NASA NTRS1992
Recent work suggests that a large degree of melting is required to segregate metal from silicates, suggesting a connection with the formation of magma oceans. At low pressures metallic liquids do not wet silicate minerals, preventing the metal from aggregating into large masses that can sink. At high pressures, above 25 GPa, the dihedral angles of grains in contact with oxygen-rich metallic liquids may be reduced enough to allow percolation of metal, but this has not been confirmed. Physical models of core formation and accretion may therefore involve the formation of magma oceans and the segregation of metal at both high and low pressures. Models of core formation involving different pressure regimes are discussed as well as chemical evidence bearing on the models. Available geophysical data is ambiguous. The nature of the 670 km boundary (chemical difference or strictly phase change) between the upper and lower mantle is in doubt. There is some evidence that plumes are derived from the lower mantle, and seismic tomography strongly indicates that penetration of subducting oceanic crust into the lower mantle, but the tomography data also indicates that the 670 km discontinuity is a significant barrier to general mantle convection. The presence of the D' layer at the base of the lower mantle could be a reaction zone between the mantle and core indicating core-mantle disequilibrium, or D' layer could be subducted material. The abundance of the siderophile elements in the mantle could provide clues to the importance of high pressure processes in Earth, but partition coefficients at high pressures are only beginning to be measured.