High-Pressure Synthesis, Crystal Structure, Chemical Bonding, and

Article Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX

pubs.acs.org/IC

High-Pressure Synthesis, Crystal Structure, Chemical Bonding, and Ferroelectricity of LiNbO3‑Type LiSbO3 Yoshiyuki Inaguma,*,† Akihisa Aimi,†,‡ Daisuke Mori,†,§ Tetsuhiro Katsumata,⊥ Masanari Ohtake,∥ Masanobu Nakayama,∥,#,¶,∇ and Masao Yonemura○,◆ †

Department of Chemistry, Faculty of Science, Gakushuin University, 1-5-1 Mejiro, Toshima-ku, Tokyo 171-8588, Japan Department of Chemistry, School of Science, Tokai University, 4-1-1 Kitakaname, Hiratsuka, Kanagawa 259-1292, Japan ∥ Frontier Research Institute for Materials Science, Nagoya Institute of Technology, Gokiso-cho, Showa-ku, Nagoya City, Aichi 466-8555, Japan # Center for Materials Research by Information Integration, Research and Services Division of Materials Data and Integrated System, National Institute for Materials Science, 1-2-1 Sengen, Tsukuba, Ibaraki 305-0047, Japan ¶ Global Research Center for Environment and Energy based on Nanomaterials Science, National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 305-0047, Japan ∇ Elements Strategy Initiative for Catalysts and Batteries, Kyoto University, 1-30 Goryo-Ohara, Nishikyo, Kyoto 615-8245, Japan ○ Institute of Materials Structure Science, High Energy Accelerator Research Organization, 1-1 Oho, Tsukuba, Ibaraki 305-0801, Japan ◆ Sokendai (The Graduate University for Advanced Studies), Shirakata 203-1, Tokai, Naka 319-1106, Japan

Downloaded via UNIV OF RHODE ISLAND on December 4, 2018 at 06:19:21 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

⊥

S Supporting Information *

ABSTRACT: A polar LiNbO3 (LN)-type oxide LiSbO3 was synthesized by a high-temperature heat treatment under a pressure of 7.7 GPa and found to exhibit ferroelectricity. The crystal structural refinement using the data of synchrotron powder X-ray diffraction and neutron diffraction and the electronic structure calculation of LN-type LiSbO3 suggest a covalent-bonding character between Sb and O. When comparing the distortion of BO6 in LN-type ABO3, the distortions of SbO6 in LiSbO3 and SnO6 in ZnSnO3, which included a B cation with a d10 electronic configuration, were smaller than those of BO6 in LN-type oxides having the second-order Jahn−Teller active B cation, e.g., LiNbO3 and ZnTiO3. The temperature dependence of the lattice parameters, second harmonic generation, dielectric permittivity, and differential scanning calorimetry made it clear that a second-order ferroelectric−paraelectric phase transition occurs at a Curie temperature of Tc = 605 ± 10 K in LN-type LiSbO3. Further, firstprinciples density functional theory calculation suggested that perovskite-type LiSbO3 is less stable than LN-type LiSbO3 under even high pressure, and the ambient phase of LiSbO3 directly transforms to LN-type LiSbO3 under high pressure. The phase stability of LN-type LiSbO3 and the polar and ferroelectric properties are rationalized by the covalent bonding of Sb−O and the relatively weak Coulomb repulsion between Li+ and Sb5+.

■

INTRODUCTION Noncentrosymmetric (NCS) structurei.e., a structure with a lack of center of symmetrygives rise to technologically important phenomena that have garnered much attention in materials science and materials engineering, namely, piezoelectricity, pyroelectricity, ferroelectricity, and nonlinearoptical (NLO) properties.1 In NSC materials, pyroelectricity and ferroelectricity are functionalities observed in NCS polar materials with electric dipole moments (spontaneous electric polarization). Among 21 NSC crystal classes, there are 10 NCS polar crystal classes; C1(1), C2(2), C3(3), C4(4), C6(6), Cs(m), C2v(mm2), C3v(3m), C4v(4mm), and C6v(6mm). Therefore, the materials design for these functionalities corresponds to how © XXXX American Chemical Society

we crystallize the materials into structures with these NSC polar crystal classes. For the stabilization of polar materials, we have paid much attention to oxides with second-order Jahn− Teller (SOJT) active cations, d0 transition-metal ions (V5+, Ti4+, Nb5+, Ta5+, Mo6+, W6+, etc.), and cations having 5s2 or 6s2 lone-pair electrons (Sn2+, Pb2+, Bi3+ etc.) because the chemical bond between their cations and oxygen results in cooperative asymmetric lattice distortion.1−13 Another method is to survey for polar structures and to stabilize the compounds with the polar structure. RhomboheReceived: September 28, 2018

A

DOI: 10.1021/acs.inorgchem.8b02767 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

pressure. Although NaSbO3 has a relatively large tolerance factor of 0.91, under ambient-pressure NaSbO3 does not adopt the Pv-type structure but the ilmenite (IL)-type structure with a small Sb−O−Sb angle of 100°.24 Mizoguchi et al.23 demonstrated that the bending of Sb−O−Sb is stabilized by the anion-based SOJT effect based on the orbital interaction between the Sb 5s and O 2p orbitals, and high pressure is needed to obtain Pv-type NaSbO3 with a relatively large Sb− O−Sb angle of 154°. We also tried to stabilize LN-type LiSbO3 under high pressure. As a result, we have successfully synthesized a polar LiSbO3 with an LN-type structure and found that this compound exhibits ferroelectricity and a second-order ferroelectric−paraelectric phase transition. The phase stability and polarity were also evaluated using firstprinciples calculation. The effects of chemical bonding including an anion-based SOJT effecton the structure and the resulting polar properties are discussed.

dral LiNbO3 (abbreviated as LN below)-type structure is a prominent example of polar ones and is adapted by a polar point group of C3v(3m). As seen in Figure S1, the cooperative cation shift based on Coulomb repulsion between the A and B cations along the c axis of the hexagonal cell against a closepacked oxygen layer results in spontaneous polarization along the c axis. Although one of the LN-type oxides, ZnSnO3, does not possess SOJT active cations, it does exhibit polarity.14 Therefore, stabilization of polar materials with LN-type structure has been extensively performed.15 On the other hand, ferroelectricity characterized by an electric polarization reversal by an applied electric field cannot necessarily be realized in all of the polar materials because of the high cohesive electric field, electronic conductive nature of the materials, and so on. For instance, although polarization switching has been reported in a thin film of LN-type ZnSnO3,16 such switching has not been realized in bulk ZnSnO3. Therefore, an additional materials design for the switching of electric polarization is needed. To realize the switching of polarization in LN-type oxides, the A cation in an octahedral site must go through a triangle site surrounded by three oxygen atoms into an adjacent octahedral site, in addition to the required shift of the B-site ions in BO6 octahedra. In the oxides with the oxidation states of cations of A2+B4+O3, A3+B3+O3, and A4+B2+O3, it is thought to be difficult to switch the polarization because of Coulomb interaction between the A cation and oxygen ions. On the other hand, because in A+B5+O3 the Coulomb interaction between the monovalent A ion and oxygen ion is thought to be relatively weaker, the polarization switching would be facile. In fact, in isostructural LiNbO3 and LiTaO3, polarization switching has been achieved.17 Because an LN-type compound with the general formula of ABX3 possesses three-dimensional corner-sharing BX6 octahedra, the same as perovskite (abbreviated as Pv below)-type compounds, the LN-type structure can be described as a derivative of the Pv-type structure.18,19 In most of the reports to date, when the tolerance factor in perovskites, t [=(rA + rX)/√2(rB + rX), where rA and rB are the ionic radii of the A ion with eight coordination and the B ion with six coordination and rX is the ionic radius of the X ion], is in the vicinity of unity, the ABX3 compounds adopt the Pv-type structure, and when t is less than 0.85, the LN-type phase is observed in general, although not always. On the basis of this expectation, we herein choose LiSbO3, which has a tolerance factor (t = 0.82) close to those of LiNbO3 and LiTaO3 (t = 0.80), as a candidate. It should be noted that LiSbO3like ZnSnO3possesses no SOJT active cations because the electronic configurations are 1s2 and 4d10 for Li+ and Sb5+, respectively. However, under ambient pressure, the compound does not adopt an LN-type structure with corner-shared SbO6 octahedra but rather possesses SbO6 octahedra sharing two edges with each other to form a zigzag chain.20,21 Goodenough and Kafalas22 have explained the phase stability of LiSbO3 as follows: The covalent bond energy between d10 B cations, such as Sb5+ and Bi5+, and oxygen ions is maximized by forming a 90° or 109° B−O−B angle because d10 ions do not possess d orbitals accessible for Β−Ο π bonding in the case of a 180° B−O−B angle. As a result, the ambient phase of LiSbO3 with a ca. 90° Sb−O−Sb angle shows greater stability than the LN-type structure with a 130−140° Sb−O−Sb bond. Meanwhile, Mizoguchi et al. 23 have successfully synthesized a Pv-type NaSbO3 under high

■

EXPERIMENTAL SECTION

Polycrystalline LiSbO3 was synthesized by a solid-state reaction. The ambient phase of LiSbO3 was first synthesized using the following starting materials: Li2CO3 (>99.99% in purity) and Sb2O3 (>99.9% in purity). An equimolar mixture of starting materials was heated at 900 °C for 8 h in air. The obtained LiSbO3 powder was put in a gold capsule (0.2 mm thickness, 3.1 mm inner diameter, and 3.2 mm depth). A pyrophyllite cube block (13 mm side length) was used as a pressure medium. A cylindrical graphite heater was placed in the cube block. The capsule was placed in the NaCl sleeve, and the sleeve was placed in the heater. Both ends of the heater were then stuffed with NaCl disks. The sample was heated at 900 °C for 30 min under a pressure of 3−7.7 GPa using a cubic multianvil-type high-pressure and high-temperature apparatus (NAMO2001; TRY Engineering) and then quenched to room temperature with a cooling rate of about 200 K/min, followed by the release of pressure. Pressure calibration was performed at room temperature using pressure fixed points: Bi I−II (2.55 GPa), Ba I−II (5.5 GPa), and Bi III−V (7.7 GPa).25 The phase identification of the obtained samples was done by the X-ray diffraction (XRD) method using a Rigaku RINT2100 diffractometer (monochromatized Cu Kα) and a PANalytical X’Pert3 powder diffractometer (monochromatized Cu Kα) in our laboratory. Scanning electron microscopy (SEM) images of an LN-type LiSbO3 sample were recorded with a Hitachi TM3030Plus tabletop microscope. Synchrotron powder XRD (SXRD) data of LN-type LiSbO3 at room temperature were collected using a Debye−Scherrertype powder diffractometer installed at the BL02B2 beamline at SPring-8.26 The step size of scanning was 0.01° in 2θ at a wavelength of λ = 0.42053 Å. The powder was packed into a glass capillary of 0.2 mm diameter. Neutron diffraction (ND) data at room temperature were collected using a SPICA time-of-flight (TOF) neutron powder diffractometer at the BL09 beamline at the Japan Proton Accelerator Research Complex. The powder was then packed into a vanadium cell of 6 mm diameter. The crystal structure was refined using the Rietveld analysis programs RIETAN-FP27 for SXRD data and Z-Rietveld28 for ND data measured using a BS bank. The coherent scattering lengths used in the refinement of the ND data are −1.90 fm for Li,29 5.57 fm for Sb,30 and 5.805 fm for O.31 In addition, a high-temperature powder XRD experiment was performed in the temperature range of 300−973 K using a Bruker AXS D8 ADVANCE diffractometer with a high-temperature stage. The phase transition was observed by differential scanning calorimetry (DSC) using a Shimadzu DSC-60 calorimeter in air in the temperature range of 300−770 K. The dielectric permittivity of the sintered sample was measured using a precision LCR meter (4284A; Agilent, Palo Alto, CA) at frequencies of 10K, 50K, 100K, 200K, 500K, and 1M Hz in the temperature range of 300−750 K. Gold electrodes were formed on both sides of the sample pellet by a direct-current sputtering method, followed by annealing at 400 °C for 30 min. The ferroelectric properties were evaluated by a P−E hysteresis loop measurement at 20 Hz using an B

DOI: 10.1021/acs.inorgchem.8b02767 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry aixACCT TF analyzer 2000 ferroelectric tester at room temperature with the same sample pellet as that used for the dielectric permittivity measurement. Second-harmonic-generation (SHG) measurements were performed on a modified Kurtz NLO system using 1064 nm light in order to confirm the noncentrosymmetry and evaluate the temperature evolution of polarity in the temperature range of 300− 650 K. The ungraded polycrystalline LN-type LiSbO3 was used for the SHG measurements. The sample was then put on a homemade sample holder with a Cryo-con 50 Ω cartridge heater. The temperature was monitored using a K-type thermocouple attached to the samples. SHG radiation was collected in reflection mode using a Continuum Minilite II YAG:Nd laser (λ = 1064 nm) operating at 10 Hz. The radiation from the sample, after passing an IR cutoff filter to remove the incident laser light, was guided into a monochromator (MC-10N; Ritsu, Japan) using an optical fiber to extract only the radiation with a wavelength of 532 nm and detected by an attached photomultiplier tube (PMT; R6427; Hamamatsu Photonics, Japan). A digital oscilloscope (RTM1052; Rohde & Schwarz, Germany) connected with the PMT via a preamplifier monitors and collects the SHG data. The intensity of incident laser light was then checked using a Si PIN photodiode (S6775; Hamamatsu Photonics, Japan). All first-principles density functional theory (DFT) calculations for the phase stability of LiSbO3 compounds under ambient pressure (abbreviated as AP below)-type (space group: Pncn), LN-type, ILtype, and Pv-type structures were performed with the Vienna Ab Initio Simulation Package (VASP)32,33 using the modified Perdew− Burke−Ernzerhof generalized gradient approximation34,35 and the projector-augmented-wave method.36 A spin-polarization calculation was used. A kinetic cutoff energy of 500 eV and appropriate k-point meshes were set to satisfy the convergence test (