Machine learning predictions for 2,4-dichlorophenoxyacetic acid removal: Breakthrough analysis using Fe2O3@PPy nanocomposite columns

• Fe 2 O 3 @PPy nanocomposites were used to adsorb 2,4-D in packed bed. • Thomas and BDST model fitted the adsorption data well. • ANN and ANFIS modelling were used to predict the breakthrough curve. The herbicide 2,4-Dichlorophenoxyacetic acid (2,4-D) poses environmental risks due to its widespread use and persistence in water bodies. This study investigated the adsorptive removal of 2,4-D from aqueous streams using Fe 2 O 3 @PPy nanocomposites in a continuous packed bed column. Column performances were assessed by changing variables including adsorbent bed height, flow rate, and influent 2,4-D concentrations. Experimental data indicated that breakthrough characteristics were influenced by bed length, concentration, and flow rate. The maximum adsorption capacity, 4.85 mg/g, was accomplished with a 3 cm bed length, 8.4 mL/min flow rate, and 30 mg/L 2,4-D concentration. The breakthrough plot data were fitted with conventional models including Adam-Bohart, Thomas, Yoon-Nelson, and BDST models. Additionally, machine learning models such as ANN and ANFIS accurately predicted the breakthrough data, with high R values of 0.99907 and 0.99927, respectively. Thus, the Fe 2 O 3 @PPy nanocomposite shows promising potential for large-scale 2,4-D removal, offering a novel and efficient solution for environmental remediation.