Superconducting qubit decoherence correlated with detected radiation events

Most quantum error correction (QEC) protocols for superconducting qubits assume spatially and temporally uncorrelated decoherence events; however, recent evidence suggests that cosmic radiation induces spatially correlated errors. We present a platform that sandwiches a superconducting transmon qubit between two microwave kinetic inductance detector (MKID) arrays, enabling real-time detection of radiation-induced phonon bursts. By synchronizing MKID event detection with single-shot measurements of qubit energy relaxation ($T_1$) and phase coherence ($T_2$), we observe statistically significant reductions in both $T_1$ and $T_2$-up to 30.5%-immediately following dual MKID events attributed to penetrating muons. Our findings directly link radiating events to correlated qubit decoherence. Furthermore, our experimental platform provides a foundation for systematic studies of radiation effects, the development of shielding and mitigation techniques, and the refinement of error-correction algorithms tailored to correlated noise sources.

💡 Research Summary

The paper presents a novel experimental platform that integrates a superconducting transmon qubit with two microwave kinetic inductance detector (MKID) arrays positioned above and below the qubit chip. This three‑layer architecture enables simultaneous, real‑time detection of high‑energy radiation events and measurement of the qubit’s coherence properties. The motivation stems from the growing recognition that cosmic‑ray particles, especially muons, can generate energetic phonon bursts in the substrate, producing quasiparticles that simultaneously affect many qubits. Such correlated errors violate the core assumption of most quantum error‑correction (QEC) protocols—that decoherence events are spatially and temporally independent—thereby threatening the scalability of superconducting quantum processors.

The MKID arrays consist of 3 × 3 aluminum resonators on sapphire, each resonator having a distinct frequency set by its interdigitated capacitor. When a radiation‑induced phonon reaches the MKID film, it breaks Cooper pairs, increasing the kinetic inductance and shifting the resonator’s frequency. The authors monitor the I‑Q response of each resonator in a pulsed readout scheme, triggering data acquisition when any resonator’s phase or amplitude exceeds a calibrated threshold. A short 15 µs buffer followed by a 700 µs observation window captures the full relaxation of the quasiparticle population, allowing the deposited energy to be inferred via a temperature calibration (valid between 100 mK and 300 mK).

Events are classified into three categories: “top‑only” (detected by the upper MKID chip), “bottom‑only” (detected by the lower chip), and “dual” (simultaneous detection on both chips). Dual events constitute roughly 25 % of all recorded events and are interpreted as signatures of particles that traverse the entire stack—most plausibly cosmogenic muons. Energy histograms reveal that dual events deposit less energy per detector on average than single‑chip events, consistent with a particle passing perpendicularly through the substrate rather than stopping within it.

Qubit coherence is probed using single‑shot protocols. For T₁, a π‑pulse excites the qubit, a waiting time equal to 33 % of the measured T₁ is inserted, and the state is read out. For T₂, a detuned Ramsey sequence places the qubit on the equator, lets it precess for a quarter of the Ramsey period (~2 µs), applies a second pulse, and reads out. After each MKID trigger, the qubit measurement sequence is initiated with a latency of 10–15 µs, and fifty such repetitions are recorded per trigger. Over long acquisition periods (>12 h), more than a thousand dual‑event shots are collected, enabling statistically robust averaging.

The results show a pronounced reduction in both T₁ and T₂ following dual MKID events. The average T₁ drops by up to 30.5 % relative to the baseline, and T₂ exhibits a comparable decline. The recovery of T₁ follows an exponential with a 1/e time constant of 38 µs, while T₂ recovers with a 1/e constant of 25 µs. In contrast, top‑only or bottom‑only events do not produce statistically significant changes in the qubit’s coherence times. This asymmetry strongly supports the hypothesis that only particles capable of penetrating both detector layers—i.e., high‑energy muons—generate substrate‑wide phonon bursts that raise the quasiparticle density throughout the qubit’s superconducting film, thereby inducing correlated decoherence.

The authors also benchmark the observed dual‑event rate (0.013 ± 0.004 events s⁻¹ cm⁻²) against the expected sea‑level muon flux (≈0.017 events s⁻¹ cm⁻²), finding good agreement. Detailed engineering choices—such as the 113.9° opening angle of the MKID arrays, the ±57° solid‑angle coverage, gold‑plated copper sample boxes for thermalization and microwave shielding, >50 dB isolation between measurement chains, and the use of a Quantum Machines OPX+ controller for synchronized pulse generation—are documented to facilitate reproducibility.

In the discussion, the authors argue that their platform provides a quantitative link between radiation events and correlated qubit errors, a critical step toward designing effective mitigation strategies. Potential avenues include optimized shielding, real‑time error flagging based on MKID triggers, and the development of QEC codes that explicitly model spatially correlated noise. By demonstrating that muon‑induced phonon bursts can measurably degrade both energy relaxation and dephasing, the work underscores the necessity of accounting for cosmic‑ray backgrounds in the roadmap toward fault‑tolerant superconducting quantum computers.