Switching Characteristics and High-Temperature Dielectric Relaxation Behaviours of Pb(Zn1/3Nb2/3)0.91Ti0.09O3 Single Crystal

This work evaluated the resistance switching characteristics in the (100)-oriented Pb(Zn1/3Nb2/3)0.91Ti0.09O3 (PZNT) single crystal. The current hysteresis can be closely related to the ferroelectric polarization and we provided a possible explanation using a model about oxygen vacancies to analyze the mechanism of switching. The obvious frequency dispersion of the relative permittivity signified the relaxer-type behavior of the sample. The value of the relaxation parameter γ = 1.48 was estimated from the linear fit of the modified Curie-Weiss law, indicating the relaxer nature. High-temperature dielectric relaxation behaviors were revealed in the temperature region of 400–650 °C. In addition, under the measuring frequency of 10 kHz, εr was tunable by changing the electric field and the largest tunability of εr reached 14.78%. At room temperature, the high pyroelectric coefficient and detectivity figure of merit were reported.


Experimental Section
The PZNT single crystal was grown using a modified Bridgman method [27]. The sample oriented along the (100) direction was prepared with dimensions of 5 mm × 5 mm × 1 mm and electroded with silver.
The polarization-electric-field (P-E) hysteresis loops was obtained using a Sawyer-Tower circuit by applying a sinusoidal input signal with a frequency of 1 Hz. The reference capacitor used in the measurement is 5 µF. The I-V curve was measured by a Radiant Technologies Precision premier II Ferroelectric Tester (Albuquerque, NM, USA). The temperature dependence of relative permittivity was measured by a precision LRC meter (Agilent E4980A, Agilent Technologies Inc., Santa Clara, CA, USA) in the temperature range of 20-650 • C with the heating rate of 2 • C/min. The biased temperature dependence of dielectric response was measured using a blocking circuit, a dc power source (Keithley 6517A, Keithley Instruments Inc., Solon, OH, USA) and a multi-frequency LCR meter (Model SR720 of Stanford Research Systems, Stanford University, Stanford, CA, USA) at frequencies of 10 kHz and linear temperature change was adopted as 1 • C/min in the biased heating from the temperature range of 25-240 • C. The tunability was measured using a blocking circuit, a dc power source (Keithley 6517A) and a multi-frequency LCR meter (Model SR720 of Stanford Research System) at room temperature.
Furthermore, the temperature dependence of pyroelectric coefficient was measured by a dynamic technique [32,33]. At a certain temperature T 0 , the sample temperature was sinusoidally modulated T(t) = T 0 + T~sin 2πft ] with frequency f = 5 mHz and amplitude T~= 1 K using a Peltier element. The pyroelectric current signal was amplified with an electrometer and the 90 • out of phase component of current with respect to the temperature modulation was measured with a lock-in amplifier. After setting to a new temperature T 0 the sample was kept at T 0 for 15 min for the signal to become stable before the pyroelectric measurement was performed. PZNT single crystal was first poled at room temperature by applying an electric field of amplitude 10 kV/cm for 5 min and then short circuited circuited at room temperature overnight.

Results and Discussion
The current-voltage (I-V) curves for the (100)-oriented PZNT single crystal obtained by decreasing the sweep range step by step are shown in Figure 1a. In the range of ±50 V, the test voltage was swept at a constant rate from 0-50 V, then to −50 V, before returning to 0 V. The numbers in the figure denote the sequence of the voltage sweeps. An asymmetry in the current can be observed in the figure.
In addition, we can find that the I-V curves show distinct hysteresis behavior, indicating the sample exhibited obvious resistance switching characteristics. Moreover, in the inset (A) of Figure 1a, the I-V segment shows an obvious diode-like rectifying I-V characteristic, indicating a diode behavior.
The I-V curves plotted on semilogarithmic scales are shown in the inset (B) of Figure 1a. When the sweeping positive bias increased from 0 V to about 33.3 V, the resistance switching was in the high-resistance state (HRS), with a low current. When the voltage was beyond 33.3 V, the current increased quickly and attained a maximum (around 5 µA) with the voltage reached 50 V, indicating that the resistance switching "turns on" and switches from the HRS to the low-resistance state (LRS). As the voltage swept back from 50 V to about −33.3 V, the resistance switching remained in the LRS. As the resistance switching continued its negative bias, the current stabilized at around 0.7 µA until the sweeping bias exceeded −33.3 V, which led to a transition from the LRS to HRS. Finally, when the sweeping voltage returned, going from −50 to 0 V, the resistance switching remained in the HRS [34,35]. In addition, the resistance switching ratio was 229 at 2.08 V.
The mechanism of resistance switching characteristics is not yet perfectly understood. Despite this, there were some phenomenological models to explain these characteristics, namely the Schottky barrier model [36,37], space charge [38,39], the electrically conducting filamentary model [40,41], and the Mott transition [42]. Jeon et al. [43] reported that resistance switching characteristics are caused by the change in the Schottky potential based on the results from the first-principles calculations. It is worth mentioning that the observed ferroelectric resistance switching behavior in our crystal would be different from that observed in some ferroelectric tunneling junctions [44,45], because the tunneling current can only be taken into account for the ultrathin ferroelectric materials, not for our crystal with a thickness of 1 mm.

Results and Discussion
The current-voltage (I-V) curves for the (100)-oriented PZNT single crystal obtained by decreasing the sweep range step by step are shown in Figure 1a. In the range of ±50 V, the test voltage was swept at a constant rate from 0-50 V, then to −50 V, before returning to 0 V. The numbers in the figure denote the sequence of the voltage sweeps. An asymmetry in the current can be observed in the figure. In addition, we can find that the I-V curves show distinct hysteresis behavior, indicating the sample exhibited obvious resistance switching characteristics. Moreover, in the inset (A) of Figure 1a, the I-V segment shows an obvious diode-like rectifying I-V characteristic, indicating a diode behavior.
The I-V curves plotted on semilogarithmic scales are shown in the inset (B) of Figure 1a. When the sweeping positive bias increased from 0 V to about 33.3 V, the resistance switching was in the high-resistance state (HRS), with a low current. When the voltage was beyond 33.3 V, the current increased quickly and attained a maximum (around 5 μA) with the voltage reached 50 V, indicating that the resistance switching "turns on" and switches from the HRS to the low-resistance state (LRS). As the voltage swept back from 50 V to about −33.3 V, the resistance switching remained in the LRS. As the resistance switching continued its negative bias, the current stabilized at around 0.7 μA until the sweeping bias exceeded −33.3 V, which led to a transition from the LRS to HRS. Finally, when the sweeping voltage returned, going from −50 to 0 V, the resistance switching remained in the HRS [34,35]. In addition, the resistance switching ratio was 229 at 2.08 V.
The mechanism of resistance switching characteristics is not yet perfectly understood. Despite this, there were some phenomenological models to explain these characteristics, namely the Schottky barrier model [36,37], space charge [38,39], the electrically conducting filamentary model [40,41], and the Mott transition [42]. Jeon et al. [43] reported that resistance switching characteristics are caused by the change in the Schottky potential based on the results from the first-principles calculations. It is worth mentioning that the observed ferroelectric resistance switching behavior in our crystal would be different from that observed in some ferroelectric tunneling junctions [44,45], because the tunneling current can only be taken into account for the ultrathin ferroelectric materials, not for our crystal with a thickness of 1 mm. In the range of ±12, ±15 and ±20 V, the I-V curve was almost symmetrical and without hysteresis, as shown in Figure 1a. Then, as the voltage increased to ±50 V, obvious hysteresis and resistance switching characteristics were observed. It was indicated that the current hysteresis and In the range of ±12, ±15 and ±20 V, the I-V curve was almost symmetrical and without hysteresis, as shown in Figure 1a. Then, as the voltage increased to ±50 V, obvious hysteresis and resistance Materials 2017, 10, 349 4 of 10 switching characteristics were observed. It was indicated that the current hysteresis and diode-like behavior can be triggered and switched at a high applied field. Figure 1b shows the P-E hysteresis loops for the PZNT single crystal with various voltages, indicating that the sample exhibits good ferroelectricity. It is worth mentioning that the P-E hysteresis loops starts at around ±45 V (not marked in the figure). According to the report of BiFeO 3 [46], because the current and ferroelectric polarization exhibited the hysteresis phenomenon at a high applied field, the current hysteresis can be closely related to the ferroelectric polarization.
The polarization mechanism of ferroelectric materials includes displacement polarization and turning-direction polarization. In the process of the preparation of samples, lead vacancies will appear inevitably due to the volatility of lead, leading to the oxygen vacancies (OVs) due to charge neutrality restrictions. It is well known that the ionization of OVs will create conducting electrons in perovskite-structure oxides during the process of preparation at high temperatures, written as: (1) where Vo•, Vo•• are single-ionized and double-ionized OVs, respectively. The current occurs by electron injection from one electrode affected by the concentration of OVs near the interface. So, according to the current study, we preferred to believe that the resistance switching characteristics of the PZNT single crystal could be caused by the change in the oxygen vacancies concentration at the metal/oxide interface by the electrically controlled electron injection.
In order to describe the processes underpinning resistance switching, the displacement and migration of OVs near the bottom electrode (BE) were introduced. Displacement was defined as the reversible movement of OVs under electrical bias (the OVs cannot get enough energy to go over the Schottky barrier), where upon removal of the applied voltage the OVs return to their initial locations. Migration occurred when the field-driven movement of the vacancies was not reversible upon removal of the bias (the OVs can get enough energy to go over the Schottky barrier) [47].
A model, shown in Figure 2a-d, was set up to explain the resistance switching characteristics of the PZNT single crystal. The yellow circles with plus signs in the model represented OVs. When the sweeping positive bias increased from 0 V to 33.3 V, the OVs could not get enough energy to go over the Schottky barrier. The positive bias displaces OVs towards BE (displacement of OVs) to enhance the electron injection, as shown in Figure 2a. However, when the bias was beyond 33.3 V, the OVs surmounted the Schottky barrier to BE (migration of OVs) because they obtained enough energy, and it greatly increased the electron injection, as shown in Figure 2b. Consequently, the current increased quickly, the resistance switching "turned on" and the switches changed their state from the HRS to the LRS. The state would remain until the bias reversal. As the voltage swept from 0 V to around −33.3 V, the state still remained at the LRS, but the current appeared asymmetric due to the displacement of OVs, as shown in Figure 2c. When the negative bias kept increasing, the resistance switching returned to the HRS because of the migration of the OVs. Finally, the resistance switching returned to the original state when the bias returned to 0 V. diode-like behavior can be triggered and switched at a high applied field. Figure 1b shows the P-E hysteresis loops for the PZNT single crystal with various voltages, indicating that the sample exhibits good ferroelectricity. It is worth mentioning that the P-E hysteresis loops starts at around ±45 V (not marked in the figure). According to the report of BiFeO3 [46], because the current and ferroelectric polarization exhibited the hysteresis phenomenon at a high applied field, the current hysteresis can be closely related to the ferroelectric polarization. The polarization mechanism of ferroelectric materials includes displacement polarization and turning-direction polarization. In the process of the preparation of samples, lead vacancies will appear inevitably due to the volatility of lead, leading to the oxygen vacancies (OVs) due to charge neutrality restrictions. It is well known that the ionization of OVs will create conducting electrons in perovskite-structure oxides during the process of preparation at high temperatures, written as: where Vo•, Vo•• are single-ionized and double-ionized OVs, respectively. The current occurs by electron injection from one electrode affected by the concentration of OVs near the interface. So, according to the current study, we preferred to believe that the resistance switching characteristics of the PZNT single crystal could be caused by the change in the oxygen vacancies concentration at the metal/oxide interface by the electrically controlled electron injection. In order to describe the processes underpinning resistance switching, the displacement and migration of OVs near the bottom electrode (BE) were introduced. Displacement was defined as the reversible movement of OVs under electrical bias (the OVs cannot get enough energy to go over the Schottky barrier), where upon removal of the applied voltage the OVs return to their initial locations. Migration occurred when the field-driven movement of the vacancies was not reversible upon removal of the bias (the OVs can get enough energy to go over the Schottky barrier) [47].
A model, shown in Figure 2a-d, was set up to explain the resistance switching characteristics of the PZNT single crystal. The yellow circles with plus signs in the model represented OVs. When the sweeping positive bias increased from 0 V to 33.3 V, the OVs could not get enough energy to go over the Schottky barrier. The positive bias displaces OVs towards BE (displacement of OVs) to enhance the electron injection, as shown in Figure 2a. However, when the bias was beyond 33.3 V, the OVs surmounted the Schottky barrier to BE (migration of OVs) because they obtained enough energy, and it greatly increased the electron injection, as shown in Figure 2b. Consequently, the current increased quickly, the resistance switching "turned on" and the switches changed their state from the HRS to the LRS. The state would remain until the bias reversal. As the voltage swept from 0 V to around −33.3 V, the state still remained at the LRS, but the current appeared asymmetric due to the displacement of OVs, as shown in Figure 2c. When the negative bias kept increasing, the resistance switching returned to the HRS because of the migration of the OVs. Finally, the resistance switching returned to the original state when the bias returned to 0 V.   In the heating temperature range from 20-650 • C at a rate of 2 • C min −1 , the temperature dependence of the relative permittivity ε r and the dielectric loss tanδ of the PZNT single crystal at various frequencies (from 1-100 kHz) are shown in Figure 3. The inset shows the dielectric loss tanδ in the temperature range between 20-300 • C. Permittivity curves for various frequencies exhibit organized (the relative permittivity ε r decreased with the increasing frequency) and the dielectric peaks are located at the temperature of about 177 • C (temperature of the maximum dielectric permittivity (T m )) and T m is unchanged with the increasing frequency, indicating a phase transition from the FE phase to the PE cubic phase. Seung-Eek Park et al. reported that the phase transition temperature was 180 • C [25]. In the temperature of T m , the values of the dielectric peak ε m could be as high as 8420 at the measurement frequency of 1 kHz. With increasing the frequency from 1-100 kHz, the value of ε m decreased from 8420 to 6135. Specially, there was another dielectric anomaly peak that took place at about 90 • C, indicating a phase transition from the FE r phase to the FE t phase. The same phenomenon had also been reported by Hosono et al. [48]. Obviously, we can observe (inset of Figure 3) the same phase transformation from the dielectric constant. However, it was worth noting that, at the high temperature range (400-650 • C), the permittivity curves presented a high-temperature relaxation phenomenon that looked like the behavior of a diffuse phase transition and the value of tanδ became large, because the space charge polarization or the conductivity of the insulating ceramics increased with the increase in temperature. In the heating temperature range from 20-650 °C at a rate of 2 °C min −1 , the temperature dependence of the relative permittivity εr and the dielectric loss tanδ of the PZNT single crystal at various frequencies (from 1-100 kHz) are shown in Figure 3. The inset shows the dielectric loss tanδ in the temperature range between 20-300 °C. Permittivity curves for various frequencies exhibit organized (the relative permittivity εr decreased with the increasing frequency) and the dielectric peaks are located at the temperature of about 177 °C (temperature of the maximum dielectric permittivity (Tm)) and Tm is unchanged with the increasing frequency, indicating a phase transition from the FE phase to the PE cubic phase. Seung-Eek Park et al. reported that the phase transition temperature was 180 °C [25]. In the temperature of Tm, the values of the dielectric peak εm could be as high as 8420 at the measurement frequency of 1 kHz. With increasing the frequency from 1-100 kHz, the value of εm decreased from 8420 to 6135. Specially, there was another dielectric anomaly peak that took place at about 90 °C, indicating a phase transition from the FEr phase to the FEt phase. The same phenomenon had also been reported by Hosono et al. [48]. Obviously, we can observe (inset of Figure 3) the same phase transformation from the dielectric constant. However, it was worth noting that, at the high temperature range (400-650 °C), the permittivity curves presented a high-temperature relaxation phenomenon that looked like the behavior of a diffuse phase transition and the value of tanδ became large, because the space charge polarization or the conductivity of the insulating ceramics increased with the increase in temperature.  Figure 3, the obvious frequency dispersion of the dielectric constant can be observed, and this phenomenon signifies the typical relaxer behavior of the present specimen. The dielectric characteristics of relaxor ferroelectric materials are well known to deviate from the typical Curie-Weiss behavior, and can be described by a modified Curie-Weiss relationship [49]: where γ and C1 are assumed to be constant. For γ = 1, a normal Curie-Weiss law was obtained, and a complete diffuse phase transition was described for γ = 2 [50]. The plots of ln(1/εr − 1/εm) versus ln(T − Tm) with a frequency of 10 kHz are shown in Figure 4. We can get the value of γ = 1.48 by fitting the experimental data. Values of γ in this work were found to vary from 1.31 to 1.62 in the frequency range from 1-300 kHz. This also supported the evidence of the relaxer nature of the PZNT single crystal.  Figure 3, the obvious frequency dispersion of the dielectric constant can be observed, and this phenomenon signifies the typical relaxer behavior of the present specimen. The dielectric characteristics of relaxor ferroelectric materials are well known to deviate from the typical Curie-Weiss behavior, and can be described by a modified Curie-Weiss relationship [49]: where γ and C 1 are assumed to be constant. For γ = 1, a normal Curie-Weiss law was obtained, and a complete diffuse phase transition was described for γ = 2 [50]. The plots of ln(1/ε r − 1/ε m ) versus ln(T − T m ) with a frequency of 10 kHz are shown in Figure 4. We can get the value of γ = 1.48 by fitting the experimental data. Values of γ in this work were found to vary from 1.31 to 1.62 in the frequency range from 1-300 kHz. This also supported the evidence of the relaxer nature of the PZNT single crystal. At 100 Hz, the temperature dependence of εr for the PZNT single crystal during the heating process under different electric fields is shown in Figure 5. The curves of the temperature dependence of εr for the PZNT single crystal were lower and wider as the electric field increased, indicating that the single crystal went through a field-induced phase transition. When the value of the applied field was 0 kV/mm, the maximum relative permittivity (εm) was 6934 at about 190 °C and the εm decreased from 6934 to 2970 with the applied field increasing from 0 to 1 kV/mm. Particularly, the Tm presented a different value under different applied electric fields, suggesting that the single crystal underwent a second-order phase transition. The values of εm and Tm are listed in Table 1. The result clearly indicates a broadening of the dielectric peak due to the diffuse ferroelectric-paraelectric phase transition.   At 100 Hz, the temperature dependence of ε r for the PZNT single crystal during the heating process under different electric fields is shown in Figure 5. The curves of the temperature dependence of ε r for the PZNT single crystal were lower and wider as the electric field increased, indicating that the single crystal went through a field-induced phase transition. When the value of the applied field was 0 kV/mm, the maximum relative permittivity (ε m ) was 6934 at about 190 • C and the ε m decreased from 6934 to 2970 with the applied field increasing from 0 to 1 kV/mm. Particularly, the T m presented a different value under different applied electric fields, suggesting that the single crystal underwent a second-order phase transition. The values of ε m and T m are listed in Table 1. The result clearly indicates a broadening of the dielectric peak due to the diffuse ferroelectric-paraelectric phase transition. At 100 Hz, the temperature dependence of εr for the PZNT single crystal during the heating process under different electric fields is shown in Figure 5. The curves of the temperature dependence of εr for the PZNT single crystal were lower and wider as the electric field increased, indicating that the single crystal went through a field-induced phase transition. When the value of the applied field was 0 kV/mm, the maximum relative permittivity (εm) was 6934 at about 190 °C and the εm decreased from 6934 to 2970 with the applied field increasing from 0 to 1 kV/mm. Particularly, the Tm presented a different value under different applied electric fields, suggesting that the single crystal underwent a second-order phase transition. The values of εm and Tm are listed in Table 1. The result clearly indicates a broadening of the dielectric peak due to the diffuse ferroelectric-paraelectric phase transition.    Dielectric tunable materials have a wide range of applications, such as in phase shifters, oscillators, filters, etc. [51]. From the above discussion, ε r was found to be tunable by changing the electric field. The ε r and the tunability of ε r are shown in Figure 6, respectively. The test was conducted under different electric fields at 10 kHz at room temperature and the applied electric fields increased from 0 to 1000 V/mm. The tunability of ε r is defined as [ε r (0) − ε r (E)] × 100%/ε r (0), where ε r (0) and ε r (E) are the ε r values when the electric field is zero and E, respectively. The results revealed that the largest tunability of ε r was 14.78%. With the increase of the electric field, the ε r decreased gradually and the tunability increased, respectively [52,53].
Dielectric tunable materials have a wide range of applications, such as in phase shifters, oscillators, filters, etc. [51]. From the above discussion, εr was found to be tunable by changing the electric field. The εr and the tunability of εr are shown in Figure 6, respectively. The test was conducted under different electric fields at 10 kHz at room temperature and the applied electric fields increased from 0 to 1000 V/mm. The tunability of εr is defined as [εr(0) − εr(E)] × 100%/εr(0), where εr(0) and εr(E) are the εr values when the electric field is zero and E, respectively. The results revealed that the largest tunability of εr was 14.78%. With the increase of the electric field, the εr decreased gradually and the tunability increased, respectively [52,53].  The pyroelectric coefficients as a function of the temperature for the PZNT single crystal are shown in Figure 7. The real part of the pyroelectric coefficients increases slowly from −576.1 to −447.4 μC/m 2 K with increasing temperatures from 18-50 °C, respectively. The imaginary part of the pyroelectric coefficients are pyroelectric losses. At a room temperature of 25 °C, the value of the pyroelectric coefficient was −463.3 μC/m 2 K, and the absolute value of the pyroelectric coefficient (463.3 μC/m 2 K) was higher than that of the graded PZT films on Pt-coated silicon substrates (202-250 μC/m 2 K) and Ba3Nb2O8 ceramic (103 μC/m 2 K) [32,54]. A useful comparative figure of merit (FOM) used in comparing pyroelectric materials is defined as FD = |P|/c(εrε0tanδ) 1/2 , where c is the heat capacity per unit volume (c = 2.5 J/cm 3 K) [55], εr is the relative permittivity, ε0 is the permittivity Figure 6. At room temperature, 10 kHz of AC bias and composition dependence of (a) the relative permittivity ε r ; (b) the tunability of ε r of the PZNT single crystal.
The pyroelectric coefficients as a function of the temperature for the PZNT single crystal are shown in Figure 7. The real part of the pyroelectric coefficients increases slowly from −576.1 to −447.4 µC/m 2 K with increasing temperatures from 18-50 • C, respectively. The imaginary part of the pyroelectric coefficients are pyroelectric losses. At a room temperature of 25 • C, the value of the pyroelectric coefficient was −463.3 µC/m 2 K, and the absolute value of the pyroelectric coefficient (463.3 µC/m 2 K) was higher than that of the graded PZT films on Pt-coated silicon substrates (202-250 µC/m 2 K) and Ba 3 Nb 2 O 8 ceramic (103 µC/m 2 K) [32,54]. A useful comparative figure of merit (FOM) used in comparing pyroelectric materials is defined as F D = |P|/c(ε r ε 0 tanδ) 1/2 , where c is the heat capacity per unit volume (c = 2.5 J/cm 3 K) [55], ε r is the relative permittivity, ε 0 is the permittivity of the vacuum and tanδ is the dissipation factor. The F D value is 8.77 × 10 −6 Pa −0.5 at 1 kHz. This result implies that the PZNT single crystal can be a promising material for pyroelectric array sensor applications. Dielectric tunable materials have a wide range of applications, such as in phase shifters, oscillators, filters, etc. [51]. From the above discussion, εr was found to be tunable by changing the electric field. The εr and the tunability of εr are shown in Figure 6, respectively. The test was conducted under different electric fields at 10 kHz at room temperature and the applied electric fields increased from 0 to 1000 V/mm. The tunability of εr is defined as [εr(0) − εr(E)] × 100%/εr(0), where εr(0) and εr(E) are the εr values when the electric field is zero and E, respectively. The results revealed that the largest tunability of εr was 14.78%. With the increase of the electric field, the εr decreased gradually and the tunability increased, respectively [52,53].  The pyroelectric coefficients as a function of the temperature for the PZNT single crystal are shown in Figure 7. The real part of the pyroelectric coefficients increases slowly from −576.1 to −447.4 μC/m 2 K with increasing temperatures from 18-50 °C, respectively. The imaginary part of the pyroelectric coefficients are pyroelectric losses. At a room temperature of 25 °C, the value of the pyroelectric coefficient was −463.3 μC/m 2 K, and the absolute value of the pyroelectric coefficient

Conclusions
In conclusion, this work has shown the resistance switching characteristics in the (100)-oriented PZNT single crystal. The current hysteresis can be closely related to the ferroelectric polarization and we provided a possible explanation with a model of oxygen vacancies to analyze the mechanism of switching. In the process of heating, the temperature of the rhombohedral-tetragonal phase transition was about 90 • C and the temperature of the FE-PE phase transition was about 177 • C. The obvious frequency dispersion of the relative permittivity signified the relaxer-type behavior of the sample, and the value of the relaxation parameter γ = 1.48, estimated from the linear fit of the modified Curie-Weiss law, indicated the relaxer nature. High-temperature dielectric relaxation behaviors were revealed in the temperature region of 400-650 • C. In addition, under the measuring frequency of 10 kHz, we found that ε r was tunable by changing the electric field and the largest tunability of ε r reached 14.78%. At room temperature, the pyroelectric coefficient and the figure of merit F D were −463.3 µC/m 2 K and 8.77 × 10 −6 Pa −0.5 , respectively.