Removal of Dimethoate Pesticide Using Double Layer Hydroxide@Graphene Oxide: Optimization Via Response Surface Methodology and Neural Networks

Abstract

Removing pesticides from water is essential to protect ecosystems and preserve water sources from pollutants. The use of nanocomposites for adsorption removal has gained attention in the last decade. In this work, Mg-Al double-layer hydroxide coated on graphene oxide (Mg-Al-LDH@GO) was synthesized and characterized using FESEM, EDS, and XRD techniques. The removal of dimethoate from aqueous solution using Mg-Al-LDH@GO as an adsorbent was optimized and modeled. The response surface method (RSM) was employed to design the experiments based on the central composite design (CCD). Four parameters affecting the adsorption removal efficiency of dimethoate, including pH, contact time, adsorbent dose, and pollutant concentration, were optimized using RSM-CCD. The results indicate that the removal process can be accurately predicted by the quadratic model. Numerical optimization results showed that when the concentration of dimethoate is 83.7 mg/L, the optimal conditions are pH 2.6, contact time 15.3 h, and adsorbent dose 27.5 mg/100 mL. Under these conditions, the removal efficiency reached 86.08%. A feed-forward neural network (ANN) model with the Levenberg-Marquardt backpropagation training algorithm was adapted to model the removal process. The performance of the ANN model showed adequate response prediction with R² values of 0.9989, 0.9803, 0.9984, and 0.9936 for training, validation, testing, and overall, respectively. The results obtained in this work demonstrate that Mg-Al-LDH@GO is a potential adsorbent for the removal of dimethoate from water.