Introduction

Recently, lead-free ceramics with large strains that can be suitable for ceramic actuator applications have attracted much attention. Among the reported lead-free counterparts, (Bi0.5Na0.5)TiO3 (BNT)-based systems usually exhibit large electric field-induced strains with the value of 0.3–0.5%. However, the electric field required for achieving large strains is very high, leading to the extremely low large-signal piezoelectric coefficient d33* (Smax/Emax) in these systems [1,2,3,4].

BaTiO3 (BT)-based piezoelectric ceramics have been considered as one of the most promising lead-free candidates because these materials exhibit not only the obvious advantages of low cost and easy processing, but also excellent piezoelectric coefficient (d33) near the multiple phase coexisted boundary. The latter case is believed to be beneficial to obtain high strain by taking into account of the strain contribution in piezoelectric ceramics [5,6,7,8,9,10]. Unfortunately, only relatively low strain values can be observed in these modified BT-based piezoelectric ceramics. It is found that in many BT-based piezoelectric ceramics, large d33* value can be only achieved in an extremely low electric field range, above which the d33* value shows an obvious decrease, leading to the low strain value [11,12,13,14,15,16,17,18]. This should be attributed to a fact that the electric field needed to switch the ferroelectric domains in BT-based systems is much lower than that of other types of piezoelectric ceramics such as Pb(Zr,Ti)O3 (PZT) and (Na,K)NbO3 (NKN) systems [19].

Acceptor doping is an effective strategy to modify the ferroelectric and piezoelectric properties of piezoelectric ceramics via the defect dipoles between the negatively charged defects (i.e., impurity defects) and positively charged defects (i.e., oxygen vacancy) [20,21,22,23]. The defect dipole is believed to cooperatively align the direction of spontaneous polarization (Ps) and can provide a restoring force to recover the switched polarization, leading to the reversible domain switching and recoverable electrostrain [24,25,26,27,28,29]. This would be beneficial for achieving large strain in BT-based ceramics. In the present study, lead-free (Ba0.865Ca0.135)(Ti0.91Zr0.09)O3 (BCTZ) + x mol% MnO2 (x = 0–1.25) system was constructed. The selected BCTZ composition is located at the tetragonal-rich side of orthorhombic–tetragonal phase boundary [10], which is conductive to achieve the hardening effect. A small amount of MnO2 was added to produce the oxygen vacancies due to the charge compensation. The influence of MnO2 on the phase structure and various electrical properties, particularly, the strain properties was investigated in detail.

Experimental procedure

The BCTZ ceramic was prepared by a conventional solid-state reaction method by using high-purity raw materials. Barium carbonate (BaCO3), calcium carbonate (CaCO3), titanium dioxide (TiO2) and zirconium dioxide (ZrO2) (AR, Sinopharm Chemical Reagent Co., Ltd., CN) were used as the raw materials. All of the stoichiometric raw powders were mixed by planetary ball milling in ethanol for 4 h. After calcination at 1250 °C for 4 h, the calcined powders with different MnO2 content (0–1.25 mol%) and 0.5 wt% polyvinyl butyral (PVB) were ball milled again for 6 h. The dried slurries were pressed into green pellets with a diameter of 10 mm and a thickness of 0.7–0.8 mm under a pressure of 150–200 MPa, and then sintered at 1320–1420 °C for 3 h in air.

The crystal structure of the as-sintered ceramic powders was examined by an X-ray diffractometer (XRD, D/MAX-RB; Rigaku, Tokyo, Japan) with Cu Kα radiation (λ = 1.5406 Å). Coupled θ–2θ scans were performed with a 2θ step interval of 0.02626° and a scanning speed of 4°/min. For the electrical measurements, the as-sintered samples were polished using silicon carbide paper, and then, the silver paste was painted on both sides of the polished sample surfaces and fired at 550 °C for 30 min. The specimens were poled at room temperature under a dc field of 3 kV/mm for 15 min in a silicone oil bath. The quasi-static d33 value was measured by using a quasi-static piezoelectric constant apparatus (YE2730A, Sinocera, Yangzhou, China), and the mechanical quality factor Qm was determined by a resonance–antiresonance method with an impedance analyzer (Impedance Analyzer PV70A, Beijing, China). Temperature-dependent dielectric properties of samples were measured using an LCR meter (E4980A, Agilent, Santa Clara, CA). The polarization versus electric field (P–E) and strain versus electric field (S–E) curves were measured by using a ferroelectric measuring system (Precision multiferroelectric, Radiant Technologies Inc., Albuquerque, NM) with an accessory laser interferometer vibrometer (AE SP-S 120E, SIOS technic, GmbH, Ilmenau, Germany). The measurement frequency of both the P–E and S–E curves was fixed at 10 Hz. The X-ray photoelectron spectrum (XPS) with a Physical Electronics PHI 5802 using Kα line of Al (hν = 1486.6 eV) as X-ray source was recorded to observe the binding state of manganese in the specimens. The Hitachi F-4600 fluorescence spectrophotometer with the scanning speed of 240 nm/min was used to measure the emission spectra for determining the oxygen vacancy concentration.

Results and discussion

Figure 1a shows the room temperature XRD patterns of BCTZ ceramics with different MnO2 contents. All samples exhibit a pure perovskite structure without any trace of secondary phase. However, the (200)C peaks show an obvious splitting corresponding to the (002)T/(200)T doublet as shown in Fig. 1b because the selected BCTZ matrix is located at the tetragonal-rich side of tetragonal-orthorhombic phase boundary [10]. With the increase in the MnO2 content, Fig. 1b, c shows that both (111)c and (200)c reflections exhibit a slight shift toward low angle side until x = 0.75, beyond which an abnormal shift toward high angle side can be observed. This suggests that the addition of a small amount of MnO2 can induce a slight lattice expansion initially, whereas a slight lattice shrinkage can be occurred at the relatively high MnO2 content.

Figure 1
figure 1

a Room temperature XRD patterns of BCTZ ceramics and the enlarged b (111)c and c (200)c reflections with different MnO2 contents

Full size image

Figure 2 shows the XPS spectra of Mn 2p3/2 orbit for x = 0.25, 0.75 and 1.25 samples, which was fitted with Gaussian distribution curves. Two distinct components around the ~ 640 eV and 641.5 eV can be assigned as Mn2+ and Mn3+ ions in 2p3/2 state, respectively [30]. This should be related to a fact that the valence of Mn ions tends to decrease with increasing temperature and it suggests that only Mn2+ and Mn3+ ions can be existed in the sample after sintering at the high temperature [31]. Similarly, Yao et al. also found that only Mn2+ and Mn3+ ions can be detected in PZT- and NKN-based ceramics [32]. With increasing the MnO2 content, although the divalent Mn2+ ions are dominant in the studied composition range, the overall integral intensities corresponding to the Mn2+ and Mn3+ ions increase monotonically. In addition, the Mn2+/Mn3+ ratio also shows a slight increase due to the change of the relative integral intensity between them. Considering that the radius of Mn2+ ions is larger than that of Mn3+, Ti4+ and Zr4+ ions (CN = 6, \({R}_{{\text{Mn}}^{2+}}=0.83\) Å, \({R}_{{\text{Mn}}^{3+}}=0.645\) Å, \({R}_{{\text{Ti}}^{4+}}=0.605\) Å, \({R}_{{\text{Zr}}^{4+}}=0.72\) Å), and the radius of Mn2+ and Mn3+ ions is smaller than that of Ba2+ and Ca2+ ions (CN = 12, \({R}_{{\text{Mn}}^{2+}}=1.22\) Å, \({R}_{{\text{Mn}}^{3+}}=1.03\) Å, \({R}_{{\text{Ba}}^{2+}}=1.61\) Å, \({R}_{{\text{Ca}}^{2+}}=1.34\) Å) [33], it is suggested that Mn ions tend to enter the B-site until x = 0.75, beyond which they preferentially enter the A-site. This induces the observed lattice expansion and shrinkage in the relatively low and high MnO2 content range, respectively.

Figure 2
figure 2

X-ray photoelectron spectroscopy analysis of the Mn 2p3/2 orbit of the BCTZ ceramics with different MnO2 content. The dots represent the experimental data, while solid lines represent the fitting data

Full size image

This also can be confirmed by the Raman spectra. Figure 3 shows the room temperature Raman spectra of BCTZ ceramics with different MnO2 contents. No obvious change of the vibrational modes except for the modes corresponding to the A1(TO1), E(TO2) and A1(LO3)/E(TO3). The modes of A1(TO1) and E(TO2) are assigned to B–O bonds, while the mode of A1(LO3)/E(TO3) is attributed to the A–O bonds [34,35,36]. It is evident that with increasing x, both A1(TO1) and E(TO2) modes exhibit a slight weakness at x ≤ 0.75, which should be resulted from the local structural disorder after the substitution of Mn ions for the B-site ions. However, the peak intensity of these two modes exhibit a slight increase with further increasing x (x > 0.75), possibly because Mn ions tend to enter A-site rather than the B-site. Figure 3 shows that the peak intensity of A1(LO3)/E(TO3) mode related to the A–O bonds also exhibits a slight increase, since Mn ions preferentially occupy the A-site as x > 0.75.

Figure 3
figure 3

Room temperature Raman spectra of BCTZ ceramics with different MnO2 contents

Full size image

It is expected that the substitution of Mn ions for B-site Ti/Zr ions can induce the oxygen vacancies due to the charge compensation, which can further form the defect dipole with the negatively charged defects (i.e., impurity Mn defects) and play an important role in various electrical properties. The existence of oxygen vacancies can be detected by photoluminescence (PL) spectra. Figure 4 shows the PL spectra of BCTZ ceramics with different MnO2 contents. Generally, the PL intensity shows a monotonic increase with increasing the oxygen vacancy content [37]. Very interestingly, with increasing x, the PL intensity exhibits an obvious increase as x ≤ 0.75. This indicates that the substitution of Mn2+/Mn3+ ions for Ti/Zr ions in this composition range is the typical acceptor doping, which can induce the oxygen vacancies. However, with further increasing x, the PL intensity shows an obvious decrease, possibly because Mn ions preferentially occupy the A-site to replace the Ba/Ca ions, which can be considered as the donor doping and is beneficial for restricting the formation of oxygen vacancies.

Figure 4
figure 4

Variation of PL spectra of BCTZ ceramics as a function of x

Full size image

Figure 5 shows the temperature-dependent relative dielectric constant of BCTZ ceramics with different MnO2 contents measured at 1 kHz. It can be seen that the temperature corresponding to the dielectric maximal (Tm) shows no obvious change with increasing MnO2 content until x = 0.75, beyond which Tm exhibits a slight decrease. It is usually believed that B-site acceptor doping shows no obvious effect on the Tm value, whereas the A-site donor doping can decrease the Tm value significantly [38,39,40]. The decrease in the Tm value after the A-site donor doping should be attributed to the local relaxation induced by the cation vacancies. This leads to the distortion of nearby BO6 octahedron and thus is responsible for the decrease in the Tm value [38]. In addition, it can be seen that the decrease in Tm at x > 0.75 is accompanied by the slight shift of the orthorhombic–tetragonal polymorphic phase transition (PPT) temperature toward high temperature range, as shown in the inset of Fig. 5. It is thus suggested that the observed soft behavior at x > 0.75 can be attributed to the synergistic effect of the occupation site change of Mn ions and the coexistence of orthorhombic and tetragonal phases.

Figure 5
figure 5

Temperature-dependent relative dielectric constant εr (εr–T) of BCTZ ceramics with different MnO2 contents. The inset shows the evolution of εr–T curves around the room temperature

Full size image

The electric field-induced polarization (P–E) and strain (S–E) curves of poled BCTZ ceramics with different MnO2 content at room temperature are displayed in Fig. 6. All samples exhibit the typical square-like P–E loops, and both the maximal polarization (Pmax) and the remanent polarization (Pr) show little composition dependence, as clearly shown in Fig. 7a. However, it can be found that the P–E loops become increasingly asymmetric with increasing x, as characterized by the separation of the absolute value between the positive coercive field (+Ec) and the negative coercive field (−Ec). This indicates the existence of an internal bias field (Ei) that defined as the difference between the value of +Ec and the absolute value of −Ec. Very interestingly, Fig. 7b shows that the Ei value exhibits an obvious increase until x = 0.75, beyond which it shows a slight decrease with further increasing x. This is in good agreement with the variation trend of oxygen vacancy content. Of particular importance is that the appearance of the Ei is accompanied by a distinct change of the strain characteristics. The typical symmetric butterfly-like S–E curves become more and more asymmetric after the MnO2 doping, which is associated with the gradual decrease in the negative strain. As a result, only a highly asymmetric sprout-shaped strain loop can be observed at x = 0.075 with the maximal strain value of ~0.242 under 2 kV/mm, which is ~82% higher than the undoped sample (only ~0.133% at x = 0). Figure 7c shows that the corresponding large signal d33* (= Smax/Emax) value also exhibits an obvious increase from only ~ 665 pm/V at x = 0 to 1210 pm/V at x = 0.75. Such a high d33* value has been rarely reported in BT-based ceramics under such a high external electric field amplitude, suggesting that the present studied composition exhibits a good potential for applications in actuators. Figure 7d displays the evolution of the small signal quasi-static d33 value and the mechanical quality factor Qm, and these two parameters exhibit an opposite trend. With increasing x, the d33 value exhibits a slight decrease with increasing x, while the Qm value shows a slight increase. This is the typical characteristic of hardening effect. The largest Qm and the lowest d33 value can be observed at x = 0.75 sample. Very interestingly, no negative strain can be observed in the corresponding strain curve. These phenomena should be undoubtedly related to the Ei established by the defect dipoles between oxygen vacancy and the acceptor doping Mn ions. After poling, the defect dipoles can be oriented along the poling field direction, acting as the pinning center to restrict the domain switching. As a result, if the external electric field direction is reversed, these preferentially oriented domains cannot be switched parallel to the reversed field direction. This also induces the observed asymmetric P–E loops. However, owing to the restriction of the preferentially oriented defect dipoles, the Pr value exhibits little composition dependence as shown in Fig. 7a. With further increasing x, an obvious soft behavior can be observed as characterized by the appearance of negative strain, the reduction in the asymmetry of S–E curves, the increase in the d33 value, the decrease in the Qm, etc., possibly because Mn ions tend to enter A site and the orthorhombic–tetragonal PPT temperature is shifted toward room temperature.

Figure 6
figure 6

Evolution of the bipolar P–E loops and S–E curves of BCTZ ceramics as a function of x measured at room temperature with the field amplitude of 2 kV/mm

Full size image
Figure 7
figure 7

Evolution of various electrical parameters as a function of x

Full size image

Conclusions

In summary, the influence of the MnO2 doping on phase structure and various electrical properties of BCTZ ceramics was investigated. It is found that a small amount of MnO2 can enter the B site to act as the acceptor defect center, leading to the formation of oxygen vacancies due to the charge compensation. After poling, the defect dipoles between the negatively charged defects and the oxygen vacancies can orient along the poling field direction, leading to the asymmetric but recoverable S–E curves. This induces an obvious enhancement of strain value and an ultrahigh electrostrain with a value of 0.242% (d33* = 1210 pm/V) at 2 kV/mm can be achieved in 0.75 mol% MnO2-doped BCTZ ceramics. However, with further increasing MnO2 content, an obvious soft behavior can be observed because of the occupation site change of Mn ions and the O–T phase coexistence. This study provides a good strategy for achieving ultrahigh strains in modified BT-based ceramics via defect engineering.