Automated and efficient quantum chemical determination and energetic ranking of molecular protonation sites.

We present an automated quantum chemical protocol for the determination of preferred protonation sites in organic and organometallic molecules containing up to a few hundred atoms. It is based on the Foster-Boys orbital localization method, whereby we automatically identify lone pairs and π orbitals as possible protonation sites. The method becomes efficient in conjunction with the robust and fast GFN-xTB semiempirical method proposed recently (Grimme et al., J. Chem. Theory Comput. 2017, 13, 1989). The protonated isomers that are found within a few seconds to minutes of computational wall-time on a standard desktop computer are then energetically refined using density functional theory (DFT), where we use a high-level double-hybrid reference method to benchmark GFN-xTB and low-cost DFT approaches. The proposed DFT/GFN-xTB/LMO composite protocol is generally applicable to almost arbitrary molecules including transition metal complexes. Importantly it is found that even in electronically complicated cases, the GFN-xTB optimized protomer structures are reasonable and can safely be used in single-point DFT calculations. Corrections from energy to free energy mostly have a small effect on computed protomer populations. The resulting protomer equilibrium is valuable, for example, in the context of electrospray ionization mass spectrometry where it may help identify the ionized species and assist the interpretation of the experiment.