V43F-01 INVITED
142Nd-182W record of terrestrial samples: Implications for Early Earth evolution
Evidence for rapid differentiation of the planet Earth comes from studies of short-lived chronometers, in particular 182Hf-182W (T1/2=9Ma) and 146Sm-142Nd (T1/2=103Ma) used to constrain the age of core formation and the silicate differentiation, respectively. Although the major fractionation between Hf and W occurred during the core formation, differentiation within the silicate earth is expected to create Hf/W fractionation, W being more incompatible than Hf. Boyet and Carlson (Science 309, 2005) showed that all terrestrial samples present 20 ppm excess of 142Nd relative to chondrites. Assuming that the Earth formed by accretion of material similar in refractory element concentrations to chondrites, they connect this signature to a global differentiation of the silicate earth during the first 30 Ma of the solar system history. Old mafic rocks from Isua, Greenland have a 142Nd excess 15 ppm higher than found in most other terrestrial rocks suggesting that the early-Earth's mantle was even more depleted than the post-3.8 Ga mantle. These anomalies provided the first indisputable evidence that differentiation of the Earth occurred while 146Sm was present and even when 182Hf was extant. W isotopic composition has been measured by MC-ICPMS at UC Davis in Isua samples for which excess in 142Nd were reported. No 182W deviation has been measured outside the error of 30 ppm (external precision). We will use these data to bring more constraints on the timing of the silicate Earth differentiation. Using a 3-stage model, if the core/mantle differentiation occurred before 25 Ma, the decoupling between 142Nd and 182W anomalies provides evidence of an early mantle differentiation (40 Ma) produced probably before the Moon's formation. Finally, we will discuss the importance of early events for the Earth's mantle evolution, trying to reconcile the modern view we have of the mantle with its long-term evolution.
V43F-02
Was Earth Initially Chondritic for its Coupled 142Nd-143Nd Signature? A Perspective From Mars.
The recent reinvestigation of the 142Nd/144Nd value of chondrites shows that there is a clearly resolvable difference between the chondritic value with ε142Nd = -0.18 and the average value for the terrestrial standard with ε142Nd = 0. This observation requires the existence of a hypothetical reservoir in Earth, with an even more negative ε142Nd than chondrites to balance the near-zero value of present-day terrestrial samples if the bulk Earth has a chondritic composition. However, the existence of this reservoir has not been confirmed, because up to now, only samples with positive ε142Nd anomalies have been measured. This implies that this reservoir, if existing, is hidden somewhere in the Earth's mantle but has not been sampled. Another way to explain the discrepancy between Earth and chondrites would be if Earth accreted from material with a Sm/Nd higher than the average for chondrites. In this case, Earth would have been characterized by an initially depleted composition similar to that observed for the depleted MORB mantle and corresponding to present-day values of ε142Nd = 0 and ε143Nd = +10.9, instead of chondritic at ε142Nd = -0.18 and ε143Nd =0. The presence of the missing Nd reservoir is difficult to test because most geochemical evidence for early differentiation has been largely erased in Earth via plate tectonics. On the other hand, the absence of plate tectonics on Mars has limited the mixing of chemical heterogeneities within the interior, permitting the observation of early differentiation processes as recorded in Martian meteorites. Given that Mars and Earth together represent ~60% of disk mass in the inner terrestrial planet region, and assuming they accreted from similar material, Mars is likely to be a valid analog to investigate the refractory element isotopic signatures of other terrestrial planets at their formation. Among SNC meteorites, ALHA84001 is unique by its very old crystallization age (~3.9-4.5 Gy) compared to the other SNC's, and is likely sampling the isotopic composition of a very early mantle reservoir in Mars. New high precision Nd isotope data of this sample give ε142Nd = - 0.23±0.05, and ε143Nd (i) = +6.1. These coupled 142Nd-143Nd systematics of ALHA84001 have been tested for the two hypotheses (Mars initially depleted or chondritic) and are inconsistent with an initially depleted Mars. By analogy, it implies that Earth was also initially chondritic for its 142Nd- 143Nd signature and that a hidden reservoir with a negative ε142Nd is required.
V43F-03
The Lu-Hf isotopic composition of CHUR and BSE: Tighter constraints from unequilibrated chondrites
The Lu-Hf isotopic system has been used increasingly in geochemistry as a chronometer and tracer of mantle and crust processes since the development of MC-ICPMS techniques [1]. Although a consensus has emerged on the value of the 176Lu decay constant, in contrast, the Lu and Hf isotopic compositions of the Chondrite Uniform Reservoir (CHUR) and Bulk Silicate Earth (BSE) have not been as well constrained. Lu-Hf isotopic compositions vary dramatically between the chondrite classes and petrologic types of the specimens that have been analyzed [1-3], which hampers a choice of Lu-Hf CHUR parameters. Chondrites are classified in three main petrologic groups: carbonaceous (CC), ordinary (OC) and enstatite chondrites (EC). They represent distinct chemical and isotopic compositions, which can be associated with reservoirs in the protoplanetary disk where the respective parent bodies have formed. They also have been subjected to various degrees of aqueous alteration (types 1 and 2) or thermal metamorphism (types 3-6) that has potentially affected their initial chemical and mineralogical characteristics. Despite the fact that numerous OCs (~50) have been analyzed for their Lu-Hf isotopic composition, nearly all of these have been equilibrated types 4-6; only 5 finds of types 3.6-3.8 unequilibrated OC have been analyzed. We have analyzed 20 new chondrites for Lu-Hf and Sm-Nd isotope systematics including (i) 13 H, L, and LL OC of types 3.0- 3.8, where their low degree of metamorphism limits the growth of phosphate (main carrier of REE) compared to the equilibrated OC; and (ii) 7 CC of types 1-3 (CI, CV, CO and CK groups). We obtained mean values (2σSE) of 176Lu/177Hf = 0.0337 ± 3, 176Hf/177Hf = 0.282802 ± 23, 147Sm/144Nd = 0.1961 ± 6, and 143Nd/144Nd = 0.512629 ± 16 from the types 1-3 OC and CC from this study and [2]. Our Lu-Hf values are higher than the previous estimates of [3] and are our best estimates for CHUR. The CC alone give higher mean values of 176Lu/177Hf = 0.0339 ± 4, 176Hf/177Hf = 0.282816 ± 32, 147Sm/144Nd = 0.1966 ± 10, and 143Nd/144Nd = 0.512639 ± 28. These last values are concordant with the Sm-Nd CHUR and BSE parameters that are currently widely used [4]. Lu-Hf and Sm-Nd isotopic systems involve refractory and lithophile elements, so that the composition of BSE should coincide with that of the CHUR. The Lu/Hf chondrite range in unequilibrated OC and CC is now constrained by ~7%, equivalent to what is found for the paired Sm-Nd system. To refine the Lu-Hf BSE estimate, we need to determine which chondrites are the best representative of BSE. For Sm-Nd isotope systematics, there is wide overlap between the chondrite groups. In constrast, CC have significantly higher Lu/Hf than OC. If we compare with other refractory and lithophile elements, CV, CK, CM and CO chondrites represent the closest composition with Earth's mantle [5]. From O and Cr isotope constraints [6], EC share a common reservoir of formation with the Earth. We will refine the BSE composition based on these observations and also present supplementary data on CI, CR and unequilibrated EC. [1] Blichert-Toft and Albarède, 1997. EPSL, 148, 243-258. [2] Patchett et al., 2004. EPSL, 222, 29-41. [3] Bizzarro et al., 2003. Nature, 421, 931-933. [4] Jacobsen and Wasserburg, 1980. EPSL, 50, 139-155. [5] Palme, 2001. Philo. Trans. R. Soc. Lond., 359, 2061-2075. [6] Trinquier et al., 2007. APJ, 655, 1179-1185.
V43F-04
Crystallization of a basal magma ocean and its geochemical consequences
The seismological evidence of partial melt at the base of the mantle suggests that more extensive melting occured in the past when the mantle was hotter. The Fe enrichment of the melt, its high density at very high pressure together with its low viscosity preserved the basal magma ocean from entrainment by solid-state convection. It is therefore an ideal hidden and unsampled reservoir for incompatible elements, even if its size is relatively small today. The \varepsilon142Nd difference between Earth and chondrites, as well as the mass balance of heat producing elements are used to constrain our thermo-chemical models of the basal magma ocean evolution through time. The best fit is obtained for an initial melt layer about 800km thick crystallizng with a decay time of 1~Gy, the remaining melt today being ULVZs. The solids crystallizing at the top of the cooling melt layer have increasing Fe content with time and when they become dense enough, they start to form stable piles which are actually observed from seismic studies. They are slowly entrained within mantle plumes providing a deep mantle component which is very diluted in incompatible elements. Though, we show that this component can provide the primitive-like noble gas signature that constrasts with that of the intensively degassed ovelying mantle.
V43F-05
Calcium Isotope Heterogeneity Among Planetary Objects
A critical issue for understanding the origin of the Earth's geochemical reservoirs is whether chondritic meteorites are an adequate model for the bulk composition of the Earth, even for refractory lithophile elements. One way to address this issue is to document the extent of isotopic homogeneity in the early solar system by studying in detail the isotopic composition of different meteorite classes, and comparing them to Earth, Mars, and the Moon. We present new Ca isotope measurements of bulk chondritic meteorites that show non-mass dependent differences among ordinary chondrites and Earth for the more abundant and lighter Ca isotopes–masses 40, 42, 43 and 44. The measurements are made using TIMS, and isotope ratios are normalized to a constant 42Ca/44Ca. The fractionation-corrected 40Ca/44Ca ratios in meteorites vary from -3 to +2 epsilon units (0.01%) relative to the terrestrial ratio. The typical analytical uncertainty is ±0.015%. The observed variations could be caused by differences in either 40Ca or 42Ca abundance. The 43Ca/44Ca ratio shows minimal variation. The observed effects are near the limits of analytical resolution, due to incompletely understood inconsistencies in Ca isotope analysis using multi-collector TIMS, but must be nucleosynthetic in origin. Hence the data suggest that the solar system was heterogeneous with respect to Ca isotopes, and some fraction of this heterogeneity was preserved through the accretion process. The Ca data add to previously documented variation in isotopes of Cr, Mo, Ba, Sm, and Nd in chondrites. Carbonaceous chondrites Allende (CV3) and Murchison (CM2) yield the highest 40Ca enrichments (+2.0 ±1.2 and +1.1 ±0.8 in units of 0.01%), whereas ordinary chondrites Saint Severin (LL6) and Bruderheim (L6) exhibit the lowest (-3.3 ±0.7 and -3.6 ±1.8). The 40Ca isotopic effects in chondrites show correlations with reported effects in \triangle 17O, 54Cr, and 53Cr. There is a weak correlation between high 40Ca and low 142Nd, but meteorites with both high and low 40Ca have 142Nd lower than terrestrial. The observed nuclear effects in Ca isotopes can be accounted for by differences in the relative contributions of Ca made during supernova explosions (explosive O-Si burning) versus Ca made in dominantly s-process nucleosynthesis such as occurs in AGB stars. There is still no single model that can account for all observed isotopic data, but evidence is mounting that incomplete mixing of a heterogeneous solar nebula must be accounted for in the assessment of the formation age and evolution of geochemical reservoirs in the Earth's interior.
V43F-06 INVITED
Integrated Nd, Hf and Pb View of Mantle Evolution
Recent developments in the 147,146Sm-143,142Nd and 176Lu-177Hf systems during the past several years have drastically altered our view of the evolutionary history of the early Earth. These new developments, in particular from 146Sm-142Nd systematics and the changes in the initial 176Hf/177Hf data for early Archean samples using the lower 176Lu decay constant, have produced a data set that, on face value, appears to be in conflict and require more complicated models for early planetary differentiation and subsequent evolution. For example, the record of Hf mantle evolution using previous decay constant of 1.94 e-11 y-1 was entirely congruent with the Nd isotopic record and broadly characterized by linear trend of radiogenic values today (εHf, εNd of +16 and +10, respectively) through slightly radiogenic values for the early Archean (εHf, εNd of ~ +4 and +2) and projecting to chondritic values within the first few 100 m.y. of Earth's history. These data seem to be explained by a rather straight-forward model with differentiation of the mantle and the development of enriched and depleted reservoirs. Using the most recent 176Lu decay constant estimates, however, the Hf isotope record of the oldest terrestrial rocks and zircons is characterized by chondritic to negative initial εHf values. In contrast, the Nd isotopic record of similar rocks are characterized by positive εNd values. The 143Nd record is now consistent with constraints from the short-lived isotope 146Sm-142Nd system which indicate very early and widespread differentiation of the mantle and possibly sequestration of the enriched component to the hidden mantle. The Pb isotopic record of galenas from the Late Archean Abitibi greenstone belt of the Superior Province also records a pervasive mantle differentiation event in the first hundred m.y. of Earth's history and therefore are broadly consistent with the 142Nd and 143Nd data. The apparent decoupling of the Hf systems with the Nd and Pb systems (if the 176Lu decay constant values are correct) must reflect processes operating during early differentiation of the planet. This differentiation is responsible for widespread fractionation of Sm/Nd and production of enriched, low Sm/Nd and depleted, high Sm/Nd reservoirs, but does not fractionate Lu/Hf effectively between these reservoirs. The 146Sm- 142Nd systematics require sequestration of the early enriched reservoir to an unseen part of the mantle. This is consistent with the positive εNd values that are characteristic of most Archean terranes. Post-4.0 Ga, the Nd and Hf isotopic systems behave congruently and record the progressive evolution of enriched and depleted reservoirs due to Lu/Hf and Sm/Nd fractionations such as we see during mantle melting processes operating on the Earth today.
V43F-07 INVITED
Rogue Mantle Helium and Neon
The canonical view of He isotope geochemistry holds that high 3He/4He ratios in basalts fingerprints undegassed mantle sources. Hawaiian basalts with unradiogenic He with 3He/4He up to 30 RA are therefore seen as originating from parts of the mantle that is still primordial, at least much more so than MORB mantle (3He/4He ~ 8 RA). This view was strongly reinforced by the discovery of solar and even planetary Ne components in oceanic basalts and gas wells. The canonical view, however, conflicts with multiple observations on ocean islands, notably Hawaiian basalts: the correlation of {187}Os/{186}Os with δ 18O combined with the presence of unusually radiogenic Hf isotope compositions for a given Nd isotope composition and the correlation between Hf and Pb isotopes are all features strongly reminiscent of ancient subducted oceanic crust and pelagic sediments in the source of the Hawaiian plume. These conflicting observations beg the question of how Hawaiian basalts, which carry the embodiment of a primordial gas signature, at the same time can provide such strong evidence of surface material recycling. I here suggest and alternative model that uses the marble cake paradigm and Shuster et al.'s data on olivine. A solution to this conundrum lies in an analogy with oil genesis: 3He and Ne do not reside in the low-melting point peridotites in which they were originally hosted but rather migrated since early in Earth history into refractory ‘reservoir' rocks. Since there can be no free gas phase percolating at pressures in excess of olivine carbonation at ~3 GPa, He must be largely redistributed by diffusion. The time scale of diffusion is the defining parameter: although over billions of years 3He diffuses across large distances, melting events are too short to efficiently strip residual refractory rocks from their high-3He/4He component. Assuming that melts begin forming over the uppermost 100 km with an upwelling rate of 10 m y-1 in plume conduits and 10 cm y-1 under mid- ocean ridges, the characteristic times of melt extraction in each of these two environments are 10,000 y and 1 My, respectively, and the maximum thickness of refractory layers contributing their He to the magmas are 10 m and 100 m, respectively. The difference in 3He/4He ratios of ocean-island and mid-ocean ridge basalts and the preservation of solar neon are ascribed to the reservoirs rocks being stretched to a different extent during melting. Old fragments of oceanic lithosphere, and possibly cumulates from the magma ocean, rather than primordial mantle ‘nuggets', should host most of the primordial He and Ne presently observed in oceanic basalts. Helium with high 3He/4He ratios may contain a component of primordial origin, but not necessarily reflect the reservoir in which it has been residing for most of the Earth's history.
V43F-08
The origin of high 3He/4He values in oceanic lavas: arguments against a heterogeneous upper mantle source
One of the contentious issues fuelling the debate on the origin of intraplate magmatism is the source of high 3He/4He ratios in oceanic lavas. The contemporary hotspot hypothesis posits that the high 3He/4He signature is coming from a distinct source reservoir that lies relatively deep in the mantle whereas others argue that there is no such a reservoir. For example, the SUMA (Statistical Upper Mantle Assemblage - Meibom & Anderson, EPSL 17, 2003) hypothesis argues that the compositional variability of oceanic lavas results from sampling upon melting of the heterogeneous upper mantle. It claims that extreme isotopic signals, such as high 3He/4He ratios, of some oceanic lavas are due to small degrees of partial melting that preferentially sample dispersed, anomalous mantle components. In contrast, large degrees of partial melting sample a larger volume, and hence many components of the heterogeneous mantle; these also promote more effective mixing of the resulting melts, producing a relatively homogeneous composition typified by that of N-type MORB. To constrain the source of high 3He/4He lavas, we analyzed a ~dozen representative samples from the axis and near-ridge seamounts in the 11o45'N to 15o00'N segment of the EPR. Axial lavas are N- to T-type MORB that came from a heterogeneous mantle source (Castillo et al., G3 1, 2000). Seamount lavas extend the compositional variability of axial lavas to both higher and lower Sr, Nd and Pb isotope values, suggesting that seamount lavas indeed result from smaller degrees of partial melting of the heterogeneous mantle beneath the region. Our preliminary results, however, also show that the majority of both axial and near-ridge seamount lavas have indistinguishable, MORB-like 3He/4He values, at 8 +/- 1 RA, over a range of He concentrations. The remainder have 3He/4He < 8 RA. Thus we conclude that the heterogeneous Pacific upper mantle cannot provide the high 3He/4He signature in oceanic lavas.