[8a-A13-1~13] 6.3 Oxide electronics

9:15 AM - 9:30 AM JST(12:15 AM - 12:30 AM UTC)

[8a-A13-2] Crystal Structure Exploration for Quasi-Two-Dimensional Oxides Using the Universal Machine Learning Potential

〇Yuta Aoki¹ (1.Matlantis Corp.)

Keywords:

oxide,two-dimensional system,machine learning potential

While quasi-two-dimensional (2D) transition metal oxides hold potential for novel properties, conventional synthesis methods are strictly limited to materials with layered parent structures. In this study, we propose a novel structure search method for quasi-2D crystal structures using TiO₂ as a model case. By thinning suboxide series Ti_nO_2n-1 down to the few-atomic-layer level, we can control the film thickness while maintaining the TiO₂ composition. Structural optimization and phonon dispersion calculations performed using a universal machine-learning potential (PFP) revealed several stable structures with no imaginary vibration modes. This approach serves not only as a crystal structure search method but also provides insights into actual synthesis processes, and is widely applicable to any oxide system possessing a similar series of suboxides.