Machine Learning-Assisted Design of CaBi2Nb2O9-based Ceramics with High Piezoelectricity and Thermal Stability

Ailin Li,Chengwen Bin,Xu Hou,Luocheng Liao,Yunya Liu,Bo Wang,Jie Wang

DOI: https://doi.org/10.1016/j.ceramint.2024.11.449

IF: 5.532

2024-01-01

Ceramics International

Abstract:Calcium bismuth niobate CaBi2Nb2O9 (CBN)-based ceramics are promising candidates for high-temperature piezoelectric applications due to its high Curie temperature. However, the traditional trial-and-error method cannot realize fast design of high-temperature piezoceramics over a wide range of composition space. Based on small sample datasets, a machine learning model was proposed to accelerate the design of CBN-based ceramics with high piezoelectricity and thermal stability. Using the trained gradient boosting regression model, the piezoelectric coefficients are predicted for CBN-based ceramics with different compositions. Based on the machine learning prediction results, we prepared the co-doped Ca1-2x(NaCe)xBi2Nb2-x(W0.75Mn0.25)xO9 (CBNCWM-x) by solid phase reaction method. The results show that co-doping changes the lattice distortion, grain size and domain structures of CBN-based ceramics, realizing the enhanced piezoelectric properties. CBNCWM-0.03 ceramics exhibit an optimum piezoelectric coefficient (d33 ∼ 17.2 pC/N) that is quite close to the predicted result by machine learning model. Moreover, the CBNCWM-0.03 ceramics possess a Curie temperature of 913 °C and maintain 90.7% of its original d33 value after annealing at 900 °C for two hours, demonstrating excellent thermal stability. The present work not only provides an approach for the rapid discovery of CBN-based ceramics with high piezoelectricity but also obtains excellent thermally stable ceramics for high-temperature applications.