The Atomic Structure of the Liquid-Vapor Interface of a Metal: An Example of the Influence of Strong Density Dependence of the Interatomic Potential

Abstract

Experimental studies and Monte Carlo simulations have established that the atomic structure of the liquid-vapor interface of a simple metal is significantly different from that of a Lennard-Jones system. In the latter the density along the normal through the interface (the longitudinal density distribution) monotonically decreases from the liquid to the vapor side of the interface; near the triple point of the system this variation in density occurs over a range of two to three atomic diameters [1]. In the former the longitudinal density distribution shows stratification on the liquid side of the interface and decreases rapidly, within a fraction of an atomic diameter, to the gas density on the vapor side of the interface [2,3]. The differences between the atomic structures of the liquid-vapor interfaces of a simple metal and the Lennard-Jones fluid can be traced to the different forms of the interatomic potential energy functions in the two cases. The potential energy of a Lennard-Jones system is rather accurately represented as a sum of density independent pair interactions. In contrast, the potential energy of a metal is strongly density dependent. In the pseudopotential representation the potential energy of a metal has two terms; a sum of density dependent pair interactions and a density dependent one body potential, it is the electron density dependence of the one body potential that is most important in a region where the liquid is inhomogeneous, as in the liquid-vapor interface. In a homogeneous liquid metal the one body potential is constant everywhere and, although it dominates the binding energy of the system, has no influence on the atomic distribution.KeywordsCapillary WaveAtomic DiameterBody PotentialSmall Momentum TransferPseudopotential RepresentationThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.