Optimization of Dispersive Liquid-Liquid Microextraction (DLLME) Based on the Solidification of Floating Organic Drop (SFO) Coupled with Ultrasonic-Assisted Extraction (UAE) for the Extraction Recovery of Deca-Bromodiphenyl Ether (BDE-209) from Surficial

This paper aims to optimize dispersive liquid-liquid microextraction based on the solidification of floating organic drop coupled with ultrasonic-assisted extraction (UAE-DLLME-SFO) of brominated diphenyl ether (BDE-209) in surficial sediment samples through genetic algorithm neural network (GANN) model and high-performance liquid chromatography (HPLC). A GANN model was established based on the orthogonalized experiment of BDE-209. The average relative deviation between the values predicted by GANN model and the experimental ones was 5.83%. GANN model established was combined with genetic algorithm to optimize the UAE-DLLME-SFO and the optimum DLLME conditions were then validated by experiments. The results indicated that the optimal UAE-DLLME-SFO for BDE-209 is to use 0.54 g of sediment sample, 13.75 mL of acetone, 10 min of UAE, 0.53 mL of disperser solvents, 35 μL of undecanol, 0% of NaCl and 14.24 min of DLLME, and the extraction recovery (ER) of BDE-209 from sediments is approximately 90% with an increase of 7.59% through this GANN model.