Clocks, engines, and information erasure in quantum dot devices

In this talk, I will explore how fluctuations can be harnessed to drive useful functionality in nanoscale electronic systems. I will begin by describing the realisation of a nanoscale clock powered by single-electron tunnelling, where we find that the clock's overall dissipation is dominated by the thermodynamic cost of measuring its ticks. I will then discuss how nonequilibrium fluctuations can be used to operate nanoscale engines and refrigerators. In particular, I will present a quantum-dot Szilard engine that extracts work efficiently under fast driving, and show how optimised protocols reveal a fundamental trade-off between power output and power fluctuations. I will also introduce an extension to Landauer's bound on the energy required to erase information in a quantum-dot device. I will finish by discussing how nanoscale systems provide a unique platform for probing the fundamental thermodynamic costs of information processing and learning.