Macrosegregation in casting and ingots

The various types of macrosegregation are described and their commercial importance discussed. Previous work on macrosegregation reported in the literature is reviewed and those theories currently most popular for explaining the various segregates are summarised. A number of commercial castings and ingots were examined. The majority were found to contain macrosegregation although its magnitude seldom exceeded ± 10% of the nominal composition. It was felt that contraction fluid flow had been the dominant segregation mechanism in operation, which was in line with the short freezing times of all the castings studied. The critical solidification time required for the gravitational interdendritic fluid flow (GIFF) mechanism to produce significant amounts of segregation was determined. An alloy of Al-5½wt%Cu was cast into a series of sand moulds of different volumes, and the macrosegregation present in the solidified ingots was related to the various observed freezing times. A total freezing time of approximately 15 mins. had to be exceeded before gravity flow segregation became appreciable. A detailed review is given of previous work concerning the nature and extent of 'GIFF', particularly with respect to position in the semi-solid zone. Inconsistencies between conclusions are highlighted and explained in terms of differences in presentation of results. It is shown that all previous work indicates that 'GIFF' is important in producing macrosegregation at high liquid volume fractions (gL > 0.80) in the mushy zone, but that none of the work by design could indicate whether flow at lower gL was important. An experiment was therefore set up to study the amount of segregation produced by 'GIFF' at various gL in the mushy zone. A quenching technique was used. Thin section, 'slice castings' of Sn-5wt%Pb were grown undirectionally horizontally for various lengths of time under identical conditions and then quenched. The progress of segregation at given reproducible positions in these castings was determined and related to the progress of the semi-solid zone across the length of the castings. It was shown that although a large part of the final macrosegregation was produced in the region just behind the dendrite tips (1 > gL > 0.8) a significant amount of segregation was produced by 'GIFF' up to approximately gL = 0.30. A detailed description of the apparatus and experimental procedure employed is given together with a summary of the experiments conducted to ensure truly reproducible growth conditions between castings. A radioactive tracer technique was used to provide direct visual evidence of fluid flow occurring deep within the semi-solid region of a casting. Radioactive material was inserted into the bulk liquid region of a solidifying 'slice casting' and the subsequent penetration of activity into the original mushy zone was studied. It was estimated from autoradiographs that 'GIFF' was still occurring up to at least gL = 0.40, thus confirming the results of the 'slice casting' experiments. A ternary Sn-Zn-Pb alloy with zero density change down its liquidus line was developed. This was a Sn-base alloy with a Zn:Pb ratio of 19.97:1 (atomic %) and a total solute content of < 15.2 at .%. The alloy was developed using a combination of a theoretical treatment and a highly accurate density measurement technique (pycnometer) . It was confirmed using the 'slice casting' apparatus that such an alloy had no susceptibility to 'GIFF' segregation. Finally, the analysis technique employed for composition measurements is described, and the operating conditions required for optimum analytical accuracy are discussed.