Researchers Link Thicker Aluminum Electrodes to Two-Thirds Lower TLS Density in Josephson Junctions

By Samudrapom DamReviewed by Susha Cheriyedath, M.Sc.Oct 20 2026

A high-throughput study of thousands of Josephson junctions combines cryogenic measurements with atomic-scale imaging to probe how fabrication shapes one of the most troublesome defect populations in superconducting quantum hardware.

Paper: Structural control of two-level defect density revealed by high-throughput correlative measurements of Josephson junctions. AI-generated abstract conceptual image created using ChatGPT/OpenAI

*Important notice: This news reports on an unedited version of an accepted paper and is awaiting final editing. Therefore, the paper should not be regarded as conclusive or treated as established information.

A paper recently published in npj Quantum Information developed a statistically rigorous methodology to correlate the density of two-level systems (TLS) detected in cryogenic measurements of Josephson circuits with their microstructure and fabrication methods.

Strongly coupled TLS

In superconducting (SC) quantum devices, unpredictable material defects that couple to circuit modes are a major limit on qubit coherence. TLS coupling to quantum circuits leads to decoherence and excess loss.

In particular, strong TLS coupling to circuit modes enables coherent energy transfer between a single defect and a quantum circuit and can render the affected qubit effectively unusable.

Strongly coupled defects can sharply reduce device reliability, making it harder to build large superconducting quantum processors with consistent performance. Weakly coupled defects mainly appear as excess dielectric loss.

Strongly coupled TLS are thought to occur in amorphous interfaces and materials within and around the commonly used aluminum/aluminum oxide/aluminum (Al/AlO_x/Al) Josephson junction (JJ), including the oxide tunnel barrier.

Researchers have tried to determine the origin and nature of such defects to improve qubit quality and prospects for larger processors. Yet the microscopic origins of TLS remain unexplained, and it remains unclear how much device fabrication can eliminate or control them.

Role of correlative measurements

The origin of strongly coupled, detrimental TLS can be investigated by correlating defect densities with microstructural characteristics and fabrication parameters.

TLS statistics in qubits and junction microstructure have been studied independently, but before this work no link had been established between the density of strongly coupled defects and structural features.

Researchers had not identified structural controls that reduce TLS occurrence. Identifying such relationships requires large, correlated datasets that meet specific criteria.

These criteria include detecting a substantial number of TLS in cryogenic measurements of Josephson circuits to estimate their probability of occurrence. They also require statistics on various microstructural parameters from materials characterization methods, such as electron microscopy.

The high-throughput, correlated approach

To generate these datasets, the researchers used junction-array resonators designed to maximize the number of strongly coupled TLS accessible to measurement. They could tune the frequency across a range containing many TLS resonances, allowing the team to gather defect statistics efficiently. The researchers then compared measured TLS densities with fabrication parameters and junction microstructure in Al/AlO_x/Al JJs.

The team paired this resonator design with a fully automated and highly efficient measurement and analysis workflow that inferred TLS densities from cryogenic circuit characterization. They used a materials characterization and statistical analysis strategy to correlate microstructural features with observed TLS densities and fabrication variations. 

The authors calibrated the automated detector with simulated resonator data to account for false-positive and false-negative detections. STEM then quantified junction microstructure features, including grain size and electrode morphology.

Testing the approach

The authors used this statistically rigorous, efficient approach to correlate TLS density detected in cryogenic measurements of Josephson circuits with microstructure and fabrication methods. The study covered about 6,000 Al/AlO_x/Al JJs, while resonator spectroscopy detected 393 distinct TLS across five samples. The microscopy dataset contained 709 STEM images from five devices.

Devices incorporating JJ arrays in several microwave resonators were fabricated to identify enough distinct, strongly coupled TLS to resolve their probability distributions of occurrence.

Measurements across resonator ensembles were needed to obtain quantitatively reliable TLS densities. The coupling strengths the authors detected, on the megahertz (MHz) scale, were consistent with junction-related TLS reported in transmon qubits.

The microstructural features of JJ arrays fabricated alongside the cryogenically measured devices were systematically analyzed.

A strong correlation between Al grain size, Al electrode thickness, and TLS density was revealed by statistically linking microstructure, fabrication parameters, and TLS occurrence. The thicker-electrode treatment reduced mean detected TLS density from 0.20 ± 0.02 to 0.067 ± 0.013 TLS GHz^-¹ μm^-², a two-thirds reduction in detectable strongly coupled TLS density.

Annealing and slower Al deposition did not produce statistically distinguishable TLS density distributions.

The samples with thicker electrodes also had larger Al grains, so the study does not establish which linked structural change directly caused the lower TLS density. The authors suggest that grain boundaries, or adjacent AlO_x, may be TLS formation sites. The structural analysis covered five devices, and the authors note that additional data across a broader range of devices will be needed to determine the limits of TLS suppression.

The spectroscopy method is primarily sensitive to strongly coupled TLS, and the experiments used resonators containing JJ arrays rather than operating qubits. So the reported reduction applies only to the detectable strongly coupled TLS population; the experiments did not directly measure gains in qubit coherence or processor reliability.

The results show that fabrication choices can alter the measured density of strongly coupled TLS in the tested Al/AlO_x/Al junctions, while the exact microscopic mechanism remains unresolved. The methodology could also help investigate the origins of these defects in Josephson quantum devices.

Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.