Mesoporous activated carbons of low-cost agricultural bio-wastes with high adsorption capacity: Preparation and artificial neural network modeling of dye removal from single and multicomponent (binary and ternary) systems

Abstract

Herein, low-cost mesoporous activated carbons (ACs) with high adsorption capacity were prepared by various agricultural bio-wastes, including Kiwi peel (KP), Cucumber peel (CP), and Potato peel (PP) and denoted as AC_KP, AC_CP, and AC_PP, respectively. The prepared ACs were analyzed by XRD, FESEM-SEM/EDAX, TGA/DTA, FTIR, BET-BJH, XPS, particle size, and zeta potential analyses and used for removing dye (methylene blue (MB)) from single (MB), binary (MB+ malachite green (MG) and MB+ rhodamine B (RhB)) and ternary (MB + MG + RhB) wastewater systems. Dye removal by the ACs was modeled by the artificial neural network (ANN). The effects of MB concentration, ACs dosage, process time, temperature, and solution pH on adsorption were fully studied. The isotherm and kinetic of dye removal followed the Langmuir and Pseudo-second order (PSO) models, respectively. The Gibbs free energy for dye removal at 298 K was −20.34, −14.10, and −15.14 kJ/mol for AC_KP, AC_CP, and AC_PP, respectively. Also, the enthalpy was 43.62, 35.26 and 25.84 kJ/mol and entropy was 0.2146, 0.1656 and 0.1327 J/mol for AC_KP, AC_CP, and AC_PP, respectively. The thermodynamic data presented that MB adsorption by the prepared ACs was endothermic, spontaneous, and physical sorption in nature. The ANN model extensively overlaps the experimental data due to the low divergence between actual and estimated data. Also, the prepared ACs have the ability for removing dyes from binary (MB + MG and MB + RhB) and ternary (MB + MG + RhB) systems.