Solute-solvent interactions in VCD
It is commonly accepted that solvation occurs on a molecular level and that, for instance, carboxylic acids or alcohols form hydrogen bonds with acceptor solvents like DMSO or acetonitrile and that self-aggregation may occur in non-polar solvents. However, detailed structural information on these hydrogen bonded structures from solution phase is unprecedented in literature as rather generic structures (carboxylic acid dimer, multimeric chains of interacting OH groups, and OH∙∙solvent interactions) are usually sufficient to qualitatively explain experimental data such as thermodynamic equilibrium constants, reaction rates or even NMR chemical shifts. Actual information on the most important, i.e. the strongest and hence the structure-determining solute-solvent cluster geometries are difficult to determine in solution, as most spectroscopic techniques are not sensitive enough to distinguish different solute-solvent binding orientations. Considering the DMSO-solvated carboxylic acid shown below as an example, IR spectroscopy is not capable of differentiating the two structures in solution phase as their computed spectra are not significantly different. Due to its structural sensitivity, VCD spectroscopy can distinguish the solute-solvent clusters, and the comparison of the experimental data with computed spectra can be used to show that the linear form is present in solution, even though the bifurcated form is computed to be significantly more stable.
Whenever such solute-solvent interactions occur, it is important to consider them explicitly in spectra calculations in order interpret the experimental data correctly. Otherwise, when solvation is not properly accounted for, AC determinations can lead to wrongs results. Therefore, I have initiated a project dedicated to a better understanding of the influence of solute-solvent interactions on VCD spectra. Its overall aim is to derive general guidelines as to how certain functional groups have to be solvated explicitly in order to achieve a good match between experimental and theoretical spectra. We are systematically analyzing the VCD spectra of small chiral molecules recorded in various solvents. The selected molecules feature only one functional group capable of a single strong hydrogen bonding at a time in order to determine group-specific preferential solvation structures. Once a representative set of such well-tested micro-solvation environments of functional groups is obtained, the molecular size is systematically increased in order to study cooperative effects. The final empirical guidelines are envisioned to be directly applicable to any (chiral) molecule, enabling the explicit consideration of solvent interactions from the beginning of the spectra analysis.
Working in solution phase does not only require the consideration of explicit solvation in the calculations of spectroscopic properties of single conformers. In fact, the conformational preferences of a structurally complex molecule are often determined by the solvent environment and may be significantly different from gas phase or micro-solvated structures. Therefore, molecular dynamics (MD) simulations play an important role in context of solvation science in order sample the conformational space of flexible molecules in the presence of solvent. While there are approaches to derive VCD spectra directly from MD trajectories, we use MD simulations to guide and support the Boltzmann weighting of single conformer spectra in the simulation of VCD spectra from high-level DFT calculations. The success of this approach could be shown in a study on a dipeptide in different solvents and we currently extend the approach towards large polyketides and other highly flexible natural products.