Emerging wurtzite ferroelectric materials have attracted significant interest due to their high Spontaneous polarization magnitude ( P_s). However, there is a limited understanding of the key factors that influence P_s. Herein, machine-learning regression models were developed to predict the P_susing datasets comprising 40 binary and 89 simple ternary wurtzite materials. Features were extracted based on elemental properties, crystal parameters and electronic properties. Feature selection was carried out using the Boruta algorithm and distance correlation analysis, resulting in a comprehensive machine learning model. Furthermore, SHapley Additive exPlanations analysis identified the average cation-ion potential ( IPi_ Aave) and the lattice parameter ( a) as significant determinants of P_s, with IPi_ Aavehaving the most prominent effect. A lower IPi_ Aavecorresponds to a lower P_sin the materials. Additionally, aexhibit an approximately negative correlation with P_s.
This multifactorial analysis fills the existing void in understanding the determinants of P_s, making a foundational contribution to the evaluation of emerging wurtzite materials and expediting the discovery of high-performance ferroelectric materials.
The dataset in this article can be accessed on the Scientific Data Bank https://www.doi.org/10.57760/sciencedb.j00213.00073. Please access the dataset of this article through a private link during the review stage https://www.scidb.cn/s/mAVvym).