INTRODUCTION
Perovskite materials with two-dimensional microstructures have attracted significant attention in recent years because of their unique shape-dependent functional properties. Anisotropic perovskite particles can be widely used as templates to texture electronic ceramics[1-3], substrates to grow crystals or deposit films[4,5], fillers in polymer systems for energy storage applications[6,7] and building blocks to fabricate functional micro/nanodevices (e.g., energy harvesting systems and field effect transistors)[8,9]. Appropriate crystal orientation, morphology and crystallinity of the particles are essential for the design of microplatelets with the desired functionality. However, anisotropic perovskite platelets are difficult to synthesize via conventional methods due to the similar surface energies of the crystal planes[10]. Until now, the most effective method for fabricating platelet perovskites has been based on the topochemical microcrystal conversion (TMC) of two-dimensional precursors with perovskite units via the deletion, insertion or exchange of individual atoms to transform localized structures[11-14].
(Bi0.5Na0.5)TiO3 is considered one of the key lead-free candidates to potentially replace commercialized toxic Pb(Zr,Ti)O3-based systems, owing to the outstanding electrical properties of its solid solutions with other perovskite components at the morphotropic phase boundary (MPB). (1-x)(Bi0.5Na0.5)TiO3-xBaTiO3 (BNT-BT) solid solutions have recently received increasing interest, owing to their substantially enhanced piezoelectric properties achieved near the MPB (i.e., 0.05 ≤ x ≤ 0.07)[15,16]. Kang et al.[17] investigated the effects of BT amount on the piezoelectric and energy harvesting properties of (1-x)BNT-xBT ceramics and found that the maximum piezoelectric charge coefficient d33 ~164 pC/N and output voltage V ~8.95 V were achieved at x = 0.06. Zhao et al.[18] studied the polar domain structural evolution of 0.94BNT-0.06BT under external fields, while Das Adhikary et al.[19] investigated its random lattice strain and relaxation, with both studies revealing the impact of these effects on the structural and electrical properties of the material. Furthermore, textured ceramics of 0.94BNT-0.06BT or similar compositions have been synthesized through templated grain growth via the use of Bi4Ti3O12[20], SrTiO3[21], BaTiO3[22], BNT[23,24] and NaNbO3[25] platelet templates. It was found that the addition of heterogeneous templates causes both lattice mismatch with the matrix and compositional changes in the final ceramics, which can deteriorate the piezoelectric properties, lower the depolarization temperature and thereby cause instability in the final textured ceramics. The utilization of BNT-BT platelets as templates instead of the above mentioned compositions could be advantageous to texture BNT-BT-based ceramics, with the aim of achieving significantly enhanced piezoelectric properties without lowering the depolarization temperature.
In this work, 0.94BNT-0.06BT microplatelets with controlled morphology and good crystallinity are fabricated by a TMC approach using Aurivillius Bi4.5Na0.5Ti4O15 (BINT) precursors. The BINT microplatelets are made of a bismuth layer-structured ferroelectric[26] and thus can be directly fabricated by molten salt synthesis. Because success in fabricating anisotropic perovskites with controlled characteristics is an important milestone in achieving their maximum functional performance, processes on both the synthesis of the BINT precursors and their topochemical conversion into 0.94BNT-0.06BT target deserve particular attention. Therefore, the main objective of this work is to fundamentally understand the formation mechanism of 0.94BNT-0.06BT microplatelets by systematically studying their compositional/structural evolution and morphological development. Furthermore, the functionality of individual 0.94BNT-0.06BT microplatelets is examined by studying their domain morphology and local functional response. The synthesis of high-quality BNT-BT anisotropic particles has the potential to enhance their performance in advanced applications and the investigation of their formation mechanism provides guidelines for the further design and synthesis of novel anisotropic perovskite crystals.
MATERIALS AND METHODS
Aurivillius BINT plate-like precursors were synthesized by molten salt synthesis. Reagent-grade Bi2O3, Na2CO3 and TiO2 powders of 99.9% purity were mixed following the stoichiometry of Equation (1) by ball-milling in ethanol for 12 h, with NaCl then added in a weight ratio of 15:1 by ball-milling for another 12 h. To study the formation process of BINT, the dried powder mixtures were heated at temperatures between 650 and 1050 °C for 1.5 h, followed by removal of the salt by washing with hot deionized water more than 15 times until no Cl- was detected by the AgNO3 solution. Next, using the BINT particles fabricated at 1050 °C as precursors, topochemical conversion of the 0.94BNT-0.06BT microplatelets was carried out according to Equation (2). Stoichiometric amounts of BINT, Na2CO3, BaCO3 and TiO2 were mixed with NaCl in a weight ratio of 1:2. To preserve the platelet morphology of the BINT precursors, Na2CO3, BaCO3, TiO2 and NaCl were firstly mixed by ball-milling for 12 h and then the precursors were added by magnetic stirring for 12 h. The dried powder mixtures were heated at temperatures between 900 and 1150 °C for 6 h to study the TMC process for 0.94BNT-0.06BT. Repeated washing using hot deionized water and filtration was performed to remove the NaCl, after which the synthesized microplatelets were dispersed by ultrasonication and rinsed five times with deionized water.
9Bi2O3 + Na2CO3 + 16TiO2 → 4Bi4.5Na0.5Ti4O15 + CO2↑ (1)
0.54BaCO3 + 1.88Na2CO3 + 5.24TiO2 + 0.94Bi4.5Na0.5Ti4O15 → 9(0.94 Na0.5Bi0.5TiO3 - 0.06BaTiO3) + 2.42CO2↑ (2)
X-ray diffraction (XRD, D/max 2400, Rigaku, Tokyo, Japan) was utilized to determine the phase structure and crystal orientation of the samples. Raman scattering spectra were collected through a LabRAM XploRA spectrometer (HORIBA Jobin Yvon S.A.S., France). Field-emission scanning electron microscopy (Helios Nanolab 600i, FEI, Hillsboro, OR, USA) combined with energy-dispersive X-ray spectroscopy (EDS) was utilized to observe the morphological and compositional features. Piezoelectric force microscopy (PFM, MFP-3D, Asylum Research, CA, USA) was utilized to obtain the height/amplitude/phase images and local electrical responses (phase-voltage hysteresis loop and amplitude-voltage butterfly curve) of the synthesized BNT-BT microplatelets.
RESULTS AND DISCUSSION
Figure 1 shows the phase formation sequence of the BINT precursors during the molten salt synthesis as a function of heat treatment. Below the melting point of NaCl (Tm ~801 °C), solid-state reactions had already taken place between the Bi2O3, Na2CO3 and TiO2 raw materials at 650 °C, so the peaks for the Bi4Ti3O12 (JCPDS card #80-2143) and BNT (JCPDS card #89-3109) intermediate phases can be clearly detected from the XRD pattern. With increasing temperature to 805 °C, the NaCl melted and the Aurivillius-structured BINT (JCPDS card #74-1316) phase began to form. Furthermore, the Bi8.5Na0.5Ti7O27 (JCPDS card #32-1044) intermediate phase, consisting of structural units of the two parent phases, Bi4Ti3O12 (m = 3) and BINT (m = 4), through sharing of the “Bi2O2” sheets along the c-axis[27], was detected. The Bi8.5Na0.5Ti7O27 peak intensities increase with increasing temperature to 825 °C and then decrease at 850 °C. Finally, all intermediate phases were consumed and completely converted into the pure BINT phase at 950 °C, which remained stable with a further increase temperature to 1050 °C.
Figure 2 demonstrates the morphology formation for the BINT precursors during the molten salt synthesis. Relatively agglomerate platelet particles with side lengths of ~0.4-2.4 μm can be detected in the molten salt at 805 °C, which can be attributed to the mixture of Bi4Ti3O12, Bi8.5Na0.5Ti7O27 and BINT phases according to Figure 1. With a further increase to 900 °C, the platelets become larger, well dispersed and more homogeneous, with side lengths of ~1.1-3.7 μm detected. Discrete and well-oriented platelet particles with side lengths of ~2.6-7.4 μm and thicknesses of ~0.2-0.3 μm were obtained at 950 °C, which were confirmed to be of the pure Aurivillius BINT phase in Figure 1. These anisotropic BINT platelets possess a structure expressed by the general formula Bi2O2(Am-1BmO3m+1), where A represents the mixture of Na+ and Bi3+ (perovskite A-site with twelvefold coordination), B represents Ti4+ (perovskite B-site with sixfold coordination) and m = 4, which means that the structure of BINT consists of four perovskite-like [(Bi2.5Na0.5)Ti4O13]2+ layers between the two (Bi2O2)2+ fluorite layers. With further increasing temperature to 1050 °C, the reaction proceeds into the particle growth stage [Figure 1], so BINT platelets with larger particle sizes (6.0-12.0 μm in side length and 0.3-0.8 μm in thickness), regular shapes and high aspect ratios were obtained, which were used as suitable precursors for the topochemical conversion of the 0.94BNT-0.06BT perovskite target.
Figure 3A displays the XRD patterns of the product(s) obtained by heating the BINT, Na2CO3, BaCO3 and TiO2 mixture at different temperatures. Pure perovskite phase was detected at 900 °C, suggesting the completed structural conversion from the Aurivillius precursor to the perovskite target. Three possible processes could finish at this temperature. The first is that (Bi2O2)2+ interlayers reacted with ambient O2- and then decomposed from the BINT crystal lattice in the form of Bi2O3, owing to the weak linkage between (Bi2O2)2+ layers and [(Bi2.5Na0.5)Ti4O13]2+ blocks. The second one is that Na+ and Ba2+ ions partially replaced the Bi3+ ions in the pseudo-perovskite blocks. The third one is that the byproduct Bi2O3 may have reacted with some Na2CO3, BaCO3 and TiO2 to form isotropic perovskite BNT-BT particles. Further investigation is needed to clarify these complicated structural evolution processes. The formed perovskite phase maintains its stability with further increasing temperature to 1050 and 1150 °C. For the product obtained at 1150 °C, close inspection of the XRD data [Figure 3B] reveals two sets of peaks, which were fitted to a mixture of rhombohedral and tetragonal phases in space groups R3c and P4bm, respectively, using the Gaussian function. Such splitting of peaks in the XRD patten suggests that the composition of the final product is in the range of the MPB, which is consistent with that reported previously for 0.94BNT-0.06BT ceramics[28,29].
Figure 3C-G present the morphological evolution of the structurally converted perovskite particles during the heating process. Many polycrystalline aggregates, which roughly preserved the shape/size of the BINT precursors and consisted of aligned perovskite grains, could be observed at 900 °C. Such a phenomenon indicates that extensive exfoliation and disintegration events occurred during the structural transformation process from Aurivillius to perovskite phase, possibly owing to the loss of epitaxy and expulsion of the Bi2O3 byproduct[13]. Furthermore, some fine isotropic perovskite particles can also be detected. These polycrystalline aggregates recrystallize with increasing heating temperature, accompanied by the disappearance of both several polycrystalline boundaries and pores inside the aggregates. The converted platelets still possessed the polycrystalline feature with rugged surfaces and irregular shapes at 1100 °C, but they became much denser. With further increasing temperature to 1150 °C, single-crystal BNT-BT microplatelets with flat surfaces, regular shapes and increased thicknesses were obtained. These platelets were ~6.1-10.5 μm in length and ~0.9-2.0 μm in thickness. The microstructure/crystallinity damage caused by the structural conversion was completely healed at this stage, which occurred through secondary recrystallization or Ostwald ripening[30]. To be specific, the polycrystalline aggregates sintered and recrystallized into the single crystals. In addition, Ostwald ripening, i.e., the growth of larger particles by consuming smaller ones, promoted further growth and reshaping of these irregular microcrystals, producing final platelets with regular shapes and increased thickness.
Figure 4A shows the XRD patterns of the aligned BINT precursors and 0.94BNT-0.06BT platelets. The BINT precursors are of the pure Aurivillius phase (JCPDS card #74-1316). The strong intensities of the (006), (008), (0010), (0016) and (0020) peaks indicate that the surfaces of the precursor platelets were parallel to (00l), suggesting that these particles are highly-oriented and thus good precursor candidates. The converted 0.94BNT-0.06BT microcrystals are identified to be of the pure perovskite phase (JCPDS card #89-3109). Only the (100)pc and (200)pc peaks can be detected from the XRD pattern, suggesting the structural compatibility between BINT (00l) and 0.94BNT-0.06BT (h00). Figure 4A indicates that highly (h00)-oriented 0.94BNT-0.06BT platelet particles can be obtained through TMC from layer-structured BINT precursors with strong (00l) orientation. EDS was conducted to determine the elemental distribution and composition of the converted microplatelets. The elemental distribution maps [Figure 4B] show the homogenous distributions of Bi, Na, Ba, Ti and O for the platelets. According to the EDS analysis, the actual chemical composition of the converted platelets approximately agrees with the nominal 0.94BNT-0.06BT composition of crystal growth. A schematic illustration of the topological mechanism responsible for the conversion from BINT precursors to 0.94BNT-0.06BT microplatelets is shown in Figure 5.
Raman scattering is sensitive to short-range order and phase structure, so the 0.94BNT-0.06BT microplatelets possess different Raman responses compared to the BINT precursors, as shown in Figure 4C. For bismuth layer-structured ferroelectrics, the Raman active modes are believed to originate from both A-site cations within the pseudo-perovskite blocks (below 200 cm-1) and the internal vibration of the TiO6 octahedra in the high-frequency region. Therefore, for the Aurivillius BINT, the most prominent band at ~270 cm-1 could be ascribed to F2g symmetry, corresponding to the interbond angle bending vibration[31]. The bands at ~544 and ~856 cm-1 are assigned to the Eg and A1g characters, respectively, reflecting two pure bond stretching vibrations of the TiO6 octahedra[31]. For the 0.94BNT-0.06BT platelets, six vibration modes can be observed at 135, 270, 529, 561, 765 and 842 cm-1, which are in good agreement with those reported in ceramics with similar compositions[32]. These Raman bands are relatively broad and were possibly caused by the presence of the disorder structural or distorted octahedral clusters at short range and the overlapping of Raman modes because of the lattice anharmonicity[33]. The Raman band near 135 cm-1 is correlated to the vibration of A-site ions (Na+, Ba2+ and Bi3+) in the ABO3 material, which is more sensitive to the phase transition induced by the symmetry change in the A-site. The Raman peak at 270 cm-1 could be ascribed to stretching arising from the bonds owing to the presence of octahedral [TiO6] clusters at short range[33]. The Raman bands near 529 and 561 cm-1 could be ascribed to the (←O←Ti→O→) stretching symmetric vibrations of the octahedral [TiO6] clusters[34]. The Raman peak at ~842 cm-1 could be associated with the presence of oxygen vacancies[29].
Figure 6A-C demonstrate the PFM height, amplitude and phase images of the 0.94BNT-0.06BT platelets converted from the BINT precursors, respectively. Relatively smooth large surfaces were also observed from the topographic PFM measurement. The phase image shows fine domains with both labyrinth and stripe shapes, suggesting the coexistence of rhombohedral and tetragonal phases in these samples. This result agrees well with the XRD patterns in Figure 3A and B. A good-symmetry phase-voltage hysteresis loop and well-shaped amplitude-voltage butterfly curve were tested at local positions of the platelets parallel to the <001>c thickness direction using local switching spectroscopy PFM (SS-PFM) by applying a sequence of DC voltage from -10 to 10 V with a superimposed AC signal of 1 V to the PFM tip. A nearly 180° phase reversal was observed from phase-voltage hysteresis loop [Figure 6D], possibly indicating sufficient polarization of the sample. Furthermore, the amplitude-voltage butterfly curve [Figure 6E] demonstrates that an amplitude of surface displacement of ~600 pm was achieved at 10 V, possibly suggesting the existence of remarkable local piezoelectricity for these microplatelets. SS-PFM can detect the electromechanical strain resulting from the converse piezoelectric effect, but such electromechanical strain on the surface can stem from several non-piezoelectric origins (i.e., artifacts), such as electrostatic interaction, charge injection, Joule heating and ion migration[35-37].
During the testing of the phase-voltage hysteresis loop and amplitude-voltage butterfly curve shown in Figure 6D and E, respectively, we utilized a stiff cantilever and observed the off-field responses to minimize the electrostatic interaction effect[35,37,38]. However, the possibility of contributions from artifact effects still cannot be excluded. Therefore, further studies are needed to confirm the presence of ferroelectricity and piezoelectricity in the 0.94BNT-0.06BT platelets, including writing up and down domains to prove that the measurement of the PFM signal was due to inversion of the polarization vector and not to the accumulation of charges. The orientation-controlled platelet morphology, pure perovskite structure, good crystallinity and remarkable local electromechanical strain make 0.94BNT-0.06BT platelets attractive to be used as building blocks for the fabrication of energy harvesting/storage devices, as templates to texture electronic ceramics and as substrates to grow crystals or deposit films.
CONCLUSIONS
Single-crystal 0.94(Bi0.5Na0.5)TiO3-0.06BaTiO3 microplatelets with preferential (h00) orientation, pure perovskite structure, controlled morphology and remarkable functional response were fabricated by topochemical conversion from Aurivillius Bi4.5Na0.5Ti4O15 precursors. The pure Bi4.5Na0.5Ti4O15 phase formed at 950 °C from intermediate phases, such as Bi4Ti3O12 and Bi8.5Na0.5Ti7O27, during the molten salt synthesis. The Aurivillius to perovskite structure transformation was completed at 900 °C, but the induced extensive exfoliation events led to the replacement of single-crystal platelets by polycrystalline aggregates. The subsequent recrystallization process, which occurred at higher temperatures via Ostwald ripening, healed such microstructural damage. The 0.94(Bi0.5Na0.5)TiO3-0.06BaTiO3 microplatelets with both single-crystal features and homogenous distributions of Bi, Na, Ba, Ti and O were achieved at 1150 °C, in which both labyrinth and stripe domains were detected, suggesting the coexistence of rhombohedral and tetragonal phases, in agreement with the XRD result. Moreover, these platelets possessed remarkable local electromechanical strain response along the <001>c direction, with an amplitude of ~600 pm detected at 10 V through SS-PFM. This work provides guidelines for the further design and synthesis of novel low-dimension functional microcrystals in the future and can also remarkably expand the application fields of (Bi0.5Na0.5)TiO3-BaTiO3 materials.
DECLARATIONS
Authors’ contributionsConception, design, writing and editing: Ma YQ, Chang YF, Li F
Materials synthesis and structural characterizations: Ma YQ, Xie H, Liu, LJ, Kou QW, Sun Y
Data analysis and interpretation: Yang B, Cao WW, Chang YF, Li F
All authors contribute to the manuscript and were involved in discussion
Availability of data and materialsNot applicable.
Financial support and sponsorshipThis work was supported by the National Natural Science Foundation of China (Nos. 52072092, 51922083 and 11572103), the Natural Science Foundation of Heilongjiang Province (No. YQ2019E026) and the Fundamental Research Funds for the Central Universities (No. HIT.OCEF.2021018).
Conflicts of interestAll authors declared that there are no conflicts of interest.
Ethical approval and consent to participateNot applicable.
Consent for publicationNot applicable.
Copyright© The Author(s) 2022.
Comments
Comments must be written in English. Spam, offensive content, impersonation, and private information will not be permitted. If any comment is reported and identified as inappropriate content by OAE staff, the comment will be removed without notice. If you have any queries or need any help, please contact us at support@oaepublish.com.