Engineering Entropy for the Inverse Design of Colloidal Crystals from Hard Shapes

TL;DR

This paper presents a method to engineer entropy for the inverse design of colloidal crystals by optimizing particle shapes to self-assemble into desired structures, including novel ones, through extensive sampling of high-dimensional shape spaces.

Contribution

It introduces an entropy-based inverse design approach that efficiently explores high-dimensional shape spaces to discover thermodynamically optimal colloidal particles.

Findings

Successfully designed particles that assemble into six target crystals.

Discovered new crystal structures with no known atomic equivalents.

Optimized thermodynamic stability of assembled colloidal crystals.

Abstract

Throughout the physical sciences, entropy stands out as a pivotal but enigmatic concept that, in materials design, often takes a backseat to energy. Here, we demonstrate how to precisely engineer entropy to achieve desired colloidal crystals. We demonstrate the inverse design of hard particles that assemble six different target colloidal crystals due solely to entropy maximization. Our approach efficiently samples $10^8$ particle shapes from 88- and 192-dimensional design spaces to discover thermodynamically optimal shapes. We design particle shapes that self assemble known crystals with optimized thermodynamic stability, as well as new crystal structures with no known atomic or other equivalent.