The adsorption of Pb2+ and Ni2+ ions utilizing modified chitosan beads: A response surface methodology and neural network modelling study

Abstract

This work investigates the application of artificial neural networks (ANN) and response surface methodology (RSM) in developing a technique for the removal of Pb2+ and Ni2+ ions from wastewater using chitosan derivative. Briefly, chitosan beads were formulated from their powdered form and thereafter grafted with aniline to provide more binding sites. The chitosan beads (CS) and grafted chitosan beads (MCS) were analyzed using infrared spectroscopy (FTIR) and a scanning electron microscope (SEM). However, during the adsorption studies, variables such as pH, adsorbent dose, contact duration, temperature, and concentration were considered as the input layer feed data, while two neurons were used as output layers, which correspond to the adsorption of Pb2+ and Ni2+ ions. The RSM and ANN models were measured using statistical metrics like average relative errors (ARE), coefficient of determination (R2), Marquart's percentage standard deviation (MPSD), mean squared error (MSE), Pearson's Chi-square (χ2), root means square errors (RMSE), and the sum of squares of errors (SSE). The ideal trained neural network depicts the training, validation, and testing phases, with R2 values of 1.0, 0.968, and 0.961, respectively. The findings, however, showed that the ANN technique is superior to the RSM-CCD model approach. the RSM-CCD model optimization results for the process variables were achieved at pH 5, a starting concentration of 100 mg/L, an adsorbent mass of 6.0 g, a reaction time of 55 min, and a temperature °f 40 °C and the greatest removal percentages for Pb2+ and Ni2+ ions were 98.14 % and 98.12 %, respectively. These findings suggest that ANN can be utilized in forecasting the removal of adsorbates from wastewater.