Computational materials design: Composition optimization to develop novel Ni-based single crystal superalloys.

[Display omitted] • A computational materials design to develop novel single crystal superalloys. • Achieving design balance of multi-objective properties. • Comprehensive design through DFT, empirical models, CALPHAD, and machine learning. • 12 alloys selected from 779,625 alloys in trade-offs between different factors. The computational materials design is performed to develop novel single crystal superalloys with balance of multi-objective properties. The entire design process is carried out by materials computation from systems selection to composition determination. The design rules are proposed to optimize compositions of Ni-based single crystal superalloys for materials characteristics, using first-principles calculations, CALPHAD calculations, theoretical models, and machine learning. By first-principles calculations, the effect of alloying elements on structural, elastic, electronic properties of Ni 3 Al are investigated and the Ni-Al-V-Cr-Nb-Mo-Ta-W-Re systems are determ5ined. Application of theoretical models and CALPHAD calculations allows the large materials exploring space to become narrow, based on materials design criterion: microstructure, density, castability, and processability. The creep resistance of remained alloys is estimated by using machine learning and creep merit index. There are 12 alloys selected from 779,625 composition combinations, reaching the balance of multiple design properties. With comparing to commercial superalloys, the selected alloy has excellent performance in trade-offs between different factors, specifically prominent in the aspect of Cr-Al space of oxidation resistance. This design procedure is expected to reduce excessive consumption of cost and time in the process of trial-and-error testing, providing a guide on developing potential Ni-based superalloys systems. [ABSTRACT FROM AUTHOR]