Machine learning-driven insights into sustainable chlorpyrifos removal using waste-leaf activated carbon: Mechanisms, regeneration, and spiking studies

This study investigates the application of activated carbon synthesized from Magnolia champaca leaves (MCAC) as a sustainable adsorbent to remove chlorpyrifos from wastewater. The material exhibited a large surface area (552.04 m²/g) with a pore volume (0.5587 cm³/g), indicating a well-developed mesoporous structure and exceptional textural properties following chemical activation with orthophosphoric acid. The Freundlich model provided a better fit, with a high R² (0.9676), revealing a multilayer adsorption process. The Langmuir model exhibited a maximum adsorption capacity of 141.89 mg/g. The adsorption kinetics followed the pseudo-second-order equation. A spontaneous and physisorption-driven adsorption process was revealed by thermodynamic analysis. The adaptive neuro fuzzy systems (ANFIS) had a high R² value of 0.9970 and lower error indices, root mean squared error (RMSE), mean squared error (MSE), and mean absolute error (MAE), indicating a slightly higher accuracy than the artificial neural networks (ANN). Additionally, the MCAC showed excellent reusability up to five adsorption-desorption cycles. Despite competing ions, it maintained high chlorpyrifos removal in various water matrices. This study showed the conversion of biomass into a valuable adsorbent capable of effectively removing the toxic chlorpyrifos from aqueous environments.