Solid State Ionics 40/41 North-Holland
OF SINGLE CRYSTAL LiNaS04
B.-E. MELLANDER Department of Physics, Chalmers University of Technology, S-412 96 Gtiteborg, Sweden
B. GRANELI Department of Reactor Physics, Royal Institute of Technology, S-100 44 Stockholm, Sweden
and J. ROOS Physik-Institut, Umversity of Ziirich, CH-8001 Ziirich, Switzerland
The ionic conductivity of single crystal LiNaSO, has been investigated for crystal orientation parallel to the a- and c-axis using complex impedance spectroscopy. The temperature range from 25 “C up to the solid-solid phase transition temperature 5 18°C has been studied. For ionic transport along the c-axis direction the ionic conductivity is only slightly larger than for transport in a direction perpendicular to the c-axis. The results of this study are compared to results for NMR measurements.
1. Introduction In the Li2S04-Na,SO, system the intermediate phase, P-LiNaSO,, is stable within a narrow concentration range [ l-4 1. The structure of this phase has been claimed to be trigonal with space group P31c [ 5 1. Recent NMR results [ 6 ] seem to contradict this, but the results of an extensive high resolution neutron diffraction study  only allow small deviations from the originally suggested structure. The ionic conductivity is low as for most low-temperature phases in sulphate systems [&lo]. At 5 18°C a first-order phase transition to the highly conducting a-phase takes place [ 8,111. In this phase the sulphate ions form a bee lattice where the lithium ions occupy tetrahedral positions while the sodium ions are in octahedral positions [ 12,13 1. There is wellfounded evidence for an enhancement of the cation mobility through a coupling to the rotation of the sulphate ions in the a-phase [4,14,15]. In the pphase, however, the rotational freedom of the sulphate ions is highly restricted. The ion dynamics of P-LiNaSO, has recently been 0167-2738/90/S ( North-Holland
03.50 0 Elsevier Science Publishers )
studied by NMR [ 16,17 ] and infrared spectroscopy [ 18 ] as well as by Raman scattering [ 19 1. We have performed measurements of the complex impedance using both single crystal and polycrystalline samples in order to study the relation between ion transport and structure.
2. Experimental Large LiNaSO, single crystals were grown by evaporation over several months at 50°C from a neutral aqueous solution containing equimolar proportions of Li,SO, and Na2S04 of Merck p.a. quality. All samples were taken from the same charge and the thin samples were cut from the same crystal, a three-sided prism, roughly 3 cm long, with one edge not fully developed. The sides of the triangular crosssection perpendicular to the c-axis was 12 mm. Samples were cut as plates in the two orientations with a disc cutter and subsequently cut to rectangular shape with a wire saw. Finally edges and corners of the thin samples were adjusted under microscope by
B.-E. Mellander et al. /Ionic conductivity of single crystal LiNaSO,
3. Results and discussion
P ? 0
Fig. 1. Complex impedance plot at 339°C for a single crystal LiNaSO, sample oriented with the c-axis parallel to the current direction.
hand polishing to give well-defined crystal sizes. The polycrystalline samples were prepared using the same chemicals which were dried, mixed and melted. The samples were allowed to cool slowly inside the furnace and then each sample was ground to a fine powder. Cylindrical pellets of 13 mm diameter and 1 to 2 mm thickness were pressed using 400 MPa pressure. The pellets were heat treated at 400°C for three days before use. The electrical conductivity was determined using complex impedance spectroscopy, a typical complex impedance plot is shown in fig. 1. Electrodes for both single crystal and polycrystalline samples were painted on the surfaces using graphite dissolved in alcohol (DAG-580, Acheson, Holland). After application of the electrodes the samples were dried in vacuum at room temperature in order to remove the remaining solvent from the graphite. The complex impedance of the samples was measured over a frequency range of 100 Hz to 100 kHz using a computer-controlled Hewlett-Packard HP4274A LCRmeter. The signal applied over the sample was 20 mV. The sample holder was purged with dry nitrogen gas during all measurements. The temperature was measured using a Platinel II thermocouple placed close to the sample.
The ionic conductivity for B-LiNaSO, single crystals with the c-axis parallel or perpendicular to the direction of the applied electric current is shown in fig. 2. It can be noted that for ion transport along the c-axis direction the ionic conductivity is only slightly larger than for transport perpendicular to the c-axis. The results for polycrystalline samples agrees well with the average value for the two crystal directions. Assuming Frenkel disorder, the conductivity in the intrinsic temperature range, where the ionic conductivity is determined by thermally activated crystal defects, can be described by UT= ( uT)~exp
where o is the conductivity, T the absolute temperature, k the Boltzmann constant, AHF the formation enthalpy for Frenkel defects and AH,,, the migration enthalpy for the mobile ion. for LiNaSO, both the lithium and the sodium ion may be mobile, i.e. we have two different mobile species which might have different enthalpies, but at present we will assume that we have only one set of enthalpies in eq. ( 1). In
Fig. 2. Temperature dependence of the ionic conductivity of single crystal LiNaSO, samples oriented with the c-axis perpendicular (circles) and parallel (squares) to the current direction.
B.-E. Mellander et al. /Ionic conductivity of single crystal LiNaSO,
the extrinsic region at low temperatures most crystal defects are due to the presence of small amounts of aliovalent impurities. In this study the In (oT) versus 1/T plots showed only minor deviations from a straight line. At low temperatures a slight curvature might indicate the onset of the extrinsic region. The measured value of AH,/2 + AH,,, is 2.15 eV for both current directions. The migration enthalpy for sodium diffusion in Na2S0,(I) is 0.5 eV . Since AH,,, for LiNaS04 is expected to be of the same order AHF can be estimated to be 3.3 eV. To determine the migration enthalpies for vacancies and interstitials controlled doping with aliovalent impurities should be performed. Very recently NMR relaxation data have been analyzed including possible relaxation by paramagnetic impurities [ 171. The activation energies to be compared with values of AHr/2+AH, from conductivity measurements are 1.9 eV for Li ions and 1.75 eV for Na ions. In the In (aT)versus 1/T plots there is a slight upward deviation from the straight line in the hightemperature region of the plot as has been observed for many other salts. This curvature may be due to defect-defect interactions, the influence of another transport mechanism or another mobile component or a temperature dependence of the formation and migration enthalpies. The curvature is, however, not very pronounced and does not seem to be an indication of extended pre-transformation effects. The phase transition to the high-temperature phase is accompanied by a distinct increase in conductivity. The ionic conductivity in the a-phase is very high, of the order of 1 R- ’ cm- ’ [ 8,111. For samples with low cell constants it is thus very difficult to measure the conductivity. Furthermore, the single crystals break when passing the first-order phase transition. We have thus in this investigation restricted the measurement to the P-phase.
Acknowledgement We would like to thank Lucien Pauli, Physik Institut der Universitat Zurich for preparing the LiNaSO, single crystals. This study has been supported financially by the Swedish Natural Sciences Research Council and Helge Ax:son Johnsons Stiftelse. References [ I] L.I. Staffansson, Acta. Chem. Stand. 26 ( 1972) 2150. [ 21 K. Schroeder, A. Kvist and H. Ljungmark, Z. Naturforsch. 27a (1972)
[ 31 K. Schroeder, Thesis, University of Gothenburg ( 1975). [ 41 A. Lunden, in: Materials for Solid State Batteries, eds. B.V.R. Chowdari and S. Radhakrishna (World Scientific, Singapore, 1986) pp. 149-160. [ 51B. Morosin and D.L. Smith, Acta Cryst. 22 ( 1967) 906.  K.-D. Junke, M. Mali, J. Roos and D. Brinkmann, Solid State Ionics 28-30 (1988) 1329. [ 7 ] B. Graneli, to be published. [ 81 A.F. Polishchuk and T.M. Shurzhal, Electrokhimiya 9 (1973) 838; Soviet Electrochem. 9 (1973) 902. [ 91 K. Singh and V.K. Deshpande, Solid State Ionics I3 ( 1984) 157. [lo] U.M. Gundusharma, C. MacLean and E.A. Secco, Solid State Commun. 57 (1986) 479. [ 111 A.-M. Josefson and A. Kvist, Z. Naturforsch. 24a ( 1969) 466. [ 121 T. Forland and J. Krogh-Moe, Acta. Cryst. 11,224 ( 1958). [ 13 ] L. Nilsson, N.H. Andersen and A. Lunden, Solid State Ionics 34(1989) 111. [ 141 A. Lunden and L. Nilsson, J. Mat. Sci. Letters 5 (1986) 645. [ 15 ] A. Lund& and J.O. Thomas, in: High Conductivity Solid Ionic Conductors - Recent Trends and Applications, ed. T. Takahashi (World Scientific, Singapore, 1989) p. 45. [ 161 K.-D. Junke, M. Mali, J. Roos, D. Brinkmann, A. Lund& and B. Graneli, Solid State Ionics 28-30 ( 1988) 1287. [ 171 K.-D. Junke, Ph.D. Thesis, University of Zurich ( 1989). [ 181 M.A. Pimenta, P. Echegut, G. Hauret and F. Gervais, Phase Trans. 9 (1987) 185. [ 191 D. Teeters and R. Frech, Phys. Rev. B 26 (1982) 5897.  M.A. Careem and B.-E. Mellander, Solid State Ionics 15 (1985) 327.