Materials defects within Josephson junctions, frequently termed two-level systems, represent a substantial limitation to the development of scalable quantum processors. Oliver F. Wolff, Harshvardhan Mantry, and Rahim Raja, working with colleagues from the Department of Physics and Department of Material Science and Engineering at The Grainger College of Engineering, University of Illinois at Urbana-Champaign, alongside researchers from the Materials Research Laboratory at the same institution, present a high-throughput methodology to identify the microstructural origins of these strongly-coupled two-level systems. Their research correlates data from over 6,000 aluminium/aluminium oxide/aluminium Josephson junctions with more than 600 atomic resolution transmission electron microscopy images, statistically linking fabrication parameters, microstructure, and the occurrence of TLS. This analysis reveals a strong relationship between aluminium electrode thickness, aluminium grain size, and TLS density, ultimately demonstrating a two-thirds reduction in TLS achieved through modified electrode fabrication, and establishing a robust, data-driven approach to defect control in quantum circuits.
Scientists are tackling a critical bottleneck in the development of scalable quantum processors, unpredictable defects within superconducting circuits. These defects, known as two-level systems (TLS), disrupt qubit coherence and limit the reliability of increasingly complex quantum devices. While their detrimental effects are well-documented, pinpointing the microscopic origins of TLS and devising effective mitigation strategies has remained a significant challenge. This work presents a high-throughput methodology to trace the structural origins of these strongly-coupled TLS in Josephson circuits, offering a pathway towards more stable and scalable quantum computation. Researchers assembled a comprehensive dataset encompassing over 6,000 Josephson junctions (JJs) fabricated with aluminium/aluminium oxide/aluminium, a common material combination, alongside more than 600 atomic resolution transmission electron microscopy images. By statistically linking fabrication parameters, microstructural characteristics, and the occurrence of TLS, they have revealed a strong correlation between the thickness of the aluminium electrodes, the size of the aluminium grains, and the density of TLS. This detailed analysis demonstrates that a two-thirds reduction in TLS density can be achieved through targeted adjustments to electrode fabrication parameters. Observed aluminium grain sizes varied considerably, influencing the density of TLS present within the junctions; smaller grain sizes consistently correlated with higher TLS densities, suggesting that grain boundaries act as preferential sites for defect formation. The team developed a workflow centred around maximising detectable defects using specifically designed resonator circuits and a fully automated analysis pipeline for inferring TLS densities from cryogenic measurements. Combining this with materials characterisation techniques, particularly scanning transmission electron microscopy, allowed for the extraction of key microstructural features. The resulting data revealed a clear relationship between aluminium electrode thickness, grain size, and TLS densities, demonstrating that controlling these fabrication parameters can effectively reduce the occurrence of detrimental defects. Atomic resolution transmission electron microscopy underpinned the methodological approach to identifying the microstructural origins of TLS within Josephson junctions. Josephson junction array resonators were fabricated and then characterised cryogenically to efficiently detect hundreds of distinct, strongly-coupled TLS, enabling the extraction of their occurrence distributions for different fabrication recipes. Scanning transmission electron microscopy was then employed to examine the microstructure of the fabricated junction arrays, with image analysis extracting key features including Al grain size, interface roughness, and layer morphology, providing a detailed characterisation of the material’s physical properties. By linking TLS density to these microstructural characteristics, the research aimed to resolve correlations and provide clues for identifying the sources of TLS. This combined approach, cryogenic measurement of TLS and detailed materials characterisation, represents a methodological innovation, moving beyond independent investigations of junction microstructure and TLS statistics. The large dataset size was critical for establishing statistically robust correlations, overcoming the challenges inherent in studying unpredictable material defects. Notably, a modification to the aluminium electrode deposition process prompted a substantial two-thirds reduction in TLS density, demonstrating a pathway for mitigating these performance-limiting defects. This decrease signifies a considerable improvement in material quality and potential qubit coherence. The ability to reduce TLS by approximately 66% represents a substantial step towards improving the reliability and scalability of superconducting quantum processors. The relentless pursuit of stable qubits has long been hampered by material defects. For years, the microscopic sources of these defects, known as two-level systems, have remained elusive, hindering progress towards scalable quantum processors. This new work represents a significant step forward, not because it offers a quick fix, but because it establishes a robust, data-driven methodology for understanding and ultimately controlling these troublesome imperfections. By linking fabrication parameters, microstructure, and the prevalence of these defects, the researchers have identified a clear connection between aluminium electrode thickness, grain size, and the density of problematic two-level systems. The observed reduction in defects following a change in fabrication is encouraging, demonstrating that mitigation is indeed possible through careful materials engineering. However, this is not a complete solution; while electrode properties are clearly important, the interplay with other materials and interfaces within the junction remains a complex area requiring further investigation. Future work will likely focus on streamlining these processes and exploring whether similar correlations exist in junctions fabricated from different materials. The broader implication is a shift towards a more ‘materials science’ approach to quantum computing, recognising that building better qubits requires not just clever design, but a deep understanding of the materials themselves.
👉 More information
🗞 Structural control of two-level defect density revealed by high-throughput correlative measurements of Josephson junctions
🧠ArXiv: https://arxiv.org/abs/2602.11469