Rapid development and the transition to a carbon-neutral society have highlighted the need for photovoltaic cells to use high-efficiency photocatalysts under visible light (>380 nm) (Liu et al., 2016; Galvao et al., 2018; González-Gutiérrez et al., 2019). Transition metal oxides (such as TiO₂ and SnO₂) have attracted considerable attention from researchers due to their excellent photocatalytic activity, high natural abundance, and cost-effectiveness (Sharon et al., 2016). Owing to two oxides with different band-gap energies (E _g) of 3.2 eV for TiO₂ and 3.8 eV for SnO₂ (Nuansing et al., 2006), respectively, the enhanced charge separation occurs in their mixture where electrons are injected from the conduction band of SnO₂ to the valence band of TiO₂, leading to better performance than individual semiconductor oxides due to their synergistic effects. To better regulate their band gap and obtain high efficiency, more efforts should be focused on building a desired heterostructure and tuning electron correlation by varying the stoichiometry of oxides Luo et al. (2018). Heterostructures were mainly fabricated by the assembly of two or more oxides that were physically (e.g., Van der Wall’s force) or chemically bonded together, exhibiting superiority of geometry and junction interface (Li et al., 2021; Liu et al., 2023). For heterostructure-based metal oxide, element doping coupled with hierarchy design achieved visible benefits due to the formation of a single phase rather than a combination of two oxide compounds. For example, the pre-synthesized SnO₂ quantum dots directed growth on titania show slow charge transfer, low active-sensitizer loading and poor uniformity control for deposition on high aspect-ratio TiO₂ nanostructures (Wang et al., 2022). Doping titania with either inorganic or metallic species can generate donor or acceptor states in the band gap of TiO₂ and impurity states that enhance visible-light adsorption (Albornoz et al., 2021; Wu et al., 2022). Non-etheless, metal oxides with doping heterostructure often suffered from thermal instability, fast recombination rates, or the requirement of an expensive ion-implantation facility for more elaborate control, and even showed much larger formation energy due to the substitution of large-radius metal ions (Singh et al., 2021; Liu et al., 2022). Many essential questions remain, such as the effect of the structure-function correction, composition, and interfaces on the photocatalytic activity, which remains challenging.

One promising strategy to obtain a heterostructure-based titania photocatalyst is to fabricate a substitutional solid solution using metal atoms with a comparable ionic radius (VI-coordinated) with titania (Reddy et al., 2008). Such a metal-doped TiO₂ solid not only retains the benefits from the quantum structures of crystalline titania during photocatalytic processes but further enhances photoinduced surface redox reactions one. Compared with the other cationic doping (such as Cr, V, and Fe), Sn ion shows a more favorable effect on photocatalytic activity in water-splitting and oxidation of organic compounds (Liang et al., 2012). To date, few examples of tin-doping titania (denoted as Ti_1-xSn_xO₂, 0 < x < 1) have been reported, especially on nanoscale systems (Asokan et al., 2010). Such compounds have been generally synthesized by solid-state reaction at high temperatures beyond 1,200°C or chemical deposition technology, such as hydrothermal (Chen et al., 2022; Zhao et al., 2016; Ruiz Esquius et al., 2020), sol-gel (Pudukudy et al., 2017; Liu et al., 2019), electrospinning (Yu et al., 2018), and polyol-mediated methods (Pelicano et al., 2022). However, these methods are limited because of the difficulty in elaborate control of doping contents and nanoparticle aggregation and the incapability of constructing hierarchical pore structures, leading to an increased probability of recombination. Besides, most compounds prepared using the above-mentioned methods have a rutile tetragonal crystalline structure with no anatase analog due to the existence of Sn (Qin et al., 2021). The latter has been proposed to have higher photoactivity than the former, but it is still limited by a less Sn-doping content (<0.3).

To address this challenge, we report a facile and effective doping method to prepare a hierarchically mesoporous Ti_1-xSn_xO₂ compound for high-performance photocatalyst under visible light irradiation. Inspired by the Ziegler-Natta catalyst mechanism, Figure 1 illustrates an inorganic polycondensation process that occurs between titanium chloride and tin chloride liquid in a liquid nitrogen bath. The ultra-low temperature effectively suppresses chloride salt hydrolysis and causes chloride ions from tin chloride to fill the empty orbital of a 6-coordinated Ti atom. Then, Ti atoms and Sn atoms were bridged to two chlorine atoms to obtain successive -Ti-Cl-Sn-Cl-Ti- chains, allowing for high Sn-contents. Subsequently, the mesoporous Ti_1-xSn_xO₂ spheres are synthesized using a low-cost aerosol-spray method. Ti_1-xSn_xO₂ nanocrystals generated from such a preformed aqueous solution are self-assembled into mesoporous spheres. Such a hierarchical heterostructure endows several advantages to the Ti_1-xSn_xO₂ compound. Firstly, mesoporous Ti_1-xSn_xO₂ spheres inherit the anatase crystalline structure from titania and small nanocrystal size, which achieves a high surface area and many active sites. Second, hierarchically porous architecture coupled with high Sn-contents can effectively promote visible-light harvest and high quantum yield. Finally, the mesoporous structure throughout the spheres endows effective transport of charge carriers, allowing for longer lifetimes of the photoelectron-hole pairs and less recombination. Such a strategy could be easily extended to other photocatalyst designs and large-scale fabrication.

Figure 2 compares the crystallinity of as-synthesized Ti_1-xSn_xO₂ nanocrystals with various Sn-doped contents. The diffraction pattern of as-synthesized TiO₂ nanocrystals match well to that of anatase structure with intense (101) (004), and (200) reflections, corresponding to the 25.4°, 37.8° and 48.1° 2θ angles (JCPD No. 04-0477). Consistently, as-synthesized SnO₂ nanocrystals show crystallized rutile structure with well-defined peaks at 25.3° and 27.4°corresponding to (101) and (110) plane (JPCD No. 84-1286), respectively. Upon increasing Sn-doped, the reflection of Ti_1-xSn_xO₂ compounds (x ≤ 0.05) almost retain the anatase phase but a shift toward lower 2θ angels (∼0.1°) at (101) peak position. While Ti_1-xSn_xO₂ (0.05 < x ≤ 0.4) exhibit two crystalline structures with strong anatase and weak rutile, indicative of crystalline phase transition. For Ti_1-xSn_xO₂ (0.6 < x ≤ 0.8), their characteristic peak positions are consistent with rutile SnO₂ (110) peaks at 26.6° and anatase TiO₂ reflection gradually disappeared.

To further confirm the Ti_1-xSn_xO₂ formed in a single crystalline phase, Figure 3 compares experimental lattice parameters (a, c) and unit cell volume (V = ac ²) of the anatase Ti_1-xSn_xO₂ (x ≤ 0.05) and the rutile Ti_1-xSn_xO₂ (0.5 < x ≤ 0.8) according to Vegard’s law (Kong et al., 2019). The Vegard’s law is an approximate empirical rule that predicts the relationship between lattice constant and the concentration of the constituent elements solid solution, as describe as: a = x 1 a 1 + 1 − x 1 a 2 (1) where a is the lattice parameter for the solid solution, a ₁ and a ₂ are lattice parameters of individual oxide, and x ₁ is the mole fraction of metal ions. Figure 3A shows the linear trend of lattice parameters a for the Ti_1-xSn_xO₂ compounds, indicating that the hetero-Sn atom-doped into the TiO₂ lattice brings the lattice expansion. Similarly, the lattice parameter c of rutile Ti_1-xSn_xO₂ with high Sn-contents (0.5 < x ≤ 0.8) retains the identical trend, as shown in Figure 3B. In contrast, the lattice parameters c of anatase Ti_1-xSn_xO₂ with low concentration of Sn-contents shows linearly decrease, implying large anisotropy compressibility occurs throughout anatase TiO₂ crystal. The different compressibility (κ, κ _a > κ _c) causes more denser cell lattice with increasing Sn contents, which is consistent with impact of the external stress on solid solution structure (Prasad and Mikula, 2000). Additionally, the phase separation may cause certain positive deviate on due to spinodal deposition. Table 1 summarized the anatase-rutile phase distribution of TiO₂, Ti_1-xSn_xO₂ (0 < x < 1), and SnO₂ according to calculation of the lattice parameters and mass fraction of rutile phase. S n 2 + + O 2 + 2 H 2 O + 2 e − → S n 4 + + 4 O H − (2) where W _R is the mass fraction of rutile phase, I _R is correction factor which indicates the difference of scattering intensities between crystalline structures, and I _A and I _R represent the integrated intensities of the anatase (101) and rutile (110) reflection, respectively. The mass fraction of rutile phase continuously increases from 14.8 wt% for Ti_0.9Sn_0.1O₂ to 60.6 wt% for Ti_0.4Sn_0.1O₂. Indeed, Sn atoms in TiO₂ lattices could provide large number of dissimilar boundaries, which would suppress the crystal growth of TiO₂ in the dealloying process, finally leading to extremely fine nanostructure in TiO₂-SnO₂ solid solution (Sayılkan et al., 2008; Wu et al., 2017). Hence, such a series of Ti_1-xSn_xO₂ (0 < x < 1) compounds underwent pure anatase, mixed anatase-rutile to pure rutile phases, respectively, which provide a fundamental understanding of the relationship between electronic structure and dye-sensitized photocatalytic degradation.

The morphology and elements analysis of Ti_1-xSn_xO₂ (0 < x < 1) compounds obtained after self-assembly are characterized by scanning electron microscope (SEM) combined with energy-dispersive X-ray (EDX), as shown in Figures 4A,B. All samples show similarly spherical morphology with a diameter range from 150 to 500 nm, exemplified by Ti_0.8Sn_0.2O₂. The bestrewed pot-holes on the spheres surface are also observed arising from the stacking of hard-sphere crystals. The uniform distribution of each element suggests that the presence of Ti and Sn atoms throughout crystalline structure. Transmission electron microscopic (TEM) image clearly displays the as-sprayed Ti_0.8Sn_0.2O₂ particles with non-porous spherical morphology before sintering (Figure 4C). After sintering, the compounds become highly porous due to removal of surfactant, and each sphere is composed of uniform nanocrystals with diameter around 5–6 nm. High-resolution TEM (HRTEM) in Figure 4D shows the ultrafine nanocrystals with an interlayer spacing of 2.32 Å, corresponding to the (200) plane of rutile TiO₂ (∼2.29 Å). This result is consistent with the XRD patterns, showing initial extension of c axis occurs during the anatase-rutile conversion at higher Sn-contents doping. Such hierarchically porous design and highly crystalline structure can provide high surface area and fast kinetics, which favors the transport of charge carriers and prevent the recombination of photoelectron-hole pairs.

Nitrogen sorption experiment was further conducted to probe the pore structure. Figure 5 shows the nitrogen sorption isotherms and pore size distribution of anatase Ti_0.8Sn_0.2O₂ solid solution with rutile-phased fraction of 20 wt%, confirming a hierarchical pore structure including micropore, mesopore, and macropores arising from particles aggregation. The significant nitrogen uptake at the relative pressures at 0.4–0.6 is consistent with a mesopore network with a narrow pore distribution of ∼10 nm (inset). Table 2 summarizes the surface area and pore structure parameters of Ti_1-xSn_xO₂ (0 < x ≤ 0.4). Upon increasing the Sn-doped contents, the surface areas of Ti_1-xSn_xO₂ compounds shows significantly decrease from 77.58 m²·g⁻¹–44.96 m²·g⁻¹ due to introduction of heavier Sn atoms. Additionally, the average pore volumes of Ti_1-xSn_xO₂ maintain approximately ∼0.15 cm²·g⁻¹.

To further understand the electronic properties depended transformation between anatase and rutile phase, the elemental composition and valence state in anatase Ti_0.95Sn_0.05O₂ and anatase-rutile Ti_0.8Sn_0.2O₂ are conducted by X-ray photoelectron spectra (XPS), as shown in Figure 6. Figure 6A shows the Ti 2p fitting curves for, which presents two Ti 2p peaks at E = 458.2 eV and E = 463.7 eV for Ti 2p _3/2 and Ti 2p _1/2 lines, respectively. Which is consistent with the values of Ti⁴⁺ reported previously (Fan et al., 2021). Meanwhile, Figure 6B represents typical Sn 3d curves for Ti_0.95Sn_0.05O₂ and Ti_0.8Sn_0.2O₂ with both peaks at 494.2 eV and 485.9 eV for Sn 3d _3/2 and Sn 3d _5/2 lines, interpolating peak positions of the metallic Sn and SnO₂ at 493.1/495.0 eV and 484.7/486.4 eV for Sn 3d _3/2/Sn 3d _5/2 lines, which can be ascribed to the Sn⁴⁺ species. These results indicate no changes in valence state of Sn and Ti during phase transformation, implying the distorted degree in tetragonal structure is neglected at low Sn concentration. Together with XPS analysis, Table 3 compares the atomic concentration ratios of O 1 s and Ti 2p for undoped TiO₂, Ti_0.95Sn_0.05O₂ and Ti_0.8Sn_0.2O₂. The concentration ratio of Ti 2p to O 1 s decrease with increasing Sn contents, suggesting the Sn-doped in the tetragonal TiO₂ lattice for formation of Ti-O-Sn bonding.

To demonstrate this intricate structure-function correction, we evaluate the photoactivity of the porous Ti_1-xSn_xO₂ (0.05 ≤ x ≤ 0.4) nanoparticles in degradation of RhB dye under visible light irradiation, as shown in Figure 7. The blank experiment showed that the concentration of RhB solution without catalyst maintain during 90 min, indicating the RhB molecules are relatively stable and hard to decompose. Before photocatalytic degradation, it is necessary to the eliminate an adsorption effect by stirring the mixtures in the dark for 1 h. Then the Ti_1-xSn_xO₂ compounds were added in RhB solution being treated in the dark for another 60 min. All the samples have strong physical adsorption abilities, but porous solid solution can enhance adsorption ability than the blank samples. After adsorption, the next photocatalytic degradation experiments kept the RhB concentration at equilibrium as the starting point. Upon increasing Sn-doped contents, the Ti_0.9Sn_0.1O₂ and Ti_0.8Sn_0.2O₂ solutions with mixed anatase-rutile phases exhibit the higher photoactivity (RhB degradation yield ≈50% and 70%, respectively after irradiation for 120 min) than that of the anatase Ti_0.95Sn_0.05O₂ solution (RhB degradation yield <10% after irradiation for 120 min). However, the photocatalytic activity of Ti_0.6Sn_0.4O₂ with higher rutile phased of 60 wt% decrease with increasing Sn content because the RhB degradation yield is only about 15% after irradiation for 120 min.

In the heterojunction composed of TiO₂ and SnO₂, the photogenerated electrons can transfer from TiO₂ to SnO₂ phase with no applied voltage (Fan et al., 2016). Because of their identical rutile phase, tetragonal crystal structures, ionic radius (0.690 for Sn⁴⁺ and 0.605 for Ti⁴⁺), and electronic characteristics, SnO₂ and TiO₂ may also form solid solutions coupling system (Sayılkan, 2007; Huang et al., 2015). Thus, it is very essential to probe the relationship between electronic structure and photoactivity in Ti_xSn_1-xO₂ solutions. Figure 8 illustrates the dependence of band gap of Ti_1-xSn_xO₂ (0 ≤ x ≤ 1) with increasing Sn-doping. Theoretical considerations by Long et al. suggest that for substitutionally Sn-doped anatase TiO₂, the band gap increases ∼0.13 eV owing to high formation energy while maintain anatase phase (Long et al., 2009). However, the substitution of Sn for rutile TiO₂ brings a red-shift of the adsorption edge deriving from the decrease of Sn 5s gap state in the conduction band (Chang et al., 2019). Consistent with the theoretical analysis, the experimental band gaps of the anatase Ti_1-xSn_xO₂ (x = 0.05 and 0.025, respectively) solid solutions indeed increase to 3.22 and 3.29 eV, respectively (Ren et al., 2020). High Sn-doped contents causes a transition from anatase to rutile phase, and the band gap of solution with mixed phase should be reasonably neutralized in between that of single phase using their mass fraction (Chen et al., 2010; Mohammad et al., 2021). However, it is worth noting that the band gap values of Ti_0.9Sn_0.1O₂ and Ti_0.8Sn_0.2O₂ decreases to 2.98 and 2.92 eV, respectively. This is possibly attributed to transition of photoexcited electrons and holes from rutile to energy states of anatase phase, which needs very low energy about 0.2∼0.25 eV. Meanwhile, we discovered that the presence of crystalline interface within mixed phases effectively reduces electron-hole recombination and creates more oxygen vacancies to enhance visible-light harvesting ability.

Continuation of this process gradually separates photoexcited electron-hole pairs, sequentially providing more photo-charges needed for the photocatalytic reaction and fast charge migration, as shown in Scheme of Figure 9. The rapid increased band gap values of Ti_1-xSn_xO₂ (0.4 ≤ x ≤ 0.8) is attributed to the more Sn doping, leading the conduction band to shift toward a higher energy, which accounts for the observed decreasing photocatalytic activity at a higher Sn content (e.g., x ≥ 0.8). Rather than facilitate charge migration and reduce charge recombination, the charge trapping may take effect as centers for electro-hole recombination and absorb the visible light, described as follow reaction 1 and 2 (Golestanbagh et al., 2018; Shi et al., 2007): T i 4 + + e − → T i 3 + (1a) S n 4 + + 2 e − → S n 2 + (2a) O 2 − → O + V o + 2 e − (3)

The visible-spectra sensitization of Sn-doped TiO₂ is attribute from charge transfer bands of peroxotitanium or the formation of Sn II) with stable lower oxidation state (Serpone., 2006). The loss of O atoms creates the formation of color center due to the charge trapped in cavity (V_o) left behind (reaction 3). Such the impacted effects on visible-light driven reactivity for color centers are even more important than the narrowing band gap caused by doping Bjelajac et al. (2018).

For the improved spectral response in the visible region, the Ti_0.8Sn_0.2O₂ solution also shows relatively stable photodegradation performance, as shown in Figure 10. The preformed powder is centrifuged and collected after every irradiation cycle and subsequently dispersed in a fresh RhB aqueous solution without any further purification. Compare with the first cycle, the photocatalyst displays a less than 10% decrease in activity after seven cycle times, endowing a cycling stability for solution with a mixed phase.

In summary, a facile and effective aerosol-spray strategy is developed to prepare high-performance photocatalyst by fabricating mesoporous Ti_1-xSn_xO₂ (0 < x < 1) solid solution. Ti_1-xSn_xO₂ nanocrystals with various Sn-doped contents are self-assembled into mesoporous spheres provide effective transport of charge carrier. As-prepared hierarchical architecture shows several advantages as a visible-light photocatalyst. First, mesoporous Ti_1-xSn_xO₂ spheres enhance the advantageous features of conventional titania and tin oxides. Second, heterostructure-based titania compounds further enhance photoinduced surface redox reactions and kinetic process in the visible region. Third, with increasing Sn-doped concentration, the mechanism between the anatase-rutile phase transition of Ti_1-xSn_xO₂ compounds and photocatalytic activity is further revealed. Therefore, as-prepared mesoporous Ti_1-xSn_xO₂ solid solution can effectively enhanced charge separation, and accelerated proton mass transfer. This demonstration of creating solid solutions offers a viable, cost-efficient method for the photocatalytic degradation of RhB, and this prototype has the potential to inspire the development of novel photocatalysts, such as splitting water, reducing CO₂, and fixing N₂.