Options
2017
Conference Paper
Titel
Identification of defect-suppressing grain boundaries in multicrystalline silicon based on measurements of as-cut wafers using advanced image processing
Abstract
The mobility and multiplication of dislocation clusters in cast Silicon depend on many factors, especially grain competition and grain boundary types. The great success of High-Performance (HP) mc-Si material in the last years is mainly based on a high occurrence of stress-absorbing grain boundaries. Where the growth of a dislocation cluster is confined by a grain boundary, the cluster displays distinct margins. Therefore cluster and grain show a mutual boundary and a corresponding vertical development. We present an image processing tool that extracts those regions with corresponding and possibly correlated development of dislocation clusters and grain structure in multicrystalline silicon (mc-Si), based on photoluminescence (PL) and optical measurements on wafers with a vertical distance in the brick of 4-8 mm. The vertical development of grain structure and PL signal is reconstructed layer by layer through the complete brick height and thus allows an insight into the whole crystal development of the whole brick and a spatially resolved macroscopic 3D-analysis on industrial scale. Grain orientations and grain boundary types were investigated within regions with corresponding development in and around dislocation clusters. Exemplary results quantitatively support the general assumption that among the dislocation-suppressing grain boundaries, random grain boundaries have the largest share with 42%.