In the case of metallothermic reduction of oxide XO_x-1(s) (X=Ti, Ta, etc.) by electro-deposited liquid metal Me(=Ca, Li or their alloys), the it accompanied with dissolved Me at the vicinity of the cathode is essential, where the oxide is placed close to the cathode. In this mechanism, the dissolved Me works effectively to reduce XO_x-1, even if these solids do not have any direct electronic contact with the cathode. In the case of Li-LiCl-KCl-Li₂O system at 723 K, colloidal microparticle dissolution of Li occurs in addition to true dissolution. Recently, the in-situ Raman spectroscopy was carried out, and obtained from colloidal Li were in agreement with the lithium cluster, Li₈. The colloidal Li formation strongly depends on the linear current density and that the current concentration should cause its formation. Although colloidal Li occurs at the liquid Li and molten salt interface, the morphology of liquid Li on the electrochemically precipitated cathode surface has not been well understood as root cause. High-speed photography with high-resolution image quality has been becoming possible by recent C-MOS image sensor which has been improved by recent digital optical technology. In this study, we investigated the morphology of liquid Li in molten LiCl-KCl-Li₂O by recording electrode surface image obtained with digital microscope at high speed.

To elucidate the mechanisms of interfacial morphology on the basis of three-dimensional shape observations, the electric resistance furnace was designed to allow the phenomena occurring within to be observed directly. The changes induced in the interface of electrodes were recorded at a rate of 500 frames per a second (fps) and a resolution of 640 × 480 pixels using a high-speed microscope (the field of view of 610 μm × 460 μm, and the length per pixel of 1.04 μm) was employed. LiCl (>99.0%) and KCl (>99.5%) were used for the melt. The eutectic mixture of LiCl–KCl (= 59:41 mol%, m.p. = 625 K) contained in a borosilicate glass crucible with flat surface was dried under vacuum at 573 K for more than one night to remove residual water and then melted. Li₂O (>99.5%) was directly added to the melt as an O^2-ion source in the concentration range of 0.1–1.0 mol%. The experimental temperature was 723 K. The interface temperature was measured with a K–type thermocouple with a glass protection tube. All the experiments were conducted in an Ar atmosphere (>99.9995%). The working electrode was a Mo rod (ϕ 1.5 mm, 99.95%) for cyclic voltammetry and it was used as a cathode for electrodeposition of liquid lithium. The counter electrode was a graphite rod (ϕ 10 mm). The Ag+/Ag reference electrode was employed, i.e. a silver wire (ϕ 1.0 mm, 99.99%) was immersed in the LiCl–KCl eutectic melt containing 0.5 mol% AgCl (>99.5%), which was set in a borosilicate tube. Electrochemical measurements were performed using an automatic polarization system (Hokuto Denko Corp., HZ-5000).

From the snapshots corresponding to each potential in cyclic voltammograms, blue “fog” was observed are observed at about −2.46 V (vs. Ag⁺/Ag) with the electrodeposited Li, notably during Li dissolution reaction. The amount of colloidal Li generation decreased by Li₂O addition. Chronoamperometry was conducted at -2.55 V vs. Ag⁺/Ag using a Mo rod electrode for 1.0 s. Photographs of the transient behavior of electrode surface in LiCl-KCl melt are shown in Fig. 1. It can be seen that droplets are formed and grow on the surface of the Mo electrode as the current is generated. It should be noted that the liquid phase precipitates grow heterogeneity on the Mo surface. Considering the “electrocapillary” theory, due to the electrostatic attraction between the droplets charge and charge in the electrode, the energy of their interface is reduced, and the wetting of the electrode should be enhanced. Nevertheless, metal droplets do not wet and spread on the electrode surface completely.