Speaker
Description
Particles and discrete objects on the scale of nanometres to micrometres, such as colloids, DNA bricks, proteins, transistors, etc, may be used as building blocks for new materials, with a variety of applications in optics, drug delivery, energy harvesting, and nanorobotics. A goal for theory and simulation is to build algorithms and design principles to find the building blocks that assemble into a structure of interest. This is a challenge for many reasons, but partly because structures that are designed to be the most thermodynamically stable, are often kinetically inaccessible. To address this challenge, we take inspiration from biology, and observe that many complex biological systems assemble hierarchically — they first build subunits, that later assemble in larger entities, in a process that can repeat over several generations. I will show that we can emulate this process in so-called “addressable” self-assembly, where each building block is distinct and has a specific location in the target structure. By choosing the interactions strengths to decrease appropriately with the scale of units considered, we can create objects that similarly assemble hierarchically, over at least five generations of hierarchy. We obtain close to perfect yield at fully equilibrium conditions, unlike previous methods which required temperature cycling. Then, I will show that we can enhance the yield even further by tuning the interactions using differentiable simulations, which allows us to unravel another design principle linked to a combinatorial property of graphs.
| Keyword-1 | fluid dynamics |
|---|---|
| Keyword-2 | computation |
