Theory and simulation of solubility and partitioning in polymers:
Application of SAFT-γ Mie and molecular dynamics simulations

Abstract

Theory, simulation and experiment were used in this work to study the solubility and partitioning in polymer systems. The recently published SAFT-γ Mie equation of state was implemented into a stand-alone program together with all algorithms for parametrising new models and predicting phase equilibria. An analysis of the transferability of low-molecular weight Mie potential parameters for predicting the miscibility of polymer mixtures and partitioning of oligomers in polymer systems revealed the need for new models optimised for polymers. A systematic overview and analysis of available and typical experimental polymer data concluded that pure component polymer melt densities and cloud point temperatures (liquid–liquid equilibria) are the best and most practical choice for parametrising new SAFT-γ Mie models. New polymer models were developed for a range of pure polymers, several binary mixtures and one ternary polymer mixture. All models showed very good agreement with the experimental data included in the model development. Good agreement was found for predicted properties and conditions not included in the parametrising process. Coarse-grained (CG) force fields were developed with the help of the SAFT-γ Mie equation of state. Excellent agreement was found for the direct translation of Mie potentials to CG force fields for modelling properties of low-molecular weight compounds and densities of polymer melts. Coarse-grained models for molecular dynamics (MD) simulations of polymer phase equilibria are more challenging to develop due to greater computational resource requirements and less perfect agreement between SAFT-γ Mie and MD force fields. The challenges were demonstrated and discussed for a polystyrene solution and a binary mixture of polystyrene and polyisoprene. The synergistic power of SAFT-γ Mie and MD simulations was used for developing coarse grained models for describing the surface of a oligomer/polymer blend. Pure component parameters were optimised within SAFT-γ Mie. The SAFT-γ Mie CG model reproduced experimental partial density surface profiles as a function of blend composition without the need to rescale length scales. Oligomer surface enrichment, wetting transition and wetting layers were correctly predicted with a single model.