Pfc/rr-92-10 an Evaluation of the Feasibility of Liquid Metal Divertors

The divertor technology has become the focus of concern for prospective steady state tokamak reactors such as the International Thermonuclear Experimental Reactor (ITER). Namely, the imposed heat flux and particle flux conditions cast doubt on the feasibility of any solid surface divertor. Thus, the old idea of liquid metals (mainly lithium and gallium) as the divertor surface protective materials has been revived, and different concepts of liquid metal divertors have been proposed. The aim of this work is to evaluate the feasibility of the existing concepts from both the physics and engineering points of view. A model is developed for the edge plasma which includes the important interaction mechanisms, such as the backscattered energy flux from the divertor plate, the plasma momentum loss due to the charge exchange process, and the correct form of the energy transmission coefficients across the plasma sheath. This model is applicable to all types of divertors. Its results show that the edge plasma temperature will be more than 100 eV for a two-null steady state ITERlike tokamak with tungsten divertors, higher than was estimated by the previous engineering models. This outcome turned out to be in agreement with the recent simulation reported by the ITER team using the much more complicated Braams code. This implies that for the imposed heat load of 70 MW the physical sputtering erosion rate of a 20 m 2 tungsten plate is about 0.56 mm/day (without taking into account nonuniform redeposition). Similarly, the erosion rate for a beryllium plate is estimated to be 2 mm/day. The hydrogen build-up in the liquid metal surface is addressed in order to determine the liquid metal divertor recycling coefficient needed for the edge plasma modeling. It is concluded that liquid lithium is a good hydrogen getter due to the precipitation of solid hydride LiH. For the condition of 70 MW per the 20 m 2 plate, this leads to an expected hydrogen recycling time of several tens of minutes. Hence, if the liquid lithium can be efficiently refreshed, the divertor may be expected to operate in a low recycling mode. However, since such efficient tritium extraction technology is probably unavailable (at the rate of about one mole of tritium per second), the tritium inventory becomes a concern as it reaches several kilograms in the divertor system. In addition, changes of lithium properties due to hydride precipitation may also occur. Thus, liquid lithium is not a favorable material for the liquid metal divertor. On the other hand, liquid gallium does not have a similar problem owing to its low hydrogen solubility and the decomposition of hydrides within the temperature window of interest. The evaluated short hydrogen recycling time of the order of milliseconds indicates that the liquid gallium divertor is suitable for high recycling operation. Simulation results show that dense (2 . 1019 m3 ) and relatively cool (57 eV) edge plasma can be reached at the gallium divertor (with two divertors for each null point). Evaluation of existing liquid metal concepts against the 1 issues of evaporation, sputtering, blistering erosion, unipolar arc erosion, MHD instability, heat transfer, and major disruption concludes that the liquid gallium droplet curtain divertor appears to be the most feasible concept. The protective film divertor concept (of film thickness 5 mm) suffers mainly from the problem of MHD instability when the film speed is higher than about 10 m/s which is needed to avoid the blistering erosion, even though the heat transfer requirement is a much lower speed of the order of 1 m/s. It is pointed out, however, that there is uncertainty about insufficiency in helium ash removal even with a large pumping panel area and perfect vacuum pumps. A new vacuum pumping scheme is thus proposed, which not only meets the vacuum pumping requirement with much smaller pumping panel area, but also offers the means to control the divertor recycling coefficient.