Proximity-induced superconductivity and Josephson critical current in quantum spin Hall systems

TL;DR

This study uses numerical optimization to analyze supercurrent profiles in quantum spin Hall insulator Josephson junctions, confirming edge-mode superconductivity and highlighting limitations in extracting detailed edge information from Fraunhofer spectra.

Contribution

It provides a fully independent numerical method to analyze supercurrent distributions, validating experimental observations of edge superconductivity in QSHI systems.

Findings

Supercurrent peaks at sample edges in QSHI phase

Qualitative agreement with experimental Fraunhofer patterns

Quantitative analysis of supercurrent distribution is challenging

Abstract

We consider recent experiments on wide superconductor-quantum spin Hall insulator (QSHI)-superconductor Josephson junctions, which have shown preliminary evidence of proximity-induced superconductivity at the edge-modes of the QSHI system based on an approximate analysis of the observed Fraunhofer spectra of the Josephson critical current as a function of the applied magnetic field. Using a completely independent exact numerical method involving a non-linear constrained numerical optimization, we calculate the supercurrent profiles, comparing our results quantitatively with the experimental Fraunhofer patterns in both HgCdTe and InAs-GaSb based QSHI Josephson junctions. Our results show good qualitative agreement with the experiments, verifying that the current distribution in the 2D sample indeed has peaks at the sample edges when the system is in the QSHI phase, thus supporting the interpretation that superconductivity has indeed been induced in the QSHI edge-modes. On the other hand, our numerical work clearly demonstrates that it will be very difficult, if not impossible, to obtain detailed quantitative information about the super-current distribution just from the analysis of the Josephson Fraunhofer spectra, and, therefore, conclusions regarding the precise width of the edge modes or their topological nature are most likely premature at this stage.