Computer-aided characterization of ferroelectric phase transitions and domain structures using polarizing light microscopy

Optical microscopy is an essential characterization technique that is used as an initial step in the study of domain structures and phase transitions of ferroelectric materials. Image analysis aided by advanced computer algorithms is urgently required to enable an improved accuracy and higher efficiency of data processing. In this study, an automatic full-angle light intensity detection algorithm is developed to study polarizing light microscopy images of domain structures and the phase evolution process is decoded using the scale invariant feature transform method. This data processing algorithm is then used to characterize the phase transition process in relaxor ferroelectric PMN-0.29PT single crystals with complex domain structures. An emerging in-plane tetragonal phase is tracked during the rhombohedral (R) to tetragonal (T) phase transition, which is associated with peaks present in the temperature-dependent dielectric curve. Compared to the traditional polarizing light microscopy characterization method, this computer-aided algorithm enables a quicker and more informative analysis of domain evolutions in ferroelectric single crystals.