Atomic-resolution three-dimensional imaging by field ion microscope with machine learning

2022 Fiscal Year Final Research Report

• Project/Area Number: 20K05325
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 29020:Thin film/surface and interfacial physical properties-related
• Research Institution: Mie University
• Principal Investigator: Nagai Shigekazu 三重大学, 工学研究科, 准教授 (40577970)
• Project Period (FY): 2020-04-01 – 2023-03-31
• Keywords: 電界イオン顕微鏡 / 機械学習 / 表面構造観察

Outline of Final Research Achievements

As a method for analyzing field ion microscope (FIM) images, we developed a system that automatically detects crystal planes using an object detection model and automatically identifies the crystal orientation of a sample using the k-nearest neighbor method. As a result, it was demonstrated that the crystal orientation of a tungsten sample can be identified with an accuracy of more than 80%. A system was constructed to extract the position of each bright spot observed in the FIM image, which reflects the atomic position, from the differential images taken continuously during field evaporation. In addition, we have implemented this system in an existing FIM system for observation, recording, and automatic extraction of atomic positions. We have found the feasibility of applying machine learning to FIM for atomic-resolution tomography microscopy.

Free Research Field

荷電粒子線工学

Academic Significance and Societal Importance of the Research Achievements

原子分解能を有する電界イオン顕微鏡は比較的簡便な構造で原子分解能像が得られるが，観察条件の選定および像解釈が困難である。本研究で実施したFIMへの最新のデジタル画像解析法と機械学習の適用により，結晶方位，結晶面，および原子位置を自動解析するシステムを構築することができた。この成果は，新たな原子分解能トモグラフィー観察手法としての可能性を示すものである。さらに本システムは簡便かつ安価であり，次世代デバイスの性能向上および特性評価への貢献が期待される。