Quasi in-situ investigation on kinking behavior of Fe-Cr-Al alloy during thermal compression.

As a rare plastic deformation mechanism, kinking has attracted substantial attention because of its additional softening effect on the matrix and its heterogeneous refinement on the original grains. However, the details of kinking deformation and the influence of crystal orientation on itself have not been sufficiently elucidated yet, thereby resulting in inadequate microstructure optimization associated with kinking. In this work, the kinking formation process in Fe-Cr-Al alloy during thermal compression deformation has been investigated by using quasi in-situ electron backscatter diffraction (EBSD). By employing the average Schmid factor (Average-SF) model, the influence of crystal orientation on kinking behavior is predicted accurately. The results suggest that the softening effect of kinking on the matrix can be attributed to the fact that the nucleation of pre kink bands (pre-KBs, misorientation <15°) leads to a pronounced reduction in geometrically necessary dislocations (GNDs) density within them. Besides, the ripening behavior of pre-KBs is closely associated with the variation of GNDs density and the overall slip activation capability of the matrix, and is essentially dependent on crystal orientation. The harder the crystal orientation, the smaller the Average-SF, the higher the increment rate of GNDs density, the larger the ripening rate pre-KBs. Under equivalent strain, orientations close to 〈111〉// compression direction (CD) have the highest probability of forming mature KBs with high angle grain boundaries (HAGBs, >15°), whereas KBs in those close to 〈001〉//CD almost always remain as pre-KBs with low angle grain boundaries (LAGBs). • The kinking formation of Fe-Cr-Al alloy is investigated by quasi in-situ EBSD. • The softening effect of kinking on the matrix is explained by EBSD and TEM. • The association of ripening of pre-KBs with GNDs density variation is revealed. • Kinking dependence on crystal orientation is predicted by Average-SF model. [ABSTRACT FROM AUTHOR]