Granule-by-granule reconstruction of a sandpile from x-ray microtomography data

Mesoscale disordered materials are ubiquitous in industry and in the environment. Any fundamental understanding of the transport and mechanical properties of such materials must follow from a,thorough understanding of their structure. However, in the overwhelming majority of cases, experimental characterization of such materials has been limited to first- and second-order structural correlation functions, i.e., the mean filling fraction and the structural autocorrelation function. We report here the successful combination of synchrotron x-ray microtomography and image processing to determine the full three-dimensional real-space structure of a model disordered material, a granular bed of relatively monodisperse glass spheres. Specifically, we determine the center location and the local connectivity of each granule. This complete knowledge of structure can be used to calculate otherwise inaccessible high-order correlation functions. We analyze nematic order parameters for contact bonds to characterize the geometric anisotropy or fabric induced by the sample boundary conditions. Away from the boundaries we find short-range bond orientational order exhibiting characteristics of the underlying polytetrahedral structure.