Modeling and optimization of photocatalytic dye degradation using tungstophosphoric acid on h-BN: RSM, ANN and ANFIS comparative approaches

Effluent dyes are a significant threat to aquatic environments, causing water pollution that endangers human health and ecosystems. This study addresses the urgent need for effective dye degradation technologies by synthesizing tungstophosphoric acid (TPA) and porous hexagonal boron nitride (h-BN) nanocomposites (TPA@h-BN). TPA was immobilized on h-BN nanosheets in varying amounts (10–40 %), with the 30 %TPA@h-BN composite showing superior photocatalytic performance in methylene blue (MB) dye degradation. The photocatalytic process was investigated using batch experiments designed by central composite design (CCD). Subsequently, response surface methodology (RSM), artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS) were employed to model the effects of reaction time, catalyst dosage, dye concentration and stirring speed. Statistical parameters were calculated for all three models, with the regression coefficients (R²) for RSM, ANN and ANFIS found to be 0.981, 0.977 and 0.928, respectively. These results indicate that the RSM and ANN models possess higher predictive capability and accuracy compared to the ANFIS model. Optimal dye degradation of 95.40 % was achieved in 240 min using 60 mg of 30 %TPA@h-BN, 15 ppm dye concentration and 550 rpm stirring. The process followed pseudo-first-order kinetics with a rate constant of 0.01303 min⁻¹.