Structure and oxide ion conductivity in Bi₂₈Re₂O₄₉, a new bismuth rhenium oxide containing tetrahedral and octahedral Re(VII)

Abstract

A new Bi–Re–O phase, Bi₂₈Re₂O₄₉, has been synthesized and characterized. Its structure, determined by neutron powder diffraction, is a superstructure of the cubic fluorite unit cell: tetragonal, I4/m, a = 8.7216(1) Å, c = 17.4177(2) Å. The structure comprises an ordered framework of linked BiO₄e trigonal bipyramids and square pyramids (e = lone pair of electrons), with discrete Re oxoanions at the origin and body-centre of the unit cell. The infra-red spectrum and Re K-edge X-ray absorption spectrum were consistent with the presence of tetrahedral ReO₄⁻ and octahedral ReO₆⁵⁻ species in the ratio of 3 : 1. This correctly provides charge balance in the overall structure. The phase is found to display high oxide ion conductivity, which may relate to the presence of the two different oxoanion species, and possible migration of O²⁻ ions between them.

---

Introduction

The substitution of small amounts of Bi in Bi₂O₃ by tetrahedrally coordinated M(VI) ions can lead to stable superstructures of the cubic fluorite unit cell. Tetragonal (I4/m; a ∼ 8.7 Å, c ∼ 17.3 Å) or monoclinically distorted variants have been reported for a Bi : M ratio of 14 : 1 with M = S, Cr, Mo and W.^1–4 The tetragonal supercell parameters are related to the fluorite unit cell size, a_f, by a = (√10/2)a_f, c = 3a_f, giving a unit cell contents of 28Bi, 2M and 48O atoms and an overall formula Bi₁₄SO₂₄ or Bi₁₄O₂₀SO₄. The basic structure comprises tetrahedral MO₄²⁻ ions at the origin and body-centre of the unit cell; the MO₄ units exhibit a degree of orientational disorder which appears to vary with M and temperature, and is linked to the transition from tetragonal (I4/m) to monoclinic (C2/m) symmetry observed for M = Mo,W.⁴ In an investigation of the possible co-substitution of M(VI) ions by M′(V) and M″(VII) species, we focussed on M″ = Re, since Bi₃ReO₈ has been reported and contains discrete ReO₄⁻ anions.⁵ However, to our surprise, the same superstructure was observed for Bi : Re ratios of 14 : 1, in the absence of M′(V) ions. Here we report the chemical nature and structure of this phase along with oxide ion conductivity data.

Experimental

A synthetic route was used which was similar to that used for Bi₁₄SO₂₄: stoichiometric amounts of Bi₂O₃ and NH₄ReO₄ were intimately ground to give a Bi : Re ratio of 14 : 1, and the mixture was heated twice (with a re-grind between firings) at 800 °C for 12 h in air. X-Ray powder diffraction (XPD) data were collected on a Siemens D5000 (Ge primary beam monochromator giving Cu-Kα₁, transmission mode, 8° PSD). Re-K-edge X-ray absorption spectra were recorded in transmission mode on Station 9.2 at the SRS, Daresbury Laboratory. A Si(220) double crystal monochromator was used, which was detuned by 40% to remove higher order harmonics. Analysis was performed as previously described⁶ using the SRS programs EXCALIB and EXBROOK for data reduction and EXCURV98⁷ for data simulation. Time-of-flight neutron powder diffraction (NPD) data were collected at 298 K using the ultra high resolution back-scattering detectors on the HRPD diffractometer at ISIS, Rutherford Appleton Laboratory. The program GSAS⁸ was employed for structure refinement. A Shimadzu 8300 spectrophotometer was used to collect FTIR data. Magnetic measurements were made on a Quantum Design PPMS using AC susceptibility (10 Oersted) and DC magnetization in a field of 1000 Oersted. Pellets for conductivity measurements were prepared in a 13 mm diameter die at 7500 kg cm⁻³. After sintering at 750 °C for 12 h, pellet densities were 85–90% of theoretical. Silver paste was used to provide contact with the faces of the pellets and the conductivity was measured in air using a Hewlett Packard 4800A vector impedance meter with frequencies 100–5 × 10⁵ Hz in the temperature range 400–600 °C.

Results and discussion

The synthesized Bi₁₄ReO_x was bright yellow in colour, in marked contrast to the pale cream observed for substituted bismuth oxides in which the substituent has no associated colour, e.g. Bi₁₄SO₂₄.¹ Tetrahedral ReO₄⁻ ions show no absorption in the visible region, and Bi₃ReO₈, with isolated ReO₄⁻ ions is, as expected, colourless.⁵ On the other hand, Re(VII) with octahedral (Li₅ReO₆, Na₅ReO₆)^9,10 or trigonal bipyramidal (Na₃ReO₅)¹¹ coordination by oxygen shows a strong yellow/orange colouration owing to charge-transfer bands. Infra-red spectroscopy, Fig. 1, revealed bands in the vicinity of 900 cm⁻¹, characteristic of tetrahedral ReO₄⁻, and 600–700 cm⁻¹, which appears to be typical of octahedral ReO₆⁵⁻ species.⁵ Splitting of the band at ∼900 cm⁻¹ suggests that the ReO₄⁻ ions are significantly distorted from regular tetrahedral symmetry.⁵

Fig. 1 FTIR spectrum of Bi₂₈Re₂O₄₉.

XPD suggested that the product was single phase with the tetragonal (I4/m) superstructure exhibited by Bi₁₄SO₂₄¹ and Bi₁₄CrO₂₄.^3,4 The synthesis conditions require the formation of Re(VII) rather than lower valent ions, which suggests a formula of Bi₁₄ReO_24.5 and a unit cell content of Bi₂₈Re₂O₄₉. Evidence from infra-red spectroscopy suggests the likelihood that 75% of the Re ions are present as ReO₄⁻ ions, and 25% as ReO₆⁵⁻ ions. The absence of reduced Re species was confirmed by magnetization measurements, which demonstrated that the phase is diamagnetic.

High resolution NPD data were collected at ambient temperature to explore the structural details of Bi₂₈Re₂O₄₉. As a result of the orientational disorder of the tetrahedral groups in the related Bi₁₄MO₂₄ compounds,⁴ and the consequent difficulty of locating the O atoms within these groups, it was considered that realistic information concerning the orientation of the mixed oxoanions in Bi₂₈Re₂O₄₉ would not be attainable even at low temperatures. Structure refinement therefore focussed on the basic Bi–O structural framework at 298 K. Using the room temperature structure of Bi₁₄CrO₂₄ (I4/m)⁴ as a starting model, refinement proceeded in a straightforward manner with the exclusion of the O atoms bonded to Re. As expected, subsequent attempts to locate these atoms proved difficult. However, given that 75% of the Re was thought to be present as ReO₄⁻, a successful refinement was achieved by making the approximation that only tetrahedral species were in the structure. On this basis, the final refinement confirmed the structure, and provided a bismuth oxide framework which is essentially the same as that found in the related Bi₁₄MO₂₄ compounds. Although less information can be obtained for the Re oxoanions, it appears that the ReO₄⁻ ions present have an orientation very similar to that reported previously for Bi₁₄MO₂₄, M = Cr, Mo, W.⁴ As observed for these phases, the Re is displaced off the origin along [001], by ∼0.17 Å. The fitted NPD profiles are shown in Fig. 2 and the refined atomic positions and other structural information are provided in Table 1. Even after allowing displacement of Re, its isotropic temperature factor remained rather high, as is the value for O5 which is bonded to it. It is likely that these values reflect disorder associated with the presence of both ReO₄⁻ and ReO₆⁵⁻ ions in the structure, but attempts at splitting the Re site further, to allow for these species, did not improve the refinement. Fig. 3 shows the structural framework, highlighting the presence of the Bi2O₄ distorted tetrahedra (trigonal bipyramids if we include the stereochemical lone-pair of electrons [e], Bi2O₄e) and square pyramids (Bi1O₄e and Bi3O₄e with e at the pyramid apex). Owing to the orientational disorder of the ReO₄⁻ units (and the fact that the ReO₆⁵⁻ units have not been incorporated), the O atoms bonded to Re (O4, O5, O6 and O7) have not been included in Fig. 3. Selected bond distances and angles are given in Table 2.

Fig. 2 Observed (+), calculated and difference NPD profiles of Bi₂₈Re₂O₄₉ at 298 K. The reflection positions are marked (|).

Fig. 3 The structure of Bi₂₈Re₂O₄₉ showing the BiO₄e trigonal bipyramids and the Bi–O bonds in the BiO₄e square pyramids. The O atoms bonded to Re are not shown.

Table 1 Refined atomic parameters for Bi₂₈Re₂O₄₉ from NPD, 298 K

Atom	Position	x	y	z	U iso × 100/Å^2	Cell occupancy
a I4/m; a = 8.7216(1) Å, c = 17.4177(2) Å; Rp = 0.0465, Rwp = 0.0548, Rexp = 0.0216.
Re	4e	0	0	0.0100(6)	4(2)	2
Bi1	8h	0.2107(3)	0.4491(2)	0	1.82(5)	8
Bi2	16i	0.2998(2)	0.1056(2)	0.1711(8)	1.66(3)	16
Bi3	4e	0.5	0.5	0.1557(2)	2.33(9)	4
O1	8g	0.5	0	0.1259(2)	2.66(6)	8
O2	16i	0.0739(2)	0.2517(3)	0.2567(1)	2.37(5)	16
O3	16i	0.3253(3)	0.6302(3)	0.0780(1)	2.16(7)	16
O4	16i	0.020(2)	0.232(2)	0.015(2)	0.3(2)	2
O5	4e	0	0	0.1126(6)	4.1(3)	2
O6	16i	0.015(3)	0.179(3)	0.036(2)	2.3(2)	2
O7	16i	0.995(2)	0.206(3)	0.012(3)	2.3(2)	2

---

Table 2 Selected bond distances (Å) and angles (°) for Bi₂₈Re₂O₄₉

Re–O4	2.03(2)	Bi2–O1	2.131(2)
Re–O5	1.79(2)	Bi2–O3	2.584(2)
Re–O6	1.76(3)	Bi2–O2	2.086(3)
Re–O7	1.84(3)	Bi2–O2	2.202(3)
O4–Re–O5	101.9(9)	O1–Bi2–O2	96.0(1)
O4–Re–O6	106.2(9)		93.0(1)
O4–Re–O7	97(2)	O1–Bi2–O3	83.7(1)
O5–Re–O6	106.2(9)	O2–Bi2–O2	96.7(1)
O5–Re–O7	102(1)	O2–Bi2–O3	79.1(1)
O6–Re–O7	99(1)		175.8(1)
Bi1–O3	2.221(3) [×2]	Bi3–O3	2.334(3) [×4]
	2.311(3) [×2]
O3–Bi1–O3	72.0(1)	O3–Bi3–O3	70.31(9) [×4]
	72.7(1) [×2]		109.0(2) [×2]
	75.4(1)
	114.9(2) [×2]

---

The Bi/O framework is seen to be essentially the same as that previously reported for the M(VI) analogous materials⁴ and consists of linked BiO₄ polyhedra with stereochemically active lone pairs of electrons. The Re–O bonds to O5,6,7 are typical of those in tetrahedral ReO₄⁻ ions (e.g. average 1.74 Å in Bi₃ReO₈⁵), whereas the Re–O4 bond is significantly longer and is similar to distances observed for octahedral ReO₆⁵⁻ species, e.g. 1.89 Å in Li₅ReO₆ and Na₅ReO₆,⁸ and La₃ReO₈.^12,13 The coordination around Bi is typical for BiO₄e arrangements and the Bi–O distances are very similar to those observed in related structures: bonds of 2.2–2.3 Å for the square pyramidal Bi1O₄e and Bi3O₄e with the trigonal bipyramidal Bi2O₄e having the equatorial bonds (2.131(2) Å and 2.086(3) Å) significantly shorter than the axial bonds (2.584(2) Å and 2.202(3) Å).

Given the difficulties in locating O atoms bonded to Re in this phase, owing to orientational disorder and the presumed presence of both ReO₄⁻ and ReO₆⁵⁻ ions, EXAFS data were collected and compared with reference materials KReO₄ (tetrahedral Re) and Li₅ReO₆ (octahedral Re). The coordination around Re was modelled as a single species and the calculations revealed two different Re–O distances: 1.74(2) Å [3.1(5) bonds per Re, Debye–Waller factor 0.004(1) Ų] and 2.15(2) Å [1.5(2) bonds per Re, Debye–Waller factor 0.002(1) Ų] and a fit index, R = 32.6%.⁶Fig. 4 shows the experimental and calculated EXAFS k³χ(k) data and their Fourier transform. Although the longer bond distance of 2.15 Å is slightly longer than expected for octahedrally coordinated Re(VII), the 1.74 Å bond is highly typical of tetrahedral ReO₄⁻ ions. The EXAFS data therefore provide good support for the structural model involving 75% ReO₄⁻ and 25% ReO₆⁵⁻. Based on a single Re species, this mixture of ions would result in an average Re coordination of 4 × 0.75 short (tetrahedral) bonds and 6 × 0.25 long (octahedral) bonds, i.e. 3 short and 1.5 long bonds, in excellent agreement with the calculated local model.

Fig. 4 Experimental (solid line) and calculated (- - - ) Re K-edge EXAFS data for Bi₂₈Re₂O₄₉: (a) k³ weighted EXAFS, (b) phase shifted Fourier transform of k³ weighted EXAFS.

Reasonably high oxide ion conductivity has recently been reported in a bismuth oxide phase containing tetrahedral sulfate groups: Bi₈SO₁₅ or BiO₁₁(SO₄).¹⁴ In this material, a high level of disorder was apparent on the oxygen sublattice, and the conductivity in the range 400–600 °C was determined to be within an order of magnitude of that observed for the Y-stabilised δ-Bi₂O₃ phase, (Bi_0.75Y_0.25)₂O₃ or Bi₃YO₆. In Bi₂₈Re₂O₄₉, although the bismuth oxide framework has no apparent disorder, the presence of both ReO₄⁻ and ReO₆⁵⁻ ions in a random fashion suggests that a conduction path may exist for the support of oxide ion migration. Conductivity measurements were therefore made between 400 and 600 °C. The complex plane impedance plots could be fitted to a single semicircle, with a Warburg impedance at low frequency that was attributed to electrode/electrolyte interfacial effects. The response was similar to that observed by us for well characterised stabilised Bi₂O₃ oxide ion conductors (e.g. Bi₃YO₆), and the resistance was assigned to bulk effects. It was assumed that under the oxidising conditions used, the electronic contribution to the conductivity was negligible, as has been found for other oxide ion conductors based on Bi₂O₃.¹⁵Fig. 5 shows plots of log (σT/S cm⁻¹ K) versus 1000/T (K⁻¹) for Bi₃YO₆¹⁶ and Bi₂₈Re₂O₄₉, from which the Arrhenius activation energies, E_a, were calculated. Bi₂₈Re₂O₄₉ produced a linear plot with E_a = 0.62 eV, which is slightly lower than the reported high temperature value for Bi₃YO₆ (E_a = 0.66 eV).¹⁶ For the temperature range studied, Bi₂₈Re₂O₄₉ exhibits conductivity very similar to that observed for the disordered phase Bi₈SO₁₅, and is higher at low temperature: at 400 °C, the conductivities are 5.4 × 10⁻⁴ S cm⁻¹ and 4.9 × 10⁻⁴ S cm⁻¹, respectively, in comparison with 2.0 × 10⁻³ S cm⁻¹ for Bi₃YO₆.¹⁶ The high conductivity, despite the high level of order within a large part of the structural framework, suggests that materials containing both tetrahedral ReO₄⁻ and octahedral ReO₆⁵⁻ units may provide interesting future targets for highly conducting solids.

Fig. 5 Plots of log (σT) versus 1000/T for Bi₂₈Re₂O₄₉ compared with Bi₃YO₆.¹⁶ The lines are linear fits to the data.

Conclusions

This study has identified a new phase with a structure closely related to that previously reported for Bi₁₄MO₂₄^1–4 materials, where M forms discrete MO₄²⁻ tetrahedral groups. When Re is the substituting cation, an interesting charge balancing mechanism is found in which 25% of the ReO₄⁻ ions are replaced by ReO₆⁵⁻ ions, and the overall composition becomes Bi₁₄ReO_24.5 or Bi₂₈Re₂O₄₉. Although NPD was able to confirm the basic structural properties, the precise oxygen arrangement around the Re ions could not be established owing to orientational disorder of the Re oxoanions, and the small concentration of the octahedral species. EXAFS data, however, provided strong support for the presence of two distinct Re–O bond distances, and the ratio of the number of bonds agreed well with that expected for the proposed ratio of ReO₄⁻ and ReO₆⁵⁻ ions. The phase exhibits high oxide ion conductivity which may be supported by a mechanism involving O²⁻ transport from ReO₆⁵⁻ to ReO₄⁻.

Acknowledgements

We thank EPSRC for the provision of financial support, and CCLRC for the provision of neutron diffraction and synchrotron facilities. We are grateful to Dr K. S. Knight for experimental assistance with the collection of NPD data.

References

1. M. G. Francesconi, A. L. Kirbyshire, C. Greaves, O. Richard and G. Van Tendeloo, Chem. Mater., 1998, 10, 626.
2. C. D. Ling, R. L. Withers, J. G. Thompson and S. Schmid, Acta Crystallogr., Sect. B, 1999, 55, 306.
3. S. A. Warda, W. Pietzuch, W. Massa, U. Kesper and D. Reinen, J. Solid State Chem., 2000, 149, 209.
4. T. E. Crumpton, M. G. Francesconi and C. Greaves, J. Solid State Chem., 2003, 175, 197.
5. A. K. Cheetham and A. R. R. Smith, Acta Crystallogr., Sect. B, 1985, 41, 225.
6. O. Helgason, J.-M. Grenache, F. J. Berry, S. Morup and F. Mosselmans, J. Phys.: Condens. Matter, 2001, 13, 10785.
7. N. Binsted, J. W. Campbell, S. J. Gurman, I. Ross and P. C. Stephenson, EXCURV98, CCLRC Daresbury Laboratory Computer Program.
8. R. B. Von Dreele, A. C. Larson and M. Lucan, GSAS, Neutron Scattering Centre, LANL, Los Alamos, NM 87545, 1995.
9. T. Betz and R. Hoppe, Z. Anorg. Allg. Chem., 1984, 512, 19.
10. L. R. Morss, E. H. Appelman, R. R. Gerz and D. Martin-Rovet, J. Alloys Compd., 1994, 203, 289.
11. E. Vielhaber and R. Hoppe, Z. Anorg. Allg. Chem., 1992, 610, 7.
12. G. Baud, J. Besse, R. Chevalier and M. Gasperon, J. Solid State Chem., 1979, 29, 267.
13. A. R. R. Smith, A. K. Cheetham and H. Fuess, Z. Anorg. Allg. Chem., 1984, 510, 46.
14. T. E. Crumpton and C. Greaves, J. Mater. Chem., 2004, 14, 2433.
15. P. Shuk, H.-D. Wiemhöfer, U. Guth, W. Göpel and M. Greenblatt, Solid State Ionics, 1996, 89, 179.
16. T. Takahashi and H. Iwahara, Mater. Res. Bull., 1978, 13, 1447.

---