Material Loss Model Calibration for Tantalum Superconducting Resonators
Guy Moshel 1, Sergei Masis 2, Moshe Schechter 3, Shay Hacohen-Gourgy 1 · 2025
Read Material Loss Model Calibration for Tantalum Superconducting Resonators by Guy Moshel 1, Sergei Masis 2, Moshe Schechter 3, Shay Hacohen-Gourgy 1 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-12141.txt.
Introduction
Guy Moshel and co-authors write a 2025 paper on testing materials for quantum circuits. Tiny flaws in a material, called two-level systems, drain energy from superconducting resonators. Knowing that loss well is key to better qubits, but past methods lean on heavy simulation.
The team gives a simple framework based on quality factor readings across heat and power ranges. They apply it to resonators made of tantalum, a rising material for quantum circuits. A plain formula model captures the loss without numeric simulation, which makes results easy to read and tune. Readings also show frequency trends that the standard flaw model does not explain, hinting at added loss paths in the material.
Catalog record
| CREATOR | Guy Moshel 1, Sergei Masis 2, Moshe Schechter 3, Shay Hacohen-Gourgy 1 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 35 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-12141
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Material Loss Model Calibration for Tantalum Superconducting Resonators Guy Moshel 1, Sergei Masis 2, Moshe Schechter 3, Shay Hacohen-Gourgy 1 Abstract Material research is a key frontier in advancing superconducting qubit and circuit performance. In this work, we develop a simple and broadly applicable framework for accurately characterizing two-level system (TLS) loss using internal quality factor measurements of superconducting transmission line resonators over a range of temperatures and readout powers. We applied this method to a series of α \alpha -Ta resonators that span a wide frequency range, thus providing a methodology for probing the loss mechanisms in the fabrication process of this emerging material for superconducting quantum circuits. We introduce an analytical model that captures the loss behavior without relying on numerical simulations, enabling straightforward interpretation and calibration. Additionally, our measurements reveal empirical frequency-dependent trends in key parameters of the model, suggesting contributions from mechanisms beyond the standard tunneling model of TLSs. \DeclareCaptionJustification justified I Introduction Superconducting qubits and resonators are the cornerstone of many advanced quantum information systems, including quantum simulations, quantum error correction, and quantum communication Devoret and Schoelkopf 2013 ; Wendin 2017 ; Houck et al. 2012 . Despite significant progress, scaling quantum processors to practical sizes remains constrained by qubit coherence times, which is currently limited by losses that far exceed those anticipated from the bulk properties of the constituent materials Siddiqi 2021 . These elevated losses suggest that critical decoherence pathways arise from uncontrolled surfaces, interfaces, and contaminants. Substantial work has been done to reduce these interfacial losses, in various material platforms Biznárová et al. 2024 ; Melville et al. 2020 ; Deng et al. 2023 . Specifically, tantalum-based superconducting qubits have achieved record lifetimes, with coherence times exceeding 0.3ms Place et al. 2021 ; Wang et al. 2022 . These results have been reproduced across diverse fabrication methods and substrates, underscoring the potential of tantalum as an important material for quantum hardware. Nevertheless, the sources of loss in high-quality-factor (Q-factor) tantalum devices remain poorly understood and present a major obstacle to further improvement. A detailed and accurate model for the loss mechanisms in tantalum is necessary. The primary loss mechanisms in superconducting devices are radiation and packaging Huang et al. 2021 ; Sage et al. 2011 , non-equilibrium quasiparticles Serniak et al. 2018 , bulk material absorption Braginsky et al. 1987 ; Gurevich and Tagantsev 1991 and two-level systems (TLSs) on the material surfaces McRae et al. 2020 . The contribution of TLSs is difficult to differentiate and accurately estimate because of the presence of multiple TLS sources within a single device and their dependence on the device geometry and fabrication method. Work had been done on various materials Burnett et al. 2018 ; Burnett et al. 2015 ; Paik and Osborn 2010 ; Lisenfeld et al. 2019 , but tantalum remained under-researched. In this work, we systematically investigate the contributions of TLSs and other loss mechanisms in tantalum superconducting resonators. By varying the readout power by many orders of magnitude, and also the temperature, we achieve precise characterization of subtle differences between the loss mechanisms in the resonators. We expand on the work of Crowley et al. 2023 , which used a similar saturation model but with added empirical parameters. Instead, we incorporate the spatial variation of the electric field in the resonators into our model, which improved the fit quality without adding any parameters to the model. In addition, we measured resonators across a wide frequency range: from 2.3GHz to 15.7GHz. This allowed us to observe frequency trends in the model parameters that were extracted for each resonator, shedding light on the underlying physics of the TLS loss, specifically the frequency dependence of the density of states. II setup We used a 2-inch, 420 \mu m thick sapphire wafer, sputtered with a 170nm thick \alpha -Ta layer by STAR cryoelectronics. The wafer was cleaned by sequential 5 minute sonication in acetone, methanol, and isopropanol, and then coated with AZ1518 photo-resist, exposed with the desired pattern using a DWL66+ laser writer by Heidelberg, and developed in TMAH (tetramethylammonium hydroxid) for 40 seconds. A hard bake was then done, and the tantalum was wet-etched in Ta-etchant 111 by Transene for 21 seconds. We examined the smoothness of the edges of the resulting tantalum structures using a scanning electron microscope (SEM), according to the characteristics in Place et al. 2021 . The etching results are shown in Fig. 2 . The wafer was then diced into the correct chip size and mounted on a designated box made of oxygen-free high-thermal-conductivity (OFHC) copper. The chip was placed on four copper pedestals at its corners and secured using ge varnish, which is thermally conductive at cryogenic temperatures. A silver-coated printed circuit board (PCB) was electrically and thermally anchored to the copper box using indium solder. The ground and signal ports of the chip were electrically connected to the PCB using aluminum wire bonds, as shown in Fig. 1 . Finally, the box was sealed with an OFHC copper lid and mounted inside the 10mK stage of a commercial BlueFors dilution cryostat. Care was taken to ensure good thermal anchoring to the cryostat. A \mu -metal shield was used to repel stray magnetic fields from the sample at low temperatures. Figure 1: The experimental setup. (a) The copper box and silver-coated PCB for chip mounting. Only the marked slot was used in this work. (b) The fabricated chip after mounting. Bright areas are tantalum and dark areas are sapphire substrate (etched tantalum). wire bond between the tantalum and PCB are visible. (c) A micrograph of one of the measured resonators. Figure 2: A SEM image of an edge of the etched tantalum, in the optimized wet process. The smooth boundary indicates valid etching. The white strip is an imaging artifact caused by substrate charging. The patterned structures were a series of \lambda/4 transmission line resonators, coupled to a common feedline in a hanging (or notch) configuration, as shown in Fig. 1 (c). The resonators and feedline were made of coplanar waveguide (CPW) segments. We measured the complex transmission signal S_{12} through the feedline to find the full lineshapes of the base mode of each resonator. By fitting each lineshape according to the procedure laid out in Probst et al. 2015 using the qkit python module, we extracted the resonance frequency and the internal and external Q-factors ( Q_{i} and Q_{e} , respectively). The measurements were carried out using a Keysight network analyzer, PNA-L model N5232B. For the temperature scans, we waited 30 minutes at each temperature point for the system to thermalize, and only then performed the power scans. We calculated the microwave power inside the feedline from the known input power using the nominal attenuator values at cryogenic temperature. The average microwave energy stored in the resonators for an impedance matched feedline, is given by McRae et al. 2020 W=\frac{2Q_{l}^{2}}{Q_{e}}\frac{P_{0}}{\omega_{0}}, (1) where Q_{l} and Q_{e} are the loaded and external quality factors of the resonator, respectively, \omega_{0} is its angular resonant frequency and P_{0} is the power in the feedline. To convert this energy to voltage, which we used in the model described below, we used V_{0}=\sqrt{2WZ_{0}}, (2) where Z_{0}=50\Omega is the resonator impedance. III model