Liquid metal batteries for large-scale stationary storage

Liquid metal batteries (LMBs) are high temperature systems consisting of liquid metal electrodes and a molten salt ionic conductor. The densities are chosen in such a way that a stable density stratification of the inmiscible layers results. LMBs were considered mainly as part of energy conversion systems in the 1960s and have only recently received renewed interest for economic large-scale storage. Typically, LMBs allow for high current densities due to the fast kinetics at liquid/liquid interfaces and the rapid mass transport in fluids.

Our work concentrates on the fluid dynamic aspects of this cell type with a special focus on the effects and properties of the Tayler instability (TI) and on electro-vortex flows. Both phenomena are driven by electromagnetic forces and should be considered for large cells. Due to the completely liquid interior of LMBs, fluid flow is an important aspect of their operation. It can be beneficial, when enhancing mass transfer in the cathode, or it might have harmful consequences, if the integrity of the electrolyte layer is disrupted. The latter case can result from the action of the current-driven TI or electrically driven vortex flows. We therefore studied the characteristics of the TI as well as some exemplary cases of electro-vortex flows using an integro-differential approach implemented in the open source library OpenFOAM. The TI occurs if a critical value of a dimensionless parameter Ha, the Hartmann number describing the ratio of electromagnetic to viscous forces, is exceeded. The critical Ha is lowest for an infinitely high vessel and corresponds to a total current of approx. 1 kA in the case of Na. Decreasing the aspect ratio increases the critical Ha and thereby the critical current since the wavelength selection for the TI becomes more and more restricted.

As mentioned above, current densities in LMBs are typically very high. A current density of 10 kA/m2 is a characteristic value for a Na|NaI-NaCl-NaF|Bi-system and results in an approximately 10 mm thick sodium layer transferred per hour from the anodic to the cathodic compartment. Depending on the design capacity and cell area, aspect ratios of the anodic compartment up to one seem imaginable. While flat enough cells will not suffer from TI induced short circuits, for taller ones stabilization measures can be applied to prevent negative consequences.

Using thin feeding lines to contact relatively large current collectors will most certainly result in inhomogeneous current density distributions in the fluid. They will generate electro-vortex flows that may again compromise the integrity of the electrolyte layer. A careful distribution of the charging current by several wires should solve the problem. Properly designed electro-vortex flows might even be used to gently stir the cathode thereby increasing mass transfer and improving cell performance.