A. Phase structure analysis of Sr_2-xCa_xNaNb₅O₁₅

Figure 1(a) shows the XRD patterns of the Sr_2-xCa_xNaNb₅O₁₅ ceramics with different Ca²⁺ contents. All ceramics are of single TB structure and no second phase could be detected, indicating that Ca²⁺ may diffuse into the TB structure lattice to form a solid solution, in which Ca²⁺ occupies A sites in the TB structure. Figure 1(b) is the magnification of Fig. 1(a) in the range from 45 to 48°. Two peaks split at about 47° are (620) and (002) peaks, respectively. When x = 0.05, the intensity of (620) peak is slightly higher than that of (002) peak. With increasing x, the intensity of (002) peak enhances gradually and then exceeds the (620) peak at x = 0.30, suggesting that some distortion of crystal structure appears when substituting Sr²⁺ with Ca²⁺. Annie SimonReference Simon and Ravez¹⁵ states that for TB structure crystals, there is a permanent anisotropy, even in centrosymmetric phases, it is not thus possible to reach the cubic symmetry, and the TB structure crystals usually have tetragonal symmetry or orthorhombic symmetry. To further study the structure distortion caused by Ca²⁺ substitution, the lattice parameters of Sr_2-xCa_xNaNb₅O₁₅ ceramics with different Ca²⁺ contents are calculated and shown in Table I. As shown in Table I, when x = 0.05, a = 12.360 Å, b = 12.354 Å and c = 3.895 Å. The difference between a and b is so negligible that the ceramics with x = 0.05 are considered to be tetragonal phase. With increasing x from 0.05 to 0.30, the value of ▵ (a-b) gradually increases from 0.006 to 0.063, indicating that the crystal structure slightly distorts from the tetragonal phase and becomes the orthorhombic phase at room temperature. This is in good agreement with the SCNN single crystals reported by Neurgaonkar et al.Reference Neurgaonkar, Oliver, Copy, Cross and Viehland⁶ In addition, the lattice parameter c and the lattice unit volume V decrease monotonically with increasing the Ca²⁺ content. The decrease of the tetragonality and the distortion of the crystal structure are all induced by substituting Sr²⁺ with Ca²⁺. The shrinkage of the unit cell can be attributed to the smaller ionic radius of Ca²⁺ (1.34Å) compared with that of Sr²⁺ (1.44Å).

FIG. 1.

XRD patterns of Sr_2-xCa_xNaNb₅O₁₅ ceramics with different x. (x = 0.05, 0.15, 0.20, 0.25, and 0.30).

TABLE 1.

The lattice parameters of obtained Sr_2-xCa_xNaNb₅O₁₅ ceramics as a function of x.

B. Microstructure analysis of Sr_2-xCa_xNaNb₅O₁₅

Figure 2 shows SEM photographs of the surface of the Sr_2-xCa_xNaNb₅O₁₅ ceramics with different Ca²⁺ contents. All ceramics show a common microstructure that is relatively dense and uniform. With increasing x from 0.05 to 0.15, the morphology becomes more homogeneous, and the number of grains with anisotropic shape decreases, leading to decrease in the average grain size. Fine-grained and relatively uniform microstructure is obtained at x = 0.15, which is very beneficial to the practical application as it is known that the fine-grained and homogenous microstructure can lead to high mechanical strength for the piezoelectric ceramics. With further increase of x to 0.25 and 0.30, the average grain size then increases slightly. The microstructure results are not consistent with those of SCNN ceramics reported by Xie et al.,Reference Xie and Akimune^10–Reference Xie, Akimune, Wang and Hirosaki¹² in which Ca²⁺ substitution had no appreciable influence on the microstructure of Sr_2-xCa_xNaNb₅O₁₅ ceramics prepared by the spark plasma sintering method. In addition, abnormal grain growth (>50 μm), which are frequently encountered in the sintering of TB compounds ceramics,Reference Kimura, Miyamoto and Yamaguchi^7–Reference Wei, Yang, Chang and Gu⁹ is not observed in any of the specimens, indicating that through careful control of the preparation process, the two-step solid-state reaction method and traditional sintering process are also effective in obtaining the Sr_2-xCa_xNaNb₅O₁₅ ceramics without abnormal grain growth and cracks. The above results show that proper amount of Ca²⁺ substitution is beneficial in suppressing the grain growth and avoiding the abnormal grain growth behavior.

FIG. 2.

Scanning electron microscopy photographs of the surface of Sr_2-xCa_xNaNb₅O₁₅ ceramics with different x: (a) x = 0.05, (b) x = 0.10, (c) x = 0.15, (d) x = 0.20, (e) x = 0.25, and (f) x = 0.30.

Figure 3 shows the relative density of the Sr_2-xCa_xNaNb₅O₁₅ ceramics with different values of x. The relative density is calculated based on the theoretical density for each individual composition, which is calculated from the unit cell volumes according to XRD results (5.027∼4.999 g/cm³). It can be seen from Fig. 3 that for all compositions, the relative density of the obtained ceramics is higher than 0.970, indicating that the two-step solid-state reaction method and traditional sintering process can also be used to obtain high-density Sr_2-xCa_xNaNb₅O₁₅ ceramics with TB structure. In addition, with increasing x, the relative density increases gradually, reaches the maximum value of about 0.991 at x = 0.15, and then decreases. The variation of relative density in Fig. 3 is consistent with the SEM result in Fig. 2. The differences between the microstructural and stoichiometric findings/ characterizations are likely to have substantial influence on the electrical properties. Details of dielectric properties are discussed in the following sections.

FIG. 3.

Relative density of the Sr_2-xCa_xNaNb₅O₁₅ ceramics with different x.

C. Dielectric properties of Sr_2-xCa_xNaNb₅O₁₅ ceramics

The temperature dependence of dielectric constant ε and dielectric loss tan δ for the ceramics with different Ca²⁺ contents measured from −60 to 350 °C at 1 kHz is shown in Fig. 4. It can be seen that for the ceramics with x ≤ 0.25, two obvious dielectric peaks can be observed, corresponding to the ferroelastic orthorhombic (mm2) to ferroelectric tetragonal (4 mm) phase transition at lower temperatures and the ferroelectric tetragonal (4 mm) to tetragonal paraelectric (4/mmm) phase transition at higher temperatures, respectively. The temperatures at which the two-phase transitions occur are abbreviated as T_e and T_c, respectively. T_e corresponds to the ferroelastic orthorhombic (mm2) to ferroelectric tetragonal (4 mm) phase transition temperature, and T_c corresponds to the ferroelectric tetragonal (4 mm) to tetragonal paraelectric (4/mmm) phase transition temperature. For the ceramics with x = 0.30, the dielectric peak related to the ferroelectric tetragonal (4 mm) to tetragonal paraelectric (4/mmm) phase transition becomes indistinct. The results are in good agreement with the literature.Reference Xie and Akimune¹⁰ For each composition, the temperature dependence of dielectric loss tan δ curve shows similar shape with that of dielectric constant ε. There are also two tan δ peaks observed for all compositions, and the tan δ peaks around T_c are more obvious than the tan δ peaks below 0 °C.

FIG. 4.

Temperature dependence of dielectric constant ε and dielectric loss tan δ for Sr_2-xCa_xNaNb₅O₁₅ ceramics measured at 1 kHz as a function of x.

ε_me and ε_m are the maximum values of the dielectric constant obtained at T_e and T_c, respectively. T_e, ε_me, T_c and ε_m of the ceramics measured at 1 kHz as a function of x are presented in Fig. 5. They all have great dependence on x. By increasing x, T_e increases gradually from −10 to 36 °C. T_e is −10 °C for the ceramics with x = 0.05, so the ceramics are tetragonal phase (4 mm) at room temperature. For the ceramics with x = 0.15, T_e is about 15 °C, which is near the room temperature, indicating that the ceramics are mixed tetragonal and orthorhombic structures. Further increasing x to 0.30, T_e increases to about 36 °C, which is above room temperature, so the ceramics are orthorhombic phase (mm2) at room temperature. Therefore it is reasonable that the tetragonality of the Sr_2-xCa_xNaNb₅O₁₅ decreases while the orthorhombicity increases when substituting Sr²⁺ with Ca²⁺. And the results are in good agreement with those of Fig. 1 and Table I. T_c also shifts to higher temperature regions gradually with increasing x. T _c is 285 °C at x = 0.05 and increases to 306 °C at x = 0.30. The above results show that Ca²⁺ substitution is beneficial to the crystal structure stability, enabling attainment of higher phase transition temperatures. In addition, ε_me and ε_m show similar trend with increasing x. They all increase at first and then decrease as x increases from 0.05 to 0.30. The maximum values of ε_me and ε_m are obtained at x = 0.15 (ε _me = 1650, ε _m = 1954).

FIG. 5.

T_{e ,}ε_me, T_c and ε_m of the Sr_2-xCa_xNaNb₅O₁₅ ceramics measured at 1 kHz as a function of x.

The diffuseness of the phase transition can also be explained by the modified Curie-Weiss law: 1/ε-1/ε _m = (T-Tm)^γ/C, where ε _m is the maximum value of the dielectric constant at the phase transition temperature. γ and C are assumed to be constant. γ is the degree of diffuseness, and C is the Curie-like constant. The degree of diffuseness γ can have a value ranging from 1 for a normal ferroelectric to 2 for an ideal relaxor ferroelectric. The plots of ln [(ε _m/ε)-1] versus ln (T-T _m) for the Sr_2-xCa_xNaNb₅O₁₅ ceramics (x = 0.05, 0.10, 0.15, 0.20, 0.25, and 0.30) at 1 kHz are shown in Fig. 6. A linear relationship is observed in all specimens. The slope of the fitting curves is used to determine the γ value. It can be seen that the γ value is between 1 and 2 for the compositions with x < 0.25, which indicates that the ceramics with x < 0.25 show intermediate relaxor-like behavior between normal and ideal relaxor ferroelectrics. When x = 0.30, the γ value is 2.11, which is higher than 2, indicating that the ceramics with 0.30 are not ferroelectrics.

FIG. 6.

Ln [(ε _m/ε)-1] as a function of Ln (T-T _m) for the Sr_2-xCa_xNaNb₅O₁₅ ceramics with different Ca²⁺ contents.

D. Ferroelectric and piezoelectric properties of Sr_2-xCa_xNaNb₅O₁₅ ceramics

Figure 7 shows the P-E hysteresis loops of all the compositions measured under the same maximum drive voltage of 60 kv·cm⁻¹. According to Fig. 7, the P-E hysteresis loops become much saturated at first and then become slimmer with increasing x from 0.05 to 0.30. Figure 8 illustrates the P_r, E_c (measured under the maximum drive voltage of 60 kv·cm⁻¹) and piezoelectric constant d ₃₃ (poled in silicone oil at 80 °C under 30 kv·cm⁻¹) of the ceramics as a function of x. P_r, E_c and d ₃₃ all increase at first and then decrease with increasing Ca²⁺ substitution level. The ceramics with x = 0.15 possess the largest P_r (7.08 μC cm^-2) and d ₃₃ (68 pC N^-1).

FIG. 7.

P–E hysteresis loops of the Sr_2-xCa_xNaNb₅O₁₅ ceramics with different Ca²⁺ contents measured under the maximum drive voltage of 50 kv·cm⁻¹ and at room temperature.

FIG. 8.

P_r, E_c and d ₃₃ of the Sr_2-xCa_xNaNb₅O₁₅ ceramics as a function of x.

According to above results, the ceramics with best properties can be obtained at x = 0.15, which is in concordance with the result of SCNN ceramics prepared by the spark plasma sintering method,Reference Xie and Akimune^10–Reference Xie, Akimune, Wang and Hirosaki¹² but different with the results of SCNN single crystals.Reference Neurgaonkar, Oliver, Copy, Cross and Viehland⁶ This is reasonable because the purity of raw materials is higher when fabricating the single crystals. In addition, except d _33, the electrical properties of the obtained ceramics with x = 0.15 (T_c = 285 °C, ε_m = 1954, and P_r = 7.08 μC cm^-2) are equivalent or higher than those of the ceramics prepared by the spark plasma sintering method (T_c = 286 °C, ε_m = 1099, and P_r = 3.0 μC cm^-2). This can be attributed to the changes of crystal structure, and the most homogeneous and densitified microstructure of the ceramics with x = 0.15. When further increasing the Ca²⁺ substitution, the ferroelectric tetragonality of the Sr_2-xCa_xNaNb₅O₁₅ decreases while the ferroelastic orthorhombicity increases, leading to the deteriorating of the electrical properties.