Correlation between phase, microstructure and electrical properties of Ba0.7Sr0.3TiO3-modified Bi0.5Na0.5TiO3-0.06BaTiO3 lead free ceramics

ABSTRACT Environmentally friendly Pb-free piezoelectric ceramics are essential for electronic industries and devices utilizations owing to their excellent piezoelectric properties which enable powerful sensor, transducer and actuator developments. In the present work, highly dense complex perovskite oxide ceramics with the formula of (1-x)[0.94Bi0.5Na0.5TiO3-0.06BaTiO3]-x(Ba0.7Sr0.3)TiO3 or (1-x)[0.94BNT–0.06BT]-xBST, where x = 0.00, 0.01, 0.02, 0.03, 0.04 and 0.05 mol fraction were fabricated by a solid-state mixed oxide method. After sintering at 1125°C for 2 h, all ceramics were characterized with attention paid on their phases, microstructures, dielectric, ferroelectric, piezoelectric and strain properties as a function of BST content. The unit all size expansion i.e. higher degree of tetragonality was detected as BST content increased. All compositions exhibited single perovskite phase and uniform microstructures containing cubic-like grains with average grain size range of all ceramics was in the order of ~ 0.75–1.57 µm, depending on the amount of BST. It was interestingly found that the addition of BST enhanced the densification, sintering ability, grain size and electrical properties of BNT-BT ceramics, indicating their potential Pb-free piezoelectric candidates for future actual applications.


Introduction
Lead-based perovskite materials such as PZT-based systems are the most widely used piezoelectric ceramics industrially manufactured in large scale in the world today.Their low-firing temperature which enables low-cost electrodes and excellent piezoelectric properties make these materials keep up with the trend of electrochromic applications and for energy harvesting applications [1][2][3][4][5][6][7].Moreover, their key fundamental investigation and technological importance have been extensively demonstrated in the literature [1][2][3][4][5][6][7][8].However, it was reported earlier that the volatilization of the lead oxide (PbO) at high temperature during heat treatment processes causing toxicity environmental pollution and uncertainty in their electrical properties is one of the major obstacles of these Pb-based ceramics [8][9][10][11][12].Hence, there has been a great deal of interest in the search for novel alternative lead-free piezoelectric materials that can take over the existed Pb-based compounds, in order to continue environmentally sustainable development.The challenges for further evolution of lead-free piezoelectric materials for electronic devices were systematically reviewed by Panda et al [9].Among Pb-based perovskite piezoelectric ceramics, BaTiO 3 or BT, Bi 0.5 Na 0.5 TiO 3 or BNT, Bi 0.5 K 0.5 TiO 3 or BKT, Bi 0.5 (Na 0.8 K 0.2 ) 0.5 TiO 3 or BNKT and their solid-solutions are widely explored as promising candidates for further technological utilizations due to nontoxic handling and individual properties of these materials [9][10][11][12][13][14][15].As is well documented, assembling various materials together is one of the useful approaches to generate novel multicomponent deriving excellent properties from each component.For example, previous works have shown that the BNKT-based systems modified with a number of oxide compounds, e.g.SrTiO 3 [13], BiFeO 3 [14] and Ba 0.7 Sr 0.3 TiO 3 [15], exhibit enhanced piezoelectric properties.On the other hand, our previous work [16] observed that the addition of LiNbO 3 significantly decreases the values of d 33 but increases the room temperature dielectric constant of BNKT ceramics, depending on the amount of added dopants.In recent years, there are several reports [17][18][19][20][21][22][23] on the investigation of BNTbased solid-solutions including BNT-BT system since they are considered to be one of the most promising candidates to replace PZT due to their excellent piezoelectric performance and large remnant polarization located at or near morphotropic phase boundary (MPB).
According to the literature, BNT is one of the relaxor ferroelectric materials and its solid-solution with normal ferroelectric BT could contribute excellent piezoelectric properties.It is known that their electrical properties depend strongly on microstructure as well as chemical compositions.It is believed that the similarity of Bi and Pb ions, in views of their large number of electrons, ionic sizes and lone pair electron, leads to the deformation of unit cell and thus to their piezoelectric properties [24,25].The BNT can be synthesized by solid state reaction technique.The advantages of technique include lower cost compared with other techniques, short reaction time, high purity of product, collection process of synthesized products is straightforward and it has good reaction efficiency [26].Furthermore, BNT indicated that nontoxic, eco-friendly piezoelectric material and attractive piezoelectric properties with high curie temperature (T C = 593 K), large remanent polarization (P r = 38 μC/cm 2 ) and high coercive field (73 kV/cm) at room temperature [27,28].However, the limitation of BNT ceramics lies in their large leakage current density and high driving electric field [8,9].Thus, utilizing chemical modification via additive is one of the interesting approaches.Recently, Supriya et al. reported Sr 2+ doped BNT system in various forms like single crystals, thin films, multi-layers, and polycrystals for enhanced piezoelectric and ferroelectric activities [29].Ren et al [30].reported large electro-strain in Pb-free BNKT ceramics modified with Ba 0.95 Sr 0.05 TiO 3 and Ta 5+ ion, where the intrinsic effect of the crystal and the external electric field could be their major causes.In connection with a solid-solution of barium strontium titanate Ba 1-x Sr x TiO 3 , related works [31] the maximum room temperature dielectric constant (ε r ~4500) was observed at x = 0.3.Moreover, the energy harvesting performance as a key component of a piezoelectric energy harvester is the piezoelectric materials themselves, which play an important role in determining the output performance of an energy harvester [32][33][34][35].Rout et al. studied the (100-x) Na 0.5 Bi 0.5 TiO 3 -xSrTiO 3 system 0≤x≤40 and reported the MPB for this system was found at x = 0.20 and room temperature value presented enhancement from 500 (x = 0) to 3800 (x = 40) at 1 kHz with increasing Sr content [36].Energy harvesting is the process by which energy is derived from external various sources such as wind energy, thermal energy, mechanical vibrations, or kinetic energy [37][38][39][40][41][42][43].Several Pb-free systems had good energy harvesting performance such as 0.94 (Bi 0.5 Na 0.5 )TiO 3 -0.06BaTiO 3 [44], (K,Na,Li)(Nb,Ta,Sb)O 3 [45], Ba 0.85 Ca 0.15 Ti 0.9 Zr 0.1 O 3 [46], Ba(Zr 0.1 Ti 0.9 )O 3 [47], (Ba,Ca)(Zr,Ti)O 3 [48], Bi 0.5 (Na 0.84 K 0.16 ) 0.5 TiO 3 /0.01ZnO[49] and (1-y)[0.995Bi0.5 (Na 0.80 K 0.20 ) 0.5 TiO 3 -0.005LiNbO 3 ]-y[(Ba 0.7 Sr 0.3 )TiO 3 ] [50].
Thus, Ba 0.7 Sr 0.3 TiO 3 composition was chosen as the desired dopant for the BNT-BT based system in our study.Interestingly, so far, there is no systematic investigation on Pb-free BNT-BT modified with such chemical formula of Ba 0.7 Sr 0.3 TiO 3 or BST.Hence, this work is the continuation of our current efforts to develop a novel alternative Pb-free piezoelectric ceramics for Pb-based ceramics replacement and we will discuss on how the different amount of BST fraction could affect the phase, structure, electrical properties and energy harvesting performance of the BNT-BT ceramics.

Experimental procedures
The perovskite piezoelectric ceramics of (1-x) [0.94Bi 0.5 Na 0.5 TiO 3 -0.06BaTiO3-x ] -x(Ba 0.7 Sr 0.3 )TiO 3 or (1-x)[0.94BNT-0.06BT]-xBST,where x = 0.00, 0.01, 0.02, 0.03, 0.04 and 0.05 mol fraction, were fabricated by a solid-state mixed oxide method.Commercial powders of Bi 2 O 3 (98%, Fluka, Sigma-Aldrich Cot., St. Louis, MS), Na 2 CO 3 (99.5%,Carlo Erba, Carlo Erba Reagenta, Peypin, France), TiO 2 (99%, Riedel-de Haën, Honeywell Specialty Chemicals Seelze GmbH, Seelze, German), BaCO 3 (98.5%,Fluka, Sigma-Aldrich Corp.) and SrCO 3 .(98%, Sigma-Aldrich, Sigma-Aldrich Corp.) were used as starting powders.The methods used for mixing, drying, and grinding of the powders were similar to those used in our earlier work [16].All starting materials were stoichiometrically weighed, ball milled for 24 h in an ethanol (99.99%) medium and then dried in an oven at 120°C for 24 h.All mixed powders were calcined at 900°C for 2 h.Calcined powders were weighted and pressed into pellets (~15 mm in diameter) under 100 MPa pressure by a uniaxial pressing machine without any binder used.Green samples were sintered in a closed alumina crucible (to minimize the loss of volatile elements) at 1125°C for 2 h with a heating/cooling rate of 5°C/min.X-ray diffractometer (XRD-Phillip, X-pert) using CuK α radiation was employed to identify crystal structure of prepared ceramics.Lattice parameters of ceramics were determined from the d spacings obtained from diffraction peaks of a tetragonal phase [16].Tetragonality (c/a) values of all ceramics were extracted using least square analysis.Bulk densities of sintered ceramics were determined using the Archimedes' method.Theoretical densities of all ceramics were calculated based on the reported densities values of BNT (5.99 g/cm 3 ) [51], BT (6.01 g/cm 3 ) [52] and BST (3.97 g/cm 3 ) [53].Linear shrinkages of all ceramics were also evaluated.A scanning electron microscope (SEM, JEOL JSM-6335F) was employed to reveal microstructural features of all samples.Their grain sizes were estimated by applying a mean linear interception method (the quantitative measurement of grain size within densely uniform structure [15]).All ceramic surfaces were carefully milled down to 1 mm thickness to achieved parallel scratch-free surfaces.Silver paste electrodes were made onto both parallel polished surfaces of each sample prior to firing at 600°C for 1 h in air.Dielectric constant (ε r ) and dielectric loss (tanδ) as a function of temperature (50-350°C) of ceramics under frequencies of 1, 10 and 100 kHz were measured by a high precision LCR meter (LCR 821, GW INSTEK).Polarization-electric field (p-E) hysteresis loops at room temperature (25°C) and at a frequency of 1 Hz were measured by a Radiant Precision ferroelectric tester under an applied electric field of 50 kV/cm.Room temperatures strain-electric field (S-E) behavior was measured by an optical displacement sensor (Fotonic Sensor model MTI-2100) under a frequency of 0.1 Hz.Maximum strain (S max ) and negative strain (S neg ) values were obtained from a bipolar curve at the maximum electric field (E max ) of 50 kV/cm.The normalized strain coefficient (d 33 * ) of all samples was also calculated.
Finally, all samples were poled in silicone oil bath at 55°C under an applied DC electric field of 40 kV/cm for 15 min.A piezoelectric coefficient (d 33 ) value of each sample was recorded from 1 day-aged samples using a d 33 meter (KCF technologies, S5865).Piezoelectric voltage constant (g 33 ) and off-resonance figure of merit (FoM) for energy harvesting of ceramics were also evaluated in this study.

Results and discussion
Based on true densities of ceramics, their relative densities were calculated with respect to the theoretical densities of BNT, BT and BST [51][52][53].After sintering at 1125°C for 2 h, well-sintered and dense ceramics with relative density values ~ 97-98% can be achieved in all compositions.As demonstrated in Table 1, it should be noted that some correlation was found between densification (i.e. the density and the linear shrinkage values) and the amount of BST addition.Generally, with increasing BST content to x = 0.04, better densification was observed.This finding was attributed to the promotion of atomic diffusion via the expense of BST.However, a fall-off in density was observed with further increasing BST content to x = 0.05.It was believed that due to the addition of higher BST content, a grain coarsening process of the ceramic was more promoted than the densification process as will be shown in the microstructure measurement part.
In order to identify phase formation of all ceramic compositions, X-ray diffraction technique was employed.XRD patterns of BNT-BT modified with various amount of BST are shown in Figure 1.In general, all samples mostly exhibited identical XRD patterns (Figure 1(a)) indicating a single phase of perovskite structure, which was consistent with literature [54][55][56][57].No diffraction peak of any impurity was detected in all patterns, reflecting a complete reaction between the starting oxide powders to form the desired products.This also implied that perovskite BST additive could be fully incorporated into the perovskite BNT-BT lattice, resulting in a formation of solidsolution ceramics.This was similar to the observed results in Refs [15,16,56].As shown in Figure 1(b), the position of (111) peak at 2θ = 38-42° gradually shifted to lower 2θ angles with increasing BST content (x), implying greater tetragonality (c/a) factors of ceramics.The observed result supported mixed rhombohedraltetragonal phase which the prediction for dominant tetragonal lattice distortion of BNT-BT system modified with tetragonal BST phase which observed in similar previous reported of BNKT-BST [15].This observation became clearer with the increasing trend in c/a values as listed in Table 1.This indicated that all BST addition could induce the lattice anisotropy of the BNT-BT system which was in analogous with those observed in BNKT-BST system [15].Similar XRD peak shifting by partial substitution of Ba 2+ for BNKT ceramics was also found by Chen et al [58].Enlarged views of (200) peak splitting at 2θ ~45-48 ° were shown in Figure 1(c).The peak split into two peaks of a tetragonal structure which were (002) and (200) planes at 2θ ~45.887° and ~ 46.559°, respectively.The change in crystal structure was mainly attributed to the difference in ionic radii of Bi 3+ (~1.17 Å), Na 2+ (~1.18 Å), Ti 4+ (~0.61 Å), Ba 2+ (~1.42 Å) and Sr 2+ (~1.26 Å) [59] as well as the influence of defects e.g.vacancies and/or interstitials mainly created through charge compensation mechanism.This finding was also in agreement with Jaita et al. [15] and Chen et al [58].
The full width at half maximum (FWHM) values were used for absolute XRD peak for all samples [60][61][62].In the present work, the FWHM value was calculated from the main peak refection of (110) as listed in Table 1.The FWHM value decreased with increasing of BST content (x), suggesting that increase in the degree of crystallinity of the samples.The crystallite size (D) was used to confirm the degree of crystallinity of all samples.The calculation was performed based on the Debye -Scherrer formula as Equation 1; Table 1.Physical properties of (1-x)Bi 0.5 Na 0.5 TiO 3 -0.06BaTiO 3 -x(Ba 0.7 Sr 0.3 )TiO 3 ceramics.where κ is the fixed number 0.9, λ is X-ray wavelength taken to be 0.15418 nm, B is the full width at half maximum (FWHM) of (hkl) reflection and θ is Bragg angle corresponding to (hkl) reflection [63][64][65][66][67][68][69][70].
The crystallite size decreased when compared with undoped sample and increased from 10.08 nm (x = 0.01) to the maximum value of 14.45 nm (x = 0.05) as presented in Table 1.The increase in crystallite size could affect the electrical properties of the ceramic samples [68].Plots of FWHM versus BST content (x) and insert presenting a relation between crystallite size (blue line) and average grain size (pink line) as a function of BST content (x) in Figure 1(d).
The Raman spectroscopy is an effective technique to determine the functional groups in solid solutions, which provides better insight into the structural study [71].Raman spectra of all compositions after sintering at 1125°C for 2 h are shown in Figure 2. In previously reported data of BNT-BT-based system, there are four main regions that could be detected in the spectrum at room temperature including the mode located below 200 cm −1 associated with the vibration of perovskite A-site, the ~300 cm −1 modes related to Ti-O vibrations, the modes ~ 450-650 cm −1 related to TiO 6 -octahedra vibration, and the modes >700 cm −1 related to the A 1 (longitudinal optical) and E (longitudinal optical) overlapping bands [36,[72][73][74].The Raman spectra of all compositions showed broad peaks due to the distorted octahedral [BiO 6 ] and [NaO 6 ] clusters or disorder in A-site of rhombohedral structure and the overlapping of Raman modes due to the lattice anharmonicity which can be observed in (Ba 0.5 Na 0.5 )TiO 3 -xBaTiO 3 based system [70][71][72][73][74][75].The increase of BST content indicated that the splitting of these peaks which was observed in the regions around 279 cm −1 and 600 cm −1 .This was due to the Ba 2+ ions in the structure increase the average ionic radius size in the lattice of the ceramics [76,77].The increase of BST concentration indicated the splitting of these peaks due to the phase transformation the mixed rhombohedral-tetragonal (for BNT-BT sample) to the tetragonal phase (for BSTdoped sample) which corresponded XRD result and observed in previous report of BaTiO 3 -modified Bi 0.5 (Na 0.80 K 0.20 ) 0.5 TiO 3 system [78].
Representative microstructures of all compositions are given in Figure 3.As expected, typical perovskite BNT-based microstructures with dense granular grain packing derived from a solid-state sintering mechanism are revealed, similar to other reports [15,16,56].It is also noted that heterogeneous microstructures consisting mainly of two ranges of grains (in respect of size and shape) were observed in all compositions, it is likely that since firing temperatures for achieving highest densification of BT (~1400°C) [79] and BNT (~1100°C) [80] are very different, this could result in grain growth behavior variation between these two phases, hence, heterogeneous microstructures, consistent with Chaisan et al [79].As data given in Table 1, the average grain size range of all ceramics is of the order of ~ 0.75-1.57µm.As x = 0.01 the average grain size range decreased when compared with unmodified BNT-BT ceramic.This was expected to be caused by the change in crystal structure as well as the solute drag mechanism due to the different diffusivity values of different species of added solutes (i.e.particularly Sr 2+ ion).However, when x > 0.01, the grain size gradually increased in the BST-doped sample up to x = 0.05.It was likely that the increase in tetragonality and increasing Sr 2+ concentration contributed to a slight increase in grain growth rate with no significant effect on grain shape.This increase in grain size trend was similarly observed in BST-doped BNKT ceramics [15].The shape of the grains was characterized by many faces (i.e.equiaxed grain morphology).This is probably because the BST phase could act as the grain growth  promoter especially for the more retarding one during atomic diffusion mechanism [81].The observation that the addition of BST results in semi-isotropic grain shape (i.e.cubic-like shape) of the perovskite BNT-BTbased ceramics is also consistent with other similar works [51,82].
Room-temperature dielectric constant (ε r ) and dielectric loss (tanδ) values of all compositions monitored at frequency of 1 kHz are listed in Table 2.In this work, the unmodified BNT-BT ceramics displayed ε r of 1663 which was close to the values reported in literature [15,16,55].When x = 0.01 of BST was added into BNT-BT ceramics, the values of ε r and tanδ increased.ε r drastically decreased from 1981 to 1095 and tanδ slightly decreased from 0.0820 to 0.0755.This observation could be attributed to the reduction of an internal stress within the larger grains via the formation of 90° domain mechanism [1,9].However, it can be seen that with further increasing BST concentration to x = 0.04, both values of ε r and tanδ slightly increased.Clearly, the increasing ε r is due mainly to the increased density and crystal structure change of the samples, in good agreement with other works [15,79].Another possible explanation is that the increasing grain size resulting in decreasing the volume fraction of grain boundaries (i.e. the sink of defects), thus more fraction of dielectric materials is predominant [79][80][81].It should be mentioned that ε r and tanδ values decreased considerably with increasing BST greater than x = 0.04 and could be attributed to both grain size and density reduction.It appears that a critical chemical modification level of the BST modified BNT-BT ceramics is x = 0.04, where the maximum value of the room temperature ε r of 1981 has been detected.Thus, in this study, it could be stated that densification and microstructure (i.e., grain size) of the ceramics, which are influenced by the addition of BST, are also responsible for the room temperature dielectric behavior of the ceramics.Although the correlation between grain size and electrical properties of perovskite ferroelectric ceramics have been widely explored for several decades.There are still controversies regarding the dependence of the piezoelectric and ferroelectric properties on the grain size.The temperature dependence (30-400°C) of dielectric constant (ε r ) and dielectric loss (tanδ) of all compositions at various frequencies (1 kHz, 10 kHz, 100 kHz and 500 kHz) are shown in Figure 4.The dielectric constant value of all composition presented highest value at frequency of 1 kHz.This is due to active presence of polarizations types (i.e.space charge, electronic, ionic).The orientational polarizations cannot follow higher applied frequencies [83].Two dielectric peaks have been observed in all compositions.The peak at lower temperature is the depolarization temperature (T d ) and that as higher temperature is dielectric maximum temperature (T m ) the BNT-based ceramics [53,84].The maximum dielectric constant (ε max ) increased with increasing of BST composition Table 2. Electrical properties of (1-x)[0.94Bi0.5 Na 0.5 TiO 3 -0.06BaTiO 3 ]-x(Ba 0.7 Sr 0.3 )TiO 3 ceramics.and indicated the optimum value of 5850 at x = 0.04 while the value dropped at x = 0.05 with the dielectric loss range 0.0652-0.0877.The increase of BST content indicated that broadness of dielectric peak around T d near 120°C observed in x = 0.03 and 0.04 suggesting a diffusive phase transition to relaxor-like behavior of the ceramic.This was due to compositional fluctuations occurring at A-and B-site of the relevant perovskite unit cell [85].The T d showed a slight decreasing trend with increasing of BST composition.However, T m tended to increase with increasing BST composition as presented in Figure 5(a).This may also be associated with the possible presence of nonpolar region in the BST doped samples [84].
The room temperature polarization electric field (p-E) hysteresis loops investigation of all compositions, measured as a function of BST addition under electric field of 50 kV/cm and frequency of 1 Hz, are shown in Figure 6.The unmodified BNT-BT ceramics exhibited a typical ferroelectric characteristic of well-saturated  hysteresis with square shape, indicating strong ferroelectric order in agreement with related works [15,79,80].As listed in Table 2, it had maximum values of P r = 25.11µC/cm 2 , E c = 26.61kV/cm.By comparison with the unmodified composition, the addition of BST (1-5 mol/%) displayed a significant effect on p-E hysteresis loops by causing narrower loops, indicating softer ferroelectric response, matched with the reduction of ε r , P r and E c values.This suggested the weakening of ferroelectric order.X-ray diffraction result indicated that the lattice distortion due to tetragonality change and lattice expansion from peak shift in BST-doped samples.The ferroelectric domains were ordered while applying the electric field, and nanodomains were uniformly aligned due to lattice expansion.The resulting smaller ferroelectric loops in x = 0.01 and 0.02 samples indicated the softer ferroelectric behavior and became more relaxor-like [86].This can be explained by the chemical fluctuation within the BNT-BT system induced via the addition of BST leading to the more randomized ferroelectric domain distribution and then, resulting in the formation of defect dipoles and possibly the polar nanoregions as suggested earlier by Jaita et al. [15] and Hussain et al [85].It should be that further increasing BST content, P max values slightly increased continuously but the p-E loops still maintain their constricted shapes (pinching), indicating dielectric softening associated with micro-polar behavior via the disordering ferroelectric mechanism [1,15].The pinched loops observed for x = 0.03-0.05samples could be due to the pinning of domain walls interacting with defects or as a result of the presence of intermediate modulated phases with inhomogeneous dipolar regions leading to the coexistence of both ferroelectric and antiferroelectric orders [87].Hence, improved understanding of their defect formation and order-disorder ferroelectric mechanism will help to guide both the development of novel Pb-free piezoceramics and the improvement of device performance for future utilization.
The corresponding strain "butterfly" or S-E loops is shown in Figure 7.All compositions showed the butterfly-shaped strain hysteresis loop [88][89][90].The maximum strain (S max ), the negative strain (S neg ) and the normalized strain coefficient (d 33 * ) were also calculated using Equation 2 as listed in Table 2.
The undoped sample has the S max of 0. The piezoelectric coefficient (d 33 ) of all ceramics is also summarized in Table 2.The values of all ceramics indicated a range of 160-277 pC/N.The unmodified BNT-BT ceramic had d 33 of 178 pC/N which was similar to the value reported by Jaita et al. [15].It is interesting to note that the addition of BST more than x = 0.03 can effectively enhance the d 33 values of BNT-BT ceramics.The optimum of d 33 was found that at the composition of x = 0.04 mol fraction with the value of 277 pC/N and dropped to 203 pC/N at composition of x = 0.05 mol fraction.Apart from their highest density observed, as already mentioned, the tetragonality enhancement of the BNT-BT ceramics modified with BST can explain better dielectric and piezoelectric properties of these ceramics.The energy harvesting properties of all ceramics were investigated by calculation.The piezoelectric voltage constant (g 33 ) is one of the important parameters for energy harvesting applications.The piezoelectric voltage constant (g 33 ) was calculated using Equation 3where ε r is the dielectric constant of the piezoelectric material, ε 0 is the dielectric constant in a vacuum = 8.854 × 10 −12 Fm −1 [91,92] and d 33 is the piezoelectric charge constant.The addition of BST of x = 0.01 caused the g 33 to slightly drop and then increased with increasing BST composition to the optimum value of 20.93 Χ 10 −3 V/mN at x = 0.05.The energy harvesting performance figure of merit can be calculated from the off-resonance figure of merit (FoM) as Equation ( 4) where d 33 is the piezoelectric charge constant, and g 33 is the piezoelectric voltage constant, respectively [93][94][95][96].As a result, the FoM value of undoped sample was 2.29 pm 2 /N.The addition of BST into BNT-BT increased the FOM value from 1.70 (x = 0.01) − 4.37 pm 2 /N (x = 0.04) with increasing of BST composition and dropped at x = 0.04 sample.Therefore, this showed that the figure of merit of the BNT-BT ceramics was significantly improved by the addition of BST.The best composition of x = 0.04 with the highest d 33 of 277 pC/N and high g 33 value of 15.79 × 10 −3 Vm/N.The relation between g 33 and FoM value is shown in Figure 5(C).The comparison of the FoM value in this work with other previous reported [44,89,94], [97]; [98] indicates that higher value.

Conclusions
Lead-free piezoelectric ceramics of (1-x)Bi 0.5 Na 0.5 TiO 3 -0.06BaTiO 3 -x(Ba 0.7 Sr 0.3 )TiO 3 with x = 0.00, 0.01, 0.02, 0.03, 0.04 and 0.05 mol fraction were successfully prepared by using a conventional mixed-oxide technique.All compositions displayed a single phase of perovskite structure with larger tetragonality obtained at higher BST content.A slight grain size increase with more densification and heterogeneous microstructures were noticeable for the ceramics modified with BST additives.The highest piezoelectric coefficient (d 33 = 277 pC/N) with good ferroelectric properties and energy harvesting properties were found for the ceramics with the addition of x = 0.04.This composition could be interesting for transducer, actuator or energy harvesting applications.

Figure 5 .
Figure 5. (a) Plots of T m and T d as a function of BST composition, (b) plot of S max and d 33 * as a function of BST composition and (c) plot of g 33 and FoM as a function of BST composition of (1-x)[0.94Bi0.5 Na 0.5 TiO 3 -0.06BaTiO 3 ]-x(Ba 0.7 Sr 0.3 )TiO 3 ceramics.
10% and d 33 * of 196 pm/V.With increasing BST concentration, the d 33 * and S max tended to increase with the increasing of BST composition and promoted the maximum value at x = 0.04 with S max = 0.26% and d 33 * = 519 pm/V.After that the value decreased at x = 0.05.Furthermore, when the added BST concentration increased, the strain curves showed drastic deviation from typical ferroelectric behavior, which was evidenced by the absence of the negative strain.This behavior was observed in other BNT-based systems [70,85,88].Plot of S max and d 33 * as a function of BST composition are shown in Figure 5(b).