Muhandisī-i Bihdāsht-i Muḥīṭ (Dec 2018)
Optimization of Removal Efficiency of An Anionic Dye Onto Magnetic Fe3O4-Activated Carbon Nanocomposite Using Artificial Neural Network
Abstract
Background and objective: Wastewaters including dyes produced by various industries have serious destructive effects on the environment. Therefore, proposing analytical and numerical mathematics methods simulating dye removal process from industrial wastewaters are great of importance. Methods: In this research, the Fe3O4-activated carbon magnetic nanocomposite was synthesized and its crystalline structure, surface, and magnetic properties were characterized by XRD, SEM, and VSM techniques. Efficiency of the composite adsorbent for decolorization of Reactive Red dye in different conditions was investigated. Then, an artificial neural network was constructed by using Matlab program to predict the removal efficiency of dye onto magnetic activated carbon and the number of neurons in a hidden layer was optimized. pH, contact time, initial dye concentration, and temperature as input parameters and dye removal percentage as an output parameter were considered. Performance of network after its training was evaluated based on the correlation factor. The experimental data were analyzed by pseudo- first- order, pseudo- second- order , and intra-particle diffusion kinetics models.The Langmuir and Freundlich models were used to describe the sorption equilibrium isotherms. Results: . The high correlation factor for testing data showed that artificial neural network model can estimate the experimental data. The intra-particle diffusion kinetics and Freundlich isotherm models best describe the experimental data for the uptake of dye. A relatively low activation energy (34.6 kJ mol-1) suggests that the adsorption involve physio sorption. Maximum adsorption capacity decreased with increasing temperature. Conclusion: Use of network prediction resulted to eliminate experiments and to improve dye removal percentage.
