Adsorption of Fe (II) from Aqueous Phase by Chitosan: Application of Physical Models and Artificial Neural Network for Prediction of Breakthrough

Removal of Fe (II) from aqueous media was investigated using chitosan as the adsorbent in both batch and continuous systems. Batch experiments were carried out at initial concentration range of 10-50 mg/L and temperature range of 20–40˚C. In batch experiments, maximum adsorption capacity of 28.7 mg/g and removal efficiency of 93% were obtained. Adsorption equilibrium data were well-fitted with Langmuir-Freundlich model and the model parameters were recovered. In column study, experiments were performed in a fixed bed of chitosan operated at continuous up-flow mode and constant temperature of 25˚C. Sharp breakthrough curves were observed at high flow rates, high inlet metal concentrations and low bed heights.  Breakthrough curves were analyzed by physical models such as Thomas and Yan’s models as well as Artificial Neural Network (ANN) method. Compared to physical models, simulation of dynamic behaviour of the system using Back Propagation Artificial Neural Network (BP-ANN) demonstrated high coincidence between experimental and predicted breakthrough curves.  The FTIR spectrum of chitosan before and after adsorption process demonstrated that hydroxyl and amino groups are the main functional groups involved in the binding of iron ions.