Earth grew by the accretion of meteoritic material. High-precision isotopic data reveal how the composition of this material changed over time, forcing revision of models of our planet's formation. See Letters p.521 & p.525
For more than half a century, scientists have estimated the bulk chemical composition of Earth by comparison with its potential cosmic building blocks, as sampled by meteorites. In a conceptual breakthrough, on page 521, Dauphas1 uses the unique isotopic content of different types of meteorite to identify those that best represent these building blocks. The author also evaluates whether the material added to Earth during its formation changed over time. On page 525, Fischer-Gödde and Kleine2 show that not even the most recently accreted 0.5% of such material consisted of the type of meteorite long thought to be a major contributor to our planet's composition. This realization challenges our understanding of how Earth obtained its inventory of volatile elements and water.
In the 1970s, Earth was shown to have a different oxygen-isotope composition from most meteorites3. The only meteorites that have a similar oxygen isotopic abundance are called enstatite chondrites, which are silicon-rich and highly reduced (most iron is in the form of metal or sulfide, rather than oxide). This similarity drove several models that based Earth's composition on enstatite chondrites4,5. However, the mismatch in the elemental composition between such meteorites and Earth's rocks led most researchers to continue using models based on more-oxidized and volatile-rich meteorites known as carbonaceous chondrites6,7.
Improvements in the ability to determine precise isotopic abundances led to the discovery that many elements can be used to distinguish between Earth and meteorites8. In 2011, a study of these isotopic differences suggested that Earth was made from a mixture of meteorite types9, not just the carbonaceous chondrites that had been the main component of most models. Dauphas takes this approach further by developing a methodology in which the isotopic disparity between different groups of meteorites and Earth can be used to track the composition of the materials that accreted to our planet throughout its formation.
The most important chemical differentiation event in Earth's history was the separation of its iron-metal core from its silicate mantle (Fig. 1). When the core formed, elements that are more soluble in metal than in silicate were selectively removed from the mantle. Some elements (such as iridium, platinum, palladium and ruthenium) are so soluble in metal that the mantle should have been effectively stripped of them during core formation. However, the observed abundances of these elements in the mantle are in the same relative proportion as those seen in primitive meteorites. Furthermore, they are depleted by a factor of only about 350 with respect to their abundance in meteorites10, compared with the million-fold depletion11 that would be expected were the mantle in chemical equilibrium with the core.
One explanation is that these elements were added back to the mantle by subsequent accretion of meteoritic material (with a mass of about 0.5% that of Earth) after core formation was complete12. Dauphas notes that, if this is so, the isotopic composition of ruthenium in the mantle tracks only the last 0.5% of the material from which our planet formed. By contrast, the mantle isotopic composition of elements that are completely insoluble in metal reflects the average composition of all the material from which Earth grew.
Using this approach, for the series of elements titanium, chromium, nickel and molybdenum (listed in order of their increasing preference for the core), Dauphas estimates that their isotopic composition in the mantle reflects the last 95%, 85%, 39% and 12% of material accreted by Earth, respectively. Then, using the isotopic differences in these elements between Earth and meteorites, the author finds that our planet formed from a mixture of meteorite types for about the first 60% of its growth and almost entirely from enstatite chondrites for the remainder. The high-precision ruthenium isotopic measurements presented by Fischer-Gödde and Kleine reinforce the conclusion that the last 0.5% of the accreted material was isotopically most like enstatite chondrites.
The disturbing aspect of this conclusion is that the chemical composition of enstatite chondrites is very different from that of rocks on Earth's surface. Consequently, if Earth is mostly made from enstatite chondrites, its deep interior must have a substantially different composition from its outer layers5. Although possible, this explanation is not easily reconciled with many lines of evidence. An alternative offered by Dauphas is that the enstatite chondrites might be leftovers of the processes that formed Earth, but whose compositions were modified by the planet-forming process. This is an intriguing suggestion, but one whose consequences will need much more investigation.
If the last 0.5% of material accreted by Earth had been composed of a particular type of volatile-rich carbonaceous chondrite, known as a CI chondrite, an amount of water equivalent in mass to Earth's oceans would have been added to the planet10. Fischer-Gödde and Kleine's measurements instead show that this late-accreted material consisted of relatively 'dry' enstatite chondrites. Water was therefore supplied to Earth throughout its growth, rather than being added late in its history through the accumulation of volatile-rich materials such as carbonaceous chondrites or comets.
The results presented in these papers lead to the troubling conclusion that the meteorites in our collection are not particularly good examples of Earth's building blocks. Although this makes estimation of the planet's bulk composition more difficult, new isotopic data and approaches to interpreting those data provide the next step towards a better understanding of how Earth formed.Footnote 1
Dauphas, N. Nature 541, 521–524 (2017).
Fischer-Gödde, M. & Kleine, T. Nature 541, 525–527 (2017).
Clayton, R. N., Onuma, N. & Mayeda, T. K. Earth Planet. Sci. Lett. 30, 10–18 (1976).
Herndon, J. M. Proc. R. Soc. A 368, 495–500 (1979).
Javoy, M. et al. Earth Planet. Sci. Lett. 293, 259–268 (2010).
McDonough, W. F. & Sun, S. Chem. Geol. 120, 223–253 (1995).
Palme, H. & O'Neill, H. St C. in The Mantle and the Core (ed. Carlson, R. W.) 1–39 (Elsevier, 2014).
Qin, L. & Carlson, R. W. Geochem. J. 50, 43–65 (2016).
Warren, P. H. Earth Planet. Sci. Lett. 311, 93–100 (2011).
Wänke, H., Driebus, G. & Jagoutz, E. in Archaean Geochemistry (eds Kröner, A., Hanson, G. N. & Goodwin, A. M.) 1–24 (Springer, 1984).
Borisov, A. & Palme, H. Geochim. Cosmochim. Acta 61, 4349–4357 (1997).
Walker, R. J. Chem. Erde 69, 101–125 (2009).