Free Piezoelectric Single Crystals

J. Am. Ceram. Soc., 98 [10] 2988–2996 (2015) DOI: 10.1111/jace.13723 © 2015 The American Ceramic Society

Journal

Seed-Free Solid-State Growth of Large Lead-Free Piezoelectric Single Crystals: (Na1/2K1/2)NbO3

Minhong Jiang,‡,§ Clive A. Randall,‡,† Hanzheng Guo,‡ Guanghui Rao,§ Rong Tu,¶ Zhengfei Gu,§ Gang Cheng,§ Xinyu Liu,§ Jinwei Zhang,§ and Yongxiang Lik ‡

Center for Dielectrics and Piezoelectrics, Materials Research Institute, The Pennsylvania State University, University Park, Pennsylvania 16802

§

Guangxi Key Laboratory of Information Materials, Department of Materials Science and Engineering, Guilin University of Electronic Technology, Guilin, Guangxi 541004, China ¶

State Key Lab of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, Wuhan 430070, China

k

The Key Lab of Inorganic Functional Materials and Devices, Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai 200050, China of PZT, and its PZN-PT counterpart ceramics.9 This is also true for the lead-free compositions.10,11 As exemplified by Na1/2Bi1/2TiO3-BaTiO3 single crystal, an excellent piezoelectric coefficient d33 483 pC/N is reported, which is almost comparable to that of PZT-5H (590 pC/N).12 Currently, nearly all of the lead-free single crystals are prepared by conventional methods, such as top-seeded solution method,13 and high-temperature melt growth method, which includes the Czochralski, zone melting and cold-crucible routes. However, these single crystal growth routes can be only adopted for a few material systems due to their technical limitations, such as the incongruent melting-induced compositional inhomogeneity, high-temperature and long-time process-induced volatility for Na2O and K2O species,14–18 as well as the highenergy consumption, high cost, and low production yields. In 1950, the solid-state crystal growth (SSCG) technology was first developed in metals and later transferred to metal oxides.19, 20 Using a seed template single crystal with a similar structure, a new single crystal will grow into the polycrystalline grains of the ceramic through a solid-state recrystallization at melting temperatures. Recently SSCG has been successfully utilized in producing crystals of Al2O3,21 Mn-Zn ferrite,22 Nd:YAG,23 BaTiO3,24,25 PMN-PT,26 PMNPZT,27 BS-PMNT,28 NBT-BT29, and NKN.10,30,31 However, in addition to the complicated preparation process for the seed crystal, the quality of the grown single crystal is also strongly affected by the seed crystal.32–35 Therefore, a seedfree method may be explored and provide an efficient way to grow the single crystals. In our recent work, we found exaggerated NKN grains (hereafter we call this a single crystal in this study), which has a dimension comparable to the macroscopic scale single crystals, are available in a relatively narrow composition range with doping low content of LiBiO3. Using this seedfree method, these single crystals can be, either easily or reproducibly, prepared by traditional grain growth process. In order to better investigate such an abnormal crystal growth phenomenon, a comprehensive study is presented here regarding to the crystal growth kinetics and the electrical properties.

Large Na0.5K0.5NbO3 (NKN) piezoelectric single crystals were obtained by seed-free solid-state crystal growth method, which is a traditional sintering grain growth process, with LiBiO3 used as a sintering aid. The largest dimension of the single crystals obtained was 11 mm 3 9 mm 3 3 mm. In addition to the LiBiO3 doping content, temperature, and time effect of the crystal growth process was systematically investigated and considered from the kinetics point of view. With the assistance of Avrami analysis, parameters relevant to the crystal growth process were determined. Laue diffraction and transmission electron microscopy suggested an orthorhombic symmetry for the single crystalline structure. Dielectric-frequency-temperature measurements revealed an orthorhombic-tetragonal and tetragonal-cubic phase transition at 155°C and 405°C, respectively, both of which are typical of first-order transitions, and have a well-defined thermal hysteresis. Rayleigh analysis was performed regarding to the extrinsic reversible and nonreversible piezoelectric properties, and the result suggested a dominant intrinsic reversible piezoelectric contribution of 91.5% under E0 = 1 kV/cm AC amplitude. Such a single crystal growth process route is low cost and a relative simple preparation process.

I.

L

Introduction

EAD-FREE

piezoelectric ceramics have been extensively studied for replacing the lead-based ceramics due to an international effort driven by the environment pollution effects of lead and certain legal restrictions on the use of toxic lead.1,2 After the first report by Saito et al.,2,3 Na0.5K0.5NbO3 (NKN) family, which is one of the most promising lead-free candidates, have been widely studied due to its high Curie temperature, low dielectric loss, relatively favorable piezoelectricity, as well as the ability to process these materials under low partial pressures of oxygen and cofire with base metal electrodes.2–8 Generally, the properties of piezoelectric single crystals are far better than that of some ceramics. For example, in a lead-based system, the piezoelectric constant d33 of PZN-8% PT single crystal is 2500 pC/N, which is 5 to 6 times to that

G. Brennecka—contributing editor

II.

Experimental Approach

(1) Sample Preparation (1x)Na0.5K0.5NbO3-xLiBiO3 (x = 0.001, 0.002, 0.003, 0.004, 0.005, and 0.006) were prepared by a solid-state reaction

Manuscript No. 36276. Received January 20, 2015; approved May 25, 2015. † Author to whom correspondence should be addressed. e-mail: [email protected]

2988

October 2015

2989

Solid-State Growth NKN Crystal

method using reagent-grade Nb2O5 (>99.5%), K2CO3 (>99%), Na2CO3 (>99.8%), Li2CO3 (>97%), and Bi2O3 (>99%) powders. Na2CO3, K2CO3, Li2CO3, and Bi2O3 were provided by Xilong Chemical Co., Ltd. (Shantou, Guangdong province, China), Nb2O5 was provided by Sinopharm Chemical Reagent Co., Ltd. (Shanghai city, China). After backing at 200°C before weighing out the designed stoichiometric ratios, the formulated powders were mixed by ball-milling in ethanol for 24 h, then dried and calcined in a temperature of 750°C for 6 h in an air atmosphere. The calcined powders were further ball milled for 8 h and then pressed into a disk-shaped pellets with 26–45 mm in diameter and 2–3 mm in thickness under a 100 MPa pressure. The pellets were sintered in a peak temperature range 1080°C–1120°C for over 3 h in air. Selective as-grown single crystals were removed, oriented, and polished. The single crystals were mechanically cut by a low speed diamond saw. The initial size of the single crystals cut is about 2–3 mm in depth. Platinum electrodes were sputtered on both (001)-oriented surfaces of each single crystal for electrical characterization.

(2) Characterization Methods The structure and chemical composition of the single crystal samples were characterized by X-ray diffraction (XRD; PANalytical Empyrean and D/MAX2500PC, PANalytical Inc., Westborough, MA) with CuKa radiation, scanning electron microscopy (SEM; FEI Nova NanoSEM 630, Eindhoven, the Netherlands), energy dispersive spectrometer (EDS; X-MaxN, OXFORD instruments, South Windsor, CT), respectively. The EDS analysis was performed with standard-less quantification. It should be noted that the Li is not available for the EDS detection. The orientation of single crystal was determined by Laue system (Multiwire Laboratory, Ltd., Ithaca, NY). Bulk ceramic disk with exaggerated grain was mechanically thinned and polished to ~100 lm thick. Specimens for transmission electron microscopy (TEM) were prepared via standard procedures including grinding, cutting, dimpling, and ion milling. Then, the disks were further thinned with an Ar-ion mill until a perforation was formed. TEM studies were performed on a Philips EM420 microscopy (Philips, Eindhoven, the Netherlands) operated at 120 kV. The temperature-dependent dielectric constant and loss tangent were measured by a LCR meter (Model HP4274A, Hewlett-Packard, Palo Alto, CA) within 25°C–500°C. Polarization-field (P–E) and schematic minor strain-field (S–E) hysteresis loops were also measured, respectively at 10 and 1 Hz with sine waves, using a modified Sawyer-Tower circuit and a linear variable differential strain transducer for the Rayleigh analysis.36–39

III.

Results and Discussion

(1) Optimizing the Crystal Growth Condition by Varying LiBiO3 Content Figure 1 shows the photograph of (1x) NKN-xLiBiO3 (x = 0.001–0.006) samples sintered at 1120°C for 10 and 15 h, respectively. An exaggerated grain growth phenomenon is observed in the samples of x = 0.004, as exemplified by the exaggerated grain and ceramic matrix areas. For other compositions, the typical polycrystalline ceramic matrix appearance is only present. Figure 2 shows the natural surface SEM pictures for such a morphology evolution. With the increasing of x, the grain size increases, and reaches a flat configuration without interior grain boundaries in x = 0.004. Then, grainy morphology appears again with further increasing of x. Therefore, the LiBiO3 content plays an important role in inducing the exaggerated grains in the matrix. The appropriate addition of bismuth oxide with high diffusion coefficient and low melting temperatures form effect liquid phase, which facilitates transportation, solution-dissolution and aids the crystal growth rates. The phase and structural details of the (1x) NKN-xLiBiO3 powder calcined at 750°C for 6 h and bulk samples sintered 1120°C for 10 h is shown in the form of X-ray powder diffraction patterns in Fig. 3. For direct comparison, each of the XRD data series is normalized to the same maximum value. A typical single phase perovskite with an orthorhombic structure is observed for all compositions. Most importantly, it is interesting to notice that there is a highly texture-like orientation along the direction, as indicated by an abnormal strong 002 peak around 32°, appears in x = 0.004 [Fig. 3(b)]. For x = 0.003 [Fig. 3(b)], a nonrandom diffraction patterns are observed, which indicate a weak polycrystalline nature; while random diffraction patterns typical of polycrystalline is revealed for x ≥ 0.005. Since the spectra for all compositions are diffracted from the same X-ray beam size, the above observations indicate that large grains or single crystal(s) dominate the diffraction with nonrandom pattern. This would be appropriately described as a lack of sufficient diffracting grains to produce random statistics. Also, it should be noted that a distinct difference of the diffraction peak intensity has been observed between the calcined powders and as-sintered ceramic pellets. This may be due to a grain growth during the sintering process since a much higher temperature is applied. Additionally, structural phase transformation may occur during sintering, which will contribute to the peak splitting as shown in Fig. 3(b), as well as the difference of the peak intensity ratios. As can be seen from Fig. 3(b), the LiBiO3 doping also has significant influence on the crystal structures. An obvious reduction in

(A)

(B)

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 1. Photograph of (1x) NKN-xLiBiO3 samples sintered at 1120°C for 15 h (A) and 10 h (B) [(a) x = 0.001; (b) x = 0.002; (c) x = 0.003; (d) x = 0.004; (e) x = 0.005; (f) x = 0.006].

2990

Journal of the American Ceramic Society—Jiang et al. (a)

(b)

(c)

(d)

(e)

(f)

Vol. 98, No. 10

Fig. 2. Natural surface SEM of (1x) NKN-xLiBiO3 samples sintered at 1120°C for 10 h [(a) x = 0.001; (b) x = 0.002; (c) x = 0.003; (d) x = 0.004; (e) x = 0.005; (f) x = 0.006].

(a)

(b)

Fig. 3. XRD patterns of (1x) NKN-xLiBiO3:(a) powder calcined and (b) bulk samples sintered.

peak splitting is observed with an increasing of doping content, which is probably associated with a phase transformation from a lower symmetry phase to a higher symmetric one.

(2) Investigation of the Crystal Growth Kinetics In order to further explore the crystal growth kinetics for such an abnormal phenomenon, the (1x)NKN-xLiBiO3 (x = 0.004) samples were prepared by varying sintering time

(A)

(B)

(C)

Fig. 4. Photographs of (1x) NKN-xLiBiO3 samples for x = 0.004 sintered at different isothermal holds (A)1100°C, (B)1090°C and (C) 1080°C for 3–24 h.

October 2015

2991

Solid-State Growth NKN Crystal (a)

(b)

Fig. 5. Plot of fraction exaggerated grain growth versus the logarithm of time of the above transformation in the (1x) NKN-xLiBiO3 (x = 0.004) system.

and temperature, as shown in Fig. 4. It can be seen that a few isolated exaggerated grain discretely distributed in the matrix at the initial stage; then these grains become larger, and finally closely packed with each other and arrange in a preferred orientation. Well-defined boundaries are observed between the big grains and the ceramic matrix. It is also interesting to note that (1) the color of the samples evolves from creamy white (or yellow) to dark green with the elongation of the dwelling time; and (2) the single crystal starts to grow from the edge toward the center. For the first case, it Table I.

Chemical Composition of LiBiO3-Doped NKN Ceramic and Single Crystal Ceramic area

Element/ at.%

O Na K Nb Bi Total

Exaggerated grain area

Sintered at 1110°C for 6h

Sintered at 1110°C for 12 h

Sintered at 1110°C for 6h

Sintered at 1110°C for 12 h

60.8 9.2 9.2 20.7 0.1 100.0

63.7 8.5 8.2 19.5 0.1 100.0

56.4 10.4 9.8 23.3 0.2 100.0

57.7 10.2 9.4 22.5 0.1 100.0

may be attributed to the substitution of K at the Nb site40 or the oxygen deficiency.17 While for the second one, a temperature gradient may exist and point from the edge to the center since the heat exchange is easier to start from the interface between the sample and atmosphere. According to an Ostwald ripening mechanism,41 the minimizing of surface-free energy leads to a single crystal growth at an expense of the smaller grains. This process is presumably triggered and processed with the aid of a liquid phase, which may appear at the interface between the single crystal and grains. After comparing Figs. 4(A)–(C), it is evident that the crystal growth rate is sensitive to both the dwelling temperature and time. To further quantify the kinetics of such a transformation, an Avrami analysis is performed. Figure 5 shows the recrystallized area fraction in the (1x) NKN-xLiBiO3 (x = 0.004) system as a function of the logarithm of time under different temperature conditions. For solid-state transformation displaying the fraction of transformation y is a function of time t and follows the well-known Avrami Eq (1)42: y ¼ 1 expðktn Þ

(1)

where k and n are time-independent constants. By convention, the rate of a transformation is taken as the reciprocal of time required for the transformation to proceed halfway

(a)

(b)

(c)

(e)

(f)

(g)

(d)

(h)

Fig. 6. The SEM micrograph of the surface, a fracture and interface of the samples grew at 1100°C. (a) Natural surface of single crystal area; (b) Interface between the single crystal area and ceramic matrix; (c, d) Natural surface of ceramic matrix area; (e–g) Fracture of single crystal; (h) the sample with fracture.

2992

Journal of the American Ceramic Society—Jiang et al.

to completion, 1 42 t0.5, or rate ¼ t0:5

(2)

So the t0.5 is about 15.5, 10.9, and 8.4 h for the samples grew under isothermal holds at 1080°C, 1090°C, and 1100°C, respectively. This observation reveals that the higher the temperature, the faster the crystallite growth rate, (a)

Vol. 98, No. 10

as indicated by the decreasing of t0.5 with increasing temperature. Furthermore, a chemical analysis was performed, as exemplified in Table I for the composition of x = 0.004, by fixing the sintering temperature at 1100°C but varying the sintering time. Both the ceramics and single crystal areas were analyzed. The results suggest that (1) for both areas a slightly

(b)

(c)

Fig. 7. The photographs of the samples (a), the Laue and XRD (b) and powder XRD (c) diffraction patterns of (001)-oriented NKN single crystals.

(a)

(b)

(c)

(d)

(f)

(g)

(e)

(h)

Fig. 8. TEM micrographs of the microstructural and crystallographic characteristics of NKN-LiBiO3 single crystal. (a) Typical domain structures viewed along [110] direction, and (b) its corresponding higher magnification image. (c) [110] zone axis diffraction pattern for a polydomain region. The diffraction spots splitting is magnified as a large bright-square in (c). (d) [110], (e) [111], (f) [001], (g) [112], and (h) [012] zone axes diffraction patterns for a monodomain region. All the zone axes are arbitrarily indexed by the pseudocubic perovskite unit cell.

October 2015


higher Na and K loss occurs at longer sintering time; (2) the K/Na ratio at a fixed dwelling time is almost equal, suggesting a uniform chemical distribution in either ceramic or single crystal region.

(3) Insight for Morphology and Microstructures Figure 6 shows the SEM micrographs of the surface, fracture, and interface of the samples grew at 1100°C for 9 h. The microstructure of the surface of the single crystals, the interface between the single crystals and ceramic matrix, and the ceramics are displayed through Figs. 6(a)–(d). It can be seen from the Figs. 6(a) and (b) that the exaggerated single crystal grows via stacking layers along the growth orientation, which agrees well with the two-dimensional layer structure growth mechanism.43 For the ceramic matrix, small grains are observed with a well faceted cube-like shape [Figs. 6(c) and (d)]. Also, liquid phase seems to appear in the matrix, and some dark areas in the ceramics in Figs 6(c)–(d) could be associated with such a liquid phase. Figures 6(e)–(g) shows the fracture section of the single crystal at different magnifications. These images further confirm that crystals grow in a layer-stacking manner with regular geometric shape and are tightly connected with each other.43 Figure 6(h) shows the picture of the sample with fracture section and natural surface. Clear boundaries between the single crystals and small grains are found to run through the matrix from the top to the bottom surface. The largest dimension of the single crystal in the present study is 11 mm 9 9 mm 9 3 mm, which is larger or comparable to those reported in the literatures. Ursic et al.44 reported that the size of NKN single crystal prepared on a [110] orientation KTaO3 seed by SSCG is of 4 mm in diameter; Inagaki et al. grew a Mn-doped NKN single crystal by floating zone and flux with size of 4.5 mm, length of 13 mm,45,46, and 6 mm 9 6 mm 9 x mm (the thickness x was not reported),47 respectively. Fisher et al. fabricated a NKN single crystals layer up to 200 lm thick on the KTaO3 (a)

(c)

2993

seed crystal.10 Zheng et al.15 also reported (Na, K)(Nb, Ta) O3 single crystal with the size of 12 mm 9 11 mm 9 11 mm prepared by top seeded solution growth (TSSG) technique. Recently, Deng et al.16 prepared (K0.25Na0.75)NbO3 (KNN25/75) single crystal with the size of 30 mm 9 10 mm by TSSG. The size of NKN single crystals by seed-free solidstate crystal growth (SFSSCG) is over the grade of 1 cm, while for AGG the size is often only micrometer dimension in thickness. Figure 7(a) shows the NKN single crystals prepared by increasing the volume of the pellet and prolonging the holding time at 1120°C. Figure 7(b) shows the XRD patterns of a single crystal [the upper right photographs in Fig. 7(a)]. Figure 7(c) shows the powder XRD patterns of the same single crystal with Fig. 7(b). An obvious difference between these two diffraction patterns [Fig. (b) and (c)] is easily identified. The powder XRD pattern Fig. 7(c) of the single crystal indicates a typical perovskite structure of the sample. Compared to the powder XRD spectrum [Fig. 7(c)], the single crystal is found to be orientated along the direction, as indicated by its XRD spectrum [Fig. 7(b)], as well as the Laue diffraction pattern [in set of Fig. 7(b)]. Further details of the microstructures are investigated with TEM on the NKN-LiBiO3 specimen sintered at 1109°C with an isothermal hold of 9 h. Figure 8(a) shows the bright-field TEM micrograph imaged from a [110]-oriented single crystal. As can be seen, well-paralleled lamellar ferroelectric domains are present, with their domain width around several hundred nanometer. The domain walls are found to be normal to the [111] crystallographic direction, further suggesting that they are on the {110} habit planes. As is well-known, typical herringbone ferroelectric domain morphology with substructures is the common pattern in NKN-based ceramics,48–53 especially typical of compositions near the polymorphic phase boundary, where a phase coexistence is observed. 50, 54 However, in the present case of NKN single crystals, close-up examination from the nanoscale insight indicates that no substructures are formed inside the lamellar domains

(b)

(d)

(e)

Fig. 9. (a) Representative TEM micrographs of NKN-LiBiO3 ceramics, and (b) its corresponding higher magnification image. Typical diffraction patterns obtained from a crystallite under (c) [001], (d) [110], and (e) [111] zone axes.

2994


[Fig. 8(b)]. In view of the electron-diffraction pattern, these {110} domain walls are found in ferroelastic nature, as revealed by the diffraction spots splitting shown in Fig. 8(c) when a polydomain region is diffracted. Subsequently, a wide tilting experiment is performed, and the diffraction patterns from the monodomain region are obtained and illustrated in Figs. 8(d)–(h) for a variety of zone axes. Besides the fundamental diffraction spots, no superlattice reflections and spot splitting are noticed. And the latter fact further verified the absence of substructures in the monodomains. In terms of crystallographic viewpoint, it is known that the virgin state of NKN-based crystals adopts either orthorhombic phase, or tetragonal phase, or a mixture of these two; both of these two phases do not show any superlattice diffraction spots.50– 52 Therefore, the TEM results shown in Fig. 8 suggest that the NKN-LiBiO3 still remains with its crystal symmetry in a phase without oxygen octahedral tilting. In addition to the NKN-LiBiO3 single crystal, the microstructure of polycrystalline ceramic matrix is also studied. As shown by the bright-field micrographs in Figs. 9(a) and (b), majority of the crystallites are found with their size in submicrometer scale. In addition to the featureless domains, the strip-like ferroelectric domains are also observed inside the grain. In view of crystallography, the electron-diffraction patterns [Figs. 9(c)–(e)] obtained from the single crystallites are consistent with those shown in Fig. 8. Again, no superlattice diffraction spots are noticed, indicating no octahedral tilting.

Vol. 98, No. 10

(4) Electrical Properties Figure 10 shows the temperature-dependent dielectric properties of an unpoled (001)-oriented NKN single crystal. Two dielectric anomalies appear in the measuring temperature range, as is expected from NKN system: the one at ~405°C corresponds to a tetragonal-cubic phase transition and the other one at 155°C is for an orthorhombic-tetragonal phase transition. Both phase transitions are of first-order nature, which can be revealed by the thermal hysteresis induced during heating and cooling process, where a temperature difference ~20°C and ~10°C are observed for orthorhombictetragonal and tetragonal-cubic phase transition, respectively. These phase transitions also leave marks on the dielectric loss spectra, as shown in Fig. 10(b). Also, a low dielectric loss

(a)

(a)

(b)

(b) (c)

Fig. 10. Variations in dielectric constant (e0 ) and loss (tand) for a NKN single crystal.

Fig. 11. (a) The schematic minor strain-field hysteresis loop adopted for the Rayleigh analysis; (b) real parts of d33 measured as a function of applied amplitude (E0); and (c) dinit and the portion of intrinsic contribution as a function of applied amplitude (E0)

October 2015

2995


~0.03 is observed in a low-temperature range. However, the loss sharply increases with the increasing of temperature. This may be due to the mobile ionic defects and a space charge polarization 17,55,56 produced by the volatile species, such as K, Na, and Bi. By employing the Rayleigh analysis,36–39 the intrinsic contributions to the electromechanical performance was further investigated. The Rayleigh region is considered as where the density and domain configuration of the domain walls remain unchanged with AC E-field (E0);approximately, this is below one-half of the coercive field (Ec). In this region, the real part of d33 (d″) linearly increases with enhancing the amplitude of E0. The induced peak strain xmax and the piezoelectric coefficient can be described as follows36–39: xmax ¼ dinit E0 þ ad E20

~205 pm/V is suggested from the intrinsic piezoelectric contribution.

Acknowledgments The authors gratefully acknowledge technical assistance from Amanda Baker, Jeff Long, Wonderling Nichole and others at The Pennsylvania State University (Penn State), and Yiqing Lu at Shanghai Institute of Ceramics in Chinese Academy of Sciences. Minhong Jiang also wishes to thank China Scholarship Council for providing the opportunity to study as a visiting scientist at Penn State. This work was also financially supported by the following projects: the National Natural Science Foundation of China (51102056), the Guangxi Natural Science Foundation (2012GXNSFGA060002, 2011GXNSFB018008, 2012GXNSFAA053198), Program for Excellent Talents in Guangxi Higher Education Institutions (2011no.40).

References

(3) 1

0

d ¼ dinit þ ad E0

(4)

where dinit and ad are the initial reversible piezoelectric response at zero electric field and the piezoelectric Rayleigh coefficient, respectively. The relative magnitude of ad is an excellent indicator of the extrinsic domain wall contribution to the piezoelectric and dielectric properties of poled ferroelectrics. The dinit and ad values can be obtained by linear fitting the plots of d0 as a function of various E0. The intersection of the fitting line is dinit, whereas ad is the slope. The real parts of d33 at various E0 are calculated from the peak to peak strain (xp-p) and the loop area (AEM) marked in Fig. 11(a) by using the Eqs. (5)–(7)36–39: d00 ¼ AEM =pE20

(5)

d ¼ xpp =2E20

(6)

d0 ¼ ðd2 d002 Þ1=2

(7)

Figure 11(b) shows the d0 (d33) value at different E0 of the NKN single crystal. Dashed lines represent the linear fitting of the experimental results. The Rayleigh coefficients, dinit and ad = 1.96 9 1016 m2/V2, are obtained from the intersection and slope of the plots in Fig. 11(b). Figure 11(c) shows that the variation in the intrinsic piezoelectric component (dinit) has a stable trend with E0. The average dinit is about 205 pm/V. At E0 = 1 kV/cm, the intrinsic piezoelectric part is 91.5%, indicating that the reversible contribution to the piezoelectricity is dominant. The reversible contribution to the piezoelectric properties decreases as shown in Fig. 11(c) due to the increasing of the real part of d33 with increasing E0, indicating that the domain wall motion begins to play a leading role with increasing of E0.

IV.

Conclusions

Large K0.5Na0.5NbO3 (NKN) lead-free piezoelectric single crystals were obtained by a SFSSCG method by doping with a small amount of LiBiO3. The largest dimension of the single crystal is able to reach at 11 mm 9 9 mm 9 3 mm. The NKN crystalline grains tend to form a self-assembled arrangement with a preferred orientation and grow by layer stacking along the growth orientation. A kinetic study was performed to better quantify the growth process. The rates of transformation are determined via an Avrami analysis. The single crystals are found to have a high-quality singlecrystalline perovskite structure with an orthorhombic symmetry. The dielectric spectra suggest two first-order transition in the temperature range from 30°C–500°C. And the electromechanical properties were further analyzed by the Rayleigh analysis under low electric fields, where an averaged dinit

X. H. Li, et al., “Phase Transitions and Electrical Properties of (1x) (K0.5Na0.5)NbO3XBiScO3 Lead-Free Piezoelectric Ceramics with a CuO Sintering Aid,” Phys. Status Solidi A, 206 [11] 2622–6 (2009). 2 Y. Saito, et al., “Lead-Free Piezoceramics,” Nature, 432 [7013] 84–7 (2004). 3 H. Ralf-Peter, A. Gerold, W. Susanne, J. Michael, and S. Hoffmann, “Characterization of Ferroelectric Domains in Morphotropic Potassium Sodium Niobate with Scanning Probe Microscopy,” Appl. Phys. Lett., 90 [25] 252905, 3pp (2007). 4 M. H. Jiang, X. Y. Liu, and G. H. Chen, “Phase Structures and Electrical Properties of New Lead-Free Na0.5K0.5NbO3–LiSbO3-BiFeO3 Ceramics,” Scr. Mater., 60 [10] 909–12 (2009). 5 K. Kobayashi, Y. Doshida, Y. Mizuno, and C. A. Randall, “A Route Forwards to Narrow the Performance Gap Between PZT and Lead-Free Piezoelectric Ceramic with Low Oxygen Partial Pressure Processed (Na0.5K0.5) NbO3,” J. Am. Ceram. Soc., 95 [9] 2928–33 (2012). 6 K. Kobayashi, Y. Doshida, Y. Mizuno, and C. A. Randall, “Possibility of Cofiring a Nickel Inner Electrode in a (Na0.5K0.5)NbO3-LiF Piezoelectric Actuator,” Jpn. J. Appl. Phys., 52 [9S1] 09KD07 (2013). 7 S. Kawada, M. Kimura, Y. Higuchi, and H. Takagi, “(K,Na)NbO3-Based Multilayer Piezoelectric Ceramics with Nickel Inner Electrodes,” App. Phys. Express, 2 [11] 111401, 3pp (2009). 8 C. Liu, P. Liu, K. Kobayashi, and C. A. Randall, “Base Metal Co-Fired (Na, K)NbO3 Structures with Enhanced Piezoelectric Performance,” J. Electroceram., 32 [4] 301–6 (2014). 9 S. E. Park and T. R. Shrout, “Ultrahigh Strain and Piezoelectric Behavior in Relaxor Based Ferroelectric Single Crystals,” J. Appl. Phys., 82 [4] 1804–11 (1997). 10 J. G. Fisher, A. Benˇcan, J. Godnjavec, and M. Kosec, “Growth Behavior of Potassium Sodium Niobate Single Crystals Grown by Solid-State Crystal Growth Using K4CuNb8O23 as a Sintering Aid,” J. Euro. Ceram. Soc., 28 [8] 1657–63 (2008). 11 J. G. Fisher, A. Benˇcan, and M. Kosec, “Growth of Dense Single Crystals of Potassium Sodium Niobate by a Combination of Solid-State Crystal Growth and Hot Pressing,” J. Am. Ceram. Soc., 91 [5] 1503–7 (2008). 12 Q. H. Zhang, et al., “Enhanced Piezoelectric and Ferroelectric Properties in Mn-Doped Na0.5Bi0.5TiO3-BaTiO3 Single Crystals,” Appl. Phys. Lett., 95 [10] 102904, 3pp (2009). 13 H. Tian, Z. Zhou, D. Gong, H. Wang, D. Liu, and Y. Jiang, “Growth and Optical Properties of Paraelectric K1yNayTa1xNbxO3 Single Crystals,” Appl. Phys. B, 91 [1] 75–8 (2008). 14 X. Q. Huo, et al., “Growth and Properties of Li, Ta Modified (K, Na) NbO3 Lead-Free Piezoelectric Single Crystals,” Phys. Status Solidi.:Rapid Res. Lett., 8 [1] 86–90 (2014). 15 L. M. Zheng, X. Q. Huo, R. Wang, J. J. Wang, W. H. Jiang, and W. W. Cao, “Large Size Lead-Free (Na,K)(Nb,Ta)O3 Piezoelectric Single Crystal: Growth and Full Tensor Properties,” Cryst. Eng. Comm., 15 [38] 7718–22 (2013). 16 H. Deng, et al., “Orientation Dependence of Electrical Properties of Large-Sized Sodium Potassium Niobate Lead-Free Single Crystals,” Cryst. Eng. Comm., 16 [13] 2760–5 (2014). 17 Y. Kizaki, Y. Noguchi, and M. Miyayama, “Defect Control for Low Leakage Current in K0.5Na0.5NbO3 Single Crystals,” Appl. Phys. Lett., 89 [14] 142910, 2pp (2006). 18 J. G. Fisher, A. Benˇcan, J. Holc, M. Kosec, S. Vernay, and D. Rytz, “Growth of Potassium Sodium Niobate Single Crystals by Solid State Crystal Growth,” J. Crys. Grow, 303 [2] 487–92 (2007). 19 A. N. Holden, “Preparation of Metal Single Crystals,” Trans. Am. Soc. Met., 42 [31] 9–346 (1950). 20 M. Hillert, “On the Theory of Normal and Abnormal Grain Growth,” Acta Metall., 13 [3] 227–38 (1965). 21 S. J. Dillon and M. P. Harmer, “Mechanism of “Solid-State” Single-Crystal Conversion in Alumina,” J. Am. Ceram. Soc., 90 [3] 993–5 (2007). 22 S. Tanji, S. Matsuzawa, N. Wakatsuki, and S. Soejima, “A Magnetic Head of Mn-Zn Ferrite Single Crystal Produced by Solid-Solid Reaction,” IEEE Trans. Magn., 21 [5] 1542–4 (1985). 23 A. Ikesue, Y. L. Aung, T. Yoda, S. Nakayama, and T. Kamimura, “Fabrication and Laser Performance of Polycrystal and Single Crystal Nd: YAG by Advanced Ceramic Processing,” Opt. Mater., 29 [10] 1289–94 (2005). 24 T. Yamamoto and T. Sakuma, “Fabrication of Barium Titanate Single Crystals by Solid-State Grain Growth,” J. Am. Ceram. Soc., 77 [4] 1107–9 (1994).

2996


25 S. M. An, B. K. Yoon, S. Y. Chung, and S. J. Kang, “Nonlinear Driving Force–Velocity Relationship for the Migration of Faceted Boundaries,” Acta Mater., 60 [11] 4531–9 (2012). 26 M. S. Kim, J. G. Fisher, S. J. L. Kang, and H. Y. Lee, “Grain Growth Control and Solid-State Crystal Growth by Li2O/PbO Addition and Dislocation Introduction in the PMN-35PT System,” J. Am. Ceram. Soc., 89 [4] 1237–43 (2006). 27 S. J. Zhang, S. M. Lee, D. H. Kim, and H. Y. Lee, “Characterization of High Tc Pb(Mg1/3Nb2/3)O3-PbZrO3-PbTiO3 Single Crystals Fabricated by Solid State Crystal Growth,” Appl. Phys. Lett., 90 [23] 232911, 3pp (2007). 28 J. B. Lim, S. J. Zhang, H. Y. Lee, and T. R. Shrout, “Solid State Crystal Growth of BiScO3-Pb(Mg1/3Nb2/3)O3-PbTiO3,” J. Electroceram., 29 [2] 139–43 (2012). 29 K. S. Moon, D. Rout, H. Y. Lee, and S. J. L. Kang, “Solid State Growth of Na0.5Bi0.5TiO3-BaTiO3 Single Crystals and Their Enhanced Piezoelectric Properties,” J. Crystal. Growth, 317 [1] 28–31 (2011). 30 C. Wang, Y. D. Hou, H. Y. Ge, M. K. Zhu, and H. Yan, “Growth of (K0.5Na0.5)NbO3 Single Crystals by Abnormal Grain Growth Method from Special Shaped Nano-Powders,” J. Euro. Ceram. Soc., 30 [7] 1725–30 (2010). 31 H. Y. Lee, J. S. Kim, and D. Y. Kim, “Fabrication of BaTiO3 Single Crystals Using Secondary Abnormal Grain Growth,” J. Euro. Ceram. Soc., 20 [10] 1595–7 (2000). 32 J. M. Saldana, B. Mullier, and G. A. Schneider, “Preparation of BaTiO3 Single Crystals Using the Modified SiO2-Exaggerated Grain Growth Method,” J. Euro. Ceram. Soc., 22 [5] 681–8 (2002). 33 Y. H. Zhen and J. F. Li, “Abnormal Grain Growth and New Core–Shell Structure in (K,Na)NbO3-Based Lead-Free Piezoelectric Ceramics,” J. Am. Ceram. Soc., 90 [11] 3496–502 (2007). 34 H. Y. Park, et al., “Microstructure and Piezoelectric Properties of 0.95 (K0.5Na0.5)NbO3–0.05BaTiO3 Ceramics,” Appl. Phys. Lett., 89 [6] 062906, 3pp (2006). 35 H. J. Lee, H. Ryu, M. S. Bae, Y. K. Cho, and S. Nahm, “Transmission Electron Microscopy Microstructure of 0.95(K0.5Na0.5)NbO3–0.05BaTiO3 Ceramics,” J. Am. Ceram. Soc., 89 [11] 3529–32 (2006). 36 R. E. Eitel, T. R. Shrout, and C. A. Randall, “Nonlinear Contributions to the Dielectric Permittivity and Converse Piezoelectric Coefficient in Piezoelectric Ceramics,” J. Appl. Phys., 99 [12] 124110, 7pp (2006). 37 C. Liu, P. Liu, K. Kobayashi, W. G. Qu, and C. A. Randall, “Enhancement of Piezoelectric Performance of Lead-Free NKN-Based Ceramics via a High-Performance Flux-NaF-Nb2O5,” J. Am. Ceram. Soc., 96 [10] 3120–6 (2013). 38 B. Peng, Z. X. Yue, and L. T. Li, “Evaluation of Domain Wall Motion During Polymorphic Phase Transition in (K,Na)NbO3-Based Piezoelectric Ceramics by Nonlinear Response Measurements,” J. Appl. Phys., 109 [5] 054107, 6pp (2011). 39 R. Eitel and C. A. Randall, “Octahedral Tilt-Suppression of Ferroelectric Domain Wall Dynamics and the Associated Piezoelectric Activity in Pb(Zr,Ti) O3,” Phys. Rev. B, 75 [9] 094106, 8pp (2007). 40 W. Wenshan, Z. Qun, and G. Zhaohua, “Investigation of the Color of KNbO3 Single Crystals Grown by the Radio-Frequency Heating Czochralski Technique,” J. Cryst. Growth, 83 [1] 62–8 (1987).

Vol. 98, No. 10

41 P. W. Voorhees, “The Theory of Ostwald Ripening,” J. Stat. Phys., 38 [1– 2] 231–52 (1985). 42 D. C. William, Materials Science and Engineering: An Introduction, 7th edition, pp.318–23. John Wiley & Sons, Inc., New York City, New York, 2012. 43 J. P. Van der Eerden, “Crystal Growth Mechanisms”; pp. 311–475 in Handbook of Crystal Growth, Fundamentals, Part A, Thermodynamics and Kinetics, Edited by D. T. J. Hurle. Elsevier Science Publishers, Amsterdam, the Netherlands, 1993. 44 H. Ursi, A. Benan, M. Skarabot, M. Godec, and M. Kosec, “Dielectric, Ferroelectric, Piezoelectric, and Electrostrictive Properties of K0.5Na0.5NbO3 Single Crystals,” J. Appl. Phys., 107 [3] 033705, 5pp(2010). 45 Y. Inagaki, K. I. Kakimoto, and I. Kagomiya, “Crystal Growth and Ferroelectric Property of Na0.5K0.5NbO3 and Mn-Doped Na0.5K0.5NbO3 Grystals Grown by Floating Zone Method,” J. Euro. Ceram. Soc., 30 [2S1] 301–6 (2010). 46 Y. Inagaki, K. I. Kakimoto, and I. Kagomiya, “Ferroelectric Domain Characterization of Orthorhombic Sodium–Potassium Niobate Piezoelectric Crystals,” J. Am. Ceram. Soc., 93 [12] 4061–5 (2010). 47 Y. Inagaki and K. I. Kakimoto, “Dielectric and Piezoelectric Properties of Mn-Doped Na0.5K0.5NbO3 Single Crystals Grown by Flux Method,” Appl. Phys. Express, 1 [6] 061602, 3pp (2008). 48 R. P. Herber, G. A. Schneider, S. Wagner, and M. J. Hoffmann, “Characterization of Ferroelectric Domains in Morphotropic Potassium Sodium Niobate with Scanning Probe Microscopy,” Appl. Phys. Lett., 90 [25] 252905, 3pp (2007). 49 R. L opez-Juarez, O. Novelo-Peralta, F. Gonzalez-Garcıa, F. Rubio-Marcos, and M.-E. Villafuerte-Castrej on, “Ferroelectric Domain Structure of Lead-Free Potassium-Sodium Niobate Ceramics,” J. Eur. Ceram. Soc., 31 [9] 1861–4 (2011). 50 J. Fu, R. Z. Zuo, and Z. K. Xu, “High Piezoelectric Activity in (Na, K) NbO3 Based Lead-Free Piezoelectric Ceramics: Contribution of Nanodomains,” Appl. Phys. Lett., 99 [6] 062901, 3pp (2011). 51 H. Z. Guo, S. Zhang, S. P. Beckman, and X. Tan, “Microstructural Origin for the Piezoelectricity Evolution in (K0.5Na0.5)NbO3-Based Lead-Free Ceramics,” J. Appl. Phys., 114 [15] 154102, 8pp (2013). 52 K. Wang and J. F. Li, “Domain Engineering of Lead-Free Li-Modified (K,Na)NbO3 Polycrystals with Highly Enhanced Piezoelectricity,” Adv. Funct. Mater., 20 [12] 1924–9 (2010). 53 S. Zhang, H. J. Lee, C. Ma, and X. Tan, “Sintering Effect on Microstructure and Properties of (K,Na)NbO3 Ceramics,” J. Am. Ceram. Soc., 94 [11] 3659–65 (2011). 54 F. Rubio-Marcos, A. D. Campo, R. L opez-Juarez, J. J. Romero, and J. Fernandez, “High Spatial Resolution Structure of (Na, K)NbO3 Lead-Free Ferroelectric Domains,” J. Mater. Chem., 22 [19] 9714–20 (2012). 55 S. Tashiro, H. Nagamatsu, and K. Nagata, “Sinterability and Piezoelectric Properties of KNbO3 Ceramics After Substituting Pb and Na for K,” Jpn. J. Appl. Phys., 41 [11B] 7113–8 (2002). 56 G. Shirane, R. Newnham, and R. Pepinsky, “Dielectric Properties and Phase Transitions of NaNbO3 and (Na, K)NbO3,” Phys. Rev., 96 [3] 581–8 (1954). h