Janne Blichert-Toft

THE LU-HF ISOTOPE GEOCHEMISTRY OF CHONDRITES AND THE EVOLUTION OF THE MANTLE-CRUST SYSTEM

Abstract We report analyses of the176Hf/177Hf ratio for 25 chondrites from different classes of meteorites (C, O, and E) and the176Lu/177Hf ratio for 23 of these as measured by plasma source mass spectrometry. We have obtained a new set of present-day mean values in chondrites of176Hf/177Hf= 0.282772 ± 29 and176Lu/177Hf= 0.0332 ± 2. The176Hf/177Hf ratio of the Solar System material 4.56 Ga ago was 0.279742 ± 29. Because the mantle array lies above the Bulk Silicate Earth in a143Nd/144Nd versus176Hf/177Hf plot, no terrestrial basalt seems to have formed from a primitive undifferentiated mantle, thereby casting doubt on the significance of high3He/4He ratios. Comparison of observedHf/Nd ratios with those inferred from isotopic plots indicates that, in addition to the two most prominent components at the surface of the Earth, the depleted mantle and the continental crust, at least one more reservoir, which is not a significant component in the source of oceanic basalts, is needed to account for the Bulk Silicate Earth Hf-Nd geochemistry. This unaccounted for component probably consists of subducted basalts, representing ancient oceanic crust and plateaus. The lower continental crust and subducted pelagic sediments are found to be unsuitable candidates. Although it would explain the Lu-Hf systematics of oceanic basalts, perovskite fractionation from an early magma ocean does not account for the associated Nd isotopic signature. Most basalts forming the mantle array tap a mantle source which corresponds to residues left by ancient melting events with garnet at the liquidus.

Evolution of the depleted mantle: Hf isotope evidence from juvenile rocks through time

Abstract The covariant behavior of Lu-Hf and Sm-Nd isotopes during most magmatic processes has long been recognized, but the details of this behavior in the depleted mantle reservoir have not been adequately examined. We report new whole-rock Hf and Nd isotope data for 1) juvenile, mantle-derived rocks, mid-Archean to Mesozoic in age, and 2) early Archean gneisses from West Greenland. Hf and Nd isotopic compositions of the juvenile rocks are well correlated, with the best fit corresponding to the equation e Hf = 1.40 e Nd + 2.1, and is similar to the collective Hf-Nd correlation for terrestrial samples of e Hf = 1.36 e Nd + 3.0. The early Archean Greenland gneisses, in contrast, have an extreme range in e Nd values ( − 4.4 to + 4.2; Bennett et al., 1993 ) that is not mirrored by the Hf isotopic system. The e Hf values for these rocks are consistently positive and have much less variation (0 to + 3.4) than their e Nd counterparts. The information from the Hf isotopic compositions of the West Greenland gneisses portrays an early Archean mantle that is relatively isotopically homogeneous at 3.8 to 3.6 Ga and moderately depleted in incompatible elements. There is no evidence that any of these gneisses have been derived from an enriched reservoir. The Hf isotopic data are in stark contrast to the Nd isotopic record and strongly imply that the picture of extreme initial isotopic heterogeneity indicated by Nd isotopes is not a real feature of the West Greenland gneisses but is rather an artifact produced by disturbances in the Sm-Nd isotope system of these rocks. Although Hf and Nd isotopic data do not uniquely constrain either the nature of the earliest crust or the timing of crustal growth, the most probable candidate for the enriched reservoir complementary to the depleted mantle in the pre-4.0 Ga Earth is a mafic, oceanic-type crust. In order to explain the predominantly positive e Hf and e Nd values for the early Archean rocks, this crust must have had a short residence time at the surface of the Earth before returning to the mantle where it was isolated from mixing with the depleted mantle for several hundred million years. The following period from 3.5 to 2.7 Ga may mark a transition during which this early formed mafic crust was mixed progressively back into the depleted mantle reservoir. While a present-day volume of continental crust at 4.0 Ga cannot be excluded on isotopic grounds, we find such a scenario unlikely based on the lack of direct isotopic and physical evidence for its existence. An important aspect of crustal growth and evolution, therefore, may be the transformation of the enriched reservoir from being predominantly mafic in the early Earth to becoming progressively more sialic through time.

A short timescale for terrestrial planet formation from Hf–W chronometry of meteorites

Determining the chronology for the assembly of planetary bodies in the early Solar System is essential for a complete understanding of star- and planet-formation processes. Various radionuclide chronometers (applied to meteorites) have been used to determine that basaltic lava flows on the surface of the asteroid Vesta formed within 3 million years (3 Myr) of the origin of the Solar System. Such rapid formation is broadly consistent with astronomical observations of young stellar objects, which suggest that formation of planetary systems occurs within a few million years after star formation. Some hafnium–tungsten isotope data, however, require that Vesta formed later (∼16 Myr after the formation of the Solar System) and that the formation of the terrestrial planets took a much longer time (62-14+4504 Myr). Here we report measurements of tungsten isotope compositions and hafnium–tungsten ratios of several meteorites. Our measurements indicate that, contrary to previous results, the bulk of metal–silicate separation in the Solar System was completed within <30 Myr. These results are completely consistent with other evidence for rapid planetary formation, and are also in agreement with dynamic accretion models that predict a relatively short time (∼10 Myr) for the main growth stage of terrestrial planet formation.

A hafnium isotope and trace element perspective on melting of the depleted mantle

Abstract New Hf isotope and trace element data on mid-ocean ridge basalts (MORB) from the Pacific Ocean basin are remarkably uniform (176Hf/177Hf≈0.28313–0.28326) and comparable to previously published data [Salters, Earth Planet. Sci. Lett. 141 (1996) 109–123; Patchett, Lithos 16 (1983) 47–51]. Atlantic MORB have 176Hf/177Hf ranging from 0.28302 to 0.28335 confirming the wide range originally identified by Patchett and Tatsumoto [Geophys. Res. Lett. 7 (1980) 1077–1080]. Indian MORB define an even wider range, from 0.28277 to 0.28337, but three exotic samples have very unradiogenic Hf isotope compositions. Their very low 176Hf/177Hf ratios, together with their trace element characteristics, require the presence of unusual plume-type material beneath the Indian ridge. All other Indian MORB have uniform Hf isotope compositions at about 0.2832, and define a small field displaced to the right of other MORB in Hf–Nd isotope space. The distinct nature of Indian MORB is best explained by the presence in Indian depleted mantle of old recycled oceanic crust and pelagic sediments. Sm/Hf ratios calculated from new high-precision rare earth element and Hf trace element data do not vary in MORB in the same way as in ocean island basalts (OIB): ratios are constant in OIB, but decrease with increasing Sm contents in MORB. The constancy of Sm/Hf in OIB is probably due to an overwhelming influence of residual garnet during melting. By contrast, the decrease of Sm/Hf in MORB is due to the effect of clinopyroxene in the residue of melting beneath ridges, an interpretation confirmed by quantitative modeling of melting. The relationship between Sm/Nd and Lu/Hf ratios in MORB does not require the presence of garnet in the residual mineralogy. The decoupling of Lu/Hf ratios and Hf isotope compositions – the so-called Hf paradox [Salters and Hart, EOS Trans. Am. Geophys. Union 70 (1989) 510] – can be explained by melting dominantly in the spinel field at shallow depths beneath mid-ocean ridges.

Heterogeneous Hadean Hafnium: Evidence of Continental Crust at 4.4 to 4.5 Ga

The long-favored paradigm for the development of continental crust is one of progressive growth beginning at ∼4 billion years ago (Ga). To test this hypothesis, we measured initial 176Hf/177Hf values of 4.01- to 4.37-Ga detrital zircons from Jack Hills, Western Australia. ϵHf (deviations of 176Hf/177Hf from bulk Earth in parts per 104) values show large positive and negative deviations from those of the bulk Earth. Negative values indicate the development of a Lu/Hf reservoir that is consistent with the formation of continental crust (Lu/Hf ≈ 0.01), perhaps as early as 4.5 Ga. Positive ϵHf deviations require early and likely widespread depletion of the upper mantle. These results support the view that continental crust had formed by 4.4 to 4.5 Ga and was rapidly recycled into the mantle.

Dating the Indian continental subduction and collisional thickening in the northwest Himalaya: Multichronology of the Tso Morari eclogites

Multichronometric studies of the low-temperature eclogitic Tso Morari unit (Ladakh, India) place timing constraints on the early evolution of the northwest Himalayan belt. Several isotopic systems have been used to date the eclogitization and the exhumation of the Tso Morari unit: Lu-Hf, Sm-Nd, Rb-Sr, and Ar-Ar. A ca. 55 Ma age for the eclogitization has been obtained by Lu-Hf on garnet, omphacite, and whole rock from mafic eclogite and by Sm-Nd on garnet, glaucophane, and whole rock from high-pressure metapelites. These results agree with a previously reported U-Pb age on allanite, and together these ages constrain the subduction of the Indian continental margin at the Paleocene-Eocene boundary. During exhumation, the Tso Morari rocks underwent thermal relaxation at about 9 ± 3 kbar, characterized by partial recrystallization under amphibolite facies conditions ca. 47 Ma, as dated by Sm-Nd on garnet, calcic amphibole, and whole rock from metabasalt, Rb-Sr on phengite, apatite, and whole rock, and Ar-Ar on medium-Si phengite from metapelites. Ar-Ar analyses of biotite and low-Si muscovite from metapelites, which recrystallized at <5 kbar toward the end of the exhumation, show that the Tso Morari unit was at upper crustal levels ca. 30 Ma. These results indicate variable exhumation rates for the Tso Morari unit, beginning with rapid exhumation while the Indian margin subduction was still active, and later proceeding at a slower pace during the crustal thickening associated with the Himalayan collision.

The Lu–Hf dating of garnets and the ages of the Alpine high-pressure metamorphism

It remains controversial whether burial and exhumation in mountain belts represent episodic or continuous processes. Regional patterns of crystallization and closure ages of high-pressure rocks may help to discriminate one mode from the other but, unfortunately, metamorphic geochronology suffers from several limitations. Consequently, no consensus exists on the timing of high-pressure metamorphic events, even for the Alps—which have been the subject of two centuries of field work. Here we report lutetium–hafnium (Lu–Hf) mineral ages on eclogites from the Alps as obtained by plasma-source mass spectrometry. We find that the Lu/Hf ratio of garnet is particularly high, which helps to provide precise ages. Eclogites from three adjacent units of the western Alps give (from bottom to top) diachronous Lu–Hf garnet ages of 32.8 ± 1.2, 49.1 ± 1.2 and 69.2 ± 2.7Myr. These results indicate that the Alpine high-pressure metamorphism did not occur as a single episode some 80–120Myr ago,,,, but rather that burial and exhumation represent continuous and relatively recent processes.

Lu–hf garnet geochronology: closure temperature relative to the Sm–Nd system and the effects of trace mineral inclusions

The highly elevated Lu/Hf of garnets with respect to other minerals, coupled with the new capability of routinely analyzing small samples (25 ng of Hf) by multiple-collector ICP-MS (MC–ICP–MS), makes the Lu–Hf garnet system a viable geochronometer. The robustness of Lu–Hf garnet-whole rock (gt-wr) ages, however, needs to be evaluated, and their closure temperature (TC) and potential effects of trace mineral inclusions need to be established. To constrain the TC of Lu–Hf relative to that of Sm–Nd in gt-wr systems, we used thermal ionization mass spectrometry (TIMS) and MC–ICP–MS techniques to determine the Lu–Hf and Sm–Nd ages of garnet-bearing rocks for which the general thermochronology had been previously established. Samples include the Huiznopala Gneiss (Hidalgo, Mexico), the Gore Mountain amphibolite (New York, USA), a xenolith from the Bearpaw Mountains (Montana, USA), and the Smith Grade Granite (California, USA). In addition to whole rocks and garnet, the Lu–Hf isotope compositions of hornblende, zircon, and monazite were also measured. Our data suggest that the TC of Lu–Hf is greater than or equal to the TC of Sm–Nd in gt-wr systems that cooled slowly (<10°C/m.y.) from granulite facies conditions. There is no single TC for Lu–Hf or even a restricted range of TC that applies to all garnets, as is the case for Sm–Nd. Leaching experiments and trace element modeling show that monazite and apatite inclusions may severely affect the Sm–Nd systematics of garnet, but they have little or no effect on the Lu–Hf system. In contrast, zircon, with its high Hf content, can strongly influence the Lu–Hf systematics of garnets and whole rocks. Zircon from two samples did not achieve Hf isotope equilibrium with the rest of the rock at the time indicated by gt-wr isochrons. Zircon is thus capable of preserving an inherited Hf component through periods of high-grade metamorphism. If present in the matrix only, such zircon will cause erroneously old Lu–Hf ages, while such zircon present only in the garnet will yield ages that are too young. If inherited zircon is distributed evenly throughout the garnet and matrix, the competing age effects will partially cancel, but the age will be too young if the rock contains a phase (e.g., hornblende) that buffers the matrix against the influence of inherited zircon Hf. For rocks that contain inherited zircon, the maximum effects on the gt-wr age must be determined before the age can be interpreted with confidence. Garnets that have significant zircon inclusion contents (i.e., 176Lu/177Hf ≲ 0.3 in this study) should be avoided for Lu–Hf gt-wr geochronology.

The role of sediment recycling in EM-1 inferred from Os, Pb, Hf, Nd, Sr isotope and trace element systematics of the Pitcairn hotspot

We present comprehensive radiogenic isotope (Os, Pb, Hf, Nd, Sr) and trace element data on basaltic lavas from Pitcairn Island and the Pitcairn seamounts and examine the origin of the enriched mantle isotopic signature (EM-1) found in these lavas. The 187Os/188Os ratios of the lavas range from 0.131 to 0.254, while those of the high-Os concentration samples (>50 pg/g) lie between 0.131 and 0.148. All 187Os/188Os ratios are higher than the bulk silicate Earth reference value of 0.127. Since ancient subcontinental lithospheric mantle (SCLM) is expected to have a 187Os/188Os ratio less than 0.127, it appears that recycled SCLM plays no role in the Pitcairn source. Variations in 187Os/188Os ratios appear to be unconnected with those of 206Pb/204Pb ratios in Pitcairn lavas, suggesting that Pb and Os isotopic variations are controlled by different factors. Modeling shows that variations in Pb isotopic compositions may mainly reflect the proportion of recycled sediment in the source, while those of 187Os/188Os ratios may reflect the proportion of peridotite mantle versus recycled oceanic crust. The occurrence of negative Nb anomalies in some of the lavas, a correlation between Nb anomaly and 87Sr/86Sr ratios (0.7036–0.7051), and extremely unradiogenic and strongly correlated Nd and Hf isotopic compositions (ϵNd of −5.9 to +1.1 and ϵHf of −5.3 to +2.2) together suggest that the Pitcairn mantle source contains a recycled continental crustal component. The slope of the ϵHf vs. ϵNd correlation is shallower for Pitcairn Island than for the Pitcairn seamounts or the global OIB array, and may be due to a variable ratio of recycled mud to sand in the Pitcairn source. A trace element mixing model also indicates the presence of small amounts of recycled pelagic and terrigenous sediment and permits variable amounts of depleted components such as recycled MORB, gabbro and depleted mantle. The 206Pb/204Pb ratios of the Pitcairn lavas vary between 17.47 and 18.10 and are very unradiogenic compared to those of other ocean island basalts. By contrast, 208Pb/204Pb ratios are high and relatively homogeneous at values of ∼39.0. This observation along with the measured Th/U ratios of the lavas, which range up to 14.1, indicate a long-term history of U loss in the Pitcairn source. In 207Pb/204Pb–206Pb/204Pb space, the data form a linear array that can be interpreted in terms of mixing between a minor recycled sediment end member and more depleted material. Lead isotopic compositions suitable for the recycled end member were investigated using a three-stage evolution model by Monte Carlo methods and suggest ages between 0.7 and 1.9 Ga for the recycled sediment. The relationships between measured Th/U and radiogenic 208Pb*/206Pb* ratios suggest that the isotopic arrays displayed by the lavas were produced by mixing, probably occurring during magma genesis.

Upwelling of deep mantle material through a plate window: Evidence from the geochemistry of Italian basaltic volcanics

87 Sr/ 86 Sr– 206 Pb/ 204 Pb–eNd–eHf space. The isotopic compositions of the two end-members of these mixing arrays are assessed by least-squares regression. The mantle-derived component ( 206 Pb/ 204 Pb = 19.8, 87 Sr/ 86 Sr = 0.7025, eNd = +8, eHf = +9) is a rather homogeneous mixture of the standard high-m (HIMU) and depleted mantle (DM) components. The crust-derived component ( 206 Pb/ 204 Pb = 18.5, 87 Sr/ 86 Sr > 0.715, eNd = � 12, eHf = � 11) accounts for the enrichment of K and other large-ion-lithophile elements in the Italian volcanics. As shown by the relationship in eHf–eNd space and the lower-thanchondritic Hf/Sm ratio, this crustal component is dominated by pelagic sediments rather than terrigenous material. The overall scarcity of calc-alkaline compositions in the Italian volcanics and the presence of a HIMU component, which is the hallmark of hot spot basalts, raise the question of how plume mantle source contributes to volcanism in a subduction environment. At about 13 Ma, the Apennine collision terminated the westward subduction of the Adria plate under the European margin and rotated the direction of convergence to the northwest. The cumulative differential of subduction between the fossil plate under Tuscany and the active plate under Sicily since the opening of the Tyrrhenian Sea amounts to at least 300 km and is large enough to rift the dipping plate and open a plate window beneath the southern part of the peninsula. This model is consistent with recent high-resolution seismic tomography. We propose that the counterflow of mixed upper and lower mantle passing the trailing edge of the rifted plate is the source of Italian mafic volcanism. Alternatively, material from a so-far unidentified plume may be channeled through the plate window. The crustal signature is probably acquired by interaction of the mantle advected through the window with the upper part of the subducted plate. INDEX TERMS: 1749 History of Geophysics: Volcanology, geochemistry, and petrology; 1025 Geochemistry: Composition of the mantle; 1040 Geochemistry: Isotopic composition/chemistry; KEYWORDS: Italian volcanism, HIMU, subduction, pelagic sediments, mixing, slab window