We are reporting preliminary results of the first study of the chemical effects of glow-discharge electrolysis (GDE) in molten salts. Typically, the anode is a short distance (10 mm) above the surface of the melt, the space being occupied by argon at reduced pressure (25 mm), and a substantial current (35 mA) is passed through the cell on application of a high voltage-around 350 V, depending on conditions. We are investigating a number of systems including "hydrate" melts such as the bisulphate eutectic, NH4HSO4-KHSO4 (83:17), and also anhydrous nitrate melts, KNOa-NaNOa (1:1) and KNOa-Ca(NOa)2 (2:1). There is no truth in an earlier report 1 that GDE is difficult to establish with molten-salt electrolytes; and therefore it should be possible to obtain interesting chemical effects similar to those reported for GDE in aqueous systems 2.
One reaction of interest is the anodic evolution of iodine from solutions of iodide in molten nitrates. The yield is unaffected by temperature but is markedly dependent on the iodide concentration, as can be seen from the data in Fig. 1, which refer to the KNOa-Ca(NOa)2 (2:1) melt. Following the usual convention for GDE 2 , the current efficiency is expressed as a "G-value" in equiv. F "1 , and from the reciprocal graph (b) it is seen that as 1/[I-] -* 0, 1/G ~ 0.33. This corresponds to a hypothetical limiting G-value of 3 equiv. F "1 as the iodide concentration tends to infinity.
This high yield of anodic oxidation (in excess of Faraday's law) is understood in terms of the iodide ions scavenging the free radicals produced in the melt by the glowdischarge. One mechanism which could account for the above results will now be discussed. Ar รท ions entering the melt from the discharge undergo a charge transfer reaction:
Assuming that iodide ions are reacting with all the available free radicals, the experimental limiting G-value can be predicted on the basis of the following reactions: NOa ~ NO2 + 0 NO2 + I'-~ NO~ + ยฝ12 (see ref. 3) O+2I'-* 0 2-+I2
(2)
(3) (4)