English

• Energy and environmental crises are becoming increasingly severe in modern society. To address these problems, the exploration of clean and renewable energy sources and the development of practical eco-friendly systems for environmental remediation are attracting increasing attention [1-3]. One promising technology is semiconductor photocatalyst systems that can split water into hydrogen and oxygen and degrade various organic pollutants to clean up the environment. These systems use solar energy as a photon source to promote the corresponding reactions and have the advantages of low external energy consumption, cost effectiveness, and sustainability [4,5]. The rational design of high-performance photocatalysts for hydrogen production or pollutant degradation has thus become a major field of research [6-8].

  Since the report by Fujishima and Honda on the photocatalytic properties of TiO₂, this compound has been extensively studied as a primary photocatalytic material [9-13]. In particular, TiO₂ mesoporous spheres (MS) with large specific surface areas and strong light scattering have been synthesized and employed as photoanodes with a fast response, low dark current, and high sensitivity [14-16]. In addition, TiO₂ MS have been widely used in catalytic pollutant degradation, solar hydrogen production, and fuel cells [17]. Nonetheless, the wide bandgap (3.2 eV for anatase and 3.0 eV for rutile), high electron-hole recombination rate, and low intrinsic electron mobility still inhibit the photocatalytic performance of TiO₂ nanostructures [18]. Various strategies to overcome these bottlenecks have been explored, such as noble metal deposition, composite formation with other semiconductors, self-doping, and defect engineering [19].

  Defects can strongly affect the lattice, surface state, dispersion, strain, and electronic structures of nanomaterials and ultimately induce changes in the photocatalytic pathways and overall performance [20,21]. Recently, the preparation of reduced TiO_2−X containing defects by surface structure engineering was confirmed as an effective strategy for improving the photocatalytic performance of materials [22]. First, reduced TiO_2-X materials featuring different colors, such as yellow, blue, red, and black, can be obtained from conventional white TiO₂ by adjusting the electronic energy band structure and expanding the optical response from the ultraviolet (UV) to the near-infrared region, leading to enhanced photocatalytic activity [23,24]. Second, oxygen vacancy are shallow donors by creating localized states in the bandgap, thereby facilitating the favorable separation of the photogenerated carriers and increasing electrical conductivity [25,26]. Third, the introduction of oxygen vacancies interferes with the lattice spacing, destroys the symmetry of the crystal structure, and alters the electronic structure of the active centers [27]. The modification of active sites also changes the material surface and affects the adsorption stability [28]. All of these effects can lead to remarkably enhanced photocatalytic performance for TiO_2-X. On the other hand, defects, especially bulk defects, can also act as charge recombination centers to limit the effective participation of photogenerated carriers in the photocatalytic reaction [29]. To date, most studies into defects only partially elucidated their beneficial or detrimental roles in photocatalysis [30-32]. Given that defects can exert both positive and negative effects on photocatalytic processes, the optimal defect concentration for maximizing the photocatalytic efficiency must be determined for each process.

  In this work, TiO_2−X MS were prepared by a facile high-temperature reduction method, which changed the electronic structure and catalytic activity through the strain effect and oxygen doping (Fig. S1 in Supporting information). The structures, chemical compositions, optical characteristics, catalytic properties, carrier localization degrees, wettabilities, and band structures of the prepared samples were investigated. The reduction temperature was found to be linearly correlated with the oxygen vacancy concentration and lattice strain. Moreover, the TiO_2−X MS with suitable lattice strain and oxygen vacancy concentration displayed excellent photocatalytic performance when applied to both hydrogen production and organic pollutant degradation. Furthermore, a novel preparation method for photocatalytic materials with controllable vacancy concentrations is illustrated for energy and environmental applications. The obtained results are anticipated to be of far-reaching significance for the design of high-performance bifunctional reduced TiO_2−X photocatalysts by combining defect engineering and strain engineering.

  The detailed nanostructure of the TiO₂ MS before and after reduction were examined by transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM). As shown in Fig. 1A, the TiO₂ MS were obtained as monodisperse spheres with a porous nanostructure and an average diameter of approximately 572.74 nm (Fig. S2 in Supporting information). The lattice fringes of the nanoparticles in the series of TiO₂/TiO_2−X MS were distinguishable by HRTEM, indicating that these nanoparticles were well crystallized, and the lattice fringes corresponding to the (101) plane of anatase TiO₂ are marked. At the same time, the plane spacing of the (101) crystal plane was determined using the Digital Micrograph software (Fig. S3 in Supporting information). The as-prepared TiO₂ MS without reduction treatment exhibited the clearest lattice fringes with a plane spacing of 0.341 nm [33]. The samples after reduction treatment displayed clear but irregular lattice fringes owing to the presence of surface defects. Correspondingly, the main Raman peak for the E_g at 143.83 cm⁻¹ showed a blue shift due to phonon confinement or the non-stoichiometric ratio (Fig. S4 in Supporting information) [34].

  Figure 1

  

  Figure 1.  (A) TEM images of the as-prepared TiO₂ MS and the TiO_2−X MS obtained at different reduction temperatures (inset: HRTEM images). (B) EPR spectra of the TiO₂/TiO_2−X MS (diamond: g = 2.003, club: g = 1.988). (C) Variation of the lattice strain measured by TEM (left y axis) and the normalized defect concentration measured by EPR (right y axis) with the reduction temperature.

  DownLoad: Full-Size Img PowerPoint

  Electron paramagnetic resonance (EPR) spectroscopy was employed to further characterize the crystal structure and electronic nature before and after the creation of defects in the reduced samples [35]. A weak EPR signal corresponding to oxygen vacancies at g = 2.003 was observed for the as-prepared TiO₂ MS, which originated from the natural defects present in the TiO₂ nanostructure and the surface absorption of O₂ from the air without light irradiation at room temperature (Fig. 1B). This signal gradually became more intense with increasing reduction temperature, as the number of unpaired electrons trapped by oxygen vacancies increased with the degree of reduction. At a reduction temperature of 300 ℃, a weak but broad signal corresponding to Ti³⁺ centered at g = 1.988 was observed [36]. According to the spin concentration measured by EPR, the defect concentrations were normalized to that at 380 ℃ (defined as 1.0 a.u.) to allow semi-quantitative evaluation. The dependences of the normalized defect concentration measured by EPR and the strain measured by TEM on the reduction temperature were compared, as shown in Fig. 1C. Increasing the reduction temperature led to an increasing concentration of defects, which induced the gradual expansion of the (101) crystal plane until the lattice strain reached 6.08% at a reduction temperature of 380 ℃. Similar positive correlations with respect to the reduction temperature were observed for the normalized defect concentration and the lattice strain. The accumulated lattice defects (Ti³⁺/O_V) presumably gave rise to unsaturated metal centers and disordered lattice structures, increasing the electrostatic repulsion between the Ti³⁺ cations and inducing gradual lattice expansion [37]. Furthermore, the oxygen vacancy concentrations were calculated from the EPR results as listed in Table S1 (Supporting information). Following reduction at 300 ℃, the amount of Ti³⁺ centers obtained was 3.53 × 10¹² spins/g (2.82 × 10¹⁷ spins/mol), indicating that one in every 2.37 × 10⁵ Ti atoms was converted to Ti³⁺. At this stage, other titanium oxides were not observed in the anatase TiO₂ phase, although X-ray diffraction data indicated the formation of Ti₃O₅ at higher temperatures (Fig. S5 in Supporting information) [38]. Fig. S6 (Supporting information) also shows the full width at half maximum (FWHM) values and particle sizes of the samples obtained at different reduction temperatures. The variation of the FWHM and grain size can be ascribed to the decrease in crystallinity caused by the defects, then the excessive accumulation of the defects induced a phase change to generate Ti₃O₅ [39].

  The TiO_2−X MS were also characterized by X-ray photoelectron spectroscopy (XPS) to assess the surface chemical states and detect the existence of Ti 2p bands and oxygen vacancies (Fig. S7 in Supporting information). As shown in Fig. S8 (Supporting information), detectable shoulders were observed at lower binding energies of 463.2 and 457.9 eV in the high-resolution Ti 2p spectra, which were ascribed to the 2p_1/2 and 2p_3/2 peaks of Ti³⁺, respectively [40]. A semi-quantitative linear relationship between the amounts of Ti³⁺ and Ti⁴⁺ and the reduction temperature was also obtained (Fig. S9 in Supporting information). The trend of developing Ti³⁺ content with reduction temperature was also consistent with that reflected in O 1s spectrum (Fig. S10 in Supporting information). With the increase of reduction temperature, the TiO_2−X obtained by high-temperature reduction exhibited a gradually enhanced signal corresponding to Ti³⁺–O at 530.5 eV, indicating an increasing trend in the concentration of oxygen defects [41]. The water contact angles were measured for the TiO₂/TiO_2−X MS and the results indicated a transition from hydrophobicity to hydrophilicity (Fig. S11 in Supporting information). This optimization of the surface structure and chemical composition influences the wettability, enabling more effective participation of the photocatalyst in hydrogen production [42]. The surface modification of the TiO₂/TiO_2−X MS was also confirmed by Fourier-transform infrared spectroscopy. The wide band at approximately 3422 cm⁻¹ corresponding to the symmetric stretching of O–H observed for the TiO_2−X MS indicates that the reduced samples possessed excellent hydrophilicity (Fig. S12 in Supporting information) [43].

  The photocatalytic activities of the TiO₂/TiO_2−X MS were investigated by water decomposition experiments. Fig. 2A presents plots of the hydrogen evolution over the photocatalysts under simulated sunlight as a function of time, revealing a positive linear correlation. At reduction temperatures up to 350 ℃, all of the reduced TiO_2−X MS exhibited enhanced photocatalytic activity compared with the TiO₂ MS, in which the hydrogen evolution followed the order of 300-TiO_2−X > 350-TiO_2−X > 250-TiO_2−X > 200-TiO_2−X > TiO₂ MS > 380-TiO_2−X demonstrating a volcano-shaped correlation between the hydrogen evolution and the reduction temperature. It is worth emphasizing that the hydrogen production obtained after reduction at 350 or 380 ℃ was inferior to that observed after reduction at 300 ℃. As shown in Fig. S13 (Supporting information), the TiO_2−X MS reduced at 300 ℃ yielded the highest hydrogen evolution rate of 523.7 µmol g^-1 h^-1, which was more than twofold higher than that measured for the as-prepared TiO₂ MS (253.4 µmol g^-1 h^-1). Even after performing the photocatalytic hydrogen generation experiment five times (25 h), the 300-TiO_2−X sample still maintained excellent stability and satisfactory photocatalytic activity, indicating that it can be considered a stable photocatalyst (Fig. 2B, Fig. S14 and Table S3 in Supporting information).

  Figure 2

  

  Figure 2.  (A) Comparison of the hydrogen evolution rates of the photocatalysts under simulated sunlight. (B) Recycling runs of 300-TiO_2−X for photocatalytic hydrogen generation.

  DownLoad: Full-Size Img PowerPoint

  Rhodamine B (RhB) was selected with high chroma and difficult biochemical degradation as a simulated organic pollutant to evaluate the photodegradation abilities of the catalysts. As shown in Figs. S15 and S16 (Supporting information), the reduction treatment using NaBH₄ exerted a considerable influence on the photocatalytic activity under simulated sunlight. The photocatalytic activity first increased with increasing reduction temperature to reach a maximum at a temperature of 300 ℃, which afforded an RhB degradation rate of 95.9% after 40 min of illumination, and then decreased as the reduction temperature was further increased. In addition, to study the degradation mechanism and role of free radicals in the photocatalytic process, classical capture experiments were performed to examine the effect of free radicals in the photocatalytic degradation of RhB over 300-TiO_2−X [44]. As shown in Fig. S17 (Supporting information), the contributions of these species to the photodegradation of RhB over 300-TiO_2−X decreased in the order of e⁻ > ^•OH > h⁺ > ^•O₂⁻. The dependences of the hydrogen production rate and RhB degradation rate on the defect density are plotted in Fig. S18 (Supporting information). It can be seen that the reduction temperature of 300 ℃ afforded the optimal defect density for superior photocatalytic performance (Table S4 in Supporting information).

  The above analysis revealed that the photocatalytic efficiency of the TiO₂/TiO_2−X MS first increased and then decreased with increasing defect concentration. To obtain a comprehensive understanding of the role of the defects, the changes in the electronic structure, energy band structure, optical properties, charge transfer kinetics, and recombination process were further explored.

  The influence of the oxygen vacancies on the light absorption ability of the photocatalysts was examined by UV–vis diffuse reflectance spectroscopy (DRS), as shown in Fig. 3A. All of the samples exhibited strong absorption with similar intensity in the UV region. In comparison to the TiO₂ MS, the visible-light absorption capacity increased continuously as the reduction temperature was increased from 200 ℃ to 380 ℃. The oxygen vacancies introduced during the reduction process cause the generation of defect energy levels and reduce the energy required for electronic transitions. Naturally, a higher degree of reduction is associated with a higher surface oxygen vacancy concentration and even a higher bulk oxygen vacancy concentration, thus increasing the visible-light absorption ability of TiO_2−X MS. The narrowing of the bandgap of TiO_2-X MS were confirmed by Tauc diagrams in Fig. S19 (Supporting information), in which the E_g bandgap decreased from 3.15 eV for the TiO₂ MS to 3.13, 3.11, 2.99, 2.84, and 2.65 eV for the TiO_2−X MS in order of increasing reduction temperature.

  Figure 3

  

  Figure 3.  (A) UV–vis DRS spectra. (B) Photoluminescence spectra. (C) Dependence of average carrier lifetime on defect concentration.

  DownLoad: Full-Size Img PowerPoint

  As shown in Fig. 3B, the main photoluminescence (PL) peak at approximately 420 nm originated from the band-to-band transition of anatase TiO₂. The green emission band centered at 525 nm corresponded to a superposition of trapped excitons and various defect-related emission bands [45]. The PL intensities were consistently lower for the reduced TiO_2−X MS than for the as-prepared TiO₂ MS. In general, a lower PL intensity reflects a lower photoelectron–hole recombination rate and a higher photocatalytic efficiency [46]. The lowest PL intensity was observed for the 300-TiO_2−X sample, indicating that an appropriate defect concentration and strain reduced the recombination of photoinduced carriers. It is worth noting that the 350-TiO_2−X and 380-TiO_2−X samples displayed markedly higher peak emission intensities than the 300-TiO_2−X despite their higher defect concentrations and lower crystalline qualities.

  Besides the light absorption properties, the photocatalytic efficiency of a photocatalyst also depends on the generation, separation, and transmission of the photoinduced carriers. Owing to the presence of shallow energy levels associated with defects near the conduction and valence band edges that protect the electron–hole pairs from recombination, the band edge states play a crucial role in determining the carrier properties of a material [47]. Therefore, we performed time-resolved photoluminescence (TRPL) to quantitatively investigate the lifetimes of the photoexcited charge carriers in the TiO₂/TiO_2−X MS, as shown in Fig. S20 (Supporting information). The decay curves were fitted using a biexponential decay model. The fitted data are shown in Table S2 (Supporting information), in which the fast neutron lifetime (τ₁) was ascribed to the capture of photocarriers by defects while the slow neutron lifetime (τ₂) corresponded to electron–hole recombination between the conduction band (CB) and valence band (VB) [48]. In accordance with the relationship between temperature and defect concentration, the TiO_2−X MS exhibited initial increases in both the fast and slow electron lifetimes with increasing vacancy concentration, reaching maximum carrier lifetimes of τ₁ = 1.34 ns and τ₂ = 12.13 ns at 300 ℃. Subsequently, the excessive defect concentration at higher temperatures caused decreases in the carrier lifetimes (Fig. 3C). The charge transfer properties were further evaluated by electrochemical impedance spectroscopy and the Nyquist semicircle plots for the different reduction temperatures were obtained, as presented in Fig. S21 (Supporting information). The lowest arc radius was observed for the 300-TiO_2–X sample, which proves that this catalyst possessed superior electronic conductivity for optimizing the charge transfer from the interior to the sample surface. As shown in Fig. S22 (Supporting information), the photocurrent intensity was significantly higher for the 300-TiO_2–X sample, indicating that an appropriate oxygen vacancy concentration and lattice strain contributed to enhanced charge separation and transfer efficiency. These results also illustrate that the structural modification induced by the appropriate defect concentration effectively inhibited charge recombination, thus providing direct and quantitative support for the optimal photocatalytic performance of the 300-TiO_2–X sample.

  The advantages of defects include improved light absorption and enhanced carrier separation resulting from local effects. Because light absorption is enhanced with increasing defect density, the influence of the defect density on photocatalytic efficiency was explored by analyzing the change in the localized defect density. Temperature-dependent PL spectra for TiO₂ MS, 300-TiO_2–X MS, 380-TiO_2–X MS are shown in Fig. 4A and Fig. S23 (Supporting information), and the variation of the emission peak energy positions with increasing temperature for the reduced TiO_2–X MS was found to be significantly different from that for the as-prepared TiO₂ MS. In the case of the TiO₂ MS, the emission peak energy position showed a slight shift phenomenon, while the reduced 300-TiO_2−X samples displayed a more obvious abnormal V-shaped behavior of blue-shift after obvious red-shift of PL peak position. For a typical compound semiconductor, the PL peak position exhibits a red shift with increasing temperature as expressed by Varshni’s equation [49]:

  (1)

  Figure 4

  

  Figure 4.  (A) Temperature-dependent PL spectra of the 300-TiO_2−X sample with an excitation wavelength of 325 nm after Gaussian fitting. (B) Possible reaction mechanism for solar-driven photocatalytic processes over the TiO_2−X MS.

  DownLoad: Full-Size Img PowerPoint

  where E_g is the energy gap, T is temperature, and α and β are constants. The temperature-dependent PL peak energy of the reduced TiO_2−X samples was similar to the phenomenon observed owing to localized states, resulting from defects in compound semiconductors [50]. The red shift observed for the reduced TiO_2−X sample in Fig. 4A was attributable to the thermal energy of the localized carriers being insufficient to overcome the localization potential, causing them to relax to lower-lying localized states before recombination. As the temperature was increased, the thermalized carriers underwent activation to higher energy states and hence the PL peak shifted to higher energy. Bandgap fluctuations resulting from carrier localization could be represented by the band tail states theoretically. On the basis of the band-tail model, the temperature-dependent emission energy curve could be fitted by the following relation [51]:

  (2)

  where K_B is the Boltzmann constant, E_g(T) is a regular temperature dependence of the band gap described by Eq. 1, and σ is the energy difference between the local energy level and the local center, representing the maximum energy range of the Gaussian distribution of the trailing band energy levels. The obtained values of the broadening parameter σ were 2.4, 2.3, 3.4, 3.2, and 1.6 meV for the reduction temperatures of 200, 250, 300, 350, and 380 ℃, respectively. A higher value of σ is associated with a greater degree of localization.

  To analyze the localization states, the absorption edge including the effects of broadening could be described by fitting the rising edge (α) and plateau of the absorption (α₀) [52]:

  (3)

  where E_B is the "effective bandgap" of the sample and ∆E is a broadening parameter equivalent to the localization energy due to oxygen vacancies. Eq. 3 was used to obtain the carrier localization energy level values (∆E) for the 200-TiO_2–X, 250-TiO_2–X, 300-TiO_2–X, 350-TiO_2–X, and 380-TiO_2–X samples as 0.0029, 0.028, 0.125, 0.108, and 0.075 eV, respectively. More importantly, the variation of the carrier localization energy with the defect concentration was similar to that of the efficiency of the photocatalysts containing defects, indicating that the localization state has an important effect on the catalytic performance.

  The detailed energy band structures of the photocatalysts were examined by density functional theory (DFT) calculations, low-temperature PL measurements, XPS valence band spectra, and Mott–Schottky curve analysis, and the complete band location is obtained (Figs. S24-S26 in Supporting information). The excellent photocatalytic performance of the reduced TiO_2–X MS was ascribed to the vicinity of the introduced defect level to the valence band and the more negative conduction band strengthening the redox properties.

  On the basis of the above results, a plausible reaction mechanism for the solar-driven photocatalytic processes over the TiO_2−X MS was proposed as illustrated in Fig. 4B. As the oxygen vacancy concentration increased from 1.88 × 10⁶ spins/Ti atom to 2.37 × 10⁵ spins/Ti atom (200-TiO_2–X to 300-TiO_2–X), the carrier localization value increased. At this point, the quantum-confined Stark effect dominates the carrier localization [53], which leads to enahanced separation of the photogenerated carriers, increasing electrical conductivity, resulting in the continuous improvement of photocatalytic performance. As the oxygen vacancy concentration further increased from 2.37 × 10⁵ spins/Ti atom to 6.22 × 10⁴ spins/Ti atom (300-TiO_2–X to 380-TiO_2–X), it can be seen that the half-height width of Luminous peak measured at same temperature widened and Luminescent peak position red shift gradually (Fig. S23B). The weakening of the Stark effect is particularly important, in addition to the decrease in crystal quality due to the increase in the concentration of oxygen vacancies. At this point, taking into account the increase in localised states, the resonance attempts of electrons and microstrip transport allow for "forbidden" transitions, the decrease in carrier localisation, and higher defect concentrations lead to an increase in the number of non-radiative complex centres, which results in enhanced light absorption and reduced carrier localisation, ultimately leading to a decrease in photocatalytic performance [54].

  In summary, a series of TiO_2−X MS with various oxygen vacancy concentrations were prepared by a simple high-temperature reduction method using sodium borohydride. Compared with the as-prepared TiO₂ MS, the reduced TiO_2−X MS exhibited excellent hydrogen production performance. In particular, the TiO_2−X MS obtained at a reduction temperature of 300 ℃ afforded a photocatalytic hydrogen evolution rate of 523.7 µmol g^-1 h^-1 under simulated sunlight and an RhB degradation efficiency as high as 95.9% after 40 min. Mechanistic analysis revealed that significant lattice strain occurred at the TiO_2−X containing oxygen defects, and electrons gather around oxygen atoms to modulate the electronic structure, and the doping of oxygen defects improves the surface hydrophilicity by increasing the humidity stability. More importantly, the Stark effect induced by the introduction of oxygen vacancies in the defect area enhances light absorption and promotes efficient carrier separation. This effect creates a conducive environment for photocatalysis. However, as the number of oxygen vacancies continues to increase, the Stark effect gradually weakens, leading to a decrease in photocatalytic efficiency. This work provides new insights for defect engineering and lattice strain modulation of photocatalyst surfaces and band structures to assist with the preparation of efficient photocatalytic systems for both hydrogen production and pollutant degradation.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Bingke Zhang: Writing – original draft, Methodology. Dongbo Wang: Writing – review & editing, Conceptualization. Jiamu Cao: Investigation. Wen He: Methodology. Gang Liu: Data curation. Donghao Liu: Methodology, Conceptualization. Chenchen Zhao: Formal analysis. Jingwen Pan: Formal analysis. Sihang Liu: Data curation. Weifeng Zhang: Investigation. Xuan Fang: Conceptualization. Liancheng Zhao: Supervision. Jinzhong Wang: Supervision, Funding acquisition.

  Acknowledgments

  This work was supported by the National Key Research and Development Program of China, (No. 2019YFA0705201), and the National Natural Science Foundation of China (No. U2032129).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110254.

  
  
• 1. [1]

     H. Nishiyama, T. Yamada, M. Nakabayashi, et al., Nature 598 (2021) 304–307. doi: 10.1038/s41586-021-03907-3

  2. [2]

     Z.Z. Li, S.J. Wang, J.X. Wu, W. Zhou, Renew. Sust. Energy Rev. 156 (2022) 111980. doi: 10.1016/j.rser.2021.111980

  3. [3]

     L.B. Xie, L.L. Wang, W.W. Zhao, et al., Nat. Commun. 12 (2021) 5070. doi: 10.1038/s41467-021-25381-1

  4. [4]

     W.X. Huang, Z.L. Wu, J.W. Tang, et al., Chin. Chem. Lett. 29 (2018) 725–726. doi: 10.1016/j.cclet.2018.05.021

  5. [5]

     B.J. Sun, W. Zhou, H.Z. Li, et al., Adv. Mater. 30 (2018) 1804282. doi: 10.1002/adma.201804282

  6. [6]

     Y.J. Zhang, Z.F. Xu, G.Y. Li, et al., Angew. Chem. Int. Ed. 58 (2019) 14229–14233. doi: 10.1002/anie.201907954

  7. [7]

     J. Chen, Y.M. Tang, L.L. Wang, et al., Chin. Chem. Lett. 33 (2022) 1468–1474. doi: 10.1016/j.cclet.2021.08.103

  8. [8]

     X.B. Chen, L. Liu, P.Y. Yu, et al., Science 331 (2012) 746–750.

  9. [9]

     W. Zhou, W. Li, J.Q. Wang, et al., J. Am. Chem. Soc. 136 (2014) 9280–9283. doi: 10.1021/ja504802q

  10. [10]

      K. Chen, J.A. Pan, W.A. Yin, C.Y. Ma, L.L. Wang, Chin. Chem. Lett. 34 (2023) 108226. doi: 10.1016/j.cclet.2023.108226

  11. [11]

      Z.Z. Li, H.Z. Li, S.J. Wang, F. Yang, W. Zhou, Chem. Eng. J. 427 (2022) 131830. doi: 10.1016/j.cej.2021.131830

  12. [12]

      N.W. Cao, M.Z. Gu, M.M. Gao, et al., Appl. Surf. Sci. 530 (2020) 147289. doi: 10.1016/j.apsusc.2020.147289

  13. [13]

      P.C. Fan, Y.H. He, J.A. Pan, et al., Chin. Chem. Lett. 35 (2024) 108513. doi: 10.1016/j.cclet.2023.108513

  14. [14]

      Y. Cao, Z.P. Xing, W. Zhou, et al., J. Catal. 356 (2017) 246–254. doi: 10.1016/j.jcat.2017.10.023

  15. [15]

      Y. Zhang, Z.Y. Zhao, S.S. Cao, et al., Appl. Catal. B: Environ. 165 (2015) 715–722. doi: 10.1016/j.apcatb.2014.10.063

  16. [16]

      B.J. Sun, J.Q. Bu, W. Zhou, et al., ACS Appl. Mater. Interfaces 13 (2021) 37545–37552. doi: 10.1021/acsami.1c10943

  17. [17]

      T. Seadira, G. Sadanandam, M. Scurrell, et al., Appl. Catal. B: Environ. 222 (2018) 133–145. doi: 10.1016/j.apcatb.2017.09.072

  18. [18]

      T.S. Zhou, J.C. Wang, S. Chen, et al., Appl. Catal. B: Environ. 267 (2020) 118599. doi: 10.1016/j.apcatb.2020.118599

  19. [19]

      H. Li, X. Li, X.J. Zhou, W.L. Sheng, X.J. Lang, Chin. Chem. Lett. 33 (2022) 3733–3738. doi: 10.1016/j.cclet.2021.10.068

  20. [20]

      Q.Y. Huang, L.L. Wang, B.B. Xiao, et al., Chin. Chem. Lett. 34 (2023) 108046. doi: 10.1016/j.cclet.2022.108046

  21. [21]

      K. Murata, Y. Mahara, A. Satsuma, et al., Angew. Chem. Int. Ed. 129 (2017) 16209–16213. doi: 10.1002/ange.201709124

  22. [22]

      X.Y. Mao, A.C. Foucher, T. Montini, et al., J. Am. Chem. Soc. 142 (2020) 10373–10382. doi: 10.1021/jacs.0c00138

  23. [23]

      Y.X. Zhao, Y.F. Zhao, R. Shi, et al., Adv. Mater. 31 (2019) 1806482. doi: 10.1002/adma.201806482

  24. [24]

      G.L. Zhu, Y.F. Shan, T.Q. Liu, et al., Nanoscale 8 (2016) 4705–4712. doi: 10.1039/C5NR07953E

  25. [25]

      A. Naldoni, M. Allieta, V. Santo, et al., J. Am. Chem. Soc. 134 (2012) 7600–7603. doi: 10.1021/ja3012676

  26. [26]

      F.S. Wang, L.B. Xie, N. Sun, et al., Nano-Micro Lett. 16 (2023) 1.

  27. [27]

      Y.Q. Zeng, Y.N. Wang, S.L. Zhang, Q. Zhong, Phys. Chem. Chem. Phys. 20 (2018) 22744–22752. doi: 10.1039/C8CP02270D

  28. [28]

      Y.X. Hu, Y.Y. Pan, Z.L. Wang, et al., Nat. Commun. 11 (2020) 2129. doi: 10.1038/s41467-020-15993-4

  29. [29]

      Y. Yin, J.C. Han, S. Jin, et al., J. Am. Chem. Soc. 138 (2016) 7965–7972. doi: 10.1021/jacs.6b03714

  30. [30]

      M. Pelaez, N. Nolan, S. Pillai, et al., Appl. Catal. B: Environ. 125 (2012) 331–349. doi: 10.1016/j.apcatb.2012.05.036

  31. [31]

      J.M. Cai, M.Q. Wu, Y.T. Wang, et al., Chem 2 (2017) 877–892. doi: 10.1016/j.chempr.2017.05.006

  32. [32]

      B.B. Dong, T.F. Liu, C. Li, F.X. Zhang, Chin. Chem. Lett. 29 (2018) 671–680. doi: 10.1016/j.cclet.2017.12.002

  33. [33]

      H. Zhang, Y. Zhang, J.W. Yin, et al., Chem. Phys. 5134 (2018) 86–93.

  34. [34]

      T. Wang, W. Li, D. Xu, et al., Appl. Surf. Sci. 426 (2017) 325–332. doi: 10.1016/j.apsusc.2017.07.153

  35. [35]

      A. Naldoni, M. Altomare, G. Zoppellaro, et al., ACS Catal. 9 (2019) 345–364. doi: 10.1021/acscatal.8b04068

  36. [36]

      L. Grabstanowicz, S. Gao, T. Li, et al., Inorg. Chem. 52 (2013) 3884–3890. doi: 10.1021/ic3026182

  37. [37]

      Y. Li, B. Yu, B. Liu, et al., Chem. Eng. J. 452 (2023) 139542. doi: 10.1016/j.cej.2022.139542

  38. [38]

      J. Park, D. Kim, S. Latthe, et al., ACS Appl. Mater. Interfaces 13 (2013) 6155–6160.

  39. [39]

      F. Göde, C. Gümüs, M. Zor, J. Cryst. Growth 299 (2007) 136–141. doi: 10.1016/j.jcrysgro.2006.10.266

  40. [40]

      J.Q. Pan, Z.J. Dong, Z.Y. Jiang, et al., Sol. Rrl. 3 (2019) 1900337. doi: 10.1002/solr.201900337

  41. [41]

      S. Oh, J. Kim, Y. Lee, et al., J. Mater. Chem. A 9 (2021) 4822–4830. doi: 10.1039/D0TA11505C

  42. [42]

      J. Park, D. Kim, S. Yoon, et al., ACS Appl. Mater. Interfaces 5 (2013) 6155–6160. doi: 10.1021/am401083y

  43. [43]

      W.F. Hu, L.B. Xie, C. Gu, et al., Coordin. Chem. Rev. 506 (2024) 215715. doi: 10.1016/j.ccr.2024.215715

  44. [44]

      X.P. He, Y.N. Lu, T. Cai, et al., J. Hazard. Mater. 457 (2023) 131842. doi: 10.1016/j.jhazmat.2023.131842

  45. [45]

      L.P. Kong, G. Liu, J. Gong, et al., Proc. Natl. Acad. Sci. U. S. A. 113 (2016) 8910–8915. doi: 10.1073/pnas.1609030113

  46. [46]

      Y.Q. Xiao, C. Feng, J. Fu, et al., Nat. Catal. 3 (2020) 932–940. doi: 10.1038/s41929-020-00522-9

  47. [47]

      S.M. Ni, F.Y. Guo, D.B. Wang, et al., ACS Sustainable Chem. Eng. 6 (2018) 7265–7272. doi: 10.1021/acssuschemeng.7b04188

  48. [48]

      C.Z. Lin, Y.P. Liu, Y.X. Sun, et al., Chin. Chem. Lett. 34 (2023) 108265. doi: 10.1016/j.cclet.2023.108265

  49. [49]

      Y.P. Varshni, Physica 34 (1967) 149–154. doi: 10.1016/0031-8914(67)90062-6

  50. [50]

      H.C. Wang, S.C. Liu, Y.C. Liu, et al., Appl. Phys. Lett. 85 (2004) 1371–1373. doi: 10.1063/1.1784033

  51. [51]

      D.B. Wang, S.J. Jiao, L.C. Zhao, et al., J. Phys. Chem. C 117 (2013) 543–548. doi: 10.1021/jp3088429

  52. [52]

      R. Martin, P. Middleton, K. O’Donnel, Appl. Phys. Lett. 74 (1999) 263–265. doi: 10.1063/1.123275

  53. [53]

      C.C. Zhao, D.B. Wang, J.M. Cao, et al., Nano. Res. 17 (2024) 1864–1874. doi: 10.1007/s12274-023-5905-6

  54. [54]

      M. Piccardo, C.K. Li, Y.R. Wu, et al., Phys. Rev. B 95 (2017) 144205. doi: 10.1103/PhysRevB.95.144205

• 

  Figure 1  (A) TEM images of the as-prepared TiO₂ MS and the TiO_2−X MS obtained at different reduction temperatures (inset: HRTEM images). (B) EPR spectra of the TiO₂/TiO_2−X MS (diamond: g = 2.003, club: g = 1.988). (C) Variation of the lattice strain measured by TEM (left y axis) and the normalized defect concentration measured by EPR (right y axis) with the reduction temperature.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (A) Comparison of the hydrogen evolution rates of the photocatalysts under simulated sunlight. (B) Recycling runs of 300-TiO_2−X for photocatalytic hydrogen generation.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (A) UV–vis DRS spectra. (B) Photoluminescence spectra. (C) Dependence of average carrier lifetime on defect concentration.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (A) Temperature-dependent PL spectra of the 300-TiO_2−X sample with an excitation wavelength of 325 nm after Gaussian fitting. (B) Possible reaction mechanism for solar-driven photocatalytic processes over the TiO_2−X MS.

  下载: 全尺寸图片 幻灯片
•