Massachusetts Institute of Technology. Dept. of Metallurgy and Materials Science. Thesis. 1969. Sc.D. / Vita. / Bibliography: leaves 83-86. / Sc.D.

Massachusetts Institute of Technology. Dept. of Metallurgy and Materials Science. Thesis. 1971. M.S. / Includes bibliographical references. / M.S.

Massachusetts Institute of Technology. Dept. of Metallurgy and Materials Science. Thesis. 1971. Sc.D. / Vita. / Includes bibliographical references. / Sc.D.

Massachusetts Institute of Technology. Dept. of Metallurgy and Materials Science. Thesis. 1974. Sc.D. / MICROFICHE COPY ALSO AVAILABLE IN SCIENCE LIBRARY. / Vita. / Includes bibliographical references. / Sc.D.

Massachusetts Institute of Technology. Dept. of Metallurgy and Materials Science. Thesis. 1969. Sc.D. / Vita. / Bibliography: leaves 141-145. / Sc.D.

Massachusetts Institute of Technology. Dept. of Metallurgy and Materials Science. Thesis. 1968. Sc.D. / Vita. / Bibliography: leaf 47. / Sc.D.

Massachusetts Institute of Technology. Dept. of Metallurgy and Materials Science. Thesis. 1973. B.S. / MICROFICHE COPY ALSO AVAILABLE IN SCIENCE LIBRARY. / Leaves with plates and diagrams unnumbered (20 leaves). / Includes bibliographical references. / B.S.

The scope of the present study was to demonstrate the capability of Flame Spray Pyrolysis (FSP) process as a unique facility for the one-step synthesis of lithium titanate (Li[sub]4Ti[sub]5O[sub]12, LTO) nanoparticles with tailored properties. FSP offers a versatile technology to produce a wide range of high-purity oxide nanoparticles with desired properties. The ability of FSP to manipulate nanoparticles' properties was demonstrated by controlling operating conditions and selecting appropriate precursors. More precisely, the effect of FSP processing conditions on LTO properties were thoroughly investigated both experimentally in a pilot-scale reactor (production rates up to 1 kg h[sup]-1) and theoretically by the development of models describing particle dynamics in the spray flame. The main aim was to obtain LTO nanoparticles of different particle sizes. The produced nanoparticles were used as active materials for the fabrication of lithium-ion battery anodes and electrochemical characterisation was performed in order to examine the influence of the particles' physical properties on the battery performance. The control of the flame synthesis parameters is crucial, since the properties of the final product depend on the nanoparticles' size distribution, morphology, extent of agglomeration, as well as phase compostition. Initially, the influence of liquid feed properties (precursor concentration and solvent) on LTO physical properties was established. LTO particle size increazsed when the precursor concentration was increased due to particle concentration increase in the flame followed by the enhancement of particle collisions and hence particle growth. Moreover, high precursor concentration caused a variation of physical properties of the precursor mixture, affecting the atomisation process, and subsequently led to the formation of larger droplets. Larger droplets generated larger particle. Additionally, the choice of solvent for the dissolution of metal precursors was proven to be an important issue for LTO synthesis by FSP. The physical properties of the solvents in relation to metal precursor properties affected the formation of LTO nanoparticles. Inhomogenous particle size distribution was observed for LTO synthesised by a precursor mixture containing isopropanoil, due to its low boiling point inducing particle formation via droplet-to-particle mechanism, whereas pure 2-ethylhexanoic acid was used, LTO particles were formed by gas-to-particle route and had homogenous size distribution. The droplets generated during atomisation by the precursor solution of pure 2- ethylhexanoic acid had the largest diameter due to the high viscosity and density of the mixture. Despite this, the obtained nanoparticles were the smallest in comparison to those obtained from other precursor solutions. In such cases, the boiling point and specific combustion enthalpy of the solvents should be taken into consideration. Apart from the liquid feed properties, the effect of FSP operating conditions (O[sub]2 dispersion gas and precursor flow rate) were also investigated in the present study. By increasing the O[sub]2 dispersion gas rate, LTO nanoparticles' diameter was decreased due to a decrease of the droplet diameter. Particle sintering was also prevented due to the faster transport of primary particles through a shorter flame. An increase of the precursor flow rate at first increased and then decreased the LTO nanoparticles' size. The initial increase of particle size occurred due to a flame temperature increase. At higher precursor flow rates, the droplets disintegrated and generated many smaller fragmented droplets due to higher temperature, which subsequently formed smaller particles. Moreover, particle growth in the spray flames was studied theoretically, and numerical models were developed. The monodisperse model developed assumed that all primary particles had the same size. However, it overestimated the primary particle diameter values. Polydispersity was taken into consideration in the development of an additional model which was solved by the quadrature method of moments. The results obtained from the polydisperse model were closer to the experimental values, both for low and high production rates. Finally, the synthesised LTO nanoparticles were used as active materials in lithium-ion battery half cells and their electrochemical behaviour was elucidated, demonstrating the effect of the particles' physical properties on their electrochemical performance. LTO of particle size 18 and 21 nm showed the best electrochemical performance with capacity retention of almost 100% after 500 cycles, whereas the smallest particle deteriorated the electrochemical performance with a capacity loss of more than 60%.

This thesis is focused on employing the finite element method (FEM) to simulate hot flat rolling process. The relevant work involves selecting a suitable constitutive equation, predicting the rolling load and roll torque, computing temperature changes and lateral deformation, simulating the evolution of substructure, modelling static recrystallisation and designing the rolling pass schedule. A practical pass schedule supplied by an aluminium company and containing reliable measured data of roll load and torque is analysed by a commercial 3-D thermornechanical coupled FEM program FORGE3 V5.3. The inverse analysis method is adopted to obtain the friction coefficient and heat transfer coefficient. The distribution of pressure, equivalent strain, the stress and damage in the roll gap in breakdown rolling are discussed. The changes of temperature and lateral profile under both laboratory and industrial rolling conditions are computed and compared with experimental measurements, the differences are then investigated. Through applying the Taguchi experimental design method, the influence of each rolling parameter on the spread, i. e. the ratio of width to thickness, the roll radius to thickness, the thickness reduction, and the deformation temperature, the relative contribution of each control parameter is quantitatively estimated and expressed as a percentage. A new spread formula is built up based on a large amount of FE analyses. The new formula is able to deal with both laboratory and industrial rolling conditions with high accuracy. Critical reviews are presented for the previous work in the modelling of subgrain size and static recrystallisation. Both empirical and physical models are applied to investigate the evolution of subgrain size, dislocation density, misorientation and the flow stress in the roll gap. The predicted subgrain size agrees very well with the experimental measurement. The difference between the use of two models are illustrated and analysed. Studies on modelling of static recrystallisation are carried out by incorporating the plastomechanical parameters, i. e. strain, strain rate and temperature, into empirical model. Various approaches are proposed to reduce the predicted volume fraction recrystallised at the surface and are verified by the comparison with measurement. Simulation results show that some of the previous work reported in the literature are erroneous. Further work in the modelling of static recrystallisation and texture evolution is detailed. The Taguchi experimental method is also applied to study the influence of the rolling parameters on the fraction recrystallised (Xv ). The study shows that rolling temperature has the greatest influence on the Xv, followed by the parameter H. 1L. The roll temperature and roll speed have little influence on the Xv. Designing a rational rolling pass schedule is critical for the control of strip profile and product quality. In the present thesis, the procedure of designing a rolling pass schedule is illustrated. The formulae used in scheduling are listed and explained. The scheduling program is then performed to check with two existing industrial schedules. The comparison shows that the rolling load, temperature and power model is reliable and shows high accuracy. A multipass simulation by the use of finite element method is also carried out and the results are compared with various model predictions. The problems in the simulation are illustrated and explained.

Damage accumulation in glass fibre woven reinforced epoxy laminates manufactured from two different fabrics have been investigated under three different loading conditions. One of the woven fabrics was non-hybrid glass using E-glass fibre yams in both warp and weft (fill) directions, the second fabric was a hybrid woven fabric using E-glass fibre yams in the warp direction and R-glass fibre yams in the weft direction. Destructive tests such as interlaminar shear, flexural and uniaxial tension tests were carried out on two different categories. In the first category four different fibre volume fractions of nonhybrid E-glass woven fabric reinforced epoxy resin laminates have been investigated. In the second category hybrid and non-hybrid woven fabric reinforced epoxy resin laminates for similar fibre volume fractions have been investigated. Acoustic Emission (AE) and Scanning Electron Microscopy (SEM) were employed as nondestructive tools to predict and characterise the damage events in the composites. All laminates were fabricated using the wet hand lay-up process to laminate the fabric layers prior to curing. Epoxy resin (L20-SL set) was the sole matrix used for all composites. Test results showed higher mechanical performance for the hybrid composites and improvements in mechanical properties for higher fibre volume fraction in the non-hybrid composites. DMTA tests were carried out on the laminates of the categories mentioned above, the test results indicated the effect of fibre surface treatment concentration on the performance of mechanical properties of woven composites. DMTA data has been used to correlate the results of ILSS, flexural and tensile tests. A model was developed in this study based on the damage event sequential process of glass woven fabric reinforced epoxy resin composites. The model is an experimental analysis model, supported by DMTA, AE, SEM and visual examination of specimen fracture surface.