Enhancement of photo-responsivity of small gap, multiple tunnelling superconducting tunnel junctions due to quasiparticle multiplication

TL;DR

This paper combines theoretical modeling and experimental validation to demonstrate how quasiparticle multiplication enhances photo-responsivity in superconducting tunnel junctions, especially at higher bias voltages.

Contribution

It introduces a set of energy-dependent balance equations modeling quasiparticle dynamics, including multiplication, and validates these models with experimental data on Al STJs.

Findings

Responsivity and signal decay time increase with bias voltage.

Theoretical predictions match experimental observations.

Quasiparticle multiplication significantly improves photon detection efficiency.

Abstract

We recently predicted the formation of a highly non-equilibrium quasiparticle (qp) distribution in low TC multiple tunnelling superconducting tunnel junctions (STJs) [1]. The situation arises through qp energy gain in cycles of successive forward and back tunnelling events in the absence of relaxation via sub-gap phonon emission. The qps can acquire sufficient energy to emit phonons, which break more Cooper pairs and release additional qps. In this paper we report theoretical and experimental studies of the effect of this process on photon detection by such an STJ. We derived a set of energy-dependent balance equations [2], which describe the kinetics of the qps and phonons, including the qp multiplication process described above. Solution of the balance equations gives the non-equilibrium distribution of the qps as a function of time and energy, and hence the responsivity of the STJ as a function of bias voltage. We compared the theoretical results with experiments on high quality, multiple-tunnelling Al STJs cooled to 35mK in an adiabatic demagnetisation refrigerator, and illuminated with monochromatic photons with wavelengths between 250 and 1000 nm. It was found that in the larger junctions with the longest qp loss time, both responsivity and signal decay time increased rapidly with bias voltage. Excellent agreement was obtained between the observed effects and theoretical modelling.