Determining the temperature-dependent London penetration depth in HTS thin films and its effect on SQUID performance

TL;DR

This paper measures the temperature-dependent London penetration depth in HTS thin films, crucial for optimizing SQUID performance, and provides empirical models validated by experimental data from 50 to 79 K.

Contribution

It offers the first comprehensive measurement and modeling of $oldsymbol{ ext{λ}}(T)$ for HTS thin films across a range of temperatures relevant for device operation.

Findings

Average λ(77 K) = 391 nm for YBCO films

Empirical λ(T) expression derived from 50 to 79 K

Estimated inductances and areas match experimental data within 7%

Abstract

The optimum design of high-sensitivity Superconducting Quantum Interference Devices (SQUIDs) and other devices based on thin HTS films requires accurate inductance modeling. This needs the London penetration depth $\lambda$ to be well defined, not only at 77 K, but also for any operating temperature, given the increasingly widespread use of miniature low-noise single-stage cryocoolers. Temperature significantly affects all inductances in any active superconducting device and cooling below 77 K can greatly improve device performance, however accurate data for the temperature dependence of inductance and $\lambda(T)$ for HTS devices is largely missing in the literature. We report here inductance measurements on a set of 20 different thin-film YBCO SQUIDs at 77 K with thickness $t = 220$ or 113 nm. By combining experimental data and inductance modeling we find an average penetration depth $\lambda(77) = 391$ nm at 77 K, which was independent of $t$. Using the same methods we derive an empirical expression for $\lambda(T)$ for a further three SQUIDs measured on a cryocooler from 50 to 79 K. Our measured value of $\lambda(77)$ and our inductance extraction procedures were then used to estimate the inductances and the effective areas of directly coupled SQUID magnetometers with large washer-style pick-up loops. The latter agree to better than 7% with experimentally-measured values, validating our measured value of $\lambda(77)$ and our inductance extraction methods.