Graduation date: 1997

Particle reinforced composites are widely used in tires, heat exchangers, thermal barrier coatings and many other applications, as they have good strength to weight ratio, excellent thermal insulation, ease of manufacturing and flexibility in design. During their service life, these composites are often subjected to harsh environments, which can degrade the thermo-mechanical properties of the constituents in the composites, affecting performance and lifetime of the composites. This study investigates performance of particle reinforced composites subjected to coupled heat conduction and thermo-elastic deformation at the macro and micro levels. A micromechanical model is used to determine the effective thermal and mechanical properties of the homogenized composite by incorporating microscopic characteristics of the composites. The constituent?s thermal conductivities of the composite are assumed to be functions of temperature and the elastic moduli to be functions of temperature and stress fields. The effective properties obtained from the micromechanical model represent average (macroscopic) properties. The effective heat conduction and thermo-elastic responses in the homogenized composites are compared with the responses of the composite with particles randomly distributed in the matrix (heterogeneous materials) which represent microscopic responses. For this purpose, two sets of finite element (FE) models are generated for composites with particle volume contents 12.5, 25, and 50%. The first FE model represents a homogenized composite panel and the effective responses from the micromechanical model are used as input for the material properties. The second FE model mimics composite microstructure with discontinuous particles randomly dispersed in a homogeneous matrix. Parametric studies on effects of conductivity ratio between particle and matrix, degree of nonlinearity, and volume fraction on the temperature distribution and steady state times have been studied. For lower volume fractions the temperature profiles of homogenized and heterogeneous composite models are in good agreement with each other. But for higher volume fractions, the detailed model showed a wavy profile whereas the effective model showed no signs of it. When the nonlinearity in thermal conductivity of the particle and matrix constituents is increased, the steady state time significantly deviates from the ones with constant constituent properties. When the volume fraction of particles in the composite increases, the steady state is reached in less time, since the thermal conductivity of particles are taken larger than that of the matrix. Effects of coefficient of thermal expansion (CTE) ratio of particle and matrix, temperature change, and volume fraction on the discontinuity of stress and strain fields at the interphase of matrix and particle have been studied. The stresses developed were more for higher CTE ratios and the magnitude of discontinuity also follows the same trend. As the volume fraction increases, the stresses developed and the magnitude of discontinuity also increase. Finally, sequentially coupled heat conduction and deformation analyses are performed on thermal barrier coating (TBC) systems to demonstrate the applicability of the micromechanical model in predicting overall thermo-elastic responses of the TBC.

This thesis outlines the development and decay of palsas in northern Sweden in relation to climatic and environmental factors. Palsas were investigated on morphology, vegetation, peat porosity, ground temperature and local climate in four bogs situated on a north-south gradient in northern Sweden. A new classification for palsa stages (embryo, young, mature, degrading and remnant stage), based on morphological and vegetational characteristics, is proposed and described in the thesis. The start of palsa growth seems to be favoured by a low vegetation height and a high cover of Sphagnum mosses. Very high porosities were measured in this peat type, resulting in a high insulation capacity preserving the frozen ground in summer. A decrease of porosities in the surface peat layer from the embryo palsa stage to the mature and degrading palsa stage was found. This contributes to the degrading of palsas, due to higher thermal diffusivities in palsas with lower porosities. The aggradation of the ice core can also cease due to the observed change in vegetation from low vegetation on embryo and young palsas to vegetation with tall Betula nana shrubs on degrading palsas. This was found to cause thicker snow cover that prevents frost penetration into the palsa core. The present climate conditions (with mean annual temperature of –1.5°C and annual precipitation of 737 mm) at the boundary of palsa distribution in Sweden were found to be unsuitable for palsa development. Palsa growth started here during a cold period in the last part of the Little Ice Age, with estimated mean annual and mean winter temperatures of –2.3°C and –10.0°C, respectively. The major conclusions are that besides air temperature, depth of snow cover, summer precipitation and hydrology, also vegetation cover and peat characteristics are important for palsa growth and decay.

The sorption/desorption behaviour of n-hexane in high molarmass linear polyethylene (PE) and branched PEs with 0.39 and5.09 hexyl branches per 100 main chain C atoms andcrystallinities between 4 and 82% at 298 K has been studied.Crystal core contents determined by Raman spectroscopy werealways lower than those determined by density measurements. Then-hexane solubilities in the copolymers depended in anon-linear manner on the content of penetrable polymercomponent and were lower for homogeneous copolymers than forheterogeneous copolymers at the same contents of penetrablecomponent. The solubility of hexane in the linear PE sampleswas proportional to the volume fraction of the penetrablepolymer and the solubility was low in comparison with that ofthe branched PE of the same crystallinity. TheCohen-Turnbull-Fujita (CTF) free volume theory was capable ofdescribing the desorption process in the PEs studied. Theconcentration dependence of the thermodynamic diffusivitypredicted by the CTF free volume theory was confirmed by thedata obtained by the differential method, and the differencesbetween the results obtained by the integral and differentialmethods were within the margins of experimental error. Thedependence of the fractional free volume of the penetrablephases on the phase composition suggests that mass transporttakes place from the liquid-like component to the interfacialcomponent and that the penetrant molecules are trapped at theinterfacial sites. The linear PE samples showed a physicallyrealistic trend with a decrease in the geometrical impedancefactor (t) with decreasing degree of crystallinity, whereas theopposite trend was obtained for the copolymers. The decrease int with increasing crystallinity in the copolymers may beexplained by the presence of wide crystal lamellae in the lowcrystallinity samples. A novel melt-extrusion method was used to createcircumferential chain orientation in pipes of crosslinked PE.The microstructure of the pipes was characterized usingdifferential scanning calorimetry (DSC), density measurements,X-ray diffraction, infrared dichroism and contractionmeasurements. The mechanical properties were assessed byuniaxial tensile tests. The maximum degree of circumferentialorientation was obtained at the inner wall of the orientedpipe. The oriented pipe material exhibited a 5-15% higherdegree of crystallinity and a greater crystal thickness thanconventionally crosslinked pipe. The circumferential and axialmoduli of the oriented, crosslinked pipe were greater than thecorresponding moduli of the non-oriented crosslinked pipe. Blends of single-site materials of linear PE andethyl-branched PE were prepared using solution- and melt-mixingmethods. The thermal properties of the blends were studied byDSC and results obtained by the two mixing methods werecompared. Data obtained for heats of melting andcrystallization, melting and crystallization peak temperaturesand melting and crystallization temperature profiles wereessentially the same for the samples obtained by the two mixingmethods. The heat associated with the high temperature meltingpeak of the blend samples obtained by both preparation methodsafter crystallization at 398 K was higher than that of thelinear polyethylene included in the blends, suggesting that apart of the branched polyethylene crystallized at 398 K. <b>Key words:</b>n-Hexane diffusion, polyethylene, free volume,solubility, sorption, desorption, mechanical properties,orientation, thermal properties, blend.

Ti-6Al-2Sn-4Zr-2Mo-0.1Si specimens were shaped by repeated cycles of heating (to 954 °C) and hammer or press forging followed by a solution anneal that varied from 968 to 998 °C. The coupons were originally extracted from billets forged below the beta trans us ( 1009 °C) and slow cooled to ambient temperature. Macroscopic and microstructural banding is observed in some forged and solution annealed coupons, that consists of regions of elongated primary alpha. More significant banding is observed subsequent to annealing at lower temperatures (968 °C), whereas much less microstructural banding is present after annealing at higher temperatures (998 °C). About the same level of banding is observed in hammer forged and press forged coupons. The observation of these bands is significant since they may lead to inhomogeneous mechanical properties. Specifically, at least some types of banding are reported to affect the high temperature creep properties of this alloy. The origin of these bands was therefore researched. Classically, banding in Ti-6242-0.1Si has been regarded as a result of adiabatic shear, chill zone formation or compositional inhomogeneity. High and low magnification metallography, electron microprobe analysis and microhardness tests were performed on forged and annealed specimens in this investigation. The composition inside the bands appears identical to that outside of the bands. The fraction of primary alpha is also found to be identical. The bands have higher microhardness. These results suggest that the bands are not related to composition gradients. The bands also do not appear to be a result of adiabatic shear or other localized deformation. The bands of this study appear to originate from the elongated primary alpha microstructure of the forged billet (from which test coupons were extracted). The deformation of the extracted coupon may be neither fully homogeneous nor sufficiently substantial and the coupon is only partly statically restored after a solution anneal. Areas not fully restored appear as "bands" with elongated primary alpha, that are remnant of the starting billet microstructure. Therefore, a source of banding in Ti-6242-0.1Si alloy, additional to the classic sources, is evident. This type of banding is likely removed by relatively high solution treatment temperatures and perhaps greater plastic deformation during forging. / Graduation date: 1992

The low thermal conductivity of gallium arsenide compared to silicon results in self-heating effects in GaAs MESFETs that limit the electrical performance of such devices for high power applications. To date, analytical thermal models of self heating in GaAs MESFETs are based on the assumption of a uniformly heated channel. This thesis presents a two dimensional analysis of the electrothermal effect of this device based on the two dimensional power density distribution in the channel under various bias conditions. The numerical simulation is performed using the finite difference technique. The results of the simulation of an isothermal MESFET without heat effects is compared with various one dimensional analytical models in the literature. Electro thermal effects into the two-dimensional isothermal MESFET model allowed close examination of the temperature profile within the MESFET. The large gradient in power distribution results in a localized heat source within the channel which increases the overall channel temperature, which shows that the assumption of a uniformly heated channel is erroneous, and may lead to an underestimation of the maximum channel temperature. / Graduation date: 1992

The ion implantation process is a very precise, controllable, and reproducible method used to enhance material properties of finished components such as ball bearings. Essentially, the target material is bombarded by accelerated ions to form a thin alloyed layer in the substrate. As the ions deposit their kinetic energy in the target it begins to heat up. To prevent thermal distortion in the finished pieces the ion implantation is performed at dose levels (dependent on the ion fluence and time duration of implantation) to insure that the target pieces stay at relatively low temperatures. Consequently, the low temperature requirement for many applications limits the economic, and probably, the physical success of ion implantation. The purpose of this study was to show the applicability of using a two-dimensional computer code developed to model plasma disruptions and subsequent energy deposition on a fusion reactor first wall to calculate surface and bulk temperature information during ion implantation. In turn the code may assist researchers pursuing development of adequate cooling for target materials in an attempt to overcome the low temperature constraint. All data supported the hypotheses that the two-dimensional code previously developed for fusion reactor applications was adequate to model the ion implantation process. / Graduation date: 1992

<p>Fluorochlorozirconate (FCZ) glasses are a member of heavy metal fluoride glasses, and are derived from a well known ZBLAN glass. In this work, halogen salts of europium (Eu²⁺) and samarium (Sm³⁺) are used as dopants in FCZ glasses. FCZ glasses doped with Eu²⁺ and Sm³⁺ can be used in high resolution x-ray imaging for tissue scanning, and have been shown to behave as storage phosphors and/or x-ray scintillators.</p> <p>Glass transition (T_g), heat capacity (C_p) and glass crystallization (T_c) properties of Eu²⁺ and Sm³⁺ doped and undoped FCZ glasses with different amounts of relative Cl concentration, that is, with respect to the total Cl and F concentration have been investigated by conventional differential scanning calorimetry (DSC) and modulated differential scanning calorimetry (MDSC) techniques. MDSC experiments were performed at different heating rates to analyze the complex transitions and to get a better resolution of any overlapping transitions. The crystallization kinetics have also been studied by applying the Kissinger technique to multiple DSC scans in order to determine the thermal stability of FCZ glass samples used in this work. The apparent activation energy for the crystallization process was obtained by the crystallization peak temperature shift method in the conventional DSC mode. The specific heat capacity (C_p) has been measured as a function of composition, and the glass transition temperature (T_g) is evaluated from the smooth change in the heat capacity curve during the glass transformation. The observation of two possible glass transitions points to the presence of two phases in these FCZ glasses with higher relative Cl content.</p> <p>Optical transmission spectra of both doped and undoped FCZ glass samples have been measured by infrared spectroscopy and optical band gaps corresponding to an absorption coefficient of 10³ cm^-1 have been determined. A good correlation between X-ray luminescence and the glass structure is observed. While the integrated photoluminescence intensity increases linearly with the Sm³⁺ concentration, the integrated X-ray luminescence increases sublinearly. The importance of appropriate annealing conditions, such as temperature, time and ambient atmosphere, and their effect on the X-ray luminescence of rare earth (RE) doped FCZ glass samples have been investigated. Annealing conditions influence the formation of BaCl₂ nanocrystals in the glass and the properties of the resulting FCZ glass ceramics.</p>

Podophyllotoxin is a natural medicine possessing an outstanding anti-tumour activity. It can be extracted from the rhizome of Podophillum peltatum (American Podophyllum). Volumetric heating of a packed bed of particles including solvent during the extraction can eliminate the solvent pre-heating time and provide uniform and quick heating of the bed. RF-assisted extraction has a potential to be a promising extraction alternative over conventional methods. The characterization and assessment of RF-assisted extraction of podophyllotoxin is crucial. Thermal properties including specific heat capacity, thermal conductivity, and thermal diffusivity of a packed bed of P. peltatum with and without ethanol solutions were determined and the associated multiples regression equations were obtained for the purpose of thermal analysis of RF-assisted packed bed extraction process and related modeling investigations.<p> The dielectric properties of the packed bed of rhizome particles were measured from 10 to 30 MHz using a precision LCR meter and a liquid test fixture. The effects of temperature, particle moisture content, volumetric concentration of ethanol and bed porosity on the dielectric constant, dielectric loss factor and power penetration depth were investigated. The dielectric loss factor significantly increased with the particle moisture content for the beds with 100% and 70% ethanol but not with 30% ethanol. The dielectric loss factor was proportional to temperature directly and to frequency inversely. With 30% ethanol (and therefore 70% water), the dielectric loss factor of the bed dramatically increased compared to 70% and 100% ethanol. Porosity had a significant effect on the dielectric constant but not on the dielectric loss factor. The power penetration depth of a packed bed with 100% ethanol was significantly larger than those of the packed bed with 30% and 70% ethanol. Empirical regression equations were developed for simulation and design of an RF-assisted packed bed extraction of podophyllotoxin.<p> A RF-transparent batch reactor was made of glass filled Teflon and the extraction kinetics of podophyllotoxin was characterized. The effects of temperature, ethanol volumetric concentration, solid/liquid ratio, RF heating and particle moisture content on the extraction rate and yield of podophyllotoxin were investigated at different extraction conditions. A generalized diffusion mathematical model taking into account three major particle geometries was developed and coupled with genetic algorithm for determination of effective diffusivity and partition coefficient through an inverse simulation approach. The approach was first verified by reported experimental data of andrographolide extraction followed by determining the effective diffusivity and partition coefficient of podophyllotoxin for different conditions. The optimum batch extraction condition was achieved with 30% ethanol-water solution at 53¢XC. A prototype was developed for RF-assisted extraction of podophyllotoxin using two optical and RF-transparent reactors with horizontal and vertical orientations. Applying the optimum conditions obtained from batch experiments, the potential of RF heating for providing a uniform temperature in the packed bed was evaluated. The effect of solvent dielectric loss factor on uniform RF heating was investigated and the chemical effect of NaCl used for increasing dielectric loss factor of the solvent on podophyllotoxin was assessed. The horizontal packed bed demonstrated a large temperature gradient across the thickness of the bed during RF heating; however, a uniform RF heating was achieved when the vertical packed bed reactor was used for RF-assisted extraction of podophyllotoxin. The concentration of 2.5 g NaCl/L of the solvent at the temperature controller set point of 40aC provided a relatively good uniform temperature of 50aC within the bed. Evaluating three flow rates of 130, 160 and 200 ml/min for the solvent of 30% ethanol with 2.5 g NaCl/L indicated that the flow rate of 160 ml/min could provide better temperature overlap of four positions of the bed height.

<p>Fluorochlorozirconate (FCZ) glasses are a member of heavy metal fluoride glasses, and are derived from a well known ZBLAN glass. In this work, halogen salts of europium (Eu²⁺) and samarium (Sm³⁺) are used as dopants in FCZ glasses. FCZ glasses doped with Eu²⁺ and Sm³⁺ can be used in high resolution x-ray imaging for tissue scanning, and have been shown to behave as storage phosphors and/or x-ray scintillators.</p> <p>Glass transition (T_g), heat capacity (C_p) and glass crystallization (T_c) properties of Eu²⁺ and Sm³⁺ doped and undoped FCZ glasses with different amounts of relative Cl concentration, that is, with respect to the total Cl and F concentration have been investigated by conventional differential scanning calorimetry (DSC) and modulated differential scanning calorimetry (MDSC) techniques. MDSC experiments were performed at different heating rates to analyze the complex transitions and to get a better resolution of any overlapping transitions. The crystallization kinetics have also been studied by applying the Kissinger technique to multiple DSC scans in order to determine the thermal stability of FCZ glass samples used in this work. The apparent activation energy for the crystallization process was obtained by the crystallization peak temperature shift method in the conventional DSC mode. The specific heat capacity (C_p) has been measured as a function of composition, and the glass transition temperature (T_g) is evaluated from the smooth change in the heat capacity curve during the glass transformation. The observation of two possible glass transitions points to the presence of two phases in these FCZ glasses with higher relative Cl content.</p> <p>Optical transmission spectra of both doped and undoped FCZ glass samples have been measured by infrared spectroscopy and optical band gaps corresponding to an absorption coefficient of 10³ cm^-1 have been determined. A good correlation between X-ray luminescence and the glass structure is observed. While the integrated photoluminescence intensity increases linearly with the Sm³⁺ concentration, the integrated X-ray luminescence increases sublinearly. The importance of appropriate annealing conditions, such as temperature, time and ambient atmosphere, and their effect on the X-ray luminescence of rare earth (RE) doped FCZ glass samples have been investigated. Annealing conditions influence the formation of BaCl₂ nanocrystals in the glass and the properties of the resulting FCZ glass ceramics.</p>