Collaborative Research: Dynamics of Short Range Order in Multi-Principal Element Alloys

几乎所有在实践中使用的金属都是合金，这意味着它们是不同类型金属原子的混合物。根据合金的不同，不同的原子类型在晶体中以有序或无序的方式排列。不同原子无序或随机排列的合金可以改善性能，包括提高强度或耐腐蚀性。最近有人提出，在一些无序合金中，实际的原子排列可能更微妙，在短距离上看起来是有序的（或非随机的），而在较长距离上则是无序的。这种被称为短程有序的现象与最近开发的一组被称为多主元素合金（mpea）的合金的特殊性能有关，这些合金在发电和国防领域的下一代应用中具有特殊的前景。然而，一个困难是短距离排序仍然难以测量。这项研究将使用一种被称为原子探针的显微镜来探测单个原子的类型和位置，但它的不确定性可能会使对细微顺序的测量无法得出结论。在这里，人工智能将用于检测数据中存在的任何短程排序。利用这种新的分析方法，该项目将测试在稳定条件下的短程排序与合金的制造方式无关的想法。在这样做的过程中，该项目将对理解如何通过加工来设计合金的短程订购产生广泛的影响，所有这些都对mpea将如何发展为下一代材料具有重要意义。在研究的同时，还将创建人工智能教学模块，向初高中学生展示和教育这种新兴技术的使用。这些内容将通过中学教师如何将以材料为中心的例子融入物理、化学和物理科学课堂的研讨会进行传播。技术概述多主元素合金（mpea），通常被称为高熵合金，是一种新兴的合金类别，具有卓越的机械性能和显著的组成设计灵活性。然而，由于缺乏关于化学短程序（SRO）的基本知识，使得mpea的合理设计受到阻碍。SRO是原子种类分布中的局部相关关系。该项目将在CrCoNi MPEA模型中描述SRO的演变，以评估两个假设；首先，mpea中的SRO达到与制造条件无关的稳定状态，其次，稳定SRO的弛豫时间由扩散动力学控制（扩散动力学本身取决于SRO）。样品将通过真空电弧熔化、直流烧结和高压扭转固结制备，以产生可测量的不同初始SRO状态。SRO表征将通过原子探针断层扫描（APT）与电子散射的对分布函数交叉相关以及能量过滤高分辨率透射电子显微镜成像来完成。APT数据集将通过机器学习方法进行分析，其中数据被建模为来自底层两两交互马尔可夫点过程的样本。一系列等温退火实验的实验数据将用于校准SRO与自扩散率相互作用的数学模型。该模型将用于开发mpea集成到服务中的时间-温度- sro图。该项目预计将提供：(1)对mpea成分稳定性的基本科学理解的成熟，这些材料可以在极端环境中投入使用。(2)一种鲁棒的机器学习APT分析方法，将该技术扩展到解决合金中高溶质聚类特征。(3)通过一个材料训练营，指导如何将材料纳入物理、化学和物理科学课程，在研究生和中学教师中培养下一代STEM劳动力。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

NON-TECHNICAL SUMMARYNearly all metals used in practice are alloys, meaning that they are a mixture of different types of metal atoms. Depending on the alloy, different atomic types arrange in either an ordered or disordered way within a crystal. Alloys where the different atoms have a disordered or random arrangement can benefit from improved properties including increased strength or corrosion resistance. It has recently been proposed that in some disordered alloys the actual atomic arrangement could be more subtle, appearing to be ordered (or non-random) over short distances and disordered over longer distances. This phenomenon, known as short range ordering, is implicated in the exceptional properties of a recently developed group of alloys known as multi-principal element alloys (MPEAs) that have particular promise for next-generation applications in power generation and national defense. However, one difficulty is that short range ordering remains problematic to measure. This research will use a microscope called an atom probe to detect the types and locations of individual atoms, but with an uncertainty that can make the measurement of subtle ordering inconclusive. Here, artificial intelligence will be used to detect any short range ordering present within the data. Using this new means of analysis, the project will then test the idea that short range ordering in stable conditions is independent of how the alloy was made. In doing so, this project will have a broad impact on understanding how to engineer short range ordering in alloys by means of their processing, all of which has implications for how MPEAs will develop as next-generation materials. Alongside the research, artificial intelligence teaching modules will be created to expose and educate middle and high school students in the use of this emerging technology. These will be disseminated through a workshop for secondary school teachers on how to integrate materials-centric examples into physics, chemistry, and physical science classrooms.TECHNICAL SUMMARYMulti-principal element alloys (MPEAs), often called high-entropy alloys, are an emerging alloy class with initial evidence of exceptional mechanical properties and significant compositional design flexibility. However, the rational design of MPEAs is hindered by a lack of fundamental knowledge about the chemical short range order (SRO), which are the local correlations in the distribution of atomic species. This project will characterize the evolution of SRO in a model CrCoNi MPEA to evaluate two hypotheses; first, that SRO in MPEAs reaches a stable state that is independent of the fabrication conditions, and second, that the relaxation time to the stable SRO is governed by diffusion kinetics (which themselves depend on the SRO). Samples will be fabricated by vacuum arc melting, direct current sintering, and high-pressure torsion consolidation to generate measurably different initial SRO states. SRO characterization will be done by atom probe tomography (APT) cross-correlated with electron scattering for a pair distribution function as well as energy filtered high resolution transmission electron microscopy imaging. The APT datasets will be analyzed by a machine learning approach where the data is modeled as a sample from an underlying pairwise-interaction Markov point process. Experimental data from a series of isothermal annealing experiments will be used to calibrate a mathematical model for the mutual interactions of the SRO and the self-diffusivity. The model will be used to develop time-temperature-SRO diagrams for MPEAs to be integrated into service. The project is expected to deliver: (1) Maturation of a fundamental scientific understanding of MPEAs’ compositional stability from which these materials can be deployed into service in extreme environments. (2) A robust machine learning APT analysis method that expands the technique to address high solute clustering characteristics in alloys. (3) The development of the next-generation STEM workforce at the graduate level as well as through secondary school teachers via a materials camp that instructs how to incorporate materials into the physics, chemistry, and physical science curriculum.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.