题目：A Deep Learning Perspective of Atomic Defects and Single Atoms Modulation in Scanning Transmission Electron Microscopy

报告人：赵晓续（Xiaoxu Zhao）

时间：4月3日（周五）16：00

地点：李薰楼249

报告摘要：

Scanning transmission electron microscopy (STEM) are powerful tools to trigger atomic-scale motions, pattern atomic defects, and lead to anomalous quantum phenomena in functional materials. Conventional ways of quantitative data analysis suffer from low yield and poor accuracy. New ideas in the field of deep learning have provided more momentum to harness the wealth of big data and sophisticated information in STEM data analytics, automatic identifying weak signals and complex features in atomic resolution microcopy images with intelligence. In our work, employing cycle generative adversarial networks (CycleGAN) and U-Nets, we propose a method based on a single experimental image to tackle high annotation costs and image noise for defect detection in a library of 2D materials. In addition, we propose an electron beam (e-beam) triggered chemical etching approach to engineer shielded metal atoms sandwiched between chalcogen layers in monolayer transition metal dichalcogenide (TMDC). Various metal vacancies have been fabricated via atomically focused e-beam in STEM. In situ sequential STEM imaging corroborated that a combined chemical-induced knock-on effect and chalcogen vacancy-assisted metal diffusion process result in atom-by-atom vacancy formation. The presence of metal vacancies strongly modified their magnetic and electronic properties, correlated with the unpaired chalcogen p and metal d electrons surrounding vacancies and adjacent distortions. These findings show a generic approach for engineering interior metal atoms with atomic precision, creating opportunities to exploit quantum phenomena at the atomic scale.

报告人简介：

赵晓续，北京大学材料科学与工程学院研究员，博士生导师，国家级青年人才，北京市杰青，国家重点研发计划青年首席科学家。赵晓续研究员于2014年获南洋理工大学一等荣誉学士学位，2018年获新加坡国立大学博士学位，2018-2020年担任新加坡国立大学博士后研究员，2020-2022年担任南洋理工大学校长博士后研究员，2022年加入北京大学。赵晓续研究员的主要研究兴趣是使用高空间以及高能量分辨的球差校正扫描透射电子显微镜/电子能量损失谱，在亚原子尺度解析和构筑低维量子材料，建立原子拓扑结构与物性之间的关联，并结合机器学习方法实现对原子尺度电镜照片的智能清洗、分类与分析等，目前以第一/通讯作者在Nature, Nat. Nanotechnol., Nat. Mater., Nat. Syn. Nat. Commun.,等期刊发表论文70余篇，引用超过17,000次，h-index 70，入选科睿唯安“高被引科学家”。赵晓续研究员于2018年获国家优秀自费留学生奖，2020年南洋理工大学Presidential Postdoctoral Fellowship，2021年入选国家级海外高层次青年人才项目，2022年入选《福布斯中国·青年海归菁英·100 人》，《麻省理工科技评论》2022 年度亚太区“35 岁以下科技创新 35 人”，等荣誉称号。主持国家重点研发计划青年科学家项目、国家重点研发计划政府间联合研发项目、国家自然科学基金面上、重大项目子课题、创新群体B骨干等多个项目。