Microstructural Evaluation of Deformed AISI 316 Steel by X Ray Peak Broadening and EBSD

ANDREA KLIAUGA; MAURIZIO FERRANTE; RAÚL E. BOLMARO

Rio de Janeiro

Congreso; 17th International Microscopy Congress - IMC17; 2010

Microstructure development by plastic deformation and possible later heat treatment include dislocation accumulation, either distributed or in dense arrays, twins, grain fragmentation, etc., and their further annihilation due to recovery and recystallization processes. Many experimental techniques are useful to evaluate, with more or less accuracy, the amount and quality of such microstructural features. Among them, a very old technique, based on X-ray peak broadening measurement, has recently evolved to become a fine tool for the evaluation of dislocation densities, grain size and twin boundary presence. The technique can be considered a “bulk” technique, in the sense that is not able to evaluate those features as field quantities but as average ones. Simultaneously, Electron Back Scattering Diffraction (EBSD) has come on age, revealing as a powerful technique to investigate, at the scale of a few tens of nanometers, the distribution of grain orientations, misorientations, phase relationships, etc.. Paradoxically, such a powerful technique still relays on complex statistical after processing for the evaluation of parameters correlated with plastic deformation and consequent dislocation storage. Because of the known inability of Transmission Electron Microscopy (TEM) to provide statistically significant information still a proper quantification is needed. The current paper shows an attempt to compare data obtained by both techniques in an AISI 316 steel deformed by various strain paths and further several heat treatments.   As Received (heat treated) AISI 316 steel was deformed by Equal Channel Angular Extrusion (ECAE) by route A by 1 and 2 passes, cold rolled at 70% reduction and, this last one, heat treated 1 h at 700 C and 800 C. Fig. 1 shows the Williamson Hall plot of As Received (0X), 1X and 2X ECAPed and 70% rolled samples, from which grain sizes, twinning and dislocation densities can be evaluated [1].   EBSD experiments were performed on similar samples after careful polishing and electropolishing to eliminate surface damage. Fig. 2 shows the EBSD map of EBSD scanning of 70% Rolled, 1h 700o C Heat treated sample. Step size 0.5 µm. Fig 3 and 4 show the low angle distribution (<15o) for Number of Boundary Segments (NBS) and Boundary Lengths (BL) normalized by number of pixels and scanned area, respectively, of several deformed and heat treated samples. The AR sample shows almost no LAB while 1X sample shows a vey well developed distribution of LABs. The heat treated sample results are somehow contradictory, with the 800o C sample showing a larger population for the smallest LAB [3-4].   Dislocation density and grain size are entangled variables despite they can apparently be separated by X ray peak broadening analysis. When dislocations are evenly distributed they exert a stress field that broadens the peaks in a peak dependent way. When they become arranged in Low Energy Dislocation Structures (LEDS) the stress field is a short range field and the influence on peak broadening is peak independent and allows the calculation of grain size. However the so calculated grain size differs from its metallographic concept. It depends on coherent domain sizes separated by small angle misorientations than not necessarily would be reflected on a continuous low angle grain boundary. Moreover, dislocation structures can not always unequivocally be considered as homogeneously distributed or concentrated in LEDS but rather in a mixed way and their influence may remain uncertain. On the other hand EBSD techniques can provide data for calculating misorientations among pixels that not always constitute a recognizable closed grain boundary. NBS and BL, once normalized by the number of scanned pixels or area, respectively, seem to be more appropriate than grain size to infer the influence of this heterogeneity on peak broadening. However they still present a mixed measure of plastic deformation and, at certain scale, it is not possible to distinguish between the relative contributions of what might be a continuous spectrum between Taylor and LEDS arrays.