New studies offer contrasting scenarios for making the Moon. One argues for a one big splat early in solar-system history; a second envisions a score of lesser blows that built up the Moon over time; and a third suggests water was involved.

Given the trove of lunar samples in hand and the power of modern laboratory analyses, you'd think that by now geochemists should have completely nailed exactly how the Moon formed. But not so — in fact, there's still lots of debate on how Earth formed.

Formation of the Moon
Artwork of a Mars-sized object colliding into the Earth early in solar system history. Many planetary scientists believe that an impact such as this threw off the debris which eventually formed the Moon.
Lynette Cook / Getty Images

Here's the basic problem: about 30 years ago, dynamicists showed that a body roughly the mass of Mars could have struck Earth a glancing blow and ejected enough debris into orbit to collect into a Moon-size object. In virtually all of those simulations, most of what ends up in the Moon came from the impactor rather than from Earth.

But the Apollo (and Luna) lunar samples, not to mention lunar meteorites, show that the Moon and Earth have very similar compositions. Apart from their lack of iron and extreme lack of water, Moon rocks match Earth's isotopic ratios for the geochemically diagnostic elements titanium, calcium, silicon, and (especially) oxygen and tungsten. This really pins the dynamicists in a corner — only in rare cases, 1% or 2% of the time, do their simulations yield a Moon with an Earthlike composition. There's also a problem of fine-tuning the impact to yield the angular momentum of the current Earth-Moon system.

I've written about possible solutions to these conundrums (or is it "conundra"?) here and here, but no one idea checks all the boxes. One can imagine that the giant impactor and proto-Earth had nearly identical compositions — but statistically and intuitively that seems unlikely.

In January 9th's Nature Geoscience, Israeli researchers Raluca Rufu, Oded Aharonson, and Hagai Perets argue that the notion of a single, giant impact is wrong. Instead, they propose that Earth endured dozens of lesser (but still potent) impacts with object ranging from 1% to 10% of its mass, each of which ejected debris into an orbiting disk. The rings quickly coagulated into moonlets, and tidal interactions with the young, mostly molten Earth then drove each of them outward. Over time they accumulated into the Moon.

Moon formation from moonlets
According to simulations by three Israeli researchers, the Moon might have assembled over time from the debris of 20 or more individual impacts with Earth.
Nature Geoscience / R. Rufu et al.

This approach yields a lunar composition that's an amalgam of many compositions, which eases the unyielding isotopic constraints. The most Earth-like contributions came from nearly head-on collisions that drilled deeply into our planet's mantle. A couple of glancing blows late in the process could have tweaked the system's angular momentum to match what exists now.

As Gareth Collins (Imperial College, London) notes in an accompanying News & Views perspective, "Lower-energy moonlet-forming impacts would leave parts of Earth unscathed. Distinct, terrestrial geochemical reservoirs may therefore have survived Moon formation." And, indeed, researchers have identified portions of Earth's mantle that are compositional mismatches to the rest of our planet.

Making the Moon: Slow or Fast?

The piecemeal assembly envisioned by the Israeli team would have taken a long time, perhaps even 100 million years — and that opens up another aspect of the lunar-formation debate. Some planetary scientists have indeed argued, mostly on geochemical grounds, that the Moon might have formed 150 to 200 million years after the beginning of the solar system. Others claim it showed up much sooner, within a few tens of millions of years.

Barboni and zircon grain
Researcher Mélanie Barboni holds a lunar sample prior to crushing it to extract zircon grains like the one in the inset.
Mélanie Barboni / UCLA

Another new analysis, published January 11th in Science Advances, maintains that the Moon came together in a hurry and had mostly solidified by 4.51 billion years ago, or 60 million years after the solar system's birth. The evidence, say Mélanie Barboni (University of California, Los Angeles) and six colleagues, is found in eight tiny grains of the mineral zircon (ZrSiO4), collected by Apollo 14's astronauts, in which they found traces of uranium, lead, and hafnium used for isotopic age-dating.

Several years ago a different research group had analyzed these same grains, and it also came up with an early formation age. But that result had wide uncertainties, owing to the techniques used. Barboni's team redid the age-dating, carefully measuring the lead isotopes that resulted from the radioactive decay of uranium-235 and -238 and also assaying the decay of lutetium to hafnium. Finally, the researchers also corrected for the lunar samples' exposure to cosmic rays, which can bias the isotopic ratios. They feel the resulting age of 4.51 billion years has an uncertainty of no more than 10 million years — and that the Moon might actually be older.

More pointedly, the Apollo 14 zircon grains presumably crystallized from the deep lunar magma ocean (LMO) that existed right after the Moon came together. This would have happened if the Moon assembled as white-hot debris after a single, catastrophic impact with Earth — but it's less likely if dozens of little cooled-off moonlets coagulated into a single whole.

Throwing Water on the Problem

As if the How and When of the Moon's formation weren't complicated enough, a third new analysis argues that — despite its extreme dryness today — the Moon likely contained a lot of water when it formed. In the same issue of Nature Geoscience, Yanhao Lin (Vrije Universiteit Amsterdam) and three others describe their experimental attempts to mimic how the Moon's magma ocean solidified. Lower density minerals would have floated to the top, forming a crust.

They find that the suite of minerals found in the lunar crust today — combined with its thickness — argue that water was part of the mix at a concentration of 270 to 1,650 ppm. This might not seem like much — but if proven true there'd be significant implications.

"A wet start of the Moon, coupled with the strong similarities between the composition of the Moon and the composition of the silicate Earth," Lin's team concludes, "suggests that equally high concentrations of water were present in the Earth at the time of the Moon-forming event."


Raluca Rufu et al. "A Multiple-Impact Origin for the Moon.” Nature Geoscience. January 9, 2017.

Gareth S. Collins. "Punch Combo or Knock-out Blow?" Nature Geoscience. January 9, 2017.

Mélanie Barboni et al. “Early Formation of the Moon 4.51 Billion Years Ago.” Science Advances. January 11, 2017.

Yanhao Lin et al. "Evidence for an Early Wet Moon from Experimental Crystallization of the Lunar Magma Ocean.” Nature Geoscience. January 9, 2017.

Ponder the true lunar origin while holding the Moon in your hands. Choose from S&T's standard Moon globe or our topographic Moon globe.


Big Splat


Image of Robert-Casey


January 13, 2017 at 12:25 pm

If the Mars sized impactor formed at L4 or L5, it would have formed from the same stuff the Earth formed from. As it was teh same distance from the Sun. Thus the isotope and chemical composition would be the same.

You must be logged in to post a comment.

Image of Ted-Forte


January 13, 2017 at 4:51 pm

Any theory that describes how the earth acquired its moon should also account for why our neighbors Venus and Mars did not. One large object impacting the protoearth seems a good deal more likely than a series of objects forming moonlets. If the multiple smaller objects is the source of our moon, why doesn't Venus have one? Wouldn't it have been in the line of fire too?

You must be logged in to post a comment.

Image of Rick


January 13, 2017 at 8:25 pm

My sympathies for those trying to work with all the complexities of the various theories.

The multiple impact theory will have to explain how so many impacts could strike at enough of a glancing angle so as to each time know loose earth material that would orbit properly (as opposed to just fall back to the earth). That would require almost a cosmic "rail gun" steadily aimed on the same trajectory.

An additional challenge to the "mulitples," as mentioned would that method create a large number of small satellites or on large moon around Venus, but more likely around Mars which is next to the asteroid belt with a huge supply of small impactors.

As to the geochemical similarities between Moon and Eaprth, a single impact could have knocked off enough Earth material which could provide the source of the samples on which we are theorizing. Just as there are sections of Earth with anomalous geologic areas, we might well find some other sections of the Moon which are less similar to Earth's geochemistry.

I think all would agree that it is still "early days" on the origin of the Moon.

Rick Littleton

You must be logged in to post a comment.

Image of J. Kelly Beatty

J. Kelly Beatty

January 17, 2017 at 7:46 am

Rick... your point about neither Venus nor Mars having a sizable moon is an excellent one!

You must be logged in to post a comment.

Image of Jacques Millet

Jacques Millet

January 14, 2017 at 8:57 am

Why the Moon couldn't have formed simply by the asymmetry of the original gas and dust blob, like double stars form?

You must be logged in to post a comment.

Image of J. Kelly Beatty

J. Kelly Beatty

January 17, 2017 at 7:59 am

Jacques... two variations of the Moon-forming event are known as the "fission theory" first proposed in 1879 by Sir George Darwin (a son of Charles) and "co-accretion theory," and your idea draws on both of them. the biggest problem in what you suggest is that although the Moon has many similarities to Earth it is extremely depleted in iron. if the two bodies formed together, the Moon should have much more iron than it does.

You must be logged in to post a comment.

Image of Jacques Millet

Jacques Millet

March 22, 2017 at 10:25 am

I guess the iron could have migrated in the center of the largest glob early enough so the Moon is left without much iron.

You must be logged in to post a comment.

Image of Edward Schaefer

Edward Schaefer

January 17, 2017 at 4:50 pm

Two speculations of why the Earth and Moon are isotopically so similar:
1) An initial hard collision betweeen two protoplanets disrupted both of them completely. Out of the resultant debris, two new objects formed, one bigger than the other. Those objects would have had similar isotopic ratios due to being well mixed. Then those object collided again to form the Moon as we know it.
2) Two objects collide somewhat obliquely, with the core of the second object impacting the Earth and spreading it iron over much of it. Because the early Earth has been liquified by the collision, the iron begins to sink into the Earth, spinning it up even more. Eventually, it is spinning so fast that it spawns the Moon though fission. As the iron continues to sink, the Earth continues to spin up at first, putting the tidal bulge ahead of the Moon and pushing it outwards quickly. Later, the iron has settled in the core of the Earth, and the Moon (being essentially a piece of the mantle) provides the tidal forced needed to slow the Earth down to its current rotation rate.

I gather the #2 runs up again the old Earth-Moon angular momentum problem. But I also recall that there are ways for the excess angular momentum to theoretically be removable.

You must be logged in to post a comment.

Image of rugeirn


January 17, 2017 at 10:48 pm

Three of the four articles given as references are behind paywalls. The honorable thing to do when referencing material behind a paywall is to state that it is not freely available in the reference. That saves us all a great deal of time and annoyance.

You must be logged in to post a comment.

Image of TheGreatOne


January 19, 2017 at 8:14 pm

If the moon had formed by a bunch of small cooled-off moonlets merging, then wouldn't it still be doing that now?

You must be logged in to post a comment.

You must be logged in to post a comment.