Extensive Evaluation of Performance of the COSMO-RS Approach in Capturing Liquid-Liquid Equilibria of Binary Mixtures of Ionic Liquids with Molecular Compounds

A conductor-like screening model for real solvents (COSMO-RS) is presented as a predictive thermodynamic tool for estimating the liquid-liquid equilibrium (LLE) phase diagrams of the binary systems composed of an ionic liquid (IL) and a molecular solute-the key property related to miscibility, thus when considering the ILs as separating agents in industrial unit operations such as extraction or extractive distillation. A review of 310 the modeling performance is demonstrated by means of comparison of the model's outcomes with an extensive collection of the experimental data (>10,000 data points) extracted from more than 120 research papers published in the last two decades. A brief summary of the structure-LLE trends observed empirically is given, followed by checking the capacity of the COSMO-RS of reproducing those trends. The impact of the computational level (COSMO vs FINE) used to optimize molecular geometries of ions and molecular solute (i.e., the model's input) on accuracy of predictions and the impact of the number of relevant conformations (per moiety) adopted in LLE calculations are discussed. The influence of the chemical family of the cation core, anion, and solute on the quality of predictions is carefully analyzed and discussed. Finally, a number of remarks for efficient utilization of the COSMO-RS model for IL-based systems are suggested.