Self-assembled quantum dots are very attractive as the building blocks for quantum light sources and udspin qubits. For instance, a single quantum dot is a robust, fast, narrow-linewidth source of single udphotons, features not shared by any other emitter. A spin qubit is implemented by a single electron udor hole confined to a quantum dot. Fundamental quantum mechanics have been explored in udexperiments with single quantum dots and spectacular success has been achieved. Future uddevelopments however demand an enhanced quantum coherence. For instance, indistinguishable udsingle photons and coherent spins are required to implement a quantum repeater. For quantum-dot-udbased single photon sources, the linewidths are in the best case typically a factor of two larger than udthe transform limit in which the linewidth is determined only by the radiative decay time. Photons udgenerated far apart in the time domain are therefore not indistinguishable. Spin coherence is udpresently limited to microsecond timescales. Improving the quantum coherence involves dealing udwith the noise inherent to the device. Charge noise results in a fluctuating electric field, spin noise in uda fluctuating magnetic field at the location of the qubit, and both can lead to dephasing and uddecoherence of optical and spin states. Here, the noise and strategies to circumvent its deleterious udeffects are explored in order to optimize the performance of solid-state quantum systems. udThis thesis is divided into five parts. The first chapter describes in detail the main experimental tool udto explore noise in the solid-state: resonance fluorescence from single quantum dots. A polarization-udbased dark-field microscope is realized allowing background-free resonance fluorescence detection udwhile operating in a set-and-forget mode. udChapter 2 investigates charge fluctuations in a semiconductor. The origin of the main source of udcharge noise in the commonly used optical field-effect devices is pinned down: charge fluctuations at uda GaAs/AlAs interface nearby the quantum dots. These defects are moved further away from the udquantum dots in an improved sample design resulting in close-to-transform limited optical udlinewidths. udEven with the improved heterostructures, the transform limit is not reached. Noise spectra of both udcharge noise and spin noise provide powerful insights into the noise inherent to the semiconductor, uddiscussed in chapter 3. A time trace of the resonance fluorescence from a single quantum dot is udtranslated into a noise spectrum. A crucial difference in their optical signatures allows the nature of udthe noise, charge or spin, to be identified. The charge noise is centred at low frequencies, the spin udnoise is centred at high frequencies. This technique is able to reveal the entire spectrum of the spin udnoise. The combined noise falls rapidly with frequency becoming insignificant above 50 kHz for the udquantum dot optical transition as signalled by transform-limited linewidths. udThe low frequency noise, charge noise, results in considerable noise in the emission frequency of the udsingle photons. This problem is solved in chapter 4 with a dynamic feedback technique that locks the udquantum emission frequency to a reference. The charge noise and its deleterious effects are highly udreduced. A frequency-stabilized source of single photons in the solid-state is realized. udThe low frequency linewidths are in the best case typically a factor of two larger than the transform udlimit. It is shown in chapter 5 that spin noise in the host material is the dominant exciton dephasing udmechanism. This applies to both the neutral and charged excitons. For the neutral exciton, the spin udnoise increases with increasing excitation power. Conversely for the charged exciton, spin noise uddecreases with increasing excitation power. This effect is exploited to demonstrate transform-udlimited linewidths for the charged exciton even when the measurement is performed very slowly. ud
展开▼
