Modeling, optimization and comparative study on abatement of fluoride from synthetic solution using activated laterite soil and fly ash

The supply of potable water by proper treatment of fluoride laden ground water is a challenging task. Batch experiments were executed with activated laterite soil and fly ash individually for removal of fluoride. The process parameters such as particle size (100–620 μm), dose of adsorbent (10–60 g/L), initial concentration of fluoride (2–12 mg/L), agitation speed (20–120 rpm), contact time (0.5–10 h) and pH (4.5–9.5) were investigated and it was observed that activated laterite soil had better fluoride removal efficiency than fly ash. At pH 4.5, contact time 10 h, particle size 100 μm, adsorbent dose 60 g/L, initial concentration of fluoride 12 mg/L, and agitation speed 120 rpm, maximum removal efficiencies of fluoride using activated laterite soil and fly ash were found as 85.91 ± 0.62% and 77.1 ± 0.39%, respectively. Kinetic study was performed and Pseudo 2nd order kinetic model was found to fit the kinetic data best. The Freundlich isotherm model fit the equilibrium data fairly well. The values of adsorption equilibrium constant as used in Freundlich isotherm model vis-à-vis adsorption capacity (KF) for activated laterite soil and fly ash were obtained as 0.1331 and 0.0772 ((mgg−1)(Lmg−1)1/n). A thermodynamic analysis was conducted to examine the process behaviour. In order to determine whether the adsorbent may be used for further cases, regeneration of the adsorbents was also done. Generally, two types of adsorption mechanisms happen like electrostatic attraction and ion exchange. The electrostatic force of attraction favours the adsorption of negatively charged fluoride ions on positively charged adsorbent surfaces. In ion exchange, fluoride ions are exchanged by hydroxyl and hydronium ion. Based on experimental data, Multi-Gene Genetic Programming (MGGP) model could accurately predict the removal efficiency of fluoride under various operating situations. Finally, using Genetic Algorithm (GA) optimization the maximum output values for both adsorbents were estimated as 99.14% and 99.02%.