In DC-cast 1xxx-series aluminum alloy ingots, the presence of Fe and Si, and cooling rates increasing from < 1°C/s in the ingot center to ~20°C/s near the surface cause the formation of metastable intermetallic Al6 Fe and AlmFe compounds in addition to the stable Al 3Fe. The highly contrasting etching/anodizing response of the Al 6Fe and AlmFe particles forming in the adjoining central and peripheral regions, respectively, creates the fir-tree zone (FTZ) defect in DC-cast ingots. This, in turn, causes streaking in Al sheets. / The need was great for a good fundamental understanding of the mechanism of metastable Al-A1mFe eutectic phase formation. Therefore, the mechanism was studied via the kinetic (experimental: Bridgman growth and DSC analyses), thermodynamic (macroscopic: phase diagram computation), and crystallographic (atomistic) approaches. / The controlled growth of Al-Al3Fe, and metastable Al-Al xFe, Al-Al6Fe, and Al-AlmFe eutectic phases was obtained in AA1050 and its binary analogue Al-0.33wt% Fe alloys via Bridgman growth at various growth velocities (V = 0.15 -60 mm/s). For each phase, by acquiring a series of growth-velocity-dependent invariant (for binary alloys) and univariant ( for AA1050) melting temperatures (T) via DSC, the binary and ternary melting and growth kinetics were determined as T vs ln(V) plots extrapolated to V = 0.001 mm/s. / Using the binary melting kinetics and tentative growth kinetics plots, the Gibbs free energy parameters of binary phases in the AA12S database were optimized, and the previously non-existent binary and ternary stable and metastable equilibrium phase diagrams (computed at V = 0.001 mm/s) and stable and metastable kinetic and dynamic phase diagrams (computed at the DC-casting velocity; V = 1 mm/s) were obtained via ThermoCalc(TM).

The present work was performed on primary A319.2 alloy to investigate the effect of magnesium and strontium additions on the heat treatment behaviour and mechanical properties of the alloy. Samples were cast in permanent mould, then cut in order to obtain coupons which were solution heat treated at 495°C for 4 and 24 hours, followed by quenching in hot water. Two different artificial aging temperatures were applied at 180°C and 220°C for times from 0.5 h up to 24 hours. The investigation consisted of a systematic measurement of macrohardness and matrix microhardness of samples under different heat treatment conditions. Tensile properties were also measured and electron microscopy study of critical samples was done. / It is seen that the alloy does not exhibit the common peak-overaged aging conditions due to the presence of several coexisting phases. It was found that multiple aging peaks are present in the age-hardening curves of Al-Si-Cu-Mg alloys. From the scanning electron microscopy studies, it is seen that addition of magnesium (~0.40 wt%) leads to the precipitation of coarse A1 5Mg8Si6Cu2 particles. Strontium modification of the high magnesium alloys leads to segregation of Cu-intermetallics in areas away from the growing Al-Si eutectic regions. It was found that strontium modification retards the precipitation of the silicon out of the matrix. / The effect of aging time and temperature on the structure of the age-hardening precipitates was studied with TEM. TEM observations reveal that the age-hardening behavior is related to the precipitation sequence of the alloy. For the Al-Si-Cu-Mg alloy, the first aging peak is the result from the high-density GP zones, while the subsequent are from metastable phases. The transition from GP zones to metastable phases may be the main reason for the formation of multiple aging peaks.

The flotation process is designed to collect hydrophobic particles by attachment on to bubbles dispersed in slurry. The effectiveness of the flotation process depends on the properties of the bubble dispersion. For example, the ability of a machine to form small bubbles at a given flow rate of air will control the quantity of surface area available for bubble-particle collision. / The gas dispersion, defined for the purposes of this thesis as bubbles suspended in slurry, is quantified by a group of "so-called" gas dispersion parameters including: superficial gas velocity (gas rate, Jg), gas holdup (epsilong), bubble size (Db) and bubble surface area flux (Sb). / Cell characterization is a commonly used term in the study of gas dispersion parameters referring to any test designed specifically to increase understanding of the gas dispersion inside a particular machine, including how the dispersion can be manipulated by adjusting the operating variables and how the dispersion parameters vary with location inside a machine. / Industrial tests were done to study mechanisms of manipulating the gas dispersion in forced air and self-aerated mechanical flotation machines, demonstrating that gas rate can be used to manipulate the gas dispersion in forced air machines, while froth depth, impeller speed and impeller submergence can be used to manipulate the dispersion in self-aerated machines. / Facilitated by the development of the multi-Jg sensor, a technique based on gas rate mapping was developed for down-the-bank sample point selection. Experience using the sensor is described. Based on industrial tests, recommendations regarding sample point selection for down-the-bank optimization are offered. / The development of an axial pressure profile sensor is described, facilitating axial investigation of variation in gas dispersion parameters. The proposed technique can be used to evaluate machine operating conditions and to resolve previously unexplained results.

Magnesium has limited formability at room temperature due to the hexagonal close-packed (hcp) structure with an axial (c/a) ratio of 1.6240 which provides only basal slip activation. Other hcp metals with different c/a ratios such as titanium have non-basal slip at room temperature. It can be postulated that changing the c/a can change the balance of deformation mechanisms and influence formability. Solutes are known to change the lattice parameters of a metal if in solid solution and hence affect deformation behavior. The objective of this project was to determine the effects of solute additions on the c/a ratio of magnesium and to attempt and assess the resulting deformation behavior. / Changes to the lattice parameters of magnesium with addition of solute elements, Ce, Li, In, and Zn, have been investigated. Binary solid solution alloys as well as multicomponent alloys were cast in permanent copper moulds and homogenized at temperatures between 375°C and 400°C. Annealed powder specimens were prepared and lattice parameters measured using a Philips PW1710 powder diffractometer with Cu-alpha radiation. The experimental X-ray diffraction (XRD) results were refined and analyzed by means of the Rietveld method while characterization of the alloys was performed with optical and scanning electron microscopy and electron probe microanalysis. / The effects of solutes on lattice parameters were explained on the basis of atom size differences and the change in electron overlap of magnesium due solute additions. It was found that addition of a lower valent solute such as Li caused a contraction of the lattice parameters and a resulting axial ratio decrease. Addition of Indium, however, resulted in a contraction of the lattice parameters, but an increase in the axial ratio due to its higher valency. Conversely, neither Zn nor Ce was found to have an effect on the axial ratio. / Characterization of the binary alloys revealed grain size can vary with the type and the amount of solute present in the pure Mg. Zinc was found decrease the grain size of Mg even at low solute content, while the Li and In concentrations had to be increased in order to achieve the same degree of grain refinement. Significant grain refinement was also observed with low Ce additions, mostly because of the presence of Ce-rich second phases. / Preliminary deformation tests were also carried out on three different alloys in order to see if changes to the axial ratio affected formability. Rolling and compression tests revealed that alloys with a finer as-cast microstructure resulted in more consistent flow curve shapes and smoother compressed surfaces.

Hot tearing is a common defect during DC casting of some commercial aluminum alloys and can result in considerable productivity loss in the cast-house. In order to better understand the hot tearing, a novel technique, dubbed DCSS for Direct Chill Surface Simulator, has been developed to reproduce the cooling conditions at the mould/liquid metal interface and to generate the unique surface and sub-surface microstructure of the DC ingot. The apparatus has been designed also to apply and measure a tensile load and the surface strain. These mechanical quantities were then used to derive the stress-strain curves that best represent the material behaviour during solidification and its capacity to resist hot tearing. / The present work describes the approach used to determine the mechanical resistance of three different aluminum-silicon binary alloys (AI-0.5wt%Si, Al-1.5wt%Si, and Al-2.5wt%Si). In addition, the present work covers the modeling of the heat transfer encountered during the test using commercial software (ProCAST(TM)) to better understand the temperature field upon solidification and tracking the solid fraction. Microstructure analyses were made to obtain various metallurgical quantities (e.g., grain morphology, size), physical measurements and thorough investigation of the torn surfaces. Information was used as inputs to the various models developed and used in the present work. / A theoretical model was updated from previous work using more realistic metallurgical quantities. This research led also to the development of a probalistic model (cellular automata) to simulate the microstructure of the cast sample. The model has been used to determine a crack propagation coefficient (CPC) that was used in the theoretical model to better represent the hot tear propagation. / Work to include even more complex coupled phenomena such as creep phenomena and microsegregation are not covered in the present scope of this work. Hypotheses are raised according to experimental work and observations made that suggest their most probable contributions. These studies have advanced to a certain degree the scientific understanding of hot tearing such as the inherent mechanical behaviour during solidification. In addition, the DCSS was used successfully to rank DC cast commercial wrought alloys in terms of hot tearing susceptibility.

No description available.

Techniques based on electrical conductivity to estimate interface level, gas holdup and bias rate in flotation columns were developed. / A conductivity probe and associated data acquisition system for locating the froth/collection zone interface in flotation columns were developed and tested. The level detection technique is based on the collection of a conductance profile around the interface and on the location of the position at which a sharp change in conductance occurs. Such a change in conductance across the interface is caused by the difference in the effective conductivity of the froth and collection zone, primarily due to their difference in gas holdup. / A conductivity cell to measure the effective conductivity of water-air, water-mineral, and water-mineral-air systems was developed. The cell consisted of two grid-electrodes covering the entire cross-sectional area of the cylinder containing the two or three phase system. Such an arrangement allowed the free movement of the phases and provided conditions for uniform potential and current (electrical) fields. It was found that Maxwell's model (1892) predicted the holdups of the non-conductive material reasonably well from the conductivity measurements. / Conditions encountered in industrial flotation columns allowed the use of conductivity to trace the flows of feed water and wash water across the interface, and therefore, to determine the bias rate. Alternative parameters to bias rate as measures of metallurgy such as fraction of feed water in the overflow water (called here feed water entrainment), feed water recovery, and conductance profiles, were explored.

The influence of boron distribution on the precipitation of Nb(C,N) and on austenite recrystallization was studied by means of a stress relaxation technique. A plain carbon steel containing 0.026% C, used as the reference material, and three other steels alloyed with 0.003% B, 0.055% Nb and 0.003% B + 0.055% Nb were employed. The microstructural evolution during relaxation was interpreted in terms of the pinning effect of precipitates on mobile dislocations. / C-shaped precipitation-time-temperature diagrams were determined for both the Nb + B and Nb steels. The presence of boron accelerates precipitation so that it begins at higher temperatures and after shorter times. Also, the combined addition of Nb and boron delays the recrystallization start time at 1000$ sp circ$C and suppresses the partial recrystallization of austenite observed at 950$ sp circ$C. / Samples were quenched or control cooled at different stages of the tests and the evolution of precipitation and the location of boron and boron compounds were studied by various microanalytical techniques. The occurrence of the non-equilibrium segregation of boron after deformation and during recrystallization at austenite grain boundaries was revealed in this way. The observations are interpreted in terms of the interaction between vacancies and boron atoms. / The strong retardation of recrystallization and the acceleration of precipitation observed when both Nb and boron are present is explained in terms of the formation of Nb-B complexes and the increase in the effective concentration of precipitate forming interstitials. At deformation temperatures of 950$ sp circ$C and above, boron appears to increase the solute drag effect of Nb, leading to greater delays in recrystallization. At lower temperatures, boron plays its role in accelerating the nucleation stage of precipitation. / Finally, changes in the size distribution of the precipitates as a function of composition, temperature and time were measured. These data are employed in the framework of the diffusion controlled nucleation and particle growth theory. Expressions for the Nb(C,N) precipitation start times and for the diffusion of Nb in austenite were obtained which are in accord with the values reported in the literature.

Following melt alloying and chlorine fluxing in reverberatory-like aluminum holding furnaces, it is common practice to allow the melt to settle for about an hour prior to casting operations. The purpose of this procedure is to allow inclusions to settle to the surface bottom or to float out to an overlying layer of dross. Mathematical modelling of this process reveals that significant natural convection currents are generated, particularly during the early part of this holding period, and that as these currents gradually diminish, so does the rate of precipitation of inclusions. A 6.25 ton pilot scale holding furnace was used to test the model's predictions of changes in metal quality with holding time. Satisfactory agreement was observed between predicted and "LiMCA"$ sp1$ measured changes in inclusion density levels of (Ti-V)B$ sb2$. Both approaches revealed exponential-like decays in inclusion density levels, with time constants in order of 10 to 60 minutes. ftn$ sp1$LiMCA--Liquid Metal Cleanliness Analyzer

The SAG Power Index test (SPI) is a tool for forecasting autogenous mill performance. Much effort has been directed towards the development of the process models relating SPI to throughput estimates, but little has been directed towards ore body modeling. Blending studies are presented showing that the SPI is not additive, affecting the geostatistical procedures. A method is given to ensure that additivity is respected during geostatistical interpolation. A procedure for relating mean SPI precision to sample spacing is given. This procedure is combined with a study of the process model error to estimate the precision of the mean throughput forecasts. A case study from Chino Mines is discussed. The relative throughput error is between 20 and 26 percent for perfect knowledge of SPI. For 100 m sample spacing at Chino, one third of the error is due to the process models and the rest to the SPI estimates.