Integrating tuning and readout techniques for scaling spin qubits

Thesis

Spin qubits are one of the more realistic propositions for the basis of an operational quantum computer. However, scaling these systems is proving to be challenging for both the academic and industrial communities, in part due to the large parameter space that a spin qubit device lives in. To get what is effectively a transistor to act as a qubit, the experimentalist has to adjust the gate voltages that modify the spin’s confinement potential to find the exact combination that results in a well-defined two-state quantum system. Even after bringing a qubit into existence, its performance is heavily parameter-dependent, requiring further optimisation of experimental control knobs.

Manually finding these parameters is time-consuming for even a single qubit. Part of the roadmap for scaling these devices past the proof-of-concept stage is automating these procedures, a goal which is slowly being tackled in the literature. This thesis contributes a series of algorithms towards the eventual goal of a fully-automated qubit tune-up process, specifically using measurement techniques that are used for qubit readout. I begin by presenting an algorithm that automatically tunes an uncalibrated quantum dot array into the double dot regime using only radio-frequency reflectometry measurements. Next, I present a package of algorithms that compensate for the effect of crosstalk between a quantum dot array and the quantum dot charge sensor used to measure its state. Third, I present a series of techniques and measurements for use in the automated discovery of qubits within a double quantum dot regime. Finally, I present a method that extracts qubit readout metrics from repeated single-shot measurements, such that any one of them can be optimised using conventional optimisation routines.

In combination, these results contribute towards the ultimate goal of a scalable, fully-autonomous tuning pipeline for heterostructure quantum dot spin qubits.

Authors: Hickie, J

• DOI: 10.5287/ora-dpdyqd4ej
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: spin qubit; quantum dot; machine learning
• Subjects: Quantum dots; Machine learning; Quantum computing