Abstract
Keywords: Dielectric behavior and diffuse phase transition have been investigated in Sb doped lead lanthanum zir- Dielectric behavior conate titanate (PLZT) ceramics to understand the role of La and Sb in ferroelectric-to-relaxor crossover. Ferroelectrics Rayleigh and Curie–Weiss type law fittings show that the La substitution gradually transforms the Phase transition PLZTs from ferroelectric to relaxor, while Sb dopants weaken the diffuseness and dielectric disper- Relaxor sion. Vogel–Fulcher model reveals that the samples at high La concentration are fully in relaxor state comparable to the undoped PLZTs. Production and hosting by Elsevier B.V. on behalf of The Ceramic Society of Japan and the Korean Ceramic Society. 1. Introduction co-dope La and Sb ions to modify the piezoelectric properties of PZT ceramics and investigate these ceramic’s phase transitions. Solid solution of Pb La (Zr Ti ) O (PLZT) is the Pérez-Delfin et al. [10] studied the relaxor behavior in MnO doped 1−x x y 1−y 3 2 1−x/4 x/4 most investigated A-site substituted lead zirconate titanate (PZT) PLZT ceramics. However, to our knowledge, careful reports on the 3+ ferroelectric material for wide applications [1–6]. The La substi- ferroelectric–relaxor transition of PLZT due to the introduction of 2+ tution for Pb of the PZTs introduces A-site vacancies, decreases aliovalent cation are still absent. In this letter, with a fixed amount oxygen vacancies, and reduces the domains to nano-size, inducing of Sb O (1.5 wt%) addition, the effect of La substitution for Pb 2 5 a gradual ferroelectric–relaxor transition. Polar relaxation like fre- on the ferroelectric-to-relaxor phase transition of PLZT x/52/48 quency dispersion is not featured in PZTs because of their normal, (2 ≤ x ≤ 16, where x is the mole fraction) was investigated system- long-range ferroelectric order. However, when aliovalent cations, atically, and the intrinsic mechanism was discussed. 3+ e.g. La , are substituted into PZT body, random distributed vacan- cies hinder the onset of long-range order (LRO). These vacancies 2. Experimental procedure and composition fluctuations create local heterogeneous structures in the form of polar nano-regions (PNRs), which can transform into All samples were prepared by the conventional solid-state reac- macroscopic domains with assistance of external electric field. At tion route. Raw materials of Pb O , La O , TiO , ZrO , and Sb O 3 4 2 3 2 2 2 5 this point, a relaxor state is formed, featured with a slim electri- were weighed carefully according to the chemical composition cal hysteresis loop that makes it a great candidate for capacitive of Pb La (Zr Ti ) O + 1.5 wt% Sb O (x = 0.02, 0.04, 1−x x 0.52 0.48 3 2 5 1−x/4 x/4 energy storage application [4,6,7]. 0.06, 0.08, 0.10, 0.12, 0.14, 0.16). 0.3 wt% excess Pb O was added 3 4 Although not all aliovalent cations can trigger the relaxor state to compensate the lead loss during sintering at high tempera- 5+ in PZTs [4,5], some dopants, e.g. Sb , can boost the piezoelec- ture. The mixtures were ball-milled in distilled water for 2 h, and tric activity of PZTs. Therefore, it is interesting to dope one or then dried and calcined in air at 900 C for 4 h. Subsequently, the 3+ more element spices along with the La substitution to investi- ball-milled powders were pressed into cylindrical disks under a gate the ferroelectric–relaxor transition in PLZTs. Several groups pressure of 25 MPa. After burning out the binder of paraffin wax at have already tried to co-dope La and other ions into PZTs. For 760 C for 3.5 h, the green compacts were sealed in crucibles and example, Helke and Lubitz [8] and Rai and Sharma [9] tried to sintered at 1250 C for 2 h. The as-prepared pellets were polished, and silver electrodes were screen-printed and fired at 800 C for Peer review under responsibility of The Ceramic Society of Japan and the Korean 10 min. Finally, the samples with diameter of 12 mm and thickness Ceramic Society. of 1.5 mm were obtained. The crystalline structure was determined by an X-ray diffractometer (XRD, PANalytical X’Pert PRO) with Cu radiation K1 at room temperature. The dielectric properties were measured using LCR meters (HP4194A and TH2816A). Commercial temperature chambers (MPC-2000A, −65 to 150 C; Nabertherm, 2187-0764 Production and hosting by Elsevier B.V. on behalf of The Ceramic 25–1300 C) were used to support the dielectric measurement at Society of Japan and the Korean Ceramic Society. various temperatures. http://dx.doi.org/10.1016/j.jascer.2013.12.006 α α tanδ δ 2 Letter / Journal of Asian Ceramic Societies 2 (2014) 1–4 4000 0.2 Pyrochlore 3000 0.15 0 10 20 La mol% 2000 0.1 2 6 10 14 4 8 12 16 1000 0.05 0 0 0.1 1 10 100 1000 f(kHz) 4 4000 0.03 ε ε α x=2 R r 0.02 20 30 40 50 60 70 ο ο 2θ θ ( ( ) ) 0.01 Fig. 1. XRD patterns of Sb doped PLZT x/52/48 (2 ≤ x ≤ 16) ceramics. Inset is the pyrochlore phase volume percentage (X ) as a function of La concentration. 2000 0 0 5 10 15 3. Results and discussion La mol% Fig. 1 shows the XRD patterns of Sb doped PLZT ceramic pow- Fig. 2. (a) Room temperature permittivity (ε ) and dielectric loss (tan ı) of Sb doped PLZT (x/52/48) as a function of frequency and (b) dependence of ε (at 1 kHz) and ders. XRD data reveals that the Sb doped PLZT ceramics were well calculated Rayleigh law fitting parameters (ε , ˛) on La concentration. crystallized. For x ≤ 4 mol%, only perovksite phase (Pe-phase) can be observed within the detection limit of XRD. For x ≥ 6 mol%, with dissimilar distortion. Yamamura et al. [19] reported that such pyrochlore phase (JCPDS 72-2370, Py-phase) was present, whose 5 7 defect Py-phase compositions can have permittivity up to 10 –10 . peak intensities increase with the increase of La concentration. To Thus, one of the following scenarios may occur in our samples: (a) estimate the volume percent (X ) of Py-phase by the ratio of the the Py-phase is a Pe-phase in a different structure; (b) the Py-phase highest peak intensities, peak (2 2 2) for the Py-phase to the sum has comparable ε with that of Pe-phase; (c) the Py-phase has a of peak (2 2 2) for the Py-phase and peak (1 0 1)/(1 1 0) for the Pe- r discontinuous distribution, e.g. in the form of columns that paral- phase, it is found that the X increases linearly from 0% to 10.6% lel to the external electric field. Any of the above assumption can in the range of x = 5–14%, and are stabilized at 10.6% at x = 16 (the zero or minimize the influence from the Py-phase to the dielectric inset in Fig. 1). Based on the facts that second phase is absent in PLZT properties. x/52/48 (0 ≤ x ≤ 20) [1,2,11] and the solubility limit of Sb in PZT is The dielectric behavior of ferroelectric materials adheres to around 2% [12], one can expect that the Py-phase is introduced by Rayleigh model up to electric field of ∼1/2 of the coercive electric the Sb dopants. Yet, due to the complication of the pyrochlore sys- field [20]. Note that the non-ferroelectric materials, e.g. Py-phase, tem (A B O , A = La, Sb, Pb, B = Zr, Ti, Sb) in PZT-based ceramics, 2 2 7−ı do not obey the model. In a random system, where restoring forces the accurate chemical composition of the Py-phase is not iden- are in normal distribution, such extrinsic contributions, e.g. domain tified. The Py-phase lattice parameters based on peak (2 2 2) are ˚ wall motion, obey the following equation before the Debye relax- calculated as 10.67 ± 0.17 A, suggesting a single composition in the ation [21–23]: samples. The influence of the Py-phase on the dielectric behavior is discussed as follows. ε(f ) = ε (1 + ˛lg(f )) (1) Fig. 2a shows the variation of permittivity (ε ) and dielectric loss (tan ı) of Sb doped PLZT ceramics with frequency at room temper- where f is the measure frequency, ε is the reversible permittivity, ature. The values of ε and tan ı are between 2200 and 3800, and and ˛ is the Rayleigh fitting coefficient. The frequency dependent from ∼0.02 to 0.04, respectively. They are very much comparable permittivity of the samples was plotted in Fig. 2a with satisfied fit- to those of PLZT ceramics at the same La doping level. Generally, ting based on Eq. (1). The fitting parameters were plotted in Fig. 2b. Py-phase is supposed to be deleterious in ferroelectric materials It is found that the reversible permittivity has very close values to because of its low permittivity [13–16]. Even a small amount of those measured at 1 kHz. The Rayleigh fitting coefficients gradually Py-phase can diminish the dielectric properties intensively [3,17]. increase from 0.015 to 0.025 for 4 ≤ x ≤ 16, indicating the reversible However, the samples in this study do not show such low ε val- domain wall motion has greater contribution to the ε at higher La r r ues. Similar phenomena were seen in the literature [9,18], where concentration. Narayanan et al. [24] found that the intrinsic permit- one paper attributed it to a single phase of perovskite structure yet tivity of PLZTs is very comparable to that of PZTs. In other words, the Intensity (a.u.) (001)/(100) (222) (101)/(110) (400) (111) X % (002)/(200) (440) (102)/(201) (211)/(121) (622) (022)/(220) ε ε ε ε r Letter / Journal of Asian Ceramic Societies 2 (2014) 1–4 3 appeared at lower temperatures of ∼40 C compared with their T (a) 4 counterparts, and then reduced to lower values at higher temper- ature in the paraelectric phase when domain wall vanishes. The 1kHz 200kHz 6 8 3+ 15000 dielectric peaks showed intense peaks in the mole fraction of La 10kHz 100kHz from 10% to 16%. Such phenomenon may have strong correlation with the impurity. Yet in the form of ceramics whose dielectric x=16 loss is extrinsic, the complexity of dielectric loss can be deter- mined by, not only the material composition, but also the metal and oxygen vacancies, local composition fluctuations, etc. Thus, 0.08 more discussion on the dielectric loss is not given in this work. It has been reported that such frequency dispersion is initialized at 0.06 La = 6 and 12 mol% in PLZT compositions of x/65/35 and x/60/40, x=16 respectively [29,30]. Thus, the ferroelectric–relaxor boundary of 0.04 1 kH z La = 10 mol%, when frequency dispersion in the samples are notice- able, is a realistic value in consideration the phase diagram of PLZTs. 0.02 6 More detailed comparisons of the phase transition parameters T 2 m and ε of this study and PLZTs are illustrated in Fig. 3b. Both Sb doped PLZTs and non-doped PLZT have decreasing peak temper- 0 50 100 150 200 250 300 350 atures as a function of La concentration. The Sb doped PLZTs has T ( C) smaller T and ε at low concentrations (x ≤ 4) and greater ones m m (b) at high concentrations (x ≥ 6). The Sb dopant may generate more 4 10 Pb vacancies that lower the values of T and ε when PLZTs are in m m with Sb 3+ the ferroelectric state. At higher mole fraction of La , Sb dopant wit hout Sb 3 10 may delay and hinder the formation of PNRs. Consequently, higher values of T and ε were found compared with those of undoped 200 m m 2 10 PLZTs. To further investigate the La concentration effect in the 1 10 ferroelectric–relaxor transition, and to describe qualitatively the shape of dielectric anomaly, Curie–Weiss (C–W) type law of relaxor is used to simulate the deviation from C–W law. 0 5 10 15 20 0 5 10 15 20 Two main features of ε (T) were used to identify a relaxor from a ferroelectric: (a) the broad permittivity peak (strong deviation from La mol% C–W law), namely the diffuse phase transition; (b) the frequency Fig. 3. (a) Temperature dependent ε and tan ı of Sb doped PLZT (x/52/48) ceramics r dispersion. The first one is usually measured with the C–W type and (b) peak temperature (T ) and peak permittivity (ε ) vs. La mole fraction. Note m m law at T ≤ T ≤ T (Burns temperature, above which relaxors are in that the PLZT data are based on compositions of x/53/47 adopted from Ref. [26–28]. paraelectric phase) [4–6]: ε (T − T ) m m increased permittivity at high La concentration is extrinsic due to − 1 = (2) r 2ı the increase of domain wall mobility. These simulation results are expected during a typical ferroelectric-to-relaxor crossover where where ı is a constant, measuring the degree of the peak dif- decreased grain/domain size eases the domain wall motion and fuseness. The fitting of the data was plotted in Fig. 4a. The fitting facilitates the domain reorientation. The frequency dependence in parameter ı is shown in Fig. 4b, where greater values of ı were m m this study, however, is weaker compared with pure PLZT. Pérez- found in higher La concentration samples, inferring broadened Delfin et al. [10] reported that the ε decrease from 4000 at 0.5 Hz to peaks and the intense composition fluctuations. 2700 at 1 MHz, leaving the value of ˛ ≈ 0.070. But the value drops to One of the methods to qualitatively demonstrate the frequency ∼0.010 in the same PLZT composition when MnO dopant (2 wt%) dispersion is using the peak temperature drift (T ) of permittivity are used. As Ahart et al. [25] and Pérez-Delfin et al. [10] discussed, peaks measured at two different frequencies. The delta T , defined the PNRs in relaxors can be easily aligned and grow with exter- as the value differences between T measured at f = 1 and 200 kHz, 5+ 4+ nal electric field, but when additional dopants, e.g. Sb , Mn , are was shown in Fig. 4b. Like the trend of ı , the values of T were m m introduced to the system, these new cations/vacancies yield addi- greater at higher La concentration from 1 K at x = 6, to the maximum tional internal electric fields that may pin and couple to the PNRs, of 24 K at x = 14. and break short coherent length (SCL) cation order. As a conse- The frequency dispersion of Sb doped PLZTs is significant at quence, the PNRs become macro-size domains, and relaxor state x ≥ 14, where dipole glass model is widely used to demonstrate disappears. Therefore, PLZT with Sb dopant is less disordered. Since the relaxor state [5,6,20]. This model adopted Vogel–Fulcher (V–F) non-ferroelectric compositions do not obey the Rayleigh model, the analysis describes the freezing of the dipoles results from cooper- frequency dependent dielectric responses also suggested that the ative interactions between moments on the nanometer scale. The Py-phase has minor to no influence in the dielectric response of Sb V–F function is given below as [5,6]: doped PLZT. The dielectric response for Sb doped PLZT x/52/48 (2 ≤ x ≤ 16) −U f = f exp (3) compositions is shown in Fig. 3a. The permittivity peak (ε ) and k (T − T ) B m f temperatures (T ) linearly decreases associating the broader per- mittivity peaks as the value of x increases. Dielectric frequency where f is the measure frequency, f is a fitting coefficient related dispersion (frequency dependent peak drifting) is not obvious until to dipole freezing time, U is the activation energy, k is the Boltz- 3+ the mole fraction of La is higher than 10%. Similar tendency were mann constant, and T is the glass freezing temperature. The fitting seen in undoped PLZTs [4,6], meaning that La ions gradually trans- of the data as shown in inset of Fig. 4b yields f = 3.9 × 10 Hz 13 ◦ ◦ form the PZTs from ferroelectric to relaxor. The dielectric loss peaks (1.51 × 10 Hz), U = 0.26 eV (0.21 eV), and T = 29.0 C (1.5 C) for T ( C) m ε ε tan δ δ ε ε m Δ Δ T (K) 4 Letter / Journal of Asian Ceramic Societies 2 (2014) 1–4 [3] G.H. Haertling, J. Am. Ceram. Soc., 82, 797–818 (1999). (a) [4] L.E. Cross, in Piezoelectricity, Ed. by W. Heywang, K. Lubitz and W. Wrsing, Springer, Berlin, Heidelberg (2008) pp. 131–155. [5] A.A. Bokov and Z.G. Ye, J. Mater. Sci., 41, 31–52 (2006). 1.5 [6] Z.G. Ye, Key Eng. Mater., 155–156, 81–122 (1998). [7] S. Tong, B. Ma, M. Narayanan, S. Liu, R. Koritala, U. Balachandran and D. Shi, ACS Appl. Mater. Interfaces, 5, 1473–1480 (2013). 1 8 [8] G. Helke and K. Lubitz, in Piezoelectricity, Ed. by W. Heywang, K. Lubitz and W. Wrsing, Springer, Berlin, Heidelberg (2008) pp. 89–130. 0.5 [9] R. Rai and S. Sharma, Ceram. Int., 30, 1295–1299 (2004). [10] E. Pérez-Delfin, J.E. Garcia, A. Vega-Garcia, F. Guerrero and J.A. Eiras, J. Eur. Ceram. Soc., 32, 1659–1665 (2012). [11] W.A. Schulze, T.G. Miller and J.V. Biggers, J. Am. Ceram. Soc., 58, 21–23 0 20 40 60 80 100 120 (1975). T-T (K) [12] Z. Ling, W. Qiu and K. Wang, Electro. Comp. Mater., 24, 31–34 (2005). [13] J.C. Champarnaud-Mesjard, M. Manier, B. Frit and A. Tairi, J. Alloys Compd., 188, (b) 120 40 174–178 (1992). 20 [14] D. Dhak, P. Dhak and P. Pramanik, Appl. Surf. Sci., 254, 3078–3092 (2008). 100 15 [15] H. Du, H. Wang and X. Yao, Ceram. Int., 30, 1383–1387 (2004). x =16 10 [16] L. Cai and J.C. Nino, J. Eur. Ceram. Soc., 27, 3971–3976 (2007). 80 x =14 [17] M.A. Mohiddon, R. Kumar, P. Goel and K.L. Yadav, IEEE Trans. Dielectr. Electr. 20 Insul., 14, 204–209 (2007). 50 100 150 60 [18] S. Dutta, R.N.P. Choudhary and P.K. Sinha, J. Alloys Compd., 426, 345–351 T ( C) (2006). 40 [19] H. Yamamura, H. Nishino and K. Kakinuma, J. Phys. Chem. Solids, 69, 1711–1717 (2008). 20 0 [20] P. Bintachitt, S. Jesse, D. Damjanovic, Y. Han, I.M. Reaney, S. Trolier- McKinstry and S.V. Kalinin, Proc. Natl. Acad. Sci. U.S.A., 107, 7219–7224 0 5 10 15 (2010). La mol% [21] D.A. Hall, J. Mater. Sci., 36, 4575–4601 (2001). [22] F. Griggio, S. Jesse, A. Kumar, O. Ovchinnikov, H. Kim, T.N. Jackson, D. Dam- Fig. 4. (a) C–W type law fitting of diffuse phase transition above T and (b) as- janovic, S.V. Kalinin and S. Trolier-McKinstry, Phys. Rev. Lett., 108, 157604 calculated diffuse coefficients ı and T vs. La concentration; Inset is the V–F law (2012). m m [23] M. Narayanan, U. Balachandran, S. Stoupin, B. Ma, S. Tong, S. Chao and S. Liu, J. fittings for PLZT x/52/48 (x = 14, 16). Phys. D, 45, 335401 (2012). [24] M. Narayanan, S. Tong, S. Liu, B. Ma and U. Balachandran, Appl. Phys. Lett., 102, x = 14 mol% (16 mol%). These values are comparable to the open lit- 062906 (2013). [25] M. Ahart, M. Somayazulu, R. Cohen, P. Ganesh, P. Dera, H. Mao, R.J. Hemley, Y. erature of relaxor PLZTs [31–33]. The above simulation indicates Ren, P. Liermann and Z. Wu, Nature, 451, 545–548 (2008). that, although Sb dopants can weaken the effects of La ions in [26] S.M. Gupta, J.F. Li and D. Viehland, J. Am. Ceram. Soc., 81, 557–564 ferroelectric–relaxor transition, at high La concentrations, 1.5 wt% (1998). [27] K. Wojcik, J. Błaszczak and J. Handerek, Ferroelectrics, 70, 39–46 (1986). Sb O doped PLZTs can be fully transformed into relaxor state sim- 2 5 [28] A. Peláiz Barranco, F. Calderón Pinar ˜ and O. Pérez Martínez, Phys. Stat. Sol. B, ilar to non-doped PLZTs. 220, 591–595 (2000). [29] G.A. Samara, J. Phys. Condens. Matter, 15, R367–R411 (2003). [30] G.A. Samara and E.L. Venturini, Phase Transitions, 79, 21–40 (2006). 4. Conclusion [31] J.L. Dellis, J. Phys. Condens. Matter, 8, 7957–7965 (1996). [32] S. Kamba, V. Bovtun, J. Petzelt, I. Rychetsky, R. Mizaras, A. Brilingas, J. Banys, J. In conclusion, we have investigated the Sb doping and La sub- Grigas and M. Kosec, J. Phys. Condens. Matter, 12, 497–519 (2000). stitution influence to the ferroelectric-to-relaxor crossover in Sb [33] F. Cordero, F. Craciun, A. Franco, D. Piazza and C. Galassi, Phys. Rev. Lett., 93, 097601 (2004). doped PLZT ceramics. The prepared samples transform from fer- roelectric to relaxor as the La concentration gradually increased Sheng Tong from 2 to 16 mol%, with a boundary at x = 10. La substitution eases ∗∗ Zhiyuan Ling the reversible movement of domain walls, and Sb dopants weaken Department of Electronic Materials Science and such effect. Although Sb dopant weakens the formation of PNRs, Engineering, School of Materials Science and Sb doped PLZTs are in pure relaxor state at high La concentration Engineering, South China University of Technology, (x ≥ 14). Our results suggest that modified relaxors with improved Guangzhou 510640, China dielectric properties while retaining their features as relaxors can be achieved by doping in pure relaxor composition, e.g. high La Corresponding author. Current address: concentration PLZT and Pb(Mg Nb )O . 1/3 2/3 3 Nanoscience and Technology Division, Argonne National Laboratory, Argonne, IL 60439, USA. Conflict of interest statement Tel.: +1 630252 4628; fax: +1 630252 4646. The authors declare that there are no conflicts of interest. ∗∗ Corresponding author. E-mail addresses: stong@anl.gov, Acknowledgments shengtg@mail.uc.edu (S. Tong), imzyling@scut.edu.cn (Z. Ling) This work was supported by Fair of Science and Technical Achievements Resulted from Cooperation of Industry, Education 30 November 2013 and Academy, Guangdong, China (2010A09020059). 22 December 2013 References 27 December 2013 [1] G.H. Haertling and C.E. Land, J. Am. Ceram. Soc., 54, 1–11 (1971). [2] G.H. Haertling, J. Am. Ceram. Soc., 54, 303–309 (1971). Available online 19 January 2014 ε ε ___ δ δ (K) ε ε lnf(Hz)
Journal
Journal of Asian Ceramic Societies
– Taylor & Francis
Published: Mar 1, 2014
Keywords: Dielectric behavior; Ferroelectrics; Phase transition; Relaxor
