Received October 27, 2005

In Final Form: December 19, 2005

Abstract:

DNA is a powerful and versatile tool for nanoscale self-assembly. Several researchers have assembled nanoparticles and colloids into a variety of structures using the sequence-specific binding properties of DNA. Until recently, however, all of the reported structures were disordered, even in systems where ordered colloidal crystals might be expected. We detail the experimental approach and surface preparation that we used to form the first DNA-mediated colloidal crystals, using 1 m diameter polystyrene particles. Control experiments based on the depletion interaction clearly indicate that two standard methods for grafting biomolecules to colloidal particles (biotin/avidin and water-soluble carbodiimide) do not lead to ordered structures, even when blockers are employed that yield nominally stable, reversibly aggregating dispersions. In contrast, a swelling/deswelling-based method with poly(ethylene glycol) spacers resulted in particles that readily formed ordered crystals. The sequence specificity of the interaction is demonstrated by the crystal excluding particles bearing a noninteracting sequence. The temperature dependence of gelation and crystallization agree well with a simple thermodynamic model and a more detailed model of the effective colloidal pair interaction potential. We hypothesize that the surfaces yielded by the first two chemistries somehow hinder the particle-particle rolling required for annealing ordered structures, while at the same time not inducing a significant mean-force interaction that would alter the self-assembly phase diagram. Finally, we observe that particle crystallization kinetics become faster as the grafted-DNA density is increased, consistent with the particle-particle binding process being reaction, rather than diffusion limited.

Download the full text: PDF | HTML