Evaluation of photocatalytic activity of ZnSe:Cu nanoparticles: synthesis, experimental measurement, kinetic study, and ANN modeling

• 87.4% DB71 degradation at natural pH (6.8) and 308 K using optimal sample ZnSe:Cu (2.25 mol%). • Pseudo-first-order kinetics dominated with a rate constant of 0.0364 min −1 under optimal conditions. • An activation energy of 35.9 kJ/mol confirmed a surface-mediated electron transfer mechanism. • A dual-output ANN model accurately predicted efficiency and kinetics (MSE = 0.3096 × 10 −3 , R ≈ 1). • Sensitivity analysis identified time (34.4%) and temperature (31.0%) as the most influential parameters. Cu NPs) containing the optimal dopant level of 2.25 mol% Cu (as established in our prior work) were synthesized via a microwave-assisted hydrothermal (MAH) method at pH 11.2. Structural characterization confirmed spherical nanoparticles (∼2 nm crystallite size via XRD) with uniform morphology (FESEM/TEM). The photocatalytic degradation of Direct Blue 71 (DB71) under UV-C irradiation was systematically optimized, revealing maximal efficiency (87.4% in 60 min) under neutral pH (6.8), 400 mg/L catalyst loading, and 308 K, corresponding to a pseudo-first-order rate constant of 0.0364 min −1 . Pseudo-first-order kinetics dominated across conditions, with an activation energy of 35.9 kJ/mol indicating surface-mediated electron transfer. A dual-output artificial neural network (ANN) model accurately predicted degradation efficiency and rate constants (MSE = 0.3096 × 10 −3 , R ≈ 1), utilizing a feed-forward backpropagation architecture with 10 hidden neurons, and logsig/purelin transfer functions. Sensitivity analysis revealed that reaction time (34.4%) was the dominant factor affecting degradation efficiency, followed by pH (26.6%) and catalyst dosage (18.9%). In contrast, temperature (31.0%) most strongly influenced reaction kinetics, highlighting the system’s responsiveness to thermal and temporal controls. This work establishes ZnSe:Cu NPs as robust photocatalysts for wastewater treatment, augmented by machine learning-driven process optimization.