| The percolation potential of iron-sulfur melts through mantle minerals was investigated experimentally at high pressures and high temperatures. To characterize percolation ability, experimental products were analyzed for phase and composition and the dihedral angle formed by quenched iron-sulfur melt in contact with solid phases was measured.; To examine the effect of phase changes on dihedral angle, the Homestead meteorite was used as a starting material. As pressure and temperature were increased, the phases in Homestead underwent mineralogic changes analogous to those inside the Earth's upper mantle. These phase changes were found to have no significant effect on the percolation ability of the iron melt. The dihedral angle of the melt phase averaged 108{dollar}spcirc{dollar}, well above the 60{dollar}spcirc{dollar} cutoff for efficient percolation. Thus, the upper mantle of Earth is not percolative by these melts.; The phase change from {dollar}gamma{dollar}-spinel to magnesium perovskite and magnesiowuestite that defines the upper mantle/lower mantle boundary in Earth was also investigated using Homestead. This phase change does significantly affect the dihedral angle, which drops to {dollar}sim{dollar}71{dollar}spcirc{dollar} for iron-sulfur melt in contact with perovskite. So, the lower mantle of Earth may be percolative under the right conditions.; A synthetic mixture of iron and sulfur in contact with magnesiowuestite was used to examine the effect of changes in pressure, temperature, and composition in the absence of phase changes. Results show that increased pressure increases dihedral angle, while increased temperature decreases dihedral angle. Changes in the composition of the solid phase have no significant effect while changes in the composition of the melt do have a significant effect at low pressure. At high pressure, melt composition does not have a significant effect. In Earth's lower mantle, the pressure effect is stronger than the temperature effect. Therefore, if the pressure effect continues down to the core-mantle boundary, the lower mantle is not percolative. If, on the other hand, the pressure effect ceases, the lower mantle would be percolative.; These results support models of core formation based on segregation by rainfall in a partially or wholly molten upper mantle, followed by diapiric sinking or percolation in the lower mantle. |
