The feasibility and development of a field emission based anisotropic vector magnetometer is presented. Within this scope current magnetic sensing technology is investigated and compared. The advantages of, and need for, a field emission based magnetic sensor are then discussed. Background theory, simulation, fabrication, testing, and future developments of field emission magnetometers are presented. The possible applications of field emission to other technologies are also investigated. The magnetic sensing device presented uses a sharp field emitting tip with a radius of the order of 100nm which is fabricated using standard silicon processing techniques on highly n-doped silicon. Under a vacuum level of 10"6 mBar and at room temperature, a potential applied to a surrounding gate electrode extracts from this tip a beam of electrons which is incident upon two separate anode electrodes. In the absence of an external magnetic field the electron current incident on each of these two electrodes is equal, while in the presence of a magnetic field the Lorenz force skews the beam towards one of the electrodes, resulting in a differential current which is proportional to the magnetic field.

Topological analysis of the Internet is needed for developments on network planning, optimal routing algorithms, failure detection measures, and understanding business models. Accurate measurement, inference and modelling techniques are fundamental to Internet topology research. A requirement towards achieving such goals is the measurements of network topologies at different levels of granularity. In this work, I start by studying techniques for inferring, modelling, and generating Internet topologies at both the router and administrative levels. I also compare the mathematical models that are used to characterise various topologies and the generation tools based on them. Many topological models have been proposed to generate Internet Autonomous System (AS) topologies. I use an extensive set of measures and innovative methodologies to compare AS topology generation models with several observed AS topologies. This analysis shows that the existing AS topology generation models fail to capture important characteristics, such as the complexity of the local interconnection structure between ASes. Furthermore, I use routing data from multiple vantage points to show that using additional measurement points significantly affect our observations about local structural properties, such as clustering and node centrality. Degree-based properties, however, are not notably affected by additional measurements locations. The shortcomings of AS topology generation models stems from an underestimation of the complexity of the connectivity in the Internet and biases of measurement techniques. An increasing number of synthetic topology generators are available, each claiming to produce representative Internet topologies. Every generator has its own parameters, allowing the user to generate topologies with different characteristics. However, there exist no clear guidelines on tuning the value of these parameters in order to obtain a topology with specific characteristics. I propose a method which allows optimal parameters of a model to be estimated for a given target topology. The optimisation is performed using the weighted spectral distribution metric, which simultaneously takes into account many the properties of a graph. In order to understand the dynamics of the Internet, I study the evolution of the AS topology over a period of seven years. To understand the structural changes in the topology, I use the weighted spectral distribution as this metric reveals differences in the hierarchical structure of two graphs. The results indicate that the Internet is changing from a strongly customer-provider oriented, disassortative network, to a soft-hierarchical, peering-oriented, assortative network. This change is indicative of evolving business relationships amongst organisations.

In order to support a wide variety of services, to different user types, and under a variety of geographic situations, telecommunications networks are typically composed of a variety of layers and heterogeneous technologies. Layers (in terms of the OSI 7 layer model) such as the transmission layer (e.g. WDM), the data link layer (also known as the transport network e.g. SDH, Ethernet) and the network layer (e.g. IP). These layers may also contain logical layers within them such as virtual paths, as well as overlay networks such as a peer-to-peer system. No single layer is independent of the adjacent layer and the provisioning requirements of one layer become the demand on the layer below. Similarly the available resources become the delivered quality of service to the layer above. This thesis is concerned with the design aspects of various layers and how they affect each other’s topology. The thesis’ main focus is topological analysis and modelling of layers, and its presents a detailed analysis of a deployed national SDH network, examining bandwidth distribution, topology, geography and the demand pattern. The thesis finds that even the strictly planned and provisioned SDH network, whose architecture contain explicit structures and hierarchy, has notable power-law traits in various metrics of the topology; traits similar to those which have been shown to exist in the Internet, as well as non-technological networks such as social graphs. There is also and examination of the protocols and architectures of the IP and SDH standards for features that affect topological development. With a better understanding of the layers, design goals and assumptions are deduced and implemented in a new topology simulator called MITIE. MITIE (Modular Inter-layer feedback Topology InvEstigation tool and simulator) is a tool designed to investigate inter-layer feedback and differs from existing topology generator in that it considers the effect of serviced demands and allows the capacity usage to affect the further development of the topology. The thesis presents results from a series of experiments with MITIE and demonstrates that as the network is re-designed to accommodate demand, it can tend to power-law compliant topologies under the correct circumstances. Such a reactive topology model could also be used to investigate the effect of topological change and the effect of increasing the number of layers (such as adding MPLS), or the use of peer-to-peer overlay networks, or the decrease of the number of layers (IP over WDM). The model could also be used to investigate link and node failure/addition and the real effect which will propagate through the rest of the multi-layer network.

Bipolar transistors are known to be the most suitable for high-voltage and power applications, due to their inherently greater current handling capability. In contrast, MOS technology is preferable for logic applications, due to its superior packing density. Therefore the 'ideal' solution to the smart power problem of integrating control elements on the same die as power switches is a marriage of the two different technologies. This results in a complex process that can only be cost effective in high volume applications. For ASIC applications and low volume product runs a less expensive compromise solution is needed. By analyzing both bipolar and MOS, low and high voltage devices, it was found that if more than one power transistor is needed on the circuit, and a single technology is to be used, then MOS power transistors are inherently easier to integrate into a low voltage process. In particular the lateral double-diffused transistor (LDMOS) with all terminal contacts on the surface is to be preferred. Analyzing a CMOS process, common processing steps were found for both the low and high-voltage devices, leading to a smart power solution that doesn't need many masking levels. By making small changes to an established n-well CMOS process, and developing a novel power transistor structure with a field oxidation separating the channel and drain, a 120 Volt n-channel power transistor could be realised within a conventional process with no additional processing steps. By adding one further masking layer, a complementary p-channel power transistor that supported -55 Volts could be fabricated. If these transistors were fabricated on a p- epitaxial layer on an n- substrate then by changing the p-channel power device structure, a breakdown voltage of -95 Volt could be achieved using only nine masking layers.

This thesis describes the design fabrication, measurement and analysis of asymmetric lightly doped drain transistors. The transistors were fabricated so that adjacent transistors in columns were of a slightly different but determinable asymmetry. This approach has been termed the progressional offset technique. The progressional offset technique has the advantage in that it can be used to fabricate transistors with different lightly doped drain lengths the same wafer. The resulting devices consist of a range of transistors with different degrees of asymmetry which can be quantified. The thesis shows that the asymmetry of these devices can be determined by simple electrical measurements. The progressional offset technique was used to show that an asymmetric lightly doped drain implant can have a large effect on the electrical properties of the devices. This occurs when the lightly doped drain implant is shadowed by the polysilicon gate during the angled implantation which is routinely used by industry. The progressional offset technique is a useful tool for analysing the effects of asymmetry on LDD transistors. It also has been shown to produce transistors with different LDD lengths on a single wafer.

Distributed-memory multicomputers will not find acceptance outwith the specialised field of high-performance numerical applications until programs written for them using systems similar to those found on conventional uniprocessors can be run with an efficiency comparable to that achieved on those uniprocessors. This work argues that the key to constructing upwardly-compatible programming systems for multicomputers based on message passing which are both efficient and usable, and which allow effective monitoring, is to require those systems to be structured, in the same way that modern programming languages require programs to be structured. It is further argued that even a well-structured message-passing system is too low-level for most applications programming, and that some more abstract system is required. The merits of one such abstraction, called generative communciation, are considered, and suggestions made for enriching standard implementations in order to improve their usability and efficiency. Finally, it is argued that the performance of any programming system for distributed-memory multicomputers, regardless of its degree of abstraction, is largely determined by the degree to which it eliminates or avoids contention. A technique for doing this, based on opportunistic combining networks, is introduced, and its effect on performance investigated using simulations.

In turbulent premixed flames the modelling of mean or filtered reac tion rate translates to the closure of flame surface to volume ratio, which is commonly referred to as the Flame Surface Density(FSD). The FSD based reaction rate closure is well established in the context of Reynolds Averaged Navier-Stokes (RANS) simulations for unity Lewis numbers. However, models for FSD in context of Large Eddy Simulations (LES) are relatively rare. In this study three-dimensional Direct Numerical Simulations (DNS) of freely propagating statistically planar premixed flames encompassing a range of different turbulent Reynolds numbers and global Lewis numbers was used. The variation of turbulent Reynolds number has been brought about by modifying the Karlovitz and the Damkohler numbers independently of each other. The DNS data has been explicitly Reynolds averaged and LES filtered for a prior assessment of existing FSD models and for the purpose of proposing new models where necessary.

The use of Renewable energy such as wind power has grown rapidly over the past ten years. However, the poor reliability and high lifecycle costs of wind energy can limit power generation. Wind turbine blades suffer from relatively high failure rates resulting in long downtimes. The motivation of this research is to improve the reliability of wind turbine blades via non-destructive evaluation (NDE) for the early warning of faults and condition-based maintenance. Failure in wind turbine blades can be categorised as three types of major defect in carbon fibre reinforced plastic (CFRP), which are cracks, delaminations and impact damages. To detect and characterise those defects in their early stages, this thesis proposes eddy current pulsed thermography (ECPT) NDE method for CFRP-based wind turbine blades. The ECPT system is a redesigned extension of previous work. Directional excitation is applied to overcome the problems of non-homogeneous and anisotropic properties of composites in both numerical and experimental studies. Through the investigation of the multiple-physical phenomena of electromagnetic-thermal interaction, defects can be detected, classified and characterised via numerical simulation and experimental studies. An integrative multiple-physical ECPT system can provide transient thermal responses under eddy current heating inside a sample. It is applied for the measurement and characterisation of different samples. Samples with surface defects such as cracks are detected from hot-spots in thermal images, whereas internal defects, like delamination and impact damage, are detected through thermal or heat flow patterns. For quantitative NDE, defect detection, characterisation and classification are carried out at different levels to deal with various defect locations and fibre textures. Different approaches for different applications are tested and compared via samples with crack, delamination and impact damage. Comprehensive transient feature extraction at the three different levels of the pixel, local area and pattern are developed and implemented with respect to defect location in terms of the thickness and complexity of fibre texture. Three types of defects are detected and classified at those three levels. The transient responses at pixel level, flow patterns at local area level, and principal or independent components at pattern level are derived for defect classification. Features at the pixel and local area levels are extracted in order to gain quantitative information about the defects. Through comparison of the performance of evaluations at those three levels, the pixel level is shown to be good at evaluating surface defects, in particular within uni- directional fibres. Meanwhile the local area level has advantages for detecting deeper defects such as delamination and impact damage, and in specimens with multiple fibre orientations, the pattern level is useful for the separation of defective patterns and fibre texture, as well as in distinguishing multiple defects.

This thesis describes four approaches developed towards the fabrication of conductive 1-dimensional copper nanostructures, or nanowires, using a DNA-templating strategy. Cu-DNA nanowires are of interest for miniaturised interconnects in microprocessors. The chemical identity of the nanowires was characterised using X-ray Photoelectron Spectroscopy (XPS) and powder X-ray Diffraction (XRD) methods. Structural investigations were performed using Fourier-Transform Infra-Red (FTIR) Spectroscopy n+ and Atomic Force Microscopy (AFM) to study the Cu -DNA interaction mode(s) and nanowire size/morphologies, respectively. The electrical properties of nanowires were elucidated using Electrostatic Force Microscopy (EFM), Conductive-AFM (C-AFM) and a two-probe semiconductor device analyser for recording current-voltage (i-V) relationships. One method describes a non-aqueous route to the formation of 1-D Cu nanostructures. + This was achieved by doping of surface-immobilised DNA (the template) with Cu ions from Cu(CHCN).PF followed by chemical reduction to the zero-valent metal using 346 phenylsilane and organic solvent, under inert conditions. Metallic copper was confirmed to have formed on DNA and copper hydroxide (Cu(OH)) was also identified as a 2 surface overlayer on this material. The final structures were ~6 nm in height and show complete coverage of the template. The material was polycrystalline due to observations of a densely packed series of Cu nanoparticles along the template axis. However, electrical studies indicated these structures to be highly resistive. Thermal annealing of the templated material resulted in a considerable structural transformation, from tightly packed particles to the formation of a sporadic array of larger clusters well separated along the structure length. 2+ Another approach involved an aqueous solution-based synthesis, whereby Cu from Cu(NO) and ascorbic acid were added to a solution of DNA. This resulted in 32 nanowires, ~7 nm in height, which were mostly continuous in morphology and 2+ noticeably absent of inter-particle boundaries. The Cu :DNA(phosphate) ratio (n) was found to be critical for the formation of smooth nanowires (n= 0.05) as opposed to aggregated assemblies of material (n> 0.1). Chemical characterisation of the reaction product confirmed the presence of metallic copper as well as a surface layer of Cu(OH) Nanowires were confirmed to be conductive and i-V measurements gave a2. conductivity value of 0.94 Scm ; the first recorded conductance for copper templated on DNA, over micron length scales. Attempts were made to passivate the nanowires formed in solution by attachment of thiol molecules (p-mercaptobenzoic acid) onto the copper surface. This modification 0 resulted in the formation of a Cu(I)-thiolate layer on Cu and significantly protected the nanowires from oxidation, resulting in the formation of almost pure metallic Cu -1 nanowires. The electrical conductivity (1.01 Scm ) was similar to that obtained for the -1 unprotected wires (0.94 Scm ). This suggested that, although the degree of oxidation was minimised significantly, other factors must be more accountable for causing resistance in these wires such as surface and/or grain boundary scattering. Finally, a physical-based approach towards metallisation of DNA was performed using a vacuum deposition method. This was achieved by suspending single molecules of DNA between two electrodes, across a trench etched into the substrate. This was followed by metallisation of the sample by evaporation of copper in a sealed vacuum chamber. Electron Microscopy data showed that nanowires lay taut across the trench and were highly continuous in coverage. This process resulted in a working two- terminal nanowire device which was conductive due to nanowire(s) bridging between insulating gaps.

In Sensorless speed drives, which employ a model reference adaptive system (MRAS) with a PI based adaptation mechanism, the gains of the adaptation mechanism play an important role into the general performance of the drives. The higher the gain, the faster the response and the more robust the drive is going to be against load disturbance. Although it is desirable to utilise an adaptation mechanism with high PI gains, it was demonstrated that high gains will cause the estimated speed to exhibit a high level of noise. More recently it was identified that the generated noise consists of high-order harmonics. As far as the literature is concerned, even though the high-order harmonic phenomenon was identified and addressed by few researchers, there was no description in their literature of the generated high-order harmonics or an assessment of their effect in terms of being problematic. Therefore, the aim of this research is to not only investigate the generation of high-order harmonics, but also to establish whether any generated noise in the estimated speed bear an effect on the overall speed estimation process. An MRAS based speed estimator is implemented to calculate the rotor flux- linkage and the speed estimates required in achieving field orientation and establishing speed control. An investigation has been carried out to gauge how the PI controller gains can influence the speed estimation process of the drive system. As well as examining the estimated speed for any trace of excessive noise and harmonics generation. It is revealed experimentally that no high-order harmonics were generated while implementing an adaptation mechanism with high PI gains. However, it was found that high PI gains do causes the estimated speed to become relatively noisy and also starts to carry some fundamental frequency components relevant to the stator’s electrical frequency. Therefore a programmable adaptive adjustment mechanism has been successfully developed not only to avoid the excessive generation of noise but also improve the speed estimation process. The performance of both of the static adaptation mechanism and the proposed adaptive one was assessed and compared.