Chapter One
Geochemistry and Secular Geochemical Evolution of the Earth's Mantle and Lower Crust BALZ S. KAMBER Laurentian University, Sudbury, Ontario, Canada
The incompatible elements U and Th are related to Pb via radioactive decay. Extraction, modification and storage of continental crust have, over time, left an isotopic record in the continental crust itself and in the depleted portion of the mantle. Ancient lower crustal xenoliths require that crust has matured by upward transport of radioactive heat-producing elements; hundreds of millions of years after formation.
Recycling of continental material has contributed in at least three ways to the generation of enriched mantle-melt sources. First, this has occurred by delamination of lower crustal segments back into the mantle. Second, sediment has been recycled back into the mantle in subduction zones, and third, since the oxygenation of the atmosphere, seawater U, weathered from the continents, has been incorporated into hydrated oceanic crust with which it has ultimately been recycled back into the mantle.
The joint treatment of the lower continental crust and the mantle in terms of their geochemistry and their isotopic evolution may seem, at first, a less than obvious choice. They are, however, related in the sense that the evidence for their evolution is largely of indirect nature, either inferred from rare xenoliths or via products of partial melting. Any joint treatment of these two geochemical reservoirs also inherently carries with it the assumption that they have, at least in part, mutually influenced each other's temporal evolution. Before attempting to condense into an opening book chapter the relevant aspects of the exhaustive body of knowledge about the geochemistry of the mantle and the much sparser information regarding the lower crust, it is necessary to remind ourselves of the evidence for their mutually related evolutions.
INTRODUCTION
The view that the Earth has suffered some form of early global, planetary-scale depletion event is deeply rooted in classic geochemical texts, including those focusing on plumbotectonics, i.e. the reconstruction of planetary differentiation from a Pb-isotope perspective (e.g. Stacey and Kramers 1975). Most early attempts at modelling the isotopic evolution of the mantle postulated one or two pervasive differentiation steps, resulting, for example, in the increase of the U/Pb ratio of the silicate portion of the Earth (the bulk silicate Earth). The notion of an early depletion event was further cemented with the observation that Archaean komatiites and high-Mg basalts, in terms of their trace-element chemistry, appeared to resemble modern ocean-island picrites, yet their radiogenic isotope character was much more depleted (e.g. Campbell and Griffiths 1993). This finding seemed to suggest an early depletion event that imparted the long-term isotopic effect with superimposed much more recent (relative to 2.7Ga) re-enrichment of the mantle, which explains the trace-element systematics but which had not yet translated into long-term isotopic evidence. More recently, the observation has been made that the bulk silicate Earth has a 142Nd/144Nd ratio different from the most common chondritic meteorites (Boyet and Carlson 2005). This has added new momentum to the idea of a very early silicate differentiation event that must have occurred within less than 1 half-life (103Ma) of the short-lived parent of 142Nd.
While the evidence for such an event appears as strong as ever, the critical question for this present treatment is whether that event was the principal cause for establishing the chemistry of the depleted mantle as it is sampled at most ocean ridges via the normal mid-ocean-ridge basalt (NMORB). Namely, if the early depletion event imparted such a fundamental geochemical signal, which over time also was manifest as a long-lived radiogenic isotope signature, then the subsequent extraction, maturing and recycling of continental crust would only have played a secondary role in modifying the chemistry of the depleted mantle. Hence, the chemistry and radiogenic isotope composition of the depleted mantle would largely tell us about the early planetary depletion event and not about the history of extraction and recycling of continental crust.
In order to address this question, it is necessary to consider elemental systematics and the radiogenic isotope evolution of those elements that are most strongly enriched in continental crust, for their extraction will be most strongly reflected by the residual depleted mantle. While average continental-crustal absolute abundances are difficult to estimate on account of the sparse occurrence of bona fide lower crustal rocks, there is nonetheless wide agreement regarding the relative enrichment of elements. The elements most strongly enriched in continental crust are largely those that behave most incompatibly during mantle melting, plus an assortment of elements that are particularly soluble in hydrous fluids, and were therefore preferentially moved into the meltsource regions of the magmas that eventually differentiated to give rise to continental crust. The best studied of these is Pb (e.g. Miller et al. 1994) but other fluid mobile elements, such as B (e.g. Ryan and Langmuir 1993), W (e.g. Kamber et al. 2005; Knig et al. 2008), Li (e.g. Chan et al. 1999) and As (Mohan et al. 2008) have also been documented. The extended trace-element diagram for average upper-continental crustal rocks, in which elements are arranged in order of incompatibility during mantle-decompression melting, illustrates not only the extraordinary enrichment of the most incompatible elements but also the strong deviations of the fluid-mobile elements from an otherwise predicable, smoothly decaying trend. Regardless of the particular significance of the elements that deviate from this trend, it is intuitively appreciable that the geoscientist interested in that aspect of mantle depletion potentially caused by the extraction of continental crust is best served by working with the elements that plot toward the left side of the abscissa of Fig. 1.1. From an isotopic point of view, it is therefore not surprising that the extent of variability in the U/ Pb and Th/Pb isotope systems in crustal and mantle rocks is of the order of several tens of percent and has formed the very basis of the mantle-rock nomenclature.
Indeed, one of the strongest pieces of evidence for the mutual chemical interaction between mantle and crust is found in the Pb-isotope composition and U-Th-Pb systematics of the source of N-MORB basalts. The present-day Pb-isotope composition of N-MORB firmly shows that, on the billion-year timescale, the time-averaged Th/U ratio of the depleted mantle source was ca. 3.6. This can be inferred from the 208Pb/206Pb ratio (Kramers and Tolstikhin 1997), which represents the decay products of the long-lived 232Th and 238U, respectively. Rather surprisingly, then, the measured elemental Th/U ratio of N-MORB is much lower, somewhere between 2.4 and 2.6. This observation is often termed the second terrestrial Pb-isotope paradox (e.g. Kramers and Tolstikhin 1997) or the kappa (as in 232Th/238U) conundrum (e.g. Elliott et al 1999). This discrepancy is not an artefact of preferential U over Th partitioning into the N-MORB parental melt because the intermediate decay product systematics of the U and Th chains support a low Th/U ratio of the source rocks, i.e. the depleted mantle itself (Galer and O'Nions 1985). The solution to this paradox is now widely believed (e.g. McCulloch 1993; Elliott et al. 1999; Collerson and Kamber 1999) to be the preferential recycling of continental U under an oxidized atmosphere since the great oxygenation event at ca. 2.3Ga (Bekker et al. 2004). This observation alone provides very robust evidence that the depleted mantle has not remained chemically inert and unchanged since an early depletion event.
The high-field-strength elements Th, U, Nb, and Ta offer further insight into the interaction between the depleted mantle and continental crust. These elements are all very incompatible and have very similar bulk partition coefficients during mantle-decompression melting. This is reflected in their close grouping in the extended trace-element diagram (Fig. 1.1). Yet the chemistry of upper continental crust shows a very distinctive deficit in Nb (and to a lesser extent Ta) relative to Th and U. This finding is very widely attributed to the preferential sequestering of Nb and Ta into a Ti-phase (e.g. rutile) in subducting slabs (e.g. Hofmann 1988). Extraction of continental crust, to the extent of its present mass of ca. 2.09 x 1025g, has severely depleted the entire mantle in Th and U. It is estimated that between 30–50% of terrestrial Th and U are harboured by continental crust. By contrast, enrichment in the equally incompatible Nb is much lower, and, hence the mantle is proportionally less depleted in this element by a factor of at least three. It should come as no surprise then that the modern N-MORB Nb/Th ratio of ca. 18 is much higher than that of chondrites of ca. 8. If this greaterthan-100% difference in a ratio that can be analysed to within 2–5% precision was caused by the early depletion event, it follows that ancient melting products of the depleted mantle should also have a ratio of ca. 18; but this is not in fact the case. For example, it is found that regardless of locality, high-Mg basalts and komatiites of the widespread 2.7Ga mantle melting event have a Nb/Th of only 12 (e.g. Sylvester et al. 1997; Collerson and Kamber 1999), much lower than modern depleted mantle melts and much closer to the chondritic value. This observation shows that, at least for the very incompatible elements, the mantle has become more depleted as a function of how much continental crust was extracted. For these elements, the early depletion event played a less important role and, therefore, they are the tools with which to most effectively reconstruct the depletion history of the mantle.
TEMPORAL EVOLUTION OF THE DEPLETED MANTLE RESERVOIR
There are two principal methods to reconstruct the depletion history of the N-MORB mantle source. The first is to search for well-preserved N-MORB-like rocks of as large an age range as possible and to study their chemical and radiogenic isotope systematics. The second is to use forward modelling to approximate the isotopic contrast displayed by modern N-MORB and average continental crust. Examples of both approaches are reviewed here.
The reconstruction approach has the obvious advantage that each temporal observation from ancient N-MORB samples provides a time capsule for the evolution from the primitive to the present-day depleted mantle. In practice, it turns out that finding well-preserved N-MORB comparable basalts is difficult. The densest array of observations is, surprisingly, from the Archean eon. Many well-preserved greenstone belts exist, ranging in age from 3.7 to 2.6Ga, and while some are clearly ensialic in origin (e.g. Blenkinsop et al. 1993), a sufficient number of uncontaminated mafic to ultra-mafic volcanic rocks are preserved. The situation for the Proterozoic is much less satisfactory. Apart from two ophiolites (Zimmer et al. 1995; Peltonen et al. 1996), the majority of other Proterozoic greenstones either formed in an arc or back-arc environment (e.g. Leybourne et al. 1997), were variably contaminated during magmatic ascent through pre-existing continental crust, or are not sufficiently well-preserved. For the Phanerozoic, the number of ophiolites and accreted ocean-floor assemblages is adequate. It must be stressed here that N-MORB of any age is particularly sensitive to continental contamination in those elemental systematics of most interest to this discussion, the systematics of those elements for which there is the most divergence between the mantle and continental crust.
In terms of suitable element systematics for reconstruction, any pair of elemental neighbours with sharply deviating behaviour on Fig. 1.1 are candidates. Namely, for elements with nearidentical bulk partition coefficients during mantle melting, a suitably large-degree melt (such as the parental melt of N-MORB) will truthfully reflect the relative concentrations in the source. Subsequent fractional crystallization (up to ca. 6% MgO) will also not greatly affect the ratio of the elements of interest. Theoretically at least, it should be possible to track mantle depletion by study of the following ratios: Th/W, Nb/Th, Ta/U, Be/B, Pr/Pb, and Zr/Li. Note that, in all these examples, the element more enriched in continental crust is the denominator and hence all ratios are expected to have increased in the depleted mantle with increasing extraction of continental crust.
In reality, a number of factors conspire to render most of these ratios less than useful for the intended purpose. Insufficient data are available for Th/W, Be/B and Zr/Li. Post-emplacement elemental mobility may affect Pr/Pb and Be/B, and the redox-sensitivity of U has affected the mantle Ta/U ratio. At present, then, the only viable ratio is Nb/Th, which was used earlier to illustrate the fact that the N-MORB source mantle has become depleted by extraction of continental crust. Jochum et al. (1991) first proposed that the reconstruction of this ratio in the depleted mantle should be a reliable monitor of the mass of continental crust that had been extracted from the mantle through time, but their limited dataset and, by modern standards, insufficient analytical precision prevented these authors from drawing a conclusion. Collerson and Kamber (1999) applied a three-fold filter to the by then much improved literature database for Nb/Th in greenstones. They eliminated most rocks that had less than 6% MgO, excluded rocks with negative slopes in CI-normalized rare earth element (REE) patterns (to screen against ocean island basalts; OIB) and rejected rocks that had lower radiogenic 143Nd/144Nd ratios than widely accepted depleted-mantle evolution curves, (such as dePaolo and Wasserburg 1976) to avoid contaminated samples.
The Nb/Th curve for the depleted mantle, depicted on Fig. 1.2(a), was converted into the continental crust mass-versus age, curve (shown on Fig. 1.2(b)), that uses a primitive mantle Nb/ Th starting value lower than in chondrites to allow for sequestration of ca. 15% Nb into the core (following Kamber et al. 2003) because Nb can become siderophile under very reducing conditions prevailing during metal removal into the core (Wade and Wood 2001). The curve suggests a sigmoidal evolution for Nb/Th in the depleted mantle, starting with relatively low ratios until 3.5Ga, then increasing strongly between 3.0 and 2.0Ga, and a slow increase ever since.
The second approach to track mantle depletion is to study the time-integrated effect of continental extraction and recycling on depleted mantle isotope systematics. Most readers are probably familiar with the long-lived 147Sm/143Nd system. Owing to the slightly higher incompatibility of Nd, continental crust has a lower Sm/Nd ratio than its mantle source and as a result, over time, will develop a lower 143Nd/144Nd ratio. The contrary situation is, of course, true for the depleted portion of the mantle. However, because neither Sm nor Nd are nearly as concentrated in continental crust as Th, and because Sm/Nd fractionation is much more modest than Nb/Th, it turns out that the present-day mantle 143Nd/144Nd ratio is not very sensitive to the extraction history and recycling rate of continental crust. Nägler and Kramers (1998) explained, in detail, that Nd-isotope systematics cannot easily discriminate between models with linear net growth of the continents or that producing the sigmoidal curve shown in Fig. 1.2 (b). However, Nd-isotope systematics do argue against very early formation of voluminous continents and subsequent recycling (to lower the average continental age to ca. 2 Ga).
(Continues...)
Excerpted from Frontiers in Geochemistry Copyright © 2011 by John Wiley & Sons, Ltd. Excerpted by permission of John Wiley & Sons. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
Wishlist
Share
Tweet
Feedback
