Crystal-Orientation-Based Analysis on Ingot-Scale Structure of Multicrystalline Silicon

The introduction of defects into multicrystalline silicon during unidirectional solidification growth is expected to be caused by phenomena on various scales, such as microscopic atomic arrangements and macroscopic stress distributions. The crystal orientation of the multicrystalline structure is extensively involved in the occurrence of twinning and the stress distribution around grain boundaries at grain boundaries. In addition, the multicrystalline structure formed by unidirectional solidification grows columnarly from nuclei at the bottom of the crucible, and the interior of the columnar structure contains a variety of grain boundaries due to continuous twinning. While the orientation relationship between individual nuclei is random, the orientation relationship due to twinning is crystallographically determined and thus contains information on the grain generation relationship. Therefore, it is expected that the macroscopic crystal orientation distribution can be used to analyze the dynamics of crystal growth. Since obtaining ingot-scale data using conventional diffraction-based methods is difficult, we have developed a method to estimate crystal orientation with a machine learning model using optical images of the surface textured by anisotropic etching as input data. We have achieved the estimation model with an average estimation error of 3.5° (95th percentile &lt; 9°) trained on the crystal orientation distribution obtained by X-ray diffraction [1-3]. In addition, to extract macroscopic features of the multicrystalline structure, we have developed a method to analyze the crystal orientation distribution by attributing the orientation relationship due to twinning to a network graph [4-7]. We will report an ingot-scale analysis of the development of the polycrystalline structure during crystal growth using these methods.<br/><br/>References<br/>[1] H. Kato et al., IEICE Tech. Report 119 [454], 81 (2020).<br/>[2] T. Kojima et al., MRM 2021, A4-O3-05.<br/>[3] K. Hara et al., Abstr. 2021 MRS Fall Meeting, DS01.15.08.<br/>[4] T. Kojima et al., Abstr. 66th JSAP Spring Meeting, 9a-W611-3, 2019.<br/>[5] T. Kojima et al., Abstr. 80th JSAP Autumn Meeting, 20a-E314-4, 2019.<br/>[6] T. Kojima et al., Abstr. 67th JSAP Spring Meeting, 14a-A205-3, 2020.<br/>[7] T. Kojima et al., Abstr. 2021 MRS Fall Meeting, CH04.09.01.