发表机构
Center for Nanophase Materials Sciences, Oak Ridge National Laboratory（橡树岭国家实验室纳米材料科学中心）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文开发了用于模拟电子束轨迹的通用电子显微镜数字孪生模型，可校准透镜、预测参数，为自动化显微镜及AI/ML集成提供物理信息工具。
AI 中文摘要
透射电子显微镜中专用成像与光谱状态的开发是新型应用（如柔性动量分辨高能量分辨率光谱）的需求，该任务因需对准和配置多种不同透镜、且各类成像与衍射平面位置存在歧义（通常未被充分记录）而变得复杂。本文开发了一种通用数字孪生模型，用于模拟电子束在整个电子显微镜中的轨迹。我们的代码包包含用于模拟镜筒的简单工具，并提供多种调整精确透镜位置和校准透镜模型的流程，使该模型可作为预测工具，让用户能快速便捷地确定设置新聚光镜或投影镜模式所需的正确透镜值。可采用自动流程更改透镜设置并测量其产生的变化以形成闭环。随着机器学习与人工智能工具的快速发展，我们认为该基于物理信息的显微镜模型可成为自动化显微镜及AI/ML集成的重要工具。
英文摘要
Novel experiments in transmission electron microscopy often require the development of specialized imaging and spectroscopy states, such as for flexible momentum-resolved high energy-resolution spectroscopy. This task is complicated by the need to align and configure many different lenses, and by ambiguity around the locations of various imaging and diffraction planes. Here we develop a versatile digital twin for simulating the electron beam trajectory throughout an electron microscope, and we develop the calibration procedures required for accurate prediction of microscope states. This allows the user to quickly and easily determine the correct lens values for setting up new condenser or projector modes. Automatic procedures to change lens settings and measure the resulting changes can be used to close the loop. With the accelerating development of machine learning and artificial intelligence tools, we also believe a physically-informed model of the microscope can serve as a valuable tool for automated microscopy and AI/ML integration.
