Controlled Dephasing of Electrons by Non-Gaussian Shot Noise

TL;DR

This study demonstrates complete electron dephasing in a quantum interferometer caused by a few electrons, revealing non-Gaussian shot noise effects and unusual visibility patterns due to strong detector coupling.

Contribution

It provides the first experimental evidence of full dephasing by few electrons and explores the impact of non-Gaussian shot noise on quantum coherence.

Findings

Visibility exhibits periodic lobe-like suppression with increasing detector current.

Dephasing depends on the partitioning of the detector current, not just shot noise variance.

Strong coupling allows only 1-3 electrons to fully dephase the interferometer.

Abstract

In a 'controlled dephasing' experiment [1-3], an interferometer loses its coherence due to entanglement with a controlled quantum system ('which path' detector). In experiments that were conducted thus far in mesoscopic systems only partial dephasing was achieved. This was due to weak interactions between many detector electrons and the interfering electron, resulting in a Gaussian phase randomizing process [4-10]. Here, we report the opposite extreme: a complete destruction of the interference via strong phase randomization only by a few electrons in the detector. The realization was based on interfering edge channels (in the integer quantum Hall effect regime, filling factor 2) in a Mach-Zehnder electronic interferometer, with an inner edge channel serving as a detector. Unexpectedly, the visibility quenched in a periodic lobe-type form as the detector current increased; namely, it periodically decreased as the detector current, and thus the detector's efficiency, increased. Moreover, the visibility had a V-shape dependence on the partitioning of the detector current, and not the expected dependence on the second moment of the shot noise, T(1-T), with T the partitioning. We ascribe these unexpected features to the strong detector-interferometer coupling, allowing only 1-3 electrons in the detector to fully dephase the interfering electron. Consequently, in this work we explored the non-Gaussian nature of noise [11], namely, the direct effect of the shot noise full counting statistics [12-15].