Thermodynamic vs. extrathermodynamic modeling of chromatographic retention

To predict a given physicochemical or biological property, and hence, to design rationally requested chemical entity, the relationships must be identified between the chemical structure and the desired property. Unfortunately, classical thermodynamics never predicts any property by itself, even so simple one like chromatographic retention. Therefore progress in understanding and describing molecular equilibrium between phases requires a combination of experimental measurements and correlations by means of empirical equations and approximate theories. In this work the retention prediction performance was tested of the well thermodynamically founded solvophobic theory of Horváth and co-workers of reversed-phase HPLC. The retention parameters of four series of analytes were modeled with regard to their chemical structure by: (1) observing the rules of classical thermodynamics; (2) applying an extrathermodynamically derived correction to the model based on the thermodynamic hermeneutics; (3) using extrathermodynamic, chemical intuition-based Quantitative Structure-Retention Relationships (QSRR). The combined thermodynamic/extrathermodynamic model with empirical correction accounting for the number of polar atoms provided an improvement of the agreement between the observed and the predicted retention parameters. However, a purely extrathermodynamic QSRR model, employing analyte descriptors from calculation chemistry, produced similar retention predictions. Both thermodynamic and QSRR models accounted well for abilities of analyte to participate in nonspecific, dispersive intermolecular interactions. Less reliable appeared descriptors of analyte polarity. The approach presented here can be further developed to search for proper polarity parameters, necessary to correctly predict complex physicochemical and biological properties of chemical compounds.