Int. J. Appl. Ceram. Technol., 12 [1] 228–234 (2015) DOI:10.1111/ijac.12150
Sodium Excess Ta-Modified (K0.5Na0.5)NbO3 Ceramics Prepared by Reactive Template Grain Growth Method Ali Hussain, Adnan Maqbool, Jin Soo Kim, Tae Kwon Song, and Myong Ho Kim* School of Advanced Materials Engineering, Changwon National University, Gyeongnam 641-773, Korea
Won Jong Kim and Sang Su Kim Department of Physics, Changwon National University, Gyeongnam 641-773, Korea
Lead-free sodium excess Ta-modified (K0.470Na0.545)(Nb0.55Ta0.45)O3 (KNNT) ceramics were synthesized by a conventional and reactive templated grain growth methods, and their degree of grain orientation, microstructure, dielectric, ferroelectric, and field-induced strain properties were systematically investigated. A high degree of grain orientation (Lotgering factor F = 80%) was obtained in textured KNNT ceramics. Results showed that textured KNNT ceramics exhibit high grain orientation, dielectric constant, and field-induced strain as compared to nontextured samples of the same composition. Room temperature unipolar fieldinduced strain of K0.5Na0.5NbO3 (KNN) ceramics was enhanced from 0.080% for nontextured sample to 0.115% for textured sample, and their corresponding dynamic piezoelectric coefficients (d33 ) were improved from 320 pm/V to 460 pm/V, respectively.
Introduction Global environmental concern over lead-based materials has driven the development of high-performance lead-free alternatives.1–5 Considerable research efforts have been devoted to lead-free potassium sodium niobate, K0.5Na0.5NbO3 (abbreviated as KNN), and materials based on KNN due to their relatively good piezoelectric response, high Curie temperature (Tc), and good biocompatibility.6–10 However, dense KNN ceramics are difficult to develop through conventional sintering method under pressureless condition, and their piezoelectric response is always poor (piezoelectric constant d33 = 80pC/N and electromechanical coupling factor kp = 34–40).8,11 Therefore, microstructural control and compositional design of KNN-based materials is required to develop lead-free ceramics with excellent properties to replace lead-based materials in electronic industry. Texture control of polycrystalline ceramics is an important approach to improve the piezoelectric properties of lead-free materials without drastically changing the composition of the ceramics.2,12–14 Formation of solid solution often improves the room temperature piezoelectric properties of lead-free ceramics, while it generally lowers Tc and therefore limits the applications. Texture control, on the other hand, hardly affects Tc as long as the ceramics retain their original compositions. *
[email protected] © 2013 The American Ceramic Society
Meanwhile, textured ceramics have additional advantages over single crystals in terms of composition flexibility and uniformity in a wide range of solid solutions besides lower production costs. There have been several texture engineering techniques, such as oriented consolidation of anisometric particles,15 templated grain growth (TGG),16,17 reactive templated grain growth (RTGG)2,12,14,18 directional solidification technology,19 screen printing multilayer grain growth technology,20–22 and so on. Among these methods, RTGG is more suitable for materials having perovskite structure. Crystallographic texturing of polycrystalline ferroelectric ceramics, such as (K,Na,Li)(Nb,Ta,Sb)O3,2 (KxNa1x)0.946Li0.054NbO3,14 Pb(Mg1/3Nb2/3)O3–PbTiO3, and Pb(Mg1/ 23 (Na0.84K0.16)0.5Bi0.53Nb2/3)0.42(Ti0.638Zr0.362)0.58O3, 24 TiO3, and Sr0.53Ba0.47Nb2O6,25 results in greatly enhanced piezoelectric properties. In RTGG process, the template particles with specific morphology and crystallographic properties are oriented in a matrix ceramic powder by tape casting technique; then, the consequent heat treatment results in the nucleation and growth of desired crystals on aligned template particles to form textured ceramics. Due to the grain orientation effect induced by the TGG process, textured ceramics can be more easily poled and thus provide much higher piezoelectric constants and electromechanical coefficients compared with their nontextured counterparts prepared conventional solid reaction method.2,14,23,24,26 In this work, we have selected nonstoichiometric sodium excess (K0.470Na0.545)(Nb0.55Ta0.45)O3 (KNNT)
www.ceramics.org/ACT
KNNT Ceramics Prepared by RTGG
ceramics for texture development, because of its environment friendly nature and high piezoelectric response in conventional form.27,28 For KNNT texture ceramics development, platelike NaNbO3 (NN) particles were synthesized from Bi2.5Na3.5Nb5O18 (BNN) by topochemical microcrystal conversion (TMC). The platelike NN particles work as templates and also act as reactant composition for texturing ceramics. The crystal structure, microstructure, dielectric, ferroelectric, and field-induced strain properties of textured ceramics were compared with nontextured samples of the same composition. It was found that textured KNNT ceramics show overall higher performance over nontextured ceramics.
Experimental Procedure Platelike NN templates particles were synthesized from bismuth layered structure ferroelectric BNN by TMC method.29 BNN precursor particles were first prepared by mixing of reagent-grade Bi2O3 powder (99.9% Sigma-Aldrich, St. Louis, MO) Nb2O5, (99.9% Junsei Chemical, Tokyo, Japan) and Na2CO3 (99.95%, SigmaAldrich, St. Louis, MO) powders in stoichiometric ratio in ethanol for 12 h using zirconia balls as the milling medium. The mixture was then mixed thoroughly with NaCl (99.95%, Junsei) in the weight ratio of oxide to salt 1:1.5 for 12 h. After drying at 90°C, the mixture was placed in a sealed alumina crucible, heated to 1100°C, and held for 4 h. After the completion of reaction, the as-synthesized product was washed several times with hot de-ionized water to remove NaCl salt. Platelike NN templates were then prepared by mixing of BNN platelets and Na2CO3 in molar ratio of BNN/Na2CO3 1:1.5 in NaCl flux. The mixture was placed in a sealed alumina crucible and heated at 950°C for 4 h. The product was washed several times with hot de-ionized water. The NN template was separated using HNO3 to remove the by-product, Bi2O3. The obtained NN particles retained the morphology of the BNN precursor well. Nonstoichiometric, sodium excess matrix powder KNNT was prepared by conventional mixed-oxide route. Commercially available carbonate powder and metal oxides, Na2CO3 (99.95%, Sigma Aldrich), K2CO3 (99.99%, Aldrich chemistry), Nb2O5, (99.9% Junsei) and Ta2O5, (99.9%, Aldrich Chemistry), were first mixed by ball milling and then calcined at 700 and 850°C for 4 h to form a perovskite structure. The calcined powders were mixed thoroughly with a solvent (60 vol.% ethanol and 40 vol.% methyl-ethyl-ketone, MEK) and triethyl phosphate (dispersant) in a ball mill for 24 h. Polyvinyl butyral (binder) and polyethylene glycol/
229
diethyl-o-phthalate (plasticizer) were added, and the mixtures were ball milled again for 12 h. The as-synthesized NN templates of 5 wt% were then added to the mixture and ball milled with a slow rotation for another 12 h to form slurry for tape casting. The slurry was tape casted to form a green sheet with a thickness of ~80 lm on SiO2-coated polyethylene film by a doctor blade apparatus. After drying, a single-layer sheet was cut, laminated, and hot-pressed at a temperature of 45°C and a pressure of 50 MPa for 2 min to form a 2-mm-thick green compact. The compacts were further cut into small samples of about 1 cm in width and 1 cm in length. The compacts were heated at 600°C for 12 h with intermediated steps of 250 and 350 for 6 and 8 h to remove organic substances prior to sintering. The samples were sintered at 1225°C for 12 h in air atmosphere and were brought to room temperature at cooling rate of 5°C/min. Nontextured KNNT ceramics were also prepared through conventional solid state reaction for comparison. The experimental details of the fabrication process have been reported and can be found elsewhere.27,28 The crystalline phases and the degree of texture development were determined using an XRD analysis with Cu-Ka radiation on the major surface (parallel to the tape-casting direction) of sintered ceramics. The bulk densities of the samples were measured by Archimedes’ method. The microstructure of the fractured surface was observed by scanning electron microscope (SEM; JP/JSM 5200, Japan). For electrical measurements, the upper and the lower extensive surfaces of the specimen were polished to become parallel and coated with a silver-palladium paste as electrode by screen printing. The specimens were poled at 100°C by immersing in silicon oil under a dc electric field of 4 kV/mm for 30 min. The dielectric constant and loss response were measured through an impedance analyzer (HP4194A, Agilent Technologies, Palo Alto, CA). Polarization versus electric field (P-E) hysteresis loops were measured in silicon oil with the aid of ferroelectric test system (Precision LC; Radian Technologies, Albuquerque, NM) at 20 Hz. Field-induced strain response was measured using a contact type displacement sensor (Model 1240; Mahr GmbH, G€ottingen, Germany) at 50 mHz.
Results and Discussion Template particles play a crucial role in the development of high-quality textured ceramics. Platelike NN templates were synthesized from BNN precursors particles by molten salt synthesis followed by TMC.29 Figure 1 presents the X-ray diffraction pattern and SEM
230
International Journal of Applied Ceramic Technology—Hussain, et al.
(a)
Vol. 12, No. 1, 2015
(b)
Fig. 1. X-ray diffraction pattern and scanning electron microscope (SEM) micrograph of NaNbO3 (NN) particle synthesized by topochemical microcrystal conversion (TMC) method at 950°C for 4 h; (a) X-ray diffraction pattern (b) SEM micrograph.
micrograph of NN templates synthesized from BNN precursors. It is clear from Fig. 1a that NN templates synthesized from BNN precursors are of single phase and all the diffraction peaks attributable to perovskite structure. The strongest (001) diffraction peak suggests that the templates are of high degree orientation. The intensity peaks were indexed on the basis of tetragonal unit cell which matched well with JCPDS no. 74-2455. Perovskite-type materials with single intensity peak (200) at 2h angle of 46.5° are characterized as a cubic phase according to JCPDS no. 75-2102; however, the splitting of (200) into (002) and (200) at 2h angle of 46.5° is the characteristic of tetragonal phase, which is identified by JCPDS no. 74-2455. The difference between the cubic phase and tetragonal phase is that in cubic phase, Na+ locates in the center of eight NbO6 octahedra, while in tetragonal phase, Na+ deviates from the center of eight NbO6 octahedra. At the higher soaking temperature, Na+ of cubic NN deviates from the center of eight NbO6 octahedra because of distortion of crystal lattice; therefore, cubic phase of NN transforms into tetragonal phase of NN. Figure 1b shows the SEM photograph of the NN particles synthesized by TMC. It is clear that NN particles have rectangular-platelike shapes. They have high aspect ratio and high degree of orientation with relatively uniform size distribution. The particle size varies in the range of 10–20 lm with thickness of 1–2 lm. These NN particles are believed to possibly be easily oriented in a matrix of fine powders by a shearing process (e.g., tape casting), resulting in well textured microstructure found in products obtained by TGG or RTGG processing. Figure 2 depicts the XRD patterns of the textured and nontextured samples of KNNT ceramics. Both samples exhibit a single-phase perovskite structure with tetragonal symmetry according to JCPDS No. 77-0037.
Fig. 2. X-ray diffraction pattern of texture and nontextured (K0.470Na0.545)(Nb0.55Ta0.45)O3 (KNNT) ceramics.
XRD patterns of nontextured KNNT sample shown in Fig. 2 suggest a random orientation, evident from the strong intensity of (111) reflection. On the other hand, relatively strong intensities of (00l) of textured specimen suggest that a large fraction of grains are aligned with their c-axis normal to the sample surface. The degree of grain orientation was estimated by the Lotgering factor (F), given by F ¼ ðp po Þ=ð1 po Þ
ð1Þ
www.ceramics.org/ACT
KNNT Ceramics Prepared by RTGG
Where p = ΣI(00l)/ΣI (hkl), I is the relative intensity of the diffraction peak, and po is the value of p for a randomly oriented sample. F varies from 0 for a randomly oriented sample to 1 for a completely oriented sample. In this work, the diffraction peaks in the range of 2h from 20° to 60° were used to calculate F. The calculated Lotgering factor F for the textured KNNT ceramic is about 0.80, indicating a high degree of grain orientation. This also suggests that the NN templates were efficient in inducing grain orientation in the KNNT ceramics. It has also been observed that the textured KNNT ceramic is well sintered and has a dense structure. The density of both samples was measured by the Archimedes’ method. The relative density of the textured KNNT ceramic was found 92%, while that of the nontextured sample of the same composition was 95%. This can be attributed to the reduced volume loss of the carbonates in raw materials during the reaction in
Fig. 3. Scanning electron microscope (SEM) micrographs of texture and nontextured (K0.470Na0.545)(Nb0.55Ta0.45)O3 (KNNT) ceramics. Arrow on the micrograph of textured KNNT sample indicates the tape casting direction.
231
RTGG, resulting from the use of the calcined powders in preparing the slurry and the green sheets samples. The SEM micrographs of the fracture surface of textured and nontextured KNNT ceramics shown in Fig. 3 present that both samples have faceted grains and compact microstructure. It can be seen that the surface of textured ceramics are composed of large grains as compared to nontextured ceramics. The reason of the large grain in textured ceramics could be the NN templates incorporation with KNNT matrix powder. The higher sintering temperature promoted epitaxial growth of the desired phase on the orientated NN templates which act as a substrate for epitaxy and as a seed for exaggerated grain growth. The epitaxial growth led to the textured KNNT grain coarsening. The large grain growth in textured ceramics could be attributed to a uniform distribution of the template particles in the green ceramic body, which was followed by the large nucleation of grain growth at each grain. This observation keeps consistency with the XRD results. Figure 4 shows the temperature dependence of the dielectric constant (e) and loss (tand) for the textured and nontextured KNNT ceramics measured at various frequencies. It is clear from the figure that dielectric constant of the textured ceramics is much higher than that of the nontextured KNNT ceramics. At measured frequency of 1 kHz, the maximum dielectric constant for nontextured sample is 6386, which increases to 10364 for the textured sample. In addition, below ferroelectric phase transition temperature (Tc), the dielectric constant of textured ceramics is less sensitive to frequency as compared to nontextured ceramics. Moreover, the Tc of both samples was observed at ~145°C, this value is quite lower than that of the pure KNN ceramics (450°C), which is due to partial substitution of Ta for Nb.28 The polymorphic phase transition (PPT) of nontextured ceramics is observed at about 35°C, while that of the textured ceramics does not show peak in the studied temperature range. It is expected that the PPT of the textured ceramics is shifted to lower temperature due to different processing and NN template addition. Similar to dielectric constant response, the dielectric loss curves also show more frequency dispersion for nontextured sample as compared to texture sample. In addition, the dielectric loss of the textures sample is slightly higher than that of nontextured sample in the studied temperature range which may be due to its lower density and introduced defects/impurities into green laminates during tape casting process. Room temperature P-E hysteresis loops of KNNT ceramics measured at 50 Hz are shown in Fig. 5.
232
International Journal of Applied Ceramic Technology—Hussain, et al.
Vol. 12, No. 1, 2015
Fig. 4. Temperature dependence dielectric constant and loss of the textured and nontextured (K0.470Na0.545)(Nb0.55Ta0.45)O3 (KNNT) ceramics at different frequencies.
Fig. 5. Polarization versus electric field (P-E) hysteresis loops of the textured and nontextured (K0.470Na0.545)(Nb0.55Ta0.45)O3 (KNNT) ceramics.
Textured KNNT ceramics exhibit overall smaller ferroelectric response as compared to nontextured ceramics of the same composition. For textured KNNT ceramics, the remanent polarization (Pr) decreased from 8.18 to 5.76 lC/cm2, and coercive field (Ec) slightly increased from 0.32 to 0.34 kV/mm. Similar response of the P-E hysteresis was also observed previously in other perovskite-type materials.30–33 This can be attributed to template particles which act as defect pinning centers
during ferroelectric switching process, so that the textured ceramics have lower ferroelectric response. Figure 6 presents field-induced bipolar and unipolar strain curves of textured and nontextured KNNT ceramics. The bipolar strain curves shown in Fig. 6a present that for textured KNNT ceramics, the negative strain (Sneg; which is associated with domain back switching, i.e., the difference between zero field and lowest strain) decreased, while the maximum strain (Smax; which is the difference between positive and negative maximum strain) increased. This behavior is consistent with the observed smaller P-E hysteresis loop of the textured ceramics. The highly oriented textured ceramics results in better field-induced strain response than those for nontextured ceramics. The unipolar field-induced strain response of the textured and nontextured KNNT ceramics measured at 2.5 kV/mm is shown in Fig. 6b. The unipolar field induced-strain increased from 0.08 0.01 for nontextured sample to 0.115 0.01 for textured sample at room temperature. The corresponding dyn amic piezoelectric coefficient (d33 = Smax/Emax) of 320 pm/V and 460 pm/V was obtained for these specimens, respectively. This value of strain and dynamic piezoelectric coefficient are lower than that of textured (K0.44Na0.52Li0.04)(Nb0.86Ta0.10Sb0.04)O3 (LF4) containing both sintering additive Li and heavy element Sb2; however, it is higher as compared to previously reported results of modified KNN-based ceramics in both textured
KNNT Ceramics Prepared by RTGG
www.ceramics.org/ACT
(a)
233
(b)
Fig. 6. Field-induced strain of the textured and nontextured (K0.470Na0.545)(Nb0.55Ta0.45)O3 (KNNT) ceramics (a) Bipolar strain (b) unipolar strain.
Table I. Dielectric and Piezoelectric Properties of Textured and Nontextured (K0.470Na0.545) (Nb0.55Ta0.45)O3 (KNNT) Ceramics Relative density Textured Nontextured Increase
92% 95% –
e 10364 6386 62%
d33
460 pm/V 320 pm/V 44%
and nontextured compositions.10,11,34–36 The dielectric and piezoelectric properties of the textured and nontextured KNNT ceramics are presented in Table I, where superior response for the textured ceramics is clearly demonstrated. It can be seen that the dynamic piezoelectric coefficient d33 at room temperature and its dielectric constant with a measuring frequency of 1 kHz at Tc increased by about 44% and 62%, respectively after texturing. These results strongly suggest that texturing can significantly enhance the piezoelectric performance of KNNT ceramics which can be considered as promising candidate materials for lead-free piezoelectric applications.
compared with nontextured KNNT ceramics prepared by conventional mixed oxide route. The results show that textured KNNT ceramics have tetragonal {00 l} orientation degree of 80% and platelike grains aligning in the direction parallel to the casting plan. The dielectric constant textured ceramics were found to have much higher value than nontextured ceramics without apparent change in the phase transition temperature (Tc). The field-induced strain increases from 0.080% for nontextured to 0.115% for textured KNNT ceramics. Furthermore, textured KNNT ceramics exhibit 44% high dynamic piezoelectric coefficient d33 as compared to the nontextured sample. These results show that RTGG method is effective for grain orientation to improve the piezoelectric response of the KNNT ceramics.
Acknowledgments This work was supported by Changwon National University in 2011–2013 and by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by Ministry of Education, Science and Technology (MEST) (2012-0009457).
Conclusions In this work, sodium excess Ta-modified KNNT ceramics were synthesized by reactive template grain growth method, and their crystal structure, microstructure, dielectric, and field-induced strain properties were
References 1. Directive 2002/95/EC of the European Parliament and of the Council of 27 January 2003, Off. J. Eur. Union. L37 19–23 (2003). 2. Y. Saito, et al., Nature. 432 84–87 (2004).
234
International Journal of Applied Ceramic Technology—Hussain, et al.
3. E. Cross, Nature, 432 24–25 (2004). 4. Y. Li, K. Moon, and C. P. Wang, Science, 308 1419–1420 (2005). 5. J. Ro¨del, W. Jo, K. T. P. Seifert, E. M. Anton, and T. Granzow, J. Am. Ceram. Soc., 92 1153–1177 (2009). 6. K. Wang and J. F. Li, Adv. Funct. Mater., 20 1924–1929 (2010). 7. D. M. Lin, M. S. Guo, K. H. Lam, K. W. Kwok, and H. L. W. Chan, Smart Mater. Struct., 17 035002–035008 (2008). 8. R. Z. Zuo, J. Ro¨del, R. Z. Chen, and L. T. Li, J. Am. Ceram. Soc., 89 2010–2015 (2006). 9. R. Z. Zuo and J. Fu, J. Am. Ceram. Soc., 94 1467–1470 (2010). 10. Y. F. Chang, S. F. Poterala, Z. P. Yang, S. Trolier-Mckinstry, and G. L. Messing, Appl. Phys. Lett., 95 2329051–2329053 (2009). 11. R. E. Jaeger and L. Egerton, J. Am. Ceram. Soc., 45, 209–213 (1962). 12. H. Yilmaz, G. L. Messing, and S. Trolier-Michenstry, J. Electroceram., 11 207–215 (2003). 13. H. Takao, Y. Saito, Y. Aoki, and K. Horibuchi, J. Am. Ceram. Soc., 89 [6] 1951–1956 (2006). 14. J. Hao, C. Ye, B. Shen, and J. Zhai, Phys. Status Solidi A, 209 1343–1349 (2012). 15. T. Kimura, J. Ceram. Soc. Jpn., 114 15–25 (2006). 16. J. A. Horn, S. C. Zhang, U. Selvaraj, G. L. Messing, and S. TrolierMcKinstry, J. Am. Ceram. Soc., 82 [4] 921–926 (1999). 17. G. L. Messing, et al., Crit. Rev. Solid State Mater. Sci., 29 45–96 (2004). 18. T. Kimura, T. Takahashi, T. Tani, and Y. Saito, J. Am. Ceram. Soc., 87 [8] 1424–1429 (2004). 19. S. Sun, et al., Appl. Phys. Lett., 84 574–577 (2004). 20. M. Wu, Y. Li, D. Wang, J. Zeng, and Q. Yin, J. Electrocerm., 22 131–135 (2009).
Vol. 12, No. 1, 2015
21. Y. Li, C. Hui, M. Wu, Y. Li, and Y. Wang, Ceram. Int., 38S S283–S286 (2012). 22. M. Wu, Y. Lia, D. Wang, and Q. Yin, Ceram. Int., 34 753–756 (2008). 23. T. Richter, S. Denneler, C. Schuh, E. Suvaci, and R. Moos, J. Am. Ceram. Soc., 91 929–933 (2008). 24. F. Gao, C. S. Zhang, X. C. Liu, L. H. Cheng, and C. S. Tian, J. Eur. Ceram. Soc., 27 3453–3458 (2007). 25. C. Duran, S. Trolier-McKinstry, and G. L. Messing, J. Am. Ceram. Soc., 83 2203–2213 (2000). 26. A. Hussain, J. S. Kim, T. K. Song, M. H. Kim, W. J. Kim, and S. S. Kim, Curr. Appl. Phys., 13 1055–1059 (2013). 27. Y. S. Sung, et al., Appl. Phys. Lett., 101 012902–012903 (2012). 28. Y. S. Sung, et al., Ceram. Int., 38 S301–S304 (2012). 29. A. Hussain, J. S. Kim, G. H. Ryu, T. K. Song, M. H. Kim, and W. J. Kim, Ceram. Int., 39 S365–S368 (2013). 30. H. Yilmaz, S. Trolier-McKinstry, and G. L. Messing, J. Electroceram., 11 217–226 (2003). 31. D. Lv, R. Zuo, and S. Su, J. Mater. Sci. Mater. Electron., 23 1367–1372 (2012). 32. Y. L. Li, C. Hui, Y. X. Li, and Y. L. Wang, J. Inorg. Mater., 27 314–318 (2012). 33. H. J. Cho, M. H. Kim, T. K. Song, J. S. Lee, and J.-H. Jeon, J. Electroceram., 30 72–76 (2013). 34. S. Zhang, R. Xia, T. R. Shrout, G. Zang, and J. Wang, J. Appl. Phys., 100 104108–104111 (2006). 35. M. Matsubara, T. Yamaguchi, K. Kikuta, and S. I. Hirano, Jpn. J. Appl. Phys., 44 258–263 (2005). 36. B. Malic, J. Bernard, J. Holc, D. Jenko, and M. Kosec, J. Eur. Ceram. Soc., 25 2707–2711 (2005).
