Immobilized PAni-TiO2 Nano-Photocatalyst Modeling for Photocatalytic Degradation of 1, 2-Dichloroethane using Response Surface Methodology (RSM) and Artificial Neural Network (ANN)

Heterogeneous photocatalytic degradation as an advanced oxidation process (AOP) is a complicated process, because many independent influential parameters can affect degradation efficiency as the response of the system. Therefore, they are difficult to be modelled using conventional mathematical modelling. This leads to uncertainty in the design and built of industrial scale photocatalytic reactors. In this study, heterogeneous photocatalytic degradation of 1,2- dichloroethane using immobilized PAni-TiO2 nano-photocatalyst was modeled by Response Surface Methodology (RSM) and Artificial Neural Network (ANN) techniques and the results were compared. Central Composite Design (CCD) was used for design of experiments. Two independent parameters of catalyst loading and catalyst composition were considered as independent parameters and 1,2-DCE degradation was considered as the response parameter. RSM suggested a quadratic model by verifying lack of fit, model summary statistics and analysis of variance (ANOVA) statistics. ANN model with different structures were developed, trained, validated and tested. The results showed that both models predicted the photocatalytic degradation of 1,2-DCE with high precision. However, the ANN model had a slightly better performance. The optimal ANN model with tansig activation function, BR training function and 2-2-1 structure was the best model for prediction of 1,2-DCE photocatalytic degradation. The design expert software optimized 1,2-DCE removal (71.42%) by optimizing independent parameters of catalyst loading and catalyst composition at 0.4 mg/cm2 and 2.4 [TiO2 :PAni], respectively.