Artificial Neural Networking for remediation of methylene blue dye using Fuller's earth clay

The present work focuses on the remediation of Methylene Blue dye (MB) by adsorption on Fuller's earth clay (FE). The optimum batch limits for MB adsorption are; pH 8; equilibration time 3 ​h; initial concentration 100 ​mg ​L−1; agitation speed 300 ​rpm; a dose of adsorbent 2 ​g ​L−1, and temperature 25 ​°C. FE clay was characterized using XRD and FESEM. XRD analysis indicated montmorillonite to be the major component of FE. FESEM images indicated clay surface to be layered with high surface area and porosity thereby making it an excellent material for dye adsorption. Classical Langmuir, Modified Langmuir, Freundlich, and BET adsorption isotherms were studied to investigate the adsorption mechanism of MB on FE. Modified Langmuir isotherm fits well for adsorption of MB on FE with an adsorption capacity of 11587 ​mg ​g−1. Kinetic studies revealed the adsorption process to follow Pseudo Second-order kinetics with the energy of activation 28 ​kJ/mol, indicating the adsorption process to be physisorption. Thermodynamic studies indicated the adsorption process to be spontaneous and endothermic. The Artificial Neural Network (ANN) model was employed in the FE system, and the model confirmed its applicability. According to error analysis, ANN was found to be more appropriate model than pseudo second-order model. Sensitivity calculations from ANN data indicated the pH of adsorption to be the most influential factor in MB adsorption with 29% relative importance. The sorption process may be practically valuable on an industrial scale to remediate MB from water assets at an economical price. [Display omitted] •Methylene Blue a cationic dye is adsorbed on Fullers earth at pH 8.•Adsorption equilibrium is achieved in 3 ​h of equilibration.•Modified Langmuir model fits best with adsorption efficiency of 11587 ​mg ​g−1.•Artificial Neural Network modelling fits best on the data with low error values.•Minimum experiments are required for evaluation of adsorption efficiency due to ANN.