Porous papers and felts fabricated from BN and Y20~ fibers have been evaluated as electrode separators for Li-A1/LiC1-KC1/FeSx cells. The physical and mechanical properties were determined, and in-cell tests exceeding 1000 hr were conducted. The papers and felts had adequate strength and flexibility for cell assembly and performed well during cell testing. Results from the property measurements are presented and the performance of these materials in the cell tests is discussed.High temperature, molten salt Li-A1/LiC1-KC1/FeSx (x ~ 1 or 2) cells are being developed by Argonne National Laboratory (ANL) for electric vehicle propulsion and for energy storage on electric utility systems (1). One of the requirements for the successful development of cells for commercial applications is the identification of a low cost, porous electrode separator. The primary function of the electrode separator is to prevent direct electron flow between the electrodes without unduly restricting ionic flow through the electrolyte. In addition, separators for Li-A1/LiC1-KC1/ FeSx cells must retain the Li-A1 and FeSx active materials within the electrodes. This requirement is significant because the Li-A1 and FeSx are in the form of powders, and penetration of the separator by these fine particles may lead to a loss in cell capacity or electrical short circuits between the electrodes. The optimum structure and properties desired in the electrode separator have not been fully determined; however, the following general requirements must be satisfied: (i) The separator must be a good electrical insulator to prevent direct electron flow between the electrodes.(ii) It should be <2 mm thick to permit close electrode spacing, and it must contain a large volume of open porosity, probably >50%. These factors minimize the cell weight and the resistance to ionic flow between the electrodes. (iii) The pore channels should be <20 ~m diam to prevent penetration of the separator by the active powders in the electrodes. (iv) The separator must be chemically stable in the cell environment at the operating temperature (about 450~(v) The meChanical strength and flexibility must be sufficient to allow easy cell assembly, and to withstand the forces exerted on the separator by the volume changes which occur in the electrodes during cycling. (vi) The separator must have a potential cost of <$22/m 2 if the cost goals for a commercial cell are to be achieved.A flexible, fibrous ceramic appears to have the greatest potential for satisfying these requirements. At present, interest is centered on two ceramics, BN and Y203; both of these have proved to be compatible with the cell environment and are available in the form of fibers * Electrochemical Society Active Member. Key words: electrode separator, boron nitride, yttrium oxide, molten salt cell.(2). Boron nitride fabric is currently being used as a separator material, and engineering test cells with BN separators have been operated for periods up to 7800 hr (1, 3, 4). However, it does not appear that the cost goa...