1 Introduction

Pb(Zr,Ti)O3 (PZT)-based ceramics have been widely applied in many fields, such as sensors, actuators, transducers, etc., due to their excellent electrical properties and relatively high Curie temperature (T C) [1, 2]. However, unfortunately, lead existed in PZT-based ceramics can cause serious harm to the health of both the environment and humans during their production, usage, and disposal operations, owing to the high toxicity of lead. Therefore, considerable attention has been paid to lead-free piezoelectric materials for their replacement of PZT-based ceramics [3,4,5], especially after the implementation of the RoHS legislation in July 2006 in Europe [6], which forbids the use of lead in electronic products. Among all lead-free materials researched so far, the (K,Na)NbO3 (KNN)-based ceramics possess relatively high values of both piezoelectric activity and Curie temperature [7,8,9], and therefore have attracted increasing attention around the world, particularly after the breakthrough made by Saito et al. in 2004 [10], when a great improvement was achieved in the piezoelectric properties of KNN-based materials, by using the reactive-templated grain growth method.

However, generally speaking, the piezoelectric properties of pure KNN ceramics are poor when compared to commercial PZT-based ceramics [11,12,13], presenting a major limitation for their practical applications. Therefore, in order to improve their piezoelectric activities, many researchers have focused on the construction of phase boundaries in KNN-based ceramics [14,15,16,17,18,19,20,21,22], as they can provide more spontaneous polarization directions [1], thus leading to better piezoelectric properties. It is well-known that pure KNN undergoes a rhombohedral–orthorhombic (R–O) phase transition at − 123 °C, and an orthorhombic–tetragonal (O–T) phase transition at 210 °C [7]. And the phase boundaries, including R–O, O–T, and rhombohedral–tetragonal (R–T) phase boundaries, are generally formed by means of shifting the temperatures of R–O and/or O–T phase transitions (namely, the T R–O and/or T O–T) to near room temperature, with the introduction of different additives [7, 14,15,16,17,18,19,20,21,22].

It is the R–T phase boundary that has been receiving more and more attention in recent years [23,24,25], since it can allow for piezoelectric properties comparable to those of certain commercial PZT-based ceramics. The construction of the R–T phase boundary requires that the T R–O and T O–T merge at room temperature. Some additives, such as AZrO3 (A = Ba, Sr, and Ca) and Bi0.5Na0.5ZrO3, play an important role in the formation of R–T phase boundaries, as they can simultaneously shift both the two temperatures to room temperature [19, 21, 22]. Very recently, some authors found that the Bi0.5Na0.5HfO3 (BNH) had similar effects on the two temperatures as AZrO3 and Bi0.5Na0.5ZrO3, and hence could bring about significant improvement in the piezoelectric properties of KNN-based ceramics [26,27,28].

Therefore, in the present work, the CaZrO3 and BNH co-modified KNN-based ceramics, with the general chemical formula of (0.98 − x)(K0.4Na0.6)NbO3–0.02CaZrO3xBi0.5Na0.5HfO3 (abbreviated as KNN–CZ–xBNH), were projected and fabricated by a conventional solid-state reaction method. A study was carried out to assess the influences of BNH content on the phase structure, microstructure, and piezoelectric properties of the ceramics. It was found that the ceramics possessed an R–T phase boundary in the composition range of 0.04 ≤ x ≤ 0.045, near which an enhanced piezoelectric activity was achieved.

2 Experimental procedure

The KNN–CZ–xBNH ceramics were prepared by a conventional solid-state reaction from adequate mixtures of K2CO3 (99%), Na2CO3 (99.8%), Nb2O5 (99.5%), CaCO3 (99%), ZrO2 (99%), Bi2O3 (99%), and HfO2 (99.99%) powders. All the raw materials were weighed in the stoichiometric amounts, and then were ball-milled in ethanol with agate media for 24 h. The slurry was dried and subsequently calcined at 850 °C for 6 h. The calcined powders were milled and dried again, and then were pressed into disks with a diameter of about 10 mm and a thickness of about 1.0 mm under a uniaxial pressure of 10 MPa after using the 3–5 wt% polyvinyl alcohol (PVA) as a binder. After burning off the PVA at 500 °C for 2 h, all the pellets were sintered at 1150 °C for 3 h in air, except those with the composition x = 0, which were sintered at 1110 °C.

An X-ray diffractometer (XRD) (XRD-6100, Shimadzu, Japan) was used to analyze the phase structures of these sintered ceramic samples, and their microstructures were observed using a scanning electron microscope (SEM) (JSM-6610, JEOL, Japan), equipped with an energy dispersive spectrometer that was used to analyze the chemical composition. The densities of the sintered samples were determined by the Archimedes water immersion method. In order to characterize the electrical properties of the ceramics, the silver paste was coated on both sides of them and then fired at 650 °C for 10 min. The temperature-dependent dielectric properties were measured via a computer-controlled precision LCR meter (E4980A, Keysight Technologies, USA). For piezoelectric characterization, the samples were poled at 40 °C in a silicone oil bath by applying a DC electric field of 3–4 kV/mm for 30 min. After aging for 24 h, the piezoelectric constant d 33 was measured on a quasi-static d 33 meter (YE2730A, Wuxi Yutian, China). A resonance–antiresonance method was employed to determine the planar electromechanical coupling coefficients (k p) of the ceramics using an impedance analyzer (E4990A, Keysight Technologies, USA) on the basis of IEEE standards.

3 Results and discussion

Figure 1a shows the XRD patterns of the KNN–CZ–xBNH ceramics, measured at room temperature and in the 2θ range of 20–60°. It can be observed that all the samples possess a single perovskite structure, indicating that the BNH has a good solubility in the KNN-based solid solutions. The good solubility between the BNH dopant and the KNN–CZ host can be further verified by the uniform elemental distribution in the ceramics, as show in Fig. 2, which was obtained using an energy dispersive X-ray analyzer (EDAX).

Fig. 1
figure 1

XRD patterns of the KNN-CZ-xBNH ceramics in the 2θ range of a 20–60° and b 44–48°

Fig. 2
figure 2

Elemental mapping of the KNN–CZ–xBNH ceramics with x = 0.04

For better identification of the structural evolution of the ceramics, the enlarged XRD patterns are shown in Fig. 1b, with the 2θ range of 44–48°. It is notable that the ceramics with compositions x ≤ 0.03 have an orthorhombic symmetry, as they show split peaks at about 46°, respectively identified as (200) and (002), and the two peaks have an intensity ratio (I 200/I 002) of about 1:2, which is a typical characteristic of orthorhombic phase in KNN-based ceramics [28]. Moreover, one can observe that the split peaks merge into a single one for the ceramics with the compositions of x ≥ 0.05, suggesting that they possess a rhombohedral structure. Nevertheless, the ceramics with compositions 0.035 ≤ x ≤ 0.045 could not be identified as any phase only given the XRD results. Therefore, for more detailed structural analysis, it is necessary to combine the XRD patterns with other experimental data, such as dielectric-temperature curves, as displayed in Fig. 3.

Fig. 3
figure 3

Dielectric-temperature curves of the KNN–CZ–xBNH ceramics in the temperature range from about − 150 to 150 °C

The dielectric-temperature curves in Fig. 3 were measured at a frequency of 10 kHz and in the temperature range of about − 150 to 150 °C. For the ceramics with compositions x ≤ 0.035, two dielectric anomalies, corresponding to the R–O and O–T phase transitions respectively, can be observed in the curves. It is clearly evident that the addition of BNH simultaneously shifts both the temperatures of the two phase transitions (i.e., the T R–O and T O–T) to room temperature, which is beneficial for the formation of R–T phase boundary, as pointed out previously. For the composition x = 0.035, its T R–O is near room temperature, while its T O–T is far above room temperature, indicating it has an R–O phase boundary at room temperature. With the further increase of BNH content, the two temperatures converge at 0.04 ≤ x ≤ 0.045. The convergence of the two temperatures is important evidence that the R–T boundary has been formed, as suggested by many references [7, 23,24,25]. Therefore, according to these results together with the previous XRD analysis, one can reasonably conclude that the ceramics possess an R–T phase boundary in the composition range of 0.04 ≤ x ≤ 0.045.

Figure 4 displays the SEM micrographs of the KNN–CZ–xBNH ceramics with x = 0, 0.02, 0.03, 0.04, and 0.05. It can be seen that the grains of the ceramics become extremely large with the addition of a small amount of BNH. The extreme growth of grains can be ascribed to the high sintering temperature, but in this work, it is the added Bi3+ and Hf4+ cations, which respectively occupy the A and B sites of the ABO3 perovskite structure, that provide the most important contributions to the extremely large grain sizes. This is because the introduction of Bi3+ and Hf4+ ions, which have valences different from those of other cations, may produce a large amount of B-site and oxygen vacancies, respectively, in the crystal lattices [26, 27]. And as is well-known, the vacancies always promote the grain growth during the sintering of ceramics. However, when the amount of added BNH exceeds a certain value, many Bi3+ and Hf4+ ions will aggregate at the grain boundaries [28], which can hinder grain growth and thereby leads to smaller grain sizes in the materials, as shown in Fig. 4d, e. Particularly in Fig. 4e, an extremely fine grain size is observed for the ceramics with the composition x = 0.05. This means that adding too much BNH will result in a lower sintering activity in the KNN-based ceramics, which generally has adverse effects on the electrical properties.

Fig. 4
figure 4

SEM micrographs of the KNN–CZ–xBNH ceramics for a x = 0, b x = 0.02, c x = 0.03, d x = 0.04, and e x = 0.05

For further evaluating the quality of the sintered samples, the densities of the KNN–CZ–xBNH ceramics have been measured, as illustrated in Fig. 5. It can be found that the density shows an increasing trend with the addition of BNH until x = 0.045, above which the density begins to decrease upon further increasing the BNH content. Apparently, it is the degradation in sintering activity (as discussed previously) that leads to the reduced density for the ceramics with relatively high BNH contents.

Fig. 5
figure 5

Density of the KNN–CZ–xBNH ceramics as a function of BNH content

To identify the effects of BNH content on the ferro-paraelectric phase transition, the dielectric constants of the KNN–CZ–xBNH ceramics were measured from room temperature to 500 °C, as shown in Fig. 6. It can be observed that the temperature of the Curie peak (approximately determined as T C) gradually declines as x increases. Furthermore, a sharp Curie peak can be seen for the ceramics with the compositions of x ≤ 0.04. However, when x ≥ 0.045, the Curie peak tends to broaden upon the further increase of BNH content, and eventually develops into a very broad and flat shape for the compositions x ≥ 0.05, exhibiting features of diffuse phase transition. This should be attributed to the inhomogeneous distribution of various cations at A and B sites of the pervoskite structure [29], which is rather difficult to be avoided due to the phase segregation of end members over a wide temperature interval, as suggested by Refs. [30, 31].

Fig. 6
figure 6

Dielectric-temperature curves of the KNN–CZ–xBNH ceramics in the temperature range from room temperature to 500 °C

Figure 7 illustrates the piezoelectric constant d 33, planar electromechanical coupling coefficient (k p), and relative dielectric constant ε r (measured at 1 kHz) of the KNN–CZ–xBNH ceramics as functions of x. It can be found that both the d 33 and k p firstly rise and then decrease alongside the increase of BNH content, and that they reach their respective maximum values of 320 pC/N and 0.40 at x = 0.045. The enhanced piezoelectric activity should be attributed to the R–T phase boundary that causes the lower energy barrier for polarization rotation, due to the involvement of more polarization states [1, 32]. However, as the crystal structure becomes a sole rhombohedral phase at x = 0.05, the ceramics experience a dramatic reduction in both the d 33 and k p. Moreover, one can see that the ε r tends to increase with x, in spite of a slight enhancement appearing at the R–T phase boundary. A similar tendency of increase in ε r with the amount of BNH has been found in other BNH-modified KNN-based ceramics as well [33]. To sum up, our results suggest that the addition of BNH is an effective method for the construction of R–T phase boundaries as well as the improvement of electrical properties in KNN-based ceramics.

Fig. 7
figure 7

d 33, k p, and ε r of the KNN–CZ–xBNH ceramics as functions of BNH content

4 Conclusion

Lead-free KNN–CZ–xBNH piezoelectric ceramics were fabricated by a conventional solid-state reaction method. In the studied composition range, all the ceramics possess a single perovskite structure, without any detectable secondary phase. The ceramics possess an R–T phase boundary in the composition range of 0.04 ≤ x ≤ 0.045, near which the materials exhibit enhanced piezoelectric activity, with maximum values of d 33 and k p, being 320 pC/N and 0.40 respectively. The formation of the R–T phase boundary is due to the fact that the introduction of BNH can simultaneously shift both the T R–O and T O–T to room temperature. Furthermore, the addition of a small amount of BNH can promote the grain growth of the ceramics during the sintering process, and even may result in extremely large grain sizes. Overall, our results suggest that the addition of BNH is an effective means for the construction of R–T phase boundaries in KNN-based ceramics, which can lead to a significant enhancement of piezoelectric properties.