A molecular Ornstein–Zernike study of popular models for water and methanol

Abstract

Some effective models of water ͑TIP3P, SPC, SPC/E, TIP4P͒ and methanol ͑OPLS, H1͒ are studied with the help of the molecular Ornstein-Zernike ͑MOZ͒ theory using the hypernetted chain ͑HNC͒ approximation. The quality of the results obtained within the HNC approximation is discussed by comparison with values from molecular dynamics ͑MD͒ simulations. The MOZ-HNC theory yields internal excess energies and dielectric constants which are about 20% smaller than the simulation results found in the literature. The relative trends of the properties observed by simulation for the different interaction models are correctly predicted. In order to calculate the rotational invariant coefficients which define the liquid structure, new MD simulations were carried out. The rotational invariant coefficients derived from the simulation and from the MOZ theory strongly differ. In particular, the center-center distribution functions show that the theory is not able to reproduce the tetrahedral structure of water. In this solvent a comparison of the O-H distribution function indicates that the MOZ theory underestimates the H bonding. The use of a spherically symmetric bridge function in the reference HNC approximation does not lead to an improvement of the MOZ results. The observed defects of the MOZ-HNC approach contrast with the good agreement found for aprotic solvents and are presumably due to the association by H bonds in water and methanol.

Order and correlation contributions to the entropy of hydrophobic solvation

“…Rotational invariants have been used in molecular Ornstein-Zernike theory with other normalisations, [30][31][32][33][34] however we choose a normalisation of 1 to simplify higher order expansions in Φ l 1 l 2 l .…”

“…Richardi et al find the quadrupolar correlation function g 220 (r) in water and methanol to be an order of magnitude smaller than dipolar correlation functions. 33 …”

Molecular Density Functional Theory of Water

“…The numerical methods that have emerged in the second part of the last century from liquid-state theories [1,2], including integral equation theory in the interactionsite [3][4][5][6][7][8] or molecular [9][10][11][12][13][14] picture, classical density functional theory (DFT) [15][16][17], or classical fields theory [18][19][20][21], have become methods of choice for many physical chemistry or chemical engineering applications [22][23][24][25]. They can yield reliable predictions for both the microscopic structure and the thermodynamic properties of molecular fluids in bulk, interfacial, or confined conditions at a much more modest computational cost than molecular-dynamics or Monte-Carlo simulations.…”

Effects of bridge functions on radial distribution functions of liquid water

“…This analytical approach often gives fair results on par with those by extensive Monte Carlo or molecular dynamics (MD) simulations, and combined with the reference interaction site model (RISM) method (Chandler and Andersen, 1972;Hansen and McDonald, 2006; Hirata and Rossky, 1981), can comprehensively describe the equilibrium properties of molecular liquids such as liquid water (Pettitt and Rossky, 1982), which plays essential roles in a variety of biochemical processes. Recent developments in the methods, algorithms and benchmarks (Lombardero et al, 1999;Lue and Blankschtein, 1995;Reddy et al, 2003;Richardi et al, 1999;Sato, 2013;Sumi and Sekino, 2006) indicate that the RISM-based integral equation approach can provide an alternative route for theoretical analyses on water and related aqueous systems with comparable reliability to more expensive, computer simulation approaches. However, it has also been observed (Lombardero et al, 1999;Lue and Blankschtein, 1995;Reddy et al, 2003;Richardi et al, 1999;Sato, 2013;Sumi and Sekino, 2006) that the descriptions of intermolecular correlations of water become less accurate at room temperature in comparison with at higher temperatures.…”