THE FLOW of ac or dc through conductance cells containing molten salts is invariably attended by appreciable polarization effects at the electrodes. The significance of the resultant impedances is accentuated by the high conductances of most molten salts and is best minimized by the use of large electrodes in cells of large cell constant, eg as in the 'capillary cell'.lb There are, however, several conducting systems, for example, molten fluorides and oxides, which are so corrosive that no stable insulating or refractory material is available for the construction of the capillary cell. Instead, all-metal cells have to be used, and these are normally of such low cell constant that the resultant electrolytic resistance is similar in magnitude to the polarization impedances.**7 The accurate elimination of these electrode impedances is therefore essential.
There is some confusion in the literature as to how this may best be achieved. The problem of polarization resistances in conductance cells has been discussed by Krager and Weisgerber,8 by Ives et al9 and by Robinson and Stokes,lO and there is general agreement that the equivalent circuit of a conductance cell can be represented as in Fig l (a), where & is the electrolyte resistance, Cd, is the series summation of the double layer capacitances at both electrodes, Rk and -W-are components of the total faradaic impedance ,I1 deGned below, and C,, represents the small 'stray' capacitances which arise, for example, between leads. For the case considered here, where R, is small, this last term C, is entirely negligible and will not be considered ftirther. (As R, becomes very small, the mutual inductance between the leads may become significant but this can be eliminated by appropriate design of the conductancebridge circuit.'? FIG. 1. Ekpivalent circuits of conductance cells.