In this thesis the fabrication and characterisation of nanostructured arrays are discussed. Nanowires, nanotubes and core-shell nanostructures are produced through electrodeposition into the pores of anodised aluminium oxide templates. The electrochemical process is optimised for all materials used. Using nickel, it is demonstrated that the dimensions of the nanowires within the template alter the wavelength position of a reflectivity minima associated with a sub-wavelength optical resonance, corresponding to enhancement of Kerr rotation. The magnetic properties are largely dictated by magnetostatic interactions resulting in an easy axis of magnetisation perpendicular to the nanowire long axis Cobalt nanowires produce a larger Kerr rotation than analogous nickel samples. The magnetic properties are altered by the uniaxial magnetocrystalline anisotropy, confirmed by x-ray diffraction to lie along the long axis of the nanowires. It is therefore possible to rotate the easy axis of magnetisation from the plane of the substrate to the nanowire long axis by increasing the aspect ratio and inter-wire separation. Cobalt nanotubes create a highly complex magneto-optical response with multiple resonances. Nickel-iron alloy nanowires were also made, and the results are compared for changing geometry as well as varying concentration levels. It is revealed through transmission electron microscopy that the grain structure is richly textured with nanocrystalline grains and high volume fraction of grain boundaries. The magnetoplasmonic properties of gold-core cobalt-shell nanowire arrays conclude the thesis. It is found that a nanometric cobalt layer is sufficient to demonstrate magnetic control over optical properties. The optical properties of the array are characteristic of the plasmonic resonances associated with the gold nanowires and the magnetic properties are similar to those of cobalt nanotubes. Combined, however, the wavelength and magnitude of the relative phase shift between P and S polarised components of incident light can be altered with the application of a magnetic field.

The extent of myocardial edema delineates the ischemic area-at-risk (AAR) after myocardial infarction (MI). Since AAR can be used to estimate the amount of salvageable myocardial post-MI, edema imaging has potential clinical utility in the management of acute MI patients. T2 weighted Cardiac Magnetic Resonance (CMR) imaging is widely used to investigate the extent of edema with recent acute MI patient. This thesis describes new approaches and methods of automatic edema segmentation and quantification with 3D visualization. An integrated approach has been developed, including the localization of Left Ventricle (LV) wall, segmentation of myocardial wall, segmentation and quantification of edema and 3D visualization and quantification. A novel automatic segmentation of LV wall is proposed. First a new LV wall localization algorithm is used to locate the centre of the blood pool region of the LV wall. Then a novel LV wall segmentation algorithm is used to segment the LV wall from the rest of anatomical structure. The advantage of the proposed method is in its ability to automatically localize the blood pool region of LV wall and the additional shape constraint which is adaptive to the data. A novel, Automatic Edema Segmentation and Quantification algorithm is presented which is developed based on a statistical mixture model. The technique takes advantage of the characteristic of the MRI signal where the signal is governed by a Rician distribution and using this information regions of edema are segmented over the rest of LV wall. A post-processing step in used to include microvascular obstruction as part of the edema region. The computational simplicity and good edema discrimination are described. Finally, a novel integrated approach to 3D visualization and quantification algorithm is presented. It extracts the information of the LV wall boundary and edema boundary. Then the information is used to generate an interactive 3D image which helps the clinician to visualize the extent of edema and its location. This edema quantification and 3D visualization method is evaluated by expert clinicians with favourable results.

Electric power industries are undergoing restructuring around the world. Under deregulation, the non-utility, independent power producers (IPP) are allowed to invest in the power system. Each year, many new investments in different types of power plants are proposed by generators all over the world. However, investment in generation projects involves large amounts of capital with a long payback period, comparable with the lifetime of the plant. Moreover, macroscopic situation of electrical industry and economic environment also need to be taken into account. Therefore, how to make an investment decision is a complex problem. This thesis considers the power plant investment issues in UK and China from microscopic and macroscopic viewpoint. The microscopic viewpoint focuses on investment appraisal approaches, different types of generation technologies, and levelised cost of generation. The macroscopic viewpoint focuses on investment environment in UK and China, such as the relationship among electricity consumption, installed capacity and GDP. The thesis begins by presenting the current situation of electricity generation in UK and China. It continues to introduce some classic and practical investment theories and project analysis tools. Then, the applications of these approaches on the investment of eight types of electricity generation technologies in the UK and pulverised coal plant in China are given. The analysis results show evidences for several investment advices on microscopic viewpoint. These advices could help investors to more clearly make their investment decision on different types of power plant. From macroscopic viewpoint, the relationship among electricity consumption, economic growth and installed capacity in the UK and China are found by applying econometric approaches. The forecast of electricity consumption and GDP of China are also given. After this, the thesis takes into account the differences of economic growth and environment in different regions of China, and classifies China's provinces into four parts: Northeast, Coastal, Central and West. The relationship between electricity consumption and economic growth in China are determined based on provincial data and panel time series approaches. These results can help investors to fully understand the electric industry investment environment of China, which lead to some policy suggestions.

This research focuses on voltage source multilevel converters in high voltage direct current (HVDC) transmission systems. The first Voltage Source Converter based HVDC (VSC-HVDC) systems with series connected IGBTs in a two-level converter represented a solution to meet industrial and economical requirements but is associated with significant drawbacks such as high dv/dt and di/dt, high switching loss, and poor output voltage and current quality. To overcome these issues, the multilevel converter was proposed for HVDC application. The Modular Multilevel Converter (M2C) was the first multilevel converter to be commercially used in the power industry. In this thesis, the M2C is investigated mainly in terms of operating principle, capacitor size and capacitor voltage ripple, capacitor voltage balancing technique and modulation scheme. The results of this investigation show that the M2C offers the following features: improved efficiency, lower supporting voltage and current in the switching devices and low dv/dt. These features make the M2C suitable for HVDC systems. Two new operational principles and modulation strategies for a Hybrid Cascaded Multilevel Converter (HCMC) are proposed in this thesis. Both modulation schemes extend the modulation index linear range and improve the output waveform quality. This gives the HCMC a higher power density than any known multilevel converter topology for the same dc link voltage and switching device rating. Simulations for both types of multilevel converter (M2C and HCMC) are supported by practical results from scaled hardware laboratory converters. Mathematical analysis and calculation of conversion loss for both types of multilevel converter and for the conventional two-level converter are performed. It is shown that both M2C and HCMC provide lower conversion loss compare to the conventional two-level converter. A control strategy for these two multilevel converters in point-to-point and multi-terminal HVDC systems is also studied. Simulation results show that these two converters are able to operate over the entire specified P-Q capability curve and are capable of riding through ac faults without imposing any over-voltage or over-current on the converter switches.

This thesis contains two projects - the development of a Raman fibre amplifier system for gas sensing and the development of a fibre-in beam-out isolator for use with fibre lasers. The first project developed a fibre Raman amplifier which amplified a 10mW narrow linewidth DFB laser at 1651nm to an output power of over 2W. To achieve this high output power from a narrow linewidth Raman amplifier mechanisms to suppress stimulated Brillouin scattering had to be employed. This amplifier system was packaged into a portable 19” rack enclosure and used to demonstrate remote, single-end, tuneable diode laser spectroscopy of methane. The packaged system detected methane concentrations of 100ppm.m at over 100m during challenging field trials. Extrapolation of lab based measurements to longer distances predicts an ultimate sensitivity of the system of 100ppm.m at greater than 200m. The second project developed a Faraday isolator to be used at the output of industrial fibre lasers. The isolator had an isolation of greater than 30dB, an insertion loss of less than 0.5dB and a return loss of greater than 50dB. The isolator used a dual Faraday rotator design with a half wave plate between the rotators to compensate for thermal stress induced birefringence. Using this approach the isolation was shown to be constant to greater than 60W. The isolator was also compensated for thermal lensing by balancing the positive thermal lens produced in the Faraday rotator with a negative thermal lens in a DKDP crystal. This reduced the thermal lens of the isolator from ~9 Rayleigh length per kW intrinsic to below 2 Rayleigh lengths per kW.

To imitate the way electrical components evolved from discrete devices to devices integrated on Si platform, the next stage for integrated circuits is to integrate photonic components with electrical components on one chip, with active devices known as optoelectronic integrated circuits (OEIC). An ideal solution for this would be to have an all-Si laser. However due to the indirect bandgap of Si this is difficult to achieve. Therefore attention has been focused on trying to integrate the existing and mature III-V laser technology with Si. The difference in lattice constant between GaAs and Si makes direct, monolithic growth of GaAs on Si difficult due to the generation of high defect densities. But the advances in quantum dot (QD) technology and in III-V buffer layer techniques have led to the improvements of direct growth integration. In this thesis an AlAs nucleation layer (NL) in the place of a GaAs nucleation layer was found to increase the photoluminescence intensity and reduce defect density in active layers. Lasers were fabricated with lower threshold current densities than similar devices with GaAs NL. Lasing operation at 1.28 μm was achieved up to 63 °C with a threshold current density of 675 A/cm2 at room temperature. In addition, Ge-on-Si substrates have been used to demonstrate the lasers on Si substrates with a very low pulsed threshold current density of 64 A/cm2, which is significantly lower than any other laser integrated with Si substrates. Also this was the first demonstration of a CW laser on Si with a threshold current density of 163 A/cm2. Lasers were operated up to 30 °C for CW devices and 84 °C for pulsed devices. The difference in threshold currents and temperature performance between CW and pulsed operation is due to high device resistances caused by a combination of poor contact resistance and the introduction of defects from the Si/Ge interface. In conclusion, lasers on Si substrates have been fabricated with low threshold current densities. A very low threshold current density of 64 A/cm2 has been achieved with a Ge-on-Si device and is the lowest result for any type of Si laser at the time of writing, which shows good potential for future integration with Si electronics.

This thesis considers how advances in optical network and optoelectronic technologies may be utilised in particle physics applications. The research is carried out within a certain framework; CERN's Large Hadron Collider (LHC) upgrade. The focus is on the upgrade of the "last-tier" data links, those residing between the last information-processing stage and the accelerator. For that purpose, different network architectures, based on the Passive Optical Network (PON) architectural paradigm, are designed and evaluated. Firstly, a Time-Division Multiplexed (TDM) PON targeting timing, trigger and control applications is designed. The bi-directional, point-to-multipoint nature of the architecture leads to infrastructure efficiency increase. A custom protocol is developed and implemented using FPGAs. It is experimentally verified that the network design can deliver significantly higher data rate than the current infrastructure and meet the stringent latency requirements of the targeted application. Consequently, the design of a network that can be utilised to transmit all types of information at the upgraded LHC, the High-Luminosity LHC (HL-LHC) is discussed. The most challenging requirement is that of the high upstream data rate. As WDM offers virtual point-to-point connectivity, the possibility of using a Wavelength-Division Multiplexed (WDM) PON is theoretically investigated. The shortcomings of this solution are identified; these include high cost and complexity, therefore a simpler architecture is designed. This is also based on the PON paradigm and features the use of Reflective Electroabsorption Modulators (REAM) at the front-end (close to the particle collision point). Its performance is experimentally investigated and shown to meet the requirements of a unified architecture at the HL-LHC from a networking perspective. Finally, since the radiation resistance of optoelectronic components used at the front-end is of major importance, the REAM radiation hardness is experimentally investigated. Their radiation resistance limits are established, while new insights into the radiation damage mechanism are gained.

No description available.

Asynchronous circuits, in comparison with synchronous circuits, have the potential to offer power and speed advantages combined with improved design reuse and composition. Continual improvements in fabrication technology increase die sizes and decrease device sizes, increasing the difficulty of clock distribution and timing validation in synchronous designs. As a consequence there has been a resurgence of interest in asynchronous circuits and design methods. This work examines aspects of asynchronous micropipeline controller design, verification and application. A micropipeline controller circuit is presented and compared with other controller circuits. A method for modelling asynchronous circuits using process algebra at an individual gate level is examined and used to verify the controller circuit. Two applications in the context of the discrete cosine transform (DCT) are then explored. The first application is an area and power efficient circuit for bit serialisation and matrix transposition. This can be used either to embed a synchronous bit-serial processing core into a bit-parallel environment or to perform matrix transposition as part of a DCT. Key elements are modelled using process algebra. The second application is an initial attempt at an asynchronous application specific processor which is used to implement the DCT, and is intended to be extendible to other signal transforms. The presented micropipeline controller was found to be superior to other controllers for linear micropipelines, which are key parts in the applications studied. The modelling method used has been found suitable for the verification of manually designed gate-level circuits. Finally the applications have illustrated that the use of asynchronous methods makes new or simpler architectures possible.

No description available.