Investigating micellar-enhanced ultrafiltration (MEUF) of mercury and arsenic from aqueous solution using response surface methodology and gene expression programming

Abstract

• Micellar enhanced ultrafiltration (MEUF) for the removal of mercury and arsenic was investigated. • Response surface methodology (RSM) was used to assess the impact of operating parameters on selected metals removal efficiency by MEUF. • The gene expression programming (GEP) model was tested to optimize the MEUF process. • Proposed RSM and GEP models were validated and compared based-on statistical analysis. Heavy metals such as mercury and arsenic in water resources have become a global concern due to their toxic, persistent, bio-accumulative, and carcinogenic characteristics. To address health-related issues as well as environmental concerns, it is important to eliminate mercury and arsenic contaminants from water streams. In this study, mercury and arsenic ions were removed from aqueous solutions through micellar-enhanced ultrafiltration (MEUF) using sodium dodecyl sulfate (SDS) and cetylpyridinium chloride (CPC) as chelating agents, respectively. Response surface methodology (RSM), based on the full factorial design (FFD) method, was applied to optimize the process parameters. The relationship between the removal of targeted heavy metals and process variables, including pressure, metal concentration, pH, and molar ratio of surfactant to metal defined by the proposed quadratic models. The models were statistically significant, as shown by analysis of variance. Experimental results presented the optimum operating parameters for > 95% mercury removal were: 2.5 bar pressure, 10 ppm mercury concentration, SDS to mercury MR of 12:1, and pH 8.0. In the case of arsenic, >90% removal was achieved with pressure at 2.5 bar, 15 ppm arsenic concentration, and CPC to arsenic MR at 8:1, and pH 8.0. Statistical validation of RSM models using predicted and experimental results indicated that the proposed model effectively provided a relationship between targeted heavy metal removal and process variables. Additionally, the MEUF performance for mercury and arsenic removal from an aqueous solution was predicted using a gene expression programming model that presented a good fit with the experimental results.