Comparative study of the Bi(Ni0.5Zr0.5)O3-PbTiO3 and Bi(Mg0.5Zr0.5)O3-PbTiO3 piezoelectric materials for energy harvesters

ABSTRACT In this study, piezoelectric properties of 0.4Bi(Ni0.5Zr0.5)O3–0.6PbTiO3 (BNZ-PT) and 0.4Bi(Mg0.5Zr0.5)O3–0.6PbTiO3 (BMZ-PT) ceramics were investigated and compared with those of Pb(Zr,Ti)O3 in energy harvester applications. Their piezoelectric properties are comparable with those of Pb(Zr,Ti)O3 based materials, although they have lower lead contents. By optimizing the sintering temperature, the piezoelectric charge coefficients (d33) were improved to 438 and 351 pC/N for 0.4BNZ-0.6PT and 0.4BMZ-0.6PT, respectively, whereas the piezoelectric voltage coefficients (g33) were improved to 25.02 × 10−3∙V∙m/N and 28.37 × 10−3∙V∙m/N for 0.4BNZ-0.6PT and 0.4BMZ-0.6PT, respectively. The generated energies were estimated by measuring the generated output voltages. Generated energy densities of 1.67 and 1.01 mJ/cm3 were obtained from 0.4BNZ-0.6PT and 0.4BMZ-0.6PT, respectively. The optimum sintering conditions were found by observing the changes in the piezoelectric constant, dielectric constant, and density as a function of the sintering temperature.


Introduction
The depletion of fossil fuels has highlighted the importance of renewable energy. Among the various renewable energy technologies, piezoelectric energy harvesting based on mechanical vibration energy is a robust renewable energy source that can convert mechanical energy into electrical energy. To present, lead-based piezoelectric ceramics have been intensively studied for industrial applications. However, owing to the environmental issues, studies on leadfree or reduced lead-content piezoelectric ceramics for piezoelectric devices are being conducted. Therefore, the world is now focusing on energy harvesting as a new energy source [1][2][3]. Energy harvesting refers to the technology of harvesting energy from natural energy sources, such as light, magnetism, vibration (piezoelectric), thermal energy, and radio frequency energy. Energy harvesting technology involves the conversion of various types of energies into usable electrical energy. Unlike other types of renewable energy sources, such as solar cells or wind power and fuel cells, piezoelectric energy harvesters are based on mechanical vibration energy. Mechanical vibration energy can be converted into electrical energy using electrostatic, electromagnetic, and piezoelectric effects [4,5]. The conversion efficiency and output property of energy harvesting devices are affected by environmental conditions. The piezoelectric energy harvesters are semi-permanent because they are less limited to environmental factors as long as the ambient vibrations are present. Therefore, piezoelectric energy harvesting is relatively less constrained by external factors than other energy harvesting technologies. In addition, it can be used wherever kinetic energy is available.
Therefore, reduced lead-content piezoelectric materials have emerged as potential materials for replacing PZT. For the first time, piezoelectric 0.4Bi(Ni 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 (BNZ-PT) and 0.4Bi(Mg 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 (BMZ-PT) ceramics were selected and compared for sensor and actuator applications. By comparing the mol% of lead, it was found that the lead composition of 0.4BNZ-0.6PT piezoelectric ceramics was 40% less than that of PZT. Thus, this piezoelectric material can be considered less harmful than PZT. In addition, 0.4BNZ-0.6PT ceramics showed similar piezoelectric properties to those of PZT ceramics. Therefore, in this study, potential piezoelectric applications, including energy harvesters of BNZ-PT and BMZ-PT ceramics, were presented. BNZ-PT and BMZ-PT piezoelectric ceramics were investigated in this study for the first time. Also, they are considered as alternatives to PZT ceramics [15,16]. BNZ-PT and BMZ-PT have improved dielectric and piezoelectric properties compared to those of other PZT-based piezoelectric materials. However, to date, only limited research has been conducted on this material. In this study, the sintering conditions of BNZ-PT and BMZ-PT were varied to determine the optimized sintering conditions. Then, the piezoelectric and dielectric properties were investigated. Furthermore, we confirmed that the optimized sintering conditions of BNZ-PT and BMZ-PT can result in improved piezoelectric and dielectric properties.

Experimental methods
The 0.4Bi(Ni 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 (BNZ-PT) and 0.4Bi(Mg 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 (BMZ-PT) ceramics were prepared using a conventional solid state reaction process. To produce BNZ-PT powders, Bi 2 O 3 (99.9%, Sigma-Aldrich Co. Ltd.), NiO (99.97%, Kojundo), ZrO 2 (99.0%, Sigma-Aldrich Co. Ltd.), PbO (99.0%, Sigma-Aldrich Co. Ltd.), and TiO 2 (99.9%, Kojundo) powders were prepared as starting materials. To produce BMZ-PT powders, Bi 2 O 3 , 4MgCO 3 ·Mg(OH) 2 · 5H 2 O (BioXtra, Sigma-Aldrich Co. Ltd.), ZrO 2 , PbO, and TiO 2 powders were prepared. These powders were ball-milled in ethyl alcohol with zirconia balls for 24 h and dried at 100 °C for 24 h. The dried powders of BNZ-PT and BMZ-PT were calcined at 780°C for 2 h in air. The calcination process was conducted at 700-900°C. After calcination, the powders had a different particle size. The powder calcined at 780°C was small and uniform. These calcined powders were mixed with 5 wt.% poly vinyl alcohol and uniaxially pressed into discs, with a diameter of 12 mm and under a pressure of 300 MPa. Then, the pressed discs were sintered in air at 1050-1200°C for 2 h, at 25°C intervals. The obtained specimens were polished to a thickness of 1 mm and silver paste was applied on both the sides. These specimens were fired at 700°C for 10 min. The volume density of the sintered specimens was measured using the Archimedes method using deionized water as medium. The crystalline structures of the sintered samples were investigated using X-ray diffraction (XRD, Bruker-AXS; New D8-Advance) analysis. The microstructures were observed using field emission scanning electron microscopy (FE-SEM, Carl Zeiss; SIGMA). The frequency-dependent dielectric constant (ɛ r ) and dielectric loss tangent of the specimens were measured using an impedance analyzer (Agilent 4294A precision) from 1 kHz to 1 MHz. The samples were poled under a DC electric field for 30 min in a silicon oil bath by employing a high voltage power supply (Hantech poling M/C). The electromechanical coupling coefficient (k p ) was estimated by measuring resonance-antiresonance frequencies (Agilent 4294A precision), following IEEE standards at room temperature. The poling temperature and electric field were varied from 20 to 100°C and 1 to 5 kV, respectively. The piezoelectric coefficient (d 33 ) values were measured using a Berlincourt quasi-static d 33 meter (YE 2730A). In particular, (100)/(001) and (200)/(002) peaks gradually split with increasing sintering temperatures. This indicates a phase transition of the 0.4BNZ-0.6PT ceramics from rhombohedral to tetragonal [17][18][19]. Similarly, for the 0.4BMZ-0.6PT ceramics in Figure 1 (b), the splitting peaks were appeared at approximately 22° and 45° with increasing sintering temperature due to a phase transition from rhombohedral to tetragonal.

Crystalline properties
Also, the XRD patterns show that both 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics were well formed with a perovskite structure. A weak pyrochlore phase peak was observed at approximately 2θ = 28°, according to the XRD JCPDS card (No. 5-543). This secondary peak was attributed to a Zr-rich phase, caused by the evaporation of Bi at high sintering temperatures. As shown in Figure 1(a,b), (002) and (200) peaks were clearly observed by increasing the sintering temperature. This means that the tetragonal structure was more clearly formed by increasing the sintering temperature for 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics. We believe that with enough thermal energy, both BNZ-PT and BMZ-PT ceramics can settle at their optimized phase structure. Figure 2 displays the bulk densities of the 0.4Bi(Ni 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 and 0.4Bi(Mg 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 piezoelectric ceramics, which were sintered from 1050 to 1200°C. As shown in Figure 2(a), the density increased with an increase in the sintering temperature, up to 1100°C. When the sintering temperature reached 1100°C, the highest density of 7.16 g/cm 3 was observed, and then the density decreased with further increase in the sintering temperature. We believe that this decrease in density was caused by the oversintering phenomenon. The bulk densities of 0.4BMZ-0.6PT are shown in Figure 2(b). Similar to the 0.4BNZ-0.6PT ceramics, the bulk density of the 0.4BMZ-0.6PT ceramics also increased as the sintering temperature was increased to 1100°C. When the sintering temperature was 1100°C, the highest bulk density value of 7.01 g/cm 3 was obtained. As the sintering temperature was increased above 1100°C and reached 1200°C, the bulk density of the 0.4BMZ-0.6PT ceramics decreased, reaching its lowest value of 6.408 g/cm 3 . Figure 3(a,b) show the frequency-dependent dielectric constant and dielectric loss of the 0.4Bi(Ni 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 and 0.4Bi(Mg 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 piezoelectric ceramics, sintered at different temperature from 1050 to 1200°C. As the frequency was varied from 1 kHz to 1 MHz, the dielectric constant ε r decreased monotonically for both 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics. As shown in Figure 3(a), for the case of 0.4BNZ-0.6PT, as the sintering temperature range was increased, the dielectric constant also increased up to 1100°C. The highest dielectric constant of 1977 was observed for the 0.4BNZ-0.6PT ceramics sintered at 1100°C, which then decreased as the sintering temperature increased. In contrast, the specimens sintered at 1200°C showed the lowest dielectric constant, with the lowest dielectric loss tangent than that of other specimens.

Dielectric characteristics and Curie temperature
As shown in Figure 3(b), for the case of the 0.4BMZ-0.6PT ceramics, the frequency-dependent dielectric constant increased as the sintering temperature was increased up to 1100°C. In general, the poling  processes were conducted to increase the dielectric polarization process. However, the main purpose for the polarization process was to obtain the piezoelectric charge coefficient, as shown in Figure 6. The dielectric constants were also strongly related with the charge coefficient. The highest dielectric constant of the 0.4BMZ-0.6PT ceramics was 1397, achieved at 1 kHz. As the sintering temperature increased, the dielectric constant also increased up to 1100 °C. The samples sintered at 1100°C showed a dielectric constant of 1397 at 1 kHz, the highest measured value. Overall, as the sintering temperature was further increased, the dielectric constant decreased. For example, the sample sintered at 1050°C showed a dielectric constant of 957, which was the lowest dielectric constant value at 1 kHz. This occurred because the sample was not sufficiently sintered. Figure 4 displays the temperature-dependent dielectric constant of the 0.4Bi(Ni 0.5 Zr 0.5 )O 3 -0.6(Pb,Ti)O 3 and 0.4Bi(Mg 0.5 Zr 0.5 )O 3 -0.6(Pb,Ti)O 3 ceramics sintered at 1100°C. The Curie temperature was measured at 1 MHz. The Curie temperature for the 0.4BNZ-0.6PT ceramics and 0.4BMZ-0.6PT ceramics was 253.8 and 250.8°C, respectively. By comparing the 0.4BNZ-0.6PT ceramics and 0.4BMZ-0.6PT ceramics, it was found that the 0.4BNZ-0.6PT ceramics showed a higher Curie temperature and maximum dielectric constant than 0.4BMZ-0.6PT ceramics. We believe that the increased Curie temperature and dielectric properties may be attributed to the increase of densification. Figure 5 shows surface images taken using field emission scanning electron microscopy (FE-SEM) obtained from the 0.4Bi(Ni 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 ceramics and 0.4Bi(Mg 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 ceramics sintered at different temperatures from 1050 to 1200°C. As shown in Figure 5(a), the 0.4BNZ-0.6PT ceramics have high porosity with very small grains, when sintered at 1050°C. The sintering mechanism and grain growth started at a temperature near 1050°C. As the sintering temperature was increased up to 1100°C, the grain size increased and became uniform. However, as the sintering temperature was further increased above 1100°C, the grains became distorted and turned to an oversintered phase. As shown in the SEM image at 1200°C, the microstructure appears dense and free from voids at some point. However, this SEM image shows irregular grain sizes, which indicates the over-growth of grain caused by over-sintering. Therefore, we believe that the irregular grain growth with increasing sintering temperature results in a decrease in densification. As shown in Figure 5(b), as the sintering temperature was increased up to 1100°C, the grain size gradually increased, and the grain size became uniform. These images also show that the porosity disappeared. However, when the sintering temperature exceeded 1100°C, the grain size tended to become distorted and unbalanced, which also seems to be due to oversintering, similar to the case for 0.4BNZ-0.6PT.   at different sintering temperatures from 1050 to 1200° C. In the case of 0.4BNZ-0.6PT, as shown in Figure 6(a), as the sintering temperature was increased up to 1100 °C, the piezoelectric charge coefficients increased up to 438 pC/N, and then decreased as the sintering temperature was increased further. It seems that the sintering temperature-dependent piezoelectric charge coefficient exhibits a behaviour similar to that of the sintering temperature-dependent bulk density and dielectric constant. In particular, the 0.4BNZ-0.6PT ceramics sintered at 1200°C showed a steeply decreased piezoelectric coefficient of 128 pC/N.

Piezoelectric properties
The piezoelectric properties of 0.4BMZ-0.6PT are shown in Figure 6(b). Similar to 0.4BNZ-0.6PT ceramics, the piezoelectric charge coefficient increased with increasing sintering temperature up to 1100°C. The piezoelectric charge coefficient showed its highest value of 351 pC/N, when sintered at 1100°C. However, as the sintering temperature was increased above 1100°C, the piezoelectric charge coefficient decreased monotonically. The reduction rate increased in the range 1175-1200°C. The specimen sintered at 1200°C showed a piezoelectric charge coefficient of 315 pC/N. It seems that this increased piezoelectric charge coefficient has a close relationship with the density and grains. Therefore, it can be argued that dipoles in the well-grown piezoelectric materials influence the piezoelectric properties of the sintered ceramics. By comparing Figure 2, a higher bulk density means a closed packed structure and improved piezoelectric properties. In addition, by considering Figure   piezoelectric ceramics is the optimized sintering temperature for density and piezoelectric properties, with a well crystallized structure. Figure 7 displays the electromechanical coupling factor (k p ) of the 0.4Bi(Ni 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 and 0.4Bi(Mg 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 ceramics sintered at temperatures from 1050 to 1200°C. As shown in the figure, the electromechanical coupling factors of the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics increased up to 1100°C, and then decreased. The shape of the electromechanical coupling factor of the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics is similar to that of the piezoelectric charge coefficient shown in Figure 6. The highest values of the electromechanical coupling factor of the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics were 57.3% and 50.8%, respectively. Figure 8 shows the poling temperature and electric field dependence of the piezoelectric coefficient of the 0.4Bi(Ni 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 and 0.4Bi(Mg 0.5 Zr 0.5 )O 3 -0.6PbTiO 3 ceramics sintered at 1100°C. As shown in Figure 8(a), the piezoelectric coefficient was increased with temperature up to 80°C, and then decreased. In addition, in Figure 8(b), the piezoelectric coefficient was increased and saturated as the electric field increased to 5 kV. The piezoelectric coefficient of the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics reached the maximum value of 440 and 355 pC/N, respectively, at a poling temperature of 80°C and electric field of 4 kV/mm. Figure 9(a) displays the schematic of the output power and energy measuring system with applied mechanical force. By employing this system, a uniform force can be applied to a piezoelectric specimen and the generated output power and energy can be measured. This system was designed and tested to obtain the actual output power and energy of the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT piezoelectric ceramics. The force was applied to the piezoelectric ceramics using a mechanical system, which applied a force of 350 N to the sample at a cycle rate of approximately 0.18 Hz [20,21]. Figure 9(b) shows equivalent circuit to obtain the generated output power. By applying this circuit, power can be calculated by multiplying open circuit voltage and short circuit current. Figure 9(c) shows energy harvesting circuit to obtaining energy. Stored voltage in the capacitor (1 μF) using full bridge rectifier is measured and calculated into energy density. By applying the energy harvesting circuit, output power and energy can be generated depending on the input forces.

Energy harvesting application
The open-circuit voltage and short-circuit current measured using this mechanical force system is shown in Figure 10(a-d). The open-circuit voltage and short-circuit current were measured using the samples sintered at 1100°C, which showed the best  characteristics among the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics sintered at temperatures between 1050 and 1200°C. Figure 10(a,c) shows the open-circuit voltage of the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics poled at 80°C and 4 kV for 30 min, respectively. The samples were subjected to a pressure energy of 350 N (0.18 Hz), during which the voltage peak values of the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics were 15.5 V and 10.3 V, respectively. Figure 10(b,d) shows the short-circuit current of the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics. In the same condition as the voltage measurement, a pressure energy of 350 N was applied at a frequency of 0.18 Hz, during which the current peak value of the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics were 1.35 μA and 1.16 μA, respectively. The generated power densities were calculated considering the sample size. The specimen diameter was approximately 10.5 mm and thickness was 1 mm. Therefore, power densities of 0.23 and 0.14 mW/cm 3 were obtained from the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics, respectively. Both 0.23 and 0.14 mW/cm 3 were enough energy sources for the sensor applications. Figure 11 shows the measured voltage across the 1 μF capacitor. The full bridge rectifier was employed to collect the positive and negative output voltages from the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics. The output voltages of 5.38 and 4.18 V were obtained for 0.4BNZ-0.6PT and 0.4BMZ-0.6PT, respectively. Additionally, the actual stored energy was calculated by measuring the stored voltage applied across the 1 μF capacitor. The stored energy was calculated using the following equation [20]: where C is the capacitance of capacitor that stored the applied voltage and V is the measured voltage across the capacitor. The stored energy density was  calculated considering the sample size. In this experiment, a load capacitor of 1 μF was employed. If the load capacitors are changed, then the applied voltage can be changed. Therefore, the generated output energy may also be changed; however, this is not true. The stored energy can be changed but the generated output energy cannot be changed. The main reason for choosing a capacitance of 1 μF was that we considered impedance matching methods. By considering impedance matching methods, optimized capacitors can be collected, and the stored energy can be optimized. Therefore, energy densities of 1.67 and 1.01 mJ/cm 3 were obtained from the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics, respectively. Meanwhile, the energy density of (33)-mode can be derived from the following (2) ~ (5) equations: ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ε T 33 S E 33 q (3) Substituting (2) to (4) into (5) gives the following equation: where t is thickness (m), s is area (m 2 ) and F is force (N). The energy density is produced by piezoelectric element is proportional to the transduction coefficient (d 33 × g 33 ), where d 33 is the piezoelectric charge coefficient and g 33 is the piezoelectric voltage coefficient, given by g 33 = d 33 /(ε 0 •ε r ) [22]. The optimized piezoelectric materials for energy harvesting should have both a high piezoelectric charge coefficient and a low dielectric constant. The piezoelectric voltage coefficients of 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics were 25.02 × 10 −3 V•m/N and 28.37 × 10 −3 V•m/N, respectively. In addition, the transduction coefficient of 10960 × 10 −15 m 2 /N and 9960 × 10 −15 m 2 /N were obtained from the 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics, respectively. We believe that the energy density of 0.4BNZ-0.6PT higher than 0.4BMZ-0.6PT is result from improving the transduction coefficient. Table 1 lists the comparison of piezoelectric properties for the lead-based energy harvesters [23][24][25][26][27][28][29][30][31]. Since energy harvesters converting environmental energy to electrical energy, in the case of low input energy, the generated output energy also relatively small [27,28]. And for the hybrid type energy harvester case, some energy can be converted to heat energy, the generated electrical energy can be small [27]. As shown in the Table 1, 0.4BNZ-0.6PT and 0.4BMZ-0.6PT piezoelectric ceramics have relatively high output voltage and energy density compared to that of the different piezoelectric ceramics.

Conclusions
For the first time, piezoelectric Bi(Ni 0.5 Zr 0.5 )O 3 -PbTiO 3 (BNZ-PT) and Bi(Mg 0.5 Zr 0.5 )O 3 -PbTiO 3 (BMZ-PT) ceramics were selected and compared for energy harvester applications. Although BNZ-PT and BMZ-PT ceramics have lower lead contents, they showed relatively high piezoelectric charge coefficient d 33 with lower dielectric constant, which can replace PZT ceramics. XRD analysis results suggested that 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics were well-crystallized to form a rhombohedral and tetragonal mixture. The uniformity of grain growth and presence of porosity were also confirmed by FE-SEM images. As a result, it was confirmed that the sintering temperature had a significant effect on the grain size and density of BNZ-PT and BMZ-PT ceramics. The piezoelectric constants for 0.4BNZ-0.6PT and 0.4BMZ-0.6PT ceramics were 438 and 351 pC/N, respectively. Generated output energy densities of 1.67 and 1.01 mJ/cm 3 were obtained from 0.4BNZ-0.6PT and 0.4BMZ-0.6PT, respectively. These results demonstrate that the proposed the 0.4BNZ-0.6PT is an outstanding candidate for utilize in the piezoelectric energy harvester with a lower lead content.

Disclosure statement
No potential conflict of interest was reported by the authors.