<p>The oxidation properties of Fe-Al alloys containing 1.5,3,4,5,7,8,12,16,20 atomic percent aluminum were investigated in dry oxygen, wet oxygen, laboratory air and alloy/(Fe+FeAl₂O₄+Al₂O₃) diffusion couples at 1173K. The reaction kinematics were determined thermogravimetrically and by layer thickness measurements. The reacted specimens were analyzed using light microscopy, electron metallography, x-ray diffraction, Auger electron spectroscopy and Mössbauer spectroscopy. Particular interest was given to the growth of Al₂O₃on these alloys. A duplex external scale consisting of Fe₂O₃at gas/scale interface, (Fe,Al)₃O₄ at scale/subscale interface and a subscale of FeAl₂O₄ platelets with Al₂O₃tips grew on alloys containing up to 5 a/o Al oxidized in dry oxygen. The growth of the FeAl₂O₄-Al₂O₃ platelet precipitates, which is interpreted by a (Fe,Al)₃O₄/Al (alloy) solid state displacement reaction, is controlled by oxygen diffusion torugh the subscale in the very early stage of oxidation. A oxidation continues, the oxygen pressre at the scale/subscale interface decreases due to increase of aluminum concentration of (Fe,Al)₃O₄ by dissolution of FeAl₂O₄ precipitates. The diffusion outward from the bulk alloy plays and important role as the velocity of precipitation front decreased by decreasing oxygen activity which leads to the development of a continuous Al₂O₃film at the subscale front in Fe-5 a/o Al alloy and isolation of the subscale from the alloy substrate. Multilayered scale consisting of Fe₂O, Fe₃O, FeO+Fe₃O, FeO+FeAl₂O₄ and subscale grew on alloys containing up to 5 a/o Al alloy oxidized in air or wet oxygen. The FeO+FeAl₂O₄ layer and subscale grew by inward oxygen migration generated by dissociation of wustite. The high rate of oxygen transfer inward with hydrogen generated by dissociation of water vapour, accelerated the oxidation processes and did not allow aluminum diffusion outward from the bulk of the alloy to affect the oxidation behavior. For Fe-7 and 8 a/o A alloys, a thin duplex scale was formed consisting of Fe₂O₃ layer at oxygen/scale interface and Al₂O₃ layer at scale/alloy interface. An oxidation model was advanced to describe the scale, development involving iron diffusion through leakage paths in the Al₂O₃ formed initially to form Fe₂O₃ layer at the scale/gas interface. Only the Al₂O₃ is formed on Fe-12, 16 and 20 a/o Al alloys. An oxidation mechanism is proposed based on inward oxygen diffusion down the grain boundaries and outward aluminum diffusion to account for the formation of hollow nodules in the scale. Alloys containing 3, 5, 7 and 8 a/o Al oxidized internally when coupled with (Fe+FeAl₂O₄+Al₂O₃) mixture. The Fe-3 a/o Al alloy oxidized linearly but the Fe-5, 7, and 8 a/o Al alloys at internal oxidation zone/alloy interface. However, an Al₂O₃ film was formed at (Fe+FeAl₂O₄+Al₂O₃)/alloy interface with no internal oxidation in Fe-12 a/o Al alloy diffusion couple. Alloy composition at which transition from internal to external scale formation of Al₂O₃ occurs, is calculated using available models and compared with the experimental results.</p> / Doctor of Philosophy (PhD)

<p>This thesis treats the discontinuous mode of precipitation in Al-22 and 28 at .8 Zn. High resolution STEM x-ray microanalysis has been used to measure the composition profiles in individual depleted lamella formed under isothermal conditions. These data, combined with local values of cell velocity and interlamellar spacing, have been used to evaluatethe kinetic and thermodynamic quantities of main concern. It is shown that the reaction is grain boundary-diffusion controlled and depending of individual grain boundary characteristics. The driving force for grain boundary migration and possible free-energy sinks are discussed. The results are compared with predictions of existing theories.</p> <p>The problem of the optimal interlamellar spacing is addressed, mainly from the experimental point of view. Experimental data for Al-Zn and Mg-Al alloys are analyzed and related to the discussion.</p> <p>The problem of the optimal interlamellar spacing is addressed, mainly from the experimental point of view. Experimental data for Al-Zn and Mg-Al alloys are analyzed and related to the discussion.</p> <p>It is demonstrated that initiation involves diffusion induced grain boundary displacements, and that during the growth process a significant amount of elastic energy is stored in the product phases. The dissolution of the reheated lamellar structure is volume diffusion-controlled.</p> / Doctor of Philosophy (PhD)

<p>A theoretical and experimental study has been made which indicates the importance of the spatial distribution of damage and a complete physical description of the operative damage processes in providing a fracture condition. A continuum damage model is presented which investigates the influence of void coalescence in accelerating damage levels, for a random array of voids growing via a ductile hole growth mechanism. A simple geometric coalescence condition was assumed. The dependence of fracture strain upon initial void volume fraction and stress state history is predicted. This approach is extended in a simulation to predict the void size distribution which develops. The Dirichlet tessellation was evaluated as a method for characterizing dispersions of points or particles. Point dispersions were generated, ranging from strongly periodic to strongly clustered, and properties of their associated tessellations were evaluated. This suggested parameters which were sensitive indicators of periodicity or clustering in a dispersion. this approach was extended to the characterization of inclusion and particle distributions in steels and several other alloys. The influence of temperature and stress on damage and fracture was studied in several continuously-cast HSLA steels. Fracture was controlled by the inclusions, the centre-line transformation products, or the interaction between these two sources of damage. Observations also suggested that the distribution of the banded transformation products could influence damage levels in the ferritic regions. This study concludes with an investigation of the creep fracture in a Ni-Sn alloy, which contained an inhomogeneous tin distribution. The regions with high tin content accumulated damage rapidly, however, ductility was promoted by the material low in tin which did not cavitate as readily. The influence of stress level on the strain dependence of damage accumulation rate, and the shift to surface crack-controlled fracture was interpreted as an effect of the enhanced relative contribution of grain boundary sliding to total strain at lower stresses.</p> / Doctor of Philosophy (PhD)

<p>For environmental and economical considerations, it is desirable to use iron ore concentrates directly without agglomeration and coal directly without coking for ironmaking. The present work is a study of the kinetics and mechanisms of iron ore reduction in ore/coal composites. An experimental system has been designed and experiments have been conducted with in-situ measurements of temperature and pressure to establish profiles within the ore/coal packing with the furnace temperature maintained at 1200 or 1300ºC. Samples were taken at various locations for chemical analysis in the partially reacted specimens. These experiments lead to a better understanding of this non-isothermal and non-isobaric system. A non-isothermal and non-isobaric mathematical model has been developed and validated by the experimental data. Mechanisms of heat transfer, mass transfer and interfacial reactions were studied using the mathematical model. The contribution of the individual kinetic steps to the rate of overall reaction were compared to evaluate the rate controlling step. The contribution of the author is the design of the experimental system which has demonstrated the non-isothermal and non-isobaric nature of the reaction system, and the development of the mathematical model. The present study may be used in new ironmaking processes and carbothermic process, such as FASTMET and cokemaking processes.</p> / Doctor of Philosophy (PhD)

<p>Growth of superlattices on V-grooved substrates has become a popular method of producing quantum wires. The success of this technique to date has been limited to the lattice-matched AlGaAs/GaAs system. The unpredictability of the morphology and defects found in lattice-mismatched systems has been a major hindrance in their development. Additionally, the misfit stresses in lattice mismatched systems may play an important role in determining the properties of the laser. In V-grooves, the singular points at the corners of the groove produce a stress distribution completely different from the case of layers grown on a planar substrate. Accurate knowledge of the stress distribution is hence necessary because a stress can affect the optoelectronic properties and lead to defects, limiting the life of the laser. The aim of the project was two fold: first to characterize the InGaAs/InP growth on (211)A and (111)B grooves using molecular beam epitaxy in terms of faceting, thickness variation, and composition variation; second to obtain an analytical and numerical stress distribution for the case of a layer with uniform composition and thickness grown on a sharp V-groove. The characterization was performed using transmission electron microscope and degree of polarization methods. Composition analysis was done with a high resolution scanning transmission electron microscope. Numerical simulation was done using a commercial finite element program, 'ABAQUS'. The results showed a clearly distinct morphology for layers grown on faceted (211)A and (111)B substrates. The differences observed in faceting, thickness variations, and composition variations in the (211)A and (111)B grooves suggest that the lower incorporation rate of group III atoms on (111)B surfaces leads to increased interfacet diffusion compared to (211)A grooves, where a higher incorporation rate produces layers with uniform composition and thickness. Furthermore, the 46.5% increase in In content found at the bottom of (111)B grooves indicates that the mean diffusion length of In is much higher than Ga for the growth conditions used in this study. The large variation in composition in (111)B grooves produced extensive defects at the bottom of the groove and at certain locations of the sidewall where the misfit exceeds the critical limit. The analytical and numerical solution of the stress distribution were in good agreement with experimental results. The stress fields obtained by these methods would be useful in helping to predict the optoelectronic properties of the quantum wire. Furthermore these models are useful in determining the composition and thickness of the layers which need to be grown to obtain specific optical properties.</p> / Doctor of Philosophy (PhD)

<p>The mechanical behaviour in uniaxial tension and compression of extruded and aged Mg-8.5wt%Al was studied. In-situ neutron diffraction was used to follow the elastic lattice strains under load in the matrix and the precipitates. The internal stresses determined from these measurements are highest in grains unfavourably oriented for both basal slip and {10Τ2} twinning, lowest in grains oriented favourably for both, and in between for grains oriented favourably for {10Τ2} twinning only. Most variations in scattered peak intensity are due to the lattice reorientation produced by {10Τ2} twinning. A critical resolved shear stress criterion is shown to apply for twinning. Intensity variations which cannot be explained by {10Τ2} twinning occurred in some grains during tensile loading. They are likely due to {10Τ1} twinning which produces c-axis compression, unlike {10Τ2} twinning. Transmission electron microscopy revealed the presence of c- and non-basal a-dislocations in the undeformed alloy. Basal slip is the most common slip system, though non-basal a-slip also occurs. Only {10Τ2} twinning was observed by TEM. Twins often traversed grains completely, despite the presence of the precipitates. Schmid factor considerations show that pure magnesium yields first by basal slip. The early portion of the stress-strain curve should thus be considered a region of rapid strain hardening due to basal dislocation pile-ups at grain boundaries. In compression {10Τ2} twinning can also occur at very low applied stress. Strengthening in the alloy before yield is explained using a Brown and Clarke mean-stress hardening model. Beyond yield, relaxation mechanisms reduce the mean stress contribution essentially to zero in tension. The mechanical and physical properties of the intermetallic were obtained from experiments on a single crystal. Property correlations have been used to estimate the fracture toughness and yield stress, assuming its behaviour is similar to that of a ceramic.</p> / Doctor of Philosophy (PhD)

<p>The most useful properties of ceramics (high temperature strength, chemical inertness and hardness at low density) are accompanied by brittleness. This is still the main factor limiting widespread application of ceramic materials. In the present thesis an energy approach to fracture of ceramics was undertaken and refined to account for a nonelastic behaviour of these materials. A chevron-notched (CN) four-point bend specimen was recognized as effective experimental arrangement for room and elevated temperature tests. Consequently, a number of theoretical studies of the specimen's performance were undertaken. Modelling of the variation of the strain energy release rate with the crack extension revealed that the subcritical crack growth in a CN specimen causes dependence of the measured fracture parameters on the experimental procedure (stressing rate, stiffness of the testing system, crack length). It appears that, as complete fracture is approached (i.e. 100% of the specimen's cross-section), the measured work-of-fracture approaches that required for crack initiation. An electrical potential drop technique for crack length measurement in the CN specimen was developed for elevated temperatures fracture studies in the ionically conducting zirconium oxide ceramics. The resistance-to-fracture versus crack extension was determined for a range of temperatures (25 to 1300℃) for stabilized zirconias and their HfO₂ solid solutions and with second phase p-Al₂O₃ particles dispersed in them. The room temperature results agreed with the literature data and model predictions. Above 1000℃ an energy input of ⁻1 J/m² is required to drive the crack through zirconium oxide ceramics. Viscoelastic effects and crack interaction with p-Al₂O₃ particles result in a total fracture energy dissipation two orders of magnitude higher.</p> / Doctor of Philosophy (PhD)

<p>Currently available theoretical formulation for the kinetics of slag-metal systems have been extended to clarify certain aspects of the system. Coupling factors for electrochemical reactions and for ionic diffusion have been defined to clarify the significance of coupling phenomena in the kinetics of multi-component systems. Theoretical equations for the interfacial reactions and the diffusion processes in both phases, in seven hypothetical pseudo-ternary slag-metal systems have been numerically solved by a finite difference method. Typical features of coupling phenomena, i.e.m acceleration and deceleration of reaction or diffusion, and up hill reaction or diffusion have been clearly demonstrated. Through numerical analysis, a modified form of "Sherwood Number", Shᵢ*≡kᵢL/ρDᵢ for a particular reaction i, is defined and found to be proper to explain the rate controlling steps for the over-all reaction. In the present analysis, the relationships between the modified Sherwood Number and rate controlling steps were found to be as follows: Sh*ᵢ > 360: diffusion control, 0.05 ≲ Sh*ᵢ ≲ 360: mixed control, Sh*ᵢ < 0.05: interfacial reaction control Thus non-trivial numerical solutions for slag-metal systems have been developed for the first time.</p> / Master of Engineering (ME)

<p>The dielectric and semiconducting properties of two types of ceramics (n-type BaTiO₃ and dielectric Ba(B'⅓ B''⅔ )O₃ where B' = Mg, Zn, Ni, and B'' = Nb₁ Ta) were characterized. Complex impedance analysis and dc conductivity measurements of samples prepared at various [Special Characters Removed] have ruled out oxygen chemisorption in favor of interfacial segregation of cation vacancies as the cause of the positive temperature coefficient of resistivity (PTCR) effect in n-type BaTiO₃ . The effect of preparation conditions, sintering atmosphere, stoichiometry, and post-sinter anneal on the defect chemistry of BaTiO₃ was studied using the electron paramagnetic resonance (EPR) technique. Several paramagnetic defects such as, Ti³⁺ , [Special Characters Removed] were detected and identified by EPR. Current-voltage characteristics (I-V ) of PTCR BaTiO₃ were analyzed in light of space-charge-limited-current, trap-filled-limited-current, Frenkel-Poole, small polaron, and double-Schottky barrier models. It was shown that for the double-Schottky barrier model, a partial stabilisation of the potential barrier is expected when the Fermi level is pinned at grain boundaries by a high density of the interface states. The deviation of I-V characteristics of BaTiO₃ in the region of the PTCR effect can be explained by dependence of the population of the interface electron states on applied voltage. Based on the Seebeck and Hall effect measurements, it was found that in the range of 100-300 K, the drift mobility of electrons in BaTiO₃ is not thermally activated, which supports the concept of conduction band electron transport rather than small radii polaron hopping. However, further study over a wider temperature range and on better quality crystals is required to unequivocally clarify the electron transport mechanism in BaTiO₃ . Phase composition, degree of cation ordering, and dielectric properties of complex perovskites with general formula Ba(B' ⅓ B''⅔ )O₃ where B' = Mg, Zn, Ni, and B'' = Nb₁ Ta were analyzed. It was shown that in Ba(Mg⅓ Ta⅔ )O₃ both intrinsic and extrinsic dielectric loss affect the Q-factor, whereas in Ba(Mg⅓ Nb⅔ )O₃ and Ba(Ni⅓ Nb⅔ )O₃ extrinsic factors such as the second phase and point defects dominate the dielectric loss at microwave frequencies.</p> / Doctor of Philosophy (PhD)

<p>Silicon nitride is sintered with the use of additives, such as Y₂O₃ and Al₂O₃, in order to enhance densification. After sintering, these additives along with SiO₂, present on the starting Si₃N₄ powder, form an intergranular amorphous phase. The presence of a glassy phase is generally thought to be detrimental to the high temperature creep properties. In the material analyzed, Kyocera SN220, this intergranular glassy phase partially devitrifies upon annealing. Thus the effect of a fully amorphous or a partially crystalline intergranular phase on the creep properties could be assessed. The creep resistance of the material is only modestly affected by partial grain, boundary devitrification in both flexural and compressive creep. However the creep life is reduced significantly.</p> <p>Extensive analytical microscopy was done on the amorphous and partially crystalline material in order to determine what changes were occurring due to anneling and due to creep deformation. A quantitative methodology was developed for electron energy loss spectroscopy to analyze the intergranular phase composition. Using this technique concentrations of light elements, such as oxygen and nitrogen, and heavier elements could be determined. Different crystalline grain boundary products were found near the surface of samples annealed in air compared to the centre of these and throughout samples annealed in an inert atmosphere. However the residual amorphous phase composition was the same regardless of annealing atmosphere or location. In addition, phases present after devitrification did not depend on the stress state. Extensive cavitation, a commonly observed effect of creep, did not occur in samples containing the maximum obtainable strain in flexure of 2.7%.</p> <p>The observed microstructural information and creep data was taken into consideration in developing a creep model. This model describes creep due to non-linear viscous flow of an amorphous intergranular phase around a hexagonal array of grains. An initial constant strain rate is predicted at low strains, followed by a decrease in strain rate as the intergranular phase is squeezed out from between grains. This decrease occurs at smaller strains in compression than in tension or flexure. The creep behaviour observed experimentally corresponds well with that predicted theoretically.</p> / Doctor of Philosophy (PhD)