightly coupled to construct an effective heterostructure. In this type, the positions of the CB and VB of semiconductor-B are both lower than those of semiconductor-A, which leads to the electrons in the CB of semiconductor-A and holes in the VB of semiconductor-B can be transferred to different sides of the heterostructure. Furthermore, the difference of the chemical potential between semiconductor-A and -B induces band bending at the junction interface. The band bending causes a built-in field, which promotes the photoexcited electrons and holes to easily move in opposite directions, leading to an efficient charge separation and greatly enhanced photocatalytic activity [52]. Semiconductors able to be excited by visible light (narrow bandgap semiconductors, such as g-C3N4 and CdS) are often coupled with UV excited semiconductors (such as TiO2 and ZnWO4) to form type-Ⅱ heterostructures, which is an effective approach to increase the photoresponse to the visible-light region. It is reported that the crystalline structure at the interface of the heterojunction has an important influence on optimizing the quantum effect of the photocatalyst. A difference in the lattice spacing between the two semiconductors is likely caused by a lattice mismatch. The lattice mismatch at the interface caused by the lattice spacing difference between the two semiconductors may induce an interface strain, and then induce an electric field. The electric field can further promote the separation of electrons and holes, thereby improving the photocatalytic activity. However, the lattice mismatch could also induce defects to capture the photogenerated e--h+, which could restrain the diffusion of e--h+ [53]. Therefore, careful selection of materials should be performed before preparing visible-light driven type-Ⅱ heterostructures. To date, a number of type-Ⅱ heterojunctions have been synthesized, such as composite nanoparticles [54-59], nanowires with quantum dots [60-62], core/shell nanowires [63-65], and so forth. These type-Ⅱ heterojunctions have achieved a higher photocatalytic hydrogen efficiency than pure photocatalytic materials. Such type-Ⅱ heterojunctions have been grown on substrates based on very commonly used CdS, g-C3N4 material systems, and different crystal faces [36, 66-72].
Compared with nanoparticles, one-dimensional (1D) nanostructures (such as nanowires and nanofibers) have attracted extensive research interest over the past decade owing to the larger specific surface area and the ability of affording a continuous transfer pathway for carrier [73, 74]. Han et al. [66] reported that when the g-C3N4/N-TiO2 nanofiber was irradiated under solar light, the photogenerated electrons and holes could transfer to the surfaces of N-TiO2 and g-C3N4, respectively (Fig. 6(a)). Thus the recombination of photoinduced electron-hole pairs can be effectively inhibited and more photogenerated electrons contribute to the photocatalytic reaction. The g-C3N4/N-TiO2 nanofiber shows a H2 evolution rate of 8931 μmol/(h·g), which is much higher than that of a pure N-TiO2 nanofiber and g-C3N4. In addition, Prof. Yu's group [70] prepared CdS@g-C3N4 core-shell nanowires by a combined solvothermal and chemisorption method (Fig. 6(b)). Because both the lowest unoccupied molecular orbital (LUMO) and the highest occupied molecular orbital (HOMO) levels of g-C3N4 are higher than the corresponding conduction band minimum (CBM) and valence band maximum (VBM) of CdS. Under visible-light illumination (λ ≥ 420 nm), the photogenerated electrons on the LUMO level of g-C3N4 can directly transfer to the CB of CdS and then take part in the photocatalytic hydrogen-generation reaction, while the photogenerated holes on the VBM of CdS can readily transfer to the HOMO of g-C3N4 and thus substantially inhibited the photocorrosion of CdS. In this case, the efficiency of the charge separation can be enhanced, resulting in improvement of the photocatalytic activity. The hydrogen production rate of the g-C3N4@CdS heterojunction is up to 4152 μmol/(h·g) at 420 nm, which is approximately 2 times higher than that over CdS nanowires. It is concluded that the synergistic effect between g-C3N4 and CdS, originating from their well-matched overlapping band structures, play an important role in enhancing the photocatalytic activity and improving the photostability of the g-C3N4@CdS nanowires. Recently, Du et al. [75] demonstrated a twin-induced zinc-blende/ wurtzite (ZB/WZ) type-Ⅱ heterojunction in Cd0.7Zn0.3S elongated nanorods, which showed alternating ZB and WZ segments along the < 111 > direction, resulted in a very high visible-light photocatalytic activity with a H2 production efficiency of 31.3 mmol/(h·g) and an apparent quantum efficiency of 65.7% at 420 nm (Fig. 6(c)).
The semiconductor pn-junction is an available structure to enhance the charge collection and separation efficiency. When p-type and n-type semiconductors contact each other, the band positions of the semiconductors would be changed accompanying a change in the electron transfer pathway [76]. As the pn-junction semiconductors are joined, electrons near the p-n interface tend to diffuse into the p-region leaving positively charged ions in the n-region, and the holes near the p-n interface begin to diffuse into the n-type region leaving ions with a negative charge. At the interface of the semiconductors, a space charge region will form because of the juxtaposition of the high concentrations of negatively and positively charged ions, then an electric field is created in the space charge region that can guide the photogenerated electrons and holes to transfer to the opposite direction. As a result, the photogenerated electrons and holes are spatially isolated, which sharply enhances the charge separation efficiency. A schematic representation of this heterojunction is provided in Fig. 7. The electric field can induce the electrons to move to the CB of the n-type semiconductors and the holes can transfer to the VB of the p-type semiconductor [77], resulting in a more efficient charge separation, rapid charge transfer to the photocatalyst and longer charge carrier lifetimes. All of these characters provide the pn-type heterojunctions with a dramatic photocatalytic performance.
TiO2 is an n-type semiconductor. The conduction band of TiO2 is slightly higher than the reduction potential of water and so it can reduce protons when it is excited by light. In recent years, TiO2 has been coupled with many different p-type semiconductors to obtain efficient heterostructured photocatalysts, such as CuxO [77-82], Ag2O [83, 84], Co3O4 [85, 86], CuAlO2 [87], TiS2 [88], NiO [89, 90], BiOX (X = Cl, Br, I) [91-93], FeTiO3 [94, 95] and Cu3P [96]. Apart from TiO2 based p-n heterojunctions, other photochemically active p-n heterojunctions have also been reported, such as Cu2S/CdS [97], MoS2/CdS [98], p-MoS2/n-rGO [99], Cu3P/CdS [100], p-CuO/n-ZnO [101], p-NiO/n-ZnO [102] and Co3O4/ZnO [103]. These nanosized photocatalysts exhibit improved photocatalytic activity for hydrogen production and degradation under UV and visible light irradiation.
Besides coupling different semiconductors to construct heterojunctions, the pn-junction in the same material can also be obtained. For example, Prof. Ye's group [104] successfully synthesized pn-homojunction g-C3N4 photocatalysts by fully mixing a certain amount of NaBH4 with pristine g-C3N4, and then subjecting them to a heat treatment at an appropriate temperature for 30 min in an Ar atmosphere. During the NaBH4 treatment, cyano terminal C≡N groups, acting as electron acceptors, were introduced into the g-C3N4 sheet edge, which resulted in the g-C3N4 exhibiting both n-and p-type conductivities (Fig. 8). This homojunction structured g-C3N4 photocatalysts (heated to 200 ℃) showed the highest performance, where a 4-times higher performance than that of pristine g-C3N4 was obtained. Furthermore, no noticeable deactivation in the H2 evolution rate was detected after four cycles of the reactions, clearly suggesting the good stability of this homojunction catalyst. Kumar et al. [105] have reported that p-ZnO has been synthesized by doping with arsenic. A ZnO homojunction synthesized by deposition of Al doped n-type ZnO and intrinsic nanowire ZnO on As-doped p-type ZnO showed clear p-n heterojunction characteristics. In addition, a p-TiO2/n-TiO2 junction can also be prepared by doping with some metal ions [106-113] or non-metal ions [114]. The positions of the CB and VB have an important influence on the ability to prepare n-type or p-type doping, and excellent photocatalytic performances can be obtained [115]. It is easier to dope an n-type material when the CB energy is lower, such as ZnO, ZnS, CdS and CdTe, while it is easier to dope a p-type material when the VB is higher, such as ZnTe [114, 116-119].
As early as 1979, Bard [120] reported the so-called Z-scheme. Since then, by mimicking natural photosynthesis, the artificial Z-scheme photocatalytic system has been widely investigated and different kinds of Z-scheme structures have been fabricated for efficient water splitting. For the above type-Ⅰ, type-Ⅱ, and pn-type heterojunctions, the obvious disadvantage is that the redox ability of photoinduced electrons and holes on each active site is weakened after charge transfer. However, the artificial Z-scheme system not only has the merit of ensuring the efficient spatial separation of electrons and holes pairs but also ensures that electrons-holes with strong reduction/oxidation abilities can be produced. Even so, Z-scheme photocatalytic systems have to face two major problems. First, photocatalysts with gathered electrons-holes will suffer from photo-reduction/photo-oxidation. Second, recombination of photogenerated electrons and holes in the Z-scheme, which ensures the efficient separation of carriers, leads to the depletion of a large number of photogenerated electrons and holes. Even then, the Z-scheme strategy still attracts much attention owing to its unique electron transfer. The artificial Z-scheme photocatalytic systems can be divided into three types, Z-scheme systems with shuttle redox mediators, solid state electron mediators, and direct Z-scheme systems.
In a typical Z-scheme photocatalytic system with shuttle redox mediators, the biomimetic Z-scheme photocatalytic system is composed of two independent semiconductors to generate hydrogen and oxygen with separate redox mediators in liquid electrolytes [121]. Upon irradiation with visible light, photogenerated holes in the VB of one semiconductor and photogenerated electrons in the CB of another semiconductor will react with the redox couple ions, such as Fe2+/Fe3+, I-/I-, and Br-/BrO3- [122-128]. As shown in Fig. 9, the photogenerated hole from the VB of the right photocatalyst and the photogenerated electron from the CB of the left photocatalyst reacts with the electron acceptor/donor (A/D) pair. Then the photogenerated electron in the CB of the right photocatalyst and the photogenerated hole in the VB of the left photocatalyst participate in the photochemical reaction processes with reactants to improve the photocatalytic activity. Particularly, Abe's group [129] reported that the rates of H2 and O2 evolution of TaON-WO3 satisfied the stoichiometry under visible light. However, the photocatalytic efficiency in this Z-scheme is limited by two factors: (1) the introduction of redox couple ions complicates the photocatalytic reaction; (2) slow diffusion of redox couple ions decreases the photocatalytic hydrogen efficiency as well as the backward reactions between the redox couple ions and e--h+ with higher energies. The photocatalytic system is an evolution from the photocatalysis composite with an indirect Z-scheme charge transfer mechanism. Great efforts have been made to restrain these negative influences.
For Z-scheme systems with solid state electron mediators, the energy levels of the solid-solid contact interface are quasi-continuous which is similar to the property of metals, such as low-resistance (Fig. 10). Thus, the solid-solid contact interface forms a known ohmic contact owing to its low contact resistance [130-132]. The ohmic contact is good for the direct recombination between photogenerated electrons from the CB of photosystem Ⅱ (PS Ⅱ) and photogenerated holes from the VB of photosystem Ⅰ (PS I), which reduces the distance of the Z-scheme electron transfer. Thus, the photogenerated holes in the VB of PS Ⅰ can be mostly reserved for the oxidation reaction, while the photogenerated electrons in the CB of PS Ⅱ can participate in the reduction reaction. At present, common solid electron mediators, such as the noble-metal (Pt, Au, Ag) particles, graphene and reduced graphene oxide (RGO), have been explored extensively. As can be seen from the HRTEM image in Fig. 11(a), Tada et al. [39] were the first to report the all-solid-state CdS-Au-TiO2 system in 2006. By photochemical deposition, Au nanoparticles (NPs) are attached on the TiO2 surface, then with subsequent ultraviolet-light irradiation, S8 molecules are selectively adsorbed on the surface of Au NPs. Next, the S8 molecules are reduced to S2- by the collected electrons of Au NPs, then Cd2+ ions react with S2- ions to form the CdS shell around the surface of Au NPs. The resulting CdS-Au- TiO2 system exhibits a higher H2 evolution activity than the Au-TiO2 and TiO2-CdS systems. In recent decades, intriguing graphene/RGO-based materials have been applied in various fields including the constructing of the Z-scheme. Iwase and co-workers [40] found that a tailor-made, photoreduced graphene oxide (PRGO) can behave as a solid electron mediator for water splitting and the proposed mechanism is depicted in Fig. 11(b). Upon excitation with visible light, photoreduced graphene oxide, as an electron conductor, transfers electrons from the CB of BiVO4 to the vacancies in the impurity levels of Ru/SrTiO3:Rh. Simultaneously, the gathered electrons of Ru/SrTiO3:Rh reduce water to H2 while the holes in BiVO4 oxidize water to O2. The work provided a new method for the design of graphitic carbon material graphene in efficient water splitting systems.
In direct Z-scheme systems, the properties of the solid-solid contact interface between two photocatalysts are vital for the resistance level of electron transfer. Hence, the design of the solid-solid contact interface by physical or chemical formation methods is important in the construction of direct Z-scheme systems. Guo et al. [133] demonstrated a nanocomposite of ZnO and Zn0.2Cd0.8S that generated H2 from water containing SO32- and S2- ions as sacrificial agents. The rate of H2 generation by the Z-scheme ZnO/Zn0.2Cd0.8S heterojunction reached 25180 μmol/(h·g) with a high apparent quantum efficiency (AQY) of 49.5% at 420 nm. The high photocatalytic hydrogen generation rate resulted from the formation of an artificial photosynthetic Z-scheme ZnO1-x/Zn0.2Cd0.8S system and oxygen vacancy abundant ZnO1-x in the heterojunction, which exhibit some similar properties to those of a conductor. These characteristics prompt the efficient separation of charge carriers (Fig. 12). Liu et al. [38] clearly demonstrated the synergistic effects of Z-scheme coupled ZnO/CdS heterostructures that exhibited an H2 evolution rate of 1805 μmol/(h·g). The rate was faster than that recorded using solely CdS or ZnO photocatalysts prepared by the parallel synthesis routes. Liu et al. concluded that upon band gap excitation of the two semiconductors, electron transfer takes place from the CB of ZnO to the VB of CdS, as observed in the lifetime of photoluminescence, and that the Z scheme-like electron transfer contributes to the enhanced performance.
Since the electronic structure of a semiconductor depends on the atomic ordering, hence the crystal structure, different crystal phases of one semiconductor compound can exhibit different band positions and/or Eg. TiO2 is found naturally in three phase structures: anatase, rutile and brookite. In particular, band gaps for anatase, rutile and brookite are 3.2, 3.0 and 3.4 eV, respectively. Many studies have shown that anatase is the most photocatalytically active phase owing to the lower recombination rate [134, 135]. The confirmed phase-dependent photocatalytic capability of TiO2 crystals offers opportunities for developing mixed-phase TiO2 photocatalysts with enhanced activities from single-phase TiO2. Xu et al. [136] prepared anatase and rutile bi-phase TiO2 nanofibers through calcining electrospun TiO2 nanofibers. In a mixed phase TiO2 nanofiber, the ratio of the rutile and anatase phases is important and the appropriate ratio results in the formation of the optimal Z-scheme photocatalyst, which is beneficial for the separation of photogenerated electrons and holes. When the sample was cooled rapidly after calcination at 500 ℃ for 1 h, anatase/rutile bi-phase TiO2 nanofibers with a roughly equal mass ratio of 55% anatase and 45% rutile were obtained, which exhibited the highest photocatalytic H2 production rate of 324 mmol/h and an apparent QE of 20.9% at 365 nm (Fig. 13).
As the world's population continues to grow and with the expansion of industry, energy consumption and environmental pollution are two major problems that need to be overcome. Photocatalytic hydrogen generation is one of the most promising methods to deal with the energy and environment crises. Nanojunctions can significantly enhance the activity of the photocatalyst. Type-Ⅰ heterojunctions can prolong the lifetime of photogenerated electrons, but it is difficult to design this heterojunction owing to both the electrons and holes being transferred to the same semiconductor. The two semiconductors of a type-Ⅱ heterojunction are irradiated simultaneously, and the photogenerated electrons and holes transfer to different sides of the heterojunction that results in the spatial separation of the electron-hole pairs. Hence, the recombination of photogenerated electrons and holes is inhibited and then many more photogenerated electrons can be provided for hydrogen evolution. The photogenerated electrons and holes in the pn-heterojunction can be transferred quickly, and the lifetime of photogenerated electrons is prolonged, which arises from the built-in internal electric field. In all type-Ⅰ, type-Ⅱ and pn-type heterojunctions, the photogenerated electrons always transfer to the CB with a lower position thereby decreasing their redox abilities. The artificial heterogeneous Z-scheme photocatalytic system, as is well known, is the most similar to natural photosynthesis for efficient water splitting. Compared with those aforementioned heterojunctions, the Z-scheme not only enables the efficient separation of photogenerated electron-hole pairs but also maintains the photogenerated electrons on the higher CB position.
In addition, the similarities and differences of these technologies are shown in Table 1 for a clear and better understanding of the heterojunction.
Meanwhile, the study of nanoheterojunctions in photocatalytic water splitting primarily still faces the following issues and future challenges.
(1) Solar energy conversion efficiency increases when one can achieve an overall water splitting under longer wavelength irradiation, as shown in Fig. 14. Therefore, we should continue to pursue more active photocatalytic systems capable of harvesting more visible photons [5, 130].
(2) The photocatalytic activity is not sufficiently high and the photocatalytic stability of prepared materials to date can only remain stable for less than 1000 h under solar illumination conditions, which makes it difficult to meet practical requirements. Therefore, a key issue is to design new photocatalysts with a wide range absorption of visible light and long lifetime of photogenerated electrons, to promote the separation and transfer of carriers and to improve the availability of photogenerated electrons.
(3) Z-scheme nanojunctions have broadened the selection range of photocatalysts. Some narrow band gap semiconductors that cannot be used for H2 evolution, such as WO3, can promote the efficient use of solar light and the resulting heterojunctions exhibit strong redox abilities when they design a Z-scheme heterojunction. Therefore, a Z-scheme heterojunction is promising because of the obvious advantages discussed above.
(4) Besides the experimental method, computer simulation and theoretical calculations also need to be combined for an in-depth understanding of the charge transfer mechanism. Scientific and technical challenges remain in the field of photocatalytic water splitting.
2017,
Vol. 38 Issue (8): 1295-1306 
PDF
2017,
Vol. 38 Issue (8): 1295-1306 PDF