(169ag) Enhanced Shockwave Synthesis through Accurate Silicon Modeling Using Chimes

Silicon nanoparticles (SiNP) and composites like silicon carbide and silicon nitride are of technological interest in the semiconductor domain. However, a precise yet high throughput synthesis capability is needed to enable integration of SiNPs into everyday technological devices. Existing synthesis techniques are inhibited by either low throughput or limited material quality. Shockwave synthesis techniques, such as explosive deposition, can achieve both the production rate and quality required, but the extreme conditions reached, which can reach 7000K and 90 GPa have thus far precluded establishing a clear picture of the underlying mechanisms and kinetics needed to enable finely tunable SiNP synthesis through experiments alone. Atomistic simulations offer a complimentary route to studying this process but rely on availability of “quantum-accurate” interatomic models capable of faithfully describing the complicated evolution that ensues under extreme conditions. Therefore, in this presentation, we discuss recent efforts to develop a machine learned interatomic potential (ML-IAP) for silicon model using ChIMES, a physics-informed framework ML-IAP designed specifically to enable large scale yet quantum-accurate simulation of materials under extreme conditions. Ultimately, this model can serve as a foundation upon which models for other silicon-containing materials (e.g., Si-C) can be generated, enabling simulation-guided refinement of high pressure (e.g. shock-) synthesis strategies for Si-based nanomaterials.