Microscopic identification and per-device characterization of the low-frequency flux-noise fluctuators in superconducting transmon qubits

Wu et al. (2025) showed that strong telegraph-like flux noise — which they attribute, as a potential mechanism, to electromagnetic radiation from a room-temperature DC-current instrument coupling into the cryogenic wiring — produces nonmonotonic, beating Ramsey fringes in frequency-tunable transmon qubits, and derived a formally exact random-telegraph-noise (RTN) expression for the resulting decoherence within small-amplitude, slow-switching approximations. Three linked issues remain open. First, the physical identity, spatial distribution, and microscopic switching mechanisms of the low-frequency flux-noise fluctuators (surface or interface spins, vortices, spin clusters, or instrument-generated radiation) are unidentified, and no quantitative, mechanism-level account separates instrument-generated from on-chip microscopic contributions. Second, no validated procedure exists to extract the individual fluctuator parameters — switching rates $\lambda_j$ and flux amplitudes $b_j$ — from measured time-domain or spectral data on an active device, which is what the exact decoherence formula requires in order to be predictive for a given device. Third, the regimes of validity of the analytic approximations (linearized flux response, truncation to the no-switch term, neglect of $e^{-\lambda t}$) outside $b\ll\Phi_0$ and $\lambda t\ll 1$ are unquantified. Resolving these would turn a phenomenological noise model into a microscopically grounded, per-device predictive account of transmon dephasing and enable targeted mitigation.

Background

A transmon is a superconducting qubit in which a Josephson junction shunted by a large capacitance forms a weakly anharmonic oscillator whose lowest two levels encode the qubit. In the frequency-tunable variant used in most modern superconducting quantum processors, the single junction is replaced by a superconducting quantum interference device (SQUID) — a superconducting loop interrupted by Josephson junctions, whose effective inductance, and hence the qubit transition frequency $\omega_{01}(\Phi_b)$, depends on the magnetic flux $\Phi_b$ threading the loop. Uncontrolled low-frequency fluctuations of this flux ('flux noise') therefore modulate $\omega_{01}$ and randomly accumulate phase between the qubit's two energy eigenstates, a decoherence process called dephasing.

The power spectral density (PSD, the distribution of fluctuation power over frequency) of flux noise is empirically close to $1/f$ over many decades of frequency in superconducting circuits — a phenomenology known since SQUID magnetometers and still lacking an accepted microscopic explanation. The standard phenomenological picture models the noise as a superposition of random telegraph noise (RTN), the bistable flipping of a discrete fluctuator between two states at some switching rate $\lambda$ (each fluctuator contributing a Lorentzian PSD), with the switching rates broadly distributed across fluctuators, since such an ensemble generically produces a $1/f$ spectrum. Koch, DiVincenzo, and Clarke proposed that dilute localized magnetic moments ('surface spins') on the superconductor and substrate surfaces generate the noise, and the areal spin densities of order $10^{17}\,\mathrm{m^{-2}}$ inferred from measured noise support this picture. de Graaf et al. identified, by on-chip electron spin resonance (ESR) on sapphire ($\mathrm{Al_2O_3}$), dilute surface spins attributed to physisorbed atomic hydrogen and to $\mathrm{O_2^-}$ at a density matching the flux-noise-inferred value — but on test resonators rather than active qubits, and without connecting the identified spins to their switching dynamics. Indeed, the identity and interaction dynamics of the fluctuators remain unsettled: alternative candidate mechanisms include stray vortices (magnetic-flux quanta trapped in the superconducting films) and magnetic moments localized at metal–insulator interfaces via metal-induced gap states (electronic states that the metal's wavefunctions induce in the adjacent insulator), and the strong discrete telegraph fluctuators occasionally seen in devices have not been traced to identified microscopic sources. Noise spectroscopy experiments on superconducting qubits (e.g., Bylander et al.) have characterized the $1/f$ background and individual fluctuators phenomenologically for over a decade without settling their origin.

A Ramsey experiment measures dephasing directly: a $\pi/2$ pulse creates a superposition of the qubit's two energy eigenstates, which accumulates phase during a free-evolution time $t$ before a second $\pi/2$ pulse converts the accumulated phase into a measurable population; sweeping $t$ traces a 'Ramsey fringe' whose envelope decays under noise. Under quasi-Gaussian small-amplitude noise the envelope decays monotonically; strong discrete telegraph flux noise instead produces a nonmonotonic, beating envelope, as previously established for charge noise. Wu et al. (2025) reported beating Ramsey fringes in frequency-tunable transmons that appear when a particular room-temperature instrument — a DC-current source powering the high-electron-mobility transistor (HEMT) amplifier of the readout chain — sits near the dilution refrigerator, and disappear when that instrument is relocated; they attribute the effect, stated as 'a potential mechanism', to electromagnetic radiation emitted by the instrument and picked up by the coaxial wiring. They model the total flux noise as a $1/f$ 'flicker' background plus a few strong RTN sources of amplitudes $b_j$ and switching rates $\lambda_j$, and reproduce the measured fringes numerically. Their appendix derives, within a linearized flux response $\omega_{01}(\Phi_b+\delta\Phi)\approx\omega_{01}(\Phi_b)+(d\omega_{01}/d\Phi_b)\delta\Phi$ valid for amplitudes $b\ll\Phi_0$ ($\Phi_0=h/2e$ is the superconducting flux quantum), a formally exact expression for the Ramsey decay factor as a product over fluctuators of single-RTN phase averages (Eq. (27) of the arXiv version; 'Eq. C9' in the published numbering), whose commonly used slow-switching simplifications rely on $\lambda t\ll 1$ for the single-shot durations.

The open problem is thus not the reproduction of the observed beating — that is achieved — but the microscopic and predictive completion of this account: what the fluctuators actually are, how to identify them on a given active device, and where the analytic model breaks down.

Problem Statement

The problem left open by Wu et al. (2025) is to convert the phenomenological RTN/1/f flux-noise model of the frequency-tunable transmon into a microscopically identified and per-device predictive account of low-frequency flux noise. Concretely: (i) determine the physical identity, spatial distribution, and microscopic switching mechanisms of the fluctuators that generate the low-frequency flux noise — both the quasi-continuous $1/f$ background and the discrete strong telegraph sources — in superconducting transmon devices, establishing quantitatively which contributions stem from instrument-generated electromagnetic radiation coupled through the measurement wiring and which from on-chip microscopic sources such as surface or interface spins, vortices, or spin clusters; (ii) develop and validate a procedure to extract, from measured time-domain or spectral data on an active device, the individual fluctuator parameters (switching rates $\lambda_j$ and flux amplitudes $b_j$) of the strong RTN sources, sufficient to predict that device's Ramsey decoherence, including beating fringes, via the exact decay-factor formula; and (iii) quantify the regimes of validity of the analytic approximations underlying the model — the linearization of $\omega_{01}(\Phi_b)$, the truncation to the no-switch ($n=0$) term, and the neglect of the $e^{-\lambda t}$ factor — beyond the regimes $b\ll\Phi_0$ and $\lambda t\ll 1$ assumed for single-shot durations, including the consequences of faster switching, larger amplitudes, and nonlinear flux response. These components jointly serve one objective: identifying the actual flux-noise sources and making their decoherence consequences predictable for a given device, which is also the prerequisite for deciding between targeted mitigations — instrument-level electromagnetic hygiene versus surface, materials, or geometric engineering — as the source paper leaves for future work.

References

1. Z.-H. Wu, L.-X. Lei, X.-F. Zhang, S.-C. Xue, S. Hu, C. Li, X. Fu, P.-X. Chen, K. Lu, M.-T. Deng, and J.-J. Wu, 'Telegraph Flux Noise Induced Beating Ramsey Fringe in Transmon Qubits', Chinese Physics Letters 42, 040501 (2025). DOI: 10.1088/0256-307X/42/4/040501; arXiv:2502.05499, https://arxiv.org/abs/2502.05499.
2. R. H. Koch, D. P. DiVincenzo, and J. Clarke, 'Model for 1/f Flux Noise in SQUIDs and Qubits', Physical Review Letters 98, 267003 (2007). DOI: 10.1103/PhysRevLett.98.267003, https://doi.org/10.1103/PhysRevLett.98.267003
3. E. Paladino, Y. M. Galperin, G. Falci, and B. L. Altshuler, '1/f noise: implications for solid-state quantum information', Reviews of Modern Physics 86, 361 (2014). DOI: 10.1103/RevModPhys.86.361; arXiv:1304.7925, https://arxiv.org/abs/1304.7925
4. S. E. de Graaf, A. A. Adamyan, T. Lindström, D. Erts, S. E. Kubatkin, A. Ya. Tzalenchuk, and A. V. Danilov, 'Direct identification of dilute surface spins on Al2O3: Origin of flux noise in quantum circuits', Physical Review Letters 118, 057703 (2017). DOI: 10.1103/PhysRevLett.118.057703; arXiv:1609.04562, https://arxiv.org/abs/1609.04562
5. D. A. Rower, L. Ateshian, L. H. Li, M. Hays, D. Bluvstein, L. Ding, B. Kannan, A. Almanakly, J. Braumüller, D. K. Kim, A. Melville, B. M. Niedzielski, M. E. Schwartz, J. L. Yoder, T. P. Orlando, J. I.-J. Wang, S. Gustavsson, J. A. Grover, K. Serniak, R. Comin, and W. D. Oliver, 'Evolution of 1/f Flux Noise in Superconducting Qubits with Weak Magnetic Fields', Physical Review Letters 130, 220602 (2023). DOI: 10.1103/PhysRevLett.130.220602; arXiv:2301.07804, https://arxiv.org/abs/2301.07804
6. Z. Wang, H. Wang, C. C. Yu, and R. Q. Wu, 'Hydrogen as a Source of Flux Noise in SQUIDs', Physical Review B 98, 020403(R) (2018). DOI: 10.1103/PhysRevB.98.020403; arXiv:1709.06208, https://arxiv.org/abs/1709.06208
7. J. A. N. Aquino and R. de Sousa, 'Flux noise in disordered spin systems', Physical Review B 106, 144506 (2022). DOI: 10.1103/PhysRevB.106.144506; arXiv:2207.10033, https://arxiv.org/abs/2207.10033
8. S. Günzler, J. Beck, D. Rieger, N. Gosling, N. Zapata, M. Field, S. Geisert, A. Bacher, J. K. Hohmann, M. Spiecker, W. Wernsdorfer, and I. M. Pop, 'Spin Environment of a Superconducting Qubit in High Magnetic Fields', Nature Communications 16, 9564 (2025). DOI: 10.1038/s41467-025-65528-y; arXiv:2501.03661, https://arxiv.org/abs/2501.03661
9. K. G. Ray, Y. Rosen, J. L. Dubois, and V. Lordi, 'Field-Dependent Qubit Flux Noise Simulated from Materials-Specific Disordered Exchange Interactions Between Paramagnetic Adsorbates', arXiv:2602.09471 (2026). https://arxiv.org/abs/2602.09471
10. J. B. Fu, Da-Wei Wang, B. Ren, Z. H. Yang, S. Hu, G. Y. Huang, S. H. Cao, D. D. Liu, X. F. Zhang, X. Fu, S. C. Xue, Y. G. Che, Yu-xi Liu, M. T. Deng, and J. J. Wu, 'Flux-noise-resilient transmon qubit via a doubly-connected gradiometric design', arXiv:2601.02137 (2026). https://arxiv.org/abs/2601.02137
11. C. Li, Z. Yang, X. Zhang, Z. Wu, S. Xue, and M. Deng, 'Systematic Characterization of Transmon Qubit Stability with Thermal Cycling', Entropy 28(3), 296 (2026). DOI: 10.3390/e28030296; arXiv:2602.07522, https://arxiv.org/abs/2602.07522
12. J. Bylander, S. Gustavsson, F. Yan, F. Yoshihara, K. Harrabi, G. Fitch, D. G. Cory, Y. Nakamura, J.-S. Tsai, and W. D. Oliver, 'Noise Spectroscopy through Dynamical Decoupling with a Superconducting Flux Qubit', Nature Physics 7, 565 (2011). DOI: 10.1038/nphys1994; arXiv:1101.4707, https://arxiv.org/abs/1101.4707
13. G. Falci, P. J. Hakonen, and E. Paladino, '1/f noise in quantum nanoscience', in Encyclopedia of Condensed Matter Physics, 2nd ed., edited by T. Chakraborty (Elsevier, 2024), vol. 1, pp. 1003-1017; arXiv:2401.11989, https://arxiv.org/abs/2401.11989
14. D. Szczęśniak and S. Kais, 'Magnetic flux noise in superconducting qubits and the gap states continuum', arXiv:2003.04885 (2020). https://arxiv.org/abs/2003.04885