Uncertainties in the composition of Earth, its core and silicate sphere

There is a need to improve our compositional models of the Earth's core in order to constrain the role of radioactive heating in thecore. Present compositional models for the mantle assume that it has overturned throughout Earth's history, such that surface processes randomly and representatively sample the bulk of the mantle, and that there is a negligible bulk compositional gradient in the mantle. Aside from a model for the light element in the core, the uncertainty in the estimate of the 5 most abundant elements in the Earth (O, Fe, Si, Mg, Ni), which make up 95% of the planet's composition (both atomic and by weight), is on the order of 5%, with key element ratios (e.g., atomic Mg/Fe and Fe/Ni) being even better constrained for the silicate Earth. Compositional models for the silicate Earth estimate the abundances of the refractory lithophile elements at 2.16 to 2.8 times that in C1 carbonaceous chondrites, which translates into a 25% spread in the concentrations of Th and U in the silicate Earth (63-83 ng/g Th and 17-22 ng/g U, respectively). However, cosmochemical constraints, both isotopic and chemical, and melting models reveal that the silicate Earth has 80 ng/g Th with about 10% uncertainty on this value. The combination of Sr, Pb and U isotope systematics shows that the silicate Earth has a chondritic Th/U ratio within uncertainties, although, these data do not preclude negligible amounts of U in the core. Combined evidence from geochemical, cosmochemical and recent petrological studies demonstrate that there is negligible potassium in the core. Therefore, based on the above and the silicate Earth's K/U value we estimate that heat-producing elements contribute 20 TW of power to the Earth. Thus, the planetary Urey number (the Earth's ratio of radioactive heat production to heat loss) is on the order of 0.5, and there is negligible heat production in the core.