Pauli principle and chaos in a magnetized disk

TL;DR

This study investigates how quantum chaos influences the magnetic response of a confined electron gas under magnetic fields, revealing distinct behaviors in regular versus chaotic regimes and potential experimental implications.

Contribution

It provides a detailed quantum analysis of electron gases in magnetic fields, linking classical chaos with quantum spectral and magnetic properties, and compares fermionic and bosonic statistics.

Findings

Magnetization oscillates non-monotonically in regular regimes and remains diamagnetic in chaotic regimes.

Time-averaged energy scales as N^2 in regular and linearly in N in chaotic regimes.

Quantum signatures of chaos affect magnetic response and energy scaling in confined electron gases.

Abstract

We present results of a detailed quantum mechanical study of a gas of $N$ noninteracting electrons confined to a circular boundary and subject to homogeneous dc plus ac magnetic fields $(B=B_{dc}+B_{ac}f(t)$, with $f(t+2\pi/\omega_0)=f(t)$). We earlier found a one-particle {\it classical} phase diagram of the (scaled) Larmor frequency $\tilde\omega_c=omega_c/\omega_0$ {\rm vs} $\epsilon=B_{ac}/B_{dc}$ that separates regular from chaotic regimes. We also showed that the quantum spectrum statistics changed from Poisson to Gaussian orthogonal ensembles in the transition from classically integrable to chaotic dynamics. Here we find that, as a function of $N$ and $(\epsilon,\tilde\omega_c)$, there are clear quantum signatures in the magnetic response, when going from the single-particle classically regular to chaotic regimes. In the quasi-integrable regime the magnetization non-monotonically oscillates between diamagnetic and paramagnetic as a function of $N$. We quantitatively understand this behavior from a perturbation theory analysis. In the chaotic regime, however, we find that the magnetization oscillates as a function of $N$ but it is {\it always} diamagnetic. Equivalent results are also presented for the orbital currents. We also find that the time-averaged energy grows like $N^2$ in the quasi-integrable regime but changes to a linear $N$ dependence in the chaotic regime. In contrast, the results with Bose statistics are akin to the single-particle case and thus different from the fermionic case. We also give an estimate of possible experimental parameters were our results may be seen in semiconductor quantum dot billiards.