The Source of Water on the Earth and Moon

A protoplanetary disk rotating around a young star. (NASA)

Francis Albarède published a review paper in the October 29^th Nature on the source of water and other volatiles in the Earth-Moon system. I read it when it came out but the LCROSS results prompted me to revisit it. Here’s a portion of Albarède’s abstract:

Accretion left the terrestrial planets depleted in volatile components. Here I examine evidence for the hypothesis that the Moon and the Earth were essentially dry immediately after the formation of Moon – by a giant impact on the proto-Earth – and only much later gained volatiles through accretion of wet material delivered from beyond the asteroid belt. This view is supported by U–Pb and I–Xe chronologies, which show that water delivery peaked ~100 million years after the isolation of the Solar System. Introduction of water into the terrestrial mantle triggered plate tectonics, which may have been crucial for the emergence of life…

Here are some possible scenarios:

1. Wet accretion: Water was incorporated during accretion of the Earth. A giant impact ejected the material that formed the Moon, but this process caused escape of much of the volatile component (water, etc.) from the Moon.
2. Dry accretion: The Earth-Moon system did not include a significant amount of water until after the Moon’s formation. Impactors, i.e., asteroids and comets, delivered water and other volatiles.
3. Wet accretion + devolatilization: The Earth-Moon system was accreted wet but lost water to degassing during an early period of global magmatism. Impactors delivered water and other volatiles.

Albarède argues for dry accretion by reviewing evidence that the inner planets, including Earth, are too close to the center of the solar system for water to have condensed during their formation. Water was added later.

Our solar system formed from a solar nebula, which was hottest towards the center (now the sun). As the nebula cooled, elements began to condense from their gaseous to their solid states. Refractory elements, including platinum group elements, Si, Al, Fe, Mg, and Ca, condensed first; next were alkali elements like Na and K, and eventually, as temperature continued to fall, the volatiles like C and water solidified. Our sun was much more violent during its early stages (the T Tauri Phase) and it produced strong solar winds which swept much of the gaseous nebula outwards. Less refractory elements that hadn’t condensed at high temperatures were pushed away from the sun. This model fits observations nicely: the rocky inner planets are composed of refractory elements and alkalis, while the gas giants and icy planets formed farther out. The nebular model also predicts that there was a “snow line” at 2 to 2.5 AU (AU = Astronomical Units, or the average distance from the Earth to the sun); any water closer to the sun than this wouldn’t have condensed before being cleared by early solar radiation. The prediction of a snow line necessitates the dry accretion model: there shouldn’t be water on the inner planets.

A proposed timeline of Earth's accretion and some of the isotope systems that constrain it from Albarède’s Figure 5. During its T-Tauri phase, the sun cleared the inner solar system of volatiles. Runaway growth marks the peak accretion of Mars-sized bodies that form the cores of giant planets and the proto-inner planets. The Hf-W system constrains core formation on Earth. The giant impact forms the Earth's Moon. The late veneer deposits water and siderophile elements.

But of course there’s water on Earth, and quite a lot: ~1.34*10⁹ km³ in the oceans, and perhaps 1-3 times that much in the mantle. If Earth accreted dry, then this water may have been added as part of the late veneer, when formation of the giant planets like Jupiter upset the orbits of planetesimals and asteroids and sent many of them hurtling into the inner solar system. This extraterrestrial material would be incorporated into the Earth’s mantle by foundering of early surface rocks or by the onset of subduction. Icy comets seem like a likely source, but the isotopic composition of Earth’s water is a better match for carbonaceous chondrites, which are concentrated in the outer part of the Main Belt, just past the snow line. They contain up to 10% water by weight.

Albarède’s paper draws on the recent literature, including much work on the H, O, Pb, W, and I-Xe isotopic systems, to make his case. He synthesizes what seem to be (based on my literature searches) the newest widely accepted findings across a number of fields. It is not, however, a consensus view. There are arguments against the standard late veneer model based on W isotope compositions and the Earth’s abundances of siderophile elements (elements that include the platinum group, Mn, Co, and Au which behave like Fe and should follow Fe during geochemical processes). We should be able to reconcile the abundances of elements deposited on Earth as part of the late veneer – not just the water, but the siderophile elements, too – and we’re not quite there yet. See, for example, Alex Halliday’s 2008 paper that discusses some of these problems.

Those issues need to be resolved, and research on siderophile element behavior during core formation is addressing them. For now, the nebular model as I understand it all but requires the late veneer to explain Earth’s water, but there’s plenty I don’t understand (yet). We’ll see.

*****

My post on the LCROSS results.

A post on Chandrayaan’s lunar water discovery.

Albarède, F. (2009). Volatile accretion history of the terrestrial planets and dynamic implications Nature, 461 (7268), 1227-1233 DOI: 10.1038/nature08477