Photocatalysis of low-density polyethylene using FKMW-ZnO NPs: optimization and predication model using a radial basis function neural network ensemble system

The present study aimed to investigate the efficiency of biosynthesized zinc oxide nanoparticles in fungal supernatant grown in kitchen wastewater with microelectronic sludge (FKMW-ZnO NPs) to be used in the degradation low-density polyethylene (LDPE) in aqueous solution. The photocatalysis process was optimized using response surface methodology as a function of four independent factors included LDPE concentrations x₁ (100–500 mg/100 mL), FKMW-ZnO NPs concentrations x₂ (10–100 mg/100 mL), time x₃ (1–6 h) and pH x₄ (4–9). The maximum photocatalysis of LDPE was 45.43% optimized with 229.96 mg LDPE/100 mL, 100 mg FKMW-ZnO NPs/100 mL at pH 7 and after one hour with R² is 0.7377. Microstructure and chemical structure analysis showed a significant change in the chemical structure of the photocatalysis of LDPE because of FKMW-ZnO NPs. The mathematical predication model using a radial basis function neural network ensemble system (RBFNNES) provided more accurate prediction model 89.2857% with R² = 0.8688. However, RBFNNES revealed that FKMW-ZnO NPs and LDPE have unstable behaviour towards the investigated factor and the interaction between these factors where the error was increasing with the increasing the time of neural network which indicates that the obtained efficiency in the optimization study might be not applicable in the large scales or in different environmental factors. More optimization with a wide range of factors is required to understand the applicability of FKMW-ZnO NPs in remediation of LDPE in the environment. Graphical Abstract: (Figure presented.).