Physics-informed unsupervised machine learning for 4D-STEM materials characterization

Physics-informed unsupervised machine learning for 4D-STEM materials characterization | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Physics-informed unsupervised machine learning for 4D-STEM materials characterization Koji Kimoto, Ovidiu Cretu, Koji Harano, Fumihiko Uesugi, Jun Kikkawa, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8846270/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Four-dimensional scanning transmission electron microscopy (4D-STEM) provides bimodal real- and reciprocal-space information and forms the basis of various STEM imaging techniques, including standard bright-field and annular dark-field imaging, ptychography, and differential phase contrast. Advancements in detection systems have rendered 4D-STEM increasingly practical, while also enabling the acquisition of large datasets. Here, we propose a comprehensive workflow that integrates 4D-STEM with unsupervised machine learning while incorporating electron microscopy knowledge. The workflow comprises three key substeps: data preprocessing, dimensionality reduction, and clustering. We detail how physical insight from electron microscopy is embedded into each substep. In the data preprocessing substep, feature extraction is performed such as noncentrosymmetry based on diffraction physics. In the dimensionality reduction substep, we develop a constrained nonnegative matrix factorization that incorporates physics-based constraints for diffraction and corresponding real-space map. In the clustering substep, diffraction similarity is defined in place of conventional cosine similarity, fully exploiting the bimodal nature of 4D-STEM data. The effectiveness of the proposed workflow is demonstrated using both simulated datasets and experimental data from two representative specimens: twist domains of monolayer MoS 2 and small crystalline precipitates in a metallic glass ZrCuAl. These results demonstrate that the proposed workflow enables advanced characterization of unknown materials that are difficult to achieve using conventional methods, highlighting the potential of physics-informed unsupervised machine learning combined with 4D-STEM. Physical sciences/Materials science Physical sciences/Optics and photonics unsupervised machine learning scanning transmission electron microscopy (STEM) nonnegative matrix factorization (NMF) principal component analysis (PCA) clustering hyperspectral imaging Full Text Additional Declarations No competing interests reported. Supplementary Files SupplementaryInformation.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8846270","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":594585687,"identity":"f19f8e3e-ec24-48fb-83a2-825b3091b2cf","order_by":0,"name":"Koji Kimoto","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7ElEQVRIie3QPQrCMBTA8VcKXVSyvlDwDJGCIAi9SnsDwaXgUinERXB19wQiiGNKB5eIB+iiCM6Kq4MptTr1YxTMn0BCyG94AdDpfjA0850BxNNTtuXXdUTkJGLNCHwIWFUPv9FZy7/cAnCIkfDJaLR1gawEDHblxDY7GyYk9GkY83TJUj/EqwdUlpOu2dlizGHIhCJtlnqAkgHlDclYEbeW2G/Sz4ipiBGSeTWhkZpFSnTULJGtiM/R8kTVLHg8rE9BMOwtcH9+tJ+pS0gSX2jFjxVSLZEfLfQgoWEtUZHiFRFg3BsRnU6n+5NeteZPqEVSmogAAAAASUVORK5CYII=","orcid":"","institution":"National Institute for Materials Science","correspondingAuthor":true,"prefix":"","firstName":"Koji","middleName":"","lastName":"Kimoto","suffix":""},{"id":594585688,"identity":"5c5f84db-fde8-4063-bf38-fbb37adc5d9c","order_by":1,"name":"Ovidiu Cretu","email":"","orcid":"","institution":"National Institute for Materials Science","correspondingAuthor":false,"prefix":"","firstName":"Ovidiu","middleName":"","lastName":"Cretu","suffix":""},{"id":594585689,"identity":"90adde83-da8f-4897-8baa-61c7fd0de239","order_by":2,"name":"Koji Harano","email":"","orcid":"","institution":"National Institute for Materials Science","correspondingAuthor":false,"prefix":"","firstName":"Koji","middleName":"","lastName":"Harano","suffix":""},{"id":594585690,"identity":"b1daef54-f343-455f-ba2a-5da7b4d5a902","order_by":3,"name":"Fumihiko Uesugi","email":"","orcid":"","institution":"National Institute for Materials Science","correspondingAuthor":false,"prefix":"","firstName":"Fumihiko","middleName":"","lastName":"Uesugi","suffix":""},{"id":594585694,"identity":"3e761cf0-91ef-4c34-8ea9-95e4d9dd4f79","order_by":4,"name":"Jun Kikkawa","email":"","orcid":"","institution":"National Institute for Materials Science","correspondingAuthor":false,"prefix":"","firstName":"Jun","middleName":"","lastName":"Kikkawa","suffix":""},{"id":594585696,"identity":"123171e2-1a7e-423d-986f-589d20869eb2","order_by":5,"name":"Kenji Nagata","email":"","orcid":"","institution":"National Institute for Materials Science","correspondingAuthor":false,"prefix":"","firstName":"Kenji","middleName":"","lastName":"Nagata","suffix":""},{"id":594585698,"identity":"ef57cc71-f27c-4eb8-8dd0-531cf5e1e3d4","order_by":6,"name":"Yuki Shibazaki","email":"","orcid":"","institution":"High Energy Accelerator Research Organization","correspondingAuthor":false,"prefix":"","firstName":"Yuki","middleName":"","lastName":"Shibazaki","suffix":""},{"id":594585700,"identity":"d9c8c3f4-b61f-4bcd-801c-e090fcdb2c53","order_by":7,"name":"Takashi Matsumoto","email":"","orcid":"","institution":"Tokyo Electron Technology Solutions Ltd","correspondingAuthor":false,"prefix":"","firstName":"Takashi","middleName":"","lastName":"Matsumoto","suffix":""},{"id":594585702,"identity":"9eb553bd-b0fc-4241-84ba-e052f5db92dc","order_by":8,"name":"Yoshiki Sakuma","email":"","orcid":"","institution":"National Institute for Materials Science","correspondingAuthor":false,"prefix":"","firstName":"Yoshiki","middleName":"","lastName":"Sakuma","suffix":""}],"badges":[],"createdAt":"2026-02-11 02:24:31","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8846270/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8846270/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106031703,"identity":"4f99d468-5e8f-4fd5-b4e8-ce60d20b8c18","added_by":"auto","created_at":"2026-04-02 15:27:15","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1832056,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8846270/v1_covered_9413eed1-05f6-488b-a3b7-2473c33769d7.pdf"},{"id":103400123,"identity":"207a3ba9-8787-483e-8d97-e00a44773dea","added_by":"auto","created_at":"2026-02-25 09:15:20","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":32943,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryInformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-8846270/v1/52a210bc201db658ffccc541.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Physics-informed unsupervised machine learning for 4D-STEM materials characterization","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"unsupervised machine learning, scanning transmission electron microscopy (STEM), nonnegative matrix factorization (NMF), principal component analysis (PCA), clustering, hyperspectral imaging","lastPublishedDoi":"10.21203/rs.3.rs-8846270/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8846270/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eFour-dimensional scanning transmission electron microscopy (4D-STEM) provides bimodal real- and reciprocal-space information and forms the basis of various STEM imaging techniques, including standard bright-field and annular dark-field imaging, ptychography, and differential phase contrast. Advancements in detection systems have rendered 4D-STEM increasingly practical, while also enabling the acquisition of large datasets. Here, we propose a comprehensive workflow that integrates 4D-STEM with unsupervised machine learning while incorporating electron microscopy knowledge. The workflow comprises three key substeps: data preprocessing, dimensionality reduction, and clustering. We detail how physical insight from electron microscopy is embedded into each substep. In the data preprocessing substep, feature extraction is performed such as noncentrosymmetry based on diffraction physics. In the dimensionality reduction substep, we develop a constrained nonnegative matrix factorization that incorporates physics-based constraints for diffraction and corresponding real-space map. In the clustering substep, diffraction similarity is defined in place of conventional cosine similarity, fully exploiting the bimodal nature of 4D-STEM data. The effectiveness of the proposed workflow is demonstrated using both simulated datasets and experimental data from two representative specimens: twist domains of monolayer MoS\u003csub\u003e2\u003c/sub\u003e and small crystalline precipitates in a metallic glass ZrCuAl. These results demonstrate that the proposed workflow enables advanced characterization of unknown materials that are difficult to achieve using conventional methods, highlighting the potential of physics-informed unsupervised machine learning combined with 4D-STEM.\u003c/p\u003e","manuscriptTitle":"Physics-informed unsupervised machine learning for 4D-STEM materials characterization","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-25 09:15:15","doi":"10.21203/rs.3.rs-8846270/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"5b4db49f-c17c-463e-9d69-89074a2b580b","owner":[],"postedDate":"February 25th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":63276691,"name":"Physical sciences/Materials science"},{"id":63276692,"name":"Physical sciences/Optics and photonics"}],"tags":[],"updatedAt":"2026-04-02T15:26:36+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-25 09:15:15","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8846270","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8846270","identity":"rs-8846270","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}