How hydrophobic hydration responds to solute size and attractions: Theory and simulations

Abstract

We focus on the hydration of a methane and spherical single and multisite C60 and C180 solutes over a range of solute-water attractions to quantify the vicinal water structure and their hydration thermodynamics using extensive molecular dynamics simulations and theory. We show that water structure near larger solutes is more sensitive to solute-water attractions compared to that near smaller ones. To understand the sensitivity, we separate the solute-water potential of mean force into a direct solute-water interaction and an indirect or solvent contribution [omega(r)]. In the absence of omega(r), water density in the solute vicinity would increase exponentially with solute-water interactions. Instead, omega(r) becomes increasingly repulsive with strengthening of solute-water attractions thereby opposing those direct interactions. We term this phenomenon "competitive expulsion," which characterizes the repulsion of a test water molecule by the hydration shell solvent waters. We develop a physically motivated theoretical approach to predict changes in omega(r) with attractions. We call this approach the modified-EXP (M-EXP) approximation owing to the similarity of ideas and especially our final expression with that of the EXP approximation of Chandler and Andersen [J. Chem. Phys. 57, 1930 (1972)]. Solute-water radial distribution functions and chemical potentials calculated using the M-EXP approach are in good agreement with simulation data. These calculations highlight the sensitivity of hydration structure and thermodynamics of bucky ball like solutes to solute-water interactions. We find that excess chemical potentials of bucky balls with standard alkane-like carbon-water interactions parameters are negative, suggesting the need for a careful calibration of those parameters for predictions of solubility, wetting, and water-mediated interactions using molecular simulations.

Structural Order of Water Molecules around Hydrophobic Solutes: Length-Scale Dependence and Solute–Solvent Coupling

“…1−11 Despite enormous efforts combining both experimental 12−17 and theoretical 18−31 tools, the precise molecular nature of hydrophobic interactions has remained elusive and debates continue. 1,31−33 The primary reason behind the complexity of a "simple" solvent like water seems to be the strong context dependence of the water-mediated interactions (e.g., properties of bulk water vs surface water vs water in a confined medium, or water near a hydrophobic vs hydrophilic molecule/surface and so on).In particular, it has been theoretically predicted by Chandler and co-workers 11,18 and later validated by a multitude of computer simulation studies 20,22,23,34 that the molecular nature of hydrophobic interactions undergoes a qualitative change with the size of the hydrophobic solutes. When the hydrophobic solutes are smaller than a nanometer, water molecules can still maintain the bulklike hydrogen-bonded network around the solute since the local density fluctuations in water can create large enough cavities to accommodate such solutes.…”

Hydrophobic interactions with coarse-grained model for water

“…Fullerene-water radial distribution function at T=300 K and ρ=1 g/cm 3 . Symbols are the simulation results for the fullerene-water radial distribution function with all-atom SPC/E water, [20] and the line is from HNC theory for coarse-grained water described by the LJG potential. …”

“…(18) and (19), the above set of MCT equations for ζ needs to be solved iteratively and self-consistently. One could use a more accurate model for F s (k, t) in terms of the velocity time correlation function of a tagged fluid particle, [34] (14), (17), (19), and (20).…”

“…. Symbols are the simulation results for the oxygen-oxygen radial distribution function of all-atom SPC/E water, [20] and the line is from HMSA theory for coarse-grained water described by the LJG potential.…”

“…In Fig. 6 we compare simulation results for the fullerene-water radial distribution function with all-atom SPC/E water, [20] and the HNC result for g(r) with coarse-grained LJG water. The agreement between the two distribution functions is satisfactory in general, although theory underestimates the heights of both primary and secondary peaks of g(r).…”