A research program by Salvatore Torquato, the principal investigator, and his colleagues in the Dept. of Chemistry & Princeton Materials Institute is aimed at studying the statistical geometry of jammed hard-particle packings. Efficient and novel algorithms are developed to generate monodisperse and polydisperse packings of hard spheres, hard ellipsoids, and hard spherocylinders in two and three dimensions. In particular, efforts are directed toward improving the Lubachevsky-Stillinger and Zinchenko packing algorithms. A randomized linear-programming algorithm will be devised to test for ``jamming'' categories in the generated hard-particle systems, as well as extensions to deal with nonlinearities in the impenetrability constraints. The generated jammed packings will be statistically characterized by using high-quality scalar order metrics and by collecting Voronoi tessellation statistics. Finally, the aforementioned tools and algorithms will be employed to investigate the space of jammed structures, including the identification of the maximally random jammed (MRJ) state and low-density jammed packings.
Important scientific advances and outcomes are likely to emerge from the proposed research. This multidisciplinary project joins the statistical physics, materials science, geometry, applied mathematics, and high-performance computing communities. Unifying these different perspectives and goals is a challenging task, but one that offers great rewards to the computational-science community in general. Because particle packings are widely used as simple models for granular materials, glasses, liquids, and other random media, the proposed research will have important implications for the development of novel materials as well as nanotechnology.
