Chemometric Analysis of Substituent Effects. X. Unification and Reparametrization of Substituent Constants in Hammett-Type Equations by Means of Alternative Interpretation of Substituent Effects (AISE)

Abstract

The paper deals with evaluation of validity of substituent scales σm, σp, σp0, σp+, and σp- in Hammett-type equations. Possible reasons of unreliability of these scales and conceivable ways of verification of individual substituent constants are discussed. On the basis of Alternative Interpretation of Substituent Effects (AISE) and four additional conditions, a regression nonlinear model has been suggested which includes the dependence of all above-mentioned substituent scales upon a single interpreting variable - the σi substituent constant. This model has been applied to the 32 most frequent substituents from each scale, i.e. altogether 160 substituent constants taken from literature. The optimization of 17 unknown regression parameters gave the residual standard deviation s = 0.180 and, after excluding 12 remote values, s = 0.061. The analysis of residua proved their standard and statistical appropriateness of the model. The agreement between literature values and predicted values of substituent constants was the best with the σm scale; on the other hand, the largest differences were encountered with the σp+ scale. The found differences above a current reliability limit of substituent constants were due first to unreliability of some literature data and next to solvent effects and (in some cases) specific properties of substituents.