NpjComput.Mater.：高效能輕質難熔高熵合金新突破：資料驅動、逐層多目標設計

海歸學者發起的公益學術平臺

分享資訊，整合資源

交流學術，偶爾風月

輕質難熔高熵合金由輕質的Al、Ti以及難熔Nb、Mo、Hf等元素構成，具有較低的合金密度，同時繼承了傳統MoNbTaWHf基難熔高熵合金高強度、高硬度、耐腐蝕等特性，在高溫結構材料領域具有重要的應用價值。但高熵合金成分空間大、元素間相互作用複雜，五元系合金中不同成分的組合就高達數十萬種。新型高效能輕質難熔高熵合金的設計難度大。同時，由於加入了輕質的Al、Zr、Ti等化合物易形成元素，合金中會出現Laves等複雜的金屬間化合物相，這些硬質相的出現會提升合金的硬度，但卻加劇了基體相與第二相之間的點蝕，顯著降低合金耐腐蝕效能，導致合金硬度和耐腐蝕效能之間存在著強烈的制約關係。因此，如何突破合金效能的限制，在高維成分空間內快速設計出高綜合性能的輕質難熔高熵合金成分是亟需解決的難題。

圖1 資料驅動的超硬超耐蝕輕質難熔高熵合金逐層多目標設計流程圖

來自中南大學粉末冶金國家重點實驗室的張利軍教授團隊，提出了一種結合特徵分析與多目標最佳化設計方法的機器學習驅動合金設計策略，可用少量已報道的實驗資料來訓練模型，進而實現輕質難熔高熵合金相結構與力學效能的準確預測。基於該方法，他們建立了AlNbTiVZrCrMoHf系輕質難熔高熵合金“成分-組織-效能”定量化關係模型，透過對合金相結構、密度、熔點、硬度和腐蝕效能進行逐步地預測和篩選，成功突破合金硬度與耐腐蝕效能的制約關係，實現了三種超硬超耐蝕新型合金成分的高效開發設計。

圖2 資料驅動攜手逐層多目標設計的高效能輕質難熔高熵合金效能與文獻資料的對比

此外，該研究還利用機器學習模型可解釋性分析技術發現：高的價電子濃度（VEC）和低的混合焓（ΔH_mix）有利於獲得高硬度的輕質難熔高熵合金。提高Al元素含量會增強固溶強化效果，而增大Ti元素含量會提升抗氯離子點蝕效能。

圖3 機器學習模型的可解釋性分析

該研究僅透過一次建模設計就實現了高效能輕質難熔高熵合金開發，並得到實驗驗證，進一步表明機器學習驅動的合金設計策略對未來新型合金的設計具有重要意義。該文近期發表於npj Computational Materials10: 256 (2024)，英文標題與摘要如下，點選左下角“閱讀原文”可以自由獲取論文PDF。

Data-driven design of novel lightweight refractory high-entropy alloys with superb hardness and corrosion resistance

Tianchuang Gao, Jianbao Gao, Shenglan Yang & Lijun Zhang

Lightweight refractory high-entropy alloys (LW-RHEAs) hold significant potential in the fields of aviation, aerospace, and nuclear energy due to their low density, high strength, high hardness, and corrosion resistance. However, the enormous composition space has severely hindered the development of novel LW-RHEAs with excellent comprehensive performance. In this paper, a machine learning (ML)-based alloy design strategy combined with a multi-objective optimization method was proposed and applied for a rational design of Al-Nb-Ti-V-Zr-Cr-Mo-Hf LW-RHEAs. The quantitative relation of “composition-structure-property” was first established by ML modeling. Then, feature analysis reveals that Cr content greater than 12 at.% is a key criterion for alloys with high corrosion resistance. The phase structure, density, melting point, hardness and corrosion resistance of the alloys were screened layer by layer, and finally, three LW-RHEAs with superb hard and corrosion resistance were successfully designed. Key experimental validation indicates that three target alloys have densities around 6.5 g/cm3, and all alloys are disordered bcc_A2 single-phase with the highest hardness of 593 HV and the largest pitting potential of 2.5 VSCE, which far exceeds all the literature reports. The successful application in this paper clearly demonstrates that the present design strategy driven by the ML technique should be generally applicable to other RHEA systems.