Grigorash, DariaDariaGrigorashMüller, SimonSimonMüllerParicaud, PatricePatriceParicaudHalddan Stenby, ErlingErlingHalddan StenbyYan, WeiWeiYanSmirnova, IrinaIrinaSmirnova2025-04-082025-04-082025-05-01Chemical Engineering Science 309: 121425 (2025)https://hdl.handle.net/11420/55264The COSMO-RS (Conductor-like Screening Model for Real Solvents) is a predictive thermodynamic model that has found diverse applications in various domains like chemical engineering, environmental chemistry, nanotechnology, material science, and biotechnology. Its core concept involves calculating the screening charge density on the surface of each molecule and letting these surface patches interact with each other to calculate thermodynamic properties. In this study, we aim to enhance the performance of the open-source implementation openCOSMO-RS by incorporating dispersive interactions between the paired segments. Several parametrizations were systematically evaluated through the extensive regression analysis using a comprehensive database of Vapor-Liquid Equilibrium (VLE), Liquid-Liquid Equilibrium (LLE) and Infinite Dilution Activity Coefficients (IDACs). Furthermore, the influence of different combinatorial terms on the model performance was investigated. Our findings indicate that incorporating dispersive interactions significantly improves the accuracy of phase equilibrium predictions for halocarbons and refrigerant mixtures.en0009-25092025Elsevierhttps://creativecommons.org/licenses/by/4.0/COSMO-RS | Dispersion | Parameterization | RefrigerantsNatural Sciences and Mathematics::541: Physical; Theoretical::541.3: Physical ChemistryJournal Articlehttps://doi.org/10.15480/882.1504010.1016/j.ces.2025.12142510.15480/882.15040Journal Article