Approach to approximating the pair distribution function of inhomogeneous hard-sphere fluids

Beyond the mean-field approximation for pair correlations in classical density functional theory: Reference inhomogeneous non-associating monomeric fluids for use with SAFT-VR Mie DFT

A free-energy functional is presented to explicitly take into account pair correlations between molecules in inhomogeneous fluids. The framework of classical density functional theory (DFT) is used to describe the variation in the density of molecules interacting through a Mie (generalized Lennard-Jones) potential. Grand Canonical Monte Carlo simulations are performed for the systems to validate the new functional. The statistical associating fluid theory developed for Mie fluids (SAFT-VR Mie) is selected as a reference for the homogeneous bulk limit of the DFT and is applied here to systems of spherical non-associating particles. The importance of a correct description of the pair correlations for a reliable representation of the free energy in the development of the equation of state is duly noted. Following the Barker-Henderson high-temperature expansion, an analogous formulation is proposed from the general DFT formalism to develop an inhomogeneous equivalent of the SAFT-VR Mie free energy as a functional of the one-body density. In order to make use of this new functional in adsorption studies, a non-local version of the DFT is considered, with specific weighted densities describing the effects of neighboring molecules. The computation of these quantities is possible in three-dimensional space for any pore geometry with repulsive or attractive walls. We showcase examples to validate the new functional, revealing a very good agreement with molecular simulation. The new SAFT-DFT approach is well-adapted to describe realistic complex fluids.

A molecular density functional theory for associating fluids in 3D geometries

A new free-energy functional is proposed for inhomogeneous associating fluids. The general formulation of Wertheim's thermodynamic perturbation theory is considered as the starting point of the derivation. We apply the hypotheses of the statistical associating fluid theory in the classical density functional theory (DFT) framework to obtain a tractable expression of the free-energy functional for inhomogeneous associating fluids. Specific weighted functions are introduced in our framework to describe association interactions for a fluid under confinement. These weighted functions have a mathematical structure similar to the weighted densities of the fundamental-measure theory (i.e., they can be expressed as convolution products) such that they can be efficiently evaluated with Fourier transforms in a 3D space. The resulting free-energy functional can be employed to determine the microscopic structure of inhomogeneous associating fluids of arbitrary 3D geometry. The new model is first compared with Monte Carlo simulations and previous versions of DFT for a planar hard wall system in order to check its consistency in a 1D case. As an example of application in a 3D configuration, we then investigate the extreme confinement of an associating hard-sphere fluid inside an anisotropic open cavity with a shape that mimics a simplified model of zeolite. Both the density distribution and the corresponding molecular bonding profile are given, revealing complementary information to understand the structure of the associating fluid inside the cavity network. The impact of the degree of association on the preferential positions of the molecules inside the cavity is investigated as well as the competition between association and steric effect on adsorption.

Mean-field theory of inhomogeneous fluids

The Barker-Henderson perturbation theory is a bedrock of liquid-state physics, providing quantitative predictions for the bulk thermodynamic properties of realistic model systems. However, this successful method has not been exploited for the study of inhomogeneous systems. We develop and implement a first-principles "Barker-Henderson density functional," thus providing a robust and quantitatively accurate theory for classical fluids in external fields. Numerical results are presented for the hard-core Yukawa model in three dimensions. Our predictions for the density around a fixed test particle and between planar walls are in very good agreement with simulation data. The density profiles for the free liquid vapor interface show the expected oscillatory decay into the bulk liquid as the temperature is reduced toward the triple point, but with an amplitude much smaller than that predicted by the standard mean-field density functional.

Random phase approximation for the non-uniform Yukawa fluid

Mean-field density functional theory can be used to estimate the free energy of non-uniform fluids. The second functional derivative with respect to density of the free energy is related to the direct correlation function of the fluid and, in principle, this can be inverted to find an improved approximation for the pair correlation function and hence the free energy, the so-called 'random phase approximation'. If the repulsive molecular interaction is approximated by the local density approximation and the attractive interaction is assumed to be of the Yukawa form, the problem reduces to that of finding the eigenvalues of Schrödinger-like equations, which, for certain models (such as the 'Φ⁴ model'), can be done analytically in the planar case. The relationship between this approach and field theoretical treatment of the vapour-liquid interface is discussed. The ultraviolet divergence of the expression can be eliminated by separating the first term in the expansion, although quantitative results still depend on the behaviour of the attractive potential in the repulsive core. In the case of a spherical droplet of radius R, correction terms to the free energy involving lnR appear due to (i) cluster translational invariance, (ii) the unstable mode corresponding to droplet growth, and (iii) capillary waves. The net effect of these terms is to modify the classical expression for the nucleation rate by a factor proportional to R ^4/3.

Models for randomly distributed nanoscopic domains on spherical vesicles

The existence of lipid domains in the plasma membrane of biological systems has proven controversial, primarily due to their nanoscopic size-a length scale difficult to interrogate with most commonly used experimental techniques. Scattering techniques have recently proven capable of studying nanoscopic lipid domains populating spherical vesicles. However, the development of analytical methods able of predicting and analyzing domain pair correlations from such experiments has not kept pace. Here, we developed models for the random distribution of monodisperse, circular nanoscopic domains averaged on the surface of a spherical vesicle. Specifically, the models take into account (i) intradomain correlations corresponding to form factors and interdomain correlations corresponding to pair distribution functions, and (ii) the analytical computation of interdomain correlations for cases of two and three domains on a spherical vesicle. In the case of more than three domains, these correlations are treated either by Monte Carlo simulations or by spherical analogs of the Ornstein-Zernike and Percus-Yevick (PY) equations. Importantly, the spherical analog of the PY equation works best in the case of nanoscopic size domains, a length scale that is mostly inaccessible by experimental approaches such as, for example, fluorescent techniques and optical microscopies. The analytical form factors and structure factors of nanoscopic domains populating a spherical vesicle provide a new and important framework for the quantitative analysis of experimental data from commonly studied phase-separated vesicles used in a wide range of biophysical studies.