|Institution:||University of New South Wales|
|Department:||Electrical Engineering & Telecommunications|
|Keywords:||Quantum computation; Quantum dot; Silicon; Electron spin; Qubit; Single-electron transistor; Metal-oxide-semiconductor; Charge sensor; Spin readout|
|Full text PDF:||http://handle.unsw.edu.au/1959.4/53732|
This thesis focuses on the development and demonstration of silicon metal-oxide-semiconductor (MOS) quantum dots (QDs) for spin-based quantum information processing. Firstly, by measuring the transport current through a MOS quantum dot, its multi-electron spin state was determined as the electron occupancy was reduced from twenty-seven electrons down to the single-electron limit. In particular, kinks observed in the electron addition energy as a function of magnetic field demonstrated that a valley-orbit excited state existed 100 μeV above the ground state. Secondly, by incorporating a silicon single-electron transistor (SET) charge sensor next to a quantum dot, the electron occupancy of the dot was probed via the sensor output signal. By applying a digitally-controlled dynamic feedback loop to the charge sensor, robust detection of the QD charge state was achieved, even in the presence of charge drifts and random charge upset events. Next, the excited states of a silicon MOS quantum dot were studied in detail. The electron occupancy and excited-state energy levels were detected using a SET charge sensor, with the aid of pulsed-voltage spectroscopy. The energy of the first orbital excited state was found to decrease rapidly as the electron occupancy increased from N = 1 to 4. By monitoring the sequential spin filling of the dot a valley splitting of ~230 μeV was extracted, which was found to be independent of electron number. Finally, by performing single-shot spin readout on a silicon MOS quantum dot, spin lifetimes were extracted for different electron occupancies and valley splitting configurations, with a maximum one-electron spin lifetime exceeding 2 seconds. We also demonstrated the ability to tune the valley splitting energy via electrostatic gate control, with a splitting that increased linearly with applied electric field over the range 0.3 - 0.8 meV. The spin relaxation rates were found to be highly dependent on the valley splitting energy, with a dramatic rate enhancement (or hot-spot) when the Zeeman and valley splittings coincided, a process that had not previously been anticipated for silicon quantum dots.