Spin Qubit Studies Using A 2x2 Quantum Dot Array In Si/SiGe

Abstract

Classical computers have long been the cornerstone of information processing, yet their capabilities are constrained by the limits of the classical laws of physics. Quantum mechanics offers a new spin on information processing, potentially providing immense speed-ups for some specialized problems. There are many approaches to building such a quantum computer, that leverages quantum mechanical principles. The most popular approach uses superconducting circuits to implement a qubit. This thesis, however, builds on the advances of the semiconductor industry. The miniaturisation of electronic devices in the last decades has enabled the fabrication of gate defined quantum dots. Such a quantum dot allows the isolation of a single charged particle that can be used to implement a qubit. More specifically this thesis employs electrons in Si/SiGe heterostructures. While most implementations so far rely on linear chains of quantum dots, scaling in a second dimension is crucial for building larger systems.

This thesis explores a 2x2 array as a proof of concept for a 2D array. This small-scale device demonstrates that charge-related properties, such as gate pitches and tunnel coupling control, remain similar when transitioning from one to two dimensions. We show that existing qubit control strategies using electric-dipole spin resonance (EDSR) and micromagnets can also be adopted for 2D arrays as long as the second dimension remains small. In larger 2D arrays, the magnetic field gradients achievable by micromagnets no longer meet the requirements for EDSR control. Additionally, the application of microwave bursts causes an unintended spin resonance shift that complicates qubit manipulation.

To address these challenges, this thesis also explores baseband control of single-spin qubits. In this scheme, single-qubit rotations are implemented using hopping gates, which use tilted quantisation axes in neighbouring quantum dots. In Si/SiGe this tilt is achieved using the strong spatial variation of the stray field of a nearly demagnetized micromagnet. Building on this, a nanomagnet-based architecture is proposed, integrating localized nanomagnets to provide magnetic field gradients for spin manipulation. This approach circumvents EDSR limitations, offering a more scalable pathway for 2D quantum dot arrays and advancing spin qubit technologies toward large-scale quantum computing.