Optimized Strain Response in (Co_0.5Nb_0.5)⁴⁺-Doped 76Bi_0.5Na_0.5TiO₃-24SrTiO₃ Relaxors

Abstract

High strain with low hysteresis is crucial for commercial applications in high precision actuators. However, the clear conflict between the high strain and low hysteresis in BNT-based ceramics has long been an obstacle to actual precise actuating or positioning applications. To obtain piezoceramics with high strain and low hysteresis, it is necessary to enhance the electrostrictive effect and develop an ergodic relaxor (ER) and nonergodic relaxor (NR) phase boundary under ambient conditions. In this work, (Co_0.5Nb_0.5)⁴⁺ doped 76Bi_0.5Na_0.5TiO₃-24SrTiO₃ (BNST24) relaxors were fabricated using the conventional solid state reaction route. X-ray diffraction patterns revealed the B-site substitution in BNST24 ceramics. By adjusting the (Co_0.5Nb_0.5)⁴⁺ doping in BNST24, we effectively tuned the T_NR-ER and T_d close to ambient temperature, which contributed to the development of the ergodic relaxor phase and enhanced the electrostrictive effect at ambient temperature. The I-P-E loops and bipolar strain curves verified the gradual evolution from NR to ER states, while the enhanced electrostrictive effect was verified by the nearly linear S-P² curves and improved electrostrictive coefficient of the BNST24-xCN relaxors. An enhanced strain of 0.34% (d^*₃₃ = 483 pm/V) with low hysteresis of 8.9% was simultaneously achieved in the BNST24-0.02CN relaxors. The enhanced strain was mainly attributed to the proximity effect at the ER and NR phase boundary of BNST24-0.02CN, while the improved electrostrictive effect contributed to the reduced strain hysteresis. Our work demonstrates an effective strategy for balancing the paradox of high strain and low hysteresis in piezoceramics.

1. Introduction

Piezoelectric ceramics with perovskite structures have been the mainstay of actuator and transducer applications for the past 60 years due to their excellent electromechanical properties and stable thermal behavior [1,2]. Pb-based piezoelectric ceramics (e.g., Pb(Zr,Ti)O₃ (PZT)) exhibit much higher strain than their lead-free counterparts, and therefore dominate the piezoelectric actuator market. Unfortunately, PZT ceramics contain more than 60 wt.% of the volatile element Pb, which can cause serious environmental problems. Moreover, WEEE/RoHS regulations have been widely applied in many countries to restrict the use of hazardous substances, including lead, in electrical and electronic equipment, driving the search for lead-free piezoelectric ceramics to replace PZT-based materials.

Among all of the reported lead-free piezoelectric ceramic families, Bi_0.5Na_0.5TiO₃ (BNT)-based ceramics are considered as potential candidates to replace commercial PZT, especially in actuator applications, due to their superior strain performance, comparable to lead-based ferroelectrics [3,4,5,6,7,8,9,10]. However, the electric field level required to induce a large strain is too high for commercial applications in these BNT-based systems [7,11,12,13]. For example, a large strain of 0.43% in 90Bi_1/2Na_1/2TiO₃-60BaTiO₃-4K_0.5Na_0.5NbO₃ ceramics at 11 kV/mm [11] and a large strain of 0.51% was achieved in <001> texture (Bi_0.5Na_0.5)TiO₃-BaTiO₃-NaNbO₃ ceramics under a driving electric field of 12 kV/mm [12]. One material with potential to overcome this problem is the (1-x) BNT-xSrTiO₃ (BNST100x) ceramic [14,15,16,17,18,19,20,21,22]. Krauss reported that a maximum strain of about 0.29% under 6 kV/mm was found at x = 0.25, which coincided with the transition from ferroelectric to anti-ferroelectric phases [14]. Duong observed a strain of 0.25% under 4 kV/mm in BNST28 and attributed this large strain to the phase evolution from nonergodic relaxor to ergodic relaxor states [15]. Furthermore, many researchers have investigated BNST100x ceramics with ST contents of around 24-26 mol%, which could form the polymorphic phase boundary (PPB). Samples near the PPB displayed large field-induced strain values, making BNST a promising material system for actuators in practical applications [16,17,18,19]. Therefore, we selected 76Bi_0.5Na_0.5TiO₃-24SrTiO₃ (BNST24) as the base composition, in an attempt to trigger large strain at a comparatively lower electrical field. However, the enormous strain in BNST-based relaxors has typically been accompanied by large strain hysteresis, which will cause displacement hysteresis and thermal consumption [8,20,21,22]. For example, the ((Bi_1/2 (Na_0.84 K_0.16)_1/2)_0.96 Sr_0.04) (Ti_1-xNb_x)O₃ (x = 0.025) ceramic exhibited a giant strain (≈0.7%) but suffered from a huge hysteresis (≈69%) [8]. Although many studies have been carried out to suppress this large hysteresis, this has often resulted in a significant reduction in strain (<0.25%) [23,24,25,26]. Pan reported that (0.94-x) Bi_0.5Na_0.5TiO₃-0.06BaTiO₃-x(Sr_0.7Bi_0.18Er_0.02)TiO₃ ceramics exhibited lower hysteresis (26%), but with a small strain (0.11%) [23]. Therefore, it is an urgent task to balance high strain and low hysteresis in BNST-based relaxors to accelerate their actual application in environmentally friendly high-precision actuators.

In relaxors, electric-field-induced strain is derived from the internal electrostrictive effect, as well as electric-field-induced phase transition. Of these, the strain generated based on the electrostrictive effect is almost hysteresis free, with the hysteresis mainly caused by the phase transition. Therefore, strain hysteresis can be suppressed by enhanced electrostriction at room temperature (RT). An effective way to improve the electrostrictive effect is to tune the phase transition temperature of the relaxors towards RT [27,28], and a giant strain can be induced by the proximity effect at two different phase boundaries [29]. According to a previous study on the relaxor ferroelectric behaviors of (1-x)(Bi_0.5Na_0.5)TiO₃-xSrTiO₃ ceramics (0 ≤ x ≤ 0.3), the virgin state of the 76(Bi_0.5Na_0.5)TiO₃-24SrTiO₃ ceramic was found to be the NR state at RT when no driving electric field was applied, and the NR state would transfer to ER state upon heating to 150 °C [30]. To take advantage of the electrostrictive property to suppress hysteresis at RT, the T_NR-ER of 76(Bi_0.5Na_0.5)TiO₃-24SrTiO₃ is supposed to be shifted to around RT. Reports [31,32] have shown that B-site complex ion doping was more effective in inducing the phase transition and improving the strain, so (Co_0.5Nb_0.5)⁴⁺ (CN) was introduced into BNT-24ST nonergodic relaxors as a dopant to induce the transition from NR state to ER state, and a giant strain could be induced by the proximity effect at the ER/NR phase boundary. The incorporation of CN is expected to tune the phase transition temperature to around RT and produce a coexistence of ER and NR phases at RT. Thus, the electrostrictive effect and strain will be enhanced, effectively breaking the restrictive relationship between strain and hysteresis. The structures, dielectric, ferroelectric and electromechanical strain performance of CN-modified BNST24 were systematically investigated to discover an optimized amount of CN doping in BNST24. An enhanced strain of 0.34% (d*₃₃ = 483 pm/V) with low hysteresis of 8.9% were successfully obtained in the optimized CN-modified BNST24, which is an optimized improvement in the strain performance compared to the recent study [24,33,34]. Herein, we proposed a strategy to balance the paradox of high strain and low hysteresis for promoting commercial applications in high precision actuators.

2. Materials and Methods

2.1. Preparation of BNST24-x CN Ceramics

The 76(Bi_1/2Na_1/2)TiO₃-24SrTi_1-x(Co_0.5Nb_0.5)_xO₃ relaxors were synthesized via a conventional solid-state reaction method employing high-purity raw materials of Bi₂O₃, Na₂CO₃, TiO₂, Nb₂O₅, Co₂O₃ and SrCO₃. These raw materials were weighed according to the stoichiometric ratio of the compositions and homogenized with the ZrO₂ balls in ethanol for 6 h. The dried slurries were pulverized and calcined at 850 °C for 2 h, and then the calcined mixture was ball milled again for 12 h and dried. The resulting powders were subsequently pressed into 10-mm-diameter disks under approximately 100 MPa. These disks were sintered in air at 1120 °C for 2 h. To minimize possible losses of volatile elements, these disks were embedded in a mixture of identical content. Finally, silver electrodes were coated on the surface of the ceramic at 580 °C for 30 min.

2.2. Characterization

The crystal structures of these ceramics were characterized via X-ray diffraction (XRD, D8-Advance, Bruker AXS Inc.,Madison, Germany) with CuK$\mathsf{\alpha }$ radiation. The microstructure and morphology were tested using scanning electron microscopy (SEM, Quanta FEG 250, Frequency Electronics, Inc., Hillsboro, OR, USA). The permittivity and dielectric loss, which are dependent on temperature and frequency, were determined via a high-precision LCR meter (TH2827, Tonghui Electronic Co., Ltd., Jiangsu, China) from RT to 500 °C at 4 frequencies (0.1, 1, 10, and 100 kHz). The electric-field-induced polarization (P-E) and strain (S-E) hysteresis loops were recorded with a ferroelectric tester (RTI-LC П, Radiant Technologies Inc., Burbank, CA, USA) connected with a laser measurement system (MTI-2100, Hottinger Baldwin Messtechnik GmbH, Darmstadt, Germany). Piezoelectric constant was quantified by employing a quasi-static d₃₃ meter (ZJ-3AN, Institute of Acoustics, Chinese Academy of Sciences, Beijing, China).

3. Results and Discussion

The XRD patterns in Figure 1 show that all compositions of the BNST24-xCN ceramics exhibited a perovskite structure without any observable secondary phases, indicating that the Co⁺³ ions and Nb⁺⁵ ions completely diffused into the crystal lattice of the BNST24 matrix. The enlarged diffraction patterns in the ranges of 39.5°–40.5° and 46°–47° are shown in Figure 1b,c, which were obtained to gain insight into the phase structure. The $\left(111\right)/(1\overline{1}1)$ and $\left(200\right)/(002)$ peak splitting, visible at both 40° and 46.5°, revealed the coexistence of rhombohedral and tetragonal phases in all of the BNST24-xCN samples. The rhombohedral phase was identified by the $\left(111\right)/(1\overline{1}1)$ peaks at about 40° and the tetragonal phase was determined by the $\left(200\right)/(002)$ peaks at about 46.5°, as reported in previous reports [21,35,36,37]. Simultaneously, the (111) and (200) peaks gradually shifted towards lower diffraction angles as the CN increased. This indicated that the Co⁺³ and Nb⁺⁵ ions replaced Ti⁴⁺ at the B-site of the perovskite unit cell. Due to the relatively larger ionic radii of Co³⁺ (0.745Å) and Nb⁵⁺ (0.64Å) compared to Ti⁴⁺ (0.605Å) at the B-site, the substituent (Co_0.5Nb_0.5)⁴⁺ ions generated lattice expansion, which then corresponded with the peak positions by shifting toward a lower angle [38,39,40,41,42].

The SEM micrographs in Figure 2a–e demonstrate the influence of various CN concentrations on the morphology and microstructure of the BNST24 ceramics. Compared to the pure BNST24 sample, BNST24-xCN samples exhibited more homogeneous and denser microstructures, where the grain size obtained by the linear intercept method first increased for 0 ≤ x≤ 0.02 and then presented a stable trend. This result showed that proper CN doping promoted grain growth, thereby improving the density and grain uniformity of the BNST24-xCN ceramics. However, the crystals have a certain solid solubility for dopants, and over-doping with CN causes dopants to aggregate at grain boundaries, and the pinning of grain boundaries would inhibit crystal growth [31]. To evaluate the distribution of CN doping in a clear manner, the plane sweeping mode of energy dispersive spectroscopy (EDS) was adopted in Figure 3b. The EDS result indicated that the Co element was distributed homogenously in the grains of the BNST24-0.02CN sample, which was consistent with the XRD results.

Temperature vs. relative dielectric constant (ε_r) and dielectric loss (tanδ) of BNST24-xCN (0 ≤ x ≤ 0.04) ceramics across various frequencies are illustrated in Figure 4a–e. These broad dielectric permittivity peaks relative to temperature and the apparent frequency dispersion of dielectric permittivity indicated that all of the BNST24-xCN samples exhibited the typical characteristics of relaxors. Two permittivity anomalies were detected for the 0 ≤ x ≤ 0.01 compositions. The first was located at T_m (the temperature where ε_r reached a maximum), reflecting the transition from paraelectric to ER phase [22], while the second anomaly was observed at a lower temperature T_NR-ER, corresponding to the transition from the ER to the NR phase. With increasing CN to 0.04, T_m showed little difference in position, but T_NR-ER varied considerably, as shown in Figure 4f. The T_NR-ER value was 140 °C for BNST24, which then shifted down to 120 °C for BNST24-0.01CN. Interestingly, no distinctive T_NR-ER could be detected for x ≥ 0.02, which implied that T_NR-ER should be below room temperature. As T_NR-ER approached room temperature, the interaction between PNRS became weakly correlated due to the increased dynamics of the PNRS, resulting in a transition from a short-range-ordered NR (weakly polar) state to an ER state with randomly oriented polar clusters (non-polar state), thereby stabilizing the ER phase at ambient temperature [29,43]. In addition, the depolarization temperature T_d where the dielectric loss exhibited a peak gradually shifted from 114 °C down to 60 °C with increasing CN from 0 to 0.04, as we can see from the inset of Figure 4f. This suggested that an appropriate concentration of CN could effectively adjust the T_NR-ER and T_d close to ambient temperature, which could promote the strain value and electrostrictive effect of BNST-based ceramics [44].

To further explore the compositional dependence of the phase evolution, the I-P-E loops of the BNST24-xCN relaxors under ambient conditions were plotted, as shown in Figure 5a–e. Parameters such as P_r, P_max, P_max-P_r and E_c derived from the P-E loops were plotted as a function of CN content in Figure 5f. As shown in Figure 5a, an unsaturated FE-like hysteresis loop with a relatively low P_r value of 18 μC/cm² was obtained for BNST. The I-E loop showed four current peaks at ±E_F and ±E_R, two for each loading quadrant. Under the cycling regime conditions, the peak at +E_F indicated the critical field that the polar order produced under a positive electric field [45]. No peaks appeared upon electric field unloading until the electric field was reversed to − E_R, suggesting that the residual polarization at E = 0, which could be confirmed by the P-E loop. The polar order produced at +E_F was recovered at −E_R. The polarization became zero at the field between −E_R and −E_F. With 0.01 CN doping, a slim and pinched P-E loop (Figure 5b) attributed to the sharp decrease in P_r and E_c (Figure 5f) was observed, suggesting the appearance of the ER phase around E = 0. The presence of a weak P_r value of 10 μC/cm² indicated a retention of the weakly polar NR state in the non-polar ER state. The I-E loop of the BNST-0.01CN ceramic also showed four current peaks at ±E_F and ±E_R, but the current peaks corresponding to ±E_R appeared during electric field unloading, implying that the polar order that occurred at +E_F (−E_F) could be recovered to a non-polar order at +E_R (−E_R) during unloading. With a further increase in CN to 0.04, the progressively decreased P_r and E_c indicated that the non-polar ER state became more stable. Of note, + E_F (−E_F) increased with a further increase in CN content, which reflected that the electric field required for the polarization mechanism became increasingly hindered. These effects suggested that the non-polar order became more stable with higher CN content at RT.

The CN doping induced a gradual transition from the NR to the ER phase, which was well verified by the bipolar strain measurements, as illustrated in Figure 6a. The deformed butterfly shape with a negative strain S_neg of 0.032% at x = 0 implied the NR phase for BNST24 at ambient temperature, which is consistent with the reported research [30]. With the introduction of CN to 0.01, the deformed butterfly-shaped strain hysteresis loop changed into a bud-like loop, accompanied by a sharp decrease in S_neg, d₃₃ and a simultaneous increase in positive strain S_pos, as presented in Figure 6b. With a further increase in CN (x ≥ 0.02), S_neg disappeared and d₃₃ continued to decrease. Moreover, Spos reached its maximum at x = 0.02, but decreased with increasing CN content. The S_neg and d₃₃ were closely correlated with domain back-switching [27], and the continuous decrease in S_neg and d₃₃ indicated that the weakly polar NR order in BNST24 was disrupted by the addition of CN, with a decreasing degree of nonergodicity and an increasing degree of ergodicity. At a critical range around x = 0.2, where ER and NR states are competitive, a large strain of 0.35% was induced by the proximity effect at the ER and NR phase boundary [29] at 70 kV/cm, as shown in Figure 7a. For x ≥ 0.03, the reduction in strain could be attributed to the phase composition in these samples, which was away from the NR-ER phase boundary, and the ER phase was dominant.

To evaluate the suitability of the materials for actuator applications, the unipolar S-E loops and normalized strain d^*₃₃ (d^*₃₃ = S_max/E_max) abstracted from the S-E loops of all investigated BNST24-xCN ceramics were plotted in Figure 7a,b. With increasing CN, the unipolar strain and normalized strain d^*₃₃ initially increased and then diminished after the maximum at x = 0.02, while the corresponding maximum strain and d^*₃₃ were 0.34% and 483 pm/V at 70 kV/cm, respectively. Combining the I-P-E loops and bipolar strain curves, the observed maximum strain and d*₃₃ indicated that large strain values could be induced by the proximity effect around the critical composition (x = 0.02) at ambient temperatures. Beyond the critical composition, either the NR phase or ER phase would dominate, thus reducing the strain response. To highlight the relationship between the strain and T_NR-ER, the d^*₃₃ and T_NR-ER abstracted from the dielectric measurements were plotted in the CN range of 0–0.04 (Figure 7b). As seen in Figure 7b, the doping of CN could decrease the T_NR-ER from 140 °C for BNST24 to RT for x = 0.02 and promote the maximum d^*₃₃ of the BNST24-xCN series. The highest d^*₃₃ was achieved near the critical zone (x = 0.02) when its T_NR-ER approached RT. However, the continued increase in CN beyond 0.02 resulted in decreased strain.

For actuator applications, large strain with low hysteresis η (η =ΔS_hyst/S_max, where the ΔS_hyst is shown in Figure 8a) is very important for the design of high-precision actuators [31,46]. The BNST-0.02CN sample presented a low η value of 8.9% at 80 kV/cm, which was mainly attributed to the intrinsic electrostrictive effect. The electrostrictive effect in dielectric ceramics could be described by the electrostrictive coefficient Q₃₃ (Q₃₃ = S₃/P²₃) [47]. To evaluate the electrostrictive effect of BNST24-xCN (0 ≤ x ≤ 0.04), the electrostrictive curves (S-P² curves) at RT for x = 0, 0.01, 0.02, 0.03 and 0.04 were carried out, as presented in Figure 8b. Evidently, the S-P² curves for x = 0 deviated from linear, while the curves with x ≥ 0.01 were almost linear. The electrostrictive coefficients were 0.0151 m⁴/C² for x = 0 and 0.0245 m⁴/C² for x = 0.02. These features indicated that the electrostrictive effect was enhanced greatly by the CN dopant, which contributed to T_NR-ER and T_d close to ambient temperature.

4. Conclusions

In this work, we utilized a conventional solid state reaction method to fabricate the BNST24-xCN ceramics (0 ≤ x ≤ 0.04). XRD and SEM were employed to investigate the structural properties, revealing the coexistence of rhombohedral and tetragonal phases and the B-site substitution of (Co_0.5Nb_0.5) ⁴⁺ in the perovskite BNST24 ceramics. The ${\epsilon }_r$-T curves indicated that an appropriate CN doping could effectively shift T_NR-ER and T_d downwards close to ambient temperatures, which contributed to an enhanced electrostrictive effect at ambient temperatures. The I-P-E and S-E loops confirmed the gradual transition from NR phase to ER phase with increasing CN doping. According to the S-P² curves of BNST24-xCN (0 ≤ x ≤ 0.04), a high Q₃₃ = 0.0245m⁴/C² was obtained at RT for the 0.02 CN composition. As a result, the composition (x = 0.02) located near the NR-ER phase boundary exhibited a low hysteresis (8.9%) and an enhanced strain value of 0.34% (d^*₃₃ ~483 pm/V). The large strain value was attributed to the proximity effect at the boundary of the ER and NR phases, and the adjustment of T_NR-ER close to RT was imperative for obtaining low hysteresis. Our work effectively balanced the conflict between high strain and low hysteresis and could stimulate further research on developing high-precision actuator applications.