Development of γ′-strengthened CoCrNi-based multi-principal element alloys via machine learning and multi-objective optimization

CoCrNi-based multi-principal element alloys (MPEAs) with face-centered cubic (FCC) structure are known for their exceptional tensile ductility and fracture toughness. However, their insufficient strength limits practical engineering applications. To overcome this limitation, this study develops an integrated strategy combining machine learning (ML) trained on thermodynamic simulation data with multi-objective optimization for compositional design. The approach aims to achieve a high γ′ phase volume fraction, high solvus temperature, and a high-entropy matrix, while suppressing the formation of detrimental precipitates. Key optimization criteria include the γ′ phase volume fraction, solvus temperature, topologically close-packed (TCP) phase content, and matrix mixing entropy, with the goal of synergistically enhancing mechanical properties. An optimized composition, Co_23.3Cr_17.4Ni_48.1Al_6.1Ti_5.1 (at.%), was identified and experimentally validated. The resulting alloy in the as-cast state exhibits an outstanding balance of strength and ductility, demonstrating a yield strength of 786 ± 15 MPa, an ultimate tensile strength of 1049 ± 6 MPa, and an elongation of 25 ± 2%. This ML and multi-objective optimization-based material design strategy not only enables efficient discovery of target compositions for γ′-strengthened CoCrNi-based MPEAs but also offers a novel pathway for advanced alloy development.