An Explainable Classification Framework for Determining and Understanding the Extraction Suitability

Liquid-liquid extraction is a separation method that relies on a solvent to transfer a product of interest from one phase (e.g., an aqueous phase) to the solvent-rich phase. To reduce solvent consumption, liquid-liquid extraction needs to be combined with another separation method (e.g., distillation) to recycle the solvent. Limited research has focused on determining and understanding the suitability of combined extraction and distillation (i.e., when combined extraction and distillation is feasible and more economical than distillation alone). In this work, we present a classification framework to predict if combined extraction and distillation is feasible and more economical than distillation based on the values of input features that include feed composition, liquid-liquid equilibrium constants, relative volatilities, and solvent price. The trained classifiers show good prediction accuracy. We further understand how specific features influence the prediction of classifiers and apply the framework to analyze the separation of acetic acid from water.