Advanced Photocatalytic Nanomaterials for Degrading Pollutants and Generating Fuels by Sunlight

Abstract

This chapter focuses on the recent development of sunlight-driven heterogeneous photocatalysts with different chemical compositions and nanostructures. Various photocatalytic nanomaterials, including metal oxides, heterojunction nanocomposites, oxynitrides, oxysulfides, and graphitic carbon nitride, are described. Their preparation methods as well as the mechanisms involved are introduced. These materials can be used to degrade pollutants and generate fuels. Photocatalytic evolution of H₂ from water and conversion of CO₂ to fuels are discussed in detail. The development of advanced photocatalytic technology involving novel nanomaterials may allow the construction of clean and facile systems for solving the global energy and environmental problems.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

Environment and energy are two of the world’s most challenging issues. For a sustainable society, it is absolutely necessary to develop efficient pollution treatment technologies and harvest clean energy. Photocatalysis has a role to play in both aspects. Photocatalysis is a natural phenomenon that promotes chemical reactions on the surface of an irradiated semiconductor. The essence of the photocatalysis is attributed to the property of photo-excited carriers (electrons and holes) with strong oxidization and reduction power. The redox reactions contribute to the decomposition of hazardous pollutants, conversion of carbon dioxide to valuable hydrocarbons, and the decomposition of water to hydrogen and oxygen.

Photocatalysts are materials that can realize the photocatalysis process. Much attention has been paid to these materials for the development of environment-friendly technology [1–9]. Titanium dioxide (TiO₂) is the most widely used photocatalyst. However, TiO₂ displays a high activity only when it is irradiated by UV light, where the light wavelength is shorter than 400 nm. It is therefore not efficient under sunlight irradiation. Suitable band engineering is required to develop new photocatalysts for solar applications (shown in Fig. 1) [2]. In this chapter, we discuss the fundamental issues that govern the design of visible-light responsive photocatalysts. We also describe some of their applications in environmental and energy aspects.

Fig. 1

Band structure control to develop visible-light-driven photocatalysts for water splitting. Reprinted with permission from Ref. [2]. Copyright 2009 American Chemical Society

2 Solar-Light-Driven Photocatalysts for Degrading Pollutants

2.1 Doped TiO₂ Photocatalysts

The band gap of bulk TiO₂ lies in the UV regime (3.0 eV for the rutile phase and 3.2 eV for the anatase phase) [10]. Solar application of TiO₂ materials is limited by its wide band gap because pure TiO₂ can only absorb a small fraction of the sun’s energy (<10%). To improve the efficiency, doping TiO₂ with metal/nonmetal atoms has proven an efficient route to broadening the photoresponse of TiO₂ to include the visible-light region. Recently, Chen et al. reviewed the modification of TiO₂ with metal/non-metal atoms in detail [10]. Herein, we will focus on the latest reported work related to doping metal/non-metal atoms into the TiO₂ framework.

2.1.1 Metal Doped TiO₂

A number of metal atoms have been doped into the framework of TiO₂ nanomaterials [11–17]. Besides the traditional methods, such as wet chemistry, high-temperature treatment, and ion implantation, various novel routes have been developed for doping metal atoms into the TiO₂ nanomaterials. These include hydro-alcohol thermal, electrospinning, and flame spray pyrolysis (FSP) techniques. Yu et al. fabricated Fe-doped TiO₂ (Fe-TiO₂) nanorods with an impregnating-calcination method using a hydrothermally-prepared titanate nanotube as a precursor and Fe(NO₃)₃ as the dopant. Fe-doping greatly enhanced the visible-light photocatalytic activity of mesoporous TiO₂ nanorods, and when the atomic ratio of Fe/Ti (R-Fe) was in the range of 0.1–1.0%, the photocatalytic activity of the samples was higher than that of Degussa P25 and pure TiO₂ nanorods. At R-Fe = 0.5%, the photocatalytic activity of Fe-TiO₂ nanorods exceeded that of Degussa P25 by a factor of more than two [11]. Wang et al. prepared mesoporous W⁶⁺-doped TiO₂ thin film photocatalysts by electrospinning and sol–gel chemistry through employing a triblock copolymer as a structure-directing agent. 3% W⁶⁺ was found to be the most suitable doping concentration, at which the recombination of photoinduced electrons and holes were effectively inhibited [12]. Li et al. fabricated V-doped TiO₂ (V-TiO₂) nanoparticles using a simple one-step FSP technique. Under visible-light irradiation, the degradation rate of 2, 4-dichlorophenol over 1% V-TiO₂ was two times higher than that of undoped TiO₂ [13]. Li et al. further utilized the one-step FSP technique to fabricate Cr-doped TiO₂ nanoparticles. The optimal Cr³⁺ concentration was found to be 1% [14]. Lorret et al. prepared nanocrystalline tungsten-doped titanium dioxide powders using a sol–gel method based on the hydrolysis of TiCl₄ in aqueous solution. Introducing tungsten into the TiO₂ framework could effectively extend light absorption of the TiO₂-based photocatalysts toward the visible-light range [15]. Dai et al. used a hydro-alcohol thermal method to fabricate Fe-doped titanium dioxide (TiO₂) microspheres with special core–shell structures. The concentration of Fe³⁺ played a key role in the photocatalytic degradation of phenol. Moreover, the 0.5 mol% Fe³⁺ doping was an optimal amount [16]. Yang et al. found that doping ruthenium, by an ion-exchange method, on the hydrothermally synthesized titania nanotube (Ti-NT) greatly enhanced the photocatalytic activity for degrading methylene blue (MB) dye under visible-light irradiation [17].

2.1.2 Non-Metal Doped TiO₂

Different nonmetal elements, such as B, C, N, F, and S, have been utilized recently to modify TiO₂ nanomaterials [18–26]. Xu et al. fabricated B-doped titania hollow spheres. They found that doping boron atoms effectively enhanced the photocatalytic activity of the hollow titania spheres in the degradation of Reactive Brilliant Red dye X-3B (C.I. Reactive Red 2) under-visible light irradiation [18]. Choi et al. reported a carbon-doped TiO₂ (C-TiO₂) photocatalyst prepared from a conventional sol–gel synthesis without using external carbon precursors. The carbon atoms from the titanium alkoxide precursor were incorporated into the lattice of TiO₂, creating mid-bandgap electronic states through controlled calcination [19]. Lu et al. demonstrated a facile route for the one-pot synthesis of visible-light responsive nitrogen-doped anatase TiO₂ sheets with dominant facets of TiN. The as-synthesized anatase TiO₂ sheets showed a strong and stable ability to generate hydroxyl radicals [27].

The UV–visible absorption spectra of the anatase TiO₂ sheets (Fig. 2) shows an additional high visible-light absorption band from 400 nm to ca. 570 nm, consistent with the yellow color of the sample (see inset of Fig. 2). Derived from the plot of the Kubelka–Munk function versus the energy of the light absorbed, assuming titania is an indirect semiconductor, the bandgap of the obtained anatase TiO₂ sheets is extrapolated to be 3.11 eV, which is nearly identical to that of pure bulk anatase TiO₂. However, such nitrogen-doped {001}-dominant anatase TiO₂ sheets show a significantly enhanced visible-light absorption [27]. Yu et al. used a one-step low-temperature hydrothermal approach to fabricate hierarchical porous F-doped TiO₂ microspheres as shown in Fig. 3. These hierarchical porous microspheres exhibited high activity in the photocatalytic degradation of 4-chlorophenol under visible-light illumination [28].

Fig. 2

UV–visible absorption spectra of (a) pure anatase TiO₂ sheets and (b) nitrogen-doped anatase TiO₂ sheets; the insets in the upper right and lower left corners are the plot of transformed Kubelka–Munk function vs the energy of light and optical photo of nitrogen-doped anatase TiO₂ sheets. Reprinted with permission from Ref. [27]. Copyright 2009 American Chemical Society

Fig. 3

SEM images of (a–d) porous F-doped TiO₂ microspheres with different magnifications; (e) an individual single microsphere showing detailed texture and porosity (ca. 1 μm in diameter); (f) EDX microanalysis spectrum of porous F-doped TiO₂ microspheres. Ref. [28]—reproduced by permission of The Royal Society of Chemistry

Yu et al. proposed a one-step low-temperature hydrothermal route to synthesize S-doped TiO₂ photocatalysts from TiS₂ and HCl. Sulfur atoms could be efficiently doped into the anatase lattice under the mild hydrothermal conditions. The S-doped TiO₂ prepared by this hydrothermal approach exhibited much higher photocatalytic activity than that obtained by the traditional high-temperature thermal annealing method for the degradation of 4-chlorophenol under visible-light irradiation [29]. Li et al. prepared a S-doped TiO₂ by treating a TiO₂ xerogel under supercritical conditions in CS₂/ethanol fluid. The TiO₂ was modified through forming S–Ti–O bonds rather than adsorbing CS₂. During liquid-phase photocatalytic degradation of MB under visible-light irradiation, the S-doped TiO₂ exhibited higher activity than that of the undoped TiO₂ and even the N-doped TiO₂ obtained via supercritical treatment. A maximum activity of nearly eight times higher than that of commercially available Degussa P25 was obtained at a 1.8% S/Ti molar ratio [22].

2.2 Nano-heterojunction (TiO₂-Based) Photocatalytic Materials

Modification of the TiO₂ band gap by doping [30–34] and development of new semiconductor materials capable of absorbing visible light [35] are the major strategies for developing visible-light photocatalysts. However, the low quantum efficiency, owing to the fast recombination of photo-generated electron–hole pairs, is still a challenge. This can be partially overcome by the construction of a heterojunction interface between semiconductors with matching band potentials. This allows electric-field-assisted charge transport from one particle to the other [36]. To date, the reported heterojunction semiconductors mainly fall into two categories: TiO₂-based photocatalysts [37, 38] and a small number of non-TiO₂-based systems [39, 40].

Recently, Yu et al. reported a cadmium sulfide quantum dots (QDs) sensitized mesoporous TiO₂ heterojunction photocatalyst [41]. It was prepared by preplanting cadmium oxide as crystal seeds into the framework of ordered mesoporous titanium dioxide and then converting CdO to CdS QDs through ion-exchange (as shown in Fig. 4).

Fig. 4

Schematic synthesis route to ordered mesoporous CdS/TiO₂. Reprinted with permission from Ref. [41]. Copyright 2009 American Chemical Society

The presence of CdS QDs in the TiO₂ framework extended its photoresponse to the visible-light region by accelerating the photogenerated electron transfer from the inorganic sensitizer to TiO₂. The new photocatalyst showed excellent photocatalytic efficiency for both the oxidation of NO gas in air and the degradation of organic compounds (MB and 4-chlorophenol) in aqueous solution under visible-light irradiation [41]. As shown in Fig. 5a, the CdO/TiO₂ sample owned a long-range order structure. The ordered structure could be well maintained even after ion-exchange with S²⁻ (Fig. 5b), though distortions of the pore channels were observed, owing to the in situ transformation of CdO to CdS. The red areas in Fig. 5c represent the S distribution, and the black areas correspond to the pores of the mesoporous CdS/TiO₂. As illustrated in the map, virtually all CdS QDs are highly dispersed on the pore walls of the mesoporous TiO₂. This confirms that CdS QDs are well-integrated into the TiO₂ mesoporous network. The nanocrystalline nature of hexagonal CdS (solid ellipses) and anatase TiO₂ (dot ellipses) are well-defined in the HRTEM image of CdS/TiO₂ as shown in Fig. 5d. These indicate that the heterojunction between CdS and TiO₂ were formed, owing to the intimate contact between CdS and TiO₂. These CdS/TiO₂ heterojunction will lead to a more efficient inter-electron transfer between the two components and improve the charge separation and, therefore, the photocatalytic activity [41, 42].

Fig. 5

a Standard TEM of CdO/TiO₂, b TEM image of CdS/TiO₂, c The chemical map of CdS/TiO₂ (red areas correspond to the S distribution) and d HRTEM image of CdS/TiO₂. Reprinted with permission from Ref. [41]. Copyright 2009 American Chemical Society

Li et al. fabricated LaVO₄/TiO₂ nanocomposite material with interconnected nanocrystal heterojunction by using a simple coupled method [43]. As shown in Fig. 6. The fringes of d = 0.352 nm matched the (101) crystallographic planes of anatase TiO₂, while the fringes of d = 0.296 nm and d = 0.272 nm matched the (012) and (202) crystallographic planes of monoclinic LaVO₄ nanoparticles. Meanwhile, interconnected fine nanoparticulate morphologies that confirmed the formation of LaVO₄/TiO₂ nanocrystal heterojunctions in the composite photocatalyst were observed [43].

Fig. 6

High-resolution TEM image of LaVO₄/TiO₂ nanocomposite. Reprinted with permission from Ref. [43]. Copyright 2009 American Chemical Society

This new type of heterojunction LaVO₄/TiO₂ nanocomposite exhibited very strong photocatalytic activity for decomposition of benzene under visible-light irradiation (450 < λ < 900 nm) with high photocatalytic stability. As shown in Fig. 7, the photocatalytic activities of T500, T400, P25, and LaVO₄ were very low under visible-light irradiation. Nevertheless, the LaVO₄/TiO₂ nanocomposite catalyst showed notably high visible-light photocatalytic activity [43]. Such enhanced photocatalytic performance of LaVO₄/TiO₂ can be attributed to the matched band potentials and the interconnected nanocrystal heterojunction of LaVO₄ and TiO₂. Figure 8 demonstrates a possible photocatalysis process for the degradation of benzene under visible-light irradiation. It includes four steps: (1) Upon visible-light irradiation, electrons and holes generated by LaVO₄ are separated. (2) Some electrons are injected into TiO₂ nanoparticles quickly because the conduction band (CB) of LaVO₄ is more negative than that of TiO₂. The formed nanostructure heterojunction on LaVO₄/TiO₂ composite can also lead to a more efficient inter-electron transfer between the two components [42]. (3) The photogenerated electrons are then captured by O₂ to yield O₂•- and H₂O₂, and then the OHo can be formed by reacting O₂•- with H₂O₂ [32]. The OH• owns a high ability to attack any organic molecules. (4) The photogenerated hole in LaVO₄ also may serve as oxidants to activate some unsaturated organic pollutants (e.g., benzene), leading to subsequent decomposition [43]. Since then, Li et al. reported a Pb(Zr_0.52Ti_0.48)O₃/TiO₂ (PZT/TiO₂) composite photocatalyst with nanostructured heterojunction prepared by a simple sol–gel method. The as-prepared PZT/TiO₂ photocatalyst with large special surface area exhibited enhanced visible-light absorption and high efficient photocatalytic activity for decomposition of ethylene under visible-light irradiation with high photochemical stability [44].

Fig. 7

Conversion of C₆H₆ (a) and the amount of produced CO₂ (b) on LaVO₄/TiO₂, T500, T400, P25, and LaVO₄ under visible-light irradiation and on LaVO₄/TiO₂ in the dark. Reprinted with permission from Ref. [43]. Copyright 2009 American Chemical Society

Fig. 8

Proposed mechanism for the visible-light photodegradation of benzene on LaVO₄/TiO₂ nanocomposite. Reprinted with permission from Ref. [43]. Copyright 2009 American Chemical Society

A FeTiO₃/TiO₂ heterojunction structure containing a FeTiO₃ nanodisc and Degussa P25 was prepared by using maleic acid as an organic linker [45]. The FeTiO₃ nanodisc was a single-crystalline ilmenite phase with its face oriented in (001) plane and grown to the (110) direction. The 5/95 FeTiO₃/TiO₂ exhibited the optimized photocatalytic activity in removing 2-propanol and evolving CO₂ in the gas phase under visible-light irradiation. Its degradation constant (k) for removing 2-propanol was 25 times that of Degussa P25. The remarkably enhanced photocatalytic activity of FeTiO₃/TiO₂ was attributed to the intersemiconductor hole-transfer mechanism due to the unique relative band positions of these two semiconductors [45]. As shown in Fig. 9, the FeTiO₃/TiO₂ system is an example of the type-B heterojunction. The photocatalytic reaction takes place based on inter-semiconductor hole-transfer . The valence band (VB) position of FeTiO₃ is very close to that of TiO₂, while its CB is much lower than that of TiO₂ (∼0.5 V lower). The VB of FeTiO₃ is rendered partially vacant by band gap excitation under visible-light irradiation. The electrons in the VB of TiO₂ can be transferred to that of FeTiO₃. Thus, the holes generated in VB of TiO₂ have a sufficient lifetime to initiate the various photocatalytic oxidation reactions [45].

Fig. 9

Energy-band diagram illustrating the type-B heterojunction of TiO₂ and FeTiO₃ with visible-light irradiation. Reprinted with permission from Ref. [45]. Copyright 2009 American Chemical Society

FeOOH/TiO₂, a heterojunction structure between FeOOH and TiO₂, was prepared by covering the surface of the similar to 100-nm-sized FeOOH particles with Degussa P25 by applying maleic acid as ail organic linker [46]. Under visible-light irradiation, such FeOOH/TiO₂ heterojunction structure showed notable photocatalytic activity for the removal of gaseous 2-propanol and evolution of CO₂.

2.3 Non-TiO₂ Photocatalytic Materials

2.3.1 Bi₂WO₆

Semiconducting materials of the Aurivillius oxides Bi₂A_n−1B _n O_3n+3 (A = Ca, Sr, Ba, Pb, Na, K, and B = Ti, Nb, Ta, Mo, W, Fe) have been extensively studied because of their layer structure and unique properties [47, 48]. Among these compounds, Bi₂WO₆, as the simplest member of the Aurivillius family of layered perovskites, has been extensively utilized as an excellent photocatalyst for water splitting and photodegradation of organic compounds under visible-light irradiation [49–51]. Kudo et al. found that Bi₂WO₆ had photocatalytic activity for O₂ evolution [52] and Zou et al. revealed that Bi₂WO₆ could degrade organic compounds under visible-light irradiation [53]. Wang et al. fabricated flower-like structured Bi₂WO₆ through the hydrothermal route without using any surfactants or templates. The products exhibited strong visible-light-driven photocatalytic performance for the treatment of RhB due to the novel hierarchical transport pores of the flower-like superstructures [54]. Subsequently, Xie et al. used the hydrothermal method to synthesize a Bi₂WO₆ hierarchical nest-like structure with the assistance of PVP [55]. Since then, new types of Bi₂WO₆ with complex morphologies, namely, flower-like, tyre- and helix-like, and plate-like shapes, were selectively synthesized via a hydrothermal process with P123 as a template by Wang et al. [56].

Recently, Zhu et al. prepared fullerene (C₆₀) modified Bi₂WO₆ photocatalyst by an absorbing process [57]. As shown in Fig. 10, the lattice structure of Bi₂WO₆ was observed from the center to the boundary (Fig. 10a). The outer boundary of Bi₂WO₆, modified by C₆₀, was distinctly different (Fig. 10b). An outer layer with an amorphous structure surrounded the surface of the Bi₂WO₆ nanosheet. The thickness of the layer was estimated to be about 1 nm, very close to the diameter of C₆₀. Therefore, it was concluded that C₆₀ was dispersed on the surface of Bi₂WO₆ with a monolayer structure [57].

Fig. 10

HRTEM images of (a) Bi₂WO₆ and (b) the C₆₀-modified Bi₂WO₆ sample. Reprinted with permission from Ref. [57]. Copyright 2007 American Chemical Society

Such new composite photocatalyst exhibited a high efficiency for the degradation of nonbiodegradable azodyes MB and rhodamine B (RhB) under visible-light (λ > 420 nm) and simulated solar light (λ > 290 nm). The enhanced photocatalytic activity for the C₆₀-modified Bi₂WO₆ could come from the high migration efficiency of the photo-induced electrons on the interface of the C₆₀ and Bi₂WO₆. The delocalized conjugated π structure of C₆₀ made the transfer of photoinduced electrons easier [58]. The schematic of photocatalytic mechanism is shown in Fig. 11 [57].

Fig. 11

Possible pathway of the photoelectron transfer excited by visible-light irradiation including photocatalytic process for C60-modified Bi₂WO₆. Reprinted with permission from Ref. [57]. Copyright 2007 American Chemical Society

More recently, Zhu et al. used a two-step process to synthesize F-substituted Bi₂WO₆ (Bi₂WO_6−X F_2X ) photocatalysts with high activity. F-substitution changed the original coordination around the W and Bi atoms. Compared with Bi₂WO₆, the photocatalytic activity of Bi₂WO_6−X F_2X increased about two times for the degradation of MB under visible light irradiation. Density functional calculations revealed that Bi₂WO_6−X F_2X has a wider valence bandwidth and lower VB position. The high activities of Bi₂WO_6−X F_2X photocatalysts come from its VB, which increase the mobility of photo-excited charge carriers and possess a stronger oxidation power [59].

Lin et al. introduced the photoelectrochemical (PEC) concept to the photocatalytic oxidation application of Bi₂WO₆ [60]. Hydrothermal combined with a spin coating technique was utilized to fabricate a Bi₂WO₆ nanoplate film electrode. As shown in Fig. 12, PCE experiments were performed in the anodic cell, using Bi₂WO₆/ITO electrode with the area of 3 cm² in 0.005 mol L⁻¹ of Na₂SO₄ electrolyte solution under visible-light illumination. The voltage applied in the EC and PEC systems was 1.2 V. During the entire experiment, the solutions in the anodic and cathodal cells were magnetically stirred. During photocatalytic oxidation process, the Bi₂WO₆/ITO electrode only worked as a photocatalyst without an applied bias [60]. The PEC system based on Bi₂WO₆ nanoplate film electrode degraded 87.2% of RhB with concentration of 5 mg L⁻¹ in 120 min, operated at low voltage and under visible-light irradiation, whereas only 36.8 and 39.5% degradation of RhB were observed for the electro-oxidation process (EC) and photocatalytic oxidation processes (PC), respectively, operated under the same condition. These results revealed a significant synergetic effect on degrading RhB via electro-oxidation and photocatalysis under visible-light irradiation [60].

Fig. 12

Schematic illustration of a PEC system: (1) visible-light lamp; (2) magnetic stirrer; (3) air compressor; (4) anodic cell; (5) cathodal cell; (6) Bi₂WO₆/ITO electrode; (7) KCl bridge; (8) Pt cathode; (9) CHI-600A potentiostat; (10) RhB solutions. Reprinted with permission from Ref. [60]. Copyright 2007 American Chemical Society

Wang et al. developed an electrospinning technique to fabricate Bi₂WO₆ nanofibrous mat with excellent photoactivity under visible-light irradiation [61]. The as-prepared nanofibers are made of single-crystalline Bi₂WO₆ nanoparticles about 100 nm in size. The diameters of the nanofibers can be controlled by tuning the weight ratio (R) of Bi₂WO₆ to poly(vinyl pyrrolidone) (PVP). As shown in Fig. 13, when the R is 0.5 (Fig. 13a), it was unfavorable for the formation of uniform nanofibers, presumably due to the low content of Bi₂WO₆ nanoparticles that could not be connected during the calcination process. The average diameter of the Bi₂WO₆ nanofibers was about 450 nm. By increasing R from 1 to 2, the average diameter of the fibers decreased from 300 to 170 nm (Fig. 13b–d) [61]. In addition to the favorable recycling characteristics, the mat with R = 2 exhibited higher photocatalytic activity in the decomposition of acetaldehyde (CH₃CHO) and aqueous ammonia than that of the sample prepared by the solid-state reaction (SSR) and the nanoparticles. Electrospinning is favorable for the effective improvement of the photocatalytic activity of catalysts. It also can provide a solution to the separation problem in conventional catalysts that are small in size. Thus, it is worth considering for the preparation of other photocatalysts [61].

Fig. 13

The SEM images of Bi₂WO₆ nanofibers with (a) R = 0.5, (b) R = 1, (c) R = 1.5, and (d) R = 2. Ref. [61]—reproduced by permission of The Royal Society of Chemistry

2.3.2 BiVO₄

BiVO₄, with a narrower band gap of 2.4 eV, is an important visible-light responsive photocatalyst, widely used in the evolution of photocatalytic O₂ and the photocatalytic degradation of organic pollutants [62, 63]. There are three naturally occurring crystal forms of BiVO₄, namely tetragonal zircon, monoclinic scheelite, and tetragonal scheelite. Monoclinic scheelite presents the best photocatalytic performance under visible-light irradiation [64]. To date, various methods of synthesizing monoclinic scheelite BiVO₄ have been developed, including aqueous [62], solid-state [65], and hydrothermal processes [66], as well as organometallic decomposition [67].

Single-crystalline BiVO₄ microtubes, with novel square cross-sections and a flower-like morphology, were synthesized by a reflux method at 80°C [68]. The as-synthesized products were almost entirely microtubes with lengths of 2–5 μm. The individual tubes had well-defined square cross-sections with side lengths of ca. 800 nm and wall thicknesses of ca. 100 nm. The clear lattice fringe indicated the high-crystallinity and single-crystalline nature of the microtubes. The d spacings were found to be 0.581 and 0.468 nm, which agreed well with the lattice spacings of (020) and (011) of monoclinic BiVO₄ (Fig. 14c, d) [68]. The as-synthesized microtubes exhibited higher photocatalytic activity under visible-light radiation than that of the reference sample prepared by a solid-state reaction. This was ascribed to the special single-crystalline tubular structure and/or flower-like morphology [68].

Fig. 14

BiVO₄ microtubes synthesized at 80°C for 6 h: a low-magnification SEM image of the product and high-magnification SEM image (inset) for a single BiVO₄ microtube, showing the hollow structure and wall thickness; b TEM and SEM image (inset) of an individual BiVO₄ microflower; c the corresponding SAED pattern taken from the rectangular part of the microtube in Fig. 14b; d HRTEM image near the nozzle of a single microtube (inset). Reprinted with permission from Ref. [68]. Copyright 2007 American Chemical Society

Yu et al. used a nanocasting route to fabricate ordered mesoporous bismuth vanadate (BiVO₄) crystals using bismuth nitrate hydrate and ammonia metavanadate as bismuth and vanadium sources and silica (KIT-6) as a template [69]. Monoclinic scheelite BiVO₄ crystals were formed inside the mesopores of silica through a mild thermal process, and BiVO₄ was obtained after the removal of the hard template (silica) by NaOH treatment. Such product exhibited a superior photocatalytic performance in the photochemical degradation of MB and photocatalytic oxidation of NO gas in air under visible-light irradiation compared to conventional BiVO₄ [69]. Wang et al. used an aqueous strategy to synthesize two novel types of BiVO₄ mesocrystals with flower-like and dendrite morphology. The primary building units perfectly aligned in three-dimensions and built up well-defined mesocrystals with sharp facets and edges. The pH value and reaction temperature had great influences on the formation of these unique mesocrystals [70]. Recently, BiVO₄ powders with unique particle architectures were synthesized using ultrasonic spray pyrolysis (USP) [71]. These nanostructured BiVO₄ with particles range from thin, hollow, and porous shells to ball-in-ball type structures, as shown in Fig. 15. The BiVO₄ powders were utilized as an oxygen evolving photocatalyst and the kinetics of O₂ formation was studied in a AgNO₃ solution irradiated with λ > 400 nm light. USP prepared BiVO₄ was found to have superior photocatalytic activity compared to commercial BiVO₄ and WO₃, likely due to the differences in particle morphology [71].

Fig. 15

a–d SEM and e and f TEM micrographs illustrating typical particle morphologies obtained during USP synthesis of BiVO₄. Reprinted with permission from Ref. [71]. Copyright 2009 American Chemical Society

Li et al. prepared a series of Cu-loaded BiVO₄ (Cu-BiVO₄) photocatalysts by the impregnation method [72]. The Cu-BiVO₄ series catalysts had significant optical absorption in the visible region between 550 and 800 nm and the absorption intensity increased with the enhancement of Cu content as shown in Fig. 16. The photocatalytic activities of Cu-BiVO₄ catalysts for the degradation of MB were found to depend largely on the Cu content and the calcination temperature. The optimum Cu loading and calcination temperature were found to be 5% and 300°C [72]. Novel Pt/BiVO₄ composite photocatalysts with visible-light activities were prepared by the impregnation method [73]. Monoclinic scheelite structures of BiVO₄ were retained after the Pt species was doped. The visible-light absorption of the BiVO₄ photocatalysts was obviously enhanced upon modifying with Pt species (Fig. 17). The platinum doped in the composite photocatalyst was present in the form of platinum chloride (PtCl₄). The Pt species doping effectively enhanced the photoactivities of BiVO₄ samples in decomposition of methyl orange under visible-light irradiation [73].

Fig. 16

UV–Vis diffuse-reflectance spectra of pure and Cu–BiVO₄ series catalysts. Reprinted from Xu et al. [72], copyright 2008 with permission from Elsevier

Fig. 17

UV–Vis diffuse absorption spectra of different photocatalysts. Reprinted from Ge [73], copyright 2008 with permission from Elsevier

2.3.3 Graphitic Carbon Nitrides (g-C₃N₄) Polymeric Photocatalyst

Among the different allotropes of carbon nitrides, the graphitic phase is usually regarded as the most stable under ambient conditions. Recently, graphitic carbon nitride (g-C₃N₄), as a conducting polymer semiconductor, has been proven an efficient photocatalyst for water splitting, even in the absence of noble metals [74]. Thermal polycondensation of common organic monomers was utilized to synthesize graphitic carbon nitrides (g-C₃N₄) with various architectures [75, 76]. As shown in Fig. 18, the graphitic planes are constructed from tri-s-triazine units connected by planar amino groups (Fig. 18a). Upon condensation at 823 K, an in-plane repeat period of 0.681 nm (for example, the distance between nitride pores) in the crystal is evident from the X-ray powder diffraction (XRD) pattern (Fig. 18b). It is smaller than one tri-s-triazine unit (0.713 nm), presumably owing to the presence of a small tilt angularity in the structure. The strongest XRD peak at 27.4, corresponding to 0.326 nm, is due to the stacking of the conjugated aromatic system. The band gap of the gold--yellow condensed graphitic carbon nitride is estimated to be 2.7 eV from its ultraviolet–visible spectrum (Fig. 18c). It implies an intrinsic semiconductor-like absorption in the blue region of the visible spectrum. Such bandgap is sufficiently large to overcome the endothermic character of the water-splitting reaction (requiring 1.23 eV theoretically) [74].

Fig. 18

Crystal structure and optical properties of graphitic carbon nitride. a Schematic diagram of a perfect graphitic carbon nitride sheet constructed from melem units. b Experimental XRD pattern of the polymeric carbon nitride, revealing a graphitic structure with an interplanar stacking distance of aromatic units of 0.326 nm. c Ultraviolet–visible diffuse-reflectance spectrum of the polymeric carbon nitride. Inset: Photograph of the photocatalyst. Reprinted by permission from Macmillan Publishers Ltd: (Nature Materials) (Ref. [74]), copyright 2009

Recently, the electronic and optical functions of polymeric g-C₃N₄ were further modified by the inclusion of metal species, such as Fe³⁺, in its structure by a simple soft-chemical method without destroying the graphitic structure of the host. The metal components strongly affected the electronic properties of g-C₃N₄ and provided the material with additional new functionalities such as mimicking metalloenzymes in H₂O₂ activation. The metal species can significantly lower the bandgap and expand the light absorption of the material further into the visible region of the electromagnetic spectrum, while keeping a sufficient overpotential for carrying out oxidation reactions [77]. As shown in Fig. 19, the optical band gap energy gradually shifts to lower energies with increasing Fe content in the Fe/g-C₃N₄ hybrid materials. This suggests a host–guest interaction between g-C₃N₄ and the metal. A change in the optical absorption was also observed for Zn/g-C₃N₄ and is probably caused by the d–p repulsion of the Zn 3d and N2p orbitals [77, 78].

Fig. 19

Optical absorption spectra of Fe/g-C₃N₄ complexes and g-C₃N₄; the bandgap is shifted toward lower energies in the hybrid materials. Arrow direction: g-C₃N₄, 1%-, 3%-, 5%-, 10%-, 15%-, and 20%-Fe/g-C₃N₄. The inset is the optical spectrum of 10%-Zn/g-C₃N₄, which also demonstrates narrowing of the g-C₃N₄ bandgap by the metal inclusion. Wang et al. [77], copyright Wiley–VCH Verlag GmbH & Co. KGaA, reproduced with permission

Photocatalytic experimental results confirmed that various organic dyes (e.g., RhB, MB, methyl orange, and p-hydroxyazobenzene) were be degraded by using H₂O₂ and Fe/g-C₃N₄. The photocatalyst can also be recovered and reused [77]. Figure 20 shows the results from the RhB-oxidation and control experiments. The overall efficiency of the process can be enhanced by photoillumination (λ > 420 nm). Various intermediates, such as hydroxylated and de-ethylated, were observed during the oxidation of RhB by H₂O₂ activated with Fe/g-C₃N₄. These can be further decomposed to smaller organic molecules, eventually being mineralized to CO₂ [77].

Fig. 20

a Concentration changes of RhB (10 mM) as a function of reaction time under different conditions: a H₂O₂ (0.01 M); b Fe/g-C₃N₄(40 mg); c Fe₂O₃(40 mg)/H₂O₂(0.01 M); d Fe/g-C₃N₄(40 mg)/H₂O₂(0.01 M); e Fe/g-C₃N₄(40 mg)/H₂O₂(0.01 M) under visible-light irradiation. b Cyclic runs of RhB (10 mM) degradation by H₂O₂ (0.05 M) activated by the Fe/g-C₃N₄ catalyst (40 mg). All reactions were carried out at neutral pH using the 10%-Fe/g-C₃N₄ catalyst. C: concentration; C₀: initial concentration. Wang et al. [77], copyright Wiley–VCH Verlag GmbH & Co. KGaA, reproduced with permission

More recently Fe/g-C₃N₄ was found to be capable of activating hydrogen peroxide for the direct oxidation of benzene to phenol in mild conditions. Fe-g-C₃N₄ is active for the direct oxidation of benzene to phenol using hydrogen peroxide. By taking advantage of the photocatalytic functions of g-C₃N₄, the yield of the phenol synthesis can be markedly improved [79]. Figure 21a shows the catalytic properties of graphitic carbon nitride. Figure 21b implies that it is able to adsorb and activate benzene chemically (Fig. 21b) and, thus, catalyzed not only Friedel–Crafts reactions of benzene but also phenol synthesis using benzene and CO₂ [75, 80].

Fig. 21

a Stacked g-C₃N₄ sheets function as an all-organic solid-state photocatalyst promoting redox reactions with visible light. b Chemical interaction of benzene and defective g-C₃N₄ via HOMO–LUMO hybridization of melem and benzene. Reprinted with permission from Ref. [79]. Copyright 2009 American Chemical Society

2.3.4 Heterojunction of Non-TiO₂ Semiconductors

Much progress has been achieved in the development of TiO₂-based heterojunction photocatalytic systems. However, there is still need to develop highly active photocatalysts working under visible light in order to put this technology into practical applications. Recently, heterojunctions of non-TiO₂ semiconductors has become a hot topic in the photocatalytic research field.

Efficient visible-light active CaFe₂O₄ (CFO)/WO₃ composite photocatalysts were prepared by Miyauchi et al. [81]. The composite containing 5 wt% CFO showed optimized performance. Surface modification was made by heating the CFO/WO₃ composite or by coating the particle surface with Ag or ITO. The photocatalytic activity was greatly improved by coating the CFO particles with a Ag or ITO layer. The CO₂ generation profile over pure WO₃, 5 wt% CFO/WO₃, Ag@CFO/WO₃ photocatalysts were presented in Fig. 22. The CO₂ generation almost stops after 40 h of irradiation over pure WO₃. Pure WO₃ cannot completely decompose acetaldehyde to CO₂ even after more than 100 h under the present experimental conditions. However, acetaldehyde can be completely decomposed to CO₂ over a 5 wt% CFO/WO₃ composite in 48 h. It is impressive that the Ag modified composite Ag@CFO/WO₃ can reach complete decomposition of acetaldehyde in 20 h, which is more than twice as fast as the CFO/WO₃ photocatalysts [81].

Fig. 22

CO₂ photogeneration profile at a long time visible irradiation. The initial acetaldehyde concentration is 500 ppm. The visible light density is 80 mW/cm². Reprinted with permission from Ref. [81]. Copyright 2009 American Chemical Society

A network-structured SnO₂/ZnO heterojunction nanocatalyst with high photocatalytic activity was synthesized through a simple two-step solvothermal method [82]. As shown in Fig. 23, the UV–vis absorption edges of the as-synthesized SnO₂ semicrystals and ZnO nanorods are located at about 305 and 380 nm, respectively. There are two prominent absorption bands for the SnO₂/ZnO sample. The former is assigned to the absorption of SnO₂ semicrystals and the latter is attributed to the characteristic absorption of ZnO nanocrystals. The absorption edges of SnO₂ and ZnO nanocrystals in the SnO₂/ZnO sample slightly shift toward blue. This indicates that the sizes of SnO₂ and ZnO in the SnO₂/ZnO sample are smaller than the corresponding values of pure SnO₂ or ZnO [82].

Fig. 23

UV–vis diffuse-reflectance spectra of the as-synthesized samples: (a) ZnO, (b) SnO₂/ZnO, and (c) SnO₂. Reprinted with permission from Ref. [82]. Copyright 2009 American Chemical Society

The photocatalytic activity of SnO₂/ZnO heterojunction nanocatalysts for the degradation of methyl orange is much higher than those of solvothermally synthesized SnO₂ and ZnO samples. Figure 24 shows the proposed band structure of the as-synthesized SnO₂/ZnO heterojunction nanocatalyst [82, 83]. Upon formation of the heterojunction between SnO₂ and ZnO, the different work functions will induce the negatively charged carriers to move from SnO₂ (the material with low work function) to ZnO (the one with high work function) until their Fermi levels align (i.e., the system reaches thermal equilibrium). Thus, an electrostatic field is created at the interface. At thermal equilibrium, the CBs and VBs of SnO₂ and ZnO bend, and a depletion layer forms around the interface, too. Under UV light irradiation, electrons (e−) in the VB can be excited to the CB while simultaneously generating the same amount of holes (h+) in the VB. The photogenerated electrons and holes can be separated under the influence of the electrostatic field induced by different work functions. Therefore, electrons move to the SnO₂ side and holes to the ZnO side. The photogenerated electrons and holes in the SnO₂/ZnO heterojunction nanocatalyst can be injected into a reaction medium and participate in chemical reactions [82].

Fig. 24

Energy-band diagram and photocatalytic mechanism of the as-synthesized SnO₂/ZnO heterojunction nanocatalyst, where vac is the vacuum level, E _f is the Fermi level, CB is the conduction band, and VB is the valence band. Reprinted with permission from Ref. [82]. Copyright 2009 American Chemical Society

Bi₂O₃ is a good n-type semiconductor and BaTiO₃ is a high dielectric and ferroelectric material, where some atoms in the lattice are movable. Based on this, Huang et al. used a milling-annealing technique to prepare a heterojunction photocatalyst, Bi₂O₃/BaTiO₃ [36]. This technique has an advantage over the direct mixing method because it can construct a tight chemically bonded interface between the coupled materials. The heterojunction semiconductors Bi₂O₃/BaTiO₃ showed better photocatalytic activities than single-phase BaTiO₃ or Bi₂O₃ for degrading methyl orange and MB. The remarkable enhancement in the photocatalytic performance of Bi₂O₃/BaTiO₃ was ascribed mainly to the electric field-driven electron–hole separation at the interface and in the two semiconductors. Besides, the fair mobility for electron and hole transportation in Bi₂O₃ and BaTiO₃, respectively, were also favorable for the high photocatalytic property [36].

Lee et al. also reported a novel BiOCl/Bi₂O₃ heterojunction-type photocatalyst [35]. The TEM image in Fig. 25a reveals that the 85/15 BiOCl/Bi₂O₃ is a well-defined longish particle of ~200 nm width. As shown in the high-resolution TEM image of Fig. 25b, the outer part of the BiOCl/Bi₂O₃ particle reveals a clear image, suggesting the presence of crystalline BiOCl. The uniform fringe, with an interval of 0.73 nm, is in good agreement with the (001) lattice plane of the tetragonal BiOCl. On the other hand, as shown in Fig. 25c, the TEM image for the core of the BiOCl/Bi₂O₃ particle is not as clear as that of outer part and several sets of mixed fringes are found, indicating the presence of mixed phases of BiOCl and Bi₂O₃ that is, the interlayer distance of 0.73 nm corresponds to the (001) lattice plane of BiOCl, and that of 0.33 nm is consistent with the (120) plane of α-Bi₂O₃. These observations indicate that the nano-sized Bi₂O₃ grains are embedded here and there inside the BiOCl matrix [35]. Though both the individual BiOCl and Bi₂O₃ show very low photocatalytic efficiency under visible-light irradiation, their heterojunctions provide unexpectedly high efficiency in decomposing organic compounds. The BiOCl/Bi₂O₃ can induce complete mineralization without formation of intermediate species by utilizing the holes generated in the VB of BiOCl. Compared to the Degussa P25, it demonstrates 5.7 times the efficiency in evolving CO₂ from gaseous 2-propanol (IP) and 10.5 times the efficiency in removing aqueous 1, 4-terephthalic acid (TA) under visible-light irradiation. In this BiOCl/Bi₂O₃ system, the BiOCl seems to work as the main photocatalyst, while the role of Bi₂O₃ is a sensitizer, absorbing visible light [35].

Fig. 25

TEM images for an 85/15 BiOCl/Bi₂O₃ particle. Typical TEM image (a), and HRTEM images for the outer (b) and inner (c) parts of the sample. Reprinted from Chai [35], copyright 2009 with permission from Elsevier

3 Solar-Light-Driven Photocatalysts for Generating Fuels

3.1 Solar-Light-Driven Photocatalysts for H₂ Evolution

The concern for the depletion of fossil fuels and the environmental problems accompanying their use fostered the research for viable alternatives. Many research efforts have been devoted to the generation of hydrogen since it is the fuel with the highest energy capacity per unit mass. Hydrogen is also a clean energy carrier because it produces neither CO₂ nor pollutants. Many reviews on photocatalytic water splitting have been published [2, 9, 84–89]. In the following sections, we focus on visible-light-driven heterogeneous photocatalytic materials, such as metal oxides, metal oxynitrides, metal oxysulfides, metal sulfides, and polymers for H₂ evolution.

3.1.1 Metal Oxide Photocatalysts

To obtain photocatalytic activity under visible-light irradiation, it is essential to control the interdependence between the electronic, microstructural, and surface properties of photocatalysts by means of a careful design of both bulk and surface properties. The strategies can be classified in five categories: (a) developing new single-phase photocatalysts; (b) tuning the band gap energy with ion doping; (c) surface modification by depositing co-catalysts; (d) sensitization; and (e) controlling the defects, size, and morphology. This section briefly reviews the recent developments in oxide photocatalysts (Table 1) that show activity under visible light.

Table 1 Overview of recently developed metal oxide photocatalysts for H₂ generation under visible-light illumination

3.1.2 Oxynitride and Oxysulfide Photocatalysts

Domen and co-workers have done extensive studies on oxynitride and oxysulfide visible-light-driven photocatalysts [87, 88, 90–99]. The hydrogen generation application for these catalysts has been reviewed in previous reports [100, 101]. Herein, we summarize the more recent publications related to the oxynitride and oxysulfide photocatalysts.

Takanabe et al. studied the photocatalytic water-splitting reactions over the (Zn_1+x Ge)(N₂O _x ) photocatalyst [102]. The photocatalyst showed high rates for overall water splitting under visible irradiation. Their results showed negligible changes in the structure and composition of the photocatalyst after the photocatalytic reaction. The photocatalytic activity was improved by metal doping in the oxynitride formulation and post-calcination after nitridation. The improvement of photocatalytic activity was ascribed to the reduction of the number of defects in the photocatalyst materials.

TaON nanotube arrays (shown in Fig. 26) were synthesized via sonoelectrochemical anodization followed by nitridation [103]. They exhibited efficient performance for photoelectrochemical generation of hydrogen from water. The photocatalytic activity of (Ga_1−x Zn _x )(N_1−x O _x ), a solid solution of GaN and ZnO, for H₂ evolution in the presence of methanol as a sacrificial reagent under visible light was investigated in detail [104]. (Ga_1−x Zn _x )(N_1−x O _x ) evolved H₂ from an aqueous methanol solution when loaded with nanoparticulate Rh_2−y Cr _y O₃ as a cocatalyst. The H₂ evolution activity was strongly dependent on the crystallinity and composition of the catalyst. The quantum efficiency for overall water splitting increased to 2.5% at 420–440 nm [105]. This represented a tenfold increase in efficiency over the highest efficiency previously obtained using nanoparticulate RuO₂ as a cocatalyst. Besides, the dispersion and size of cocatalyst nanoparticles were identified as important factors affecting the degree of enhancement for stoichiometric water splitting. The results of photocatalytic reactions and photoelectrochemical measurements suggested that the rate-determining step for overall water splitting using (Ga_1−x Zn _x )(N_1−x O _x ) was the H₂ evolution process [105].

Fig. 26

a FESEM images of TaON nanotube arrays on Ta foil. The insets show the cross sectional image of Ta₂O₅ NT arrays. b HRTEM and FFT pattern of TaON NTs. Ref. [103]—reproduced by permission of The Royal Society of Chemistry

Lee et al. developed a zinc germanium oxynitride, a solid solution between ZnO and ZnGeN₂, through a reaction of GeO₂ and ZnO under an NH₃ flow [106]. The samples nitrided for 5–15 h under these conditions exhibited a single phase of wurtzitic (Zn_1+x Ge)(N₂O _x ) and were responsive to visible light with a band gap of ca. 2.7–2.8 eV. Nitridation for 15 h afforded (Zn_1+x Ge)(N₂O _x ) with the highest photocatalytic activity for overall water splitting. A variety of cocatalysts were also examined, and Rh_2−x Cr _x O₃ was identified as the most effective cocatalyst for (Zn_1+x Ge)(N₂O _x ), which caused an increase in the activity for hydrogen evolution. Modification of the optimized (Zn_1.44Ge)(N_2.08O_0.38) sample by loading with Rh_2−x Cr _x O₃ (3.0 wt% Rh, 0.2 wt% Cr) resulted in an effective photocatalyst for overall water decomposition with a quantum efficiency of ca. 0.20% at 420 nm.

A rose-red color Nb₂Zr₆O_17−x N _x oxynitride photocatalyst was synthesized by thermal ammonolysis of Nb₂Zr₆O₁₇ at 1073 K [107]. TEM images of the Nb₂Zr₆O_17−x N _x sample showed prismatic pseudo orthorhombic shaped particles with clear edges and an average particle size in the range of 80–90 nm. The oxynitride Nb₂Zr₆O_17−x N _x gave a quantum yield of 13.5% in the production of hydrogen from the decomposition of hydrogen sulfide under visible-light irradiation. A d⁰–d¹⁰ complex photocatalyst, zinc, and titanium spinel oxynitride (Zn _x TiO _y N _z ) reduced H⁺ to H₂ in the presence of a sacrificial electron donor under visible-light irradiation [108]. Ogisu et al. reported a lanthanum–indium oxysulfide visible-light-driven (420 < λ < 480 nm) photocatalyst for water splitting. Loading with Pt is effective for promoting H₂ evolution [99].

3.1.3 Metal Sulfide Photocatalysts

Metal sulfides photocatalysts have been widely studied due to their outstanding performance in hydrogen generation via photocatalysis. The VB usually consists of S 3p orbitals the level of which is more negative than O 2p. The photocorrosion problem is usually solved by adding sacrificial reagents such as S²⁻ and SO₃ ²⁻ into the water-splitting system. Many visible-light-driven metal sulfide photocatalysts have been summarized in a previous report [2]. Table 2 lists the new photocatalysts reported in the last 3 years.

Table 2 Sulfide photocatalysts for H₂ evolution from aqueous solutions in the presence of sacrificial reagents

3.1.4 Polymeric Photocatalysts

Synthetic polymer semiconductors such as polyparaphenylene have also been used for hydrogen production under UV illumination [109]. Recently, a metal-free polymeric, visible-light-driven photocatalyst for hydrogen production was reported [74]. It was a graphitic carbon nitride (g-C₃N₄) synthesized via a thermal polycondensation of cyanamide. The bandgap of g-C₃N₄ was estimated to be 2.7 eV from its ultraviolet–visible spectrum, showing an intrinsic semiconductor-like absorption in the blue region. Different thermal condensation enabled the finer adjustment of the electronic and optical properties. The photocatalyst produced H₂ from water containing triethanolamine as a sacrificial electron donor upon light illumination (λ > 420 nm) in the absence of noble metal catalysts such as Pt. No N₂ evolution was observed for this catalyst, even after a very long irradiation time, indicating excellent stability due to the strong binding of N in the covalent carbon nitride. This is the first polymeric photocatalyst that is cheap and commonly available. It will open new avenues for organic semiconductors as energy transducers.

However, the quantum yield of the above system (0.1% at 420–460 nm) must be improved. The efficiency of hydrogen production over g-C₃N₄ can be improved by tailoring its nanostructure. A mesoporous structure can enhance the light harvesting ability and mass transfer due to its large surface and multiple scattering effects. Wang et al. advanced g-C₃N₄ by generating a nanoporous structure into the polymeric matrix to improve its structural and electronic functions for solar energy conversion [110]. The photocatalyst mpg-C₃N₄ has a 3D porous framework (shown in Fig. 27), exhibiting an improved efficiency by an order of magnitude. The improved catalytic efficiency was due to the large surface area, which was a basic requirement for a heterogeneous (photo)catalyst to be chemically productive. This example shows excellent artificial photosynthesis over mesoporous polymer semiconductors.

Fig. 27

a TEM image of mpg-C₃N₄, showing a 3D porous framework constructed from tri(s)triazine units. The stacking distance of 0.332 nm is evident by the intense electron diffraction ring(inset), providing high partial crystallinity of the wall. b TEM image of bulk g-C₃N₄. Reprinted with permission from Ref. [110]. Copyright 2009 American Chemical Society

The ordered mesostructure permits the structural orientation of guest molecules in the periodic nanopores, which enhances the selectivity and activity in photocatalysis [69, 111, 112]. Very recently, highly ordered porous g-C₃N₄ materials (shown as ompg-C₃N₄ in Fig. 28) were synthesized via a SBA-15 template route [113]. The photocatalytic activity was evaluated by photochemical reduction of water in the presence of an electron donor with visible light [113]. The total evolution of H₂ reached 2.1 mmol during the course of 25 h visible-light irradiation. The H₂ evolution on the ordered mesoporous C₃N₄ was about five times higher than that of bulk g-C₃N₄ [110]. Such a structure is promising as a host semiconductor scaffold for the design of hybrid visible-light photocatalyst. Furthermore, the photocatalyst surface can be functionalized easily via surface reaction or deposition. Various cocatalysts such as chromophoric antenna molecules can be coassembled into the ordered mesoporus carbon nitride, generating new biomimetic photocatalyst systems.

Fig. 28

a SAXS patterns of ompg-C₃N₄ and SBA-15 template. The inset shows the corresponding 2D SAXS image of ompg-C₃N₄. b, c Typical TEM images of ompg-C₃N₄. The insets show the corresponding fast Fourier transforms of the patterns. Reprinted with permission from Ref. [113]. Copyright 2009 American Chemical Society

3.2 Solar-Light-Driven Photocatalysts for Valuable Hydrocarbon Evolution from CO₂

In recent years, carbon dioxide emissions from the burning of fossil fuels have grown to ~2.5 × 10¹⁰ metric tons per annum. This presents a significant environmental challenge for the twenty-first century. In nature, CO₂ is removed from the environment by photosynthesis. The energy obtained from sunlight is ultimately used to convert CO₂ into glucose, a sugar molecule that stores solar energy in the form of chemical energy. However, the efficiency of energy transformation is low. Even under the optimal artificial conditions (microalgae in full sunlight), the energy efficiency is only about 7% [8].

Carbon sequestration is considered a promising interim solution to global warming. It involves the capture and storage of fossil fuel-derived CO₂ emissions to prevent their release into the atmosphere. The captured CO₂ is stored in the oceans or in depleted gas and oil fields. The main drawback of the technology is the temporary nature of the storage. CO₂ stored in the ocean, for example, will inevitably return to the atmosphere in periods estimated from hundreds to thousands of years. Another drawback of ocean storage is the acidic characteristic of dissolved CO₂ and the effects of pH change in seawater on the local environment. Storage of large amounts of non-converted, concentrated CO₂ in oil and gas reservoirs needs continuous monitoring for an infinite time. A sudden release of CO₂ could be lethal, as demonstrated in the 1986 Lake Nyos disaster in Cameroon [114]. Carbon capture and storage will therefore not be commercially available until the cost and safety issues are satisfactorily addressed.

The conversion of CO₂ to useful fuels by physiochemical means not only reduces CO₂ in the atmosphere, but also eases our dependence on oil. The conventional approach involves the thermal hydrogenation of CO₂ into hydrocarbons under relatively high temperatures and pressures [115]. The major problem with the catalytic reduction of CO₂ is that huge amounts of H₂ are required as the reducing agent and in addition fossil fuels are consumed to provide the heat needed for the reaction to proceed.

An attractive alternative to thermal hydrogenation is photocatalytic reduction, especially if this utilizes sunlight. In this approach, CO₂ from industrial waste gases is converted to valuable fuels, such as methane and methanol. These products can be easily transported, stored and used in industry or, in the case of methanol, as a gasoline-additive for automobiles. Moreover, they can be transformed into other useful chemicals by using conventional technologies. This is a perfect solution to both the global warming and energy shortage problems. This section reviews the use of photocatalysts to produce valuable fuels from the virtually free resources of carbon dioxide, water and sunlight. The potential products are methane, methanol, or even longer chain hydrocarbons via a Fisher-Tropsch type condensation. From the viewpoints of both energy and environment, the conversion of CO₂ to fuels by solar energy is an ideal solution to the current global warming and energy crises.

For solar-driven catalytic conversion of CO₂ to fuels to be practical, highly efficient photocatalysts are required. Titania (TiO₂) has been considered the most appropriate candidate due to its powerful oxidizing nature, superior charge transport properties, and corrosion resistance. Earlier studies, however, could only achieve low CO₂ conversion rates in spite of using UV illumination for band gap excitation. Anpo et al. carried out a series of studies on Ti-zeolites and Ti-mesoporous materials [116, 117]. Powdered TiO₂ was also used by Adachi and co-workers as a photocatalyst for the reduction of CO₂ with H₂O. A total hydrocarbon (methane, ethene and ethane) generation rate of about 1.7 μL/(h g) was achieved under xenon lamp illumination when copper-loaded titania nanoparticles were dispersed in CO₂-pressurized water [118]. Tan et al., using titania pellets, obtained a maximum rate of about 0.25 μmol/h of methane from the irradiation of moist carbon dioxide by monochromatic ultraviolet light (253.7 nm wavelength) [119, 120]. Using UV irradiation of a hydrogen (90%), water, carbon dioxide combination, a rate of 4.1 μmol/(h g) was obtained by Lo and co-workers [121].

Recently, numerous studies on the preparation of solar-light-driven photocatalysts for hydrocarbon formation were reported. They can be classified into two categories: TiO₂-based photocatalysts and composite photocatalysts.

3.2.1 TiO₂-Based Photocatalysts

Metal doped TiO₂ catalyst sensitized with N3 dye was employed to photoreduce CO₂ with H₂O under concentrated natural sunlight to fuels in an optical-fiber photoreactor [122]. A methane production rate of 0.617 μmol/(g h) was achieved on N3-dye-Cu(0.5 wt%)-e(0.5 wt%)/TiO₂ coated onto optical fibers under an average solar light intensity of 20 mW/cm². The N3 dye substantially improved the photoactivity of Cu(0.5 wt%)-Fe(0.5 wt%)/TiO₂ catalyst toward methane production under concentrated natural sunlight due to its full visible-light adsorption. The photocatalyst was stable up to 6 h.

Recently, N-doped TiO₂ nanotubes with copper and platinum nanoparticles loaded onto the surfaces were developed to realize efficient solar conversion of carbon dioxide and water vapor to methane and other hydrocarbons [123]. The experiments were conducted in outdoor sunlight at University Park, PA (shown in Fig. 29). Using outdoor global AM 1.5 sunlight, 100 mW/cm², a hydrocarbon production rate of 111 ppm cm⁻² h⁻¹, or ~160 μL/(g h), was obtained when the nanotube array samples were loaded with both 52% Cu and 48% Pt nanoparticles. The authors pointed out that the efficiency of the catalyst was still quite low, but were optimistic that further work could improve it.

Fig. 29

a Digital photograph of the reaction chambers kept under natural sunlight for photocatalytic CO₂ conversion. b Spectral irradiance recorded from 12:39 p.m. to 3:52 p.m. for an experiment conducted on September 1, 2008 at University Park, PA. Reprinted with permission from Ref. [123]. Copyright 2009 American Chemical Society

3.2.2 Composite Photocatalysts

Carbon dioxide can be reduced with water to organic compounds over a hybrid catalyst under concentrated sunlight [124]. The catalyst was Pt-loaded K₂Ti₆O₁₃ coupled with an Fe-based catalyst supported on a dealuminated Y-type zeolite (Fe-Cu-K/DAY) [124]. The Pt/K₂Ti₆O₁₃ catalyst decomposed water to produce H₂ and the Fe-Cu-K/DAY catalyst reduced CO₂, with resulting organic compounds of CH₄, HCOOH, HCHO, CH₃OH, and C₂H₅OH. The Pt/K₂Ti₆O₁₃ catalysts can be combined with another CO₂ hydrogenation catalyst of Cu/ZnO [125]. The generation of CH₃OH over this composite photocatalyst under concentrated sunlight means successful photocatalytic conversion carbon dioxide to fuels. These studies suggest that sunlight-driven photocatalytic processes have potential for organic compound evolution from CO₂ and water.

Recently, a NiO/InTaO₄ photocatalyst with a band gap of 2.6 eV was developed by Chen’s Group [126]. The NiO cocatalyst was loaded by incipient-wetness impregnation with an aqueous solution of Ni(NO₃)₂. The product was calcined at 350°C for 1 h in air, and then pretreated by H₂ reduction at 500°C for 2 h and subsequent O₂ oxidation at 200°C for 1 h. This catalyst was able to reduce CO₂ to methanol under visible-light illumination. A 1.0 wt% NiO-InTaO₄ photocatalyst in 0.2 M KHCO₃ gave the highest activity (1.394 μmol/(h g)) (shown in Fig. 30). The reduction–oxidation pretreatment had a positive effect on the activity of the catalyst.

Fig. 30

Methanol yield on NiO-InTaO₄ in 0.2 M KHCO₃ aqueous solution under visible-light irradiation. Reprinted from Pan and Chen [126], copyright 2007 with permission from Elsevier

In summary, the ultimate goal is to design advanced catalysts with high photon efficiencies. For the most promising catalysts the rate limiting steps in the conversions of water and CO₂ need to be determined. Relationships between the photocatalytic efficiency and the characteristics of the catalyst such as morphology, pore structure, surface area, surface electronic states and band gap must be investigated.

4 Summary and Future Prospects

Various photocatalytic semiconductor nanomaterials with great potential for energy and environmental applications have been prepared. Despite the tremendous research efforts on the synthesis and modification of photocatalysts, many problems still exist. This is particularly true with regard to low photocatalytic efficiencies and the lack of understanding of the intrinsic mechanisms for the systems. Thus, the design of novel photocatalytic materials with higher efficiency is a perennial subject of interest in the field of photocatalysis. As more and more attention is paid to energy issues, the development of practical systems for H₂ evolution from water and the photo-reduction of carbon dioxide to fuels cannot be overemphasized. Advanced photocatalytic nanomaterials may be the key to a sustainable future.