9.2 Mantle Hf evolution

9.2.1 Hf zircon analysis

Patchett et al. (1981) investigated the usefulness of initial Hf isotope ratios as a tracer of mantle Hf through time, using crustal rocks with a clear mantle derivation. In order to calculate accurate initial ¹⁷⁶Hf/¹⁷⁷Hf ratios for many rock bodies without making numerous isochron determinations, Patchett et al. principally analysed zircon separates. These provided excellent material for Hf isotope analysis for several reasons:

1. Hf forms an integral part of the zircon lattice, which is therefore very resistant to Hf mobility and contamination.

2. The very high Hf concentrations in zircon (ca. 10 000 ppm) yield very low Lu/Hf ratios and consequently minute age corrections.

3. There are large quantities of zircon separates previously prepared for U–Pb analysis, which yield accurate dates for the same material.

4. Any metamorphic overprinting or zircon inherited from a previous crustal history are clearly revealed by the U–Pb data.

Initial ¹⁷⁶Hf/¹⁷⁷Hf ratios of presumed mantle-derived igneous rocks of various ages were plotted on a hafnium isotope evolution diagram by Patchett et al. (1981). All of the igneous rocks with a mantle-derived signature lay within error of, or above, the chondrite evolution line (Fig. 9.4). However, only one sample, a meta-tholeiite dyke cutting the Suomussalmi)Kuhmo greenstone belt in Finland, had an initial ¹⁷⁶Hf/¹⁷⁷Hf above a linear depleted mantle evolution line drawn from the primordial solar system value to the most radiogenic MORB analysis (Fig. 9.4).

Fig. 9.4. Diagram of Hf isotope evolution over geological time. Initial ratios of uncontaminated mantle-derived magmas show them to be derived from a slightly depleted source relative to chondrites. Data from Patchett et al. (1981).

One explanation for the fan of Hf data in Fig. 9.4 is their derivation from a heterogeneous mantle exhibiting variable trace element depletion of Hf relative to Lu through space and time. An alternative would be to derive magmas from a more depleted homogeneous source (such as defined by the dashed evolution line in Fig. 9.4) and subject them to contamination by older crustal basement. Patchett et al. (1981) preferred this model for the 1.4 Byr-old Silver Plume and 1.0 Byr-old Pikes Peak batholiths of Colorado, which have Nd and Hf initial ratios near the chondritic evolution line, and were argued to contain large fractions of 1.7 Byr-old crust by DePaolo (section 4.2.2). Two 1.8 Byr-old post-tectonic granites intruded into the Archean craton of north Finland had spectacularly low initial ratios, corresponding to , Hf values of !10 and !12. These two samples are clearly of crustal derivation on the basis of Pb and Sr isotope data, and were selected to demonstrate the effects of crustal reworking on Hf isotope systematics.

It is important to remember that when this paper was published (in 1981) the only conclusive Nd isotope evidence for depleted mantle in the Proterozoic was provided by the work of DePaolo on the Front Ranges of Colorado (see section 4.2.2). Nd isotope evidence from komatiites for Archean depleted mantle (section 4.4.3) was not yet available. Hence the major significance of this paper at the time.

9.2.2 Archean sediments

Because Hf involves greater analytical difficulties than Nd, the latter has been the primary tool for studying continental growth through geological time (section 4.4). However, the resistance of zircon to weathering and to metamorphic re-setting means that Hf isotope analysis of this mineral provides a good test for Nd isotope data in studies of the Earth’s early evolution, involving both ancient sediments and mantle-derived igneous rocks. These data have then been used to constrain models of crustal growth or recycling in the early Archean (section 4.4.2).

Stevenson and Patchett (1990) made the first Lu)Hf study on Archean and Proterozoic sedimentary zircons from the Canadian, Wyoming, North Atlantic and South African cratons. The samples were mainly from the Late Archean and Early Proterozoic, and are summarised on a diagram of Lu)Hf model age versus stratigraphic age (Fig. 9.5), using the old CHUR value. A significant number of samples gave model ages in excess of stratigraphic age (i.e. above the ‘concordia’ line). The excess of model age over stratigraphic age was less in older sediments, leading Stevenson and Patchett to argue that there were ‘no great quantities of continental crust in the early Earth providing ancient zircons...’. However, a significant number of points lie below the concordia line in Fig. 9.5, indicating that the chondritic mantle model is an under-estimate of the provenance ages of these samples.

The use of a depleted mantle model implies provenance ages ca. 200 Myr older than the CHUR model, suggesting a significant volume of Early Archean crust. Nevertheless, there is no evidence for a ‘big bang’ model in which the volume of Early Archean crust was as great as that at the present day. Consequently, the data provide some support for a model involving progressive crustal growth with time, but are nevertheless ambiguous.

Fig. 9.5. Model age versus stratigraphic age diagram for sedimentary zircons to constrain crustal growth models for the Archean. ( ! ) = Hf data; ( Q ) = Nd data. Dashed line accents the upper envelope of the data. After Stevenson and Patchett (1990).

Amelin et al. (1999) extended the study of Hf in detrital zircons to the early Archean. This was based on zircons from the Jack Hills meta-conglomerate in western Australia, which have yielded U–Pb ages as old as 4.28 Byr. The greater sensitivity of MC-ICP-MS (section 2.5.2) allowed the Hf analysis of single grains rather than bulk zircon separates, and therefore permitted a better understanding of this detrital zircon population with variable U–Pb ages. Results for 37 individual grains fell into three groups on a plot of , Hf against lead 207/206 age (Fig. 9.6). However, two of the groups lay on an isotopic growth line (dashed) with a ¹⁷⁶Lu/¹⁷⁷Hf ratio (0.022) typical of mafic rocks. This implies that the zircon grains in these two groups (whose elemental chemistry is indicative of felsic magmas) might be derived from granitoid rocks of various Early Archean ages, derived by melting of mafic crustal rocks.

Fig. 9.6. Plot of , Hf against 207/206 lead ages for 37 detrital zircon grains from the Jack Hills meta-conglomerate of the Narryer gneiss complex of western Australia. Error bar indicates average precision. Modified after Amelin et al. (1999).

Back projection of the Hf isotope crustal growth line to an intersection with the chondritic evolution line allowed the formation age of this mafic crustal terrane to be estimated. The result obtained using the conventional chondritic evolution line suggested a crustal formation age of around 4.1 Byr for this terrane. However, the new low decay constant would imply a chondritic evolution line about 3 epsilon units higher, and therefore a crustal formation age of around 4.4 Byr (Amelin et al., 2000). This implies that at least some crustal fragments were in existence almost immediately after the last giant impact event led to the formation of the Moon (section 15.5.4). On the other hand, the new high decay constant implies a chondritic line 2 - 3 epsilon units lower than the old value. This would imply that mantle depletion was occurring very early in Earth history.

9.2.3 Western Greenland

Another important application of Hf isotope data was to test Nd evidence for early crustal evolution in western Greenland. Nd isotope analysis of Amitsoq gneisses by Bennett et al. (section 4.4.4) had previously implied a large range of , Nd values, based on initial ratios calculated at the U–Pb ages of the rocks. This included some strongly positive , values, suggesting strong mantle depletion in the Early Archean. To test these results, Vervoort et al. (1996) analysed zircon separates from a selection of the samples analysed by Bennett et al. However, the , Hf values fell in a narrower range than , Nd, suggesting to Vervoort et al. that whole-rock Sm–Nd systems in some of these rocks had probably been disturbed.

In a continuation of this study, Vervoort and Blichert-Toft (1999) analysed Hf in both whole-rock samples and additional zircon separates. The results bore out the earlier work, showing that whole-rock Hf–Nd isotope systematics in the West Greenland samples depart from the Mantle Array defined by juvenile samples younger than 3.5 Byr (Fig. 9.7). Hence, the whole-rock Hf isotope data support the earlier data on zircon separates in implying that the Nd isotope system was somewhat disturbed. It should also be noted at this point that these conclusions apply to samples from the heavily reworked southern area of the Itsaq gneiss terrane (as it is now called), as well as the less reworked northern area (see section 4.4.4).

Fig. 9.7. Plot of initial , Hf against , Nd (using values of t from U–Pb ages) for Amitsoq gneisses ( ! ) compared with a variety of juvenile samples less than 3.5 Byr old ( " ). After Vervoort and Blichert-Toft (1999).

When initial Hf isotope ratios are plotted against U–Pb ages, the resulting Hf isotope evolution plot shows essentially linear depletion with time, supporting the model of progressive continental growth through time (Fig. 9.8). An exception to this linear evolution trend is exhibited by some zircon data from the Amitsoq gneisses, with ¹⁷⁶Lu/¹⁷⁷Hf values less than 0.1, which are shown by solid symbols. However, these samples are the most susceptible to perturbation by any errors in the decay constant, because their evolution lines depart most strongly from the chondritic evolution curve. In view of the uncertainties in the decay constant, these samples should be given less significance in interpreting the depleted mantle composition against time, suggesting that this is essentially a linear evolution trend.

Fig. 9.8. Hf isotope evolution plot showing initial ratios of juvenile mantle-derived rocks. Samples with low Lu/Hf ratios are shown by solid symbols. After Vervoort and Blichert-Toft (1999).

Notwithstanding the results previously obtained, Blichert-Toft et al. (1999) reversed their earlier position on gradual Hf mantle depletion in the Early Archean as a result of new Hf isotope analysis on the Isua supracrustal sequence of western Greenland. As described in section 9.1, this sample suite, and one from western Africa, gave apparent Lu–Hf isochron ages consistently lower than Sm–Nd isochron ages on the same whole-rock samples. This led Blichert-Toft et al. (1999) to invoke Hf mobility in these samples, whose effects they tried to ‘see through’ by a fairly complex modelling procedure. The result of this modelling was to create strongly positive , Hf and , Nd values, implying strong mantle depletion. However, in a scientific comment on this paper, Villa et al. (2001) pointed out that the use of the new Lu decay constant causes the Sm–Nd and Lu–Hf isochron ages to agree within error, avoiding the necessity for wholesale Hf isotope re-setting in these rocks. In the meantime, however, Albarede et al. came up with a different explanation of the Hf data. This was based on the Lu/Hf versus Sm/Nd diagram, which has been quite widely used in Hf isotope studies.

Firstly, Albarede et al. plotted measured ratios of Lu/Hf against Sm/Nd for Early Archean rocks of western Greenland, and compared them with these ratios in modern volcanic rocks (Fig. 9.9). They showed that a moderate number of Amitsoq gneisses and Isua metavolcanic rocks lay along the array of modern volcanic rocks, whereas samples of the Garbenschiefer unit of the Isua sequence lay well off the array. They interpreted this to mean that the latter unit (which is REE-depleted) has disturbed Lu–Hf and Sm–Nd systems, whereas the samples that plot within the modern array are undisturbed.

Fig. 9.9. Plot of measured concentration ratios of Lu/Hf and Sm/Nd for Amitsoq gneisses ( " ) and Isua metavolcanics ( ! ), compared to a ‘Mantle Array’ of modern volcanic rocks. After Albarede et al. (2000).

Having thus ‘screened’ the Early Archean samples for disturbance, Albarede et al. then plotted the ‘successful’ samples on a graph of time-integrated Lu/Hf versus Sm/Nd ratios (Fig. 9.10). The significance of the plotted ratios of Lu/Hf and Sm/Nd is that these are model values necessary to generate the initial isotope ratios of the samples (at the time of igneous crystallisation given by their U–Pb age) by isotopic evolution in a closed system starting from a chondritic composition 4.56 Byr ago. In essence, this is equivalent to plotting , Hf against , Nd. When the western Greenland Early Archean samples are compared with modern oceanic volcanics, lunar samples, and Martian (SNC) meteorites, it appears that the Greenland rocks show similar variations to other solar system bodies (the Moon and Mars) which were characterised by very heterogeneous isotope signatures in their early history. Hence, Albarede et al. inferred that the Greenland magmas were sampling a heterogeneous terrestrial mantle.

Fig. 9.10. Plot of model (time-integrated) Lu/Hf versus Sm/Nd ()) necessary to generate calculated initial Hf and Nd isotope ratios of igneous rocks at the time of their crystallisation. The ratios are chondrite normalised using the new Hf decay constant. ( <> ) = Moon; ( Î ) = Mars; ( ! ) = Isua; ( " ) = Amitsoq; After Albarede et al. (2000).

Because the argument put forward by Albarede et al. is quite complex, it is difficult to evaluate. However, there are a couple of hidden assumptions. The first is that colinearity with modern volcanics in Fig. 9.9 implies a lack of disturbance. This need not be true if the manner of disturbance of the gneisses was isotopic homogenisation between different rock types, as was suggested by Moorbath et al. (section 4.4.4).

The second assumption is that the spread of time-integrated Lu/Hf and Sm/Nd ratios in the Greenland rocks is similar to those of the other solar system bodies, and may therefore have originated in the same way. However, an examination of Fig. 9.10 shows that some of the Greenland rocks (in the top left quadrant) have much lower time-integrated Sm/Nd ratios than Lunar or Martian samples. These compositions represent sources that are very enriched relative to the chondritic evolution line (i.e. , Nd values well below the chondritic line). It is much more likely that these are disturbed crustal signatures rather than enriched mantle reservoirs. However, if we admit that these disturbed crustal rocks could nevertheless lie on the modern volcanic array in Fig. 9.9 then samples in the top-right quadrant of Fig. 9.10 could likewise be disturbed. Hence, the present author agrees with Kramers (2001) that the balance of evidence does not support the existence of strongly depleted sources in the Early Archean Earth.

9.2.4 Mantle depletion and recycling

Patchett and Tatsumoto (1980c) made the first Hf isotope measurements on selected MORB and OIB samples (that had previously been analysed for ¹⁴³Nd/¹⁴⁴Nd and ⁸⁷Sr/⁸⁶Sr). These data, augmented by Patchett (1983), show that ¹⁷⁶Hf/¹⁷⁷Hf very closely parallels ¹⁴³Nd/¹⁴⁴Nd in OIB (Fig. 9.11). However, MORB samples display a proportionally greater degree of spread in ¹⁷⁶Hf/¹⁷⁷Hf (60% of the total range for oceanic basalts) than ¹⁴³Nd/¹⁴⁴Nd (only 30% of the total range). Hence the MORB arrays in Fig. 9.11 are nearly three times steeper than the OIB arrays. Patchett and Tatsumoto attributed these differences to stronger fractionation of Lu/Hf than Sm/Nd and Sr/Rb in very trace-element depleted source regions such as MORB, due to the greater incompatibility displayed by Hf relative to Lu than Nd/Sm or Rb/Sr.

Fig. 9.11. Hf versus Sr and Nd isotope diagrams showing early data for oceanic volcanics. OIB samples define a Mantle Array, but MORB samples show some decoupling of Hf systematics from Sr and Nd. After Patchett (1983).

Compared with their radiogenic Hf isotope signatures (indicative of a depleted source), many MORB samples are enriched in Hf/Lu (corresponding to a low Lu/Hf ratio). This phenomenon has also been observed for other incompatible element systems such as Sm/Nd and U/Th, and is generally attributed to sequential extraction of very small degree melts (sections 6.2.2 and 13.3.1); however, it is more marked in the Lu)Hf system than the Sm)Nd system. Salters and Hart (1989) attributed this ‘hafnium paradox’ to residual garnet in the MORB source. Because Hf is not a true rare earth element, its chemistry is not coherent with that of the heavy REE. This is illustrated by the mineral)magma partition coefficients shown in Fig. 9.12. Hence there is an opportunity for more extreme fractionation in Lu/Hf than in Sm/Nd ratio when low-degree melting occurs at depths greater than 80 km, within the garnet stability zone.

Fig. 9.12. Partition coefficients for REE and Hf between minerals (cpx, garnet) and kimberlite magma. After Fujimaki et al. (1984).

Gruau et al. (1990) analysed 3.5 Byr-old komatiites from the Onverwacht Group of South Africa to see whether Archean depleted mantle displayed extreme ranges of Lu/Hf depletion due to garnet fraction. They compared Al-depleted and Al-enriched komatiites, whose contrasting chemistry is attributed to garnet fractionation in the mantle source. However, calculated initial Hf and Nd isotope ratios were indicative of similar degrees of mantle depletion (, from 0 to +2) in both Al-enriched and Al-depleted types. Hence there is no evidence for widespread garnet fractionation in the early Earth, such as might have resulted from a magma ocean. Consequently the major-element composition of the magma was attributed to garnet fractionation in the source during the komatiite melting event itself, a conclusion supported by Blichert-Toft and Arndt (1999).

Although the concept of garnet fraction from an ancient magma ocean has been rejected, Salters and Hart (1991) argued that variable garnet contents in ancient melting events could nevertheless explain the partial decoupling of Hf)Nd isotope systematics in MORB. When garnet is present in the residue from melting, this residue develops high Lu/Hf ratios, and hence, over time, a radiogenic Hf isotope signature. On the other hand, the melting residues of (garnet-free) spinel peridotites will have lower Lu/Hf ratios, resulting in less radiogenic Hf isotope signatures over geological time. Johnson and Beard (1993) termed these two types of depleted mantle signature DMM-I and DMM-II respectively, and showed that the latter type was also represented in the source of Tertiary basalts in the southern Rio Grande rift (USA). They argued that the DMM-I and DMM-II components were generated by ancient depletion events, and that intermediate MORB compositions were generated by mixing between them (Beard and Johnson, 1993).

Salters (1996) presented an enlarged Hf–Nd data set for MORB samples in order to re-examine the ‘hafnium paradox’ in ocean ridge magmatism. Trace element fractionation during melting is quantified using the ) value, which compares the actual Lu/Hf and Sm/Nd ratios in MORB samples with the time-integrated ratios necessary to generate observed isotopic signatures by evolution from a 2 Byr old chondritic source (Fig. 9.13).

Fig. 9.13 Plot of time integrated ratios of Sm/Nd and Lu/Hf ()) necessary to generate observed isotope ratios in MORB ( " ) and OIB ( !) over 2 Byr. Curves show model melting conditions which can explain the ) ratios as calculated. After Salters (1996).

Almost all MORBs and OIBs have significantly more extreme ) (Lu/Hf) values than ) (Sm/Nd) in Fig. 9.13, which implies melting in the garnet stability zone. In addition, average ) values (interpreted as garnet signatures) from four different ridges showed a positive correlation with the water depth over the ridge (Fig. 9.14), which in turn is inversely dependent on the rate of magma generation at the ridge. In other words, the ridge with the lowest rate of magma generation had the largest garnet signature. This is the opposite of what would be predicted for peridotite melting, since ridges with high magma generation rates should begin melting at greater depth allowing a longer melting interval in the garnet zone.

Fig. 9.14. Plot of calculated ) values for Sm/Nd (Fig. 9.13) against water depth for a selection of ridge segments. ) Lu/Hf shows a similar pattern. After Salters (1996).

To explain this paradox, Hirschmann and Stolper (1996) argued that Lu/Hf fractionation in MORB results from melting of garnet pyroxenite (rather than garnet peridotite) in a marble cake mantle. According to this model, the ‘garnet signature’ from melting of pyroxenite is preserved under deep ridges with low melting rates, but under shallow ridges this signature is diluted by large-scale melting of spinel peridotite. Evidence from other isotopic systems may help to resolve this debate (sections 8.3.6 and 13.3.4).

The advent of MC-ICP-MS offered the opportunity to test the earlier work on MORB samples by the analysis of larger sample suites at levels of precision nearly an order of magnitude better. At first, it appeared that the large spread of Hf isotope analyses originally measured by Patchett et al. on Atlantic MORB samples might not be reproduced (e.g. Nowell et al., 1998; Kempton et al., 2000). However, subsequent work by Chauvel and Blichert-Toft (2001) was able to reproduce most of the range of Hf isotope ratios seen in the early work on MORB.

Hf analyses of OIB samples were presented by Salters and Hart (1991) and Salters and White (1998) in order to establish the locations of the end-member components proposed by Zindler and Hart (section 6.4.2). The results (Fig. 9.15) showed that EMI and EMII were strongly colinear. On the other hand, the isolation of HIMU below the main trend suggests a possible connection with the DMM-II component, which was attributed by Johnson and Beard (1993) to ancient depleted spinel peridotite. In other words, if we take into account the location of the HIMU field below the main OIB array, the degree of decoupling between Hf and Nd isotope signatures in OIB mirrors that seen in MORB.

Fig. 9.15. Hf)Nd isotope diagram, showing fields for geochemically important ocean islands, along with the estimated compositions of end-members. The Leucite Hills represent subcontinental lithosphere. Modified after Salters and Hart (1991).

9.2.5 Sediment recycling

Another geological environment where Lu/Hf can undergo strong fractionation relative to Sm/Nd is the sedimentary system. Patchett et al. (1984) plotted Lu/Hf ratios against Sm/Nd for various types of marine sediment (Fig. 9.16). While ¹⁴⁷Sm/¹⁴⁴Nd ratios are more-or-less constant at ca. 0.12 ) 0.14 in most samples analysed, ¹⁷⁶Lu/¹⁷⁷Hf is strongly fractionated between sandstones and clays. Patchett et al. attributed this fractionation to the very strong affinity of Hf for zircon, which, because of its resistance to mechanical and chemical attack, becomes enriched in sand-grade sediments. Hf is correspondingly depleted in the fine-grained clay fraction.

Fig. 9.16. Plot of Lu/Hf versus Sm/Nd ratio in different sediment types, showing that large fractionations in Lu/Hf are not accompanied by significant changes in Sm/Nd. After Patchett et al. (1984).

The sorting of marine sediments according to grain size is expected to yield low-Lu/Hf sands and turbidites on the continental shelf and continental slope, medium-Lu/Hf shales and clays, and very-high-Lu/Hf red clays and Mn nodules in the deep ocean, where terrigenous sediment is lacking. These variations in sediment Lu/Hf ratios were in turn predicted to generate large variations in Hf isotope ratio after recycling and storage in deep mantle reservoirs. Hence, Patchett et al. (1984) calculated that any individual sediment type subducted (e.g. red clay, average pelagic sediment, or turbidite) would yield distinctive isotopic compositions after 2 Byr of residence in the mantle (Fig. 9.17). Since such dramatic divergences are not seen in OIB samples, Patchett et al. concluded that the magnitude of sediment recycling into the mantle must be strictly limited.

Fig. 9.17. Trends in OIB data predicted to result from the subduction of different types of sediment into the deep mantle, followed by storage for 1 – 2 Byr. After Patchett et al. (1984).

Island-arc basalts offer a means of monitoring the composition of material actually being recycled into the mantle, in order to test theoretical models such as Patchett et al. (1984). Hf)Nd isotope data were presented by White and Patchett (1984) for arc basalts sampling depleted and enriched sources (solid symbols in Fig. 9.18). These data fall within the field of OIB samples, showing that old sedimentary material presently being subducted into the mantle has appropriate Hf)Nd systematics to explain the composition of OIB magmas. Therefore, contrary to the prediction of Patchett (1984), this implies that non-extreme mixtures of the various sediment types can explain the composition of OIB sources with moderate ease.

Fig. 9.18. Hf–Nd isotope plot showing the composition of Island Arc Basalts ( ! , and marked fields) relative to MORB, OIB and subcontinental lithosphere. After Salters and Hart (1991).

This conclusion was supported by MC-ICP-MS analysis of Hf in a much larger suite of sediment samples by Vervoort et al. (1999). This work showed that Hf–Nd isotope systematics in the global sedimentary system were more coherent than previously expected. Apart from a very few extreme samples, the vast majority of sediments (of a wide variety of ages) lie along the same trend as the OIB mantle array (Fig. 9.19). This is particularly true for sediments from active margins, which are much more likely to be recycled into the mantle than passive margin sediments. Therefore, it is concluded that Hf isotope data provide a weaker constraint on sediment recycling into the mantle than originally expected, so that models involving terrigenous or pelagic sediment recycling into different OIB reservoirs are no longer ruled out.

Fig. 9.19. Hf–Nd isotope data for a large suite of sediments of different ages ( ! ) to show variation relative to the field of oceanic volcanics. After Vervoort et al. (1999).

With the advent of new high-precision Hf data sets from MC-ICP-MS, more subtle trends have been observed within the OIB data set, which can give a better understanding of the origins of OIB sources from various types of recycled crustal material. For example, a detailed Hf isotope study of Hawaiian lavas (Blichert-Toft et al., 1999) revealed two en echelon arrays formed respectively by the Koolau volcano and by a composite of several other volcanoes (Fig. 9.20). These arrays cut across the main Nd–Hf isotope array in OIB, with significantly lower slopes, and were attributed to a significant fraction of recycled pelagic sediment in parts of the Hawaiian plume, mixed in different proportions with a component of recycled oceanic lithosphere. This model is supported by osmium and oxygen isotope evidence (section 8.3.4). A similar shallow trend, consistent with a subducted sediment component, was subsequently found in Pitcairn samples (Eisele et al., 2002). Since Pitcairn is the most extreme example of the EMI mantle reservoir, this supports the model of pelagic sediment recycling into this OIB source (section 6.5.3).

Fig. 9.20. Plot of , Hf versus , Nd compositions for several Hawaiian volcanos relative to the main trend of the Hf–Nd OIB array. ( ! ) = Koolau volcano; ( " ) = Lanai & Kahoolawe; ( <> ) = Haleakala. After Blichert-Toft et al. (1999).

References

Albarede, F., Blichert-Toft, J., Vervoort, J. D., Gleason, J. D. and Rosing, M. (2000). Hf–Nd isotope evidence for a transient dynamic regime in the early terrestrial mantle. Nature 404, 488–90.

Albarede, F., Simonetti, A., Vervoort, J. D., Blichert-Toft, J. and Wafa, A. (1998). A Hf–Nd isotopic correlation in ferromanganese nodules. Geophys. Res. Lett. 25, 3895–8.

Amakawa, H., Ingri, J., Masuda, A. and Shimizu, H. (1991). Isotopic compositions of Ce, Nd and Sr in ferromanganese nodules from the Pacific and Atlantic Oceans, the Baltic and Barents Seas and the Gulf of Bothnia. Earth Planet. Sci. Lett. 105, 554)65.

Amakawa, H., Nozaki, Y. and Masuda, A. (1996). Precise determination of variations in the ¹³⁸Ce/¹⁴²Ce ratios of marine ferromanganese nodules. Chem. Geol. 131, 183–95.

Amelin, Y., Lee, D.-C. and Halliday, A. N. (2000). Early–middle Archaean crustal evolution deduced from Lu–Hf and U–Pb isotopic studies of single zircon grains. Geochim. Cosmochim. Acta 64, 4205–25.

Amelin, Y., Lee, D.-C., Halliday, A. N. and Pidgeon, R. T. (1999). Nature of the Earth’s earliest crust from hafnium isotopes in single detrital zircons. Nature 399, 252–5.

Beard, B. L. and Johnson, C. M. (1993). Hf isotope composition of late Cenozoic basaltic rocks from northwest Colorado, USA: new constraints on mantle enrichment processes. Earth Planet. Sci. Lett. 119, 495)509.

Bizzarro, M., Baker, J. A., Haack, H., Ulfbeck, D. and Rosing, M. (2003). Early history of Earth’s crust–mantle system inferred from hafnium isotopes in chondrites. Nature 421, 931–3.

Blichert-Toft, J. and Albarede, F. (1997). The Lu–Hf isotope geochemistry of chondrites and the evolution of the mantle–crust system. Earth Planet. Sci. Lett. 148, 243–58.

Blichert-Toft, J., Boyet, M., Telouk, P. and Albarede, F. (2002). ¹⁴⁷Sm–¹⁴³Nd and ¹⁷⁶Lu–¹⁷⁶Hf in eucrites and the differentiation of the HED parent body. Earth Planet. Sci. Lett. 204, 167–81.

Blichert-Toft, J., Albarede, F. Rosing, M., Frei, R. and Bridgwater, D. (1999). The Nd and Hf isotopic evolution of the mantle through the Archean. Results from the Isua supracrustals, West Greenland, and from the Birimian terranes of West Africa. Geochim. Cosmochim. Acta 63, 3901–14.

Blichert-Toft, J. and Arndt, N. T. (1999). Hf isotope compositions of komatiites. Earth Planet. Sci. Lett. 171, 439–51.

Blichert-Toft, J., Frey, F. A. and Albarede, F. (1999). Hf isotope evidence for pelagic sediments in the source of Hawaiian basalts. Science 285, 879–82.

Chauvel, C. and Blichert-Toft, J. (2001). A hafnium isotope and trace element perspective on melting of the depleted mantle. Earth Planet. Sci. Lett. 190, 137–51.

David, K., Frank, M., O’Nions, R. K., Belshaw, N. S. and Arden, J. W. (2001). The Hf isotope composition of global seawater and the evolution of Hf isotopes in the deep Pacific Ocean from Fe–Mn crusts. Chem. Geol. 178, 23–42.

Dickin, A. P. (1987a). La)Ce dating of Lewisian granulites to constrain the ¹³⁸La $-decay half-life. Nature 325, 337)8.

Dickin, A. P. (1987b). Cerium isotope geochemistry of ocean island basalts. Nature 326, 283)4.

Dickin, A. P. (1988). Mantle and crustal Ce/Nd isotope systematics. Nature 333, 403)4.

Dickin, A. P., Jones, N. W., Thirlwall, M. F. and Thompson, R. N. (1987). A Ce/Nd isotope study of crustal contamination processes affecting Palaeocene magmas in Skye, northwest Scotland. Contrib. Mineral. Petrol. 96, 455)64.

Dixon, D., McNair, A. and Curran, S. C. (1954). The natural radioactivity of lutetium. Phil. Mag. 45, 683)4.

Duchene, S., Blichert-Toft, J., Luais, B., Telouk, P. and Albarede, F. (1997). The Lu–Hf dating of garnets and the ages of the Alpine high-pressure metamorphism. Nature 387, 586–9.

Eisele, J., Sharma, M., Galer, S. J. G., Blichert-Toft, J., Devey, C. W. and Hofmann, A. W. (2002). The role of sediment recycling in EM-1 inferred from Os, Pb, Hf, Nd, Sr isotope and trace element systematics of the Pitcairn hotspot. Earth Planet. Sci. Lett. 196, 197–212.

Faure, G. (1977). Principles of Isotope Geology. Wiley, 464 p.

Fujimaki, H., Tatsumoto, M. and Aoki, K. (1984). Partition coefficients of Hf, Zr and REE between phenocryst phases and groundmass. Proc. 14th Lunar Planet. Sci. Conf., J. Geophys. Res. 89 (supp.), B662)72.

Godfrey, L. V., Lee, D.-C., Sangrey, W. F., Halliday, A. N., Salters, V. J. M., Hein, J. R. and White, W. M. (1997). The Hf isotopic composition of ferromanganese nodules and crusts and hydrothermal manganese deposits: implications for seawater Hf. Earth Planet. Sci. Lett. 151, 91–105.

Gruau, G., Chauvel, C., Arndt, N. T. and Cornichet, J. (1990). Aluminum depletion in komatiites and garnet fractionation in the early Archean mantle: Hafnium isotopic constraints. Geochim. Cosmochim. Acta 54, 3095)101.

Herr, W., Merz, E., Eberhardt, P. and Signer, P. (1958). Zur bestimmung der $ halbwertszeit des ¹⁷⁶Lu durch den nachweis von radiogenem ¹⁷⁶Hf. Z. Natur. 13a, 268)73.

Hirschmann, M. M. and Stolper, E. M. (1996). A possible role for garnet pyroxenite in the origin of the ‘garnet signature’ in MORB. Contrib. Mineral. Petrol. 124, 185–208.

Holmes, A. (1932). The origin of igneous rocks. Geol. Mag. 69, 543)58.

Johnson, C. J. and Beard, B. L. (1993). Evidence from hafnium isotopes for ancient sub-oceanic mantle beneath the Rio Grande rift. Nature 362, 441)4.

Kempton, P. D., Fitton, J. G., Saunders, A. D., Nowell, G. M., Taylor, R. N., Hardarson, B. S. and Pearson, G. (2000). The Iceland plume in space and time: a Sr–Nd–Pb–Hf study of the North Atlantic rifted margin. Earth Planet. Sci. Lett. 177, 255–71.

Kramers, J. (2001). The smile of the Cheshire Cat. Science 293, 619–20.

Lee, D.-C., Halliday, A. N., Hein, J. R., Burton, K. W., Christensen, J. N. and Gunther, D. (1999). Hafnium isotope stratigraphy of ferromanganese crusts. Science 285, 1052–4.

Makishima, A. and Masuda, A. (1994). Ce isotope ratios of N-type MORB. Chem. Geol. 118, 1)8.

Makishima, A., Nakamura, E., Akimoto, S., Campbell, I. H. and Hill, R. I. (1993). New constraints on the ¹³⁸La $-decay constant based on a geochronological study of granites from the Yilgarn Block, Western Australia. Chem. Geol. (Isot. Geosci. Section) 104, 293)300.

Marshall, B. D. and DePaolo, D. J. (1982). Precise age determination and petrogenetic studies using the K)Ca method. Geochim. Cosmochim. Acta 46, 2537)45.

Marshall, B. D. and DePaolo, D. J. (1989). Calcium isotopes in igneous rocks and the origin of granite. Geochim. Cosmochim. Acta 53, 917)22.

Masuda, A., Shimizu, H., Nakai, S., Makishima, A. and Lahti, S. (1988). ¹³⁸La $-decay constant estimated from geochronological studies. Earth Planet. Sci. Lett. 89, 316)22.

Nakai, S., Shimizu, H. and Masuda, A. (1986). A new geochronometer using lanthanum-138. Nature 320, 433)5.

Nelson, D. R. and McCulloch, M. T. (1989). Petrogenetic applications of the ⁴⁰K)⁴⁰Ca radiogenic decay scheme ) a reconnaissance study. Chem. Geol. (Isot. Geosci. Section) 79, 275)93.

Nir-El, Y. and Lavi, N. (1998). Measurement of the half-life of ¹⁷⁶Lu. Appl. Radiat. Isot. 49, 1653–5.

Norman E. B. and Nelson M. A. (1983). Half-life and decay scheme of ¹³⁸La. Phys. Rev. C 27, 1321)4.

Nowell, G. M., Kempton, P. D., Noble, S. R., Fitton, J. G., Saunders, A. D., Mahoney, J. J. and Taylor, R. N. (1998). High precision Hf isotope measurements of MORB and OIB by thermal ionisation mass spectrometry: insights into the depleted mantle. Chem. Geol. 149, 211–33.

Patchett, P. J. (1983). Hafnium isotope results from mid-ocean ridges and Kerguelen. Lithos 16, 47)51.

Patchett P. J., Kouvo O., Hedge C. E. and Tatsumoto M. (1981). Evolution of continental crust and mantle heterogeneity: Evidence from Hf isotopes. Contrib. Mineral. Petrol. 78, 279)97.

Patchett P. J. and Tatsumoto M. (1980a). A routine high-precision method for Lu)Hf isotope geochemistry and chronology. Contrib. Mineral. Petrol. 75, 263)7.

Patchett, P. J. and Tatsumoto, M. (1980b). Lu)Hf total-rock isochron for the eucrite meteorites. Nature 288, 571)4.

Patchett P. J. and Tatsumoto M. (1980c). Hafnium isotope variations in oceanic basalts. Geophys. Res. Lett. 7, 1077)80.

Patchett P. J. and Tatsumoto M. (1981). Lu/Hf in chondrites and definition of a chondritic hafnium growth curve. Lunar Planet. Sci. 12, 822)4, Lunar Planet. Inst.

Patchett P. J., White W. M., Feldmann H., Kielinczuk S. and Hofmann A. W. (1984). Hafnium/rare earth element fractionation in the sedimentary system and crustal recycling into the Earth’s mantle. Earth Planet. Sci. Lett. 69, 365)78.

Pettingill, H. S. and Patchett, P. J. (1981). Lu)Hf total rock age for the Amitsoq gneisses, West Greenland. Earth Planet. Sci. Lett. 55, 150)6.

Pettke, T., Lee, D.-C., Halliday, A. N. and Rea, D. K. (2002). Radiogenic Hf isotopic compositions of continental aeolian dust from Asia, its variability and its implications for seawater Hf. Earth Planet. Sci. Lett. 202, 453–64.

Piotrowski, A. M., Lee, D.-C., Christensen, J. N., Burton, K. W., Halliday, A. N., Hein, J. R. and Gunther, D. (2000). Changes in erosion and ocean-circulation recorded in the Hf isotopic compositions of North Atlantic and Indian Ocean ferromanganese crusts. Earth Planet. Sci. Lett. 181, 315)25.

Russell W. A., Papanastassiou D. A. and Tombrello T. A. (1978). Ca isotope fractionation on the Earth and other solar system materials. Geochim. Cosmochim. Acta 42, 1075)90.

Salters, V. J. M. (1996). The generation of mid-ocean ridge basalts from the Hf and Nd isotope perspective. Earth Planet. Sci. Lett. 141, 109–23.

Salters, V. J. and Hart, S. R. (1989). The hafnium paradox and the role of garnet in the source of mid-ocean-ridge basalts. Nature 342, 420)2.

Salters, V. J. and Hart, S. R. (1991). The mantle sources of ocean ridges, islands and arcs: the Hf-isotope connection. Earth Planet. Sci. Lett. 104, 364)80.

Salters, V. J. M. and White, W. M. (1998). Hf isotope constraints on mantle evolution. Chem. Geol. 145, 447–60.

Salters, V. J. M. and Zindler, A. (1995). Extreme ¹⁷⁶Hf/¹⁷⁷Hf in the sub-oceanic mantle. Earth Planet. Sci. Lett. 129, 13–30.

Sato J. and Hirose T. (1981). Half-life of ¹³⁸La. Radiochem. Radioanal. Lett. 46, 145)52.

Scherer, E. E., Cameron, K. L. and Blichert-Toft, J. (2000). Lu–Hf garnet geochronology: closure temperature relative to the Sm–Nd system and the effects of trace mineral inclusions. Geochim. Cosmochim. Acta 64, 3413–32.

Scherer, E. E., Cameron, K. L., Johnson, C. M., Beard, B. L., Barovich, K. M. and Collerson, K. D. (1997). Lu–Hf geochronology applied to dating Cenozoic events affecting lower crustal xenoliths from Kilburn Hole, New Mexico. Chem. Geol. 142, 63–78.

Scherer, E., Munker, C. and Mezger, K. (2001). Calibration of the lutetium–hafnium clock. Science 293, 683–8.

Shimizu, H., Tanaka, T. and Masuda, A. (1984). Meteoritic ¹³⁸Ce/¹⁴²Ce ratio and its evolution. Nature 307, 251)2.

Stevenson, R. K. and Patchett, P. J. (1990). Implications for the evolution of continental crust from Hf isotope systematics of Archean detrital zircons. Geochim. Cosmochim. Acta 54, 1683)97.

Tanaka, T. and Masuda, A. (1982). The La)Ce geochronometer: a new dating method. Nature 300, 515)18.

Tanaka, T., Shimizu, H., Kawata, Y. and Masuda, A. (1987). Combined La)Ce and Sm)Nd isotope systematics in petrogenetic studies. Nature 327, 113)17.

Tanaka, T., Shimizu, H., Kawata, Y. and Masuda, A. (1988) Reply to: Dickin, A. P. (1988). Mantle and crustal Ce/Nd isotope systematics. Nature 333, 403)4.

Tanaka, T., Usui, A. and Masuda, A. (1986). Oceanic Ce and continental Nd: multiple sources of REE in oceanic ferromanganese nodules. Terra Cognita 6, 114 (abstract).

Tatsumoto, M., Unruh, D. M. and Patchett, P. J. (1981). U)Pb and Lu)Hf systematics of Antarctic meteorites. Nat. Inst. Polar Res. Tokyo.

Vervoort, J. D. and Blichert-Toft, J. (1999). Evolution of the depleted mantle: Hf isotope evidence from juvenile rocks through time. Geochim. Cosmochim. Acta 63, 533–56.

Vervoort, J. D., Patchett, P. J., Blichert-Toft, J. and Albarede, F. (1999). Relationships between Lu–Hf and Sm–Nd isotopic systems in the global sedimentary system. Earth Planet. Sci. Lett. 168, 79–99.

Vervoort, J. D., Patchett, P. J., Gehrels, G. E. and Nutman, A. P. (1996). Constraints on early Earth differentiation from hafnium and neodymium isotopes. Nature 379, 624–7.

Villa, I. M., Kamber, B. Z. and Nagler, T. F. (2001). Comment on “The Nd and Hf isotopic evolution of the mantle through the Archean. Results from the Isua supracrustals, West Greenland, and from the Birimian terranes of West Africa” by Blichert-Toft et al. (1999). Geochim. Cosmochim. Acta 65, 2017–21.

White, W. M. and Patchett, J. (1984). Hf)Nd)Sr isotopes and incompatible element abundances in island arcs: implications for magma origins and crust)mantle evolution. Earth Planet. Sci. Lett. 67, 167)85.

White, W. M., Patchett, J. and Ben Othman, D. (1986). Hf isotope ratios of marine sediments and Mn nodules: evidence for a mantle source of Hf in seawater. Earth Planet. Sci. Lett. 79, 46)54.