Effect of Sintering Conditions on the Electrical Properties of Lead-Free Piezoelectric Potassium Sodium Niobate-Based Ceramics

Abstract

Lead-free piezoelectric 0.92(K_0.445Na_0.5Li_0.05)NbO₃-0.08BaZrO₃ (KNLN-BZ) ceramics were prepared via conventional sintering. Single-step, 2-step and 3-step temperature-controlled conditions were designed. The structure and electrical properties of ceramics obtained using different temperature-controlled procedures were systematically studied. It was found that ceramic prepared using the 3-step method with a holding time of 20 h showed the highest electrical properties. The Curie temperature was approximately 286 °C, and the dielectric constant and dielectric loss at room temperature were 1350 and 4.5% at the frequency of 1 kHz, respectively. The highest remanent polarization, piezoelectric strain and piezoelectric coefficient, d₃₃*, were obtained 60 °C, indicating a phase transition between ferroelectric phases. Although the ceramics did not show excellent piezoelectric properties, the 3-step sintering method can be considered an effective method to optimize the electrical performances of KNN-based ceramics. Combined with an appropriate composition, ceramics with excellent electrical properties could be obtained. This study provides a path to enhance the density and electrical properties for KNN-based ceramics with simple composition and potential for industry application.

1. Introduction

Lead-free piezoelectric K_0.5Na_0.5NbO₅ (KNN) ceramics have been widely studied regarding their excellent piezoelectric properties and high Curie temperatures in the past decades [1,2,3,4,5,6]. The formation of a rhombohedral–tetragonal (R–T) phase boundary has been recognized as an effective path to obtain superior electrical properties for piezoelectric ceramics [7,8,9]. For KNN systems, A-site substitutions, such as Li⁺, Ag⁺, Ba²⁺ and Sr²⁺ ions, can shift the orthorhombic–tetragonal phase transition temperature (T_O-T) downward, near room temperature [10,11], and B-site substations of Zr⁴⁺, Fe³⁺ and Sb⁵⁺ can shift the rhombohedral–orthorhombic phase transition temperature (T_R-O) upward, near room temperature [12,13,14,15,16]. Thus, the coexistence of rhombohedral and tetragonal phases at room temperature is promising because of the combination of appropriate A-site and B-site substitutions in KNN materials. It has been confirmed that a R–T phase boundary can be formed in (KNaBa)(Nb,Zr)O₃ (generally denoted as KNN-BZ) ceramics, with a 6–8% BaZrO₃ addition [17]. However, the electrical performance of this composition is very dependent on sintering conditions. The relatively narrow sintering temperature zone and the volatilization of alkali ions at high temperature easily leads to poor compactness of ceramics, which seriously affects the densification and electrical properties of samples [18]; therefore, the exploration of an appropriate sintering process is significant for KNN-BZ ceramics.

There are many types of sintering methods used to prepare KNN-based ceramics, such as hot-press sintering, spark plasma sintering, microwave-assisted sintering, cold sintering processes, and so on [19,20,21,22]. Although these ceramics show improved electrical properties, the cost is very high and not suitable for industry production application. To obtain a high crystalline degree of ceramics using the single-step treatment in the conventional sintering process, sintering temperatures tend to be high. At high temperatures, especially near the melting temperature, a liquid phase occurs, which may generate pores due to the mismatch between the existed solid phase and emerging amorphous phase; this degrades the densification and electrical performances of the obtained ceramics [23]. However, if the amount of the liquid phase could be carefully controlled, it might not cause degradation or densification of the ceramics, but reversely improve the growth of crystalline grains, contributing to a high density. In the initial stage of ceramic sintering, the liquid phase starts to form and the grains grow fast; if the temperature rapidly drops and holds for long time, a small amount of liquid phase is retained, which is favorable to the densification of ceramics [23]. Carefully controlling the duration of the critical temperature that forms the liquid phase and low-temperature sintering should be a promising path to obtain dense and stoichiometric KNN-based ceramics [23].

In this study, the traditional single-step method was first used to determine the melting point of KNN/BZ. Thereafter, according to the melting temperature of ceramics, 2-step and 3-step temperature-controlled procedures were designed. The density, structure and electrical properties of ceramics prepared using different temperature-controlled procedures were investigated.

2. Experiments

Lead-free piezoelectric 0.92(K_0.445Na_0.5Li_0.05)NbO₃-0.08BaZrO₃ (KNLN-BZ) ceramics were prepared using the conventional sintering process. Reagent-grade oxide/carbonate powders supplied by China National Pharmaceutical Group Corporation, K₂CO₃ (99%), Na₂CO₃ (99.8%), Li₂CO₃ (98%), Nb₂O₅ (99.99%), BaCO₃ (99%) and ZrO₂ (99%) were chosen as starting materials. They were all weighed according to the stoichiometry and placed in a 500 mL nylon jar. These powders were mixed in ethanol with zirconia balls (φ2 mm:φ10 mm = 1:1) via ball milling (planetary ball mill, QM-3SP2, Nanjing University Instrument, China) at 400 rev/min for 14 h. Next, the mixture was dried at 90 °C for 4 h in a drying oven and uniaxially pressed into disks (30 mm diameter), followed by calcination at 800 °C for 2 h in a muffle furnace (KSL-1700X, Kejing Materials Co., Ltd., Hefei, China). The calcined disks were ground and then ball-milled for 14 h in the same manner used before calcination. The secondary ball-milled slurry was dried, sieved and pressed into 8 mm diameter disks with polyvinyl alcohol (PVA) under a pressure of 200 MPa. The green compacts were sintered using single-step, 2-step and 3-step methods. First, the traditional single-step method was used. The green ceramics were heated at 600 °C for 4 h to eliminate the binder, sintered in the temperature range of 1080–1220 °C for 2 h in air, and then naturally cooled to room temperature in the furnace. In this study, the melting point of KNLN/BZ composites was found to be approximately 1200 °C, when traces of pinholes were observed on the surface of the ceramics. Consequently, the 2-step and 3-step sintering methods were designed based on a temperature of 1200 °C, when the liquid phase started to form. The 2-step method was set as follows. The green ceramics were heated at 600 °C for 4 h at 3 °C/min, the same as in the single-step method. Next, the temperature was first raised to 1200 °C at a higher heating rate of 10 °C/min and held for 5 min; second, the temperature was decreased to 1100 °C at a rate of 10 °C/min and held for 7 h; finally, the temperature was naturally cooled to room temperature. The 3-step sintering process was set as follows. After the PVA was burned off, the temperature was first raised to 1100 °C (slightly lower than the melting point) at 3 °C/min and held for 0 min to further improve the densification of ceramics. Second, the temperature was increased to 1200 °C at a higher rate of 10 °C/min and held 5 min; third, the sintering temperature was decreased to 1100 °C at 10 °C/min and held for 5 h; and finally, the temperature was naturally cooled to room temperature. The density of ceramics obtained using the 3-step method was higher than that obtained using the 2-step method, but lower than that obtained using the single-step method. The 3-step method was attempted with the holding time increased from 5 h to 20 h. Ceramics prepared using different methods were denoted as single-step, 2-step, 3-step/5 h and 3-step/20 h. To measure their electrical properties, ceramics were polished, coated with silver paste on both surfaces and then fired at 600 °C for 1 h as electrodes.

The densities of ceramic samples were measured using the Archimedes drainage method. X-ray diffraction (XRD) using Cu Kα radiation (D/MAX-2400, Rigaku Co., Tokyo, Japan) was used to characterize the phase compositions of as-prepared samples, which had a 2θ range from 20° to 80°. The morphology was observed using field-emission scanning electron microscopy (FE-SEM, FEI Quanta FEG 250, Thermo Fisher Scientific Co., Waltham, MA, USA). The effective dielectric constant and dielectric loss of the ceramics were measured using a LCR meter (HP 4980A, Agilent Technology Inc., Colorado Springs, CO, USA) under an applied frequency range between 1 kHz and 1 MHz. The temperature dependence of samples’ dielectric properties were characterized using the same LCR meter and a high-temperature resistant furnace. The test temperature range was from room temperature to 500 °C. A ferroelectric testing system (TF Analyzer 2000E, aixACCT systems GmbH, Aachen, Germany) was used to characterize the piezoelectric strain, the polarization–electric field (P–E), and the strain–electric field (S–E) hysteresis loops. After the disks were poled in a silicon oil bath at 150 °C for 15 min by applying a direct current electric field triple of coercive field of samples, the piezoelectric constant, d₃₃, of the ceramic samples was measured using a piezoelectric testing system (ZJ-1, CAS, Beijing, China).

3. Results and Discussion

Figure 1 shows the XRD patterns of 0.92KNLN-0.08BZ ceramics sintered using single-step, 2-step and 3-step methods. Ceramics prepared using single-step and 2-step sintering methods had single perovskite structures without a visible secondary phase. However, a very weak diffraction peak in the angle of 27° appeared in the ceramic prepared using the 3-step sintering method, which originated from the composition difference due to different heat-treatment conditions. Upon further inspection, obvious peak splitting appeared in the sample prepared using the single-step method and the 3-step/20 h method, which was attributed to the good crystallization of ceramics. The ceramic prepared using the 3-step/20 method showed more obvious peak splitting, indicating a higher crystalline degree.

Figure 2 shows the surface morphologies of 0.92KNLN-0.08Z ceramics sintered using different methods. Figure 2a is a SEM image of the ceramic obtained using the single-step method; the ceramic had a dense microstructure, which was related to the higher sintering temperature. Figure 2b,c are SEM images of ceramics obtained using 2-step and 3-step/5 h methods, respectively. There were some pores on the surfaces of the two ceramics, but with relatively uniform grains. In addition, the pores of the ceramic prepared using the 3-step/5 h sintering method were reduced, as shown in Figure 2c. As a result of prolonging the holding time from 5 h to 20 h, the ceramics prepared using the 3-step method had a denser micro-structure, as shown in Figure 2d. The formation of many square grains also manifested good crystallization, consistent with the analysis result of XRD.

Table 1. Density values for ceramics prepared using different sintering methods.

Sintering Method	Single-Step	2-Step	3-Step/5 h	3-Step/20
Density (%)	98.89	95.98	97.57	99.04

lists the densities of all ceramics sintered using different methods. The ceramic prepared using the single-step method showed a higher density (98.9%) than those prepared using 2-step and 3-step/5 h methods. The sample obtained using the 2-step method had the lowest density (95.8%). As the density of ceramic obtained using the 3-step/5 h method was 97.6% and increased again, prolonging holding time was attempted in the 3-step process; holding time was increased from 5 h to 20 h. As expected, the density of the ceramic prepared using the 3-step/20 h was 99.1%, similar to that prepared using the single-step method.

Figure 3 provides the frequency-dependent dielectric properties of lead-free piezoelectric KNLN-BZ ceramics sintered under different conditions. The ceramic prepared using the 3-step/20 h method showed a higher dielectric constant and lower dielectric loss than the other ceramics. The inset shows the temperature-dependent dielectric properties of the ceramic prepared using the 3-step/20 h method. The Curie temperature (T_C) 286 °C corresponded to the maximum permittivity peaks of the 0.92KNLN-0.08 BZ ceramics. Another abnormal dielectric peak, at 60–70 °C, was observed, which originated from the ferroelectric-ferroelectric phase transition in the KNN system.

The piezoelectric coefficients, d₃₃ of 0.92KNLN-0.08BZ ceramics prepared by different methods were given in Figure 4. It was found that the specimen sintered using the 3-step/20 h method showed the highest piezoelectric coefficient, d_33, which was 161 pC/N. A high dielectric constant should also contribute to the enhancement of piezoelectric properties. In addition, the dense microstructure may have been responsible for the enhancement of piezoelectric properties.

There were no differences among the ferroelectric properties of all ceramics sintered under different conditions, which show the same polarization–electric field (P–E) hysteresis loops, maximum polarization (P_max) values and remanent polarization (P_r). The related results are not shown here. We chose the ceramic sintered using the 3-step/20 h method to study the thermal stability of ferroelectric properties; its P–E loops at different temperatures and the variation of P_max and P_r with temperature are shown in Figure 5. Figure 5a shows that the KNLN-BZ ceramic prepared using the 3-step/20 h showed well-saturated ferroelectric polarization between 30 °C and 180 °C (the measurement instrument’s maximum temperature), manifesting good ferroelectric properties at high temperatures near 200 °C. Figure 5b clearly indicates that decreases in P_max and P_r were not monotone with temperature. An obvious peak appeared at 60–70 °C, which might be attributed to the ferroelectric phase transition; this is consistent with the results of temperature-dependent dielectric properties. The phase transition at this temperature led to an increase in polarization. The P_max and P_r at room temperature were 19.3 μC/cm² and 12.7 μC/cm², respectively; at 180 °C they decreased to 15.3 μC/cm² (20.7% change) and 7.51 μC/cm² (40.9% change), respectively, in the 30 °C to 180 °C temperature range. The dramatically decreased P_r manifested obvious domain switching.

Figure 6a shows the unipolar piezoelectric strain of the 0.92KNLN−0.08BZ ceramic sintered using the 3-step/20 h method at temperatures ranging from 30 °C to 180 °C. The strains did not show a continuous decrease with temperature, but abnormally increased at 60 °C. The highest strain was approximately 0.12%. This result could be attributed to the ferroelectric phase transition at 60 °C between tetragonal and rhombohedral phases, which was consistent with ferroelectric properties. The piezoelectric coefficient, d₃₃*, at different temperatures, was calculated as the function of d₃₃* = S/E and is plotted in the inset of Figure 6a. The d₃₃* first increased and decreased with temperature increases; the highest piezoelectric coefficient, d₃₃*, of 400 pm/V, also appeared at 60 °C. The d₃₃* could remain at 300 pm/V even when the temperature was as high as 180 °C. The strain response was compared to that of LF4T textured ceramics and several well-known piezoelectric ceramics in Figure 6b, which shows normalized d₃₃* at different temperatures. The ceramics show higher d₃₃* at temperatures lower than 80 °C, but the thermal stability should be further improved by modifying the process.

4. Conclusions

Lead-free piezoelectric 0.92KNLN-0.08BZ ceramics were sintered under different conditions. It was found that ceramic prepared using the 3-step/20 h method showed the highest electrical properties. Dielectric properties were independent of the sintering process; all ceramics had similar Curie temperatures of 286 °C and dielectric constants of 1340. Different dielectric losses might have originated from differences in crystalline degree and densification. All ceramics had similar ferroelectric properties. Examination of the thermal stability of ferroelectric polarization and electric field-induced strain indicated that the highest remanent polarization and piezoelectric strains were obtained at 60 °C, and that good ferroelectric properties and piezoelectric strain responses could be retained at 180 °C. Although these ceramics did not show superior electrical properties, especially piezoelectric properties, the effectiveness of the 3-step sintering process regarding the improvement of electrical properties of KNN-based piezoelectric ceramics was confirmed. This study provides a path to enhance the densification and electrical performance of KNN-based ceramics using a simple composition that satisfies the requirements of industrial applications.