A pre-operational annealing process is under investigation as a potential means of preventing a sudden reduction of ionic conductivity in a β"—alumina solid electrolyte (BASE) during use. On the basis of tests described below, the sudden reduction of ionic conductivity, followed by a slow recovery, has been found to occur during testing of the solid electrolyte and electrode components of an alkali metal thermal-to-electric converter (AMTEC) cell. This conductivity reduction may be observed quite infrequently; at lower operating temperatures, T<1,073 K, it is not usually observed at all, while at T=1,123–1,173 K, hundreds of hours may pass before conductivity reduction occurs. Only on tests running at higher operating temperatures for thousands of hours is this phenomenon regularly exhibited. The reduction of ionic conductivity would degrade the performance of an AMTEC cell. A pre-operational annealing process would help to sustain performance.
The tests involved, among others, maintaining BASE at a temperature between 1,050 and 1,170 K in the presence of low-pressure sodium vapor — an environment similar to that in a sodium AMTEC cell. It was observed that after a few tens to hundreds of hours in this environment, the ionic conductivity of BASE can fall suddenly to a lower value, and thereafter recover during several tens of hours. The decrease in ionic conductivity has been attributed to the formation of nanometer-thickness cracks, which would be high-resistance paths with respect to conduction of ions. The recovery of ionic conductivity has been attributed to closing of the cracks through sintering of β"—alumina grains.
The approach to annealing is based on the following line of reasoning: Some NaAlO2 is incorporated into BASE during manufacture to facilitate sintering the ceramic to near 100-percent density. It is hypothesized that the cracks are initiated by the formation of voids, which, in turn, form as a result of (1) a reaction in which NaAlO2 is converted to Na2O and either β— or β" — alumina and (2) some or all of the Na2O escapes by vaporizing. If this hypothesis is correct, then it should be possible to perform a pre-operational anneal to (1) force the conversion of NaAlO2 and the formation of voids and (2) sinter the BASE to close the voids. Sintering should strengthen the material and thereby help to prevent the formation of voids and microscopic cracks during operation. In addition, if the loss of NaAlO2 during operation contributes to the loss of subsequently recoverable ionic conductivity, then the removal of NaAlO2 in the pre-operational sintering process should help to stabilize the high-temperature ionic conductivity of BASE.
In an experiment, samples of β"—alumina ceramic containing ≈1 weight percent of NaAlO2 were packed in a mixture of lithia-stabilized β" —alumina powder and sodium β—alumina powder in a loosely capped crucible, and annealed at a temperature of 1,400 °C (1,673 K) in a vacuum for various times up to 100 hours. (The powder mixture was used to establish fixed low sodium Na2O activity characteristic of the β/β" phase boundary. The Na2O activity of this mixture is much lower than that above NaAlO2 and facilitates the loss of Na2O from the NaAlO2 in the ceramic as well as sintering to close voids at grain boundaries.)
The β"—alumina samples were found to have lost an average of between 0.4 to 0.5 percent of mass during the 100-hour anneal; this amount of loss is consistent with the conversion of most of the NaAlO2 to β— or β"—alumina. (Complete conversion would entail a mass loss of 0.5 percent.) Qualitative tests showed the mechanical characteristics of the samples to have been improved by the anneal.
At this time, high-temperature tests of limited duration have indicated the superiority of the treated BASE, but reproducible tests over thousands of hours are necessary to confirm that microcracking has been eliminated. The ionic conductivity of the treated BASE is also measured to be higher than untreated BASE at 1,073 K in low-pressure sodium vapor. Microcracking resulting in loss of conductivity was not observed with treated BASE in one high-temperature experiment, but this result must be duplicated over very long testing times to be sure of the effect. Shorter annealing times (10 to 20 hours) were found to result in significantly less loss of mass; it may be necessary for the packed powder mixture to evolve some Na2O before the Na2O can leave the ceramic.
This work was done by Roger Williams, Margie Homer, Margaret Ryan, Roger Cortez, Virgil Shields, and Adam Kisor of Caltech for NASA's Jet Propulsion Laboratory. For further information, access the Technical Support Package (TSP) free on-line at www.techbriefs.com/tsp under the Materials category. NPO-20919.