Search

Alzheimer’s disease (AD), the most common cause of dementia, affects more than 520,000 people in the UK. It is a progressive disease and its causes still remain unclear. Besides the traditional neuropsychological test, imaging biomarkers are playing an increasingly important role for AD detection and especially for early stage diagnosis including prediction of the transition from mild cognitive impairment (MCI) to AD. The use of structural imaging such as T1 MRI for AD diagnosis has been widely explored, however, the application of PET imaging to AD is relatively novel, with new tracers appearing. In addition, the development of machine learning and deep learning algorithms can help researchers and clinicians build computer-based image analysis tools, which can be used to further enhance the AD diagnosis results by utilising image features that cannot be easily interpreted directly by human eyes. This thesis focuses on the development of such image analysis methods for the prediction of MCI to AD conversion. We first propose the use of novel features, based on complexity measurement, from volumetric amyloid PET images, which we use as input to a nonlinear Support Vector Machine (SVM) classifier. We tested this method using the novel PET features alone or in combination with standard PET metrics and MRI-derived features, obtaining state of the art accuracy. Our second main contribution involves the development of a deep learning model using two parallel 3D Convolutional Neural Networks (CNNs) to combine the information from amyloid PET and T1 MRI. Deep learning methods in this area are typically limited by the small size of available PET datasets. We show that prediction accuracy is improved by combining “diagnostically similar” datasets together as a form of augmentation (progressed MCI + pAD against stable MCI, as well as progressed MCI + pAD against stable MCI + healthy volunteer). Our third major contribution explores a new method for augmentation using data synthesis through a conditional Generative Adversarial Network (cGAN), which can generate new amyloid PET images from volumetric T1 MRI scans, for which there is much larger availability. The generated amyloid images are used to augment the original progressed MCI dataset to address the problem of data imbalance. The results showed that the accuracy for MCI to AD prediction using generated data augmentation method outperformed traditional augmentation methods substantially.

This dissertation investigates the growth and magnetic properties of magnetic thin films deposited on semiconductor GaAs and the insulator MgO, which could be useful for devices such as Spin-FET and MRAM. CoFeB amorphous films were grown on both GaAs and MgO. We have studied the origin of the uniaxial magnetic anisotropy (UMA) and perpendicular magnetic anisotropy (PMA) with TEM, VSM and XMCD. Our results demonstrated that the orbital moment of Co atoms play an important role to both UMA and PMA. The origin of UMA in Fe/GaAs (100) system with Cr interlayers is explored. The values of UMA in the Fe/GaAs systems were found to be dependent on the thickness of Cr interlayer by the SQUID-VSM measurements. RHEED patterns and TEM images offered the morphology and crystalline structure information of different layers in the samples. Our results show that the UMA disappears when the interlayer Cr forms continuous film of around 5 ML. This offers direct evidence for the first time that the origin of UMA is from the interface bonding rather than the lattice mismatch related film stress. Finally, Fe films were deposited onto GaAs (100) substrate with a heavy metal element Au interface layer. The enhancement of UMA was found in the Fe/Au/GaAs system by the VSM measurements. The XMCD results show that the orbital moment of Fe is enhanced when the Au interlayer is under 0.5 ML, which leads to the enhancement of the UMA in the Fe/Au/GaAs system. Results from the three different systems provide an important understanding of the research into the interface magnetic properties of ferromagnetic metal/semiconductors. These interesting discoveries are very useful for the development of next generation electronic devices like MRAM and SpinFET.

Since the introduction of medical implants into human bodies, corrosion and wear have been regarded as key issues for their long-term durability. There has been a recent renewed interest in the use of large diameter metal-on-metal (MoM) hips, primarily because of the reduced volumetric wear compared with the wellestablished polyethylene-on-metal joints. Long term durability of MoM joints relies on control of both their corrosion resistance (relating to ion release) and wear behaviour (relating to creation of nanometre-scale wear debris). Concerns about the potential risk of released metal ions to the biological environment (patient) are of great importance. In this respect tribocorrosion is a serious consideration in joint performance. An integrated electrochemical cell on a reciprocating tribo-meter was employed to evaluate the corrosion and tribocorrosion behaviour in a protein rich solution (Bovine Serum), a cell culture solution (DMEM) and a saline solution (NaCI) with the attempt to isolate the organic species effects. Three commonly used orthopaedic materials were involved in this study. A High Carbon Cobalt-Chromium-Molybdenum alloy, a Low Carbon Cobalt-Chromium-Molybdenum and UNS S31603 Stainless Steel (316L). A range of electrochemical methods were used in the assessment of materials under biotribocorrosion systems and results were supported by surface analysis and bulk solution analysis techniques. The material degradation rate is strongly dependent upon the charge transfer (corrosion), the mechanical damage (tribology) and also their interactions (tribocorrosion) in these simulated biological environments.

This thesis describes an experimental and theoretical investigation of the effects of an electric field on two kinds of phase change process in industry: nucleate boiling in heat exchangers and crystallization from industrial melts. The first is called electrohydrodynamic (EHD) enhancement of nucleate boiling heat transfer at low superheat surfaces in heat exchangers. The major emphasis of the work has been to probe the compound effect of EHD and passive enhanced sur-faces on nucleate boiling. Field potential distribution for six different surfaces has been predicted and analyzed using a computational fluid dynamics (CFD) method. The analysis has shown that the shapes of passive enhanced surfaces potentially create considerable increases in local field gradient to produce such favourable con-ditions. The critical bubble radius, bubble nucleation rate and bubble movement in the presence of an electric field have been also studied and discussed. Equations are derived showing that a local electric field gradient will enhance nucleation rate and decrease critical bubble radius, both leading to improved heat transfer rate. In the mean time, experimental work for two typical low superheat surfaces, Hitachi Thermoexcel-HE and Wieland Gewa-T tubes was carried out using a single tube test rig. Furthermore, comparisons for current experimental results and those for other surfaces are made. Major improvements are shown in the performance of heat exchangers fitted with enhanced tube surfaces by applying up to 30 kV between those surfaces and a surrounding electrode. In parallel with this, the effect of an electric field on melt crystallization has been studied. The experimental samples were selected from industry; the simplicity of laboratory preparation, the dielectric properties and the polymorphic behaviour were typical and attractive for the research purpose. A large number of experiments under different conditions for the effects of an electric field on the crystallization of a dielectric with polymorphic behaviour and lower molten temperature has been carried out for the first time. The experiments are assessed by time-temperature measurement and differential scanning calorimetry (DSC), respectively. The results have shown that the crystallization of the melt on cooling is enhanced under the action of an electric field. For the effect of an electric field on the melt mixture, a novel approach of applying the Taguchi method to the experimental DSC results has been used. Also, a theoretical analysis for dielectric crystallization in the presence of an electric field has been carried out.

Reactive separation is proposed as a less energy-intensive chemical process with functions of catalytic reaction and separation. It offers an active chemical conversion combined with selective separation to obtain the desirable products. Catalytic membranes are one of the successful examples and the design of it requires understanding the structure-activity relationship between catalyst and membranes. To guide the process design, this thesis gives computational insight on the following aspects: catalyst design focused on selectivity and activity in aqueous environments, and a porous structure suitable for incorporating catalytic active sites and responsible for separation. For catalyst design, the aim is to understand the role of the local chemical environment on reaction activity and selectivity. It is promising to tune the reaction activity by manipulating its solvation environments and coordination environment for the catalyst. By understanding the interplay between catalytic performance and the local chemical environment, a catalytic active site with a well-defined structure is proposed for further catalytic membrane design. Additionally, a well-established membrane for gas separation is studied to characterize cavity formation with tunable membrane contents. The study encompasses fundamental aspects of catalysis, including electronic structure, kinetic principles, and temperature-dependent separation processes, while analyzing detailed polymer structures. Firstly, the aldol condensation reaction of cyclopentanone (CPO) over metal oxide MgO is investigated. Functionalized groups grown on an oxide catalyst highlights the tunable kinetic performance and promotional role of water in association with functionalized groups. Density Functional Theory (DFT) calculations show the strong adsorbate-adsorbate interaction results in hindered C-C coupling step, which contributes to low reaction rates observed on the functionalized surface. The role of water is identified in facilitating CPO adsorption and reducing the kinetic barrier of the kinetic-relevant step. Water thereby promotes the reaction rate by alleviating serious adsorbate-adsorbate interactions on the functionalized surface. This study paves a way to understand the solvent effect on the heterogeneous catalyst. With an emphasis on denitrification in wastewater treatment, palladium has been proposed as an active catalyst for wastewater treatment, but the detailed mechanism and the solvent effect are not well understood. To address this, DFT calculations were employed to investigate the detailed reaction mechanism and kinetics for nitrite reduction. This study revealed that the overall activity is primarily determined by the activity of NO and a comprehensive kinetic model is developed to explain the observed reaction order. Interestingly, water played a promotional role in the nitrite hydrogenation reaction over the palladium catalyst. DFT analysis demonstrates that water facilitates a lower-energy pathway for hydrogenation by shuttling surface hydrogen to the reactants. The tunned free energy pathway aligns with the experimental findings on the reaction order. The solvent effect is associated with the micro-solvation environments created by a polymer (n-isopropyl acrylamide) overlayer that shows temperature-dependent behavior. Growing this NIPAM overlayer on the catalyst support alters the selectivity and activity of palladium along with the tuned solvation environment. Molecular dynamics simulations propose a lower activation barrier introduced by NIPAM due to polymer-water interactions. The presence of polymer in the aqueous phase forms a structured water layer on the palladium, favoring the reaction pathway through the proton-shuttling mechanism. These findings enhance our understanding of designing and manipulating the micro-solvation environments for a tunable solvent effect. As revealed on metal catalysts, the binding energy of NO provides crucial insights to promote reaction activity and selectivity. The binding energy determines the activation of NO bond hydrogenation is highly associated with the structure of the active sites. The investigation extends to the activity of single atoms in nitrite reduction powered by renewably generated electricity and examines the denitrification mechanisms across dispersed Co and La-based catalysts. The enhanced charge transfer is identified on NO bonded with La. and different mechanisms are investigated including entropic stabilization and the solvent effect. The study suggests that atom-dispersed metal is one of the promising catalysts toward a high activity and selectivity of NO reductions which dominates the denitrification reaction. The catalytic membrane requires a stable polymer structure to selectively filter out the ions. We conducted a simulation on membrane synthesis using a triptycene-based membrane (TPBO) with tunable cavity distribution based on the content of triptycene. By manipulating the content of the triptycene unit, the free volume distribution corresponding to the triptycene molecule content is obtained. These structure-properties relationships reveal the flexibility of TPBO-based polymers for gas separation. Our study on polymer structure provides a foundation acknowledge for further membrane design which is critical to catalytic membranes. Overall, this dissertation suggests a synergetic combination that takes advantage of polymer structures and single-atom catalysis to achieve enhanced activity and selectivity in reactive separation. This approach holds great promise for advancing catalytic technologies and addressing challenges in fields such as separation science and environmental sustainability. This study on catalysis research has been extensively focused on understanding the fundamental principles governing catalyst performance, with a growing recognition of the significance of the local chemical environment. Fine-tuning and manipulating this environment presents novel avenues for catalyst design.

Renewable energy adoption is a popular topic to release the stress of climate change caused by greenhouse gas. Electricity is ideal secondary energy for clean primary energy such as nuclear, wind, photovoltaic, and so on. To extend the application of electricity and reduce fossil energy consumption by transportation sectors, electric vehicles (EVs) become promising technology that can further inspire the development of renewable energy. Battery as the core in an EV provides the energy to the motor and all on-board electric equipment. The battery charger is mainly composed of a power factor correction (PFC) and isolated DC-DC converter. Therefore, power electronics equipment plays an important role in automotive products. Meanwhile, in recent years, the market capacity for wide band-gap devices, SiC MOSFET, continues to increase in EV applications. Dual active bridge (DAB) is an excellent candidate for isolated DC-DC converter in EV battery charger. The characteristics include an easy control algorithm, galvanic isolation and adjustable voltage gain. Different modulation strategies are developed to improve the performance and stability by using multiple phase shift (MPS) control. This thesis focuses on the utilization of different modulation strategies to realize smooth transition among MPS control in full operational range with securing zero-voltage-switching (ZVS) to eliminate the crosstalk in the hard-switching process. The influence of MPS control on ZVS resonance transient is also addressed to find out the accurate minimum required energy of the inductor to finish the ZVS transition. Furthermore, a general common-mode voltage model for DAB is proposed to analyze the impact of MPS control on the common-mode performance.

Over the period 1980 – 2015, foreign direct investment (hereafter FDI) has become more significant in connecting East Asia and the global market. Meanwhile, income inequality has been growing in this region. Although existing literature has achieved noticeable progress in identifying the inequality-inducing effect of FDI inflows both theoretically and empirically, evidence is still inconclusive. This dissertation contributes to the scholarship by providing new time-series evidence from East Asia confirming the inequality-inducing effect of FDI inflows. But this dissertation makes another important contribution by introducing FDI policies to the scholarship. By employing the nested analysis approach, this dissertation sheds light on how FDI policy liberalization conditions the distributional effect of FDI inflows. Time-series cross-section analyses indicate that FDI inflows have deteriorated income inequality in East Asia, and this relationship is conditional on FDI policy liberalization. Liberalized FDI policies intensify the inequality-inducing effect of FDI inflows. Building on case studies on China and Korea, I demonstrate that the conditional effect of liberalized FDI policies can be achieved via two channels: regional distribution and sectoral distribution of FDI inflows. The Chinese case shows that an increasingly equal regional distribution of FDI inflows due to policy liberalization leads to rising income inequality. The Korean case indicates that when liberalized policies open more economic sectors to foreign investors, FDI inflows become more skewed toward service sectors, also resulting in growing income disparities. Thus, the positive relationship between FDI inflows and income inequality is not merely an economic phenomenon, but rather a joint product of political and economic forces. As FDI policy liberalization becomes a dominant policy paradigm, and FDI inflows continue to be a significant way of economic integration between East Asia and the world, income gaps are likely to further deteriorate in this region. How to make FDI inflows more beneficial to society has become a pressing issue for the governments in this region.

Field permeability test using NCAT Permeameter were conducted at four project sites before and after the application of fog seal. Due to the fact that it is difficult to estimate the thickness of pavement tested, instead of permeability, the infiltration rate is calculated to estimate the effectiveness of fog seal on reducing the permeability of pavement.;Cores from I-79, I-64 and Route-19 projects were collected and sent to the WVUATL for bulk specific gravity (Gmb) measurements. According to AASHTO T 166, when the percent of water absorbed by the specimen exceeds 2%, AASHTO T 275 or T 331 should be used to determine the bulk specific gravity. There is a proposal to reduce this threshold to one percent which is supported by the data and analysis in this study. Paired student test, and a line of equality regression analysis were used to analyze the data provided by the contractor and collected in the WVUATL, the AASHTO T166 is proven to be repeatable, and CoreLok is a viable device for determining the Gmb and corresponding air voids of field cores. Based on density data from the I-79 project, provided by the contractor and WVDOH, the thin-lift gauge had smaller correction factors than standard nuclear gauges. Developing correction factors based on the ration of means may be better than using the difference of means methods which is currently used by the WVDOH.;Laboratory permeability tests were conducted on all field cores following Florida Testing Method 5-565. In general the permeability of the field cores on this project is consistent with the rule of thumb that permeability becomes an issue when the air voids of the sample are greater than eight percent. The regression model of the field cores permeability indicates that at eight percent air voids the permeability is about 80x10-5 cm/sec which is less than the recommended criteria of 100x10-5 cm/sec to 125x10-5 cm/sec.;In addition, eight HMA mixtures including 9.5 mm fine and coarse, 12.5 mm coarse, 19 mm fine and coarse, 25 mm fine, and 37.5 mm fine and coarse, were evaluated. For each mixture, gyratory compacted samples were produced and tested at target air voids contents of 4, 7, 9 and 11 percent, with two samples at each target. Laboratory permeability tests were performed to determine how air voids content, nominal maximum aggregate size, and aggregate gradation affect the permeability of HMA mixtures. The model developed by Vivar et al. was found to be suitable for data collected in this study, and new coefficients were suggested. The results of multi regression analysis indicated that air voids and nominal maximum aggregate size significantly affect the permeability, but aggregate gradation does not.;Comparing the permeability data from field cores and laboratory compacted pills, it is clear cores have higher permeability than laboratory compacted pills when the air voids are lower than approximately 10 percent, and the pills have a higher permeability at more than 10 percent air voids. This observation is consistent with previous research.