Prediction of solid-liquid-gas equilibrium for binary mixtures of carbon dioxide + organic compounds from approaches based on the COSMO-SAC model

Six predictive approaches based on the Peng-Robinson (PR) equation of state (EOS), conductor-like screening model segment activity coefficient (COSMO-SAC), and mixing rules were applied to model solid-liquid-gas equilibrium for 21 binary mixtures of CO₂ and an organic compound. The accuracy of these approaches in predicting equilibrium temperatures at given pressures (635 experimental data with T = 220 ∼ 413.97 K and P = 0.05 ∼ 48.35 MPa), liquid phase compositions, and liquid molar volumes was examined and compared to provide an overview on their performance. The recently developed PR + COSMO-SAC EOS was found to be most accurate, with deviations of 6.25 K in temperature, 0.071 in liquid mole fraction, and 21% in liquid molar volume. The performance of these models can be very different for the solid containing different functional groups. Nevertheless, the PR + COSMO-SAC EOS could provide useful a priori predictions with only input of experimental heat of fusion and melting temperature of the solid.