This thesis investigates the operation of the current transfonner (CT) when sensing retum-to-zero current pulses in power electronic circuitry. The CT's output signal is nonnally rectified when sensing current pulses and the effects of the different rectification techniques on peak current and average current droop are evaluated. Initially, the various current sensing techniques and their application in power electronics circuits are reviewed. The CT and both diode and synchronous rectification are then reviewed in more detail. Operation of the CT with diode rectification (DR) and natural resetting is investigated. Three operating modes are identified. These are the discontinuous magnetizing current, continuous magnetizing current and discontinuous secondary current modes. The error (droop) in the average output signal obtained is found to be predominantly defined by CT core losses. Coefficients are given for correcting the error due to droop, provided that the discontinuous secondary current mode is avoided. Diode rectification with the dual CT arrangement is also investigated. Operation of the CT with synchronous rectification (SR) and natural resetting is then investigated. The SR topologies possible using a discrete MOSFET are categorized. During experimentation the arrangement used to drive the MOSFET's gate is found to be important if distortion is to be minimized. It also is found that the average current droop is dependent on the oscillatory behaviour of the resetting circuit and has an effectively random component. The magnitude of this component is defined by the voltage drop exhibited by the SR MOSFET's intrinsic anti-parallel diode. SR is then implemented using a commercially available analogue switch. The problems detailed with the use of a discrete MOSFET are largely alleviated. Another benefit is that the increased restriction on maximum duty factor imposed by introducing a discrete MOSFET is also eased. However, whichever SR technique is implemented, an operational amplifier is used and the transient response of this circuit element is important. A method of minimizing droop by indirect sensing of the CT's peak core flux excursion is then presented. A corresponding correcting voltage is applied in series with the CT's output terminals during a current pulse. The magnitude of this voltage is based on the magnitude of the resetting voltage sensed during previous switching cycles. A circuit is implemented and simulated. Experimental results are presented. A switched-mode circuit operating at a frequency higher than that of the main power circuit is then used to apply the correcting voltage with the objective of reducing the power drawn. Again, the circuit is implemented and simulated and experimental results are presented.

ABSTRACT Solid oxide fuel cells (SOFCs) are of major interest in fuel cell development due to their high energy conversion efficiency, wide range of fuels and environmental friendliness. One important obstacle for their industrial development is their processing difficulties. These difficulties have recently been addressed by employing a novel technique namely electroless nickel - yttria-stabilised zirconia (YSZ) co-deposition which eliminates multi-layer processing and high temperature sintering. The novel work carried out in this research programme investigates the effects of different processing parameters on the co-deposited anodes for SOFCs. In particular, YSZ particle size, electroless bath agitation method, electroless bath pH and substrate surface condition are investigated. These variables were investigated for their effect on (i) the ceramic to metal ratio – important in terms of matching the coefficient of thermal expansion of the anode and substrate, as well as providing electronic conductivity, and (ii) the porosity content in the deposited layers – required for fuel and exit gas penetration through the anode. The experimental work was based on a full factorial Design of Experiment (DoE) approach and consisted of three phases – namely, designing, running and analysing. A 16 run 24 full factorial DoE with five replications was constructed with YSZ particle sizes of 2 and 10 µm; bath agitation of air bubbling and mechanical stirring; bath pH of 4.9 and 5.4; and substrate surface treatment of hydrofluoric acid etching and mechanical blasting. A total of 80 samples were analysed for nickel content by energy dispersive X-ray analysis and porosity content by Archimedes buoyancy measurement. The DoE was analysed by the ANOVA statistical tool in Minitab 15 software. The co-deposition conditions that produced anodes with (i) the lowest volume percentage of nickel and (ii) the highest level of porosity were determined. Linear regression models for both nickel to YSZ content and porosity responses were built to estimate the correlation between experimental and predicted data. The coefficient of determination, R2 for nickel to YSZ content indicated a reasonable correlation between experimental and predicted values while the regression model for porosity response was less reliable. One anode containing 50 vol.% nickel recorded an electronic conductivity at 400oC in air that is comparable to the published data. Another series of tests at higher temperatures (up to 800oC) in air and nitrogen resulted in encouraging electronic conductivities being recorded.

This thesis investigates a novel method for manufacturing a tubular solid oxide fuel cell cathode. The work involved depositing a lanthanum nickel ferrite / lanthanum strontium manganite cathode onto an yttria stabilized zirconia electrolyte using an electroless co-deposition technique. The lanthanum nickel ferrite is a promising cathode material but suffers from degradation at the high temperatures encountered during the sintering process which is required during conventional cathode processing. The novel technique employed in this work does not involve these high temperatures so the investigation was focused on whether co-deposition could be employed to use the lanthanum nickel ferrite (LNF). The experimental work involved co-deposition of conventional cathode materials – lanthanum strontium manganite and yttria stabilized zirconia onto both planar and tubular sections of alumina substrates, under the environment of both acid and alkaline baths. This was followed by repeating the procedure using lanthanum nickel ferrite onto tubular and planar yttria stabilized zirconia substrates. The performance of these co-deposited cathodes was characterized using optical and scanning electron microscopy, energy dispersive X-ray analysis and electrochemical analysis. These were planar fuel cells so as to allow basic testing of the cells. The work thus demonstrated that electroless co-deposition of tubular cathodes incorporating LNF was successful – both via SEM / EDX characterisation and basic electrical testing. Factors which affected the coating deposition and performance were also investigated and a comprehensive overview of the development of solid oxide fuel cells is also detailed.

The rapid progress in Adaptive Optics (AO) imaging, in the last decades, has had a transformative impact on the entire approach underpinning the investigations of retinal tissues. Capable of imaging the retina in vivo at the cellular level, AO systems have revealed new insights into retinal structures, function, and the origins of various retinal pathologies. This has expanded the field of clinical research and opened a wide range of applications for AO imaging. The advances in image processing techniques contribute to a better observation of retinal microstructures and therefore more accurate detection of pathological conditions. The development of automated tools for processing images obtained with AO allows for objective examination of a larger number of images with time and cost savings and thus facilitates the use of AO imaging as a practical and efficient tool, by making it widely accessible to the clinical ophthalmic community. In this work, an image processing framework is developed that allows for enhancement of AO high-resolution retinal images and accurate detection of photoreceptor cells. The proposed framework consists of several stages: image quality assessment, illumination compensation, noise suppression, image registration, image restoration, enhancement and detection of photoreceptor cells. The visibility of retinal features is improved by tackling specific components of the AO imaging system, affecting the quality of acquired retinal data. Therefore, we attempt to fully recover AO retinal images, free from any induced degradation effects. A comparative study of different methods and evaluation of their efficiency on retinal datasets is performed by assessing image quality. In order to verify the achieved results, the cone packing density distribution was calculated and correlated with statistical histological data. From the performed experiments, it can be concluded that the proposed image processing framework can effectively improve photoreceptor cell image quality and thus can serve as a platform for further investigation of retinal tissues. Quantitative analysis of the retinal images obtained with the proposed image processing framework can be used for comparison with data related to pathological retinas, as well as for understanding the effect of age and retinal pathology on cone packing density and other microstructures.

Animal models are often used to test the efficacy and safety of clinical applications employing focused ultrasound that range in various stages of research, development and commercialization. The animals are usually subjected to conditions that cause pain, distress and euthanasia. Access to cadaveric models is not easy and affordable for all research institutions, whereas conservation and changes of their physical properties over time can be a delimiting factor for translational research. The above set the motivation for this project, which its primary objective is to design and develop appropriate tissue mimicking phantoms using a simplistic and cost effective methodology. These phantoms are expected to contribute in reducing the need for animal testing and allow researchers to get hands experience with tools that will promote and accelerate testing in focused ultrasound thermal protocols. The main requirements for these phantoms are to be geometrically accurate, compatible with magnetic resonance imaging (MRI) and to be composed of materials that approximate the acoustic and thermal properties of the replicated tissues. Throughout the duration of the project three ultrasonic composite phantoms (head, femur bone-muscle and breast-rib) were developed. The acoustic properties of candidate materials were assessed using pulse-echo immersion and through transmission techniques. The thermal properties were estimated by observing the rate of heat diffusion following a sonication in the soft tissue parts with MR thermometry. Acrylonitrile butadiene styrene (ABS) was used to replicate bone tissue, where its acoustic attenuation coefficient was found to be 16.01 ± 6.18 dB/cm at 1 MHz and the speed of sound at 2048 ± 79 m/s. Soft tissue parts consisted out of agar-based gels doped with varying concentrations of additives that controlled the relative contribution of acoustic absorption (evaporated milk) and scatter (silica dioxide) to total attenuation independently. Brain tissue phantom (2 % w/v agar - 1.2 % w/v SiO2 - 25 % v/v evaporated milk) matched an attenuation coefficient of 0.59 ± 0.05 dB/cm-MHz whereas muscle and breast mimicking phantom (2 % w/v agar - 2 % w/v SiO2 - 40 % v/v evaporated milk) were estimated of inducing an attenuation coefficient of the order of 0.99 ±0.08 dB/cm-MHz. The speed of sound for the brain and muscle/breast recipe were estimated at 1485 ± 12 m/s and 1529 ± 13 m/s respectively. The thermal conductivity of the brain phantom was estimated to be 0.52 ± 0.06 W/mº-C and 0.57 ± 0.10 W/mº-C for the muscle/breast phantom. The acoustic and thermal properties of candidate materials were within range of the replicated tissues extracted from literature, except the speed of sound in ABS compared which was lower compared to bone (~3000 m/s). Three dimensional models of bone parts (skull, femur, rib) were reconstructed in Standard Tessellation Language (STL) format by segmenting bony tissue of interest from adult human computed tomography (CT) images. The STL bone models were 3D printed in ABS using a fused deposition modelling (FDM) machine. The final composite phantoms were fabricated by molding the agar based soft tissue phantoms inside/around the ABS bone phantoms. The functionality of all three composite phantoms was assessed with focused ultrasound sonications applied by a 1 MHz single element transducer while temperature was monitored with 1.5 Tesla MRI scanner. A spoiled gradient recalled (SPGR) pulse sequence was used to produce phase images that were analyzed using a custom coded software developed in Matlab that employed proton-resonance frequency shift (PRFS) thermometry.

The pronounced development in the field of solar cells has been driven by the increasing interest in “green” energy generation in recent decades. Nevertheless, to increase the deployment of solar cells the energy conversion efficiency has to be improved further. The highest energy conversion efficiency has been recorded using a Silicon solar cell. However, there are limitations such as the high reflection from the solar cell surface that limits further improvement of the energy conversion efficiency. The large refractive index contrast between air and the material of the solar cell leads to high reflection. As a consequence, reducing the reflection from the solar cell surface is a priority. This research aims at reducing reflection from the solar cell surface. To achieve this goal, modeling based analysis of a micro pillar array texturing pattern and a new and exciting texturing pattern (the hut-like pattern) are presented. The simulation method used for this study is the Finite Difference Time Domain (FDTD) method. In the discussion, the effect of key structural parameters on the reflection is analyzed to obtain an in-depth understanding of the patterns. Additionally, the inter-dependence between the different structural parameters under study is considered during the discussion. The analysis shows that the reflection from a micro pillar array solar cell decreases as the Height (H) increases. The H by Diameter (H/D) ratio analysis presented in this work determines that there is a convergence in the reflection when the H/D ratio is high. This can be useful especially for designers with low precision fabrication equipment who can target higher H/D ratio to ensure a low reflection. The high surface-to-volume ratio when the H/D ratio is high can lead to high surface recombination. High surface recombination is a major problem in textured solar cell since it diminishes the electrical performance. An alternative is to use the hut-like pattern that combines the benefits of lowering the reflection with a low surface-to-volume ratio. The low surface-tovolume ratio is expected to have a positive effect on the surface recombination (i.e. lower surface recombination). The results show excellent optical performance of the pattern. Additionally, the hut-like pattern provides a reflection lower than other texturing patterns such as pyramid, nanowires and micro pillars. Furthermore, the results of a fabricated proof of concept hut-like sample are presented which highlights the excellent optical performance of the pattern.

Gallium nitride (GaN) based technology has been heavily researched over the past two decades due to its ability to deliver higher powers and higher frequencies that are demanded by the market for various applications. One of GaN’s main advantages lies in its ability to form heterojunctions to wider bandgap materials such as Aluminium Gallium Nitride (AlGaN) and Aluminium Nitride (AlN). The heterostructure results in the formation of the so called 2 dimensional electron gas (2DEG), which exhibits high electron densities of up to 6E13 cm−2 and high electron mobilities of up to 2000 cm2/V·s that enable the devices to support high current densities. Furthermore, it supports very high breakdown fields of 3.3 MV/cm due to its wide bandgap of 3.4 eV. The main objective of this work was to further advance the transistor technology using simple, cost effective and reliable techniques. The AlN/GaN material system exhibits higher sheet carrier concentrations compared to the conventional ternary AlGaN barrier, but introduces additional challenges due to its reduced thickness of 2-6 nm compared to 18-30 nm of AlGaN. The additional challenges of the thin AlN binary barrier include strain relaxation, high gate leakage currents and high Ohmic contact resistances due to its high bandgap of 6.2 eV. In this work, a thin (5 nm) in-situ SiNx passivation layer was employed to reduce the strain relaxation, reduce gate leakage currents and improve Ohmic contacts resistances. The optimised Ohmic contact annealing condition resulted in an Ohmic contact resistance of 0.4 Ω·mm and a sheet resistance of 300 Ω/

Separating and sorting cells and micro-organisms from a heterogeneous mixture is a fundamental step in biological, chemical and clinical studies, enabling regenerative medicine, stem cell research, clinical sample preparation and improved food safety. Particle and cell manipulation by ultrasound acoustic waves provides the capability of separation of cells on the basis of their size and physical properties. Offering the advantages of relatively large microfluidic volumes in a label-free, contactless and biocompatible manner. Consequently, the discovery of alternative methods for precise manipulation of cells and particles is of highly demand. This thesis describes a novel approach of ultrasound acoustic manipulation of particles and cells. The principle of operation of the dynamic acoustic field method is described accompanied with acoustic separation simulations. Furthermore, the complete fabrication and characterisation of two types of ultrasound devices is given. The first one is a bulk acoustic wave (BAW) device and the second is a surface acoustic wave (SAW) device. Successful experiments using the BAW device for sorting different diameter particles with a range from 5 to 45 μm are demonstrated, also experiments for sorting particles depending on their density are presented. Moreover, experiments of the proposed method for sorting porcine dorcal root ganglion (DRG) cells from a heterogeneous mixture of myelin debris depending on their size are displayed. Experimental results of sorting cells depending on their stiffness are demonstrated. Experiments using the fabricated SAW device for sorting different diameter particles in a constant flow with a range from 1 μm to 10 μm are presented. Furthermore, experiments of the proposed method for sorting live from dead Htert cells depending on their mechanical properties, i.e. stiffness are displayed. As a side project a new idea for dynamic acoustic manipulation by rotating the acoustic field is demonstrated. The basic principles of this method and the simulations for verifying this concept are displayed. Experiments for sorting 10 μm from 3 μm polystyrene particles are presented, with two different types of the dynamic acoustic rotating field being examined.

Over the last decades of research on sustainable energy, thermoelectric generation has been identified as a potential energy harvesting solution for a wide range of applications. Nowadays, the commercial thermoelectric technology is almost entirely based on tellurium alloys, it mainly addresses room temperature applications and it is not compatible with MEMS and CMOS processing. In this work, silicon-germanium based micro-devices have been designed, developed and characterized with the aim of addressing the heat recovery needs of the automotive industry. The micro-scale of the fabricated devices, together with the full compatibility with silicon micro-processing, also profiles an interesting potential for application in the autonomous sensor field. Most importantly, the configuration and the fabrication processes of such silicon-based generators constitute a platform to transfer the results of decades of promising material investigations and engineering into practical micro-scaled thermoelectric generators. The room temperature characterization of the manufactured micro-generators revealed power factors up to 13.9x10-3 μW/(cm2K2) and maximum output power density up to 24.7 μW/cm2. In such temperature range, the micro-devices manufactured in this work are still not as performing as the state-of-the-art bismuth-telluride based technology. However, at around 300 C, the developed micro-modules are predicted to produce a maximum power output of 1.2-1.5mW under 10 C temperature gradient, which corresponds to 35-45% of the room temperature performance of the only commercial bismuth telluride based micro-devices. The results show that silicon-germanium micro-modules could potentially compete with the state-of-the-art commercial micro-devices, being better performing at higher temperature, but also offering the advantage of being a sustainable MEMS and CMOS compatible option for autonomous sensors integration.

Nanowire transistors (NWTs) represent a potential alternative to Silicon FinFET technology in the 5nm CMOS technology generation and beyond. Their gate length can be scaled beyond the limitations of FinFET gate length scaling to maintain superior off-state leakage current and performance thanks to better electrostatic control through the semiconductor nanowire channels by gate-all-around (GAA) architecture. Furthermore, it is possible to stack nanowires to enhance the drive current per footprint. Based on these considerations, vertically-stacked lateral NWTs have been included in the latest edition of the International Technology Roadmap for Semiconductors (ITRS) to allow for further performance enhancement and gate pitch scaling, which are key criteria of merit for the new CMOS technology generation. However, electrostatic confinement and the transport behaviour in these devices are more complex, especially in or beyond the 5nm CMOS technology generation. At the heart of this thesis is the model-based research of aggressively-scaled NWTs suitable for implementation in or beyond the 5nm CMOS technology generation, including their physical and operational limitations and intrinsic parameter fluctuations. The Ensemble Monte Carlo approach with Poisson-Schrödinger (PS) quantum corrections was adopted for the purpose of predictive performance evaluation of NWTs. The ratio of the major to the minor ellipsoidal cross-section axis (cross-sectional aspect ratio - AR) has been identified as a significant contributing factor in device performance. Until now, semiconductor industry players have carried out experimental research on NWTs with two different cross-sections: circular cylinder (or elliptical) NWTs and nanosheet (or nanoslab) NWTs. Each version has its own benefits and drawbacks; however, the key difference between these two versions is the cross-sectional AR. Several critical design questions, including the optimal NWT cross-sectional aspect ratio, remain unanswered. To answer these questions, the AR of a GAA NWT has been investigated in detail in this research maintaining the cross-sectional area constant. Signatures of isotropic charge distributions within Si NWTs were observed, exhibiting the same attributes as the golden ratio (Phi), the significance of which is well-known in the fields of art and architecture. To address the gap in the existing literature, which largely explores NWT scaling using single-channel simulation, thorough simulations of multiple channels vertically-stacked NWTs have been carried out with different cross-sectional shapes and channel lengths. Contact resistance, non-equilibrium transport and quantum confinement effects have been taken into account during the simulations in order to realistically access performance and scalability. Finally, the individual and combined effects of key statistical variability (SV) sources on threshold voltage (VT), subthreshold slope (SS), ON-current (Ion) and drain-induced barrier lowering (DIBL) have been simulated and discussed. The results indicate that the variability of NWTs is impacted by device architecture and dimensions, with a significant reduction in SV found in NWTs with optimal aspect ratios. Furthermore, a reduction in the variability of the threshold voltage has been observed in vertically-stacked NWTs due to the cancelling-out of variability in double and triple lateral channel NWTs.