Assessment of COSMO-RS for Predicting Liquid-Liquid Equilibrium in Systems Containing Deep Eutectic Solvents

Deep eutectic solvents (DESs) have gained significant attention as green solvents for liquid-liquid extraction applications. Predictive thermodynamic models can identify suitable DESs for a specific separation problem. To this end, we conducted a comprehensive evaluation of the conductor-like screening model for real solvents (COSMO-RS) to predict liquid-liquid equilibrium (LLE) in systems containing DESs. The TZVP and TZVPD-FINE parameter sets of COSMO-RS were considered, and the salts were modeled as associated or fully dissociated ions. The results showed that COSMO-RS predicted LLE in all systems with an average root mean square deviations (RMSDs) below 10% and represented lower average RMSD in systems with nonsalt DESs than salt-based DESs. COSMO-RS predictions using the TZVPD-FINE parameter set were slightly better than the TZVP parameter set. Nevertheless, the COSMO-RS predictions of the solute distribution between the phases were not accurate enough compared to experimental data using either of the parameter sets. COSMO-RS largely overestimated the solute partition coefficients for aromatic and heterocyclic solutes in systems containing alkanes and salt-based DES. On the other hand, acceptable RMSDs and reliable predictions of partition coefficients were observed for alcohol solutes in systems containing alkanes and salt-based DESs. Given the complexity of describing the interactions and LLE in quaternary systems containing salts, further improvements in the COSMO-RS model are still needed. This study highlighted the potential and shortcomings of COSMO-RS as a tool to utilize DES for liquid-liquid extraction applications.