Three-Dimensional Niobium Coaxial Cavity with $\sim0.1\,$second Lifetime

TL;DR

This work demonstrates that a SRF-inspired surface treatment—comprising bulk BCP, high-temperature bake, a brief BCP flush, followed by mid-temperature annealing at 460–600°C—drastically lowers TLS-related losses on Nb 3D coaxial cavities. The optimized path, particularly a 600°C mid-temperature anneal, yields a record $Q_{int}$ of ≈$3\times10^9$ at base temperature and single-photon regime, with internal lifetimes on the order of tens to hundreds of milliseconds and remarkable stability against cooldowns and air exposure. XPS analysis links the improvement to a reduced Nb2O5 fraction and increased NbO, suggesting a surface oxide reconfiguration that stabilizes performance and is compatible with Nb-based quantum devices. Overall, the study bridges SRF surface processing with quantum Nb circuits, offering a practical route to long-lived Nb qubits and high-coherence 3D cavity elements for quantum information processing.

Abstract

We report on the internal quality factor of a three-dimensional niobium quarter-wave coaxial cavity, with mid-temperature annealing, exhibiting $Q_{\rm int} \gtrsim 3\times10^9$ at the single-photon level below 20\,mK, which corresponds to an internal photon lifetime of $τ_{\rm int}\sim90\,\mathrm{ms}$. Moreover, $Q_{\rm int}$ of the mid-temperature annealed cavities remains almost unchanged even after several cooldown cycles and air exposure. These results suggest that stable low-loss niobium oxides might be formed by mid-temperature annealing on the surface of three-dimensional niobium cavity. This surface treatment could be applicable to the fabrication of 2D superconducting circuits and help improve the lifetime of Nb-based superconducting qubits.

Figures (8)

• Figure 1: (a) A photo of the niobium cavity used in this study. (b) The simulated electric field distribution for the $\lambda$/4 mode of the coaxial cavity with input voltage $V_{\rm in}=0.1\,\mathrm{V}$ from the extetnal coupling port. (c) Characterization of the internal quality factor $Q_{\rm int}$ of cavity B2 after a mid-temperature anneal for 3 hours at 600 $^\circ\mathrm{C}$. The internal quality factor as a function of average photon number $\bar{n}$ in the cavity is fitted using the standard TLS model (Eq.(1)) and is shown in the red dashed line with $F\delta^0_{\rm TLS} = 1.82\times10^{-10}$, $n_c = 2.0\times10^6$, $\beta=0.34$, and $Q_{\rm res}=4.77\times10^{9}$. $Q_{\rm int}$ (points) is extracted from both $S_{11}$ reflection spectra (upper inset) and ring-down measurements (lower inset). Ring-down measurements are performed both with (orange) and without (magenta) a Traveling Wave Parametric Amplifier (TWPA) since the quantum-limited amplifier saturates under a high-power signal. Note that this data was obtained in the second cooldown, after exposing the cavity to air for 10 hours, leading insignificant degradation of $Q_{\rm int}$ compared to the first cooldown. (d) A schematics of the low-temperature part of the measurement setup employed for this experiment.
• Figure 2: Temperature dependence of the internal quality factor $Q_{\rm int}$ after each surface treatment. Dashed lines are fits to the temperature-dependent TLS model [Eq. \ref{['TLSmodel-T']}] using low-power data. Results are shown for cavities A2 and B2 after : 100$\,\mu\mathrm{m}$ BCP + 900 $^\circ\mathrm{C}$ for 3 hours + BCP flush (circles); and subsequent mid-temperature annealing at 460 $^\circ\mathrm{C}$ or 600 $^\circ\mathrm{C}$ for 3 hours (triangles). These measurements were performed at an average photon number of $\bar{n}\sim10^4$.
• Figure 3: Temperature dependence of the normalized frequency shift $\delta f / f_r$ (a) and internal quality factor $Q_{\rm int}$ (b) for cavity B2 measured across multiple cooldown cycles, before and after 600 $^\circ\mathrm{C}$ mid-temperature annealing.
• Figure 4: Comparison of $Q_{\rm int}$ as a function of average photon number $\bar{n}$ in cavity A2 after 460 $^\circ\mathrm{C}$ annealing over multiple cooldown cycles. The cavity was exposed to air for approximately 2 hours between each cooldown cycles, and it confirms the stability of the cavity performance against air exposure. $Q_{\rm int}$ is extracted from $S_{11}$ spectral measurements with the VNA.
• Figure 5: XPS spectrum of the Nb $3d$ region after a BCP flush, and subsequent 600 $^\circ\mathrm{C}$ annealing with 1 hour air and 1 day air exposures. Peaks corresponding to ${\rm Nb_2O_5}$, ${\rm NbO_2}$, NbO, ${\rm NbO_x} (0<x<1)$, and metallic Nb are indicated with filled areas.