JOURNAL OF MATERIALS SCIENCE 29 (1994) 3691-3701 Conductivity measurements of various yttria-stabilized zirconia samples J. VAN HERLE, A. J. McEVOY, K. RAVINDRANATHAN THAMPI Institut de Chimie Physique II, Ecole Polytechnique F#d#rale de Lausanne, 1015 Lausanne, Switzerland Samples of yttria-stabilized zirconia manufactured by the following fabrication procedures, were obtained from commercial sources: (i) hot isostatic pressing; (ii) tape casting; (iii) vacuum plasma spraying, and (iv) calendering. The ionic conductivities of these samples were measured by (a) impedance spectroscopy; (b) the four-point probe method; (c) the current-interruption technique, and (d) the van der Pauw technique. The tape-cast and hot pressed samples showed good and very reproducible conductivity values. The vacuum plasma sprayed samples showed an anisotropy in their conductivity, with the cross-plane value being several times lower than the in-plane value. A simple model based on the porous microstructure of these samples can explain this observation. Sintering of the plasma sprayed samples minimized the anisotropy and significantly improved their conductivity values. The calendered samples also showed a similar anisotropy in their conductivity data when they were inadequately sintered. 1. Introduction Yttria-stabilized zirconia (YSZ) is the most widely used solid electrolyte in solid oxide fuel cell (SOFC) applications [1, 2]. Despite its relatively high resistiv- ity compared to alternative solid electrolytes, like doped CeO2 or BizO 3 [3-6], it is still used because of its superior chemical stability under extreme condi- tions of temperature and gas environment (oxidizing and reducing on opposite faces of electrolyte). Besides, for samples that are thin enough (< 300 gm), its conductivity is still sufficient to allow high current densities ( > 2 Acm- 2), at not too low temperatures (900 ~ YSZ is fabricated by commercial manufacturers and university laboratories in a variety of ways [1, 2, 7, 8]. Fabrication techniques include hot isostatic pressing (HIP) 1,9], tape casting (TC) 1-10-13], vacuum plasma spraying (VPS) 1-14-20], atmospheric plasma spraying (APS) [21,22], sputtering [3,23-25], calendering 1-26, 27], sol-gel processing 1-28], electrochemical va- pour deposition (EVD) [29-31], chemical vapour de- position (CVD) 1-31,32], vapour phase electrolytic deposition (VED) [33, 34], slurry coating [18, 35, 36], spray pyrolysis [35, 37, 38], laser physical vapour de- position (LPVD) [39-41], slip casting [36, 37, 42] and ion plating 1,20, 37, 43]. A partial overview of depos- ition techniques is covered in 1,44]. Solid oxide fuel cells with electrolytes that originate from one or other of these fabrication procedures may show big differences in their power output character- istics [18, 37, 45-47]. Of course, electrolyte resistivity constitutes only one of the losses that occur in a solid oxide fuel cell, and normally is not solely responsible for performance differences. Indeed, the losses occurring at the electrodes (polarization losses) are generally recognized as being more important than the pure ohmic loss of the electrolyte [48, 49]. This provides, in fact, another argument in favour of the continued use of YSZ as an electrolyte, in spite of the availability of other, better conducting (but less stable) electrolytes [3-6]. As long as electrode polarization losses are dominant in overall cell performance, little can be gained by replacing the electrolyte material. However, one fact seems to emerge from the literat- ure: whenever the electrolyte membrane of a test cell is not perfectly dense, the maximum obtainable output of the cell is low. On comparing experimental data of cells with a differently fabricated YSZ electrolyte with otherwise identical electrodes, this fact becomes clear [18, 45-47]. It is pronounced for cells with a YSZ electrolyte made by VPS [18, 21,45,46, 50], sputtering [23, 25], ion plating [37, 45, 47] or APS [21, 46]. Some values are listed in Table I. As an example, Fig. 1 shows a current-voltage curve of a VPS cell tested in the laboratory. The maximum power output is less than 100 mW cm-2 at 900 ~ whereas the envisaged working point of SOFC should be at least over 200 mW cm- 2 at this temper- ature, considering the thickness of the YSZ layer. Impedance spectroscopy measurements showed that, in this case, electrolyte resistance is responsible for 90% of the loss. This illustrates the point that electro- lyte ohmic polarization can be solely responsible for poor overall cell performance. Nevertheless, this experiment shows the high poten- tial of the VPS technique: the high energy impact of the plasma spray produces a very intimate electrode-electrolyte contact, resulting in low elec- trode polarization. In fact, the high potential of plasma spraying techniques has been demonstrated, 0022-2461 9 1994 Chapman & Hall 3691