pK_a prediction in non-aqueous solvents

Area of Science:

• Physical Chemistry
• Computational Chemistry

Background:

• Acid dissociation constants (pKa) are fundamental chemical properties, primarily studied in aqueous solutions.
• Limited data and predictive models exist for pKa values in non-aqueous solvents, hindering broader chemical research.

Purpose of the Study:

• To develop and validate a computational method for predicting pKa values in diverse non-aqueous solvents.
• To assess the accuracy of this method across various solvents and molecular types, including drug molecules.

Main Methods:

• Utilized COSMO-RS (Conductor-like Screening Model for Real Solvents) for solvation energy calculations.
• Applied a reference-based approach, correcting pKa values from one solvent to another using calculated solvation energies.

Main Results:

• Achieved average deviations below 1 pKa unit from experimental data in six out of ten tested solvents.
• Demonstrated comparable accuracy for smaller molecules like amino acids and drug candidates.
• Identified specific solvents and molecular characteristics that influence model performance, noting higher errors for larger molecules.

Conclusions:

• The developed method provides a reliable approach for estimating pKa values in non-aqueous solvents.
• The technique can be extended to calculate proton transfer energies and solvation energies, contributing to a better understanding of chemical processes in different media.