Effect of Dimensionality on the Percolation Threshold of Overlapping Nonspherical Hyperparticles

Abstract

A set of lower bounds on the continuum percolation threshold $η_c$ of overlapping convex hyperparticles of general nonspherical (anisotropic) shape with a specified orientational probability distribution in $d$-dimensional Euclidean space have been derived [S. Torquato, J. Chem. Phys. {\bf 136}, 054106 (2012)]. The simplest of these lower bounds is given by $η_c \ge v/v_{ex}$, where $v_{ex}$ is the $d$-dimensional exclusion volume of a hyperparticle and $v$ is its $d$-dimensional volume. In order to study the effect of dimensionality on the threshold $η_c$ of overlapping nonspherical convex hyperparticles with random orientations here, we obtain a scaling relation for $η_c$ that is based on this lower bound and a conjecture that hyperspheres provide the highest threshold among all convex hyperparticle shapes for any $d$. This scaling relation exploits the principle that low-dimensional continuum percolation behavior encodes high-dimensional information. We derive a formula for the exclusion volume $v_{ex}$ of a hyperparticle in terms of its $d$-dimensional volume $v$, surface area $s$ and {\it radius of mean curvature} ${\bar R}$ (or, equivalently, {\it mean width}). These basic geometrical properties are computed for a wide variety of nonspherical hyperparticle shapes with random orientations across all dimensions, including, among other shapes, various polygons for $d=2$, Platonic solids, spherocylinders, parallepipeds and zero-volume plates for $d=3$ and their appropriate generalizations for $d \ge 4$. We then compute the lower bound and scaling relation for $η_c$ for this comprehensive set of continuum percolation models across dimensions. We show that the scaling relation provides accurate {\it upper-bound} estimates of the threshold $η_c$ across dimensions and becomes increasingly accurate as the space $d$ increases.