Prediction of solubility in nonideal multicomponent systems using the UNIFAC group contribution model

There is a need to identify suitable blends of solvents to dissolve drugs. Empirical approaches, such as trial-and-error and response surface, require several solubility measurements. In this study the UNIFAC method was used to predict solubility in highly nonideal multicomponent systems in which only the solute enthalpy of fusion and melting point must by measured. UNIFAC combines a group contribution approach with the UNIQUAC model for activity coefficients. Parameters characterizing interactions among constituent groups of a molecule have been previously determined from binary vapor pressure data. These tabulated group parameters are used to predict activity coefficients for newly synthesized compounds. These coefficients, together with the ideal solubility, permit a prediction of solubility. The solubility of 4-hexylresorcinol in ethyl acetate, ethyl myristate, and hexane mixtures was both measured and calculated using UNIFAC. The predicted solubilities were within 10% of the experimental solubilities for all but 3 of 21 mixtures. Since the method accounted for positive and negative deviations from ideality in a hydrogen-bonding system of molecules having different sizes, it shows great potential for use in pharmacy.