Local signatures of emerging global order in complex crystal growth

Designing and building functional materials requires an understanding of how different crystal structures grow—depending on their complexity, local motifs, and symmetry. Currently, the manner in which particle-particle interactions lead to the emergent formation of any given crystal structure remains a mystery. Using molecular dynamics simulations of particles interacting via isotropic pair potentials, we assemble a diverse set of complex crystal structures [1] and observe the crystallization process. With a machine-learning-powered order parameter [2], we classify particles by their local environments into different phases and crystalline positions. We examine the progression of crystal growth on a local level to illuminate the emergence of long-range order from short-range interactions.

[1] J. Dshemuchadse, P. F. Damasceno, C. L. Phillips, M. Engel, S. C. Glotzer, in preparation (2020).
[2] M. Spellings, S. C. Glotzer, AIChE Journal 64 (6), 2198–2206 (2018).