Isotopic traces from earliest Earth history and evolutionary trends

A window on fractionation events in the early history of the crust and mantle is provided by radioactive isotope systematics with a relatively short half-life, such as 235U-207Pb and 146Sm-142Nd and by Hf isotopes in very ancient detrital zircon grains, which act as time capsules.

In a case study of Pb isotopes in ~ 3.7 Gyr metasediments and gneisses in the Godthaabsfjord area, West Greenland, a large discrepancy in the 207Pb/204Pb ratios for similar 206Pb/204Pb ratios is revealed (Fig. 27.15). Chemical metasediments (cherts and banded-iron formations, BIF) as well as gneisses and clastic metasediments define arrays that yield 207Pb/206Pb ages of ~ 3.7 Gyr, but the chemical metasediment data are offset to higher 207Pb/204Pb values by a large amount, ~ 0.5. To explain this discrepancy, a differentiation event at 4.3 Gyr is envisaged, in which a high-\ enriched material, with \ = 10.3, is extracted from a BSE with an average \ = 7.7 (Fig. 27.15). The 207Pb/204Pb and 206Pb/204Pb ratios increase faster in this material than in the BSE. As 235U was abundant in early Earth history, the difference in the 207Pb/204Pb ratios is particularly pronounced. Fractionation of BSE-like (or slightly depleted) matter gives rise to ages for the Amitsoq and South of Isua gneisses between 3.8 and 3.7 Gyr, and the enriched material provides Pb to the chemical sediments. Kamber etal. (2003) argued that the enriched material was a protocrust segment preserved at the Earth's surface from 4.3 to 3.7 Gyr and subsequently destroyed. Late bombardment of the Earth (by analogy with the Moon, Section 21.5) could have played a role in the destruction of this Hadean crust, as is also implied by the 146Sm-142Nd data in West Greenland gneisses (see below).

Less pronounced but nevertheless significant variations in 207Pb/204Pb values at a given 206Pb/204Pb value are quite common in the Archaean continental crust, leading to a broadening in the growth curve array (Fig. 27.11). In contrast, Pb-isotope compositions in Archaean volcanic rocks portray a much more uniform (or much better mixed) reservoir, with respect to both the U-Pb and Th-U-Pb prehistories, than the Archaean crust (Tilton, 1983). In the generation of tholeiitic and ultramafic melts, or their early differentiation, no fractionation of the Th/U ratio is expected (both are highly incompatible and melt fractions are large), so that the Th/U ratio of their mantle source can be estimated from the 208Pb/204Pb versus 206Pb/204Pb array. For the Archaean data this gives a value of 3.32, which is much higher than the present-day value of about 2.6 (Section 27.2).

Fig. 27.15 Lead-isotope data for chemical metasediments (banded-iron formations and cherts) and for the Amitsoq and South of Isua gneisses (SOI), West Greenland. Reference lines: chemical sediments (thickbroken line); gneisses (thick solid line). Both portray ages close to 3.75 Gyr. The arrays are offset in their 207Pb/204Pb ratios. In the simplified model shown, a system with \x = 7.7 starts with Canyon Diablo primordial Pb (CD) and accumulates radiogenic 206Pb and 207Pb from 4.55 Gyr onwards (open diamonds, shown at 100 Ma intervals from 4.4 Gyr). At 4.3 Gyr, differentiation produces a subsystem with \x = 10.3 (solid diamonds). Gneisses are chiefly produced from the first system and the Pb in chemical sediments is derived from the second. After Kramers (2007). Reproduced by permission of the Geological Society, London.

An important observation relevant to the formation and early evolution of terrestrial materials is that the 142Nd/144 Nd ratios in all terrestrial samples exceed the chondritic value by +0.2 s 142 units (Fig. 27.16(a), (b)). Assuming a chondritic-Earth model, this would imply that the Earth's mantle acquired a subchondritic Sm/Nd ratio during the first ~ 100 Myr of Earth history, when the parent 146Sm was still live. Assuming a chondritic-Earth model, such fractionation could have resulted from the extraction of enriched crust-like material, which would be then apparently isolated from Earth's accessible reservoirs (EARs). This hidden reservoir could possibly be the D" layer (Chapter 19).

Further, after much uncertainty it has been confirmed that in the Godthaabs-fjord area, West Greenland, the ~ 3.7 Gyr gneisses as well as metasediments have 142Nd/144Nd ratios higher by ~ 0.1 s142 units than those of all the other

Fig. 27.15 Lead-isotope data for chemical metasediments (banded-iron formations and cherts) and for the Amitsoq and South of Isua gneisses (SOI), West Greenland. Reference lines: chemical sediments (thickbroken line); gneisses (thick solid line). Both portray ages close to 3.75 Gyr. The arrays are offset in their 207Pb/204Pb ratios. In the simplified model shown, a system with \x = 7.7 starts with Canyon Diablo primordial Pb (CD) and accumulates radiogenic 206Pb and 207Pb from 4.55 Gyr onwards (open diamonds, shown at 100 Ma intervals from 4.4 Gyr). At 4.3 Gyr, differentiation produces a subsystem with \x = 10.3 (solid diamonds). Gneisses are chiefly produced from the first system and the Pb in chemical sediments is derived from the second. After Kramers (2007). Reproduced by permission of the Geological Society, London.

Metasediments

Orthogneisses

Metabasalts

Amphibolite enclave

Acasta Barberton

Metasediments

Orthogneisses

Metabasalts

Amphibolite enclave

Acasta Barberton

jBelingwe }Kostomuksha

Chondrites

Komatiites

Pitcairn (EM I ) OIBs

Indian ocean Pacific ocean

MORBs Carbonatites i i-jjfti1! W Greenland

> South Africa ^ Colorado Montana

Abee i

Sharp ji-

Bruderheim {

Allende

Terrestrial rocks

External error e

Chondrites

Sharp ji-

Fig. 27.16 Meteoritic and terrestrial 142Nd/144Nd ratios. The ratios are expressed in the e142 notation (Eqn 19.1b). The grey vertical bars show the average external precision obtained on repeat analyses of the same sample. Error bars on the individual points show the 2a deviation obtained for multiple measurements of each sample. See Section 21.3 and Tables 3.3 and 28.1 for the Sm-Nd systematics data. (a) Several samples from the Godthaabsfjord area, West Greenland, indicate a more depleted source than other terrestrial samples. (b) The terrestrial samples are all depleted relative to the chondritic composition. After Boyet et al. (2003), Boyet and Carlson (2005, 2006) and Caro et al. (2003, 2006), © Elsevier Science 2006, reproduced by permission (both (a) and (b)).

terrestrial rocks measured; these included samples from many Archaean cratons (Fig. 27.16(a)). Irrespective of which precise mechanism is invoked to explain this difference, it requires that some reservoir, depleted in incompatible trace elements, has remained isolated since the earliest few hundred Myr of Earth history to give rise to this regional anomaly. The analogy to the above-discussed Pb-isotope case is only partial, as the reservoir required there should have been enriched.

Zircon has proved to be a most useful mineral for studying differentiations in the earliest history of the Earth. It can be accurately dated, with U/Pb systematics, and because of its intrinsic high-Hf content and low Lu/Hf ratios, the 176Hf/177Hf ratios have not substantially changed within it. Thus the 176Hf/177Hf ratios of the parental magmas can be readily reconstructed. Further, zircon is a chemically and mechanically highly resistant mineral, which is not destroyed in sediment transport (Fig. 27.17).

The most ancient terrestrial matter hitherto found consists of detrital zircon grains in metasediments in the Mt Narryer and Jack Hills areas of the Yilgarn craton, western Australia. Among the most ancient zircons there are samples with 176Hf/177Hf ratios lying above the bulk silicate Earth development line. This shows the occurrence of depleted-magma-source regions as early as ~ 4.5 Gyr ago. However, most data, particularly from the somewhat younger detrital zircons, plot variably below this line. This points to enriched rock provinces, with Lu/Hf ratios below BSE values, that originated a few hundred Myr before they partially melted and the zircons were formed. From a comparison with the 176Hf/177Hf-ratio development slopes in Fig. 27.17, such rock provinces appear on average to have had the Lu/Hf ratio of continental crust generated from a BSE-type mantle.

Non-detrital zircons from Archaean gneiss terrains do not show a large scatter in 176Hf/177Hf ratios around the BSE development line but plot mainly below it (Fig. 27.17). These zircons thus probably crystallized either from mantle-derived evolved melts (e.g. the Itsaq, Barberton and Pilbara samples) or via the remelting of pre-existing enriched crust (the Acasta samples). Thus, even if depleted domains were generated in the mantle in the first few hundred Myr of Earth history, they appear to have been obliterated by the time of the early Archaean (~ 3.8 Gyr), presumably by remixing.

Numerous Nd-isotope data on whole-rock samples, spanning all the 3.8 Gyr geological history, illustrate an increasing diversity of s143 in the DMM and continental crust as time progresses (Fig. 27.18). This is expected as the differentiated reservoirs age. In the oldest rocks, the Sm-Nd systematics have been modified mostly by episodes of metamorphism, so that careful selection is needed before applying these data to Nd-isotope evolution modelling (Nagler and Kramers, 1998). The upper bound of the data envelope in Fig. 27.18 approximately defines the growth in

0.2807

0.2805

6 0.2801

0.2799

0.2797

0 0

Post a comment