Time-resolved noise characterisation tool to track fluctuating noise effects in superconducting qubits
Abhishek Agarwal, Lachlan Lindoy, Deep Lall, Yannic Rath, Ivan Rungger
Abstract
Superconducting qubits have seen rapid increases in their coherence in the last few decades. However, instability in qubit parameters and their noise environment remains as a major challenge that limits the device performance. Qubit metrology to collectively measure and characterise different sources of noise and fluctuations is essential for accurately benchmarking devices but can be challenging, and typically different noise sources are characterised independently. In order to characterise the combined noise, we develop a tool that uses few-measurement data to detect and diagnose noise fluctuations, as well as a time series segmentation tool to further disambiguate different sources of fluctuations. We demonstrate the tool by computing time resolved properties of individual superconducting qubits. We find large changes of the dominating noise contributions, such as qubit frequency fluctuations, over both long time-scales of hours and days, and also over very short milli-seconds time-scales. We show that this information can be used alongside accurate noise models to learn physical properties of the coupled noise sources as well as benchmark device stability.