• Effect of chemical heterogeneities on plasticity and strength is studied using atomic simulations in high-entropy alloys. • Elemental anisotropy/difference factor is proposed, and then used to evaluate the short range order and predict the HEA properties. • Elemental anisotropy factor for optimal high performance is accurately estimated by machine learning. • By a machine learning approach coupled with MD simulations, elemental anisotropy ranges from 2.9 to 3. The chemical element distributions always strongly affect the deformation mechanisms and mechanical properties of alloying materials. However, the detailed atomic origin still remains unknown in high-entropy alloys (HEAs) with a stable random solid solution. Here, considering the effect of elemental fluctuation distribution, the deformation behavior and mechanical response of the widely-studied equimolar random CoCrFeMnNi HEA are investigated by atomic simulations combined with machine learning and micro-pillar compression experiments. The elemental anisotropy factor is proposed, and then used to evaluate the chemical element distribution. The experimental and simulation results show that the local variations of chemical compositions exist and play a critical role in the deformation partitioning and mechanical properties. The high strength and good plasticity of HEAs are obtained via tuning the chemical element distributions, and the optimal elemental anisotropy factor ranges from 2.9 to 3 using machine learning. This trend can be attributed to the cooperative mechanisms depending on the local variational composition: massive partial dislocation multiplication at an initial stage of plastic deformation, and the inhibition of localized shear banding via the nucleation of deformation twinning at a later stage. Using the new insights gained here, it would be possible to create new metallic alloys with superior properties through thermal-mechanical treatment to tailoring the chemical element distribution.
Read full abstract