This thesis presents a study of the electron-phonon interaction in an n-doped weakly coupled semiconductor superlattice (SL). Two experiments were performed which studied different aspects of this interaction. Firstly, a coherent phonon optics chip was designed. This was used in an experiment where a phonon beam was passed through the SL while an electrical bias was applied to it. The experiment provided a sensitive measurement of the effects caused by bias in the SL on the phonon beam. Secondly, a train of strain pulses was passed through the SL and the charge transferred in the device due to the strain was investigated. A coherent phonon optics chip was formed using a semiconductor superlattice as a transducer structure and a p-i-n photodiode as a coherent phonon detector on the opposite side of the substrate. The doped weakly coupled superlattice structure, which is the main subject of investigation in this thesis was grown between the transducer and detector structures. Optical access mesas were processed on both sides of the substrate to allow the application of bias to both the doped superlattice and the p-i-n structures. A photocurrent pump-probe experiment was then performed using a femtosecond laser to excite the transducer structure and activate the detection mechanism. The application of bias to the weakly coupled SL was found to cause a small attenuation to the 378 GHz phonon beam passing through it. An investigation of the possible causes of this attenuation ruled out several trivial explanations, suggesting that it was caused by the interaction between electrons and phonons in the structure. The active control of phonon amplitude by electrical means has not previously been demonstrated and may offer exciting new prospectives for phonon devices and experiments. The coherent phonon optics technique was shown to be very sensitive and it will be a useful technique to increase our understanding of future acousto-electric devices. The electrical signal that acoustic excitation caused in the SL device was investigated using a pulse shaping technique in combination with an amplified femtosecond laser. A Fabry-Perot cavity was used in the laser path to create a train of equally spaced laser pulses with an adjustable pulse spacing. Focusing these pulses on an aluminium film transducer creates a train of equally spaced acoustic pulses simulating a monochromatic acoustic wave packet. The SL was processed and electrically contacted so that the charge transferred through it due to the acoustic pulse train could be monitored using a 12.5 GHz-bandwidth digital oscilloscope. The variation in charge transfer seen as a function of the DC bias applied to the device and as a function of the total energy of the acoustic pulse train was investigated. The behavior was compared to a theoretical model developed in the style of previous theories of electrical conversion in SLs excited by electromagnetic waves. The dependencies of the charge transfer on the bias and energy of the pulse train were well reproduced in the theory. The theory predicted that magnitude of the signal in the superlattice was independent of the frequency of the acoustic pulse train. This was verified by measuring the frequency dependence of the signal seen for a variety of transducer films. The frequency dependencies seen were well explained through simulations presuming the device response was independent of train frequency. This confirms the predictions of the theory. Both the experiments detailed in this thesis have helped increase our understanding of the nature of electron-phonon interactions in superlattices. It is hoped that a fuller understanding of these interactions may be instrumental in the creation of exciting new acousto-electrical devices.

4 p. / The adoption of solar photovoltaic (PV) energy systems to serve as an energy source for residential, commercial and agriculture applications is growing. Early use of solar PV energy as an alternative energy source to fossil fuels became popular in the 1970’s during the rise of the environmental movement. The cost of solar power in 1977 was $76.00 per watt. A combination of factors including public awareness, demand for solar, availability of product and service, and improving technology has dropped the cost per solar watt. In 2015, the cost of solar power was $0.613 per watt (Shahan, 2014). Energy rebates offered by local, state, and federal agencies has made the adoption of solar energy more affordable.

The research work conducted and presented forward in this document is the evaluation of real time values obtained using three recording devices at two independent locations and implementing them as recorder devices in Eskom’s power system. The research work conducted was presented at an IEEE International Conference (ICIT2013) and Appendix A shows the accepted paper presented. A derived model within a simulation software package known as DIgSILENT PowerFactory is created and Electromagnetic Transient (EMT) studies are performed and then compared to the real time values obtained using the OMICRON CMC 356’s. Transformers are normally energised via a circuit breaker which is controlled by an auxiliary closing contact. By applying system voltage at a random instant in time on the transformer windings may result in a large transient magnetizing inrush current which causes high orders of 2nd harmonic currents to flow under no load conditions. A philosophy known to mitigate these currents is to energise the transformer by controlling each individual phase 120 degrees apart with the first pole closing at the peak on the voltage waveform. Transients produced due to 500MVA transformers been introduced into the power system at a certain space in time can cause nuisance tripping’s at the particular location where the respective transformer is energised. OMICRON EnerLyzer is the software tool used for the Comtrade recordings at both locations. Four independent case studies are generated within EnerLyzer software and the relevant Comtrade files are extracted for the four independent case studies relative to Transformer1 and Transformer2 switching’s. TOP software, which is a mathematical tool used to analyse Comtrade files, is then used to analyse and investigate the four case studies. Results from DIgSILENT PowerFactory are then generated according to the derived model. The results extracted depict three scenarios, indicating a power system that is weak, strong and specifically a power system that correlates to the actual tripping of a Static VAr Compensator (SVC). The results are all formulated and then evaluated in order to produce a conclusion and bring forward recommendations to Eskom in order to effectively ensure the Dedisa/Grassridge power system is reliable once again.

This thesis discusses experiments that have been performed to investigate ultrafast acoustoelectric effects in semiconductor devices. Current commonly employed techniques to generate ultrafast acoustic pulses and detect them with spectral resolution require a powerful pulsed laser system that is bulky, expensive and complicated. If the acoustic pulses could instead be generated and detected by electrical methods, picosecond acoustic techniques could become more readily available as a tool for other users. This thesis focusses on the electrical detection of acoustic pulses with spectral resolution. In many of the key experiments described in this thesis a picosecond strain pulse was generated optically on the opposite face of the sample to the semiconductor device of interest. The strain was generated either in a thin Al film thermally deposited on the sample surface, or directly in the GaAs substrate. Acoustic phonons generated by this method propagated across the substrate to the device. Transient voltages across the semiconductor device caused by the incident phonons were detected using a high frequency real-time oscilloscope. The first evidence of heterodyne mixing of coherent acoustic phonons with microwaves was obtained, for frequencies up to about 100 GHz. First, it was confirmed that Schottky diodes can produce a fast transient voltage in response to an incident acoustic wavepacket. The detection process occurs at the semiconductor-metal interface, and is due to the deformation potential. Bow-tie antenna fabricated directly onto the GaAs substrate proved to be ineffective at coupling microwaves from free space to the Schottky diode. A waveguide-coupled beam-lead Schottky diode provided by e2v had a sufficient response to the incident microwaves to proceed with the mixing experiments. The microwave local oscillator signal was mixed with a tunable narrow frequency band acoustic signal that was produced using a Fabry-Perot etalon external to the laser cavity. The intermediate frequency components were in the range of 1-12 GHz, which could be detected on the oscilloscope. Mixing was performed using both the fundamental frequency acoustic wave and the second harmonic generated in the sample. Semiconductor superlattices were also investigated as electrical detectors for ultrafast acoustic pulses. In this case, the transient voltage measured across the device contained an unexpected contribution in the form of a peak with a width of approximately 2 ns. This signal is too slow to be caused by a strain pulse and too fast for a heat pulse. It is proposed that this peak is caused by long-lived phonon modes from the centre of the mini-Brillouin zone being confined in the superlattice due to Bragg reflections. The peak caused by confined phonons and the two peaks caused by heat pulses also present in the detected signal were investigated for a range of experimental conditions. This allowed comparisons to be made to previous works. A similar superlattice structure had a very different response to the incident acoustic wavepacket. The polarity of the transient voltage detected was inverted and there was no evidence of an electronic response to the confined phonon modes, which would have been present in both samples. It is proposed that the barriers of the NU1727 superlattice sample are thicker than expected, and this strongly affects the electron transport through the structure. This thesis shows that semiconductor devices can be suitable for the electrical detection of ultrafast strain pulses. For this technique to reach its full potential, it is also necessary to be able to generate these strain pulses electrically. A step recovery diode has been considered for this purpose as part of the suggested future work.

The aim of this project is the conception, implementation, and application of a simulation tool for the accurate modeling of electromagnetic fields within inhomogeneous materials with complex shapes and the propagation of the resulting fields in the surrounding environment. There are many methods that can be used to model the scattering of an electromagnetic field, however one of the most promising for hybridisation is the Boundary Element Method (BEM), which is a surface technique, and the Unstructured Transmission Line Modeling (UTLM) method, which is a volume technique. The former allows accurate description of the scatterer's boundary and the field's radiation characteristics, but cannot model scattering by materials characterized by a non-uniform refraction index. The latter, on the contrary, can model a very broad range of materials, but is less accurate, since it has to rely on approximate absorbing boundary conditions. A method resulting in the hybridisation of BEM and UTLM can be used to construct a tool that takes into account both the interaction with non-uniform tissue and propagation in its environment. The project aims to describe in detail the implementation of the novel method, and deploy it in a heterogeneous distributed computing environment.

No description available.

The magnetic behaviour (magnetization curves, Meissner effect, hysteresis losses, remanent flux and flux line cutting) for high $T\sb{c}$ samples of Nd and $YBa\sb{2}Cu\sb{3}O\sb{7-x}$ at 77 K, semi-reversible PbIn and hysteretic VTi at 4.2 K have been investigated and compared. The magnetic behaviour of the high $T\sb{c}$ samples and of the PbIn are remarkably similar but contrasts dramatically with the phenomena observed in the VTi. A hump structure appears in the low field region of the initial magnetization of the high $T\sb{c}$ samples. This phenomenon occurs because the sintered samples consist of a compact agglomeration of small irregularly shaped grains which are electrically coupled before the hump structure but are isolated after this feature. We show that the appropriate calibration should be based on the diamagnetic response of the uncoupled grains. A large ratio of $j\sb{c\Vert}$ to $j\sb{c\perp}$ (the critical current densities $\Vert$ and $\perp$ to the flux line density) accounts for the observations on the VTi in the flux cutting regime whereas $j\sb{c\Vert} \approx j\sb{c\perp}$ is indicated by the behaviour of the high $T\sb{c}$ and PbIn samples.

We have developed a novel experimental approach for the study of the interaction of sheets of non parallel flux lines in hysteric type II superconductors. We continuously monitor the evolution of the components of the magnetic flux density $\Vert$ and $\perp$ to $H\sb{a}$ (i.e. $\langle B\sb{z}\rangle$ and $\langle B\sb{y}\rangle$) as $H\sb{a}$ is raised to various intensities, and then reduced to zero. In our investigation of a high $T\sb{c}(YBa\sb{2}Cu\sb{3}O\sb{7-x}$) ceramic, $H\sb{a}$ exceeded $H\sb{*}$, the first full penetration field. We applied the phenomenological Clem general critical state model to the analysis of our extensive observations. A computer program was developed to solve the four coupled differential equations of this theory with appropriate physical constraints for the situations prevailing in our experiment. This analysis provides detailed insight into the evolution of the intricate configurations of the magnetic flux density $\vec B$(x), the critical current density, $\vec J$(x), and electric field $\vec E$(x) patterns as the injected and trapped flux lines are made to interact, unpin, migrate and undergo flux cutting processes. The model is seen to generate the variety of complicated measured curves of $\langle B\sb{y}\rangle$ and $\langle B\sb{z}\rangle$ vs $H\sb{a}$ very satisfactorily. (Abstract shortened by UMI.)

The purpose of this study is to investigate the conduction electrons' interaction mechanisms over the wide temperature range in order to find possible deviations from the existing theories. The temperature dependence of the electrical resistivity of two scattering systems, the crystalline and the amorphous, in the range of 1.7--300 K is studied. In the pure crystalline metals, the nearly-free-electron model can qualitatively and quantitatively account for the temperature dependence of the electrical resistivity. In the low-temperature range, the electron-electron scattering is the dominant scattering mechanism. Above the Debye temperature, the phonon-electron scattering is the dominant scattering mechanism. The temperature dependence of the electrical resistivity for the amorphous metallic alloys can be described qualitatively by the Ziman model. The scattering mechanism of the conduction electrons of the samples can be categorized in the weak scattering limit regime. A small, but significant correction must be added to the Ziman model in order to account for the additional scattering contributions that exist in the wide temperature range. In the very-low-temperature range the electron-electron interaction effect is the major correction to the temperature dependence of the electrical resistivity. In the low-temperature range, the weak localization effect is the major correction to the temperature dependence of the electrical resistivity. In the high-temperature range, the temperature dependence of the electrical resistivity is linear with a small but significant correction. For some samples the magnetic contribution is the major correction, for others the multiphonon scattering is the major correction. For the resistivity minima at temperatures below 20 K could be due to the Kondo effect. A more plausible explanation, however, is the electron-electron interaction effect. The Ziman model with the proper corrections is found to be an adequate and successful model in accounting for the temperature dependence of the electrical resistivity of all of the studied amorphous metallic alloys.

The applicability of the weak-localization theory to highly ordered quasicrystals raises the question of whether or not the long-range order in these alloys can be reconciled with the electronic disorder. This study did not detect any unusual structure-induced contribution to the resistivity at low temperatures other than those known for metallic glasses. The temperature dependence of the resistivity in icosahedral quasicrystals of high structural quality showed that the transport behaviour of these alloys at low temperatures can be satisfactorily explained in terms of conventional weak-localization and electron-electron interaction theories. The temperature dependence of the resistivity in weakly disordered (low-resistivity) alloys can be explained qualitatively in terms of the classical theories extended to liquids and disordered alloys. The experimental evidence shows the existence of a close relationship between the icosahedral quasicrystalline structure and the weak localization tendency of the electronic states at the Fermi level. To explain the temperature dependence of the resistivity at high temperatures additional models based on the concept of the band transition and hopping have to be invoked. In the high-resistivity stable icosahedral Al-Cu-Ru alloy the insulating-like (electron hopping) behaviour was found to dominate the electron transport even at low temperatures. This causes the deviations from the weak-localization theory and is due to the enhancement of the density of localized electronic states at the Fermi level. Consequently, the Anderson localization is collapsed in the vicinity of the metal-insulator transition. The values of the correlation gap in the Al-Cu-Fe-Mn icosahedral system suggest a considerably different nature for the pseudogap around the Fermi level in amorphous and icosahedral phases. However, unless an ideal quasicrystal belongs to the metal-insulator transition region (if it exists), the weak-localization theory will be the most appropriate tool to investigate the low temperature electron transport properties in icosahedral phases.