Doping-induced Polar Defects Improve the Electrocaloric Performance of ${\mathrm{Ba}}_{0.9}{\mathrm{Sr}}_{0.1}{\mathrm{Hf}}_{0.1}{\mathrm{Ti}}_{0.9}{\mathrm{O}}_3$

In materials science, intentional doping has been widely used to improve the properties of a variety of materials. However, such an approach is not yet exploited in the fast-growing field of electrocaloric materials, which represent a serious alternative for next-generation cooling systems. Here we demonstrate with ${\mathrm{Ba}}_{0.9}{\mathrm{Sr}}_{0.1}{\mathrm{Hf}}_{0.1}{\mathrm{Ti}}_{0.9}{\mathrm{O}}_3$, an ecofriendly ferroelectric material, that doping with 2% of $\mathrm{Cu}$ introduces defect dipoles into the ferroelectric matrix and results in (i) enhancement of the adiabatic temperature change ΔT by up to 54% while maintaining performance after a large number (up to ${10}^4$) of electric field cycles, (ii) suppression of the parasitic irreversibility of ΔT between on-field and off-field states, and (iii) an alternative design of refrigeration cycle with a prepoled sample, allowing a two-field-step process showing both conventional (ΔT > 0) and inverse (ΔT < 0) responses when the field is sequentially varied. We also demonstrate that doping significantly increases the energy storage density (by up to 72%). The defect engineering approach therefore offers a path for designing ferroelectrics with improved electrocaloric performances. Beyond ferroelectrics, this strategy could also be promising in other solid-state caloric materials (magnetocalorics, elastocalorics, etc.).

Figure 1

${\mathrm{Ba}}_{0.8}{\mathrm{Sr}}_{0.2}{\mathrm{TiO}}_3$–${\mathrm{BaHf}}_{0.2}{\mathrm{Ti}}_{0.8}{\mathrm{O}}_3$ (BSHT) phase diagram and dielectric and polarization behavior without and with intentional $\mathrm{Cu}$ doping. (a) The phase diagram of the BSHT system constructed using the dielectric permittivity versus temperature results. (b) and (c) Temperature dependence of dielectric permittivity (at 100 Hz; cooling rate, 3 °C/min) and specific heat capacity of BSHT-0 and BSHT-2 samples. Room-temperature PE loops (at 10 Hz). (d) The normal PE loop of BSHT-0. (e) The appearance of the pinched hysteresis loop for BSHT-2. (f) The shifted single hysteresis loop for prepoled BSHT-2.

Figure 2

Electrocaloric response and defect dipole consequences. (a) The direct temperature change (ΔT) in the field off regime for BSHT-0 and BSHT-2 samples measured under different electric fields at room temperature. (b) and (c) Directly measured values of ΔT, as a function of temperature, under the application (E_on) and removal (E_off) of the electric field (10 kV cm⁻¹). The insets show the directly recorded heat flow signal as the electric field is turned on and off. Dashed line is drawn to show the isothermal process. (d) Two idealized refrigeration cycles proposed for BSHT-0 and BSHT-2. (e) The PE loop and direct |ΔT| values of poled BSHT-2. The insets show the schematic configurations in a single domain pattern with defect dipoles and the connection between the Landau free energy (F) and polarization (P) under external electric field in prepoled BSHT-2. (f) The signal of direct ECE measurements for poled BSHT-2. (g) An idealized refrigeration cycle that utilizes both negative and positive ECE (steps i–viii) based on a prepoled BSHT-2 sample compared with the conventional cycle (steps I–IV). The cycle begins (step i) with the adiabatic application of a low electric field (E_d) that causes cooling of the material via the negative ECE (ΔT_n = |T_a − T_d|). Then, the cold body absorbs heat from the environment under constant E_d (step ii) and the system returns to T_a. In step iii, the adiabatic application of a large electric field (ΔE = E_M − E_d) results in an increase in the temperature of the material. The material then returns to the original T_a by heat exchange with a sink under constant E_M (step iv). In step v, E_M decreases adiabatically to E_d, leading to a drop in the temperature to T_c via a positive ECE (ΔT_p = |T_a − T_c|). In step vi the system returns to the original T_a by absorbing heat from the load. The cycle closes via adiabatic removal (decrease) of E_d (step vii) followed by heat exchange with a sink (step viii). Overall, the final temperature change is enhanced to ΔT = ΔT_n + ΔT_p. Note: All the temperature terms in (d) and (g) are used for better illustration; they are not actual values. Specifically, T_a is the initial temperature of the sample and T_b, T_c, T_d, and T_e are the corresponding temperatures induced by the ECE.

Figure 3

Electrocaloric response after cycling and aging. The durability test to confirm the cycling and cooling stability of the BSHT-2 and poled BSHT-2 samples under 30 kV cm⁻¹. (a)–(c) P_max, P_r, and E_c of BSHT-2 as a function of the number of cycles under 30 kV cm⁻¹. (d) The direct and indirect ΔT changes for BSHT-0 and BSHT-2 samples as a function of the number of cycles for 10 and 30 kV cm⁻¹, respectively. (e)–(g) The P_max, P_r, E_c, and E_d for prepoled BSHT-2 as a function of the number of cycles under 30 kV cm⁻¹. (h) The direct and indirect ΔT change for prepoled BSHT-2 samples as a function of the number of cycles for 10 and 30 kV cm⁻¹, respectively. The data in the right hand area is from cycled samples after aging for 24 h. The waveform of the electric field for fatigue testing is shown in the inset.