| Summary: | An attempt has been made in this thesis to elucidate the effects of growth velocity, temperature gradient and impurity additions on the structure of the aluminium-silicon eutectic, and in particular on the transition from an (unmodified) flaky to a (modified) fibrous morphology. Samples of pure eutectic alloy have been unidirectionally solidified over a range of growth rates and at two temperature gradients, and specimens examined by optical, transmission electron and scanning electron microscopy in order to characterise the flake to fibre transition. It has been found that the transition occurs over a range of growth rates, with, increasing temperature gradient favouring the fibrous morphology. Transmission electron microscopy has confirmed the presence of twins in the flake silicon, but twins have not been observed in the fibrous silicon, although this is not considered to be conclusive evidence that twins were not present. Measurements have been made of the growth undercooling as a function of velocity for the flaky silicon, and it has been found that the flake silicon eutectic can be grown at undercoolings of 15°C and over. In order to compare the effects of different modifying elements, ingots of pure alloy have been doped with various impurity elements and thermally analysed, after which the microstruetures have been examined using optical and scanning electron microscopy. It has been found that sodium, potassium and strontium modify the eutectic while other group II and the group V elements are ineffective. Comparison of growth undercoolings has showed that the modified eutectic grows at undercoolings of 3°C and over, the undercooling under given conditions increasing in the order strontium, potassium, sodium. Lithium has been found to produce an anomalous microstructure. Phosphorous has been found to refine the primary silicon, and crystallographic particles inside the primary crystals have been shown by microprobe analysis to contain phosphorous and aluminium. A hot stage has been designed and constructed in which experiments using eutectic droplets melted onto silicon single crystal substrates have been carried out, Measurements have been made of undercoolings for nucleation of eutectic on silicon showing that silicon does not nucleate the eutectic. The droplets have been found by X-ray analysis to be polycrystalline, so no epitaxial relationship has been found for eutectic aluminium and primary silicon. Growth layers formed by cooling modified and unmodified droplets from above the eutectic temperature have been compared, and it has been concluded that these layers had been formed by layer growth of the {111} silicon substrate. Twinned primary silicon crystals have been observed projecting upwards from the growth layer and it has been concluded that growth using twin plane re-entrant edges is more kinetically favourable than {111} layer growth. The principle of temperature gradient zone melting has been used to compare the migration rates of modified and unmodified droplets across a silicon single crystal substrate in an applied temperature gradient. It has been found that modification by strontium reduces the migration rate. Finally, it has been proposed that morphological transitions in this system depend on the difference between the kinetic undercoolings of the aluminium and silicon phases, and that they do not occur at fixed total undercoolings.
|
|---|
