Packing in the Spheres

Original entry: William Bonificio, AP 225, Fall 2009

Reference

Packing in the Spheres David A. Weitz, Science , vol 303 968 (2004)

Keywords

close, packing, spheres, oblate, ellipsoid

Summary

The paper discusses how nonspeherical objects pack compared to spherical ones. Surprisingly, ellipsoidal objects take less energy to reach a close packed structure. Some reasons for this unintuitive result are discussed.

Excluded volume (shown in yellow) between particles of different geometries. Purple particles are micelles (not to scale)

Soft Matter Example

Colloidal physics often deals with the problem of packing. For the most part, scientists will approximate all object that are being packed as spheres and then calculate the packing efficiency with that basis. One such type of packing that can occur is when the spheres still have a completely random orientation, but are packed as tightly as they can be, this is called random close packing, or RCP. For spheres, the volume fraction for RCP is 0.64.

If we allow the spheres to be ordered, and pack them as tightly as possible they take a configuration called hexagonal close packed, or HCP. In this configuration the spheres form hexagonal layers, and each layer above will fall into the crevices formed by the layer below. The volume fraction for spheres in HCP is 0.74.

The final kind of packing discussed is random loose packing, RLP. In this packing the spheres are in a random orientation, but have been packed in a fluid that proved buoyancy. One can imagine dumping marbles in an empty bucket (RCP) versus dumping marbles in a bucket full of water (RLP). The volume fraction for spheres in RLP is 0.56.

What are the packing values for shapes that are not spheres, such as monodisperse ellipsoids? It appears that the volume fraction of RCP for ellipsoids is generally 0.74 and decreases as the ellipsoid approaches the spherical shape. This is a surprising result that suggests that spheres are perhaps an anomaly with a very low RCP Volume fraction.

The reason for the anomaly suggested by the paper is that spheres in RCP have several nearest neighbors that are all jammed into place. The sphere can only be translated by the forces from these neighbors, not rotated. However, ellipsoids can experience a torque along with a translation. The result of this is several more nearest neighbors needed to balance the forces, and much closer packing.