Abstract
Ferroelectric materials in various applications can be ascribed to its functional properties, which originates from the spontaneous polarization forming the ferroelectric domain configuration. Hence, it is of great importance to characterize the complex polarization state to interpret the associated phenomena in ferroelectric materials. In particular, recent progress on highperformance materials with exotic nanoscale polarization heterogeneity requires advanced technology on probing local structure for ferroelectrics. In this work, we employed a high-throughput scanning convergent beam electron diffraction (SCBED) method to investigate the local structure of (Ba 0.78 Ca 0.22 )(Ti 0.88 Sn 0.12 )O 3 ceramics by using a transmission electron microscope. The obtained patterns were further analyzed by the Principal Components Analysis (PCA) and Bayesian machine-learning algorithm to recognize polar axes and their spatial distribution. The results show the coexistence of cubic, tetragonal, orthorhombic and rhombohedral crystal symmetries among nanodomains, which suggest a reduced polarization anisotropy. Our work may provide an effective approach to characterize the complex polarization state in ferroelectric materials, which enables the establishment of structure-property relationship for high-property functional dielectric materials.