In this research, a novel CoFe2O4-GO (Graphen Oxide) resulting from the combination of high applicable magnetic and organic base materials and synthesized with a simple and fast co-precipitation route was synthesized for the REEs (Rare Earth Elements) extraction. This adsorbent could remove the La3+, Ce3+, Nd3+ and Eu3+ by maximum adsorption capacity of 625, 626, 714.2, 1111.2 mg/g at optimized pH = 6, respectively. A data-driven model was obtained using Group Method of Data Handling (GMDH)-based Neural Network to estimate the adsorption capacity of these LREEs as a function of time, pH, temperature, adsorbent ζ (zeta)- potential, initial concentration of lanthanides ions, and ε which is defined by the physico-chemical properties of lanthanides. The results clearly indicated that the model estimate the experimental values with good deviation (mostly less than 10%) and it can be used for the prediction of the results from other similar researches with less than 25% deviation. The results of sensitivity analysis indicated that the adsorption capacity is more sensitive to pH of the solution, temperature, and ε. Finally, the desorption studies showed an excellent removal efficiency (97%) at least for three adsorption-desorption cycles. These results claimed that the CoFe2O4-GO is a highly efficient adsorbent for the REEs extraction.
Keywords: Adsorption; Graphene oxide; Machine learning; Magnetic adsorbent; Spinel ferrite nanoparticles.
Copyright © 2022 Elsevier Ltd. All rights reserved.