(62c) Inverse Design of Self-Assembled Materials

Technologies ranging from clean energy devices to plant-based meat and printable organs would be transformed by the rapid design of complex functional materials. Towards the goal of readily designing functional materials, researchers have synthesized a broad range of interacting components, including de novo proteins and patchy colloidal particles. These advances have the potential to transform materials engineering. However, exhaustively searching the space of all combinations of these complex components to find desired functions would require billions of years. Instead, I leverage automatic differentiation, the tool underlying much of the dramatic success in machine learning and non-convex optimization, to search the vast design space of functional materials. I demonstrate quantitative control over non-equilibrium functions in self-assembly, including tunable transition rates and on-demand disassembly. As an initial model for potential drug delivery applications, I design a reaction that induces the release of a small particle trapped in a virus-like colloidal shell.