Nanostructured Layers for Catalytic ApplicationsView this Special Issue
in the Surface of Titanium Dioxide: Generation, Properties and Photocatalytic Application
Titanium dioxide (TiO2) is the most investigated crystalline oxide in the surface science of metal oxides. Its physical and chemical properties are dominantly determined by its surface condition. Ti3+ surface defect (TSD) is one of the most important surface defects in TiO2. According to publications by other groups and the studies carried out in our laboratory, the formation mechanism of TSD is proposed. The generation, properties, and photocatalytic application of TSD are overviewed; the recent exploration of TSD is summarized, analyzed, and evaluated as well in this paper.
Titanium dioxide (TiO2) has been studied extensively in the field of surface science due to the wide range of its applications and the expectation for insights into surface properties on the fundamental level . These studies have been motivated in part by the discovery that TiO2 is a photocatalyst with relatively high efficiency for the decomposition of water [2–8] and the degradation of organic species [9–19]. The photocatalytic activity of TiO2 is often dependent on the nature and density of surface defect sites. Liu et al.  have shown that the dominant defects in TiO2 surfaces are Ti3+ defects and oxygen vacancies. Much work has been focused on the TSD and the possible effect of the TSD in TiO2. It has been shown that the chemistry of stoichiometric TiO2 surfaces differs markedly from surfaces containing Ti3+. Ti3+ is considered to be an important reactive agent for many adsorbates; hence many surface reactions are influenced by these point defects . Sirisuk et al.  reported that Ti3+ sites play an essential role in photocatalytic process over TiO2 photocatalyst. Sakai and coworkers [23–25] reported that TSD in anatase is an important parameter controlling the hydrophilic property. Also, as a support with high-surface area of anatase, TSD plays a significant role in enhancing the dispersion and stability of supported metal such as gold cluster and cobalt via the strong interaction between defect site and metal cluster . Furthermore, for practical applications, pristine TiO2 is not a good candidate, because it is only active under ultraviolet (UV) irradiation in order to overcome the band gap of above 3.0?eV. Reduced TiO2 (TiO2-x ), which contains the Ti3+, however, has been demonstrated to exhibit visible light absorption [27–31]. It is therefore highly important to have a comprehensive understanding of the methods and the techniques of Ti3+ generation and monitoring as well as Ti3+ property exploration.
TSD is not hard to be produced because its generation does not need harsh preparation conditions. The reported methods to produce TiO2 containing TSD include UV irradiation [29, 32, 33], heating TiO2 under vacuum [34, 35], thermal annealing to high temperatures (above 500?K) , reducing conditions (C , H2 ), plasma-treating , laser , and high-energy particle (neutron [40–42], Ar+ , electron , or ?-ray ,) bombardment. Averaging spectroscopic techniques like X-ray photoelectron spectroscopy (XPS) [46–48], electron spin (or paramagnetic) resonance (ESR or EPR) [49–59], and temperature-programmed desorption (TPD) [22, 60] have given valuable information for monitoring Ti3+ defects in TiO2 surfaces. In addition, TiO2 with TPD shows not only the properties of acidic oxide but also reducing characteristics. Besides photocatalysis, it can also be applied to gas absorption [61, 62], photochromism [30, 31], biology , and energy storage [28, 64].
Since Ti3+ strongly influences the surface chemistry of TiO2, a detailed picture of TSD may help to understand reactivity and overall material performance in photocatalytic applications. Up to now, the knowledge of TiO2 has been reviewed extensively, but that of TiO2 containing TSD has not yet been mentioned until now. According to the publications by other groups and studies carried out in our laboratory, we will present a general overview of the subject on the generation, properties, and photocatalytic application of Ti3+ in the surface of TiO2.
2. Generation of Ti3+ Surface Defects
2.1. Formation Mechanism of Ti3+ Surface Defects
TSD can be generated by reduction of Ti4+ ions. There are two typic processes for Ti4+ reduction to Ti3+.
One is that a Ti4+ ion receives a photoelectron, which is usually generated due to UV irradiation on TiO2. As shown in Figure 1 , photogenerated electrons and holes are produced in TiO2 under UV irradiation. The electrons can be trapped and tend to reduce Ti4+ cations to Ti3+ state, and the holes oxidize O2- anions for the formation of O- trapped hole or even oxygen gas . The charge transfer steps are as follows:
Another process for Ti4+ reduction to Ti3+ is usually accompanied by a loss of oxygen from the surface of TiO2. TSD can be introduced deliberately by annealing in vacuum condition, thermal treatment under reducing atmosphere (H2, CO), or by bombardment using electron beam, neutron, or ?-ray. In these processes, Ti4+ ions receive electrons from these reducing gases or lattice oxygens which are usually removed from stoichiometric TiO2. Figure 2  shows the EPR intensity of Ti3+ and oxygen vacancies versus H2 treatment temperature during the H2 treatment. Liu et al.  proposed that the interaction between H2 and TiO2 fell into three types. Firstly, hydrogen interacted physically with the adsorbed oxygen on the surface of TiO2 at a temperature below 300°C. Secondly, when the temperature was higher than 300°C, electrons were transferred from the H atoms to the O atoms in the lattice of TiO2. Then, the oxygen vacancies were formed when the O atom left with the H atom in the form of H2O. Thirdly, when the temperature was up to 450°C, the interaction between H2 and TiO2 proceeded more drastically, in which the electrons transferred from oxygen vacancies to Ti4+ ions, and then Ti3+ ions were formed. In this case, the higher the temperature is, the more Ti3+ ions were produced. Furthermore, when the temperature increased to 560°C, more energy was supplied, and the electrons already in the oxygen vacancies were driven away and transferred to Ti4+. This result is in accordance with the EPR data which indicated the intensity of oxygen vacancies decreased and that of Ti3+ increased.
Chen et al.  reported that some Ti4+ ions in the surface of TiO2 were reduced into Ti3+ state by carbon formed from pyrolysis of titanyl organic compounds. The mechanism for this reduction reaction is that the carbon from organic component carbonizing could reduce Ti4+ to Ti3+ at high temperature, which is similar to Liu’s report. However, Chen et al. may hold a wrong idea that TSD is the same as oxygen vacancies. According to the publications by Berger et al.  and Liu et al. , the TSD is quite different from the oxygen vacancies because both TSD and oxygen vacancies can be generated, respectively. Here, we can see that the principal character of the second reduction process different from the first one is that Ti4+ ions are not reduced by photoelectrons generated by UV irradiation, but reduced by electron donators such as H2, C, or lattice oxygen in TiO2.
2.2. Generation Methods for Ti3+ Surface Defects
2.2.1. UV Irradiation
Figure 1 is a scheme for the generation of TSD in anatase TiO2 powder at 140?K and below by UV irradiation. As discussed in Section 2.1, photogenerated electrons and holes are produced by UV radiation in TiO2, and the electrons can be trapped and tend to reduce Ti4+ cations to Ti3+ state. However, it has been shown that only a limited fraction of electrons are actually localized as in Ti3+ state, while the major fraction remains in the conduction band after UV excitation . After discontinuation of UV exposure, the electrons trapped in Ti3+ can be stored for hours when TiO2 particles are electronically isolated and kept at low temperatures. A high quantum yield and much longer life of the Ti3+ ions can be achieved in wet TiO2 gels. Kuznetsov et al.  reported that the quantum yield of Ti3+ as high as 22–25% has been independently obtained from the absorption delay and extinction measurements. These Ti3+ centers can be produced by the wet TiO2 gels under the UV irradiation in the spectral range between 3.25 and 4.4?eV. All photo-induced Ti3+ centers are chemically active and long lived. Under a prolonged UV-laser irradiation, their lifetime can be extremely long and exceeds months at room temperature in the absence of oxygen . The electrons are stored in the gel network as small polarons, Ti3+ centers, whereas the holes are stored in the liquid phase as H+ ions or radicals. Some other research groups also reported the generation of Ti3+ defects by UV irradiation [68, 69].
2.2.2. Annealing or Calcination
Vacuum annealing and calcination are widely used to generate TSD [34–36, 63, 70–72]. Guillemot et al.  reported that low-temperature vacuum annealing could create a controlled number of TSD ranging from low concentration (<3% Ti3+/Ti4+) to high concentration (around 21% Ti3+/Ti4+) at 323 and 573?K, respectively. Figure 3 shows the variation of the Ti3+ to Ti4+ ratio with vacuum annealing temperature. Nevertheless, TSD stability has been ascertained by first UHV annealed at 523?K and stored under different conditions. In fact, in relation to storage conditions, one night under UHV or one week under a laboratory atmosphere, the initial Ti3+/Ti4+ only decreased from 9.1% to 8.7 or 3.4%, respectively.
Xu and coworkers  prepared TiO2 ultrafine particles by the colloid chemical method. They found that as the calcining temperature decreased, the size of TiO2 ultrafine particles decreased, and the contents of TSD and the number of hydroxyl active species increased, which were considered to be essential to the photocatalytic activity of the samples. Huizinga and Prins  reported that the reduction of Pt/TiO2 at 573?K led to the formation of a Ti3+ ESR signal. After reduction at 573?K, 0.3% of the total number of Ti4+ ions in TiO2 was reduced to Ti3+. They found that the reduction of the TiO2 by hydrogen was catalyzed by the platinum. The reduction mechanism has been proposed as follows:
2.2.3. Particles (Electron, Neutron, and ?-Ray) Bombardment
Zhang et al.  prepared TiO2/Si films on silicon substrates by the DC reactive sputtering method. The TiO2/Si structures with different TiO2 film thickness were irradiated by electron beams. It is found that the number of Ti3+ ions increased and Ti4+ ions decreased after the irradiation. The relative abundance can be calculated. The number of Ti3+ increased to 10% and that of Ti4+ decreased from 98% to 90%. This implies that in the transition layer, a fraction of Ti4+ ions turned to Ti3+ ions and the chemical composition changed as follows:
Jun et al.  reported the improvement of the photoactivity of TiO2 which underwent electron beam (EB) treatment (1?MeV) as a function of the absorbed radiation dose (MGy). The radiation-induced effects on the TiO2 crystal structure, for example, change of the Ti3+/Ti4+ ratio, were investigated. As shown in Figure 4, the quantitative analyses of Ti4+ and Ti3+ surface states in EB-treated TiO2 at different radiation doses are presented by XPS data. It shows that a maximum decrease of Ti4+ amount treated at higher EB doses and approximately the same amount of Ti3+ on the surface state in Ti2p up to about 3?MGy is observed. Both Ti2p1/2 and Ti2p3/2 states showed a change in Ti3+ and Ti4+, and the Ti3+ states increased by about 15% as a consequence of EB treatment. The Ti3+ state on TiO2 surface is important because it can play a similar role as observed in TiO2 doped with metal atoms, which can trap the photo-generated electrons and thereafter leave behind unpaired charges to promote photoactivity.
Figure 5(a)  shows the ESR signals of O- anion radicals and Ti3+ cations on TiO2 powder after ?-ray irradiation of 150?MR under N2 atmosphere. Before ?-ray irradiation, no signal could be observed in the ESR signals, while after ?-ray irradiation, the ESR signals attributed to the O- anion radicals and associated Ti3+ were observed. It increased with the increase of the ?-ray irradiation dosage shown in Figure 5(b) .
Huang et al.  reported that upon exposure to ?-radiation, a concentrated TiO2 sol changed from colorless to deep blue with an absorption maximum at 540?nm. The absorption has been assigned to trapped electrons or Ti3+ ions in the solid matrix based on its spectroscopic similarity to the samples irradiated with UV light. The origin of the trapped electrons during ?-ray irradiation may be traced to a series of reducing species produced by the high-energy electrons, which in turn are the direct result of ?-ray irradiation. The absorption intensity is linearly related to the duration of exposure to ?-ray irradiation.
The reduction effect (from Ti4+ to Ti3+) in rutile induced by neutron irradiation is also reported. Lu et al.  found that Ti4+ ions reduced to Ti3+, Ti2+, and even Ti+ after the neutron irradiation by means of characteristic techniques of UV-VIS-IR, XPS, XRD, and AFM.
TiO2 irradiated by UV light can create TSD with ease. However, a special gel structure, which has a poor chemical stability, is definitely needed for a high quantum yield of TSD. Thus, this method is subjected to the certain restriction in the practical photocatalytic application. Annealing and calcination can effectively control concentration of TSD and morphology of TiO2. Thus, both of them are extensively studied and used although they are still tedious, for example, high temperature, high vacuum, and atmosphere frequently needed. The bombardment method is atypical, but important.
3. Properties of Ti3+ Surface Defects
3.1. Structural Properties of Ti3+ Surface Defects
Lu et al.  described the structure of ideal rutile single crystal and rutile with insufficient oxygen. In the ideal rutile single crystal, each titanium ion is located in the center of oxygen octahedron. The symmetry of the oxygen octahedron is rhombic symmetry (). The parallel () and vertical () Ti4+-O2- bonding lengths are 1.988 and 1.944?Å (see Figure 6), and the bonding angle in the vertical plane is about 80.83°. When a host Ti4+ ion is changed to Ti3+ ion, the local electrostatic balance is broken, and an OV (oxygen vacancy) should be introduced because of charge compensation. In order to determine the position of the OV as well as the local structure of the [Ti3+-OV] center in the reduced rutile crystal, the structure model is established, and the optical spectra is calculated by using the crystal field theory. In this structure model, the axial OV is located on the nearest position of central Ti3+ ion. When an OV appears on the nearest position of central Ti3+ ion along the direction of the oxygen octahedron, the local symmetry would change from to C2v. Since the effective charge of the OV is positive, the central Ti3+ is expected to shift away from the OV along the direction by an amount due to the electrostatic repulsion. Similarly, the four O2- ions on the vertical plane would also move towards the OV by an amount due to the electrostatic attraction (see Figure 6). Considering the much larger distance between the OV and the remaining O2- ion of the direction, the displacement of the only axial O2- ion may be much smaller than or and can be omitted for simplicity.
This model is in accordance with those proposed by Weyl and Forland  and Breckenridge and Hosler . In their models, Ti3+ ions are also on sites adjacent to the oxygen vacancy. In addition, Weyl and coworkers have drawn a schematic picture of the structure of Ti3+ ions (Figure 7). Figure 7  shows schematically what happens when a crystal of Ti4+ loses one oxygen atom, and the two electrons of the O2- ion change two Ti4+ to Ti3+ ions. These Ti3+ ions are strongly polarized and a state is assumed, in which its extreme can be described as two Ti4+ ions and two additional electrons. The latter may assume the position of the missing O2- ion. The intensive light absorption in such partly reduced crystals is the result of the strong distortion of the outer orbitals of the Ti3+ ions. Ti3+ ions in this state absorb light more intensively than undeformed Ti3+ ions.
3.2. Optical Properties of Ti3+ Surface Defects
Optical properties of TSD lie in two aspects. One is that Ti3+ ions are the origin of the blue coloration or coloration center [27, 58, 64, 74] and the other is that Ti3+ species has a characteristic visible absorption spectrum [27–31, 75–78]. Ookubo et al.  developed a method that examined the quantitative relation between degree of blue coloration and the concentration of the Ti3+ ions. The method has been proposed based on the phenomenon that a Ti3+ salt was hydrolyzed slowly with the presence of urea in an aqueous solution. With this method, a blue TiO2 was obtained, and the degree of the coloration has been controlled by changing the period of refluxing the solution in the synthetic procedure. In addition, numerous studies have verified that the Ti3+ species induced oxygen vacancy states between the valence and the conduction bands, which would contribute to the visible response. Xiong et al.  identified the conversion of Ti4+ to Ti3+ in the TiO2/Cu2O bilayer film after the visible-light irradiation by the UV-vis diffuse reflectance measurement. Because Ti4+ has no response to visible light while Ti3+ does, the presence of Ti3+ ions leads to a weaker absorbance of the bilayer film in the short wavelength range while stronger in the long wavelength range. It was also found that the transparent TiO2 film turned blue under the irradiation (Figure 8(a)).
Bityurin et al.  proposed several models fitting the experimental data on the kinetics of UV laser-induced darkening in TiO2 gels. Samples were exposed to various average laser intensities in the range of 0.05~0.5?W/cm2 by using beam attenuators. After irradiation of a sample by a laser beam, the modified area looked like a dark spot on the transparent material surface (Figure 8(b)). The dark color spot was believed to be caused by the transformation of Ti4+ to Ti3+. TiO2 sol was also shown to change from colorless to deep blue with an absorption maximum at 540?nm upon exposure to ?-radiation . The absorption has been assigned to trapped electrons or Ti3+ ions in the solid matrix based on its spectroscopic similarity to the samples irradiated with UV light.
3.3. XPS Properties of Ti3+ Surface Defects
According to the standard binding energy of Ti2p3/2 in TiO2, that for Ti3+ is usually located at 457.7?eV and that for Ti4+ is at 459.5?eV. The O (1?s) binding energy for TiO2 is 529.3?eV . Generally speaking, the binding energy peak of Ti2p in stoichiometry TiO2 is not broad and no shoulder peak. However, that for TiO2 containing Ti3+ in the surface is usually tortured and turns into much broader. In this case, a fitting skill such as Gaussian of the Gauss-Lorentzian fitting can be taken to obtain the information of Ti3+ and Ti4+ in the surface of reduced TiO2. The relative contents of Ti3+ in the surface of reduced TiO2 can also be obtained according to the XPS peak areas.
Greenlief et al.  reported the Ti2p binding energy data of TiO2 thin films with different monolayers on polycrystalline Pt after vacuum annealing by XPS technique. The measured binding energy 458.9?eV was the Ti2p3/2 peak for the TiO2 thin films with coverages more than 3 monolayers (1 monolayer is assumed to be an evenly dispersed film 2.6?Å thick). The O (1?s) binding energy for these TiO2 thin films were 530.3?eV. The O/Ti ratio (as determined by XPS peak areas) was stoichiometric (2.0) for these TiO2 thin films. After vacuum annealing, the ratio dropped to between 1.0 and 1.5, and a mixture of Ti3+ and Ti4+ was obtained. The binding energy 458.9?eV was observed and attributed to Ti3+ state.
Recently, our group  found that the TSD can be generated in TiO2/Cu2O bilayer film under visible-light irradiation. By means of XPS (Figure 9), it can be seen that the binding energy of Ti2p3/2 in bilayer film without the irradiation was 458.7?eV (Figure 9(a)). It is difficult to simulate this peak since the peak was not so broad, and there was no shoulder peak. The binding energy of Ti2p3/2 in the bilayer film after the irradiation was 458.1?eV. Compared with the peak for Ti2p3/2 in the same bilayer film without the irradiation, the one after the irradiation was much broader and there was 0.6?eV shift. This Ti2p3/2 spectrum can be simulated with Gaussian simulation. The peaks at 457.8?eV and at 459.2?eV were attributed to Ti3+ and Ti4+, respectively (Figure 9(b)). The result is in accordance with the reported data . The content of Ti3+ ions was as high as 74% in TiO2/Cu2O bilayer film after visible-light irradiation according to the comparison of the peaks area.
3.4. EPR Properties of Ti3+ Surface Defects
EPR is a highly sensitive technique which allows investigation of paramagnetic species having one or more unpaired electrons either in the bulk or at the surface of various solids. EPR technique is already widely used by many researchers to characterize TSD. The data of EPR signals assigned to Ti3+ in some published paper are shown in Table 1.
Tijana et al. also presented some published data of EPR signals of Ti3+ in the book edited by Kokorin and Bahnemann . They proposed that Ti3+ in the surface and that in the bulk of TiO2 can be distinguished by the differences of their EPR parameters. The values of the -factors for surface Ti3+ particles are significantly lower than those usually found in bulk TiO2. Similar result was also reported by Nakaoka and Nosaka . There is a small change in the -values of the axially symmetrical -tensor (previously identified as interstitial interior Ti3+ ions) upon the sample heating ( = 1.957, = 1.990 for untreated sample and = 1.961, = 1.992 for sample heated at 700°C for 5?h). They attributed the first signal to the photo-generated electron trapped on the surface Ti3+ ions, and the second one to the inner Ti3+ ions.
In addition, Tijana et al. found the difference of EPR signal shapes for surface and inner Ti3+ ions. According to the characters of EPR signals, they proposed two general types of traps (Ti3+ ions): internal, having a narrow axially symmetric EPR signal, and surface, with broad EPR lines. Magnetic resonance parameters of the internal, interstitial (inner) Ti3+ ions slightly vary due to the different delocalization of the unpaired electron density and symmetry of the local surroundings (presence of vacancies and impurities in the nearest coordination sphere). It also happens for the surface electron trap. In this case, -values and the linewidth of the surface Ti3+ ions mainly depend upon surface modification.
3.5. Temperature Program Reduction (TPD)
Temperature-programmed desorption using carbon dioxide (CO2-TPD) as a probe can be employed to monitor surface defects in TiO2 [22, 60, 84, 92]. CO2-TPD analysis indicates a signal of CO2, which desorbs from the TiO2 surface between the ranges of 123 and 253?K and results in two peaks: one peak at about 170?K attributed to CO2 molecules bound to a regular five-coordinate Ti4+ site which was a perfect TiO2 structure, and another peak at about 200?K corresponding to CO2 molecules bound to Ti3+ which was a defected TiO2 structure [60, 84] (Figure 10). The positions of the peaks are shifted slightly from those reported by Sirisuk and coworkers . Thompson et al.  have observed the characteristic two-step desorption process which is indicative of the presence of both Ti3+ defects and nondefective TiO2 sites. Moreover, they have measured the activation energy for both of the desorption peaks of CO2. For CO2 desorption from the fully oxidized (or perfected) surface, where the CO2 is bound to Ti4+ sites, a zero-coverage activation energy for CO2 desorption of 48.5?kJ/mol is measured. It is in approximate agreement with theoretical calculations value of 65?kJ/mol . For CO2 desorption from the Ti3+ sites produced by annealing TiO2 (110) in vacuum, a zero-coverage activation energy of ~54?kJ/mol was measured. This indicates that Ti3+ sites bind CO2 slightly more strongly (TPD peak at 200?K) than the fivefold coordinated Ti4+ sites (TPD peak at 170?K) do.
4. Photocatalytic Application
4.1. Photocatalytic Reaction Mechanism of Ti3+ Surface Defects
The photocatalytic reaction mechanisms are widely studied. The principle of TiO2 photocatalytic reaction is straightforward. Upon absorption of photons with energy larger than the band gap of semiconductor materials, electrons are excited from the valence band to the conduction band, producing electron-hole pairs. These charge carriers migrate to the surface and react with the chemicals adsorbed on the surface to decompose these chemicals. This photodecomposition process usually involves one or more of radicals or intermediate species such as , O2-, H2O2, or O2, which play important roles in the photocatalytic reaction. The photocatalytic activity of a semiconductor is largely controlled by (i) the light absorption properties, for example, light absorption spectrum and coefficient, (ii) reduction and oxidation rates on the surface by the electron and hole, (iii) and the electron-hole recombination rate .
As for TiO2 containing TSD, its photocatalytic activity is definitely dominated by TSD. TSD improves the photocatalytic activity of pure TiO2 from the following two aspects: (i) it extends the photoresponse of TiO2 from UV to visible light region, which leads to visible-light photocatalytic activity; (ii) it provides important reactive agents for many adsorbates and results in the reduction of an electrohole pair recombination rate .
Park et al.  proposed the mechanism of TSD participating in photocatalytic reaction. When the TiO2 photocatalysts were irradiated, e-/h+ pairs were formed. In the absence of the electron and hole scavengers, most of them recombined with each other within a few nanoseconds. If the scavengers or surface defects were present to trap the electron or hole, e-/h+ recombinations could be prevented, and the subsequent reactions caused by the electrons and holes were dramatically enhanced. In this case, electrons donors reacted with holes. The electron can be trapped by Ti4+ to generate an isolated Ti3+ ion. In the presence of O2, the Ti3+ sites easily react with O2, leading to the formation of radicals such as , , and . The charge transfer processes may be as follows:
4.2. Photocatalytic Application of Ti3+ Surface Defects
4.2.1. Ion-Doped TiO2 Photocatalyst
A great deal of research has been focused on ion-doped TiO2 with both transition metal and nonmetal impurities in order to lower the threshold energy and improve photocatalytic activity. Doping with transition metals has shown both positive and negative effects. Indeed, these metal-doped photocatalysts have been demonstrated to suffer from thermal instability, and metal centers acting as electron traps have reduced the photocatalytic efficiency .
Stimulated by the report of Asahi et al. in 2001 , there has been an explosion of interest in TiO2 doping with nonmetal ions, N, C, S, B, especially with nitrogen. Livraghi et al.  reported the origin of photocatalytic activity of N-doped TiO2 under visible light. As shown in Figure 11, the material contains single-atom nitrogen impurities that form either diamagnetic () or paramagnetic () centers. Both types of Nb centers give rise to localized states in the band gap of the oxide. The relative abundance of these species depends on the oxidation state of the solid since upon reduction, electron transfer from Ti3+ ions to Nb results in the formation of Ti4+ and . The presence of Ti3+ ions is helpful to form the N paramagnetic centers at the expense of the Ti3+ ions oxidized to Ti4+. The photocatalytic activity of the N-doped TiO2 catalyst was thought to be the synergistic effect of nitrogen and Ti3+ species [76, 97].
Sun et al.  reported that the N-doped TiO2 catalyst showed higher photocatalytic activity for degradation of 4-CP than pure TiO2 under not only visible but also UV irradiation due to the presence of Ti3+. Qin et al.  demonstrated that the enhancement of methyl orange and 2-mercaptobenzothiazole photodegradation using the N-doped TiO2 catalysts is mainly involved in the efficient separation of electron-hole pairs owing to the presence of Ti3+ and the improvement of the organic substrate adsorption in catalysts suspension and optical response in visible-light region. However, it is found that excessive Ti3+ acted as a recombination center for holes and electrons, which is the reason for an optimal content of Ti3+ in the nitrogen-doped TiO2. The N-doped TiO2 with N/Ti proportioning of 20?mol% calcined at 400°C exhibited the highest visible-light activity. Therefore, the presence and optimal content of Ti3+ might be the critical factors for the improvement of the photoactivity.
Carban-doped TiO2 has also shown much better photocatalytic activity due to the presence of TSD [37, 66, 82, 88, 92]. Chen et al.  reported that the C from pyrolyzing process of titanyl organic compounds could reduce the part of Ti4+ in the surface of TiO2 into Ti3+, and some OH groups were formed on crystal surface with the presence of water from pyrolysis. This allowed TiO2 to possess abundant TSD, which acted as the active center for both photocatalytic reaction and matching OH at its surface. When Ti3+/OH ratio in the surface of nanometer crystalline TiO2 approaches 1, there is more effective photocatalytic activity. Li et al.  reported that the carbon-doped TiO2 with high surface area and good crystallinity showed an obvious enlarged range of absorption up to 700?nm and had much better photocatalytic activity for gas phase photo-oxidation of benzene under artificial solar light than pure TiO2. The visible-light photocatalytic activity is ascribed to the presence of oxygen vacancy state because of the formation of Ti3+ species between the valence and the conduction bands in the TiO2 band structure, which results in the as-synthesized carbon-doped TiO2 responsive to the visible light.
4.2.2. Self-Doped TiO2 Photocatalyst
Here, self-doped TiO2 photocatalyst can be interpreted as Ti3+-doped TiO2, in which nothing but Ti and O ions exists. As discussed in Section 2.2, self-doped TiO2 photocatalyst can be produced by UV irradiation, heating under vacuum, thermal annealing to high temperatures (above 500?K), reducing conditions (C, H2), plasma treating, laser, and high-energy particle (neutron, Ar+, electron, or ?-ray,) bombardment. Abundant presented data and numerous valuable conclusions indicate that TSD in self-doped TiO2 photocatalysts is responsible for the enhancement of photocatalytic activity [20, 34–36, 38–40, 42–44, 70, 83, 89, 99–102].
In addition, our group reported the high photocatalytic degradation of methylene blue by TiO2/Cu2O composite film  and photocatalytic water splitting by TiO2/Cu2O bilayer film owing to the presence of Ti3+ . For photocatalytic degradation reaction, electrons excited from TiO2/Cu2O composite film under visible light were transferred from the conduction band of Cu2O to that of TiO2. The formed intermediate state of Ti3+ ion was observed by X-ray photoelectron spectroscopy (XPS) on the TiO2/Cu2O composite film. Additionally, the accumulated electrons in the conduction band of TiO2 were transferred to oxygen on the TiO2 surface for the formation of O2- or , which combines with H+ to form H2O2. The evolved H2O2 with FeSO4 and EDTA forms Fenton reagent to degrade methylene blue. With regard to photocatalytic water splitting process, the photogenerated electrons from the conduction band of Cu2O were captured by Ti4+ ions in TiO2, and Ti4+ ions were further reduced to Ti3+ ions. The Ti3+ ions have a long lifetime and bear the photogenerated electrons as a form of energy. The electrons trapped in Ti3+ ions as stored energy lead to evolve H2 from H2O. The electron transfer processes may be as follows:
5. Concluding Remarks
An overview on current literature on the subject of TSD in TiO2 has been provided here, but the works that have been done and the progresses that have been achieved on this subject are far from being completely resolved. While some of these results reviewed here might be turned out to be of mere fundamental interest and irrelevant for the particular environment and applications, some might help to understand the behavior of this material. Since TiO2 is used in so many different fields, TSD would attract more attention. It is expected that this paper will help to link the more fundamental and more applied lines of research on this subject.
This work was financially supported by the Project of Hubei Provincial Department of Education (XD201010406), the Doctor Foundation of Huanggang Normal University. Professor Y. Yu thanks the National Natural Science Foundation of China (no. 20973070), the National Basic Research Program of China (no. 2009CB939704), the Key Project of Ministry of Chinese Education (no. 109116), and self-determined research funds of CCNU from the colleges’ basic research and operation of MOE for financial support.
U. Diebold, “The surface science of titanium dioxide,” Surface Science Reports, vol. 48, no. 5–8, pp. 53–229, 2003.View at: Google Scholar
J. Tang, J. R. Durrant, and D. R. Klug, “Mechanism of photocatalytic water splitting in TiO2. Reaction of water with photoholes, importance of charge carrier dynamics, and evidence for four-hole chemistry,” Journal of the American Chemical Society, vol. 130, no. 42, pp. 13885–13891, 2008.View at: Publisher Site | Google Scholar
A. Wold, “Photocatalytic properties of titanium dioxide (TiO2),” Chemistry Materials, vol. 5, no. 3, pp. 280–283, 1993.View at: Google Scholar
Y. Du and J. Rabani, “The measure of TiO2 photocatalytic efficiency and the comparison of different photocatalytic titania,” The Journal of Physical Chemistry B, vol. 107, no. 43, pp. 11970–11978, 2003.View at: Google Scholar
G. Li, N. M. Dimitrijevic, L. Chen, J. M. Nichols, T. Rajh, and K. A. Gray, “The important role of tetrahedral sites in the phase transformation and photocatalytic activity of TiO2 nanocomposites,” Journal of the American Chemical Society, vol. 130, no. 16, pp. 5402–5403, 2008.View at: Publisher Site | Google Scholar
G. B. Raupp and J. A. Dumesic, “Adsorption of CO, CO2, H2, and H2O on titania surfaces with different oxidation states,” The Journal of Physical Chemistry, vol. 89, no. 24, pp. 5240–5246, 1985.View at: Google Scholar
N. Sakai, R. Wang, A. Fujishima, T. Watanabe, and K. Hashimoto, “Effect of ultrasonic treatment on highly hydrophilic TiO2 surfaces,” Langmuir, vol. 14, no. 20, pp. 5918–5920, 1998.View at: Google Scholar
N. Sakai, A. Fujishima, T. Watanabe, and K. Hashimoto, “Enhancement of the photoinduced hydrophilic conversion rate of TiO2 film electrode surfaces by anodic polarization,” The Journal of Physical Chemistry B, vol. 105, no. 15, pp. 3023–3026, 2001.View at: Google Scholar
N. Sakai, A. Fujishima, T. Watanabe, and K. Hashimoto, “Quantitative evaluation of the photoinduced hydrophilic conversion properties of TiO2 thin film surfaces by the reciprocal of contact angle,” The Journal of Physical Chemistry B, vol. 107, no. 4, pp. 1028–1035, 2003.View at: Publisher Site | Google Scholar
C. Kormann, D. W. Bahnemann, and M. R. Hoffmann, “Preparation and characterization of quantum-size titanium dioxide,” The Journal of Physical Chemistry, vol. 92, no. 18, pp. 5196–5201, 1988.View at: Google Scholar
R. F. Howe and M. Grätzel, “EPR study of hydrated anatase under UV irradiation,” The Journal of Physical Chemistry, vol. 91, no. 14, pp. 3906–3909, 1987.View at: Google Scholar
U. Diebold, J. Lehman, T. Mahmoud et al., “Intrinsic defects on a TiO2(110)(1 × 1) surface and their reaction with oxygen: a scanning tunneling microscopy study,” Surface Science, vol. 411, no. 1-2, pp. 137–153, 1998.View at: Google Scholar
G. Lu, A. Linsebigler, and J. T. Yates, “ defect sites on TiO2(110): production and chemical detection of active sites,” The Journal of Physical Chemistry, vol. 98, no. 45, pp. 11733–11738, 1994.View at: Google Scholar
J. M. Pan, B. L. Maschhoff, U. Diebold, and T. E. Madey, “Interaction of water, oxygen, and hydrogen with TiO2 (110) surface having different defect densities,” Journal of Vacuum Science and Technology A, vol. 10, no. 4, pp. 2470–2476, 1992.View at: Google Scholar
S. Dohshi, M. Anpo, S. Okuda, and T. Kojima, “Effect of γ-ray irradiation on the wettability of TiO2 single crystals,” Topics in Catalysis, vol. 35, no. 3-4, pp. 327–330, 2005.View at: Google Scholar
C. M. Greenlief, J. M. White, C. S. Ko, and R. J. Gorte, “An XPS investigation of TiO2 thin films on polycrystalline Pt,” The Journal of Physical Chemistry, vol. 89, no. 23, pp. 5025–5028, 1985.View at: Google Scholar
G. B. Hoflund, H. L. Yin, A. L. Grogan et al., “Surface study of the oxidative and reductive properties of TiO2(001),” Langmuir, vol. 4, no. 2, pp. 346–350, 1988.View at: Google Scholar
R. F. Howe and M. Grätzel, “EPR observation of trapped electrons in colloidal TiO2,” The Journal of Physical Chemistry, vol. 89, no. 21, pp. 4495–4499, 1985.View at: Google Scholar
E. Abi Aad and A. Aboukaïs, “Characterisation by EPR spectroscopy,” Catalysis Today, vol. 56, no. 4, pp. 371–378, 2000.View at: Google Scholar
S. J. DeCanio, T. M. Apple, and C. R. Dybowski, “Electron spin resonance study of reduced Rh/TiO2 and TiO2,” The Journal of Physical Chemistry, vol. 87, no. 2, pp. 194–196, 1983.View at: Google Scholar
S. J. DeCanio, J. B. Miller, J. B. Michel, and C. Dybowski, “Electron spin resonance and nuclear magnetic resonance study of the reduction of Rh/TiO2 systems at 298 K,” The Journal of Physical Chemistry, vol. 87, no. 23, pp. 4619–4622, 1983.View at: Google Scholar
M. Grätzel and R. F. Howe, “Electron paramagnetic resonance studies of doped TiO2 colloids,” The Journal of Physical Chemistry, vol. 94, no. 6, pp. 2566–2572, 1990.View at: Google Scholar
J. Dumas, M. Continentino, J. J. Capponi, and J. L. Tholence, “Electron paramagnetic resonance study of the warwickites ,” Solid State Communications, vol. 106, no. 1, pp. 35–38, 1998.View at: Google Scholar
T. M. Salama, H. K. Ebitani, H. Hattori, and H. Kita, “Electron paramagnetic resonance studies of reduced platinum supported on titanium oxide: effect of hydrogen adsorption-desorption on paramagnetic platinum(I) and titanium(III) species,” Chemistry of Materials, vol. 6, no. 1, pp. 21–26, 1994.View at: Google Scholar
A. Ookubo, E. Kanezaki, and K. Ooi, “ESR, XRD, and DRS studies of paramagnetic ions in a colloidal solid of titanium oxide prepared by the hydrolysis of TiCl3,” Langmuir, vol. 6, no. 1, pp. 206–209, 1990.View at: Google Scholar
T. L. Thompson, O. Diwald, and J. T. Yates Jr., “CO2 as a probe for monitoring the surface defects on TiO2(110)-temperature-programmed desorption,” The Journal of Physical Chemistry B, vol. 107, no. 42, pp. 11700–11704, 2003.View at: Google Scholar
G. Lu, A. Linsebigler, and J. T. Yates Jr., “Photooxidation of CH3Cl on TiO2(110): a mechanism not involving H2O,” The Journal of Physical Chemistry, vol. 99, no. 19, pp. 7626–7631, 1995.View at: Google Scholar
A. Linsebigler, G. Lu, and J. T. Yates Jr., “CO photooxidation on TiO2(110),” The Journal of Physical Chemistry, vol. 100, no. 16, pp. 6631–6636, 1996.View at: Google Scholar
A. Fujishima, T. N. Rao, and D. A. Tryk, “Titanium dioxide photocatalysis,” Journal of Photochemistry and Photobiology C, vol. 1, no. 1, pp. 1–21, 2000.View at: Google Scholar
Z. Xu, J. Shang, C. Liu, C. Kang, H. Guo, and Y. Du, “The preparation and characterization of TiO2 ultrafine particles,” Materials Science and Engineering B, vol. 63, no. 3, pp. 211–214, 1999.View at: Google Scholar
T. Huizinga and R. Prins, “Behavior of centers in the low- and high-temperature reduction of Pt/TiO2, studied by ESR,” The Journal of Physical Chemistry, vol. 85, no. 15, pp. 2156–2158, 1981.View at: Google Scholar
W. A. Weyl and T. Forland, “Photochemistry of rutile,” Industrial and Engineering Chemistry, vol. 42, no. 2, pp. 257–263, 1950.View at: Google Scholar
N. Abd El-Shafi and M. M. Morsi, “Optical absorption and infrared studies of some silicate glasses containing titanium,” Journal of Materials Science, vol. 32, no. 19, pp. 5185–5189, 1997.View at: Google Scholar
H. Tachikawa, T. Ichikawa, and H. Yoshida, “Hydration structure of titanium(III) ion. ESR and electron spin-echo study,” Journal of the American Chemical Society, vol. 112, no. 3, pp. 977–982, 1990.View at: Google Scholar
D. R. Park, J. Zhang, K. Ikeue, H. Yamashita, and M. Anpo, “Photocatalytic oxidation of ethylene to CO2 and H2O on ultrafine powdered TiO2 photocatalysts in the presence of O2 and H2O,” Journal of Catalysis, vol. 185, no. 1, pp. 114–119, 1999.View at: Google Scholar
K. Suriye, P. Praserthdam, and B. Jongsomjit, “Control of surface defect on TiO2 nanocrystal using various calcination atmospheres as the first step for surface defect creation and its application in photocatalysis,” Applied Surface Science, vol. 253, no. 8, pp. 3849–3855, 2007.View at: Publisher Site | Google Scholar
W. Kongsuebchart, P. Praserthdam, J. Panpranot, A. Sirisuk, P. Supphasrirongjaroen, and C. Satayaprasert, “Effect of crystallite size on the surface defect of nano-TiO2 prepared via solvothermal synthesis,” Journal of Crystal Growth, vol. 297, no. 1, pp. 234–238, 2006.View at: Publisher Site | Google Scholar
A. I. Kokorin and D. W. Bahnemann, Eds., Chemical Physics of Nanostructured Semiconductors VSP-Brill, Academic Publishers, Utrecht, The Netherlands, 2003.
Y. Nakaoka and Y. Nosaka, “ESR Investigation into the effects of heat treatment and crystal structure on radicals produced over irradiated TiO2 powder,” Journal of Photochemistry and Photobiology A, vol. 110, no. 3, pp. 299–305, 1997.View at: Google Scholar
M. A. Henderson, “Evidence for bicarbonate formation on vacuum annealed TiO2(11O) resulting from a precursor-mediated interaction between CO2 and H2O,” Surface Science, vol. 400, no. 1–3, pp. 203–219, 1998.View at: Google Scholar
S. Yamazaki, S. Tanaka, and H. Tsukamoto, “Kinetic studies of oxidation of ethylene over a TiO2 photocatalyst,” Journal of Photochemistry and Photobiology A, vol. 121, no. 1, pp. 55–61, 1999.View at: Google Scholar
P. Supphasrirongjaroen, P. Praserthdam, J. Panpranot, D. Na-Ranong, and O. Mekasuwandumrong, “Effect of quenching medium on photocatalytic activity of nano-TiO2 prepared by solvothermal method,” Chemical Engineering Journal, vol. 138, no. 1–3, pp. 622–627, 2008.View at: Publisher Site | Google Scholar