Recent Developments of Advanced Ti³⁺-Self-Doped TiO₂ for Efficient Visible-Light-Driven Photocatalysis

Abstract

Research into the development of efficient semiconductor photocatalytic materials is a promising approach to solving environmental and energy problems worldwide. Among these materials, TiO₂ photocatalysts are one of the most commonly used due to their efficient photoactivity, high stability, low cost and environmental friendliness. However, since the UV content of sunlight is less than 5%, the development of visible light-activated TiO₂-based photocatalysts is essential to increase the solar energy efficiency. Here, we review recent works on advanced visible light-activated Ti³⁺-self-doped TiO₂ (Ti³⁺–TiO₂) photocatalysts with improved electronic band structures for efficient charge separation. We analyze the different methods used to produce Ti³⁺–TiO₂ photocatalysts, where Ti³⁺ with a high oxygen defect density can be used for energy production from visible light. We categorize advanced modifications in electronic states of Ti³⁺–TiO₂ by improving their photocatalytic activity. Ti³⁺–TiO₂ photocatalysts with large charge separation and low recombination of photogenerated electrons and holes can be practically applied for energy conversion and advanced oxidation processes in natural environments and deserve significant attention.

1. Introduction

Environmental pollution and sustainable energy development are controversial issues we all face today [1]. In keeping with the pace of technological development, the development of renewable energy technologies must be studied in depth. Solar energy is the most abundantly available energy source. An amount as large as almost 99.9% of the energy used on Earth is produced by the sun [2,3,4]. Harnessing this enormous amount of solar energy would provide sustainable and environmentally friendly energy. It is also a promising option for economically feasible technology development.

Semiconducting photocatalysts are an eco-friendly and promising technology that converts solar energy into chemical energy. It can be used in various fields such as photolysis of harmful chemicals [5,6,7,8], artificial photosynthesis [9,10,11,12], photocatalytic water splitting [1,13,14,15,16,17] and electrochemical energy conversion [18,19,20,21,22,23]. Semiconducting photocatalysts using solar energy can solve environmental problems and have been actively researched around the world; as such, they have the potential to become a sustainable energy source [24,25,26]. Recently, various materials have been studied, including conjugated polymers (CPs) [27,28,29,30], 2D-layered materials [25,26,31,32,33,34], semiconductor quantum dots (QDs) [35,36,37,38,39], graphene quantum dots (GQDs) [40,41], isolated single-atom site catalyst (ISAS-catalyst) [42,43,44,45] and MOF-based materials [46,47,48]. However, semiconducting photocatalysts containing TiO₂ remain the most popular. There are several advantages of TiO₂ photocatalysts, including their low cost, good chemical and physical stability, lack of toxicity to humans and high photoreactivity [49,50,51,52,53].

However, there are also disadvantages that limit the practical application of TiO₂. First, the recombination of light-generated electrons and holes easily occurs, hindering the efficient use of absorbed energy [54,55,56]. Second, the TiO₂ surface is generally hydrophilic, so its absorption capacity for aromatic organic pollutants is low, limiting its responsiveness [54,57,58,59,60]. Though there are many other issues to overcome, the wide bandgap (~3.2 eV) of TiO₂ requires ultraviolet light for photoactivity [24,54]. Only 5% or less of sunlight is UV light [24,53,59]. This means that TiO₂ cannot use the visible-near IR energy that comprises most sunlight. Therefore, great efforts have focused on research to improve charge separation, surface reactivity and light absorption to overcome the disadvantages of TiO₂ [12,61,62,63,64]. Increasing solar energy efficiency and using more energy are major challenges for expanding the light-activated region into the visible region.

Adjusting the band structure to harness visible light can result in higher photon harvesting. To achieve this goal, attempts have been made to understand and modify the physical and chemical properties of TiO₂, such as the slow-light effect in periodic TiO₂ photonic crystals [65] and the multi-reflection effect in semi-hollow TiO₂ spheres [66]. Attempts have also been made to improve TiO₂ by classic chemical element doping [67,68,69,70,71], dye sensitizing [72,73], formation of semiconducting junctions [74,75,76,77], localized surface plasmon resonance (LSPR) photosensitization [78,79], stoichiometric adjustment or coordination of surface atoms or surface oxygen vacancies [61,80,81,82,83] and tuning inter-particle interactions of TiO₂ nanocrystals [84,85]. In particular, Ti³⁺-self-doped TiO₂ (Ti³⁺–TiO₂) such as blue-black TiO₂ has been found to induce high efficiency in visible-light-driven photocatalysis [86,87,88,89,90,91,92,93,94]. Nonetheless, improvements in photogenerated charge separation in Ti³⁺–TiO₂ remain challenges for a high photon harvesting efficiency. Thus, Ti³⁺–TiO₂ has been developed with modifications of the electronic structures to extend TiO₂ photocatalyst to highly efficient visible light harvesting.

In this review, we focus on the development of advanced Ti³⁺–TiO₂ for efficient solar energy harvesting of TiO₂ photocatalysts. Ti³⁺–TiO₂ is oxygen-deficient (TiO_2−x) and was first reported in 2011 by Chem et al. They produced defects on the TiO₂ surface and called the resulting material ‘disorder-engineered black TiO₂.’ This method reduced the light absorption band gap of to 1.5 eV, suggesting the possibility of using it for visible light catalysis by increasing solar absorption efficiency. [92] Since then, TiO₂ has been reduced to produce blue TiO₂ [87], and great efforts have been put into developing blue-black TiO₂ to use visible light, and much progress has been made. Figure 1 shows the schematic of the nanoparticle structure of Ti³⁺–TiO₂ and the band location for the insertion of each dopant.

We provide the latest reports on synthesis methods of Ti³⁺–TiO₂-based photocatalysts and their application efficiencies. First, we analyze the physical concept of Ti³⁺–TiO₂ photocatalysts and their production and advancement. We discuss the shortcomings in the production of Ti³⁺–TiO₂ and recent reports on the use of other materials to enhance the absorption of visible light. Moreover, many efforts have been discussed to overcome the disadvantages of photocatalytic efficiencies of Ti³⁺–TiO₂ due to low charge separation and high recombination of photogenerated electrons and holes. Finally, we present some perspectives on the future of modified Ti³⁺–TiO₂ photocatalysts.

2. Preparation of Ti³⁺–TiO₂

Ti³⁺–TiO₂ (TiO_2−x) has been demonstrated to enhance visible light absorption [95]. Ti³⁺-self doping causes defect levels due to oxygen vacancies in the lattice structure of self-doped TiO₂. The defect levels overlap the conduction band of TiO₂ and can induce vacancies in the electronic state just below the conduction band [96]. As a result, the band gap is reduced, resulting in enhanced visible light absorption. Methods for producing Ti³⁺–TiO₂ have been a focus of many studies. In general, high pressure and high temperature are used to reduce oxygen to form oxygen vacancies. Ti³⁺–TiO₂ doped with high oxygen vacancies is super hydrophilic. Thus, Ti³⁺–TiO₂ on hydrophilic substrates can result in strong bonding to the surface, thus increasing the coating ability.

According to recent trends, there are many reduction methods of TiO₂, which we divide into five main branches: hydrogenation, hydrothermal reaction, alkaline metal reduction, sol-gelation and phase-selective reduction. As generations go through, the preparation method of Ti³⁺–TiO₂ is not always superior, but it is considered to be a little more relaxed and more sophisticated.

2.1. Hydrogenation

Hydrogenation using hydrogen gas to react with oxygen atoms in the lattice of TiO₂ is the most common method for Ti⁺³ doping. Ti³⁺ is self-doped in an abundance of TiO₂ by high thermal or electromagnetic energy [97,98]. Hydrogenation requires harsh synthetic conditions and dangerous production processes (i.e., high heat and pressure, risk of explosion of hydrogen gas) [92,97,98,99]. To overcome this, other reducing agents that can provide hydrogen such as NaBH₄ and TiH₂ have been introduced as alternatives to hydrogen gas [98,100,101,102]. Additionally, increasing the rich Ti³⁺ doping involves higher heat, higher pressure and longer reaction times. Well-crystallized samples can effectively enhance visible and infrared absorption and exhibit hydrophilicity by oxygen doping with H in the lattice. However, in hydrogenated TiO₂, a high concentration of hydrogen atoms was required for efficient photocatalytic activity [103].

Additionally, if the temperature increases under firing conditions, phase conversion from anatase to rutile may occur. This can result in a band gap shift from anatase (~3.2 eV) to rutile (~3.0 eV). Phase transformation requires 600 to 700 degrees of energy in the absence of impurities, but the presence and amount of oxygen vacancies can lower the firing temperature. The increased oxygen deficiency causes relaxation of the large oxygen sublattice and facilitates the rearrangement and modification of atoms from anatase to rutile. In a high-pressure reducing atmosphere, heat transfer is also advantageous because of the high material density. Fast heat transfer will further promote anatase-to-rutile phase transformation [104]. However, the abundant doping of Ti³⁺ results in an irregular crystalline phase, called a disordered phase. Disordered crystals may behave differently from the normal physicochemical properties of anatase and rutile crystals.

2.2. Hydrothermal Reaction

A method of manufacturing Ti³⁺–TiO₂ with abundant defects has been reported using TiCl₃ and TiF₄ as precursors. Ti³⁺ can easily be oxidized under high heat and pressure conditions. Introduction of Ti⁴⁺ can inhibit oxidation of Ti³⁺ in this reaction, leading to abundant doping [105]. In addition, Fang et al. synthesized various reduced TiO₂ samples using Zn powder as a reducing agent and HF as a solvent in a simple one-pot hydrothermal process. However, Ti³⁺ introduced by Zn reduction is not stable in air and can be easily oxidized [106].

This method, which has been slightly improved under hydrogenation conditions, still requires high heat and pressure. It has an advantage in that the reaction proceeds in a liquid state that is easier to handle than hydrogen gas. However, it is unstable in the air and hydrogen saturation is not effective. A new solvent is needed to overcome this stability issue.

2.3. Alkaline Metal Reduction

Researchers tired of using high heat and pressure gases and liquids. In general, attempts have been made to reduce TiO₂ by using alkali metal, the most commonly used solid reducing agent. It is possible to obtain Ti³⁺–TiO₂ by simply replacing the high heat and pressure and simply mixing the powder. In 2013, Yin et al. produced gray TiO₂ rutile nanowires by reducing the TiO₂ surface by directly introducing aluminum metal at a temperature of 700 °C [107]. There have also been attempts to generate Ti³⁺ under milder conditions. In 2017, Zhang et al. sufficiently reduced TiO₂ using Na/NaCl powder under room temperature and an argon atmosphere and treated it with Ru to synthesize nanoparticles of TiO₂ with Ti³⁺ and Ru. Crystalline TiO₂ was milled with Na and NaCl powder at room temperature for 0.25 to 4 h under an argon atmosphere. It can be seen from Figure 2 that the color of P25 becomes darker as it undergoes harsh reaction conditions. Ti³⁺-rich TiO₂ nanocrystals can be used as effective support for Ru particles and the Ru/Ti³⁺–TiO₂ catalysts showed excellent performance in catalytic hydrogenation of N-methyl pyrrole [108].

2.4. Sol-Gelation

Since then, people have shown how to synthesize TiO₂ doped with Ti³⁺ abundantly, starting with the precursor building block, rather than doping Ti³⁺ into the already prepared TiO₂. Zhang et al. synthesized TiO₂ single-crystal nanorods using a sol-gelation method and succeeded in obtaining Ti³⁺-self-doped blue TiO₂ single-crystal nanorods by further annealing at 350 °C. Blue TiO₂ showed a 97.01% higher decomposition rate of RhB than did white TiO₂ under visible light, and the photocatalytic hydrogen generation rate increased. As a result of these investigations, the band gap was reduced to 2.61 eV due to the synergistic effect of the 1D-shaped single crystal structure and this increased the photocatalyst’s visible light absorption ability and photocatalytic reactivity [109]. In 2018, Yao et al. synthesized N/Ti³⁺-doped TiO₂ in one step using TiN. Additionally, using BiOBr, N/Ti³⁺-doped TiO₂ and heterojunctions were formed to enhance sonocatalytic activity. Doping of BiOBr and N/Ti³⁺ leads to effective charge separation of electron—electron pairs, increasing the methylene blue removal efficiency by 1.81 times [110].

In general, the grown TiO₂ crystal phase is anatase. However, depending on the reaction conditions, rutile anatase may also be present. Among the synthetic conditions, there is a possibility that BiOBr promoted rutile growth [110]. Optimal performance is achieved by mixing anatase and rutile in appropriate proportions (85:15). However, it is generally found that the photocatalytic efficiency of anatase is higher. Due to the high density of localized states in anatase, it is used for surface adsorbed hydroxyl radicals and slow charge carrier recombination. Rutile has a high charge recombination rate and consequently has a reduced ability to adsorb functional species, which is detrimental to photocatalytic performance [104].

2.5. Phase-Selective Reduction

There are many ways to produce Ti³⁺–TiO₂, but almost all of these require high temperature and high pressure. In addition, it is not yet possible to selectively reduce anatase and rutile phase on TiO₂ crystals in harsh conditions of high temperature and high pressure. In 2016, Zhang et al. first developed a method for selectively reducing the crystal phase of Degussa P25 TiO₂ nanoparticles at room temperature and atmospheric pressure using a simple Li-Ethylene diamine (EDA) solution. They melted Li metal in EDA solution at room temperature and pressure to produce a blue electride solution of a strong reducing agent to reduce TiO₂ nanocrystals. The Li-EDA solution is unique in reducing only the rutile phase while the anatase phase remained the same as in the composition of Degussa P25. They confirmed that Ti³⁺ was selectively introduced into TiO₂ to significantly improve visible and near-infrared absorption and increase the charge-hole charge separation efficiency through type II band gap alignment (Figure 3). These effects promoted a strong hydrogen-generating surface reaction. Therefore, when using phase-selective disorder engineering, 0.5 wt% Pt and 3.46-mmol were used to show high stability and a high hydrogen generation rate of 13.89 mmol·h⁻¹·g⁻¹ [87]. In addition, in 2019, Hwang et al. discovered that when P25 is treated with Na-EDA solution, only the anatase phase is reduced. Engineered A_d/R_o TiO₂ selectively treated with Na-EDA to reduce the anatase phase contained only Ti³⁺ defect sites in the anatase phase and the internal energy band gap was narrowed. This indicates enhanced visible light activity. Hwang et al. subsequently reduced CO₂ using Ti³⁺-self-doped A_d/R_o TiO₂ that selectively reduced only the anatase phase to produce CH₄ using light in the visible region [111]. The phase-selective method of reducing TiO₂ using a metal-EDA electride solution has attracted attention as a promising method of self-doping rich Ti³⁺ on the surface of TiO₂ at room temperature and atmospheric pressure to improve the visible light activity of TiO₂. The developed Blue TiO₂ (BTO) was used for organic arylation reactions, the yield of which was supported with untreated P25 (40%) and BTO (63.4%) [112]. In addition, the BTO had excellent reactive oxygen species (ROS) generation and showed strong efficacy in decomposing algae (Chlamydomonas) under sunlight [113].

The phase-selective reduction of P25 does not require heat or pressure. In addition, while selectively reducing rutile and anatase TiO₂ crystal phases, it became possible to impart selectivity [114]. With the selective reduction of rutile and anatase, their physicochemical properties are clearly different. We expect to pioneer new applications in various fields by applying this method.

3. Enhancement of Photogenerated Charge Separation and Photocatalytic Activity in Advanced Ti³⁺–TiO₂

Photocatalysts exhibit catalytic activity when introduced to light. The photogenerated electrons and holes move to the photocatalytic surface. The moved electrons and holes react with adsorbed electron acceptors and donors, respectively, to complete the catalytic reaction. Abundant Ti³⁺ doping can lead the excitation region from ultraviolet to visible light. Ti³⁺-doped TiO₂ can narrow the wide band gap of TiO₂ for harvesting visible light and can provide an increase in electronic conductivity due to doped Ti³⁺ defects. However, after photoexcitation, the generated electrons and holes may undergo significant electron recombination in the depletion region of Ti³⁺-doped TiO₂ [115,116,117].

To overcome this and ultimately enhance photocatalytic activity in all applications, deformation studies with several different materials have been conducted. Performance can be improved by using Ti³⁺–TiO₂ in combination with other materials or elements. Studies using various other materials have been conducted to overcome the previously mentioned limitations of Ti³⁺–TiO₂ deformation.

We focus on the synthetic methods of advanced Ti³⁺–TiO₂ with high photogenerated charge separation and efficiency and classify them into four categories according to recent developments. Ti³⁺–TiO₂ modification has been extensively studied, including metal and nonmetal doping, semiconducting coupling, and stoichiometry modification to prepare and stabilize composites with other materials.

3.1. Metal-Doping

It has long been a major concern to increase the efficiency of TiO₂ photocatalysts and expand their use in the visible region using easily accessible metal dopants. Metal dopants can be mixed with Ti³⁺ atoms or replaced by Ti atoms, and the band gap can be adjusted in such a state. Moreover, a new band was produced as a conduction band to help better separate photogenerated charges after photoexcitation. Metal dopants act as an electron acceptor as well as a hole acceptor in the valence band. The introduction of metal dopants is still a promising option for Ti³⁺–TiO₂. We summarized these efforts in

Table 1. Summary of metal-doped Ti³⁺self-doped TiO₂ photocatalysts.

Catalysts	Light	Application Target	Efficiency	Ref.
Plasmonic Ni/Ti3+ TiO2/SiO2 nanospheres	300 W Xe lamp	Methylene blue	0.11 ± 0.04 μmol·L−1·min−1	[69]
Pt single atom-anatase(001) Ti3+ TiO2 nano sheets	325 nm, 365 nm	Hydrogen evolution reaction	400 µmol·h−1·g−1	[89]
Zn-assisted Ti3+ TiO2	500 W Tungsten halogen lamp	Formic acid	100% 90 min−1	[106]
Li-EDA TiO2–Pt	150 W Xe lamp	Hydrogen evolution reaction	350 µmol·h−1·g−1	[114]
Bi3+/Ti3+ TiO2	Xenon lamp	Flumequine under HSO5−	65%	[118]
Co2+/Ti3+ TiO2	300 W Xe lamp	Acid Orange 7	100% 5 h−1	[119]
Cu/Ti3+ TiO2	300 W Xe lamp	F. graminearum and B. dothidea spores	100%	[120]
RE-doped/Ti3+ TiO2/Sep	300 W Xe lamp	Orange G	72%	[121]

.

3.1.1. Surface Plasmon Effects on Ti³⁺–TiO₂

LSPR-based photosensitization can promote the absorption of visible or near infrared rays [122,123]. If the frequency of the incident photon coincides with the natural frequency of the vibrating surface electrons, resonant group oscillations of the valence electrons can occur to promote absorption of Ti³⁺-self-doped TiO₂ visible light [124]. Usually, LSPR excitation occurs in the visible region when Au [125], Ag [126], and Pt [127] are combined with TiO₂ [128]. Attempts to combine Ti³⁺-self-doped TiO₂ and LSPR metals have continued. However, in almost all cases, noble metals should be used for plasmon resonance [125,126,127]. The use of precious metals is still valid. In 2019, phase-selective room-temperature solution engineering was used to enhance HER performance by attaching Pt particles to TiO₂ richly doped with Ti³⁺ [114]. This indicates that the simple phase-selective room-temperature solution engineering method does not lag behind TiO₂ manufactured in other harsh conditions of different reducing conditions.

Recently, He et al. used Ti³⁺ and Ni for plasmon-mediated carrier transfer to enhance the degradation of methylene blue in visible light. They deposited Ni–TiO₂ structures on SiO₂ spheres based on the modified nanosphere lithography method. It was found that plasmon-generated hot electrons and Ni holes can be transferred to TiO₂ in a heterogeneous structure (Figure 4). The transferred hot electrons and holes occupy oxygen vacancies or produce Ti³⁺. TiO₂ self-doped with Ti³⁺ showed enhanced removal performance of methylene blue of 0.11 ± 0.04 μmol·L⁻¹·min⁻¹ in visible light [69].

Furthermore, it is valuable to be able to achieve the same effect as LSPR by using inexpensive non-precious metal transition metals other than Ni. We believe that developing a method of applying transition metals will become the main approach to metal-doped Tt³⁺ TiO₂ in the future.

3.1.2. Single Atom Site Doping Effects on Ti³⁺–TiO₂

Isolated single atom site (ISAS) catalysts are a new approach in the field of catalysis due to their increased catalytic activity and excellent selectivity. Hejazi et al. created a stable platinum single atom site by grafting Pt single atoms to the Ti³⁺-self-doped TiO₂ anatase (001) nanosheet by hydrogenation [89]. After reducing the thin Ti³⁺layer in the surface Ar/H₂ environment, it was used as a Pt single atom support and tightly adhered from the diluted aqueous Pt solution. The amount of Pt deposition can be adjusted by considering the heat treatment temperature in Ar/H₂. This showed a 150-fold improvement in hydrogen generation rate over that observed using the same amount of platinum nanoparticles (Figure 5) [89]. Despite many issues, including the stability of ISAS catalysts, this is a very attractive approach. With a small amount of metal, a photocatalytic performance of tens to hundreds of times improvement is observed. In addition, we believe that a single atom leads to a widely applicable approach for many. The development of the ISAS catalyst using Ti³⁺ as support can provide a new approach for Ti³⁺-self-doped TiO₂ for absorbing visible light.

3.1.3. Transition or Rare Metal Grafting on Ti³⁺–TiO₂

In general, the bonding of transition metals and TiO₂ is a classic method. Similarly, much effort has been put into grafting transition metals into Ti³⁺–TiO₂. Recently, research on photocatalysts for removing microorganisms and viruses was actively conducted in addition to the removal of existing classical organic or inorganic contaminants. Zhang et al. reported cell lysis of E. coli and agropathogenic fungal spores by preparing classical Cu and Ti³⁺–TiO₂ in a new way. It was reported that anhydrous ethanol can be used as a reducing agent in the hydrothermal process leading to the formation of metallic copper and Ti³⁺. They confirmed the photocatalytic disinfection effect on five agricultural pathogenic fungal spores, including F. graminearum and B. dothidea spores, by fluorescence staining using a FungaLight CFDA-AM/propidium iodide yeast vitality kit. Almost all bacteria were killed upon irradiation for 3 h. In Figure 6, killed bacteria are identified by fluorescent staining. Both copper and other metals are expected to be effective in bacterial and viral cell lysis when combined with Ti³⁺-self-doped TiO₂ [120].

Recently, grafting of rare earth metals such as Bi (rather than classic commercially available metals such as Cu, Zn and Co) into Ti³⁺–TiO₂ has been reported. Gul et al. increased the solar reactivity by grafting Bi to Ti³⁺–TiO₂. They performed extensive doping with Ti³⁺ and Bi-by-Bi doping and sol-gel technology and showed better photocatalytic performance than undoped TiO₂ for flumequine decomposition [118].

As shown in Figure 7, Zhou et al. introduced doping with six rare earth metals (i.e., La, Ce, Pr, Nd, Eu and Gd) and sepiolite with Ti³⁺–TiO₂ to improve visible light activity and provide low band gap energy. In addition, this strategy resulted in a high adsorption capacity of contaminants through the sepiolite and improved the photocatalytic performance. Figure 7 shows the schematic structure for modification of the conduction band maximum and the photodegradation effect of Orange G for each metal. They found that, upon doping of rare earth metal ions, the structure and chemical properties of the nanocomposite of TiO₂ depended heavily on the radius of the rare earth metal ion. Rare-earth ions with a valence of +3 (i.e., RE⁺³) can be abundantly doped with Ti³⁺ because the rare earth metal ions are much larger than that of Ti⁴⁺. However, this can cause a charge imbalance that can contribute to the adsorption of OH⁻. It is also reported that the empty 4f level of rare earth metals can act as a scavenger for light-generating electrons. Of the six rare earth metals, the Eu complex showed the best photocatalytic performance [121]. The combination of common transition metals and rare earth metals with Ti³⁺-self-doping can be an excellent way to enhance photocatalytic performance, and promising materials for cell lysis of viruses and bacteria can be developed.

3.2. Nonmetal Doping

Much effort has been made to expand the applications of TiO₂ photocatalysts with metal and nonmetal based dopants into the visible light regime and increase the photocatalytic efficiency. In particular, many common elements such as nitrogen, carbon, sulfur and phosphorus have been used for grafting. The grafting of nonmetallic elements such as these can lower the surface energy for adsorption of the photocatalytic surface [129]. Unlike metal dopants, nonmetal dopants can be directly involved in the band of TiO₂. In addition, nonmetallic elements with a different electronic environment from oxygen can have positive and negative effects on the band gap energy. We believe that such modification through band engineering is a unique possibility for nonmetal dopants. We summarize the recent Ti³⁺ and nonmetallic dopant grafting in

Table 2. Summary of nonmetal doped Ti³⁺self-doped TiO₂ photocatalysts.

Catalysts	Light	Application Target	Efficiency	Ref.
N/Ti3+/C TiO2	500 W Xe lamp	Rhodamine B	100% 1 h−1	[70]
N/Ti3+ TiO2 nanotube	400 W Halogen lamp	Rhodamine B	100% 4 h−1	[130]
N/Ti3+ TiO2 spheres	300 W Xe lamp	Methylene orange	100% 3 h−1	[131]
S/Ti3+ TiO2	150 W Xenon lamp	Hydrogen evolution reaction	22.5 µL·h−1·cm2	[132]
S/Ti3+ TiO2–Ag nanorods	Xe lamp	Hydrogen evolution reaction	209.2 μmol·h−1·g−1	[133]
C–N–S-tridoped Ti3+ TiO2	300 W Xe lamp	Hydrogen evolution reaction	149.7 μmol·h−1·g−1	[134]

.

3.2.1. N-Doped Ti³⁺–TiO₂

Nitrogen is one of the most frequently used nonmetal elements for adjusting the valence band of TiO₂. Doping with nitrogen may cause the maximum upward movement of the valence band, and visible light absorption may also be improved by controlling the red-shift of the absorption edge. Nitrogen can improve both carrier density and electron conductivity in the neutral region [129]. Additionally, it can act as an absorption site for organic pollutants. Recently, efforts have been made to enhance visible light absorption and enhance photocatalytic performance using Ti³⁺ and N dopants together. By doping N and Ti³⁺ together, Ti³⁺ in the band decreases the conduction band, and N increases the valence band to reduce the band gap energy and enhance the photocatalytic performance. Jia et al. improved the charge separation by introducing carbon after N doping, introducing carbon components, and using carbon as a site to accept the separated charge after photoexcitation. This showed a significantly faster Rhodamine B decomposition rate than conventional nitrogen-doped Ti³⁺-self-doped TiO₂ (Figure 8) [70].

3.2.2. Sulfur-Doped Ti³⁺–TiO₂

Among nonmetal dopants, sulfur has the same valence as oxygen, has adequate energy, and may not cause charge imbalance in TiO₂. The energy band occupied by the charge imbalance can act as a carrier recombination site and degrades the carrier transport process [129,135]. ⁴ Based on these advantages, sulfur has been noted as another promising nonmetallic dopant to drive the performance of TiO₂ photocatalysts. Active research with Ti³⁺ is ongoing. Ji et al. introduced S²⁻ into anatase TiO₂ together with Ti³⁺, which significantly lowered the band gap energy and increased photocatalytic hydrogen production. They confirmed that, by introducing sulfur and Ti³⁺, the band gap decreased from 3.2 eV for the existing anatase TiO₂ to 2.0 eV after co-doping [132]. As shown in Figure 9, Meng et al. added plasmonic Ag particles to TiO₂ nanorods introduced with sulfur and Ti³⁺ to synergistically reduce the band gap through impurity dopants and expand LSPR into the visible light regime. These are photocatalytic reactions where phenol’s visible-light-driven photocatalytic decomposition rate and hydrogen production rate are as high as 98.67% and 209.2 μmol·h⁻¹·g⁻¹, approximately 2 and 5 times higher than TiO₂ alone [133].

3.2.3. Multi-Doped Ti³⁺–TiO₂

Several elemental dopants can be introduced at the same time. Yan et al. reported black TiO₂ nanosheets with Ti³⁺ by doping nitrogen, carbon and sulfur. The doped C, N, S elements can result in additional impurity levels above the valence band. The mass concentrations of doped C, N and S are 14.20%, 4.80% and 3.30%, respectively, and these worked synergistically with Ti³⁺. They used visible light to photocatalytically decompose methyl orange up to 92.13%. The amount of hydrogen generated by the photocatalyst was 149 μmol·h⁻¹·g⁻¹ [134]. An attempt to use multiple materials without being limited to a single material may provide a new way to improve photocatalytic performance.

3.3. Semiconducting Coupling

Semiconducting heterogeneous catalytic junction enhances various properties by adding a new band to Ti³⁺–TiO₂. The newly added band may additionally absorb visible light, promote optical junction carrier separation and transfer and enhance interactions with the target material. Semiconducting heterojunctions are advantageous for absorbing visible light of photocatalysts and increase photocatalytic activity by promoting separation and transfer of photojunction carriers of semiconductors [136,137]. Previous bands of semiconductor materials such as CdS [138], Ag₃PO₄ [139], Cu₂O [140] and WO₃ [141], have been combined with TiO₂. Recently, g–C₃N₄ has shown overall water splitting band potential and high visible light response, leading to the current trend [142,143]. Likewise, when Ti³⁺–TiO₂ is combined with semiconducting materials, better synergy can be expected than before. We looked at the recent trends of Ti³⁺–TiO₂ and organized them in

Table 3. Summary of semiconducting coupling with Ti³⁺self-doped TiO₂ photocatalysts.

Catalysis	Light	Application Target	Efficiency	Ref.
Ti3+ TiO2/gC3N4	Xe lamp	Hydrogen evolution reaction	3748.46 μmol·h−1·g−1	[90]
Ti3+ TiO2/WO3–Ag	Solar	CO2 to CO	1166.72 μmol·h−1·g−1	[91]
Ti3+ TiO2/rGO	500 W Xe lamp	Methylene blue	100% 2 h−1	[93]
Ti3+ TiO2/meso-gC3N4	300 W Xenon lamp	Phenol	100% 90 min−1	[94]
Ti3+ TiO2/WO3	420 nm cutoff	Toluene	100% 60 min−1	[103]
Ti3+ TiO2 nanowires/rGO	300 W Xe lamp	Waste oil	100% 5 h−1	[102]
N/Ti3+ TiO2/BiOBr	120 mW lamp	Methylene blue	100% 50 min−1	[110]
Ti3+ TiO2/MoS2/Ag	500 W Xe lamp	Bisphenol A	100% 2 h−1	[144]
Ti3+ TiO2/Ag3PO4 QD	300 W Xe lamp	Methylene orange	100% 100 min−1	[145]
Ti3+ TiO2/CdS QD	300 W Xe lamp	Methylene blue	100% 150 min −1	[146]
Ti3+ hollow TiO2/MoS2/CdS	300 W Xe lamp	Hydrogen evolution reaction	8950 μmol·h−1·g−1	[147]
N/Ti3+ TiO2/MoS2	300 W Xe lamp	Methylene orange	100% 2 h−1	[76]
Ti3+ TiO2/Ce3+ CeO2	300 W Xe lamp	Methylene blue, Methylene orange	100% 3 h −1	[148]
Ti3+ TiO2/FeOx	420 nm LED	Oxygen evolution reaction	410 μmol·h−1·g−1	[149]
Ti3+ TiO2/g–C3N4 hollow nanosphere/MoS2 QD	300 W Xe lamp	Hydrogen evolution reaction	1524.37 μmol·h−1·g−1	[150]
Ti3+ TiO2/g–C3N4 nanospheres/Ag	300 W Xe lamp	Methylene orange	100% 3 h−1	[151]
Ti3+ TiO2/g–C3N4	Xe lamp	NOx	75%	[152]

.

3.3.1. Transition Metal Dichalcogenide or Quantum Dot/Ti³⁺–TiO₂

Transition metal dichalcogenide (TMD) materials based on quasi-dimensional (2D) structures have recently attracted attention because of their excellent electronic and optical properties. These materials have great potential for high-tech applications and have different properties compared to conventional bulk counterpart materials. The state density of the TMD material increases in a semi-continuous stepwise fashion. Therefore, TMD materials can exhibit new electronic and optical properties [25,26]. MoS₂ is a representative TMD material, and it is a popular material in HER photocatalysts and electrocatalysts. This is because the sandwiched MoS₂ atomic layer can provide various unsaturated bonds at the edge of the layer. There have been attempts to enhance the photocatalytic performance with recent Ti³⁺–TiO₂ using these new electronic and optical properties. There have been attempts to dope the MoS₂ layer and other semiconducting or metal materials over spear-shaped black Ti³⁺–TiO₂. Attempts have been made to bond N/Ti³⁺–TiO₂ with spear and MoS₂ and to apply hollow Ti³⁺–TiO₂ and MoS₂/CdS together [76,147]. All of the materials were able to clearly identify the vertically grown layer and showed excellent performance in removing HER and Methylene orange. Figure 10 shows the schematic structure and mechanism of action of hollow black TiO₂/MoS₂/CdS. In addition to TMD metal, there have been attempts to combine Ti³⁺–TiO₂ with a layered material represented by bismuth oxyhalides. Yao et al. confirmed the well-grown TiO₂ and BiOBr nanostructures by synthesizing N/Ti³⁺–TiO₂ and BiOBr in one step. These materials showed high efficiency in removing methylene blue due to increased visible light absorption [110].

In addition to 2D-based layered materials, zero-dimensional quantum dots using TMD materials are also attracting attention because of their small size, large specific surface area and shortened charge travel distance. Many recent TiO₂ quantum dot complexes have attracted considerable attention. Coupling of TiO₂ and semiconducting quantum dots can improve the charge separation of photo-generated electron-electron pairs and smaller quantum dot materials are uniformly dispersed on the TiO₂ surface, preventing aggregation of quantum dots and providing more active sites [153,154]. However, the synergy of the existing TiO₂/QD materials was not satisfactory [155,156]. Therefore, a catalyst using visible solar energy was achieved with the introduction of Ti³⁺–TiO₂. Likewise, there have been recent attempts to modify the MoS₂ and CdS QDs through conjugation of Ti³⁺–TiO₂ as shown in Figure 11 [146,150]. Abundant Ti³⁺self-doping sufficiently improves the visible light response, and QD MoS₂ and CdS can promote photocatalytic carrier separation through a built-in electric field to provide a rich specific surface area to improve photocatalytic performance.

In addition to TMD materials, there have also been attempts to integrate Ti³⁺–TiO₂ with QD. Ag₃PO₄ attracted attention because of its high quantum yield and band gap energy (2.45 eV) suitable for absorbing visible light. However, Ag₃PO₄ was only used as a photocatalyst due to fast charge recombination and relatively large particle size. Overcoming these shortcomings with QD Ag₃PO₄ and bonding with Ti³⁺–TiO₂ increased visible light responsiveness and improved photocatalytic performance (Figure 12) [145]. The introduction of quantum dots suggests a new paradigm to solve problems that could not be solved with the existing classical materials. Richly doped Ti³⁺ can serve as an effective support for quantum dot materials and as an electron carrier delivery site. Thus, richly doped Ti³⁺ has very promising and potential possibilities, requiring more in-depth research into this area.

3.3.2. Carbon-Based Nonmetallic Semiconductors/Ti³⁺–TiO₂

Carbon-based nonmetallic semiconducting materials are attracting great attention in the energy industry because of their attractive properties such as high conductivity and easy accessibility. The two-dimensional electronic network of sp² carbon atoms exhibits high electron mobility and conduction [157]. In the past few years, carbon-based materials have been extensively explored as photocatalysts. Recently, graphitic-carbon nitride (g–C₃N₄) has attracted much attention as a visible light active catalyst due to its appropriate band gap (2.69 eV). In particular, TiO₂ and Z-structure photocatalyst band alignments are similar to the plant photosynthesis process and can provide high photocatalytic activity through more powerful charge separation than type I and type II band alignment [158,159]. Ti³⁺–TiO₂ can strengthen this Z-structure with abundant Ti³⁺ doping in the conduction band. Figure 13 shows the Z structure diagram between TiO₂ and g–C₃N₄. Recently, studies on enhancing the performance of visible light-activated photocatalysts have been actively conducted through the combination of Ti³⁺–TiO₂ and g–C₃N₄. There have been attempts to improve charge separation by doping plasmon metal between Ti³⁺–TiO/g–C₃N₄. Cao et al. introduced Ag-doped black TiO₂ nanosphere structures and g–C₃N₄. Figure 14 shows a schematic of the Ti³⁺–TiO₂/Meso-g–C₃N₄ manufacturing method. They showed that Ag could act as an electron-conducting bridge between TiO₂ and g–C₃N₄, contributing to electron donors and acceptors in the Z-structure [151].

3.3.3. Metal Oxide/Ti³⁺–TiO₂

Heterojunctions of metal oxide and TiO₂ are a classic (and the easiest) method to enhance TiO₂ photocatalytic activity after metal doping. To date, new heterojunctions of various metal oxides and TiO₂ have been studied. Heterojunctions of Ti³⁺–TiO₂ and metal oxides are also of great interest. Although many metal oxides are candidates, heterojunctions of WO₃ and Ti³⁺–TiO₂ have recently been developed. WO₃ can achieve Z-structure band alignment with Ti³⁺–TiO₂. The resulting Z-structure has a band structure similar to natural photosynthesis [158,159]. This can strongly increase the charge separation efficiency. In addition, two different redox sites can suppress the reverse reaction of radical production. In 2020, Nguyen et al. developed a catalyst that introduced blue Ti³⁺–TiO₂ and WO₃ produced by lithium (Li–EDA) of ethylenediamine, a strong reducing agent of the previous superbase. They show 100% selectivity in reducing CO₂ to CO, and that blue Ti³⁺–TiO₂/WO₃ formed a perfect Z-structure (Figure 15) [91]. Consequently, the introduction of a complete Z-structure can be a powerful modification of Ti³⁺–TiO₂. This is a direct way to overcome the weakness of charge separation and recombination and dramatically enhance the performance of photocatalysts.

There have also been attempts to make heterojunctions with CeO₂, which has a band gap energy similar to TiO₂. CeO₂ has recently attracted attention due to the special redox properties between Ce (III) and Ce (IV) oxidation states, which include high thermal stability, excellent oxygen storage capacity and easy conversion. Xiu et al. developed Ti³⁺/Ce³⁺ self-doped TiO₂/CeO₂ nanosheets by reducing TiO₂ and CeO₂ together. They showed that Ti³⁺ and Ce³⁺ complement each other and enhance electron separation (Figure 16) [148].

3.4. Stoichiometry Modification

There have been attempts to change the stoichiometry of materials beyond controlling Ti³⁺-self-doped intrinsic defects to enhance the visible light absorption of TiO₂ photocatalysts. Introduction of defect states such as Ti³⁺ is sensitive to the chemical state and spatial distribution of electronic structure modifiers. Therefore, photocatalytic activity may be reduced through the formation of an undesirable band structure. It is important to create favorable defects at the atomic level to improve absorption in the visible light band. However, it is still difficult to correct structural defects at the atomic level and produce an optimal TiO₂ material. A study on correcting atomic-level structural defects is ongoing beyond the recent Ti³⁺–TiO₂ work. Efforts have been made to replace hydrogen vacancies in Ti³⁺–TiO₂ with atomic levels of hydrogen [81], carbon [82] and nitrogen [82,83]. We summarized in

Table 4. Summary of stoichiometry modification of Ti³⁺self-doped TiO₂ photocatalysts.

Catalyst	Color	Application Target	Efficiency	Ref.
Ti3+2H TiO2	Red	–	–	[81]
Ti3+CN TiO2	–	Oxygen evolution reaction	4.125 μmol·h−1·g−1	[82]
Ti3+N TiO2	Yellow-green	Methylene orange	100% 2 h−1	[83]

.

In general, reduced Ti³⁺–TiO₂ has a blue-black color, whereas substituted TiO₂ has a different color. Red TiO₂ appears when the space for a single oxygen atom is filled with two hydrogen atoms, called Ti³⁺_H TiO₂ [81]. As shown in Figure 17, the color and atomic/band structure of TiO₂ was changed according to the vacancies in the crystal structure. In addition, the introduction of heterogeneous elements results in a new band state. The band gap is reduced due to changes in the conduction band and valence band, and both show excellent activity in visible light.

We believe that engineering the oxidation state of Ti atoms (beyond Ti³⁺, which is simply reduced) and the surrounding electronic environment through complete stoichiometric control is a new paradigm for developing effective TiO₂ visible photocatalysts. In the future, we can expect to expand and improve the photocatalytic performance through visible light for TiO₂ photocatalysts via uniformly controlled stoichiometric modifications.

4. Conclusion and Outlook

TiO₂-based photocatalytic reactions were first reported in the early 1930 s. They were first used to bleach dye using reactive oxygen species generated under UV light. TiO₂ photocatalyst (called a “photosensitizer”) has been studied extensively, and it is expected to play a role in solving the limited energy problem [24]. In particular, efforts have been made to improve the absorption band of TiO₂ photocatalyst into the visible light regime to efficiently use solar energy. These efforts continued to Ti³⁺–TiO₂, and since the report of black Ti³⁺–TiO₂ reduced by Ward et al. in 2011, Ti³⁺–TiO₂ research has made great progress [92]. We first studied the development of TiO₂ reduction to produce abundant Ti³⁺ for overcoming the high heat and pressure of the hydrogenation reaction (as the first generation of Ti³⁺–TiO_2-based photocatalysts). Then, due to their long reaction times and risks, increasing mild and easy to handle methods have been developed as the next generation. A lithium-ethylenediamine solution (Li-EDA) has been introduced for self-doping Ti³⁺ effectively at room temperature and atmospheric pressure, as well as controlling the crystalline phase of reduced TiO₂. Selective reduced crystal phases with different physicochemical properties will pioneer photocatalytic utilization for several properties [87]. However, although Ti³⁺ is abundantly doped such that the band region for visible light is expanded, there are still limitations. The interaction and reactivity between target materials and photocatalysts, as well as effective suppression of charge separation and recombination, are still challenges that need to be overcome. The best way to improve the limited performance of these single materials is with several different composite materials.

In this review, we focused on the types and methods of Ti³⁺–TiO₂ and the materials introduced and categorized them into four categories according to recent developments. Ti³⁺ doping can be used as sufficient support for an isolated single-atom site catalyst and can stabilize the catalyst atom. The isolated single-atom catalyst has excellent selectivity and reactivity with a small amount of catalyst [89]. Introducing metals other than plasmonic precious metals will make commercialization easier. The introduction of LSPR using isolated single atom site catalysts and non-precious metals in metal dopants is the most effective modification method. Going further, bimetallic sites, which have been extensively studied recently, are being proposed as a new way to solve the selectivity and stability issues that a single atom catalyst cannot. There are different advantages depending on the metal combination, and there are even advantages to not using precious metals. It has recently been shown that bimetallic site catalysts can also be applied with TiO₂ [160]. However, this approach requires more research of bimetal sites.

The introduction of nonmetal dopants is a way to engineer the fundamental band structure of TiO₂. Modification of the band structure can make it easier to expand this approach to visible photocatalysts, and most nonmetallic dopants can improve the decomposition of harmful substances through strong interactions with organic harmful substances. Another approach is doping two or three heterogeneous elements beyond single element nonmetallic doping. The introduction of a suitable semiconducting material can be used to overcome the shortcomings of the aforementioned TiO₂ photocatalysts. Beyond the classic 3D dopant, the introduction of 2D TMD and 0D QD materials shortens both the large specific surface area and charge travel distance. Because of this, it is possible to improve the charge separation of electron—electron pairs, which could not be solved before. The improved charge separation has high expectations for HER and water splitting. The introduction of metal oxides can also be a great alternative. Due to their excellent accessibility and stability, carbon-based materials can be promisingly combined with Ti³⁺ TiO₂. The Z-structure has a structure similar to that of photosynthesis. The formation of Z-structures through the introduction of metal oxides and carbon-based materials is a powerful approach to improving photocatalytic performance. Similar to a plant’s photosynthetic structure, the photocatalyst can replace various functions of plants in the future. This research field has potential in the energy industry.

Furthermore, advances in material engineering technology can dominate atomic defects. This presents a new base of TiO₂ photocatalysts through heteroatom insertion into the TiO₂ grid beyond Ti³⁺ single atom doping [81,82,83]. It is expected that the next generation of Ti³⁺–TiO₂ is in stoichiometric modification. If the electronic environment is completely controlled by controlling the oxidation state of the Ti atom and the surrounding atoms rather than simply doping the reduced Ti³⁺, there is a possibility to produce a TiO₂ photocatalyst suitable for each application method. TiO₂ has been studied as a practical photocatalytic material. However, in addition to applications that can be directly applied, in-depth studies on the electron environment, oxidation state of various dopants in the crystal structure, enhanced charge separation, and defect-level band energy are feasible. By investigating the atomic state and bonding level in the final crystal structure, it is possible to identify the principles and mechanisms of theoretical photocatalytic applications. Such an understanding is essential to improving photocatalytic performance.

Advanced Ti³⁺–TiO₂ approaches have recently been extended to visible light in solar energy harvesting and air purification. In particular, we believe that by applying air purification [152], diseases caused by viruses and bacteria [120] in the air—that have emerged as among the biggest problems faced by humanity—can be prevented by using Ti³⁺–TiO₂-based photocatalysts, a visible photocatalyst that is harmless to humans. For this, it is necessary to study the interactions of cells with photocatalysts in a view point of biochemistry, as well as conduct theoretical studies about classical interactions between organic and inorganic and their mechanisms. Nonetheless, the application of visible light photocatalysts in this industry is still insufficient. It is essential to explore sufficient functions required for commercialization [91], further study of Ti³⁺–TiO₂-based materials and the development of more effective heterogeneous materials. This makes Ti³⁺–TiO₂ a viable option for visible photocatalysts in the future.