9:30 AM - 9:45 AM
△ [22a-M206-3] Prediction of dielectric properties of doped oxides using first-principles calculation and machine learning
Keywords:dielectric material, materials informatics, first-principles calculation
Rutile-type TiO2 co-doped with Nb and In has attracted attention as it shows a huge permittivity. In this study, the permittivity of the co-doped model structures based on the rutile oxides was calculated by the first-principles calculation, and a prediction model was constructed by machine learning. Some co-doped models showed a dielectric constant exceeding that of the base material, suggesting the possibility of new dielectric materials. It was found that the current machine learning model needs improvement.