In these studies we examined the structure, thermodynamics, and dynamics of water in both bulk and in peptide solutions using simulation and experiment. With these techniques we were able to carefully examine the molecular mechanisms that give rise to several observable equilibrium properties of water as a function of temperature and pressure. First we critically examined the implications of a recent experiment that concluded that liquid water does not form a tetrahedral network, as is commonly believed, but rather organizes into chains. We found that by imposing an asymmetric charge distribution on a water model, (an appropriate test of the chain-water hypothesis) the resulting structure of the water does not agree well with experimental data. Hence, we provided further evidence that this new interpretation of non-tetrahedral water is not consistent with experimental data. Secondly, through the use of systematically derived isotropic water-like potentials, we determined that the thermodynamic anomalies that occur in water as its temperature is lowered depend on the presence of explicit (directional) orientational interactions. However, waters diffusion anomaly can be observed without orientational interactions, as long as the pair density correlations (average structure) of water are conserved. The results of this study prompted a more general investigation into whether static properties of a liquid, such as pair structure and entropy, are sufficient variables to predict the macroscopic time scales (diffusion constant) of the liquid. We determined that some (but not all) entropic metrics are effective in describing the trends of diffusion with temperature and density, and that in particular, the pair correlation function contains a sufficient amount of information to describe diffusion trends over large regions of the phase diagram.

In our studies on the dynamics of hydration water in concentrated peptide solutions, we found that water molecules near peptides with both hydrophilic and hydrophobic regions (amphiphilic) exhibited translational motion that evolved on two distinct time scales. In contrast, the water near a purely hydrophilic peptide diffused on one single timescale. Through simulation we determined that in the amphiphilic peptide solution, the water forms hydration layers (an inner and outer layer) that exchange more slowly with one another than in the purely hydrophilic peptide solution. Ultimately as the temperature is lowered, the outer hydration water becomes more mobile than average, while the inner layer is slower than average, and this gives rise to the separation in timescales. Thus the presence of both the hydrophilic and hydrophobic regions in the solution introduces a larger perturbation on the water network. Lastly, through these solution simulation studies we found that our results were inaccurate when modeled with a fixed-charge empirical force field. We suggest that accurate simulations of multi-component solutions require that polarization effects be included in the force field, and we detail how polarization parameters can be derived.