Thesis (Ph. D.)--University of Missouri-Columbia, 1998. / Typescript. Vita. Includes bibliographical references (leaves 107-111). Also available on the Internet.

Sintering, a process of forming dense solid bodies from powder compacts remains the most important route for processing of ceramics. The process of sintering involves formation and growth of necks during initial stage, coarsening, relative particle rotation, filling of connected pores in intermediate stage, filling of isolated pores during final stage sintering and rapid grain growth towards the end of densification. The processes involve a combi-nation of grain boundary diffusion, surface diffusion, grain boundary migration and grain boundary sliding. Studies of interfacial processes during sintering are still of interest since modifying interface structure offers a means to tailor low and high temperature mechanical properties of ceramics. Many of the studies in literature on single phase systems are based on geometric changes during sintering. Sintering has been modelled as 1D or 2D array of spheres. The simplest of these consist of a contacting pair of spherical particles. Early models studied changes in size and shape of the necks during initial stage sintering and associated mass transport mechanisms. There have been studies on coarsening that report shrinkage rates of smaller particles is a system of two particles with different radii. In both the cases of neck growth and coarsening, thermodynamic variables as given by dihedral angle (relative grain boundary to surface energy of the system) and kinetic parameters of grain boundary surface diffusivity have been found to influence the size and shape evolution with time. Also, there have been studies comparing self similar geometries at different absolute length scales such as a system of micro and nano sized particles, which show different sintering behaviour depending on the absolute particle size. There have been studies on multi particle arrays both linear and closed. Early studies on linear arrays observed rearrangement of particles and relative rotation due to non spherical shape and bond angle of an array of three particles. Also there was a study that predicted rearrangement due to differential shrinkage in an assembly containing a combi-nation of large and small particles. Similar observations were also made on closed arrays of four or more particles both in 2D and 3D. Formation of high energy local configurations such as six grain boundaries (GBs) meeting at a line were found, followed by the topological transitions such as formation of new GBs or elimination of existing ones, leading to specific features in sintering behaviour. Geometrical evolution during final stage sintering is critical for forming dense final products. While most studies related the shrinkage behaviour to shape of the pore (convex or concave) and the number of grains surrounding a pore, later the absolute size of the pore was observed to be an important parameter. In 2D simulations and experiments large convex pores were found to shrink due to mass transport from surrounding GBs. In 3D simulations, pores with large coordination number as high as 32, pore shrinkage was observed followed by gradual reduction in coordination number and final elimination. Also studied are evolution of pore -GB configuration in case of small pores as separation of these from GB and entrapment into grains will freeze further shrinkage. In addition to the geometry related changes are also crystallography related microstructural changes. Crystallographic arrangement at the atomic scale leads to anisotropy of interfacial energies and diffusivities, that effect microstructural evolution. The presence of positive and negative ions in ionic solids can result in additional features such as charged and neutral planes Crystallography can affect the rotation of powder particles in initial stage sintering to subtle differences in microstructure evolution during grain growth in final stage sintering. Conversely crystallography has to be related to diffusion at interfaces. The rotation of spheres is governed by energetics. The final configuration corresponds to local energy minima in misorientations between the spheres and the single crystal plate. This technique is useful in finding a number of crystallography related aspects such as low energy GBs and equilibrium shapes of metal droplets. Rotation of unconstrained crystal related to neighbouring crystal has also been observed in thin films. Surface energy anisotropy has often been studied using topography of annealed sur-faces studied using atomic force microscopy (AFM). While low energy stable surfaces show perfectly flat surfaces, planes close to a stable plane form terrace and ledge structures whereas unstable planes form hill and valley structures. A method of “inverse Wulff shape” of pores trapped in single crystals has been used to find relative stability of sur-face planes using a combination of electron back scattered diffraction (EBSD) and AFM. Crystallography is very much related to the phenomenon of abnormal grain growth that occurs during later stages of sintering. Similarly, polycrystal assemblies have shown varying GB migration velocities for different crystallographic planes. Most recently, 3D EBSD has been used to study crystallography of GBs in sintered polycrystalline materials. In the present study, we address two specific issues. The first is related to the effect of microstructure of polycrystalline powder particles on initial stage sintering, where we compare sintering between particles with same particle size but different grain sizes. The second is related to the crystallographic aspects of interfaces in sintered materials with specific reference to yttria stabilized cubic zirconia. The present study is mostly confined to pressure less (free) sintering where the only driving force is the reduction in interfacial energy of the system. The effect of polycrystalline nature on initial stage sintering is investigated and com-pared with the behaviour of single crystal particles. We extend the model by Coble on single crystals to polycrystalline particles containing space filling tetrakaidecahedral grains with an identical grain size. The grain boundaries within particles are considered to be additional sources for mass to be plated at the neck and the flux equations are suitably modified. A model was developed to characterize the variation with time in the growth rate (x/R), where x and R are radii of the neck and particle respectively. The model indicated that the neck growth rate for polycrystalline spheres was faster compared to single crystals towards end of initial stage sintering (large value of x/R). There is large scope for extending the model further for complex geometries, diffusion distances and grain size distributions. Sintering experiments were conducted with annealed 2D random arrays of spheres of zirconia with two different grain sizes and a particle size of 40 m. Two different forms of zirconia (8YCZ and 3YTZ) were used as model systems for a few and a large number of grains in a particle respectively. The experimental results were limited, but broadly consistent with the new model. However necks were found to grow to a value f x=R = 0:12 and they did not grow further. In the second part of our study, grain boundaries in yttria stabilized cubic zirconia were studied in the context of macroscopic crystallographic parameters of misorientations of grains on either side of the grain boundary and crystallographic coordinates of grain boundary planes. Our aim was to study the evolution of misorientations and grain bound-ary planes during sintering process, starting from formation of necks during the initial stage to grain boundary migration during later stages. Orientation imaging microscopy based on an EBSD technique in an SEM was carried out on fully dense samples and also on porous samples obtained by interrupting sintering before attaining full density. The fraction of CSL misorientations on nearly dense cubic zirconia with grain sizes varying from submicrocrystalline 0.61 to 10 m was close to a random distribution. The number fraction of necks with CSLs formed in porous cubic zirconia with microcrysatlline particles was slightly higher than a random distribution. However, the present study covers only nearly dense-microcrystalline, nearly dense- submicrocrystalline, porous - microcrystalline regime , but misorientation information could not be obtained experimentally in a low density - submi-crocrystalline regime that is critical for sintering process. We also studied the distribution of grain boundary planes in fully dense 8YCZ with a grain size of 2.8 m by a stereological method using 2D OIM data. The overall distribution of grain boundary planes showed very weak anisotropy with slight maxima with 1.1 multiples of random distribution (MRD) at {100} planes, which is consistent with observations in literature on larger grain sizes. Interestingly, the planes that were abundant were not low energy surface planes (also mentioned in literature), in clear contrast with other ceramics studied in literature. The distribution of grain boundary planes was also plotted for specific misorientations, including those around low index axes of [100], [110], [111] and low misorientations. The grain boundary character distribution (GBCD) shows a high frequency of occurrence in position of pure twists about [100] and symmetric tilts at certain low misorientations . The highest frequency of occurrence was observed for coherent twin 3 on {111} plane and symmetric tilt (higher order twin) 11 on {113} plane, both corresponding to low energy GBs reported in literature in bicrystal experiments. With pure twists on {100} for rotations about [100] axis and pure tilts with {11w} or {1ww} planes for rotations about [110], both the criteria for specialness based on surface planes forming GB or symmetric tilts are found to be valid for specific cases. Notable is the frequency of occurrence of coherent twin 3 on {111} and 11 on {113}, that was 4.8 MRD for microcrystalline 8YCZ and as high as 7.8 MRD for submicrocrystalline 8YCZ samples, which is much higher than frequency of occurrence of any GB plane in any oxide studied in literature.

Polycrystalline silicon photovoltaic (PV) modules have the advantage of lower manufacturing cost as compared to their monocrystalline counterparts, but generally exhibit both lower initial module efficiencies and more significant early-stage efficiency degradation than do similar monocrystalline PV modules. For both technologies, noticeable deterioration in power conversion efficiency typically occurs over the first two years of usage. Estimating PV lifetime by examining the performance degradation behavior under given environmental conditions is, therefore, one of continual goals for experimental research and economic analysis. In the present work, accelerated lifecycle testing (ALT) on three polycrystalline PV technologies was performed in a full-scale, industrial-standard environmental chamber equipped with single-sun irradiance capability, providing an illumination uniformity of 98% over a 2 x 1.6m area. In order to investigate environmental aging effects, time-dependent PV performance (I-V characteristic) was evaluated over a recurring, compressed day-night cycle, which simulated local daily solar insolation for the southwestern United States, followed by dark (night) periods. During a total test time of just under 4 months that corresponded to a year equivalent exposure on a fielded module, the temperature and humidity varied in ranges from 3 degrees C to 40 degrees C and 5% to 85% based on annual weather profiles for Tucson, AZ. Removing the temperature de-rating effect that was clearly seen in the data enabled the computation of normalized efficiency degradation with time and environmental exposure. Results confirm the impact of environmental conditions on the module long-term performance. Overall, more than 2% efficiency degradation in the first year of usage was observed for all thee polycrystalline Si solar modules. The average 5-year degradation of each PV technology was estimated based on their determined degradation rates.

M.Sc. / Current solar cell research programmes in general aim to develop a high conversion efficiency photovoltaic (PV) module from high quality thin films. In this study, Cu (In,Ga)Se2 (CIGS) thin films were grown and characterized. These films were grown by selenization of Cu-In-Ga precursors. These precursors were prepared by co-sputtering In and (Cu, Ga). All the precursors were grown on Mo coated soda lime glass substrates. The selenization was conducted under different conditions in Ar/H2Se atmosphere, i.e. taking different values of flow rate of H2Se (5.00, 1.00, 0.25 mol%) in Ar, temperature (350, 450, 550 ºC) and time (10, 20, 30, 40, 50, 60 min). At each selenization condition, two samples were placed at different positions in the chamber. The structural properties of the produced films were analyzed by the techniques of X-ray Diffraction (XRD) for phases, Scanning Electron Microscopy (SEM) for morphology and Energy Dispersive Spectroscopy (EDS) for the bulk composition. The effect of temperature variation, the effect of flow rate variation and the effect of time variation were analyzed by comparing the structural properties as analyzed by the techniques mentioned. All in all this specific study delivers important information about the sensitivity of Cu(In,Ga)Se2 (CIGS) thin films to the temperature, gas flow rate and exposure time of the selenization step. / Doctor C.A. Engelbrecht Professor Vivian Alberts

A estrutura da florencita foi refinada pelo método de mínimos quadrados, utilizando dados experimentais obtidos através do método de Debye-Scherrer. A coleta dos dados das intensidades integradas foi feita através da leitura do difratograma de pó por um microdensitômetro óptico automático, e empregando métodos de análise numérica para fazer a integração da função (2&#952,Y), tabelada a pontos eqüidistantes. Foram observados 14 picos de difração e as reflexões que se superpunham contribuindo para as intensidades dos picos foram identificadas e suas contribuições levadas em conta através de seus fatores de multiplicidade. O refinamento foi feito com o programa POWLS (Powder Least Squares) e inicialmente foram fornecidos os parâmetros posicionais dos átomos da Goyazita, que é isomorfa com a florencita. As intensidades observadas foram corrigidas pelos fatores de Lorentz-polarização e adsorção. O índice de discordância R atingido para os 14 picos de difração observados foi de 0.097. A fórmula molecular da florencita é CeAl3(PO4)2(OH)6. O composto cristaliza no sistema hexagonal com parâmetros de rede ao=6.96Å co=16.33Å &#945= &#946 = 90° &#947=120° V=685.07&#1973. O grupo espacial é R3m com Z=3 e densidade calculada igual a 3.67g.cm-3. / The crystal structure of the florencita was refined by least squares using experimental data obtained with the Debye-Scherrer method. An automatic optical microdensitometer was used for the data collection from powder difractogram and numerical analysis methods for the integration of the function (2&#952,Y) which is tabulated at equidistant points. 14 diffraction peaks were observed, reflections which superpose contributing to the same peak were identified and their contributions were taken in account using multiplicity factors. The program POWLS (Powder Least Squares) was used for the refinement and initially the positional parameters of the atoms of the Goyazite, which is isomorfous with the florencita were used. Intensities were corrected for the Lorentz, polarization and absorption factors. The final R factor for the 14 peaks was of 0.097. The molecular formula of the florencitais CeAl3(PO4)2(OH)6. It crystallizes in the hexagonal system, space group R3m with cell ao=6.96Å co=16.33Å &#945= &#946 = 90° &#947=120° V=685.07&#1973.

Made available in DSpace on 2017-07-21T20:42:40Z (GMT). No. of bitstreams: 1 Luiza Kupchak.pdf: 1787533 bytes, checksum: ed9f8e72989aa4cdbca14d73ae5bce65 (MD5) Previous issue date: 2007-02-23 / The polycrystalline ceramics study the base of SnO2 presents great interest due to the inherent properties of emiconductor, which allows the application in some types of components, such as, protective gas sensors, varistors, protects films for different refractory uses and for glasses fusion of heavy metals. The addition of ZnO to the SnO2, makes with that these ceramics get high densification when added the low concentrations. In this work it of sintering in the initial period of training was studied kinetic, of the SnO2 with addition of 0,6% and 1,2% in mol of ZnO. The used methods of study had been the considered ones for Venkatu-Johnson, Woolfrey-Bannister, Wang-Raj and Young-Cutler. With the application of these models it was determined the energy of activation and the possible operating mechanisms in the system. The samples had been gotten by means of mixture of oxides, isostatic pressing and sintering in dilatometer, under taxes of constant heating of 2,5, 5 and 10 ºC.min- 1 until the temperature of 1350 ºC to air. The techniques used to carry through the characterization of the sintered samples had been to the ones of pycnometry of helium, diffraction of rays X and electronic microscopy of sweepings. It was verified that the addition of ZnO to the SnO2 made possible the ceramics attainment with superior relative densities 90%. The values of energy of activation gotten for the SnO2 with 0,6% in mol of ZnO, for the diverse studied models are of: 500 kJ.mol-1; 555 kJ.mol-1; 540 kJ.mol-1; 491 kJ.mol-1, respectively. For the SnO2 with 1,2% in mol of ZnO, the values of activation energy supplied by the models had been of: 444 kJ.mol-1; 555 kJ.mol-1; 424 kJ.mol-1; 491 kJ.mol-1, respectively. The results for the dominant mechanism of the sintering gotten from the used models did not allow to determine precisions from the operating diffusional mechanism in the studied compositions. / O estudo de cerâmicas policristalinas a base de SnO2 apresenta grande interesse devido às propriedades inerentes a este semicondutor, as quais permitem a aplicação em vários tipos de componentes, tais como, sensores de gás, aristores, filmes protetores para diferentes usos e refratários para fusão de vidros de metais pesados. A adição de ZnO ao SnO2, faz com que estas cerâmicas obtenham elevada densificação quando adicionadas a baixas concentrações. Neste trabalho estudou-se a cinética de sinterização no estágio inicial, do SnO2 com adição de 0,6% e 1,2% em mol de ZnO. Os étodos de estudo utilizados foram os propostos por Venkatu-Johnson, Woolfrey-Bannister, Wang-Raj e Young-Cutler. Com aplicação destes modelos foi determinada a energia de ativação e os possíveis mecanismos atuantes no sistema. As amostras foram obtidas por meio de mistura de óxidos, prensagem isostática e sinterização em dilatômetro, sob taxas de aquecimento constante de 2,5, 5 e 10 ºC.min-1 até a temperatura de 1350 oC ao ar. As técnicas utilizadas para realizar a caracterização das amostras sinterizadas foram às de picnometria de hélio, difração de raios X e microscopia eletrônica de varredura. Verificou-se que a adição de ZnO ao SnO2 possibilitou a obtenção de cerâmicas com densidades relativas superiores a 90%. Os valores de energia de ativação obtidos para o SnO2 com 0,6% em mol de ZnO, para os diversos modelos estudados são de: 500 kJ.mol-1; 555 kJ.mol-1; 540 kJ.mol-1; 491 kJ.mol-1,respectivamente. Para o SnO2 com 1,2% em mol de ZnO, os valores de energia de ativação fornecidos pelos modelos foram de: 444 kJ.mol-1; 555 kJ.mol-1; 424 kJ.mol-1; 491 kJ.mol-1, respectivamente. Os resultados para o mecanismo dominante da sinterização obtido a partir dos modelos utilizados não permitiu determinar de forma precisa o mecanismo difusional atuante nas composições estudadas.

Doctor of Philosophy (PhD) / This thesis aims to establish a sound scientific methodology for the effective and efficient polishing of thermally stable PCD composites (consisting of diamond and SiC) for cutting tools applications. The surface roughness of industrial PCD cutting tools, 0.06 μm Ra is currently achieved by mechanical polishing which is time consuming and costly because it takes about three hours to polish a 12.7 mm diameter PCD surface. An alternative technique, dynamic friction polishing (DFP) which utilizes the thermo-chemical reactions between the PCD surfaces and a catalytic metal disk rotating at high peripheral speed has been comprehensively investigated for highly efficient abrasive-free polishing of PCD composites. A special polishing machine was designed and manufactured in-house to carry out the DFP of PCD composites efficiently and in a controllable manner according to the requirements of DFP. The PCD polishing process and material removal mechanism were comprehensively investigated by using a combination of the various characterization techniques: optical microscopy, SEM and EDX, AFM, XRD, Raman spectroscopy, TEM, STEM and EELS, etc. A theoretical model was developed to predict temperature rise at the interface of the polishing disk and PCD asperities. On-line temperature measurements were carried out to determine subsurface temperatures for a range of polishing conditions. A method was also developed to extrapolate these measured temperatures to the PCD surface, which were compared with the theoretical results. The material removal mechanism was further explored by theoretical study of the interface reactions under these polishing conditions, with particular emphasis on temperature, contact with catalytic metals and polishing environment. Based on the experimental results and theoretical analyses, the material removal mechanism of dynamic friction polishing can be described as follows: conversion of diamond into non-diamond carbon takes place due to the frictional heating and the interaction of diamond with catalyst metal disk; then a part of the transformed material is detached from the PCD surface as it is weakly bonded; another part of the non-diamond carbon oxidizes and escapes as CO or CO2 gas and the rest diffuses into the metal disk. Meanwhile, another component of PCD, SiC also chemically reacted and transformed to amorphous silicon oxide/carbide, which is then mechanically or chemically removed. Finally an attempt was made to optimise the polishing process by investigating the effect of polishing parameters on material removal rate, surface characteristics and cracking /fracture of PCD to achieve the surface roughness requirement. It was found that combining dynamic friction polishing and mechanical abrasive polishing, a very high polishing rate and good quality surface could be obtained. The final surface roughness could be reduced to 50 nm Ra for two types of PCD specimens considered from pre-polishing value of 0.7 or 1.5 μm Ra. The polishing time required was 18 minutes, a ten fold reduction compared with the mechanical abrasive polishing currently used in industry.

Part I - We have designed a binary metallic alloy for nanoscale resonator applications. We used magnetron sputtering to deposit films with different stoichiometries of aluminum and molybdenum and then characterized the microstructure and physical properties of each film. A structure zone map is proposed to describe the dependence of surface and bulk structure on composition. We then fabricated proof of principle resonators from the Al-32 at%Mo composition, selected for its optimized physical properties. An optical interferometer was used to characterize the frequency response of our resonators. Part II - We investigate the growth of faceted polycrystalline thin films with modeling and simulations. A new analytic model is derived for the case of orientation dependent facet growth velocity and the dependence of growth on initial grain orientations is explicitly calculated. Level set simulations were used to both confirm this analytic model and extend it to include various angular flux distributions, corresponding to different deposition methods. From these simulations, the effects of self-shadowing on polycrystalline film growth are quantitatively evaluated. / Materials Engineering

This study examines the lapping machining process used during the manufacturing of polycrystalline diamond compacts (PDCs). More specifically, it is aimed at improving the productivity of the process by developing a better understanding of the parameters that affect the material removal rate (MRR) and MRR uniformity of lapped PDC samples. Experiments that focused on several controllable lapping parameters were performed to determine to what extent they affected the process. It was determined that the MRR can be modeled with the Preston equation under certain ranges of pressure and speed. It was also found that using a hard and rigid sample holder produces higher MRRs than soft and flexible sample holders. The results have also shown that MRRs in excess of 300 micrometers per hour can be achieved while using 10 grams of diamond abrasive per PDC per hour of lapping. The productivity of the lapping process can also be improved by placing the maximum allowed PDC samples in a concentric circle on the edge of the sample holder. The MRR uniformity between samples lapped on the same sample holder was found to be dependent on the sample holder material. This thesis is composed of six chapters. The first chapter introduces the need for PDC's as extreme cutting tools, the manufacturing process of PDC's, and the lapping process. The second chapter discusses the motivation behind this research and the primary objectives that were established. The third chapter details the materials and the experimental procedure, and the fourth chapter presents the results. The fifth chapter discusses the results, and the sixth chapter presents conclusions and information on possible future work.

Silicon interposers with TSVs (through-silicon-vias) have been developed in single-crystalline silicon wafer to achieve the high I/O (Input/Output) density. However, single-crystalline silicon interposers suffer a few problems such as cost, electrical performance and reliability. To overcome these shortcomings, an entirely different approach using polycrystalline silicon interposers with thick polymer liners are proposed by Georgia Tech Packaging Research Center, aiming to achieve lower cost silicon interposers with high performance and reliability. The objective of this research is to explore and demonstrate thin polycrystalline silicon as a suitable interposer material to achieve high performance and high reliability TPVs (through-package-vias) in polycrystalline silicon materials with lower cost. Three fundamental challenges were defined, including: 1) low resistivity of the polycrystalline silicon, resulting in high electrical loss; 2) reliability problems resulting from CTE (coefficient of thermal expansion) mismatch between silicon and Cu, and 3) handling and processing of thin silicon panels. A three-dimensional EM (electromagnetic) model was developed to simulate insertion loss and crosstalk of TPVs and compared with TSVs. It has been shown thick polymer liner is effective in addressing the fundamental challenge of low resistivity for the polycrystalline silicon material, leading to better electrical performance of TPVs than TSVs. Parametric studies indicate that thicker sidewall liners result in better electrical performance. A two-dimensional axisymmetric model was established to simulate the first principal stresses in silicon and shear stresses in TPV under thermal cycling. TPVs with thick polymer liners present both smaller principal stresses and shear stresses than TSVs due to the low modulus of polymer. Parametric studies suggest that sidewall liners act as stress buffers and thicker liners result in better mechanical performance. Design guidelines based on simulation results were used in TPV demonstration and test vehicle fabrication. Fracture strength of polycrystalline silicon panel has been fundamentally studied with four-point bending tool and Weibull plot. Surface polymer liners on both sides were introduced to improve the handling of thin silicon panels. Quantitative study showed higher characteristic fracture strength for the panel with surface liners than raw silicon panel. Low cost and double-side processes have been developed for TPV fabrication including UV (ultraviolet) lasers for TPV formation, double laser method for liner fabrication and electroless Cu for seed formation. Key steps and mechanisms for aforementioned processes were summarized and discussed. Polycrystalline silicon interposers with TPVs and up to four metal RDLs (re-distribution layers) were designed, fabricated and characterized. Measurement results showed low insertion loss for both TPVs and CPW (co-planar waveguide) transmission lines. Good model to hardware correlation was also observed. Reliability test vehicles of polycrystalline silicon interposers were also designed and fabricated for thermal cycling test. TPVs survived 4000 cycles without significant resistance changes. SEM imaging on the cross-section of the samples confirmed no Cu or silicon cracking. Magnified images around corner also suggested good adhesion at Cu/liner and silicon/liner interfaces.