Polymer Blend Model

Most commercial products made of polymers are, in fact, a mixture of more than one component and, in many instances, of more than one polymer. Mixing polymers and additives is essential to enable processing and to obtain good product properties, yet the mixing behavior of polymer blends is still not well understood, and what little is known has been learned empirically. Because of the time and expense involved in experimental determination of multi-component phase diagrams, a model capable of predicting the miscibility of polymer mixtures would provide a powerful tool for academic and industrial researchers.

We have developed a new model for polymer mixtures that builds upon the classical Flory-Huggins free energy for polymer solutions and mixtures, taking into account component compressibilities. The model requires only pure component parameters as input, allowing for the prediction of complex phase behaviors such as lower critical solution transition (LCST) behavior often observed in polymer-containing mixtures. Combined with group contribution approaches, the model allows for a priori prediction of phase diagrams for any binary polymer mixture, provided the components are weakly interacting (i.e., absent hydrogen bonding or electrostatic interactions). The model has successfully predicted the qualitative phase behavior of over 30 binary polymer mixtures, and been extended to multi-component polymer blends to obtain spinodal curves for three component systems, employing only pure component properties of coefficients of thermal expansion, densities and solubility parameters.