3.1 Piezoelectric properties
The relative densities, εT33/εo, tan δ, d33, and kp values of the KNNS-(0.04-x)SZ-xBAZ piezoceramics (0.0 ≤ x ≤ 0.04) are shown in Fig. 1(a), and these samples were sintered at 1090°C–1100°C for 3 h. All the samples have comparatively large relative densities (≥ 94.5% of the theoretical density). Hence, all the specimens were well densified. The εT33/εo of the piezoceramic (x = 0.0) was approximately 1864, and its variation was not significant when x was less than 0.03. However, a very large εT33/εo of approximately 3836 was observed for the sample (x = 0.03) because the TT−O−R was developed at approximately RT for this sample, which will be discussed later. The sample (x = 0.04) exhibited similar results. The tan δ of the sample (x = 0.0) was approximately 4.5%, and it slightly decreased with an increase in x. The d33 of the piezoceramic (x = 0.0) was 360 pC/N and increased with increasing x, and the largest d33 of 650 pC/N was observed in the piezoceramic (x = 0.03). However, it decreased to 523 pC/N for the specimen (x = 0.04). Figure 1(b) shows the d33 values for KNN-related lead-free ceramics that have been reported in the literature. The largest d33 value of 650 ± 20 pC/N was observed in the (K, Na)(Nb, Sb)O3-(Bi, Na, K)ZrO3-Fe2O3-AgSbO3 piezoceramics [36]. Therefore, the d33 value observed from the sample (x = 0.03) is almost the same as the largest d33 value reported in the literature. The kp of the sample (x = 0.0) was approximately 0.44, which increased with increasing x, and the largest kp of 0.52 was observed from the sample (x = 0.02). The sample (x = 0.03) exhibited a similar kp of 0.51. The kp value is proportional to dij2/εT33. Therefore, the large kp of the piezoceramic (x = 0.02) is ascribed to its small εT33 and relatively large d33, and the large kp of the sample (x = 0.03) is attributed to its large d33 value. The d33 of the sample (x = 0.03) was measured at various temperatures to investigate the temperature dependence of the piezoelectricity of this sample (Fig. 1(c)). The d33 slightly decreased with an increase in the measuring temperature, but a relatively high d33 (500 pC/N) was observed at 120°C, implying that this sample preserved comparatively high piezoelectricity up to 120°C. Figure 1(d) shows the variation in the d33 value with respect to the poling temperature for the sample with x = 0.03. The sample poled at 20°C displayed the largest d33 of 650 pC/N, and a similar result was observed for the sample poled at 30°C. However, it decreased when the poling temperature exceeded 30°C. It has been reported that KNN-based specimens show better piezoelectric properties when they are poled near the phase transition temperature because many structures coexist at the phase transition temperature [54–56]. The piezoceramic (x = 0.03) exhibited the largest d33 value of 650 pC/N when it was poled at 20°C–30°C because this sample has a TT−O−R near RT, which will be shown later. In addition, the polarization–electric field (P–E) hysteresis curves of these piezoceramics were also measured (Figs. S1(a)–(f)]. The piezoceramic (x = 0.03) shows a normal P-E curve and has a comparatively high saturated polarization (PS) of 21.6 µC/cm2, a remnant polarization (Pr) of 16.3 µC/cm2, and a coercive electric field (EC) of 0.65 kV/mm (Fig. S1(d)), indicating that it has good ferroelectric properties. Moreover, other piezoceramics also exhibited similar ferroelectric properties (Fig. S1(f)).
3.2 Structural properties
Figures 2(a)–(d) show the SEM images of the thermally etched surface of the KNNS-(0.04-x)SZ-xBAZ piezoceramics (0.0 ≤ x ≤ 0.04). All the samples had a dense microstructure, resulting in a large relative density (Fig. 1(a)). The sample (x = 0.0) has a microstructure consisting of two types of grains: large grains with an average grain size of approximately 20 µm and small grains with an average grain size of 1 µm, as shown in the inset of Fig. 2(a). As BAZ was added, the samples began to exhibit large grains without small grains (Figs. 2(b)–(d)), suggesting that the BAZ assisted the grain growth of the samples. Grain size did not change with an increase in BAZ content, and the average grain size of the samples with 0.01 ≤ x ≤ 0.04 was 23–25 µm. SEM images were also obtained from the fractured surfaces of the samples, and they showed similar results (Figs. S2(a)-(e)). It has been generally accepted that samples with large grains exhibit better piezoelectric properties [57–59]. Hence, these results suggest that the addition of BAZ can enhance the piezoelectric characteristics of the samples by increasing their grain size.
The piezoelectric properties of the KNN-based ceramics were considerably influenced by the crystal structure of the samples. Piezoceramics generally exhibit large piezoelectric properties when they have a T-O-R (or R-T) multi-structure. Hence, it is important to clarify the crystal structure of the KNNS-(0.04-x)SZ-xBAZ ceramics with 0.0 ≤ x ≤ 0.04. According to the XRD patterns, all the samples have a homogeneous perovskite phase, without a secondary phase (Figs. S3(a)–(e)). However, it is not possible to determine the crystal structure of these samples using the normal XRD pattern. Hence, the XRD peaks at 66.5°, which were obtained by the slow-speed scanning method, were deconvoluted using the Voigt function to investigate the crystal structure of the samples (0.0 ≤ x ≤ 0.04), as displayed in Figs. 3(a)–(e). The sample with x = 0.0 shows rhombohedral (220)R and (2–20)R peaks, strong orthorhombic (004)O, (400)O, and (222)O peaks, and weak tetragonal (202)T and (220)T peaks, indicating that this sample has a T-O-R multi-structure. However, the intensity of the orthorhombic peaks was much larger than those of the rhombohedral and tetragonal peaks for the specimen (x = 0.0) (Fig. 3(a)). As the amount of BAZ increased, the intensity of the orthorhombic peaks decreased, and that of the tetragonal structure was enhanced, but the variation in the rhombohedral peak intensity was not significant. For the samples (0.0 ≤ x ≤ 0.02), however, the proportion of the orthorhombic structure was still larger than those of other structures (Figs. 3(a)–(c)). However, each structure had a similar proportion in the sample (x = 0.03), as illustrated in Fig. 3(d). Moreover, it can be suggested that this sample has an ideal T-O-R multi-structure because the rhombohedral, orthorhombic, and tetragonal structures have similar proportions. Finally, the sample (x = 0.04) exhibited an R-T multi-structure, with a large proportion of the tetragonal structure (Fig. 3(e)).
Rietveld refinement was also performed to determine the detailed crystal structure of the piezoceramic (x = 0.03). Various models were used for the Rietveld analysis (Figs. 4(a)–(c)): O-R multi-structure, T-O multi-structure, and T-O-R multi-structure. Table 1 shows the atomic coordinates, site occupancies, R-values, and lattice parameters of these models. The T-O-R multi-structure that consists of 31% R3m rhombohedral, 33% Amm2 orthorhombic, and 36% P4mm tetragonal structures shows the lowest R-value, as shown in Fig. 4(c). The samples with x = 0.0, 0.01, and 0.02 also have a T-O-R multi-structure, but the amount of orthorhombic structure is larger than those of the other structures, as shown in Figs. S4(a)-(e). Therefore, Rietveld analysis confirmed that the piezoceramic (x = 0.03) has an ideal T-O-R multi-structure, in which the three structures have similar proportions. Moreover, it can be suggested that the presence of an ideal T-O-R multi-structure can contribute to the large d33 value of this sample. In addition, the piezoceramic (x = 0.04) has an R-T multi-structure, as shown in Figs. S5(a) and (b), which is identical to the R-T morphotropic phase boundary structure of PZT-based piezoceramics. Hence, this piezoceramic also has a comparatively large d33 of 523 pC/N.
Table 1
Atomic coordinates, site occupancies, lattice parameters, and R-values of the KNNS-0.01SZ-0.03BAZ piezoceramic for O-R, T-O, and T-O-R multi-structure models.
Phase | Structural model (SG) | Site label | x | y | z | Site occupancy | Lattice parameter (Å) | R factor |
O-R multi-structure | Orthorhombic (Amm2) (69%) + Rhombohedral (R3m, H) (31%) | K/Na/Sr /Bi/Ag | 0(-) | 0(-) | 0(-) | 0.114(-)/0.123(-)/0.002(-) /0.004(-)/0.004(-) | a = 3.9894(2) b = 5.6153(3) c = 5.6223(3) α = β = γ = 90° | Rp/Rwp/Rexp Rb/Rf 4.70 / 6.09 / 3.00 5.49 / 2.80 |
Nb/Sb/Zr | 0.5(-) | 0(-) | 0.5250(37) | 0.224(-)/0.016(-)/0.010(-) |
O1 | 0(-) | 0(-) | 0.5205(94) | 0.250(-) |
O2 | 0.5(-) | 0.2249(67) | -0.2105(84) | 0.500(-) |
K/Na/Sr /Bi/Ag | 0(-) | 0(-) | 0.4742(57) | 0.076(-)/0.082(-)/0.002(-) /0.004(-)/0.002(-) | a = b = 5.6289(18) c = 6.8957(42) α = β = 90° γ = 120° | Rp/Rwp/Rexp Rb/Rf 4.70 / 6.09 / 3.00 5.48 / 3.36 |
Nb/Sb/Zr | 0(-) | 0(-) | 0(-) | 0.150(-)/0.010(-)/0.007(-) |
O1 | 0.5057(-) | -0.5057(-) | 0.4805(51) | 0.500(-) |
T-O multi-structure | Tetragonal (P4mm) (44%) + Orthorhombic (Amm2) (56%) | K/Na/Sr /Bi/Ag | 0(-) | 0(-) | -0.0483(31) | 0.057(-)/0.062(-)/0.001(-) /0.002(-)/0.002(-) | a = 3.9768(1) c = 3.9801(1) α = β = γ = 90° | Rp/Rwp/Rexp Rb/Rf 4.26 / 5.76 / 3.01 4.06 / 2.86 |
Nb/Sb/Zr | 0.5(-) | 0.5(-) | 0.5(-) | 0.112(-)/0.008(-)/0.005(-) |
O1 | 0.5(-) | 0.5(-) | 0.0155(55) | 0.125(-) |
O2 | 0.5(-) | 0(-) | 0.4476(63) | 0.250(-) |
K/Na/Sr /Bi/Ag | 0(-) | 0(-) | 0(-) | 0.114(-)/0.123(-)/0.002(-) /0.004(-)/0.004(-) | a = 3.9932(2) b = 5.6134(2) c = 5.6176(2) α = β = γ = 90° | Rp/Rwp/Rexp Rb/Rf 4.26 / 5.76 / 3.01 3.75 / 2.62 |
Nb/Sb/Zr | 0.5(-) | 0(-) | 0.4686(58) | 0.224(-)/0.016(-)/0.010(-) |
O1 | 0(-) | 0(-) | 0.4829(199) | 0.250(-) |
O2 | 0.5(-) | 0.2499(88) | 0.1981(89) | 0.500(-) |
T-O-R structure | Tetragonal (P4mm) (36%) + Orthorhombic (Amm2) (33%) + Rhombohedral (R3m, H) (31%) | K/Na/Sr /Bi/Ag | 0(-) | 0(-) | -0.0680(117) | 0.057(-)/0.062(-)/0.001(-) /0.002(-)/0.002(-) | a = 3.9772(1) c = 3.9805(1) α = β = γ = 90° | Rp/Rwp/Rexp Rb/Rf 4.06 / 5.38 / 3.00 2.75 / 2.06 |
Nb/Sb/Zr | 0.5(-) | 0.5(-) | 0.5(-) | 0.112(-)/0.008(-)/0.005(-) |
O1 | 0.5(-) | 0.5(-) | 0.0218(63) | 0.125(-) |
O2 | 0.5(-) | 0(-) | 0.4741(69) | 0.250(-) |
K/Na/Sr /Bi/Ag | 0(-) | 0(-) | 0(-) | 0.114(-)/0.123(-)/0.002(-) /0.004(-)/0.004(-) | a = 3.9948(2) b = 5.6147(2) c = 5.6189(2) α = β = γ = 90° | Rp/Rwp/Rexp Rb/Rf 4.06 / 5.38 / 3.00 3.13 / 2.85 |
Nb/Sb/Zr | 0.5(-) | 0(-) | 0.5094(29) | 0.224(-)/0.016(-)/0.010(-) |
O1 | 0(-) | 0(-) | 0.5724(-) | 0.250(-) |
O2 | 0.5(-) | 0.2011(130) | 0.2585(-) | 0.500(-) |
K/Na/Sr /Bi/Ag | 0(-) | 0(-) | 0.4467(27) | 0.076(-)/0.082(-)/0.002(-) /0.002(-)/0.002(-) | a = b = 5.6339(11) c = 6.8980(25) α = β = 90° γ = 120° | Rp/Rwp/Rexp Rb/Rf 4.06 / 5.38 / 3.00 3.06 / 2.51 |
Nb/Sb/Zr | 0(-) | 0(-) | 0(-) | 0.150(-)/0.010(-)/0.007(-) |
O | -0.5109(-) | -0.5109(-) | 0.4841(-) | 0.500(-) |
3.3 Temperature dependence of dielectric properties and domain structure
Figures 5(a)–(e) show the εT33/εo versus temperature plots for the KNNS-(0.04-x)SZ-xBAZ ceramics with 0.0 ≤ x ≤ 0.04, ranging between − 60°C and 90°C. Variations in the TT−O and TO−R peaks are shown in these figures. For the sample (x = 0.0), the TT−O and TO−R are approximately 70°C and 0.0°C, respectively, as shown in Fig. 5(a). The TT−O decreases and the TO−R increases with an increase in x, and they meet at approximately 25°C for the sample with x = 0.03, resulting in the formation of TT−O−R at RT. When x exceeded 0.03, the orthorhombic phase disappeared and a broad TR−T was formed at approximately 10°C for the sample with x = 0.04. Therefore, the results of the εT33/εo versus temperature curves are similar to those of the XRD analysis. In addition, the inset of each figure also shows the εT33/εo versus temperature curve of the corresponding sample that includes TC. The TC of the sample (x = 0.0) was approximately 154°C, and it increased with an increase in x to 185°C for the sample (x = 0.04). In particular, the TC of the sample (x = 0.03) was comparatively high at 182°C. Therefore, this sample can maintain a high d33 value of up to 120°C.
The εT33/εo values of the KNNS-(0.04-x)SZ-xBAZ piezoceramics (0.0 ≤ x ≤ 0.04) were measured at different frequencies to investigate the relaxor properties of these samples. Relaxor ceramics generally have nanodomains with a small domain boundary energy [32–34]. Hence, relaxor piezoceramics with nanodomains have been reported to exhibit high piezoelectric characteristics [32–34]. Figure 6(a) shows the εT33/εo values measured at various frequencies for the sample (x = 0.0). A change in TT−O was not observed with an increase in frequency, implying that this sample is a normal ferroelectric ceramic. For the sample (x = 0.03), however, the TT−O−R was enhanced with an increase in the measuring frequency, as shown in Fig. 6(b). The temperature difference (ΔT) between the TT−O−R measured at the lowest frequency and the TT−O−R measured at the highest frequency was approximately 11.3°C for the sample with x = 0.03. Therefore, the sample with x = 0.03 is considered to be a relaxor ceramic, and this ceramic is expected to have ferroelectric nanodomains. The εT33/εo values were also measured at various frequencies for the samples (x = 0.01, 0.02, and 0.04), as shown in Figs. S6(a)-(c). The ΔT value of the specimen (x = 0.01) was 4.5°C, and it increased with the increase in x. Therefore, it is considered that the samples with x = 0.01, 0.02, and 0.04 also exhibit relaxor characteristics. TEM analysis was performed on the piezoceramics (x = 0.0, 0.03) to investigate the domain structure of these samples. Figure 6(c) shows the TEM bright-field image for the sample with x = 0.0. The domain size of this sample was approximately 50 nm × 400 nm, indicating that this sample had relatively large domains. A TEM bright-field image was also observed from the piezoceramic (x = 0.03), as shown in Fig. 6(d). Nanodomains with a size of 2 nm × 15 nm were formed in this sample, possibly because this sample showed relaxor properties. This result suggests that the presence of ferroelectric nanodomains also contributes to the large d33 value of this sample.
3.4 planar-type actuator
A KNNS-0.01SZ-0.03BAZ thick film (x = 0.03) with dimensions of 20 mm × 20 mm × 0.3 mm was fabricated to produce a planar-type piezoelectric actuator. The crystal structure of this thick film was determined using Rietveld analysis of the XRD pattern, as shown in Fig. 7(a), and was identified as a T-O-R multi-structure consisting of R3m rhombohedral (29.7%), Amm2 orthorhombic (32.9%), and P4mm tetragonal (37.4 %) structures. Hence, the crystal structure of this thick film was the same as that of the KNNS-0.01SZ-0.03BAZ piezoceramic. The microstructures of the KNNS-0.01SZ-0.03BAZ thick films sintered at various temperatures were also studied. The thick film sintered at 1080°C shows two types of grains: large and small grains with average grain sizes of 25 µm and 0.5 µm, respectively (Fig. S7(a)). An SEM image of the thick film sintered at 1090°C is displayed in Fig. 7(b), showing a dense microstructure with large grains with an average grain size of 25 µm, which is similar to the microstructure of the KNNS-0.01SZ-0.03BAZ piezoceramic. The thick films densified at 1100°C and 1110°C also displayed an equivalent microstructure (Figs. S7(b) and (c)]. Therefore, it can be concluded that the structural properties of the KNNS-0.01SZ-0.03BAZ thick films sintered at temperatures higher than 1080°C are similar to those of the KNNS-0.01SZ-0.03BAZ piezoceramic.
Figure 7(c) shows the relative densities, εT33/εo, tan δ, d33, and kp values of the KNNS-0.01SZ-0.03BAZ thick films sintered at various temperatures. The thick film sintered at 1080°C showed a low relative density (90% of the theoretical density), but the thick films sintered at temperatures higher than 1080°C exhibited a large relative density (≥ 93% of the theoretical density). Hence, the KNNS-0.01SZ-0.03BAZ thick film must be densified at temperatures higher than 1080°C. The εT33/εo of the thick film sintered at 1080°C is relatively small at 1666, probably because of its low density, and it increased with increasing sintering temperature to 3390 for the thick film sintered at 1110°C. The tan δ of the sample sintered at 1080°C was approximately 3.3%, and it increased slightly with an increase in sintering temperature. The d33 value of the thick film sintered at 1080°C was low (410 pC/N), but increased with an increase in the sintering temperature. The thick film sintered at 1090°C exhibited the largest d33 value, 630 pC/N. This d33 value is slightly smaller than that of the KNNS-0.01SZ-0.03BAZ piezoceramic (650 pC/N), but the difference is not large. The thick films sintered at 1100°C and 1110°C exhibited slightly reduced d33 values of 610 pC/N. The kp value showed an equivalent trend, and a maximum kp value of 0.51 was obtained from the thick film densified at 1090°C. Therefore, the optimum sintering temperature of the KNNS-0.01SZ-0.03BAZ thick film was considered to be 1090°C. This thick film was also sintered at 1090°C for various amounts of time; the thick film sintered for 6.0 h showed the best piezoelectric properties (Fig. S7(d)), and this thick film was used to fabricate planar-type piezoelectric actuators.
The planar-type actuator was fabricated using a KNNS-0.01SZ-0.03BAZ thick film with dimensions of 20 mm × 20 mm × 0.3 mm, as shown in the schematic diagram in Fig. 8(a). The COMSOL program was used to simulate the actuating characteristics. Figure 8(b) shows a simulated image of the KNNS-0.01SZ-0.03BAZ planar-type actuator, along with its strain. The accelerations of this actuator were measured at various frequencies and applied voltages, as shown in Fig. 8(c). The maximum accelerations were obtained at 610 Hz, indicating that 610 Hz is the resonance frequency of the actuator. The maximum acceleration increased with an increase in the applied electric field, and a large acceleration of 335 G was obtained under a comparatively small electric field of 120 V/mm (36 V) at 610 Hz. The maximum accelerations, which were measured at various applied voltages and 610 Hz, are displayed in the inset of Fig. 8(c) as black circles. A similar resonance frequency of 620 Hz was obtained from the simulation (Fig. S8(a)). Accelerations of this actuator were calculated at various applied voltages, and the maximum accelerations, which were calculated at the resonance frequency and various applied voltages, are displayed as red circles in the inset of Fig. 8(c). The calculated accelerations are similar to the measured values. Figure 8(d) shows the change in the displacement as a function of the frequency obtained at various applied electric fields for the KNNS-0.01SZ-0.03BAZ planar-type actuator. The maximum displacements were also obtained at 610 Hz and are shown in the inset of Fig. 8(d) as black circles. The displacement increased with an increase in the electric field, as illustrated by the black circles in the inset of Fig. 8(d). The largest displacement of 231 µm was obtained with the application of a low electric field of 120 V/mm. Displacements were also simulated (Fig. S8(b)), and the maximum displacements at different applied voltages are indicated by red circles in the inset of Fig. 8(d). The simulated displacements were also similar to the measured displacements, implying that the experimental results were in good agreement with the calculated results. The PZT-based planar-type haptic actuator with a circular shape showed a large displacement of 427 µm with an applied electric field of 750 V/mm [60], which is almost six times larger than the electric field used in this study. The displacement of the planar-type KNNS-0.01SZ-0.03BAZ actuator can be further enhanced by increasing the applied electric field, indicating that this actuator is capable of generating a larger displacement than the PZT-based planar-type actuator. Therefore, the KNNS-0.01SZ-0.03BAZ lead-free piezoceramic is a promising material for piezoelectric actuators.