Phenol removal by HRP/GOx/ZSM-5 from aqueous solution: Artificial neural network simulation and genetic algorithms optimization

Abstract

Abstract In this study, horseradish peroxidase (HRP) and glucose oxidase (GOx) were immobilized on mesoporous ZSM-5 nanoparticles by using glutaraldehyde as cross linking agent, and the prepared biocatalyst was characterized using SEM and EDAX mapping analysis. The resulted HRP/GOx/ZSM-5 biocatalyst was used for phenol removal from aqueous solution. In order to prevent the deactivation of HRP in the presence of excess H2O2, required H2O2 was produced by GOx in situ to activate HRP, which led to an increase in removal efficiency of phenol about 30%. Investigations on the removal efficiency of phenol for both immobilized and free enzymes indicated that immobilized enzymes have higher activity and less sensitivity to pH variation compared with free ones. In addition, the effect of parameters such as temperature, pH, HRP/GOx ratio, phenol and glucose concentration on the removal efficiency of phenol was investigated. Finally, an artificial neural network (ANN) was developed to model and express the relationship between removal efficiency of phenol and aforementioned parameters. The optimized values of parameters were determined by optimizing the resulted ANN model using genetic algorithm (GA).