During a study of mechanical properties of recrystallized B-free Ni3Al single crystals, regularly spaced parallel traces within individual grains were discovered on the surfaces of thin recrystallized sheets, see Fig. 1. They appeared to be slip traces, but since we could not find similar observations in the literature, a series of experiments was performed to identify them. We will refer to them “traces”, because they contain some, if not all, of the properties of slip traces. A variety of techniques, including the Electron Backscattering Pattern (EBSP) method, was used to ascertain the composition, geometry, and crystallography of these traces. The effect of sample thickness on their formation was also investigated.In summary, these traces on the surface of recrystallized Ni3Al have the following properties:1.The chemistry and crystallographic orientation of the traces are the same as the bulk. No oxides or other second phases were observed.2.The traces are not grooves caused by thermal etching at previous locations of grain boundaries.3.The traces form after recrystallization (because the starting Ni3Al is a single crystal).4.For thicknesses between 50 μm and 720 μm, the density of the traces increases as the sample thickness decreases. Only one set of “protrusion-like” traces is visible in a given grain on the thicker samples, but multiple sets of “cliff-like” traces are visible on the thinner ones (See Fig. 1 and Fig. 2).5.They are linear and parallel to the traces of {111} planes on the surface, see Fig. 3.6.Some of the traces terminate within the interior of the grains, and the rest of them either terminate at or are continuous across grain boundaries. The portion of latter increases with decreasing thickness.7.The grain size decreases with decreasing thickness, the decrease is more pronounced when the grain size is comparable with the thickness, Fig. 4.8.Traces also formed during the recrystallization of cold-rolled polycrystalline Cu thin sheets, Fig. 5.
Fe-based superconducting transition temperature modeling by machine learning: A computer science method