Electrochemically doped/surface modified TiO2 nanotube arrays
This chapter describes some fundamental features of photoelectrocatalytic processes, including the basic concepts of the technique, the phenomena at the electrode/electrolyte interface and the development of new materials employed in the last few years related to the specific applications. The nanostructured materials used in the photoelectrochemical field can be called photoanodes (n-type) when oxidation reactions take place at the interface, and photocathodes (p-type) when the reduction is the main process [1, 2]. This chapter focuses on photoanode materials and how their surface influences the applications of this technique.
Photoelectrocatalysis could be described as a multidisciplinary field, involving surface science, electrochemistry, solid-state physics and optics. The basic concept is that when a semiconductor surface is irradiated by light (hν ≥ Eg) there is generation of electron/hole pairs (e−/h+) by the promotion of an electron from the valence band (lower energy level) to the conduction band (higher energy level). The electrons are forwarded to the counter electrode under positive bias potential (n-type) in order to minimize the recombination of these pairs due to the short life-time. When immersed in electrolyte the adsorbed water molecules and/or hydroxyl ions react with the holes on the valence band to generate hydroxyl radicals (●OH), which are a powerful oxidizing agent (+2.80 V) [3-5].
The first findings, from 1839, found that the photoelectrochemistry field was stimulated by the Becquerel effect . They observed a photocurrent flow of electrons due to illumination of a material connected by two electrodes immersed in solution. In 1972, the work of Fujishima and Honda had a huge impact on this field. They studied the use of a TiO2 semiconductor on the photoelectrolysis of water (water splitting) under anodic bias potential in a photoelectrochemical (PEC) cell [7, 8]. Nowadays, photoelectrocatalysis is an emerging field with many applications, such as organic compounds oxidation [9-11], inorganic ions reduction [12, 13], disinfection [14, 15] and production of electricity and hydrogen [16-18].
The development of this technique is intimately related to a better understanding of materials’ surfaces and properties. Highly ordered nanomaterial arrays have promoted a revolution in applications of these materials as nanotubes, nanowires, nanofibres, nanorods, nanowalls, etc. . The main applications of the technique include the degradation of unwanted environmental pollutants (organic and inorganic compounds) and converting sunlight directly into an energy carrier [4, 19, 20].
This work presents an overview of the fundamentals of photoelectrocatalysis and the huge contribution made by nanostructured architectures, as well as explaining the efficiency of the technique as a treatment method for organic and inorganic compounds and for water splitting.
2. Photoelectrocatalysis: Basic concepts
Advanced oxidation processes (AOPs) have been proposed as alternative methods for the degradation of recalcitrant organic compounds in water , air  and soil  in recent years . AOPs are based on the generation of hydroxyl radicals (●OH) as highly oxidant species, which are responsible for the oxidation of the major pollutants [4, 21]. Among the AOPs, heterogeneous photocatalysis deserves particular attention . The method is based on the use of a semiconductor (mostly TiO2) irradiated with light energy equal to or greater than its band-gap energy. Since 1972 it has been known that is possible to promote photoelectrolysis of water (water splitting) under anodic bias potential . Since then, photocatalysis has been explored to promote organics oxidation [9-11], inorganics reduction [12, 13], disinfection of water containing biological materials [14, 15] and production of electricity and hydrogen [16-18].
A semiconductor material is characterized by two energy bands separated by the band-gap energy, Eg. A semiconductor at absolute zero is insulating, because the valence band (lower energy level) is completely occupied and the conduction band (higher energy level) totally empty (Figure 1). To become conductive, charge carriers need to be created, usually by photoexcitation. The basic concept is that when a semiconductor surface is irradiated by light (hν ≥ Eg) there is generation of an electron/hole pair (e−/h+) by promotion of an electron from the valence band (VB) to the conduction band (CB) (Equation 1) [5, 24].
The oxidizing nature of the holes (h+) in the valence band means they generate ●OH radicals by the oxidation of H2O molecules or OH− ions adsorbed on the semiconductor surface, and are also able to oxidize organic molecules directly. The photoexcitation of TiO2 and possible oxidation of an organic compound (RX) are represented in Equations 1−4 [21, 25].
Although heterogeneous photocatalysis is a well understood process, and despite its promising results in water decontamination, its practical exploitation has been restricted by its low photonic efficiency, which is mainly due to recombination of the e−/h+ pair, as shown in Equation 5 [25, 26].
Therefore, there are considerable efforts being made to obtain new processes able to separate charge carriers and minimize their recombination rate [26, 27]. The combination of electrochemical and photocatalysis processes (photoelectrocatalysis) offers the opportunity to separate photo-generated e−/h+ pairs by gradient potential [28, 29]. Specifically, when the photocatalyst is attached to a conductive substrate (photoanode), there is the possibility to apply an anodic bias potential to the semiconductor and to modify the substrate/electrolyte interface. This alternative improves the efficiency of charge separation by driving the photogenerated electrons via the external circuit to the counter electrode [26, 28-30]. Figure 2 illustrates the mechanism of photoelectrocatalysis.
Furthermore, the great goal is to avoid the removal of photocatalyst suspensions. The immobilization of the photocatalyst particles on a solid substrate is usually applied on photoelectrocatalysis and therefore the process dispense next filtration step [28, 29].
It is interesting to understand why photoelectrocatalysis is efficient in charge separation. When a semiconductor is in contact with an electrolyte there is formation of a junction semiconductor/electrolyte interface, which determines the electron hole separation kinetics. The junction in a redox electrolyte causes a change in the electrochemical potential (Fermi level) due to discrepant potentials at the interface . Thus, the equilibration of this interface needs the flow of charge from one phase to another, and a band-bending is created within the semiconductor phase. The amount of band-bending in this Schottky junction will depend on the difference of the Fermi levels of semiconductor and electrolyte. The region where there is bending is called the space charge layer (SCL), which is characterized by the accumulation of electrons or holes at the surface [5, 19, 24, 31]. Figure 3 shows the behaviour of these charges in the semiconductor before and after this equilibration when it is in contact with an electrolyte.
Another method to control the Fermi level (and therefore the band-bending) is by applying a bias potential . For any given semiconductor and electrolyte, there is an exact potential for which the potential drops between the surface and the bulk of the electrode is zero; in other words, there is no space charge layer . Because the band edges are flat, this potential is called flat-band potential, Vfb (Figure 4). The application of any potential greater than the flat-band potential will increase the band-bending at the n-type semiconductor electrode, such as TiO2. In this case electrons are depleted and holes enriched at the surface, as we can see in Figure 4. When TiO2 is irradiated, it is observed that the photogenerated holes have an oxidizing power equivalent to the potential of the valence band edge, and are able to oxidize an RED molecule, whose formal potential is more negative than the valence band. In the case of TiO2, the H2O can be oxidized producing ●OH radicals. The electron in the conduction band flows via an external circuit to the counter electrode, where reduction reactions may occur, such as the reduction of H+ ions to H2 (Figure 2). It is important to note that in photo(electro)catalysis, the greater the band-bending (and therefore the SCL) the faster the electron/hole separation occurs, and then the recombination of charges is minimized [5, 19, 24, 31].
Thus, considering the high oxidative power of ●OH that is easily generated by irradiation of the TiO2 surface, an increased number of applications of photoelectrocatalysis has developed with the aim of promoting the degradation of organic pollutants to CO2 and minerals.
3. The degradation of organic compounds on thin films
The presence of recalcitrant organic pollutants such as pesticides, hormones, pharmaceuticals, phenols, surfactants and dyes in water and wastewater has been described in the literature as one of the most serious problems for human beings and the environment [32, 33]. The great concern is mainly that the genotoxic and mutagenic properties of these pollutants can cause bioaccumulation problems and transportation that is magnified in the food chain . They have therefore received great attention since they are released into the environment through a variety of human and industrial activities. Conventional techniques such as adsorption, precipitation, flocculation and reverse osmosis simply transfer organic pollutants from different phases or concentrate them in one phase, without actually removing them .
Different methodologies have been proposed to promote the complete degradation of organic matter. Among them, the use of advanced oxidative processes (AOPs) has been seen as an efficient alternative for pollutant degradation and has received a great deal of attention from several researchers. The
Over the past decades, electrochemical methods such as electrocoagulation, electrocatalysis oxidation and reduction, electro-Fenton, photoelectro-Fenton, photocatalysis and photoelectrocatalysis (Figure 5) have been pointed out as good alternatives to promote the degradation and mineralization of organic pollutants, since they combine the advantages of hydroxyl radicals formation and the efficiency of electrochemistry [21, 36].
In Electrochemically Mediated Oxidative Advanced Processes (EOAPs), hydroxyl radicals can be generated by direct electrochemistry (anodic oxidation) or indirectly through electrochemical generation of Fenton’s reagent. In photoelectrocatalytic oxidation the ●OH is generated heterogeneously by direct water discharge on specific anodes such as DSA and BDD electrodes . During the electro-Fenton reaction the hydroxyl radicals are generated homogenously via Fenton’s reaction .
Photoelectrochemical methods have been intensively investigated as promising alternative methods not only to remove organic pollutants but also to decrease toxicity, since they degrade substances in a short period of time. The degradation mechanism of photocatalysis can be classified into five steps: (1) transfer of reactants in the fluid phase to the surface; (2) adsorption of the reactants; (3) reaction in the adsorbed phase; (4) desorption of the products; and (5) removal of products from the interface region .
The key to obtaining success with photocatalytic and photoelectrocatalytic methods is the development of novel efficient materials as working electrodes, which present good optical, mechanical, electronic, electrochemical and catalytical properties . The choice of the synthesis method to produce the semiconductor material is of fundamental relevance, as it will determine the efficacy of the PEC treatment. All factors related to the surface material will influence the success of photoelectrochemical processes as morphological and structural features (particle size, surface area), good charge separation (e−/h+), suitable photonic efficiency and band-gap energy level .
3.1. Synthesis of thin film semiconductor materials
Emerging technologies providing feasible alternatives for the development of new materials have been the subject of several studies. Titanium dioxide is the most used material and can be prepared in the form of powder, crystals or thin films. To obtain good-quality materials there are many methods described in the literature, based on precipitation and co-precipitation [41, 42], solvothermal , sol-gel , microemulsion , electrochemical  and gas-phase methods .
Heterogeneous photocatalysis started with the use of TiO2 semiconductors in a slurry system (suspension of fine powder). The most efficient powder reported in the literature is the Degussa P25, which is a combination of rutile and anatase allotropic phases in the ratio 3:1. There are many advantages of using this powder: it provides high surface area showing excellent photocatalytic activity because of the adsorptive affinity of organic compounds on the surface of anatase . However, a post-treatment filtration step is required to separate it from the solution, which limits practical application as this is a time-consuming and costly process. Moreover, the suspended particles tend to aggregate, especially at high concentrations, which makes the separation more complicated and limits application in continuous flow systems .
Since 1993, the immobilization of TiO2 on a substrate has offered an alternative way of using powder and started a search for thin films [28, 47]. Several researchers have anchored photocatalysts onto a variety of surfaces, such as glass (ITO and FTO), silica gel, metal, ceramics, polymer, thin films, fibres, zeolite, alumina clays, activated carbon, cellulose, reactor walls and others . To support TiO2 there must be four main criteria: strong adherence, stability of the catalyst, high specific surface area to promote strong adsorption of the pollutant on the electrode surface . The substrate material has a great influence on the electron transfer along the film. It is reported that conducting glasses have a relatively poor connection within the film; on the other hand, metal substrates present a lower impedance because there is a reduction of charge transfer resistance leading to better PEC activity .
The photocatalytic activity of a TiO2 system mainly depends on its intrinsic properties, such as particle size, surface area, film thickness, crystallinity and crystal phase [33, 48]. For this purpose, many different techniques emerge from the need for immobilization, since the photocatalytic activity of the film is highly dependent on the preparation method . For instance, the most reported preparation routes are sol-gel , chemical vapour deposition , electrodeposition , sol-spray , and hydrothermal . Besides the preparation routes, the coating techniques also influence the resulting material properties. Deposition methods such as dip-coating , spin coating  and even the development of new coating methods based on conventional dip and spin coating  have been shown to be simple and able to produce stable materials.
When compared to other methods, the advantages of the sol-gel technique are easy control of deposits, reliability and reproducibility, resulting in good-quality nanostructured thin films . In fact, successful formation of the desired crystal phase is directly related to the starting material, composition, and deposition, as well as the annealing temperature. The crystal morphology has a direct relation to the light absorption as incident light affects photoelectrocatalytic efficiency. Film thickness can affect the efficiency of both light energy conversion and electron transfer; thick films may lower efficiency as these processes have a higher resistance . It has been also shown that the pH of the original solution can influence particle size . It is known  that acidic conditions favour the formation of smaller particles, while at higher pH values larger particle size is observed. The use of sol-gel methods has inspired a great number of studies on the development of new semiconductors for the suppression of electron/hole recombination and enhancement of the photosensitivity of titania for successful application . Therefore, the use of nanoporous thin films for photoelectrochemical purposes has been widely described in studies on the removal of organic matter such as dyes , phenol , tetracycline , toxic metals  and microorganisms . Annealing temperature has been intimately related to the crystal structure formation because phase transfer is temperature dependent. For many uses, including photoelectrocatalysis and solar cells, the most desired crystal structure is anatase, because this structure shows a higher charge carrier mobility than rutile [19, 63]. However, in many cases of photocatalysis, combinations of anatase and rutile have been used due to the higher photocatalytic activity that these display compared to pure anatase (probably due to the smaller band-gap energy of rutile (Eg=3.0 eV vs. anatase Eg=3.2 eV) absorbing more visible light radiation).
The use of mesoporous TiO2 thin films has also been studied. According to the definition of IUPAC, porous solids can be classified into three groups based on their pore diameter, namely microporous (5–20 Å), mesoporous (20–500 Å), and macroporous (>500 Å) materials . The success of mesoporous materials depends on the availability of precursor materials and the precision of control over the hydrolysis reaction, as well as the choice of an appropriate surfactant. All these parameters interfere with the obtaining of highly organized materials. In order to obtain mesoporous materials with good photocatalytic features it is necessary to use an appropriate method to produce films with a large surface area, pore-wall structure and crystallinity .
Other thin-film semiconductors have been used in the degradation of such organic compounds as WO3 , ZnO  and Fe2O3 [68, 69]. The anodic growing of tungsten trioxide thin film has been described as a good alternative to TiO2, mainly because of its intrinsic characteristics like lower band-gap energy of Eg=2.8−3.0 eV and higher photoactivity . Iron oxide (α-Fe2O3) has the desirable property of narrowing the band gap (Eg=2.2 eV), as well as low cost, electrochemical stability and low toxicity . ZnO (Eg=3.2 eV) has good properties for use as a photocatalyst, such as high photocatalytic efficiency, low cost and environmental friendliness . It can also be used for degradation and disinfection purposes, as it can degrade dirt and inhibit the growth of microorganisms .
3.2. Operational characteristics on the PEC systems
The basic photoelectrochemical reactor setup consists of three conventional electrodes (working, reference and counter electrode) immersed in an aqueous electrolyte contained within a vessel for the potentiostatic mode. A two-electrode system (working and counter) can also be used when current density is used to supply the system. The vessel containing the aqueous electrolyte is transparent to light or fitted with an optical window, usually quartz, that allows light to reach the photoactive electrode .
Besides material properties, some operational parameters such as pH, biased potential, initial concentration of analyte and electrolyte composition have a direct influence on the degradation of organic pollutants. The point of zero surface charge (pzc) of the TiO2 at the electrode/electrolyte interface will determine the adsorption of the pollutant in relation to the pH and pKa of the pollutant. In acidic conditions TiO2 is positively charged, while in basic conditions it is negatively charged, according to the equations below [25, 33, 73]:
The influence of biased potential on the degradation rate must be optimized as a function of the flat band potential. Generally, when the potential is increased, the degradation rate increases as well until no more gain is observed because electrons and holes have a good separation and recombination rate is minimized [33, 74]. Current density can be applied instead of potential, as it requires a much simpler arrangement of two electrodes, lowering costs and favouring the photoelectrocatalytic application on large-scale reactors [4, 43].
The initial pollutant concentration, especially for wastewaters and coloured solution, will limit the photoanode activation by light . Moreover, at high concentration the photoelectrochemical efficiency is decreased and longer treatment periods will be required to achieve complete pollutant removal. Depending on the pollutant, it is possible to promote the degradation at high concentrations [4, 25].
Recent investigations prove that light intensity and lamp irradiance are critical factors in photoelectrochemical systems. It has been reported in the literature that the higher light intensity achieved, the faster the degradation rate will be . Zainal and colleagues  demonstrated that a 100 W UV lamp was almost equivalent to a 300 W halogen lamp, probably due to the higher intensity of the halogen lamp.
When the degradation is conducted in the presence of different electrolytes, there will be significant change in the degradation rate. In the presence of chloride, the degradation is improved because there will be generation of chlorine radicals, with a high oxidizing power which is not observed in sulphate and nitrate mediums .
The PEC reactor also plays an important role in the efficiency of photoelectrochemical methods. Different materials (glass, quartz and Teflon) and shapes are employed on these systems. The photoanode irradiation can be used either externally or internally . The reactor could be rectangular or cylindrical, although the latter makes greater use of light and hence better performance. There are single chamber reactors and double-vessel reactors, also known as H-type .
4. Strategies to enhance the PEC efficiency
Several photocatalysts have been applied in photoelectrocatalysis, among them TiO2, WO3 , ZnO , CdS, Fe2O3 [68, 69] and SnO2. Over the years considerable effort has been devoted to the improvement of the materials used in photocatalysis. TiO2 has become one of the most common materials used in materials science  as it is environmentally friendly, low cost, has a long lifetime of electron/hole pairs, presents a compatible energy position of BV and BC, and has good chemical and thermal stability and superior catalytic stability [20, 76]. Among these features, the band edge positions relative to H2O oxidation represent a very important characteristic that improves the applicability of TiO2 in photo(electro)catalysis to decompose H2O to H2 and O2 and also to create ●OH radicals . There are many transition metal oxides with semiconductor properties, but many of them do not have suitable electronic properties (energy position of bands edges) for useful electron transfer reactions.
Some of the main applications of TiO2 photoelectrocatalysis have involved water-splitting [16, 77, 78] inactivation of microorganisms [14, 79] and degradation of contaminants in water [10, 33, 39, 78, 80]. Although it is the most suitable material for such applications, titanium dioxide has some limitations that hinder its use in technological applications. For example, it is activated only under ultraviolet irradiation (λ ≤ 387 nm), and thus the use of sunlight is limited because it provides up to 5% of UV light; it also presents recombination of electron/hole pairs. In order to obtain a better utilization of the photocatalytic properties of TiO2 and to achieve more responsiveness to the visible wavelengths, the preparation of nanostructured materials and their surface modification or doping (band-gap engineering) has emerged as a potential method.
Thus, in order to increase the efficiency of photoelectrocatalysis, organized nanostructured materials, especially those involving electrochemical methods of preparation, have attracted attention. The main advantages are discussed below.
4.1. Nanostructured morphologies
Nanostructured materials represent an important challenge of current science, and the new materials have presented special physical and chemical properties. Recently, one-dimensional (1D) nanostructures such as rods, belts, wires and tubes have become a focus of intensive research, mainly due to their high surface area (ideal for catalysis as it facilitates reaction/interaction between the devices and the interacting media) and other exceptional properties such as electrical properties: charge carrier transfer is mainly governed by the quantum confinement phenomenon .
The discovery of carbon nanotubes by Iijima in 1991 , with their variety of interesting properties, boosted research focused on the synthesis of tubular nanostructures of other materials. Among the various nanotube materials, titanium dioxide nanotube arrays are of particular interest because of their many applications, for example in photo(electro)catalysis [10, 78, 83-87], sensors [88, 89], biosensors , dye-sensitized solar cells [91, 92], hydrogen generation by water photoelectrolysis [77, 78, 93], photocatalytic reduction of CO2 [94, 95] and biomedical-related applications [96, 97].
In recent years, a great number of investigations have focused on the photocatalytic activity of TiO2 nanomaterials and effective ways to improve their photocatalytic efficiency. Various nanostructures have been reported, such as nanowires , nanofibres , nanorods [100, 101], and nanowalls , but TiO2 nanotubes are certainly the most promising and explored architecture.
4.1.1. TiO2 nanotube arrays
As previously mentioned, TiO2 is a widely studied materialdue to its versatility, and in photoelectrocatalysis it is undoubtedly the most explored semiconductor. The use of the TiO2 nanotubes morphology has allowed advances in photo(electro)catalysis due to specific improvement of properties and will be further discussed.
For photoelectrocatalytic applications TiO2 nanotubes (TiO2 NTs) present interesting properties, such as large internal surface area, which can be easily filled with liquid enabling intimate contact with electrolytes and excellent charge transport [39, 94]. Due to its high structural organization, the nanotubes architecture exhibits excellent electron percolation pathways for vectorial charge transfer between interfaces, thereby minimizing the recombination of charges. Figure 6 illustrates an image of scanning electron microscopy (SEM) of TiO2 NTs prepared under electrochemical anodization. As the TiO2 film grows on the metal surface (is not deposited) there is a good electrical connection between the oxide and the metal. Zhu and colleagues  found charge carrier recombination much slower in the TiO2 NTs films than in the nanoparticulate TiO2 films in dye-sensitized solar cells.
Additionally, the morphological parameters of the architecture can be precisely controlled when the material is prepared by electrochemical anodization. The control of the nanotube dimensions is important because each application may require morphological surfaces with particular characteristics. For example, Liu et al.  found that the photoelectrocatalytic activity shows a dependence on the length of the nanotube arrays. They studied the degradation of phenol at TiO2 NTs electrodes with different tube lengths under UV irradiation and applied potential. It was verified that a short nanotube array shows better photoelectrocatalytic activity than a long nanotube array, which can be explained by the reduced recombination effects. However, the photocatalytic degradation (no applying potential) showed that longer nanotubes were more efficient because they favour light trapping.
More information can be obtained in some excellent reviews found in the literature, dealing with preparation, properties, strategies to increase the photoactivity and applications of TiO2 NTs [19, 20, 39, 81, 94, 104-107]. Titania nanotubes can be synthesized in two forms: powder form and self-organized nanotube arrays grown on a substrate of metallic titanium. Several techniques for the preparation of TiO2 NTs have been reported, such as hydro/solvothermal methods , sol-gel , template-assisted methods  and electrochemical anodization [39, 105, 106]. The growth of TiO2 NTs by electrochemical anodization in a fluorinated-based electrolyte is less expensive and simpler that most of these methods and allows precise control of dimensions, presenting a more orderly arrangement of nanotubes .
The first self-organized oxide obtained by anodization in electrolytes containing hydrofluoric acid was reported by Zwilling and colleagues in 1999, where a nanoporous structure was achieved . In 2001, Gong and colleagues  developed the first generation of highly ordered and vertically oriented nanotube arrays of 500 nm length. The structure was obtained by electrochemical oxidation of titanium in a HF aqueous electrolyte. The fabrication of TiO2 NTs films was performed in a two-electrode electrochemical cell using aqueous electrolytes containing 0.5-3.5 wt. % HF and voltages varying from 3 to 23 V. They found that at low voltage (3 V), porous films are obtained and at higher voltage (23 V) the nanotube structure was destroyed. The ideal conditions were 0.5 wt. % HF electrolyte applying 20 V for 20 min.
In 2005, Cai and colleagues  developed the second synthesis generation of titania nanotubes. They found that adequate control of the electrolyte pH can decrease the oxide chemical dissolution rate; thus, the tube length is enhanced using aqueous buffer electrolyte. The pH of a KF-containing electrolyte is adjusted to 4.5 using additives such as sulphuric acid, sodium hydroxide, sodium hydrogen sulphate, and/or citric acid. This usually obtains TiO2 NTs of 4.4 μm in length.
The third synthesis generation of titania nanotube arrays, initially reported by Ruan and colleagues  in 2005, involves improvements in nanotube-array length using non-aqueous electrolytes or polar organic solvents such as formamide, N-methylformamide, dimethyl sulphoxide, and ethylene glycol mixed with HF, NH4F or KF to provide fluoride ions [112, 115-117]. Ruan and colleagues  also studied the anodization of titanium in polar organic solvent using mixtures of dimethyl sulphoxide (DMSO) and hydrofluoric acid. TiO2 nanotube arrays of 2.3 µm length were obtained in DMSO+4.0% HF electrolyte applying 20 V for 70 h.
The fourth synthesis generation of TiO2 NTs was developed by Richter and colleagues  and Allam et al. , and is characterized by the fabrication of nanotube arrays by Ti anodization using fluoride-free HCl aqueous electrolytes. The mechanism of TiO2 NTs formation on Ti substrate is well studied in the literature [94, 105, 106].
126.96.36.199. Mechanism of formation of nanotubes by electrochemical anodization
The production of oxide films on metal surfaces by oxidation in an electrolytic process can be called electrochemical anodization. In practice, a metallic electrode compatible with oxide growth is connected to the positive pole (anode) of a dc power supply and the cathode, usually a platinum piece (or another material, such as carbon for example) is connected to the negative pole (Figure 7). The electrodes are placed in an electrolytic solution and when a potential is applied in the system the metal reacts with oxygen ions from the electrolyte, growing an oxide film on the surface. The electrons resulting from the oxidation travel through the external circuit to reach the cathode, where they can react with H+ ions and generate bubbles of H2 . The key point that determines the form of the oxide is the composition of the electrolyte. The TiO2 NTs, in this case, can be achieved in electrolytes containing fluoride ions, with adjustments of applied potential and anodization time.
The growth of self-organized TiO2 NTs (as well as porous structures of other metals such as Zr, Nb, W, Ta, and Hf) by electrochemical anodization in fluoride-containing electrolyte is governed by competition between steps that occur simultaneously.
First, there is the formation of oxide on the metal surface. In this step, there is a field-assisted oxidation of Ti metal to form Ti4+ species which will react with O2- (from H2O). After the formation of an initial oxide layer, further oxide growth is controlled by field-assisted ion transport, where O2- anions migrate through the oxide layer until they reach the metal/oxide interface, where they react with the metal [94, 106, 120].
In another step, Ti4+ ions migrate from the metal through the oxide by field-assisted transport until they reach the oxide/electrolyte interface. Then, small pits are formed due to the localized dissolution of the oxide by the high electrical field, which act as pore-forming centres.
The key step is the chemical dissolution of oxide by fluoride ions at the as-formed pits, forming soluble fluoride complexes. The Ti4+ ions field transported at the oxide/electrolyte interface are also complexed [94, 106, 120].
If the chemical dissolution is too high or too low, there is no formation of nanotubes. The dissolution rate can be adjusted by varying the concentration of F− and pH (more acidic pH and higher concentrations of F− increases the chemical dissolution) . This was the principle used to obtain longer and smoother nanotubes, leading to the second and third generations of TiO2 NTs.
When the rate of pore growth at the metal–oxide interface becomes identical to the rate of oxide dissolution at the pore–bottom–electrolyte interface, the thickness of the barrier layer remains unchanged, although it moves further into the metal, making the pore deeper [94, 106, 120]. Commonly, the wall thickness of TiO2 NTs varies from 5 to 30 nm and the pore size from 20 to 350 nm (tube diameter is reported to be linearly dependent on the applied anodic potential during growth [106, 121]). The length often varies from 0.2 to 1000 μm; the aspect ratio, defined as the ratio between length and diameter of the tube, can be controlled from about 10 to approximately 20,000 by selection of appropriate anodization variables .
4.1.2. Nanostructured arrays of other semiconductors
Nanostructured architectures are also fabricated by electrochemical anodization for other semiconductors of interest in photoelectrocatalysis, such as ZnO, WO3 and Fe2O3.
Prakasam and colleagues  prepared nanoporous film of Fe2O3 by submitting a Fe foil to electrochemical anodization in electrolyte composed of 1% HF+0.5% ammonium fluoride+0.2% 0.1 M nitric acid (HNO3) in glycerol (pH 3) at 10°C. LaTempa and colleagues  produced α-Fe2O3 (hematite) nanotubes by potentiostatic anodization of iron foil in an ethylene glycol electrolyte containing NH4F and deionized water. Hematite has a band gap of ≈2.2 eV (indirect) and can absorb light at λ ≤ 560 nm; it can therefore be activated in a large part of the solar spectrum.
Lai et al.  prepared WO3 nanotubes by electrochemical anodization of W foil in electrolyte composed of 1 M of sodium sulphate+0.5 wt.% of ammonium fluoride at 40 V. The WO3 is photoactive when irradiated by visible light due to its small band-gap energy (2.4 eV to 2.8 eV) and has attracted scientific interest in photo(electro)catalysis. Some reviews [29, 70] have explored the use of WO3 photoanodes mainly in photoelectrochemical water splitting.
Park and colleagues  reported a synthesis of ZnO nanowires by electrochemical anodization on a Zn foil using as electrolyte 5 mM KHCO3 aqueous solution. ZnO has a similar band gap and band positions of TiO2 (Eg about 3.2 eV), but higher quantum efficiency than TiO2. On the other hand it has limited applications due to its photocorrosion in acidic medium .
4.2. Band-gap engineering
Despite all the improvements made to TiO2 as a photoactive catalyst, the material still presents problems, such as activation with UV irradiation (λ≤387 nm), due to its wide band gap (Eg=3.2 eV). Thus, the use of solar energy is limited since the activation of TiO2 occurs only from UV light, which corresponds to a small fraction (≈5%) of the sun’s energy compared to visible light (45%) . In this sense, efforts have been directed at shifting the optical response of titanium dioxide from the UV to the visible spectral range, which would be of great utility in photo(electro)catalysis and other applications of TiO2. This modification of optical properties of semiconductors has been called band-gap engineering [19, 39, 94, 107].
Modification of TiO2 properties has been achieved mainly by (i) doping with different transition metal ions (such as Cr , Co , W , Zr  and Fe ) and with different anions (such as N , F , S , B , C ) that replace oxygen in the crystal lattice, and (ii) by surface decoration, which includes coupling with other semiconductors and deposition of particles of noble metals [14, 84, 134-137].
However, these arrangements frequently increase only the absorption and do not properly improve material properties such as the stability of the semiconductor under illumination, efficiency of the photocatalytic process, and the wavelength range response. One example is the CdS, which absorbs a good portion of the visible radiation but is usually unstable and photodegrades with time .
|N||Anodization of Ti–N alloy*|
Anodization in nitrogen-containing electrolyte*
Electrodeposition in nitrogen-containing electrolyte
|, , , , , , , |
|C||Anodization in carbon containing electrolyte*||, |
|F||Anodization in containing electrolytes*|||
|B||Anodization in boron-containing electrolyte*|
Electrodeposition in boron-containing electrolyte
|, ,  |
|W||Anodization of Ti–W alloy*|
Anodization in tungsten-containing electrolyte*
|, , |
|Zr||Anodization in zirconium-containing electrolytes*|
Electrodeposition in zirconium-containing electrolyte
|, , |
|La||Electrodeposition in lanthanum-containing electrolyte|||
|Si||Anodization in silicon-containing electrolyte*|||
|Nb||Anodization of Ti–Nb alloy*|||
|Ag||Electrodeposition in silver-containing electrolyte||, |
|Pt||Electrodeposition in platinum-containing electrolyte||, |
|Pd||Electrodeposition in palladium-containing electrolyte||, |
|CdS||Electrodeposition in Cd and S-containing electrolyte||, |
|CdTe||Electrodeposition in Cd and Te-containing electrolyte|||
|Cu2O||Electrodeposition in Cu-containing electrolyte|||
In order to make materials more photoactive under visible light and more stable under certain conditions, and to have lower band-gap energy, the doping of TiO2 with several metals and non-metal compounds has also been explored: Table 1 shows a summary of the electrochemical methods adopted to promote doping/surface modification of TiO2 nanotubes,with the related references.
4.2.1. Doped TiO2 nanomaterials
Asahi et al. , in a 2001 study, developed a method for TiO2 visible light activation through doping of C, N, F, P, or S for O in the anatase TiO2 crystal using calculated densities of states (DOSs). They found that the substitutional doping of N was the most effective method because nitrogen p states contribute to band gap narrowing by mixing with O 2p states. Nitrogen can be easily introduced into the TiO2 structure, due to its comparable atomic size with oxygen, small ionization energy and high stability.
There are two main ways to perform anion doping in TiO2 by electrochemical techniques: (i) electrodeposition and (ii) adding a precursor of the element into the electrolyte during electrochemical anodization to oxide formation. It should be noted that for this the TiO2 film must be immobilized on a conductive substrate, as in the case of TiO2 NTs grown on metallic titanium.
In 2006, Shankar and colleagues  described a simple way to introduce N atoms into TiO2. N-doped thin films were fabricated by anodic oxidation of a pure titanium sheet in electrolyte composed of 0.07 M HF, NH4NO3 (from 0.2 to 2.5 M) and NH4OH to adjust the pH to 3.5. The material showed optical absorption in the visible wavelength range from 400 to 530 nm. The XPS data confirmed that all the incorporated nitrogen is substitutional on the oxygen site, and the proportions of N atoms in TiO2−xNx were x=0.23, x=0.09 and x=0.02. The N-doped samples exhibited a shift in absorption toward the visible spectra from 400 to 510 nm. Antony and colleagues  prepared N-doped TiO2 NTs by anodizing Ti foils in ethylene glycol+NH4F+water mixture containing urea as a nitrogen source. They used various concentrations of urea and achieved different N concentrations in TiO2 film, determined by X-ray photoelectron spectroscopy (XPS). There was nitrogen incorporation in TiO2lattice mainly in substitutional form (substitution of O2− ions by N3− ions). The doped samples showed visible light response, and the calculated optical band gaps were 3.27, 3.21, 2.75 and 2.77 eV for pristine TiO2, TiO1.85N0.115, TiO1.813N0.14and TiO1.84N0.121, respectively. Zhou et al.  fabricated N-doped using the same methodology, via anodic oxidation of Ti in electrolyte composed of ammonium fluoride (NH4F) and triethylamine (C6H15N). Nitrogen was successfully introduced into the TiO2 lattice replacing oxygen atoms, and as a result there was a shift of TiO2 band edge from 380 nm to 405 nm in N-doped TiO2.
Kim et al.  produced N-doped TiO2NTs by anodization of a high-purity TiN alloy with approximately 5 at.% of N in a glycerol+water (50:50 vol%)+0.27 M NH4F electrolyte. XPS data of the sample surfaces indicated 2−3 at.% of N atoms present as Ti–O–N in the nanotubes. They found that the nanostructured layer grown on TiN alloy showed decreased UV response compared with pure TiO2 NTs film, but showed a strongly increased photoresponse in visible light spectra.
Li and colleagues  used an electrochemical technique to dope TiO2 with nitrogen atoms, in two steps. N-doped TiO2 NTs were prepared by electrochemical anodization in glycerol electrolyte, followed by electrochemical deposition in NH4Cl solution. The optimal conditions in electrodeposition were: voltage of 3 V, reaction time 2 h, and NH4Cl concentration of 0.5 M. Both the photoelectrochemical properties and photocatalytic activity under visible light irradiation were enhanced after N doping into TiO2 nanotube arrays.
By using the aforementioned electrochemical techniques for the nitrogen, it is also possible to perform doping with other non-metals, such as C and B, for example. Milad and colleagues  achieved carbon-doped titanium oxide nanotubular arrays via anodic oxidation of titanium foil at 20 V in acidic (0.5 M H3PO4+0.14 M NaF) and organic media (ethylene glycol+0.3 wt% NH4F) with 0.5 and 1 wt% carbon source (polyvinyl alcohol). Approximately 2.75% and 8.45% carbon was incorporated into the TNT in the acidic and organic electrolyte, respectively. The highest photocurrent density was observed for the sample with the higher amount of carbon atoms incorporated. Krengvirat et al.  produced carbon-incorporated TiO2 by anodic oxidation in EG containing 0.5 wt% NH4F+1 wt% water. The interstitial carbon arising from the pyrogenation of ethylene glycol electrolytes induced a new C
Besides anion doping, there are numerous papers that investigate the effect of doping with metal ions in the TiO2 lattice. The metal ions can occupy two different positions in the TiO2 matrix, which are substitutional and interstitial, depending on the ionic radius of the metal. The dopant occupies the interstitial sites if the dopant radius is much smaller than the matrix cation, in this case, titanium. If the dopant has similar ionic radius of Ti, the substitutional mode is adopted . In metal-doped TiO2, new energy states can be formed either within or beyond the VB and CB, decreasing band-gap energy. However, transition metals may also act as recombination sites and may cause thermal instability in the anatase phase of TiO2 [7, 27].
Tungsten-doped TiO2 NTs were prepared by Gong et al.  in glycerol/fluoride electrolyte containing sodium tungstate via the electrochemical oxidation of a Ti substrate. XPS data showed that the W6+ ions were loaded into TiO2 lattice by displacing Ti4+ ions and forming W–O–Ti bonding. Thus, the UV−Vis spectra of W-doped samples show red shift and decrease the band-gap energy from 3.18 eV (bare TiO2 NTs) to 2.97 eV (W-doped TiO2 NTs). These findings can be attributed to the fact that the conduction band of the W-doped samples was reformed in the presence of W6+ ions. Das and colleagues  prepared tungsten-doped TiO2 NTs by electrochemical anodization of Ti–W alloys. The sample containing 9% W presented band-gap energy of 2.83 eV and higher visible photocurrents than undoped samples.
Liu and colleagues  produced Zr-doped TiO2 NTs. They prepared TiO2 NTs by electrochemical anodization in 0.14 M NaF and 0.5 M H3PO4 electrolyte, and made the zirconium doping by electrodeposition in 0.1 M Zr(NO3)4 electrolyte, varying the applied potential. When the amount of zirconium in TiO2 was small (lower potentials of deposition) zirconium entered into the lattice of TiO2, acting as defect positions, improving separation of charges. At higher Zr amounts, zirconium atoms were partially unable to enter into the TiO2 lattice, acting as recombination sites on the TiO2 surface, decreasing the photocatalytic efficiency. Using a similar approach, Nie and colleagues  produced lanthanum-doped (La-doped) TiO2 NTs. After the preparation of TiO2 NTs, they executed a cathodic electrochemical process using lanthanum nitrate solution as the La source. The material became visible photoactive, and the band gap was decreased from 3.32 eV (undoped) to 3.03 eV (La-doped NTs).
Another approach reported in the literature focuses on the incorporation of more than one anion (or an anion and a cation) in the structure of TiO2, which is called codoping. Su et al.  prepared N-F-codoped TiO2 NTs by electrochemical anodization of Ti in oxalic acid+NH4F electrolyte. N-doping into TiO2 resulted in the creation of surface oxygen vacancies, and F-doping produced several beneficial effects, such as the creation of surface oxygen vacancies, which enhance the surface acidity, and creation of Ti3+ ions, which reduce electron/hole recombination. Zhou and colleagues  produced B,N-codoped TiO2 nanotube arrays. Sun et al.  produced Si–W codoped TiO2 NTs using a one-step anodization process with the presence of silicotungstic acid in the electrolyte, and the doped samples presented visible photocurrent 2.5 times larger than bare TiO2 NTs. Xua and colleagues  produced passivated n–p co-doping of niobium and nitrogen into TiO2 lattice by anodizing Ti–Nb alloys and posterior N-doping. Liu et al.  produced N/Zr-codoped TiO2nanotube arrays in a two-step process. Firstly they prepared the TiO2 NTs by electrochemical anodization and then accomplished doping using electrochemical deposition in Zr(NO3)4 and NH4Cl electrolyte. The doped materials presented increased photoactivity under UV and visible light; the visible light sensitivity was caused by N-doping, and Zr-doping was responsible for enhancing the charge separation.
Although several mechanisms have been proposed for doping from experimental and theoretical data, it is not possible to clearly understand the role of dopants and therefore there is no consensus in the scientific community [7, 166]. Table 1 shows a summary of the electrochemical strategies fordopingTiO2 nanotubes, with the related references.
4.2.2. Composite semiconductor as photocatalysts
The coupling of two semiconductors with appropriate energy CB and CV can reduce the recombination of e-/h+ pairs due to the transfer of carriers from one semiconductor to the other, as can be seen in Figure 8. Furthermore, depending on the band-gap energy of the semiconductor used, the composite can be activated in the visible region .
There are few papers that report on the coupling of semiconductors by electrochemical techniques. In this case, the composite is produced by a two-step process. CdS is the most used semiconductor to coupling with TiO2 due to its small band-gap energy (Eg=2.4 eV). Li and colleagues  produced CdS nanoparticles-modified TiO2 nanotube arrays by electrodeposition via direct current. In the electrodeposition step, they used as electrolyte a mixed solution of 0.01 M CdCl2 in dimethyl sulphoxide (DMSO) with saturated elemental sulphur. CdS was cathodically electrodeposited at the optimum constant DC density of 0.5 mA cm−2 for 5–15 min. They found that the photocurrents of CdS/TiO2 NTs were much larger than those of pure TiO2 NTs. Under UV−Vis irradiation, both semiconductors are excited and as the conduction band of TiO2 is more anodic than that of the CdS there is efficient electron transfer between the CdS and TiO2. Thus, the photogenerated electrons are injected from the conduction band (CB) of CdS to the CB of TiO2; at the same time, the holes transfer from the valence band (VB) of TiO2 to the VB of CdS. In heterojunctions such as CdS/TiO2 there are less electron/hole recombinations and enhanced light absorption, both UV and visible. Zhang et al.  prepared water-soluble CdS quantum dots (QDs) and deposited on highly ordered TiO2NTs by various methods, including cyclic voltammetric (CV) electrodeposition. The QDs were prepared using 0.01 mol L−1 cadmium nitrate and 0.01 mol L−1 sodium sulphide dissolved in 6 × 10−5 mol L−1
Feng and colleagues  prepared a heterojunction of CdTe/TiO2 NTs. CdTe is a direct band-gap semiconductor with Eg=1.5 eV, absorbing almost across the visible spectrum. After the preparation of TiO2 NTs, CdTe nanoparticles were pulse electrodeposited in a conventional three-electrode system (with the TiO2 NTs as working electrode) in electrolyte solution containing 0.08 mol L−1 CdSO4 and 0.05 mol L−1 NaTeO3. The pulse on–off time ratio was 0.2:1, with a running voltage of −1 V. A red shift of 50 nm was observed in CdS/TiO2 NTs composite and the calculated optical band gap was 1.5 eV. The positions of CB and VB in relation to the TiO2were similar to the CdS; there was electron injection from the photoexcited CdTe to TiO2 CB, and the photogenerated holes moved from the TiO2 VB to the CdTe VB, preventing the recombination of charges.
Tsui and colleagues  studied the modification of TiO2 NTs with Cu2O by electrodeposition. Cu2O is a p-type semiconductor with a direct band gap of 1.95–2.2 eV. The junction between p-type Cu2O and n-type TiO2 in principle enhances the separation of electron/hole pairs; the Cu2O is also visible-light responsive. Electrodeposition of Cu2O was performed using the as-prepared TiO2 NTs with working electrode using a three-step pulse plating method (−0.5 V for 5 ms, −0.3 V for 0.5 ms, and 0 V for 5 s) from a solution containing 0.02 M Cu(CH3COO)2 and 0.1 M NaCH3COO (pH 5.7). The Cu2O/TiO2 composite presented visible light absorption and the band gap values obtained were 3.27 eV for TiO2 and 2.21 eV for Cu2O/TiO2 heterojunction. However, Cu2O on TiO2 NTs dissolves under intense light, limiting the use of Cu2O in photoelectrochemical devices.
4.2.3. Metal deposition
The decoration of TiO2 by dopants of nanoparticles of noble metals (such as Ag, Au, Pt, and Pd) has attracted attention in order to enhance the photoactivity of the material. Due to different Fermi levels of TiO2 and the metal nanoparticles, a Schottky barrier can be formed in the new material. Therefore, there is a rectification of the charge carrier transfer where the energetic difference at the semiconductor/metal interface drives the e− from the CB of the TiO2 into the metal nanoparticles. In other words, the metal acts as an electron trap, promoting interfacial charge transfer and therefore minimizing recombination of the e-/h+ pairs, as shown in Figure 9 .
Xie and colleagues  produced Ag-loaded TiO2 NTs using pulse current deposition technique in 0.01 M AgNO3 and 0.1 M NaNO3 electrolyte, using the as-prepared TiO2 NTs as working electrode. They applied −15 mA cm−2 of pulse current with 0.1 s on-time and 0.3 s off-time. Highly dispersed Ag nanoparticles of 10–40 nm were deposited on TiO2. TiO2 NTs and Ag/TiO2 NTs showed a similar maximum photocurrent density λ (imax 330 nm), but Ag/TiO2 NTs displayed much more intensive photocurrent response, which can be explained by the Schottky barrier formation separating the charge carriers more efficiently. Zhang and colleagues  prepared N-doped TiO2 NTs and loaded Ag nanoparticles on the TiO2 surface by electrochemical deposition using 0.2 g L-1 AgNO3 in 2.5 g L-1 EDTA solution applying −0.1 V for 1−20 s.
Xing et al.  produced Pt-nanoparticles-decorated TiO2 NTs by cyclic voltammetry electrodeposition in 19.3 mM H2PtCl6 solution from −0.4 to 0.5 V at a scan rate of 10 mV s−1 (controlling the number of cycles). Yin and colleagues  also prepared Pt/TiO2 NTs using an electrochemical approach, but using AC electrodeposition at 2–4 V for 5−30 min in solution containing 1 mmol L–1 of H2PtCl6.
In the paper of Qin and colleagues  Pd particles were deposited onto the TiO2 NTs electrode by a pulse electrodeposition technique in PdCl2 (2 g L−1) electrolyte solution (pH 1.5). Cheng et al.  prepared Pd/TiO2 NTs through an electrochemical deposition method at a constant potential of −0.8 V using PdCl2 solution (1 mM) in 0.5 mol L−1 NaCl electrolyte. The Pd/TNTs sample displayed absorption between 540 nm and 700 nm and presented transient photocurrent density of about 0.094 mA cm−2, higher than that of TNTs (0.067 mA m−2) under xenon lamp irradiation, indicating that decoration with Pd improves the charge separation, according to the Schottky barrier formation mechanism.
All these materials have been demonstrated to massively improve photoelectrocatalytic oxidation processes. Works dealing with water contaminated by a wide range of compounds are discussed below and summarized in Table 2.
5. Application of nanostructured materials in photoelectrocatalysis
As the complexity of contaminants increases, the efficiency of photoelectrocatalytic treatment methods needs to be enhanced by the use of different strategies, as they pose a potential risk to the environment. Most reported work tackles the oxidation of organic pollutants, such as dyes of different classes and industry uses, hormones, pharmaceuticals, pesticides, etc. Oxidation of biological microorganisms such as bacteria and fungus has also been investigated. In all these studies, oxidation is promoted by ●OH action generated at the interface photoanode/electrolyte. As discussed previously, these hydroxyl radicals are generated on n-type semiconductors when the holes (h+) on the electrode surface react with water and/or hydroxyl ions.
The reduction of inorganic contaminants has been studied as well. The main contaminants described have been bromide, nitrate, nitrite and CO2. In this case, the reduction takes place at a p-type semiconductor . The reduction of toxic metals (Cr6+ to Cr3+) has also been described  in a photoelectrocatalytic process where the cathode is Pt but is conjugated in a system where the organic molecules are oxidized simultaneously in a photocathode such as Ti/TiO2 and the electrons are forwarded to the counter electrode, where the reduction of Cr (VI) takes place . Solar conversion of CO2 to hydrocarbon fuels seems promising to reduce global warming for improved sustainability. Solar fuels include hydrogen, carbon monoxide, methane and methanol .
More recently, the application of semiconductor materials has received a great deal of attention in a re-emerging field: the generation of hydrogen as a clean energy carrier. Studies have described the direct water splitting process and the degradation of organic pollutants in order to obtain hydrogen . For this purpose, the use of n and p-type semiconductor materials using the photoelectrocatalysis method was investigated. The choice of the semiconductor material for hydrogen generation purposes depends on the valence and conduction-band energy levels, which are pH dependent (Figure 10).
The lower edge of the conduction band needs to be greater than the energy level for H2 evolution (according to Equation 11). For water-splitting purposes (Figure 10), the upper edge of the valence band needs to have enough energy to promote the H2O/O2 reaction (Equation 12), while for simultaneous organic-pollutant removal the energy level must be more electropositive than the OH−/●OH level for hydroxyl radical formation (Equations 11,12):
The use of solar light for hydrogen generation purposes has been desirable for the same reasons as for PEC purposes. Hence, the development of photoanodes that absorb light in the visible region (λ>400 nm) is necessary, and could be achieved by lowering the photoanode band-gap energy.
5.1. TiO2 nanomaterials applied to water treatment
The use of TiO2-nanostructured materials in the removal of contaminants is undoubtedly a successful system in the treatment of wastewater. The use of nanotubes obtained from different routes has been described as an efficient alternative method to promote higher discoloration and partial mineralization of main organic pollutants, as they have a high and homogenous surface area and suitable photocurrent values.
The degradation of organochlorinated compounds [169, 170], pesticides [171, 172], aromatic amines , biological microorganisms [14, 15], hormones (endocrine disrupters) [173, 174], flameretardants  and mainly dyes  has been reported with high efficiency shown by nanotube materials acting as photoanodes in photoelectrocatalytic treatment.
TiO2 NTs have proved to be more photoactive and to improve the efficiency of PEC degradation of pentachlorophenol under biased potential, with sodium sulphate as electrolyte (0.01 mol L−1) and low pH of the original solution. The photoelectrocatalytic processes have been shown to be more efficient than electrocatalytic, photolytic and photocatalytic techniques . Quan and colleagues  also observed the synergistic effect of photoelectrocatalysis compared to photocatalytic and electrochemical processes aiming at the degradation of pentachlorophenol in aqueous solution. They also reported that TiO2 NTs under UV irradiation promoted higher mineralization than a conventional sol-gel film electrode.
The photoelectrocatalytic degradation of pesticides has been performed by TiO2 thin films. Philippidis and colleagues achieved 82% of degradation of the pharmaceutical compound imidacloprid using Ti/TiO2 electrodes prepared by the immobilization of P25 powder onto Ti substrate. The degradation efficiency increased with increased applied potential, following the ﬁrst-order kinetics model after three hours of treatment. The method was proved to be more efficient than photocatalysis (63% removal) and photolysis (5% removal) operating under UV irradiation . The pesticide Dipterex has been removed by using TiO2 as a photoanode, prepared by a sol-gel method depositing over a nickel net. The method promoted a chemical oxygen demand (COD) removal and organophosphorous conversion of up to 82.6% and 83.5%, respectively, after 2 h of treatment under UV light .
The incomplete reduction of azo dyes and nitroaromatic compounds can usually promote aromatic amine formation, which can be released into the environment as potential carcinogens. This has been reported in drinking water treatment plants . The use of TiO2 NTs as photoanodes was proposed by Cardoso and colleagues. The method is efficient since it promotes the complete degradation and mineralization of 4,4-oxydianiline after 2 h of photoelectrocatalytic treatment under UV irradiation .
The PEC degradation of 4,4‘-dibromobiphenyl used in flame retardants in the textile, and electronic industries, and in additives in plastics, has been performed using TiO2 NTs as photoanodes. This class of compounds is described as toxic to human health and the environment. The photoelectrocatalytic process was more efﬁcient than the photocatalytic and electrolytic process alone. Different anodes were compared: TiO2, Zr/TiO2 and Zr,N/TiO2 NTs. The photoelectrocatalytic efficiency was signiﬁcantly affected by the properties of the catalysts and the best performance was observed with TiO2 doped with nitrogen and zirconium, as it had a higher photocurrent under UV irradiation by a 125 W mercury lamp .
Biological microorganisms can cause the contamination of water by spreading potential pathogens. TiO2 nanotube arrays and Ag-loaded TiO2 NTs have been employed in the disinfection of water containing
Endocrine disrupters have been reported as a class of compounds which can mimic or inhibit the natural actions of the endocrine system in animals and humans, such as synthesis, secretion, transport and binding. They can be either natural or synthetic compounds that come from different sources, such as pharmaceutical compounds, personal care products, disinfectionproducts and surfactants . The literature  reports the removal of Bisphenol A from wastewater using TiO2 NTs in a photoelectrocatalytic oxidation process under UV light and applied potential of +1.2 V. The removal was confirmed by HPLC/DAD analysis. The degradation of carbamazepine has been conducted with Ti/TiO2 electrodes prepared by pulsed laser deposition. After 120 min of treatment, 73.5% pollutant removal was achieved, and 21.2% mineralization. Although complete degradation was not achieved the by-products were not toxic in the presence of
Different activities in the textile, paper, pharmaceutical, leather and food industries, among others, release a huge amount of dyes in effluents that can reach drinking water treatment plants if they are not appropriately treated. There are serious concerns over these compounds – many are potential carcinogens, or have xenobiotic or toxic properties that can harm the environment and living organisms .
The PEC oxidation of methyl orange , methylene blue  and rhodamine B  dyes has been reported. The photoelectrochemical method promoted 100% discoloration and high reduction of the toxicity of dispersed and indigoid organic dyes [185-187].
Recently, the main target of PEC studies has been the visible light activation of materials . The relevance of reactors for photoelectrocatalytic treatment has also been described. It has been mentioned that the use of solar cells to supply the energy in PEC systems could reduce the cost of batch reactors by making it unnecessary to purchase electricity –electricity costshave been pointed out as the main disadvantage of this process .
For hydrogen production, a lot of photocatalysts have been studied in the literature,though mainly TiO2 and modified TiO2. Lianos described the use of TiO2 supported on ITO and FTO and TiO2 doped with N, C and S as well as the use of photocatalysts combined with noble metals such as Pt, Pd and Au and the coupled semiconductors TiO2/SnO2, TiO2/WO3, TiO2/RuO2, TiO2/V2O5 in an attempt to use visible light irradiation . Pure TiO2 nanotube arrays have also been described in photoelectrochemical water splitting and simultaneous degradation of methylene blue . The PEC experiments were conducted using an artiﬁcial sunlight simulator. The higher photoconversion efﬁciency for hydrogen generation and the degradation efficiency of MB were attributed to the better electron transfer process observed for two-step TiO2 NTs over one-step TiO2 NTs. CdS/TiO2 nanotubes for photoelectrochemical hydrogen production have also been described: the doped material presented a better performance in the H2 generation rate than the pure TiO2 NTs under solar light illumination .
Zhao and colleagues carried out simultaneous photoelectrochemical destruction. They obtained contaminant and nickel recovery on the cathode. The deposition of TiO2 film was performed by dip-coating . Paschoal and colleagues promoted the photoelectrochemical reduction of bromate under Ti/TiO2 coated as a photocathode. Photoelectrocatalytic reduction of BrO3− to Br− can reach 70% at neutral pH under biased potential of −0.20 V after 75 minutes of treatment . Table2 shows a summary of the selected studiesusingdopedand undoped TiO2 photoanodes used in photoelectrocatalytic applications.
|TiO2 NTs||Organics degradation||, , , , , , |
|Water splitting||[194, 195], , |
|TiO2 NTs coupled with other semiconductors||Organics degradation||, , , , , , |
|Anion-doped TiO2 NTs||Organics degradation||, , , , |
|Water splitting||, , |
|Cation-doped TiO2 NTs||Organics degradation||, , |
|TiO2 NTs coupled with noble metals||Organics degradation||, |
|Water splitting||, |
|Disinfection||, , |
|TiO2 thin film||Organics degradation||, , |
|Water splitting||, |
|Doped TiO2 thin film||Organics degradation||, , , , |
5.2. Application of doped, decorated and composite of TiO2 nanomaterials PEC
N-doped TiO2 coatings prepared by radiofrequency magnetron sputtering has been employed on the degradation of the antibiotic chlortetracycline under 0.6 A of current intensity and solar simulator irradiation during 180 min, leading to 99% degradation. This is more efficient than pure Ti/TiO2. This process has also shown to be efficient in the inactivation of faecal coliform, which is an indicator pathogen . Wu and Zhang  prepared nitrogen-doped double-wall TiO2 NTs, which under simulated solar light presented a high photoelectrochemical water splitting performance due to the high surface areas and absorbance in the visible light region. Sun et al.  prepared N-doped TiO2 NTs, which presented better efficiency in Rhodamine B PEC degradation.
Boron-doped TiO2 NTs have also been studied as photoanodes prepared by chemical vapour deposition. The electrode was applied in the degradation of methyl orange dye under visible light irradiation promoting 100% discoloration under applied potential of +2.0 V and UV irradiation . In the studies by Lu and colleagues  and Li et al.  boron-doped TiO2 NTs were prepared and applied in the PEC degradation of atrazine and phenol, respectively.
TiO2 has been doped with nickel and used as a photocatalyst in the degradation of Acid Red 88 dye. The photoanode powder was prepared by the sol-gel method and 95% COD and TOC removal was obtained after 35 min of treatment under UV and solar irradiation. The colour removal was 72% for photocatalytic treatment and 97% for photoelectrocatalytic treatment under +1.6 V . Gong and colleagues prepared W-doped TiO2 NTs and applied these in simultaneous Rhodamine B degradation and production of hydrogen ; tungsten-doped TiO2 films were also applied in dodecyl-benzenesulfonate removal by PEC .
Arrays of porous iron-doped TiO2 as photoelectrocatalyst with controllable pore size have been synthesized by using polystyrene spheres as templates. It was found that photoelectrochemical hydrogen generation was favoured by a shift in the ﬂat-band potential from −0.38V to −0.55 V vs. SCE and an increase of photocurrent by 80% .
Pt-deposited TiO2 photoanodes have been prepared by a sol-gel method, where the amount of Pt was shown to interfere with the photoelectrochemical response for glucose oxidation. The increased Pt lowered the photocurrent but the overall oxidation efficiency of the PEC process was better than the PC process, for both TiO2 and Pt-TiO2 ﬁlms . Ye et al.  prepared TiO2 NTs sensitized by palladium quantum dots (Pd QDs), which exhibit highly efficient photoelectrocatalytic hydrogen generation. Zhang and colleagues  prepared TiO2 NTs loaded with Pd nanoparticles, and the PEC activity was investigated with degradation of methylene blue and Rhodamine B.
CdS-ZnS/TiO2 composite material has been investigated in the production of electricity. The band-gap energy can be tuned between that of ZnS (3.5 eV) and that of CdS (2.3 eV) by varying Cd (or Zn) content. Photocatalytic and photoelectrocatalytic processes in basic electrolyte with ethanol as a sacrificial electron donor was also investigated. The performance of CdS-ZnS, Pt/(CdS-ZnS), Pt/(CdS-ZnS)/TiO2 and Pt/TiO2 photoanodes was compared and 75% CdS–25% ZnS over pure TiO2 presented better electrocatalyst effect than 100% CdS over TiO2 . CdS nano-crystallites-decorated TiO2 nanotube array photoelectrodes were prepared through anodization and electrodeposition strategies. Enhancement of photoelectrocatalytic degradation of Rhodamine B was achieved under Xenon light irradiation .
Georgieva and colleagues described the use of bicomponent anodes of TiO2/WO3 for the photoelectrocatalytic oxidation of organic species. WO3 is a promising additive for TiO2 since it modifies its photochemical properties in a favourable manner, both with respect to reduced recombination and visible light activity because of the lower band-gap energy. The coupling of semiconductor oxides leads to electron and hole transfer between the two materials in opposite directions, thus limiting recombination of the photogenerated species in the same material . These materials have been employed in the degradation of 2,3-dichlorophenol under visible light irradiation , the removal of the hair dye Basic Red 51 under UV and visible light source  and the PEC oxidation of indigo carmine dye .
The use of heterojunctions was studied by Christensen and colleagues, who conducted the PEC degradation of
The PEC degradation of flame-retardants has been described under macroporous silicon/graphene (MPSi/Gr) heterostructure. The experiments were conducted under visible light irradiation and compared to photocatalytic degradation. The photoelectrocatalytic degradation five times faster than PC degradation .
CdTe nanotubes have been produced by using ZnO as a template on an ITO surface. These were then used with the photoelectrocatalytic degradation of the Acid Blue 80 dye. This study provided a good strategy for the design of visible light-responsive photocatalysts that can be recycled and possess high efﬁciency, extremely low mass and high chemical stability .
The PEC remediation of 2,4-dichlorophenol by visible-light-enhanced WO3 has also been described. The degradation process achieved 74% pollutant removal after a period of 24 hours, monitored by both chemical analysis and a bacterial biosensor (
For hydrogen production, photocatalysts reported in the literature apart from TiO2 include ZnO, Fe2O3, and SrTiO3, which has the energy levels necessary to create active radical species that could efficiently carry out photodegradation process . Under visible light irradiation some n-type materials have been described: nanoporous WO3, α-Fe2O3 or haematite and nanocrystalline BiVO4 .
The Cu/Cu2O system as photocathode has been described in relation to nitrate removal under UV irradiation and biased potential. The material was prepared by electrodeposition and long-term stability was achieved. 93% nitrate removal was achieved after 75 min under the best experimental conditions. Nitrate reduction on Cu/Cu2O photoelectrodes occurs in the cathodic compartment cell via electrons generated under UV irradiation, as expected for a p-type electrode, leading to 42% of remaining nitrite and 52% gaseous nitrogen derived, respectively .
Zanoni and colleagues employed TiO2 NTs in the photoelectrocatalytic oxidation of an organic synthetic dye (reactive black 5) and the simultaneous hydrogen generation. The photoanode was irradiated with UV light and biased at +1.0 V. Complete dye degradation and 72% mineralization was achieved after 2 h of treatment. The estimated overall hydrogen generation was around 44%, which corresponds to 0.6 mL cm−2 .
6. Final remarks
Photoelectrocatalysis is an emerging field with many applications, such as organics oxidation, inorganics reduction, biological materials and production of electricity and hydrogen.
The technique could be described as a multidisciplinary field, where the basic concept is the irradiation by light (hν≥Eg) of the semiconductor surface. There is the generation of electron/hole pairs (e−/h+) by the promotion of an electron from the valence band (lower energy level) to the conduction band (higher energy level). The electrons are forwarded to the counter electrode under positive anodic bias (n-type) in order to minimize the recombination of these pairs due to the short life-time. When immersed in electrolyte the adsorbed water molecules and/or hydroxyl ions react with the holes on the valence band to generate hydroxyl radicals (●OH), which are a powerful oxidizing agent.
Titanium dioxide (TiO2) is a classic example of an n-type semiconductor widely used as a catalyst for heterogeneous photocatalysis and photoelectrochemical applications. It has received a great deal of attention due to its good chemical and thermal stability, non-toxicity, low cost, high photoactivity and other advantageous properties. It is a typical n-type semiconductor mainly composed of anatase and rutile allotropic forms whose band-gap energy is 3.2 and 3.0 eV, respectively. The anatase phase is the desired form as it is more photoactive than the other forms.
The degradation of organic pollutants by photoelectrocatalysis has been described in the literature as one of the most effective treatments among advanced oxidative processes (AOPs) in the oxidation of recalcitrant compounds, as they are harmful to the environment and human health. The contamination of water is an increasing concern because pollutants can accumulate in the environment and are mutagenic and genotoxic.
The architecture of nanostructures used in the electrode construction has deeply influenced the results of PEC. Nanotube, nanowire, nanofibre, nanorod, and nanowall morphologies can be easily obtained by electrochemical methods. These kinds of nanostructures have improved efficiently organic contaminants degradation, especially due to their high surface area and ability to minimize charge recombination. The use of nanotube arrays has received a great deal of attention especially because it is the structure with the highest surface area/geometric area ratio; moreover, it is of a highly oriented and organized nature, leading to efficient charge transport as it has a unique and effective direct interfacial direction, decreasing the charge recombination effect. Among all TiO2 NTs preparation routes, the electrochemical anodization method presents the greatest advantages, since they are cheaper, simpler and allow precise control of dimensions, presenting highly ordered nanotube arrays. The first generation of nanotube materials applied in PEC materials were obtained in aqueous solutions with the addition of HNO3, H2SO4 and H3PO4 to HF acid as electrolyte. The second generation of nanotube arrays was obtained in buffered electrolytes. Aiming for better quality and performance, the third generation was obtained in organic medium as ethylene glycol, diethylene glycol, glycerol and NH4F. Non-fluoride-based electrolytes are classified as the fourth generation, where HCl, H2O2 and a combination of both are used as electrolyte. Nanotube array photoanodes have presented good results on the water decontamination of organic contaminants and also water disinfection.
Recently, studies have addressed the challenge of obtaining PEC materials which can be activated by visible light, with the aim of using solar light to promote photoactivation, not only to reduce cost but also to establish an environmentally friendly method. For this purpose, different strategies are discussed in the literature to improve photoactivity and shift the PEC material absorption to the visible region, such as the use of photoanodes decorated with Ag and Pt, or combinations of semiconductors like ZnO/TiO2, CdS/TiO2, WO3/TiO2 in order to obtain composite and bicomponent materials; doping with metals (Fe, Mn, Cr), non-metals (B, C, Si) and co-doping (N-F, N-C) has also been thoroughly described.
Therefore, the use of TiO2 and other materials is of huge relevance to photoelectrocatalysis applied to water treatment, and the success of photoanodes and photocathodes depends on the synthesis process and a better understanding of materials’ properties.
The importance of photoelectrocatalysis has been discussed, with emphasis on recent advances in TiO2-based materials and strategies of electrochemical synthesis and modification. Currently, TiO2 nanotube arrays occupy a prominent position. These can be prepared by electrochemical anodization of titanium plates in fluoride-containing electrolytes. In the search for catalysts that can be photoactivated with visible radiation, doping or modification of these materials can be easily performed by electrochemical techniques. The use of these photocatalysts immobilized on conducting substrates employed in photoelectrochemical reactors is a viable strategy for increasing the efficiency of water splitting or to promote efficient degradation of organic compounds.