“…At this stage, the crust composition treatment and thermodynamic closures remain simple while being qualitatively coherent (and probably quantitatively at the first order) to regain the expected physical behavior of the coupling at the interface between the pool and its crust. Indeed, it is shown in [20] that, for a given pool composition, the variation of the liquidus temperature of the oxide pool remains small (less than 50 K). Plus, we show in the following that for usual reactor case calculations, the metallic pool layers do not lose enough power in the crust in such a way that the interface temperature of its facing oxidic crust stays well above the solidification temperature.…”