Crystal Structure Classification Using Machine Learning

X-ray diffraction refers to the phenomenon in which when incident x-ray is irradiated upon a matter, strong x-ray is directed in several particular directions different from the incident x-ray. The intensity and direction of x-ray diffraction differs based on the type and structure of atoms forming the matter, so studying the x-ray diffraction using these characteristics can yield to knowledge of the microstructure of matter. Today, x-ray diffraction is widely used in research on matter apart from basic science, as well as industrial fields such as the semiconductor and metal industries.

Our development of crystal structure automatic classification has the purpose of automatically analyzing the crystal structure of arbitrary matter using x-ray diffraction patterns. Using ICSD (Inorganic Crystal Structure Database), a large database containing information on the crystal structure of around 180,000 inorganic compounds, we are in the process of developing an automatic crystal structure analyzer by obtaining the x-ray diffraction pattern of each matter and inputting this information through machine learning.