[11p-E206-1~12] 13.2 Exploratory Materials, Physical Properties, Devices

4:30 PM - 4:45 PM JST(7:30 AM - 7:45 AM UTC)English Presentation

[11p-E206-12] A Unified High-κ Dielectric Screening Pipeline Based on a Machine Learning Hamiltonian

〇(B)Doyun Kim^1,2, Dongik Park², Juhan Hong^1,3, Jaewon Bae², Chanyoung Park^2,3 (1.Seoul Natl. Univ., 2.NanoForge AI, 3.KAIST)

Keywords:

High-k dielectrics,machine learning Hamiltonian,materials screening

High-κ gate dielectrics are essential for semiconductor device scaling, but candidate oxides must combine a high static dielectric constant with a wide band gap. We present a unified high-κ dielectric screening pipeline based on a machine learning Hamiltonian (MLH). The predicted Kohn-Sham Hamiltonian is used to obtain the band gap, electronic dielectric contribution, and ionic dielectric contribution in a unified manner. This enables internally consistent screening of oxide materials for high-κ applications. We validate the pipeline on dynamically stable oxides held out from training and then apply it to a broader oxide screening study.