The crystal structure and electrical properties of the oxide ion conductor Ba₃WNbO_8.5

The structural and electrical properties of the hexagonal perovskite derivative Ba₃WNbO_8.5 have been investigated. Ba₃WNbO_8.5 crystallises in a hybrid of the 9R hexagonal perovskite and palmierite structure as recently reported for the novel oxide ion conductor Ba₃MoNbO_8.5. Ba₃WNbO_8.5 is also an oxide ion conductor and appears to exhibit oxide ionic conduction with negligible electronic conductivity over a wider pO₂ range than Ba₃MoNbO_8.5. A neutron diffraction study has revealed that at 20 °C the average structure of Ba₃WNbO_8.5 contains just 13% of W(1)/Nb(1)O₄ tetrahedra within the average structure of Ba₃WNbO_8.5 in comparison to 50% of Mo(1)/Nb(1)O₄ tetrahedra in Ba₃MoNbO_8.5. The presence of (M/Nb)O₄ tetrahedra with non-bridging apical oxygen atoms is an important prerequisite for the ionic conduction observed in the Ba₃MNbO_8.5 system (M = W, Mo). The strong reduction in the ratio of (M/Nb)O₄ tetrahedra to (M/Nb)O₆ octahedra upon replacement of W⁶⁺ for Mo⁶⁺ results in a reduction in the ionic conductivity by an order of magnitude at 450 °C. The bulk conductivities converge upon heating so that at 600 °C the bulk conductivity of Ba₃WNbO_8.5, σ_b = 0.0017 S cm⁻¹, is comparable to that of Ba₃MoNbO_8.5 (σ_b = 0.0022 S cm⁻¹). The results demonstrate that other members of the Ba₃MM′O_8.5 family can support oxide ion conductivity and further studies of hexagonal perovskite derivatives are warranted.

---

Introduction

There has been increasing interest in oxide ion conductors in recent years due to their application in oxygen sensors, oxygen separation membranes and as electrolytes in solid oxide fuel cells (SOFCs).^1–4 The development of SOFCs as alternative systems for the generation of electric power is important due to their advantages over conventional methods of power generation, including significantly higher efficiencies of energy conversion, achieving efficiencies of greater than 85% for combined heat and power (CHP) applications.^5,6 In addition, SOFCs have the advantage of fuel flexibility, are extremely reliable and are more environmentally friendly, producing reduced emissions of pollutants such as CO₂ and NO_x gases.⁷

A variety of materials that exhibit oxide ion conductivity have been reported, including fluorite-related structures,^4,8 La₂Mo₂O₉ (LAMOX) based materials,⁹ Bi₂O₃ and Bi₄V₂O₁₁ based oxides,¹⁰ germanate apatite oxide systems,¹¹ perovskites such as strontium and magnesium doped lanthanum gallates (LSGM),¹² Na_0.5Bi_0.5TiO₃ (NBT)¹³ and the perovskite derivative NdBaInO₄.¹⁴ The major challenge in the commercialisation of SOFC technology has been the high temperature of operation (800–1000 °C) required to achieve sufficient conductivities from conventional zirconia-based electrolytes.¹⁵ The use of such high temperatures leads to slow start-up times, thermal degradation over time and necessitates the use of expensive materials.¹⁶ Reducing the operating temperature of the SOFC to an intermediate range (400–600 °C) would reduce costs, shorten start-up times, widen the choice of materials used in the fuel cell and extend the lifetime of the SOFC.¹⁷ As a result there is increasing interest in the development of electrolyte materials with sufficiently high oxide ion conductivities at lower temperatures. It is therefore important to discover and characterise new classes of oxide ion conducting materials.

Recently we have reported a novel oxide ion conductor, Ba₃MoNbO_8.5.¹⁸ Ba₃MoNbO_8.5 is the first hexagonal perovskite derivative reported to exhibit significant oxide ion conductivity and has a bulk conductivity of 2.2 × 10⁻³ S cm⁻¹ at 600 °C, comparable to that of conventional SOFC electrolyte materials. Ba₃MoNbO_8.5 exhibits oxygen transport numbers of 0.99 in air/O₂ and 0.92 in air/5% H₂ in Ar at 600 °C, suggesting that Ba₃MoNbO_8.5 is an oxide ion conductor with negligible electronic conductivity in air/O₂ and that a small amount of electronic conduction is observed in air/5% H₂ in Ar.¹⁸ Ba₃MoNbO_8.5 crystallises in a hybrid of the 9R hexagonal perovskite and palmierite structure. The crystal structure contains Mo/NbO_x in a mixture of octahedral and tetrahedral coordination geometry and has intrinsic oxygen vacancies.¹⁸ Ba₃MoNbO_8.5 presents an unusual structural rearrangement upon increasing the temperature which is caused by modification of the oxygen/vacancy distribution on the pseudo-cubic layers.¹⁹ The change in the O(2) and O(3) site populations leads to the concomitant increase of the ratio of (Mo/Nb)O₄ tetrahedra to (Mo/Nb)O₆ octahedra and enhances the oxide ionic conductivity.

Here, the crystal structure and electrical properties of the novel tungsten derivative Ba₃WNbO_8.5 are reported. The material Ba₃WNbO_8.5 also crystallises in a hybrid structure which is intermediate between that of the 9R-polytype and palmierite structures and exhibits oxide ionic conduction with negligible electronic conductivity over a wider pO₂ range than Ba₃MoNbO_8.5.

Experimental

Ba₃WNbO_8.5 was prepared by the solid-state reaction of stoichiometric amounts of the starting materials BaCO₃ (99.98%, Sigma-Aldrich), WO₃ (99.9%, Sigma-Aldrich) and Nb₂O₅ (99.99%, Sigma-Aldrich). The starting materials were ground, pressed into a pellet and calcined in an alumina crucible for 10 h at 900 °C. The pellet was then reground, pelleted and heated at 1300 °C for 10 hours before being cooled to room temperature at a rate of 5 °C min⁻¹. The latter heating step was repeated until a phase pure product was obtained.

X-ray powder diffraction patterns were collected at room temperature on a PANalytical Empyrean Powder diffractometer with Cu K_α1 radiation. Data were recorded in the range 5° < 2θ < 70°, with a step size of 0.0131°. Time of flight (TOF) neutron powder diffraction data were collected at room temperature (∼20 °C) on the Polaris diffractometer at the ISIS pulsed neutron source, Rutherford Appleton Laboratory, UK. ∼5 g of Ba₃WNbO_8.5 were loaded into an 8 mm diameter cylindrical vanadium can and data acquired in all 5 detector banks (covering a continuous d-spacing range from ∼0.2 Å < d < 25 Å) for ∼300 μA h integrated proton beam to the ISIS target (corresponding to ∼2 hours exposure). The morphology of the sintered pellets was monitored using a scanning electron microscope (SEM) (Carl Zeiss Gemini 300). The sample was covered with a thin layer of sputtered gold.

A pellet of ∼10 mm diameter was prepared from a powdered sample of Ba₃WNbO_8.5 and sintered for 10 hours at 1300 °C to achieve ∼92% of the theoretical density. Pt electrodes were painted on both sides of the pellet using a Pt-paste (Metalor 6082). Impedance spectroscopy measurements were recorded in the frequency range 0.1 Hz to 1 MHz with a Solartron 1260 impedance analyser. Data were recorded upon cooling from 700 °C to 450 °C in a sealed tube furnace under a dry flow of air, nitrogen, oxygen and 5% hydrogen/nitrogen. The impedance data were measured every 15 °C with 2 hours of equilibration time at each temperature. The geometrical factor (the ratio of surface area to thickness of the pellet) was used to correct the data obtained and the data was treated with the ZView software.

Results and discussion

Preliminary characterisation

The SEM micrographs (Fig. S1†) show that the sample exhibits a grain size of ∼5 to 30 μm. No secondary phases can be detected from backscattered micrographs thus confirming the purity of the synthesised sample. The X-ray diffraction pattern of Ba₃WNbO_8.5 could be indexed with the space group R3̄m H (giving a = 5.8551(2) Å and c = 20.9884(5) Å). This space group has been previously reported for Ba₃W_1.33Nb_0.66O_8.66.²⁰ There was no evidence of impurities.

X-ray diffraction data were also recorded after Ba₃WNbO_8.5 was annealed at 600 °C for 24 hours in flowing O₂, N₂ and 5% H₂/N₂. The X-ray diffraction data show no evidence of impurities or change in crystal structure post annealing. The X-ray diffraction patterns of a sample of Ba₃WNbO_8.5 annealed at 600 °C for 24 hours in flowing air and 5% H₂/N₂ is displayed in Fig. 1. This is similar to the Ba₃MoNbO_8.5 compound which also demonstrates a wide stability range under reducing conditions.

Fig. 1 X-ray diffraction patterns of the “as prepared” Ba₃WNbO_8.5 (black line) and after 24 hours annealing in 5% H₂ in N₂ at 600 °C (blue line).

Electrical properties

A typical impedance spectrum of Ba₃MoNbO_8.5 recorded in dry air at 600 °C is presented in Fig. 2. An electrode response is observed in the low frequency region (<10 Hz) at all temperatures above 500 °C, indicative of ionic conduction in a material with partially blocking electrodes.²¹ The electrode response was found to become less resistive when the oxygen content of the atmosphere increased, strongly indicating oxide ion mobility.

Fig. 2 Complex impedance plot recorded in dry air at 625.5 °C. An electrode response is observed in the low frequency region. The inset shows the complex impedance plot recorded in dry air at 450 °C. The numbers and corresponding filled circles indicate selected frequencies (in Hz), while the red line is the equivalent circuit fitting.

The bulk and grain boundary responses overlap at higher frequency. Equivalent circuit analysis was performed to extract the individual bulk, grain boundary and electrode responses at each temperature (the model used is displayed in Fig. S2†). The electrode response could be fit with a short Warburg element indicating finite length diffusion. The bulk and grain boundary responses have respective capacitance values of C_b ∼ 6.5 pF cm⁻¹ and C_gb ∼ 0.18 nF cm⁻¹.

The Arrhenius plot of the bulk and grain boundary conductivities of Ba₃WNbO_8.5 and Ba₃MoNbO_8.5 are presented in Fig. 3(a). For Ba₃WNbO_8.5 the grain boundaries constitute the most resistive part, thus dominating the total resistivity of the material. The grain boundary conductivity is more than one order of magnitude lower than the bulk conductivity. This could be a result of the grain boundary having different fractional occupancies of O(2) and O(3) or different oxygen stoichiometry than the bulk. The ionic conductivity of Ba₃MoNbO_8.5 is known to be sensitive to small changes in the fractional occupancies of O(2) and O(3).¹⁹ The grain boundary conductivity of Ba₃WNbO_8.5 is 7.5 × 10⁻⁵ S cm⁻¹ at 600 °C with an activation energy of 1.48 ± 0.015 eV. The bulk conductivity is 0.0017 S cm⁻¹ at 600 °C. A change in slope is observed at temperatures >525 °C, with the bulk activation energy lowering from 1.94 eV to 1.40 eV. A similar change in slope has previously been reported for Ba₃MoNbO₈.^18,19 At 450 °C the bulk conductivity of Ba₃WNbO_8.5 is approximately an order of magnitude lower than that of Ba₃MoNbO_8.5. Upon heating the bulk conductivities of Ba₃WNbO_8.5 and Ba₃MoNbO_8.5 converge (σ_b = 0.0017 S cm⁻¹ and 0.0022 S cm⁻¹ for Ba₃WNbO_8.5 and Ba₃MoNbO_8.5 respectively).

Fig. 3 (a) Arrhenius plots of the bulk (σ_b) and grain boundary (σ_gb) conductivities of Ba₃WNbO_8.5 and Ba₃MoNbO_8.5. (b) Arrhenius plots of the total conductivities of Ba₃WNbO_8.5 in a range of different atmospheres. Linear fits to the data are shown.

Fig. 3(b) displays the Arrhenius plot of the total conductivity of Ba₃WNbO_8.5 in a range of different atmospheres. The results suggest that Ba₃WNbO_8.5 exhibits negligible electronic conductivity over the pO₂ range measured. This is in stark contrast to Ba₃MoNbO_8.5, where an electronic component to the bulk and grain boundary conductivity is seen in 5% H₂/N₂.¹⁸

Crystal structure of Ba₃WNbO_8.5

W⁶⁺ is more stable than Mo⁶⁺ under reducing conditions and possesses a similar ionic radius (0.59 Å and 0.60 Å for Mo⁶⁺ and W⁶⁺ respectively). Upon substituting W⁶⁺ for Mo⁶⁺ in the oxide ion conductor La₂Mo₂O₉ (LAMOX) the conductivity decreases only very slightly as the W concentration increases. This reduction in the ionic conductivity is observed at all temperatures²² and demonstrates similar electrical properties for the Mo⁶⁺ and W⁶⁺ substituted phases. The slight difference in the ionic conductivity of LAMOX and W⁶⁺ substituted LAMOX phases are attributed to subtle changes in the crystal structure²² rather than the different electrical properties of W⁶⁺ and Mo⁶⁺. The bulk conductivity of Ba₃WNbO_8.5 is an order of magnitude lower than that of Ba₃MoNbO_8.5 at 450 °C but converges at higher temperatures. Similar behaviour is observed in Ba₃Mo_1−xNb_1+xO_8.5−x/2 and is a result of a change in M(1), M(2) and O(3) fractional occupancies upon increasing x. A neutron diffraction study of Ba₃WNbO_8.5 has been performed in order to further investigate differences in the electrical properties and crystal structures of Ba₃MoNbO_8.5 and Ba₃WNbO_8.5.

Rietveld refinement was performed using the GSAS/EXPGUI package.^23,24 Modelling of the background was performed using a shifted Chebyshev polynomial function and the peak shapes were fitted using a pseudo-Voigt function convoluted with back to back exponentials. Ba₃W_1.33Nb_0.66O_8.66 was previously reported as a 9R polytype perovskite, with trimers of (W/Nb)O₆ face-sharing octahedra connected by corner-sharing in the stacking sequence (hhc)₃ (space group R3̄m).²⁰ In this model the barium atoms are in position 3a (Ba(1)) and 6c (Ba(2)). The W and Nb atoms share the 6c (W(1)/Nb(1)) and the 3b (W(2)/Nb(2)) sites. The oxygen atoms occupy two distinct positions, 18h (O(1)) and 9e (O(2)). This model is closely related to the hybrid palmierite-9R polytype model proposed for Ba₃MoNbO₈, with the only difference being that the oxygen in position 6c (or in the split position 36i) is missing. Employing the Ba₃W_1.33Nb_0.67O_8.66 model provided acceptable statistical factors (χ² = 8.30, R_p = 6.41% and wR_p = 5.62%). However, examination of the fitted histograms revealed significant discrepancies between observed and calculated profiles, suggesting the 9R structural model was incorrect.

An excellent Rietveld fit was obtained with the hybrid 9R polytype – palmierite model previously reported for Ba₃MoNbO_8.5 (ref. 18) (space group R3̄m H; a = 5.8570(1) Å, c = 20.9964(5) Å, V = 623.77(3) ų, χ² = 2.37, R_p = 5.75% and wR_p = 3.18%). The Rietveld refinement fit from the neutron diffraction data taken from the 90° detector bank is presented in Fig. 4. The Ba₃WNbO_8.5 structure is composed of intertwined 9R polytype and palmierite domains, with mixed (W/Nb)O₄ tetrahedra and (W/Nb)O₆ octahedra (Fig. 5). The different crystal structures reported for Ba₃WNbO_8.5 and Ba₃W_1.33Nb_0.67O_8.66 (ref. 20) are potentially a result of the change in stoichiometry. The structural parameters obtained from Rietveld refinement of the Ba₃WNbO_8.5 neutron diffraction data are tabulated in Table S1.†

Fig. 4 Fitted room temperature TOF powder neutron diffraction pattern from Ba₃WNbO_8.5 collected in the Polaris 2θ ∼ 90° detector bank. Black crosses indicate the observed data, the red line the Rietveld fit, the blue line the difference between the observed and the calculated profiles, the green line the fitted background function and the pink vertical bars mark the reflection positions.

Fig. 5 Crystal structure of Ba₃WNbO8.5. (a) The hybrid structural model formed by the superimposition of the 9R-polytype (left) and the palmierite (right) sub-units representing the average structure of the system. (b) Bond lengths and angles of the (W/Nb)O₄ tetrahedra and (W/Nb)O₆ octahedra. Colours in (a and b) indicate: green Ba(1)/Ba(2), blue W(1)/Nb(1), purple W(2)/Nb(2), red O(1), orange O(2) and yellow O(3).

Anisotropic atomic displacement parameters were refined for all atoms. Disorder of the oxygen atom, O(3), within the (Mo/Nb)O₄ tetrahedra was evidenced by large U₁₁, U₂₂, and U₁₂ values. This was modelled by using a single isotropic U factor for O(3) and splitting the site as shown in Table S1† and Fig. 5. The disorder of O(3) onto the 18h positions arises as a result of the short distance between Ba(2) and W(1)/Nb(1) so that the disorder stabilizes the structure, resulting in longer and more realistic Ba(2)–O(3) and W(1)/Nb(1)–O(3) bond lengths. Similar structural disorder has been reported for Ba₃MoNbO_8.5 (ref. 18) and the mixed conducting hexagonal perovskite Ba₇Y₂Mn₃Ti₂O₂₀.²⁵ The thermal ellipsoids of O(1), O(2), Ba and W/Nb atoms are displayed in Table S1.† The W/Nb cations exhibit highly anisotropic thermal displacement along the c-axis in both the 6c and 3b sites (Table S1†) Similar U_ij values have been found for Mo/Nb in Ba₃MoNbO_8.5, and have been ascribed to the different coordination environments of the Mo/Nb cations.

The Ba fractional occupancies refined to within ±1% of the full occupancy and thus were fixed in further refinement at 1.0. Refinement of the individual oxygen site occupancies resulted in an overall oxygen stoichiometry of 8.5. The oxygen (O(1)) at position 18h is fully occupied while the O(2) and O(3) sites are partially occupied. The results demonstrate that the replacement of Mo⁶⁺ with W⁶⁺ is accompanied by large changes in the fractional occupancies of O(2), O(3), M(1) and M(2). At room temperature the majority of the W(1)/Nb(1) ions in Ba₃WNbO_8.5 exhibit octahedral coordination with just ∼13% of W(1)/Nb(1)O₄ tetrahedra within the average structure. In contrast there are ∼50% Mo(1)/Nb(1)O₄ tetrahedra within the average structure of Ba₃MoNbO_8.5.¹⁹ The W⁶⁺ preferentially occupies the M(2) site so that there is an increase in the fractional occupancy of the M(2) site upon replacement of Mo⁶⁺ with W⁶⁺ (Table S1†). It has previously been reported that there is a strong correlation between the oxide ionic conductivity and the number of (Mo/Nb)O₄ tetrahedra within the average structure of Ba₃MoNbO_8.5. The magnitude of oxide ionic conductivity observed is enhanced as the ratio of (Mo/Nb)O₄ tetrahedra to (Mo/Nb)O₆ octahedra increases. Oxide ion migration occurs via the partially occupied O(2) and O(3) sites in Ba₃MoNbO_8.5. The increase in the number of (Mo/Nb)O₄ tetrahedra most likely offers more low energy transition paths for transport of the O²⁻ ions enhancing the conductivity.¹⁹ The ratio of (Mo/Nb)O₄ tetrahedra to (Mo/Nb)O₆ octahedra strongly decreases upon replacement of W⁶⁺ for Mo⁶⁺ so that the ionic conductivity is an order of magnitude lower in Ba₃WNbO_8.5 at 450 °C (Fig. 3(a)).

However upon heating, the bulk ionic conductivities of Ba₃WNbO_8.5 and Ba₃MoNbO_8.5 converge (Fig. 3(a)) so that by 600 °C similar ionic conductivities are observed for both compounds. A similar result has been reported for the Ba₃Mo_1−xNb_xO_8.5−x/2 solid solution.²⁶ A variable temperature neutron diffraction study of Ba₃MoNbO_8.5 has previously shown that above 300 °C the oxygen/vacancy distribution changes as the temperature increases so that the ratio of (Mo/Nb)O₄ tetrahedra to (Mo/Nb)O₆ octahedra increases upon heating with a concomitant increase in the ionic conductivity.¹⁹ A change in slope is evidenced in the Arrhenius plot of both Ba₃WNbO_8.5 and the Ba₃MoNbO_8.5 at 540 °C and 510 °C respectively.

The variable temperature neutron diffraction study of Ba₃MoNbO_8.5 demonstrated that above ∼500 °C the rate of change in the ratio of (Mo/Nb)O₄ tetrahedra to (Mo/Nb)O₆ octahedra slows down.¹⁹ The gradient of the Arrhenius plot of the bulk conductivity of Ba₃WNbO_8.5 above 540 °C is greater than that observed for Ba₃MoNbO_8.5. This indicates that above 540 °C there is a larger rate of increase in the number of tetrahedra within the crystal structure of Ba₃WNbO_8.5. A likely explanation for the convergence of the ionic conductivity at 600 °C is therefore that the same structural rearrangement of (M/Nb)O₄ tetrahedra and (M/Nb)O₆ octahedra occurs in Ba₃MNbO_8.5 (M = Mo and W) upon heating but the rate of change in the number of tetrahedra above ∼500 °C is greater for M = W⁶⁺ and by 600 °C the number of (M/Nb)O₄ tetrahedra within the average structure of Ba₃MNbO_8.5 (M = Mo and W) are at similar levels. The same result is observed for the Ba₃Mo_1−xNb_xO_8.5−x/2 solid solution²⁶ and suggests that the structure of the hexagonal perovskite derivative Ba₃MNbO_8.5 has the propensity to rearrange to an optimum ratio of (M/Nb)O₄ tetrahedra and (M/Nb)O₆ octahedra upon heating to 600 °C. Variable temperature neutron diffraction are warranted to investigate this further.

Selected bond lengths and angles of Ba₃MoNbO_8.5 and Ba₃WNbO_8.5 are displayed in ESI Tables S2 and S3† for comparison purposes. The structure of Ba₃MNbO_8.5 (M = Mo, W) contains three equal M(1)–O(1) bonds and three equal O(1)–M(1)–O(1) angles (α) forming the M(1)O(1)₃ unit, which is common to the M(1)–O(1)₃O(3) tetrahedron and the M(1)–O(1)₃O(2)₃ octahedron. M(1)–O(1)₃O(3) is then defined by the M(1)–O(3) distance and the O(1)–M(1)–O(3) angle (β), obtained by the average of the possible angles given by the O(3) split positions. The M(1)–O(1)₃O(2)₃ octahedron is defined by the M(1)–O(1) bond length, α, the M(1)–O(2) bond lengths and the O(1)–M(1)–O(2) and O(2)–M(1)–O(2) angles (γ and δ respectively) (Fig. S3†). Upon replacement of W⁶⁺ for Mo⁶⁺ there are clear changes to the bond lengths and angles of the M(1)O_x polyhedra. The M(1)–O(1) bond length increases whereas the M(1)–O(2) and M(1)–O(3) bond lengths both decrease in magnitude. The angle α decreases whereas the angles β, γ and δ increase. These changes reflect the relaxation of the structure as the number of oxygen vacancies at the O(3) site increases. The changes of the bond lengths and angles demonstrate that the replacement of W⁶⁺ for Mo⁶⁺ leads to displacement of the M(1) atom closer to the O(2)/O(3) sites, in respect to its equilibrium position in Ba₃MoNbO_8.5, with subsequent relaxation of the surrounding lattice (Table S2†). The magnitude of the displacement can be obtained from the distance between M(1) and the [O(1)–O(1)–O(1)] face, given by D = –(M(1)–O(1))cos(β).¹⁹

Displacement of the metal atoms adjacent to the mobile oxygen sites has been reported in various oxide ion conductors^27,28 and can be associated with the energetics of the ionic transport as displacement of d-metal cations from oxygen vacancies is thought to lower the motional enthalpy required for the mobility of the oxide ions.²⁹ The motional enthalpy, ΔH_m, directly contributes to the activation energy of the conduction process and is the sum of two components: a barrier energy for the ionic hopping and a relaxation energy needed for the lattice relaxation around the conductive sites.³⁰ The M(1) polyhedra in Ba₃MoNbO_8.5 are already distorted¹⁸ and it has been previously reported that the M(1) displaces away from the O²⁻ ions at the O(2)/O(3) sites upon heating above 300 °C.¹⁹ Upon substituting W⁶⁺ for Mo⁶⁺ the M(1) atom shifts in the opposite direction, i.e. towards the O(2) and O(3) positions. Such motion of M(1) towards the partially occupied oxygen sites results in the increase of the bulk activation energy from 1.21 eV for Ba₃MoNbO_8.5 to 1.94 eV for Ba₃WNbO₈ as more energy is needed for the relaxation of the metal lattice around the oxygen vacancies and the O(2)/O(3) sites.

Out-of-centre distortions in the coordination polyhedra of d⁰ transition metals are also influenced by electronic effects, which mutually contribute to the overall displacement.³¹ Electronic distortions, such as second order Jahn–Teller distortions, are generated in the presence of transition metals with large formal charges and small cationic radii. A decrease in the second order Jahn–Teller effect would be expected upon substituting W⁶⁺ for Mo⁶⁺ since W⁶⁺ is a moderate distorter, while Mo⁶⁺ is a strong distorter.³² The M(1) displacement towards the O(2)/O(3) sites evidenced for Ba₃WNbO_8.5 with respect to the M(1) “equilibrium” position in Ba₃MoNbO_8.5 is hence a result of both the electronic and structural effects induced by the substitution of W⁶⁺ for Mo⁶⁺.

Conclusions

In summary the electrical and structural properties of the novel oxide ion conductor Ba₃WNbO_8.5 have been elucidated. W⁶⁺ is more stable than Mo⁶⁺ under reducing conditions so that replacing W⁶⁺ for Mo⁶⁺ reduces the electronic component in low pO₂. This is similar to tungsten doping studies of LAMOX where the presence of W⁶⁺ leads to an increase in the stability range for La₂Mo_0.5W_1.5O₉ at 800 °C.³³

In Ba₃MNbO_8.5 (M = Mo, W) the ionic migration is thought to proceed via tetrahedral and octahedral interchange occurring through a cooperative interstitialcy-like motion of the oxygen atoms between the occupied and vacant oxygen sites within the pseudo-cubic BaO_2.5 layers. At 20 °C in Ba₃WNbO_8.5, the O(3) site presents a much lower fractional occupancy than observed for Ba₃MoNbO_8.5 thus reducing the concentration of charge carriers, resulting in a bulk oxide ionic conductivity an order of magnitude lower than that reported for Ba₃MoNbO_8.5 at 450 °C. Upon heating to 600 °C the bulk conductivities of Ba₃MoNbO_8.5 and Ba₃WNbO_8.5 converge. Given that the ionic conductivity has been shown to be sensitive to the ratio of (M/Nb)O₄ tetrahedra to (M/Nb)O₆ tetrahedra,^19,26 it's highly likely that the crystal structure of Ba₃MNbO_8.5 hexagonal perovskite derivatives rearranges to an optimal ratio of tetrahedra : octahedra upon heating. Finally the results demonstrate that oxide ion conductivity can be established in other members of the Ba₃MM′O_8.5 family and the synthesis of new materials where M = a hexavalent cation and M′ = a pentavalent cation could result in new materials with applications as SOFC electrolytes, oxygen sensors and/or oxygen separation membranes.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This research was supported by the Northern Research Partnership and the University of Aberdeen. We also acknowledge Science and Technology Facilities Council (STFC) for provision of beamtime at ISIS.

References

1. B. C. H. Steele, Mater. Sci. Eng., B, 1992, 13, 79–87.
2. J. B. Goodenough, Annu. Rev. Mater. Res., 2003, 33, 91–128.
3. L. Malavasi, C. A. J. Fisher and M. S. Islam, Chem. Soc. Rev., 2010, 39, 4370–4387.
4. S. J. Skinner and J. A. Kilner, Mater. Today, 2003, 6, 30–37.
5. O. Yamamoto, Electrochim. Acta, 2000, 45, 2423–2435.
6. E. D. Wachsman and K. T. Lee, Science, 2011, 334, 935–939.
7. S. C. Singhal, Solid State Ionics, 2000, 135, 305–313.
8. A. Navrotsky, J. Mater. Chem., 2010, 20, 10577–10587.
9. P. Lacorre, F. Goutenoire, O. Bohnke, R. Retoux and Y. Laligant, Nature, 2000, 404, 856–858.
10. N. M. Sammes, G. A. Tompsett, H. Näfe and F. Aldinger, J. Eur. Ceram. Soc., 1999, 19, 1801–1826.
11. E. Kendrick, M. S. Islam and P. R. Slater, J. Mater. Chem., 2007, 17, 3104–3111.
12. T. Ishihara, H. Matsuda and Y. Takita, J. Am. Chem. Soc., 1994, 116, 3801–3803.
13. M. Li, M. J. Pietrowski, R. A. De Souza, H. Zhang, I. M. Reaney, S. N. Cook, J. A. Kilner and D. C. Sinclair, Nat. Mater., 2014, 13, 31–35.
14. K. Fujii, Y. Esaki, K. Omoto, M. Yashima, A. Hoshikawa, T. Ishigaki and J. R. Hester, Chem. Mater., 2014, 26, 2488–2491.
15. K. T. Lee, H. S. Yoon and E. D. Wachsman, J. Mater. Res., 2012, 27, 2063–2078.
16. K. R. Kendall, C. Navas, J. K. Thomas and H. Z. Loye, Solid State Ionics, 1995, 82, 215–223.
17. J. W. Fergus, J. Power Sources, 2006, 162, 30–40.
18. S. Fop, J. M. S. Skakle, A. C. McLaughlin, P. Connor, J. T. S. Irvine, R. I. Smith and E. J. Wildman, J. Am. Chem. Soc., 2016, 138, 16764–16769.
19. S. Fop, E. J. Wildman, J. T. S. Irvine, P. A. Connor, J. M. Skakle, C. Ritter and A. C. McLaughlin, Chem. Mater., 2017, 29, 4146–4152.
20. S. Kemmler-Sack and U. Treiber, Z. Anorg. Allg. Chem., 1981, 478, 198–204.
21. J. T. S. Irvine, D. C. Sinclair and A. R. West, Adv. Mater., 1990, 2, 132–138.
22. S. Georges, O. Bohnké, F. Goutenoire, Y. Laligant, J. Fouletier and P. Lacorre, Solid State Ionics, 2006, 177, 1715–1720.
23. A. C. Larson and R. B. Von Dreele, Los Alamos National Laboratory Report LAUR 86-748, 1994.
24. B. H. Toby, J. Appl. Crystallogr., 2001, 34, 210–213.
25. X. Kuang, M. Allix, R. M. Ibberson, J. B. Claridge, H. Niu and M. J. Rosseinsky, Chem. Mater., 2007, 19, 2884–2893.
26. S. Fop, E. J. Wildman, J. M. S. Skakle, C. Ritter and A. C. Mclaughlin, Inorg. Chem., 2017, 56, 10505–10512.
27. M. S. Islam, J. Mater. Chem., 2000, 10, 1027–1038.
28. M. S. Islam, J. R. Tolchard and P. R. Slater, Chem. Commun., 2003, 13, 1486–1487.
29. J. B. Goodenough, A. Manthiram, M. Paranthamam and Y. S. Zhen, Mater. Sci. Eng., B, 1992, 12, 357–364.
30. J. B. Goodenough, Annu. Rev. Mater. Res., 2003, 33, 91–128.
31. M. Kunz and D. I. Brown, J. Solid State Chem., 1995, 115, 395–406.
32. K. M. Ok, P. S. Halasyamani, D. Casanova, M. LLunell, P. Alemany and S. Alvarez, Chem. Mater., 2006, 18, 3176–3183.
33. D. Marrero-Lopez, J. Canales-Vazquez, J. C. Ruiz-Morales, J. T. S. Irvine and P. Nunez, Electrochim. Acta, 2005, 50, 4385–4395.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ta08989a

---