Digital alchemy for the inverse design of patchy particles

Patchy particles are colloidal building blocks whose shape and interaction anisotropy offer a strategic path to the self-assembly of complex structures with desirable properties. The many possible anisotropy dimensions of patchy particles leads to an enormous design space, making inverse design a crucial tool for the field. Of the inverse design methods available, digital alchemy (DA) is particularly attractive for patchy particles. DA is a generalized ensemble method that allows inverse design by treating particle attributes (e.g., shape and patchiness) as thermodynamic variables thereby allowing in situ updates to them. In this work, we extend the DA framework to the inverse design of triblock Janus patchy particles that self-assemble target crystal structures. We design symmetric triblock Janus particles that self-assemble the open 2D kagome and 3D pyrochlore lattices. To highlight the generality of our method, we also design asymmetric triblock Janus particles to self-assemble a snub square lattice. We find that particles designed based upon local valence considerations fail to assemble the target structure, whereas particles designed via DA succeed, highlighting the ability of the method to find nontrivial solutions to a deceptively complex design problem.