Amy E. Larsen and David G. Grier
December 2, 1996
Sub-micron diameter latex spheres colloidally suspended in water can form into regular arrays known as colloidal crystals. Unlike most conventional solids, colloidal crystals can be forced into a metastable superheated state. Analysis of the structure and dynamics of these metastable crystals using digital video microscopy provides clear evidence for a surprisingly strong and long-ranged attraction between similarly charged spheres. Such unexpected and unexplained attractions may influence the properties of natural and industrial suspensions.
Charge-stabilized colloidal suspensions of extremely uniformly-sized polymer spheres were developed in the 1950's as media for paints and other surface coatings. These suspensions have attracted considerable attention in recent years for their utility as model systems with which to study the mechanisms of structural phase transitions . Depending on their concentration and chemical environment, colloidal spheres can form up into regular crystalline arrays or devolve into fluid disorder. Transformations between the ordered and disordered states are phase transitions analogous to melting and freezing of atomic matter. Unlike atoms, however, colloidal spheres can be tracked with a conventional light microscope. Their structural transformations thus provide unparalleled opportunities to investigate the microscopic mechanisms of phase transitions.
This article describes, conversely, how the study of phase transitions sheds new light on the fundamental properties of colloidal suspensions. In particular, we find evidence in an exotic colloidal melting transition for long-range attractive interactions between like-charged microspheres. When combined with direct measurements of the spheres' pairwise interaction potential, these observations strongly suggest that such attractive interactions may be responsible for a variety of other unexplained phenomena observed over the past decade in bulk colloidal suspensions.
The most striking of these anomalies include large stable voids in otherwise homogeneous suspensions [2, 3] and equilibrium phase separation between colloidal fluids of different densities [4, 5]. Neither should be possible in a system with purely repulsive interactions, although explanations based on impurity effects  have proved difficult to exclude. Persistent quantitative discrepancies between predicted [7, 8, 9, 10, 11] and observed [12, 13] melting points of colloidal crystals are no less disturbing, but might reflect relatively minor shortcomings in the theory. The long-ranged attractions we observe would account naturally for these effects, but are qualitatively inconsistent with the long-accepted theory for colloidal interactions. We suggest, therefore, that the theory for colloidal interactions requires substantial revision.