arXiv:2609.27065
Thermodynamic Uncertainty Relations in Chaotic Andreev Billiards
I. R. A. C. Lucena, T. J. A. Mori, M. M. Soares, D. Bazeia, A. R. Rocha
Abstract
We investigate how particle-hole symmetry, quantum interference, and tunnel barriers shape thermodynamic uncertainty relations in chaotic Andreev billiards. Using random-matrix theory and the Mahaux-Weidenmüller scattering approach, we study charge conductance and shot noise across the four Altland-Zirnbauer symmetry classes, from the single-channel extreme quantum limit to the multichannel semiclassical regime and from ideal to opaque contacts. We characterize thermodynamic precision through two complementary ensemble observables: either by averaging the sample-resolved noise-to-conductance ratio, or from the separately averaging noise and conductance. Their pronounced discrepancy in the extreme quantum regime reveals the non-self-averaging character of mesoscopic transport and persists over a broad range of barrier transparencies. For ideal contacts, the first provides a sensitive fingerprint of the Altland-Zirnbauer symmetry class. In the opaque regime, this hierarchy changes. Despite these strong symmetry- and barrier-dependent effects, the standard thermodynamic uncertainty relation remains satisfied throughout all regimes investigated. Our results establish thermodynamic uncertainty as a symmetry-sensitive probe of universal transport in chaotic normal-superconducting systems.
AI Summary
The paper asks how particle–hole symmetry, time-reversal and spin-rotation symmetries, and contact transparency affect thermodynamic uncertainty relations (TURs) in chaotic Andreev billiards. It models a coherent cavity coupled to normal leads and superconductors with phase difference $\pi$, using random-matrix Hamiltonians and the Mahaux–Weidenmüller scattering construction across Altland–Zirnbauer classes D, DIII, C and CI. Conductance and zero-temperature shot noise are computed from $10^6$ realizations per class, comparing single-channel and multichannel cases, $N=1$ and $N=10$. The TUR analysis distinguishes averaging sample-wise noise-to-conductance ratios, $\mathcal Q_{\mathrm R}$, from taking the ratio of ensemble-averaged noise and conductance, $\mathcal Q_{\mathrm{MS}}$. These differ substantially at $N=1$, with discrepancies persisting across broad transparency ranges. Both satisfy the conventional bound $\mathcal Q\geq2$ in the investigated regimes. The detailed TUR comparison focuses on time-reversal-preserving classes CI and DIII, at bias $0.05\,\mathrm V$ and inverse temperature approximately $2000\,\mathrm{eV}^{-1}$. Time-reversal-induced interference contributions are extracted by subtracting corresponding time-reversal-broken ensembles. They change sign as transparency varies. For ideal contacts and $N=1$, their relative magnitudes in $\mathcal Q_{\mathrm R}$ are approximately $51\%$ for DIII and $17\%$ for CI. At transparency $\Gamma=0.1$, $\mathcal Q_{\mathrm{MS}}$ is more sensitive, with corresponding fractions of $37\%$ and $12\%$. The reduced relative interference contribution is attributed to an enhanced Andreev-modified background. The treatment is restricted to coherent, noninteracting universal transport; finite-temperature and finite-energy transport, continuous symmetry crossovers, dephasing and interactions are identified as extensions.
Central claim may not hold: The displayed TUR observables retain an extra voltage factor after substituting I = GV and therefore are not dimensionless as required for comparison with the bound 2. The quoted bias and temperature correspond to e|V|/(k_B T) approximately 100 rather than thermodynamic linear response, and the finite-temperature TUR conclusion is not accompanied by a finite-temperature noise calculation or an approximation-error estimate. The paper does not justify identifying its Bogoliubov transmission-block formulas with electrical charge conductance and noise in the stated hybrid geometry, leaving the physical observable underlying its central TUR claim unestablished. The paper incorrectly interprets disagreement between an average ratio and a ratio of averages as a breakdown of the law of large numbers and central limit theorem.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.