Multiscale Self-Assembly of Quantum Dots into an Anisotropic Three-Dimensional Random Network

One of the well-known challenges in design of nanomaterials is to simultaneously achieve material properties pertaining to few-atom scale and bulk properties through which the material connects to other materials or interacts with devices. This is difficult because properties arising from physics at different scales are often mutually exclusive. An important example is the 30 year-old problem of realizing a connected-but-confined Si nanostructure embedded in a dielectric matrix (e.g., SiO₂) that simultaneously brings together quantum-dot (QD)-like optical properties and good electrical conduction. Here, we solve this problem through creation of a hierarchically self-assembled anisotropic random network of Si QDs: At the atomic scale, QDs are formed, which sparsely interconnect without inflating their diameters to form an isotropic random network, and larger scales, this network becomes anisotropic, preferentially growing in the vertical direction to form nanowire-like structures. We report simultaneous achievement of good electrical conductivity (~0.1 S/cm) and a bandgap tuneable over the visible light range (from 1.8 to 2.7 eV).
In order to determine how to self-assemble such a topology without using advanced control over dynamical details of the system, we developed a toy model of the stochastic deposition process, from which we related the intended topology to parameters governing stochastic growth and determined the experimental conditions that can give rise to it. Monte Carlo and Molecular Dynamics simulations are performed to guide our methodology and fabrication was done using magnetron sputter deposition. The two leverages that we used for multiscale self-assembly were as follows: (i) We keep the substrate “cold” and adjust how “hot” the deposited particles are. This way, we create spatio-temporal temperature gradients on the surface and thereby, we control surface diffusion and promote vertical growth in the microscale resembling nanowires. (ii) We fine-tune the thin-film stoichiometry in order to control the phase-separation. This way, we control the nominally unstable medium that QDs are embedded in and limit further inflation of their diameters in the atomic scale. This way we show that self-assembly under nonequilibrium conditions and nonlinear dynamics sweeps aside a large number of factors that influence the details of thin-film growth, but provides a couple of simple “rules” (with clearly identifiable corresponding experimental conditions) to determine the final morphology. The generality and material-independence of this methodology is strongly suggestive of possibility to apply it to solve a variety of other nanomaterial problems, which also pertain to multiple scales.