PREDICTION OF RP-HPLC LOG-P FROM SEMIEMPIRICAL MOLECULAR-PROPERTIES OF DIPHENYL ETHER AND PHENOPYLATE HERBICIDES

The relationships between octanol/water partition coefficients (log P) estimated by reversed-phase high-performance liquid chromatography (R P-HPLC) and various bulk and electronic properties, calculated by mole cular and quantum mechanics methods, of diphenyl ether (DPE) and pheno pylate herbicides were examined. The three molecular parameters, van d er Waals volume, dipole moment, and energy of the highest occupied mol ecular orbital, together accounted for 78-83 % of the variation in the log P data. On the average, the predicted values using regression equ ations deviated from the observed values by only O.13 and O.19 log uni t in DPE and phenopylate classes, respectively. The same three molecul ar parameters appeared in the model when the structures of two classes of herbicides were included in the regression analysis. This research suggests that one can successfully predict the RP-HPLC partition coef ficients from the semiempirical molecular properties calculated from m olecular and quantum mechanical techniques. In addition, the RP-HPLC e stimated log P values and those calculated from the fragment additivit y method using the CLOGP computer program were highly correlated (r = 0.91).