Spin qubits in silicon fin field-effect transistors

Quantum computing holds the promise of solving certain tasks faster than any classical computer. The fundamental building block of a quantum computer is the qubit, which can be realised in various different systems including trapped ions, superconducting circuits or semiconductor quantum dots. The latter platform is used to host spin qubits, where the spin state of a single electron or hole is used to encode quantum information. Spin qubits in silicon offer distinct advantages over other materials in terms of scalability by leveraging well-established complementary-metal-oxide-semiconductor technology from the classical computer industry.
While spin qubits typically operate at milikelvin temperatures, classical control electronics are maintained at room temperature, leading to a wiring bottleneck and challenges in cooling, which hinder scalability. Silicon, in this regard, presents a solution: by elevating the spin qubit operation temperature above 1$\,$Kelvin, where cooling power is dramatically increased, it becomes feasible to integrate classical control electronics and quantum hardware in the same package, paving the way for a truly scalable unit cell. Recent advancements have demonstrated the operation of electron spin qubits in silicon at elevated temperatures. Hole spins, an alternative to electron spins, possess unique advantages such as fast, all-electric spin control via a strong intrinsic spin-orbit interaction, absence of valleys and sweet spots for charge and nuclear spin noise. However, despite their potential, hole spin qubits in silicon still lag behind their electron counterparts in crucial aspects, including operation at elevated temperatures and the demonstration of entangling two-qubit gates, which are essential for universal quantum computing.
This thesis explores the potential of hole spin qubits hosted in silicon fin field-effect transistors, one of the industry’s standard transistors. The device design is optimised for quantum applications through the introduction of a self-aligned gate layer and its ability to host quantum dots is demonstrated. The devices are used to define hole spin qubits, which are driven all-electrically via electric-dipole spin resonance, and their spin state is read out using Pauli spin blockade. A driving speed exceeding 100$\,$MHz and a qubit coherence time $T^∗_2$ greater than 400$\,$ns is observed and a single-qubit gate fidelity of ∼99% is reported at 1.5$\,$K. The robustness of these hole spin qubits against further increase of temperature is demonstrated by operating the qubit even above 4$\,$K, albeit with a reduction of coherence and a whitening of the noise spectrum. Furthermore, the anisotropy of hole spins in silicon is investigated by measuring and modelling the magnetic field orientation-dependence of the effective g-factor and the qubit driving speed and coherence, showcasing sweet spots in the qubit quality factor. Disentangling the influence of two driving mechanisms enables control over the type of driving by applying the microwave driving tone to different gate electrodes. Additionally, a formalism is developed to model the anisotropic nature of the exchange interaction between two neighbouring qubits, enabling the extraction of the full exchange matrix. The model for anisotropic exchange allows sweet spots for fast, high-fidelity two-qubit gates to be predicted and in a close-to-optimal configuration a CROT operation is demonstrated in just 24$\,$ns. Furthermore, a conceptually different drive to the standard Rabi drive, the phase drive, is explored. By incorporating a second far-detuned driving tone, resonant Rabi oscillations can be suppressed and reappear at tunable sidebands, offering opportunities for global driving schemes and noise mitigation. In summary, the interesting physics of hole spins in silicon is investigated, and key technological milestones for hole spin qubits in silicon fin field-effect transistors are demonstrated throughout this work.