Skip to main content

Earth's moon had a magma ocean for 200 million years

Earth's moon was once so hot that its mantle was completed molten into magma.
Earth's moon was once so hot that its mantle was completed molten into magma.
(Image: © NASA/Goddard Space Flight Center/Conceptual Image Lab)

Earth's newborn moon may have possessed an ocean of magma for 200 million years, much longer than scientists thought, a new study finds.

These new findings may shed light on the formation of Earth and the rest of the solar system, researchers said.

Scientists think Earth arose about 4.5 billion years ago, with the moon born a short time later. The leading explanation for the moon's origin is that it resulted from the collision of two protoplanets, or embryonic worlds. One of those was the nascent Earth, and the other was a Mars-size object called Theia. 

Related: How the moon formed: 5 wild lunar theories

"An important outcome of this scenario is that the early moon, which accreted from the debris of this giant impact, was very hot — hot enough for its rocky mantle to be largely molten and form what we call a magma ocean," study lead author Maxime Maurice, a planetary scientist at the German Aerospace Center in Berlin, told Space.com.

Previous research found evidence for such a magma ocean on the early moon in the composition of the lunar crust, which is best explained as the floating residue of molten rock that crystallized under lunar conditions, Maurice explained.

"While the idea of a primordial magma ocean on the moon is largely accepted, the time it took to solidify was not very clear," Maurice said. "Previous models suggested it was fairly rapid — some tens of millions of years."

In the new study, the researchers developed a model for the solidification of the ancient lunar magma ocean that "considers the influence of many processes occurring back then, some of them disregarded until now," Maurice said. One such process was mantle convection — the way eddies can form and churn in solid mantle, "which, on the Earth, causes volcanism," he said.

Related: Earth's mantle and crust are in a fiery battle to the death … of supercontinents

All in all, they found the lunar magma ocean may have solidified over the course of 150 million to 200 million years, lasting about 10 times longer than previously thought. They also estimated the moon formed between 4.4 billion and 4.45 billion years ago, which is 50 million to 100 million years later than previously considered, Maurice said. This age closely matches that of Earth's core.

"There is a strong link between the moon-forming event and the formation of the Earth's core, because the giant impact likely resulted in large scale melting in the Earth's mantle, which significantly helped the formation of the core," Maurice said.

If the moon is as old as the new model estimates, that suggests the era of giant collisions such as the one that gave birth to the moon, which marked the last stage of planetary formation, was still active about 150 million years after the birth of the solar system, Maurice said. 

The scientists detailed their findings online July 10 in the journal Science Advances.

Correction: A previous version of this article incorrectly stated that mantle convection occurs in molten rock. This process occurs in solid mantle, not molten rock.  

Follow Charles Q. Choi on Twitter @cqchoi. Follow us on Twitter @Spacedotcom and on Facebook.

Join our Space Forums to keep talking space on the latest missions, night sky and more! And if you have a news tip, correction or comment, let us know at: community@space.com.

  • rod
    The article says, "All in all, they found the lunar magma ocean may have solidified over the course of 150 million to 200 million years, lasting about 10 times longer than previously thought. They also estimated the moon formed between 4.4 billion and 4.45 billion years ago, which is 50 million to 100 million years later than previously considered, Maurice said. This age closely matches that of Earth's core...If the moon is as old as the new model estimates, that suggests the era of giant collisions such as the one that gave birth to the moon, which marked the last stage of planetary formation, was still active about 150 million years after the birth of the solar system, Maurice said. "It was very exciting to see that our findings had these kinds of broad implications."

    FYI. Models developed for accretion and collisions like the postulated Theia with the proto-earth vary widely for the period 4.5E+9 to 3.7E+9 years ago. 'Impact bombardment chronology of the terrestrial planets from 4.5 Ga to 3.5 Ga', https://ui.adsabs.harvard.edu/abs/2020Icar..33813514B/abstract, "...Our calculated lunar, martian and mercurian chronologies use the impacts recorded onto the planets from dynamical simulations rather than relying on the decline of the population as a whole. We present fits to the impact chronologies valid from 4500 Ma to ca. 3700 Ma by which time the low number of planetesimals remaining in the dynamical simulations causes the impact rate to drop artificially."

    Pinning down the time of Theia impact with proto-earth and reconciling with different accretion and collision models is not easy. Radiometric dating methods must be reconciled too for the magma ocean and cooling period. 'The timing of lunar solidification and mantle overturn recorded in ferroan anorthosite 62237', https://ui.adsabs.harvard.edu/abs/2020E%26PSL.53816219S/abstract, "Ferroan anorthosite suite (FAS) rocks are widely interpreted to represent primordial lunar crust. Despite their importance in pinpointing the timing of lunar crust formation, robust chronological investigations for this rock type are scarce. Here, we report the Ar-Ar, Rb-Sr, and Sm-Nd isotopic systematics for the FAS troctolitic anorthosite 62237. The Ar-Ar isotopic system has been reset by a thermal event at 3710 ± 48 Ma, and the Rb-Sr isotopic systematics has been disturbed such that a Rb-Sr isochron age cannot be determined."

    Important issues when reading about the giant impact origin for the Moon.
    Reply
  • rod
    FYI, here is a bit more demonstrating some challenges in dating Theia impact with the proto-earth.

    'The Hf-W Isotopic System and the Origin of the Earth and Moon', https://ui.adsabs.harvard.edu/abs/2005AREPS..33..531J/abstract, "The Earth has a radiogenic W-isotopic composition compared to chondrites, demonstrating that it formed while 182Hf (half-life 9 Myr) was extant in Earth and decaying to 182W. This implies that Earth underwent early and rapid accretion and core formation, with most of the accumulation occurring in ~10 Myr, and concluding approximately ~30 Myr after the origin of the Solar System. The Hf-W data for lunar samples can be reconciled with a major Moon-forming impact that terminated the terrestrial accretion process ~30 Myr after the origin of the Solar System. The suggestion that the proto-Earth to impactor mass ratio was 7:3 and occurred during accretion is inconsistent with the W isotope data. The W isotope data is satisfactorily modeled with a Mars-sized impactor on proto-Earth (proto-Earth to impactor ratio of 9:1) to form the Moon at ~30 Myr."

    So we have the new report in the space.com article indicating the giant impact could take place nearly 150 million years after the solar system formed, and earlier reports like the 2005 source indicating 30 million years after the solar system formed. The giant impact model with Theia and proto-earth can be enjoyable to investigate and study :)---Rod
    Reply
  • dfjchem721
    The oldest known terrestrial rocks on earth date to 4.4 bya*, indicating a cool surface ca. 150 mys after its formation. These rocks are dated from zircon crystals which trapped U-238 (half-life 4.5 billion years), and those crystals were embedded in verified terrestrial rocks. These data are highly reliable, and represent an extremely accurate date for rocks on earth which must have formed after the molten surface solidified.

    Their presence in rock on earth dating to 4.4 bya strongly suggest that the earth had a largely cool and solid crust during the time the moon supposedly had a molten ocean. It is diificult to reconcile the hard data from the ultra-precise U-238 decay in zircon crystals with this new report which "estimated the moon formed between 4.4 billion and 4.45 billion years ago....."


    * http://www.geology.wisc.edu/~valley/zircons/Wilde2001Nature.pdf
    Reply
  • rod
    Post #4 shows it is difficult to develop the evolutionary prehistory (i.e. sequence of events in the early solar system) from 4.56E+9 years old radiometric dated meteorites (since Clair Patterson days in mid-1950s), Earth rocks dated 4.4E+9 years old, and when Theia hit the proto-earth to create the Moon and the Moon orbiting Earth with the postulated magma ocean.
    Reply
  • Torbjorn Larsson
    The article seems not to have been published.

    But for what it is worth, Maurice has published on late dates for Moon impacts for years https://meetingorganizer.copernicus.org/EPSC2018/EPSC2018-688.pdf ]. It seems the new and rapidly accepted dating of the Moon forming impact (of 50 Myrs after formation) has left little impression.

    Given that, what would be a good explanation for observations? Here is a model that explains much more, such as expected freezing time for mares relative to observed, the near and far side dichotomy, and the deep South Polar Aitken impact material: https://phys.org/news/2020-06-scientists-explanation-strange-asymmetry-moon.html .

    The moon, being smaller, likely cooled down faster and geologically froze. The apparent early dynamism of the moon challenges this idea. New data suggests this is because radioactive elements were distributed uniquely after the catastrophic moon‐forming collision.

    The heat from the radioactive decay of these elements can melt the rocks they are contained in, which may partly explain their co‐localisation.

    This study shows that, in addition to enhanced heating, the inclusion of a KREEP component to rocks also lowers their melting temperature, compounding the expected volcanic activity from simply radiogenic decay models. Because most of these lava flows were emplaced early in lunar history, this study also adds constraints about the timing of the moon's evolution and the order in which various processes occurred on the moon.

    Reply
  • Torbjorn Larsson
    dfjchem721 said:

    The oldest known terrestrial rocks on earth date to 4.4 bya*, indicating a cool surface ca. 150 mys after its formation. These rocks are dated from zircon crystals which trapped U-238 (half-life 4.5 billion years), and those crystals were embedded in verified terrestrial rocks. These data are highly reliable, and represent an extremely accurate date for rocks on earth which must have formed after the molten surface solidified.

    Indeed - I think the latest works are indicating a solidification within 10 Myrs - so Tellus likely had a crust pre-Theia and Earth had one early post-Theia as well.

    Some models of Moon seem to imply the smaller body had a harder time solidify its crust, e.g. Maurice et al line of work. But for what it is worth the new KREEP paper may solve the late mare as well as the Moon surface dichotomy problems.

    rod said:
    Pinning down the time of Theia impact with proto-earth and reconciling with different accretion and collision models is not easy. Radiometric dating methods must be reconciled too for the magma ocean and cooling period.

    rod said:
    'The Hf-W Isotopic System and the Origin of the Earth and Moon', https://ui.adsabs.harvard.edu/abs/2005AREPS..33..531J/abstract, ... The W isotope data is satisfactorily modeled with a Mars-sized impactor on proto-Earth (proto-Earth to impactor ratio of 9:1) to form the Moon at ~30 Myr."

    rod said:

    Post #4 shows it is difficult to develop the evolutionary prehistory (i.e. sequence of events in the early solar system) from 4.56E+9 years old radiometric dated meteorites (since Clair Patterson days in mid-1950s), Earth rocks dated 4.4E+9 years old, and when Theia hit the proto-earth to create the Moon and the Moon orbiting Earth with the postulated magma ocean.

    It has been and remain problematic, but the new 50 million year impact https://www.nature.com/articles/s41561-019-0398-3 ] dating seems quickly accepted in literature https://www.nature.com/articles/s41561-019-0398-3/metrics ]. It is dated directly on the Moon rocks with modern methods, and has high precision and robustness.

    Here we present high-precision trace element composition data from inductively coupled plasma mass spectrometry for a wide range of lunar samples. Our measurements show that the Hf/W ratio of the silicate Moon is higher than that of the bulk silicate Earth. By combining these data with experimentally derived partition coefficients, we found that the 182W excess in lunar samples can be explained by the decay of the now extinct 182Hf to 182W. 182Hf was only extant for the first 60 Myr after the Solar System formation. We conclude that the Moon formed early, approximately 50 Myr after the Solar System, and that the excess 182W of the silicate Moon is unrelated to late accretion.


    It is also coincident with the new models on fast system formation https://www.sciencemag.org/news/2020/06/missing-mass-planet-formation-found-young-disks-gas-and-dust ] and no "late bombardment" https://www.sciencemag.org/news/2020/01/cataclysmic-bashing-giant-planets-occurred-early-our-solar-systems-history ], so in at least this respect the new data seem to fit nicely together. The problem is in the details (such as the KREEP paper models - we'll see how it fares).

    To find out how much material is available for planet formation, researchers have used the Atacama Large Millimeter/submillimeter Array (ALMA) in Chile to weigh the disks around young stars between 1 million and 3 million years old. Past studies found that some lacked the mass to form even a single Jupiter-size world. The results suggested astronomers were either overlooking some hidden reservoir of material or they were looking too late in the planet-forging process, after growing protoplanets had already vacuumed up much of the material.

    The answer, says Łukasz Tychoniec, a graduate student at Leiden Observatory and lead author of the new paper, is that “we need to look earlier instead of for missing mass.”
    Reply
  • rod
    Interesting information provided in post #7. I note this "It has been and remain problematic, but the new 50 million year impact https://www.nature.com/articles/s41561-019-0398-3 ] dating seems quickly accepted in literature https://www.nature.com/articles/s41561-019-0398-3/metrics ]. It is dated directly on the Moon rocks with modern methods, and has high precision and robustness."

    If the postulated 50 million years collision date for Theia with the proto-earth is based upon the date assigned relative to the date assigned for the origin of the solar system, that is 4.56E+9 years ago (Clair Patterson work), thus the giant impact event between Theia and the proto-earth took place 4.51E+9 years ago. So this must be the accepted age for the giant impact model event with Theia according to post #7, it seems and when the Moon formed. However, note the article in this thread and what was clearly reported here by space.com.

    "All in all, they found the lunar magma ocean may have solidified over the course of 150 million to 200 million years, lasting about 10 times longer than previously thought. They also estimated the moon formed between 4.4 billion and 4.45 billion years ago, which is 50 million to 100 million years later than previously considered, Maurice said. This age closely matches that of Earth's core."

    There is a real difference between the giant impact model Theia hit date of 4.51E+9 years ago as post #7 suggest vs. 4.4E+9 to 4.45E+9 years ago for when the Moon formed according to the new report with a magma ocean potentially lasting 200 million years. We also have zircons like the Jack Hills zircon said to be 4.375E+9 years old.

    The giant impact model pre-history presented for when Theia hit the proto-earth does not have supporting radiometric ages published that show rocks dated 4.51E+9 years ago taken from the Moon or Earth.
    Reply
  • rod
    FYI. Here is some more reporting on this model for dating the when the Earth and Moon formed and how long the lunar magma ocean lasted.

    https://phys.org/news/2020-07-younger-age-earth-moon.html
    My note. The reference paper has many details on how this new lunar age was calculated. ‘A long-lived magma ocean on a young Moon’, https://advances.sciencemag.org/content/6/28/eaba8949, “Abstract. A giant impact onto Earth led to the formation of the Moon, resulted in a lunar magma ocean (LMO), and initiated the last event of core segregation on Earth. However, the timing and temporal link of these events remain uncertain. Here, we demonstrate that the low thermal conductivity of the lunar crust combined with heat extraction by partial melting of deep cumulates undergoing convection results in an LMO solidification time scale of 150 to 200 million years. Combining this result with a crystallization model of the LMO and with the ages and isotopic compositions of lunar samples indicates that the Moon formed 4.425 ± 0.025 billion years ago. This age is in remarkable agreement with the U-Pb age of Earth, demonstrating that the U-Pb age dates the final segregation of Earth’s core.”

    My note. The advances.sciencemag.org report does show it is difficult to reconcile various radiometric ages obtained from lunar samples with this new magma ocean model and age for the origin of the Moon and when Theia hit the proto-earth in the giant impact model. This is refreshing to read.

    “Comparison to chronology of lunar samples. The LMO solidification time scale of up to ~200 Ma inferred from our model seems inconsistent with the chronology of LMO products. Crystallization ages for ferroan anorthosites (FANs), representing the LMO’s flotation crust, range from ~100 to ~200 Ma after the beginning of the solar system…The Lu-Hf isotopic evolution of urKREEP has also been investigated using lunar zircons separated from KREEP-rich highland breccias (23). However, the zircon and bulk rock data cannot simultaneously be fitted in a single model, because they do not plot along a common isotopic evolution line for urKREEP. Also, the zircon data can only be fitted to the LMO’s Hf isotopic evolution for an unrealistically old onset time of LMO crystallization of 4.567 Ga (i.e., the age of the solar system). As such, and unlike for the bulk rock data, the zircon data are difficult to reconcile with a realistic model for the LMO’s thermal and isotopic evolution. One problem with the lunar zircon data may be that their Hf isotopic compositions require large downward corrections of ε176Hf for the effects of cosmic ray exposure (23). These corrections are inherently uncertain, and so, the lunar zircons may not accurately record the isotopic evolution of the LMO (see Materials and Methods). Age of the Moon. Our model not only provides the formation time of urKREEP but also predicts the time at which the LMO started to crystallize, which closely approximates the Moon’s formation time (that we treat as a free parameter). In the model above, the age of the Moon is 4.44 to 4.36 Ga (Fig. 4C, red histogram, and table S5). This age can be further refined by also including the well-dated lunar meteorite Kalahari 009, which has a precise age of 4.369 ± 0.007 Ga (24) and an elevated initial ε176Hf, indicating that it derived from a high-Lu/Hf source region (25).”

    My note. Zircons are used to date the Earth’s oldest rocks now, Jack Hills at 4.375 billion years old so this new model for the magma ocean of the Moon and Theia time of impact and origin of the Moon dated to 4.44 billion years ago, has difficulty with lunar zircons that suggest the Moon formed near the same time as the solar system, 4.567 billion years ago. This is consistent with other lunar samples from Apollo dated too, some >= 4.5 billion years (see the NASA ADS Abstract service).

    Perhaps the final chapter on Theia impact with the proto-earth to create our Moon, remains to be written :)
    Reply