COSMO-RS Evaluation and UNIFAC Model Extension for Ionic Liquids and Dimethyl Carbonate/Methanol Systems

Dimethyl carbonate (DMC)–methanol (CH3OH) separation is an important process in the chemical industries, requiring accurate thermodynamic models for process design and optimization. Although various predictive models, including COSMO-RS and UNIFAC, have been widely applied to analyze ionic liquid (IL)-based extractive distillation, their accuracy in IL-DMC/CH3OH systems has not been systematically evaluated until now. In this study, the COSMO-RS model is first evaluated based on the extensively collected experimental database, followed by the extension of both the original and modified UNIFAC (orig. and mod. UNIFAC) models. Comparative analyses of the performance of COSMO-RS and the extended orig. and mod. UNIFAC models on the training set and test set are conducted in detail. It is concluded that the COSMO-RS generally provides qualitative and semiquantitative prediction, while the extended UNIFAC models demonstrate significantly superior quantitative performance, providing reliable thermodynamic guidance for the computer-aided screening and design of ILs in DMC/CH3OH systems.