Modeling and AI optimization of visible-light-driven acid orange 25 degradation using an Ag/N/ TiO2/persulfate system

Industrial wastewater contamination from azo dyes is a serious environmental and public health concern that requires efficient and sustainable treatment technologies. In this study, the synthesis, comprehensive characterization, and enhanced photocatalytic performance of Ag/N/TiO 2 for the degradation of Acid Orange 25 in the presence of persulfate under visible and solar light were investigated. Structural analyses (XRD, FTIR, FESEM-EDX, DRS) confirmed uniform synthesis, nitrogen incorporation, effective photodeposition of silver, and spherical particles (< 30 nm). Band gap narrowing from 3.17 to 2.54 eV indicated improved visible-light absorption due to nitrogen and silver doping. AI optimization showed that, at pH 4.28, 0.7 g L −1 photocatalyst, and 2.22 mM persulfate, nearly complete degradation of 94 mg L −1 Acid Orange 25 could be achieved using Ag/N-TiO 2 in the presence of persulfate within 32 min. Radical scavenging experiments revealed that positive holes (h + ) played the dominant role, with the contribution of reactive species following the order: h + > SO 4 –· > · OH. The kinetic results exhibited pseudo-first-order behavior and were consistent with the Langmuir–Hinshelwood model within the concentration range of 20–80 mg L −1 . The quadratic model provided the best fit to the experimental data, and Sobol sensitivity analysis ranked dye concentration, reaction time, pH, photocatalyst dosage, and persulfate concentration as the most influential parameters, with the strongest interaction observed between pH and dye concentration. Monte Carlo simulation confirmed the robustness of the optimized model under 5% uncertainty. These findings demonstrate the potential of Ag/N/TiO 2 in the presence of persulfate for efficient azo dye degradation under visible-light and sunlight irradiation.