Abstract

The free energy of solvation, ?G s 0, in octanol of organic compunds is quantitatively predicted from the molecular structure. The model, involving only three molecular descriptors, is obtained by multiple linear regression analysis from a data set of 147 compounds containing diverse organic functions, namely, halogenated and nonhalogenated alkanes, alkenes, alkynes, aromatics, alcohols, aldehydes, ketones, amines, ethers and esters; covering a ?G s 0 range from about -50 to 0 kJmol -1. The model predicts the free energy of solvation with a squared correlation coefficient of 0.93 and a standard deviation, 2.4 kJmol -1, just marginally larger than the generally accepted value of experimental uncertainty. The involved molecular descriptors have definite physical meaning corresponding to the different intermolecular interactions occurring in the bulk liquid phase. The model is validated with an external set of 36 compounds not included in the training set.