Photocatalytic degradation of textile reactive dye using artificial neural network modeling approach

This study describes the application of TiO2, ZnO, TiO2-ZnO-based systems modified with 1.5 and 2.5% wt. Fe using the impregnation method for the maxilon blue 5G dye discoloration. Specific surface area (BET method), X-ray diffraction, thermogravimetric analysis, and photoacoustic spectroscopy characterization techniques were used in this work. The presented results indicate that structure alterations, textural properties, nominal metallic charges, and photocatalyst thermal treatment affected the photocatalytic activity of the system. Moreover, all of the synthesized catalysts yielded complete discoloration of the maxilon blue 5G dye within 180min of reaction. The reaction time required for total discoloration was strongly affected by the temperature of photocatalysts calcination. In general, increasing the catalysts nominal metallic charge improves the adsorption process. A neural model was developed for each studied catalyst using MLP network with three intermediate layers, backpropagation learning algorithm, and sigmoid activation function implemented in FORTRAN. The three models presented the best results with three neurons in the intermediate layer. Therefore, the neural networks can be successfully employed to model the discoloration process involving the synthesized catalyst (R-2 varying between 0.98 and 0.99).