Machine learning and DFT dual-guidance of carbon dots implanted SrTiO3 hollow nanosphere for efficient all-pH-value photocatalysis

The photo-Fenton-like reaction holds significant promise for treating low-concentration antibiotic wastewater, yet its progress is hindered by the weak adsorption and activation ability of peroxydisulfate (PDS) and limited pH application range. This work employs density functional theory and machine learning models in tandem to craft optimal photocatalysts for tetracycline (TC) degradation across all pH ranges. Our investigation reveals that carbon dots (CDs) modified SrTiO₃ hollow nanospheres exhibit robust PDS adsorption and activation capabilities, facilitating electron transfer from the photocatalyst to SO₄^•-. Additionally, the abundant functional groups on CDs confer a protective effect on SrTiO₃, shielding it from corrosion by strong acids and bases. Consequently, CDs-SrTiO₃ demonstrates excellent photocatalytic performance across the entire pH spectrum, particularly in alkaline and extremely acidic conditions. Furthermore, CDs deposition enhances the solar utilization, specific surface active site amount, and hydrophilicity of SrTiO₃. Theoretical calculations and experimental characterizations elucidate the degradation mechanism and pathways of TC, while machine learning models optimize the experimental parameters. This work provides valuable insights into the rational design and adept preparation of high-quality photocatalysts suitable for a wide pH range towards wastewater treatment.