Ultrasonic-assisted tetrabromobisphenol A-bis-(2,3-dibromo-2-methylpropyl ether) extraction process from ABS polymer supported by machine learning

Abstract

The extensive use of electronic and electrical equipment and ever-increasing rate of renewal inevitably generate numerous e-waste plastics which need to be disposed of properly, and eliminating the brominated flame retardants (BFRs) present in them can significantly alleviate their environmental impact. In the present study, TBBPA-bis-(2,3-dibromo-2-methylpropyl ether) (TBBPA-DBMP), a novel tetrabromobisphenol A-derived BFR, was removed using an extraction process assisted by ultrasonic in alcohol extractants. Response Surface Methodology (RSM) achieved by support vector regression (SVR) was used to predict the optimum extraction conditions (extraction time, extraction temperature, liquid to solid ratio) in methanol and ethanol solvent, respectively. The maximum predicted extraction efficiencies of TBBPA-DBMP were 98.1% under the optimized conditions of 178 min, 180 mL/g, 66 °C in methanol and 99.5% under the optimized conditions of 150 min, 178 mL/g, 66 °C in ethanol. And the validity of predicted optimal extraction conditions was verified. The plastic matrix was capable of being recovered well after extraction (recovery rate > 90%). Moreover, shrinking core model was referred to conduct the kinetic analysis of the extraction, which demonstrates that the interface transfer and internal diffusion are the rate-controlling step in the extraction of TBBPA-DBMP in both methanol and ethanol extractants. • TBBPA-DBMP was extracted from ABS polymer with ultrasonic-assistance. • RSM via machine learning predicted the optimum extraction conditions. • The plastic recovery rate was above 90% after extraction. • Interface transfer and internal diffusion are rate-controlling step in the process.