Fig. 8: Preparation and lifetime of decay-protected fluxon states.
a Measured spectrum of the lowest energy plasmon mode versus magnetic flux of device C. The top panel was measured without fluxon preparation and shows the non-periodic spectrum similar to Fig. 5 at 7mK for device B. The middle panel was measured right after a 1 ms long resonator pump tone with \(\bar{n}=100-2000\). Here the periodicity is restored, and the fictitious phase particle is locked to the global minimum of the potential landscape. Near the flux frustration point, the ground state initialization is not perfect as indicated by two visible lines (black arrows). The spectrum in the bottom panel shows the same measurement with only \(\bar{n}=70-80\) resulting in an excited fluxon state, i.e., the fictitious phase particle is generally found in the lowest energy neighboring well rather than the global minimum. b Measurement sequence used for the data in panel (c). First, we actively determine the fluxon state (yellow) immediately after the excitation pulse (green) and repeat the sequence until the fluxon excitation is successful. Then, we continue to measure the state every Δt = 30 s until it has decayed via a quantum tunneling event. The fluxon readout (yellow) is based on the fluxon-dependent plasmon frequency away from half flux. Applying a plasmon π pulse (blue) is only successful if the plasmon level is on resonance, i.e., in a certain flux well. The resulting plasmon excitation is measured via the standard dispersive readout for 15 μs ≈ T1 (purple) in the single photon limit to avoid photon-induced fluxon tunneling and repeated 4000 times for high SNR. c Fluxon decay from the excited to the ground state measured for three different flux bias configurations (dots) and exponential fits (lines). The data is based on ≈80 decay events each, conducted over up to 1 week of raw measurement time for the longest decay time.
