The discrepancy in nucleation rate densities between simulated and experimental hard spheres remains staggering and unexplained. Suggestively, more strongly sedimenting colloidal suspensions of hard spheres nucleate much faster than weakly sedimenting systems. In this work, we consider first the effect of sedimentation on the structure of colloidal hard spheres by tuning the density mismatch between solvent and colloidal particles. In particular, we investigate the effect on the degree of the fivefold symmetry present. Second, we study the size of density fluctuations in these experimental systems in comparison to simulations. The density fluctuations are measured by assigning each particle a local density, which is related to the number of particles within a distance of 3.25 particle diameters. The standard deviation of these local densities gives an indication of the fluctuations present in the system. Fivefold symmetry is suppressed by a factor of two when sedimentation is induced in our system. Density fluctuations are increased by a factor of two in experiments compared to simulations. The change in fivefold symmetry makes a difference to the expected nucleation rates, but we demonstrate that it is ultimately too small to resolve the discrepancy between experiment and simulation, while the fluctuations are shown to be an artefact of 3d particle tracking.
Polydispersity shifts the Hard Sphere coexistence region to higher packing fractions. Experiments need to take this shift into account in order to report accurate volume fractions. This is done implicitly in the strongly sedimenting data22–24 as they measure the coexistence region for their polydisperse systems and then rescale their measured volume fractions to the freezing line of monodisperse Hard Spheres. This leads to a collapse of the nucleation rates at a given distance from the freezing line.14 After this procedure, experiments and simulations still disagree,35 suggesting that further mechanism is present.