Ferroelectric materials have been extensively studied theoretically and experimentally for many decades. Their ferroelectric, piezoelectric, pyroelectric, dielectric and electro-optical properties offer great promise in various applications such as non-volatile random access memory devices, non linear optics, motion and thermal sensors, and tunable microwave devices. Advanced applications for high dielectric constant insulators and nonvolatile memories in semiconductor industry have led to a meteoric rise of interest in the ferroelectrics recently. As most studied and technically important ferroelectric materials, lead zirconate titanate (PZT) and barium strontium titanate (BST) are widely investigated to understand their properties for potential device applications. Using radio frequency magnetron sputtering, single crystalline PZT and BST thin films have been achieved on SrTiO3 substrates, and been characterized for their structural and electrical properties. Eyeing their different potential applications, ferroelectric, pyroelectric and dielectric properties of PZT and BST thin films were studied. In addition, the introduction of bridge layers (nucleation or buffer layers) grown by molecular beam epitaxy (MBE) has been employed to facilitate the heterostructure growth of PZT thin films on GaN and BST thin films on sapphire substrates. Highly (111)-oriented perovskite PZT thin films were achieved on silicon-doped GaN (0001)/c-sapphire with a PbTiO3/PbO oxide bridge layer. And (001)-oriented BST thin films were grown on a-plane sapphire with an MgO/ZnO bridge layer. This dissertation also discusses the realization of PZT ferroelectric field effect transistors (FeFET). Two different 1T FeFET structures were successfully fabricated and their electrical properties were examined. Ferroelectric behavior was observed in the plot of source-drain current versus gate voltage where it exhibited a large counterclockwise hysteresis with 50% current modulation.

Owing to the large electric-field-dependent permittivity, ferroelectric thin films have attracted a great deal of attention on applications in miniature tunable microwave components with high performance and cost reduction, such as phase shifters, tunable oscillators, delay lines, and antennas. These tunable devices require large change in the dielectric constant with applied field and a low loss at microwave frequencies. As one of the promising ferroelectric materials, barium strontium titanate thin film, especially Ba0.5Sr0.5TiO3 (BST) films, have raises great research interests due to its high dielectric constant, which is tunable in an external electric field, combined with relative low loss at microwave frequencies. Tunable microwave components, such as phase shifter, based on the BST films have been widely investigated. Since the polarization, the significant characterization of ferroelectrics, is very sensitive to distortion in crystal structure of ferroelectrics, strain can be effectively utilized to tailor the dielectric properties of BST films. Due to the lattice-mismatch from the substrate and various deposition conditions, epitaxial BST thin film usually contains residual strain generated during film growth. Strain control by improved deposition technique and implementing thermal treatment as well as choosing suitable substrate has attracted intensive attentions in ferroelectric film fabrication. Theory predicts that high dielectric properties can be achieved when free strain or slightly tensile strain left in the BST thin film at room temperature. Microwave application, such as phase shifter, also expects the enhanced tunability by an applied electric field. In this dissertation, single crystalline BST thin films deposited by radio frequency magnetron sputtering on SrTiO3 and DyScO3 substrates were studied. The crystal structure characteristics, including lattice parameters and film strain, were determined using X-ray diffraction. A new growth technique, three-step technique, was introduced and implemented into BST thin film deposition. The application of this new technique in deposition dramatically reduced the compressive strain in the films. We use microwave measurements on coplanar waveguides to evidence the improvement on dielectric properties achieved by tailoring the film strain. Additionally, we studied the BST film deposited by pulsed laser deposition (PLD) with introducing a sputtered seed layer of BST thin film. Compared with the BST film directly deposited on the substrate by PLD deposition, the films with a seed layer showed a large enhancement on the dielectric constant and tunability. The discussion on the change in film strain and dielectric performance of the PLD deposited films further proved the influence of film strain on dielectric properties. We discussed the design, fabrication, and measurement of coplanar waveguide transmission lines as phase shifters fabricated BST films. The thin BST films (~700 nm) on DyScO3 substrates deposited by sputtering demonstrated that the three-step deposition technique improved differential phase shift and microwave figure of merit to a great extent. The introduction of the sputtered seed layer into the PLD deposition of a thicker BST film (~2.15 μm) showed a dramatically enhancement on differential phase shift and microwave figure of merit. The enhanced performance on different series of BST films in microwave frequencies is consistent with the improvement on crystal structure, especially with the change in film strain.

The main objective of this thesis is to efficiently disperse and align SWNTs in two different polymer matrices to obtain an orthotropic composite whose strength, stiffness and electrical properties depend on the orientation of the SWNTs. The SWNTs are successfully dispersed and aligned in a polyimide matrix and a polymer blend of UDMA/HDDMA. In-situ polymerization under sonication is used to disperse the SWNTs in polyimide matrix and sonication is used to disperse SWNTs in the UDMA/HDDMA matrix. In both cases, an electric field is used to align the SWNTs in the polymer matrices. In the polyimide, the SWNTs are aligned by electrospinning technique, and in (UDMA/HDDMA) the SWNTs are aligned by applying an AC electric field, while the composite is cured.The electrical and mechanical properties of randomly dispersed SWNT polyimide composites and SWNT/UDMA/HDDMA composite are measured. The dielectric constant and storage modulus of SWNT polyimide composite increased with SWNT concentration. Low percolation (0.06 wt%) and an increase of 113% in storage modulus with 0.2 wt% SWNTs, both indicate good dispersion of SWNTs in the polyimide matrix. The dielectric constants, conductivity for the unaligned SWNT/UDMA/HDDMA composite are isotropic. The electrical and mechanical properties of the randomly dispersed SWNT polyimide composite and SWNT/UDMA/HDDMA composite are used as references when analyzing the aligned counter parts. Different characterization methods are used to assess the alignment of the SWNTs in the polyimide and (UDMA/HDDMA) matrices. A variety of characterization techniques, i.e. microscopy, Raman spectroscopy, electrical conductivity, dynamic dielectric spectroscopy and dynamic mechanical analysis, indicate preferential alignment of SWNTs in two types of polymers: Polyimide and (UDMA/HDDMA). Optical microscope images showed alignment of the SWNTs in the UDMA/HDDMA composite. Inspection of the Raman spectra on aligned SWNT polyimide composite fibers and aligned SWNT/UDMA/HDDMA composite indicates a decrease in the intensity of the tangential peak of the SWNT with increase in the polarizer angle. The difference in the perpendicular and parallel Raman peaks indicate preferential alignment of SWNTs in both the polymer matrices. In the aligned polyimide composite, percolation transition is at 0.2 wt% SWNT concentrations when dielectric constant is measured parallel to the aligned SWNTs. But percolation transition is at 0.65 wt% SWNT concentrations when dielectric constant is measured perpendicular to the aligned SWNTs. Electrical measurements on aligned SWNT polyimide and UDMA/HDDMA composite are highly anisotropic. In both cases, the dielectric constant values parallel to the direction of SWNT alignment are higher than the values perpendicular to the direction of SWNT alignment. To analyze the resulting anisotropy in the dielectric constant, Bruggeman's effective medium approach is used. The effective medium theory predicts the effective dielectric constant of a composite with aligned anisotropic inclusions. The effective dielectric constant, perpendicular to the aligned inclusions and parallel to the aligned inclusions is estimated. The dielectric constant values of aligned SWNT polyimide and aligned SWNT/UDMA/HDDMA composites are compared to the experimental results. Both the values from the theory and experiment show anisotropy in dielectric constant. The theory indicated that the dielectric constant parallel to the aligned inclusions is highly influenced by the dielectric constant of the inclusion and the dielectric constant perpendicular to the aligned inclusions is highly influenced by the dielectric constant of the polymer matrix. Results from the different characterizing techniques indicate that SWNTs are successfully aligned in the polyimide matrix and (UDMA/HDDMA) matrix by electrospinning technique and by an AC electric field respectively.

Engineers know well from an early point in their training the trials and tribulations of having to make design tradeoffs in order to optimize one performance parameter for another. Discovering tradeoff conditions that result in the elimination of a loss associated with the enhancement of some other parameter (an improvement over a typical tradeoff), therefore, ushers in a new paradigm of design in which the constraints which are typical of the task at hand are alleviated. We call such a design paradigm “tuning” as opposed to “trading off”, and this is the central theme of this work. We investigate two types of microwave electronic devices, namely GaN-based heterostructure field effect transistors (HFETs) and tunable ferroelectric-ferrite-based microwave phase shifters. The “tuning” associated with these types of devices arises from the notion of an optimal 2DEG density, capable of achieving higher performance in terms of electron velocity and enhanced reliability in the case of the HFET, and the coupling of ferroelectric and ferrite materials in tunable microwave phase shifters, capable of achieving high differential phase shifts while at the same time mitigating the losses associated with impedance mismatching which typically arise when the phase is tuned. Promises and problems associated with HFET devices based on the intriguing InAlN/GaN material system will be described. We focus on the fundamental problem associated with the induction of the large density of carriers at the interface, namely the disintegration of an excess of longitudinal optical phonons (hot phonons) in the channel. We use microwave measurements in conjunction with stress tests to evidence the existence of an optimal 2DEG density wherein the hot phonon effect can be “tuned,” which allows for enhanced high frequency performance as well as device reliability. Next, we focus on the design, fabrication, and measurement of tunable phase shifters consisting of thin films of BaxSr1-xTiO3 (BST), which has the advantage of having high dielectric tunability as well as relatively low microwave loss. We discuss the design, fabrication, and measurement of a simple coplanar waveguide (CPW) type of phase shifter as well as a more complicated “hybrid” phase shifter consisting of a ferrite (YIG) in addition to BST. The use of such a bilayer allows one to “tune” the impedance of the phase shifters independently of the phase velocity through careful selection of the DC biasing magnetic fields, or alternatively through the use of an additional piezoelectric layer, bonded to YIG whose permeability can then be tuned through magnetostriction.

The processing and properties of novel polymer-based nano-composite (PNC) dielectrics for capacitor applications has been studied. PNCs were fabricated via a vacuum based deposition technique and their micro/nano-structure, chemical and dielectric properties investigated. After process development and optimisation, co-deposited Al and nylon-6 PNCs had a dielectric constant k∼7 at an approximate Al volume fraction of 0.3 that agreed with analytical predictions if it was assumed that the Al transformed to an oxide in-situ and/or after deposition. The significant effect of absorbed water vapour and temperature on PNC dielectric properties was revealed using different types of post-deposition heat treatment. Alternately-deposited PNCs consisting of Al or Ag 2-20 nm layers sandwiched between nylon-6 layers were fabricated in which the overall PNC Al or Ag volume fraction was controlled by varying the nominal Al or Ag layer thickness. Ag layers comprised of discrete nano-islands that produced a nano-capacitor network effect that increased k to ∼11. In the case of Al layers, when the layer thickness was ≥ 5 nm, corresponding to a nominal volume fraction of 0.1, Al (core)-oxide (shell) nanoparticles were formed and the PNC dielectric constant increased to ∼19. The detailed nano-structure of the core-shell particles was studied using various types of transmission electron microscopy (TEM), and the elevations in dielectric constant ascribed to multiple-interface polarisation effects dependent on the formation of the core-shell structure. PNCs based on alternate deposition of Ti sandwiched in nylon-6, and then both Ti and Ag in nylon-6 were also fabricated, with k reaching ∼73 for Ag+Ti/nylon-6 PNCs. As well as Ti-based core (metal)-shell (oxide) particles, the Ag volume fraction was sufficiently high in the 10 nm nylon-6 layers to again form a nano-capacitor network that contributed to the overall device capacitance and effective dielectric constant. Again, various types of high magnification TEM were critical in resolving the Ti-based core-shell structure and its role in high-k behaviour. The vacuum-based alternate deposition technique has been developed to offer ease of operation, reliability, flexibility and applicability to chemically different filler and matrix systems in the fabrication of high-k PNC based capacitors, in which high-k performance relies critically on the formation of core (metal)-shell (oxide) particles in both Al and Ti based systems.

With the emergence of the Internet of Things (IoT), wireless body area networks (WBANs) are becoming increasingly pervasive in everyday life. Most WBANs are currently working at the IEEE 802.15.4 Zigbee standard. However there are growing interests to investigate the performance of BANs operating at higher frequencies (e.g. millimetre-wave band), due to the advantages offered compared to those operating at lower microwave frequencies. This thesis aims to realise printed conductive traces on flexible substrates, targeted for high frequency wearable electronics applications. Specifically, investigations were performed in the areas pertaining to the surface modification of substrates and the electrical performance of printed interconnects. Firstly, a novel methodology was proposed to characterise the dielectric properties of a non-woven fabric (Tyvek) up to 20 GHz. This approach utilised electromagnetic (EM) simulation to improve the analytical equations based on transmission line structures, in order to improve the accuracy of the conductor loss values in the gigahertz range. To reduce the substrate roughness, an UV-curable insulator was used to form a planarisation layer on a non-porous substrate via inkjet printing. The results obtained demonstrated the importance of matching the surface energy of the substrate to the ink to minimise the ink de-wetting phenomenon, which was possible within the parameters of heating the platen. Furthermore, the substrate surface roughness was observed to affect the printed line width significantly, and a surface roughness factor was introduced in the equation of Smith et al. to predict the printed line width on a substrate with non-negligible surface roughness (Ra ≤ 1 μm). Silver ink de-wetting was observed when overprinting silver onto the UV-cured insulator, and studies were performed to investigate the conditions for achieving electrically conductive traces using commercial ink formulations, where the curing equipment may be non-optimal. In particular, different techniques were used to characterise the samples at different stages in order to evaluate the surface properties and printability, and to ascertain if measurable resistances could be predicted. Following the results obtained, it was demonstrated that measurable resistance could be obtained for samples cured under an ambient atmosphere, which was verified on Tyvek samples. Lastly, a methodology was proposed to model for the non-ideal characteristics of printed transmission lines to predict the high frequency electrical performance of those structures. The methodology was validated on transmission line structures of different lengths up to 30 GHz, where a good correlation was obtained between simulation and measurement results. Furthermore, the results obtained demonstrate the significance of the paste levelling effect on the extracted DC conductivity values, and the need for accurate DC conductivity values in the modelling of printed interconnects.

This work deals with the linear analysis of plasma waves, especially with the methods of solution of a hot plasma dispersion relation. There are cited some results achieved in the space plasma research and mapped current numerical methods of their analysis. Besides, this work introduces a new numeric procedure, computer code PDRS (Plasma Dispersion Relation Solver), which allows finding solution of a dispersion function of a cold or hot plasma with general distribution function. It also demonstrates the usage of the PDRS methods on real examples of waves in space plasma based on the spacecraft Cluster measurement.

A thesis submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Doctor of Philosophy. August 2014. / The process of the computational discovery of materials for future technologies is a combination of numerical techniques and general scientific intuition to select elements and combine in order to form novel types of materials. Modern ab initio methods based on density functional theory are capable of predicting with a high level of accuracy the most stable ground state atomic configurations of any given material. Once the ground state configurations are established, the electronic, optical and mechanical properties of the novel bulk nitrides may be determined. Electronic properties of C3N4, CN2, SiN2, GeN2, C2N2(NH), Si2N2(NH), Ge2N2(NH) and Sn2N2(NH) are analysed by computing the Kohn-Sham band structures. The optical properties are investigated by calculating the real and the imaginary parts of the frequency-dependent dielectric constant. The mechanical properties are determined by calculating elastic constants, Young’s modulus, Poisson’s ratio, Vickers hardness, shear and bulk moduli.

Student Number : 9803381J - MSc Dissertation - Faculty of Engineering and the Built Environment / Although pulsed power has been used in many parts of the world over the last few decades to initiate high-voltage discharges through rock, no systematic test work on South African ore bodies and related rock types has been done so far. As part of CSIR Miningtek’s integrated approach of combining underground comminution with a novel Tore© hydrotransport system, which has been shown to operate well with coarse particles up to 10 mm, various rock types were fragmented in single discharge mode under laboratory conditions. The work was conducted at the University of the Witwatersrand’s high-voltage laboratory with a custom-designed test rig. The rig configuration was based on a critical review and analysis of the literature and on assessments of existing test facilities elsewhere. Core samples with diameters ranging from 16 to 48 mm were cut from test specimens with thicknesses ranging from 8 to 48 mm. Rock types included Ventersdorp Contact Reef, Carbon Leader, Elsburg Formation, UG2 and Merensky, as well as pure quartz, shales, lava and dykes. A six-stage Marx generator provided a voltage rise time of 2 000 kV/μs to create a discharge through the rock, in preference to a discharge through the surrounding water, which acts as an insulator at ramp-up times faster than 0,5 μs. High-speed photography, and an analysis of the voltage and current signals for various rock types and for water alone, were used to quantify the potential benefits of rock breaking by electric discharge. It was found that some Kimberlite specimens and mineralised gold-bearing reefs were much easier to fragment than hanging wall or footwall material. Merensky reef appeared to be more susceptible than the less brittle UG2 material. A correlation was derived between the dynamic resistivity of various rock types, measured at 16 MHz excitation frequency, and the electrical breakdown strength at which discharge took place. The fragments created had a more cubical shape than would be created by conventional impact crushing. However, the high voltage requirements of about 30 to 35 kV per millimetre of rock thickness would necessitate not only efficient mechanical and electrical contact between the electrodes and the rock, but also considerable safety features for underground installations. The clearly identified, preferential fracturing of reef rock types, compared with the hanging or footwall materials, suggests that the greater benefit of electric rock breaking may lie in primary rock breaking as a mining method, rather than in secondary comminution of broken rock to enable hydraulic transportation by pipeline to surface.

[pt] É notória a necessidade de encontrar novas tecnologias para atuação de sistemas robóticos tão eficazes quanto a do músculo natural. Os atuadores tradicionais possuem grande agilidade e força quando comparados aos músculos naturais, mas suas dimensões e peso são elevados em relação à força que são capazes de exercer, e demandam muita energia para cumprirem suas tarefas. Manipuladores robóticos menores e mais baratos poderiam existir se pudessem utilizar músculos naturais para impulsioná-los. Ao mesmo tempo, as indústrias gastariam menos com os custos envolvidos em energia e compra desses manipuladores. Este trabalho estuda o comportamento de músculos artificiais baseados no polímero acrílico VHB4905, através da construção de uma bancada de teste com sensor de força, incluindo especificação de todos os equipamentos e o desenvolvimento de circuitos de alta tensão (até 10kV) para acioná-los. Durante o desenvolvimento e implementação do circuito foram encontrados problemas intrínsecos à manipulação de tensões da ordem de vários kV. Esses problemas foram apontados e soluções satisfatórias foram implementadas, de forma a tornar os experimentos possíveis. Modelos matemáticos de algumas das principais configurações possíveis para atuadores foram desenvolvidos. Os modelos desenvolvidos para uma das configurações típicas foram comparados com resultados experimentais com um erro máximo absoluto de 1% (26,7mN) do valor real. Experimentos em atuadores de molduras fixas foram feitos com resultados de 223% de deformação da região ativa, com desempenho muito superior ao dos músculos naturais. A partir de um dos modelos desenvolvidos, implementou-se um controlador PID compensado que gerou melhores resultados a entradas em degrau que o PID padrão, o qual não leva em consideração a não linearidade e a alta sensibilidade do atuador quando submetido a tensões próximas da tensão de quebra do dielétrico. A eficácia da técnica de controle proposta foi comprovada experimentalmente. / [en] It is well known the needs of finding new technologies for robotic systems actuations, with the same efficiency of the natural muscles. The common actuators have better agility and force when compared to natural muscles, but the dimensions and weight are bigger and for that the demand of energy necessary for the actuation is higher. Smaller and cheaper robot manipulators could exist if they were able to use natural muscles to drive them. At the same time, industries would spend less money with energy and manipulators. This work studies the behavior of artificial muscles based on dielectric elastomers (VHB4905) through the development of a test bench with force transducer, including the specification of all the equipments and the development of a high voltage circuit (10kV maximum). During the development and implementation of the circuit, problems inherent to high voltage manipulation were found. Those problems were shown and tolerable solutions were taken, so that the experiments were feasible. Mathematic models of some of the main configurations for actuators were developed. One of those models (from a typical configuration) was compared with experimental results with a maximum absolute error of 1% (26.7mN) of the real value. Experiments with fixed frame actuators were made with 223% of strain, showing a much higher performance compared to natural muscles. With one of the mathematic models, a PID controller with adjustable gains was developed and presented better results, for a step response, when compared to a standard PID controller. This last one do not take into account the non-linearities and for that it behaviors with great sensibility when subjected to high voltages (close to dielectric breakdown). The effectiveness of the proposed control technique was proved experimentally.