J O U R N A L O F C H E M I S T R Y Materials Communication NaBi3V2O10: a new oxide ion conductor Derek C. Sinclair, Craig J.Watson, R. Alan Howie, Janet M. S. Skakle, Alison M. Coats, Caroline A. Kirk, Eric E. Lachowski and James Marr Chemistry Department, University of Aberdeen, Meston Walk, Aberdeen, UK, AB24 3UE c=5.5312(6) A ° , a=84.542(12), b=113.318(11) and c= 112.267(12)°. Table 1 shows the first thirty lines of the indexed The new phase NaBi3V2O10 is reported; it was synthesised by oxide reaction at 600 °C and is triclinic with a=7.2026(10), powder pattern.Given that the volume of an oxygen ion can be estimated as ca. 22A ° 3, the cell volume of 238.53(4) A ° 3 is in b=7.0600(9), c=5.5312(6) A ° , a=84.542(12), b=113.318(11) and c=112.267(12)°; it is an oxide ion conductor with good agreement with that expected for the proposed formula with Z=1, giving support to the suggested unit cell.In conductivity of ca. 1.5 mS cm-1 at 675 °C. addition, selected area electron diVraction (SAED) studies have also been found to be consistent the unit cell proposed by the VISSER program. DiVerential thermal analysis showed the presence of a large endotherm at 575 °C on heating which was fully reversible on Yttria-stabilised zirconia (YSZ) is commonly employed as a thermal cycling, suggesting NaBi3V2O10 undergoes a poly- solid electrolyte in many technological applications such as morphic phase transition at this temperature.Although this solid oxide fuel cells and oxygen pumps.1 Although YSZ is an has been confirmed by high temperature XRD, as yet, we have excellent solid electrolyte there remains much interest, from no information on the symmetry or crystal structure of the both a fundamental and an industrial view point, in trying to high temperature polymorph.find new oxide ion conductors2,3 which have superior electrical ac Impedance measurements on a pellet sintered at 675 °C properties compared with YSZ.Over the last eight years, there and coated with Au paste electrodes were collected on both has been considerable interest in doped Bi4+yV2-yO11-y solid heating and cooling in air between 25–675 °C. Complex solutions4,5 known by the acronym BIMEVOX, where ME impedance plane, Z*, plots consisted of a single, semi-circular corresponds to the dopant ion, owing to their high oxide ion arc and a low frequency electrode ‘spike’, as shown in Fig. 1. conductivity. During a phase diagram study of the composi- The associated capacitance of the arc was calculated to be ca. tional range of BINAVOX solid solutions6 within the 3–5 pF cm-1 using the relationship vRC=1 (where v=2pf and Na2O–Bi2O3–V2O5 ternary system, we discovered a new phase, is the angular frequency) at the arc maximum.This capacitance whose composition was determined via electron probe microanalysis (EPMA) to be NaBi3V2O10.7 Here, we report the synthesis of this new phase, a fully indexed X-ray pattern based on a primitive, triclinic cell and preliminary conductivity data Table 1 Indexed X-ray diVraction pattern for triclinic NaBi3V2O10 that suggests NaBi3V2O10 is an oxide ion conductor.with a=7.2026(10), b=7.0600(9), c=5.5312(6) A° , a=84.542(12), b= NaBi3V2O10 was prepared by conventional solid state syn- 113.318(11) and c=112.267(12), V=238.53(4) A °3. A full listing (101 lines) is available from the authors on request thesis. Bi2O3 (99.99%), V2O5 (99.6%) and Na2CO3 (99.99%) reagents were dried at 300 °C overnight and stored in a dobs dcalc h k l I/I0 D(2h) desiccator prior to use. A reaction mixture of stoichiometry NaBi3V2O10 totalling 3–4 g was weighed from the starting 6.5256 6.5199 0 1 0 15.3 -0.0119 reagents and mixed into a paste with acetone using an agate 6.1454 6.1361 1 0 0 72.4 -0.0220 mortar and pestle, dried and fired in Au foil boats.A combi- 5.6451 5.6415 -1 1 0 19.5 -0.0103 5.0719 5.0688 0 0 1 14.6 -0.0105 nation of X-ray powder diVraction (XRD) data and EPMA 4.9816 4.9773 -1 0 1 22.6 -0.0156 results showed that a single phase yellow powder could be 4.5150 4.5113 -1 1 1 9.5 -0.0162 prepared by heating the reaction mixture at 600 °C for 24 h, 4.1327 4.1332 0 -1 1 29.2 0.0025 with an intermediate regrind after 12 h.EPMA analysis showed 3.8844 3.8821 0 1 1 1.1 -0.0139 that there was no evidence of any secondary or unreacted 3.8137 3.8139 1 1 0 9.1 0.0013 phases on a micrometre scale.Quantitative EPMA analysis 3.5657 3.5659 -1 -1 1 8.7 0.0008 3.4631 3.4641 -1 2 0 6.8 0.0078 on twenty seven points of a sintered pellet determined the 3.4278 3.4293 -2 1 1 22.6 0.0109 composition to be 15.9 mol% Na2O, 33.6 mol% V2O5 and 3.3195 3.3224 1 0 1 96.7 0.0235 50.5 mol% Bi2O3 which is in good agreement with the starting 3.2867 3.2891 -2 1 0 29.2 0.0203 composition of NaBi3V2O10. 3.2479 3.2456 -2 0 1 100.0 -0.0197 On heating above ca. 700 °C, the yellow powder became 3.0836 3.0867 -1 2 1 71.1 0.0298 brown and extra reflections associated with a secondary phase 3.0650 3.0681 2 0 0 9.2 0.0300 2.8577 2.8607 -2 2 1 15.7 0.0338 appeared in the XRD patterns.The reaction mixture melted 2.8180 2.8207 -2 2 0 86.5 0.0309 at ca. 755 °C and formed a purple–brown coloured solid on 2.7531 2.7523 -1 0 2 41.8 -0.0100 cooling. XRD analysis showed the minor phase in powders 2.7041 2.7048 1 1 1 12.7 0.0090 heated above 700 °C and the major phase cooled from the 2.6775 2.6769 1 -2 1 18.6 -0.0074 melt to be a c-polymorph of the BINAVOX solid solution. 2.6268 2.6289 -1 1 2 15.9 0.0285 Detailed phase studies are currently in progress and will be 2.5671 2.5660 -2 -1 1 2.1 -0.0152 2.5326 2.5344 0 0 2 19.3 0.0266 reported elsewhere. 2.5133 2.5159 1 2 0 36.8 0.0386 The program VISSER8 was used in an attempt to index the 2.4877 2.4886 -2 0 2 5.2 0.0149 XRD pattern of NaBi3V2O10; results suggested that the most 2.4458 2.4464 2 1 0 17.5 0.0101 probable solution was a primitive triclinic cell, which was 2.3952 2.3969 -3 1 1 5.2 0.0286 refined to give a unit cell, a=7.2026(10), b=7.0600(9), J.Mater. Chem., 1998, 8(2), 281–282 281Fig. 2 Z* plots for NaBi3V2O10 in various atmospheres at 650 °C. Fig. 1 Z* plot for NaBi3V2O10 in air at 505 °C. Selected frequencies 0.3 Hz is identified by the filled symbol in N2 and air data.in filled circles are identified by the logarithm of the frequency, e.g. 2=102 Hz. value was temperature independent over the measured range and is consistent with a bulk or intra-granular response. The presence of a low frequency spike with an associated capacitance of 1–5 mF in Fig. 1 is attributable to ionic polarisation and diVusion-limited phenomena at the electrode and supports the idea that the conduction is mainly by means of ions.At higher temperatures in air, ca. 600 °C, the low frequency response consists of a broad semi-circular arc with an associated capacitance of ca. 10-5 F, consistent with a charge transfer reaction occurring at the sample/electrode interface. The resistance associated with this process can be estimated from the diameter of the low frequency semi-circular arc in Z* plots.In order to establish if the material was an oxide ion conductor the gas atmosphere at 650 °C was changed sequentially from laboratory air to flowing oxygen to flowing nitrogen before reverting to laboratory air. The oxygen partial pressure of the atmosphere had a dramatic eVect on the low frequency Fig. 3 Arrhenius plot of the bulk conductivity in air.Open and closed response, as shown in Fig. 2. On changing the atmosphere circles represent heating and cooling data, respectively. from laboratory air to flowing oxygen the resistance associated with the charge transfer process occurring at the electrode/ fully reversible on thermal cycling. As the data do not obey sample interface decreased from a value of ca. 1.25 kV to a the Arrhenius law it is diYcult to calculate any activation (constant) value of ca. 0.25 kV after ca. 1 h. In flowing N2 the energy for the bulk conduction process, however, the data resistance associated with the charge transfer process increased clearly start to approach a plateau above 600 °C. rapidly and after 1 h the low frequency response consisted of The bulk conductivity of NaBi3V2O10 is two orders of an inclined-spike at an angle of ca. 45°. Such a response is magnitude lower than that of YSZ at ca. 600 °C;1 as yet, indicative of a Warburg-like response and suggests that the however, we have only studied the parent material. It may be rate-limiting step controlling the overall impedance at low possible to enhance the conductivity of NaBi3V2O10, especially frequencies involves the diVusion of electroactive species below the phase transition temperature of 575 °C by stabilising to/from the electrode/sample interface. Given the dependence the high temperature polymorph via chemical doping, as is the of this process on oxygen partial pressure in the surrounding case with ZrO2- and BIMEVOX-based solid electrolytes.atmosphere, the diVusing species must be oxygen-based, presumably O2 molecules. This therefore indicates that We wish to thank Professor Tony West for useful discussions, NaBi3V2O10 is predominantly an O2- ion (rather than an Na+ the University of Aberdeen for a studentship (C.J.W.) and ion) conductor. EPSRC for financial support for the EPMA facility. The changes in electrode behaviour were reproducible on switching between the various atmospheres whereas the bulk resistivity was independent of oxygen partial pressure, as References shown by the high frequency, non-zero intercept in Fig. 2. 1 B. C. H. Steele, Solid State Ionics, 1984, 12, 391. Although we need to perform concentration (emf ) cell measure- 2 H. L. Tuller and A. S. Nowick, J. Electrochem.Soc., 1975, 122, 255. ments in order to prove that NaBi3V2O10 is an oxide ion 3 T. Ishihara, H. Matsuda and Y. Takita, J. Am. Chem. Soc., 1994, conductor, the impedance behaviour described above is com- 116, 3801. 4 F. Abraham, J. C. Boivin, G. Mairesse and G. Nowogrocki, Solid pelling evidence that this material is predominantly an oxide- State Ionics, 1990, 40/41, 934. ion conductor. 5 C. K. Lee, B. H. Bay and A. R. West, J.Mater. Chem., 1996, 6, 331. Bulk conductivity values were calculated from the reciprocal 6 C. J. Watson, A. Coats and D. C. Sinclair, J. Mater. Chem., 1997, of the low frequency intercept of the high frequency semi- 7, 2091. circular arc with the Z¾ axis of Z* plots and are shown in the 7 C. J.Watson, M.Sc. Thesis, University of Aberdeen, 1997. form of an Arrhenius plot, Fig. 3. There is no discontinuity in 8 J.W. Visser, J. Appl. Crystallogr., 1969, 2, 89. the Arrhenius plot around the transition temperature at ca. 575 °C, instead the data yield a sigmoidal curve which is nearly Communication 7/07760B; Received 28th October, 1997 282 J. Mater. Chem., 1998, 8(2), 281–282