ANN-based investigation of the electric field dependence of electrocaloric performances in Mn/Y co-doped Ba0.67Sr0.33TiO3 ceramics

Abstract

The electrocaloric effect (ECE) offers a promising avenue for developing energy-efficient solid-state refrigeration technologies. This study employs an artificial neural network (ANN) approach to investigate the electric field dependence of the electrocaloric effect (ECE) in Mn/Y co-doped Ba_0.67Sr_0.33TiO₃ ceramics. By analysing isothermal polarization data, the ANN model accurately predicts key electrocaloric metrics, including entropy change (ΔS), temperature change (ΔT), and heat-carrying capacity (ΔQ), across a range of electric fields. The co-doping of Mn and Y enhances the microstructure, reduces leakage current, and broadens the temperature span of the ceramics, thereby improving the performance of the electrocaloric effect (ECE). The doped ceramics exhibit considerable ECE around room temperature and a relatively broad electrocaloric temperature range. The figures of merit, including the refrigerant capacity and temperature-averaged entropy change, increase with the applied field strength. Furthermore, the study investigates the field dependence of entropy change ΔS and confirms the second-order character of the electric phase transition through master curve analysis. The ANN method is demonstrated to be a rapid and accurate tool for characterising electrocaloric materials, reducing experimental time, and enhancing the optimisation of novel materials. These findings highlight the potential of Mn/Y co-doped Ba_0.67Sr_0.33TiO₃ ceramics in developing environmentally friendly and energy-efficient refrigeration technologies.