Optimizing the removal of strontium and cesium ions from binary solutions on magnetic nano-zeolite using response surface methodology (RSM) and artificial neural network (ANN)

Abstract

The feasibility of using magnetic nano-zeolite (MNZ) to remove cesium and strontium from their binary corrosive solutions was investigated by considering the multi-variant/multi-objective nature of the process. RSM (Response Surface Methodology) and ANN (Artificial Neural Network) were used to model and optimize the removal system and assess sensitive parameters that can affect the process reliability. MNZ is characterized by its high surface area and cation exchange capacity and possesses good regeneration behavior for both elements using citric acid. Its stability is comparable to other sorbents in acidic media and the stability increases in alkaline media, where dissolution rate follow first order reaction on heterogeneous sites. MNZ removes both contaminants simultaneously with small tendency toward Cs, where MNZ is suggested for application in pre-treatment of highly contaminated alkaline solutions. The percentage removal, decontamination factors, and separation factors have different dependency on the effluent/process conditions; this dependency is the same for both contaminants. Sorption kinetics is initially controlled by external mass transfer through the boundaries then intra-particle diffusion dominates the reactions. The process sensitivity to pH changes is attributed to changes in structural elements -species distribution at the solid/aqueous interface. Cs+ and Sr+2 are exchanged with Na+ and H+, regardless the effluent pH value, and with Al and Fe cations at specific pH. Isosteric heat of sorption calculations indicated that the total heat needed to complete the reaction was considerably reduced by operating the process at optimized temperature.