Z-Scheme g-C3N4/BiVO4 photocatalysts via hydrothermal tuning: charge transfer mechanism and machine learning optimization

Background Escalating global water pollution from intensified industrial and agricultural activities has highlighted the urgent need for sustainable wastewater treatment. Semiconductor photocatalysis is a promising solution, yet conventional type-II heterojunctions often suffer from sluggish charge separation and weak redox capacity, limiting their practical application. Methods To address these challenges, a series of Z-scheme g-C 3 N 4 /BiVO 4 (CN/BV) heterojunctions were rationally engineered via hydrothermal temperature modulation (100-180 °C), enabling precise control of crystallinity, morphology, and band alignment. Advanced characterizations (XRD, TEM, UV-vis DRS, and XPS) were combined with EPR analysis, DFT calculations, and XPS valence band studies to elucidate the structure–property–activity relationship. Furthermore, machine learning models were integrated with experimental data to predict norfloxacin (NOR) degradation performance under diverse operational conditions. Significant Findings The optimized CN/BV composite synthesized at 120°C exhibited outstanding photocatalytic activity, achieving 94.41 % degradation of norfloxacin (10 mg/L) within 120 min under visible light (λ > 420 nm), with an apparent rate constant ( k = 0.02618 min⁻¹), 4.83 times higher than BiVO 4 and 2.44 times higher than g-C 3 N 4 . EPR and radical trapping analyses identified •OH and •O₂⁻ as the dominant reactive species, while mechanistic evidence confirmed a direct Z-scheme charge transfer pathway enabling efficient carrier separation and strong redox capability. Integrating this rationally designed photocatalyst with machine-learning prediction ( R 2 = 0.8558, RMSE = 0.0510) establishes a data-driven framework for performance optimization. This study provides both a mechanistic insight and an intelligent design paradigm for next-generation photocatalysts targeting efficient degradation of emerging contaminants.