The classic Stokes-Einstein relation ($D=k_BT/(6\pi\eta r)$) it is assumed that the particle undergoing Brownian motion is spherical, and experiences ordinary Stokes flow between (or during) consecutive impulses. In the nonspherical case, however, translational motion (driven by an external field for example) is typically correlated with chiral motion; consider, for example, the helical motion of a maple seed pod that develops under gravity. This suggests that there should be some correlation between angular and translational Brownian increments, unless angular increments are expected to be considerably larger or smaller than translational ones. This also suggests, as angular diffusion can likely arise also through anisotropic forces exerted on the boundary of an asymmetric particle, that angular diffusion (or the ratio of angular to translational diffusivities) is non-universal and depends on the shapes of particles and interfacial properties of constituent materials. Is this in fact the case?
