Phase-Resolved Observation of Amplitude–Phase Coupling in a Strongly Driven Superconducting Qubit

Note (September 2026): This preprint is superseded by a new manuscript and dataset (doi:10.5281/zenodo.22803589); several of its claims are withdrawn there. [Update note, April 2026]A more comprehensive version of this work, with substantially increasedstatistics (12 phase points, 1000 shots/point, additional supplementaryexperiments including isenergetic and gap-insertion tests), has beenposted: "Phase-Resolved Observation of Amplitude-Phase Coupling in aStrongly Driven Superconducting Qubit" (DOI: [10.5281/zenodo.19816115]). The Berryconnection interpretation discussed in the present record has beenrevised in the subsequent work; readers interested in the mechanismanalysis should consult the accompanying mechanism paper(DOI: [10.5281/zenodo.19816695]).[End of update note] We report a differential measurement comparing the phase response of a superconducting transmonqubit under two control trajectories: a closed-loop Figure-8 pulse (non-zero area in the IQcontrol plane) and a constant-amplitude Rectangular pulse (zero area). Measurements were performedat five drive amplitudes on a Rigetti Ankaa-3 processor (qubit #25) via AWS Braket. Threetrajectory-dependent features are observed that are not trivially explained by Kerr nonlinearityalone: (1) progressive divergence of optimal control phase between trajectories, (2) trajectorydependentmodulation visibility with a crossover between low and high amplitudes, and (3) apparentsuppression of phase modulation for the Figure-8 trajectory at amplitude A = 0.8 whilethe Rectangular trajectory retains clear sinusoidal modulation. Observations 1 and 2 are robustand statistically clear. Observation 3 is suggestive but requires further statistical validation withincreased shot counts. These results are interpreted in the context of amplitude-phase coupling(APC) with measured coupling strength κ ≈ 0.9 rad/a u , previously reported in Ref. [1]. We proposefour follow-up experiments—including an isenergetic control test—to discriminate geometriccontributions from dynamic nonlinearities.