Memory erasure in small systems

TL;DR

This paper investigates heat dissipation during memory erasure in nanoscale systems, revealing that less heat than Landauer's bound can be dissipated, indicating the need to generalize Landauer's principle at small scales.

Contribution

It demonstrates that at the nanoscale, heat fluctuations allow for erasure with less heat than the Landauer bound, challenging classical thermodynamic limits.

Findings

Full erasure can occur with less heat than Landauer's bound.

Heat fluctuations are significant at the nanoscale.

Proposes an experiment to verify the theoretical predictions.

Abstract

We consider an overdamped nanoparticle in a driven double-well potential as a generic model of an erasable one-bit memory. We study in detail the statistics of the heat dissipated during an erasure process and show that full erasure may be achieved by dissipating less heat than the Landauer bound. We quantify the occurrence of such events and propose a single-particle experiment to verify our predictions. Our results show that Landauer's principle has to be generalized at the nanoscale to accommodate heat fluctuations.