HR: 14:40h
AN: V23D-05 [Abstracts]
TI: Yes, There Really is a Lead Paradox
AU: * Hofmann, A W
EM: hofmann@mpch-mainz.mpg.de
AF: Max Planck Inst. for Chemistry, Postfach 3060, Mainz, 55020
Germany
AU: Hemond, C
EM: chhemond@univ-brest.fr
AF: UMR 6538 "Domaines oceaniques" I.U.E.M., Place Nicolas Copernic, Plouzané, 29280
France
AU: Sarbas, B
EM: sarbas@mpch-mainz.mpg.de
AF: Max Planck Inst. for Chemistry, Postfach 3060, Mainz, 55020
Germany
AU: Jochum, K P
EM: kpj@mpch-mainz.mpg.de
AF: Max Planck Inst. for Chemistry, Postfach 3060, Mainz, 55020
Germany
AB:
In the early days of isotope geochemistry, Allègre (1969) noted that 207Pb/204Pb-206Pb/204Pb ratios of oceanic basalts,
including MORB, differ from the 4.57Ga "geochron", such that the U/Pb ratios of their sources must have increased at a
younger age. Allègre called this the "lead paradox" because it appeared inconsistent with the depleted nature of MORB
sources, which should have decreased U/Pb ratios, because uranium is expected to be more incompatible than lead. However,
experimental partitioning data for lead in silicates suggest that Pb is also highly incompatible, and this might mean that
the paradox never existed in the first place. Hofmann et al. (1986) examined the relationship of lead relative to other trace
elements in a suite of 30 MORB and 11 OIB samples and concluded that, during mantle melting, the bulk partition coefficient
of Pb is similar to that of Ce. However, Sims & DePaolo (1997) questioned the validity of this assessment, but to our
knowledge no systematic reevaluation of Pb partitioning has been undertaken. We reassess Pb partitioning in over 500 OIB
samples (selected from http://georoc.mpch-mainz.gwdg.de/) and in specific MORB regions. We use logarithmic abundance plots of
Pb versus a series of other elements (U and REE). The slopes of these correlations range from >1.0 to <1.0. The
correlation of the element with a slope closest to 1.0 is linear and its partition coefficient is closest to that of Pb. This
method applies to both partial melts and their residues. Element pairs with slopes different from unity will not be strictly
linear, but the deviations from linearity are small and of little significance. The new results, based on vastly larger and
more representative set of data, are remarkably similar to those of the original paper by Hofmann et al. (1986). The bulk
partitioning of lead is closest to that of cerium, with Ce/Pb = 28 +/- 9 in most OIB. Equivalent MORB evaluations show that
lead partitioning varies slightly, resembling Ce in some cases, and Pr in others. Ce/Pb ratios also vary slightly (28 at
equatorial MAR, 25 at 24°MAR, 25 at 10-24°N MAR, and 17 at 18-20°S CIR). This means that Pb partitioning
during mantle melting is controlled most likely not by silicate minerals but by residual sulfides. These results show that
the lead paradox is alive and well: Lead is much more compatible than uranium in both OIB and MORB producing environments.
The paradox can be explained in part by the completely different behavior of lead during subduction related melting and
production of continental crust, perhaps by volatile-related mobilization and transfer of lead. However, the problem remains
that the bulk continental reservoir is probably too radiogenic in Pb to balance the Pb composition of the bulk MORB and
OIB-source reservoirs (Rudnick and Goldstein, 1990). We speculate that the necessary unradiogenic reservoir may exist in the
D" layer at the base of the mantle. It may have formed early in Earth history by partial melting in the presence of residual
Ca-perovskite, which has very high partition coefficients for U and Th (Corgne et al., 2005). In contrast with Corgne et al.,
who suggest a Ca-perovskite-bearing, high-U/Pb, residual reservoir in the mantle, we propose that a low-U/Pb D" layer formed
by downward migration of dense melts formed in the deep mantle in equilibrium with Ca- perovskite.
DE: 1020 Composition of the continental crust
DE: 1025 Composition of the mantle
DE: 1040 Radiogenic isotope geochemistry
SC: Volcanology, Geochemistry, Petrology [V]
MN: Fall Meeting 2005