The electronic structure of indium arsenidegallium arsenide self-assembled quantum dots in a high magnetic field

The electronic energy levels of dome-shape InAs self-assembled quantum dots (SAQD) grown by the Stranski-Krastanow mode on GaAs substrates are similar to those obtained from a two-dimensional harmonic-oscillator. A simple selection rule allows transitions only that preserve angular momentum, depicted with atomic-like orbital labels s, p, d, f, etc. This electronic structure was examined with photoluminescence (PL) and photoluminescence excitation (PLE) techniques. As well, in magnetic fields up to 28 Tesla applied parallel to the growth direction, SAQD energy-level degeneracies were lifted. The number of branches observed is correlated to the angular momentum. The ground state (GS) level at zero angular momentum is shifted quadratically under the magnetic field and the behavior could be explained with the Fock-Darwin (F-D) spectral model. The effect of annealing at temperatures from 825°C to 900°C in 25°C steps on the SAQD electronic structure was also examined with the PL technique combined with an applied magnetic field in the Faraday configuration. The PL lines were similar to the F-D spectral lines with their degeneracy lifted by the applied magnetic field. These lines exhibited ten (anti-)crossings: three each at 10 T and 28 T, four at 18 T, while the inter-level spacing and the FWHM were reduced with increasing annealing temperature. Thus an increase in the observed (anti-)crossings resulted for the higher anneal temperatures. The in-plane excitonic reduced-mass was inferred from the systematic splitting of the PL p-branches in a magnetic field. The reduced-mass for all the annealed QD samples was about 0.066 m0 +/- 0.012m0 which decreased slightly with anneal temperature. An 8-band k*p model predicted a similar reduced-mass at low alloying of gallium, but an incorrect trend was observed as the alloying increased with annealing temperature. Unrealistic reduced-masses at 50 percent gallium content were reached. This discrepancy is explained assuming the F-D model is a single (independent) bulk particle picture neglecting many-body effects, and also the k*p model assumes identical disks before and after annealing. The SAQDs were in fact inhomogeneous shallow domes whose height is reduced with annealing temperatures. It is an attempt to reduce the effect of many-body interactions such as exchange, configuration and screened coulomb interactions dominant in the PL technique, the PLE technique was use. In this technique, a single level in a collection or 'ensemble' of dots is excited with tuned laser-light and only the Coulomb interactions are assumed to be important. The PLE peaks were found to be blue-shifted relative to PL peaks. Furthermore, under the influence of a magnetic field, two PLE peaks were observed that corresponded to the p and d energy states. However, three 'd' lines were expected and is hypothesized that one of the d lines remained degenerate. Moreover, the carrier dynamics observed in PLE spectra are much more difficult to interpret than that of the PL spectra. Applying the same method, the analysis of the p-branch peaks suggested an in-plane reduced-mass of ∼0.084m0 +/- 0.002m0, higher than obtained from PL measurement. Since the effective mass is normally associated with the mobility of the carriers, this would imply that the excitons in the PLE measurement are less mobile than in PL. This is despite the reduced many-body effects, suggesting that some extra interactions in the PL excitation may actually enhance the carrier mobility. Given the current interest in devices such as QD infrared photo-detectors and the necessary controls on the number of charge carriers in these devices, a single-layer and 25-layer SAQD samples with doping in the top cap layer were compared to un-doped sample using PLE at various detection energies. No absorption signatures appeared for the doped single layer, whereas they were recovered in the 25-layer doped sample. Evidently either dopants or injected carriers diffused into the QD layers beneath the cap. This diffusion and its influence is expected to be decreasing with depth. Finally, the number of injected charge-carriers in doped GaAs barriers interleaving 50 SAQD layers was studied in order to understand the influence on their electronic structure. From the relation between the dot density and the dopant dose, two to twenty-two charge carriers were estimated to be present in the barriers of each QD. The PLE results indicated that as this number was increased, direct radiative recombination from the higher levels decreased. In addition to Auger scattering and multi-phonon scattering, the enhanced scattering by the dopants impurities appears to add further decay channels toward the lower-energy recombination. This suggests that the PLE technique is sensitive for characterizing the doping effects in SAQD materials. Some fundamental questions regarding the optical and electronic properties of InAs/GaAs SAQD have been answered in this dissertation and the results can be used to support the future development of opto-electronic devices at the nano-scale level.