Unveiling the Constraints of COSMO-SAC for PEG-Water Liquid-Liquid Equilibrium Prediction

This work studies the constraints of the COSMO-SAC model for predicting the liquid-liquid equilibrium (LLE) of polyethylene glycol (PEG) and water mixtures. The research extends the COSMO-SAC methodology to include polymer end groups in the σ-profile construction and employs a global optimization approach for LLE prediction across the entire composition range. The influence of various parameters, including end groups, hydrogen bond energies, and volume effects, on the LLE prediction is analyzed. The results indicate that while the COSMO-SAC model can qualitatively predict the phase behavior of PEG-water systems, including, for the first time, the closed-loop diagram, significant deviations from experimental data are observed for the LCST. The results obtained suggest that further improvements are required in both the combinatorial and residual terms to enhance the predictive accuracy of COSMO-SAC for complex polymer-solvent systems.