Machine Learning Based Distinguishing Between Ferroelectric and Non‐Ferroelectric Polarization–Electric Field Hysteresis Loops
Qicheng Huang,Zhen Fan,Lanqing Hong,Shengliang Cheng,Zhengwei Tan,Guo Tian,Deyang Chen,Zhipeng Hou,Minghui Qin,Min Zeng,Xubing Lu,Guofu Zhou,Xingsen Gao,Jun-Ming Liu
DOI: https://doi.org/10.1002/adts.202000106
2020-01-01
Advanced Theory and Simulations
Abstract:The polarization-electric field (P-E) hysteresis loop is one of the most important criteria for identifying ferroelectricity. However, aP-Eloop with apparent hysteresis window can be generated from non-ferroelectric sources such as leakage current. So far distinguishing between ferroelectric and non-ferroelectric loops is still performed in a manual way, which can be error prone and time consuming, particularly when the loops are not easily distinguishable and the number of loops to be identified is large. Here, two machine learning (ML) approaches are developed, one using the polarization values along theP-Eloops as the input dataset (termed as "value-based" approach) and the other using the loop images as the input dataset (termed as "image-based" approach), to identify theP-Eloops as ferroelectric or non-ferroelectric. The value- and image-based ML approaches achieve identification accuracies as high as 93.08% and 87.42%, respectively. In addition, it is tested that both approaches complete an identification of about 160 loops in very short time (approximate to 1.0 s). The high accuracy and efficiency therefore demonstrate that the ML approaches significantly outperform the manual way for distinguishing ferroelectric from non-ferroelectricP-Eloops, which may greatly facilitate the research on ferroelectrics.