Extending quantum dot arrays into three dimensions

Abstract

Gate defined germanium spin qubits appear to demonstrate many attractive properties for building block of a future quantum computer. In this thesis we explore beyond the necessities, demonstrating three key results. In chapter4, a resonating valence bond is realised on a 2×2 qubit array, utilising precise control over the exchange interaction. The rich intrinsic physics provided by the spin orbit coupling gives many possibilities for controlling Ge spin qubits. While it can enable all electrical control using microwaves, we demonstrate that fast, high fidelity, and long coherence all electrical control can even be achieved with baseband pulses. Through the development of hopping based spin qubits, we exploit the g-tensor misalignment between quantum dots in order to produce quantum gates by diabatically and coherently shuttling spins between quantum dots. In chapter5 we show that the g-tensor anisotropy can be exploited for low power, baseband qubit control, with fast gate times and high fidelity. We obtain high single qubit gate fidelities, and measure two-qubit gate fidelities of FCZ = 99.3% using IRB, and FCZ = 98.1% using GST. It is certainly a step in the right direction; we don’t require strong magnetic fields, or microwave pulses, however this approach does depend on precise timing control, and more work is needed to determine how gates in a scaled computer would be timed.

In chapters 6, 7, 8, the toolbox of spin qubits is expanded once more and we show that spin qubits need not be confined to planar architectures, but can be scaled in the out of plane direction. Further to this, they support coherent spin physics. We establish methods to tune multi-well systems and realise a three-dimensional dot array. Bilayers promise higher connectivity and future research may indicate whether they can enable hopping-based qubits where the quantization axis can be consistently controlled by engineering the strain of the quantum wells. The extra tuning complexity introduced by additional layers may offset these gains; whether it would stand to benefit the community to adopt multilayer heterostructures remains an open question. When considering wiring bottlenecks this may reveal an efficient approach to circumventing the problem, and to achieve connectivities which can efficiently simulate natural systems it may be necessary to have three-dimensional quantum dot arrays.