CAREER: Understanding Grain Boundary Strength via Adaptive Electron Backscatter Diffraction and Multiscale Analysis

人们使用金属和合金已有几千年的历史，并通过开发新的制造方法不断提高它们的强度和性能。这种改进金属的过程很大程度上依赖于猜测和检查。尽管人们已经变得更善于猜测和更快地检查，但智能地设计更坚固、更安全、更有弹性的金属的基础知识仍然缺失。改善金属和合金的部分挑战在于，它们的性能取决于数万亿个比几个原子大不了多少的缺陷的类型、密度和分布。这些缺陷的两种重要类型是位错，它允许金属变形，以及晶界，它在金属中移动时作为位错的障碍。晶界的特性影响了位错在材料中移动的容易程度，进而影响了金属的强度。然而，在晶界的性质和位错运动的屏障强度之间的直接联系还没有建立起来。通过在百万分之一米甚至更小的超小长度尺度上观察晶界，该项目将建立这种直接联系。为了做到这一点，电子显微镜，能够成像材料到单个原子的水平，将被用来观察当材料被弯曲时，位错是如何在材料的晶界附近积累的。人工智能（AI）将被内置到电子显微镜中，以便快速和自动地探索数以万计的晶界，以获得对晶界特征如何与其强度相关联的统计理解。这种理解将有助于指导开发更强、更安全、更持久的新金属和合金。这项工作的范围更广，包括将来自代表性不足社区的高中生纳入暑期实习研究项目。这些实习生将与该项目支持的研究生密切合作，调查金属的强度，并将制定包含虚拟现实元素的课程计划，以便在下一学年带回课堂。暑期实习还将包括参观诺贝丽斯研究中心，这是一家总部位于亚特兰大附近的全球人工智能公司。长期以来，人们已经认识到晶界在决定金属和合金的力学行为和失效敏感性方面的核心作用。然而，由于对决定单个晶界强度的因素的不完全理解，理解晶界特征变化如何影响材料性能的努力受到了阻碍。本项目的目的是确定决定晶界强度的特征，这里定义为晶界对位错传播的阻挡强度。将开发一种新的自适应网格重新划分电子背散射衍射（AR-EBSD）方法，结合在线处理和自动自适应网格重新划分，快速采样数万个晶界，需要建立一个可以应用机器学习方法的库。这种方法将与透射电子显微镜（TEM）表征和原子模拟相结合，以将晶界强度与位错传递机制联系起来。这种耦合方法将促进对晶界空间的前所未有的探索，就研究的晶界数量而言，允许建立严格的晶界强度函数。此外，在本研究过程中开发的多尺度电子显微镜技术将广泛应用于研究环境和极端条件下材料变形的材料表征工具箱。此外，一个针对来自弱势群体的高中生的“可视化科学”暑期实习项目将为弱势群体参与STEM研究创造一条渠道。这些实习生将与该项目支持的研究生一起研究金属的韧性断裂行为，学习机械测试和表征技术，并参观总部位于亚特兰大附近的全球人工智能公司诺贝丽斯研究中心。为了扩大这一项目的影响力，实习生们还将制定包含虚拟现实元素的课程计划，并将其带回高中课堂。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 SUMMARYPeople have been using metals and alloys for thousands of years and have constantly improved their strength and performance through developing new ways to make them. This process of improving the metals has relied largely on guess and check. Although people have become better at guessing and faster at checking, the foundational knowledge to intelligently design stronger, safer, more resilient metals is still missing. Part of the challenge in improving metals and alloys is that their performance is dependent on the type, density, and distribution of trillions of defects that are not much bigger than a few atoms. Two important types of these defects are dislocations, which allow metals to deform, and grain boundaries, which act as barriers to dislocations as they move through the metal. The character of a grain boundary influences how easily dislocations can move through the material, and in turn, affects the strength of the metal. However, a direct link between the character of a grain boundary and how strong of a barrier it is to dislocation motion has not been established. By looking at grain boundaries at ultra-small length scales of one millionth of a meter and smaller, this project will establish that direct link. To do so, electron microscopes, capable of imaging materials down to the level of individual atoms, will be used to see how dislocations accumulate in the material near grain boundaries while the material is being bent. Artificial intelligence (AI) will be built into the electron microscopes in order to rapidly and automatically explore tens of thousands of grain boundaries to obtain a statistical understanding of how grain boundary character is connected to its strength. This understanding will be instrumental in guiding the development of new metals and alloys that are stronger, safer, and longer lasting in application. The broader outreach of this work includes integrating high school students from underrepresented communities in a summer internship research program. These interns will work closely with graduate students supported by the program to investigate the strength of metals and will also develop lesson plans incorporating virtual reality elements to take back to their classes in the following school year. The summer internship will also include visits to the Novelis research center, a global Al company with research headquarters near Atlanta.TECHNICAL SUMMARYThe central role of grain boundaries has long been recognized in dictating the mechanical behavior and failure susceptibility of metals and alloys. However, efforts to understand how variations in grain boundary characteristics affect material properties have been hampered by an incomplete understanding of what determines the strength of individual grain boundaries. The purpose of this project is to determine the characteristics that dictate grain boundary strength, here defined as the barrier strength that grain boundaries pose to dislocation propagation. A new adaptive remeshing electron backscatter diffraction (AR-EBSD)-based approach will be developed, combining in-line processing and automated adaptive grid remeshing to rapidly sample the tens of thousands of grain boundaries needed to build a library to which machine learning approaches can be applied. This approach will be coupled with transmission electron microscopy (TEM) characterization and atomistic simulations to correlate grain boundary strength with dislocation transfer mechanisms. This coupled approach will facilitate an unprecedented exploration of grain boundary space in terms of the number of grain boundaries investigated, allowing rigorous grain boundary strength functions to be established. In addition, the multiscale electron microscopy techniques developed over the course of the proposed work will be a widely applicable addition to the materials characterization toolbox in investigating material deformation under ambient and extreme conditions. Furthermore, a pipeline for underrepresented minorities to engage in STEM research will be created by a “visualizing science” summer internship program for high school students from underrepresented communities. These interns will work with graduate students supported by this program to investigate the ductile fracture behavior of metals, learn mechanical testing and characterization techniques, and visit the Novelis research center, a global Al company with research headquarters near Atlanta. To enhance the broader impact of this program, the interns will also develop lesson plans incorporating virtual reality elements to take back to their high school classes.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.