The field of heterogeneous photocatalysis has expanded rapidly within the last four decades, having undergone various developments especially in relation to energy and the environment. It can be defined as the acceleration of photoreaction in the presence of a catalyst. The two most significant applications of photocatalysis have been in solar water splitting and the purification of air and water containing low concentrations of pollutants. The multidisciplinary nature of the field has also increased significantly and includes semiconductor physics, surface sciences, photo and physical chemistry, materials science and chemical engineering [1].

Heterogeneous photocatalysis can be described as the acceleration of photoreaction in the presence of a catalyst. In the contexts of history and research, interest in heterogeneous photocatalysis can be traced back to many decades when Fujishima and Honda discovered in 1972 the photochemical splitting of water into hydrogen and oxygen in the presence of TiO₂ From this time, extensive research, much of it published, has been carried out to produce hydrogen from water in oxidation reduction reactions using a variety of semiconductor catalyst materials.

In recent years, interest in photocatalysis has focused on the use of semiconductor materials as photocatalysts for the removal of ambient concentrations of organic and inorganic species from aqueous or gas phase systems in environmental clean-up, drinking water treatment, industrial and health applications. This is because of the ability of TiO₂ to oxidize organic and inorganic substrates in air and water through redox processes In this context, TiO₂ has not only emerged as one of the most fascinating materials in both homogeneous and heterogeneous catalysis, but has also succeeded in engaging the attention of physical chemists, physicists, material scientists and engineers in exploring distinctive semiconducting and catalytic properties.

Inertness to chemical environment and long-term photostability has made TiO₂ an important material in many practical applications, and, in commercial products ranging from drugs to foods, cosmetics to catalysts, paints to pharmaceuticals, and sunscreens to solar cells in which TiO₂ is used as a desiccant, brightener, or reactive mediator [2]. The U.S. Food and Drug Administration permits up to 1% TiO₂ as an inactive ingredient in food products. While there are no known health effects associated with the use of TiO₂, a recent study found that 3–6 year old children are the most affected group of people that consume TiO₂ particles from food products. Many new properties of TiO₂ have been explored during the past few years [3]. It should be stated that regulatory frame work for the use of TiO₂ in food products are yet to be firmly established in many countries, especially developing nations. The catalyst itself is unchanged during the process and no consumable chemicals are required. This results in considerable savings and simpler operation of the equipment involved.

Large bandgap semiconductors like TiO₂ are commonly investigated in the rutile (bandgap 3.0 eV) and anatase (bandgap 3.2 eV) phases and TiO₂ response to UV light has not only led to photocatalysis research [4,5,6] but also to an extensive investigation of TiO₂ superhydrophilicity its use in environmental remediation and solar fuel production. Bandgap excitation of TiO₂ causes charge separation followed by scavenging of electrons and holes by surface adsorbed species, Figure 1:

Visible photocatalysis can thus be utilized by doping TiO₂ with a sensitizing dye or short bandgap semiconductors. By controlling the surface treatment and medium conditions, it is possible to fine-tune the photocatalytic properties of TiO₂ to desired applications including the effective mineralization of contaminants in air, as self-cleaning glass and ceramic tiles. On the other hand, the practical use of TiO₂ for remediation of chemical contaminants in wastewater is a challenge because of catalyst poisoning or deactivation. In addition, there remain some challenges with respect to extending the photocatalytic response of TiO₂ in the visible region [6], an area that merits fundamental research.

In spite of the extensive efforts to dope TiO₂ with C, N, S and transition metal ions such as Ag, Au, Fe, Ru, photocatalytic activity of TiO₂ in the visible region has remained quite low, an issue that has been addressed in recent articles highlighting issues and challenges associated with the application of photocatalysis. Generally, two or more phases are involved in a photocatalytic reaction—a light source and a semiconductor material are used to initiate the photoreaction while the catalyst system can simultaneously carry out oxidation and reduction reactions using long wavelength, UV light as well as sunlight. As a method for contaminant control in water and air, Figure 2, heterogeneous photocatalysis using semiconductors such as titanium dioxide is more efficient than conventional methods. This is because as the photocatalytic process gradually breaks down the contaminant molecules, no residue of the original material remains and therefore no sludge requiring disposal to landfill is produced. The catalyst itself is unchanged during the process and no consumable chemicals are required. This results in considerable savings and a simpler operation of the equipment involved. Additionally, because the contaminant is attracted strongly to the surface of the catalyst, the process will continue to work at very low concentrations. Taken together, these advantages mean that the process results in considerable savings in water production cost and keeping the environment clean.

In order to activate the degradation process, pure TiO₂ requires photo-excitation with light at wavelengths exceeding the band gap of the active anatase phase of 3.2 eV, that is, wavelengths of <387 nm. Indeed, commercially available photocatalytic water treatment plant using artificial UV-light as the energy source is available and can be considered as a developed market.

However, UV in natural sunlight represents only 5%–8% of the solar spectrum at sea level and this provides a limitation and the requirement of artificial illumination of the catalysts in order to achieve degradation of the organic material in times that are practical for water treatment processing. This greatly restricts the application of the technology to areas where abundant electricity supplies are available. The need for more stable, efficient catalysts which can be activated by natural sunlight is now the goal of researchers. In order to achieve significant improvements in catalytic activity using visible light, precise control of the stoichiometry of the catalytic metal oxides and mixed metal oxides, addition of dopants, particle shape and pore topology are all critical factors for catalyst developers.

Advances in the manipulation of matter at the nanoscale are leading to potential solutions for the depollution of water in remote locations. In recent years, the development of nanoscale metal oxides has greatly increased the catalytic activity by virtue of the high specific surface area available for reaction of the smaller particles. In addition, the strategy to alter the band gap of the catalyst is an important approach as this determines the portion of the solar spectrum the catalyst absorbs and, consequently, the amount of energy that is converted to reactive species. Photocatalysts that have high activity using wavelengths of light in the visible spectrum (380 nm < λ < 500 nm) have been demonstrated where potentially greater amounts of energy is available (solar peak energy is around 460 nm). This has been achieved by altering the band gap of, for example, TiO₂. Doping of TiO₂ with transition metal ions (for example, V, Cr, Mn, Fe and Ni) as well as with Ag, Au and Ru, have been demonstrated to red-shift the TiO₂ absorption band from the UV into the visible region, resulting in a great increase in the efficiency of solar-light photocatalysis.

Many industrial waste streams are not suitable for biological processing due to their inherent toxicity, but their treatment by traditional non-catalytic chemical processes or by incineration is energy intensive [7]. On the other hand, TiO₂ photocatalysts has been shown to decompose organic contaminants in water efficiently because a strong oxidizing ability is generated when the TiO₂ is irradiated by appropriate band gap illumination [8]. Increased attention is being paid to heterogeneous solution phase catalysis in order to develop an environment-friendly technology to purify polluted air and water without electricity or other energy consuming sources.

Heterogeneous photocatalytic reactions are carried out either in a slurry-type reactor where the catalyst particles are suspended in the contaminated water or in an immobilized-type reactor where the catalyst particles are immobilized onto the surface of various inert substrates of various types and configurations [9,10,11] and the process can be explained on the basis of the following mechanism.

A photocatalyst harnesses UV radiation from sunlight or artificial light and uses the energy to break down different substances including organic materials, organic acids, estrogens, pesticides, dyes, crude oil, microbes (including viruses) and chlorine resistant organisms, inorganic molecules such as nitrous oxides (NO_x) and, in combination with precipitation or filtration, can also remove metals such as mercury [13,14,15]. Due to this universal applicability, photocatalysis with nanoparticles as catalysts is used to reduce air pollution, in building materials, for self-cleaning surfaces in addition to water purification. Titanium dioxide (TiO₂) is the most common photocatalyst and comparably little research has been conducted on zinc oxide, ZnO, which could be a viable alternative for some applications. To avoid free nanoparticles in water, TiO₂ nanoparticles are usually immobilized on a substrate [16] or integrated into thin-films and other materials. For the activation of TiO₂, UV irradiation from sunlight or artificial light is needed, with UVB more efficient than UVA. To allow activation by visible light, TiO₂ can be modified with a second semiconductor, dyes, nitrogen, carbon or sulphur. For example, TiO₂ doped with nitrogen demonstrated superior photocatalytic activities compared to commonly used unmodified TiO₂ nanoparticles in both chemical compound degradation and bactericidal reactions [17].

With respect to a semiconductor oxide such as TiO₂, photocatalytic reactions are initiated by the absorption of illumination with energy equal to or greater than the band gap of the semiconductor. This produces electron-hole (e⁻/h⁺) pairs as in Equation (1), Figure 3: (1) where cb is the conduction band and vb is the valence band. Thus, as a result of irradiation, the TiO₂ particle can behave either as an electron donor or acceptor for molecules in contact with the semiconductor. The electron and hole can recombine, releasing the absorbed light energy as heat, with no chemical reaction taking place. On the other hand, they can participate in redox reactions with adsorbed species as the valence band hole is strongly oxidizing while the conduction band electron is strongly reducing. On the semiconductor surface, the excited electron and the hole can participate in redox reactions with water, hydroxide ion (OH⁻), organic compounds or oxygen leading to mineralization of the pollutant.

In fact, research shows that the charges can react directly with adsorbed pollutants, but reactions with water are predominant since the water molecules are more abundant than contaminant molecules. Consequently, oxidation of water or OH⁻ by the hole produces the hydroxyl radical (·OH), a powerful oxidant. For the purpose of comparison, it is important to compare the oxidation potential of hydroxyl radical (·OH) which is 2.8 V relative to the normal hydrogen electrode. Apart from electron and their holes, other substances that may be used for water disinfection include ozone (2.07 V), H₂O₂ (1.78 V), HOCl (1.49 V) and chlorine (1.36 V). OH radicals are able to rapidly attack pollutants on the semiconductor surface and, as such, are the most important radicals formed in TiO₂ photocatalysis. An important reaction of the conduction band electron is reduction of adsorbed oxygen to oxygen radicals and this prevents the electron from recombining with the hole and results in an accumulation of oxygen radicals that can also participate in degrading contaminants in solution [11,18].

Published work indicate that heterogeneous photocatalytic process involves at least five separate reaction steps [19] and include (1) diffusion of reactants to the surface of semiconductor, (2) adsorption of reactants onto the surface of semiconductor, (3) reaction on the surface of semiconductor, (4) desorption of products from the surface of the semiconductor and (5) diffusion of products from the surface of the semiconductor. There are two routes through which OH radicals can be formed—the reaction of the valence-band holes with either adsorbed H₂O or with the surface OH groups on the TiO₂ particle. These processes have been summarized using appropriate equations [20].

Current and past research in photocatalytic materials has investigated several photocatalysts and their properties. Ideally, a photocatalyst should possess the following properties: photoactivity, biological and chemical inertness, stability toward photo-corrosion, suitability for visible or near UV light energy harnessing, low cost and lack of toxicity [21]. TiO₂ as a semiconductor photocatalyst, has excellent pigmentary properties, high ultraviolet absorption and high stability which allow it to be used in different applications such as electro-ceramics, glass and in the photocatalytic degradation of chemicals in water and air. It has been used in the form of a suspension, or a thin film in water treatment [22,23] and has different crystalline forms, the most common forms being anatase and rutile while the third, brookite, is uncommon, unstable and hence does not feature in discussion of catalyst materials.

Anatase TiO₂, the most stable form of TiO₂ [24], can be converted to rutile by heating to temperatures above 700 °C. In photocatalytic applications, research has shown that anatase is more efficient than rutile because of its more open structure compared with rutile. Degussa P-25 is the commercially available form of TiO₂ and consists of two forms of approximating 25% rutile and 75% anatase, and has been used in many studies of photocatalytic degradation because of its chemical stability, ready availability, reproducibility and activity as a catalyst for oxidation processes [13,25]. However, studies are being carried out to develop existing or prepare new materials which can be used under solar energy and hence shortening the time needed for degradation.

This development includes increasing the surface area, the active sites, the absorption of photon and reducing the band gap energies of the semiconductor. For industrial applications, high activity, resistance to poisoning and stability for prolonged use at elevated temperatures, mechanical stability, and resistance to attrition, physical and chemical stability in various conditions, are required characteristics of catalysts [21]. In addition, in the degradation of organic compounds, the redox potential of the H₂O/·OH electrode, OH⁻ → ·OH + e⁻; E₀ = −2.8 V, must lie within the band gap of the photocatalyst as illustrated in Figure 1. Nowadays, many semiconductors are used as photocatalysts because of their narrow band gap between the valence and conduction bands and they need to absorb energy equal to or more than this energy gap. This involves the movement of electrons from e⁻/h⁺ or negatively charged electrons to positively charged hole pairs [26].

Many semiconductors have enough band-gap energies for the effective catalysis of many chemical reactions and this includes materials such as TiO₂, WO₃ and ZnO. Though it is well known that metal oxides are usually less active catalysts than noble metals in the majority of applications, metal oxides are more suitable since they are more resistant to poisoning and deactivation. In addition, combining two or more metal oxide catalysts could improve or enhance catalytic activity. TiO₂ is the most frequently used photocatalyst because of its photostability and low cost, combined with its biological and chemical inertness and resistant to photo and chemical corrosion. On the other hand, binary metal sulfide semiconductors such as CdS and PbS are regarded as insufficiently stable for catalysis and are toxic. ZnO is also unstable in illuminated aqueous solutions while WO₃ has been investigated as a potential photocatalyst, but it is generally less active catalytically than TiO_2. However, these can be combined with other semiconductors including TiO₂ to achieve greater photocatalytic efficiency or stability [27,28,29,30].

There are many semiconductor support materials that have been investigated. Usually, semiconductor supports are classified by their chemical nature and these can be organic or inorganic supports. They play an important role in immobilizing active catalyst, increase the surface area of catalytic material, decrease sintering and improve hydrophobicity, thermal, hydrolytic and chemical stability of the catalytic material. Examples include glass, carbon fibers and woven fiber cloths and these have been studied as support materials in groundwater denitrification and in the photocatalytic oxidation of water pollutants as well as in other applications. When fibrous supports are applied, loss of pressure is low and pore diffusion resistance is significantly lower than with pellet shaped catalysts. Glass possess an advantage as a catalyst support because of its transparency to UV light in photocatalytic applications [30].

Studies by Teoh, Kamat and others [1,9,12,31,32,33], show that interest in photocatalysis has also centered on ZnO, which has similar energy characteristic to TiO₂. They report that ZnO however suffers from photo-corrosion problems upon excitation in solution. Hematite (α-Fe₂O₃) on the other hand, which is low cost, abundant and has narrow bandgap for harnessing solar energy, suffers from rapid charge recombination and a short charge carrier diffusion length. As a result of these draw backs, renewed interest in this material has focused on its modification with cationic dopants such as Cr and Mo to improve its charge transport properties [34] or doping with Si to reduce the charge diffusion path length. Other reviews have also talked about Tungsten trioxide WO₃ as another narrow bandgap material, 2.7 eV, and, as such, it has received renewed interest. Like hermatite, Tungsten trioxide, has the disadvantage of a low electron conduction band, although studies claim that coupling with Pt cocatalyst has been observed to be useful in promoting alternative multiple-electron reduction processes with lower reduction potentials. This has increased the use of WO₃ as one of the very few highly visible-light-active single-phase oxide photocatalysts [29,30,31,34,35,36,37,38,39,40,41].

In many cases, the superiority of other photocatalyst materials over P25 is based on a selection of test reactions or only a single organic substrate and this has its problems as all of the important factors cannot be validated in a single experiment or reaction. An example is the pairing of test substrate degradation mechanistic Langmuir-Hinshelwood kinetic model, direct charge transfer, hydroxyl radicals, and superoxide radicals with many physicochemical characteristics of the photocatalyst such as the crystallinity, hydroxyl group density, size, specific surface area, surface defects, surface energetic and aggregation [12]. Even among flame-made TiO₂ of similar crystallinity, about 80% anatase and 20% rutile, and exact anatase-rutile content, differences in the photocatalytic activity of P25 were observed [42] due mainly to the class of organic substrate [43] used and the materials intrinsic efficiencies for ·OH generation and direct charge transfer, which governed the dominant degradation mechanism for each of the substrate functional groups. In a broader context, Ryu and Choi [44] compared the photocatalytic performance of commercial variants of TiO₂ and observed that one photocatalyst cannot meet all photocatalytic requirements, a fact borne out by several studies [45].

With respect to the synergy between anatase and rutile, the dominant phases of TiO₂ studies [46] have suggested that polymorphic characteristics of these materials was one of the reasons for the high photocatalytic efficiency of P25 and that the heterojunction of anatase-rutile promoted efficient charge separation as evidenced by studies indicating discrete trapping of electrons and holes across the different crystal phases. Any synergistic effect, however, is difficult to measure because of the differences in the physicochemical characteristics of pure anatase and rutile (surface area, crystallite size, defect content, surface hydroxylation, etc.), compared with the individual components in the mixture. Ohtani and other studies [46,47,48] on the other hand, isolated the individual components of P25 by selective acid dissolution and assessed the single components for a range of photocatalytic reactions. The studies concluded, at least for P25, a less probable synergistic effect beyond the photocatalytic activity of the individual components. By carefully synthesizing highly crystalline and physically similar anatase and rutile nanoparticles, Kho [49,50] and other studies demonstrated higher rates of hydrogen evolution from aqueous methanol across a wide range of mixed anatase-rutile compositions. The effect, however, was only observed when a strong inter-particle contact existed between the anatase and rutile nanocrystals, supporting the view of an efficient charge separation across heterojunctions in these materials [1].

Doping of TiO₂ with fine noble metals such as silver (Ag), gold (Au) and palladium (Pd) has been carried out extensively as a means of enhancing charge separation. Doping has been reported [51] to establish a barrier through equilibration between the photocatalyst and metal deposits and is governed by the difference in work function of the deposits and the electron conduction band of the photocatalyst and does not influence the mechanism of the specific reaction [1]. Previous studies noted differences in the enhancement by Pt and Ag as well as other noble metals, on TiO₂ over a range of organic classes arising from the combination of substrate or functional group specificity and the dark catalytic oxidation effects [12,32,33,51,52]. The latter included the catalytic oxidation of formic acid, oxalic acid, methanol, and the opening of aromatic rings, with each reaction induced by the catalytic properties of the metals. Studies have shown that when a photocatalyst is illuminated, a type of synergistic effect between the catalytic and photocatalytic processes is possible and in fact studies have shown these effects do exist. [33,52,53].

With respect to new materials for use in photocatalytic work, Teoh et al. [1] reported in a recent article that a graphene oxide material was demonstrated to act as a mediator for electron transport and for visible-light water splitting in photocatalysis. It has to be stated that prior to this development, electron mediators were largely based on aqueous redox couples such as Fe²⁺/Fe³⁺ and CO₂⁺/CO₃⁺ as well as Au and Ag. A recent study by Wei and coworkers [54] using hybrid Fe₂O₃-Pd nanoparticle photocatalyst grown by epitaxial growth of Fe₂O₃ on Pd nanoparticles, showed improved photocatalytic efficiency, although this was limited by the presence of organic stabilizers which are difficult to remove. Other materials considered in recent work include ZnS, ZnO, CdSe, Fe₂O₃ and InP [55]. Some of these materials were reported to show some photocatalytic promise, but with uncertainty over their efficiencies and spectral characteristics [56]. Fe₂O₃ is particularly interesting because of its stability against photo/chemical corrosion at neutral or basic pH and has band gap energy of about 2.0 to 2.2 eV corresponding to the absorption of 564 to 620 nm light [54].

Investigations of photcatalytic degradation processes including the mechanisms have usually been carried out by using model compounds. This can strongly influence the photocatalytic performance of a material according to Teoh and co-workers [1] because of the effect of the structure of such model compounds and other factors on the degradation process. The study show that, for example, while monitoring the decolourization of dyes in solution using spectroscopic techniques is simple and convenient, using a dye as the model compound can yield unreliable results as this may not represent the intrinsic photocatalytic activity of the semiconductor, especially for visible-light-activated materials for which dyes are a good example and these include methyl orange, congo red, alizarin yellow, etc.

Cationic and anionic azo-dyes can undergo photolysis or sensitize the semiconductor and thereby inject an electron into its conduction band [1,12]. In addition, because of their complex structure, dyes have complex photodegradation mechanistic pathways thus making determining accurate photocatalytic efficiencies a very difficult exercise [56]. In addition, studies [1] have indicated that dyes do not necessarily require carbon oxidation for decolorization to occur. Although photocatalysis can play a significant role in treating dye-polluted systems [57], the choice of a model pollutant substrate should be undertaken with due consideration when assessing a photocatalytic activity under visible illumination. This reviewer has carried out photocatalytic work using azo-dyes as model substrates and would agree with the conclusion reached regarding the use of model compounds activated by visible light to validate photocatalytic degradation mechanisms and efficiency.

Research into photocatalyst doping has spanned several decades. Usually doping involves the use of metals or non-metals and is designed to extend the photocatalytic activity of a semiconductor lower energy excitation. Technically, doping is the introduction of foreign elements into the parent photocatalyst without giving rise to a new crystallographic forms, phases or structures and the aims are [1] to enhance the net separation of photogenerated charges and thereby efficiently harness the wide visible-light component of about 43% in the solar spectrum as opposed to the narrow ultraviolet component of 5%. It is thus an area of increasing research activity in photocatalysis.

Although discovered as a modification technique in the early 1980s, doping has become a standard and routine method in the design of water-splitting photocatalysts [12]. Modifying pure photocatalyst materials with metal ions, especially d block metal ions such as Fe³⁺ and Cr³⁺ results in the insertion of impurity energy levels between the parent conduction and valence bands. In this case, the inserted energy levels provide sub-bandgap irradiation from which electrons can be excited from dopant d-band to conduction band or from valance band to dopant d-band by lower energy photons than are required by the pure photocatalysts. However, there are a number of issues associated with metal ion doping, related to the efficiency of subsequent photocatalytic processes. Whereas the presence of metal dopants was found in some cases to enhance charge separation as well as interfacial charge transfer [58,59] in many other cases, the metal dopants actually resulted in rapid charge recombination through their switchable redox states, for example, Fe³⁺ ↔ Fe²⁺ and thereby reducing the electron diffusion length and lifetime [1,9,12,60]. This is one of the reasons for the reduced quantum efficiency in many photo response-extended doped photocatalysts.

Several studies have indicated that doping creates structural defects that could be sources for charge recombination and in this sense are potentially negative in their effects. The only exception being doping using high-energy RF and magnetron sputtering, which provide the uncommon existence of tetravalent dopants such as Fe⁴⁺ and Cr⁴⁺ that match the valency of Ti⁴⁺ in TiO₂ [61]. Co-doping with a conjugate metal cation pair such as Rh³⁺/Sb⁵⁺ can preserve charge equality of the doped photocatalysts and result in improved and extended photocatalytic ability. However, this may not necessarily prevent structural defects arising from the differences in cationic radii between dopants and the host photocatalyst. In a recent study, Serpone and co-workers [62] provided a substantiated alternative argument that the extended photo-response instead of arising from d-band insertion, may actually originate from the color centers and oxygen vacancies that arise from the doping [63,64].

Doping involving the use of non metals has also been investigated and studies have reported that the introduction of non-metal such as N, C and S impurity energy levels above the parent photocatalyst valence band has been a popular doping technique and although doping wide bandgap photocatalysts such as TiO₂ and ZrO₂ with F [64,65,66,67,68,69] does not result in coloration, their enhanced photocatalytic efficiencies, even under sub-bandgap excitation, likely results from enhanced stoichiometric defects and this is useful in photocatalytic processes. Higher photocatalytic activity, stability [28] and nontoxicity of dopant ions are among the advantages of nonmetal over metal dopants and, in general, TiO₂ and several other materials have been studied including the effect of impurity levels in the dopant on photocatalytic efficiency. In addition, results obtained using TiO₂ and oxides such as WO₃, Ta₂O₅ and Ba₅Ta₄O₁₅ have been published [69,70,71]. By studying nitrogen doped TiO₂ using real-time fluorescence imaging, Tachikawa and coworkers [72] confirmed the absence of ·OH under sub-bandgap irradiation but retained the formation of O₂^•–, which, in turn, became the dominant radicals. Whereas full band gap excitation from the valance to the conduction band may result in the original oxidative potential of the parent photocatalysts, the quantum efficiency is hampered because of the pronounced charge recombination at dopant-induced defect sites [36].

Similar to metal-doping strategies, anionic co-doping, for example involving N/F-TiO₂ or N/F- and C/N-ZnWO₄ [73,74,75] has been used to lower charge defects, but in many cases, the quantum efficiencies remain unmatched compared with the excited photocatalysts. Many studies, for example [76,77], have also reported that new conduction and valence bands are produced in a semiconductor when significant levels of foreign elements are introduced to the photocatalyst. Research using this band gap engineering approach has been carried out successfully and resulted in the emergence of new photocatalytic materials for water-splitting purposes including Bi³⁺, Ti⁴⁺ and W⁶⁺ [78,79]. From the various studies carried out on doping and the effect of doping, it is obvious that the exploration of new photocatalytic doping materials will continue to be driven mainly by efforts in water splitting reactions although these materials will find relevance in other photocatalytic applications such as PbBi₂Nb₂O₉, for the treatment or remediation of aqueous pollutants, CaBi₂O₄ for the oxidation of gaseous pollutants and ZnGa₂O₄ for the reduction of carbon dioxide and similar materials [77,80,81].

Generally, for all catalyst materials, a high surface area is an advantage in terms of a greater concentration of active sites per square meter and this generally leads to higher reactivity. The smaller the particle size, the larger the surface area, and the higher the expected activity [81]. This can be explained in terms of an increase in the number of active sites per square meter as well as greater absorbance of the pollutant on the catalyst surface [80]. It has often been reported however that while the activity of the TiO₂ catalyst is affected by its external surface [82], photocatalytic activity is not a function of the surface area alone. For example, while an increase in the extent of heat treatment applied leads to a reduction in the surface area, the impact of this reaction on the photocatalytic activity is not as significant as that of the phase transformation. The increase in crystallite size and changes in the microstructure with a decrease in surface area greatly influence the photocatalytic property of the catalyst [83,84].

One of the ways to increase the efficiency of photocatalytic reaction is to couple semiconductor nanoparticles to noble metal co-catalysts to enhance the quantum yield of the electron transfer processes through improvement in charge separation in the semiconductor, discharging photogenerated electrons across the interface and providing a redox pathway with low over-potential. The interfacial charge transfer processes are influenced by the presence of a metal co-catalyst or surface bound molecular relays. If metal nanoparticles are coupled to the semiconductor nanoparticle, then they readily accept and shuttle electrons to an acceptor molecule at the interface easily. The discharge capacity of metal nanoparticles plays an important role in dictating Fermi level equilibration between semiconductors and metal nanoparticles [1,9,12].

It follows from these studies that in order to enhance the photoactivity of titanium oxide particles, interfacial charge-transfer reactions need to be enhanced. Improved charge separation and inhibition of charge carrier recombination is essential in improving the overall quantum efficiency for interfacial charge transfer. This can be achieved by modifying the properties of the particles by selective surface treatment and several approaches have been taken to achieve this including surface modification of the semiconductor particles with redox couples or noble metals and the coupling of two semiconductor particles with different electronic energy levels. To this end, catalytic thin films, foams and hybrid nanoparticles of TiO₂ as well as gels have been made and used in various catalytic applications [85,86,87,88] as described in this review.

In thin film deposition, a thin layer of material is prepared on a suitable substrate or a previously deposited layer. “Thin” is a relative term, since a film may look operationally thin or thick compared to the wavelength of the light used. Compared to conventional thin film forming processes such as evaporation, or sputtering, sol-gel processing makes it possible to control the microstructure of the deposited film including the pore volume, pore size, and surface area [89] and uses mainly alkoxide, organic/ inorganic salts, metal oxides and other salts in a liquid phase chemical reaction. The method provides excellent chemical homogeneity and the possibility of deriving unique metastable structures at low reaction temperatures [90]. The liquid phase reaction enhances the uniformity of the product produced and the process can be operated under ambient conditions. The reactant concentration, temperature and type of solvent will affect the quality of the final product.

In the production of TiO₂ particles, calcinations at high temperatures are necessary to change the crystalline structure of TiO₂ that possesses photo catalytic activity and the specific surface area, particle size, and other properties of TiO₂ can be altered after calcinations at different temperatures. Thus thin films formed by dipping or spinning, use little raw materials and may be processed quickly without cracking and thus overcoming most of the disadvantages of sol-gel processing. In addition, large substrates may be accommodated and it is possible to uniformly coat both sides of planar and axially symmetric substrates such as pipes, tubes, rods and fibers which are not easily handled by more conventional coating processes.

There are two main ways to synthesize gels at room temperature: spin coating and dip coating. Spin coating is used for applications where relatively flat substrates or objects are coated with thin layers of material. On the other hand, the dip-coating process, involves immersion, start-up, deposition, evaporation and drainage. The substrate is slowly dipped into and withdrawn from a tank containing the sol, with a uniform velocity, in order to obtain a uniform coating. The gel samples are dried at 100 °C and then fired at 450 °C to remove organics from the film pyrotically. Once the sample has reached the desired thickness through multiple coatings, it is crystallized by heating. The maximum thickness of crack free coatings that can be achieved is 0.5 microns. Sol-gel and “doctor blade” TiO₂ films vary in thickness from 0.2 microns to 0.8 microns.

A foam is a substance that is formed by trapping pockets of gas in a liquid or solid. Particle sizes, porosity differ for various materials depending on the synthetic routes chosen, we have synthesized TiO₂ foams (reported in nano-letters) of 100–500 nm in thickness. TiO₂ films similarly have various thicknesses ranging from a few microns to hundreds or even more depending on the synthetic conditions. We have produced sol-gel and “doctor blade” TiO₂ films varying in thickness from 0.2 microns to 0.8 microns. Generally, the porosity of the foam is larger than that of films (depending on the film and process conditions). Foams usually offer more diffusion pathways than films especially for solution phase degradation. Films on the other hand are usually easier to integrate into structured photocatalytic reactors and light absorption is better than in films. Foams usually get “clogged” at some stage during the photocatalytic process. The synthetic route for TiO₂ foam as reported in the literature [91]. The thickness of the TiO₂ foam produced by this process was 100 nm and it should be pointed out that particle size of 500 nm is possible depending on the synthetic process. The TiO₂ foam offers extended porosity which is important for “light harvesting”, Figure 4.

The TiO₂ foams have been used as catalyst materials in the degradation of low concentrations of benzene, toluene and xylene (BTX) in contaminated air. The degradation of anionic and cationic dyes in industrial effluents [8,82,86] as shown in Figure 5, an example given with methyl orange has been undertaken.

A metal can affect the surface properties of a semiconductor by generating a “barrier” which acts as an electron trap of the metal in contact with the semiconductor surface. Similarly, doping of transition metal ions to semiconductors improves the trapping of electron and inhibits e⁻/h⁺ recombination because the relatively low photoactivity of TiO₂ is believed to be due to the fast recombination of photogenerated electrons and holes. In recent years, researchers have focused on the combination of different kinds of metallic oxide particles. Oliviero et al. [89] showed that Ru/CeO₂ particle catalysts have very different behavior in the catalytic wet air oxidation of acrylic, succinic and acetic acids depending on the support morphology and the metal/support contact. The use of metal nanoparticles and hybrids has also been researched [92] and the synthetic route for TiO₂/RuO₂/SiO₂ nanoparticles is reported elsewhere [92]. These materials have been used in the degradation of low concentrations of anionic and cationic azo-dyes in industrial effluents, Figure 6.

Recent work by Wei et al. [54] using hybrid Fe₂O₃/Pd nanoparticulate photocatalyst grown by epitaxial growth of Fe₂O₃ on Pd nanoparticles showed improved photocatalytic efficiency, although this was limited by the presence of organic stabilizers which are difficult to remove. Other materials considered in recent work include ZnS, ZnO, CdSe, Fe₂O₃ and InP [93]. Some of these materials were reported to show some photocatalytic promise, but with uncertainty over their efficiencies and spectral characteristics. Fe₂O₃ is particularly interesting because of its stability against photo and chemical corrosion at neutral to basic pH and has a band gap energy of about 2.0 to 2.2 eV corresponding to the absorption of 564 to 620 nm light.

Significant advances have been made in reactor design for various photocatalytic reactions. Published literature contains several types and designs of photocatalytic reactors [1,12,92,93,94] .Apart from the conventional reactor design parameters, the major challenge in the design of a photocatalytic reactor is the efficient illumination of the catalyst. For a high activity in the reactor, a large area has to be illuminated. Therefore, in immobilized systems the thickness of the supported catalyst layer should be small enough to enable the light to reach all the catalyst. Therefore, one of the most important parameters in design consideration for photocatalytic reactors is light intensity and how the reactor is to be designed to accommodate various light sources. Types of reactors developed include fixed bed and slurry bed. With regard to design considerations, various models have been reported to deal with irradiation of the photocatalyst and how the photocatalyst affects the irradiation—the degree of absorption, scattering, the nature and type of particles, surface of the photocatalyst, etc. Light intensity is an important reactor design factor when considering light sources because the major requirement of any photocatalytic process is that the catalyst, pollutant and source of illumination must be in close proximity. This partly because studies have shown that the rate of photocatalytic degradation has a nonlinear dependence on light intensity and is proportional to I^0.5 at high intensity and I^1.0 at low intensity [93]. It has also been shown that an excess of light promotes a faster electron-hole recombination and the low activation energy of photocatalytic reactions (5–20 kJ/mol), compared with ordinary thermal reactions. In slurry type systems, catalyst loading is an important design parameter for the effective use of reactor space and photocatalyst and a wide range (0.15 to 8 g/L) of optimal catalyst loading in aqueous suspensions has been reported for different photocatalyst systems and reactor configurations [1,12,91,92,93,94,95,96,97,98,99], Figure 7:

Another example of a design that has been used in the degradation of organic in polluted air is shown in Figure 8. This reactor is filled with silica beads coated with titania paste and is externally irradiated by UV irradiation and used in the photocatalytic degradation of organic pollutants in air or water.

The design and selection of a photocatalytic reactor depends on the experimental conditions and the application and two main types have been extensively researched-reactors that use a catalyst as suspension form and reactors that use a thin film catalyst either coated to the reactor wall or coated in spherical glass beads. Both reactors can be designed as an immersion wall or flat wall. Immersion well photoreactors are normally used at laboratory scale for evaluation purposes and are operated in batch or continuous mode where oxidant flow and temperature can be easily monitored. The light source can be either single or multiple with or without reflectors. Catalyst in these reactors can be used as thin film or as suspension form, with suspension system more preferable as they are more efficient than thin film reactors due to the large surface area in contact with the substrates. This allows more number of photons to hit the catalyst and large adsorption capacity results. Other reactors are flat wall and tubular photoreactors, which are simple and easy to design and generally use solar energy. In tubular reactors, reflectors are used to concentrate the light and enhance the photoreaction.

Selected applications of photocatalysis are given in Table 1, Table 2, Table 3. Heterogeneous photocatalysis has been demonstrated as a low cost and sustainable technology for the treatment of a host of pollutants in air and water including organics and heavy metals, etc., with Japan, USA, India and China as major users of this technology as partly demonstrated by the number of research publications in this area, Figure 9 [1]. Unlike reverse osmosis, nano and ultrafiltration, photocatalysis is a cheap and a potential “stand alone” technology for water treatment. As photocatalysis makes use of sunlight or UV radiation, the technology is inexpensive, environmentally friendly and can be applied worldwide. It requires minimal equipment, is highly deployable and appropriate for developing countries and remote sites with no access to electricity. Photocatalysis has also been used successfully in many developing nations to destroy pathogens [5,92] and algal blooms in fresh water supplies. Photodisinfection sensitized by TiO₂ has been used to degrade the green algae, treat humic substances [100] which act as substrates for bacterial growth, inhibit bacterial degradation of impurities in natural water, and aid the transport of metals in the environment and complex with Fe, Pb, Mn, making it harder to remove them. Specific examples of applications are as follows.

The widespread application of nanoparticle TiO₂ should, in theory, have beneficial impacts on the health of the general public and thus on the quality of life [3]. Since nanoparticle TiO₂ particles are inexpensive and may be integrated into different materials, photocatalytic systems and surfaces, they are not limited to large-scale applications in water treatment facilities. They may also be applied in homes, hospitals, or offices for disinfection or the degradation of water and air pollutants [6,9,10,11]. The use of nanoparticle TiO₂ for water treatment and disinfection is expected to have a positive effect on the environment, as it can replace more toxic substances such as organic biocides. This has a significant implication for developing nations. It can also improve the quality of the water released from water treatment plants by assisting traditional treatment methods to target more substances and thereby obtain a higher efficiency of the whole process [14,15,16].

TiO₂ has been extensively studied worldwide resulting in more than 13,600 publications between 2010 and 2011. Opportunities for TiO₂ now lie in areas such as energy conversion and storage. TiO₂ nanostructures have shown promise in designing Li-ion and Na-ion storage batteries and shape- and size-controlled TiO₂ nanostructures will continue to provide the base architecture to construct light-harvesting assemblies and facilitate photo-induced charge separation processes. As stated earlier in this review, pure nanoparticle TiO₂ is only activated by UV-light and this means that for indoor applications it is important to develop a TiO₂ based catalysts that can absorb visible wavelengths. Furthermore, the most efficient system set-up and substrate materials have to be determined for different applications in private and public facilities to assure maximal longevity, efficiency, and functionality of the photo-catalyst. A significant problem that still needs to be addressed is suppressing charge recombination in semiconductor materials. Studies have shown that manipulation of interfacial charge transfer can assist in improving photocatalytic and solar cell conversion efficiency. This may be achieved by surface modification surface modification and/or with a co-catalyst [12].

Another major challenge is that nanoparticle TiO₂ does not only destroy all organic materials but also the organic matrix in which the nanoparticles are embedded. This limits its application to inorganic environments. Nanoparticle TiO₂ is expected to have minimal negative effect both on human health and the environment since it is usually immobilized in/on a substrate material, for example, metals, tiles, or glass beads. However, nanoparticle TiO₂ can accumulate if its use is widespread and could potentially have health impacts for workers exposed to nanoparticle TiO₂ dust. There are currently no regulations related to the use of nanoparticle TiO₂ for water treatment, but standards on test methods for photocatalytic water purification are under development. Stricter water treatments standards would require new treatment methods and could thus further the application of photocatalytic systems with nanomaterials.

This review covered mainly TiO₂ semiconductor as well as RuO₂ and SnO₂. The principles and methods covered in this review apply to all photocatalyst materials in general. It is now possible to tailor catalysts to specific applications and there are ways of enhancing photocatalyst activity and improving performance. As well as destruction of viruses and bacteria, heterogeneous photocatalysis has been used to decompose natural organic matter, volatile organic compounds in water, air and soil and there are applications in consumer goods, food and medicine. Extending the wavelength for TiO₂ application is an area of increasing research. One of the ways of doing this is by doping, as already described in this review and this merits continuing research efforts.