Prediction of Phase Behaviors of Polymer–Solvent Mixtures from the COSMO-SAC Activity Coefficient Model

Abstract

An efficient method for prediction of the vapor–liquid and liquid–liquid phase behaviors of polymer–solvent mixtures is proposed using the COSMO-SAC activity coefficient model. In particular, we examine various approaches for generating the screening charge distribution of homepolymers and copolymers from quantum mechanical calculations and propose a novel method to generate such data efficiently using a finite number of repeating units. The free volume effects known to be important in polymer solutions are considered through the free volume model of Elbro and co-workers. We have examined this model using 2249 vapor–liquid equilibria data points from 25 homopolymers, 8 copolymers, and 48 solvents. The overall average deviation (AAD%) of solvent activity is found to be about 16%, which is comparable to that from other UNIFAC-based models. The liquid–liquid equilibrium of polymer–solvent mixtures can be predicted in qualitative agreement with experiment. The proposed method is capable of describing the changes in solvent activity due to temperature, polymer molecular weight, and polymer tacticity. Our results suggest that the proposed method is a useful complementary to group contribution method for phase behavior of polymer solutions when group parameter or density information of polymers is unavailable.