Stationary states of activity-driven harmonic chains

TL;DR

This paper investigates the stationary states of a harmonic chain driven by active reservoirs with different active force dynamics, revealing Gaussian velocity fluctuations, energy equipartition, and boundary-specific non-Gaussian features.

Contribution

It demonstrates that diverse active driving mechanisms lead to universal Gaussian velocity fluctuations and energy equipartition in the bulk of the chain.

Findings

Velocity fluctuations are Gaussian with uniform kinetic temperature.

Bulk energy equipartition occurs in the thermodynamic limit.

Boundary effects cause non-Gaussian velocity and current distributions.

Abstract

We study the stationary state of a chain of harmonic oscillators driven by two active reservoirs at the two ends. These reservoirs exert correlated stochastic forces on the boundary oscillators which eventually leads to a nonequilibrium stationary state of the system. We consider three most well-known dynamics for the active force, namely, active Ornstein-Uhlenbeck process, run-and-tumble process and active Brownian process, all of which have exponentially decaying two-point temporal correlations but very different higher order fluctuations. We show that irrespective of the specific dynamics of the drive, the stationary velocity fluctuations are Gaussian in nature with a kinetic temperature which remains uniform in the bulk. Moreover, we find the emergence of an `equipartition of energy' in the bulk of the system -- the bulk kinetic temperature equals the bulk potential temperature in the thermodynamic limit. We also calculate the stationary distribution of the instantaneous energy current in the bulk which always shows a logarithmic divergence near the origin and asymmetric exponential tails. The signatures of specific active driving become visible in the behavior of the oscillators near the boundary. This is most prominent for the RTP and ABP driven chains where the boundary velocity distributions become non-Gaussian and current distribution has a finite cutoff.