Machine Vision-Assisted Selective Adsorption and Photothermal Catalytic Activity from Tunable Built-in Electric Field of Ti³⁺, N, S-Codoped Hollow TiO₂ Nanostructures

To modulate an electric field and oxygen vacancies within the same semiconductor instead of different heterojunctions, using hollow inlaid TiO2 magic dices as a model, metal ions Ti3+ and nonmetal elements N and S were simultaneously doped into the sole semiconductor through the one-pot ultrasonic solvothermal method. By adjusting the reaction conditions, including the selection of alcohols, the ratio of ethylene glycol to acetic acid, and the content of methionine, a large-surface-area and mesoporous hollow inlaid magic dices structure was successfully constructed and fully characterized. On the catalyst surface, due to the uneven charge distribution caused by cationic and anion species and multihierarchical mesoporous morphology, it was found that there was great adsorption capacity and selective adsorption properties for different types of dyes with assisted machine vision monitoring after thousands of machine learning. Mechanism and kinetic investigations indicated that the selective adsorption of RhB on 0.25TNS obeyed the Langmuir isothermal adsorption model with the pseudo-second-order equation. Inside the catalyst lattice, codoping of cation and anion ions accompanied by abundant oxygen vacancies led to the formation of a tunable internal built-in electric field with the doping amounts and a Ti–S–Ti–N-Ti electronic bridge, accelerating the separation of charge carriers and significantly enhancing the photothermal catalytic activity. This study provides important insights and a method for the simultaneous doping of anions and cations into the same substance to form a tunable built-in electric field.