A deep-learning technique for phase identification in multiphase inorganic compounds using synthetic XRD powder patterns

Abstract

Identifying the composition of multiphase inorganic compounds from XRD patterns is challenging. Here the authors use a convolutional neural network to identify phases in unknown multiphase mixed inorganic powder samples with an accuracy of nearly 90%.