| Abstract Scope |
Refractory compositionally complex alloys (RCCAs) are emerging as promising candidates for next-generation high-temperature materials. A key challenge in RCCA development lies in navigating their high-dimensional compositional space to identify alloys with optimal properties. In this work, we present a fast and robust framework for RCCA composition design, based on nine refractory elements (Ti, V, Cr, Zr, Nb, Mo, Hf, Ta, and W). Thermodynamic properties of binary and ternary systems were calculated using the CalPHAD for single-phase solid solutions and DFT for Laves and B2 phases. To capture the relationship between composition and mechanical performance, we developed a theory-informed machine learning (ML) model. The ML model predicting temperature-dependent yield strength achieved a Rē value of 0.98 over the full temperature range. The influence of each elemental constituent on six key properties—including compressive ductility—was systematically analyzed. These results provide valuable insights that can inform experimental design and accelerate RCCA discovery. |