🔬 condensed matter

Impact of etches on thin-film single-crystal niobium resonators

This study demonstrates that varying acid cleaning treatments on molecular beam epitaxy-grown single-crystal niobium thin films significantly degrade their electrical and superconducting properties by inducing hydrogen diffusion and surface hydrocarbon crystallite formation, leading to substantial reductions in resistivity ratios and resonator quality factors.

Original authors: H. Wang, T. Banerjee, T. G. Farinha, A. T. Hanbicki, V. Fatemi, B. S. Palmer, C. J. K. Richardson

Published 2026-10-01

📖 5 min read🧠 Deep dive

CC0 1.0

Original authors: H. Wang, T. Banerjee, T. G. Farinha, A. T. Hanbicki, V. Fatemi, B. S. Palmer, C. J. K. Richardson

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the quest to build quantum computers, scientists are trying to create machines that can solve problems far beyond the reach of today's most powerful supercomputers. At the heart of these machines are tiny circuits that store information in a fragile state called superconductivity, where electricity flows with absolutely no resistance. To keep this delicate state alive, the circuits must be kept colder than the depths of outer space. However, even in this frozen environment, tiny amounts of energy are lost, causing the information to fade away. This loss often comes from the very surfaces of the materials used to build the circuits. Just as a dirty window distorts light, a microscopic layer of contamination or a rough surface on a metal film can absorb energy and ruin the performance of the quantum device. Researchers are constantly searching for the perfect way to clean and prepare these surfaces to ensure the circuits can hold their information for as long as possible.

A team of researchers recently investigated how different cleaning methods affect a specific type of metal used in these circuits: niobium. They grew a thin, perfectly ordered layer of niobium on a sapphire wafer, creating a single-crystal film that is as smooth and uniform as possible. They then cut this film into tiny strips and resonators, which are devices designed to vibrate at specific frequencies, much like a guitar string. To test how clean the surface was, they subjected identical samples to a series of different acid baths. Some samples were treated with a standard mixture of sulfuric acid and hydrogen peroxide, known as piranha, which is excellent at removing organic grime. Others were treated with this same mixture followed by a buffered oxide etch, a solution designed to strip away layers of oxide. A final group was treated with a long soak in hydrofluoric acid, a powerful chemical often used to clean metals.

The results revealed that not all cleaning methods are created equal. The samples treated with the standard piranha acid alone performed well, showing low electrical resistance and high-quality vibrations. However, the samples that underwent longer treatments or were exposed to hydrofluoric acid suffered a dramatic decline in performance. Their electrical resistance increased significantly, and their ability to vibrate without losing energy dropped by a factor of twenty or more. Perhaps most strikingly, the surfaces of these damaged samples began to change over time. Within a day of being stored in a dry box, tiny triangular mounds began to appear on the metal, growing larger over several days. These formations looked like snowflakes and were found to be composed of hydrocarbons, a type of organic compound containing carbon and hydrogen.

The researchers used advanced imaging and chemical analysis to understand what was happening beneath the surface. They found that the niobium itself had not turned into a different chemical compound, nor had it formed a solid layer of hydride, which is a compound of niobium and hydrogen. Instead, the evidence pointed to hydrogen atoms from the acid treatments seeping into the metal and dissolving within its crystal structure. This invisible absorption of hydrogen, even in amounts too small to form a new solid phase, was enough to disrupt the flow of electricity and degrade the performance of the resonators. The triangular mounds that appeared later were likely a sign that the hydrogen was interacting with the surface, causing these hydrocarbon deposits to nucleate and grow. The study suggests that while these acids are intended to clean the metal, the hydrofluoric acid and long exposure times are actually introducing a hidden impurity that harms the material's superconducting properties.

This discovery highlights a delicate balance in the fabrication of quantum devices. The goal is to remove surface contaminants without introducing new problems. The researchers found that the single-crystal nature of their niobium film allowed them to see these effects more clearly than in previous studies using rougher, multi-grained metals. The triangular shapes of the defects were uniform because the underlying crystal structure guided their growth, a phenomenon that might be harder to spot in less ordered materials. The findings indicate that the hydrogen diffusing into the film was the primary culprit, acting as a silent disruptor that lowered the quality of the resonators and increased electrical resistance.

The study concludes that while acid cleaning is necessary, the specific type and duration of the treatment are critical. The long exposure to buffered oxide etch and the use of hydrofluoric acid led to the worst outcomes, creating surface defects and degrading the material's ability to function as a superconductor. The team did not find a single "perfect" cleaning method that maximized performance for every condition, but they did identify that certain aggressive treatments are counterproductive. The presence of hydrogen, even below the level where it forms a distinct hydride compound, is sufficient to cause significant damage. This work serves as a cautionary tale for the field, reminding scientists that the chemicals used to prepare a surface can leave behind invisible traces that fundamentally alter the behavior of the quantum circuits they are meant to improve.

Technical Summary: Impact of Etches on Thin-Film Single-Crystal Niobium Resonators

Problem Statement
Superconducting qubits rely on low-loss materials to maximize coherence times. While aluminum is the standard, refractory metals like niobium (Nb) offer larger superconducting gaps but present challenges regarding surface oxides and chemical processing. Different oxidation states of Nb can lead to high microwave loss, necessitating etching protocols to remove surface contaminants. However, the impact of specific acid cleaning treatments—particularly those involving hydrofluoric acid (HF)—on single-crystal Nb films is not fully understood. Previous studies on polycrystalline Nb have established that hydrogen incorporation can modify electrical resistance and form defects, but the specific effects on high-quality single-crystal thin films used for resonators require further investigation.

Methodology
The study utilized a single-crystal Nb (111) thin film (400 nm thick) grown via molecular beam epitaxy (MBE) on a c-plane sapphire substrate. The film exhibited high uniformity and a relaxed strain state.

Fabrication: Samples were patterned into DC resistivity test strips and coplanar waveguide (CPW) quarter-wave resonators (5.3–6.0 GHz) using standard photolithography and inductively coupled plasma reactive ion etching (ICP-RIE).
Cleaning Protocols: Following identical solvent cleans, samples underwent different acid treatments:
- NC: No acid clean (baseline).
- P: Heated piranha (60°C, 15 min).
- PSB: Heated piranha followed by a short Buffered Oxide Etch (BOE) (5 min).
- PLB: Room temperature piranha followed by a long BOE (90 min).
- HF: Room temperature undiluted 49% HF (5 min).
Characterization:
- Morphology: Optical microscopy and Atomic Force Microscopy (AFM) were used to inspect surface defects.
- Chemical Analysis: X-ray Photoelectron Spectroscopy (XPS) analyzed surface composition, oxidation states, and contamination. Raman spectroscopy was also employed to search for niobium hydrides.
- Electrical/RF Performance: DC resistivity was measured from 3.5 K to room temperature. Microwave loss was characterized via internal quality factors ( $Q_i$ ) of resonators at temperatures below 20 mK, measuring power dependence from single-photon levels up to $\sim 10^5$ photons.

Key Results

Surface Morphology: Samples treated with HF and the long BOE protocol (PLB) developed extensive surface crystallites over time (approx. 24 hours) when stored in a dry box. These triangular, 3-fold symmetric features grew up to 80 µm in length and were elevated mounds (~50 nm high). They appeared to form via inhomogeneous nucleation near etched features, with larger crystallites growing at the expense of smaller ones, leaving "clear fields" around them.
Chemical Composition (XPS):
- Acid cleans generally reduced carbon contamination compared to the NC sample.
- The PLB sample showed sulfur contamination (from piranha) in both defect-free and defective regions.
- Crucially, XPS revealed no significant difference in the oxidation state of Nb between the defective and defect-free regions of the PLB sample; both were primarily metallic Nb and $Nb_2O_5$ .
- No direct XPS evidence of hydrogen was found (as XPS cannot detect H directly), and Raman spectroscopy showed no distinct peaks for niobium hydrides.
DC Resistivity:
- Samples treated with HF and PLB showed a significant degradation in DC properties.
- The Residual Resistivity Ratio (RRR) dropped by a factor of 2 for HF (19.7) compared to the NC baseline (39.2), while the PLB sample showed a reduction to 25.5.
- Room temperature resistivity increased by over 10% for HF and PLB samples.
- The transition to the superconducting state became stepped rather than sharp for HF and PLB samples, with a reduced critical temperature ( $T_c$ ) of 7.9 K and 8.3 K, respectively, compared to 9.2 K for NC.
Microwave Loss:
- The internal quality factor ( $Q_i$ ) of resonators was severely impacted by HF and PLB treatments.
- At low photon numbers, the PLB sample exhibited a $Q_i$ of $\sim 5 \times 10^4$ , a drastic reduction compared to the NC baseline ( $\sim 0.5 \times 10^6$ ) and the P/PSB samples ( $\sim 1.5 \times 10^6$ ).
- The HF sample was not measured for RF performance due to its similar DC degradation and surface defects.

Significance and Claims
The authors conclude that the observed degradation in both DC conductivity and microwave quality factors correlates with the formation of surface crystallites, which appear to be hydrocarbons. The data suggests that hydrogen diffuses into the niobium film during HF and extended BOE treatments.

The paper posits that:

The hydrogen incorporation occurs at levels below the saturation threshold required to form observable niobium hydrides (consistent with the lack of hydride peaks in Raman and XPS).
The hydrogen likely occupies interstitial spaces in the BCC Nb lattice, acting as a single-phase mixture.
The reduction in RRR and $T_c$ , along with the increased microwave loss, is consistent with hydrogen loading effects previously documented in polycrystalline Nb, but observed here in a single-crystal context.
The single-crystal nature of the film allows for the observation of these interactions without the masking effects of grain boundaries found in polycrystalline studies.

The study does not claim to have optimized etch procedures for maximum $Q_i$ but rather highlights the sensitivity of single-crystal Nb to specific acid combinations, particularly those involving HF or prolonged BOE exposure, which can introduce hydrogen and surface defects detrimental to superconducting device performance.

More like this