9-7-2012 Creation and Destruction of Morphotropic Phase Boundaries through Electrical Poling : A Case Study of Lead-Free ( Bi 1 / 2 Na 1 / 2 ) TiO 3-BaTiO 3
Cheng Ma,Hanzheng Guo,Scott P. Beckman,Xiaoli Tan
2017-01-01
Abstract:We report the first direct evidence that the morphotropic phase boundary in ferroelectric materials, along with the associated strong piezoelectricity, can be created, destroyed, or even replaced by another morphotropic phase boundary through phase transitions during electrical poling. The real-time evolution of crystal structure and domain morphology during the poling-induced phase transitions in (Bi1/2Na1/2)TiO3BaTiO3 is observed with in situ transmission electron microscopy. These observations elucidate the microstructural origin of the macroscopic piezoelectricity’s dependence on the poling field and previously unexplained strain behaviors. This study demonstrates that the ferroelectric-to-ferroelectric transitions during the poling process can completely alter the morphotropic phase boundaries and, hence, must be comprehensively investigated when interpreting the microscopic mechanism of macroscopic piezoelectric behaviors. Disciplines Ceramic Materials | Materials Science and Engineering Comments This article is from Physical Review Letters 109 (2012): 107602, doi:10.1103/PhysRevLett.109.107602. Posted with permission. This article is available at Iowa State University Digital Repository: http://lib.dr.iastate.edu/mse_pubs/25 Creation and Destruction of Morphotropic Phase Boundaries through Electrical Poling: A Case Study of Lead-Free ðBi1=2Na1=2ÞTiO3-BaTiO3 Piezoelectrics Cheng Ma, Hanzheng Guo, Scott P. Beckman, and Xiaoli Tan* Department of Materials Science and Engineering, Iowa State University, Ames, Iowa 50011, USA (Received 15 March 2012; published 7 September 2012) We report the first direct evidence that the morphotropic phase boundary in ferroelectric materials, along with the associated strong piezoelectricity, can be created, destroyed, or even replaced by another morphotropic phase boundary through phase transitions during electrical poling. The real-time evolution of crystal structure and domain morphology during the poling-induced phase transitions in ðBi1=2Na1=2ÞTiO3-BaTiO3 is observed with in situ transmission electron microscopy. These observations elucidate the microstructural origin of the macroscopic piezoelectricity’s dependence on the poling field and previously unexplained strain behaviors. This study demonstrates that the ferroelectric-to-ferroelectric transitions during the poling process can completely alter the morphotropic phase boundaries and, hence, must be comprehensively investigated when interpreting the microscopic mechanism of macroscopic piezoelectric behaviors. DOI: 10.1103/PhysRevLett.109.107602 PACS numbers: 77.65. j, 61.05.jm, 77.80.B , 77.84.Cg Amaterial with coexisting energetically comparable states often displays extraordinary responses to external stimuli, such as large magnetostriction [1], giant electrocaloric effect [2], and colossal magnetoresistance [3]. In the case of piezoelectricity, ultrahigh piezoelectric responses are realized in compositions at the morphotropic phase boundary (MPB), where multiple ferroelectric phases with different crystal structures coexist [4–9]. Polycrystalline MPB materials must be poled by electric fields in order to exhibit piezoelectricity [4]. Under poling fields, the energy landscape of the constitutive polar phases in the MPB composition may be altered differently. Therefore, knowledge about the real-time evolution of crystal structures and domain morphologies during and after electrical poling is essentially important in understanding the physics of piezoelectricity. However, in the past six decades, both theoretical and experimental studies have mainly focused on the virgin state of materials, prior to the application of a poling field [4–9]. Although different models on the piezoelectricity enhancement associated with MPBs have been proposed to describe interesting microscopic behaviors at zero field, the essential understanding of the microstructural mechanisms cannot be established without a complete picture of the in situ structural evolution under poling fields [10]. As a result, many fundamental questions have remained unanswered for decades: Will MPBs survive the strong poling fields? Can the poling field create an MPB in originally single-phase ferroelectrics? If poling fields can create or destroy MPBs, will such changes be preserved after the fields are removed? How does the change in MPBs during poling impact the macroscopic piezoelectric property? Conclusive answers to these questions demand in situ real-time structural characterizations. There are two principal difficulties that hinder such studies. First, the crystal structures of different perovskite ferroelectric phases are frequently too similar to be well resolved by x-ray or neutron diffractions. It is particularly difficult to distinguish a material with coexisting phases from one single phase with a more complicated crystal structure. For example, there has been a continued debate on the microstructure of the MPB in PbðZr;TiÞO3 (single monoclinic phase vs mixed rhombohedral and tetragonal phases) [5,6], the most important and widely studied piezoelectric compositions. Second, in situ application of electric fields in ferroelectric materials inevitably leads to a large piezoelectric strain and causes a complicated shifting of diffraction peaks [10]. Moreover, electric-field-induced domain texture must be taken into account in the structural analysis. These complications make the phase identification using x-ray or neutron diffractions particularly challenging. These issues can be alleviated by using electric-field in situ transmission electron microscopy (TEM). Unlike x-ray or neutron diffraction, which obtains average structural information of materials, TEM is capable of precisely resolving the crystal structure of each individual phase among mixed phases through selected area electron diffraction. Furthermore, in situ TEM visualizes the real-time domain morphology changes under an electric field. However, electric-field in situ TEM requires very delicate sample preparation, and thus the number of such studies is still very limited [11–15]. In this Letter, direct evidence from in situ TEM experiments and corresponding piezoelectricity measurements provides conclusive answers to the questions raised above; we demonstrate that MPBs, and the enhanced piezoelectric response associated with them, can be either destroyed or created irreversibly during poling. Our first-principles calculations reveal that the ferroelectric phases are energetically close to each other, which corroborates the experimental conclusions. PRL 109, 107602 (2012) P HY S I CA L R EV I EW LE T T E R S week ending 7 SEPTEMBER 2012 0031-9007=12=109(10)=107602(5) 107602-1 2012 American Physical Society One of the most extensively studied lead-free piezoelectrics, the ð1 xÞðBi1=2Na1=2ÞTiO3-xBaTiO3 solid solution [9,16–18], was selected to demonstrate the concept. In the virgin state, these ceramics crystallize in the rhombohedral R3c ferroelectric phase for x 6%, the tetragonal P4bm relaxor ferrielectric phase for 5% x 11%, and the tetragonal P4mm ferroelectric phase for x 10% [19–21]. Therefore, a R3c=P4bm MPB at 5% x 6% and a P4bm=P4mm MPB at 10% x 11% exist in the unpoled state. Compared with results reported on poled ceramics [22–25], an irreversible P4bm-to-P4mm phase transition is speculated to occur during poling for compositions with x 7%. We have recently confirmed this transition by in situ structural studies for x 1⁄4 7% with poling fields up to 2:5 kV=mm [14]. However, the stability of the induced P4mm phase in x 1⁄4 7% against poling fields above 2:5 kV=mm is unknown. Any additional ferroelectric-to-ferroelectric transitions at higher fields may create an MPB in this composition. By the same token, electric-field-induced phase transitions in 5% x 6%, if any, could possibly destroy the existing MPB. To verify these speculations, our study focuses on three compositions: x 1⁄4 5:5%, 6%, and 7%. The preparation of ðBi1=2Na1=2ÞTiO3-BaTiO3 ceramics and TEM specimens has been described elsewhere [20]. The electric-field in situ TEM technique is illustrated in Supplemental Material [26] and described in previous studies [12,13]. For each of the three compositions, over ten corroborating in situ TEM experiments were conducted. The piezoelectric coefficient d33 was measured on specimens poled at room temperature under fields from 1.5 to 6:5 kV=mm with increasing sequence of field levels. The first-principles calculations were performed on the base compound ðBi1=2Na1=2ÞTiO3 by using the density functional theory methods encoded in the ABINIT software package [27]. The exchange-correlation energy was approximated by using the local density approximation [28]. In place of the all-electron ion potentials, projector augmented wave pseudopotentials were employed [29]. The plane wave expression of the wave function was truncated at 816 eV, and a Monkhorst-Pack [30] mesh was used to sample the Brillouin zone. For these calculations the forces on the atoms were converged to better than 10 meV= A. The supercells used in the calculation were based on a superlattice constructed of alternating layers of BiTiO3 and NaTiO3 stacked in the [111] and [001] directions. The in situ TEM results for x 1⁄4 6% are displayed in Fig. 1. At zero field, consistent with previous studies [19,20], approximately 40% of the grains exhibit mixed P4bm and R3c phases (see Supplemental Material [26]), and the rest of the grains consist of P4bm nanodomains only [Fig. 1(a)]. The P4bm and R3c symmetries are unambiguously determined from diffraction patterns: The P4bm symmetry is characterized by the presence of 1=2fooeg (o and e denote odd and even Miller indices, respectively) and the absence of 1=2fooog superlattice diffraction spots [Fig. 1(e)] and the R3c symmetry by the presence of 1=2fooog and the absence of 1=2fooeg spots [19–21,31–34]. A grain with FIG. 1. TEM observations along the [112] zone axis of an origi