U13C-01
Super-chondritic Sm/Nd in Mars, Earth, and the Moon
Differences in the isotopic abundance of 142Nd in planetary materials reflect the chronology and mechanisms by which crustal reservoirs formed during or shortly after accretion. Interpretation of these differences assumes that the terrestrial planets have a composition identical to that of chondritic meteorites. The ca. 20 ppm difference in 142Nd/144Nd between chondrites and the Earth's upper mantle may thus indicate very early (<30 Ma) formation and recycling of the first terrestrial crust [1]. However, it has also been suggested that the Earth may in fact be slightly non-chondritic in bulk composition. We present high-precision 142Nd data for 16 martian meteorites showing that Mars also has a non-chondritic composition. Meteorites belonging to the Shergottite group define a planetary isochron yielding an age of differentiation of 40±18 Ma for the martian mantle. This isochron does not pass through the chondritic reference value (ε142Nd=-21±3 ppm; 147Sm/144Nd=0.1966 [2]). The Earth, Moon and Mars appear to have all accreted in a portion of the inner solar system with 5% higher Sm/Nd when compared with material accreted in the asteroid belt. Such chemical heterogeneities in the accretion disk may have arisen from turbulent sorting of objects such as chondrules [3], which typically have high Sm/Nd (e.g. [4]). [1] Boyet, M. & Carlson, R. W., Science 214, 427-442 (2005). [2] Carlson, R. W., Boyet, M. & Horan, M., Science 316, 1175-1178 (2007). [3] Hewins, R. H. & Herzberg, C. T., Earth Planet. Sci. Lett. 144, 1-7 (1996). [4] Amelin, Y. & Rotenberg, E., Earth Planet. Sci. Lett. 223, 267-282 (2004).
U13C-02
Volatiles in the Lunar Volcanic Glasses, Evidence for the Presence of Indigenous Water in the Moon's Interior
The composition and origin of the lunar volatiles has been the subject of considerable interest and one of the most important unsolved questions regarding the formation of the Moon. Volatiles influence mantle melting, magma crystallization and volcanic eruption, and their abundances and spatial distribution provide important constraints on models for the thermal and chemical evolution of the Moon's interior. The general consensus today is that the Moon formed and evolved through a single or series of catastrophic heating events in which most of the highly volatile elements were either stripped or evaporated away. Hydrogen being the lightest element is thought to have been completely lost during this period. Most of the geochemical inferences about the deepest section of the moon have been based on studies of the most primitive melts erupted on the Moon's surface, the lunar volcanic glasses. We reported new volatile contents (CO2, H2O, F, S, Cl) for the lunar volcanic glasses, which by virtue of SIMS analysis provide improved detection limits by almost an order of magnitude. All volcanic glasses have CO2 below or within the 2 standard deviation of the detection limit for our runs (6 ppm CO2) and will not be considered further. The volatile contents after background correction have the following ranges: 4-46 ppm (±2 ppm 2σ) for H2O; 4-40 ppm (±0.2 ppm 2σ) for F; 115-576 ppm (±3 ppm 2σ) for S; and 0.06-2 ppm (±0.03 ppm 2σ) for Cl. Two outstanding features of the data is the significant correlation among H2O, Cl, F and S contents, and the clear relationship between the volatile and the major and trace element contents of the glasses. Most importantly, the correlations among the volatiles suggest that the measured H2O content in the volcanic glasses is indigenous. To test this hypothesis we determined concentration profiles for the volatiles within a single selected glass bead. Our data define concentration profiles with decreasing volatile contents from the core to the rim of the glass bead. This is especially significant for H2O, which decreases from ~30 ppm in the center to ~14 ppm in the margin of the bead, supporting the hypothesis of the indigenous origin of H2O subsequently affected by degassing processes during melt decompression and eruption. To our knowledge this is the first definitive confirmation of the presence of indigenous H2O in primitive lunar basalts. Our results suggest that, contrary to the prevailing ideas, the bulk Moon is not uniformly depleted in highly volatile elements, and the presence of water in particular must be included to constrain models for the thermal and chemical evolution of the Moon's interior.
U13C-03
Geodynamic modelling of volatile-rich mantle melts
Many difficulties remain in modelling the evolution of mantle chemistry, particularly with respect to the treatment of silicate melts. Given plausible scales of mantle heterogeneity and reasonable melt migration velocities (relative to mantle flow), it is likely that most silicate melts are not in thermodynamic equilibrium with their source rock. These effects are more significant for chemically derived melts, where time scales for diffusive equilibrium are much greater than that of thermal equilibration. Further, two-phase continuum approaches break down in this regime, as partial melts give way to dike formation, fingering instabilities, and lensing. Here we describe a new method for the theoretical treatment of the dynamic and chemical stability of silicate melts in the mantle. Our method explores the feedback between dynamic transport (due to gravitational forcing and differing material properties between the solid and melt) and chemical reactions that convert solid to melt (and vice-versa) due to thermodynamic disequilibrium. This method can be used to address the stability of silicate melts in a wide variety of tectonic environments under many different chemical conditions.
U13C-04
Magma dynamics with the Enthalpy Method
Forward models of magma genesis and transport through the mantle are an important tool for studying the dynamics of plate boundaries because they have the potential to track geochemical signals in magma from depth to the surface or, at least, to the Moho. To accomplish this, however, requires a self-consistent thermodynamic closure for the conservation equations for mass, momentum, composition and enthalpy of the magma/mantle system (e.g. McKenzie '84). Such a closure should simultaneously account for all modes of melting and freezing, allowing magma to interact with a cold thermal boundary layer, the lithosphere. I have implemented a model in which I assume a two-component system that is in local thermodynamic equilibrium everywhere within the domain. I thus close the system of equations by prescribing a binary phase diagram that should approximate the mantle. This approach is known as the Enthalpy method (e.g. Alexiades and Solomon '93). It differs from previous treatments of magma dynamics in that it captures polybaric, polythermal, two-component melting and freezing with one parameterization: the phase diagram. I have implemented this system in 2D and configured it to model a mid-ocean ridge. Melting occurs by adiabatic decompression at a rate determined self-consistently through conservation of energy and the phase diagram. Melt segregates from the mantle matrix and rises buoyantly. The existence of a freezing boundary, where rising magma reaches its solidus temperature, results in lateral melt transport towards the ridge axis (Sparks and Parmentier '94). Similar effects are expected in arc models, and should be accessible with the same general approach. A disadvantage of the Enthalpy method is that it requires the system to be in equilibrium everywhere, which makes it computationally difficult to solve, and is at odds with observations of disequilibrium of major elements in primitive basalts from ridges. An advantage is that both melting and freezing are treated according to a standard parameterization: the two-component phase diagram. While still fairly simple to model, two components allow for univariant melting, which is not the case with a single component (Sramek et. al '06). In the ridge model, I have chosen to use a binary loop phase diagram, however this choice is not due to any constraint from the method. http://www.damtp.cam.ac.uk/user/rfk22
U13C-05
The noble gas "subduction barrier" revisited
It has been proposed that an important subduction of atmospheric noble gases in the mantle occurred during Earth's history, on the basis of the measurements of light xenon isotopes in CO2 well gases. Moreover, the fact that the 38Ar/36Ar ratio is atmospheric in all oceanic basalts, even for uncontaminated samples (e.g. with high 20Ne/22Ne), may also suggest that a massive subduction of atmospheric argon occurred, if the primitive Earth had a solar-like 38Ar/36Ar. This also implies that the atmosphere suffered a massive gas loss accompanied by mass fractionation (e.g. hydrodynamic escape) after mantle degassing or that a late veneer with an atmospheric composition occurred. Such a hypothesis is explored for rare gases, by developing a model in which degassing and air subduction started ~4.4Ga ago. In the model, both radiogenic and non-radiogenic isotopic ratios are used to constrain the subduction flux and the degassing parameters. It is shown that subduction and massive contamination of the entire mantle is possible but implies that the 40Ar/36Ar and the 129Xe/130Xe ratios were higher in the past than today, which is not observed in Archean samples. It also implies that the sediments and the altered oceanic crust do not lose their noble gases during subduction or that the contaminated mantle wedge is mixed by the convective mantle.Moreover, such a model has to apply to the OIB source since it shows the same isotopic signature of argon and xenon. A scenario where the isotopic composition of the argon and xenon were acquired before or during accretion is therefore preferred to the subduction hypothesis (e.g. irradiation by solar wind for argon).
U13C-06
The Magnesium isotopic composition of the ocean and its consequences on chemical geodynamics
It has been long recognised that the input of Mg in the ocean by river is removed by precipitation of Mg-rich bearing phases, either directly from the ocean such as dolomite or through hydrothermal circulation in the oceanic crust. By subduction, a leakage of Mg towards the mantle becomes a part of the global geochemical cycle and unlike the large-ion lithophile elements, carbon or sulfur, Mg is poorly stripped off from the subducting slab. The fluxes (either riverine input or heat budget and hydrothermal output) suggest a residence time of Mg in the ocean in the order of 10 Ma but the removal of Mg from the whole mantle at ridges is far greater (20 to 50 times) than the re-incorporation of Mg through subduction zones. Nevertheless, none of these fluxes will have a significant impact of the Mg inventory of the mantle since less than 2% of Mg has seen the surface during the whole Earth's history. Therefore, any Mg isotopic heterogeneity in mantle rock could be viewed has primordial or the result of processes related to the production and/or the transfer of these rocks to the surface, including late stage metasomatism. The recent report of a range of 4‰ in δ26Mg for mantle material is rather surprising and great care has to be taken when considering Mg-isotopic composition of silicate material because of unaccounted chemical bias during the analysis and related inaccuracy of the measurements. Nevertheless, 2 viable mechanisms could account for such a range in mantle rocks. First, a late stage metasomatism and associated large kinetic isotopic fractionation, as suggested by Pearson et al. (doi:10.1016/j.chemgeo.2005.09.029). However, Mg is a major element and it has to be replaced for mass/charge balance purpose by the diffusion of Fe in the opposite direction. The lack of relationship between Forsterite content and δ26Mg in mantle olivine suggests that late stage metasomatism and kinetic fractionation associated with diffusion are unlikely to induce large Mg isotopic heterogeneity in the mantle, but this mechanism should be more precisely investigated. Second, the introduction of fractionated Mg from the ocean. We find the Mg isotope composition of seawater to be constant, with a δ26Mg = -0.82±0.10 ‰ relative to the DSM3 standard, about 0.5‰ lighter than chondrite. The re-incorporation of Mg by the formation of serpentine occurs in the top 10% of the oceanic crust. In addition, dolomites are even further depleted in heavy isotopes than seawater. Considering the modern production rate of oceanic crust and the Mg isotopic oceanic budget (e.g. Tipper et al., doi:10.1016/j.epsl.2006.07.037), an anomaly of up to -0.4‰ in δ26Mg is carried by the top of the slab. Since magmatic differentiation has been ruled out as a process introducing Mg isotopic fractionation (Teng et al., doi:10.1016/j.epsl.2007.06.004) and that the range in Mg isotopes in chondritic material is small, the recycling of the oceanic crust and overlying sediment is likely to be the dominant mechanism responsible for variation in the mantle.
U13C-07
Chemical Erosion in the Himalayas for the Past 4 Million Years Studied by Pb and Nd Isotopic Stratigraphies
A record of the 143Nd/144Nd ratio of Northern Indian Ocean seawater and associated detrital sediments has been obtained for the past 4 Ma using a new method of differential dissolution on marine sediments. We studied ODP Sites 758 and 757 both located on the Ninetyeast Ridge. At Site 758, the εNd and δ18O curves fluctuate in conjunction during the glacial-interglacial periods. The largest variation occurred during the last 20 kyrs where εNd varies from -7.5 at Last Glacial Maximum to -10.5 during the Holocene, whereas the amplitude of the variation of the seawater signal is less than 1 εunit 3 Ma ago. The correlation between maxima and minima of δ18O and εNd is excellent (r=0.92). However, a detailed comparison of the two signals show that εNd and δ18O vary simultaneously during warming while εNd is delayed with respect to δ18O during cooling. The southern ODP Site 757 shows little variations in Nd isotopic ratio. We thus interpret the εNd fluctuations at Site 758 as being linked to the erosion regime in the Himalayas-Tibet rather than to variations in the intensity of the deep ocean conveyor belt which flows from the South to this area. Such variations were probably governed by storage of more ice in the Himalaya-Tibet highlands during glacial periods and by ice melting and the enhancement of monsoon rainfall during interglacials. A simple quantitative model assuming that seawater Nd is a mixture of Nd that was chemically eroded in the Himalaya-Tibet with Nd coming from Indonesian island arcs indicates that chemical erosion in the Himalaya-Tibet was 2 to 4 times more intense during interglacial than during glacial periods. We have shown by chemical tests that the same technique of partial dissolution can be also applied for lead isotopes. At Site 758, the lead isotopes ratios fluctuate following glacial interglacial alternance, but within a very small range. Indeed, 206Pb/204Pb, 207 Pb/204Pb , 208Pb/204Pb varies from 18.9 to 19.0, 17.735 to 17.770, 39.23 to 39.42 respectively. These variations are extremely well correlated suggesting a binary mixture. Using the lead isotopic ratio from the various parts of Himalaya-Tibet (Gariépi et al., 1985), we suggest that the two component are the Trans-Himalaya-Belt (THB) and the High-Himalaya-Chain (HHC), the HHC signature being transported by the Ganges and the THB one by the Bramaputra. During Glacial chemical erosion on the HHC was reduced by a factor of 6 to 7 while the chemical erosion on the THB was reduced by a factor of 2 only. This suggests that very large permanent glaciers were present on the Himalayas during Glacials, there were probably not as large in the Southern Tibet.