Low-loss Nb on Si superconducting resonators from a dual-use spintronics deposition chamber and with acid-free post-processing

TL;DR

The paper addresses magnetic impurities that degrade superconductivity and investigates whether a magnetron sputter chamber used for magnetic materials can still yield high-quality Nb-on-Si superconducting resonators. It combines substrate preparation variants, resist-strip baths, and acid-free post-processing, with comprehensive materials characterization (SIMS, XPS, AFM) and cryogenic microwave measurements to assess surface contamination and device losses. The key findings show no detectable magnetic impurities in Nb bulk or Nb–Si interfaces, and CPW resonators achieve internal quality factors near $Q_i \approx 10^6$ with a $g=3$ μm gap; acid-free post-processing can match acid-treated performance, and AZ300T strip baths with HF dipping can reach near-state-of-the-art $\,\delta_{LP}$ around $0.83\times10^{-6}$. These results demonstrate the viability of dual-use deposition tools for superconducting nanofabrication, broaden access to materials exploration, and suggest pathways for integrating magnetic materials into hybrid devices.

Abstract

Magnetic impurities are known to degrade superconductivity. For this reason, physical vapor deposition chambers that have previously been used for magnetic materials have generally been avoided for making high-quality superconducting resonator devices. In this article, we show by example that such chambers can be used for this purpose; with Nb films sputtered in a chamber that continues to be used for magnetic materials, we demonstrate compact (\SI{3}{\micro\meter} gap) coplanar waveguide resonators with low-power internal quality factors near one million. We achieve this using a resist strip bath with no post-fabrication acid treatment, which results in performance comparable to previous strip baths with acid treatments. We also find evidence that this improved resist strip bath provides a better surface chemical template for post-fabrication hydrogen fluoride processing. These results are consistent across three Si substrate preparation methods, including a \SI{700}{\celsius} anneal. These results will inform nanofabrication for other superconducting materials and the integration of magnetic materials for hybrid systems.

Figures (14)

• Figure 1: Secondary-ion mass spectrometry (SIMS) profiles of magnetic impurities (Fe, Co, Ni, and Cr) as a function of depth in a 60nm Nb film. The raw intensity of impurity ions was normalized to the total secondary ion counts. (a) A Nb film deposited with only superconductors present in the chamber (DP1). (b) A Nb film deposited with co-loaded magnetic targets having been used in the chamber (DP2). In both cases, no magnetic contamination is detected in the bulk of the film or at the substrate-metal interface. Differences in signal are only observed at the metal-air (MA) interface. (c) Additional measurements focused on the metal-air interface of similar Nb samples for both the control and DP2 films. Solid filled shapes represent control samples and unfilled shapes represent DP2 samples.
• Figure 2: Summary of resonator measurements from DP1 performed across three measurement setups, referred to as Meehl, NIST, and SU for different substrate preparation methods with the 1165 sample treatment. (a) Optical microscope device image of CPW resonator. (b) A sample Thermal resonator measurement at $\langle n\rangle \approx 10^5$ photons on the Meehl cryostat at 17mK. Black and purple points are the measured magnitude and phase of $S_{21}$, respectively, plotted on the left and right axis, indicated by the arrows. The dashed lines are the fit for the data, from equation \ref{['eq:S21-fit']}. (c) Three examples of loss $\delta_{i}$ vs. average photon number $\langle n\rangle$ for different samples. Colors and shapes represent the sample type and cryostat used for measurements, respectively, as shown in the panel (d) legend. (c) Box plots of low power losses $\delta_{LP}$ for the devices measured with the different substrate preparation methods. Colors and shapes represent the substrate preparation and the fridge, respectively. HF dose refers to the duration of the post-fabrication HF treatment. Star symbols plot the median HP losses for each device. Plus symbols plot the median difference between LP and HP losses for each device. HP measurements were not taken for the fourth device from the left. Some combinations of preparation and measurement protocols were repeated for more than one sample die. The shaded band is the window of 25th to 75th percentile performance of all of the resonators plotted, with the dashed line being the median.
• Figure 3: Performance of AZ300T resist strip. First three datasets use the 1165 strip bath with no post-fabrication acid treatment. The remainder all use the AZ procedure. Films deposited with both DP1 and DP2 are shown. E-beam labels Nb films deposited by electron-beam evaporation. The E-beam sample with blue markers had an additional in situ anneal step for UHV oxide desorption. The shaded band is the window of 25th to 75th percentile performance of the acid-treated resonators in Fig. \ref{['fig:boxplots']}, with the dashed line being the median. Vertical dashed line separates resonators without post-HF, with post-HF, and with post-BOE treatments. The two post-BOE samples were aged for 4 and 6 weeks, as indicated by the arrows.
• Figure 4: Scaling plot comparing our median losses to similar studies in the literature for Nb CPW resonators mcrae_materials_2020jones_grain_2023alghadeer_surface_2023altoe_localization_2022oh_exploring_2024torres-castanedo_formation_2024zheng_nitrogen_2022kowsari_fabrication_2021wang_impact_2024drimmer_effect_2024manzo-perez_physical_2025. Solid markers indicate post-fabrication acid treatment and open markers indicate no acid treatment. The median LP losses for DP1, 1165 samples with post-HF treatment in our work were $\delta_{LP}=1.37\times 10^{-6}$ for BOE, Anneal, and Thermal treatments combined, marked with a teal square with a thin black outline. The median LP losses for AZ samples without post-HF in our work were $1.52\times10^{-6}$ for DP1 and DP2 combined, marked with an empty teal diamond. The median LP losses for AZ samples with post-BOE in our work were $0.48\times10^{-6}$ for DP1 and DP2 combined, marked with a teal square and thick black outline. All of our samples correspond to $g=3$$\mu$m; some points are offset horizontally for visibility. Gray dashed lines approximate constant surface loss tangent with varied CPW resonator gap width $g$. Our losses are on par with state-of-the-art devices in the field, showing that our films are of comparable quality. Similarly, the markers for Torres-Castanedo, 2024 are slight offset from $g=3µm$, as well as Kowasari, 2023 and Zheng, 2022 from $g=2µm$. We remark that Zheng, 2022, zheng_nitrogen_2022 used a rather deep trench of $\sim500nm$ which likely significantly reduces the metal surface participation relative to the other works, and Torres-Castanedo, 2024, torres-castanedo_formation_2024 uses a fluorine-based etch chemistry rather than chlorine.
• Figure S1: SIMS profiles of (a) hydrogen, (b) carbon, (c) fluorine, and (d) oxygen contamination for the three different substrate preparation methods with the 1165 recipe, normalized to total secondary ion counts. Dashed black lines plot the Si profile, indicating the Si substrate. Analysis focuses on differences in the bulk of the Nb film and the substrate-metal interface. In particular, we see qualitative differences in the amount of fluorine and carbon present across the three samples.