Summary

In this modelling process, it has been shown that the observed inventories and isotope characteristics of noble-gas and other incompatible trace elements in the Earth's mantle and continental crust can be reproduced by a transport-balance model based on a single convecting mantle and starting from accreting matter that had, on average, a chondrite-like composition. An early-formed D" layer, consisting of sunken oceanic crust and including a chondritic regolith component, is essential for this model to be successful.

Such a D" layer could have formed in the late stages of the Earth's accretion by the subduction of basaltic/komatiitic crust loaded with chondritic regolith (as proposed by Tolstikhin and Hofmann, 2005). The subducted material (containing metal as a major constituent of the regolith and therefore intrinsically more dense than the silicate mantle) could have accumulated at the base of the mantle above the core and stabilized there (as proposed by e.g. Samuel and Farnetani, 2003; Mao et al., 2004). This scenario has important geochemical consequences and enables one to reconcile mass balances with whole-mantle convection in the framework of a chondritic Earth model.

The model D" is an important reservoir containing ~ 20% of the terrestrial incompatible trace elements. These include heat-generating U, Th and K, and therefore a significant portion of radiogenic heat is generated at the base of the silicate mantle, stimulating mantle convection and increasing the heat flow from the depleted mantle.

Because of its early formation and efficient isolation, the D" layer is the major store for early-generated radiogenic noble gas isotopes, as well as solar and planetary ones (Q). Fluxes of mass and/or species from the D" layer into the converting mantle and the mixing of D"-derived species with the radiogenic components generated in the mantle and a recycled air component can explain the observed noble-gas isotope patterns, which range from almost solar in some plume-related samples to those typical of MORBs. A small amount of solar-particle-implanted material, less than 0.0005 Me, is enough to maintain the mantle's noble-gas flow throughout geological history.

The model predicts an extremely high degree of mantle degassing, constrained by a low 136Xe(Pu)/136Xe(Pu, U) ratio, < 0.3, in the mantle Xe, especially taking into account that the major portion of 136Xe(Pu) is transferred into the mantle from the D" layer, and so the intrinsic mantle ratio (the ratio for the degassed mantle in the absence of fluxes from D" and the atmosphere) should be much less than the observed value.

For the degassing to reach the levels required by the model, the mantle fractionation rate in the Hadean and the corresponding flux of liquid silicates would need to have been on average two to three orders of magnitude greater than their present-day values. With the exception of helium and (in some cases) neon, the heavier non-radiogenic mantle noble-gas species are mainly recycled atmospheric components.

Although the Sm/Nd and Lu/Hf fractionation of D" relative to that of the BSE is weak, its early isolation causes a substantial isotopic time-integrated effect and thus allows the signatures of 147Sm-143Nd and 176Lu-176Hf in the mantle and crustal reservoirs to be reconciled.

Much work is still needed to verify the proposed origin, history and significance of D" as an important geochemical reservoir, including the modelling of terrestrial regolith accumulation on the surface of an early basaltic crust, its subduction and stabilization at the base of a convecting mantle. The search for such deep ancient reservoir(s) through new isotopic tracers (such as 182W and 142Nd) and noble-metal isotope systematics (e.g. Pt-Os and Re-Os; Brandon et al., 2003) is important. We have merely demonstrated that a D" mass (2.3 x 1026 g) in agreement with seismic observations, and the one-fifth proportion of regolith material to crustal material required to stabilize D", allows fits to geochemical data in an Earth model with whole-mantle convection.

0 0

Post a comment