Low-leakage superconducting-qubit measurement with sub-100-ns total duration

Superconducting quantum processors rely on dispersive readout to extract qubit states, but this process is often slow and prone to measurement-induced state transitions (MIST), which cause leakage into non-computational states. The authors investigate whether high-fidelity, low-leakage readout can be achieved in the small qubit-resonator detuning regime, where multiphoton resonances are typically dense and problematic.

To optimize readout speed and fidelity, the authors employ a dedicated filter resonator coupled to the readout resonator. By setting the effective resonator decay rate to 30.8 MHz and tuning the dispersive shift close to the optimal signal-to-noise ratio (SNR) condition, they enable rapid signal acquisition and passive photon depletion. They characterize the system using the chi-kappa-power method and benchmark MIST using a leakage-sensitive measurement sequence, comparing their results against Floquet simulations to understand the dynamics of multiphoton resonances.

The researchers achieved a total measurement duration of 97 ns, defined as the time required for measurement-induced errors on subsequent operations to fall below 10^-4. They report an assignment error of 0.17% with a 58-ns measurement pulse, without requiring active depletion. The per-measurement leakage rate was found to be 2.7 x 10^-5, which is only twice the background rate and significantly lower than the measurement-induced relaxation rate. Floquet simulations confirm that while the system traverses several multiphoton resonances, the couplings are weak and the crossings are traversed diabatically, preventing significant leakage.

This work demonstrates that the small-detuning regime—often avoided due to concerns about dense multiphoton resonances—can actually support fast, high-fidelity, and low-leakage readout. By leveraging large resonator decay rates and rapid passive depletion, this approach simplifies the control requirements for quantum error correction by eliminating the need for complex active depletion pulses.