Pumice-supported ZnO-photocatalyzed degradation of organic pollutant in textile effluent: optimization by response surface methodology, artificial neural network, and adaptive neural-fuzzy inference system.

A heterogeneous photocatalysis was adopted to treat textile industry effluent using a combination of pumice-supported ZnO (PUM-ZnO) photocatalyst and solar irradiation. The visible light-responsive PUM-ZnO photocatalyst was prepared via the impregnation method and characterized using various spectroscopic techniques. The photocatalytic degradation process was modeled via response surface methodology (RSM), artificial neural network (ANN), and adaptive neuro-fuzzy inference system (ANFIS), while the optimization of the three independent parameters significant to the photocatalytic process was carried out by a genetic algorithm (GA) and RSM methods. The low standard error of prediction (SEP) of 0.56-1.75% and high coefficient of determination (R) greater than 0.96 for the models developed indicated that they adequately predicted the photodegradation process with high accuracy in the order of ANFIS > ANN > RSM. The process optimization results from the developed models showed that GA performed better than RSM. The best optimal condition (3.29 g/L catalyst dosage, 45.85 min irradiation time, and 3.13 effluent pH) that resulted in maximum degradation efficiency of 99.46% was achieved by the ANFIS model coupled with GA (ANFIS-GA).