P1-1249Applications of Homology Analysis to TEM/STEM imagesKey Words: TEM/STEM, Homology, NanostructuresMicro/nanostructuresofmaterialsinfluencetheirphysicalpropertiesandcharacteristics.Therefore,additionallytoexplorationofnewmaterialsandsynthesismethods,developmentsofmeasurementtechniquesarecrucialforcharacterizationofthemicro/nanostructures.Transmissionelectronmicroscopy(TEM)allowscharacterizationofnanostructureswithatomic-levelspatialresolutionandtimeresolutionontheorderofmicroseconds.⚫Extraction of descriptors with homology analysis⚫Betti number enables quantitative evaluation of relation between nanostructure and ionic conductivity⚫Persistent homology allows interpretable classification of nanostructures with effective descriptorsConclusionBetti numbersPersistent homology⚫Application of homology analysis to various structures such as crystalline and amorphous materials⚫Application of homology analysis to various images such as high-resolution TEM/STEM Future PlanIntroductionTheme underDiscussionEnvironment-Controlled Microscopy Group, GREEN Ayako HASHIMOTOE-mail：： HASHIMOTO.Ayako@nims.go.jpBetti number β0: number of connected componentsBetti number β1 : number of circles or loopsBetti number β2 : number of cavitiesHomology analysis: CHomPhttp://chomp.rutgers.edu/software.htmlProf. Mischaikow (The State University of New Jersey)Inσ0(Scm-1)E(eV)Bettinumber: β0CeO2 connectivityIonicconductivityY. Wen, A. Hashimoto, A.S.B.M. Najib, A. Hirata, and H. Abe, Appl. Phys. Lett., 118 (2021) 054102.QuantitativeevaluationofrelationbetweennanostructureandionicconductivityY. Wen, H. Abe, A. Hirata, and A. Hashimoto, ACS Appl. Nano Matt., 4 (2021) 13602.R: Pt G: Ce50nmADF-STEMW: PtB: CeO2STEM-EDS mapby NIMSH.Abe groupADF-STEM Metal-oxide composite :Pt#CeO21:12:13:1CO:O2= 0:1100 nmSample preparation: FIB（JIB-4000, JEOL）Observation: STEM (JEM-2100F, JEOL)N=1Birth of holeDeath of holerbrdFiltrationHomologyanalysis:HomCloudI.Obayashietal.,J.Phys.Soc.Jpn.91,091013(2022).https://homcloud.dev/index.en.htmlEguchi, Y. Wen, H. Abe and A. Hashimoto, Nanomaterials 14, 1413 (2024).Temperature, gas ratioE=1.24eVE=1.2eVE=0.74eVDisorder of bendsScale factorNumber of small arc (< 4 nm)Width [nm]Number of small arcsWidthEffective descriptorsPC1coefficientCharacteristic size↓Small arc↑Large arcGas ratio (CO:O2)(a)(b)(c)(d)(e)(f)(g)(h)(i)(j)(k)(l)Temperature (oC)0:11:12:13:1500600 700100 nm0-D hole1-D hole2-D holeTopologyBetti numberfor blackApplication to Pt#CeO2nanocompositesStructural change with gas ratioIonic conductivity: Impedance spectroscopyStructure-ionic conductivity relationshipFast trans.Slow trans.PtPtCeO2CeO2CeO2phase = fast pathway for oxygen ion transportation Persistent homologyTracking N-D hole’s appearanceand disappearanceunder continuous expansion, filtrationAppearanceDisappearanceBirth valueDeath value(rb,rd)Persistence diagram (PD)Application to Pt#CeO2nanocompositesClassification of nanostructuresMachine learningClassification with effective descriptorsGeometric+Topologicalinfo.Extraction of interpretable featureInterpretable classification of nanostructures with effective descriptorsPD: extraction of interpretable featuresIngeneral,TEM/scanningTEM(STEM)providegeometricfeaturesofnanostructuressuchaslength,area,andtheirdistribution.However,alternativefeaturesarerequiredtoevaluatetherelationbetweennanostructuresandpropertiesquantitatively.Inthisstudy,weappliedhomologyanalysistoSTEMimagesinordertoextractneweffectivedescriptorsrepresentingnanostructuresforquantitativeevaluation.
元のページ ../index.html#49