[7a-N321-1~8] 3.8 Terahertz technologies

9:30 AM - 9:45 AM JST(12:30 AM - 12:45 AM UTC)

[7a-N321-1] Application of Machine Learning to the Time Waveform of a Transmitted THz Wave
through a Single-layer Graphene

〇Limin Zhao¹, Suguru Yamauchi², Shin-ichiro Yanagiya^2,3, Yasuhide Ohno², Masao Nagase², Yasuo Minami¹ (1.Nihon Univ., 2.Tokushima Univ., 3.pLED Inst., Tokushima Univ.)

Keywords:

Machine Learning,Terahertz Time Domain Spectroscopy (THz-TDS),Principal Component Analysis(PCA)

We used THz time-domain spectroscopy (THz-TDS), which has attracted much attention in recent years, to perform imaging and machine learning classification on single-layer graphene prepared by SiC pyrolysis. 15 x 15 mm² areas were measured at 1 mm intervals with THz-TDS. Transmittance images were obtained from the power spectra. The obtained time waveforms were normalized and classified into Air, Sample, and Holder using CNN and trained with decision trees (1000 trees, 200 depth) and multiple learning rates, with the best classification results obtained when the learning rate was 0.0001.