Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry

G Model

ARTICLE IN PRESS

JECS-11289; No. of Pages 11

Journal of the European Ceramic Society xxx (2017) xxx–xxx

Contents lists available at www.sciencedirect.com

Journal of the European Ceramic Society journal homepage: www.elsevier.com/locate/jeurceramsoc

Feature article

Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry Xing Liu, Feng Li, Peng Li, Jiwei Zhai ∗ , Bo Shen, Baihui Liu Key laboratory of Advanced Civil Engineering, Functional Materials Research Laboratory, School of Materials Science & Engineering Tongji University, 4800 Caoan Road, Shanghai 201804, China

a r t i c l e

i n f o

Article history: Received 5 April 2017 Received in revised form 22 May 2017 Accepted 22 May 2017 Available online xxx Keywords: NBT-based ceramics Bi nonstoichiometry Structural transitions Field-induced strain

a b s t r a c t In this study, the Bi-nonstoichiometric 0.99Bix (Na0.8 K0.2 )0.5 TiO3 -0.01SrTiO3 (BNKST) ceramics with x = 0.5–0.535 mol (Bi50-Bi53.5) were prepared by a conventional solid-state reaction method. The effects of Bi excess on structural transition and ferroelectric stability of BNKST ceramics were systematically investigated by the Raman spectra, dielectric analyses and electromechanical measurements. The introduction of excess Bi3+ could significantly break the long-range ferroelectric order and favor the presence of relaxor phase, then the ferroelectric-relaxor transition temperature (TFR ) can be effectively tuned to around room temperature by Bi nonstoichiometry, giving rise to an enhanced room-temperature strain property. The positive strain Spos and dynamic piezoelectric constant d33 * of Bi52.5 critical composition reach 0.33% and 440 pm/V, respectively at 6 kV/mm. The high recoverable strain of Bi52.5 sample can be attributed to the electric-field-induced reversible relaxor-ferroelectric phase transition. The present work may be helpful for further understanding and designing high-performance NBT-based lead-free ceramics for piezoelectric actuator applications. © 2017 Elsevier Ltd. All rights reserved.

1. Introduction Piezoelectric ceramics have been extensively used in ultrasonic transducers, sensors and various electronic devices since their specific function in converting the electrical energy and mechanical energy [1,2]. Lead zirconate titanate (PZT)-based ceramics are the most widely used piezoceramics due to their superior piezoelectric and ferroelectric properties [1,2]. However, there is an increasing demand in developing the lead-free piezoelectric ceramics to replace PZT in recent years due to the toxicity of lead oxide [3–5]. Extensive researches are focusing on three perovskite-type lead-free piezoelectric families: K0.5 Na0.5 NbO3 (KNN), Na0.5 Bi0.5 TiO3 (NBT) and BaTiO3 (BT) [3–6]. Despite several promising findings of lead-free high-performance materials, PZT still works as the dominating piezoelectric materials since one single lead-free compound cannot replace PZT across the whole diverse range of piezoelectric applications [1–6]. Therefore the tailored lead-free systems are sought for individual applications [7,8]. In the case of piezoelectric actuator applications, the performance is largely governed by the electric-field-induced strain [4,5,7,8]. Recently, significant improvement in strain property

∗ Corresponding author. E-mail address: [email protected] (J. Zhai).

(S ∼ 0.45%) was achieved in Na0.5 Bi0.5 TiO3 -BaTiO3 -K0.5 Na0.5 NbO3 (NBT-BT-KNN) pseudo-ternary system as reported by Zhang et al. [9]. Such performance was partially competing with or even surpassing PZT. The large strains have also been developed in other NBT systems such as Na0.5 Bi0.5 TiO3 -Ba(Al0.5 Ta0.5 )O3 (NBT-BAT), Na0.5 Bi0.5 TiO3 -Ba0.85 Ca0.15 Ti0.9 Zr0.1 O3 (NBT-BCTZ), (NBT-KBT-KNN), Na0.5 Bi0.5 TiO3 -K0.5 Bi0.5 TiO3 -K0.5 Na0.5 NbO3 Na0.5 Bi0.5 TiO3 -K0.5 Bi0.5 TiO3 -Ba0.7 Sr0.3 TiO3 (NBT-KBT-BST), etc [10–13]. NBT-based materials are appealing not just from the application point of view but also from the scientific perspective. The underlying physical mechanisms of the large strain properties are still controversial. Several structural models have been suggested, which are based on the X-ray diffraction (XRD), transmission electron microscopy (TEM), Raman spectroscopy, piezoresponse force microscopy (PFM), etc [14–18]. Zhang et al. firstly proposed that the large strains of NBT-BT-KNN system come from the volume change during the field-induced antiferroelectric-ferroelectric phase transition [9]. However, Jo et al. suggested that the large recoverable strain is due to the reversible relaxor (nonpolar)ferroelectric (polar) phase transition without any notable volume change [19,20]. The ferroelectric-relaxor transition temperature (TFR ) can be determined from dielectric spectra using poled samples [5,19,20]. Schütz et al. studied the local structural sensitivity of (Li, Nd)-codoped NBT-KBT ceramics using the in situ

http://dx.doi.org/10.1016/j.jeurceramsoc.2017.05.042 0955-2219/© 2017 Elsevier Ltd. All rights reserved.

Please cite this article in press as: X. Liu, et al., Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry, J Eur Ceram Soc (2017), http://dx.doi.org/10.1016/j.jeurceramsoc.2017.05.042

G Model JECS-11289; No. of Pages 11

ARTICLE IN PRESS X. Liu et al. / Journal of the European Ceramic Society xxx (2017) xxx–xxx

2

Raman spectroscopy [16]. They suggested that the loss of macroscopic ferroelectric order and the peculiar strain properties around TFR can be explained by the loss of orbital hybridization between the 6s2 orbitals of bismuth lone pair and the oxygen p orbitals [16]. Recently, Tan et al. investigated the field-induced relaxor-ferroelectric phase transition of (Sr, Nb)-codoped NBT-KBT ceramics by in situ TEM [17]. They found that the giant strain is related to a reversible transition from the mixed R3c and P4bm nano-sized domains to single R3c lamellar domains [17]. Liu et al. investigated the domain evolution and nanoscale structure of NBTBT-ST ceramics by PFM technique, and the interconversion between the ergodic PNRs and ferroelectric domain was directly observed [18]. These investigations also reveal that the complexity of the crystal structure is an obstacle to fully understand the transition dynamics of NBT-based materials. By introducing a low molar concentration of dopants into NBTbased ceramics, the TFR can be shifted to room temperature and the strain properties will be significantly enhanced [5]. Recently, it was found that the TFR can be tailored by a modulation of the mole ratio of Na, K and Bi, i.e., altering the A-site stoichiometry [21–23]. According to these reports, the giant recoverable strain can be achieved in the nonstoichiometric NBT-BT and NBT-KBT systems with excess in Bi and/or deficiency in (Na, K) [21–23]. On the contrary, the deficiency in Bi and/or excess in (Na, K) will cause the high leakage current and the deterioration of ferroelectric properties [24,25]. The Bi3+ is considered as a replacement candidate of Pb2+ due to their electronic configuration similarities. Both of them have lone electron pair, large ionic radius and large number of electrons, which may contribute to the structural distortions and high piezoelectric response [16,26]. Thus the role of Bi3+ deserves further studies. However, the detailed investigations about the effects of Bi nonstoichiometry on structure and electromechanical properties of NBT-based ceramics are still absent. In this work, we found that the ferroelectric-relaxor transition temperature (TFR ) in NBTbased ceramics can be effectively tuned by Bi nonstoichiometry, and a high room-temperature recoverable strain was obtained in a critical composition. The composition- and temperature-induced structural transitions were systematically investigated, and the mechanism for generating high electrostrain was also discussed.

2. Experimental procedure The 0.99Bix (Na0.8 K0.2 )0.5 TiO3 -0.01SrTiO3 (BNKST) was selected as the research system. Such composition was chosen based on our recent work [27]. The slight addition of SrTiO3 could improve the microstructure and tailor the TFR . The Bi content varies from 0.5 mol to 0.535 mol, abbreviated as Bi50-Bi53.5. The ceramics were fabricated by the normal solid-state sintering method. The raw materials Na2 CO3 , K2 CO3 , Bi2 O3 , SrCO3 and TiO2 (purity >99.9%) were weighed in the appropriate ratio and ball-milled for 12 h. After drying, the mixtures were calcined at 850 ◦ C for 4 h and ball-milled again for 12 h. The dried powders were pressed into disk-shaped samples with 10 mm in diameter and 1 mm in thickness using 8 wt% PVA as the binder. The samples were sintered at 1110–1150 ◦ C for 3 h. During sintering, the samples were buried in the powders of the same composition to minimize the loss of volatile elements. The relative densities were evaluated by the Archimedes method. The phase structures were analyzed by X-ray diffraction (XRD, Bruker D8 Advanced, Germany) with Cu K˛ radiation. The sintered ceramics were polished and thermally etched at 950–1000 ◦ C for 1 h to observe the microstructure morphology using the SEM (JSM, EMP-800; JEOL, Tokoyo, Japan). Raman spectroscopy was measured by a Horiba Lab-Ram iHR550 spectrometer equipped with a Linkam THMSE 600 heating stage. All the samples were wellpolished before the Raman measurements. For electrical properties

characterizations, the silver paste was painted on both surfaces of ceramic pellets and fired at 600 ◦ C for 30 min. The ceramics were poled in silicone oil bath at room temperature under a DC electric field of 5 kV/mm for 15 min. Temperature and frequency dependence of dielectric properties were measured by a computercontrolled high-precision LCR meter (Agilent E4980A, Agilent, Palo Alto, CA) with a heating/cooling rate of 2 ◦ C/min. The piezoelectric constant d33 values were measured by a quasi-static d33 meter (ZJ-6A; Institute of acoustics, China). The polarization curves were tested by the FE test system (Precision Premier II; Radiant Technologies Inc, Albuquerque, NM) connected with a miniature plane-mirror interferometer and the accessory laser interferometric vibrometer (SP-S 120/500). 3. Results and discussion 3.1. Microstructure Fig. 1 shows the SEM micrographs of Bi50-Bi53.5 samples. Fig. 2 shows the relative density and average grain size (via a linear intercept method) of all samples, inset shows the optimum sintering temperature (Tsinter ) as a function of Bi content. All the samples exhibit highly dense and homogeneous microstructure without apparent pores, corresponding to the high relative densities of 97–98%. With the increase of Bi content, the optimum Tsinter decreases gradually due to the low melting point of Bi2 O3 (∼820 ◦ C), while the grain size increases from 0.27 ␮m for Bi50 sample to 0.6 ␮m for Bi52.5 sample, further increase of Bi content leads to the slight decrease of grain size. It is suggested that the moderate excess of Bi may act as sintering aid for grain growth and promote sintering of the ceramics (Bi50-Bi52.5), while the excessive Bi may also aggregate at grain boundary and inhibit the grain growth (Bi53-Bi53.5). 3.2. Phase structure Fig. 3(a) shows the room-temperature XRD patterns of Bi50Bi53.5 samples. Fig. 3(b) and (c) provides the locally magnified (111) and (200) diffraction peaks to give an insight into their phase structure. The XRD data was measured using powders of crushed pellets because the well-polished surfaces will present peak splitting on diffraction patterns due to the ferroelastic 90◦ domains induced by mechanical stress, which is often misunderstood as noncubic lattice distortion [28]. It can be seen that the Bi50-B52.5 samples exhibit pure perovskite structure, while slight impurity phase of Bi2 Ti2 O7 (PDF:32-0118) exists in Bi53 and Bi53.5 samples as reflected by a small undesired peak around 2␪ of 30◦ . The formation of this secondary phase is due to the high deviation of Bi from unity [22,23]. All the studied compositions exhibit pseudocubic symmetry, and an expanded view on (111) and (200) peaks further supports the fact that there is no obvious long-range noncubic distortion. The noncontact and nondestructive Raman spectroscopy is a powerful microprobe for the complex structures due to its coherence length down to the nanometer scale and its sensitivity to local symmetry [29,30]. It has been widely used in clarifying the complex phase transition sequence or establishing phase diagrams of ferroelectric materials [16,29–34]. Fig. 4(a) and (b) shows the roomtemperature Raman spectra of unpoled and poled Bi50-Bi53.5 samples, respectively. All the spectra were fitted by the Lorentzian function using PeakFit software to obtain the peak location, linewidth and intensity of each mode [31–33]. Eight Raman modes can be discerned in the wavenumber range of 100–1000 cm−1 , which can be assigned as P1-P8 modes, respectively. These modes can be divided into three main regions [16,32–35]: (1) modes below 200 cm−1 related to the vibrations of perovskite A-site; (2) modes in

Please cite this article in press as: X. Liu, et al., Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry, J Eur Ceram Soc (2017), http://dx.doi.org/10.1016/j.jeurceramsoc.2017.05.042

G Model JECS-11289; No. of Pages 11

ARTICLE IN PRESS X. Liu et al. / Journal of the European Ceramic Society xxx (2017) xxx–xxx

3

Fig. 1. SEM images of polished and thermally etched surfaces of Bi50-Bi53.5 samples.

Fig. 2. The relative density and average grain size as a function of Bi content; inset shows the optimum sintering temperature (Tsinter ) of each composition.

the wavenumber range of 200–400 cm−1 related to Ti-O vibrations; (3) the high-wavenumber region above 400 cm−1 associated with the vibrations of TiO6 oxygen octahedra. The unpoled Bi50-Bi53.5 samples exhibit similar Raman spectra. After poling, clear changes in P3 and P6 modes for Bi50-Bi52 samples can be distinguished as marked by the dark arrows in Fig. 4(b). However, the Raman spectra of poled Bi52.5-Bi53.5 samples exhibit no changes compared to that of unpoled ones. In order to render more evident the local structural evolution with Bi excess, the intensity ratios of P3 and P2 modes (IP3 /IP2 ), P6 and P2 modes (IP6 /IP2 ) for unpoled and poled Bi50-Bi53.5 samples are shown in Fig. 5(a) and (b), respectively. It can be seen that the intensity ratio exhibits an obvious inflection point around Bi52.5, suggesting the transition nature of this composition. The poling treatment causes a pronounced reduction of IP3 /IP2 and IP6 /IP2 ratios for Bi50-Bi52 samples. By comparison, the IP3 /IP2 and IP6 /IP2 ratios of Bi52.5 and higher-Bi-content samples change little after poling. Such phenomenon demonstrates the field-induced irreversible phase transition in Bi50-Bi52 sam-

ples, while the poling process cannot induce any local distortions in Bi52.5-Bi53.5 samples [5,27]. Fig. 4(c) and (d) show the in situ temperature dependence of Raman spectra for poled Bi50 and Bi52.5 samples, respectively. Note that all the measured Raman spectra have been corrected by the Bose-Einstein population factor: n(w) + 1 = 1/[1−exp(−w/kB T)], where  and kB are Planck’s and Boltzmann’s constants, respectively. Then the temperature dependence of Raman modes can only result from the intrinsic changes of local structure after the Bose-Einstein correction [35,36]. The wavenumber, line-width and intensity of P3 and P6 modes for Bi50 and Bi52.5 samples are summarized in Fig. 6(a)–(c) and (d)–(f), respectively. With the increase of temperature, the P3 and P6 modes of Bi50 sample display notable discontinuities around 70 ◦ C as marked by the dashed lines, which is a signature of the local structural transitions [16,29–35]. It should be noted that this temperature is also close to the ferroelectric-relaxor transition temperature (TFR ) of Bi50 sample determined by the dielectric spectra.

Please cite this article in press as: X. Liu, et al., Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry, J Eur Ceram Soc (2017), http://dx.doi.org/10.1016/j.jeurceramsoc.2017.05.042

G Model JECS-11289; No. of Pages 11 4

ARTICLE IN PRESS X. Liu et al. / Journal of the European Ceramic Society xxx (2017) xxx–xxx

Fig. 3. Room-temperature XRD patterns (a), and locally magnified (111) (b) and (200) (c) diffraction peaks of Bi50-Bi53.5 samples.

Fig. 4. Raman spectra of unpoled (a) and poled (b) Bi50-Bi53.5 samples. Temperature dependence of Raman spectra for poled Bi50 (c) and Bi52.5 (d) samples measured in situ. The multiple peak deconvolutions of unpoled (a) and poled (b)(c) Bi50 sample, poled Bi52.5 (d) sample are also provided.

Thus the sudden changes of Raman modes for poled Bi50 sample around this specific temperature may reveal the loss of long-range ferroelectric order (established by pre-poling) and the presence

of relaxor phase [5,16]. By comparison, the P3 and P6 modes of poled Bi52.5 sample only experience the continuous peak softening accompanied by the peak broadening and reduction in intensity

Please cite this article in press as: X. Liu, et al., Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry, J Eur Ceram Soc (2017), http://dx.doi.org/10.1016/j.jeurceramsoc.2017.05.042

G Model JECS-11289; No. of Pages 11

ARTICLE IN PRESS X. Liu et al. / Journal of the European Ceramic Society xxx (2017) xxx–xxx

Fig. 5. Intensity ratios of P3 and P2 modes (IP3 /IP2 ) (a), P6 and P2 modes (IP6 /IP2 ) (b) for unpoled and poled Bi50-Bi53.5 samples.

with increasing temperature, suggesting the enhanced structural disorder with no phase transitions [31–35]. 3.3. Dielectric analyses Temperature dependence of dielectric properties for poled Bi50, Bi51, Bi52.5 and Bi53 samples measured at 1, 10 and 100 kHz upon heating and cooling (i.e., depoled) are shown in Fig. 7(a)–(d), respectively. Two dielectric anomalies can be seen for all samples during the heating and cooling process: a lower-temperature shoulder around 110 ◦ C (Ts ) with strong frequency dispersion, a diffused higher-temperature dielectric peak around 300 ◦ C (Tm ) with weak frequency dispersion. Jo et al. proposed that the two dielectric peaks are related to the transformations of two types of PNRs (low-temperature PNRs and high-temperature PNRs) [5,20]. The dielectric shoulder at Ts is related to the thermal evolution of lowtemperature PNRs with different symmetries, while the dielectric

5

peak at Tm is due to the mutual transition of the two types of PNRs [5,20]. The composition dependence of Tm and the structural evolutions around Tm are shown in Fig. 8. The Tm varies stably for Bi50-Bi52.5 samples and decreases suddenly for Bi53 and Bi53.5 samples. The dielectric spectra upon heating and cooling process exhibit an obvious dielectric thermal hysteresis, which is a characteristic of the first-order diffuse phase transition [11,37]. However, the thermal hysteresis gradually disappears for Bi53 sample. Pu et al. found that the dielectric hysteresis between heating and cooling process should be attributed to the transition of low-temperature PNRs to high-temperature PNRs [11]. It implies that the excess of Bi favors the presence of high-temperature PNRs at lower temperature (indicated in Fig. 8), which may weaken the contribution of transformation between the two types of PNRs and result in the narrowing of thermal hysteresis range [11]. During the heating process, an obvious difference of dielectric spectra compared to the cooling process is revealed by a low-temperature dielectric anomaly with frequency-independent characteristic. This anomaly is corresponding to the ferroelectricrelaxor transition (marked by TFR ), where the field-induced ferroelectric phase converts back to the relaxor phase [5,19,20]. The TFR indicates the thermal boundary below which the materials can be poled and behave like normal ferroelectrics with a definite macro-piezoelectricity [20]. Above TFR the long-range ferroelectric phase is broken into the relaxor phase due to a thermally enhanced local random field [20]. The TFR as a function of Bi content is shown in Fig. 8. Of particular interest is that the TFR shifts from 76 ◦ C for Bi50 sample to 34 ◦ C for Bi52.5 sample, and disappears for Bi53 sample, indicating that the TFR can be effectively tuned by Bi nonstoichiometry. The Bi52.5 might be the critical composition located at the ferroelectric-relaxor transition boundary, and the improved electromechanical properties can be expected for Bi52.5 sample. The discrepancy of TFR determined by Raman spectra and dielectric measurements can be attributed to their different detection length scales [38]. The dielectric constant (1 kHz) above Tm is fitted by the quadratic law to further investigate the dielectric relaxor characteristics [39,40]: 1 1 (T − Tm )n = + ε εm C

,

where εm is the maximum dielectric constant at Tm , n is the degree of diffuseness, C’ is constant. n has a value ranging from 1 for nor-

Fig. 6. The wavenumber, line-width and intensity of P3 (a)–(c) and P6 (d)–(f) modes as a function of temperature for poled Bi50 and Bi52.5 samples.

Please cite this article in press as: X. Liu, et al., Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry, J Eur Ceram Soc (2017), http://dx.doi.org/10.1016/j.jeurceramsoc.2017.05.042

G Model JECS-11289; No. of Pages 11 6

ARTICLE IN PRESS X. Liu et al. / Journal of the European Ceramic Society xxx (2017) xxx–xxx

Fig. 7. Temperature dependence of dielectric constant (εr ) and loss tangent (tanı) for poled Bi50 (a), Bi51 (b), Bi52.5 (c) and Bi53 (d) samples measured at different frequencies upon heating and cooling. Insets show the linear fitting of dielectric constant above Tm according to the quadratic law.

Fig. 8. The TFR and Tm as a function of Bi content extracted from dielectric spectra on heating process.

mal ferroelectrics to 2 for ideal relaxor ferroelectrics [39,40]. The fitted dielectric curves are shown in the insets of Fig. 7. The n values of Bi50, Bi51, Bi52.5 and Bi53 samples are 1.84, 1.92, 1.87 and 1.86, respectively, indicating that all the samples approach the ideal relaxor ferroelectrics. 3.4. Room-temperature electromechanical properties The room-temperature polarization-electric field (P-E) hysteresis loops, bipolar and unipolar strain (S-E) curves, and current-electric field (I-E) loops of Bi50-Bi53 samples measured at 6 kV/mm and 10 Hz are shown in Fig. 9(a)–(d), respectively. The corresponding polarization, strain and piezoelectric constant val-

ues are summarized in Fig. 10(a)–(c), respectively. The Bi50 and Bi51 samples display the square-like P-E hysteresis loops, butterfly shaped bipolar S-E curves, and two sharp polarization current peaks (denoted as P1 ) on I-E curves with the features of macrodomain switching, which are the typical characteristics of normal ferroelectrics [5,19,41]. The unipolar S-E curves of Bi50 and Bi51 samples exhibit a small hysteresis with low normalized strain values (d33 * = Sunipolar /Emax ) of <260 pm/V, indicating the large amount of intrinsic contributions to the piezoelectric response as confirmed by their relatively high quasistatic d33 values of >150 pC/N [41,42]. With the increase of Bi content, the P-E loops of Bi52 and Bi52.5 samples exhibit the pinched shape along with the appearance of additional two current peaks (denoted as P2 ) on I-E curves. The difference (Pmax -Pr ) between the maximum polarization (Pmax ) and remanent polarization (Pr ) values as shown in Fig. 10(a) increases abruptly at Bi52.5 composition. Then the P2 current peaks on IE curves, which represent the degeneration of ferroelectric phase to relaxor phase upon the removal of external field [43], shift to around zero electric field. The unipolar S-E curve of Bi52.5 sample displays a large hysteresis, indicating the dominating contributions of extrinsic effects [41,42]. Additionally, the Bi52.5 sample develops the sprout-shaped bipolar S-E curve, accompanied by the significantly enhanced positive strain value (Spos ) of 0.33% and dynamic d33 * value of 440 pm/V. However, the quasistatic d33 and negative strain (Sneg ) values, which represent the macro-ferroelectricity and irreversible domain switching [10,12,27], nearly disappear for Bi52.5–Bi53.5 samples. Thus it is reasonable to infer that the excess of Bi content leads to the destabilization of long-range ferroelectric order and favors the presence of disordered relaxor phase [5,19,20]. The ferroelectric order can be established or disrupted reversibly in relaxor matrix under a sufficiently high electric field due to their competitive free energies, thus generating a high recoverable strain

Please cite this article in press as: X. Liu, et al., Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry, J Eur Ceram Soc (2017), http://dx.doi.org/10.1016/j.jeurceramsoc.2017.05.042

G Model JECS-11289; No. of Pages 11

ARTICLE IN PRESS X. Liu et al. / Journal of the European Ceramic Society xxx (2017) xxx–xxx

7

Fig. 9. The polarization (P-E) hysteresis loops (a), current-electric field (I-E) loops (b), bipolar (c) and unipolar (d) strain (S-E) curves of Bi50-Bi53 samples.

[5,19,20]. The Bi52.5 is confirmed as a critical composition separating the ferroelectric phase region and relaxor phase region in Bi-excess BNKST system. Fig. 11 shows the frequency dependence of dielectric constant (εr ) and loss tangent (tanı) of unpoled (insets) and poled Bi50Bi53.5 samples. It can be seen that the poled Bi50-Bi52 samples exhibit apparent resonance and anti-resonance peaks compared to the unpoled ones, which is indicative of the field-induced irreversible phase transition and the residual piezoelectricity after poling [44]. However, the peaks decay gradually with the increase of Bi content and nearly disappear for Bi52.5-Bi53.5 samples. Such phenomenon further illustrates that the Bi50-Bi52 samples exhibit the ferroelectric or weak-ferroelectric phase, while the Bi52.5Bi53.5 samples present the dominating relaxor phase [44].

3.5. Structural model of high recoverable strain Based on the Raman spectra, dielectric analyses and electromechanical properties of Bi-excess BNKST ceramics, a schematic model is presented in Fig. 12 to explain the Bi nonstoichiometry induced structural transitions and high recoverable strains. For the stoichiometric Bi50 sample, the field-induced butterfly shaped bipolar S-E curve, well-saturated P-E hysteresis loop and high quasistatic d33 values are mainly due to the intrinsic contribution, i.e., the converse piezoelectric effect [41,42,45]. During the poling process, the randomly-oriented ferroelectric domains of Bi50 sample are aligned along the direction of applied field as much as possible, causing an irrecoverable lattice distortion and macro-piezoelectricity [42]. However, it is widely accepted that the strain derived from the converse piezoelectric effect in polycrystalline ceramics is always limited [42]. A common strat-

Fig. 10. The Pmax , Pr and Pmax -Pr (a), Spos and Sneg (b), d33 and d33 * (d) values as a function of Bi content extracted from polarization and strain curves.

egy used to enhance the intrinsic contribution includes tailoring the chemical compositions within a polymorphic phase transition

Please cite this article in press as: X. Liu, et al., Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry, J Eur Ceram Soc (2017), http://dx.doi.org/10.1016/j.jeurceramsoc.2017.05.042

G Model JECS-11289; No. of Pages 11 8

ARTICLE IN PRESS X. Liu et al. / Journal of the European Ceramic Society xxx (2017) xxx–xxx

above. The dynamically active PNRs will grow up into the longrange ferroelectric order by merging small ones when the applied external electric field overcomes the effects of local random field. However, this polar ferroelectric state is unstable upon the removal of external field, then the system will relax back to the original state with the assistance of quenched random fields [19]. Thus the high room-temperature strain in Bi52.5 critical composition is assumed to derive from the extrinsic contribution, i.e., the electric-field-induced reversible relaxor-ferroelectric phase transition [41,42,45,46]. In addition, the notable hysteresis indicated in S-E curves of Bi52.5 sample is mainly due to the large nucleation energy barrier that has to be overcome for the long-range ferroelectric phase to be formed in the disordered relaxor matrix. [47,48]. It is worth noting that the electromechanical properties of Bi53 and Bi53.5 samples decay drastically compared to that of Bi52.5 sample, which can be attributed to the higher polarization dynamics of PNRs and larger local random fields, making it more difficult to reestablish the long-range ferroelectric order under the same external electric field [49]. 3.6. Temperature dependence of electromechanical properties

Fig. 11. The frequency dependence of εr and tanı of unpoled (insets) and poled Bi50-Bi53.5 samples.

zone (i.e., two or more ferroelectric phases coexist at room temperature) [3,5,42]. For the Bi-nonstoichiometric BNKST ceramics, the introduction of excess Bi3+ disrupts the charge balance and strengthens the local structural disorder, then the long-range ferroelectric domains are broken into the micro-domains or highly dynamic PNRs [5,19,20], accompanied by the pronounced reduction of piezoelectric and ferroelectric properties as mentioned

The P-E hysteresis loops, bipolar and unipolar S-E curves of Bi50 and Bi52.5 samples measured at different temperatures are shown in Fig. 13(a)–(c), (d)–(f), respectively. The extracted polarization (Pmax and Pr ) values, positive strain (Spos ) and normalized d33 * values are shown in Fig. 13(g)–(i), respectively. The Pmax values of Bi50 and Bi52.5 samples keep relatively stable in the studied temperature range because it is mainly related to the poling state of the field-induced ferroelectric long-range phase [10,12,41]. With the temperature increases to around TFR , the P-E loop of Bi50 sample transforms from rectangle shape to pinched shape, and the bipolar S-E curve transforms from butterfly shape to sprout shape, accompanied by the reduction of Pr values and the significant increase of Spos and d33 * values, suggesting the loss of ferroelectricity and the occurrence of ferroelectric-relaxor phase transition [5]. The

Fig. 12. Model of the Bi nonstoichiometry induced structural transitions and recoverable large strains in BNKST ceramics.

Please cite this article in press as: X. Liu, et al., Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry, J Eur Ceram Soc (2017), http://dx.doi.org/10.1016/j.jeurceramsoc.2017.05.042

G Model JECS-11289; No. of Pages 11

ARTICLE IN PRESS X. Liu et al. / Journal of the European Ceramic Society xxx (2017) xxx–xxx

9

Fig. 13. The P-E hysteresis loops, bipolar and unipolar S-E curves of Bi50 (a)–(c) and Bi52.5 (d)–(f) samples measured at different temperatures. Temperature dependence of Pmax and Pr (g), Spos (h) and d33 * (i) values of Bi50 and Bi52.5 samples.

large strain response (Spos = 0.437% and d33 * = 641 pm/V) of Bi50 sample at TFR is due to the temperature-induced ergodic relaxor phase which can bring the system back to the original state upon the removal of external field [5,19,20]. However, for Bi52.5 sample, the P-E hysteresis loops and strain curves decay gradually, the Spos and d33 * values only decrease progressively with increasing temperature, which is ascribed to the dominating ergodic relaxor characteristic of Bi52.5 sample as mentioned above. It should also be noted that the strain properties of Bi50 sample at TFR are larger than that of Bi52.5 sample at room temperature, although both the two states locate at the ferroelectric-relaxor transition boundary. Such phenomenon can also be observed in many other NBT-based ceramics, which may be caused by the different size or dynamics of PNRs in the local structure [50–53]. Fig. 14(a) and (b) shows the I-E hysteresis loops measured at different temperatures for Bi50 and Bi52.5 samples, respectively. At room temperature, the I-E curve of Bi50 composition only presents two current peaks (P1 ), indicating the domain switching behavior for typical ferroelectrics [43]. When the temperature approaches TFR , another two current peaks (P2 ) shift to around zero electric field for Bi50 sample, revealing the presence of ergodic relaxor phase [43]. For Bi52.5 sample, the I-E curve at room temperature is similar to that of Bi50 sample at 80 ◦ C, suggesting that the two states are both in the critical transition region (ferroelectric-relaxor) [5,43]. Moreover, the P2 peaks of Bi52.5 sample shift to the higher-field region, while the P1 peaks shift to lower-field region, and all the current peaks decay progressively with increasing temperature, which can be attributed to the increasing relaxor phase content and the weakening of ferroelectricity at high temperatures [43].

Fig. 14. The I-E curves of Bi50 (a) and Bi52.5 (b) samples measured at different temperatures.

Please cite this article in press as: X. Liu, et al., Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry, J Eur Ceram Soc (2017), http://dx.doi.org/10.1016/j.jeurceramsoc.2017.05.042

G Model JECS-11289; No. of Pages 11 10

ARTICLE IN PRESS X. Liu et al. / Journal of the European Ceramic Society xxx (2017) xxx–xxx

4. Conclusions In summary, the Bi-nonstoichiometric BNKST ceramics were prepared and the effects of Bi excess on structure and ferroelectric stability were investigated by the Raman spectra, dielectric analyses and electromechanical measurements. The introduction of excess Bi3+ can significantly break the charge balance and disrupt the long-range ferroelectric order, leading to the downward shift of TFR and the enhanced room-temperature strain response. The Bi52.5 critical composition exhibited a relatively high strain of 0.33% and d33 * value of 440 pm/V along with the nearly vanished macro-ferroelectricity, which can be ascribed to the electricfield-induced reversible relaxor-ferroelectric phase transition. The temperature dependence of polarization/strain (P-E and S-E) loops and polarization current (I-E) curves were performed to further investigate the effects of temperature on structural transitions and electromechanical properties. Our work may provide an effective approach to design NBT-based lead-free piezoelectric ceramics with large strains through Bi nonstoichiometry.

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

Acknowledgements The authors would like to acknowledge the National Natural Science Foundation of Chinaunder grant No. 51332003 and 51372171.

[27]

[28]

References

[29]

[1] N. Setter, R. Waser, Electroceramic materials, Acta Mater. 48 (2000) 151–178. [2] D. Damjanovic, Ferroelectric, dielectric and piezoelectric properties of ferroelectric thin films and ceramics, Rep. Prog. Phys. 61 (1998) 1267–1324. [3] J. Wu, D. Xiao, J. Zhu, Potassium-sodium niobate lead-free piezoelectric materials: past, present, and future of phase boundaries, Chem. Rev. 115 (2015) 2559–2595. [4] J. Rödel, W. Jo, K.T.P. Seifert, E.M. Anton, T. Granzow, Perspective on the development of lead-free piezoceramics, J. Am. Ceram. Soc. 92 (2009) 1153–1177. [5] W. Jo, R. Dittmer, M. Acosta, J. Zang, C. Groh, E. Sapper, K. Wang, J. Rödel, Giant electric-field-induced strains in lead-free ceramics for actuator applications-status and perspective, J. Electroceram. 29 (2012) 71–93. [6] W. Liu, X. Ren, Large piezoelectric effect in Pb-free ceramics, Phys. Rev. Lett. 103 (2009) 257602. [7] J. Rödel, K.G. Webber, R. Dittmer, W. Jo, M. Kimura, D. Damjanovic, Transferring lead-free piezoelectric ceramics into application, J. Eur. Ceram. Soc. 35 (2015) 1659–1681. [8] C.H. Hong, H.P. Kim, B.Y. Choi, H.S. Han, J.S. Son, C.W. Ahn, W. Jo, Lead-free piezoceramics-where to move on, J. Materiomics 2 (2016) 1–24. [9] S.T. Zhang, A.B. Kounga, E. Aulbach, H. Ehrenberg, J. Rödel, Giant strain in lead-free piezoceramics Bi0.5 Na0.5 TiO3 -BaTiO3 -K0.5 Na0.5 NbO3 , Appl. Phys. Lett. 91 (2007) 112906. [10] W. Bai, Y. Bian, J. Hao, B. Shen, J. Zhai, The composition and temperature-dependent structure evolution and large strain response in (1-x)(Bi0.5 Na0.5 )TiO3 -xBa(Al0.5 Ta0.5 )O3 ceramics, J. Am. Ceram. Soc. 96 (2013) 246–252. [11] Y. Pu, M. Yao, H. Liu, T. Frömling, Phase transition behavior, dielectric and ferroelectric properties of (1-x)(Bi0.5 Na0.5 )TiO3 -xBa0.85 Ca0.15 Ti0.9 Zr0.1 O3 ceramics, J. Eur. Ceram. Soc. 36 (2016) 2461–2468. [12] J. Hao, B. Shen, J. Zhai, H. Chen, Phase transitional behavior and electric field-induced large strain in alkali niobate-modified Bi0.5 (Na0.80 K0.20 )0.5 TiO3 lead-free piezoceramics, J. Appl. Phys. 115 (2014) 034101. [13] P. Jaita, A. Watcharapasorn, N. Kumar, S. Jiansirisomboon, D.P. Cann, Lead-free (Ba0.70 Sr0.30 )TiO3 -modified Bi0.5 (Na0.80 K0.20 )0.5 TiO3 ceramics with large electric field-induced strains, J. Am. Ceram. Soc. 99 (2016) 1615–1624. [14] M. Hinterstein, M. Knapp, M. Hölzel, W. Jo, A. Cervellino, H. Ehrenberg, H. Fuess, Field-induced phase transition in Bi1/2 Na1/2 TiO3 -based lead-free piezoelectric ceramics, J. Appl. Cryst. 43 (2010) 1314–1321. [15] C. Ma, H. Guo, S.P. Beckman, X. Tan, Creation and destruction of morphotropic phase boundaries through electrical poling: a case study of lead-free (Bi1/2 Na1/2 )TiO3 -BaTiO3 piezoelectrics, Phys. Rev. Lett. 109 (2012) 107602. [16] D. Schütz, M. Deluca, W. Krauss, A. Feteira, T. Jackson, K. Reichmann, Long-pair-induced covalency as the cause of temperature- and field-induced instabilities in bismuth sodium titanate, Adv. Funct. Mater. 22 (2012) 2285–2294. [17] X. Liu, X. Tan, Giant strains in non-textured (Bi1/2 Na1/2 )TiO3 -based lead-free ceramics, Adv. Mater. 28 (2016) 574–578. [18] D. Liu, C. Tian, C. Ma, L. Luo, Y. Tang, T. Wang, W. Shi, D. Sun, F. Wang, Composition, electric-field and temperature induced domain evolution in

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39] [40] [41]

[42]

[43]

[44]

lead-free Bi0.5 Na0.5 TiO3 -BaTiO3 -SrTiO3 solid solutions by piezoresponse force microscopy, Scr. Mater. 123 (2016) 64–68. W. Jo, T. Granzow, E. Aulbach, J. Rödel, D. Damjanovic, Origin of the large strain response in (K0.5 Na0.5 )NbO3 -modified (Bi0.5 Na0.5 )TiO3 -BaTiO3 lead-free piezoceramics, J. Appl. Phys. 105 (2009) 094102. W. Jo, S. Schaab, E. Sapper, L.A. Schmitt, H.J. Kleebe, A.J. Bell, J. Rödel, On the phase density and its thermal evolution of lead free (Bi1/2 Na1/2 )TiO3 -6 mol% BaTiO3 , J. Appl. Phys. 110 (2011) 074106. Y. Guo, M. Gu, H. Luo, Y. Liu, R.L. Withers, Composition-induced antiferroelectric phase and giant strain in lead-free (Nay Biz )Ti1-x O3(1-x) -xBaTiO3 ceramics, Phys. Rev. B 83 (2011) 054118. F. Ni, L. Luo, W. Li, H. Chen, A-site vacancy-induced giant strain and the electrical properties in non-stoichiometric ceramics Bi0.5+x (Na1-y Ky )0.5-3x TiO3 , J. Phys. D: Appl. Phys. 45 (2012) 415103. J.S. Yun, J.S. Park, K.T. Lee, Y.H. Jeong, J.H. Paik, J.H. Cho, Effects of A-site vacancies and lanthanum aluminate substitution on the dielectric and electromechanical properties of (1-y)Bi0.5+x (Na0.78 K0.22 )0.5-3x TiO3 -yLaAlO3 lead-free ceramics, Ceram. Int. 42 (2016) 1015–1019. M. Li, M.J. Pietrowski, R.A.D. Souza, H. Zhang, I.M. Reaney, S.N. Cook, J.A. Kilner, D.C. Sinclair, A family of oxide ion conductors based on the ferroelectric perovskite Na0.5 Bi0.5 TiO3 , Nat. Mater. 13 (2014) 31–35. M. Li, H. Zhang, S.N. Cook, L. Li, J.A. Kilner, I.M. Reaney, D.C. Sinclair, Dramatic influence of A-site nonstoichiometry on the electrical conductivity and conduction mechanisms in the perovskite oxide Na0.5 Bi0.5 TiO3 , Chem. Mater. 27 (2015) 629–634. B.W. Eerd, D. Damjanovic, N. Klein, N. Setter, J. Trodahl, Structural complexity of (Na0.5 Bi0.5 )TiO3 -BaTiO3 as revealed by Raman spectroscopy, Phys. Rev. B 82 (2010) 104112. X. Liu, J. Zhai, B. Shen, F. Li, Y. Zhang, P. Li, B. Liu, Electric-field-temperature phase diagram and electromechanical properties in lead-free (Na0.5 Bi0.5 )TiO3 -based incipient piezoelectric ceramics, J. Eur. Ceram. Soc. 37 (2017) 1437–1447. ˇ M. Otoniˇcar, S.D. Skapin, B. Janˇcar, D. Suvorov, Structural diversity of the (Na1-x Kx )0.5 Bi0.5 TiO3 perovskite at the morphotropic phase boundary, J. Appl. Phys. 113 (2013) 024106. E.K.H. Salje, U. Bismayer, Hard mode spectroscopy: the concept and applications, Phase Transit. 63 (1997) 1–75. J. Kreisel, A.M. Glazer, G. Jones, P.A. Thomas, L. Abello, G. Lucazeau, An x-ray diffraction and Raman spectroscopy investigation of A-site substituted perovskite compounds: the (Na1-x Kx )0.5 Bi0.5 TiO3 (0 ≤ x ≤ 1) solid solution, J. Phys.: Condens. Mater. 12 (2000) 3267–3280. N. Klein, E. Hollenstein, D. Damjanovic, H.J. Trodahl, N. Setter, A study of the phase diagram of (K,Na,Li)NbO3 determined by dielectric and piezoelectric measurements, and Raman spectroscopy, J. Appl. Phys. 102 (2007) 014112. D. Rout, K.S. Moon, S.J.L. Kang, I.W. Kim, Dielectric and Raman scattering studies of phase transitions in the (100-x)Na0.5 Bi0.5 TiO3 -xSrTiO3 system, J. Appl. Phys. 108 (2010) 084102. C. Chen, H. Zhang, H. Deng, T. Huang, X. Li, X. Zhao, Z. Hu, D. Wang, H. Luo, Electric field and temperature-induced phase transition in Mn-doped Na1/2 Bi1/2 TiO3 −5.0 at.%BaTiO3 single crystals investigated by micro-Raman scattering, Appl. Phys. Lett. 104 (2014) 142902. K. Datta, A. Richter, M. Göbbels, R.B. Neder, B. Mihailova, Atomistic origin of huge response functions at the morphotropic phase boundary of (1-x)(Na0.5 Bi0.5 )TiO3 -xBaTiO3 , Phys. Rev. B 90 (2014) 064112. M. Acosta, L.A. Schimitt, L.M. Luna, M.C. Scherrer, M. Brilz, K.G. Webber, M. Deluca, H.J. Kleebe, J. Rödel, W. Donner, Core-shell lead-free piezoelectric ceramics: current status and advanced characterization of the Bi1/2 Na1/2 TiO3 -SrTiO3 system, J. Am. Ceram. Soc. 98 (2015) 3405–3422. S. Yoon, H.L. Liu, G. Schollerer, S.L. Cooper, P.D. Han, D.A. Payne, S.W. Cheong, Z. Fisk, Raman and optical spectroscopic studies of small-to-large polaron crossover in the perovskite manganese oxides, Phys. Rev. B 58 (1998) 2795–2801. F. Craciun, C. Galassi, R. Birjega, Electric-field-induced and spontaneous relaxor-ferroelectric phase transitions in (Na1/2 Bi1/2 )1-x Bax TiO3 , J. Appl. Phys. 112 (2012) 124106. A.M. Mazuera, P.S. Silva, J.A.D. Rodrigues, P.S. Pizani, Y.R. Barcelay, M. Venet, M. Algueró, Origin of discrepancy between electrical and mechanical anomalies in lead-free (K,Na)NbO3 -based ceramics, Phys. Rev. B 94 (2016) 184101. L.E. Cross, Relaxor ferroelectrics, Ferroelectrics 76 (1987) 241–267. K. Uchino, S. Nomura, Critical exponents of the dielectric constants in diffused-phase-transition crystals, Ferroelectr. Lett. 44 (1982) 55–61. G. Dong, H. Fan, J. Shi, M. Li, Composition- and temperature-dependent large strain in (1-x)(0.8Bi0.5 Na0.5 TiO3 -0.2Bi0.5 K0.5 TiO3 )-xNaNbO3 ceramics, J. Am. Ceram. Soc. 98 (2015) 1150–1155. J. Fu, R. Zuo, Giant electrostrains accompanying the evolution of a relaxor behavior in Bi(Mg,Ti)O3 -PbZrO3 -PbTiO3 ferroelectric ceramics, Acta Mater. 61 (2013) 3687–3694. Y. Ehara, N. Novak, S. Yasui, M. Itoh, K.G. Webber, Electric-field-temperature phase diagram of Mn-doped Bi0.5 (Na0.9 K0.1 )0.5 TiO3 ceramics, Appl. Phys. Lett. 107 (2015) 262903. F. Li, J. Zhai, B. Shen, X. Liu, K. Yang, Y. Zhang, P. Li, B. Liu, H. Zeng, Influence of structural evolution on energy storage properties in Bi0.5 Na0.5 TiO3 -SrTiO3 -NaNbO3 lead-free ferroelectric ceramics, J. Appl. Phys. 121 (2017) 054103.

Please cite this article in press as: X. Liu, et al., Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry, J Eur Ceram Soc (2017), http://dx.doi.org/10.1016/j.jeurceramsoc.2017.05.042

G Model JECS-11289; No. of Pages 11

ARTICLE IN PRESS X. Liu et al. / Journal of the European Ceramic Society xxx (2017) xxx–xxx

[45] W. Zhao, R. Zuo, J. Fu, M. Shi, Large strains accompanying field-induced ergodic phase-polar ordered phase transformations in Bi(Mg0.5 Ti0.5 )O3 -PbTiO3 -(Bi0.5 Na0.5 )TiO3 ternary system, J. Eur. Ceram. Soc. 34 (2014) 2299–2309. [46] N.H. Khansur, M. Hinterstein, Z. Wang, C. Groh, W. Jo, J.E. Daniels, Electric-field-induced strain contributions in morphotropic phase boundary composition of (Bi1/2 Na1/2 )TiO3 -BaTiO3 during poling, Appl. Phys. Lett. 107 (2015) 242902. [47] W. Zhao, R. Zuo, J. Fu, X. Wang, L. Li, H. Qi, D. Zheng, Enhanced rhombohedral domain switching and low field driven high electromechanical strain response in BiFeO3 -based relaxor ferroelectric ceramics, J. Eur. Ceram. Soc. 36 (2016) 2453–2460. [48] C.W. Ahn, G. Choi, I.W. Kim, J.S. Lee, K. Wang, Y. Hwang, W. Jo, Forced electrostriction by constraining polarization switching enhances the electromechanical strain properties of incipient piezoceramics, NPG Asia Mater. 9 (2017) 346.

11

[49] J.E. Daniels, W. Jo, J. Rödel, V. Honkimäki, J.L. Jones, Electric-field-induced phase-change behavior in (Bi0.5 Na0.5 )TiO3 -BaTiO3 -(K0.5 Na0.5 )NbO3 : a combinatorial investigation, Acta Mater. 58 (2010) 2103–2111. [50] S.T. Zhang, A.B. Kounga, E. Aulbach, W. Jo, T. Granzow, H. Ehrenberg, J. Rödel, Lead-free piezoceramics with giant strain in the system Bi0.5 Na0.5 TiO3 -BaTiO3 -K0.5 Na0.5 NbO3 . II. Temperature dependent properties, J. Appl. Phys. 103 (2008) 034108. [51] Y. Guo, H. Fan, J. Shi, Origin of the large strain response in ternary SrTi0.8 Zr0.2 O3 modified Bi0.5 Na0.5 TiO3 -Bi0.5 K0.5 TiO3 lead-free piezoceramics, J. Mater. Sci. 50 (2015) 403–411. [52] R.A. Malik, A. Hussain, A. Maqbool, A. Zaman, T.K. Song, W.J. Kim, M.H. Kim, Giant strain, thermally-stable high energy storage properties and structural evolution of Bi-based lead-free piezoceramics, J. Alloys Compds. 682 (2016) 302–310. [53] W. Bai, D. Chen, Y. Huang, P. Zheng, J. Zhong, M. Ding, Y. Yuan, B. Shen, Z. Ji, Temperature-insensitive large strain response with a low hysteresis behavior in BNT-based ceramics, Ceram. Int. 42 (2016) 7669–7680.

Please cite this article in press as: X. Liu, et al., Tuning the ferroelectric-relaxor transition temperature in NBT-based lead-free ceramics by Bi nonstoichiometry, J Eur Ceram Soc (2017), http://dx.doi.org/10.1016/j.jeurceramsoc.2017.05.042