Thermodynamic driving force for molecular diffusion-lattice density functional theory predictions (Q861379)

The authors are experts in molecular diffusion thermodynamics. The article under review is a part of a sequence of articles published in J. Chem. Phys. 105 (1996), also 111 (1999), also 112 (2000), also 116 (2002), also 121 (2004), and in Chem. Eng. Sci. 60 (2005), J. Phys. Chem, Chem. Phys. 8 (2006), and other articles with different coauthors. These articles generally start with discussion of the simplest linear model known as Fick's first law, published in 1855, where the diffusive flux of the \(n\)th species is assumed to be directly proportional to the concentration of the \(n\)th species. The proportionality constant is the transport diffusivity of these species (\(-D_n\)), which in fact may not turn out to be constant during a typical diffusive flow. For vapors the pressure gradient is generally used to describe diffusive flux. However, now the proportionality ``constant'' turns out to be dependent on thermal energy. For such ``non-ideal'' problems the fugacity gradient is used, with supposedly better results. Fugacity \(f\) is defined by a formula relating to the rise in chemical potential \(\Delta\mu_n\). Thus: \(\Delta\mu_n= Kt\ln(f_n/P_n)\). Here \(P_n\) is the partial pressure of the \(n\)th species of vapor, while \(kT\) is the thermal energy of the entire multi-component vapor. Linear equations which are very popular can be at best considered approximations to the Maxwell-Stefan equations. Since so many approaches are possible; it is convenient to bypass the question: ``gradient of what is responsible for the molecular diffusion?'' Let us denote this ``what'' by \(\Gamma\). Now the authors produce a detailed analysis of \(\Gamma_n\). First they prove that an integral along a path describing the process \(\int c\Gamma_n\) exists, where \(c\) is the integrating factor, depending on the path. We know that integrating factors exist for such paths. For example, the classical Clausius formula defining entropy uses \(T^{-1}\) as the integrating factor, where \(T\) is the absolute (Kelvin) temperature. The authors argue that diffusion is very similar to heat transfer. Specifically they study how diffusion is path dependent and produce an integrating factor within their framework of lattice density functional theory. Various paths are considered in detail, and it is shown how some proposed forces result in well known diffusion equations. The authors have deliberately bypassed probabilistic arguments, including Monte-Carlo, only mentioning the fact that the equations of state may have been derived by such probabilistic approaches. This article provides interesting information about the state of art in studies of molecular diffusion theories. It is well-written, contains several interesting graphs, and a lengthy list of references, with references of publications dated past 1960 consisting almost exclusively of the USA studies.