Bi3n+1Ti7Fe3n–3O9n+11 Homologous Series: Slicing Perovskite

Jan 8, 2016 - ... separated by (001)p interfaces containing quasi-1D anatase fragments and stabilized owing to the stereochemical activity of the Bi3+...
22 downloads 4 Views 4MB Size
Article pubs.acs.org/IC

Bi3n+1Ti7Fe3n−3O9n+11 Homologous Series: Slicing Perovskite Structure with Planar Interfaces Containing Anatase-like Chains Dmitry Batuk,† Alexander A. Tsirlin,‡,§ Dmitry S. Filimonov,∥ Konstantin V. Zakharov,⊥ Olga S. Volkova,⊥,#,∇ Alexander Vasiliev,⊥,#,∇ Joke Hadermann,† and Artem M. Abakumov*,†,∥,○ †

Electron Microscopy for Material Science (EMAT), University of Antwerp, Groenenborgerlaan 171, B-2020 Antwerp, Belgium Experimental Physics VI, Center for Electronic Correlations and Magnetism, University of Augsburg, 86159 Augsburg, Germany § National Institute of Chemical Physics and Biophysics, 12618 Tallinn, Estonia ∥ Chemistry Department, Moscow State University, 119991 Moscow, Russia ⊥ Physics Department, Moscow State University, 119991 Moscow, Russia # Theoretical Physics and Applied Mathematics Department, Ural Federal University, 620002 Ekaterinburg, Russia ∇ National University of Science and Technology ″MISiS″, 119049 Moscow, Russia ○ Center for Electrochemical Energy Storage, Skolkovo Institute of Science and Technology, Nobelya str. 3, 143026 Moscow, Russia ‡

S Supporting Information *

ABSTRACT: The n = 3−6 members of a new perovskite-based homologous series Bi3n+1Ti7Fe3n−3O9n+11 are reported. The crystal structure of the n = 3 Bi10Ti7Fe6O38 member is refined using a combination of X-ray and neutron powder diffraction data (a = 11.8511(2) Å, b = 3.85076(4) Å, c = 33.0722(6) Å, S.G. Immm), unveiling the partially ordered distribution of Ti4+ and Fe3+ cations and indicating the presence of static random displacements of the Bi and O atoms. All Bi3n+1Ti7Fe3n−3O9n+11 structures are composed of perovskite blocks separated by translational interfaces parallel to the (001)p perovskite planes. The thickness of the perovskite blocks increases with n, while the atomic arrangement at the interfaces remains the same. The interfaces comprise chains of double edge-sharing (Fe,Ti)O6 octahedra connected to the octahedra of the perovskite blocks by sharing edges and corners. This configuration shifts the adjacent perovskite blocks relative to each other over a vector 1/2[110]p and creates S-shaped tunnels along the [010] direction. The tunnels accommodate double columns of the Bi3+ cations, which stabilize the interfaces owing to the stereochemical activity of their lone electron pairs. The Bi3n+1Ti7Fe3n−3O9n+11 structures can be formally considered either as intergrowths of perovskite modules and polysynthetically twinned modules of the Bi2Ti4O11 structure or as intergrowths of the 2D perovskite and 1D anatase fragments. Transmission electron microscopy (TEM) on Bi10Ti7Fe6O38 reveals that static atomic displacements of Bi and O inside the perovskite blocks are not completely random; they are cooperative, yet only short-range ordered. According to TEM, the interfaces can be laterally shifted with respect to each other over ±1/3a, introducing an additional degree of disorder. Bi10Ti7Fe6O38 is paramagnetic in the 1.5−1000 K temperature range due to dilution of the magnetic Fe3+ cations with nonmagnetic Ti4+. The n = 3, 4 compounds demonstrate a high dielectric constant of 70−165 at room temperature.



INTRODUCTION Exploring the parameter space of solids opens up intriguing opportunities for searching new materials with useful functionalities. The most fundamental parameter of any crystalline material is its crystal structure. Therefore, the discovery of hitherto unknown structure types extends our abilities to combine various elements with different oxidation states and chemical bonding into one structure. Among oxide materials, the perovskite ABO3 structure is most adaptable to the chemical nature of constituting elements. It can host a large variety of cations with different formal charges, ionic sizes, and covalency/ionicity of metal−oxygen bonding. This flexibility is achieved by deformations of the 3D framework of corner© XXXX American Chemical Society

sharing BO6 octahedra and/or by introducing anion or cation vacancies. Besides, the diversity of the perovskite-based structures is greatly expanded by the formation of intergrowth structures and homologous series, where 2D perovskite slabs (modules) are interlaced with modules of other simple inorganic structure types, such as rock salt, fluorite, α-PbO, etc. High-temperature superconductivity and tunneling magnetoresistance in the Ruddlesden−Popper phases (perovskite− rock-salt intergrowths) and ferroelectricity in the Aurivillius phases (perovskite−α-PbO intergrowths) are, perhaps, the Received: October 24, 2015

A

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Imaging of Paul Scherrer Institut (LNS PSI, Villigen, Switzerland). A powdered sample of about 10 g was placed in a vanadium container of 8 mm in diameter. The NPD pattern was collected at the wavelength 1.8857 Å in the 2θ range of 4−165° at 300 K. The crystal structure analysis was performed with the JANA2006 program.8 Images presenting crystallographic data were created using the VESTA program.9 Cation composition of the samples was confirmed by energydispersive X-ray (EDX) analysis conducted on a JEOL 5510 scanning electron microscope equipped with an INCAx-sight 6587 system (Oxford Instruments). About 60 spectra were collected for each sample, and the Bi-M, Ti-K, and Fe-K emission lines were used for quantification. Specimens for transmission electron microscopy (TEM) investigation were prepared by grinding the samples in an agate mortar with ethanol and depositing a few drops of the suspension on holey carbon TEM grids. Electron diffraction (ED) patterns were acquired using an FEI Tecnai G2 transmission electron microscope operated at 200 kV. High angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images and annular bright-field (ABF) STEM images were recorded on a probe aberration corrected FEI Titan3 80−300 microscope operated at 300 kV. The distribution of Fe and Ti in the structures was analyzed using atomic resolution electron energy loss spectroscopy (EELS) data recorded in the STEM mode. The STEM-EELS measurements were performed on an FEI Titan3 60−300 microscope equipped with a Gatan Enfinium ER spectrometer and operated at 120 kV. Individual elemental distribution maps were generated by placing an integration window over the backgroundsubtracted absorption edges of the elements (Ti-L2,3, Fe-L2,3). All STEM experiments were conducted with a probe convergence semiangle of about 21 mrad and a probe current of about 50 pA. Theoretical high-resolution HAADF-STEM and ABF-STEM images were calculated using the QSTEM program.10 To analyze the oxidation state, distribution, and local coordination environment of the Fe cations in the Bi10Ti7Fe6O38 structure, Mössbauer spectroscopy was conducted on a sample enriched for 10% of 57Fe in the B sublattice (Fe + Ti). The measurements were carried out in a transmission mode using a constant acceleration Mössbauer spectrometer with a 57Co/Rh γ-ray source. Velocities were calibrated with a standard α-Fe absorber; isomer shifts were related to α-Fe. The resulting spectra were processed using UnivemMS and custom software. Magnetic susceptibility was measured using the vibrating sample magnetometer (VSM) insert of Quantum Design PPMS. Measurements above 400 K were performed using the oven setup in high vacuum (10−5 Torr). The dielectric permittivity of the ceramics was studied as a function of temperature at frequencies of 1−20 kHz on cylindrical samples of 8 mm in diameter and 1.3 mm in height on a capacitance bridge Andeen Hagerling 2700A with field strength on the sample ∼15 V. Electron localization function (ELF) distribution was obtained from the electronic structure of Bi10Ti7Fe6O38 calculated using the TBLMTO-ASA code with the local-density approximation for the exchange-correlation potential.11,12 The ordered arrangement of Fe and Ti was assumed. The 8 × 8 × 8 k mesh was used for integration over the first Brillouin zone.

most famous examples of functional materials with the perovskite-based layered structures.1 Formation of a new structure due to recombination of modules with different atomic arrangements can often be considered as association of defects (typically planar) in one of the prototype structures. These defects form a long-range ordered periodic arrangement introducing configurational and compositional changes.2 For example, the Ruddlesden−Popper structures, An+1BnO3n+1, can be derived from the ideal perovskite structure ABO3 by periodic removal of the (BO2) layers from a regular −[AO−BO2]− perovskite stacking of atomic layers. This operation reduces the B-cation content and creates double AO−OA layers, which form modules with the rock-salt type structure. Potentially, finding a new kind of planar defect in a simple structure can provide clues about new and unexpected building principles that can be used to create novel families of modular materials. For instance, a discovery of superconducting cuprates Hg2Ba2Y1−xCaxCu2O8‑δ with double HgO layers was motivated by the transmission electron microscopy (TEM) observations of such layers as planar defects in HgBa2Can−1CunO2n+2+δ compounds containing single HgO layers.3 Atomic resolution TEM observations on the structure of crystallographic shear planes in “Pb2Fe2O5” triggered a discovery of the anion-deficient perovskite-based homologous series AnBnO3n−2.4,5 These and other similar studies demonstrate that TEM is an inexhaustible source of inspiration for material science, as it is capable of providing structural information at the atomic scale. In a series of recent publications on the local structure of the Ti and Nd codoped bismuth ferrite, (Bi0.85Nd0.15)(Ti0.1Fe0.9)O3, unusual planar defects were discovered in the perovskite matrix, and their atomic structure was derived from advanced highresolution TEM (Figure S1 of the Supporting Information).6,7 These defects can be classified as nonconservative antiphase boundaries, where the BO6 octahedra share edges and together with the adjacent octahedra from the perovskite matrix create S-shaped tunnels hosting double A-cation columns. In this contribution, we demonstrate that in the Bi−Ti−Fe−O system this type of interfaces can be used to slice the perovskite structure into uniform blocks of variable thickness, giving rise to a new family of layered perovskites with a general composition Bi 3n+1 Ti 7 Fe 3n−3 O 9n+11 or A 3n+1 B 3n+4 O 9n+11 (AB4O11 + 3nABO3) (A = Bi, B = Fe, Ti). We present a synthesis and detailed crystal structure investigation of the n = 3 member of the series, demonstrate the existence of the higher members with n = 4, 5, 6, and propose tentative structures for the lower members with n = 1, 2.



EXPERIMENTAL SECTION

The samples were synthesized using a conventional solid-state reaction method from Bi2O3 (Aldrich, 99.9%), TiO2 (anatase, Aldrich, ≥99.9%; rutile Aldrich, 99.9%), and Fe2O3 (Sigma-Aldrich, ≥99.98%). Sample compositions were chosen according to a general formula Bi4Ti7O20 + 3(n − 1)BiFeO3 with n = 1−6. The starting materials were thoroughly ground and compressed into pellets. The solid-state reaction was conducted in air in several annealing steps. All the samples were first annealed at 800 °C for 10 h, then 3−5 times at 1000 °C for 20 h. After each annealing, the samples were reground and pressed into pellets again. The X-ray powder diffraction (XPD) data were acquired on a Huber G670 Guinier diffractometer (Cu Kα1 radiation, curved Ge(111) monochromator, image plate). The neutron powder diffraction (NPD) data for Bi10Ti7Fe6O38 were collected on the high-resolution powder diffractometer HRPT at the Laboratory for Neutron Scattering and



RESULTS Crystallographic and Compositional Characterization. The XPD patterns (Figure S2 of the Supporting Information) together with the TEM data confirm the formation of new perovskite-based structures in the Bi3n+1Ti7Fe3n−3O9n+11 samples for n ≥ 3. The interpretation of the XPD patterns was assisted by electron diffraction. The ED patterns of the three main zones for the n = 3 Bi10Ti7Fe6O38 structure are shown in Figure 1. They can be indexed on a body-centered orthorhombic lattice with the cell parameters related to the parameter of the perovskite subcell ap as a ≈ 3ap, b ≈ ap, c ≈ 8ap. The only B

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

rows reveals that these reflections are accompanied by streaks along c*. In some crystals, these streaks are well pronounced (see Figure S3 of the Supporting Information). Apparently, the streaking is a signature of the stacking faults confined to the (001) layers. They induce lateral displacements over a vector R = [±1/3,u,0], which can be derived from a condition g·R = integer for the unstreaked reflections g = 3h,0,l. The quality of the Bi3n+1Ti7Fe3n−3O9n+11 XPD patterns rapidly degrades with n (Figure S2 of the Supporting Information). Reflections gain strong anisotropic broadening and displace from their proper positions, indicating a formation of numerous defects, which hamper data interpretation and structure analysis. Nevertheless, the XPD pattern of the n = 4 Bi13Ti7Fe9O47 material could still be indexed on an A-centered monoclinic lattice with the unit cell parameters a = 11.8381(3) Å, b = 3.85923(7) Å, c = 41.3915(9) Å, β = 95.346(2)° (a ≈ 3ap, b ≈ ap, c ≈ 10ap), as shown in Figure S4 of the Supporting Information. Unambiguous indexation and refinement of the cell parameters for the materials with higher n was not possible. Cation composition of the Bi3n+1Ti7Fe3n−3O9n+11 samples examined using EDX analysis was found to be in a good agreement with the nominal composition (Table S1 of the Supporting Information). 57Fe Mössbauer spectroscopy on Bi10Ti7Fe6O38 reveals the isomer shift characteristic of the Fe3+ cations in an octahedral coordination environment (see details further in the text), which implies that oxygen vacancies are not present in noticeable concentration in this compound. Crystal Structure Refinement of Bi10Ti7Fe6O38. We refined the crystal structure of the n = 3 Bi10Ti7Fe6O38 member of the Bi3n+1Ti7Fe3n−3O9n+11 series against combined powder XPD and NPD data. The initial structure model was constructed within the unit cell determined from XPD and ED data (a ≈ 3ap, b ≈ ap, c ≈ 8ap) and the most symmetric space group Immm. Atomic positions were assigned using prior knowledge on the atomic arrangement of the perovskite blocks and the separating interfaces, which is also consistent with our TEM observations. The refinement of this model (model 1, idealized model) readily provided satisfactory agreement between the experimental and calculated XPD and NPD profiles. Drastically different neutron scattering lengths for the Fe and Ti atoms (9.45 fm for Fe, −3.44 fm for Ti) allowed the refinement of the occupancy factors for the B positions. In order to avoid correlations between the occupancy factors and atomic displacement parameters (ADPs), isotropic ADPs for all six B positions were restricted to be the same, and the Fe/Ti occupancy factors were refined, keeping the overall Fe and Ti content in accordance with the Bi10Ti7Fe6O38 formula. In the final stage of the refinement, we used common ADPs for the oxygen atoms and individual ADPs for the Bi atoms. Crystallographic data and parameters of the refinement are listed in Table 1. The powder profiles after the refinement are shown in Figure 2. Coordinates of the atomic positions, occupancy factors, and ADPs are given in Table 2. The main interatomic distances are provided in Table 3. In the course of the refinement, we noticed that the Bi atoms and some oxygen atoms acquire abnormally high ADPs of about 0.05−0.1 Å2. The refinement of ADPs for these positions in the anisotropic approximation revealed strong anisotropy of their thermal ellipsoids occurring along different crystallographic directions for different positions. One could expect that this behavior reflects static atomic displacements, which cannot be properly taken into account within the Immm space group. Therefore, we tested a number of orthorhombic (I2mm, Im2m,

Figure 1. ED patterns of Bi10Ti7Fe6O38 taken along the main crystallographic zone axes. Arrowheads indicate the h/2,0,l, h − odd diffuse intensity lines in the [010] ED pattern.

reflection condition observed in the ED patterns is hkl: h + k + l = 2n due to the I-centering, implying the most symmetric space group Immm or its subgroups. Indexing the XPD profile is consistent with ED and provides unit cell parameters a = 11.8500(2) Å, b = 3.85046(7) Å, c = 33.0767(7) Å. On the [010] ED patterns, one can notice that, alongside regular reflection rows, there are lines of modulated diffuse intensity at the h/2,0,l, h − odd positions. Our further TEM investigation demonstrates that this diffuse intensity is associated with shortrange ordered displacements of the Bi atoms. Also, careful inspection of the reflections in the h0l, h ≠ 3n reciprocal lattice C

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry Table 1. Crystallographic Data and Parameters of the Rietveld Refinement for Bi10Ti7Fe6O38 (Model 1) formula space group a, Å b, Å c, Å V, Å3 Z radiation 2θ range, step, deg. RF Rp, Rwp radiation 2θ range, step, deg. RF Rp, Rwp

Table 2. Atomic Parameters for Bi10Ti7Fe6O38 (Model 1) position

Bi10Ti7Fe6O38 Immm 11.8511(2) 3.85076(4) 33.0722(6) 1509.28(4) 2 neutrons, λ = 1.8857 Å 7−147, 0.05 0.063 0.032, 0.042 X-rays, λ = 1.54056 Å 4−100, 0.02 0.068 0.027, 0.037

a

A1 A2a A3a B1b B2b B3b B4b B5b B6b O1 O2 O3 O4 O5 O6 O7 O8 O9 O10 O11 O12 O13 O14 O15

site

x/a

y/b

z/c

Uiso, Å2

8m 4i 8m 2a 4e 4i 8m 4j 4i 4e 2b 4i 8m 8m 4j 4i 8m 8m 4j 8m 4f 4j 8m 4i

0.1574(2) 1/2 0.2380(2) 0 0.329(1) 0 0.3360(6) 1/2 0 0.173(1) 1/2 0 0.3377(8) 0.1628(8) 1/2 0 0.3330(8) 0.1678(9) 1/2 0 0.3656(9) 0 0.3211(8) 1/2

1/2 1/2 1/2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1/2 1/2 1/2 1/2 1/2

0.07090(8) 0.0576(1) 0.20215(7) 0 0 0.1295(3) 0.1175(2) 0.1862(3) 0.2406(3) 0 0 0.9364(3) 0.0590(3) 0.1180(2) 0.1253(4) 0.1846(4) 0.1785(3) 0.2415(2) 0.2474(4) 0 0 0.1234(4) 0.1123(2) 0.2039(4)

0.0345(9) 0.066(2) 0.0341(8) 0.0074(6) 0.0074(6) 0.0074(6) 0.0074(6) 0.0074(6) 0.0074(6) 0.0284(5) 0.0284(5) 0.0284(5) 0.0284(5) 0.0284(5) 0.0284(5) 0.0284(5) 0.0284(5) 0.0284(5) 0.0284(5) 0.0284(5) 0.0284(5) 0.0284(5) 0.0284(5) 0.0284(5)

a

A = Bi. bOccupancy factors for the B positions: B1: 0.28(1) Fe + 0.72(1) Ti; B2: 0.348(7) Fe + 0.652(7) Ti; B3: 0.402(7) Fe + 0.598(7) Ti; B4: 0.562(7) Fe + 0.438(7) Ti; B5: 0.587(7) Fe + 0.413(7) Ti; B6: 0.398(7) Fe + 0.602(7) Ti.

Table 3. Selected Interatomic Distances in Bi10Ti7Fe6O38 (Model 1) bonds

Figure 2. Results of the combined Rietveld refinement of the Bi10Ti7Fe6O38 structure against the NPD and XPD data at room temperature. Experimental profiles are shown together with the calculated ones and the difference curves. Vertical bars indicate the positions of the Bragg peaks.

Imm2, and I222), monoclinic (I2/m, Im, and I2 in different settings), and triclinic (I1̅ and I1) subgroups of the Immm group. Neither of them completely eliminated high ADPs, resulting only in a minor improvement of the fit, which could be attributed solely to the increasing number of refinable parameters. The only remaining space group that matches the

length, Å

A1−O3 × 2 A1−O4 × 2 A1−O5 × 2 A1−O13 × 1 A1−O14 × 1

2.692(2) 2.903(7) 2.477(5) 2.547(9) 2.376(9)

A2−O2 × 2 A2−O4 × 4 A2−O6 × 2 A2−O12 × 2 A2−O14 × 2

2.708(3) 2.722(7) 2.952(10) 2.483(8) 2.788(9)

A3−O8 × 2 A3−O9 × 2 A3−O9 × 1

2.363(5) 2.468(5) 2.173(8)

B1−O1 × 2 B1−O3 × 2 B1−O11 × 2

2.05(1) 2.103(11) 1.92538(4)

B2−O1 × 1 B2−O2 × 1 B2−O4 × 2 B2−O12 × 2

1.85(2) 2.031(12) 1.956(8) 1.974(4)

bonds

length, Å

B3−O3 × 1 B3−O5 × 2 B3−O7 × 1 B3−O13 × 2

2.18(2) 1.967(10) 1.82(2) 1.936(2)

B4−O4 × 1 B4−O5 × 1 B4−O6 × 1 B4−O8 × 1 B4−O14 × 2

1.934(10) 2.052(12) 1.960(8) 2.019(10) 1.941(1)

B5−O6 × 1 B5−O8 × 2 B5−O10 × 1 B5−O15 × 2

2.02(2) 1.995(9) 2.02(2) 2.012(5)

B6−O7 × 1 B6−O9 × 2 B6−O10 × 2 B6−O15 × 1

1.85(2) 1.989(9) 1.966(3) 1.84(2)

observed reflection conditions is I212121, but it is not consistent with the crystal structure, as it is not a subgroup of Immm. Therefore, we concluded that the structure has a high degree of D

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 3. Crystal structure of Bi10Ti7Fe6O38 refined against the room-temperature NPD and XPD profiles: model 1 with idealized atomic positions and model 2 with split Bi and some O positions.

Figure 4. [010] high-resolution HAADF-STEM images of the Bi3n+1Ti7Fe3n−3O9n+11 (n = 3−6) structures.

along b and c, O11 and O12 − along a and c, and O15 − along the a axis. The splitting for all oxygen positions, except O12, is easily accomplished by moving the atom away from the mirror planes. For all the A positions and O12 atom, the splitting along the a axis was achieved by duplicating the atomic positions (denoted with the prime symbol in Table S3 of the Supporting Information). The splitting restores reasonable values for the ADPs and decreases the reliability factors from RF = 0.063 to RF = 0.043 for the NPD pattern and from RF = 0.068 to RF = 0.057 for the XPD pattern. Crystallographic data and refinement parameters for the “split” model are listed in Table

disorder related to random static displacements of atoms, while the average atomic positions still obey the Immm symmetry. This conclusion is also unequivocally confirmed by our TEM observations (see below). To account for these static atomic displacements, we introduced a “split” Immm model (model 2), where atoms are displaced from their high symmetry positions in accordance to the apparent anisotropy of their ADPs. The A1 and A2 positions were split along the a and c axes; the A3 atom was split along the a axis only. The O1, O2, O3, O11, O12, and O15 positions with highly anisotropic ADPs were split as follows: the O1 and O2 positions − along b, O3 − E

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

significantly different Z. Complementary high-resolution HAADF-STEM and ABF-STEM images Bi10Ti7Fe6O38 are shown in Figure 5. Comparing these images, one can identify

S2 of the Supporting Information. The powder profiles after the refinement are shown in Figure S5. Positional parameters, occupancy factors, and ADPs are given in Table S3, and main interatomic distances are listed in Table S4. The two refined models of Bi10Ti7Fe6O38 are shown in Figure 3. The perovskite blocks consist of three octahedral layers (n = 3). The periodic interfaces separating the blocks are confined to the (001) planes. They are based on edge-sharing BO6 octahedra that form infinite zigzag chains propagating along the b axis. The chains are connected to the octahedra of the perovskite blocks by sharing edges on one side and corners on the other. Upon going from one chain to another along the a axis, the edge- and corner-sharing connectivity alternates, giving rise to a 3-fold perovskite periodicity, 3ap, of the structure in that direction. The edge-sharing connectivity between the octahedra of the chains as well as between the chains and the perovskite blocks eliminates a fraction of the A positions and displaces neighboring perovskite blocks relative to each other over a vector 1/2[110]p. Together, the octahedra of the chains and the perovskite blocks form S-shaped tunnels that accommodate double columns of the A3 positions. They have a highly asymmetric coordination environment typical for the stereochemically active lone pair cations, such as Bi3+. The stereochemical activity of Bi3+ cations in the A1 and A2 positions is also pronounced, as evidenced by 2−4 shortened A−O distances of 2.40−2.55 Å (Table 3). In the model 1, the Bi3+ cations appear strongly underbonded. The bond valence sum (BVS) for the Bi cations inside the perovskite blocks is 2.18(2) and 2.31(1) for the A1 and A2 positions, respectively, whereas, for the A3 cations inside the Sshaped tunnels, BVS is 2.50(2). Apparently, this underbonding promotes random displacements of the Bi cations and the neighboring oxygen atoms that are seen in the “split” structure. These displacements create shorter Bi−O contacts of ∼2.1−2.3 Å (Table S4 of the Supporting Information), so that Bi cations acquire more asymmetric coordination with a low number of relatively short Bi−O bonds. However, a large number of the Bi−O separations due to simultaneous “splitting” of the Bi positions and multiple oxygen positions preclude detailed analysis of the Bi oxygen coordination in the “split” structure. Transmission Electron Microscopy. High-resolution HAADF-STEM images of the Bi3n+1Ti7Fe3n−3O9n+11 (n = 3− 6, Figure 4) structures confirm that they are built of uniform perovskite blocks separated by the interfaces similar to those found in (Bi0.85Nd0.15)(Ti0.1Fe0.9)O3. In these images, the intensity scales as I ∼ Z1.6−1.9 (Z is an average atomic number of the column), so that the Bi columns appear as bright dots, (Fe,Ti)-O columns as faint dots, and the O columns are not visible. The perovskite blocks appear as slabs with a square pattern of the Bi and (Fe,Ti)-O columns, and the interfaces can be recognized by the rows of alternating double Bi and double (Fe,Ti)-O columns. The thickness of the perovskite blocks steadily increases with n, confirming the formation of the Bi3n+1Ti7Fe3n−3O9n+11 homologous series, where n represents the number of the octahedral layers in the perovskite blocks. To access the information on the local structure of Bi10Ti7Fe6O38, we employed the ABF-STEM imaging. In this microscopy technique, light O columns can be visualized in the presence of heavy elements, such as Bi. ABF-STEM images demonstrate an absorptive type of contrast; i.e., atomic columns appear as dark dots on a bright background. The Z-dependence of the ABF-STEM signal for heavy elements scales as ∼Z1/3, which allows simultaneous visualization of atomic columns with

Figure 5. Complementary high-resolution [010] HAADF-STEM and ABF-STEM images of Bi10Ti7Fe6O38. The insets demonstrate simulated images of the size of 2a × 1c calculated using crystallographic data for model 1 for a thickness of 10 nm.

the projected atomic species. The light O columns are not seen on the HAADF-STEM image, while Bi columns appear brightest. On the ABF-STEM image, the Bi columns correspond to the darkest spots, whereas the (Fe,Ti)-O and pure O columns are represented by faint dots of nearly equal darkness. The experimental images are in a fairly good agreement with the theoretical ones (shown as insets in Figure 5) calculated using crystallographic data for model 1, thus corroborating the refined crystal structure. The HAADF-STEM images of Bi10Ti7Fe6O38 reveal that defects are formed in the material on at least three different levels: (1) Stacking faults due to insertion of extra layers into the perovskite blocks. (2) Stacking faults in the pattern of the interfaces due to their mutual displacements along the a axis. (3) Locally ordered displacements of the Bi atoms, showing, however, no long-range periodicity. Similar to many other homologous series, Bi10Ti7Fe6O38 demonstrates an occasional increase in the thickness of the perovskite blocks. The insertion of extra (BO2) and (AO) layers creates lamellas of the n = 4 structure (Figure 6a). However, in Bi10Ti7Fe6O38, these defects are sparse. Streaking of the h0l, h ≠ 3n reflections in the [010] ED patterns (Figure 1 and Figure S3 of the Supporting Information) implies a presence of stacking faults with the R = [±1/3,u,0] lateral displacements. The [010] HAADF-STEM image in Figure 6b demonstrates that these stacking faults are attributed to the interfaces between the perovskite blocks. The F

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

interfaces results in streaking of the reflections on the ED patterns and causes anisotropic broadening of the reflections on the XPD patterns, which, in turn, hinders precise structure analysis of the materials with high n. Besides, this kind of disorder can at least partially contribute to the smearing of the scattering density at the split Bi and O positions. In the refined “split” structure, all the Bi atoms demonstrate pronounced random displacements from their ideal positons (i.e., positions in model 1). However, these displacements are not completely random. For the A2 positions, they demonstrate local ordering and give rise to the diffuse intensity lines at the h/2,0,l h − odd positions in the [010] ED pattern. The displacements can be directly observed in the [010] HAADFSTEM images (Figure 7). The adjacent A2 atomic columns

Figure 6. Two types of stacking faults in the n = 3 Bi10Ti7Fe6O38. (a) Insertion of lamellas with n = 4 structures. (b) Stacking faults with lateral displacement of the interfaces relative to each other over the vector R = [±1/3,0,0]. White lines highlight positions of the Bi atomic columns in the perovskite blocks and interfaces. The arrowheads mark the perovskite blocks, where the adjacent interfaces are displaced.

internal periodicity of the interfaces along the [100]p direction is 3ap. It is defined by the alternating corner- and edge-sharing connectivity of the double octahedral chains to the octahedra of the perovskite blocks. The neighboring interfaces can be positioned relative to each other in three ways: precisely on top of each other corresponding to R = 0 or with a shift of 1ap along either [100]p or [1̅00]p directions corresponding to the displacement vectors R = [±1/3,0,0] (Figure S6 of the Supporting Information). An ordered structure with no lateral displacement of the interfaces (i.e., R = 0) adopts an orthorhombic symmetry. This arrangement is dominant in Bi10Ti7Fe6O38. Cumulative lateral displacement (i.e., either only R = [1/3,0,0] or only R = [−1/3,0,0]) will impose monoclinic symmetry, which we observed in the next member n = 4 Bi13Ti7Fe9O47. Disorder in the lateral displacements of the

Figure 7. High-magnification [010] HAADF-STEM image demonstrating short-range ordered displacements of Bi3+ cations in Bi10Ti7Fe6O38. (a) An experimental image, as acquired. (b) The same image with masked A1 and A3 atomic columns to highlight offcenter displacements of A2 cations. Arrows indicate the direction of the displacements. G

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry form pairs of bright dots aligned along the c axis and separated by ∼ap. The upper and lower columns in the pairs are shifted in the opposite directions along the a axis, which can be seen as a rotation of the pairs about the b axis (Figure 7). On going from one A2 column pair to another along the a axis, the displacement direction alternates. The repeat period of these displacements corresponds to a doubled a parameter of the orthorhombic Immm unit cell. However, the ordered sequences of these displacement along the a axis are just several unit cells in length. Therefore, instead of sharp reflections corresponding to this ordering, lines of diffuse intensity appear on the [010] ED pattern at h/2,0,l, h − odd. The HAADF-STEM data acquired as a time series (Video 1 and its description in the Supporting Information) reveals that the displacements of the A cations are cooperative. Absorbing energy from the electron beam, the A cations wiggle around their idealized positions (as in the model 1). Remarkably, along the b direction (i.e., along the viewing direction), the A-cation displacements obey the translational symmetry of the structure, because, in each individual HAADF-STEM image of the video, the corresponding atomic columns appear as bright sharp dots without any smearing. Displacements in one atomic column affect the displacements in the neighboring columns. Thus, adjacent A2 columns displace in opposite directions that is visible as a rotation of the A2 atomic column pairs about the b axis. One can also notice that the rotation of the A2 pairs induces displacements in the neighboring A1 columns. The displacements in the A3 atomic columns are very subtle and difficult to distinguish because of scan distortion of the HAADF-STEM images. Unlike the A positions, the B cation columns do not demonstrate noticeable displacements throughout the time series. We analyzed local distribution of Fe and Ti in Bi10Ti7Fe6O38 using atomic resolution electron energy loss spectroscopy (STEM-EELS). The corresponding elemental maps corroborate the structure refinement (Figure 8). The maps

and Ti at the interface is similar to that observed in Ti, Nd codoped bismuth ferrite, (Bi0.85Nd0.15)(Ti0.1Fe0.9)O3.7 Mössbauer Spectroscopy. In order to confirm the refined distribution of Fe and Ti in the Bi10Ti7Fe6O38 structure, we analyzed the local coordination environment of Fe atoms using the Mössbauer spectroscopy. Spectra acquired at 298 and 78 K are demonstrated in Figure 9. At both temperatures, the spectra

Figure 9. Mössbauer spectra for Bi10Ti7Fe6O38 at 298 and 78 K. The spectra are fitted with two paramagnetic doublets D1 (green) and D2 (blue).

can be fitted with two overlapping paramagnetic doublets D1 and D2 (Table S5 of the Supporting Information) with the isomer shift of 0.4−0.5 mm/s corresponding to the Fe3+ cations in octahedral coordination.13 The doublets demonstrate different quadrupole splitting: ∼0.4 mm/s for D1 and 0.7− 0.8 mm/s for D2. We attribute this difference to the different coordination environment of the Fe atoms. In the Bi10Ti7Fe6O38 structure, the octahedra around the B1, B2, and B3 positions are connected to the neighboring octahedra only by sharing corners, whereas octahedra around the B4, B5, and B6 positions share at least one edge with the adjacent octahedra (Figure 3). From the refined structure (Table 2), one can calculate that 30% of Fe is located in the B1, B2, and B3 positions, and 70% in the B4, B5, and B6 positions. This result is in a good agreement with the contribution of the doublets to the Mössbauer spectra, i.e., 61−68% for D1 and 32−39% for D2. The discrepancies might be attributed to fitting uncertainties, deformations of the first coordination sphere of the cations (i.e., variations in the (Fe,Ti)−O bond lengths and angles), and local variations in the second coordination sphere (i.e., whether the adjacent octahedra are occupied by Fe or Ti). Magnetic and Dielectric Measurements. The Bi10Ti7Fe6O38 sample demonstrates an overall paramagnetic behavior (Figure S7 of the Supporting Information), but its inverse magnetic susceptibilities follow the Curie−Weiss law above 400 K only. The Curie−Weiss temperature θ obtained from the χ = C/(T + θ) fit is as high as 466 K, and the effective magnetic moment derived from the Curie constant C equals to 14.7 μB/f.u. or 6.0 μB/Fe, in good agreement with the expected moment of 5.92 μB/Fe for high spin Fe3+ (Figure 10). At T < θ, antiferromagnetic exchange interactions between the Fe3+ ions result in deviations from the paramagnetic Curie−Weiss behavior. However, no sign of long-range magnetic order is observed.

Figure 8. STEM-EELS map showing the distribution of Fe and Ti in the n = 3 Bi10Ti7Fe6O38 structure.

demonstrate that Ti is concentrated in the central octahedral layer of the perovskite blocks, corresponding to the B1 and B2 atomic positions, containing 72% and 65% Ti, respectively. At the interfaces, the B4 and B5 positions form characteristic triangles of the Fe-rich atomic columns. At the center of the interface, these triangles are connected to the Ti-rich columns of the B6 positions. The observed partial ordering between Fe H

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 10. Inverse magnetic susceptibility of Bi10Ti7Fe6O38 measured in the applied fields of 2 and 5 T, and the Curie−Weiss fit above 400 K with the paramagnetic effective moment of 14.7 μB/f.u. and the AFM Curie−Weiss temperature θ = 466 K.

The dielectric permittivity ε in the Bi3n+1Ti7Fe3n−3O9n+11 family is comparable to that of the chemically and structurally related compounds, i.e., Bi2Ti4O12, TiO2.14,15 Almost no signal was observed for the imaginary part of the permittivity, implying poor electric conductivity of the samples. The real part ε′, matching total ε, rises almost monotonously with temperature, reaching values of 70 for Bi10Ti7Fe6O38 and 165 for Bi13Ti7Fe9O47 at room temperature (Figure 11). The high

Figure 12. (a) Schematic illustration of an interface between the perovskite blocks in the A3n+1B3n+4O9n+11 homologous series. Layers 1, 2 and 6, 7 belong to the perovskite blocks; layers 3−5 constitute the interface. (b) Atomic arrangement within the layers viewed along the stacking direction (i.e., crystallographic c direction). Orange and light red circles represent perovskite A positions. Blue and green circles represent perovskite B positions, which are coordinated with the octahedra filled with the same color. For clarity, the O positions are not shown in (a), and in (b), they are indicated as black circles. Red dotted lines delimit repeat units in the structure. Gray ovals in (a) highlight double BO6 octahedral chains at the interfaces.

Figure 11. Temperature dependences of dielectric permittivity of Bi10Ti7Fe6O38 and Bi13Ti7Fe9O47 measured at frequencies of 1 kHz (solid line), 5 kHz (dashed line), and 20 kHz (dashed−dotted line).

dielectric constant of the structurally related Bi2Ti4O12 was attributed to the displacements of Bi3+ ions in the S-shaped tunnels and the off-center motion of the Ti4+ ions in the distorted oxygen octahedra.14 Both of these mechanisms can be applied to explain the high dielectric permittivity of the Bi3n+1Ti7Fe3n−3O9n+11 compounds.

The composition of the layers is given for the 3ap × 1ap repeat unit in the ab plane, which arises owing to the alternating corner- and edge-sharing connectivity of the double octahedral chains to octahedra of the perovskite blocks. Therefore, the composition of the (AO) and (BO2) perovskite layers is represented here as (A3O3) and (B3O6). In the sequence, n stands for the number of the (BO2) layers between the interfaces, i.e., the number of the octahedral layers in the perovskite blocks. The first member of the series with a single (BO2) layer corresponds to the composition A4B7O20 (Figure 13). The composition of other homologues is derived by adding 3(n − 1)ABO3 (n > 1) to the A4B7O20 formula, in accordance with the increasing thickness of the perovskite blocks (Figure 12). In spite of the plausible tentative crystal



DISCUSSION Our XPD and TEM data unequivocally demonstrate the formation of a new perovskite based homologous series Bi3n+1Ti7Fe3n−3O9n+11. Although the composition of the series seems to be rather complex, it can be explained by considering the stacking sequence of atomic layers constituting the structure. The atomic arrangement of each layer viewed along the [001] stacking direction is demonstrated in Figure 12. For the general A3n+1B3n+4O9n+11 formula, the repeat sequence of the atomic layers along the c axis can be derived as follows −[( B3O6 −A3O3)n − 1−B3O6 −A 2B1O4 −B2O6 −A 2B1O4 ]− I

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 13. Crystal structure models of the A3n+1B3n+4O9n+11 homologues for n = 1−4.

structure and simple chemical composition of the first two members of the series, our attempts to prepare these compounds using high-temperature solid-state reaction were not successful. For Bi4Ti7O20 (n = 1), annealing at 800−1000 °C produces a mixture of Bi2Ti4O11 and Bi4Ti3O12. For Bi7Ti7Fe3O29 (n = 2), it results in a mixture of phases with the main contributions of Bi10Ti7Fe6O38 (n = 3) and Bi2Ti2O7. More advanced techniques, such as soft chemistry, mechanochemical activation, or high-pressure synthesis, may be required for obtaining the first two members of this homologous series. One can notice that the atomic arrangement within the layers constituting the interfaces in A3n+1B3n+4O9n+11 (layers 3−5 in Figure 13) is based on the perovskite-like grid, and the formation of the interfaces can be represented as a set of configurational and compositional changes in the perovskite matrix. Indeed, the interfaces eliminate some A positions, and displace the neighboring perovskite blocks over 1/2[110]p. In this context, the interfaces can be compared to the crystallographic shear (CS) planes in the AnBnO3n−2 perovskite-based homologous series. In these materials, the 1/2[110](101)p crystallographic shear (CS) planes slice the perovskite structure into uniform {110}p-shaped 2D slabs and accommodate anion deficiency by converting the corner-sharing BO6 octahedra into edge-sharing BO5 tetragonal pyramids along the planes.5 The notation indicates that the perovskite blocks are displaced over the 1/2[110]p vector along the (101)p crystallographic plane. If the CS planes are strictly confined to the {110}p crystallographic planes, they do not alter the A/B cationic ratio but reduce the oxygen content. In the A3n+1B3n+4O9n+11 series, the 1 /2[110]p displacement is applied to the {001}p plane. In contrast to the AnBnO3n−2 series, it does not change the coordination number of the B cations: they all remain octahedrally coordinated. In the AnBnO3n−2 series, the terminal oxygen atoms linked only to one B cation would be positioned too close to the A cations, so these terminal atoms are eliminated, resulting in an oxygen vacancy, which transforms BO6 octahedra into BO5 pyramids (Figure 14). In the A3n+1B3n+4O9n+11 series, such terminal O atoms remain; they are directly bonded to the Bi cations (A3 and O9 positions with d = 2.16 Å in Bi10Ti7Fe6O38, Figure 15). These terminal atoms are confined to the (B2O6) (= BO3) atomic layers (layer 4 in

Figure 14. Comparison of the interfaces between the perovskite blocks in the A3n+1B3n+4O9n+11 (a) and AnBnO3n−2 (b) homologous series. White squares mark the oxygen vacancies, transforming BO6 octahedra into BO5 square pyramids.

Figure 13), which become anion-excessive compared to the (B3O6) (= BO2) perovskite composition. Nevertheless, the interfaces in A3n+1B3n+4O9n+11 still reduce the oxygen content, as reflected by the O/B atomic ratio of 2.857 in the first member A4B7O20; the O/B ratio approaches 3 at n → ∞. In this case, the reduction of the oxygen content is attributed to the edgesharing connectivity of the BO6 octahedra between the (B3O6) and (A2BO4) layers and between the (B2O6) and (A2BO4) layers. Nevertheless, these interfaces require B cations with higher formal charge in order to compensate for the A-cation J

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

structures can be stabilized by lone pair cations adopting asymmetric oxygen coordination.17−20 Although the S-shaped tunnels in the A3n+1B3n+4O9n+11 structures have a new and very unusual topology for perovskite-based structures, similar structure fragments can be found in other complex oxides, such as Bi2Ti4O11 and PbTi3O7 (Figure 16).21,22 The Bi2Ti4O11 and PbTi3O7 structures are the

Figure 15. (left) Local coordination environment of the A positions in Bi10Ti7Fe6O38 as refined in the idealized model 1. (right) ELF isosurfaces showing the electron localization around the Bi3+ cations. Note the characteristic lobe-shaped ELF distribution around the A3 positions attributed to the lone electron pair domains that complete the coordination environment of the Bi3+ cations. There are also signs of electron localization around the A1 and A2 positions, but they do not have well-defined lobe shapes because of atomic displacements unaccounted for in this structure model.

deficiency caused by the edge-sharing. Thus, such interfaces can serve to accommodate donor dopants in perovskites with the lone electron pair A cations, such as Ti4+ cation in the Ti-doped (Bi,Nd)FeO3. The characteristic feature of the interfaces in the A3n+1B3n+4O9n+11 homologues is the S-shaped tunnels populated by the Bi3+ cations with lone electron pairs. Owing to specific covalent cation−anion interactions, which involve mixing of the hybridized [Bi6s−O2p]* and Bi6p states, the lone pair cations often demonstrate stereochemical activity and occupy positions with an asymmetric coordination environment.16 In the Bi10Ti7Fe6O38 structure, the A3 position at the interfaces has incomplete five-fold oxygen coordination with four longer basal Bi−O distances of 2.36−2.47 Å and one shorter apical distance of 2.17 Å (Figure 15, left). The adjacent A3O5 tetragonal pyramids are connected by common edges into zigzag chains, leaving empty space for the lone pair domains at the tails of the S-shaped tunnels. This is confirmed by our ELF calculations showing characteristic lobe-shaped domains of electron localization for these positions (Figure 15, right). The lone pair domains complete the coordination of the Bi cations and help relieving configurational strain associated with the formation of the interfaces. The A3n+1B3n+4O9n+11 homologues series is a new manifestation of the so-called “chemical scissor” concept, where complex interfaces between the modules of simple parent

Figure 16. Crystal structures of Bi10Ti7Fe6O38, Bi2Ti4O11, PbTi3O7, and TiO2 (anatase). Rectangles highlight the S-shaped tunnels in Bi10Ti7Fe6O38, Bi2Ti4O11, and PbTi3O7, and triangles mark double chains of the edge-sharing octahedra.

limiting members of the MxBi2−2xTi4−xO11−4x (M − divalent cation) family, which arises from introduction of CS planes into the Bi2Ti4O11 structure.23 Ignoring small differences in the atomic displacements, the octahedral fragments creating the Sshaped tunnels in Bi2Ti4O11 are very similar to those of Bi10Ti7Fe6O38. However, the distribution of the tunnels in these structures is different. In Bi2Ti4O11, the tunnels have the same configuration and form chess-board patterns. In Bi10Ti7Fe6O38, there are two types of mirror-related tunnels, which line up in an alternating fashion to form planar interfaces. This way, the A3n+1B3n+4O9n+11 homologues can be considered as an intergrowth between the perovskite blocks and polysynthetically twinned fragments of the Bi2Ti4O11 structure. The interfaces in the A3n+1B3n+4O9n+11 compounds are based on the chains of double edge-sharing octahedra. These structure fragments are the main building blocks of the TiO2 anatase structure (Figure 16), where they run along the a and b crystallographic directions, forming a criss-cross pattern. Thus, for a more intuitive description, the A3n+1B3n+4O9n+11 homoK

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry logues can be described as modular structures built up of 2D perovskite and 1D anatase modules.

XPD profiles for the Bi3n+1Ti7Fe3n−3O9n+11 samples with n = 3−6, ED pattern of Bi10Ti7Fe6O38 with stacking faults, Le Bail fit of the XPD pattern of the Bi13Ti7Fe9O47, results of the EDX analysis, crystallographic data for the “split” Bi10Ti7Fe6O38 structure and the corresponding Rietveld refinement profiles, schematic illustration of the lateral displacements of the interfaces, fitting parameters for the Mössbauer spectra, detailed description of Video 1, and temperature dependence of magnetic susceptibility for Bi10Ti7Fe6O38 (PDF) Crystallographic data data for two refined models of the Bi10Ti7Fe6O38 structure: (1) with idealized atomic positions and (2) with split Bi and some O positions (CIF) Video 1 showing the time series of the [010] HAADFSTEM images for Bi10Ti7Fe6O38 and for Bi10Ti7Fe6O38 unveiling the dynamics of Bi atomic columns in the structure (AVI)



CONCLUDING REMARKS We have demonstrated the existence of a new family of layered perovskite compounds Bi3n+1Ti7Fe3n−3O9n+11. In these compounds, the perovskite blocks are separated by the periodically placed (001)p interfaces that shift the adjacent perovskite blocks over 1/2[110]p. The interfaces are based on anatase-like chains of edge-sharing BO6 octahedra, which, together with the octahedra of the perovskite blocks, delimit S-shaped tunnels occupied by double columns of the Bi3+ cations. The thickness of the perovskite blocks increases with n, while the atomic arrangement at the interfaces remains nearly the same. Apparently, the stereochemical activity of the lone pair Bi3+ cations plays an important role in stabilizing the interfaces. In the tunnels, they occupy positions with highly asymmetric oxygen coordination, where their coordination environment is completed by the lone pair domains. Inside the perovskite blocks, the Bi3+ cations also demonstrate stereochemical activity of their lone electron pairs. Together with the adjacent oxygen atoms, they cooperatively displace from their average positions, forming short-range ordered patterns of atomic displacements. This can be attributed to a limited number of O atoms, which can simultaneously form covalent bonding with the Bi3+ cations, and to the deformations of the perovskite matrix induced by the periodic interfaces. We expect that the A3n+1B3n+4O9n+11 materials will demonstrate large chemical flexibility and tolerate diverse chemical substitutions. The A positions in the perovskite blocks should be capable of accommodating alkali, alkali-earth, and rare-earth cations. The A positions in the S-shaped tunnels are suitable for the lone electron pair cations, such as Bi3+ and Pb2+. These substitutions should open up a possibility to tune the overall charge of the A sublattice and vary the cation composition of the B sublattice, so that the composition and properties of the A3n+1B3n+4O9n+11 homologues can be tuned. Although the n = 3 Bi10Ti7Fe6O38 member is paramagnetic due to the dilution of magnetic cations with nonmagnetic Ti4+, our preliminary Mössbauer experiments on the n = 4 Bi13Ti7Fe9O47 compound demonstrate that it transforms to a magnetically ordered state. A conjecture that the mixing of the d0 Ti4+ and d5 Fe3+ cations impedes concerted off-center polar cation displacements in Bi10Ti7Fe6O38 also looks plausible. Full potential of the new A3n+1B3n+4O9n+11 homologous series is still to be explored, and this work is now underway. Another degree of compositional flexibility can be expected if the interfaces deviate from the (001)p orientation, for example, by introducing periodic sideway steps decreasing the A/B ratio, such as those observed by TEM in (Bi0.85Nd0.15)(Ti0.1Fe0.9)O3.24 Mastering these stepped interfaces to form long-range ordered arrangements, similar to those found in the (Pb,Bi)1−xFe1+xO3−y perovskites modulated by crystallographic shear planes,25 will generate new homologous series, where the composition depends not only on the spacing between the interfaces but also on their orientation with respect to the perovskite matrix.





AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: +7 (495) 280 14 81. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We are grateful to the Laboratory for Neutron Scattering and Imaging of Paul Scherrer Institut (LNS PSI, Villigen, Switzerland) for granting beam time at the HRPT diffractometer and to Dr. Denis Sheptyakov for the technical support during the experiment. We are also grateful to Valery Verchenko for his help with magnetization measurements. The work has been supported by the Russian Science Foundation (grant 14-13-00680). A.A.T. was partly supported by the Federal Ministry for Education and Science through a Sofja Kovalevskaya Award of Alexander von Humboldt Foundation.



REFERENCES

(1) Raveau, B. Prog. Solid State Chem. 2007, 35, 171−173. (2) Ferraris, G.; Makovicky, E.; Merlino, S. Crystallography of Modular Materials; Oxford University Press: Oxford, U.K., 2003. (3) Radaelli, P. G.; Marezio, M.; Perroux, M.; de Brion, S.; Tholence, J. L.; Huang, Q.; Santoro, A. Science 1994, 265, 380−383. (4) Abakumov, A. M.; Hadermann, J.; Bals, S.; Nikolaev, I. V.; Antipov, E. V.; Van Tendeloo, G. Angew. Chem., Int. Ed. 2006, 45, 6697−6700. (5) Abakumov, A. M.; Hadermann, J.; Batuk, M.; D’Hondt, H.; Tyablikov, O. A.; Rozova, M. G.; Pokholok, K. V.; Filimonov, D. S.; Sheptyakov, D. V.; Tsirlin, A. A.; Niermann, D.; Hemberger, J.; Van Tendeloo, G.; Antipov, E. V. Inorg. Chem. 2010, 49, 9508−9516. (6) Wang, L. Q.; Schaffer, B.; MacLaren, I.; Miao, S.; Craven, A. J.; Reaney, I. M. J. Phys. Conf. Ser. 2012, 371, 012036. (7) MacLaren, I.; Wang, L.; Morris, O.; Craven, A. J.; Stamps, R. L.; Schaffer, B.; Ramasse, Q. M.; Miao, S.; Kalantari, K.; Sterianou, I.; Reaney, I. M. APL Mater. 2013, 1, 021102. (8) Petříček, V.; Dušek, M.; Palatinus, L. Z. Kristallogr. - Cryst. Mater. 2014, 229, 345−352. (9) Momma, K.; Izumi, F. J. Appl. Crystallogr. 2011, 44, 1272−1276. (10) Koch, C. Determination of Core Structure Periodicity and Point Defect Density Along Dislocations. Ph.D. Thesis, Arizona State University, Phoenix, Arizona, 2002. (11) Andersen, O. K.; Pawlowska, Z.; Jepsen, O. Phys. Rev. B: Condens. Matter Mater. Phys. 1986, 34, 5253−5269.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.5b02465. L

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry (12) Perdew, J. P.; Wang, Y. Phys. Rev. B: Condens. Matter Mater. Phys. 1992, 45, 13244−13249. (13) Menil, F. J. Phys. Chem. Solids 1985, 46, 763−789. (14) Liu, J.; Duan, C.-G.; Yin, W.-G.; Mei, W. N.; Smith, R. W.; Hardy, J. R. J. Chem. Phys. 2003, 119, 2812−2819. (15) Tobar, M. E.; Krupka, J.; Ivanov, E. N.; Woode, R. A. J. Appl. Phys. 1998, 83, 1604−1609. (16) Walsh, A.; Payne, D. J.; Egdell, R. G.; Watson, G. W. Chem. Soc. Rev. 2011, 40, 4455−4463. (17) Yang, T.; Abakumov, A. M.; Hadermann, J.; Van Tendeloo, G.; Nowik, I.; Stephens, P. W.; Hemberger, J.; Tsirlin, A. A.; Ramanujachary, K. V.; Lofland, S.; Croft, M.; Ignatov, A.; Sun, J.; Greenblatt, M. Chem. Sci. 2010, 1, 751−762. (18) Abakumov, A. M.; Hadermann, J.; Van Tendeloo, G.; Antipov, E. V. J. Am. Ceram. Soc. 2008, 91, 1807−1813. (19) Mayerová, Z.; Johnsson, M.; Lidin, S. Angew. Chem., Int. Ed. 2006, 45, 5602−5606. (20) Jo, V.; Kim, M. K.; Lee, D. W.; Shim, I.-W.; Ok, K. M. Inorg. Chem. 2010, 49, 2990−2995. (21) Kahlenberg, V.; Böhm, H. Acta Crystallogr., Sect. B: Struct. Sci. 1995, 51, 11−18. (22) Kato, K.; Kawada, I.; Muramatsu, K. Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem. 1974, 30, 1634−1635. (23) Michiue, Y.; Yamamoto, A.; Onoda, M.; Sato, A.; Akashi, T.; Yamane, H.; Goto, T. Acta Crystallogr., Sect. B: Struct. Sci. 2005, 61, 145−153. (24) MacLaren, I.; Wang, L.; Craven, A. J.; Ramasse, Q. M.; Schaffer, B.; Kalantari, K.; Reaney, I. M. APL Mater. 2014, 2, 066106. (25) Abakumov, A. M.; Batuk, D.; Hadermann, J.; Rozova, M. G.; Sheptyakov, D. V.; Tsirlin, A. A.; Niermann, D.; Waschkowski, F.; Hemberger, J.; Van Tendeloo, G.; Antipov, E. V. Chem. Mater. 2011, 23, 255−265.

M

DOI: 10.1021/acs.inorgchem.5b02465 Inorg. Chem. XXXX, XXX, XXX−XXX