Density-Gradient Control over Nanoparticle Supercrystal Formation

With the advent of DNA-directed methods to form "single crystal" nanoparticle superlattices, new opportunities for studying the properties of such structures across many length scales now exist. These structure-property relationships rely on the ability of one to deliberately use DNA to control crystal symmetry, lattice parameter, and microscale crystal habit. Although DNA-programmed colloidal crystals consistently form thermodynamically favored crystal habits with a well-defined symmetry and lattice parameter based upon well-established design rules, the sizes of such crystals often vary substantially. For many applications, especially those pertaining to optics, each crystal can represent a single device, and therefore size variability can significantly reduce their scope of use. Consequently, we developed a new method based upon the density difference between two layers of solvents to control nanoparticle superlattice formation and growth. In a top aqueous layer, the assembling particles form a less viscous and less dense state, but once the particles assemble into well-defined rhombic dodecahedral superlattices of a critical size, they sediment into a higher density and higher viscosity sublayer that does not contain particles (aqueous polysaccharide), thereby arresting growth. As a proof-of-concept, this method was used to prepare a uniform batch of Au nanoparticle (20.0 ± 1.6 nm in diameter) superlattices in the form of 0.95 ± 0.20 μm edge length rhombic dodecahedra with body-centered cubic crystal symmetries and a 49 nm lattice parameter (cf. 1.04 ± 0.38 μm without the sublayer). This approach to controlling and arresting superlattice growth yields structures with a 3-fold enhancement in the polydispersity index.