New York Paper - Prevention of Columnar Crystallization by Rotation during Solidification (with Discussion)

That the quiescence of a liquid while it is solidifying should favor the formation of columnar crystals, normal of the cooling surface, is seen readily on considering the mechanism of solidification. First, each particle of any composite liquid, whether it be an aqueous solution or a molten metal, in solidifying splits up into two parts, different in composition and hence in fusibility. One part is infusible at the existing temperature, and hence solidifies, and in general attaches itself to the inclosing walls of metal which have already solidified; A, Fig. 1. The other part is fusible at the existing temperature and hence remains molten. In the case of carbon steel, the part of each drop which actually solidifies is poorer in carbon than the drop itself was before it began to solidify, and this impoverishment of the solidifying half-drop enriches the other half-drop in carbon, and thus makes it unfreezable at the existing tern- By this mechanism there arises during solidification a "littoral" or shore layer of liquid B, Fig. 1, bathing the already solidified walls, and more fusible than either those walls or the great remaining central mass of liquid or " deep sea " C, from which it separates them. It is essential that we grasp clearly this conception of a fusible littoral molten layer coating the already solidified walls and separating them from the deep sea. Meanwhile heat is flowing rapidly outward through these walls, its escape cooling them, so that if any given particle of the deep sea metal could get past this littoral layer and come into contact with the solid walls, it would in turn solidify, and like its predecessors would split up in solidifying into a less fusible half-drop which would attach itself to those . walls and a more fusible one which would remain molten. Thus we see