Abstract

Suppose we want to create a material with a certain unusual property. One strategy is to start with a model of an existing material without that property, and bias its dynamics to sample unlikely trajectories for which the atypical property is pres ent. Looking at the biased trajectories, it may be possible to spot some choice of local interactions that would achieve the required effect. I will describe an instance of this in the realm of self-propelled spherical colloids. Here, biasing the ensemble to reduce colloidal collisions creates states in which the propulsion directions have polar order: accordingly, collisions can be reduced by introducing polar interactions. While this particular outcome is relatively obvious, the method is generalizable in principle to more complex cases where genuinely new design principles might emerge.

Coauthors: Takahiro Nemoto, Étienne Fodor, Robert L. Jack, Julien Tailleur

Reference: Optimizing active work: dynamical phase transitions, collective motion and jamming. T. Nemoto et al, arXiv 1805.02887