Prediction of methyl orange removal by iron decorated activated carbon using an artificial neural network

Abstract

ABSTRACT Date Stones were used as a bio-degradable waste source for preparing iron impregnated activated carbon. The prepared activated carbon-containing oxides of iron were characterized using SEM, XRD, FTIR, and BET. The specific surface area of the iron decorated activated carbon was 738.65 m2/g. The XRD confirmed the presence of magnetite and hematite while the SEM images assured the presence of pores. The prepared activated carbon was used to remove methyl orange from wastewater. Genetic Algorithm was used to develop a model which could predict the removal efficiency of the dye. The ANN model was validated and the effect of different parameters like adsorbent dosage (0.1–1 g/L), initial dye concentration (2–20 mg/L), pH (2–11), time (10–55 min) and temperature (30–75°C) was estimated both experimentally and predicted using the model. The adsorption process follows the Freundlich isotherm and pseudo-second-order kinetic model. The values of 1/n and KF obtained from the Freundlich isotherm designate good adsorption capacity. Both experimental and model-predicted data agrees with the kinetic model. The adsorption rate is proportionate to the square of the number of vacant adsorption sites. From the thermodynamic study, the positive worth of ΔH° indicates the energy-absorbing nature of the surface assimilation method and the process is endothermic in nature. The low values of each ΔG° (−200 to 0 kJ/mol) and ΔH° correspond to physical surface assimilation. A positive worth of ΔS° reflects the inflated randomness at the solid-aqueous interface with some structural changes in adsorbate and adsorbent.