Prediction of the Experimental Data for Removal of Organic Pesticides by Carbon Nanoparticle Synthesized from Pomegranate Peel using Artificial Neural Networks

Background and purpose: The present study is aimed to investigate the prediction of the experimental data for the removal of agricultural pesticides including three herbicides Trifluralin, Glyphosate, and 2,4-Dichlorophenoxyacetic acid from aqueous solution by carbon nanoparticles synthesized from pomegranate peel using artificial neural network. Materials and Methods: Removal studies were conducted under the different experimental conditions in pH = 4-8, contact time of 0-25 minutes, and the initial concentrations in the range of 50-250 mg/L. In the present study, artificial neural network, back propagation algorithm, and Levenberg Marquardt training approach were used. Results: The results showed that the removal of agricultural pesticides Trifluralin, Glyphosate and 2,4D depended on pH such that the optimal removal efficiency observed for pesticides Trifluralin, Glyphosate, and 2,4D in pH=8 was 92.6, 78, and 92%, respectively. The optimal adsorbent weight was also found to be 0.5 g for pesticides Trifluralin, Glyphosate, and 2,4D so that the removal efficiency was equal to 97, 98.8 and 98.4% within 20 minutes. In the initial concentration of 50 mg/L, the removal efficiency was respectively equal to 88, 94, and 92% for Trifluralin, Glyphosate, and 2,4D. The results also showed that the experimental data followed from both isotherm models. Conclusions: The artificial neural network successfully predicts the data, and there is a good agreement between experimental and predicted data