The New Science of Moon Formation

We’d like to thank Speakly for supporting PBS.

Einstein once asked if “the moon exists only when I look at it?".

It was rhetorical objection to the idea that measurement in quantum mechanics causes reality to become real.

But there was a time when the moon didn’t exist, and then hours later suddenly did.

At least, according to the latest simulations of its formation.

No one was there to see the Moon form, but let’s side with Einstein and assume that this was a real event that actually happened.

Without direct observation, we have to get clever to figure out what went down.

And there’s good reason to be curious.

Earth’s moon is very, very unusual.

For one thing, it’s gigantic - by far the largest relative to its planet in the whole solar system.

Excluding Pluto, which isn’t a planet anymore so whatever.

The moon's composition is also unusual - it has a relatively tiny iron core for its size.

On the other hand, the elements that it does have look like they could have been scraped off the surface of the Earth, which you would absolutely not expect for a celestial body that wasn’t scraped off the surface of the Earth.

Before we get into how such a body could have formed, let’s talk about how we even know all of this, and why these properties of the moon are so unusual.

There was this magical time half a century ago - it wasn’t very long, just a few years - when humans actually visited the moon.

The Apollo program may have had a political side, but it also provided a trove of scientific data about our celestial little sibling.

For example, seismometers were set up at every landing site, which allowed geologists to tune into the moon’s vibrations.

Impacts from asteroids or even the spent fuel stage of the Apollo 17 Saturn V rocket cause seismic waves - moonquakes.

They resonated through the lunar interior.

From the nature of the waves that reach the seismometer we can reconstruct the moon’s interior.

And it turns out the insides of the moon are weird.

Most notably, its inner core of iron is really small, only about 20% of the moon's diameter - surrounded by a long-since frozen mantle.

By comparison, Earth’s core is closer to 50% of its diameter.

Apollo astronauts brought back a half-ton of material from the Moon’s surface, which adds the second layer to the mystery.

Each element on the periodic table can have different numbers of neutrons - they come in different isotopes.

The ratio of isotopes for a given element varied across the protoplanetary disk that our solar system formed from.

Each world in our system has a signature set of isotopic ratios.

Mars is different to Earth is different to some random asteroid.

But in the Apollo samples, isotopes of oxygen in the lunar silicate rocks are identical to those on Earth to a few parts per million.

OK.

So the moon has an isotopic composition identical to Earth’s crust, but it has an abundance of heavy elements that’s very different, evidenced by its tiny iron core.

We’re going to need more data.

And we can get that by just looking at the moon.

The types of rock on the lunar surface also speak to its history.

The moon clearly has darker stuff and lighter stuff.

The dark regions are basaltic flows, where volcanic magma oozed onto the surface and froze, similar to the dark rocks of Hawaii and Iceland.

These so-called ‘mare,’ Latin for ‘seas’, fill many relatively flat impact craters in lowlands, mostly on the near side of the moon.

Meanwhile, the light spots are anorthositic rocks.

This is also an igneous rock like basalt, but it forms differently.

Basalt forms ‘extrusively’, solidifying on the surface, while anorthosites form ‘intrusively’ inside the magna.

They are made of light elements like calcium, silicon, and oxygen and are less dense than the magma that makes them.

Because of this, when bits of anorthosite start to freeze within the still-molten magma they tend to float to the top.

Together, these rocks tell us that the moon was once covered by a layer of liquid magma - and that would have been for tens to hundreds of millions of years.

Over that time it would have formed a light anorthositic crust that was later splotched by darker basalt as eruptions pushed some magma back up on top.

And then to finish it off, add a couple billion years of impact craters.

So then, how do we make a gigantic moon from Earth-like material with a really dinky core and completely covered with a magma ocean?

Let’s talk about some of the ways to make a moon, and see if we can find one that we like.

Maybe the most obvious way is to do it like we make planets.

Planets pull themselves together gravitationally out of the protoplanetary disk leftover from the Sun’s formation.

But in that process they have their own ‘circumplanetary disks’.

In this environment, planets and moons grow simultaneously until the disk is exhausted, and this is the kind of formation we expect produced the big moons of Jupiter.

This mode of formation would explain the matching isotopic ratios of the Earth and moon, because they both would have formed from the same general region of the protoplanetary disk.

But if this were the whole story then we’d expect the moon to have roughly the same proportion of different elements, not just the same isotope ratios for a given element.

And the moon’s tiny core tells us this is not the case.

Another piece of evidence against this formation mechanism is that the rotation axis of the earth and the plane of the moon’s orbit don’t line up.

If they formed from the same disk, they should line up.

OK, so the moon probably did not form in a circumplanetary disk at the same time as the Earth.

So what else we got?

If you don’t make your moon at home, the other way to get one is to order out.

Plenty of moons in the solar system show evidence of forming elsewhere before being gravitationally captured by their planet But that’s harder than it sounds.

The biggest challenge is slowing down your potential moon during a close flyby and stopping it from just whizzing right back off into space.

You can do that with atmospheric drag, and the gas giants capture asteroids that way.

It’s also a possible explanation for Mars’s sad little potato moons.

But this is pretty implausible for a moon like ours, which completely dwarfs our atmosphere.

Another way to capture a moon is to drop in two at the same time.

The complicated gravitational ballet of a binary - or really trinary system - can leave one body in orbit and kick the other out with even higher velocity than it entered with.

This sort of ‘exchange’ was probably the source Neptune’s moon Triton, with its huge tilt and backwards orbit.

Capture was actually one of the most popular explanations for our moon’s origin until, you guessed it, the Apollo missions killed it.

Those isotopic ratios really are a smoking gun for the moon and earth formed from the same stuff.

But how did it form from the same stuff if it didn’t form in Earth’s circumplanetary disk?

The last, most likely scenario, and by far the most awesome scenario is the giant impact hypothesis.

The basic idea is that two proto-planets were forming in the same orbit and collided.

There was a larger one that we’ll call Proto-earth, and probably a Mars-ish sized body that we call Theia after the mother of the moon in Greek mythology.

Theia most likely formed at the L4 or L5 Lagrange point, leading or trailing proto-earth by a sixth of an orbit.

When the solar system was about 100 million years old and the planets had already eaten up most of their raw materials, Theia’s orbit destabilized – probably from the gravitational tugs of the other planets.

It began to drift toward the earth.

Let’s take a moment to imagine what that would have looked like … A giant planet getting bigger and bigger and bigger on the sky until… Well, for decency let’s zoom out and look at a simulation of what may have happened.

Theia smashes into proto-Earth, and the debris forms a disk around the liquified planet.

Much of that debris eventually rains back onto the Earth, but some recedes outward and over months to years pulls itself together into the moon with its own gravity.

The giant impact hypothesis solves a lot of our problems – the isotopic similarity of the earth and moon would come from mixing during the impact.

The moon’s iron core is small because the lighter stuff from both planets got sprayed into space, while Theia’s iron core was absorbed by the Earth.

And a collision like this would have easily liquefied rock, giving us our lunar magma ocean.

Besides explaining the anomalies of the moon, this hypothesis gives us some bonus answers about the Earth’s own weirdnesses.

Earth has a particularly robust iron core, which is responsible for our planet’s strong protective magnetic field.

Absorbing Theia’s core may have really helped out there.

Finally, Earth’s own rotational axis should have started out the same as its orbital axis, but now it’s tilted.

This collision would have been more than enough to knock our tilt off axis, depending on the angle of impact.

This story seems to fit a lot of the data, but there are lots of uncertainties.

What were the real masses of the bodies?

Their relative speeds?

Was it a glancing hit, or a head-on collision?

And how do you even test something like this?

While we can’t go and smash proto-planets into each other to test these different scenarios - at least not in the real universe … yet - we can obliterate worlds to our hearts content in computer simulations.

Using massive hydrodynamic simulations run on supercomputers, scientists explore the wide range of outcomes for different giant impact scenarios.

If a given setup for a virtual collision between a proto-Earth and Theia leads to something like the Earth-moon system, then it’s a sign we’re on the right track.

Up until now, typical simulations have played out like I described - Theia crashes into the moon, spraying up a big tail of debris that splashes around the earth and and also forms a big disk in orbit.

Over months to years that forms the moon.

While we can test pretty much any scenario this way, there’s still a big limitation - and that's the resolution of the simulation.

We can’t simulate every atom, or even every pebble, but we can approximate the system as say hundreds of thousands of particles all interacting gravitationally, and-or fluid cells interacting hydrodynamically.

You can have confidence in the result of these approximations if increasing the resolution doesn’t change the outcome too much.

But apparently, that’s not the case yet, because a very new simulation at a vastly higher resolution has found something completely different to anything we’ve seen before.

Earlier this year, Dr. Jacob Kegerreis at NASA Ames and Durham University in the UK and his collaborators ran simulations with as many as a hundred million little matter particles, a thousand times more detailed than standard simulations.

Let’s take a look at their best simulation to see what they discovered.

We starts with a pretty typical scenario- a very nearly earth-sized proto-earth, and a roughly Mars-sized Theia.

The collision happens at about a 45 degree angle at very nearly the escape velocity- not a head on collision, but still a major collision.

While the proto-earth survives the collision, Theia is basically obliterated.

The debris from Theia shears an enormous amount of the early earth’s mantle off with it, creating a plume towering thousands of kilometers into space.

But now, we see something no one has ever seen before.

For the briefest period, no more than a few hours, there are two moons.

Only in these most detailed simulations does this happen.

While the largest falls back into the earth, its gravity plays an important role in helping raise the smaller one into a wider, stabler orbit.

These simulations show that in just under two days’ time, Theia can hit the earth, make two satellites, and then destroy one while leaving the other to become our moon.

That’s very different to the previous picture of a ring of material pulling together over many months.

The new simulation gives us something very like our own system - a moon that’s around 1% of the Earth’s mass, with an outer layer heated to 4000K - plenty to give us our magma ocean.

The surface of this virtual moon contains a lot more material from proto-earth than other models, while the interior is almost entirely from Theia.

However the iron content is low, just like in the real moon.

So, like I said.

One minute there’s no moon - just a giant planet hurtling towards you - and a few hours later the moon is in place - according to this simulation, with all the properties of the actual moon.

Of course, there’s still a lot of uncertainty.

Just because a few simulations give us something that looks like the moon doesn’t mean this is exactly what happened.

But given the quality of the simulation, we could argue that this is a more plausible scenario than the earlier ideas.

Imagine what we could learn if we threw even more resources at the problem.

For example, we could add magnetic fields, which both bodies should have had prior to collision.

We’ll never have a perfect description of the moon’s formation, but as evidence mounts and simulations get better we can continue to narrow down the range of possibilities for the moon’s formation until we’re left with a general picture of what likely happened.

And if these simulations lead to predictions - like, for example, the nature of the moon’s interior - then we can test our theory.

Perhaps soon we’ll be able to confirm that cosmic cataclysm led to the weirdest moon in our local patch of space time.

We’d like to thank Speakly for supporting PBS.

If time travel were invented and you thought “Great!

Now I can go back to the 1920’s and debate interpretations of quantum mechanics with Einstein and Schrödinger!” – well, then you’d need to learn German.

Unless you already know it… but even in this current time period, if you’ve wanted to learn another language, then I’d like to introduce you to Speakly.

Offering 8 different languages to choose from, Speakly was created by 2 polyglots who both speak 7 languages.

They researched thousands of language learners and created a unique method that teaches words and sentences based on their relevance in real-life situations.

This means that you don’t learn anything that you actually can’t use to speak the language.

Available on both web and mobile platforms, you’ll learn new vocabulary, with speaking exercises, writing exercises, listening comprehension exercises, and even music recommendations in the language that you’re learning so that you don’t get bored.

There’s a link in the description to learn more.

Hello and happy new year!

I’m excited to get into another year of trying to understand this ridiculous universe with you.

Another good place to try to understand the universe is Curt Jaimungal’s wonderful Theories of Everything podcast.

As it happens, he interviewed me recently on topics ranging from consciousness and free will to godel incompleteness and dualities in physics.

I get quite a bit more opinionated than I usually do on Space Time, so it’s at least colourful, and perhaps even interesting.

If you want to, you could also let Curt know in the comments that space time sent you.

Link in the description.

OK, today we’re doing comment responses for the episode on supercritical fluids - that bizarre hybrid state of matter between liquid and gas.

mitchblahman13 asks if I can clarify what liquid metallic hydrogen is - one of the states of hydrogen in the gas giants.

Sure thing.

Firstly, solid metallic hydrogen is a theoretical state of hydrogen at very high pressures in which H2 molecules form a lattice with metallic properties - namely the electrons can travel freely through the lattice.

Liquid metallic hydrogen is when you have the same electron-sharing properties due to the high density, but the temperature is high enough to break the lattice bonds.

It still has the high conductivity of a metal, but can flow.

ChaosPotato and Jason Bouvette ask similar questions: could you swim or float in an open-top boat on an ocean of supercritical fluid.

My intuition says no, unfortunately.

Swimming requires substantial viscosity so you have something to push against to move forward.

Supercritical fluids have viscosities closer to gases, so to swim you’d need really gigantic hands.

Although some animals do have hands large enough to swim in a gas.

They’re called birds.

An open-top boat probably wouldn’t work because the boundary between the supercritical ocean and the gaseous atmosphere is not well defined.

The buoyancy of such a boat comes from the fact that it’s filled with the medium of the atmosphere, which needs to be substantially less dense that the medium of the ocean - enough of a difference to make up for the fact that the material of the boat itself is more dense than both media.

But in the blurry supercritical fluid-gas boundary, the change in density between the base and rim of the boat is relatively small.

So in order to float the boat has to be large compared to the transition scale between media, which I think means gigantic.

Or you could have a closed surface filled with lower-density gas, in which case you could easily float at the boundary.

This would be something between a zeppelin and a submarine, which sounds super cool and steampunk.

Roland Pihlakas requests that I please briefly summarise again, how is supercritical fluid different from gas.

Happy to, Roland.

The main difference is density.

Density is liquid-like, while viscosity and absence of surface tension is gas-like.

The high density means it interacts much more readily with whatever its in contact with than a gas does due to the large number of particles per unit volume, and can carry with it far more dissolved matter and heat.

The low viscosity and surface tension means it can get into tiny spaces much more easily than a liquid - allowing it to actually get to the particles you want to dissolve.

Together, these properties make supercritical fluids amazing solvants, oxydizers, heat carriers, etc.

as well as making them extremely corrosive.

Lord Marcus asks if there’s a supercritical phase where solids become indistinguishable from liquids.

Gareth Dean correctly responds - not really, because liquids and gases are both fluids - no rigid bonds between particles, which means an intermediate state between liquids and gases is meaningful.

Solids, however, do have rigid bonds which means they can’t flow, and so there’s no true intermediate state with liquids.

That said, there are cases where the solid and liquid states are so mixed that the effect feels like a hybrid.

Examples include gels - a crystal lattice of fiber networks filled with fluid, or colloids - aka slimes - solid particulates suspended in a liquid.

By the way, that whole thing about glass really being a liquid is false.

It’s an amorphous solid.

On the topic of solid-liquid hybrids, Vitaliy Vuychych and Sascha Wust point out that cats are simultaneously solid and liquid.

While it’s true that cats can be brought into an equilibrium between solid and liquid, this means raising their temperature to the cat melting point.

While that has been shown to place the cat in a state of being simultaneously solid and liquid, it’s no longer in a state of being simultaneously alive and dead.

It’s just one of those.