Abstract

This work introduces a novel photocatalytic methodology that leverages the synergistic action of visible light and H₂O₂ with a green catalyst for the efficient degradation of tris(1-chloro-2-propyl) phosphate (TCPP). A PPy-g-C₃N₄-MIL88B photocatalyst was synthesized with a narrowed bandgap, high structural stability, minimized iron leaching, and enhanced charge transfer properties. Comprehensive characterization using HAADF-FE-SEM, TEM, FTIR, XRD, XPS, BET, PL, and UV-DRS confirmed the successful fabrication of the photo-Fenton catalyst. UV-DRS analysis revealed a 28% bandgap reduction, from 3.2 eV (MIL-88B) to 2.3 eV (PPy-g-C₃N₄-MIL88B). The PPy-g-C₃N₄-MIL88B/H₂O₂/LED system was subsequently synthesized and employed for TCPP degradation, where five by-products (BPs) were detected and quantified, each showing significant elimination within 20 minutes. Toxicity assays demonstrated improved ecological safety, as the LC₅₀ (96 h) for fathead minnows increased from 7.24 mg/L (TCPP) to 19.12 mg/L (BP175) and 58.23 mg/L (BP323) after treatment. Kinetic modelling through response surface methodology and central composite design identified a reduced quartic model with strong significance (p < 0.0001). ANOVA-based optimization indicated near-complete TCPP degradation (99.29%) under optimal conditions of 59 min irradiation, 100 mg/L catalyst dosage, pH 5, and 15 L/min airflow. Finally, the applicability of PPy-g-C₃N₄-MIL88B was validated in real water matrices from four Canadian lakes and rivers, achieving over 65% contaminant removal. More, three machine learning models were employed, including KNN (K=5), Random Forest, and linear regression for data modelling and degradation efficiency prediction. Among the applied models, Random Forest showed the best performance in predicting TCPP degradation efficiency with the least MSE (15.196). These findings highlight the potential of PPy-g-C₃N₄-MIL88B as a sustainable photocatalyst for addressing emerging water pollutants.