(9h) Controlling Active Brownian Motion By State-Dependent Rotational Diffusion

The motion of Active Brownian particles (ABPs) is determined by a coupling between stochastic and deterministic forces. At short times, deterministic forces propel ABPs in a given direction. Yet, over time, stochastic forces randomize the motion by inducing the ABPs, and thus the direction of their motion, to rotate. As a result, active Brownian motion is ballistic at short times and diffusive over long times. The critical timescale at which the motion becomes diffusive, that is the relaxation time Ï„_R, goes as the inverse of the rotational diffusion coefficient (1/D_R).

In this talk, we show that controlling D_R over time and space brings about a new type of anomalous diffusion that opens up new research avenues in both fundamental and applied directions. We do so by means of experiments and Langevin dynamics simulations. In particular, we report on the first experimental realization of Janus colloidal particles exhibiting a state-dependent D_R (i.e., space- and time-dependent). Experiments and simulations show that when ABPs travel through alternating regions of high and low D_R, the distribution of displacement probability can exhibit a markedly non-Gaussian behavior. We find that the departure from Gaussian statistics is maximized when the relaxation times in the different regions are well separated and for high values of the ratio L/v, where L is the region size and v is the ballistic velocity.

If the ABPs experience a space-dependent D_Rwith a certain sensorial delay, then they also tend to segregate and accumulate in regions of high D_R. We studied the case in which D_Rchanges periodically, according to the ABP position, with a period equal to the sensorial delay, during which D_R remains constant. For vanishingly small delays the homogeneous distribution is retrieved, whereas increasingly longer delays result in a higher degree of persistent segregation, which eventually reaches a maximum at delays on the order of L/v. Interestingly, the extent of maximum segregation admits a universal scaling function that also depends on L/v. For delays >> L/v the dynamics becomes even more nuanced as the degree of segregation does not attain a constant value in the limit of long times and instead oscillates with a period equal to the delay.