Brownian dynamics study of the self-assembly of ligated gold nanoparticles and other colloidal systems

Abstract

We carry out Brownian Dynamics Simulations to study the self-assembly of ligated gold nanoparticles for various ligand chain lengths. First, we develop a phenomenological model for an effective nanoparticle-nanoparticle pair potential by treating the ligands as flexible polymer chains. Besides van der Waals interactions, we incorporate both the free energy of mixing and elastic contributions from compression of the ligands in our effective pair potentials. The separation of the nanoparticles at the potential minimum compares well with experimental results of gold nanoparticle superlattice constants for various ligand lengths. Next, we use the calculated pair potentials as input to Brownian dynamics simulations for studying the formation of nanoparticle assembly in three dimensions. For dodecanethiol ligated nanoparticles in toluene, our model gives a relatively shallower well depth and the clusters formed after a temperature quench are compact in morphology. Simulation results for the kinetics of cluster growth in this case are compared with phase separations in binary mixtures. For decanethiol ligated nanoparticles, the model well depth is found to be deeper, and simulations show hybrid, fractal-like morphology for the clusters. Cluster morphology in this case shows a compact structure at short length scales and a fractal structure at large length scales. Growth kinetics for this deeper potential depth is compared with the diffusion-limited cluster-cluster aggregation (DLCA) model. We also did simulation studies of nanoparticle supercluster (NPSC) nucleation from a temperature quenched system. Induction periods are observed with times that yield a reasonable supercluster interfacial tension via classical nucleation theory (CNT). However, only the largest pre-nucleating clusters are dense and the cluster size can occasionally range greater than the critical size in the pre-nucleation regime until a cluster with low enough energy occurs, then nucleation ensues. Late in the nucleation process the clusters display a crystalline structure that is a random mix of fcc and hcp lattices and indistinguishable from a randomized icosahedra structure. Next, we present results from detailed three-dimensional Brownian dynamics simulations of the self-assembly process in quenched short-range attractive colloids. Clusters obtained in the simulations range from dense faceted crystals to fractal aggregates which show ramified morphology on large length scales but close-packed crystalline morphology on short length scales. For low volume fractions of the colloids, the morphology and crystal structure of a nucleating cluster are studied at various times after the quench. As the volume fraction of the colloids is increased, growth of clusters is controlled by cluster diffusion and cluster-cluster interactions. For shallower quenches and low volume fractions, clusters are compact and the growth-law exponent agrees well with Binder–Stauffer predictions and with recent experimental results. As the volume fraction is increased, clusters do not completely coalesce when they meet each other and the kinetics crosses over to diffusion-limited cluster-cluster aggregation (DLCA) limit. For deeper quenches, clusters are fractals even at low volume fractions and the growth kinetics asymptotically reaches the irreversible DLCA case.