Introduction
For us, recycling involves great effort — sorting empty bottles, collecting old newspapers and taking the broken fridge to the recycling site. The Earth, however, can recycle its waste naturally. The Earth's scum consists of ocean sediments and the rigid layer beneath composed of oceanic crust and the uppermost oceanic mantle. The movement of the tectonic plates ensures that this scum is efficiently disposed off at subduction zones where one plate dives beneath another, back into the deeper parts of the mantle. Within the mantle, these materials then get refashioned into several possible products: dead residues at the core–mantle boundary, enriched regions, or hot low-density plumes that rise through the mantle to melt and create new crust at ocean islands such as Hawaii or Samoa. The isotopic compositions of lavas that ooze from such ocean islands have long been thought to provide evidence of the recycling of subducted material into upwelling plumes1. On page 64 of this issue, Chauvel and co-authors2 reproduce the hafnium and neodymium isotopic compositions of ocean-island basalts by simulating the mixing of ancient oceanic crust, as well as sediments, with the ambient mantle. Their results suggest that both subducted oceanic crust and sediments are important recycled components in the sources of ocean-island lavas.
In most of the Earth's rocks, Hf and Nd isotopes co-vary in a predictable way: the 143Nd/144Nd ratios of mantle-derived magmas and continental rocks, plotted against the ratios of 176Hf/177Hf, fall on a straight line, forming the so-called 'terrestrial array'3. However, marine sediments are exceptional in this respect, as they have higher Hf ratios than other rocks at a given Nd ratio, the cause of which has been debated4, 5. Because of this, if only marine sediments were recycled into the mantle, the mantle Nd and Hf isotopic compositions would deviate from the terrestrial array. This would give a different signature to that typically found for ocean-island lavas. However, oceanic crust does not fit the bill as the source for ocean-island basalts either. Because of the way elements redistribute in the mantle during melting, the material that forms oceanic crust would evolve to give ocean- island basalts with Hf isotopic ratios that are relatively low at a given Nd ratio — that is, the plot would lie below the terrestrial array. Such arguments have been used as evidence against the recycling of ancient oceanic crust or sediment into the source of ocean-island volcanoes7.
Chauvel and co-authors propose a simple solution to this problem. If neither subducted oceanic crust nor marine sediment alone will generate compositions that plot on the terrestrial array as required, mixtures of the two perhaps would. To test this idea, they simulate the composition of subducted crust as a random mixture of oceanic basalt and sediment in different proportions and from different times in Earth's history. The result is a band of compositions that mimics the terrestrial array, implying that both the oceanic crust and the pile of sediments that accumulates on top of it are subducted into the mantle. There, they remain together, but distinct, for billions of years. When they melt along with the surrounding mantle, they give rise to ocean-island lavas with the Nd and Hf isotopic composition required to explain the terrestrial array.
Models simulating the physical aspects of this process, although providing some support for the hypothesis, also pose some difficulties. Such models predict relatively high temperatures at the surface of subducting slabs, which may cause sediments to rise up to shallower levels before the slab sinks deep into the mantle8. It is not clear, therefore, whether significant amounts of sediment can make their way into the deeper parts of the mantle. However, if they do manage to reach the deep mantle, the models predict that they will be able to mature there and be mixed and melted to form ocean-island magmas in the fashion suggested by Chauvel and co-authors (Fig. 1).
Figure 1: Oceanic crust (black line) collects sediment as it travels from the mid-oceanic ridge (A).
Upon subduction (B), the crust is recycled (C) and remixed into the mantle (D), and can be brought up passively at mid-oceanic ridges or through plumes (E) at ocean island hotspots. Chauvel and co-authors2 provide quantitative constraints on the relative amounts of sediment, oceanic crust and upper mantle that need to be mixed to explain the Hf–Nd terrestrial array. Image provided by J. P. Brandenburg and based on work in ref. 11.
Full size image (28 KB)The Nd–Hf modelling exercise conducted by Chauvel and co-authors is elegant in its simplicity, but several geochemical aspects remain to be tested. Marine sediments vary enormously in their elemental and isotopic composition from place to place9. Chauvel and co-authors chose only a single sedimentary composition. Whether this composition is indeed representative of the global average will have to await a survey of the Nd and Hf isotopic compositions of marine sediments subducting at oceanic trenches worldwide. In the modelling by Chauvel and co-authors, time is considered a prime determinant for generating isotopic heterogeneities in the mantle2. However, it could very well be that local compositional variations are more important. In addition, other studies have invoked chemical processes internal to the mantle and not external sources such as sediments or oceanic crust, to explain mantle heterogeneity10, 11.
Thus, the debate regarding the external versus internal source of mantle heterogeneity and the ultimate fate of subducting slabs continues. It is likely to be resolved by a better quantification of fluxes into the mantle at subduction zones, improved models that consider composition as well as convection, better resolution of seismic images of the mantle, and consideration of the full suite of geochemical tracers relevant to the problem of recycling within the deep Earth.