Probing Excited-State Dynamics of Transmon Ionization
or
Can we see how the transmon becomes “ionized”?
Tldr; Yes! In a collaboration between the Blok-Lab at the University of Rochester and the Blais group at University of Sherbrooke, we measured the dynamics of multi-photon resonances that occur during high-power readout of a superconducting qubit using a high dimensional transmon qudit.
Why is this important?
Good readout is essential for quantum computers, yet strong measurement drives can trigger unwanted transitions into highly excited states. While observed by superconducting circuit labs worldwide, the process was initially poorly understood: excited states and rotating terms are commonly neglected in models for transmon readout.
We now know that excitations can arise from multi-photon resonances owing to Stark Shifts induced by high-power (readout) drives. This phenomenon is sometimes called transmon ionization as it resembles multiphoton ionization in atoms and molecules and is also known as MIST (for Measurement-Induced State Transition) or DUST (Drive-induced Unwanted State Transitions).
Until now, transmon ionization has been observed indirectly, by measuring leakage out of the qubit subspace. Our work measures the dynamics of this process directly using a transmon qudit, providing new insight into both the mechanism and how to prevent it.
Expert take aways:
In our paper, now published in Physical Review X (PRX), we use a ten level transmon qudit to:
– Pinpoint the exact excited state of the transmon after ionization
– Verify that this process is a Landau-Zener-type transition
– Use pulse shaping to control adiabaticity and reduce the fraction of population transferred
One interesting find is that the critical photon number is highly sensitive to the offset-charge especially in typical (Ej/Ec ~ 55) transmons. You likely seen this as readout fidelity fluctuations on your devices. This sensitivity allowed us to measure the instantaneous offset charge by measuring the critical photon number. Not only does this provide an extremely accurate test of the theoretical model but also suggest improvements in readout qubits in surface code architectures by reducing their charge sensitivity (e.g. with higher Ej/Ec).
Bottom line: transmon ionization is a coherent Landau-Zener process that we can measure with a transmon qudit and this knowledge can help inform theory and mitigation strategies
Thank you to collaborators Philippe Gigon, Ben D’Anjou, and Alexandre Blais and special shoutout to first-author Zihao Wang who drove this project and recently graduated from our lab. Congrats Zihao!
Full article in Physical Review X:
https://journals.aps.org/prx/abstract/10.1103/8tdv-hgmb
DOI: 10.1103/8tdv-hgmb
Image credit: Anny De Lafont