Abstract:Modern experimental methods enable the creation of self-assembly building blocks with tunable interactions, but optimally exploiting this tunability for the self-assembly of desired structures remains an important challenge. Many studies of this inverse problem start with the so-called fully-addressable limit, where every particle in a target structure is different. This leads to clear design principles that often result in high assembly yield, but it is not a scaleable approach -- at some point, one must grapple with ``reusing'' building blocks, which lowers the degree of addressability and may cause a multitude of off-target structures to form. While the existence of off-target structures does not necessarily preclude high yield assembly of the target structure(s), it massively complicates the design process because it is often unclear if or how these off-target structures can be controlled. Here, we solve a key obstacle preventing robust inverse design in the ``semi-addressable regime'' by developing a highly efficient algorithm that enumerates all structures that can be formed from a given set of building blocks. By combining this with established partition-function-based yield calculations, we successfully inverse design for target structures while minimizing the required number of distinct building block species. Furthermore, we show that it is almost always possible to find semi-addressable designs where the entropic gain from reusing building blocks outweighs the presence of off-target structures and even increases the yield of the target. Thus, not only does our enumeration algorithm enable robust and scalable inverse design in the semi-addressable regime, our results demonstrate that it is possible to operate in this regime while maintaining the level of control often associated with full addressability.

Algorithmic design of self-assembling structures

Inverse Design for Self Assembly via On-the-Fly Optimization

Self-assembling kinetics: Accessing a new design space via differentiable statistical-physics models

Inverse optimization techniques for targeted self-assembly

Designing athermal disordered solids with automatic differentiation

Inverse design of two-dimensional structure by self-assembly of patchy particles

Thermodynamic stability versus Kinetic Accessibility: Pareto Fronts for Programmable Self-Assembly

Designing self-assembling kinetics with differentiable statistical physics models

The Role of Pressure in Inverse Design for Assembly

Design principles for non-equilibrium self-assembly

Variational design principles for nonequilibrium colloidal assembly

Accessing Semi-Addressable Self Assembly with Efficient Structure Enumeration

Designing the self-assembly of arbitrary shapes using minimal complexity building blocks

Designing patchy interactions to self-assemble arbitrary structures

Fast Generation of Potentials for Self-Assembly of Particles

A Bottom-Up Design Strategy for Controllable Self-Assembly Based on the Isotropic Double-Well Potential

Discovering optimal kinetic pathways for self-assembly using automatic differentiation

Breadth versus depth: Interactions that stabilize particle assemblies to changes in density or temperature

Dimensionality and design of isotropic interactions that stabilize honeycomb, square, simple cubic, and diamond lattices

Multi-objective optimization for targeted self-assembly among competing polymorphs