A universal approach for continuum solvent pKa calculations: are we there yet?

This paper reviews several pK(a) calculation strategies that are commonly used in aqueous acidity predictions. Among those investigated were the direct or absolute method, the proton exchange scheme, and the hybrid cluster-continuum (Pliego and Riveros) and implicit-explicit (Kelly, Cramer and Truhlar) models. Additionally, these protocols are applied in the pK(a) calculation of 55 neutral organic and inorganic acids in conjunction with various solvent models, including the CPCM-UAKS/UAHF, IPCM, SM6 and COSMO-RS, with a view to identifying a universal approach for accurate pK(a) predictions. The results indicate that the direct method is unsuitable for general pK(a) calculations, although moderately accurate results (MAD <3 units) are possible for certain classes of acids, depending on the choice of solvent model. The proton exchange scheme generally delivers good results (MAD <2 units), with CPCM-UAKS giving the best performance. Furthermore, the sensitivity of this approach to the choice of reference acid can be substantially lessened if the solvation energies for ionic species are calculated via the IPCM cluster-continuum approach. Reference-independent hybrid approaches that include explicit water molecules can potentially give reasonably accurate values (MAD generally similar to 2 units) depending on the solvent model and the number of explicit water molecules added.