Site Selectivity of Hydride in Early-Transition-Metal Ruddlesden

Aug 16, 2018 - In the double-layered material, hydride anions show a clear preference for the bridging-apical site within the perovskite slabs. In con...
19 downloads 0 Views 3MB Size
Download PDF
Article Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX

pubs.acs.org/IC

Site Selectivity of Hydride in Early-Transition-Metal Ruddlesden− Popper Oxyhydrides Olivier J. Hernandez,† Gregory Geneste,‡ Takeshi Yajima,§,∥ Yoji Kobayashi,§ Masatoshi Okura,§ Kouhei Aidzu,§ Ced́ ric Tassel,§ Serge Paofai,† Diptikanta Swain,†,# Clemens Ritter,∇ and Hiroshi Kageyama*,§,⊥ †

Univ Rennes, CNRS, ISCR (Institut des Sciences Chimiques de Rennes) - UMR 6226, F-35000 Rennes, France CEA, DAM, DIF, F-91297 Arpajon, France § Department of Energy and Hydrocarbon Chemistry, Graduate School of Engineering, Kyoto University, Nishikyo-ku, Kyoto 615-8510, Japan ∥ Institute for Solid State Physics, The University of Tokyo, Kashiwa, Chiba 277-8581, Japan ⊥ CREST, Japan Science and Technology Agency (JST), Kawaguchi, Saitama 332-0012, Japan # Solid State and Structural Chemistry Unit, Indian Institute of Science, Bangalore 560012, India ∇ Institut Laue-Langevin, 71 avenue des Martyrs CS 20156, 38042 Grenoble Cedex 9, France

Inorg. Chem. Downloaded from pubs.acs.org by UNIV OF WINNIPEG on 08/18/18. For personal use only.

‡

S Supporting Information *

ABSTRACT: Layered perovskite titanium oxyhydrides have been prepared by lowtemperature topochemical CaH2 reduction from Ruddlesden−Popper Srn+1TinO3n+1 phases (n = 1, 2) and structurally characterized by combined synchrotron X-ray and neutron diffraction data refinements. In the single-layered Sr2TiO3.91(2)D0.14(1) material, hydride anions are statistically disordered with oxides on the apical site only, as opposed to known transition-metal oxyhydrides exhibiting a preferred occupation of the equatorial site. This unprecedented site selectivity of H− has been reproduced by periodic DFT+U calculations, emphasizing for the hydride defect a difference in formation energy of 0.24 eV between equatorial and apical sites. In terms of electronic structure, the model system Sr2TiO3.875H0.125 is found to be slightly metallic and the released electron remains mostly delocalized over several Ti atoms. On the other hand, hydride anions in the double-layered Sr3Ti2O6.20H0.12 material show a clear preference for the bridging apical site within the perovskite slabs, as confirmed by DFT calculations on the Sr3Ti2O6.875H0.125 model system. Finally, the influence of the B-site chemical nature on the hydride site selectivity for early 3d transition metals is theoretically explored in the single-layered system by substituting vanadium for titanium. The V3+ electronic polaron is suggested to play a role in stabilizing H− on the equatorial site in Sr2VO4−xHx for x = 0.125.

■

layered perovskites.6 A good correlation is found between the anion charge in a given site (estimated from the experimentally observed anion chemical occupancies) and the calculated bond strength sum (b). For instance, nitride (N3−) ions in Sr2TaO3N with n = 1 RP structure type selectively occupy the equatorial site, which can be understood by a larger value of bequatorial (vs baxial). For the same reason, axial site preference is observed for monovalent halide ions in layered perovskite oxyhalides (e.g., Ca2CuO2Cl2, K2NbO3F, and Sr3Co2O5Cl2). A growing number of transition-metal oxyhydrides have been recently reported, in particular with perovskite-related structures (see the reviews in refs 7 and 8). The small mass, low valence, and lability of the hydride (H−) ion offer new possibilities for materials design: for example, as hydride anion

INTRODUCTION Mixed-anion compounds exhibit intriguing properties that arise from the difference in anionic characters such as valence, ionic radius, electronegativity, and polarizability. The observed properties range from nontoxic pigments in (Ca1−xLax)Ta(O2−xN1+x),1 high Tc superconductivity in LixHfNCl and La(O1−xFx)FeAs,2,3 and a high-capacity conversion electrode in FeOF4 to visible-light-responsive water splitting in (ZnO)1−x(GaN)x.5 Here, anion order−disorder phenomena are expected to play a crucial role in their properties. Nevertheless, the control of anion distributions is still unsatisfactory as opposed to cation ordering and thus still remains a challenge. In 2006, Fuertes demonstrated that Pauling’s second crystal rule (PSCR) allows one to predict the distribution of anions in oxyhalides and oxynitrides with diverse structural types and compositions, including n = 1−3 Ruddlesden−Popper (RP) © XXXX American Chemical Society

Received: June 14, 2018

A

DOI: 10.1021/acs.inorgchem.8b01645 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 1. Crystal structures of (a) Sr2TiO4 (n = 1) and (b) Sr3Ti2O7 (n = 2) and the hydridized compounds (c) Sr2TiO4−xDx and (d) Sr3Ti2O7−xHy. Sr sites are shown in green, Ti sites in light blue, O sites in red, and H/D sites in blue.

conductors9−11 and precursors enabling multistep topochemical reactions.12−14 The anion order−disorder in perovskite oxyhydrides can be somewhat categorized by the d-electron count; ATi(O,H)3 (d0−0.6)15−18 and SrCrO2H (d3)19 adopt a cubic perovskite with a random anion distribution, whereas the d2 configuration in SrVO2H accounts for the trans-VO4H2 octahedron, leading to doubly degenerate dxz/dyz orbitals as the low-lying state.20 Remarkably, all the currently known layered transition-metal perovskite oxyhydrides (e.g., LaSrCoO3H0.7,21 Sr2Mn(O,H)4,22 Sr3Co2O4.33H0.84,23 Sr2VO3H, and Sr3V2O5H220) have an opposite tendency in comparison to oxyhalides; indeed these former compounds display preferential occupancy of the hydride ion at the equatorial site, despite its smaller valence (vs O2−).24 In this study, we demonstrate that titanium RP oxyhydrides Sr2Ti(O,H)4 and Sr3Ti2(O,H)7 can be synthesized using a lowtemperature CaH2 topochemical reduction from the corresponding n = 1, 2 members of RP oxides, Sr2TiO4 and Sr3Ti2O7. The Rietveld structural analyses of high-resolution synchrotron X-ray and neutron powder diffraction data reveal a distinct site selectivity between the n = 1 and n = 2 systems. Experimental results are confirmed and further tentatively rationalized by state of the art DFT calculations using model systems that are closely related to the synthesized materials. In addition, the influence of the B-site nature on the hydride site selectivity in the single-layered system is formally explored for an early 3d transition metal, substituting vanadium for titanium.

■

mol excess of CaH2, pelletized in a nitrogen-filled glovebox, sealed in an evacuated Pyrex tube (15 cm3) at pressures below 2 × 10−2 Pa, and heated for 3−14 days at 480 °C. The obtained products were washed with 0.1 M NH4Cl/MeOH in air to remove residual byproducts CaO and excess CaH2. For the RP n = 1 oxyhydride material, deuterated samples were specifically prepared according to the same procedure by using CaD2 and ND4Cl/MeOD. Characterization. Transmission electron microscopy (TEM) images were taken using a Hitachi NAR-9000 (300 kV) or JEOL JEM-2010 (200 kV) microscope. Scanning electron microscopy (SEM) images were taken in a Hitachi S-3400N microscope, with an operating voltage of 30 kV. Laboratory X-ray diffraction patterns were measured on a Bruker D8 ADVANCE diffractometer equipped with a Cu Kα source. For structural refinements, high-resolution synchrotron X-ray powder diffraction (SXRPD) experiments were performed at room temperature using a large Debye−Scherrer camera installed at the beamline BL02B2 of the Japan Synchrotron Radiation Research Institute, SPring-8. The structures of Sr2TiO4 and Sr3Ti2O7 and the hydridized compounds Sr2TiO4−xDx and Sr3Ti2O7−xHy are given in Figure 1. The incident beam from a bending magnet was monochromated to 0.420346 Å. The finely ground powder samples were sieved (32 μm) and put into a Pyrex capillary of 0.1 mm inner diameter. The sealed capillary was rotated during measurements so as to minimize the effect of potential preferential orientation. The diffraction patterns were collected in a 2θ range from 0 to 30° for Sr2Ti(O,D)4 and from 0 to 68° for Sr3Ti2(O,H)7, with a step interval of 0.01°. High-resolution neutron powder diffraction (NPD) experiments were carried out at room temperature using BT-1 (λ = 1.5403 Å) at the National Institute of Standards and Technology (NIST) for the n = 1 deuterated sample (∼5 g) and D2B at the Institute Laue-Langevin (λ = 1.594 Å) for the n = 2 hydridized sample (∼12 g). The collected SXRPD and NPD data were simultaneously analyzed by combined Rietveld refinements using the FullProf program.25 A Thompson−Cox−Hastings pseudo-Voigt profile function was used. Anisotropic peak broadening due to microstructural effects was modeled either through the Stephens phenomenological model26 (in case of microstrain effects) or using the Scherrer formula and a spherical harmonics expansion of the crystallite shape (in the case of size effects). Additionally, the width and shift of several selected isolated peaks were relaxed. Thermogravimetric analysis (TGA) was conducted using a TGDTA 2000S instrument (Bruker AXS), by heating the sample (about 30 mg) at a rate of 10 °C/min to 800 °C under flowing O2 gas. During the TGA experiment, released gaseous products were monitored by quadrupole mass spectroscopy (MS) using a BrukerAXS MS9610 spectrometer. Density Functional Theory Calculations. DFT calculations were performed with the ABINIT code,27 using the projector augmented wave (PAW) method.28,29 We used LDA+U calculations (U = 3 eV, applied on Ti 3d orbitals). Tests performed using U = 4 eV (see the Supporting Information) gave similar results. All of the calculations involving defects (hydrides, vacancies) were spinpolarized, while those performed on the stoichiometric oxides were

EXPERIMENTAL SECTION

Synthesis. In order to examine the grain size effect on the reactivity of the CaH2 reduction, the precursor oxides Sr2TiO4 and Sr3Ti2O7 were synthesized by two routes: a citric acid method and a high-temperature solid-state reaction. For the citric acid method, SrCO3 (Rare Metallic, 99.9%, 65 mmol), SrNO3 (Rare Metallic, 99.9%, 65 mmol) and citric acid (Wako, 98.0%, 156 mmol) were dissolved in 250 cm3 of methanol, and this solution was added slowly to a 1/1 solution of titanium isopropoxide (Aldrich, 97%) and anhydrous ethylene glycol with stirring at 25 °C, where the Sr:Ti molar ratios were set as 2:1 and 3:2 for Sr2TiO4 and Sr3Ti2O7, respectively. We found that gels formed when the liquid was allowed to evaporate. White powders obtained by drying the gels at 300 °C were calcined in air at 800 °C for 24 h. For the high-temperature solid-state reaction, mixtures of stoichiometric amounts of SrCO3 (Rare Metallic, 99.9%) and TiO2 (Rare Metallic, 99.9%) were pelletized and calcined in air at 1200 °C for 24 h. The products were then reground and calcined at the same temperature for 48 h. Before reduction, the obtained powders were preheated at 120 °C overnight under vacuum to eliminate adsorbed water molecules. The dried specimens of Sr2TiO4 and Sr3Ti2O7 were mixed with a 3 and 6 B

DOI: 10.1021/acs.inorgchem.8b01645 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 2. Mass spectroscopy of (a) oxyhydrides Sr2Ti(O,D)4 (n = 1) and (b) Sr3Ti2(O,H)7 (n = 2) during the decomposition process under flowing Ar gas. not. The plane-wave cutoff was 25 Ha. We performed full structural optimizations (atomic positions and lattice vectors) until all of the Cartesian components of the atomic forces (respectively stress tensor components) were below 2 × 10−4 Ha/bohr (∼0.01 eV/Å) (respectively 2.0 × 10−6 Ha/bohr3). We first optimized the perfect Sr2TiO4 (n = 1) and Sr3Ti2O7 (n = 2) systems using 14- and 24-atom unit cells, for which the first Brillouin zone was sampled by a 6 × 6 × 2 k-point mesh. Then different defects (H and VO, described below) were studied using a supercell constructed by doubling the previous cells along the two lateral directions, leading respectively to 56- and 96-atom supercells. The structural optimizations of these supercells were performed using a 3 × 3 × 2 k-point mesh in accord with that used to study the perfect systems. We have checked, in the case of the n = 1 system, that a refined mesh (6 × 6 × 4) leads to optimized structures of apical H− and equatorial H− with the same energy difference, within 0.01 eV. The LDA+U calculations concerning H and VO in the n = 1 system were also compared to GGA+U calculations (with same U = 3 eV), which provided similar results. In several cases, the electronic density of states was computed, on a refined k-point mesh (24 × 24 × 8 for the perfect systems, 9 × 9 × 6 for the supercells). Simulation of the hydride incorporation was performed by simply replacing one O atom of the supercell by one H. This led to two model compounds of formulas Sr 2 Ti O 3 . 8 7 5 H 0 . 1 2 5 and Sr3Ti2O6.875H0.125, which are very close to the experimental stoichiometries. Formally, such substitution liberates one electron that populates a state at the bottom of the conduction band. The different possible sites for H were explored. In addition, we studied the neutral O vacancy by simply removing one O atom from the supercell (which liberates in that case two electrons). All of the systems presented in the article were thus, except when explicitly mentioned, globally neutral (no charged supercell used), or in other words, the defects were studied in their neutral charge state. This appears to be in good agreement with the experiments, certainly for Sr2TiO4−xDx and to a lesser extent for Sr3Ti2O7−xHy (see below), since no other point defectsthat would be necessary to compensate the charge of ionized or partially ionized defectsare present in the samples in a concentration as large as that of the H−. A few calculations using charged supercell were nevertheless performed and are presented with more details in the Supporting Information, as well as optimized lattice constants and energies of the different configurations studied.

conventional high-temperature solid-state reactions, the SEM images showed typical grain sizes ranging from 2 to 3 μm. The XRPD patterns revealed that all of the Sr2TiO4 (I4/ mmm) and Sr3Ti2O7 (I4/mmm) precursor samples obtained by both synthetic routes are of single phase. The cell parameters of Sr2TiO4 obtained by the citrate method (a = 3.88479(3) Å, c = 12.5919(1) Å) and solid-state reaction (a = 3.88445(2) Å, c = 12.5931(1) Å) are almost the same as those recently reported (3.88656(7), 12.5975(4) Å).30 On the other hand, while the cell parameters of Sr3Ti2O7 obtained by a solid-state reaction (a = 3.89885(2) Å, c = 20.3453(1) Å) appear very close to the reported values (a = 3.9026(1), c = 20.3716(4) Å from NPD data with λ = 1.893 Å and dmin = 0.96 Å),31 the lattice constant c of Sr3Ti2O7 obtained by the citrate method (a = 3.89932(8) Å, c = 20.3107(6) Å) is significantly lower. However it is worth recalling that the presence of stacking faults along the c axis in RP phases is well-known (whose occurrence depends upon the synthetic route used) and may affect the shape of the reflections and incidentally the refined unit cell parameter values.32 After the CaH2 reduction of Sr2TiO4 and Sr3Ti2O7, the white samples changed to dark blue, as encountered for ATiO3−xHx.15,16 The XRPD patterns of reduced Sr2TiO4 and Sr3Ti2O7 were all adequately refined with the same tetragonal space group (I4/mmm) as for the parent oxides. The dependence of the cell parameters and cell volume on reaction time as obtained by Le Bail refinements of laboratory XRPD data of the n = 1, 2 samples are plotted in Figures S2 and S3. For the n = 1 compounds, we observed in the course of the reduction an increase of the a axis while the c axis remained unchanged. The sample with smaller particle size (synthesized by the citrate method) showed a larger change in the a axis in comparison to the sample with a larger particle size (synthesized by a solid-state reaction). Most importantly, all diffraction peaks in the reduced sample obtained by a solidstate reaction have an asymmetric shape due to the existence of a shoulderlike peak at lower angles (see Figure S4), whereas the samples obtained by the citrate method do not present such an obvious asymmetric shape of the diffraction peaks. This means that the solid-state samples were reduced inhomogeneously: i.e., the inside of the particle was weakly reduced in comparison with the surface. For the n = 2 compound, we observed an increase of the a axis and a decrease of the c axis, the former change being more pronounced for the material synthesized by the citrate method. However, high-resolution SXRPD and NPD patterns of

■

RESULTS AND DISCUSSION CaH2 Reduction: Cell Parameter Changes versus Particle Size Dependence. For the starting materials, the TEM images of the materials prepared by the citrate method showed that the grains are less than 200 nm in size (Figure S1). In contrast, for the starting materials obtained by C

DOI: 10.1021/acs.inorgchem.8b01645 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

metric parent oxides. We observed a large mass gain of 1.85 wt % for Sr2TiO4−xDx (cf. Figure 3a) leading to the formula Sr2TiO3.65D0.35, where we neglected the impurity phases and assumed a negligible amount of anionic vacancies as observed for SrTiO3−yHy.16 Using average impurity amounts of 3.29 wt % for SrTiO3‑yDy (with y ≈ 0.6) and 0.96 wt % for TiD2 as obtained from the final combined SXRPD/NPD refinement (see below), we estimated the hydride concentration in Sr2TiO4−xDx to be x ≈ 0.27 from TGA, leading to the formula Sr2TiO3.73D0.27. The TGA curve for Sr3Ti2O7−xHy in Figure 3b shows that the relative mass increase is 3.95 wt %. Assuming again a negligible amount of anionic vacancies gives the chemical composition Sr3Ti2O5.76H1.24. Furthermore, taking into account a slight amount of TiH2 (0.58 wt % from Rietveld refinements, see below) leads to the formula Sr3Ti2O5.82H1.18. It is however clear for both 214 and 327 complex systems that the estimation of the hydride content by TGA under an oxidizing atmosphere is indirect and depends on the presence of impurity (remaining perovskite and more importantly TiH2 that transforms into TiO2) whose weight fraction, determined by Rietveld refinement, is itself subject to rather large uncertainties. In addition, the calculation of the H content from TGA results is made using the strong assumption that vacancies are absent. In Sr3Ti2O7−xHx, vacancies are observed (see the structural characterization section), obviously bringing significant additional errors on the TGA x value. It is worth noting that both Sr2TiO4 and Sr3Ti2O7 oxides undergo oxygen reduction of less than 0.1 wt % by conventional chemical methods such as high-temperature treatment under a H2 flow.33 Structural Characterization of Sr 2 TiO 4−xD x and Sr3Ti2O7−xHy. We analyzed the high-resolution SXRPD and NPD data by first performing Le Bail refinements (in order to properly model the reflection profiles) and then Rietveld refinements. For Sr2TiO4−xDx, we built a structural model on the basis of the Sr2TiO4 structure (space group I4/mmm).30 Using our experience in structural analysis of perovskite oxyhydrides,7,15 a combined SXRPD/NPD three-phase Rietveld refinement was carried out in the final stage of our investigations to account for the TiD2 and SrTiO3−yDy impurity phases (0.96 and 3.29 wt %, respectively). The refinement where all atoms were refined anisotropically converged successfully, providing satisfactory reliability factors (RBragg = 0.80 and 1.31% for SXRPD and NPD data; see Table 1 and Figure 4 for the Rietveld plots). As shown in Figure 1,

reduced samples obtained by the solid-state method turned out to be of much better quality than those obtained by the citrate method owing to the presence of much fewer impurity peaks, more symmetrical peak profiles, and much better signal to noise ratio. These observations led us to finally use for subsequent analyses Sr2TiO4 and Sr3Ti2O7 samples reduced respectively for 7 and 8 days from precursors obtained by the citrate and solid-state methods, unless otherwise stated. This reaction period was selected because the highest change in lattice parameters was observed according to the laboratory XRPD data. Existence of Hydride Ion. Hydrogen gas release under an inert atmosphere at high temperature has been reported in ATiO3−xHx oxyhydrides.15,16 When the reduced Sr2TiO4 and Sr3Ti2O7 specimens were heated under an Ar atmosphere, an intense peak derived from hydrogen gas was observed in the MS signal (Figure 2). More specifically, the reduced Sr2TiO4 and Sr3Ti2O7 exhibit one sharp peak at around 600 and 520 °C, respectively, implying a successful synthesis of the oxyhydrides Sr2TiO4−xDx and Sr3Ti2O7−xHy. However, a quantitative estimation of hydrogen content is not possible without calibration with standards. The clear peak indicates that the amount of released hydride ions is quite significant. In order to estimate the anion stoichiometry of Sr2TiO4−xDx and Sr3Ti2O7−xHy, TGA data were collected under O2 gas, from which we observed a considerable mass increase (cf. Figure 3). After the TGA measurements, the RP oxyhydride specimens should transform into their respective stoichio-

Table 1. Refined Structural Parameters of Sr2Ti(O,D)4 at Room Temperature (SXRPD/NPD Combined Rietveld Refinement)a atom

site

occ

x

y

z

Ueq (Å2)

Sr Ti O1 D1 O2 D2

4e 2a 4c 4c 4e 4e

1 1 1 0 0.938(5) 0.070(6)

0 0 0 0 0 0

0 0 0.5 0.5 0 0

0.35573(4) 0 0 0 0.15915(9) 0.15915(9)

0.0071(2) 0.0072(6) 0.0084(7) 0.0084(7) 0.0110(6) 0.0110(6)

a

All atoms were refined anisotropically. Uij values for O and D on the same site are set to the same least-squares parameter. Crystal data: space group I4/mmm, a = 3.89892(2) Å, c = 12.58797(8) Å. SXRPD: RBragg = 0.804%, Rwp = 4.58%, Rp = 5.91%, χ2 = 1.30. NPD: RBragg = 1.32%, Rwp = 7.07%, Rp = 7.01%, χ2 = 1.12. Global user-weighted χ2 = 1.21.

Figure 3. Thermogravimetric data collected during the oxidation process of (a) Sr2TiO4−xDx and (b) Sr3Ti2O7−xHy under O2 flow. D

DOI: 10.1021/acs.inorgchem.8b01645 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

For Sr3Ti2O7−xHy, we initially built a structural model on the basis of the Sr3Ti2O7 structure (space group I4/mmm).31 The combined refinement against SXRPD and NPD data was based on this model, taking into account that the anionic sites may be occupied by oxide, hydride, and vacancy, gives acceptable agreement factors (RBragg = 5.66 and 5.51% for SXRPD and NPD data, respectively; see Table 2 and Figure 5 for the Table 2. Refined Structural Parameters of Sr3Ti2(O,H)7 at Room Temperature (SXRPD/NPD Combined Rietveld Refinement: Model Including the Presence of Vacancy on the O1 Bridging Apical Site)a atom

site

occ

x

y

z

Uiso (Å2)

Sr1 Sr2 Ti O1 H1 O2 H2 O3 H3

2b 4e 4e 2a 2a 4e 4e 8g 8g

1 1 1 0.557(20) 0.12(4) 1 0 0.910(8) 0

0.5 0.5 0 0 0 0 0 0 0

0.5 0.5 0 0 0 0 0 0.5 0.5

0 0.18501(7) 0.09915(16) 0 0 0.1975(2) 0.1975(2) 0.09463(18) 0.09463(18)

0.0047(4) 0.0102(4) 0.0047(4) 0.017(5) 0.017(5) 0.0068(10) 0.0068(10) 0.0069(8) 0.0069(8)

a All atoms were refined isotropically. Uiso values for O and H on the same site are set to the same least-squares parameter. Crystal data: space group I4/mmm, a = 3.90128(4) Å, c = 20.3219(3) Å. SXRPD: RBragg = 5.66%, Rwp = 13.4%, Rp = 16.9%, χ2 = 5.23. NPD: RBragg = 5.54%, Rwp = 15.0%, Rp = 18.0%, χ2 = 4.20. Global user-weighted χ2 = 4.71.

Rietveld plots), where all atoms were refined isotropically. Despite the fact that only one impurity phase was slightly detected (0.58 wt % of TiH2), refinement results are more unstable and less accurate, in comparison with those for Sr2TiO4−xDx, due to stronger peak overlapping, with worse profiles (even for the solid-state synthesized sample) and more anisotropic peak broadening. Sr3Ti2O7 displays the three inequivalent anionic sites O1, O2, and O3 (see Figure 1). The O1 and O3 sites, bridging-apical and equatorial sites, respectively, are in the perovskite layer, and the O2 site (apical site) is within the SrO rock salt layer. The results of the Rietveld refinement indicate with a rather high confidence that the hydride species occupy only the O1 site (bridging apical) among the three possible anionic sites. The occupancies of oxide and hydride ions for the O1 site are 0.665(15) and 0.335(15), respectively, assuming again the absence of any vacancy on this site. By relaxation of the chemical constraints for this site, no improvement in the structural agreement factors (RBragg) is obtained but the profile R factors appear slightly better (4.71 versus 4.77 for the global χ2). The chemical occupancies are significantly modified mostly through a decrease of hydride content correlatively to the presence of vacancies (56(2)% of oxygen, 11(4)% of hydride, and 33% of vacancy for the O1 site). For the O3 site, the presence of a slight amount of anion vacancy (8%) and absence of any hydride incorporation (the occupancy factor for H3 becomes highly negative once refined) has been unambiguously found, with an occupancy for the oxide anion of 0.918(8). The occupancy of oxide anion at the O2 site converged to 1 within the standard uncertainties, indicating that this site does not undergo any anion exchange. We note that in general, in our previous studies with BaTiO3−xHx, SrTiO3−xHx, and CaTiO3−xHx, we did not see

Figure 4. Combined SXRPD/NPD Rietveld plot of Sr2TiO4−xDx, showing observed (red circle), calculated (black line), and difference (blue line) profiles. The ticks represent the positions of Bragg reflections. SrTiO3−yDy and TiD2 impurities have been considered in the refinement.

Sr2TiO4−xDx is characterized by two inequivalent oxygen sites: the O1 site located in the single perovskite layer (equatorial site) and the O2 site within the SrO rock salt layer (apical site). We could not detect appreciable hydride anion or vacancy at the O1 site; the occupancy factor for oxygen on O1 converged to a value close to 1 (namely 1.019(6) without any constraints applied on chemical occupancies), while the site occupancies at the O2 site are 0.938(5) for O2− and 0.070(6) for D−, respectively (again without any occupancy constraints), r e s u l t i n g i n t h e fin a l r e fin e d s t o i c h i o m e t r y o f Sr2TiO3.91(2)D0.14(1) (that reasonably compares with the Sr2TiO3.73D0.27 formula estimated from TGA; see above). The refined chemical occupancies from the combined SXRPD/NPD Rietveld analysis indicate with certitude that no vacancy is present on any anionic site, as seen previously in ATiO3−xHx.15,16 The refinement also yields a very crude value for the hydride composition of the SrTiO3−yDy impurity phase (y = 0.57(15)), owing to the very small number of weak observable Bragg peaks for this phase. E

DOI: 10.1021/acs.inorgchem.8b01645 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

anion compounds with an RP structure often show anion order, strongly depending on the ionic size and electronegativity of the anion.34 For instance, oxynitrides and oxyfluorides with layered perovskite structures (e.g., Sr2TaO3N35 and RbLaNb2O6F36) have a preferential anion distribution despite similar ionic radii for anions. In comparison to oxygen (χO = 3.5), the less electronegative nitrogen (χN = 3.0) tends to occupy the anion site in the perovskite layer, whereas the more electronegative fluorine (χF = 4.0) tends to occupy the site of the rock salt layer. This is understood by the fact that the more electronegative anion is preferentially coordinated to the more electropositive metal. In RP type An+1BnO3n+1 compounds with B = transition metal, the A cation is generally less electronegative than the B cation, and this rule of thumb is realized for the title compounds (χSr = 1.0, χTi = 1.5). The anion in the AO rock salt layer is coordinated to five A cations and one B cation, whereas the anion in the An−1BnO3n−1 perovskite layer is coordinated to four A cations and two B cations. These coordination environments result in the occupancy of anions with higher electronegativity in the AO rock salt layer. The hydride ion has similar ionic size but a much smaller electronegativity (χH = 2.1) with regard to the oxide ion (χO = 3.5). Hence, according to pure electronegativity considerations, the hydride ions in the RP structure are expected to occupy the anion site within the perovskite layer (equatorial site) as encountered in oxynitrides. In fact, as underlined in the Introduction, the hydride anion in other known transition-metal oxyhydrides with RP structures effectively occupies the anionic site in the perovskite layer, as seen in the present work for Sr3Ti2O7−xHy. The selective hydride occupancy at the O2 apical site in Sr2TiO4−xDx is also quite interesting in terms of hydride migration during the synthesis. In nonlayered perovskite oxides (n = ∞), the oxide anions should diffuse along the BO6 octahedral edge.37 In iron and cobalt oxides with the RP structure, oxygen is selectively removed from the perovskite layer when the compound is reduced.38,39 Owing to this selective removal of oxygen from the perovskite layer, oxygen in the RP structure generally diffuses along the BO 6 octahedron edge within the perovskite layer as reported in the n = ∞ perovskite oxides.40 Thus, considering the traditional ionic migration pathway in perovskite oxides, hydride incorporation into the AO rock salt layer observed in Sr2TiO4−xDx should be accompanied by hydride incorporation into the perovskite layer. To account for the selective incorporation into the AO rock salt layer in the n = 1 Sr2TiO4−xDx oxyhydride, an alternative pathway for hydride diffusion involving only O2 sites should exist. The formation and migration of interstitial anions are generally unfavorable.37 However, interstitial oxygen and its migration in the AO rock salt layer has been found in several n = 1 RP oxides such as La2NiO4+δ and Pr2NiO4+δ.41,42 As in those cases, hydride might migrate through the SrO layer during the formation of Sr2TiO4−xDx and might result in selective hydride incorporation into the SrO layer. Since the mechanism and pathway of hydride diffusion are not clear even in ATiO3−xHx, further experimental studies such as high-temperature NPD experiment and in situ NPD during the formation of Srn+1Tin(O,H)3n+1 (n = 1, 2, ...) are crucial to understand the hydride incorporation mechanism and the distinct site selectivity observed for n = 1, 2. Nonetheless, in order to get better insights into the origin for this unprecedented difference in hydride site selectivity in the newly synthesized Sr2TiO4−xDx

Figure 5. Combined SXRPD/NPD Rietveld plot of Sr3Ti2O7−xHy, showing observed (red circle), calculated (black line), and difference (blue line) profiles. The ticks represent the positions of Bragg reflections. TiH2 impurity has been considered in the refinement.

any evidence of measurable amounts of vacancies; the anion sites were filled with either O2− or H−. Here, however, with the layered n = 2 systems we observe vacancies in addition to oxide and hydride depending on the refinement assumptions about vacancy presence (see above): Sr3Ti2O6.34(5)H0.33(2) (0.33 vacancies per formula unit, exclusively on the O3 site) and Sr3Ti2O6.20(5)H0.12(4) (0.68 vacancies per formula unit, on the O3 and O1 sites). On top of this, hydride selectively occupies the apical positions: for the n = 2 case the O1 “bridging apical” site within the perovskite slab (but not the O2 apical site), and for the n = 1 case the apical O2 site. Site Selectivity of Hydride Incorporation. The present titanium oxyhydrides with the RP structure consist of an alternate stacking of rock salt layers and perovskite layers. In this study on the n = 1, 2 oxyhydride phases, a preferential site selectivity is observed; for n = 1 the hydride anion is located at the apical site within the rock salt layer, while for n = 2 it is incorporated into the bridging apical site of the perovskite slab. As we discuss below, the site selectivity in Sr3Ti2O7−xHy is reasonable given the electronegativity of hydrogen (χH = 2.1), while that for Sr2TiO4−xDx is not readily explained. MixedF

DOI: 10.1021/acs.inorgchem.8b01645 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry and Sr3Ti2O7−xHy (besides PSCR or electronegativity considerations), we have undertaken state of the art DFT calculations, whose results are discussed hereafter. DFT Calculations. Our LDA+U calculations show that for the hydride ion in the n = 1 system the apical site is clearly more stable than the equatorial site by a significant energy difference of 0.24 eV (see Figure 6 for a representation of the

Figure 7. Electronic density of states of Sr2TiO4 (upper panel) and Sr2TiO3.875O0.125 with the hydride ion in the apical site. The Fermi energy (0 eV in both cases) is shifted to the bottom of the conduction band when the compound is hydridized (this figure illustrates the metallization consecutive to hydride insertion).

(height along c of the titanium octahedron) remains similar to that in the (oxide) perfect system (4.02 and 3.96 Å, respectively). For the n = 2 system, DFT calculations reveal that the two neighboring Ti atoms are also significantly pushed away along c by the (bridging apical) hydride (see the Supporting Information). Note that, in the latter (n = 2), a set of calculations has been done using charged supercells (see the Supporting Information): we find that the different sites for the charged defects exhibit the same energetic order as for the neutral ones, suggesting that the released (delocalized) electron(s) interact weakly with the defect. In order to try to rationalize these results for early transition metals in general, we also performed similar calculations on the Sr2VO4 compound, in which the V4+ ions are in a 3d1 configuration that constitutes a key electronic distinctiveness with regard to the d0 titanium oxides. In contrast to Sr2TiO4 and Sr3Ti2O7, which are charge-transfer insulators, Sr2VO4 is a Mott insulator,43 a property well reproduced by our LDA+U calculation (see Figure 8; the band gap in Sr2VO4 separates occupied 3d states of V at the top of the valence band from unoccupied 3d states of V at the bottom of the CB). We stress here that these calculations on Sr2VO4 are done for comparison; they do not pretend to accurately describe the complex electronic structure of the strongly correlated electron Sr2VO4 compound (which would require to go beyond the standard DFT+U), but we may expect that charge distribution and bonding will be correctly reproduced by DFT+U. We have adopted an antiferromagnetic (AFM) order that corresponds to alternating magnetic moments along the two lateral Cartesian directions and performed full structural optimizations of the perfect supercell and of the supercell with one H substituting one O (in both possible sites), for U = 5 and 3.5 eV (J = 0 eV) and U = 5 eV (J = 0.7 eV). In all cases, the equatorial site was found to be more stable than the apical site for the hydride ion, by more than 0.14 eV, in contrast to our results for Sr2Ti(O,H)4 and in good agreement with the results obtained for the Sr2VO4−xHx materials (with similar hydride content) synthesized by the solid-state high-pressure technique.43 Moreover, in the vanadate, the electron liberated by introducing the hydride ion is found to be localized on a single

Figure 6. View of Sr2TiO3.875H0.125, with H− in the apical site, the model compound used to reproduce by DFT+U the hydridizing properties of Sr2TiO4. Green, light blue, red, and blue spheres respectively denote Sr, Ti, O, and H atoms. The lattice vectors of the 56-atom supercell (periodically repeated in the three directions) are indicated with black arrows.

corresponding system). GGA+U calculations give similar results, as well as LDA+U with U = 3 and 4 eV and J = 1 eV (see the Supporting Information). In the n = 2 system, the bridging apical site is also clearly more stable than the two other sites for the hydride ion (0.17 eV more stable than the apical site, 0.28 eV more stable than the equatorial site, from LDA+U). This is in good agreement with the experimental results. Our DFT calculations indicate also that the neutral oxygen vacancy is the most stable at exactly the same sites as the hydride ion. The hydride-substituted compound is found to be slightly metallic (Figure 7): the electron liberated by substituting O by H (in the globally neutral supercell) populates a state at the bottom of the conduction band (CB). As a whole, for the n = 1 system, the introduction of H in place of O (in its most stable site, i.e., apical) produces a slight lateral expansion (by ∼0.01 Å along ⟨100⟩) and a small longitudinal compression (by ∼0.015 Å along [001]) of the unit cell (the O vacancy has the same effect). It is worth underlining that such a unit cell change upon hydridization is (i) totally corroborated by the experimental values (see ref 27 and the Supporting Information: ∼+0.015 Å for a and ∼−0.01 Å for c from Sr2TiO4 to Sr2TiO4−xDx) and (ii) opposite to what is experimentally obtained for Sr2VO4−xHx with comparable x values (∼0.13). 43 Within the titanium octahedron, the DFT-optimized supercell shows that the Ti− H local distance is considerably longer than the Ti−O distance in TiO6 octahedra by 0.18 Å (2.16 versus 1.98 Å; see Figure 6 and the Supporting Information). Additionally, the Ti atom is pushed away along c by the hydride anion toward the other apical (oxide) site (Ti−Oap = 1.86 Å) within the same octahedron, as the distance between both latter apical sites G

DOI: 10.1021/acs.inorgchem.8b01645 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

apical site) for n = 2. Degrees of anion disordering of 7 and 12% are observed in the n = 1, 2 RP oxyhydrides, respectively. This site selectivity has been well reproduced by DFT+U calculations, which also suggest that hydridized Sr2TiO4 and Sr3Ti2O7 may be metallic. In ATiO3−xHx (n = ∞), high electrical conductivities of 102−104 S/cm have been reported in the epitaxial thin film form.45 By analogy with ATiO3−xHx, electrical conductivity is expected in those RP oxyhydrides. Fabrication of an epitaxial thin film of An+1Bn(O,H)3n+1 is desirable to clarify its electronic properties and the relationship with the observed site selectivity. Finally for Sr2V(O,H)4, DFT calculations suggest that the site selectivity (hydride localization on the O1 equatorial site) is opposite to that observed in Sr2Ti(O,H)4, highlighting the potential role of electronic localization in non-d0 systems such as the formation of a small V3+ electron polaron.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.8b01645. Reaction period dependence of cell parameters for CaH2 reduced Sr2TiO4 and Sr3Ti2O7, XRPD patterns for Sr2TiO4 and Sr3Ti2O7 synthesized by solid-state reactions as well as their reduced derivatives (Sr2TiO4−xHx and Sr3Ti2O7−xHy oxyhydrides), and details about DFT calculations (PDF) Accession Codes

CCDC 1850059−1850060 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by emailing [email protected], or by contacting The Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223 336033.

■

AUTHOR INFORMATION

Corresponding Author

Figure 8. Electronic densities of states of (top) Sr2TiO4, (middle) Sr3Ti2O7, and (bottom) Sr2VO4, projected on the p states of O (l = 1) and the d states of Ti/V (l = 2). This figure illustrates that the titanates are charge-transfer insulators (gap O 2p−Ti 3d) while the vanadate is instead a Mott insulator (gap V 3d−V 3d).

*E-mail for H.K.: [email protected]. ORCID

Diptikanta Swain: 0000-0003-4048-5017 Hiroshi Kageyama: 0000-0002-3911-9864

V ion, next to the H−, as a small polaron V3+ (3d2), as also theoretically proposed in the titanate oxyhydride BaTiO3−xHx,44 suggesting for the vanadate oxyhydride that the electronic polaron may play a role in stabilizing the equatorial site.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by CREST (JPMJCR1421) from the JST and Kakenhi project (JP16H06439, JP16H02267) from the MEXT and the JSPS Core-to-Core Program (A) Advanced Research Networks. CNRS/JSPS supported also this bilateral work between Kyoto and Rennes through PRC No. 0684 (2013-2014). Fruitful discussions with Dr. Guilhem Dezanneau (CNRS, CentraleSupélec Paris) are warmly acknowledged. D.S. thanks the Région Bretagne for his SAD postdoc grant “RINSITU”.

■

CONCLUSION The reduction of the n = 1, 2 Ruddlesden−Popper phases Sr2TiO4 and Sr3Ti2O7 with CaH2 yields oxyhydride phases, as reported for ATiO3−xHx. Structural refinements revealed that the obtained chemical compositions are Sr2TiO3.91(2)D0.14(1) (n = 1) and Sr3Ti2O6.20(5)H0.12(4) (n = 2), assuming the presence of vacancy on the O1 bridging apical site for the latter. The location of the hydride ion strongly depends on the number n. For n = 1, the hydride anion occupies only the apical site, which is behavior different from that of other reported transition-metal oxyhydrides with the RP structure, where the hydride species occupies the equatorial site. In contrast, the hydride is positioned within the perovskite layer (bridging

■

REFERENCES

(1) Jansen, M.; Letschert, H. P. Inorganic yellow-red pigments without toxic metals. Nature 2000, 404, 980−982. (2) Ito, T.; Fudamoto, Y.; Fukaya, A.; Gat-Malureanu, I. M.; Larkin, M. I.; Russo, P. L.; Savici, A.; Uemura, Y. J.; Groves, K.; Breslow, R.;

H

DOI: 10.1021/acs.inorgchem.8b01645 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Cation Substitution: SrVO2H versus SrV1−xTixO1.5H1.5. Inorg. Chem. 2018, 57, 2890−2898. (19) Tassel, C.; Goto, Y.; Kuno, Y.; Hester, J.; Green, M.; Kobayashi, Y.; Kageyama, H. Direct Synthesis of Chromium Perovskite Oxyhydride with a High Magnetic-Transition Temperature. Angew. Chem., Int. Ed. 2014, 53, 10377−10380. (20) Denis Romero, F.; Leach, A.; Möller, J. S.; Foronda, F.; Blundell, S. J.; Hayward, M. A. Strontium Vanadium Oxide-Hydrides: “Square-Planar” Two-Electron Phases. Angew. Chem., Int. Ed. 2014, 53, 7556−7559. (21) Hayward, M. A.; Cussen, E. J.; Claridge, J. B.; Bieringer, M.; Rosseinsky, M. J.; Kiely, C. J.; Blundell, S. J.; Marshall, I. M.; Pratt, F. L. The Hydride Anion in an Extended Transition Metal Oxide Array: LaSrCoO3H0.7. Science 2002, 295, 1882−1884. (22) Tassel, C.; Goto, Y.; Watabe, D.; Tang, Y.; Lu, H.; Kuno, Y.; Takeiri, F.; Yamamoto, T.; Brown, C. M.; Hester, J.; Kobayashi, Y.; Kageyama, H. High-Pressure Synthesis of Manganese Oxyhydride with Partial Anion Order. Angew. Chem. 2016, 128, 9819−9822. (23) Helps, R. M.; Rees, N. H.; Hayward, M. A. Sr3Co2O4.33H0.84: An Extended Transition Metal Oxide-Hydride. Inorg. Chem. 2010, 49, 11062−11068. (24) Poulsen, F. W. Speculations on the existence of hydride ions in proton conducting oxides. Solid State Ionics 2001, 145, 387−397. (25) Rodriguez-Carvajal, J. Recent advances in magnetic structure determination by neutron powder diffraction. Phys. B 1993, 192, 55− 69. (26) Stephens, P. W. Phenomenological model of anisotropic peak broadening in powder diffraction. J. Appl. Crystallogr. 1999, 32, 281− 289. (27) Gonze, X.; Amadon, B.; Anglade, P. -M.; Beuken, J. -M.; Bottin, F.; Boulanger, P.; Bruneval, F.; Caliste, D.; Caracas, R.; Côté, M.; Deutsch, T.; Genovese, L.; Ghosez, Ph.; Giantomassi, M.; Goedecker, S.; Hamann, D. R.; Hermet, P.; Jollet, F.; Jomard, G.; Leroux, S.; Mancini, M.; Mazevet, S.; Oliveira, M. J. T.; Onida, G.; Pouillon, Y.; Rangel, T.; Rignanese, G. -M.; Sangalli, D.; Shaltaf, R.; Torrent, M.; Verstraete, M. J.; Zerah, G.; Zwanziger, J. W. . ABINIT: Firstprinciples approach to material and nanosystem properties. Comput. Phys. Commun. 2009, 180, 2582−2615. (28) Torrent, M.; Jollet, F.; Bottin, F.; Zerah, G.; Gonze, X. Implementation of the projector augmented-wave method in the ABINIT code: Application to the study of iron under pressure. Comput. Mater. Sci. 2008, 42, 337−351. (29) Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 50, 17953. (30) Kawamura, K.; Yashima, M.; Fujii, K.; Omoto, K.; Hibino, K.; Yamada, S.; Hester, J. R.; Avdeev, M.; Miao, P.; Torii, S.; Kamiyama, T. Structural Origin of the Anisotropic and Isotropic Thermal Expansion of K2NiF4-Type LaSrAlO4 and Sr2TiO4. Inorg. Chem. 2015, 54, 3896−3904. (31) Elcombe, M. M.; Kisi, E. H.; Hawkins, K. D.; White, T. J.; Goodman, P.; Matheson, S. Structure determinations for Ca3Ti2O7, Ca4Ti3O10, Ca3.6Sr0.4Ti3O10 and a refinement of Sr3Ti2O7. Acta Crystallogr., Sect. B: Struct. Sci. 1991, 47, 305−314. (32) Hungria, T.; MacLaren, I.; Fuess, H.; Galy, J.; Castro, A. HREM studies of intergrowths in Sr2[Srn−1TinO3n+1] Ruddlesden− Popper phases synthesized by mechanochemical activation. Mater. Lett. 2008, 62, 3095−3098. (33) Oswald, H. R.; Felder-Casagrande, S. Structure and properties of perovskite-related Sr-Ru-Ti-O phases. Solid State Ionics 1993, 63− 65, 565−569. (34) Pauling, L. THE PRINCIPLES DETERMINING THE STRUCTURE OF COMPLEX. J. Am. Chem. Soc. 1929, 51, 1010− 1026. (35) Clarke, S. J.; Hardstone, K. A.; Michie, C. W.; Rosseinsky, M. J. High-Temperature Synthesis and Structures of Perovskite and n = 1 Ruddlesden−Popper Tantalum Oxynitrides. Chem. Mater. 2002, 14, 2664−2669.

Hotehama, K.; Yamanaka, S.; Kyriakou, P.; Rovers, M.; Luke, G. M.; Kojima, K. M. Two-dimensional nature of superconductivity in the intercalated layered systems LixHfNCl and LixZrNCl: Muon spin relaxation and magnetization measurements. Phys. Rev. B: Condens. Matter Mater. Phys. 2004, 69, 134522. (3) Kamihara, Y.; Watanabe, T.; Hirano, M.; Hosono, H. Iron-Based Layered Superconductor La[O1‑xFx]FeAs (x = 0.05−0.12) with Tc = 26 K. J. Am. Chem. Soc. 2008, 130, 3296−3297. (4) Lin, C.-F.; Fan, X.; Pearse, A.; Liou, Sz-C.; Gregorczyk, K.; Leskes, M.; Wang, C.; Lee, S. B.; Rubloff, G. W.; Noked, M. Highly Reversible Conversion-Type FeOF Composite Electrode with Extended Lithium Insertion by Atomic Layer Deposition LiPON Protection. Chem. Mater. 2017, 29, 8780−8791. (5) Maeda, K.; Teramura, K.; Lu, D.; Takata, T.; Saito, N.; Inoue, Y.; Domen, K. Photocatalyst releasing hydrogen from water. Nature 2006, 440, 295. (6) Fuertes, A. Prediction of Anion Distributions Using Pauling’s Second Rule. Inorg. Chem. 2006, 45, 9640−9642. (7) Kobayashi, Y.; Hernandez, O.; Tassel, C.; Kageyama, H. New chemistry of transition metal oxyhydrides. Sci. Technol. Adv. Mater. 2017, 18, 905−918. (8) Kageyama, H.; Hayashi, K.; Maeda, K.; Attfield, J. P.; Hiroi, Z.; Rondinelli, J. M.; Poeppelmeier, K. R. Expanding frontiers in materials chemistry and physics with multiple anions. Nat. Commun. 2018, 9, 772. (9) Kobayashi, G.; Hinuma, Y.; Matsuoka, S.; Watanabe, A.; Iqbal, M.; Hirayama, M.; Yonemura, M.; Kamiyama, T.; Tanaka, I.; Kanno, R. Pure H− conduction in oxyhydrides. Science 2016, 351, 1314− 1317. (10) Bai, Q.; He, X.; Zhu, Y.; Mo, Y. First-Principles Study of Oxyhydride H− Ion Conductors: Toward Facile Anion Conduction in Oxide-Based Materials. ACS Appl. Energy Mater. 2018, 1, 1626−1634. (11) Fjellvåg, Ø S.; Armstrong, J.; Vajeeston, P.; Sjåstad, A. O. New Insights into Hydride Bonding, Dynamics, and Migration in La2LiHO3 Oxyhydride. J. Phys. Chem. Lett. 2018, 9, 353−358. (12) Masuda, N.; Kobayashi, Y.; Hernandez, O.; Bataille, T.; Paofai, S.; Suzuki, H.; Ritter, C.; Ichijo, N.; Noda, Y.; Takegoshi, K.; Tassel, C.; Yamamoto, T.; Kageyama, H. Hydride in BaTiO2.5H0.5: A Labile Ligand in Solid State Chemistry. J. Am. Chem. Soc. 2015, 137, 15315− 15321. (13) Yajima, T.; Takeiri, F.; Aidzu, K.; Akamatsu, H.; Fujita, K.; Yoshimune, W.; Ohkura, M.; Lei, S.; Gopalan, V.; Tanaka, K. M.; Brown, C.; Green, M. A.; Yamamoto, T.; Kobayashi, Y.; Kageyama, H. A labile hydride strategy for the synthesis of heavily nitridized BaTiO3. Nat. Chem. 2015, 7, 1017−1023. (14) Mikita, R.; Aharen, T.; Yamamoto, T.; Takeiri, F.; Ya, T.; Yoshimune, W.; Fujita, K.; Yoshida, S.; Tanaka, K.; Batuk, D.; Abakumov, A. M.; Brown, C. M.; Kobayashi, Y.; Kageyama, H. Topochemical Nitridation with Anion Vacancy-Assisted N3−/O2− Exchange. J. Am. Chem. Soc. 2016, 138, 3211−3217. (15) Kobayashi, Y.; Hernandez, O. J.; Sakaguchi, T.; Yajima, T.; Roisnel, T.; Tsujimoto, Y.; Morita, M.; Noda, Y.; Mogami, Y.; Kitada, A.; Ohkura, M.; Hosokawa, S.; Li, Z.; Hayashi, K.; Kusano, Y.; Kim, J.; Tsuji, N.; Fujiwara, A.; Matsushita, Y.; Yoshimura, K.; Takegoshi, K.; Inoue, M.; Takano, M.; Kageyama, H. An oxyhydride of BaTiO3 exhibiting hydride exchange and electronic conductivity. Nat. Mater. 2012, 11, 507−511. (16) Sakaguchi, T.; Kobayashi, Y.; Yajima, T.; Ohkura, M.; Tassel, C.; Takeiri, F.; Mitsuoka, S.; Ohkubo, H.; Yamamoto, T.; Kim, J.; Tsuji, N.; Fujihara, A.; Matsushita, Y.; Hester, J.; Avdeev, M.; Ohoyama, K.; Kageyama, H. Oxyhydrides of (Ca,Sr,Ba)TiO3 Perovskite Solid Solutions. Inorg. Chem. 2012, 51, 11371−11376. (17) Yamamoto, T.; Yoshii, R.; Bouilly, G.; Kobayashi, Y.; Fujita, K.; Kususe, Y.; Matsushita, Y.; Tanaka, K.; Kageyama, H. An Antiferro-toFerromagnetic Transition in EuTiO3−xHx Induced by Hydride Substitution. Inorg. Chem. 2015, 54, 1501−1507. (18) Amano Patino, M.; Zeng, D.; Blundell, S. J.; McGrady, J. E.; Hayward, M. A. Extreme Sensitivity of a Topochemical Reaction to I

DOI: 10.1021/acs.inorgchem.8b01645 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry (36) Kobayashi, Y.; Tian, M.; Eguchi, M.; Mallouk, T. E. IonExchangeable, Electronically Conducting Layered Perovskite Oxyfluorides. J. Am. Chem. Soc. 2009, 131, 9849−9855. (37) Saiful Islam, M. Ionic transport in ABO3 perovskite oxides: a computer modelling tour. J. Mater. Chem. 2000, 10, 1027−1038. (38) Patrakeev, M. V.; Bahteeva, J. A.; Mitberg, E. B.; Leonidov, I. A.; Kozhevnikov, V. L.; Poeppelmeier, K. R. Electron/hole and ion transport in La1‑xSrxFeO3‑δ. J. Solid State Chem. 2003, 172, 219−231. (39) Kharton, V. V.; Naumovich, E. N.; Vecher, A. A.; Nikolaev, A. V. Oxide Ion Conduction in Solid Solutions Ln1‑xSrxCoO3‑δ(Ln = La, Pr, Nd). J. Solid State Chem. 1995, 120, 128−136. (40) Shilova, Y. A.; Patrakeev, M. V.; Mitberg, E. B.; Leonidov, I. A.; Kozhevnikov, V. L.; Poeppelmeier, K. R. Order−Disorder Enhanced Oxygen Conductivity and Electron Transport in Ruddlesden−Popper Ferrite-Titanate Sr3Fe2−xTixO6+δ. J. Solid State Chem. 2002, 168, 275− 283. (41) Parfitt, D.; Chroneos, A.; Kilner, J. A.; Grimes, R. W. Molecular dynamics study of oxygen diffusion in Pr2NiO4+δ. Phys. Chem. Chem. Phys. 2010, 12, 6834−6836. (42) Chroneos, A.; Parfitt, D.; Kilner, J. A.; Grimes, R. W. Anisotropic oxygen diffusion in tetragonal La2NiO4+δ: molecular dynamics calculations. J. Mater. Chem. 2010, 20, 266−270. (43) Bang, J.; Matsuishi, S.; Hiraka, H.; Fujisaki, F.; Otomo, T.; Maki, S.; Yamaura, J.; Kumai, R.; Murakami, Y.; Hosono, H. Hydrogen Ordering and New Polymorph of Layered Perovskite Oxyhydrides: Sr2VO4−xH. J. Am. Chem. Soc. 2014, 136, 7221−7224. (44) Liu, X.; Bjørheim, T. S.; Haugsrud, R. Formation and migration of hydride ions in BaTiO3−xHx oxyhydride. J. Mater. Chem. A 2017, 5, 1050−1056. (45) Yajima, T.; Kitada, A.; Kobayashi, Y.; Sakaguchi, T.; Bouilly, G.; Kasahara, S.; Terashima, T.; Takano, M.; Kageyama, H. Epitaxial Thin Films of ATiO3−xHx (A = Ba, Sr, Ca) with Metallic Conductivity. J. Am. Chem. Soc. 2012, 134, 8782−8785.

J

DOI: 10.1021/acs.inorgchem.8b01645 Inorg. Chem. XXXX, XXX, XXX−XXX