G Model
ARTICLE IN PRESS
JECS-10564; No. of Pages 8
Journal of the European Ceramic Society xxx (2016) xxx–xxx
Contents lists available at www.sciencedirect.com
Journal of the European Ceramic Society journal homepage: www.elsevier.com/locate/jeurceramsoc
Enhanced rhombohedral domain switching and low field driven high electromechanical strain response in BiFeO3 -based relaxor ferroelectric ceramics Wanli Zhao a , Ruzhong Zuo a,∗ , Jian Fu a , Xiaohui Wang b , Longtu Li b , He Qi a , Donggeng Zheng a a b
Institute of Electro Ceramics & Devices, School of Materials Science and Engineering, Hefei University of Technology, Hefei 230009, PR China State Key Lab of New Ceramics and Fine Processing, Department of Materials Science and Engineering, Tsinghua University, Beijing 100084, PR China
a r t i c l e
i n f o
Article history: Received 21 January 2016 Received in revised form 27 February 2016 Accepted 5 March 2016 Available online xxx Keywords: Relaxor ferroelectric Strain Phase transition Domain switching
a b s t r a c t A low-field driven high electromechanical strain response (large-signal d33 * of ∼1100 pm/V at 3.5 kV/mm) was found in a new (0.75-x)BiFeO3 -0.25PbTiO3 -xPb(Mg1/3 Nb2/3 )O3 ternary relaxor ferroelectric ceramic. In-situ synchrotron X-ray diffraction measurements suggested that such a large strain be ascribed to a collective effect of electric field induced PNRs’ growth, domain switching and rhombohedral-tetragonal (R–T) phase transition with increasing fields, among which ergodic PNRs’ growth into R ferroelectric microdomains dominated the fastest increase of strains and the formation of the maximum strain hysteresis. Most interestingly, obviously enhanced R domain orientation along the electric field direction was believed to make a unique contribution as a result of its slightly reduced lattice distortion, compared with other Bi-containing relaxor ferroelectrics. An illustration of the domain morphology evolution in an ergodic relaxor was depicted to disclose the formation of strain hysteresis based on the delay in dynamics during electric cycling. © 2016 Elsevier Ltd. All rights reserved.
1. Introduction Bismuth-based perovskite ferroelectric materials have attracted lots of attention in recent years either in the field of hightemperature piezoelectric materials [1–3] or large-strain ceramic actuators [4–9]. Particularly, giant strains of ∼0.4% were reported in a few Bi-based perovskite relaxor ferroelectric systems such as Asite complexly occupied (Na0.5 Bi0.5 )TiO3 (BNT) based compositions [4–8], and B-site complexly occupied Bi(Mg0.5 Ti0.5 )O3 (BMT) based compositions [9,10]. The generation of large strains was usually found to be associated with the evolution of the dielectric relaxor behavior and mainly ascribed to the reversible phase transition from an ergodic relaxor state to a long-range ferroelectric state [5–11]. This mechanism has been recently updated by contributing the transition from polar nanoregions (PNRs) to short-range polar order, to the c-axis oriented growth of PNRs [12]. Unfortunately, the field strength required for triggering giant strains in these Bi-based perovskite ferroelectrics is very large (∼6–8 kV/mm), such that the
∗ Corresponding author. E-mail addresses:
[email protected],
[email protected] (R. Zuo).
value of large-signal d33 * (normalized strain Smax /Emax ) is usually less than 600 pm/V. As an important member of the Bi-based perovskite family, BiFeO3 (BF) owns a high Curie temperature (Tc = 1100 K) and a rhombohedral (R) structure with an R3c space group at room temperature (RT) [13,14]. A huge shift of Bi3+ and Fe3+ ions, a counter rotation of oxygen octahedra along the (1 1 1) direction and a similar electronic structure of Bi3+ to Pb2+ have enabled it to have intrinsically high spontaneous polarization [13,14]. A large polarization value of ∼100 C/cm2 was reported for BF thin films due to the strain effect introduced by the lattice mismatch between the film and the substrate [15]. This strain effect was believed to transform the structure of BF films into a mixed phase state with stripe-shaped R phase embedded in tetragonal (T) matrix [16,17], resulting in a large electric field induced strain of over 5%. On the other hand, theoretical predictions based on first-principles calculations or ab initio calculations have shown that the large polarization value of BF is inherently high and relatively insensitive to strain [18,19]. This conclusion was later on proved by the measurements on good single crystals as well as bulk ceramics [20,21]. However, only a small strain (∼0.07%) was obtained in BF ceramics measured at 60 kV/cm, i.e., well below the coercive field of 100 kV/cm [22]. Recently, a large peak-to-peak bipolar strain of up
http://dx.doi.org/10.1016/j.jeurceramsoc.2016.03.002 0955-2219/© 2016 Elsevier Ltd. All rights reserved.
Please cite this article in press as: W. Zhao, et al., Enhanced rhombohedral domain switching and low field driven high electromechanical strain response in BiFeO3 -based relaxor ferroelectric ceramics, J Eur Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.03.002
G Model JECS-10564; No. of Pages 8
ARTICLE IN PRESS W. Zhao et al. / Journal of the European Ceramic Society xxx (2016) xxx–xxx
)
BF (7 0 /30 PT )(M PB
PT NB) PM (MP ) /32
(68
0.0 PT 1.0 0.1 0.9 0.2 0.8 0.3 0.7 T 0.4 0.6 0.5 0.5 0.6 0.4 MPB 0.7 0.3 0.8 0.2 0.9 0.1 R 1.0 0.0 BF0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 PMN Fig. 1. Schematic phase diagram of (0.75-x)BF-0.25PT-xPMN ternary system.
to 0.36% was obtained in pure BF ceramics by applying a large electric field at a low frequency [23]. Such a large strain shows a strong frequency dependence, indicating a strong extrinsic contribution [24]. Ex-situ X-ray diffraction (XRD) measurements have shown a strong (111) domain texture in the poled BF sample [25], in which an obvious remanent strain would make this large strain less useful as a result of irreversible domain switching. In this work, a new ternary system (0.75-x)BiFeO3 -0.25PbTiO3 xPb(Mg1/3 Nb2/3 )O3 ((0.75-x)BF-0.25PT-xPMN) was constructed. All composition points were located in the proximity of a theoretically predicted morphotropic phase boundary (MPB) line for the BF-PTPMN ternary system, as plotted in Fig. 1 [26,27]. On the one hand, the addition of PT can reduce the conductivity and the coercive field of BF. On the other hand, the addition of a canonical PMN relaxor can enhance the disordered distribution of different ions at A and/or B sites of the BF perovskite lattice, inducing the formation of local random fields. The mechanism of generating large strains and the origin of strain hysteresis were further explored by means of in-situ high-resolution synchrotron XRD in combination with the calculation of lattice parameters and unit cell volumes under different external electric fields. 2. Experimental (0.75-x)BF-0.25PT-xPMN (x = 0.2–0.75) ceramics were prepared by a conventional solid-state reaction method using high-purity (>99%) Bi2 O3 , Fe2 O3 , TiO2 , PbO, Nb2 O5 and (MgCO3 )4 ·Mg(OH)2 ·5H2 O as raw-materials. The starting powders were weighed according to the stoichiometric formula and ball-milled for 6 h in ethanol with zirconia balls. The dried powders were calcined twice at 820 ◦ C for 2 h and then ball milled again for 8 h. The powders were pressed into pellets with 10 mm in diameter and the pellets were sintered in sealed crucibles at 1050–1200 ◦ C for 2 h. The major surfaces of the sintered pellets were well polished, then coated with silver paste and finally fired at 550 ◦ C for 30 min. The phase structure of crushed samples was examined by a conventional powder X-ray diffractometer (XRD, D/Max-RB, Rigaku, Tokyo, Japan) using Cu K˛ radiation at RT. Dielectric properties for both unpoled and poled samples were measured as a function of temperature and frequency by an LCR meter (Agilent E4980A, Santa Clara, CA). The polarization (P), polarization current density (J) and strain (S) under bipolar or unipolar electric fields (E) were measured by using a ferroelectric measuring system (Precision multiferroelectric, Radiant Technologies Inc., Albuquerque, NM) with an accessory laser interferometer vibrometer (AE SP-S
(100)
x=0.75
Intensity (a.u.)
2
(110) (111)
(200)
(211) (210)
(a)
(b) x=0.75 x=0.50
x=0.72
x=0.35
x=0.68
x=0.20
x=0.60
38.5
x=0.50
(200)
39.2
(c)
x=0.45
x=0.75
x=0.35
x=0.50
x=0.25 x=0.20 20
(111)
25
30
x=0.35 x=0.20
35
40
45
2θ (deg.)
50
55
60 45.0 45.9
Fig. 2. (a) XRD patterns of (0.75-x)BF-0.25PT-xPMN ceramics at RT and the locally scanned (b) (1 1 1) and (c) (200) diffraction peaks.
120E, SIOS Metechnik, GmbH, Ilmenau, Germany). For in-situ synchrotron XRD measurement, well-polished samples with ∼40 m in thickness and ∼8 mm in diameter were annealed in air at 550 ◦ C for 2 h to eliminate the strains. After that, gold electrodes were sputtered onto two major sides of the ceramic disks. High-resolution XRD was taken at Shanghai Synchrotron Radiation Facility (SSRF) using beam line 14B1 ( = 1.2378 Å). An X-ray beam with dimensions 0.3 mm × 0.3 mm was used to irradiate the samples in the direction parallel to the electric field direction with increasing the electric field amplitude from 0 kV/mm to 5 kV/mm, and then decreasing it from 5 kV/mm to 0 kV/mm. The diffraction peak profiles were fitted using a pseudo-Voigt peak shape function and the crystal symmetry of the samples can be described according to the peak splitting. 3. Results 3.1. Composition dependence of crystal structure and microstructure Fig. 2 shows RT XRD patterns of (0.75-x)BF-0.25PT-xPMN ceramics. It is obvious from Fig. 2(a) that all the studied compositions possessed a pure perovskite structure without apparent secondary phases. Sharp and narrow peak profiles with no obvious splitting suggested a pseudocubic (PC) structure for samples with x < 0.75. For the x = 0.75 sample, the splitting of the (1 1 1) peak, as more obviously seen from Fig. 2(b), was observed while the (2 0 0) peak maintained a single one (Fig. 2(c)), indicating an R symmetry for this composition. Therefore, the addition of BF into 0.25PT-0.75PMN tended to decrease the lattice distortion of the R phase. Moreover, a gradual shift of the diffraction peaks to lower angles was observed with increasing the PMN content, indicating a slight lattice expansion. This is probably due to relatively large ionic radii of Pb2+ , Mg2+ and Nb5+ compared with those of Bi3+ and Fe3+ ions (CN = 12, RPb = 1.49 Å > RBi = 1.45 Å; CN = 6, RMg = 0.72 Å > RNb = 0.64 Å > RFe = 0.55 Å) [28]. Fig. 3 shows the grain morphology of (0.75-x)BF-0.25PT-xPMN ceramics sintered at their optimum temperatures. It can be seen that the addition of PMN promoted the densification of BF-PT binary system. The relative density slightly varied from ∼95% to ∼97.5% as x changed from 0.20 to x = 0.75. Moreover, the grain size changed obviously as a result of the addition of PMN. It decreased from ∼6.1 m at x = 0.20 to ∼3.6 m at x = 0.75. The variation of grain size and sample density in (0.75-x)BF-0.25PT-xPMN ternary system is probably attributed to the fact that the relative content of Bi2 O3 and PbO may influence both the amount of liquid phases
Please cite this article in press as: W. Zhao, et al., Enhanced rhombohedral domain switching and low field driven high electromechanical strain response in BiFeO3 -based relaxor ferroelectric ceramics, J Eur Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.03.002
G Model JECS-10564; No. of Pages 8
ARTICLE IN PRESS W. Zhao et al. / Journal of the European Ceramic Society xxx (2016) xxx–xxx
3
Fig. 3. SEM micrographs of (0.75-x)BF-0.25PT-xPMN ceramics sintered at their optimum sintering temperatures: (a) x = 0.2, x = 0.35, x = 0.45, (d) x = 0.60, (e) x = 0.68 and x = 0.75.
and the temperature of forming liquid phases during sintering. The improvement of the sample density and the refinement of grain morphology with increasing x would benefit to the enhancement of the insulation properties. 3.2. Evolution of dielectric and ferroelectric properties in (0.75-x)BF-0.25PT-xPMN The temperature and frequency dependence of the dielectric permittivity (ε) for both unpoled and poled (0.75-x)BF-0.25PTxPMN ceramics is shown in Fig. 4(a–f), respectively. Fig. 4(g) shows a comparison of dielectric permittivity before and after poling using the x = 0.25 sample as an example. On the one hand, the frequency dependence of the maximum dielectric permittivity increased with increasing the PMN or BF content in the ternary system, revealing the enhancement of the relaxation degree of this ternary system, as further illustrated by two typical parameters: Trelax = Tm,1MHz -Tm,10kHz , and the diffuseness degree using the modified Curie-Weiss law [29], as shown in Fig. 4(h). On the other hand, an additional dielectric anomaly was detected after poling for samples with x = 0.20–0.35 and 0.72–0.75 as indicated by arrows (Figs. 4(d-g)). The temperature Tf,r at the field induced dielectric anomaly was generally called as the depolarization temperature Td . By comparison, no obvious distinction before and after poling could be detected for 0.40 ≤ x ≤ 0.70 samples, meaning that these
samples would exist in an ergodic relaxor state at RT. A freezing temperature Tf was obtained by fitting the measured dielectric permittivity versus temperature curves at different frequencies for unpoled samples to the Vogel-Fulcher relationship [30], as shown in Fig. 4(i). According to the definition of these two critical temperatures [9,30], the coexistence of ergodic and nonergodic phases should occur between Tf,r and Tf . Figs. 5(a-d) depict P-E loops as well as the corresponding J-E curves measured at 1 Hz at RT. Obviously, square P-E loops together with single sharp J-E peaks were observed for compositions with x <0.30 and x >0.70, indicating these samples are nonergodic at RT. The large Pmax value (>35 C/cm2 ) for all investigated compositions (Fig. 5(e)) denotes that a complete long-range ferroelectric order state was established under the applied electric fields. Low Pr values of <10 C/cm2 were obtained for samples with 0.40 ≤ x ≤ 0.65 and simultaneously P–E loops became very slim, corresponding to flat and broad J–E peaks. All of these are typical features for ergodic relaxors. By comparison, a split J–E peak (as indicated by P1 and P2 ) was observed for samples with x = 0.30, 0.35, 0.68 and 0.70, which was usually ascribed to the coexistence of nonergodic and ergodic relaxors [9,31]. The electric field values at J–E peaks of different samples are plotted in Fig. 5(f), which can more clearly indicate the phase transition at RT with changing the PMN content. For polarization current platforms (ergodic relaxors), there are no corresponding electric field values shown in Fig. 5(f).
Please cite this article in press as: W. Zhao, et al., Enhanced rhombohedral domain switching and low field driven high electromechanical strain response in BiFeO3 -based relaxor ferroelectric ceramics, J Eur Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.03.002
G Model
ARTICLE IN PRESS
JECS-10564; No. of Pages 8
W. Zhao et al. / Journal of the European Ceramic Society xxx (2016) xxx–xxx
Poled 3
8
Unpoled 2.5 (g)
(d) f
4
(b)
(e)
Tf,r
2.0 1 kHz 1.5
30
x=0.45 x=0.60 x=0.68
8
f
1.0
1 MHz x=0.25 45 60 75 90 o
Temperature ( C) 1.8
4
30
(h)
20
γ
3
ε (x10 )
0
Tf,r
Poled
o
x=0.20 x=0.25 x=0.35
0 15 (c)
o
T ( C)
x=0.70 x=0.72
10 5
Tf,r
0 100 200 300 400
10
1.6
(f)
100 o200 300 400
ΔTrelax ( C)
Unpoled
(a) 12
ε (x10 )
4
0 320 (i) 240 Tf 160 80 RT 0 Tf,r 0.25 0.40 0.55 0.70
Temperature ( C)
PMN content, x
2
0.8
(b)
(d)
(c)
x=0.45 x=0.72
(e)
40 30 20 10 0 (f) -2.7
P max Pr E1 E2
-1.8
0.0
-0.8
x=0.35 x=0.70 2
(a)
x=0.30 x=0.68
P ( μC/cm )
25 0 -25
x=0.20 x=0.60
E (kV/mm)
J ( μA/cm ) P ( μC/cm 2)
Fig. 4. Temperature and frequency dependence of the dielectric permittivity for (a–c) unpoled and (d–f) poled (0.75-x)BF-0.25PT-xPMN ceramics as indicated, (g) a comparison of the dielectric permittivity for the x = 0.25 sample before and after poling, (h) the Trelax and values, and (i) the Tf and Tf,r values with varying the PMN content.
-0.9
P2 P1 -4 -2 0 2 4
-4 -2 0 2 4
Electric field (kV/mm)
0.0 0.3
0.5
0.7
PMN content, x
Fig. 5. (a–b) P-E hysteresis loops and (c-d) J-E curves of (0.75-x)BF-0.25PT-xPMN ceramics measured at a frequency of 1 Hz, (e) the variation of the maximum polarization Pmax and remanent polarization Pr , and (f) the electric field values E1 and E2 at the P1 and P2 peaks, respectively.
3.3. Compositional dependence of electrostrains in (0.75-x)BF-0.25PT-xPMN Fig. 6(a, b) show the S–E curves from non-first cycles under bipolar and unipolar electric fields at RT, respectively. It should be noted that the measurement between each cycle was done discontinuously with an interval of few seconds. Therefore, the achieved maximum unipolar strain is repeatable although a small remanent strain can be observed in Fig. 6(b). According to our previous studies [10,32], the formation of this small remanent strain should be due to the time effect of field induced long-range ferroelectric order back to the initial ergodic state. It could completely decay with time very shortly after the electric field was totally released. The positive strain (Spos ), negative strain (Sneg ) and the d33 * value were plotted as a function of the PMN content, as shown in Fig. 6(c). Typical butterfly bipolar S-E loops were observed for samples with x ≤ 0.25 and x ≥ 0.72, showing an apparent Sneg . A further increase of the PMN or BF content in the ternary system resulted in a decrease of the Sneg value and an increase of the Spos value. The Sneg value reached nearly zero in the range of x = 0.35–0.68. During the above process,
the bipolar strain loop changed from a butterfly shape to a sprout shape. The change in Sneg and the strain shape should be ascribed to the appearance of ergodic relaxor phases with increasing the intersubstitution of PMN and BF. The maximum Spos value of ∼0.47% and ∼0.33% under an electric field of 5 kV/mm (d33 * = ∼940 pm/V and ∼550 pm/V, respectively) were reached in x = 0.35 and x = 0.68 samples, respectively. Although these two samples are just located at the ergodic and nonergodic phase boundary, a much higher Spos or d33 * value appeared in the x = 0.35 sample. The PMN-rich composition (x = 0.68 sample) only owned a comparable electrostrain or d33 * value to that of previously reported BNT-based and BMT-based relaxor ferroelectrics [5–9]. It is worthy of note that the appearance of large electrostrains was accompanied by a drastic decrease in the Sneg value. Similar phenomena have been also observed in other Bi-containing perovskite relaxor ferroelectrics [5,9]. Furthermore, Fig. 6(d) shows the effect of the applied electric field magnitude on unipolar strains and d33 * values of the x = 0.35 sample. The strain hysteresis (Hys.) was calculated from the ratio of the widest part of each unipolar S-E loop over the maximum strain value under corresponding electric fields (Hys. = Wmax /Smax ). The
Please cite this article in press as: W. Zhao, et al., Enhanced rhombohedral domain switching and low field driven high electromechanical strain response in BiFeO3 -based relaxor ferroelectric ceramics, J Eur Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.03.002
G Model
ARTICLE IN PRESS
JECS-10564; No. of Pages 8
0
2
4
0.2 0.0 0
0.0
-0.2 0.2
2
1
2
3
4
5
Spos
*
d33
Sneg 0.3
0.4
0.5
0.6
PMN content, x
0.7
5
6
7
2
3
4
5
60 30
Eef 1
*
0 4
x=0.35
0.0 0
9 6 3 0
3
0.2 0.1
Electric field (kV/mm) (c) 0.4 0.2
1
Electric field (kV/mm) (e)
0.3
-8
Electric field (kV/mm) (b)
0.0 0
6
7
Electric field (kV/mm) (f) x=0.35
0
Hys. (%)
-2
4
x=0.35
Es
2
8
0.6
S33 0.4
Sv 0
1
2
3
4
5
S31 6
7
Strain (%)
-4
d33 x10 (pm/V)
12
(d)
2
Strain (%)
x=0.45 x=0.72
0.2
0.0
-0.2
Strain (%)
5
0.4
0.2
0.4
x=0.35 x=0.68
0.6
dS/dE (10 ) m/V Strain (%)
(a)
x=0.25 x=0.60
*
0.4
x=0.20 x=0.50
d33 x10 (pm/V)
Strain (%)
W. Zhao et al. / Journal of the European Ceramic Society xxx (2016) xxx–xxx
0.2 0.0 -0.2
Electric field (kV/mm)
Fig. 6. (a) Bipolar S-E loops and (b) unipolar S-E loops for (0.75-x)BF-0.25PT-xPMN ceramics measured at RT at a frequency of 1 Hz, (c) the variation of Spos , Sneg and d33 * with the PMN content, (d) the field dependent unipolar strain loops and the corresponding d33 * values, (e) the strain hysteresis and the dS/dE value as a function of electric field and (f) unipolar strains S33 and S31 and volume strain Sv of the x = 0.35 sample. Here S33 and S31 are the strains measured parallel and perpendicular to the electric field, respectively.
rate of the increase of strains in response to electric fields (dS/dE) during the period of increasing the field was achieved by making the derivative of the lower-half unipolar strain loop with respect to the electric field. Both Hys. and dS/dE are shown in Fig. 6(e). It can be seen that the absolute strain values and d33 * exhibited an obvious electric field dependence. A significantly enhanced electrostrain value of ∼0.55% could be generated under an electric field of 7 kV/mm for the x = 0.35 sample. A high d33 * value of ∼1100 pm/V at a relatively low electric field of 3.5 kV/mm was obtained in the x = 0.35 sample, which would make BF-based relaxor ferroelectrics more interesting than other Bi-containing perovskite materials. Moreover, both Hys. and dS/dE also exhibited a strong dependence on the applied electric field, reaching their maximum values at the same electric fields of ∼2.2 kV/mm for the x = 0.35 sample. This indicates that both of them should be dominated by the similar physical mechanism. As the electric field is above 2.2 kV/mm, both Hys. and dS/dE started to decrease. Compared with the field value Eef at the maximum dS/dE or Hys., the d33 * value reached the maximum at Es ∼ 3 kV/mm and then declined with further increasing the electric field. In the low field range, Spos was very small and almost linearly changed with the electric field up to 1 kV/mm (dS/dE is a constant before 1 kV/mm). After that, Spos started to increase quickly within a certain field range (1–3 kV/mm), and finally again increased almost linearly at higher fields. However, its volume strain Sv was calculated by using the formula Sv = S33 + 2S31 , as shown in Fig. 6(f). It can be seen that the volume strain was nearly zero during loading the electric field before E reached 2 kV/mm. After that, it started to increase relatively quickly and finally exhibited a slow and linear increase once again. That is to say, the changes generated in the low field and high field ranges did not exhibit an obvious volume effect. As for the x = 0.35 sample, the strain hysteresis was nearly zero as the electric field was below 1 kV/mm, meaning that a slight and linear strain (∼1.2% of total strains) probably came from the electrostrictive effect. Afterwards, the strain and its hysteresis value rapidly increased with electric field, indicating the beginning of the contribution of ergodic state to ferroelectric phase transition at 1 kV/mm. It became fastest with electric field at ∼2.2 kV/mm and should approximately end at a slightly higher electric field
(∼3 kV/mm) where d33 * values started to decay. That is to say, the electric field induced ergodic relaxor to ferroelectric phase transition was dominated in the field range of 1–3 kV/mm, generating strains of ∼55.7% of the total ones. As the field was above 3 kV/mm, the remaining strain proportion should be ascribed to possible R-T ferroelectric phase transition and further domain switching of ferroelectric macrodomains, taking up ∼43.1% of total strain values. The E = Es -Eef values might denote the difficulty of the subsequent ferroelectric phase transition (ferroelectric R microdomains to T macrodomains) and the ferroelectric T macrodomain switching. The E = 0.8 kV/mm in current material system is smaller than that for other Bi-based perovskite relaxor ferroelectrics [9,32]. 4. Discussion 4.1. In-situ synchrotron XRD during loading and unloading electric fields To make clear what really contributes to such a large electrostrain in the present study, the structural evolution of the x = 0.35 sample under the application of an in-situ electric field was investigated as shown in Fig. 7. An initial PC symmetry at E = 0 kV/mm could be confirmed by single (1 1 1) and (2 0 0) diffraction peaks, as more clearly seen in Fig. 7(b). As E < 1 kV/mm, only a slight increase of the diffraction peak intensity was observed while the peak positon kept almost unchanged. The change at this stage should be ascribed to the increase of the number of stably existing PNRs at the expense of the nonpolar matrix because the application of external electric fields tended to lower the energy fluctuation of the system. As 1 ≤ E ≤ 2.5 kV/mm, both peak position and intensity were altered with increasing electric field. The shift towards lower angles indicates a slight expansion of the crystal lattice. An asymmetric (1 1 1) peak started to be observed from 1 kV/mm up to 2.5 kV/mm and simultaneously a single (2 0 0) peak was maintained, suggesting that an R symmetry appeared at 1 kV/mm and then became totally resolved at 2.5 kV/mm. In the field range of E = 0–1 kV/mm, PNRs are so small that their intrinsic R symmetry cannot be resolved by x-ray till the number of PNRs is large enough
Please cite this article in press as: W. Zhao, et al., Enhanced rhombohedral domain switching and low field driven high electromechanical strain response in BiFeO3 -based relaxor ferroelectric ceramics, J Eur Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.03.002
G Model
ARTICLE IN PRESS
JECS-10564; No. of Pages 8
W. Zhao et al. / Journal of the European Ceramic Society xxx (2016) xxx–xxx
6
(a)
(111) (b) x=0.35
E=5 kV/mm (h) (111)t (002)t
(200)
5.0 kV/mm 4.5 kV/mm 4.0 kV/mm 3.5 kV/mm 3.0 kV/mm 2.5 kV/mm 2.0 kV/mm
Intensity (a.u.)
Intensity (a.u.)
2.0 kV/mm 3.0 kV/mm 4.0 kV/mm
Increase
1.0 kV/mm
Decrease
0.0 kV/mm
1.5 kV/mm
E=4 kV/mm
(f)
E=3 kV/mm
measured fitted
R T
(e)
E=2.5 kV/mm
(-111)r (200)r (d)
E=1 kV/mm
(111)r (c)
1.0 kV/mm 0.5 kV/mm 0.0 kV/mm
31.0
(g)
(111)pc
31.1 35.7 36.0 36.3 2θ (deg.)
(200)t
31.04
E=0 kV/mm PC
(200)pc
31.12
35.8
36.2
2θ (deg.)
0.35
aR
(a)
4.01
aT
0.28
cT
0.21
4.00
aPC
0.14
3.99
Δv/v
0.07
0.012 0.010
0.00
(b) 60
δR
0.008
δT
0.006 0.004
45
fR
30
0.002 0.000
Δv/v (%)
4.02
0
1
fR (%)
o
δr, δt
at 1 kV/mm. However, PNRs are still weakly correlated at this stage. With further increasing the electric field up to 2.5 kV/mm, which is necessary for overcoming the local random field, PNRs started to grow up quickly by merging small ones [9], leading to both the fastest increase of strain with electric field and the largest strain hysteresis near 2.2 kV/mm (Fig. 6(e)). It seems that there is no detectable orientation of PNRs along the direction of electric fields before 2.5 kV/mm (see Fig. 7(c–e)). As E > 2.5 kV/mm, R ferroelectric microdomains started to exhibit an obvious orientation along the electric field direction, which was believed to contribute to the continuous increase of d33 * up to a special field Es (∼3 kV/mm). Therefore, the field induced relaxor to ferroelectric phase transition can be generally considered as the PNRs’ growth (1–2.5 kV/mm) first and then orientation (2.5–3 kV/mm) within the field range of 1 ∼ 3 kV/mm. Such a large extrinsic contribution of the field forced PNRs growth, taking up ∼35.3% of the total strain, would be responsible for the formation of the largest strain hysteresis approximately at 2.2 kV/mm. It also made the strain increment to deviate from a linear response to the electric field as E ≥ 1 kV/mm. As E = 3 kV/mm, the (1 1 1) diffraction line further shifted to lower angles and its intensity continuously increased. At the same time, the (2 0 0) diffraction lines started to split, meaning part of R phases transformed into T phases. The (1 1 1) peak intensity was extremely enhanced probably owing to the orientation of R domains, and reached the maximum approximately at 4.5 kV/mm as a result of the completeness of both the R domain switching and the number of remaining R phases. With an increase of the applied electric field, the induced ferroelectric T phases were also found to exhibit an obvious domain orientation along the electric field direction in addition to the increase of their amounts. The further R to T phase transition and c-axis orientation of T domains would definitely continue to contribute to the formation of electrostrains in addition to R domain orientation as E was above 3 kV/mm, such that a giant strain of ∼0.55% was obtained at 7 kV/mm as shown in Fig. 6(d). During 3 ≤ E < 5 kV/mm, a downward bending of the strain curve (Fig. 6(d)) and a decrease of dS/dE were observed (Fig. 6(e)), indicating that the R-T phase transition and further T domain switching
Lattice parameters (A)
Fig. 7. Evolution of (a) (1 1 1) and (b) (2 0 0) diffraction lines during loading and unloading electric fields for the x = 0.35 sample and the (c–h) the (1 1 1) and (2 0 0) diffraction peaks fitted by using PeakFit software during loading different electric fields as indicated.
2
3
4
5
15
Electric field (kV/mm) Fig. 8. The calculated (a) lattice parameters and unit cell volume variation and (b) the lattice distortion ␦R , ␦T for R and T phases, respectively and the volume fraction FR of R domains oriented along the electric field direction for the x = 0.35 sample.
contributed to the strain at a lower rising rate than the electric field induced growth of PNRs into R microdomains. The (1 1 1) and (2 0 0) diffraction peak profiles of the x = 0.35 sample under different electric fields were fitted by a pseudo-Voigt function (Figs. 7(c-h)). The calculated lattice parameters and cell volume strain as a function of electric field are shown in Fig. 8(a). It can be seen that an R phase could be resolved from a PC matrix from 1 kV/mm and became single till 2.5 kV/mm. Part of R phases were transformed into T phases from 3 kV/mm. The coexistence of R and T phases remained up to 5 kV/mm in the studied electric field range. The variation of unit cell volume v/v was calculated using v/v = (vR · FR + vT · FT -vPC(0) )/vPC(0) where vR , FR and vT , FT stand for the unit cell volume and the volume fraction of R and T phases, respectively, and vPC(0) is the unit cell volume of PC phase at zero field. In the field range of 1–2.5 kV/mm, the mixed two phases are PC and R. As a result, vT and FT in the above formula should be replaced by vPC and FPC , respectively. The results indicated that
Please cite this article in press as: W. Zhao, et al., Enhanced rhombohedral domain switching and low field driven high electromechanical strain response in BiFeO3 -based relaxor ferroelectric ceramics, J Eur Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.03.002
G Model JECS-10564; No. of Pages 8
ARTICLE IN PRESS W. Zhao et al. / Journal of the European Ceramic Society xxx (2016) xxx–xxx
7
Fig. 9. (a–j) Illustration of domain morphology and phase structure with first increasing and then decreasing the electric field, and (k) the unipolar strain loop indicating the corresponding step during electric cycling for the x = 0.35 sample.
v/v is almost zero before 2 kV/mm, keeping a good consistency with the measurement of the macroscopic volume strain (Fig. 6(f)). This would further indicate that the electric field forced PC to R transition is inherently a size effect caused by the growth of PNRs, instead of a real phase transition. It started to increase visibly after 2.5 kV/mm and reached ∼0.19% at 5 kV/mm with the same order of magnitude as measured in Fig. 6(f), suggesting the occurrence of R to T phase transition. Moreover, the lattice distortion ıR and ıT for R and T phases and the volume fraction (fR ) of R domains oriented along the electric field direction were plotted as a function of electric field, as shown in Fig. 8(b). It can be seen that ıT exhibited an increase with electric field as usually found in a few relaxor ferroelectrics [33,34]. Interestingly, poling field did not induce an increase of ıR , but a slight reduction, which would facilitate the (1 1 1) orientation of R domains, as observed in Fig. 7. The fR value was calculated by fR = I(1 1 1) /(I(111) + I(-1 1 1) ). It can be seen that the poling field has resulted in a fast increase in the proportion of R domains oriented along the electric field direction (∼28% at 2.5 kV/mm to ∼70% at 5 kV/mm). It is worthy of note that fr is ∼27% at 2.5 kV/mm, hinting that R domains formed under this electric field are almost randomly distributed (fr approximates to 25% for a completely random R domain). Therefore, a strong (1 1 1) texture of R domains at higher fields was believed to largely contribute to giant strains observed in BF-PT-PMN ternary system. By comparison, other Bi-based relaxor ferroelectrics usually exhibited little increase of fr with increasing the electric field prior to the R to T phase transition [9,34]. This unusually observed phenomenon might be inherently correlated with the structure of BF in which a huge shift of Bi3+ and Fe3+ ions and a counter rotation of oxygen octahedra were reported to occur along the (1 1 1) direction [13,14]. 4.2. Models of domain morphology evolution for an ergodic relaxor under the application of an external electric field As shown in Fig. 7, an increase of the (1 1 1) peak intensity was firstly observed as E was reduced down to 4 kV/mm owing to the reversibility of R-T phase transition. Afterwards, the decrease of the (1 1 1) peak intensity and the blurring of the (2 0 0) peak splitting occurred as E was below 4 kV/mm as a result of the weakening of R domain orientation. At E = 2 kV/mm, the (2 0 0) diffraction line looked wide together with an asymmetric (1 1 1) peak, suggesting that R and T phases still occur. As the electric field was completely released, single (1 1 1) and (2 0 0) diffraction peaks could
be observed, meaning that an initial PC phase was recovered. These observations demonstrated that the phase structure is different at the same electric field during loading and unloading, particularly in the field range of 1–3 kV/mm where the largest strain hysteresis was observed, although the electric field induced changes proved to be reversible. The above mentioned transition process from initial relaxor state to the final long-range ferroelectric state is schematically illustrated in Fig. 9(a–j). Each figure represents the end state of the corresponding stage during electric cycling. These stages were also displayed in a corresponding unipolar S-E loop, as shown in Fig. 9(k). It is obvious that the evolution of domain morphology during electric cycling demonstrated an obvious delay, which could be basically ascribed to the formation of the observed strain hysteresis. The delay of the transition process in response to the electric field might be related to the existence of local random fields. Woodward et al. [26] found that the R phase in BF-PT ceramics consist of ∼50 nm-sized domains, much larger than the PNRs in PMN relaxor [35]. The introduction of PMN would obviously increase the local random field and further make PNRs growth hard. However, a weaker local random field would make the backward switching process lazier, tending to increase the strain hysteresis. Therefore, a medium local random field in the x = 0.35 sample is probably the reason for the simultaneous achievement of both relatively low driving field for giant strains (only at 3.5 kV/mm for d33 * ∼1100 pm/V) and low strain hysteresis (∼45% at 7 kV/mm) compared with BNT or BMT based relaxor ferroelectrics [4–6]. 5. Conclusions Significantly enhanced electrostrains were found to be accompanied by the evolution of dielectric relaxor behavior in a new (0.75-x)BF-0.25PT-xPMN solid solution system. A high electrostrain of ∼0.22%–0.55% was obtained in the field range between 2.5 kV/mm and 7 kV/mm, generating a maximum normalized strain d33 * of ∼1100 pm/V at 3.5 kV/mm in the x = 0.35 sample. A collective effect of the electric field induced growth of PNRs (PNRs to ferroelectric R microdomains), the switching of domains and the RT phase transition (ferroelectric microdomains to macrodomains) was identified as the physical mechanism of generating large strains by means of in-situ synchrotron XRD, among which the PNRs growth under a relatively low electric field contributed to ∼35.3% of the total strains and dominated the formation of strain hysteresis. Most interestingly, an obvious (1 1 1) orientation of R domains
Please cite this article in press as: W. Zhao, et al., Enhanced rhombohedral domain switching and low field driven high electromechanical strain response in BiFeO3 -based relaxor ferroelectric ceramics, J Eur Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.03.002
G Model JECS-10564; No. of Pages 8
ARTICLE IN PRESS W. Zhao et al. / Journal of the European Ceramic Society xxx (2016) xxx–xxx
8
prior to the R-T phase transition was believed to make a unique contribution to the strain in this study. The above-mentioned transition process proved to be reversible for an ergodic relaxor but exhibited a serious delay in dynamics during the period of increasing and decreasing the electric field, as schematically illustrated by the evolution of the domain morphology. Acknowledgments Financial support from the National Natural Science Foundation of China (Grant Nos. 51472069, U1432113, 50402079) and the Anhui Provincial Natural Science Foundation (1508085JGD04) is gratefully acknowledged. References [1] M.R. Suchomel, P.K. Davies, Predicting the position of the morphotropic phase boundary in high temperature PbTiO3 -Bi(B B )O3 based dielectric ceramics, J. Appl. Phys. 96 (2004) 4405–4410. [2] C.J. Stringer, T.R. Shrout, C.A. Randall, I.M. Reaney, Classification of transition temperature behavior in ferroelectric PbTiO3 -Bi(Me Me )O3 solid solutions, J. Appl. Phys. 99 (2006) 024106. [3] L.P. Kong, G. Liu, S.J. Zhang, W.G. Yang, Origin of the enhanced piezoelectric thermal stability in BiScO3 -PbTiO3 single crystals, Appl. Phys. Lett. 106 (2015) 232901. [4] X.M. Liu, X.L. Tan, Giant strains in non-textured (Bi1/2 Na1/2 )TiO3 -based lead-free ceramics, Adv. Mater. 28 (2016) 574–578. [5] J.G. Hao, B. Shen, J.W. Zhai, C.Z. Liu, X.L. Li, X.Y. Gao, Large strain response in 0.99(Bi0.5 Na0.4 K0.1 )TiO3 -0.01(Kx Na1-x )NbO3 lead-free ceramics induced by the change of K/Na ratio in (Kx Na1-x )NbO3 , J. Am. Ceram. Soc. 96 (2013) 3133–3140. [6] J. Shi, H.Q. Fan, X. Liu, Q. Li, Giant strain response and structure evolution in (Bi0.5 Na0.5 )0.945-x (Bi0.2 Sr0.70.1 )x Ba0.055 TiO3 ceramics, J. Eur. Ceram. Soc. 34 (2014) 3675–3683. [7] J.E. Daniels, W. Jo, J. Rödel, J.L. Jones, Electric-field-induced phase transformation at a lead-free morphotropic phase boundary: case study in a 93%(Bi0.5 Na0.5 )TiO3 -7%BaTiO3 piezoelectric ceramic, Appl. Phys. Lett. 95 (2009) 032904. [8] W. Jo, R. Dittmer, M. Acosta, J.D. 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. [9] W.L. Zhao, R.Z. Zuo, J. Fu, Temperature-insensitive large electrostrains and electric field induced intermediate phases in (0.7-x)Bi(Mg1/2 Ti1/2 )O3-x Pb(Mg1/3 Nb2/3 )O3- 0.3PbTiO3 ceramics, J. Eur. Ceram. Soc. 34 (2014) 4235–4245. [10] W.L. Zhao, R.Z. 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. [11] 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. [12] R.Z. Zuo, F. Li, J. Fu, D.G. Zheng, W.L. Zhao, H. Qi, Electric field forced c-axis oriented growth of polar nanoregions and rapid switching of tetragonal domains in BNT-PT-PMN ternary system, J. Eur. Ceram. Soc. 36 (2016) 515–525. [13] F. Kubel, H. Schmid, Structure of a ferroelectric and ferroelastic monodomain crystal of the perovskite BiFeO3 , Acta Crystallogr. B 46 (1990) 698–702. [14] W.M. Zhu, H.Y. Guo, Z.G. Ye, Structural and magnetic characterization of multiferroic (BiFeO3 )1-x (PbTiO3 )x solid solutions, Phys. Rev. B 78 (2008) 014401.
[15] J. Wang, J.B. Neaton, H. Zheng, V. Nagarajan, S.B. Ogale, B. Liu, D. Viehland, V. Vaithyanathan, D.G. Schlom, U.V. Waghmare, N.A. Spaldin, K.M. Rabe, M. Wuttig, R. Ramesh, Epitaxial BiFeO3 multiferroic thin Film heterostructures, Science 299 (2003) 1719–1722. [16] J.X. Zhang, B. Xiang, Q. He, J. Seidel, R.J. Zeches, P. Yu, S.Y. Yang, C.H. Wang, Y.H. Chu, L.W. Martin, A.M. Minor, R. Ramesh, Large field-induced strains in a lead-free piezoelectric material, Nat. Nanotechnol. 6 (2011) 98–102. [17] J. Zhou, M. Trassin, Q. He, N. Tamura, N. Kunz, C. Cheng, J.X. Zhang, W.I. Liang, J. Seidel, C.L. Hsin, Y.H. Chu, J.Q. Wu, Directed assembly of nano-scale phase variants in highly strained BiFeO3 thin films, J. Appl. Phys. 112 (2012) 064102. [18] J.B. Neaton, C. Ederer, U.V. Waghmare, N.A. Spaldin, K.M. Rabe, First-principles study of spontaneous polarization in multiferroic BiFeO3 , Phys. Rev. B 71 (2005) 014113. [19] C. Ederer, N.A. Spaldin, Effect of epitaxial strain on the spontaneous polarization of thin Film ferroelectrics, Phys. Rev. Lett. 95 (2005) 257601. [20] D. Lebeugle, D. Colson, A. Forget, M. Viret, P. Bonville, J.F. Marucco, S. Fusil, Room-temperature coexistence of large electric polarization and magnetic order in BiFeO3 single crystals, Phys. Rev. B 76 (2007) 024116. [21] V.V. Shvartsman, W. Kleemann, R. Haumont, J. Kreisel, Large bulk polarization and regular domain structure in ceramic BiFeO3 , Appl. Phys. Lett. 90 (2007) 172115. [22] T. Rojac, M. Kosec, B. Budic, N. Setter, D. Damjanovic, Strong ferroelectric domain-wall pinning in BiFeO3 ceramics, J. Appl. Phys. 108 (2010) 074107. [23] T. Rojac, M. Kosec, D. Damjanovic, Large electric-field induced strain in BiFeO3 ceramics, J. Am. Ceram. Soc. 94 (2011) 4108–4111. [24] D. Damjanovic, Logarithmic frequency dependence of the piezoelectric effect due to pinning of ferroelectric-ferroelastic domain walls, Phys. Rev. B 55 (1997) 649–652. [25] T. Rojac, A. Bencan, B. Malic, G. Tutuncu, J.L. Jones, J.E. Daniels, D. Damjanovic, BiFeO3 ceramics: processing, electrical, and electromechanical properties, J. Am. Ceram. Soc. 97 (2014) 1993–2011. [26] D.I. Woodward, I.M. Reaney, R.E. Eitel, C.A. Randall, Crystal and domain structure of the BiFeO3 -PbTiO3 solid solution, J. Appl. Phys. 94 (2003) 3313–3318. [27] B. Noheda, D.E. Cox, G. Shirane, J. Gao, Z.G. Ye, Phase diagram of the ferroelectric relaxor (1-x)PbMg1/3 Nb2/3 O3 -xPbTiO3 , Phys. Rev. B 66 (2002) 054104. [28] R.D. Shannon, Revised effective ionic radii and systematic studies of interatomie distances in halides and chaleogenides, Acta Crystallogr. A 32 (1976) 751–767. [29] K. Uchino, S. Nomura, Critical exponents of the dielectric constants in diffused-phase-transition crystals, Ferroelectr. Lett. 44 (1982) 55–61. [30] D. Viehland, S.J. Jang, L.E. Cross, M. Wuttig, Freezing of the polarization fluctuations in lead magnesium niobate relaxors, J. Appl. Phys. 68 (1990) 2916–2921. [31] G. Viola, H.P. Ning, X.J. Wei, M. Deluca, A. Adomkevicius, J. Khaliq, M.J. Reece, H.X. Yan, Dielectric relaxation, lattice dynamics and polarization mechanisms in Bi0.5 Na0.5 TiO3 -based lead-free ceramics, J. Appl. Phys. 114 (2013) 014107. [32] D.G. Zheng, R.Z. Zuo, Relaxor-normal ferroelectric phase transition and significantly enhanced electromechanical strain behavior in Bi(Ni1/2 Ti1/2 )O3 -PbTiO3 -Pb(Mg1/3 Nb2/3 )O3 ternary system close to the morphotropic phase boundary, J. Eur. Ceram. Soc. 35 (2015) 3485–3493. [33] B. Noheda, Z. Zhong, D.E. Cox, G. Shirane, S.E. Park, P. Rehrig, Electric-field-induced phase transitions in rhombohedral Pb(Zn1/3 Nb2/3 )1-x Tix O3 , Phys. Rev. B 65 (2002) 224101. [34] 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. [35] D.S. Fu, H. Taniguchi, M. Itoh, S.Y. Koshihara, N. Yamamoto, S. Mori, Relaxor Pb(Mg1/3 Nb2/3 )O3 : a ferroelectric with multiple inhomogeneities, Phys. Rev. Lett. 103 (2009) 207601.
Please cite this article in press as: W. Zhao, et al., Enhanced rhombohedral domain switching and low field driven high electromechanical strain response in BiFeO3 -based relaxor ferroelectric ceramics, J Eur Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.03.002