| Abstract: | Accurate prediction of the acidity dissociation constant (K a) is a challenge for the theory of proton-transfer reactions, making first-principles prediction of pK a within 0.5?pK units of experimental values a benchmark of broad interest. In the present contribution, the defined-sector explicit solvent in continuum cluster model, which considers the structure-to-chemical affinity relationship of the carboxyl functional group, is presented. The model demonstrates predictable solvent networks based on established ‘preferred’ conformations found in a training set. Predictability within 1?kcal?mol–1 accuracy is shown for a full set of carboxylic acid systems with varying functionality. |
