Observations related to healing of interface damage and cavity nucleation during superplastic flow

Two critical observations related to the nucleation of superplastic cavities are reported. Cavities preexisting in a fine grain Al-Mg alloy prior to superplastic deformation have been minimized, but not completely eliminated by a high temperature annealing treatment. This has reduced the density of emerging cavities during subsequent deformation, however nucleation of cavities occur continuously with superplastic strain, starting possibly from partially healed defects of imperceptible (nanoscale) size. The observations raise doubts about the often-used cavity nucleation criterion based on a thermodynamically stable void size. Diffusional cavity growth model, used in conjunction with this assumed "stable" size artificially enlarges its perceived contribution, but still predicts less growth than the observed cavitation levels in superplastic metals. Alternate growth mechanism for submicron size cavities is suggested.