Using COSMO-SAC models in the mixing rules of the translated-consistent Peng-Robinson equation-of-state: a rigorous benchmarking across a broad and diverse database

Among the various continuum solvation models available, the COSMO-type ones are the most accurate option as they incorporate a robust statistical thermodynamic formulation combined with quantum-based calculations, ensuring a good level of accuracy and a high predictive power. Therefore, COSMO-type models offer an interesting alternative to group contribution-based approaches like UNIFAC, whose applicability relies on the availability of functional groups and their associated interaction parameters. When it comes to calculating thermodynamic data at low and high pressures, UNIFAC is typically used as the excess Gibbs energy model within the mixing rules of cubic equations of state, such as PSRK, VTPR, and UMR-PRU. In view of the advantages associated with the COSMO-based approach, this work aims to systematically analyze the use of COSMO-based models within mixing rules of cubic equations of state. For this purpose, three versions of the COSMO-SAC model (COSMO-SAC 2002, 2010, and dsp) were implemented (and re-parameterized) alongside the translated-consistent Peng-Robinson equation of state ( tc -PR EoS) for a comprehensive benchmarking study based on a large and diverse experimental database of 160 binary mixtures. This database comprises data on vapor–liquid equilibria at both low and high pressures, as well as critical points and mixing properties, such as enthalpy and heat capacity. By combining the tc -PR EoS with more advanced COSMO-SAC models (2010 and dsp), it was possible to effectively represent the phase behavior of a wide range of mixtures, consistently reproducing complex phase diagrams at least from a qualitative point of view and, in many cases, also quantitatively.