The Earth and its moon are unique in our Solar System. Earth is the only rocky planet with a large moon, and only the dwarf planet Pluto has a moon that’s so similar in size to its host planet. The Moon is also remarkably similar to the Earth in terms of its composition, suggesting they formed from the same pool of material instead of the Moon forming elsewhere and having been captured.
This collection of properties led to a number of ideas about how the Moon formed, all of which failed to fit the data in various ways. Eventually, however, scientists came up with an idea that seemed to get most of the big picture right: a collision between the Earth and a Mars-sized object happened early in the Solar System’s history, creating a cloud of debris that coalesced into the Moon.
While that got the major features of our two-body system right, there were still some subtle differences that weren’t resolved by the impact model. Now, a team of Japanese researchers say that there’s a way to tidy up some of these loose ends: having the impact take place while the Earth was covered in a molten magma ocean.
Close, but not quite right
Smashing a Mars-sized object into an Earth-sized planet can readily produce something that looks a lot like the Earth-Moon system. If the impact occurs obliquely, you end up with the right amount of angular momentum in the orbits of the two bodies. These events dump much of the impactor’s iron into the Earth’s core, and the Moon ends up with a lot of material from the proto-Earth, explaining their similar composition.
But there are a couple of nagging details that don’t quite work out. For example, the Moon has more iron oxide than the Earth, suggesting that the two bodies that collided had different compositions. But the ratios of various isotopes that we’ve tested show that the Earth and Moon are nearly identical, which suggests that they must have started out with an extremely similar composition. Reconciling these two facts has not been easy.
But the researchers behind the new work decided to rethink the physics of the collisions, rather than the composition of the bodies colliding. Early in our Solar System’s history, even after the major planets formed, there was still a lot of debris left over. Many of these objects would ultimately be swept up by the planets, creating a steady stream of impacts. The energy released by these impacts could be enough to melt the planet’s surface, creating a magma ocean, either globally or near the equator. So, what if the Moon-forming impact occurred while Earth was in a magma ocean phase?
To find out, the researchers first had to update the software used for modeling these impacts. This tended to treat both objects involved in the impact as collections of solid material. So, they modified the existing code so that it could handle boundaries in density, which occur at the boundary between the magma ocean and the solid mantle underneath it. With that in place, they started smashing things into each other.
A big splash
Upon a typical impact, the smaller body comes apart, forming an arc that’s a bit like an orange peel opening up (without an orange inside). The core of the impactor ends up leaving most of the rest of the planet and forms clumps that fall into the Earth, migrating into the core and leaving the debris iron-poor. Meanwhile, much of the magma ocean is peeled off the Earth, ending up in a spray that both leads and trails the remains of the impactor. Shortly after, another clump of the impactor splashes into the remains of the magma ocean, putting more of it into orbit around the remains of the Earth. In the end, nearly half the magma ocean ends up being ejected into space.
The research team performed multiple runs, with varied initial conditions, in order to get a sense of the range of results that these collisions could produce. To be a realistic solution, the collisions need to produce a system with the right angular momentum, the right mass, and a chemistry that matches what we know of the distribution of elements in the two bodies. The latter condition includes things like an iron-rich core on Earth, excess iron oxide in the Moon, and similar isotope ratios.
Two of these aren’t issues. Most of the collisions end up with the majority of the metallic iron in the Earth, and all of them produced the right angular momentum in the final system. This means that a huge range of conditions are consistent with these physical constraints on the model. Less limiting was the mass, where results ranged from about half the mass of the Moon to 1.4 times the Moon’s mass. Typically, you get a larger debris disk if the collision is oblique and at lower speed, so that limits the conditions of the collision slightly.
But a key difference in the simulations was the presence of a magma ocean. The simulations showed that the presence of the magma ocean causes a dramatic change in the shock heating that occurs upon collision. If the magma ocean is deeper than 500km, a lot of the ocean gets put into orbit. This is partly because the already-hot magma ocean is heated more efficiently by the collision. Since more of the proto-Earth gets put into orbit, a higher fraction of the Moon ends up composed of that material—or, more specifically, of material from the magma ocean.
The higher fraction of material from the proto-Earth means that the Moon formed with an isotope fraction similar to that of the Earth. In addition, iron oxide has a relatively low melting point, which should place more of it in the magma ocean. That explains the Moon’s relative abundance of this material. Having more of the Moon originating in the proto-Earth means the constraints on the object that collided with it aren’t as tight. As a result, the researchers suggest that the impactor could have originated from two different classes of starting material, rather than the same material as the Earth formed from.
So, the new idea seems to take the dominant explanation for the formation of the Moon—a giant impact—and refines it a bit. In doing so, it tidies up a few of the inconsistencies between existing impact models and the data we’ve developed over the last few decades. In many ways, this provides a great example of how science typically operates: rather than discarding ideas when they don’t quite match up with reality, scientists generally refine the ideas by tweaking them to better match the data.