Solvent-terminated dispersive liquid-liquid microextraction (ST-DLLME) as a simple, fast, and low-cost technique was developed for simultaneous extraction of Cd2+ and Cu2+ ions in aqueous solutions. Multiobjective evolutionary algorithm based on decomposition with the aid of artificial neural networks (ANN-MOEA/D) was used for the first time in chemistry, environment, and food sciences to optimize several independent variables affecting the extraction efficiency, including disperser volume and extraction solvent volume, pH, and salt addition. To perform the ST-DLLME operations, xylene, methanol, and dithizone were utilized as an extraction solvent, disperser solvent, and chelating agent, respectively. Non-dominated sorting genetic algorithm versions II and III (NSGA II and NSGA III) as multiobjective metaheuristic algorithms and in addition central composite design (CCD) were studied as comparable optimization methods. A comparison of results from these techniques revealed that ANN-MOEA/D model was the best optimization technique owing to its highest efficiency (97.6% for Cd2+ and 98.3% for Cu2+). Under optimal conditions obtained by ANN-MOEAD, the detection limit (S/N = 3), the quantitation limit(S/N = 10), and the linear range for Cu2+ were 0.05, 0.15, and 0.15-1000 μg L-1, respectively, and for Cd2+ were 0.07, 0.21, and 0.21-750 μg L-1, respectively. The real sample recoveries at a spiking level of 0.05, 0.1, and 0.3 mg L-1 of Cu2+ and Cd2+ ions under the optimal conditions obtained by ANN-MOEA/D ranged from 94.8 to 105%.
Keywords: Artificial neural networks; Cadmium ions; Copper ions; Multiobjective evolutionary algorithm based on decomposition; Solvent-terminated dispersive liquid–liquid microextraction.