Self-Consistent Noise Computation in Quantum Dot Systems

The power, efficiency, and reliability of all classical heat engines is fundamentally constrained by thermodynamic uncertainty relations (TURs), which however can be broken in the quantum regime. Motivated by this, we explore current flucuations (noise) in realistic models of strongly interacting non-equilibrium quantum dot systems. Surprisingly, we find that even at high temperatures -- where noise and current are in a perturbative regime -- their ratio (Fano factor) can be sensitive to higher order corrections, which necessitates a treatment beyond bare perturbation theory for quantitative predictions.

To address this, we develop a renormalizataion group scheme based on the Nakajima-Zwanzig memory kernel and additionally apply Fermi liquid theory, which allows us to compute noise even at low temperatures. We use this to compute the transition of the noise between the cotunneling and Kondo regime and analyze above mentioned TURs.