[17p-S2_204-1~18] 23.1 Joint Session N "Informatics"

6:15 PM - 6:30 PM JST(9:15 AM - 9:30 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[17p-S2_204-18] Topological Data Analysis Applied to Morphological Classification of Nanomaterials using FE-SEM Images

〇(B)Nodoka Takatsuka¹, Yukiko Hirose¹, Emi Minamitani², Kazuto Akagi³, Tohru Sugahara^1,2 (1.Kyoto Inst. Tech., 2.SANKEN, 3.Tohoku Univ.)

Keywords:

nanomaterial,Topological Data Analysis,surface morphology

Surface morphology governs the performance of nanomaterials, yet quantifying complex structures like entangled nanorods remains challenging. Addressing limitations in conventional metrics and black-box deep learning, this study proposes an objective evaluation method using Topological Data Analysis (TDA). MoO_x nanorod films were prepared using metal-organic decomposition. FE-SEM images were analyzed using persistent homology to track the "birth" and "death" of 0-dimensional (H₀, gaps) and 1-dimensional (H₁, loops) features. The resulting persistence diagrams were vectorized for Principal Component Analysis (PCA). The PCA score plot successfully classified surface morphologies into distinct groups, with the first principal component capturing variations in nanorod density and length. The average persistence diagrams quantitatively linked these physical traits to specific topological signatures. This method offers a robust, interpretable framework for evaluating nanomaterials without massive training datasets.