Optimization of Solvent Terminated Dispersive Liquid–Liquid Microextraction of Copper Ions in Water and Food Samples Using Artificial Neural Networks Coupled Bees Algorithm

• Published in: Bulletin of environmental contamination and toxicology Vol. 100; no. 3; pp. 402 - 408
• Main Authors: Farajvand, Mohammad, Kiarostami, Vahid, Davallo, Mehran, Ghaedi, Abdolmohammad
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.03.2018; Springer Nature B.V
• Subjects: Acetone; Algorithms; Aquatic Pollution; Artificial neural networks; Bees; Chelating agents; Chelating Agents - chemistry; Chelation; Copper; Copper - analysis; detection limit; Dispersion; Earth and Environmental Science; Ecotoxicology; Environment; Environmental Chemistry; Environmental Health; Environmental quality; Extraction processes; Food Contamination - analysis; Ions; Limit of Detection; Liquid Phase Microextraction - methods; liquid-phase microextraction; multivariate analysis; Neural networks; Neural Networks (Computer); Optimization; pH effects; Pollution; Quantitation; Soil Science & Conservation; Solvents; Solvents - chemistry; Spectrophotometry, Atomic; system optimization; Waste Water Technology; Water Management; Water Pollutants, Chemical - analysis; Water Pollution Control; Xylene; Bees algorithm; Solvent terminated dispersive liquid–liquid microextraction; Response surface methodology; Copper ions
• ISSN: 0007-4861; 1432-0800; 1432-0800
• DOI: 10.1007/s00128-017-2263-7

A multivariate method based on solvent terminated dispersive liquid–liquid microextraction was developed for the determination of Cu 2+ ions in aqueous samples. In the proposed approach, di-2-ethylhexylphosphoric acid, xylene and acetone were used as chelating agent, dispersive and extraction solvents, respectively. The effects of various factors on the extraction efficiency such as extraction and dispersive solvent volumes, salt addition and pH were studied using central composite design (CCD) and artificial neural networks coupled bees algorithm (ANN-BA). Upon comparison of these techniques, ANN-BA model was considered to be better optimization method due to its higher percentage relative recovery (about 5%) as compared to the CCD approach. The linear range and the limits of detection (S/N = 3) and quantitation (S/N = 10) were 0.22–140, 0.08 and 0.22 µg L −1 , respectively. Under the optimal conditions, the recoveries for real samples spiked with 0.1 and 0.3 mg L −1 were in the range of 85–98%.