Data-driven insights into oxidation and phase stability behavior of refractory high-entropy alloys

Refractory high-entropy alloys are promising candidates for ultra-high-temperature structural applications; however, their practical use is limited by insufficient oxidation resistance and phase instability during prolonged thermal exposure. These two degradation mechanisms are strongly interconnected but are often investigated independently, hindering a comprehensive understanding of high-temperature alloy performance. In this work, a data-driven framework is employed to analyze the coupled behavior of oxidation resistance and phase stability in refractory high-entropy alloys. A literature-derived dataset comprising alloy compositions, oxidation conditions, and experimentally reported specific mass gain values was constructed. Oxidation resistance was quantified using specific mass gain, while phase stability was represented by a semi-quantitative indicator derived from reported microstructural observations. Machine-learning analysis was implemented in MATLAB using ensemble-based regression models to extract dominant trends and controlling variables. The results reveal that oxidation resistance and phase stability are governed by distinct but complementary factors: aluminum and chromium strongly influence oxidation behavior, while refractory elements such as niobium play a critical role in phase instability.