Assessing the Partitioning of the Electrostatic Solvation Energy Following the Interacting Quantum Atoms Approach

Abstract

The Interacting Quantum Atoms (IQA) approach allows for a decomposition of the global molecular energy into atomic contributions which tend to the energies of isolated atoms at the limit of non-interaction. Within a molecule, one can discriminate between the own energy of an atom (ie., the one corresponding to the electron movement, repulsion and attraction to a given nucleus) and its interaction with the rest of the atoms in the system. It is possible to group atomic contributions, giving rise to fragment energies as system descriptors.A crucial issue to correctly model a molecule is to adequately reproduce its environment. To this end, the use of implicit solvation models in combination with quantum mechanical methods may provide an accurate description while preserving an affordable computational cost. It is therefore desirable to include solvation contributions for the group description of molecules within a particular solvent. Nevertheless, it is well known that solvation free energy cannot in principle be decomposed into independent contributions, since the contiguous groups highly influences the solvation sphere of a given atomic group.In this work we aim to assess the goodness of an IQA-like group partitioning of the electrostatic contribution to the solvation energy.1 For this purpose, we carry out COSMO-HF/aug-cc-pVTZ calculations followed by IQA analyses on more than 400 neutral and ionic solutes from the MNSol database. Structure-activity trends can be outlined from our results, attending to the constancy or dispersion of the fragment contributions and their relationship with structural or substituent effects. Overall, the dispersion of the fragment solvation energies for neutral species is moderate, although the reconstructed molecular solvation energies from fragment contributions are not satisfactory. Further improvements of IQA-based solvation energies shall account for correlation effects, a more adequate fragment selection, the inclusion of extra parameters and a larger dataset of solute molecules.