Three-dimensional structural measurement and material identification of an all-solid-state lithium-ion battery by X-Ray nanotomography and deep learning
M. Kodama , A. Ohashi , H. Adachi , T. Miyuki , A. Takeuchi , M. Yasutake , K. Uesugi , T. Kaburagi , S. Hirai
{"title":"Three-dimensional structural measurement and material identification of an all-solid-state lithium-ion battery by X-Ray nanotomography and deep learning","authors":"M. Kodama , A. Ohashi , H. Adachi , T. Miyuki , A. Takeuchi , M. Yasutake , K. Uesugi , T. Kaburagi , S. Hirai","doi":"10.1016/j.powera.2021.100048","DOIUrl":null,"url":null,"abstract":"<div><p>Three-dimensional measuring method of the material distribution of an all-solid-state lithium-ion battery (ASSLiB) cathode, by synchrotron radiation high-resolution X-ray computational tomography (nanotomography, nano-CT) and deep learning is proposed in this study. The cathode of the ASSLiB comprised materials with high X-ray absorption coefficients, such as LiCoO<sub>2</sub> and LiNi<sub>0.5</sub>Co<sub>0.2</sub>Mn<sub>0.3</sub>O<sub>2</sub>. Such high absorption coefficients imparted difficulties in obtaining a high-resolution, high-contrast image and in identifying materials with conventional CT value method. The method proposed in this study was effective in acquiring a high-resolution image with fewer artifacts and measured the heavy materials at a high signal-to-noise ratio. We used deep learning with a customized U-net, enabling high accuracy and ultra-high-speed material identification. Using this method, constituent materials were successfully identified in three dimensions. This material identification technique showed great potential for application to other techniques such as focused ion beam–scanning electron microscopy.</p></div>","PeriodicalId":34318,"journal":{"name":"Journal of Power Sources Advances","volume":null,"pages":null},"PeriodicalIF":5.4000,"publicationDate":"2021-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1016/j.powera.2021.100048","citationCount":"13","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Power Sources Advances","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2666248521000032","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"CHEMISTRY, PHYSICAL","Score":null,"Total":0}
引用次数: 13
Abstract
Three-dimensional measuring method of the material distribution of an all-solid-state lithium-ion battery (ASSLiB) cathode, by synchrotron radiation high-resolution X-ray computational tomography (nanotomography, nano-CT) and deep learning is proposed in this study. The cathode of the ASSLiB comprised materials with high X-ray absorption coefficients, such as LiCoO2 and LiNi0.5Co0.2Mn0.3O2. Such high absorption coefficients imparted difficulties in obtaining a high-resolution, high-contrast image and in identifying materials with conventional CT value method. The method proposed in this study was effective in acquiring a high-resolution image with fewer artifacts and measured the heavy materials at a high signal-to-noise ratio. We used deep learning with a customized U-net, enabling high accuracy and ultra-high-speed material identification. Using this method, constituent materials were successfully identified in three dimensions. This material identification technique showed great potential for application to other techniques such as focused ion beam–scanning electron microscopy.