文献类型：期刊文献

英文题名：A thermodynamics-informed and data-driven framework for accelerated discovery of ultra-high-temperature oxidation-resistant Al-Co-Cr-Fe-Ni high-entropy alloys at 1300 °C

作者：Shi, Jiayang[1];Shi, Junmiao[1,2];Xu, Wenhu[1];Lu, Tiwen[1];Tian, Fuqiang[1];Zhao, Zheng[1];Cai, Huangyue[3];Zhang, Xiancheng[1]

机构：[1]East China Univ Sci & Technol, Key Lab Pressure Syst & Safety, Minist Educ, Shanghai 200237, Peoples R China;[2]Harbin Inst Technol, State Key Lab Precis Welding & Joining Mat & Struc, Harbin 150001, Peoples R China;[3]Shanghai Jiao Tong Univ, Sch Mat Sci & Engn, Shanghai Key Lab Adv High Temp Mat & Precis Formin, Shanghai 200240, Peoples R China

年份：2026

卷号：270

外文期刊名：CORROSION SCIENCE

收录：;EI(收录号：20262620999310);WOS:【SCI-EXPANDED(收录号:WOS:001808732000001)】；

基金：This work was supported by the National Natural Science Foundation of China (52375146) .

语种：英文

外文关键词：High-entropy alloys; Ultrahigh-temperature oxidation; CALPHAD; Machine learning

摘要：Ultrahigh-temperature (e.g., 1300 degrees C) oxidation of high-entropy alloys (HEAs) is governed by coupled phase stability, elemental activities, and oxide-scale/interfacial integrity, yet oxidation-resistant compositional design is constrained by sparse experimental feedback, inefficient screening, and limited benchmarking against engineering references. Here, we introduce a thermodynamics-informed, data-driven framework for the Al-Co-Cr-Fe-Ni system that integrates 1300 degrees C oxidation measurements with CALPHAD-derived equilibriumphase and thermodynamic descriptors to build a Delta W predictor for rapid global screening. A hyperparametertuned random-forest regressor predicts oxidation weight gain (Delta W), achieving R2 = 0.990/0.940 and MAE = 0.046/0.105 mg & sdot;cm-2 on the training/test sets; five-fold cross-validation yields R2 = 0.931 +/- 0.029 and MAE = 0.133 +/- 0.023 mg & sdot;cm-2, confirming robust generalization. Interpretability analyses identify Delta Hmix, Al/Cr contents, and the B2 (beta-NiAl-type)-FCC (gamma) phase balance as key determinants of Delta W, and screening shows that low Delta W concentrates within a beta/gamma dual-phase synergy window rather than following a monotonic "higher B2/ lower FCC" trend. Representative non-equiatomic alloys from this window form continuous, dense, adherent alpha-Al2O3 scales with markedly shallower Al depletion during long-term exposure near 1300 degrees C. Head-to-head comparisons versus a reactive-element (Y/Hf) microalloyed NiCoCrAl benchmark and an equiatomic reference show a decrease in parabolic rate constant from 5.6 & times; 10-1 2 to 3.7 & times; 10-1 2 cm2 & sdot;s-1 , together with markedly improved spallation resistance. The framework is expected to accelerate bond-coat design and screening under extreme thermal exposure near 1300 degrees C in advanced hot-section protection systems.