pKa Prediction in Non-Aqueous Solvents

Jonathan W. Zheng,Emad Al Ibrahim,William H. Green
DOI: https://doi.org/10.26434/chemrxiv-2024-vx797-v2
2024-05-13
Abstract:Acid dissociation constants (pKa) are widely measured and studied, most typically in water. Comparatively few datasets and models for non-aqueous pKa values exist. In this work, we demonstrate how the pKa in one solvent can be accurately determined using reference data in another solvent, corrected by solvation energy calculations from the COSMO-RS method. We benchmark this approach in ten different solvents, and find that pKa values calculated in six solvents deviate from experimental data on average by less than 1 pH unit. We observe comparable performance on a more diverse test set including amino acids and drug molecules, with higher error for large molecules. The model performance in four other solvents is worse, with some MAEs exceeding 3 pH units; we discuss how such errors arise due to both model error and inconsistency in data calibration. Finally, we demonstrate how this technique can be used to estimate the proton transfer energy between different solvents, and use this to report a value of the proton’s solvation energy in formamide, a quantity that has does not have a consensus value in literature.
Chemistry
What problem does this paper attempt to address?
This paper mainly discusses the problem of predicting the acid dissociation constant (pKa) in non-aqueous solutions. Currently, most pKa data and models are focused on aqueous environments, with relatively little data and models for non-aqueous environments. The researchers proposed a method to accurately predict pKa in different solvents using the COSMO-RS solvation model and reference data in water, and estimated the transfer free energy of the proton between different solvents through thermodynamic cycles. This method performs well in six solvents, with an average deviation of less than 1 pH unit, but has larger errors in four other solvents, possibly due to model inaccuracies and inconsistent data calibration. The paper also discusses how the inconsistency of experimental data leads to prediction errors of the model. Using this method, they also estimated the solvation energy of the proton in formamide, which lacks a consensus value in the literature.