9月の学内共同研究促進トークは、情報機能素子科学研究室のBermundo Juan Paolo Soria先生にご講演いただきました。
Asst. Prof. Bermundo Juan Paolo Soria (Information Device Science Laboratory)
TITLE: Machine-learned Fermi level prediction of solution processed ultrawide bandgap amorphous oxide semiconductors
Ultrawide bandgap amorphous oxide semiconductors, in particular amorphous gallium oxide (a-Ga2Ox), have interesting properties such as transparency (ultrawide bandgap), excellent uniformity (amorphous), and high breakdown voltage which makes them promising materials for thin film transistors, sensors, and solar-blind photodetectors. Nevertheless, the combination of its low carrier concentration, ultrawide bandgap, and high defect density especially when deposited using solution process render it insulating. Thus, controlling the electron carrier concentration, represented by the Fermi level (EF) position, is necessary. Traditionally, experimental methods that are complex, time consuming, and resource intensive need to be performed to determine and optimize EF. Here, I’ll talk about how a supervised machine learning approach can be utilized as a rapid and cost-effective alternative to effectively predict the EF of a-Ga2Ox.