≡ Menu

Fine-Tuning Mechanisms for Water Delivery

We’ve long been interested in how the Earth got its oceans, with possible purveyors being comets and asteroids. The idea trades on the numerous impacts that occurred particularly during the Late Heavy Bombardment some 4.1 to 3.8 billion years ago. Tuning up our understanding of water delivery is important not only for our view of our planet’s development but for its implications in exoplanet systems with a variety of different initial conditions.

Image: This view of Earth’s horizon was taken by an Expedition 7 crewmember onboard the International Space Station, using a wide-angle lens while the Station was over the Pacific Ocean. Credit: NASA.

But the picture becomes more complex when we compare regular hydrogen atoms (one proton, one electron) with ‘heavy hydrogen,’ or deuterium atoms. The latter have a neutron in addition to a proton in the nucleus. A recent paper in the Journal of Geophysical Research digs into isotope ratios, the ratio of deuterium to ordinary hydrogen atoms, commonly referred to as the D/H ratio. One reason asteroids are favored by some scientists as the likely source of the bulk of Earth’s water is that asteroidal water has a D/H in the neighborhood of 140 parts per million. Contrast that with cometary water, which runs from 150 ppm to as high as 300 ppm.

When we examine Earth’s oceans, we find a D/H ratio close to that found in asteroids. The new study, from Jun Wu and colleagues at the School of Molecular Sciences and School of Earth and Space Exploration at Arizona State University, takes aim at the asteroid explanation, not by way of discounting it but rather of finding other sources making a contribution to Earth’s water.

“It’s a bit of a blind spot in the community,” said ASU’s Steven Desch, a co-author of the new study. “When people measure the [deuterium-to-hydrogen] ratio in ocean water and they see that it is pretty close to what we see in asteroids, it was always easy to believe it all came from asteroids.”

We are learning, however, that too uncritical an acceptance of D/H ratios may oversimplify the issue. For the hydrogen in Earth’s oceans is not necessarily representative of hydrogen deeper inside the planet, where D/H ratios close to the boundary between the core and mantle show considerably less deuterium. This may indicate a non-asteroidal source for at least some of the hydrogen.

Another telling point is that helium and neon, showing isotopic signatures inherited from the original solar nebula, have also been found in Earth’s mantle. Contrasting hydrogen at the core-mantle boundary with what we see in Earth’s oceans and factoring in these noble gases may change our thinking. The Wu study considers the formation of the planets in the earliest days of the Solar System, when small, often colliding planetary embryos up to the size of Mars went through gradual planetary accretion.

The new model works like this: As larger embryos formed largely from water-laden asteroids, they began to develop into planets. On Earth, decaying radioactive elements melted iron within the emerging world, pulling in asteroidal hydrogen and sinking to form a core. Collisions among planetesimals would meanwhile have created enough energy to melt the surfaces of the larger embryos like the Earth into magma oceans.

Hydrogen and noble gases from the solar nebula would be drawn in to create an early atmosphere. The nebular hydrogen, lighter than asteroidal hydrogen, would have dissolved into the molten iron of the magma ocean, eventually being drawn into the mantle, along with hydrogen from other sources.

What the authors argue is that this process created a slight enrichment of hydrogen in the molten iron and left a higher ratio of deuterium behind in the magma (the process is called isotopic fractionation). Hydrogen is attracted to iron, while the heavier isotope, deuterium, less attracted to iron, would have remained in the magma which would eventually become Earth’s mantle. We would end up with lower D/H ratios in the core than in the mantle and oceans. The authors argue that while most of Earth’s water is asteroidal, some of it did in fact come from the solar nebula.

The process is complex, and also takes in impacts from smaller embryos and other objects that continued to add water and mass until Earth reached its final size. The authors provide this synopsis as their caption for Figure 1 (above), which I’ll reproduce verbatim but break into sections for reasons of readability:

  • (a) Earth accreted from embryos with chondritic [asteroidal] levels of water concentrations and D/H ratios.
  • (b) These embryos differentiated and stored relatively light hydrogen in their cores, raising the D/H of hydrogen in their mantles.
  • (c) The largest embryo accreted a proto-atmosphere and sustained a magma ocean into which nebular hydrogen diffused.
  • (d) The largest embryo’s magma ocean crystallized and overturned, mixing light hydrogen into the mantle, but incompletely.
  • (e) As smaller embryos were accreted, their mantles joined the proto‐Earth’s mantle, and their cores merged with the proto‐Earth’s core.

The result is:

  • (f) Earth’s mantle today contains approximately three oceans of water in its mantle and surface, with average D/H ≈ 150 × 10−6, and ~4.8 oceans’ worth of hydrogen in its core, with D/H ≈ 130 × 10−6. Mantle plumes can sample low‐D/H material from the core‐mantle boundary.

Wu says this: “For every 100 molecules of Earth’s water, there are one or two coming from [the] solar nebula.” But we now see the potential for including gas left over from the formation of our Solar System in the question of water delivery and accumulation. In other stellar systems where water-bearing asteroids may not be as abundant, this implies that water could still be in place from the system’s own stellar nebula. Wu adds: “This model suggests that the inevitable formation of water would likely occur on any sufficiently large rocky exoplanets in extrasolar systems. I think this is very exciting.”

The paper concludes:

Our comprehensive model for the origin of Earth’s water considers, for the first time, the effects of isotopic fractionation as hydrogen dissolved into metal and was sequestered into the core. Based on the behaviors of proxies, we consider it likely that the D/H ratio of the core is ~10% lighter than the mantle. We hypothesize that Earth accreted a few to tens of oceans of water from chondrites, mostly carbonaceous chondrites. Drawing on the latest theories of planet formation, which argue for rapid (<2 Myr) formation of planetary embryos, we favor ingassing of a few tenths of an ocean of solar nebula hydrogen into the magma oceans of the embryos that formed Earth.

The paper is Wu et al., Journal of Geophysical Research: Planets 09 October 2018 (full text).


Comments on this entry are closed.

  • Bruce D. Mayfield November 8, 2018, 16:09

    The term “ocean” is used here as a unit of measure, as in these statements:

    “We hypothesize that Earth accreted a few to tens of oceans of water from chondrites.” and
    “… we favor ingassing of a few tenths of an ocean of solar nebula hydrogen into the magma oceans of the embryos that formed Earth.”

    Just how much water is in 1 ocean? As much as is in the Earth’s present global ocean? 1 Pacific? an Artics worth?

    • Paul Gilster November 8, 2018, 17:37

      I was assuming 1 ocean = all the water in Earth’s oceans today. I’ll see if I can find anything in the paper that clarifies this.

      • Bruce D. Mayfield November 10, 2018, 1:59

        Section (f) of the article’s illustration shows that your assumption was correct Paul.

    • Nicky November 8, 2018, 21:19

      1.4*10^21 kg water

    • Robin Datta November 9, 2018, 1:00

      1,335,000,000 Volume (km³)
      % Ocean Volume 100.0
      1.37 billion km³
      1.332 x 10⁹ km³

      • Bruce D. Mayfield November 9, 2018, 13:18

        Thanks for those replies. So they’re taking about multiples of our total global ocean. Then, if there was so much water in play then there must have also have been a great deal of water loss, or wouldn’t this planet be a water world with no dry land?

  • Nicky November 8, 2018, 21:29

    This mechanism is indeed very plausible. It would work even better on super-Earths. Pebble accretion model has shown super-Earths grow into Earth-mass before the dissipation of nebula.
    Primordial hydrogen ingassing is way more efficient on an Earth-mass planetary embryo (in the case of super-Earths) than on a Mars-mass embryo (in the case of Earth and Venus) due to stronger gravity and longer magma ocean. This implies that most super-Earths might contain as much as tens to hundreds of oceans of water.

    • Bruce D. Mayfield November 9, 2018, 13:29

      If that’s correct then most super-Earths would be water worlds with no land. With their increased surface gravity I would guess that erosion of any land protruding above the waves would be even more effective than here on Earth.

  • Michael C. Fidler November 8, 2018, 23:06



    How Snowballs From Space May Have Sowed Seeds of Life


    Sea plankton ‘found living outside International Space Station.


    So I guess all those oceans brought in some wild plankton slime. The cosmonauts need to clean their boots and gloves before the come in from the spacewalks…

    • Robin Datta November 9, 2018, 16:29

      For 1 inch of global water depth equivalent incoming in each in 10,000 years on a 4.5 billion year old earth gives us an ocean of average depth of 7.1 miles: has some water departed the scene of the deluge?

      • Bruce D. Mayfield November 9, 2018, 23:17

        That must be the case Robin. Imagine way back when a Venus-sized proto Earth was hit by a Mars-sized body, leading to our Earth-Moon life friendly system. There must have been several “oceans” worth of water boiled off and swept away by the solar wind after that blast.

        This might be yet another reason having a large moon is an important factor in improving the odds for having a truely Earthlike planet complete with big seas and land masses. If no giant impact Earth might never had had any land at all.

        • Abelard Lindsey November 10, 2018, 11:41

          Yes, and since that giant impact is assumed to be relatively rare, the vast majority of even Earth-sized planets in their HZ’s are going to be water worlds. Super Earths will have really deep oceans (100km or more) with the larger ones being mini-Neptunes. We like to think of Earth as the “water” planet. But what is significant about the Earth is actually how little water it has. Of course, all of this has implications with regards to the availability of habitable planets in the galaxy (as in not many).

          • Nicky November 10, 2018, 21:14

            It is unlikely that a moon-forming impact can destroy all the water on Earth. Yes that impact can vaporize all the volatile into atmosphere, but the timescale is too short (only a few Myr) to have any significant escape, which requires over tens to hundreds of million of years.
            In contract, the impact probably delivered more volatile.

            • Bruce D. Mayfield November 11, 2018, 15:47

              No one was suggesting that “all” water would be lost in such massive impacts, but I take your point that the Earth would have ended up with more total water. What matters however is how to come up with the right amount of water at the surface.

              Of course, that’s from the point of view of a land creature.

  • Michael November 9, 2018, 8:17

    Just wondering if the early cloak of hydrogen gas our planet had helped with formation of water. If rocks impact our early hydrogen atmosphere there would have been a reaction with the hydrogen gas and the oxides in the rock at high temperatures and pressures of entry to form water.

  • ljk November 9, 2018, 14:42

    How many comets would it take to cover an Earth with water?

    And how often would they have to keep replenishing such a planet until it could sustain its own oceans?

  • AlexT November 11, 2018, 3:41

    It is not clear why D/H relation in asteroids and in Solar Nebula should have some difference, I thought that in Solar System asteroids and planets are composed from the same Solar Nebula’s material…

  • Geoffrey Hillend November 14, 2018, 19:01

    I thought hydrogen was an atmophile so most of it went into the atmosphere. There are hydroxide anions in the silicate materials in the mantel but I don’t think a H D/H ratio in the mantel has any application to how much hydrogen came from the solar nebula. The D/H ratio only applies to our atmosphere where the heaver deuterium remains behind but why would the light hydrogen stay in the core. It seems to me that it should be the heaver deuterium. Both hydrogen and deuterium have to escape from the mantle through volcanoes anyway. Hydrogen is not a siderophile, and not a metal so I don’t see why it would remain in the iron core. Hydrogen would escape as a gas from the molten surface into the atmosphere and space.

  • Geoffrey Hillend November 14, 2018, 19:19

    The idea that deuterium is attracted to iron more than hydrogen is not supported by the laws of chemistry and quantum mechanics. All attraction between different elements in chemistry is the result of the ionic and covalent electron bonds between different elements in molecules, the negatively charged electrons in the electron shells of one element attracting the opposite positive charges of the protons of another element in their nucleus to make molecules. Hydrogen and deuterium are isotopes, so they both have only one electron, so they should have the same attraction to Iron. The only difference between the isotopes of hydrogen and deuterium is that deuterium has a neutron, but hydrogen does not. The neutron is in the nucleus of deuterium, and the neutron is neutrally charged or without charge, so it does not affect the attraction to iron any more than hydrogen. https://en.wikipedia.org/wiki/Goldschmidt_classification