Correlation of gas–liquid partition coefficients using a generalized linear solvation energy relationship

In a number of previous communications, we reported on the utility of the solvatochromic linear solvation energy relationship (LSER) method for the correlation of a number of solute and solvent-dependent properties. In those studies, it was our practice to examine the effect of a variety of solvents on a given solute or a number of solutes in a given solvent. Here we report on a novel generalized LSER in which the solute and solvent were both simultaneously varied so as to assess the validity of the entire LSER concept and define its limits. The Hildebrand solubility parameter, δ_H, the Kamlet–Taft solvatochromic parameters, π*, α, β and the solute molar volume, V₂, were used as the explanatory variables. The gas–liquid partition coefficient (K) was the property of interest. We have found that the correlation using the generalized linear solvation energy relationship is statistically as good as the previous LSER correlations despite the use of a far smaller number of freely adjustable parameters. Furthermore, the new approach is able to give reasonable predictions of K values of systems not included in the data set upon which the regression is based.