Fluid–fluid phase separations in nonadditive hard sphere mixtures

• Published in: The Journal of chemical physics 1995-01, Vol.102 (3), p.1349-1360
• Main Authors: Jung, Jinkyung, Jhon, Mu Shik, Ree, Francis H.
• Format: Article
• Language: English
• Subjects: 360602 - Other Materials- Structure & Phase Studies; AGGLOMERATION; BINARY MIXTURES; CALCULATION METHODS; CHEMICAL COMPOSITION; COLLISIONS; DENSITY; DISPERSIONS; DISTRIBUTION FUNCTIONS; EQUATIONS; EQUATIONS OF STATE; FLUCTUATIONS; FLUIDS; FUNCTIONS; HARD-SPHERE MODEL; INTEGRAL EQUATIONS; MATERIALS SCIENCE; MIXTURES; MOLECULE COLLISIONS; MONTE CARLO METHOD; PERCUS-YEVICK EQUATION; PHASE STUDIES; PHYSICAL PROPERTIES; SAMPLING; SHAPE; SIZE; VAN DER WAALS FORCES; VARIATIONS
• ISSN: 0021-9606; 1089-7690
• DOI: 10.1063/1.468921

We investigated the phase stability of a system of nonadditive hard sphere (NAHS) mixtures with equal diameters, d, between like species and an unequal collision diameter, d(1+α), between unlike species. It is based on an analytic equation of state (EOS) which refines an earlier expression [J. Chem. Phys. 100, 9064 (1994)] within the mixed fluid phase range. The new EOS gives a reliable representation of Monte Carlo EOS data over a wide range of density, composition, and nonadditivity parameters (α). Comparisons with available computer simulations show that the new EOS predicts satisfactory phase boundaries and the critical density line. It is superior to results derived from integral equations (the Percus–Yevick, the Martynov–Sarkisov, and the modified Martynov–Sarkisov) and analytic theories (the MIX1 model, the van der Waals one-fluid model, and the scaled particle theory). The present study shows that, unless α exceeds 0.026, the fluid phase will remain fully miscible up to the freezing point of pure hard spheres. We have also investigated structural aspects of the phase stability by Monte Carlo computations. The radial distribution functions, the local mole fraction, and coordination numbers for like and unlike pairs of hard spheres exhibit significant number dependencies close to the fluid phase boundary. They provide precursory signals to an impending phase change. Finite systems used in the Monte Carlo sampling limit fluctuations in sizes and shapes of heterogeneous clusters. The observed number dependence simply reflects this fact.