Ensemble Machine‐Learning‐Based Analysis for In Situ Electron Diffraction. Issue 4 (9th February 2022)

• Record Type: Journal Article
• Authors: Ge, Mengshu; Liu, Xiaozhi; Zhao, Zhicheng; Su, Fei; Gu, Lin; Su, Dong
• Is Part Of: Advanced theory and simulations. Volume 5:Issue 4(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-02-09
• Subjects: electron diffraction -- in situ TEM -- machine learning -- structural analysis; Science -- Simulation methods -- Periodicals; Science -- Methodology -- Periodicals; Engineering -- Simulation methods -- Periodicals; Engineering -- Methodology -- Periodicals; 507.21
• DOI: 10.1002/adts.202100337
• Languages: English
• ISSNs: 2513-0390

Abstract

Abstract: In situ transmission electron microscopy is an important characterization approach for exploring the structural dynamics of materials. However, the recorded high resolution in situ videos normally have tremendous amount of data, which is challenging for quantitative analysis. In case of in situ electron diffraction (ED), the classical analysis method only tracks changes of the integral profile and ignores important information of position, intensity, and distribution angle due to the lack of a proper data processing tool. In this work, an ensemble machine‐learning‐based framework which enables the fore‐and‐after tracking of all diffraction spots of an in situ ED video is established. As demonstrated in the case of the lithiation of the Co3 O4 nanoparticles, the method precisely quantifies the changes of lattice parameters and then unveils the inhomogeneous structural evolution induced by the insertion of lithium ions. This method is generally applicable to analyze in situ ED data derived from any dynamical processes, enlarging the capability to grasp the key information from massive in situ data. Abstract : In situ electron diffraction is an important characterization method for exploring the structural dynamics of materials. This work introduces an ensemble machine‐learning‐based framework which enables tracking the movement of all diffraction spots from in situ electron diffraction. This method can be generally applied to analyze in situ ED data derived from any dynamical processes.