Multiobjective Optimization for Targeted Self-Assembly among Competing Polymorphs

Abstract

Most approaches for designing self-assembled materials focus on the thermodynamic stability of a target structure or crystal polymorph. Yet in practice, the outcome of a self-assembly process is often controlled by kinetic pathways. Here we present an efficient machine-learning-guided design algorithm to identify globally optimal interaction potentials that maximize both the thermodynamic yield and kinetic accessibility of a target polymorph. We show that optimal potentials exist along a Pareto front, indicating the possibility of a trade-off between the thermodynamic and kinetic objectives. Although the extent of this trade-off depends on the target polymorph and the assembly conditions, we generically find that the trade-off arises from a competition among alternative polymorphs: The most kinetically optimal potentials, which favor the target polymorph on short timescales, tend to stabilize a competing polymorph at longer times. Our work establishes a general-purpose approach for multiobjective self-assembly optimization, reveals fundamental trade-offs between crystallization speed and defect formation in the presence of competing polymorphs, and suggests guiding principles for materials design algorithms that optimize for kinetic accessibility. Published by the American Physical Society 2025