Aluminum-Based Superconducting Tunnel Junction Sensors for Nuclear Recoil Spectroscopy

TL;DR

This work demonstrates that aluminum-based superconducting tunnel junctions (Al-STJs) can serve as a viable platform for BeEST material studies, enabling systematic disentanglement of detector-material effects from potential sub-MeV sterile-neutrino signals. Through three fabrication iterations, the authors optimize geometry, membranes, and barrier transmissivity to improve signal response and suppress backgrounds, while calibrating with a UV laser and performing $^7$Be recoil spectroscopy. The final iteration achieves energy resolutions of 2–4 eV FWHM below 100 eV and 2.96 eV FWHM at 50 eV, indicating sufficient intrinsic detector performance for BeEST spectral analysis. These results establish Al-STJs as a complementary testbed to Ta-STJs for assessing material-dependent features in recoil spectra, thereby reducing systematics in future BeEST measurements and informing detector design for higher-statistics searches.

Abstract

The BeEST experiment is searching for sub-MeV sterile neutrinos by measuring nuclear recoil energies from the decay of $^7$Be implanted into superconducting tunnel junction (STJ) sensors. The recoil spectra are affected by interactions between the radioactive implants and the sensor materials. We are therefore developing aluminum-based STJs (Al-STJs) as an alternative to existing tantalum devices (Ta-STJs) to investigate how to separate material effects in the recoil spectrum from potential signatures of physics beyond the Standard Model. Three iterations of Al-STJs were fabricated. The first had electrode thicknesses similar to existing Ta-STJs. They had low responsivity and reduced resolution, but were used successfully to measure $^7$Be nuclear recoil spectra. The second iteration had STJs suspended on thin SiN membranes by backside etching. These devices had low leakage current, but also low yield. The final iteration was not backside etched, and the Al-STJs had thinner electrodes and thinner tunnel barriers to increase signal amplitudes. These devices achieved 2.96 eV FWHM energy resolution at 50 eV using a pulsed 355 nm (~3.5 eV) laser. These results establish Al-STJs as viable detectors for systematic material studies in the BeEST experiment.

Figures (6)

• Figure 1: Top-view and cross-sectional primitive of an Al-AlOx-Al STJ on an SiN membrane with Nb wiring, Nb quasiparticle plugs, Au bond pads, and Au thermalization layer. This primitive can be duplicated to produce chips with arrays of STJ pixels. The Al base electrode is read out by the right pad and requires a Nb plug. Anodization of the junction edges (see Section \ref{['al.fabrication']}) is shown in black, and is used to insulate the wiring from the base electrode when reading out the top electrode using the left pad. Layers not to scale.
• Figure 2: The current-voltage characteristics of the first Al-STJs on SiN at 100 mK have very high dynamic resistance above $180~\mu$V and surprisingly high thermal currents of 40--100 nA.
• Figure 3: First-run Al-STJ nuclear spectrum and fit to the BeEST model. The devices demonstrate $\sim$8 eV of electronic noise, but nonetheless successfully resolve the primary peaks expected from $^7$Be EC decay (labeled, see text for description).
• Figure 4: 128-pixel array of (130 $\mu$m)$^2$ STJ sensors on SiN membranes. The array is back-lit to highlight the transparent membranes.
• Figure 5: Current-voltage characteristics of an Al-STJ on an SiN membrane at 100 mK. No thermal leakage current was visible, but an exponential current increase at $1\Delta_\text{Al}/e$ indicates some flux trapping in this run.