Abstract
Kinetic inductance devices (KIDs) are superconducting resonators with high kinetic inductance sensitive to external energy perturbations. KIDs made with superconductors having $T_c$ far below one Kelvin are of particular interest for sensing minuscule signals, such as light dark matter detection and millimeter wave telescopes for astronomy and cosmology. In this work, we report the promising performance of KIDs fabricated with Hafnium from heated sputter depositions. The KIDs have $T_c$ lower than $249 \pm 2$ mK, the internal quality factor ($Q_i$) of the resonators exceeds $10^5$ and the temperature dependence of the resonances can be described by the generalized Mattis-Bardeen model with a disorder parameter $\Gamma=4\times10^{-3}\Delta_0$. Post-fabrication annealing at temperatures above the deposition temperature can further reduce the $T_c$ and $\Gamma$ without reducing $Q_i$, leading to improvements in energy resolution.
Introduction
Kinetic Inductance Devices (KIDs) are superconducting resonators that serve as highly sensitive detectors. When energy, such as a dark matter particle or a photon, interacts with the detector's substrate, it breaks superconducting Cooper pairs into quasiparticles. This process alters the material's kinetic inductance, resulting in a measurable shift in its resonant microwave frequency.
Detecting extremely low-energy events, such as sub-GeV light dark matter interactions, requires detectors with sub-eV energy thresholds. Achieving this necessitates superconductors with a low critical temperature ($T_c$), as a lower $T_c$ reduces the energy required to break a Cooper pair, thereby increasing the number of quasiparticles generated per unit of deposited energy. While Aluminum ($T_c = 1.2$ K) is a standard material, its $T_c$ is too high for certain advanced architectures. This research investigates Hafnium (Hf) as a scalable alternative with a $T_c$ well below 1 K, high kinetic inductance, and a high quality factor.
Methods and Device Fabrication
The research team fabricated lumped-element KIDs by depositing 250 nm thick Hafnium films onto silicon substrates using a heated plasma sputtering process at 200 °C. The films were patterned using chlorine-based plasma etching. To explore optimization, some chips underwent post-fabrication annealing at 300 °C in an Argon environment for 30 minutes.
Characterization was performed in a dilution refrigerator at temperatures as low as 10 mK. Microwave transmission ($S_{21}$) was measured using a vector network analyzer to extract resonant frequencies and quality factors. The temperature dependence was modeled using a generalized Mattis-Bardeen theory, incorporating a disorder parameter ($\Gamma$) to account for energy state broadening below the superconducting gap.
Key Results and Insights
The study demonstrated that Hafnium is an excellent candidate for low-$T_c$ KIDs. The initial deposition yielded a $T_c$ of $249 \pm 2$ mK, which was further reduced to $233 \pm 2$ mK after annealing. Crucially, the internal quality factor ($Q_i$) remained high, exceeding $10^5$ at 10 mK, which is more than sufficient for high-resolution sensing.
Annealing also increased the surface kinetic inductance from $4.8 \pm 0.1$ pH/sq to $5.3 \pm 0.1$ pH/sq and reduced the disorder parameter ($\Gamma$) from $4 \times 10^{-3}\Delta_0$ to $1.5 \times 10^{-3}\Delta_0$. This reduction in disorder is significant because sub-gap states in highly disordered films can "soak up" quasiparticles, dampening the detector's response. By lowering $\Gamma$, annealing theoretically enhances the detector's energy sensitivity.
Conclusion and Future Directions
Hafnium KIDs, with a $T_c$ around 250 mK, are ideally suited for architectures where Aluminum serves as a phonon absorber. While the current results are promising, future work will focus on reducing film thickness to further increase kinetic inductance. Additionally, direct energy calibrations, quasiparticle lifetime measurements, and comprehensive noise characterization are planned to fully validate these sensors for next-generation dark matter and astrophysical experiments.
Summary updated