Microstructure and Texture Evolution During Thermomechanical Processing of Al_0.25CoCrFeNi High-Entropy Alloy

Abstract

Microstructure and texture evolution was studied in a Al_0.25CoCrFeNi high-entropy alloy after cold work and subsequent isochronal annealing treatments. The alloy was synthesized by arc-melting and the resulting ingots were cold rolled up to two different thickness reductions: 50 pct and 80 pct. The alloy was then annealed for 1 hour at 973 K, 1073 K, 1173 K, and 1273 K. The final microstructures were analyzed using X-ray diffraction, scanning electron microscopy, and electron backscatter diffraction and the results were correlated with hardness measurements. The sample reduced by 80 pct recrystallized at a lower temperature than the sample reduced by 50 pct, confirming the influence of the deformation degree on the rate of nucleation and growth of new grains. The recrystallized microstructures revealed equiaxed grains with a high density of twins, which is typical of metals with low stacking fault energy. Twins were also observed in the as-rolled samples, suggesting that twinning precedes shear banding during deformation of this alloy. This accounts for the strength of the {110} 〈112〉 brass texture component in the cold rolled samples.

Appendix

Dictionary Indexing (DI) relies on a physics-based forward model to perform accurate quantum mechanical simulation of the electron scattering process.[42] The deterministic Schrödinger’s equation, which produces the characteristic Kikuchi bands is fused with the stochastic inelastically scattered electrons, which form the background. The dynamical signal from elastic scattering modulates the background signal of the inelastically scattered electrons to produce the typical electron backscatter patterns. Unlike the commercial softwares, which rely on the Hough transform feature detection, the DI method performs an exhaustive search of the “pattern dictionary” using the normalized dot product as the similarity metric. The dictionary is just a collection of simulated patterns for identical diffraction geometry but different orientations, which are sampled from a set of uniformly distributed orientations covering the fundamental zone. The orientation of the experimental pattern is labeled the same as the orientation of its closest match in the dictionary.

Once a solution is found, the results are further refined using the forward model. This is necessary to alleviate the artifacts due to the discrete sampling of orientation space in the DI approach. For a well calibrated system, this procedure was shown to produce an orientation accuracy of ~ 0.2 deg for a range of noise in the diffraction pattern.[43] Two samples reduced up to 80 pct presented a heavily deformed microstructure, implying the use of a novel dictionary-based indexing[32,44,45,46] to solve their Kikuchi patterns. The results are shown in Figure A1.

Fig. A1

Comparison between IPF maps indexed by OIM-TSL and the new dictionary indexing of the as-rolled (left) and annealed at 973 K (right) samples reduced by 80 pct