The hole spin in a semiconductor quantum dot

Extensive research on semiconductor quantum dots (QDs) has been a hot topic in the semiconductor community over the past 20 to 30 years and is still ongoing. In the late 1980s the term "quantum dot" was introduced to describe a semiconductor nano-structure. Some of the motivating prospects driving the research are low-threshold QD lasers, single dots for medical markers, lighting technologies for TVs or single spins for spintronic applications, e.g. quantum information processing. The size and the structure of a QD can vary from a few nanometres in colloidal dots (also known as nanocrystals) to a few hundred nanometres in lithographically defined electrostatic devices. The material components and the fabrication methods can differ a lot between the individual types of QDs. One feature all different kinds of QDs have in common is the restriction of the carrier motion in all three dimensions, which is induced by confinement. That property is the origin of the name zero-dimensional ("0D") structures. A second term often used describes the QD as an "artificial atom". The strong confinement establishes discrete energy states for the localized single carriers inside the QD, which resembles the properties of carriers in atoms.
The QDs investigated in this thesis are self-assembled InAs QDs in a semiconductor heterostructure, laying the focus on the confined positive charged carriers, the holes. The spin properties of the individual quantum states are characterized with advanced optical spectroscopy techniques.
The following thesis is split into four parts. The first part motivates the search for coherent single hole spins and explains how to get from a bulk semiconductor to a single spin. After a short introduction of semiconductor self assembled quantum dots, their optical properties and bandstructure, the requirements to perform single spin physics are described. The advantage to choose the hole spin for a spin qubit instead of the electron spin, regarding their decoherence properties is discussed.
The second section of the introduction covers the experimental techniques and improvements to current systems paving the way to a highly coherent spin qubit via the hole spin and high quality data. The new device structure as well as the sophisticated technique of resonance fluorescence detection are explained here. A description of the laser frequency locking mechanism and a power stabilization concludes the chapter.
In the second part the first experiments of this thesis on coherent hole spins are presented. With the spectroscopic measurement method of coherent population trapping (CPT) long decoherence times are achieved. Charge noise is determined as a hole spin dephasing mechanism. Despite the very promising results the experiment suffers from two disadvantages. First the measurement method via resonant absorption spectroscopy in combination with the unstable measurements conditions offers a very poor signal to noise ratio. Secondly the low frequency charge fluctuations, inherent in the sample, promote dephasing and induce shifts in the CPT resonance position from scan to scan.
The third part covers different approaches to address the noise issue of part two.
The optical linewidth and the noise are closely related in solid state emitters: The linewidth broadening is caused by spin and charge noise in the quantum device.
First, low frequency charge fluctuations are reduced by a feedback scheme, which stabilizes the emission frequency of the quantum dot to a stabilized reference. The feedback loop minimizes the fluctuations in the emission frequency, even over several hours, and eliminates the charge noise in the quantum dot to a large extent. This method realises a frequency stabilized source of single photons in the solid-state.
The next chapter introduces a new sample design in order to reduce spectral fluctuations. The n-i-p device growth sequence is inverted, which prevents the usual contamination of the QDs by the C-doping. The characteristics of the ultra clean p-doped samples are narrow linewidths in combination with high count rates. The "transform-limit" is reached with a fast scanning method. In the sample a voltage dependent blinking behaviour of the positively charged exciton is discovered.
The story of low-noise samples and noise control continues in the next chapter. Transform-limited linewidth of the neutral and the negatively charged exciton are presented. For the neutral exciton this is even true for slow measurements lasting several seconds. For already low-noise structures the residual linewidth broadening is only caused by the nuclear spin noise. A two colour experiment provides control over the nuclear spins, which dominate the exciton dephasing.
In the last part the interaction of the hole spin with its environment is investigated.
The hole spin states interact in an in-plane magnetic field with an external electric field. The interactions result in a tunable hole g-factor, showing a linear dependency over a large electric field range. In contrast the electron g-factor is not influenced by the electric field at all. Theory reproduces the hole g-factor dependence, which arises from a soft hole confining potential, an In concentration gradient and a strong dependence of the material parameters on the In concentration.
The last chapter demonstrates the anisotropic behaviour of the hyperfine interaction between nuclear spins and the hole spin. In the experiment, again with the measurement method of coherent population trapping, a low-noise sample and resonance fluorescence spectroscopy are combined. The resulting high signal to noise ratio and the ultra narrow CPT dip enable the measurement of very precise values for the energy splitting of the hole spin states. This is leading to the main result: a minimal hole hyperfine interaction in an in-plane magnetic field, proofing a decoupling from the hole spin and the nuclear spins.