Can colloidal swarms settle faster than isolated particles? (LMC)
Abstract
Usually, the more concentrated a particle dispersion is, the slower it settles in natural gravity or in a centrifuge. Yet, we have recently shown that this must not perforce be true: in fact, the sedimentation velocity $v(\phi)$ of a suspension of particles interacting via strong attractive forces depends non-monotonically on particle volume fraction. While at high particle concentration $v(\phi)$ does decrease with $\phi$, like in the case of hard spheres, at sufficiently low $\phi$ the suspension settles even faster than a single particle. This evidence, obtained for a system where depletion interactions can be finely tuned and quantified, is favorably compared to recent numerical results suggesting the occurrence of such a “promoted” sedimentation regime, which utterly contrasts with standard “hindered” settling. Our results, however, also highlight an important and previously unnoticed consequence of promoted settling on the kinetics of sedimentation. In fact, the settling front, which for hindered settling takes on a time-invariant, shock-wave profile, conversely spreads with time, becoming liable to thermal instabilities that may lead to “stratification” of the profile into distinct concentration bands. Promoted settling may actually be much more pronounced for systems of “patchy” particles interacting via short range anisotropic attractions that do not lead to phase separation, but rather to finite clustering effects. In fact, we suggest that our results may be specifically relevant for ultracentrifuge investigations of protein association effects, where promoted settling is sometimes observed and usually interpreted using semiphenomenological “clustering models”. Indeed, by investigating the specific case of $\beta$-lactoglobulin A (BLGA), a globular protein displaying strong attractive interactions in a narrow temperature and pH range, we show that the structural and dynamic information provided by light scattering measurements performed at equilibrium can be exploited to quantitatively predict the settling kinetics actually observed in an ultracentrifuge, without resorting to any detailed “chemical association” models.
