Machine learning guided design and deformation mechanisms of dual-phase Al0.4Co0.5VxFeNi high-entropy alloys

This study employs machine learning to guide the design of Al 0.4 Co 0.5 V x FeNi high-entropy alloys, and the Al 0.4 Co 0.5 V 0.6 FeNi composition is selected for experimental validation. Based on phase stability criteria, including valence electron concentration and atomic size mismatch, this alloy is identified as an FCC + BCC dual-phase high-entropy alloy. Machine learning predicts a tensile strength of approximately 1082 ± 133 MPa and an elongation of about 20.2 ± 3.9% at V = 0.6, results that align closely with experimental measurements. The as-cast alloy exhibits a eutectic dual-phase structure, comprising an FCC matrix and an Al-rich B2 phase. Following cold rolling deformation, this dual-phase structure is retained, while diffraction peaks broaden, grains undergo significant refinement, and high density dislocations are introduced. During deformation, the FCC phase accommodates strain through twinning, whereas the BCC phase undergoes dislocation cross-slip. Strain incompatibility at phase boundaries further induces interfacial sliding and gives rise to nanoscale periodic structures. After 30% cold rolling, the yield strength increases markedly from 771 MPa in the as-cast state to 1553 MPa, and the tensile strength rises from 1121 MPa to 1640 MPa. This work establishes a comprehensive framework for the composition design and deformation strengthening of dual-phase HEAs, systematically elucidating the synergistic interplay among phase stability, microstructural evolution, and mechanical performance.