Article pubs.acs.org/IC
Influence of Structural Disorder on Hollandites AxRu4O8 (A+ = K, Rb, Rb1−xNax) Geneva Laurita, Rosa Grajczyk, Matthew Stolt, Italo Coutinho, Arthur W. Sleight, and M. A. Subramanian* Department of Chemistry, Oregon State University, Corvallis, Oregon 97331, United States S Supporting Information *
ABSTRACT: Structural disorder can play an important role in the electrical properties of correlated materials. In this work we examine the average and local disorder in hollandites AxRu4O8 (A+ = K, Rb, Rb1−xNax) through neutron total scattering techniques. Samples with A+ = Rb, Rb1−xNax exhibit the largest amount of local disorder as evidenced by higher atomic displacement parameters, and as a result, a weakened temperature dependence of the resistivity is observed upon cooling as compared to KxRu4O8. All samples exhibit anisotropic resistivity that is dominated by metallic conductivity at lower temperatures, and this is corroborated by Pauli paramagnetic behavior throughout the measured temperature regime.
■
INTRODUCTION
this has a measurable effect on the temperature dependence of the electrical resistivity.
The hollandite crystal structure comprises a family of interesting compounds due to the one-dimensional nature of the structure. Materials with quasi-one-dimensional (Q1D) structures can exhibit novel properties such as charge density waves in K0.3MoO3, spin density waves in (TMTSF)2PF6, ballistic conduction in carbon nanotubes, and superconductivity in Nb3Se4.1−4 The idealized tetragonal hollandite structure of AB4O8 with symmetry I4/m (no. 87) consists of double chains of edgeshared BO6 octahedra (Figure 1a)5,6 which result in a onedimensional (1D) tunnel structure that can house a variety of A+ cations.7−10 The B-site and O atoms occupy variable 8h Wyckoff sites at (x, y, 0), while the A-site cation occupies the fixed 2b Wyckoff site at (0, 0, 0.5) central to the tunnel structure. The KRu4O8 hollandite is of interest as a strongly correlated Q1D oxide.7−10 This material has been termed a “clean” electrical conductor with a reported mean-free-path of 1300− 2800 Å in single crystals at low temperatures.8,10 In addition to the Q1D nature of the tunnel structure, there is a large interest in studying 4d and 5d metal oxide materials, which exhibit strong spin−orbit coupling. Oxides of ruthenium exhibit a wealth of interesting electronic and magnetic properties, such as spin-glass behavior in Ca2Ru2O7, superconductivity in Sr2RuO4, and simultaneous metallic conductivity and ferromagnetism in SrRuO3.11−14 Here we present the characterization of polycrystalline samples of ARu4O8 hollandites, where A+ = K, Rb, Rb1−xNax, through neutron total scattering techniques, four probe resistivity measurements, and magnetic susceptibility measurements. We find the A-site exhibits structural disorder in the form of large anisotropic atomic displacement parameters, and © XXXX American Chemical Society
■
METHODS
Polycrystalline samples of nominal KRu4O8 and RbRu4O8 were prepared by thoroughly grinding stoichiometric amounts of RuO2 and excess A2CO3 to account for the volatility of the alkali salts, where A+ = K (15% excess) and Rb+ (30% excess), in an agate mortar and pestle. The samples were then annealed under N2 to promote the formation of mixed valent Ru3+/4+ at 1123 K for 12 h in an alumina boat followed by an intermediate grinding and additional 12 h anneal under N2 at 1123 K to ensure formation of the hollandite phase. The first annealing treatment resulted in small needle-like crystals (Figure 1b), which were ground for subsequent heatings, molten salt ion exchange, and bulk characterization. Polycrystalline Rb1−xNaxRu4O8 was prepared through the molten salt ion exchange method. Approximately 0.2000 g of RbRu4O8 was combined with NaNO3, which acted as both the Na+ precursor and molten salt solvent, in a 1:5 metal to Na+ molar ratio, and the ion exchange was performed at 623 K for 12 h in a ceramic crucible. Samples were rinsed with deionized water to remove any exchanged or excess salts. Powder neutron diffraction experiments were performed on the Nanoscale Ordered Materials Diffractometer (NOMAD) at the Spallation Neutron Source (SNS) located at Oak Ridge National Laboratory (ORNL) with a collection time of approximately 2−4 h per sample.15 Samples were loaded in a quartz capillary, and an empty capillary was collected and subtracted as background. The pair distribution function (PDF), G(r), was obtained by the transformation of the normalized total scattering function, S(Q), according to the equation:
g (r ) − 1 =
1 2π rρΣb2 2
∫Q
Q max min
(S(Q ) − 1)Q sin(Qr ) dQ
(1)
Received: December 16, 2015
A
DOI: 10.1021/acs.inorgchem.5b02897 Inorg. Chem. XXXX, XXX, XXX−XXX
Inorganic Chemistry
Article
■
RESULTS AND DISCUSSION Structural Analysis. Rietveld refinements (Figure 2) on neutron diffraction data indicate all samples can be indexed to
Figure 2. Rietveld fits of the neutron diffraction data for samples (a) KxRu4O8, (b) RbxRu4O8, and (c) Rb1−xNaxRu4O8. All samples are indexed to space group I4/m, and refined lattice and atomic parameters are in good agreement with literature values.8
the hollandite structure with space group symmetry I4/m, and refined crystallographic parameters are summarized in Table 1. Table 1. Select Crystallographic Data (Space Group I4/m, no. 87) from the Rietveld Refinement of the Powder Neutron Diffraction Dataa refined composition
K0.8(1)Ru4O8
Rb0.79(3)Ru4O8
Rb0.61(3)Na0.39(3)Ru4O8
a (Å) c (Å) volume (Å 3) A U11,22 (Å2) A U33 (Å2) Ru Uiso (Å2) O1 Uiso (Å2) O2 Uiso (Å2) Ru x position Ru y position O1 x position O1 y position O2 x position O2 y position Rw (%)
9.8996(2) 3.1196(2) 05.73(11) 0.04(6) 0.07(2) 0.0067(2) 0.0087(9) 0.0087(7) 0.3511(3) 0.1676(4) 0.1505(5) 0.2004(4) 0.5445(4) 0.1587(5) 5.9
10.0023(2) 3.1175(2) 11.98(3) 0.032(4) 0.134(4) 0.0099(3) 0.0105(4) 0.0111(4) 0.3511(4) 0.1661(5) 0.1537(6) 0.2034(5) 0.5433(5) 0.1578(6) 6.9
9.9994(4) 3.1162(3) 311.658(9) 0.056(5) 0.14(2) 0.0091(4) 0.0070(4) 0.0117(5) 0.3512(4) 0.1658(5) 0.1541(6) 0.2019(5) 0.5447(6) 0.1587(7) 7.0
Figure 1. (a) Idealized hollandite structure AB4O8 (space group I4/m) consisting of corner-shared double chains of edge-shared BO6 octahedra. (b) One-dimensional nature of the tetragonal unit cell reflected in the long, needle-like crystals obtained from standard solid state routes under N2 (g) flow. where r is the peak position in Å, ρ is the number density in atoms per Å3, b is the coherent neutron scattering length of each atom in barns, Q is the magnitude of the scattering vector in inverse angstroms, G(r) = [g(r) − 1], Qmin = 0.5 Å, and Qmax = 31.5 Å−1. Rietveld refinements were performed on the reciprocal-space diffraction data to analyze the average structure using GSAS software with the interface EXPGUI.16,17 The local structure was investigated via analysis of the real-space PDF using the PDFgui software suite.18 Correlated motion in the PDF was modeled using the δ2 parameter, which accounts for a 1/r2 dependence of peak sharpening. Crystal structures were visualized using the VESTA suite of programs.19 Four-probe resistivity and magnetic susceptibility measurements were collected on bulk polycrystalline samples upon warming with a Quantum Design Physical Properties Measurement System (QD PPMS) from 5 to 300 K. Samples were prepared for resistivity measurements by pressing into bars and sintering at 873 K under N2 flow for 12 h. Ag electrodes were painted onto the samples, and Cu wires were used to perform the measurement. Samples were packed into a capsule with cotton and loaded in a sample straw, and zero-fieldcooled (ZFC) DC magnetization data were collected using the ACMS mode with a magnetic field of 0.5 T. A correction was applied to the susceptibility data to account for diamagnetic contributions from the straw and cotton.
A-site ADP cross-terms are fixed by symmetry so that U12 = U13 = U23 = 0. a
Lattice parameters for KxRu4O8 are in good agreement with the literature.8 Larger a lattice parameters and unit cell volumes are observed for Rb-containing samples, and this follows with the expected trend for the ionic radii of K+ (1.51 Å for CN 8) as compared to Rb+ (1.61 Å for CN 8).20 For samples of composition KxRu4O8 and RbxRu4O8 it was found that the Asite was not fully occupied, but upon Na+-ion exchange with RbxRu4O8, full A-site occupancy was realized. Na+ exchange with the Rb-hollandite resulted in a Na+ content of 40%, and B
DOI: 10.1021/acs.inorgchem.5b02897 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry this is the first report of Na substitution into the RbRu4O8 hollandite. Due to the large channel in the hollandite structure, atomic displacement parameters (ADPs, reported in Uiso) were modeled anisotropically for the A-site cations. Refinements indicated a displacement of all A-site cations parallel to the RuO6 tunnel structure, illustrated for each sample in Figure 3.
refinements to the values obtained from the Rietveld analysis. Refinements of the local structures over a r-range of 1.5−10 Å against neutron PDFs (Figure 4) indicate smaller ADPs than
Figure 3. Refined average structures of samples AxRu4O8 showing the anisotropic nature of the displacements of the A-site cations along the Ru−O tunnel structure.
Figure 4. Fits of the neutron PDFs for nominal samples (a, b) KRu4O8, (c, d) RbRu4O8, and (e, f) Rb1−xNaxRu4O8 over an r-range of 1.5−10 Å. The local bonding is consistent with I4/m symmetry. However, a poorer fit is observed in the range less than 3 Å in RbRu4O8 and Rb1−xNaxRu4O8, indicating the local, first coordination disorder in these samples may not be not fully captured with the I4/m description.
As the inclusion of an additional parameter to model anisotropic ADPs only slightly improved the fits (Rw = 5.9 vs 5.8 for KxRu4O8, Rw = 7.1 vs 6.9 for RbxRu4O8, and Rw = 7.1 vs 7.0 for Rb1−xNaxRu4O8), a Hamilton R-ratio test was employed to determine the significance of these results. Due to the large number of data points (2796 for NOMAD bank 5), it was determined that the additional parameter modeling anisotropic atomic displacement was indeed statistically significant at
