Investigating the Origin of Two-Level Systems in Amorphous Oxides Using Molecular Dynamics

Superconducting qubits are a leading platform for quantum computing, but their performance is limited by relatively short coherence times. A major source of noise is from surface oxides on the interfaces of the superconductor, which contain two-level systems (TLSs) which couple to the qubit and induce decoherence. While it is well established that oxides adversely affect the coherence times of qubits, the microscopic origin of the TLS remains unclear. Building off previous work [1,2], we utilise machine learning potentials to reduce finite-size effects inherent to ab initio calculations on amorphous samples. From this work, we aim to illuminate the origin of TLSs in a wide family of amorphous oxides beyond the aluminium paradigm.

[1] A.C. Tyner, J.T. Heath, T.C. Thann, V.P. Michal, P. Krogstrup, M. K. Svendsen, and A.V. Balatsky, Advanced Quantum Technologies, e2500170 (2025)

[2] J.T. Heath, A.C. Tyner , T.C. Thann, V.P. Michal, P. Krogstrup, M. K. Svendsen, and A.V. Balatsky, arXiv:2503.08767 (2025)