Computational chemistry applied to the design of chiral stationary phases for enantiomeric separation

Abstract

AbstractA general computational scheme for the rational design of chiral stationary phases for the chromatographic separation of enantiomers has been established. The developed scheme was based on applying different interaction models (force field methods versus semi empirical quantum chemical methods), different docking algorithms (systematic grid search methods versus interactive methods guided by rules based on binding modes) and different levels of approximations (rigid versus flexible docking) to a representative test problem containing the 3,5‐dinitrobenzoyl group. The computational methods in use covered the most sophisticated methods which could presently be applied to problems of such a size (about 80 atoms). It has been shown that the current computational approaches using rigid body approximations for the docked molecules and simple molecular mechanics (not taking pi‐“effects” into account) are invalid in view of the required predictive precision of about 1–2 Kcal/mole for the differential binding energy. Another surprising result was the failure of the commonly used systematic search methods in determining the most favorable binding modes. Based on our calculations on the representative test problem we propose a new arrangement for the most stable complexes without parallel stacking of the aromatic pi‐donor and the 3,5‐dinitrobenzoyl pi‐acceptor systems.