Reconstruction of Incomplete X-Ray Diffraction Pole Figures Using Deep Learning

Abstract

Abstract X-ray diffraction crystallography allows non-destructive examination of crystal structures. Furthermore, it has low requirements regarding the surface preparation, especially compared to electron backscatter diffraction. However, up to now, X-ray diffraction is highly time-consuming in standard laboratory conditions since we have to record intensities on multiple lattice planes by rotating and tilting the sample. In this article, we propose a method based on deep learning that allows faster experimentation due to accurate reconstructions of pole figure regions, which we did not probe experimentally. To speed up the development of our proposed method and further machine learning algorithms, we introduce a GPU-based simulation for data generation. Furthermore, we present a pole widths standardization technique using a custom deep learning architecture that makes algorithms more robust against influences from the experiment setup and material.