Beating of Aharonov-Bohm oscillations in a closed-loop interferometer

TL;DR

This study demonstrates beating in Aharonov-Bohm oscillations in a GaAs/AlGaAs interferometer without strong spin-orbit interaction, attributing the phenomenon to multiple longitudinal modes rather than Berry-phase effects.

Contribution

It provides evidence that AB oscillation beating can occur independently of spin-orbit interaction, challenging previous interpretations involving Berry-phase pickup.

Findings

Beating observed in a low-SOI GaAs/AlGaAs interferometer.

Nodes of beats with h/2e period show parabolic distribution.

Fourier spectrum can be explained by multiple longitudinal modes.

Abstract

One of the points at issue with closed-loop-type interferometers is beating in the Aharonov-Bohm (AB) oscillations. Recent observations suggest the possibility that the beating results from the Berry-phase pickup by the conducting electrons in materials with the strong spin-orbit interaction (SOI). In this study, we also observed beats in the AB oscillations in a gate-defined closed-loop interferometer fabricated on a GaAs/AlGaAs two-dimensional electron-gas heterostructure. Since this heterostructure has very small SOI, the picture of the Berry-phase pickup is ruled out. The observation of beats in this study, with the controllability of forming a single transverse subband mode in both arms of our gate-defined interferometer, also rules out the often-claimed multiple transverse subband effect. It is observed that nodes of the beats with an h/2e period exhibit a parabolic distribution for varying the side gate. These results are shown to be well interpreted, without resorting to the SOI effect, by the existence of two-dimensional multiple longitudinal modes in a single transverse subband. The Fourier spectrum of measured conductance, despite showing multiple h/e peaks with the magnetic-field dependence that are very similar to that from strong-SOI materials, can also be interpreted as the two-dimensional multiple-longitudinal-modes effect.