Collect. Czech. Chem. Commun. 1995, 60, 829-840

Chemometric Analysis of Substituent Effects. VI. A Study of ortho Effect in Dissociation of 2,6-Disubstituted Benzoic Acids

Jiří Kulhánek and Oldřich Pytela

Department of Organic Chemistry, University of Pardubice, 532 10 Pardubice, Czech Republic


Ten 2,6-disubstituted benzoic acids have been synthesized containing all possible combinations of the following substituents: CH3, OCH3, Cl, and NO2. The dissociation constants of these acids have been measured by potentiometric titration in methanol, acetone, dimethyl sulfoxide, dimethylformamide, acetonitrile, pyridine, and 1,2-dichloroethane. The experimental data obtained together with the pK values of 2-substituted benzoic acids in the same solvents have been analyzed from the point of view of ortho effect and additivity of disubstitution. The mutual interaction between substituents was found to represent only 0.12% of the variability due to substitution and to contribute to the overall variability of data less than the interaction between the substituent and solvent by a factor of about 13. The analysis of data by the method of multiple linear regression revealed a contribution of steric effects beside the effects transmitted through the aromatic skeleton. The 2- and 6-substituents effects are additive within the validity of the Hammett equation, and an addition of a multiplicative term describing interactions between the substituents is statistically insignificant. Nonlinear regression has been adopted in the additive model with multiplicative term to find the inner substituent constants including all the effects of substituents from ortho position: the term describing the interaction between 2- and 6-substituents is statistically insignificant in this model. An application of the method of conjugated deviations revealed two statistically significant latent variables. The first one explains 91.5% of the variability of data and is connected with the substituent effects transmitted through the aromatic skeleton. The second one explains 7.5% of variability of data and predominantly reflects the steric effects of substituents.