Finite-element method for calculation of the effective permittivity of random inhomogeneous media

The challenge of designing new solid-state materials from calculations performed with the help of computers applied to models of spatial randomness has attracted an increasing amount of interest in recent years. In particular, dispersions of particles in a host matrix are scientifically and technologically important for a variety of reasons. Herein, we report our development of an efficient computer code to calculate the effective (bulk) permittivity of two-phase disordered composite media consisting of hard circular disks made of a lossless dielectric (permittivity epsilon(2)) randomly placed in a plane made of a lossless homogeneous dielectric (permittivity epsilon(1)) at different surface fractions. Specifically, the method is based on (i) a finite-element description of composites in which both the host and the randomly distributed inclusions are isotropic phases, and (ii) an ordinary Monte Carlo sampling. Periodic boundary conditions are employed throughout the simulation and various numbers of disks have been considered in the calculations. From this systematic study, we show how the number of Monte Carlo steps needed to achieve equilibrated distributions of disks increases monotonically with the surface fraction. Furthermore, a detailed study is made of the dependence of the results on a minimum separation distance between disks. Numerical examples are presented to connect the macroscopic property such as the effective permittivity to microstructural characteristics such as the mean coordination number and radial distribution function. In addition, several approximate effective medium theories, exact bounds, exact results for two-dimensional regular arrays, and the exact dilute limit are used to test and validate the finite-element algorithm. Numerical results indicate that the fourth-order bounds provide an excellent estimate of the effective permittivity for a wide range of surface fractions, in accordance with the fact that the bounds become progressively narrower as more microstructural information is incorporated. Future directions of the active field of computational studies of the structure-property relations for composite systems are briefly discussed.