The penetration of liquid copper into iron grain boundaries has been experimentally analysed. During the penetration, Cu diffuses into the iron grains. In this case, a Kirkendall effect is created, leading to vacancies which diffuse to the grain boundaries where they condense. A grain boundary crack is formed, which will be filled with liquid Cu. The driving force for penetration of the Cu-liquid into the grain boundary is the condensation of vacancies and the surface/interface free energy change. A theoretical model describing the penetration rate is derived and a 41 good agreement between theory and experiments was achieved. The model is general and can be used to explain liquid metal embrittlement.

High temperature tensile properties during solidification have been determined for two Fe±Ni alloys, in order to understand the hot cracking (tearing) mechanism. Tensile tests were made on ªin situº solidified samples. Hot cracking susceptibility was determined by using the transition temperature for area reduction (hot ductility), i.e. the transition temperature between a brittle and a ductile fracture. True strain and ultimate tensile stress were also measured. The results from the tensile tests were compared to the liquidus and solidus temperatures, which were measured by thermal analysis using differential scanning calorimetry (DSC).The Fe-2 %Ni alloy solidifying in the ferritic mode, was not particular sensitive to hot cracking because the transition temperature for area reduction (ductility) occurred just below the solidus temperature. The cooling rate had a minor effect on the transition temperature. Primarily intergranular fractures were formed in the alloy when it solidified to ferrite.The Fe-10 %Ni alloy solidifying in the austenitic mode, was very sensitive to hot cracking as the transition temperature for area reduction (ductility) was found at 1275±1300 C at the lowest cooling rate of 10 C/min. This is about 200 C below the solidus temperature. The cooling rate had a strong effect on the transition temperature for area reduction. Increasing cooling rate (100 C/min) increased the transition temperature by 100 C to 1375±1400 C. The fracture mode of the alloy solidifying to austenite was also dependent on the cooling rate. The low cooling rate gave intergranular fractures with visible grains, while the medium cooling rate gave interdendritic fractures with visible dendritic structure. The experimental results could not be explained by the formation of COMMUNICATIONS 66

Hot cracking formation and its mechanism in invar alloys (Fe-36mass%Ni) during continuous casting was investigated. The invar alloy is very sensitive to hot cracking due to its low transition temperature of brittle to ductile fracture even though its solidification interval is narrow. At the cooling rate of 10°C/min the transition temperature of brittle to ductile fracture is about 113°C below the solidus temperature. An increased cooling rate in invar alloy increases transition temperature of brittle to ductile fracture. It is increased to about 43°C below the solidus temperature when the cooling rate is increased to 100°C/min. A mechanism of hot cracking formation in invar alloys has been proposed. Hot cracks in invar alloys with a fast cooling rate are formed between the primary dendrites due to the equiaxed solidification structure. However, at slow cooling rate, hot cracks are formed between the grain boundaries due to the columnar structure.

The hot crack sensitivity in metals is suggested to be caused by the supersaturation of vacancies created during the solidification process. Equations have been derived to predict the nucleation and growth of cracks by the condensation of vacancies. The transition temperature from brittle to ductile fracture was found to be related to the decrease in the supersaturation of vacancies due to an annealing process. The hot crack sensitivity was observed to be related to the supersaturation of vacancies, the diffusion rate, and the structure coarseness. The effect of surface active elements such as phosphorous and sulphur in steel alloys is discussed.

Today's available theories for hot crack formation are based on the fact that hot cracks are formed in the presence of liquid films in the interdendritic areas or at the grain boundaries, which are exposed to tensile stress. Copper is well known to cause hot shortness in steels. In order to study how the liquid embrittles the material, high-temperature tensile tests were performed at two strain rates during the penetration of liquid copper into Fe-10%Ni. The penetration distance was measured in samples that were exposed to strain without fracturing. The ductility (area reduction), strain and ultimate tensile stress were determined. Microprobe analysis was performed on the fractured samples. The transition temperature of ductility was found at 1400-1450 • C without copper penetration whereas it occurred at 1025-1078 • C during the pene-tration of copper, i.e. copper starts to embrittle at a temperature below its melting point. The microprobe measurements showed that the diffusion rate of copper into Fe-10%Ni was enhanced when the lattice was strained. The results are discussed in terms of a new theory concerning vacancy formation and condensation as the dominating mechanism for hot crack formation during solidification.