Comprehensive Evaluation of COSMO-RS for Predicting Ternary and Binary Ionic Liquid-Containing Vapor–Liquid Equilibria

To develop efficient ionic liquid (IL)-based separation processes like extractive distillation, the knowledge of IL-involved vapor–liquid equilibria (VLE) is necessary. Since experimental measurements are usually time-consuming and expensive, reliable models for VLE prediction are highly desirable. In this work, the predictive performance of COSMO-RS is systematically evaluated for ternary and binary IL-containing VLE systems. A large experimental VLE database is established from the literature, including 6531 and 2265 data points for ternary and binary systems, respectively. The ternary and binary VLE systems are divided into five and three subsets based on the varieties of molecular solutes. The COSMO-RS performance in capturing the effects of IL and solute structures, IL dosage, and temperature on VLE is analyzed. COSMO-RS provides generally qualitative and in some cases quantitative description of the VLE behaviors and thus could be used as a valuable tool for prior IL screening and process analysis for related separations.