Phase measurements in Aharonov-Bohm interferometers

TL;DR

This paper investigates phase measurements in Aharonov-Bohm interferometers with quantum dots, deriving conditions under which the intrinsic quantum phase can be accurately extracted from conductance data.

Contribution

The paper provides theoretical models that relate measurable conductance to the quantum dot phase, including criteria for open interferometers to accurately reflect the intrinsic phase.

Findings

Relation $|t_{QD}|^2 ightarrow ext{sin}^2( ext{phase})$ allows direct phase measurement

Closed ABI conductance can reveal both amplitude and phase of transmission

Open ABI phase $eta$ depends on opening details, with criteria for accurate phase extraction

Abstract

In this paper we address measurements of the resonant quantum transmission amplitude $t_{QD}=-i|t_{QD}|e^{i\alpha_{QD}}$ through a quantum dot (QD), as function of the plunger gate voltage $V$. Mesoscopic solid state Aharonov-Bohm interferometers (ABI) have been used to measure the "intrinsic" phase, $\alpha_{QD}$, when the QD is placed on one of the paths. In a "closed" interferometer, connected to two terminals, the electron current is conserved, and Onsager's relations require that the conductance ${\cal G}$ through the ABI is an even function of the magnetic flux $\Phi=\hbar c\phi/e$ threading the ABI ring. Therefore, if one fits ${\cal G}$ to $A+B\cos(\phi+\beta)$ then $\beta$ only "jumps" between 0 and $\pi$, with no relation to $\alpha_{QD}$. Additional terminals open the ABI, break the Onsager relations and yield a non-trivial variation of $\beta$ with $V$. After reviewing these topics, we use theoretical models to derive three results on this problem: (i) For the one-dimensional leads, the relation $|t_{QD}|^2 \propto \sin^2(\alpha_{QD})$ allows a direct measurement of $\alpha_{QD}$. (ii) In many cases, the measured ${\cal G}$ in the closed ABI can be used to extract {\it both} $|t_{QD}|$ and $\alpha_{QD}$. (iii) For open ABI's, $\beta$ depends on the details of the opening. We present quantitative criteria (which can be tested experimentally) for $\beta$ to be equal to the desired $\alpha_{QD}$: the "lossy" channels near the QD should have both a small transmission and a small reflection.