Metastability of Superheated Colloidal Crystals

Figure 2: Video microscope view of the (110) face of a superheated metastable FCC colloidal crystal in contact with a low density fluid. Superimposed traces indicate the trajectories of selected spheres in 1/30 sec intervals. These trajectories clearly distinguish well-localized spheres in the crystal from freely diffusing spheres in the fluid. Some spheres, such as the one indicated with the arrow, collide with and temporarily attach to the crystal-fluid interface.

Suspensions of purely repulsive monodisperse spheres with pairwise additive interactions are believed to exist in three equilibrium phases: fluid, face-centered cubic (FCC) crystal or body-centered cubic (BCC) crystal. Experiments on charge-stabilized colloidal suspensions have revealed other states, however, including equilibrium liquid-vapor phase separation [25,26], reentrant solid-liquid transitions [27], stable void structures consistent with solid-vapor and liquid-vapor phase separation [26], and metastable superheated crystals [14,28]. These additional states are most easily explained if the pair potential includes an attractive component [19,20,25,26,27,29] or if the system develops a many-body cohesion [30].

Metastable crystals' structure and dynamics, in particular, reveal long-ranged attractions consistent in range and magnitude to the measured wall-induced attraction [14]. These crystals, an example of which is shown in figure 2, are made by compressing a low density suspension of spheres against glass walls through an electrohydrodynamic instability [28]. The crystals should melt in a matter of seconds once the compressing field is turned off, at a rate limited only by diffusion. Instead, some crystalline regions persist for as long as an hour, their facets and interfacial fluctuations attesting to a large stabilizing latent heat.

These observations might seem only to confirm that nearby walls induce attractions among pairs of spheres since the crystals are formed against glass surfaces. These are three-dimensional crystals, however, and extend away from confining walls into the bulk of the suspension. Indeed, their surfaces are so far from the nearest walls that pair interaction measurements at comparable separations reveal no attraction [14]. Thus, the wall-induced attraction cannot be responsible for the crystals' three-dimensional structure. This is strong circumstantial evidence that the spheres themselves can engender a many-body cohesion whose range and strength are comparable to the wall-induced pair attraction.

The suggestion that bulk colloidal suspensions can experience many-body cohesions even in the absence of nearby bounding walls is consistent with observations of other anomalous phase transformations in bulk colloidal suspensions. It should not be viewed as evidence for pairwise attractions, however, since the direct interaction measurements described in Section 2 largely rule out this interpretation.

In this sense, observations on metastable superheated colloidal crystals serve as a bridge between pair interaction measurements and measurements of bulk phase behavior. They suggest that a similar mechanism may be responsible for the anomalous behavior in both classes of experiments. If this is the case, then the failure so far to measure long-ranged attractions in isolated pairs of spheres suggests that all of the anomalous behavior observed in bulk suspensions arises from the breakdown of pairwise additivity in macroionic interactions. The interesting thing about this failure is that it is not subtle - it leads to qualitatively new behavior when compared with mean-field theory's predictions. It is evident even in systems for which the DLVO theory has long been assumed to be adequate.