Particle motion in granular materials: three dimensional imaging of slow flows and compaction

Abstract

Granular materials have been a subject of study for centuries. Their bulk properties are well known and are quite reproducible. However, it is not well understood how motions at the grain level relate to bulk behaviors. In this thesis, we describe the use of a 3D imaging technique to determine the motions of individual grains in known geometries.

We use a method known as the Refractive Index-Matched Scanning (RIMS) method to locate and track centers of individual grains in dense granular piles. This method enables us to capture grain scale rearrangements where other techniques, such as displacement field imaging, may fail. We may also track motions of grains with respect to their nearest neighbors in order to measure local flows.

We first apply the RIMS method to the study of a gentle compaction process, known as thermal cycling. We track the centers of grains between temperature cycles, capturing cycle-to-cycle displacements. The tracks are used to generate dynamic Voronoi volumes about the centers of grains at each cycle. We are able to observe fluctuations in the shapes of the Voronoi volumes which correlate strongly with subsequent motion of grains. We find that the grains move preferentially toward the centroid of the vertices of their respective Voronoi cells.

We then study grain motions in quasistatic flows in a split-bottom geometry. We observe nearest neighbor separation events during both steady and cyclic shearing processes. We find a critical strain beyond which there is a qualitative change in the breakage of contacts between neighbors. Cyclic shear flows with amplitudes below this critical strain settle into a nearly reversible flow pattern, while those with amplitudes above the critical strain remain in a persistent diffusive, irreversible state.

Overall, the RIMS method is a powerful tool for probing the structure of slow granular flows. We are now able to examine particle level rearrangements which were previously explored in simulations. Furthermore, with a modification of the image processing technique, it is possible to apply the RIMS method to the study of other sorts of granular flows, such as those with bidisperse grains, or even those with detectable orientations.