William F McDonough
Department of Geology
University of Maryland, College Park, MD 20742

Abstract: Estimating the composition of the Earth and that of its differentiated parts (core, mantle, curst) requires integrating physics and chemistry of the Earth with cosmochemical insights; this approach is not without controversy nor considerable uncertainties. The earth's composition is the integrated product of that which condensed and accreted from a collapsing, co-rotating solar nebula, with accreted materials having had residence in or transited across the 1AU region of the solar system. The guiding principals for estimating the planet's composition are (1) to constrain models to be consistent with physical observations of the planet (i.e. density profile (seismology), moment of inertia (geodesy), magnetosphere and paleomagnetic record (geomagnetism)), (2) to use chondrites (undifferentiated meteorites) to constrain the relative abundances of refractory elements (high temperature condensates, ~40 elements) in planets and to provide perspective on redox and volatility processes during parent body formation, and (3) to use mantle and crustal rocks to identify the behavior of elements during differentiation. The composition of the crust and mantle (silicate earth) is initially established from these principles followed by a model for the bulk planet. With these results the core composition is calculated based on a mass balance relationship between bulk planet and the silicate shell.

The principle uncertainties in developing compositional models for the earth and its reservoirs (core, mantle, curst) come from (1) variations in the compositions of crust and mantle samples, (2) variations in the compositions of chondritic meteorites, and (3) the intrinsically biased nature of the samples we have for analyses. There is a constant concern as to whether we have sampled/identified/characterized the full range of components in the earth (i.e., are there hidden reservoirs with an important complement of elements?). The chemical variability in mantle samples leads to a range of compositional models derived from reconstructing mantle melting trends. This, in turn, yields uncertainties in the absolute concentrations of elements in the earth. An important goal of these models is establishing the absolute concentrations of the refractory elements, given the accepted paradigm of planets having chondritic proportions of these elements. Underpinning our models are the critical observations provided by several radiogenic isotope systems (i.e., 146Sm- 142Nd, 147Sm-143Nd, 176Lu-176Hf, 206Pb-208Pb) that demonstrate the silicate earth to have chondritic proportions of refractory elements. Thorium and uranium are refractory elements and therefore knowing their absolute abundances places tight constraints on the planetary production of geoneutrinos. Present modeling sets the planetary abundances of refractory elements at ~1.8 times that of C1 chondrites abundances, with ~80 and ~20 ng/g of Th and U, respectively, in the silicate earth and no Th and U in the core.

A critical concern regarding the composition of the planet focuses on the role of radioactive elements in the energy budget of the earth. There is a desire by some to have radiogenic heating in the core in order to find an energy balance for the earth. The power needed to drive the geodynamo, as based on the heat flow from the core with no radioactive heating, is believed to be insufficient over the history of the earth, as constrained from the existence of ancient rocks with a record of strong magnetic fields and their implications for the presence of an inner core. In contrast, however, other models posit that less power is needed to drive the geodynamo because of how ohmic losses are treated; these models are finding that no radiogenic heating is required to drive the geodynamo. Petrological experiments that report on the possibility of radioactive elements being sequestered into in the core are simply plausibility arguments that permit, but do not require, the presence of such components in the core or other regions of the earth's deep interior. Plausibility arguments such as these must be coupled with corroborating evidence that are free of negating chemical consequences.

A second concern comes from the perceived limited contribution of radioactive heating to the total power output of the planet. It is generally recognized that the global heat flow appears to be twice as strong as that derived from radiogenic heating of the earth (i.e., a Urey number of ~0.5, based on conventional compositional models of the earth (as above). However, more recent modeling has concluded that the widely accepted value of 44 TW for the surface heat loss is too high and that this value is more likely ~31 TW. Consequently, this finding implies a higher Urey number for the above compositional model for heat-producing elements in the earth. An alternative compositional model has a Urey number of ~1 and a core composition that contains K and U.

Testing compositional models for the earth requires insights from a wide range of geophysical, geochemical and cosmochemical studies. Results from the KamLAND experiment, as well as from other geoneutrino detectors, are critical in establishing the abundances and distribution of radioactive elements in the earth, which in turn provide additional insights into the absolute abundances of the refractory and volatile elements in the planet. I will summarize a series of recent experiments that we conducted that specifically examine the potential of extracting long-live radioactive elements into the core.