[15p-P06-1~15] 23 Joint Session N "Informatics" (Poster)

[15p-P06-13] Improvement of Efficiency of Dielectric Thin Film Material Development by 2D Measurement Data and Study of Linkage with Existing Data

〇Takahiro Nagata¹, Hiori Kino¹, Takashi Kono¹, Noriaki Ozaki², Jun Ikeda² (1.NIMS, 2.Murata co. ltd.)

Keywords:

2D-XRD,Variational AutoEncoder,dielectric thin film

We have proposed an automatic analysis method for 2-D X-ray diffraction data, which is an important method for understanding physical properties in the development of thin film materials, using a non-negative matrix factorization to classify the data and a variational autoencoder to visualize the relationships between them. We applied this method to high-dielectric thin film materials and compared the results with material maps created from the NIMS database AtomWork-Adv. As a result, we found that the metastable structure of the thin film shows a different trend in contrast to the region where the structure and dielectric constant trends are consistent.