| A reformulation of thermodynamic theories for predicting phase transitions in solutions, especially those with strong solute-solvent interactions, is proposed. It is based on providing an account of such interactions through a specific ratio of solute-solvent association, with a characteristic interaction energy, e.g., hydrogen bonding energy, which is different from those of other, non-polar interactions in the solution. Objective methods for estimating the parameters required for prediction from the theory are also described. The efficacy of the underlying concepts and methodologies has been verified by incorporating them in lattice models for solutions, and also in the group contribution method, UNIFAC.; The initial formulations have been obtained with the assumption of a constant solvent/solute association ratio. Molecular modeling methods have been used for estimating the required parameters. Methods for estimating these parameters from experimental measurements, such as spectroscopy, light scattering and osmotic pressure, have also been identified. Predictions from lattice (ART-L) and UNIFAC (ART-U) models have been compared with experimental data for a number of solutions, such as acrylonitrile/water, succinonitrile/water, and polyacrylonitrile/water/dimethyl formamide. Even with the simplifying assumptions which have been used in its initial implementation, the predictions of the current theory have been found to be in excellent agreement with experimental liquid-liquid phase separation data. The refinements and generalizations which should follow the initial formulation have been discussed, with an example of the enhancement in predictive capabilities which can be expected. |
