Modeling natural bentonite, thermal-modified bentonite and iron-modified bentonite with artificial neural network, sorption kinetics and sorption isotherms for simulated sorption tetracycline

In this research, modification for the natural bentonite clay was achieved using thermal activation and iron modification. The ability of different adsorbents (natural bentonite, thermal-activated bentonite and iron-modified bentonite) in removing tetracycline from contaminated water was examined utilizing artificial neural network consisting of three layers and 315 batch experiments. The analysis of sensitivity proved that iron-modified bentonite had more influence in the present sorption process with relative importance of 28.67%. To study equilibrium isotherm and adsorption kinetic, two models were used which were pseudo-first-order and pseudo-second-order kinetic equations, and Freundlich and Langmuir models. The best input for Langmuir model was iron-modified bentonite as well as the equilibrium data of natural bentonite, while thermal-modified bentonite was indicated using the Freundlich model. For the three adsorbents, the adsorption kinetic associated with the pseudo-second-order equation was used.