[19p-B12-1~10] The present and future of crystalline Si solar cells

[19p-B12-1] [Highlight] Toward development of high-performance silicon ingot by multicrystalline informatics

〇Noritaka Usami¹, Kentaro Kutsukake², Abderahmane Boucetta¹, Takuto Kojima³, Tetsuya Matsumoto³, Hiroaki Kudo³, Yusuke Noda¹, Tatsuya Yokoi¹, Yasuo Shimizu⁴, Yutaka Ohno⁴ (1.Grad. Eng. Nagoya Univ., 2.RIKEN AIP, 3.Grad. Info. Nagoya Univ., 4.IMR Tohoku Univ.)

Keywords:multicrystalline silicon, machine learning, image processing

We attempt to pioneer “multicrystalline informatics” to establish general grain boundary physics through collaboration of data collection from large quantities of practical multicrystalline silicon wafers, machine learning and computation. In this presentation, we will report on methodology to extract various numerical data from 3 dimensional structures of crystal defects, predication of crystal orientations and grain boundary properties by machine learning, and so on.