A two phase composite disc containing α-alumina and yttria-stabilized zirconia (YSZ) was die pressed and sintered at 1600oC first. The disc was then converted to β"-alumina and YSZ composite by the previously reported vapor phase method in which the sintered samples of α-alumina + YSZ were exposed to Na2O vapor [1, 2]. After the vapor phase conversion treatment, the samples were ion-exchanged in molten LiNO3 to form Li-β”-alumina + YSZ composites. Fig.1a shows the measured ion exchange fraction (Li+/(Li++Na+)) vs. ion exchange time at 300oC and 370oC. The measured ion exchange fraction agrees relatively well with that by Yao and Kummer in single phase β-alumina . The partially ion-exchanged Li-β”-alumina is expected to be a predominantly Li+-ion conductor 
Ionic conductivity of both Na-β”-alumina + YSZ and Li-β”-alumina +YSZ was measured using electrochemical impedance spectroscopy (EIS) with Au electrodes. Fig. 1b shows the Arrhenius plots of the total ionic conductivity measured by EIS on a fully converted sample and after ion exchange for 24 h at 300oC. The left inset shows EIS spectra (10 Hz to 1 MHz) at different scales. Over the range of temperatures investigated, the grain boundary arc was not observed. Thus, the measured conductivity is attributed to the total conductivity. Both samples exhibit an Arrhenius behavior. The activation energy for the converted sample is 0.13 eV while that for the ion exchanged sample is 0.18 eV. The room temperature conductivities were estimated by extrapolation. The estimated total ionic conductivities at room temperature are 3.6 x 10-3 and 7.5 x 10-4 Scm-1 for the sample before and after ion exchange, respectively.
Fig. 1c shows the measured OCV of a LMO+C||Li+- β"-alumina + YSZ ||Li cell at room temperature over a period of 12 hours using a LiNO3 ion exchanged disc (0.4 exchange fraction, 1 mm thickness). The OCV was measured on the cell assembly without any charging process. The measured initial OCV of 3.4 V was slightly lower than theoretical OCV of ~3.5 V reported for this type of cathode. Since the electronic conductivity of this type of solid electrolyte is expected to be extremely low, self-discharge is not expected. The voltage drop here thus indicates possibly changing activity of the anode due to non-optimized cell assembly, such as the sealing method. EIS spectra obtained on the cell at room temperature over a frequency range from 1 Hz to 1 MHz showed a very large and incomplete arc. The high frequency intercept is ~350 Ω. This is expected to be mainly the electrolyte resistance. From the disc dimensions, the estimated ionic conductivity is ~1.78 x 10-4 Scm-1.
The measurements indicates the battery performance was predominantly limited by the electrode polarization. Future work will involve the fabrication of an all solid-state battery with cathode containing a mixture LMO (or other materials such as sulfur), powder of Li- β”-alumina and carbon. Future work will also involve using a much thinner electrolyte disc, made by tape-casting, sintering, vapor phase conversion, and ion exchange.
Acknowledgments: This work was supported by the National Science Foundation under Grant Number DMR-1407048.
. A. V. Virkar, J. F. Jue, and K-Z. Fung, US Patent No. 6,117,807.
. P. Parthasarathy and A. V. Virkar, J. Electrochem. Soc., 160 A2268-A2280 (2013).
. Y.-F.Y. Yao and J.T. Kummer, Journal of Inorganic and Nuclear Chemistry, 29 2453-2475 (1967).
. G.C. Farrington and W.L. Roth, Electrochimica Acta, 22 767-772 (1977).
Fig.1. (a) Measured Li+ ion exchange fraction as functions of temperature and time, (b) measured total conductivity with insets showing typical EIS spectra, (c) measured open circuit voltage (OCV) of the LMO+C||Li+-β"-alumina + YSZ ||Li cell at room temperature over a period of 12 h. Inset show the cell strucutre.