Multi-level charge fluctuations in a Si/SiGe double quantum dot device
ORAL
Abstract
A candidate explanation for charge noise in semiconductor quantum dot devices is the fluctuation
of an ensemble of discrete two-level charge noise sources. A large variety of such fluctuators results
in Gaussian 1/f noise; however, significant deviation from Gaussianity is routinely observed that
may inform models of the underlying contributors to charge noise. Here we analyze multi-level
charge fluctuations present in a Si/SiGe double quantum dot device over a range of device voltage
and temperature values to probe the response of the charge fluctuations to changing operating con-
ditions. We perform algorithmically assisted drift detection and change point detection to detrend
the data and remove a slow fluctuator component, as a preprocessing step. We then perform model
comparison on the residual, post-processed time series between different n-level fluctuator (nLF)
factorial hidden Markov models (FHMMs) (2x2LF, 3LF, and 4LF), finding that although in most
voltage configurations the 2x2LF model is preferred, there is a particular region of device space
where the 4LF model outperforms the other models, indicating an emergent conditional rate de-
pendence between the constituent 2LFs. Finally, we fit a phenomenological, detailed balance model
to the extracted 2x2LF rate data, yielding lever arm estimates in the range of −2μeV/mV up to
4μeV/mV between the two 2LFs and nearby gate electrodes.
of an ensemble of discrete two-level charge noise sources. A large variety of such fluctuators results
in Gaussian 1/f noise; however, significant deviation from Gaussianity is routinely observed that
may inform models of the underlying contributors to charge noise. Here we analyze multi-level
charge fluctuations present in a Si/SiGe double quantum dot device over a range of device voltage
and temperature values to probe the response of the charge fluctuations to changing operating con-
ditions. We perform algorithmically assisted drift detection and change point detection to detrend
the data and remove a slow fluctuator component, as a preprocessing step. We then perform model
comparison on the residual, post-processed time series between different n-level fluctuator (nLF)
factorial hidden Markov models (FHMMs) (2x2LF, 3LF, and 4LF), finding that although in most
voltage configurations the 2x2LF model is preferred, there is a particular region of device space
where the 4LF model outperforms the other models, indicating an emergent conditional rate de-
pendence between the constituent 2LFs. Finally, we fit a phenomenological, detailed balance model
to the extracted 2x2LF rate data, yielding lever arm estimates in the range of −2μeV/mV up to
4μeV/mV between the two 2LFs and nearby gate electrodes.
*SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525
–
Presenters
-
Dylan Albrecht
- Sandia National Laboratories