The formation of texture through thermomechanical treatment was investigated in Mg-18.8 at％ Sc shape memory alloy to enhance its superelasticity at room temperature. The samples were cold rolled in an α phase or in a β phase and then finally heat treated at 690℃ followed by water quenching to obtain a β phase. In the case of cold rolling in the α phase, a basal-plane texture was formed, while no preferential texture was observed along in-plane direction. After the final heat treatment, {011}<uvw>_β transformation texture was obtained, according to Burgers relationship, indicating no improvement of the superelasticity along in-plane direction. In the case of the cold rolling in the β phase, a weak {111}<011>_β recrystallization texture was obtained. The sample showed about 0.65％ superelastic tensile strain along rolling direction, while that along transverse direction (//~<113>β) showed only about 0.43％. This trend is in good agreement with the orientation dependence on the transformation strain, but, the obtained superelastic strain was much lower than the expected value, which is due to the weak texture and suggests the existence of a strong grain constraint in the Mg-Sc shape memory alloy.

The lamination cladding of Nb and Al plates with an inserted Ti foil and heat treatment at 973 K, which was higher than the Al melting point, was applied to the development of Nb-Al intermetallic compounds. It was revealed that the uniform region consisting of fine particles with a grain size of below 1 µm and Al was formed in the vicinity of the Nb substrate and gradually grew toward the interior of molten Al, in which TiAl₃ particles distributed. EDX analysis and X-ray diffraction examination revealed that the constituent phase of the fine particles is a ternary (Ti,Nb)Al₃ of the D0₂₂ crystal structure. The molten Al containing Ti may contribute to the formation of the fine particle regions in contact with the Nb substrate. By heating at 973 K, the front of the fine particle regions proceed toward the molten Al, accompanied by the decomposition of the TiAl₃ particles distributed in the molten Al, resulting in the uniform growth of the fine particle regions without the TiAl₃ particles.

Hydrogen in materials degrades mechanical properties, which is widely recognized as hydrogen embrittlement. To understand hydrogen embrittlement, it is necessary to clarify the accumulation behavior of hydrogen under stress. The neutron radiography and neutron tomography techniques are applied to examine whether hydrogen accumulation behavior can be visualized directly. Palladium specimens with and without hydrogen were prepared for the neutron imaging experiment under stress. Solute hydrogen has caused distinct contrast change in both the neutron radiography and neutron tomography. Hydrogen distribution at a notch-tip in a loaded specimen has not been visualized in the tomographic cross-sectional images. It can be inferred that this is fundamentally attributable to low spatial resolution of the present imaging set-up, and possibility to visualize hydrogen distribution due to loading is discussed.

Industrial properties of aluminum alloys are usually increased by having additives in a liquid state which leads to crystal grain refinement. Conventional methods to confirm the effect of the miniaturization are binarization of electron microscope images and electron probe micro analyzer (EPMA) which require enormous man-hours. The purpose of this research is to establish a method for classifying crystals in significantly shorter period. The proposed method is crystal classification per pixel on microscopic images by applying deep learning. In the experiment, the proposed method was performed on Al-Si-Mg alloys with an abundance ratio of 88:11:1, and the results were compared with the correct values created by the conventional EPMA and manual methods. The percentage of accuracy (the intersection over union (IoU)) came out lower than 30％ in Mg, but approximately 90％ in Si. Therefore, the study shows that this method can be applied with high accuracy if the element is presented in about 10％ of the image.