Transport properties of polydisperse hard-sphere fluid: effect of distribution shape and mass scaling

Abstract

A model polydisperse fluid represents many real fluids, such as colloidal suspensions and polymer solutions. In this study, we consider a concentrated size-polydisperse hard-sphere fluid with size derived from two different distribution functions, namely, uniform and Gaussian, and explore the effect of polydispersity and mass scaling on the transport properties in general. A simple analytical solution based on the Boltzmann transport equation is also presented (together with the solution using Chapman–Enskog (CE) method) using which various transport coefficients are obtained. The central idea of our approach is the realisation that, in polydisperse systems, the collision scattering cross-section is proportional to a random variable z which is equal to the sum of two random variables \(\sigma _i\) and \(\sigma _j\) (representing particle diameters), and the distribution of z can be written as the convolution of the two distributions \(P(\sigma _i)\) and \(P(\sigma _j)\). In this work, we provide expressions for transport coefficients expressed as an explicit function of polydispersity index, \(\delta \), and their dependence on the nature of particle size distribution and mass scaling is explored. It is observed that in the low polydispersity limit, the transport coefficients are found to be insensitive to the type of size distribution functions considered. The analytical results (for diffusion coefficients and thermal conductivity) obtained using the CE method and our simple analytical approach agree well with the simulation. However, for shear viscosity, our analytical approach agrees for \(\delta \le 20\%\), while it agrees up to \(\delta \approx 40\%\) with the result obtained using the CE method (in the limit \(\delta \rightarrow 0\)). Interestingly, the effect of scaling mass (i.e., mass proportional to the particle size and thus a random variable) produces no significant qualitative difference.