Using the unusual weak N…CO bond as a solvation probe

Most chemical (or biochemical) reactions take place in a liquid solvent. Water is the natural solvent of biochemical reactions, and is increasingly used for organic synthesis. As water is not an inert solvent, modelling such a complex environment and evaluating the solvation effects on the electronic structure of the solute is a challenge. The unusual weak bond is used here as a probe of the solvation effects. Several solvation models were tested: a continuum, 1027 TIP3P water molecules and a microsolvation model complemented by a continuum or by TIP3P molecules. First, we show that the combination of the topological analysis of the electron localisation function (ELF) and the theory of atoms in molecules (AIM) is a robust way to evaluate and rationalise the strength and the accuracy of a given solvation model. These analyses demonstrate that a polarisable continuum model (PCM) is less accurate than a quantum mechanics/molecular mechanics (QM/MM) calculation where only the probe molecule is included in the QM region. We also show that solvating the solute and its first solvation shell embedded in a PCM leads to the same polarisation effect as a costly QM/MM calculation with 30 H2O included in the QM part and approximately 1000 classical water molecules. Finally, beyond this work, we show here that the combined ELF and AIM analyses can open up new opportunities for the electronic description of environment effects, for example in dynamical calculations.