Arunkumar Bupathy (WPI-SKCM²): Designing Competing Structures for Selective Self-Assembly and Machine Learning Techniques for Coarse-Grained Many-Body Potentials

Multi-component self-assembly mixtures offer the ability to encode multiple target structures from the same set of interacting components. Earlier studies have focussed on selective retrieval via changes to the composition of the mixture, through seeding procedures or by strengthening specific interactions. This requires preparing the system in an initial state that favours the formation of the desired target structure, which may not be always possible, especially in closed systems. In this work, we show the design of a multi-component self-assembly mixture that can form one of two pre-defined structures, with high selectivity, through simple temperature protocols alone [1]. Our simulations show that to avoid spurious aggregation, the components should have different bonding neighbours in the two target structures, and that the component library itself should be maximally shared by the two structures to improve selectivity of retrieval. We demonstrate one possible way in which selectivity can be improved through secondary aggregation products which we term ”vestigial aggregates”.

In the remainder of the talk, I will introduce machine learning techniques for developing coarse-grained descriptions of complex many-body interaction potentials [2]. These methods can not only speed up numerical simulations but also help understand the nature of the interactions and to predict phases, such as those of knotted fields in chiral liquids and magnets.

References:

[1] Arunkumar Bupathy, Daan Frenkel and Srikanth Sastry. Temperature protocols to guide selective self-assembly of competing structures. Proceedings of the National Academy of Sciences 119, e2119315119 (2022).

[2] Giuliana Giunta, Gerardo Campos-Villalobos and Marjolein Dijkstra. Coarse-grained many-body potentials of ligand-stabilized nanoparticles from machine-learned mean forces. ACS Nano 2023, 17, 23391-23404.