A laboratory model for iron snow in planetary cores

Top-down solidification has been suggested in the liquid cores of small planets, moons, and large asteroids. An iron snow is then thought to exist, consisting of the crystallization of free iron crystals at the top of these cores and of their settling in a stably stratified ambient, until they remelt in a hotter, deeper region. This inward crystallization and associated buoyancy flux may sustain dynamo action by convection below the remelting depth. However, thermal evolution models are up-to-now oversimplified, assuming a constant-in-time and homogeneous-in-space buoyancy flux at the bottom of the snow zone. We have shown from analog experiments that the buoyancy flux is heterogeneous in time and space, with intense snow events, corresponding to an explosion of frazil-ice,  separated by quiescent periods where the snow zone supercools. We found that a wide range of crystal sizes exists, with large crystals overshooting the convection region and challenging the thermodynamic equilibrium hypothesis underlying the evolution models. The spatio-temporal variability of the energy source obviously impacts the shape and intensity of the generated magnetic field, which may provide alternative explanations for the observed and surprising features of Mercury's and Ganymede's magnetic fields.