Accelerated design of AgNbO3-based ceramics with high energy storage performance via machine learning

Abstract

Silver niobate-based lead-free antiferroelectric (AFE) ceramics exhibit tremendous potential in energy storage applications, but large-scale experimental investigations are always required to achieve high performance. In this work, we employed a machine learning method to accelerate the development of high-performance AgNbO₃-based AFE ceramics using the (Ag_{1−x−3y−4z−3m−3n−3t}Na_xLa_yCe_zNd_mSm_nTb_t)(Nb_1−sTa_s)O₃ component system as a representative, in which supporting vector regression with the radial basis function kernel model and lowest cross-validation error was assigned as the preferred machine learning model. Based on the predicted outcomes, the highest recoverable energy storage density of 7.0 J cm⁻³ was successfully achieved in the (Ag_0.94Sm_0.02)(Nb_0.6Ta_0.4)O₃ ceramic experimentally, which was close to its predicted value of 6.76 ± 0.55 J cm⁻³, indicating the good reliability of this machine learning technique. Moreover, the (Ag_0.94Sm_0.02)(Nb_0.6Ta_0.4)O₃ ceramic exhibited excellent temperature, frequency and cycling stabilities, thereby possessing outstanding practical application potential. This study exhibits a paradigm for the rapid achievement of a good energy storage performance in AgNbO₃-based and other related materials through the employment of machine learning techniques.