Ion interactions with non-polar and amphiphilic solutes in water

Despite its extensive use in, for example, the concentration and crystallization of proteins, the salting out phenomenon remains poorly understood, with many--sometimes contradictory--explanations found in the literature. Following our earlier work using isotope substitution neutron scattering on an aqueous tertiary butyl alcohol (TBA) with added NaCl that examined in detail the molecular-level interactions in the system, and suggested that attempts to understand salting out through ion perturbation of the non-polar hydration shell may not be appropriate, we report here on two further sets of high resolution structural experiments that detail the structural interactions between ion, solvent and amphiphile in a wider range of relevant systems. First, a set of X-ray absorption spectroscopy experiments probed the effects on the hydrophobic hydration shell of Kr of adding a range of different salts (specifically Na2SO4, NaClO4, NaCl, NaNO3 and Mg(ClO4)2). The bottom line from these experiments is that the hydration shell is essentially unaffected by the added ions, underlining further the stability of the shell to such perturbations. The second set of experiments was a further isotope substitution neutron scattering study of the molecular-level solution structures of aqueous TBA, this time for a range of different added ions (CsF, NaBr, and NaCl at three concentrations), for which we have in addition extracted equilibrium constants to quantify the relative strengths of the various interactions. From this second set of results, we can conclude again that the solvent structure is essentially unperturbed by both the various ions and the amphiphile, and also identify a number of ion-specific effects. Comparing our results with those obtained from simulations that are not constrained by the experimental data underlines how sensitive structural conclusions are to the assumed potential functions, and that drawing conclusions from simulations not constrained to fit experimental data can lead to seriously erroneous conclusions.