Inference of transport phenomena in quantum devices

Thesis

This thesis is concerned with charge transport in electrostatically defined quantum dot devices. Such devices display a wide range of transport phenomena in both open and closed configurations. The transport regime can be tuned experimentally by controlling the voltages applied to gate electrodes, but the precise electrostatic landscape which determines the transport regime is unknown. This uncertainty is given by variations in device fabrication, material defects, and sources of electrostatic disorder. The research chapters of this thesis consider a range of transport regimes in quantum dot devices, and infer properties of the device using both experimental and theoretical techniques. The first research chapter considers the detection of single charge transport events through a double quantum dot. By fitting an open quantum systems model to the sub-attoampere currents measured, tunnel rates are inferred. The second results chapter considers an electrostatic simulation of a quantum dot device and how it can be accelerated using deep learning. This accelerated model is then used in the third results chapter, along with experimental measurements of the transport regime, to inform a Bayesian inference algorithm and produce a set of disorder potentials to narrow the gap between simulation and reality. The final results chapter develops a differentiable quantum master equation solver which is used for parameter estimation in a theoretical study of transport in single and double quantum dots.

Authors: Craig, D

• DOI: 10.5287/ora-nbx0jzppy
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: Bayesian inference; quantum transport; open quantum systems; machine learning; quantum dots
• Subjects: Materials; Physics; Algorithms; Quantum dots