[16p-D63-1~17] CS.9 Code-sharing Session of 6.5 & 7.5

2:30 PM - 2:45 PM JST(5:30 AM - 5:45 AM UTC)

[16p-D63-6] High-precision extraction of multilayer stack thin film structure parameters using Noise2Noise denoising for 4D-XPS Spectrum Big data

〇Satoshi Toyoda¹, Masashi Yoshimura², Sumida Hirosuke³, Mineoi Susumu³, Yoshigoe Akitaka⁴, Suzuki Satoru⁵, Kazushi Yokoyama⁵, Minowa Takuya⁶, Ogura Atsushi^6,7, Machida Masatake¹ (1.Scienta Omicron, 2.SPring-8 Service, 3.Mazda, 4.JAEA, 5.Univ. of Hyogo, 6.Meiji Univ., 7.MREL)

Keywords:

semiconductor,X-ray photoelectron spectroscopy,surface analysis

We are developing a method to visualize the chemical reactions and functional properties at the interfaces of multilayer thin films with high resolution using X-ray photoelectron spectroscopy (XPS). This method involves automatic analysis of 4D-XPS spectral big data to generate high-resolution 4K-equivalent videos. To accelerate the analysis, we utilize GPGPU and address the instability caused by increased Poisson noise using Noise2Noise deep learning technology. This presentation focuses on the precise extraction of film thickness and elemental concentration in multilayer thin film structures.