A study of the metallographic structures of near-eutectic alloys of the Al-CuAl₂ system has been carried out, with the major intention of determining the mode of nucleation and growth of the eutectic. It is shown that both nucleation and growth characteristics of the eutectic vary with alloy composition. In eutectic and aluminium-rich alloys, groups of eutectic grains nucleate from an initiating surface of CuAl₂, which must first be precipitated from the melt. Further growth of the eutectic usually occurs in the form of columnar grains. In copper-rich alloys, nucleation occurs on the surfaces of primary CuAl₂ dendrites, but columnar grains of eutectic do not develop. In these alloys, eutectic growth is completely dominated by the dendrite pattern; this distinctive growth behaviour can be related to local variations in melt composition during freezing.
The variations in nucleation behaviour are shown to result from the relative nucleation capacities of the primary phases. CuAl₂ provides a very favourable surface for the nucleation of aluminium, but aluminium is a poor nucleant for CuAl₂. This difference also accounts for the appearance of haloes of CuAl₂ around aluminium dendrites, and the absence of similar haloes around CuAl₂ dendrites.
The Al-CuAl₂ system was selected for study as a typical example of an alloy which forms a normal eutectic structure, and the experimental findings are discussed in the light of an extensive survey of the literature on binary metallic eutectic structures. The present state of knowledge in this field can be summarised as follows;
1. A system of classification has been developed by which eutectic microstructures can be designated as either "normal" or "anomalous". There is a considerable mass of information concerning the wide variety of microstructural features which can occur within each class. A method of classifying these relatively minor features is proposed.
2. There are two theories which account for the formation of an "ideal" lamellar eutectic. A theory advanced by Scheil, based on the relative velocities of growth of the two eutectic phases, provides a theoretical basis for the distinction between normal and anomalous eutectic structures. A later theory advanced by Tiller is based on an analysis of the diffusion conditions at the solid-liquid interface. The most important of the predictions from this theory is a quadratic relationship between inter-lamellar spacing and rate of freezing of the eutectic.
3. A theory of organic eutectic solidification was advanced by Kofler, of which some aspects can be applied to metallic eutectics. The theory is based on experimental measurements of the relative velocities of growth of the two eutectic phases, and leads to quantitative predictions of the limits within which normal eutectic growth is possible. Scheil's theory is an application of these predictions to metallic eutectics.
Although the literature contains a number of proposals concerning the mechanisms by which the two phases of a eutectic are nucleated together, none have been of general application. It is considered that the principles set forth in this thesis can be applied to all systems which form a normal eutectic structure. The experimental work also leads to some expansion of Scheil’s theory, and some of Kofler’s work is applied in the explanations of halo formation. With reference to the more detailed microstructural features, it is suggested that eutectic colony formation does not occur under the experimental conditions used, and some lamellar defects of a new type are noted.