[17a-PB02-1~9] 23.1 Joint Session N "Informatics"

[17a-PB02-5] Machine learning assisted high-throughput analysis of Raman imaging spectra and application for thickness identification of graphene sheets

〇(B)Riku Gotoh^1,2, Asako Yoshinari^1,2, Seiya Suzuki^2,3,4, Yasunobu Ando⁵, Tarojiro Matsumura⁵, Masato Kotsugi¹, Naoka Nagamura^1,2,4 (1.Tokyo Univ. of Science, 2.NIMS, 3.JAEA, 4.JST PRESTO, 5.AIST)

Keywords:graphene

Raman microscopy is a useful analysis tool for graphene, whose properties are affected by stress and the number of layers. However, it has been practically difficult to perform peak fitting and extract spectral information manually from spatial resolved spectral big datasets. Therefore, we have developed a high-throughput analysis method for spectral big datasets using an automatic peak fitting package, "EM Peaks", based on machine learning. We adopted this method to identify the spatial distribution of the layer number of CVD graphene.